The conservation and rational use of water resources - not to mention the problem of supplying populations with adequate safe water- stands as one of the most ominous environmental problems, mankind has already to cope with at the turn of this century .

The " water problem "is of an outstanding complexity and concern as peoples both in developped and developping countries have ambiguous links with water ressource and aquatic ecosystems. On one’s hand, they entirely rely on this ressource for drink, domestic and industrustrial uses. At the opposite, mankind, up to now, has not payed much attention for safeguarding the water quality, at its renewal as a vital resource, the major running waters and aquifers beeing permanently spoiled by a growing number of pollutants. Last but not least, man is currently playing to " russian roulette " with the water cycle. Instead of securing the water supply by a proper managment of terrestrial ecosystems, man’s activity, especially deforestation, and the developpement of an intensive and unsustainable agriculture, decrease permenently the water availabillity in increasing the runnoff and slowing down the seepage into aquifers...

I- The global water reserves:

In spite of its apparent overwhelming abundance of water on the earth as 71 % of its overall surface is covered by the ocean, the freshwater is relatively scarce in the continental biosphere ( Fig. 1) standing as a small part of the overall hydrosphere volume (table I).

II- The global hydrological cycle and water balance(Fig.2)..

Solar energy fuels the water cycle into which water passes from one form ( vapor, liquid and solid) to another and moves from the ocean to land masses and conversely.

The renewal time of the differents kind of waters varies widely. The average residence time of water is only of 8 days in the atmosphere and 16 days in the rivers. At the opposite, the residence time of groundwater spans from centuries, to thousand and even to hundredths thousand of years (as well as for ice caps and oceans).

Basically, four fundamental processes activate the water cycle: evaporation (Ev) at the surface of the oean and others water bodies - and its variant the evapotranspiration (ETP) throughout plant biomasss which is the major source of water vapor during the growing season in the most of terrestrial ecosystems-precipitations (P) which bring down atmospheric water to the earth surface , runoff (Q), which is the main source of freshwater and actuates the hydrological cycle on the continents, and - last but not least regarding water balance and availability to man- seepage, which repletes the groundwater aquifers The 9/10th of the total precipitation fall back on the ocean and only 10% reaches the continents(Fig.3)..

At global level, on average , about 119. 000 km³ of precipitation fall annually on the dry land and supply all form of land water. The majority of this amount evaporates directly so that only 35% of this amount , about 45.000 km³ is returned to the ocean by the way of river and underground flow, by glacial runoff or runoff in hydrologically closed regions such as endorheic arid basin (Fig.4).

III- Availability of the water ressource for mankind and its growing demand of water supply.

Taking apart polar ice caps runoff, especially the one’s of Antarctica, which is the biggest, about 40.700 km³ per year reach the ocean throughout the rivers flow. However, only a minor part of this volume is actually available to man. If one’s discards peak flow related to floodin g periods, the stable flow is only of 14.000 km³ at global level. As human populations are not equally distributed on the continents surface, the the total amount of water actally available to meet human needs is lower , amounting to about 9000 km³ per year, big river occuring in areas where only very few peoples live. In 2000, some 56% of this total is already " consummed " by the various economic activities: domestic, industrial and agricultural, irrigation beeing the first source of water consumption.( fig 5) .

On other hand, this absolute value of water balance id not equally shared on the continent taking into account the number of people which lives on them.( tableII);

The reading of this table shows up that large differences occur from one continent to another. These differences are even stronger than it could be expected due to the occurrence of big rivers in continental areaswhere verybfew people live. For instance, in subsaharian Africa, despite the Congo river stands as the second biggest river flow in the world after the Amazon, its basin is few densely populated. At the opposite, people are overcrowded in others parts of the same continent where they already face major water supply problems.

Indeed, the water ressource availability appears more and more as the absolute limit to human demographic growth. At the dawn of the XXIth century, the water crisis- amongst the other major environmental problems to which mankind is confronted - shows obviously that in a number of countries; the demographic explosion as reached such levels that it already impede their sustainable développement. Far worse, if we pay attention to the problem of water availability for food production, in some 50 overcrowded countries struggling with the water crisis, any man who adds to the present population is already in excess as there is no more water left for irrigating the crops necessary to feed this supplementary mouth. Namely, as per now this additionnal population would starve if grain imports were to be cut.......

The water supply problem seems really far more concerning if one’s take into account the population s growth prospect in the developping world for the forthcoming decennial (Table III)

The water availability for people in a given country may be assessed by refrence to the Flow Unit, namely the number of existing people versus 1 million cubic meters of water per year.

If the population per flow unit remains low, say about 100 people per flow unit, it is not even a problem. At 300 people per flow unit, as in southern Europe, better managment is required to assure quality and proper distribution. As population pressure increases, especially as per 600 people per flow unit, signs of stress appear, beyond 1000 people per flow unit, countries

experience water shortages. At 2000 people per flow unit, under current technologies, extreme scarcity occurs ( Fig. 6) .

According to the Table III, a number of world region would outpass the 300 people per flow unit in 2025 and some of them as North Africa - as well as the midle East-are already above the 1000 people per FU threshold and would exceed 2000 in 2025! But even on continents where the average water supply would seem secured, the situation of some countries is soon very tight. For example, in Kenya, despite the abundant water resources of Africa taken as a whole, already cope with 1500 people per FU and this number would exceed 4000 in 2025!

IV- Man’s impingment on water supply.

Not only the water crisis is worldwide already fueled by the demographic explosion, but it is worsened throughout a number of anthropogenic mismanagments and perturbations of the natural ecological processes. Therefore water availability is impaired both at quantitative and qualitative level.

1)Man made disruption of the Water Cycle. Hydrological characteristic of water cycle defines the absolute limits of human societies access to water resources. Paradoxically, society’s activities generate additionnal forms of limitations to water supply and quality. Both in developed an developing countries, a number of ecosystems mismanaments and unwise use of natural resources are currently disturbing the water cycle, subsequently decreasing the amount of water supply.

In the industrialised countries, large scale alterations of rural landscape related to the développement of an industrial agriculture which has eradicated edge rows, the relentless developpment of roads and car parks, buildings construction, urbanization, resort housing developpement, not to mention a strongly growing artificial sylviculture whith an extension of clear cuts practices, strongly impinge on the water cycle. Both have strongly increased runoff versus infiltration with a lowered repletion of aquifer and last but not least a significative increase in the frequency and importance of flooding periods. Additionnally systematic mismanagment of rivers, due to chenelization and embankment have dryed out flood plains which formerly buffered the river water flow, storing water during the flood season and releasing it to the river in dry times.

In developing countries, the impingment of burgeonning populations on the water cycle is -may be- far worse. A number of developing countries have undergone, during the last decennials, a large scale alteration of their natural environment due to their attempt to keep the pace of food production with the ones of their population growth. In spite they have not achieved the security in their food production, they have already proceeded to widespread deforestation, soil erosion and overgrazing of their natural ranges at an extent up to now unknown in the human history. This unwise exploitation of their natural resources has already lead to an increasing number of the so-called " natural desasters ", indeed man made, and subsequently a growing number of ecological catastrophies. For example, what occured in December 1999 in northern Venezuela exemplifies demonstratively the kind of havoc generated by such an inconscious entirely " natural " but the first of all a by-product of the relentless deforestation until quite recently of the Sierra de Avila which is as well the water tower of the Caracas area and of the whole coastal range,!

In tropical ecosystems, vegetation destruction generates another disastrous impact on the water cycle. As the evapotranspiration is the major input of atmospheric water vapour in the atmosphere, any decrease in forest cover will negatively impinge on rainfall. Isotopic methods have demonstrated that in Amazonia or Congo watershed area, up to 85% of rainfalls are not brought by the atlantic depressions but by self condensation of the vapour generated by the evapotranspiration of their rainforests ( Fig 7 ). It has been computed that if ever Amazonia were entirely clearcut, the average annual precipitations would drop from over 1800 mm to under 500 mm, leading to a climatic situation similar to that is as a permanent climatic constraint, changes in vegetation cover impinge more heavily on the climate and lead to an increasing dersertification.

2)Water pollution is the second major way througth which man is worldwide impinging hardly on water resources. Though water pollution occurs since the very beginning of the neolithic first cities it has tremendously increased during the XXth century. The Table IV summarizes the major categories of water pollutants.

A wide number of chemicals can be found in natural waters. It has been computed that more than 120.000 chemicals are currently on sale in the world. Many of them are found both in surface and groundwaters. Indeed, in spite of policies that have spurred the completion of urban waste water treatment plants, and the subsequent decrease of water pollution by oxygen demanding organic matter observed downstream of the big cities in developed countries (figure 8), the xenobiotic pollution is still growing as shown up by the monitoring of water quality in the OECD countries. For example, up to 50 various pesticides can be detected in french groundwaters samples in Western Europe, is such that surface waters are usually no more adequate for supplying urban water drinkable networks! In France up to 70% of the urban communities relies already on underground water what renders more worrying the present aquifer pollution.

IV- Conclusion. Man impingment on the hydrological cycle, both quantitatively on the overall amount of water availability and from a qualitative standpoint on water resources - throughout pollutions- is such that it jeopardizes at world scale the ability to meet the needs for safe water, industry and agriculture in the forthcoming decennials. At the eve of the XXI^th century, the extent of the water crisis reaches such a global dimension that it could threatens any sustainable development both in the developed and developing world. The magnitude of this challenge is such that only immediate, heroic steps could address it properly, which would require urgent and deep changes in the ways present mankind currently manages the global environment.

F. Ramade, " Dictionnaire Ecyclopédique des Sciences de l’ Eau ", Ediscience International, 1998, 800 p

Figure 1 : Distribution of water on earth: volumes of the various biospheric reservoirs.

Figure 2 : Overall fluxes and stocks of water in the ecosphere.

Figure 3 The major processes involved in the global water cycle

Figure 4 : The freshwater flow global balance on the continents.

Figure 5 : The freshwater global consumption increase

Figure 6 : Population pressure and water resources problem: from the availability to resource scarcity;

Figure7 : Importance of forest cover to the hydrological cycle in tropical areas: the Congo river watershed overall water balance;

Figure 8 : BOD variations downstream from some major cities from indsustrialised countries during the last decennials.

Despite the growing importance of performing rapid microbiological tests, the use of standard agar growth plates for detection and enumeration of viable microrganisms, complemented by specialized media for isolating microbial species, still remains the analytical methodology used.

With this type of of analysis, generation of results is dependant upon an incubation period, typically 3 to 5 days up to 2 weeks, resulting in a retrospective response to microbiological contamination.

In order to improve the ability to respond, as quickly as possible, to potential problems and move towards real-time process control, manufacturers have looked towards new technologies as a mean to reduce the time for sample analysis and data production.

The main challenge is to not only produce a quicker method of analysis, but also to ensure that the sensitivity of detection is maximised. The technology should have the ability to detect very low levels of microorganisms, down to single cell copies.

With any technology using cell growth as part of the operating protocol, the analytical speed is governed by the rate of growth of the organism, including the initial lag phase. After periods of stress, the lag phase can be particularly important while the cells recover; this in turn affects the ability of any technology requiring cell growth to be performed rapidly.

CHEMSCAN ^® RDI SYSTEM

In order to address the need for a truly rapid assay with a high level of sensitivity, Chemunex has developped the ChemScan^® RDI system (Figure 1).

The system is based on direct fluorescent labelling of microorganims, coupled with an ultrasensitive laser scanning and counting system. The high level of sensitivity of the system means that a single microbial cell can be directly detected, which removes the need for cell growth and multiplication. Results are therefore not affected by any log phase.

To run analysis, samples are filtered through a 25 mm diameter, 0.4mm porosity track edge polyester membrane which retains the microorganisms on its surface. The membrane is transferred to an absorbent pad saturated with a micro-organisms fluorescent marker and incubated. The membrane is subsequently transferred to the solid phase laser scanning cytometer.

The total surface area of the membrane is scanned by a laser spot of 6mm diameter at a wavelength of 488nm (Figure 2). The individuals scan lines are 2.2mm apart to ensure that every section of the membrane is illuminated twice by overlapping consecutive scan lines. The software discriminants exclude any signal that is not seen by two consecutive scans in the same position. Three photomultiplier (PMT) tubes detect the fluorescent light emitted by the labelled cells. The wavelength windows for the three PMT tubes are set for the emission spectrum of the fluorescein at 500-530 nm (green), 540-585 nm (amber) and 590-605 nm (red). The signals produced by the photomultipliers are processed using a series of software discriminants that enable the instrument to differentiate between valid signals (labelled microorganisms) and background noise (electronic, optical or autofluorescent particles). The discriminants analyse the fluorescence spectrum of each signal by color ratio, specific fluorescent signal intensity and distribution of fluorescent intensities across the detected event (Figure 3). The membrane scan and signal processing are completed in 3 minutes.

The results of the scan are displayed as a direct cells count in the form of a membrane scan map, identifying the position of each labelled cell on the membrane surface (Figure 4). Because the entire process does not destroy the cells, the operator can visually confirm the presence of every labelled cell and confirm the morphology by transferring the membrane to an epifluorescent microscope that is linked to the cytometer and-by manipulating the cytometer- returning to the x-y coordinates of the fluorescent event on the membrane (Figure 5). This can provide a useful tool during the validation of the method.

CHEMSCAN ^® RDI APPLICATIONS

A. Detection of total viable micro-organisms.

Within the pharmaceutical, biotech and cosmetics industries, in-process control is a vital part of manufacturing. The ability to spot problems as they arise is the key to improving quality and reducing wastage. In the case of process intermediates or excipients, such as water, knowing their microbiological quality is essential for the manufacture of pharmaceuticals particularly for injectable products.The microbiological quality of pharmaceutical-grade water is usually defined as containing less than 10 organisms in 100 ml and European and US Pharmacopoeia require that testing for these very low levels of microorganisms should be untertaken using time consuming plate count methods.

Chemunex therefore developped protocol to detect all viable microorganisms in water using the ChemScan^® RDI system.

Viable microorganisms are labelled using Fluorassure^® reagents. These are based on non-fluorescent substrates that liberate free fluorochrome into the cytoplasm when enzymatically cleaved by esterases (Figure 6). As only viable cells (including spores) have the ability to perform this cleavage, they can be easily detected by the laser scanning technology. Even under conditions of stress, such as nutrient depleted conditions of high purity water or in the presence of growth inhibitors, the cleavage will still occur within viable organisms. Table 1 shows a typical range of bacteria, yeasts or fungi vegetative and spore forms that are detected with the ChemScan^® RDI.

After the samples are filtered through the membrane, this latter is transferred to an absorbent pad saturated with an activation reagent for 60 minutes at 37°C, followed by Fluorassure^® Labelling Solution and labelled for 30 minutes at 30°C then transferred to the analyser (Figure 7). As no pre-incubation step is required, sample preparation and laser scanning take under 90 minutes to perform using a sample collection.

In addition to its speed to result, the new system has a sensitivity that is at least equal to that of the culture plate: one microorganism in a sample. Based on the membrane filtration there is no limit on the sample’s volume, from just a few microliters to hundreds of milliliters.

The validation of the ChemScan^® RDI technology for water-system testing has been performed from multicentric studies and from a number of pharmaceutical sites that undertook parallel monitoring of their water facilities with plate count and the laser scanning method.

The correlation between direct counts from ChemScan^® RDI and results for colony forming units (cfu) on culture plates has been evaluated in pure cultures and process samples and showed the correlations to be high (Figure 8). The degree of correlation (where the slope = 0.992, r = 0.957) also held even at low cell concentrations, down to 1 to 10 microorganisms per sample - a region in which it breaks down for many alternative rapid techniques.

Comparative analysis of routine quality control samples at several pharmaceutical plant sites has shown that the accuracy, precision, sensitivity and range of the ChemScan^® RDI at least matches those of the plate count method when used for in-process water samples. The results indicate that the system has a large detection range that has been shown by in-house studies to be 1 to more than 40.000 cells by filter.

These studies have shown the benefits associated with the ability of the ChemScan^® RDI to directly detect and count all viable micro-organisms in a sample including spores, anaerobes, stressed or fastidious organisms. This has been demonstrated by the capability of this technology to detect filter deterioration in a water system, not detected by the plate, before the contamination level rose above set parameters and caused process failure. This enabled immediate resolution of the process failure, which would not have been possible with traditional culture plate techniques. The detection of cells missed by culture plates was confirmed by subsequent correlation to limulus amebocyte lysate (LAL) endotoxin test. When the LAL result passed the action limit, an ultrafiltration filter was replaced and all results (ChemScan counts, LAL and plates) returned to basal values (Figure 9). This analysis showed that positive samples had been correctly flagged by the system, well in advance of the corresponding LAL data.

The ChemScan^® RDI offers pharmaceutical laboratories the opportunity to switch to a far more proactive and corrective approach when assessing the microbiological quality of process waters.

B. Rapid detection and enumeration test of E. coli in drinking water.

The control of the microbiological quality of water, particularly with regard to fecal contamination is a daily routine in all drinking water laboratories. It includes compliance testing of the final product, monitoring of the raw (groundwater/surface waters) and partially purified water during the production process and "emergency testing" following construction works and major breakdowns.

Escherichia coli is still the major indicator of fecal contamination. Numerous tests have been introduced throughout the years for this micro-organism, amongst others tests based on lactose fermentation (e.g. mFC agar, tergitol 7 agar) and, more recently, on the demonstration of the b-glucuronidase activity (Figure 10). The latter offers the advantage of speed (18-24 h with reference to 24-48 h for lactose fermentation based tests). Although for monitoring purposes, a rapid result may not be really necessary, it probably is for compliance testing and certainly for " emergency testing". In the latter case, the ideal detection should be maximum 4 hours to ensure sufficient time for collection and transport of samples.

To reach the goal of a 4 hours, a test should be independant from bacterial multiplication as stressed and injured bacteria will fail to do so; consequently an instrument capable to detect single bacterial cells is required.

It is why an application has been developped to detect, in a specific way, E. coli in drinking water with the ChemScan^® RDI. This work has been initiated and developped in the laboratory of Prof. H. Nelis by Dr. S. Van. Poucke of the University of Ghent,Belgium in a strong collaboration with Chemunex.

The approach which was followed was to demonstrate b-glucuronidase activity in both single cells and in small microcolonies using a two-stage enzymatic labelling procedure in which the target enzyme b-glucuronidase is selectively induced before the actual labelling takes place.

The separation of enzyme induction and cell labelling greatly enhances speed with reference to procedures combining both steps.

The 100 ml water sample was vacuum filtered through the 25 mm, 0.4mm track-edge polyester membranes. The filter was subsequently placed on an an absorbant pad soaked with an Enzyme Inducer Cocktail and both incubated at 37°C for 3 hours. After incubation the filter was transferred to a second cellulose pad impregnated with the Labelling Buffer containing Fluorescein-di-b-D-glucuronide. The filter-pad combination was left for 30 minutes on a glass or ceramic plate cooled to 0°C with the aid of a circulating water:ethylene:glycol mixture kept at -7°C. Finally the membrane filter was removed from the pad and introduced in the ChemScan^® RDI.

Although ChemScan^® RDI is able to detect single cells, some multiplication cannot be avoided in the course ot the 3 hours required to induce as sufficient level of b-glucuronidase activity. As a result, a mixture of single cells and small microcolonies (2 to 20 cells) are found on the membrane filter (Figure 11). Both single cells and microcolonies were, however, detected as single fluorescent events and, hence, equally contributed to the count.

So a 3.5 hour test for the detection and enumeration of E. coli was sucessfully validated after analysis of 155 naturally contaminated and uncontaminated well, tap, out-of-pump, surface and sewage spiked water samples.

Evidence for the good specificity of the method came from the analysis of samples in which neither ChemScan^® RDI nor the reference method detected E.coli,but from which high number (10 to 10⁴ per 100 ml) of b-glucuronidase-containing bacteria were recovered on R2A agar supplemented with X-gluc. Although these did not grow on selective media for E. coli incubated at 37°C (CC-agar) or 44.5°C (m FC agar), they could potentially have been detected as single cells in the laser scanning cytometer . However, in none of these samples bacteria were detected by ChemScan^® RDI because of their intrinsically lower b-glucuronidase activity with reference to that of E. coli and/or because there were les efficiently induced.

Statistical analysis of the counts obtained in the reference and with the ChemScan^® RDI method did not indicate significant differences. Correlation coefficients (R) and slopes (a) of the linear regression lines obtained on data pairs of positive samples were R = 0.834 and a = 0.929 against CC-agar (n =60) (Figure 12) and R = 0.841 and a = 0.906 against m FC agar (n = 61) (Figure 13). Moreover, preliminary experiments with pure cultures have shown that there are no differences in detectability between ChemScan^® RDI method and plate counts on mFC agar, CC-agar.

In conclusion the ChemScan^® RDI based method combines speed with sensitivity and specificity and is applicable for detection of waterborne E. coli in 3.5 hours in all sample types tested. The tests yiels results that agree and are equivalent with those obtained in reference methods.

C. Detection of Cryptosporidium and Giardia with ChemScan^® RDI.

Current procedures for the detection of protozoan parasites Cryptosporidium parvum and Giardia lamblia are time-consuming, tedious and relatively inefficient. The actual recovery is usually low and often may be less than 1 percent.

The seriousness of the infections caused by Cryptosporidium particularly in the immunocompromised populations, the ills, elderly or young people, has focused researchers on more rapid and efficient techniques for its detection.

The process for detection can be splitted into three basic steps: sample concentration-separation of cysts and oocysts from contaminating debris and detection.

- Samples concentration: As oocysts tend to occur in low numbers in water, a system that allows their efficient recovery from large volumes of water is required. Recently the development of a new method for Cryptosporidium testing describes the use of a new filtration device based on a 1mm polysulfone cartridge filter from Pall-Gelman called Envirocheck.

- Separation of cysts and oocysts: Initial methods for separating targets organisms from contaminating background debris particles involved the use of density centrifugation where samples were "floated" on sucrose or Percol solutions of known specific gravity. This procedure has been shown inefficient and variable for detecting oocysts in environmental or field samples. The use of immunomagnetic separations (IMS) where magnetic beads coated with antibodies are used to separate the oocysts has been shown more reliable and efficient.

- Detection of cysts and oocysts: The most widely used procedure for detecting cysts and oocysts is by staining with monoclonal antibodies labelled with fluorescein isothiocyanate (FITC) and examination with epifluorescence microscope. The examination of slides is a laborious task that can take several hours to complete for each sample and lead to target organisms to be missed if the microscope use is not performed carefully.

ChemScan^® RDI has been successfully used to detect cysts and oocysts previously labelled with monoclonal antibodies on the surface of 2mm porosity filtration membranes through which concentrated waters samples were filtered. After the 3 minutes scan of the membrane, this latter is placed on an epifluorescence microscope equipped with the motorized stage. A rapid screening of any particles with characteristics of cysts and oocysts is usually performed within five minutes or less.

Thus a specific application has been developped which allows the detection of Cryptosporidium oocysts and Giardia cysts in less than 3 hours from concentrated waters samples with ChemScan^® RDI. (Table 2). In these conditions, after labelling, oocysts and cysts can be easily identified with the microscope (Figure 14).

From spiking experiments in deionized water or in river water, it was shown that the ChemScan^® RDI can detect and localize with high accuracy cysts and oocysts on membrane’s surface and sometimes can detect more oocysts than direct epifluorescence microscopy (Figure 15).

From extensive validations studies conducted in several water companies as Compagnie Générale des Eaux it was shown that there were good equivalences between the ChemScan^® RDI Cryptosporidium and Giardia applications and the traditional microscopy approach (Figures 16). (Data kindly provided by Dr. M-R de Roubin, Anjou Recherche).

Thus, the ChemScan^® RDI is an efficient method for screening water concentrates for the presence of Cryptosporidium oocysts and Giardia cysts. The use of a such system allows to have a much more standardized and reproducible and reliable method and significantly reduces the amount of time spent per operator performing epifluorescence microscopy.

Polyether toxins are produced by many species of marine microalgae, most commonly by marine dinoflagellates from approximately a dozen genera. Phylogenetic alignment of the dinoflagellates that produce biologically active polyether metabolites along a hypothetical evolutionary tree does not reveal clear patterns in structural homology of polyethers that can be related to morphological or ultrastructural criteria, such as radial versus bilateral symmetry or the presence of cellulose thecal plates. Furthermore, there is no evident association between polyether structural type and dinoflagellate habitat, although production of these metabolites may be more common among benthic species.

These polyether compounds are responsible for certain human intoxication syndromes linked to seafood consumption, such as ciguatera fish poisoning (CFP), diarrhetic shellfish poisoning (DSP) and neurotoxic shellfish poisoning (NSP). The toxins associated with CFP and NSP exert their toxic effects primarily through their role as potent ion-channel effectors, after site-specific binding to Na+ and Ca++ channels. Members of the DSP toxin complex are toxic due to their ability to inhibit protein phosphatases (PP1A and PP2A-types), thereby affecting smooth mussel contractions.

Also included in the polyether toxin group are the "fast-acting toxins" that cause rapid death (within minutes) of experimental mice after intraperitoneal injection of pure compounds or crude extracts of toxigenic dinoflagellates. Gymnodimine and spirolides are produced by marine dinoflagellates, and pinnatoxin, a closely related compound isolated from bivalve molluscs, is strongly suspected to be of dinoflagellate origin as well, although the causative organism has not been identified. These "fast acting toxins" all share a cyclic imine functionality, which is believed to be the cause of the potent pharmacological activity. Deactivation of the compounds by opening the cyclic imine ring is evidence of this mechanism. Gymnodimine is essentially non-toxic via oral administration to mice, whereas spirolides have an oral toxicity comparable to that of the neurotoxin saxitoxin — one of the most lethal known biotoxins. The human health significance of these compounds is poorly defined, although pinnatoxin was associated with a mass intoxication of shellfish consumers in China.

Based upon structural similarity, the polyether dinoflagellate toxins can be divided into two major sub-classes - the linear and macrocyclic polyethers, and the ladder-frame type. There have been many recent advances in structural elucidation of dinoflagellate polyethers. For example, at least 20 congeners of the ladder-frame ciguatoxin have been identified, and the structure of maitotoxin, one of the most complex secondary metabolites (M.W. 30,000) occurring as a single sub-unit, has finally been described. In a few cases, a plausible biosynthetic pathway has been proposed using stable isotope label incorporation followed by NMR. The structure of brevetoxin B is believed to have been assembled from two-carbon units derived from polyketide biosynthesis, although the origin of the methyl groups — exclusively from acetate or also from methionine via the Krebs Cycle — has been controversial.

Relatively little is known about the structural and functional relationships of these polyketide metabolites. There has been considerable speculation as to the functional significance of dinoflagellate polyethers, particularly those with potent biological activity. The hypothetical functions that have been suggested for these secondary metabolites in dinoflagellates include a role in chemical defense against predators, allelopathic interactions with competing species within the same ecological niche, hormonal and enzyme regulation, pheromones/sexual attraction involved in mating induction, ion channel regulation for membrane transport, and possible involvement in sequestration or catabolism of other metabolites. It is also conceivable that these biologically active secondary metabolites may also act as multiple effectors in eliciting a variety of cellular responses. For most dinoflagellate polyethers, evidence of functional significance is merely speculative or supported by rather weak (and often indirect) data. For example, in eco-evolutionary terms, it is premature to conclude that a "toxin" acts as a chemical defence mechanism against in situ competitors in nature, based upon its high lethality (i.e., low LD₅₀) towards conventional bioassay subjects (e.g., toxin administered intraperitoneally into laboratory mice). Nevertheless, it is difficult to accept that such a complex biosynthesis evolved (or was retained) without functional significance — as a relict pathway to primary or intermediary metabolism or as a dead end for the production of metabolic waste products. Biosynthetic pathways to polyether metabolites require the coordinated interplay of an array of biosynthetic enzymes, regulatory genes and the expenditure of considerable metabolic energy (in the form of ATP, nucleotide co-factors, and other high-energy intermediates) to produce these elaborate chemical structures.

Identification of the critical biosynthetic pathways for dinoflagellate polyethers will assist in defining a suite of genetic characteristics that are directly linked to the evolutionary history of the organism. Knowledge of the biosynthetic pathways and mechanisms employed provides information on the number and nature of the genes required for biosynthesis and how they interact in the assembly of the molecule. Such information can then be used to determine if any biosynthetic steps or assembly mechanisms are common or if they are unique to the organism. From an evolutionary perspective, biosynthetic genes for a particular secondary metabolite are often clustered on the genome, lending further credence to the idea that the metabolite is designed for a specific purpose and hence serves a useful function for the producing organism. Although it is clear that polyketide biosynthesis is the source of dinoflagellate polyethers, it is not known if the polyketide synthase is a single multi-domain complex (Type 1 synthase), or whether polyketides are assembled by a complex consisting of several enzymes (Type 2 synthase).

Comparison of polyether toxin profiles among natural dinoflagellate populations typically reveals a high degree of structural polymorphism and geographically distinct patterns. In clonal isolates in culture, the toxin cell quota may vary markedly over the culture cycle in response to physiological changes. Yet the toxin composition is usually preserved upon transfer into clonal culture, and tends to be quite refractory to environmental perturbations. This suggests a strongly defined genetic template. The polyether toxins are constitutively produced, thus they are not classic "stress" metabolites.

Current efforts are focused on establishing the timing and sequence of key cell cycle events involved in the biosynthesis of polyether toxins. The metabolic cascade leading to the synthesis of two groups of dinoflagellate polyethers — the linear DSP toxin derivatives produced by the benthic dinoflagellate Prorocentrum lima and the macrocyclic imine spirolides synthesized by the planktonic species Alexandrium ostenfeldii — were investigated using photoperiod-induced cell synchronisation techniques. Batch cultures of Prorocentrum lima in exponential growth phase were synchronized by transfer to darkness for 17 days. After dark synchronization, cultures were transferred back to the original 14:10 h light-dark photoperiod regime. Cells were harvested for DSP toxin analysis by LC-MS (liquid chromatography with mass spectrometry), and double-stranded (nuclear) DNA was quantified by flow cytometry. The cell populations became asynchronous within approximately three days after transition from darkness to the photoperiod, perhaps due to the prolonged division cycle (5-7 days) that is not tightly phased by the photoperiod. Unlike other planktonic Prorocentrum spp., cytokinesis in P. lima occurred early in the dark and ceased by "midnight". No toxin production was evident during the extended period of dark synchronization nor during the initial period when NH₄ was consumed as the major nitrogen source. An increase in DSP toxins usually occurred in the light, marked by a rise in DTX4 levels that preceded an increase in the cellular concentration of OA and DTX1 (delayed by 3-6 h). Thus, DTX4 synthesis was initiated in the G1 phase of the cell cycle and persisted into S phase ("morning" of the photoperiod), whereas maximal OA and DTX1 production occurred later during S and G2 phases ("afternoon"). No toxin production was measured during cytokinesis, which happened early in the dark.

In the gonyaulacoid dinoflagellate Alexandrium ostenfeldii the effects of physiological status on spirolide production were studied by comparing nutrient-replete growth of a toxic strain of A. ostenfeldii (SH98) isolated from Nova Scotia with that of a non-toxic clone (BAHME 146) after dark-induced cell synchronization. After 5 days of dark adaptation, cultures were transferred back to a 14:10 light—dark photocycle for measurement of chlorophyll a (extracted and in vivo), cell number, cell size, cellular DNA, and spirolide concentration at 2 h intervals through three L/D cycles. Although complete cell synchronization was not achieved, cell size variation was related to the photocycle, with size increasing in the light and decreasing in the dark. Mean cell numbers decreased during the dark, but net growth was positive during the experiment (µ=0.18 div d^-1). Analysis by liquid chromatography-mass spectrometry (LC-MS) showed that the spirolide profile did not vary significantly, consisting primarily of the des-methyl-C derivative (>90% molar), with analogues C, C3, D, D3 and des-methyl-D as minor constituents. For the toxigenic clone, the total spirolide concentration per unit culture volume was directly related to the concentration of cells and chlorophyll a, but there was a dramatic increase in cell quota of spirolides at the beginning of the dark phase and a corresponding decrease in the light.

Results of the experiments on the production of polyether metabolites using photo-period induced cell synchronization techniques indicated that biosynthesis is affected by light-dependent events in the cell cycle. However, the differences in the sequence and phasing of production of spirolides and DSP toxins by their respective dinoflagellates does not elucidate a common mechanism for these polyketide-derived metabolites.

Regulation of polyether toxin biosynthesis in the cell cycle is currently been studied at the transcriptional level by differential display of RNA expression. Although it is not possible to definitively ascribe a functional role to the polyether toxins, and gene regulation of toxin production remains poorly understood, hypotheses concerning their evolutionary significance and biogeographical distribution will be addressed.

Although cyanobacteria display a typical prokaryotic cell structure, they have long been, and often still are, misleadingly designated ´ blue green algae, cyanophyceae or microalgae ª. Their capacity, unique in the microbial world, to perform oxygenic photosynthesis like higher plants and algae is in part at the origin of this erroneous interpretation of their taxonomic position. Generally photoautotrophs, these micro-organisms, which may have appeared on our planet 3.5 billion years ago, require few nutrients and colonize very diverse terrestrial and aquatic ecosystems. Under favourable environmental conditions of light, temperature, nutrients (in particular phosphorus and nitrogen supply), in the absence of turbulence and serious competition with other natural populations (viruses, bacteria, actinomycetes, etc.), a number of cyanobacterial species, generally those capable of synthesizing gas vesicles which provide cells with buoyancy, may dominate. This leads to deeply coloured water blooms or even to scum formation, often deleterious to the environment due to the production of secondary metabolites which may be hepato-, neuro-, dermato- or cytotoxic, and provide water with unpleasant odours and taste. Toxic cyanobacteria are widely distributed throughout the five continents. Based on studies performed in different countries, it has been estimated that 50-70% of the cyanobacterial blooms are toxic and that hepatotoxins, in particular microcystins, are more frequent than neurotoxins. A detailed study performed by Vézie and collaborators in 1994, confirmed that 60% of the sites analysed in Brittany (France), including drinking water reservoirs, contained microcystins. The most frequent species of cyanobacteria found to be toxic in nature are the unicellular Microcystis aeruginosa and the filamentous Oscillatoria agardhii (also called Planktothrix agardhii), Nodularia spumigena, Anabaena flos-aquae and Cylindrospermopsis raciborskii. However, the same species may also be non toxic, rendering impossible the evaluation of the risk based only on the microscopic examination of the cell morphology of the cyanobacteria found in natural samples.

Three main groups of chemical structures, cyclic peptides, alkaloids and lipopolysaccharides, have been described as cyanotoxins. Microcystins and nodularins are the most frequent hepatotoxic cyclic peptides found in blooms from fresh and brackish waters. These products which contain seven (microcystins) or five amino acids (nodularins), including a very unusual amino acid called Adda ([2S,3S,8S,9S]-3-amino-9-methoxy-2,6,8-trimethyl-10-phenyldeca-4,6-dienoic acid), have a MW of 800-1100 Da. Intracellularly located, these soluble peptides are released upon cell lysis and are very stable. Inhibitors of the serine and threonine phosphatases 1 and 2A involved in complex cellular regulatory circuits, they participate in tumour promotion. Their primary target is liver and they lead to death by haemorragic shock within 1-3 h. The alkaloid neurotoxins, anatoxin-a (MW  165) and homoanatoxin-a (MW  179), are secondary amines that bind to acetylcholine receptors causing paralysis and death by respiratory arrest in 2-30 min. Other harmful alkaloids are the organophosphorus compound anatoxin-a(S) (MW  252) that inhibits acetylcholinesterase activity, and sulphated or non-sulphated carbamates (saxitoxins or paralytic shellfish poisoning). Initially found in tropical and subtropical waters of Australia, cylindrospermopsin (MW  415) is an alkaloid hepatotoxin that mainly affects the liver and inhibits protein synthesis in general. Finally, comparatively poorly studied are the lipopolysaccharides which are pyrogenic and toxic components of the cyanobacterial outer cell envelope. Very little is known about the biosynthetic pathways of alkaloid neurotoxins and lipopolysaccharides. Hepatotoxic cyclic peptides are synthesized by a nonribosomal process involving peptide synthetases, and recent genetic studies have demonstrated that some peptide synthetase modules, such as McyB, are specific for the synthesis of microcystins. Active research in that field is presently performed in several laboratories in Europe, Australia and the United States.

Several methods have been developed for the detection, identification and quantification of cyanotoxins. Different bioassays, including for example brine shrimps, Daphniae, bacteria and mammalian cells in culture, can be used to detect microcystins, nodularins and saxitoxins. Mouse assay, the use of which is undesirable and not permitted in some countries, is so far the only bioassay that detects anatoxin-a but it is generally neither very sensitive nor specific. Enzymatic and immunological methods, such as the inhibition of eukaryotic protein phosphatases and acetylcholinesterase for microcystins or anatoxin-a(S), respectively, as well as ELISA tests for microcystins and nodularins, are sensitive and rapid. Physiochemical methods such as HPLC-PDA, LC/MS, Maldi-TOF, etc. are useful to identify and quantify cyanotoxins but the equipment cost can be prohibitive and these methods require the use of standards not commercially available for all types of cyanotoxins.

The main exposure routes for cyanotoxins include direct skin contact in recreational waters, oral absorption of drinking waters, inhalation of aerosols and intravenous uptake following renal dialyses. The latter was at the origin of a tragedy in a hospital in Caruaru in Brazil in 1996 where 60 patients died following haemodialyses performed using water contaminated with microcystins. In 1998, a provisional guideline value of 1 mg.l^-1 for drinking water quality was adopted by the World Health Organization for microcystin-LR. There exist, however, more than 60 analogues of microcystins and microcystin-LR is not the most common one in some regions. No guideline values have been set up for other cyanotoxins. Some basic rules to follow and measures to avoid incidents have been established by experts in the field and described in a book edited by J. Bartram and I. Chorus, and published on the behalf of the World Heath Organization in 1999. Their recommendations include the choice of the drinking water source, groundwaters being often of better quality than surface waters, the treatment of water with appropriate processes, and the avoidance of cell lysis during transportation and treatment of water. Unpleasant odours, turbidity or an unusual colour of waters indicate their inappropriateness for recreational purposes and warning signs should be posted on sites. Microcystins have been reported to inhibit plant photosynthesis. Therefore, both for the protection of humans and safe agriculture, the quality of the irrigation waters should be carefully and regularly checked. Several other measures should also be followed to avoid incidents. First, for the determination of the risk level, not only the identification of the cyanobacterial population but also toxin assessment are required. Secondly, it would be wise to regularly monitor physico-chemical parameters, to survey animal deaths and determine the periodicity of bloom formation and cyanobacterial components in critical sites. Thirdly, the public should be educated, informed and motivated to participate and help in the survey of bloom formation. To avoid or eliminate cyanobacterial blooms, the reduction of external nutrient inputs must be seriously taken into consideration. Excessive use of inorganic and organic fertilizers, discharge of industrial and domestic waters, should be avoided, and run-off and erosion of fertilized agricultural lands and deforested areas should be kept under control. A number of chemical treatments based on the precipitation of phosphorus with flocculants and inhibitors of cell growth have been used with different degrees of success, depending on the polluted sites. All treatments have to be used with precaution as toxins are usually released upon cell lysis. Mechanical treatments involving mixing, filtration of water or withdrawal of the phosphorus-rich bottom sediments of water bodies have also been successful. Finally, physical treatments (barriers or booms) and biological ones (introduction of predatory fishes and higher plants) may help in preventing bloom propagation and development.

To be able to prevent and more efficiently control the formation of cyanobacterial blooms, and to determine the risk level in a satisfactory way, research needs to be developed further in different areas. A collection of reference strains is required including axenic strains and environmental samples from sites which frequently or occasionally form blooms, as well as standards for the different cyanotoxins. Simple and rapid methods to predict water blooms, and to detect, identify and quantify cyanobacteria and their toxin contents in environmental samples, should be also established so that the risks can be reliably determined at a minimum cost and, if possible, on site. At a more basic level, research on the ecophysiology of cyanobacteria, the biosynthetic pathways of toxin synthesis and the regulatory processes that control their production has to be intensified.

Some years ago R. Rippka and co-workers started to establish, in our laboratory, a collection of cyanobacterial strains. This collection, designated PCC for Pasteur Culture Collection, includes nowadays more than 700 axenic strains from highly diverse ecosystems. They are representative of at least 50 different genera and include 99 reference strains. The PCC represents the largest collection of pure cyanobacterial cultures in the world. About 450 of the PCC strains are presented in a catalogue available on the Web (http://www.pasteur.fr/recherche/banques/PCC). Observation by light microscopy is by no doubt the quickest method for the detection of the presence of cyanobacteria in environmental samples, but identification based on morphology can easily be misleading. More reliable means are thus required for strain identification. In our laboratory, we presently focus our research on the design and use of molecular DNA probes specific to cyanobacteria. These probes target either 16S ribosomal RNA sequences (16S rRNA), internal transcribed spacer (ITS) regions of the ribosomal RNA operon, or specific genes, such as those encoding phycobiliproteins, pigments found exclusively in cyanobacteria, cryptomonads and red algae. After DNA amplification by the Polymerase Chain Reaction (PCR) technique, the products or amplicons can be separated and the length of the DNA fragments compared by gel electrophoresis before or after digestion by restriction enzymes (Restriction fragment length polymorphism or RFLP). Another powerful method targeting 16S rRNA sequences consists of in situ hybridization of whole cells using rDNA specific probes labelled with horseradish peroxydase and detected by tyramide signal amplification. Examples of the use of these methods will be presented. The quantification of cyanobacteria in natural samples can be done by the determination of the cell dry weight or chlorophyll a. However, to be accurate, these determinations should be restricted to samples in which cyanobacteria are dominant. Direct cell counts can be done by light microscopy or by epifluorescence since cyanobacteria are autofluorescent. Flow cytometry is an alternative system for cell counting but requires costly equipment and homogeneous cell suspensions. The latter two methods coupled to in situ hybridization with one or several fluorescent substrates would permit both identification and quantification of one or several living organisms in a single sample collected from the environment.

The areas of research we wish to further develop in our laboratory are the detection and identification of toxic strains in natural samples by using i) in situ hybridization targeting mcy messenger RNAs that encode modules of the peptide synthetase specific for the synthesis of microcystins in Microcystis strains and ii) in situ immunodetection of Mcy and/or microcystins with specific antibodies. Finally, we shall focus studies on the detection methods and biosynthesis pathways for neurotoxins, a field of research still in its infancy.

References

1. Codd G.A., Ward C.J. and S.G. Bell. (1997). Cyanobacterial toxins: occurrence, modes of action, health effects and exposure routes. Archives of Toxicology Supplement 19. pp. 399-410. Seiler J.P. and E. Vilanova. Springer-Verlag, Berlin.

2. Dittman E., Christiansen G., Neilan B.A., Fastner J., Rippka R. and T. Börner. (1999). Peptide synthetase genes occur in various species of cyanobacteria. In: "The Phototrophic Prokaryotes". Peschek G.A., Löffelhardt W. and G. Schmetterer (Eds). Kluwer Academic/Plenum Publishers Corporation, New York. Chap. 72, pp. 615-621.

3. Nishizawa T., Asayama M., Fujii K., Harad K.-I. and M. Shirai./ (1999). Genteic analysis of the peptide synthetase genes for a cyclic heptapeptide microcystin in Microcystis spp. J. Biochem. 126: 520-529.

4. Schönhuber W., Zarda B., Eix S., Rippka R., Herdman M., Ludwig W. and R. Amann. (1999). In situ identification of cyanobacteria with horseradish peroxidase-labeled, rRNA-targeted oligonucleotide probes. Appl. Environ. Microbiol. 65:1259-1267.

5. Vézie C., Brient L., Sivonen K., Bertru G., Lefeuvre J.C. and M. Salkinoja-Salonen. (1997). Occurrence of microcystin-containing cyanobacterial blooms in freshwaters in Brittany (France). Arch. Hydrobiol. 139:401-413.

6. Book : "Toxic cyanobacteria in water: a guide to their public health consequences, monitoring and management." (1999). I. Chorus and J. Bartram (Eds). pp. 416. Published on behalf of the World Health Organization

Waterborne transmission of Cryptosporidium parvum is now recognized as a source in outbreak of cryptosporidiosis. Oocysts which are the exogenous stages of the parasite are resistant to inactivation by drinking water desinfectants currently used. This ubiquitous parasite is infectious for both mammals and humans. Although cryptosporidiosis causes self limiting diarrhea in immunocompetent individuals, persisting and life-threatening diarrhoea may develop in immunocompromized patients, especially in AIDS patients. Over the past decade, waterborne cryptosporidiosis has been responsible of many outbreaks especially documented in North America and in the United Kingdom.

Microsporidia are intracellular protozoan parasites of a wide variety of vertebrates and invertebrates. In humans, microsporidiosis is mainly associated with a chronic diarrhea in AIDS patients due to the 2 species : Enterocytozoon bieneusi and Encephalitozoon intestinalis. Although the mode of transmission remains unclear, there are some arguments to suggest that microsporidiosis could be a waterborne pathogen. Such a transmission of microsporidia has been recently identified in France.

The environmental sources of transmission of C. parvum

It has been suggested for long that the transmission of Cryptosporidium was related to either close contact with infected persons or animals, or consumption of contaminated food or water.

Person to person transmission is well known particularly in children, and several outbreaks of cryptosporidiosis have been reported in day care centers. Outbreaks observed were generally associated with high rates of cryptosporidiosis. In all cases outbreaks were due to person to person transmission and associated with a household contact with a diarrhea patient or an asymptomatic carrier. The risk was not limited to the period of symptoms since children excreted oocysts for 1 or 2 weeks and in some cases for more than 50 days after recovery. Hospital cross infection from patient to patient or from patient to staff members have been documented, especially for those caring for infected immunocompromized patients. A close contact with wild and domestic animals may be responsible for transmission in man. Cryptosporidium has been detected in about 80 different mammalian species including cattle, horse, sheep, goat and pigs and inter species transmission has been demonstrated. In contrast, companion animals such as dogs and cats may be infected, but are occasionally responsible of human transmission.

Foodborne transmission is linked to consumption of certain foods such as raw milk, offal, raw fresh saussages, fresh fruits or vegetables and even prepared foods. Fresh apple cider consumption has been found responsible of an outbreak in Maine in 1994. Two other outbreaks were also reported in 1996 in New York State and in Connecticut related to consumption of unpasteurized apple juice or cider.

Waterborne cryptosporidiosis has been documented by numerous outbreaks recorded in the past decade, especially in the United States and the United Kingdom. The most important was observed in Milwaukee, WI. in 1993, where 403,000 persons were infected. Outbreaks due to drinking water were also reported in Sheffield, Ayshire, Thanet Island, Oxforshire in the United Kingdom and recently in Japan. In all cases, outbreaks share water treatment deficiencies. Swimming pool associated outbreaks of cryptoporidiosis have been observed. Fecal accidents, sewage contaminations and inoperate filters were the most frequent causes of recreational water contamination. Ground water as well as surface water from river, lake, spring have been implicated as sources of contamination related to livestock, grazing lands, human sewage. The risk seems greatly increased by heavy rain, snow run-offs or discharges.

A heterogeneity among the isolates of C. parvum has been demonstrated from isoenzyme together with genomic DNA analysis. Isoenzyme typing of isolates from different locations has identified 2 isoenzymes of phosphoglucomutase and hexokinase which segregate between human or animal origin of isolates. Using PCR and RFLP analysis, 2 major genotypes of C. parvum have been identified, one isolated exclusively in humans designated genotype1and one isolated in livestock animals, wild animals and humans designated genotype 2. Genotype 1 isolates have a limited infectivity for animals while genotype 2 isolates may be transmitted to man. These subtypes appear to have variable importance in the waterborne outbreak and the significance of the distribution of the 2 genotypes remains unclear. In addition, intra genotypic variations within the genotypes have been found. In the waterborne outbreak of 1995 in the southwest part of England , all the patients were infected by Genotype 1 and 93.5% in the outbreak of 1997 (McLauchlin, 1999). Surprisingly, genotypes 2 had the highest distribution in the southwest of England and a large proportion of water supplies were provided from surface water in this area. This distribution of the genotypes suggests that different cycle of transmission not completely known may exist or a non water transmission may explain human outbreaks.

Dispersion and survival of C. parvum in water of environment

The agricultural sources of water contamination by C. parvum oocysts are considered the most important. Infected livestock have a considerable potential for contaminating aquatic environment. Blewett (1989), showed that infected calves and lambs could excrete more than 1010 oocysts in their stools daily for up to 14 days. Adult cattle may also be infected. This ensures a high level of contamination in the environment which favours waterborne transmission. Studies on the prevalence of Cryptosporidium in surface and ground water from streams, lakes, springs used as source of drinking water, have been conducted between 1987 and 1993 in North America. Among the 395 samples of water studied the average of positivity varyied from 9.1% to 100% and ranged from 0 to 251.7 oocysts per liter (Fayer, 1997). However, there is no direct relationship between the number of oocysts found in water and the risk of contamination. Therefore counting oocysts in water is not sufficient to maesure the risk of transmission and the viability or the infectivity of oocysts should be determined. Survival of C. parvum oocysts in water may be affected by the the contact time, temperature, pH, and also the treatment by disinfectants. Oocysts may survive in human and cattle feces and Robertson in 1992 showed that 34% of oocysts were still alive in cow semi-solid feces maintained for 176 days at ambient temperature. An increase of temperature diminished the survival time of oocysts as well as a decrease under 0°C. In a recent study, we showed that the percentage of survival oocysts in tap water maintained at 10°C was 83% after 2 weeks decreasing to 22.4% after 8 weeks, and were 96.4% and 80.5% for those maintained at 4°C in the same conditions. The infectivity of oocysts maintained at 10°C for 8 weeks decreased by 43.1% and only by 19.3% at 4°C (Brasseur, unpublished data). Among oocysts frozen at -22°C for 21 h, 67% were not viable and 100% after immersion in liquid nitrogen (Robertson, 1992). C. parvum oocysts are resistant to commonly used disinfectants. Chlorine-based disinfectants have a low level of effectiveness. Finch (1994) reported that a 4 h contact-time of oocysts with 15 mg/l of free chlorine decreased the infectivity by 47% and Korich (1990) showed that a concentration of 80 mg/l for 90 min was needed to obtain an inhibition of infectivity of more than 99%. Ozone disinfection of water appears to be more promizing. Perrine (1990) showed that a contact-time of 6 min with an ozone residual concentration of 0.4 mg/l inactivated more than 99% of C. parvum oocysts Oocyst inactivation could be obtained using exposure of water to UV light (Lorenzo-Lorenzo, 1993)

Host risk factors.

In human, DuPont in 1995, showed that an oral ingestion of oocysts of Cryptosporidium as few as 30 oocysts initiated an infection in 20% of healthy adult volonteers, whereas 100% of those who ingested 1000 oocysts were infected. The infectivity of 2 C. parvum isolates from calf and one from horse of the genotype C investigated in healthy adult volonteers showed significant differences in the ID₅₀ : 87, 1042 and 9 respectively (Okhuysen, 1999). Such data are not available in immunocompromized patients, but one might expect that a very low dose of oocysts could be infective. A close correlation has been found between lymphocyte CD4+ counts and the persistance of cryptosporidiosis in HIV infected patients (Flanigan, 1992). Other host risk factors have been identified, such as malnutrition, measles, and other viral infections increasing the suceptibility to cryptosporidiosis. Evaluation of the risk are based on filtration of water specimens, identification and oocyst counts, determination of the percentage of excystation and evaluation of the infectivity of sporozoites excysted using culture on enterocytic cell lines and animals models. Other methods using PCR or RT-PCR are in progress to increase the sensitivity of the detection of C. parvum oocysts in water.

The environmental sources of transmission of Microsporidia

The sources of human microsporidial infection and the ways of transmission are not clearly established. Humans and animals infected may released Microsporidia into the environment via stool, urine or respiratory secretions. Although, there is no direct proof of transmission from animals to humans or from humans to animals, E. intestinalis has been identified in donkey, dog, pig, cow and goat and E. bieneusi in pig, dog and monkey (Didier, 1998). Person to person transmission may be significant, but waterborne transmission is probably another way of infection for humans. Recent studies using molecular methods showed the presence of E. intestinalis and E. bieneusi in surface water and ground water. They have also been identified in raw sewage and effluents. Their presence in recreational water and in swimming pool has been suspected as a source of infection. The resistance of microsporidial spores to desinfectants or physical agents is not known and the factors of risk are ascertain. One waterborne outbreak of intestinal microsporidiosis in 200 persons has been reported in France which occured in summer 1995 Cotte, 1999).

Waterborne transmission of Cryptosporidium oocysts is considered a major risk for humans and public health policies are needed to limit the contamination of surface waters and to ensure an efficient disinfection of drinking water to limit the contamination especially for immunocompromized individuals. Waterborne transmission of Microsporidia is possible but its importance in human contamination has to be determined. Further studies are needed to evaluate the risk of waterborne transmission of Microsporidia

References

The consumption of bottled mineral water has increased dramatically in recent years, primarily in countries like the USA, and the UK where mineral water sales were very low. The reason for this increase in the sales of bottled mineral water is due to the public’s perception that bottled water is more wholesome, and tastes better than tap water. For these reasons the European Union has set rigorous microbiological controls on the production and sale on bottled mineral water. Moreover, all natural mineral waters have a microbial flora. The question then arises as to the microbiological quality of these waters.

Public water is submitted to disinfection because the sources of these waters are commonly contaminated with pathogenic organisms, and it would be impossible to use them without prior treatment. However, European legislation prohibits disinfection of mineral waters by any means. Therefore, unlike most foods, which can be packaged after heating or the addition of preservatives, cooled or frozen and even irradiated with gamma sources, bottled mineral water contain a large population of microoganisms. The very presence of bacteria worries health authorities that naturally impose rigorous testing bottled mineral water. Bottled mineral waters can be studied from a fundamental point of view or it can be studied from a practical point of view.

Bottled mineral water can be defined as microbiologically wholesome water that originates from an underground water table, and which is tapped from natural springs or boreholes, although it is becoming increasingly rare that bottlers use natural springs as a source of bottled mineral water. Some mineral waters sources use artesian boreholes, but many boreholes have pumps to bring water to the surface.

There is sufficient evidence to indicate that aquifers have an autochthonous flora composed primarily, if not exclusively of prokaryotic microorganisms (Bacteria and Archaea), and associated bacteriophages. For some time, it was assumed that bottled mineral water contained the flora that originated exclusively from the source. This is not entirely true and it is now established that the heterotrophic bacteria in the finished product are partially derived from the pipes and filling systems. Therefore, it is more correct to define the autochthonous flora of a bottled mineral water as the bacteria that are found in the aquifer and that survive and grow in the finished water, the bacteria which have colonized the pipes, and the organisms that are associated with the borehole. It is not conceivable to think that many of the organisms in a untapped aquifer, and which necessarily contains a high proportion of autotrophic organisms remain viable under completely different conditions at the surface, and after exchange of natural gases by gases at atmosphere pressure. Moreover, we do not know which populations were introduced into the borehole environment during drilling and maintenance of a functional borehole system and have become established. At this time we know a little about the normal flora of the source (tap closest to the borehole head), along the piping system and in the finished product.

Several studies show that the number of heterotrophic mesophilic bacteria (heterotrophic plate count) recovered on low nutrient media is very low at source and immediately after bottling. After a few days at room temperature the total HPC increases dramatically to numbers that may reach 10⁵ colony forming units per ml. One study, at least, showed that there was an alteration in the populations immediately after bottling and those recovered during a year of storage, indicating that the environment changed and some populations were selected for during this period. One fundamental question that has not been answered satisfactorily is the state of the bacteria at source and immediately after bottling. The initial increase in the HPC after bottling could be due to multiplication of low numbers of heterotrophic bacteria derived from the source and the bottling plant, or it could be due to resuscitation of viable but non-culturable bacteria after bottling. Most studies have not been able to answer this question and the results are somewhat contradictory.

The identification of the normal flora of bottled water has lagged somewhat behind what we would hope for. First of all, bacterial populations vary among bottled mineral water because of differences in physico-chemical properties. Many of the species classified were chosen because they belong to a group the investigators are interested in and does not reflect the major populations in the mineral water. This is especially true, for example, with the recent description of several new species of the genus Pseudomonas. While Pseudomonas senso stricto are important components of bottled mineral water and the water at source, other organisms seem to constitute the major heterotrophic populations of some mineral waters. Recent developments in fluorescence in situ hybridization (FISH) and small subunit rDNA sequence analysis of whole populations are answering interesting questions on the microbial flora of mineral water. Not only are organisms being detected that we knew were present in these waters, but groups formerly believed to constitute minor or transient populations are now known to constitute major heterotrophic and facultatively autotrophic populations. One case in point is the predominance of beta-Proteobacteria in some waters, and the paucity of gamma-Proteobacteria in one of the waters examined.

One may ask why mineral waters are not disinfected to eliminate all of the microorganisms, but there is no straightforward answer. It may be a question that can only be answered by asking a question. Why should water that is wholesome and which poses minimal or no risk to normal humans be disinfected?

Very soon many of the questions about the health risks of bottled mineral waters will be answered using modern molecular techniques. These techniques will also be able to answer several questions about the development of microbial populations in this very unique ecosystem.

Introduction

Strategy and Methods for the evaluation of microbial quality of drinking water has not significantly changed during the last century. The main parameters used are one or two indicators of fecal pollution (coliformes, fecal coliformes or Escherichia coli) and mesophilic, heterotrophic colony forming units. The same parameters are used for testing of different kinds of waters (drinking water resources, finished drinking water and recreational water as well as water for irrigation). Microbiological testing of drinking water remained the same while chemical testing increased dramatically in the past twenty years, in spite of the fact that microbial problems in drinking water are far more often experienced than chemical problems (1).

The methods have been criticized for different reasons: many of them are time consuming and results are often obtained some days after the water has been consumed, only. Additionally, absence of the bacterial indicators not always coincides with the absence of pathogenic viruses or protozoa. The adoption of new methods has therefore been requested (2). Besides the use of chromogenic precursors in the growth medium for a more rapid detection of E. coli (3) the use of molecular methods has been proposed (4). The latter methods have the potential of being rapid, results can be obtained within some hours to one working day enabling the adaptation of the treatment process of drinking water. They can be used for the detection of a particular bacterial strain as well as for groups of bacteria, depending on the selection of probes.

The OECD-Workshop "Molecular Methods for Safe Drinking Water" in Interlaken 1998 discussed the possible use and benefits of the molecular methods. In the final recommendations of the workshop (5) a critical revision of the concept of microbial drinking water testing as well as improvements and standardization of the molecular methods have been requested.

Microbial pollution in the field of drinking water is often handled in parallel to the experience with chemical pollutants (6). But as lined out in table 1 significant differences exist between the two forms of water pollution. Whereas chemical pollution of drinking water resources only rarely leads to acute illness in developed countries, microbial pollution is typically recognized by epidemic illness among the consumers. Theoretically, a single pathogenic organism can lead to infection and illness, although for the majority of microbial pathogens this probability is low to extremely low.

Another peculiarity with biological pollution is, that infection is not always manifested by illness. But the infected individual is a source for potential secondary infection. A previous infection may result in a varying degree of immunity, which can last from some weeks to months or years. This can lead to the paradoxical situation, that a population using drinking water with a small, but constant concentration of a pathogen might show less cases of illness than a population with high quality drinking water for most of the time and sporadic breakthrough of the same pathogen (7). Different treatment steps in the drinking water production lead to an aggregation of microorganisms resulting in a few parts of volume containing substantially higher concentrations of organisms than the bulk drinking water (8).

Especially bacterial pathogens can multiply in food, another feature of biological water pollution. Even small numbers of bacteria in drinking water with low probability of infection can reach high degree of illness after growth in food prepared with such water.

In view of these properties of biological pollution in water and the large number of possible pathogens, safe drinking water cannot be obtained simply by using "new and improved" methods for testing the end product. In this respect a lot can be learned from the food industry, where the concept of "Hazard Analysis Critical Control Point (HACCP) is used for an optimal end product.

Structures of Quality management for drinking water

Table 2 gives an overview of the steps needed for safe drinking water. One has to start with the characterization of the water resources used. With respect to the microbiological aspects, this means a sound knowledge about possible contamination by pathogenic organisms as well as the extent of variation in this contamination throughout the year. Within this survey of the water resources possible pollution sources and incidents have to be evaluated. Since feces of humans and warm-blooded animals are the most important source of pathogens, determination of the number of E. coli can be an acceptable approach (9). But it has to be kept in mind, that survival and transport of bacteria can be quite distinct from that of viruses and protozoa. It might therefore be necessary to include other indicators in this evaluation. This is especially important, where ground water without any treatment is used for drinking water. In Switzerland, for instance, almost one third of the consumers is served with ground water or well water receiving no treatment.

Based on the results of this survey and the treatment goals the minimal number of pathogen removal can be derived. This will allow designing treatment processes for the drinking water. The disinfection efficiency of the individual steps need to be determined under production conditions. Additionally, parameters which allow to easily determine whether this efficiency is reached always during production have to be identified. As an example the certification of UV-treatment plants according to the Austrian or the German regulations will be discussed.

Testing the disinfection efficiency by determinining the reduction of the number of E. coli is not a good idea, since this organism belongs to the group of most sensitive bacteria for many disinfectants. Also filtration does not affect viruses the same way as bacteria.

The hydraulics of treatment systems is crucial for their efficiency and not always looked after with the necessary caution. Therefore evaluation under production conditions is extremely important.

If several steps are used for the treatment, they should be selected according to the multiple barrier approach, which means that even under conditions of poor function of one step, the full treatment is still able to obtain an acceptable performance.

The importance of the microbiological drinking water quality justifies that also the final product is tested with some microbiological parameters. After treatment, those organisms which are the most resistant to this treatment, should be selected. The quality of the finished drinking water should not only be determined in the waterworks but also at the consumers taps. In this way it is possible to pick up deterioration within the distribution system. Such deteriorations can be due either to recontamination or re-growth in the pipes. Besides measurements of the heterotrophic colony counts indicators for fecal pollution should be used.

Disinfection efficiency and control: UV-plants as example

Germany and Austria have installed certification requirements for UV disinfection of drinking water (10,11). Installation of such plants requires previous certification. Also in Switzerland newly installed UV-plants are required to fulfill one of the certification requirements (12). Both certification programs include biodosimetric determination of the disinfection efficiency under production conditions.

UV radiation with wavelength around 260 nm induces damage of DNA. The repair mechanisms of the microorganisms are not able to cope with DNA damage above a certain level and die. Intermediate levels of DNA damage, induced by insufficient doses of UV-light, on the other hand, lead to a temporary decrease of the number of colony forming units (cfU). But after some time, original levels of colony count can be obtained after insufficient UV treatment with a temporary reduction of cfU’s of several orders of magnitude (13). This illustrates that the use of colony counts after UV disinfection will produce misleading results.

Another critical point in UV disinfection is the UV-transmission of the water, which influences the radiation dose applied to the microorganisms in the water. It is important to realize, that from the hygienically important parameter is not the mean, but the minimal UV-dose for each water volume passed through the reactor. This minimal dose cannot be derived mathematically; it has to be experimentally determined by evaluating the hydraulics of the reactor.

In the certification process the disinfection efficiency is determined using spores of Baccillus subtilis (14). Since spores are metabolically not significantly active, repair mechanisms are not important and can be neglected. The water entering the UV-plant is inoculated with the spores and their concentration is determined before and after UV-treatment (Fig. 1). From the reduction of viable spores the UV-dose can be derived from a dose-response curve measured previously in the lab. Variation of the light intensity and the flow rate will allow to define the range of flow rate with turbulent mixing within the reactor.

By using well defined UV radiation chambers it is possible to compare the UV sensitivity of B. subtilis spores with that of other microorganisms (15). The respective elimination of those organisms within the UV treatment plant can be derived from the biodosimetrically determined UV dose. Measuring UV light intensity in the reactor will serve as a control for process performance. The measuring point should be positioned such that changes in the UV transmission of the water will influence the reading of the light meter (Fig. 2).

The result of the experimental determination of the disinfection efficiency will be represented in a graph showing the maximal flow rate as a function of the UV-transmissions of the water which will ensure a certain, minimal radiation dose (Fig. 3). If Operation remains within these limits, the predetermined reduction of the number of viable pathogens can be assured.

UV disinfection is also a good example that molecular methods will not always be the better solution. A PCR detection method, able to detect a few cells of E. coli was applied to culture samples after UV irradiation. After UV treatment of E. coli the isolated DNA gave not only the expected band in the agarose gel, but produced smears which inhibited quantification at UV doses above 200 J/m² (16).

Conclusions

Besides the development of molecular methods for the testing of drinking water, the concept of the quality management must be adapted according to the scheme outlined above. The molecular method will help to achieve this task at some control points, whereas conventional or even non-microbiological measurements will better suit at other points. In routine testing of drinking water the molecular methods will therefore not change the system completely. Their importance will be much higher for epidemiology and investigation during outbreaks (17).

References

1.Moore, A.C., Herwaldt, B.L., Craun, G.F., Calderon, R.L., Highsmith, A.K and Juranek, D.D.: Waterborne disease in the United States 1991 and 1992. J. Amer. Water Assoc. 86, 87 — 99 (1994).

2.Young, P.: Safe drinking water: A call for global action. The American Academy of Microbiology warns of serious consequences from deteriorating water quality. ASM News 62, 349 — 352 (1996).

3.Feng, P.C.S., and.Hartman, P.A.: Fluorogenic assays for immediate confirmation of E.coli., Appl. Environ. Microbiol. 43, 1320-1329 (1982)

4.Atlas, R.M. and Steffan, R.J.: Polymerase chain reaction: applications in environmental microbiology. Ann. Rev. Microbiol. 45, 137-161 (1991)

5.Anonymous: Molecular Technologies for safe drinking water: The Interlaken workshop OECD, Paris, 1999.

6.Farland W. H. and Gibb, H. J.: US Perspective on balancing chemical and microbial risks of disinfection. In: Craun, G. F. (Ed) Safety of water disinfection: Balancing chemical and microbial risks. ILSI Press, Washington DC, 1993.

7.Frost, F. and Craun, G.: The importance of acquired immunity in the epidemiology of Cryptosporidiosis and Giardiasis. In: Proceedings of the OECD Workshop Interlaken 1998, http://www.eawag.ch/publications_e/proceedings/oecd/proceedings/Frost.pdf (1999)

8.Gale P., Van-Dijk P.A.H. and Stanfield, G.: Drinking water treatment increases micro-organism clustering: The implications for microbiological risk assessment. Aqua 46, 117-126 (1997)

9.Payment, P.: Waterborne viruses and parasites: resistance to treatment and disinfection In: Proceedings of the OECD Workshop Interlaken 1998, http://www.eawag.ch/publications_e/proceedings/oecd/proceedings/Payment.pdf (1999)

10.ÖNORM M 5873, Vornorm, Wien 1996

11.Anonymous: UV-Desinfektionsanlagen für die Trinkwasserversorgung — Anforderungne und Prüfung, Arbeitsblatt W 294, DVGW 1997

12.Snozzi, M., Haas, R., Leuker, G., Kolch, A. and Bergman, R.: Prüfung und Zertifizierung von UV-Anlagen. Gas, Wasser, Abwasser 79, 380 — 385 (1999)

13.Mechsner, K. und Fleischmann, T.: Vergleichende Untersuchungen zur Wiederverkeimung des Wassers nach Ultraviolettdesinfektion. Gas, Wasser, Abwasser 72, 807-811 (1992).

14.Sommer, R. and Cabaj, A.: Evaluation of the efficienty of a UV plant for drinking water disinfection. Wat. Sci. Tech. 27, 357 — 362 (1993).

15.Sommer ,R., Weber , G., Cabaj, A., Wekerle, J., Keck , G. und Schauberger, G.: UV - Inaktivierung von Mikroorganismen in Wasser. Zbl. Hyg. 189, 214 — 224 (1989).

16.Zimmermann, B. and Snozzi, M.: PCR detection of E. coli after disinfection with UV or ozone. In: Proceedings of the OECD Workshop Interlaken 1998, http://www.eawag.ch/publications_e/proceedings/oecd/proceedings/Zimmermann.pdf (1999)

17.Bartram, J.: Policy and administrative issues. In: Proceedings of the OECD Workshop Interlaken 1998, http://www.eawag.ch/publications_e/proceedings/oecd/proceedings/Bartram.pdf (1999)

Figure 1: Schematic representation of the determination of disinfection efficiency of UV treatment systems. 1: back flow inhibition, 2. flow control, 3. control of UV transmission by addition of sodium thiosulfate, 4. addition of spores of B. subtilis, 5. statix mixing chambers, 6. measurement of UV transmission, 7. pressure measurement, 8. temperature measurement, 9. flow measurement, 10. sample outle before UV reactor, 11. UV reactor, 12. sample outlet after reactor, 13 valve.

Figure 2: Measured UV light intensity at the control point of a reactor as a function of UV transmission of the water. The light intensity is given in arbitrary units, UV transmission (T₁₀₀) as percent transmission with using a 100 mm light pass.

Figure 3: Maximal flow rates leading to a minimal UV dose of 400 J/m² at different UV transmissions of the water. The dashed area indicates safe operating conditions.

Legionellosis is caused by a group of gram-negative bacilli including 43 species of Legionella. Legionellae require for growth special media supplemented with L-cysteine, soluble iron and a pH adjusted to 6.9. When cultured on charcoal yeast-extract agar, the colonies present a characteristic "cut-glass" appearance. Some serogroups and strains of Legionella are more virulent than others, L. pneumophila accounts for 90% of all the legionellosis. L. pneumophila serogroup 1 is the most frequently identified species isolated either from the environment and patients with legionellosis.

Legionellae survive and multiply in aquatic habitats. It is a common inhabitant of natural waters, such as lakes, ponds, streams and soils. L. pneumophila tend to grow as part of the biofilm in such environments. Legionella enter and colonize a variety of man-made aquatic habitats including cooling towers and evaporative condensers, as well as the plumbing systems of hospitals, hotels, spas and homes. Environmental conditions that promote the growth of the organism include:

- water temperatures between 20 and 50°C. The optimal growth range is 35-40°C,

- stagnant water,

- sediments,

- algae and protozoa that supply essential nutrients for growth of Legionella.

Legionella are particularly adapted to warmer conditions and their cellular fatty acid is similar to that of known thermophilic bacteria. Legionellae are recovered from waters of rivers and lakes with temperature from 5.7 to 63°C, but they multiply easily up to 42°C. Amoebae and other protozoa act as natural hosts and amplifiers for Legionella in the environment. Legionella growing in the presence of cyanobacteria have a doubling time of 2.7 h which is twice as fast as that reported for the growth of Legionella in laboratory defined media. Since Legionella can be incorporated into amoebic cysts, the effects of disinfectants, low pH and heat is therefore considerably reduced; bacteria sequestered in amoebic cysts becoming resistant to chlorine concentrations that are bactericidal for free-living organisms. If protozoa are the natural host of Legionella in the environment, human macrophages are accidental secondary hosts.

Legionnaires' disease is pneumonic with an incubation period from 2 to 10 days, an attack rate of 1% and an overall case-fatality of 10-20%. Pontiac fever is a nonpneumonic, flu-like disease that is caused by Legionella. This disease has a high attack rate (>90%) and a short incubation period of 48-72h. Complete recovery usually occurs in 2-5 days without medical intervention. The major mechanism of human infection appears to be direct transmission from the environment by inhalation of the bacterium in aerosolized contaminated water; other routes of transmission as ingestion may exist. The most susceptible people to legionellosis include the smokers, aged, alcoholics and immunosuppressed individuals.

Most of the early outbreaks have been traced to aerosols contaminated with these organisms from either cooling towers or evaporative condensers, while recent outbreaks have been traced to potable water services and components such as water heaters, showers, faucets, whirlpool baths, and respiratory therapy equipment. Domestic hot water systems that have water heaters operating below 60°C and deliver water below 50° C provide suitable conditions for the growth of Legionella are considered as potential bacteria amplifiers. Although legionellosis appears to be seasonal with most of the epidemics occurring in the summer, sporadic cases, especially in hospitals, occur throughout the year.

The greater the density of Legionella to which a person is exposed, the more likely the disease will occur. Although the infectious dose for humans has not been clearly defined, some authors have suggested "trigger" levels. In experimental legionellosis in guinea pigs, highly virulent organisms required lower densities for infection than less virulent organism.

The most effective control of legionellosis is prevention of transmission. The ultimate methods from preventing human infections of Legionella would be completely eliminate the bacterium from the environment. This is an impossible task because of the ubiquity of legionellae. The maintenance of clean water system is of critical importance in reducing the risk of legionellosis. Tower maintenance procedures must be mandated by national or local laws and health codes. Any control measures must include microbiological monitoring for Legionella as part of the quality assurance/quality control program to insure effectiveness.

In conclusion, design and good housekeeping procedures that prevent amplification and dissemination of Legionella should be preferentially formulated and implemented before systems are operated, and continued rigidly thereafter. Although this practice will not guarantee that a system or individual component will be free of legionellae, it should reduce the chance of it becoming heavily infected with the bacteria because of their need to receive nutrients from other organisms such as algae and protozoa.

References

Abu Kwaik Y, Gao LY, Stone BJ, Venkataraman C, Harb OS. Invasion of protozoa by Legionella pneumophila and its role in bacterial ecology and pathogenesis. Appl Environ Microbiol 1998;64:2127-3133.

Edelstein PH. Legionnaires' disease. N Engl J Med 1998;338:200-201.

Fliermans CB. Ecology of Legionella: from data to knowledge with a little wisdom. Microbial Ecology 1996;32:203-228.

Yu VL. Could aspiration be the major mode of transmission for Legionella? Amer J Med 1993;95:13-15.

The high population densities and activities often present in coastal areas result in pollution, including that of contaminated wastewater. However, recreational activities such as bathing and shellfish harvesting require good water quality. In order to evaluate the potential efficacy of proposed regulations and to devises strategies for their implementation, techniques are needed to predict the transport of microbial pathogens in water and shellfish. Pathogenic microorganisms such as bacteria and viruses (more than 100 different viruses are excreted in human faeces) are abundant in human waste (Metcalf et al., 1995) which is often discharged into natural waters with little or no treatment (Tree et al., 1997; Shin & Sobsey, 1998). In developed countries, however, sewage is more often treated before discharge into receiving waters.

Although treatment can result in large decreases in microorganism concentrations, considerable microbial fluxes are apparent in coastal areas effected by rivers or direct inputs. In fact, few data are available on enteric bacteria or viruses present in wastewaters and rivers (Pinto et al., 1996; Dupray et al., 1999).

Importance of microbial load discharged in coastal areas

Investigations carried out on several wastewater treatment plants and rivers on French coasts allowed to determine the quality and quantity of microbial discharges. Analyses were performed in primary treatment plants, in biological treatment plants (activated sludge) and rivers. Results are reported in table 1.

Table 1 :Frequency of pathogenic microorganisms in wastewaters and rivers from two locations : (1) Brittany coast (n = 6); (2) South France Lagoon (n = 24).

* results from PCR (Dupray et al., 1999), ** results from RT-PCR (Le Cann et al., submitted), nd : not done.

Most pathogenic microorganisms are present in wastewaters. These results are consistent with previous reports, where opportunistic pathogens or real pathogens were measured in wastewaters even after disinfection (Ashbolt et al., 1995 ; Maier et al., 1995). Activated sludge treatment allows the microbial load to be decreased by one logarithm, both for pathogens and other bacteria. The presence of various pathogens such as Salmonella, Clostridium, Cryptosporidium, hepatitis A virus, etc., are often reported in literature.

Faecal fluxes discharged in French coastal areas by rivers and sewage depuration outfalls were calculated. We observed very large fluxes (> 109 Faecal Coliforms/s) from sewage treatment plants, while contaminated rivers presented a mean flux of about 108 FC/s. The lower fluxes (< 107 FC/S) were measured in rivers where no sewage was discharged. In this case residual pollution could result from agriculture activities and sediment release. In winter, with rainy weather, the fluxes can be 10 to 100 times higher.

Microbial behaviour in the sea

Microorganism behaviour depends on physical and chemical factors such as dilution, temperature, sunlight radiation, salinity, predation, etc.

Several authors have demonstrated that nutrients help bacteria to resist damage from sunlight or osmotic stress (Gourmelon et al., 1994; Trousselier et al., 1998). Bacterial metabolism can be dramatically affected by hostile environmental conditions (Colwell et al., 1985). The latter demonstrated that bacteria can evolve in a viable but non-culturable state for which the usual detection techniques, such as plate counts, are not suitable. Viable cells can keep their plasmid content and virulence (Pommepuy et al., 1996) and under favourable conditions bacteria can be revitalized and recover an active metabolism. Once intestinal bacteria are exposed to seawater, different genetic defence circuits will be activated to allow them to combat the deleterious effects of harsh environmental conditions they have been suddenly exposed to.

Viruses are inert particles whose metabolism is not active outside of animal or human cells. Therefore, their fate in marine water mainly depends on external factors (temperature, solar radiation, dispersion, sedimentation, microbial predation). Some differences have been observed among different viral strains but little is currently known about viral fate in the sea, especially for the various bacteriophages which are potential candidates as viral indicators. In any case, their resistance to environment adverse conditions is higher than the one of bacterial indicators such as i.e. Escherichia coli.

Engineers involved in coastal management have previously used a "bacterial decay-rate" to evaluate the impact of bacterial contamination on receiving waters. This decay-rate, called "T90", is the time necessary for culturable bacterial counts to decrease by a logarithm in concentration. T90 values have been introduced into contamination models. These models needed to integrate factors such as the environmental conditions which influence bacterial behaviour. However, because faecal coliforms can evolve in a viable but non culturable state, the existing T90 models are currently regarded as unacceptable. Attempts to improve such models by correlation of the physiological state of bacteria have begun. Such models will enable the effects of different environmental conditions to be simulated. No predictive model yet exists to evaluate the behaviour of human pathogenic viruses or bacteriophages in the sea.

Health related microorganisms and their detection in Shellfish harvesting area

Outbreaks have been reported even from shellfish meeting bacterial standards (Le Guyader et al., 1996). Several projects are exploring the potential of bacteriophages as an alternative to pathogen direct detection, and subsequently for a better identification of viral hazard (Dore & Lees, 1995). Even if this virus presents an interest as an indicator, its occurrence at the same time as the enteric viruses must be demonstrated.

An other way to assess viral risk is the direct detection of the real pathogen by using molecular tools. Recent developments, demonstrating the success of molecular techniques (RT-PCR), have been reviewed for measuring microbial marine pollution (Pommepuy & Le Guyader, 1998). Until now, many difficulties have hampered attempts to detect bacteria and viruses such as astrovirus (AV), calicivirus (CV), Hepatitis A virus, and rotavirus (RV) which replicated poorly, or not at all, in cell culture. As microbial detection methodologies are improving, more accurate data on the numbers and types of enteric viruses present in environmental samples are generated. The development of PCR is a new research tool, especially for the study of the behaviour and fate of viruses which cannot be recovered in cell cultures (Le Guyader et al., 1998). PCR has the advantage of amplifying nucleic acid directly from tissues without previous culture or strain isolation from the sample. This is of particular interest for the detection of non-culturable microorganisms.

The PCR method is highly selective and allows detection even when there are only a few cells in a sample. Sensitivity is highly dependent on the sample and the presence of inhibitors. However, obstacles arise from the small numbers of pathogenic microorganisms present in environmental samples. Judging from detection of human virus in sewage and shellfish, the techniques are promising at the sensitivity and specificity levels. But more research has to be done to improve quantification and even sensitivity. Moreover, these techniques must be simplified for a diagnostic use.

Shellfish contamination: some examples of coastal contamination.

An EC Directive (91/492/EEC) has defined the health conditions for production and marketing of live bivalves. All Member States are required to classify their harvesting areas into one of three categories according to the level of faecal indicators present in shellfish samples. Shellfish from category A can be placed directly on the market. They must meet a standard of no more than 230 Escherichia coli per 100 g of shellfish flesh (or 300 FC/100 g) as well as other standards for specific pathogens (such as Salmonella), chemicals and algal biotoxins. Shellfish from category B must be purified before marketing, and shellfish from category C must be placed again in clean water for 2 months prior to marketing.

Sanitary quality is monitored by faecal coliforms, which are poor indicators of the presence of other pathogens such as Vibrio, enteric viruses or stressed bacteria. Recent investigations of outbreaks have shown that meeting faecal coliform standards in shellfish does not provide a guarantee for consumer safety since bacteria behave differently than viruses in marine environments, especially with respect to survival times (viruses generally survive longer). This limits the ability of bacterial monitoring to predict viral contamination in bivalve shellfish and ensure public health protection. Molecular techniques offer a better way to investigate such contamination. Recent progresses in this research area have provided data on viral contamination of the marine environment (Le Guyader et al., 1998).

Microorganisms discharged in coastal areas are accumulated in shellfish because of their filter feeding activity. Under favourable conditions, a single large oyster may filter up to 5 liters of water per hour. Thus, such an animal is an effective filter and concentration device of particles which may be present in growing areas, particles which will persist in the animal for a long time. If depuration is not efficient enough, viruses would remain in the product (Schwab et al., 1998; Doré & Lees, 1995). Thus, shellfish are a unique food stuff, partly because of the characteristics of the animal, but also because of the areas where they are grown and of the eating habits: these animals are grown in seawater, and this environment is the main source of contamination.

Numerous studies have been carried out in coastal areas over long periods of time, demonstrating the variable water quality directly influenced by variable abiotic environmental conditions (fluxes, currents, presence of mud and silt...). Pathogenic bacteria have often been found: Salmonella spp., Clostridium perfringens, Candida albicans, Vibrio parahaemoliticus, Campylobacter spp., Staphylococcus.. Some French harvesting areas have been investigated:

• Lagoon (Mediterranean coast)
• La Fresnaye (North Brittany coast)
• Fort de France Bay (Martinique),
• La Baule Bay (South Brittany coast): sediments, cockles and mussels were sampled from November 1992 to April 1993 and analyses performed using semi-nested PCR (enterovirus and Hepatitis A virus) and plate counts (Faecal Coliforms). Results are reported on Table 2.

Table 2 : Distribution of positive results (La Baule Bay): HAV (Hepatitis A virus); EV (enterovirus);

(-) negative results;* faecal coliforms > 300 CF/100 ml; (Le Guyader et al., 1993)

Relationship between gastroenteritis in population and shellfish contamination.

The microbial contamination is submitted to fluctuations varying from year to year and season to season, (increasing in autumn with a peak in December and/or January), (Miossec et al., 1998). Most of the time the evolution of bacterial contamination in shellfish (expressed in E. coli i.e. faecal indicator) is not correlated with viral contamination. Moreover, no evident relationship is observed between winter epidemic of acute gastroenteritis and bacterial contamination. However, viral contamination seems to be correlated to the local incidence rate per 100 000 inhabitants for acute diarrhoea (figure 2).

Figure 2: Comparison between winter epidemic of acute gastroenteritis and shellfish viral contamination in Mediterranean lagoons (Miossec et al, 1998)

These results allowed to hypothesize that diarrhoea epidemic in the French population would contribute to the viral contamination of the marine environment. Letrilliart et al., (1997) demonstrated that the major epidemic of acute diarrhoea occurring every winter in France, has a viral etiology, and that transmission from person to person is the main risk factor. However, only a few sporadic outbreaks have been linked to shellfish consumption (Charlet & Ferchaud, 1994 ; Miossec et al., 1998). Nevertheless, the free exchange of shellfish on the European market provides a dissemination route for enteric viruses in populations accustomed to consuming such products. Studies have to be developed to appreciate these hazards using a risk assessment approach (Miossec et al, 1999).

Conclusions

Studies carried out in France and in other countries have demonstrated the importance of microbial flux discharges in coastal areas. Urban sewages are mainly responsible for the microbial load. Agricultural practise can also provide local contamination. Among microbial flora, the presence of pathogens such as Salmonella, Hepatitis A virus and calicivirus has been determined. These microorganisms are demonstrated to be responsible for some gastroenteritis cases due to shellfish consumption.

Microbial discharges induce microbial contamination in coastal areas (shellfish and sediment). Recently, molecular tools enable rapid detection of pathogens and faecal contamination, were developed and used to detect real pathogens. Behavior of faecal bacteria in sea water depends on water quality and physical parameters. However, stressed bacteria evolve in a viable but non culturable stage where pathogenicity can be maintained and viruses can persist in the environment for long periods of time.

Acknowledgements:

Results were obtained from collaboration with: M.P. Caprais, A. Derrien, E. Dupray, H. Kopecka, P. Le Cann, F. Le Guyader, L. Haugarreau, D. Menard and L. Miossec.

Bibliography

Ashbolt N.J. et al., 1995, Wat. Sci. Tech., 31: 263-269.

Charlet F. & Ferchaud R., 1994, B.E.H., 3: 169-170.

Colwell R.R. et al., 1985, Biotechnology, 3: 817-820.

Doré W. & Lees D., 1995. Appl. Environ. Microbiol., 61: 2830-2834.

Dupray E. et al., 1999, in Pollutions diffuses, Actes Colloques, Ifremer, 169-178.

Gourmelon M. et al., 1994, J. Appl. Bacteriol., 77: 105-112.

Le Guyader F et al., 1994, Appl. Environ. Microbiol. 60: 3665-3671.

Le Guyader F. et al., 1996, Appl. Environ. Microbiol. 62: 4268-4272.

Le Guyader F. et al., 1998, Wat. Sci. Tech. 38: 45-50.

Letrillard et al., 1997, B.M.J., 315:1645-1649.

Maier A. et al., 1995, Wat. Sci. Tech., 31: 141-145.

Metcalf T.G. et al., 1995, Annu. Rev. Microbiol. 49: 461-87.

Miossec et al., 1998, J. Shellfish Research, 5: 1661-1664.

Miossec et al.,1999, in SRA Europe Conf. 355-365.

Pinto R.M. et al., 1996, Appl. Environ. Microbiol. 62: 1811-1813.

Pommepuy M. et al., 1996, Appl. Env. Microbiol 62: 4621- 4626.

Pommepuy M. & Le Guyader F., 1998, Curr. Op. in Biotech., 9: 292-299.

Schwab et al., 1998, Wat. Sci. Techn. 38: 83-86.

Shin G.A., Sobsey M.D., 1998, Wat. Sci. Techn. 38: 151-154.

Trousselier M. et al, 1998, Oceanol. Acta 21: 965-981.

Abstract:

Protozoan parasites have emerged as frequent contaminants in surface and groundwater and are associated with a high risk of waterborne diseases particularly for the immunocompromised. Major waterborne outbreaks due to Cryptosporidium and Giardia have been reported, either transmitted by recreational or drinking water, and waterborne transmission of Microsporida and Cyclospora is highly suspected.

Although recent technological advances in the detection of protozoa in water improved our understanding of the epidemiology of these micro-organisms, information on their viability, infectivity and resistance to disinfectants are dramatically lacking. This represents a major limiting factor for assessment and application of efficient preventive measures.

Résumé

La contamination hydrique par les protozoaires pathogènes est maintenant reconnue comme étant associée à des risques épidémiques importants, notamment pour Cryptoporidium et Giardia. De même, la transmission hydrique est également probable pour d'autres parasites "émergents" tels que Cyclospora et les microsporidies, avec un risque accru chez les patients immunodéprimés.

Malgré les améliorations techniques apportées à la détection et l'identification des protozoaires dans l'eau, nos connaissances restent très limitées sur la viabilité, l'infectivité et la résistance de ces parasites aux désinfectants, ce qui limite considérablement l'évaluation et l'application de moyens de prévention efficaces.

Over the late decades numerous parasitic diseases outbreaks due to protozoa have been reported. Cryptosporidium and Giardia lamblia are the most prevalent parasitic diseases throughout the world and the two latter had been responsible for waterborne outbreaks infecting hundreds of thousands of people²⁴. For other protozoan parasites such as Isospora, Microsporidum, Cyclospora and Toxoplasma gondii the role of water as vehicle of infection is probable but its importance has not been properly estimated. These events have led to a regain of interest and research on these pathogens mainly because of their potential morbidity and mortality in immunocompromised patients.

Most of these protozoa have common characteristics which may explain their potential for contaminating the environment and producing waterborne outbreaks: (i) large amount of parasites are shed daily in the environment by the infected hosts, via feces and/or urine (several millions per day for Giardia, Cryptosporidium or Toxoplasma gondii), (ii) excreted parasitic stages are small in size and do not multiply in the environment (iii) they are usually highly resistant and can remain infective for months in water (iv) most are zoonotic parasites, which implies a permanent risk of contamination of water resources from a domestic or wild animal reservoir.

Cryptosporidium and Giardia are considered as the most frequent contaminant found in drinking water^{5, 18, 21}. Both are responsible for diarrhea which can be life threatening for the immunocompromised patients. Since the outbreak in Milwaukee of water-borne cryptosporidiosis affecting over 400,000 people¹⁵, there have been at least 100 smaller outbreaks associated with these parasites. Cryptosporidium can infect a wide range of mammals and birds. Human and other mammals, especially calves, are important reservoirs of infection. Several studies also suggest that some water-borne epidemics of giardiasis were of zoonotic origin, although the host-specificity of Giardia spp. is still controversial. The high potential of Cryptosporidium and Giardia in outbreaks may be related to the relatively low number of cyst or oocysts capable of causing infection and the fact that these parasitic stages are directly infective, without a period of external maturation. With an infectious dose of 30 oocysts of Cryptosporidium, infection was observed in 20% of immunocompetent volunteers⁷; Giardia is evermore infective since 10 cysts can cause a 100% infection rate⁹. The infecting dose in immunocompromised patients is not known but is expected to be even lower. Because of the lack of effective therapy for cryptosporidiosis, this underlines the determinant role of preventive measures that can reduce the risk of environmental and water contamination

Microsporidiosis is an emerging infection the epidemiology of which is still poorly understood. Microsporidia are intracellular parasites that infect invertebrate and vertebrate hosts. There are more than 1000 species among which 12 have been reported to infect humans. Enterocytozoon bieneusi is essentially prevalent in HIV-infected patients but has also been occasionally reported in organ transplant recipients and in immunocompetent patients. It is responsible for chronic diarrhea and wasting in patients with severe immunodeficiency. Other species of microsporidia infecting humans are thought to be less prevalent although serological studies revealed the presence of high levels of specific antibodies directed against Encephalitozoon cuniculi in 4.7% of blood donors. Routes of transmission and sources of human infections are unknown. Spores of microsporidia are shed in feces for E. bieneusi as well as in urine and respiratory secretion for Encephalitozoon sp. Since several species of microsporidia that infect humans also infect wild and domestic mammals, an environmental contamination by these parasites is very likely³. However, it is still questionable wether microsporidiosis can be considered as a waterborne disease. Several lines of evidence are in favour of this mode of transmission. A case-control study performed in HIV-infected patients showed that the only two factors associated with intestinal microsporidiosis, were swimming in pools and male homosexuality, suggesting that the mode of transmission is fecal-oral and possibly through contaminated water¹¹. Two studies provided evidence of contamination of surface water by human pathogenic microsporidia including E. bieneusi, but the rate of spore recovery was low^{6, 22}. However, one recent study strongly argues in favour of a potential risk of waterborne outbreak of microsporidiosis. This occurred in 1995 in the region of Lyon (France) where 200 persons were infected by E. bieneusi (attack rate, 1% in HIV-positive patients/month). There was no evidence of fecal contamination of water but the major factor associated with diagnosis of microsporidiosis during the outbreak was living in an area corresponding to one of the three water distribution subsystems of the town⁴. Indeed, these results indicate that surface water may represent a potential source of human contamination by this parasite but the risk for humans seems lower than for Cryptosporidium or Giardia.

Cyclosporiasis is also a "new" pathogen with a not well defined life cycle¹⁶. Disease is characterised by mild to severe nausea, anorexia, abdominal cramping, and watery diarrhea. Humans appear to be the sole host for the latter and a distinct seasonality has been observed in endemic areas around the world. Routes of transmission are still unknown, although the fecal-oral route, either directly or via water, is probably the major one. Several outbreaks suggested transmission of Cyclospora by ingestion of contaminated berries and market vegetables¹⁷. However, oocysts have also been demonstrated in wastewater suggesting that water contaminated by feces may also act as a vehicle of transmission²⁵.

Close to Cyclospora, Isospora belli is another coccidian parasite which cause intestinal disease in human. Infections are acquired by the ingestion of sporulated oocysts in contaminated food or water¹⁴. Symptoms of infection in immunocompetent patients include diarrhea, steatorrhea, headache, fever, malaise, abdominal pain, vomiting, dehydration, and weight loss. Symptoms are more severe in AIDS patients and recurrences are common. Isosporiasis is endemic in tropical and subtropical regions but at present, no major waterborne outbreaks have been reported.

Toxoplasmosis was not considered as a major waterborne transmissible disease until the occurrence in March 1995 of a large waterborne outbreak in the western Canadian province of British Columbia². Mapping studies of cases, and case-control studies showed significant associations between acute infection and residence in the distribution system of one reservoir supplying water to Greater Victoria. The epidemic curve appeared bimodal, with peaks that were preceded by increased rainfall and turbidity in the implicated reservoir. A municipal water system that uses unfiltered, chloraminated surface water was the likely source of this large outbreak although oocysts were not demonstrated in the water. A contamination of the water supply by oocyst excreted by infected cougars was likely¹.

Since waterborne transmission of protozoa has become more prevalent in recent years, the need for screening water for these organisms has emerged. A major problem associated with testing for these organisms is the lack of reliable methodologies and baseline information on the prevalence of these parasites in various water sources. Several techniques have been developed, with the aim of detecting low concentration of parasites. Most are based on the sequential use of a combination of concentration methods including filtration, flotation, centrifugation, immunocapture, and identification methods^{8, 23}. Presently, most of these technique have major limitations, mainly due to their low recovery efficiency and difficulty for specific identification of the parasites. The use of immunomagnetic separation techniques combined with the immunofluorescence antibody assay using monoclonal antibodies have markedly improved the sensitivity of detection of Giardia and Cryptosporidium in environment water. More recently PCR-based methods and their combination with immunocapture concentration methods were introduced for parasite identification, allowing a reduction of the threshold of detection to 1 to 10 parasites into 1000 liters^{10, 12}. However, the cost of each determination is still very high, thus limiting the use of these techniques in routine practice. A better feasibility and cost-effectiveness is expected from the development of microelectronics and chip devices.

A crucial point in the risk assessment of waterborne protozoa is the determination of the viability and infectivity of the parasites that are recovered from the environment. None of the currently available techniques, including staining with fluorogenic vital dyes, in vitro excystation, cell culture or inoculation to animals is fully reliable or sufficiently sensitive ; moreover, these techniques cannot be combined with PCR-based detection methods. The reverse-transcriptase PCR method could be a valuable alternative if cysts, oocysts or spores produce detectable mRNA ; it has been successfully applied for detecting low numbers of viable Cryptosporidium oocysts into environmental water concentrates¹².

In conclusion, the key to control and prevention of waterborne protozoan infection is elimination or reduction of the cysts or oocysts in the environment. Since most of these parasites are resistant to chemical disinfectants used to treat water supplies^{13, 19, 20}, physical removal of the parasite appears to be the only efficient process, at least in combination with disinfection. Although this does not guarantee a complete removal of the parasite it may reduce the risk of transmission to a very low and acceptable level. In this strategy of prevention, the monitoring of purification processes is determinant. The major goal for the near future will be to provide cost-effective standardised methods which may be applied for multiple parasite routine monitoring for assessing safety of water supply.

REFERENCES

1. Aramini JJ, stephen C, Dubey JP, Engelstoft C, Schwantje H, Ribble CS. Potential contamination of driking water with Toxoplasma gondii oocysts. Epidemiol Infect, 1999 ; 122:305-315.
2. Bowie WR, King AS, Werker DH, Isaac-Renton JL, Bell A, Eng SB, Marion SA. Outbreak of toxoplasmosis associated with municipal drinking water. Lancet 1997 ; 350:173-7
3. Bryant RT, Schwartz DA. Epidemiology of microsporidiosis. In "The microsporidia ans microsporidiosis". M.Wittner and L.M. Weiss ed., ASM Press Ed., 1999; pp 502-516.
4. Cotte L, Rabodonirina M, Chapuis F, Bailly F, Bissuel F, Raynal C, Gelas P, Persat F, Piens MA, Trepo C. Waterborne outbreak of intestinal microsporidiosis in persons with and without human immunodeficiency virus infection. J Infect Dis 1999 ; 180:2003-8.
5. Craun GF, Hubbs SA, Frost F, Calderon RL, Via SH. Waterborne outbreaks of cryptosporidiosis. J. AWWA, 1998 ; 90: 81-91.
6. Dowd SE, Gerba CP, Pepper IL. Confirmation of the human-pathogenic microsporidia Enterocytozoon bieneusi, Encephalitozoon intestinalis, and Vittaforma corneae in water. Appl Environ Microbiol 1998 ; 64:3332-5
7. DuPont HL, Chappell CL, Sterling CR, Okhuysen PC, Rose JB, Jakubowski W. The infectivity of Cryptosporidium parvum in healthy volunteers. N Engl J Med 1995 ;332:855-9
8. Fricker CR, Crabb JH. Water-borne cryptosporidiosis: detection methods and treatment options. Adv Parasitol 1998;40:241-278
9. Gibson CJ, Haas CN, Rose JB. Risk assessment of waterborne protozoa: current status and future trends. Parasitology, 1998 ; 117: S205-S212.
10. Hallier-Soulier S., Guillot E. An immunomagnetic separation polymerase chain reaction assay for rapid and ultra-sensitive detection of Cryptosporidium parvum in drinking water. FEMS Microbiol Lett., 1999 ; 176: 285-289.
11. Hutin YJ, Sombardier MN, Liguory O, Sarfati C, Derouin F, Modai J, Molina JM. Risk factors for intestinal microsporidiosis in patients with human immunodeficiency virus infection: a case-control study. J Infect Dis 1998 ; 178:904-7
12. Kaucner C, Stinear T. Sensitive and rapid detection of viable Giardia cysts and Cryptosporidium parvum oocysts in large-volume water samples with wound fiberglass cartridge filters and reverse transcription-PCR. Appl Environ Microbiol 1998 ; 64:1743-9
13. Korich DG, Mead JR, Madore MS, Sinclair NA, Sterling CR. Effects of ozone, chlorine dioxide, chlorine, and monochloramine on Cryptosporidium parvum oocyst viability. Appl Environ Microbiol 1990 ;56:1423-1428
14. Lindsay DS, Dubey JP, Blagburn BL. Biology of Isospora spp. from humans, nonhuman primates, and domestic animals. Clin Microbiol Rev 1997 ; 10:19-34
15. Mac Kenzie WR, Hoxie NJ, Proctor ME, Gradus MS, Blair KA, Peterson DE, Kazmierczak JJ, Addiss DG, Fox KR, Rose JB, et al. A massive outbreak in Milwaukee of Cryptosporidium infection transmitted through the public water supply. N Engl J Med 1994 ; 331:161-167
16. Ortega YR, Sterling CR, Gilman RH. Cyclospora cayetanensis. Adv Parasitol 1998;40:399-418.
17. Ortega YR, Roxas CR, Gilman RH, Miller NJ, Cabrera L, Taquiri C, Sterling CR. Isolation of Cryptosporidium parvum and Cyclospora cayetanensis from vegetables collected in markets of an endemic region in Peru.Am J Trop Med Hyg 1997 ;57:683-6.
18. Ortega YR, Adam RD. Giardia: overview and update. Clin Infect Dis 1997 ; 25:545-549
19. Peeters JE, Mazas EA, Masschelein WJ, Villacorta Martiez de Maturana I, Debacker. Effect of disinfection of drinking water with ozone or chlorine dioxide on survival of Cryptosporidium parvum oocysts. Appl Environ Microbiol 1989 ;55 :1519-1522
20. Robertson LJ, Campbell AT, Smith HV. Survival of Cryptosporidium parvum oocysts under various environmental pressures. Appl Environ Microbiol 1992 ; 58:3494-3500
21. Rose JB. Environmental ecology of Cryptosporidium and public health implications. Annu Rev Public Health 1997;18:135-161
22. Sparfel JM, Sarfati C, Liguory O, Caroff B, Dumoutier N, Gueglio B, Billaud E, Raffi F, Molina JM, Miegeville M, Derouin F. Detection of microsporidia and identification of Enterocytozoon bieneusi in surface water by filtration followed by specific PCR. J Eukaryot Microbiol 1997 ; 44:78S
23. Smith HV. Detection of parasites in the environment. Parasitology, 1998, 117: S113-S141.
24. Steiner TS, Thielman NM, Guerrant RL. Protozoal agents: what are the dangers for the public water supply?Annu Rev Med 1997;48:329-340
25. Sturbaum GD, Ortega YR, Gilman RH, Sterling CR, Cabrera L, Klein DA. Detection of Cyclospora cayetanensis in wastewater. Appl. Evironn. Microbiol. 1998, 64:2284-2286.

The use of surface waters for agriculture irrigation and of costal waters for growing shellfish bypasses the barriers otherwise erected by drinking water technology. In the last decades tourism has emerged as a booming bussiness; often the tourism is associated with bathing, swimming or practicing aquatic activities during which the association between water and people is very close.

Pathogenic and other microorganisms enter the aquatic environment mainly through municipal wastewaters, treated or partially treated discharges. Many human pathogens are discharged into surface waters from several sources (fig.1). Among them are bacteria, viruses, and parasites. Actually, all possible pathogens present in human waste might find their way into the surface waters. Among them are pathogens belonging to very different biological groups and entailed with different biological properties, a circumstance that has to be considered when it comes to the discussion of the persistence of a pathogen, its resilience against disinfection, and the significance of indicator organisms.

The hygienic quality of surface waters is usually estimated by determining the concentration of faecal indicator bacteria, like faecal coliforms, coliform bacteria (total coliforms), or enterococci (faecal streptococci), and eventually of the bacteriophages.

Three types of bacteriophages have been proposed as specific indicators of viral contamination: the somatic coliphages, the F+specific RNA , and the Bacteroides fragilis phages.

With some exceptions, indicator bacteria are not pathogens, through they reflect the degree of faecal pollution of the water. Many pathogens faecally excreted are not constantly present in the waste waters, or not at the same concentrations over time. Waste water treatement procedures eliminate microorganisms according to their physical and chemical properties. ?

Shellfish (clams, oysters, mussels, etc.) grown in faecally contaminated waters accumulate bacteria and viruses during their filter feeding activity. Human pathogens persist in their feeding organs and are ingested with the shellfish. Shellfish-associated illnesses are well known events, reported repeatedly, over many years. The causative agents were mainly determined to be the enteric viruses ( HAV, Norwalk virus and Small Round Structured Virus).

Many countries have introduced standards of bacterial indicators for either the waters in which the shellfish are reared, for the flesh in the shellfish, or for both. The International trade should take the responsibility of the possible outbreaks due to contaminated seafood.

The main causative agents of the waterborne deseases are enteric viruses: adenoviruses, astroviruses, caliciviruses, enteroviruses, hepatitis A and E viruses and rotaviruses (fig.2). As there are no generally accepted microbial indicators for enteric viruses, there is a clear need for methods and indicators that can be used for direct virological examination of water supplies to better assess health risks due to enteric viruses. There is now sufficient evidence that the bacterial indicators in use are not satisfactory for the assessment of water quality with a view to avoiding health risks. New quality standards should therefore be considered. Enteroviruses could be considered as indicators of the enteric viruses in the environment. The rate of enterovirus shed in raw sewage ranges from 100 to 1000 pfu/100ml. Suspended solids-associated viruses settle onto botom sediments, where the concentration may be 10-10 000x times greater than those found in water. Treatement of sludge by drying, pasteurization, anaerobic digestion, and composting reduces but does not eliminate enteroviruses.

Human enteroviruses, from the Picornaviridae family, are represented by 66 different serotypes (fig.3). They are human pathogens with a wide spectrum of clinical manifestations (fig.4) ranging from non-specific acute infections to serious diseases involving the central nervous system, heart and other tissues. Infected individuals shed virus into feaces and hence in the sewage, usually for several weeks (4), the viruses survive well in sewage and can be readily isolated from concentrated or unconcentrated waters (2,3). Enteroviruses are small, non enveloped viruses, with organized capsid around the single stranded RNA molecule( fig.5).

These viruses, with some exceptions, are culturable in cell culture. During the last years, molecular biology methods have been developed to detect the majority of enteroviruses from clinical and environmental samples (2,3). Molecular techniques, such as RT-PCR offer several advantages (high sensitivity, specificity, low cost and rapidity) for detection of low amounts of viral particles in environment (5). Although this technique connot discriminate between infectious and non-infectious viruses, it offers a rapid way to screen multiple samples. The use of combined technologies of cell culture and RT-PCR increase the equivalent volumes, reduce the effects of toxicity, reduce inhibitory effects on the PCR and maximise the detection of infectious virus using PCR (fig.6).

Enterovirus isolation was possible from sewage, surface waters, sediments, coastal waters and seafood during all seasons but was more frequent in the autumn.

Laboratory studies quantified the relative survival of F+coliphages, B.fragilis phages, hepatitis A virus, poliovirus, and rotavirus in seawater and seawater-sediment mixtures at 5°C and 25°C (1).

F+ coliphages were inactivated faster than the three enteric viruses at 25°C in seawater, but its inactivation was comparable to those enteric viruses in seawater-sediment mixture at both experimental conditions. In contrast, B. fragilis phages survived comparably to or better than enteric viruses in all conditions. Survival rates were significantly greater at 5°C than at 25°C for all five viruses.

The presence of enterovirus genomes in waters may be regarded as an indicator of more or less recent contamination with viral pathogen. It is a useful indicator because of its resistence, its rapidity and sensitivity of detection. A clear correlation was demonstrated between the genome detection and infectious enterovirus detection, even if the latter was detected at lower rates in the treated wastewaters (2). The nucleic acid quantification of the detected genomes in natural water samples and the determination of the threshold for the presence of infectious enterovirus would be the best way to estimate the potential health hazard for the public due to the viral contamination.

1) H. Chung and M.D. Sobsey. Comparative survival of Indicator Viruses and Enteric Viruses in Seawater and Sediment. Wat. Sci. Tech. 1993, 27, 425-428.

2) C. Gantzer, A. Maul, J.M.Audic, and L. Schwartzbrod. Detection of Infectious Enteroviruses, Enterovirus Genomes, Somatic Coliphages, and Bacteroides fragilis Phages in treated wastewaters. Appl. and Environ. Microbiol., 1998, 64, 4307-4312.

3) H. Kopecka, S. Dubrou, J. Prevot, J. Marechal, J. M. Lopez-Pila.

Detection of naturally occuring enteroviruses in waters by reverse transcription, polymerase chain reaction, and hybridization. Appl. and Environ. Microbiol., 1993, 59, 1213-1219.

4) J.L.Melnick. Enteroviruses: Polioviruses, coxsackieviruses, echoviruses and newer enteroviruses. In Fields BN, Knipe DM, ChanochRM.,eds. Fields Virology 2nd ed.vol 1. New York: Raven Press, 1990: 549-605.

5) T.G. Metcalf, J.L. Melnick, and M.K. Estes. Environmental Virology: From Detection of Virus in Sewage and Water by Isolation to Identification by Molecular Biology - A trip of Over 50 Years. Annu. Rev. Microbiol., 1995, 49, 461-487.

More than 100 enteric viruses have been found in human stools and most may also be present in environmental samples. Although these viruses may have pathogenic potential, only a few have been reported to cause environmentally transmitted diseases. Hepatitis A and E viruses (HAV and HEV), rotaviruses, astroviruses, and caliciviruses, including the Sapporo calicivirus, the Norwalk virus and Norwalk-like viruses in particular, are the viruses most frequently implicated in waterborne diseases. With the exception of rotaviruses, all these viruses are non-enveloped plus-sense single-stranded RNA viruses, and belong to the picornaviridae (HAV), astroviridae or caliciviridae families. HEV, which assigned to the caliciviridae, is now an unassigned genus. Rotaviruses belong to the reoviridae family and their genomes consist of 11 segments of double-stranded RNA. They are the commonest cause of diarrhoea and are responsible for more than 25% of all deaths from diarrhoea world-wide. Caliciviruses and astroviruses generally cause less serious transient gastroenteritis. HAV infection is one of the most common causes of infectious jaundice world-wide but can now be prevented by an efficient vaccine. Hepatitis caused by HEV is more severe than that caused by HAV and HEV infection frequently has a fatal outcome in pregnant women. This virus, encountered mainly in developing countries, commonly contaminates water supplies. Enteric adenoviruses (adenoviridae) are naked double-stranded DNA viruses, often found in water. They have been shown to cause gastroenteritis and, in some cases, pharyngoconjunctivitis.

Some other viruses are thought to cause water-borne diseases. Human faecal parvovirus-like virus, a naked single-stranded DNA virus, and coronaviruses, enveloped plus-sense single-stranded RNA viruses, may be excreted by humans with gastroenteritis.

Enteroviruses (picornaviridae) are human pathogens frequently found in environmental waters. Since the start of vaccination programmes against poliomyelitis, these viruses are no longer considered to be a serious threat to human health. However, they are still of interest for medical and public health reasons. They are also a possible indicator and model for studies of environmental contamination with viruses of faecal origin. Studies of enterovirus genomes may lead to the development of molecular biology techniques for differentiating enteroviruses of human origin from those originating from domestic animal species (pigs or cows). This may help to determine the origin of faecal pollution. Human enteroviruses have been pushed into the linelight by vaccination campaigns aimed at eradicating poliomyelitis and poliovirus. Vaccine—derived poliovirus and other enteroviruses are the objects of the accurate virus surveillance required to monitor the disappearance of the poliovirus species from the environment. Enterovirus isolates provide us with the opportunity to study the characteristics and evolution of enteric viruses present in water. Studies of the genetic and phenotypic evolution of poliovaccine strains towards pathogenic viruses, and of the long-term circulation of these vaccine-derived strains, enable us to evaluate the possibility that these strains sustain a reservoir of pathogenic viruses. Genetic variability including mutations and genetic exchanges between polioviruses and between enteroviruses, has been evidenced. Interferences between enteroviruses and perhaps with other enteric viruses in the same ecological niche, may modulate the viral populations present in the digestive tract, faeces and liquid waste. Detailed analysis of the determinants that differentiate serious epidemic enterovirus strains from related strains isolated from sporadic cases may make it possible to assess the risk that these viruses are a potential source of new pathogenic agents and of new water-borne diseases.

During the last century the activated sludge process has become a globally applied key component of modern municipal and industrial waste water treatment plants. It is a matter of common knowledge that biological waste water treatment is mainly driven by prokaryotic activity. However, despite its importance for ameliorating anthropogenic damage to the environment, relatively little is known about the microbiology of activated sludge. For decades, light microscopy in combination with simple staining techniques was the only tool available for microbiologists who were interested to directly study bacteria within their natural environment. However, due to the limited number of morphotypes among prokaryotes standard microscopic techniques can not be used to reliably identify bacteria. Thus, classical cultivation techniques were widely applied to indirectly study the composition of the complex microbial consortia present in activated sludge. Since the majority of microorganisms are notoriously recalcitrant to cultivation, only less than 20% of the microscopically detectable activated sludge microorganisms can be identified using this approach. Furthermore, the inherent selectivity of all cultivation methods strongly biases the composition of the bacterial isolates towards those species best adapted to the offered growth conditions (Wagner et al., 1993).

With rRNA in situ hybridization techniques culture-independent examination of the microbial population structure and dynamics became possible. In combination with confocal laser scanning microscopy and digital image analysis these techniques allow to accurately quantify probe-identified bacterial populations and to study the spatial distribution of bacterial species within activated sludge flocs or biofilms (Wagner et al., 1998; Schmid et al., in press). In the top-to bottom approach a battery of group-specific probes is applied for in situ hybridization for rapid, low-resolution community-structure analysis (Wagner et al., 1993). The composition of bacterial groups with high in situ abundance is further resolved by applying appropriate genus- and species-specific probes. High resolution studies of the microbial community structure of activated sludge are performed best by the bottom-up (or full-cycle rRNA) approach. Using the latter approach almost complete analyses of the microbial community structure of activated sludge and biofilms can be achieved.

Results from applying the full-cycle rRNA approach to activated sludge of an industrial sewage treatment plant will be presented. The treatment plant is connected to a rendering plant and receives sewage with extremely high ammonia-concentrations (>1000 mg/l). Due to an intermittent aeration regime the activated sludge performs quasi-simultaneous nitrification and denitrification. After DNA extraction using three different methods, 16S rDNA libraries were established to obtain a phylogenetic inventory. Comparative sequence analysis of 96 16S rDNA clones revealed the presence of 44 operational taxonomic units (OTUs). Surprisingly, 16S rDNA sequences of ammonia-oxidizing bacteria were not retrieved despite efficient ammonia-oxidation of the activated sludge at the time of sampling. Concerning nitrite-oxidizers, two cloned 16S rDNA sequences were unambiguously affiliated to the genus Nitrospira.

Based on the retrieved sequences specific rRNA-targeted oligonucleotides were designed for fluorescence in situ hybridization (FISH). Using group- and OTU-specific probes, 89% of those activated sludge bacteria detectable in situ by a bacteria-specific probe mix (Daims et al., 1999) could be identified. The in situ abundance of some but not all OTUs differed significantly from their relative frequencies in the 16S rDNA libraries. Consistent with the results from the 16S rDNA libraries, quantitative FISH analysis demonstrated the in situ dominance of beta-subclass Proteobacteria in the activated sludge. Within this subclass Azoarcus-related and Zoogloeae-like organisms were numerically most important. Application of probes specific for beta-subclass ammonia-oxidizers (Wagner et al., 1995; Mobarry et al., 1996; Juretschko et al., 1998) showed a high in situ abundance of Nitrosococcus mobilis, which was previously considered to occur exclusively in brackish water. Among the nitrite-oxidizers, Nitrospira-like bacteria occurred in significant numbers while members of the genus Nitrobacter could not be detected in situ. These results demonstrate that unexpected ammonia- and nitrite-oxidizers are important for efficient nitrification in waste water treatment plants. Almost all nitrifying activated sludge systems analyzed up to now by FISH do contain significant amounts of Nitrospira-like organisms while Nitrobacter is only present in negligible amounts (Wagner et al., 1996; Juretschko et al., 1998). These results are of relevance for the design and control of nitrifying waste water treatment plants since Nitrobacter is still considered as model organism for nitrite-oxidation. More generally the observed differences between the FISH results and the composition of the 16S rDNA libraries strongly questions the suitability of PCR-based techniques for quantitative microbial community structure analysis.

The in situ identification of microorganisms rarely allows to infer their physiological properties. While practically nothing is known about the physiology of yet not culturable bacteria, even bacterial species which have been extensively studied in the laboratory can show different physiological traits within their natural environment. Consequently, it is for example impossible to identify those bacteria responsible for denitrification in activated sludge by using rRNA-targeted oligonucleotide probes. In order to obtain information on the physiological and biochemical activity of bacteria within their environment we combined FISH and microautoradiography (Lee et al., 1999). After incubation of activated sludge with radioactively labeled substrates the microorganisms are fixed by addition of paraformaldehyde. Subsequent application of FISH and microautoradiography allows to directly identify those bacteria which took up labeled substrate. Using this technique we could identify Azoarcus-related microorganisms as numerically important denitrifiers in activated sludge. Furthermore, we are currently studying the in situ physiology of the yet not culturable Nitrospira-like nitrite-oxidizers in activated sludge and biofilms. First results demonstrate that these bacteria can fix CO₂ under aerobic conditions but are not able to take up acetate. In addition to Nitrospira, many more still uncultured or even not recognized bacteria are responsible for efficient nutrient removal in sewage treatment. Molecular surveys of plants with enhanced biological phosphorous removal (Bio-P) have shown that this process is not catalyzed by its model organism Acinetobacter, but rather involves not further characterized members of the beta-subclass of Proteobacteria as well as certain Gram-positive bacteria with a high DNA G+C content (Wagner et al., 1994a; Bond et al., 1999; Hesselmann et al., 1999). We are now intensifying the hunt for the Bio-P organism(s) by combining the full-cycle rRNA approach with microautoradiography. Recently, a not yet culturable bacterium capable to catalyze the anaerobic oxidation of ammonium were identified as a novel, deep-branching member of the Planctomycetales (Strous et al., 1999). Using the full-cycle rRNA approach we could demonstrate genus level-diversity of bacteria capable to anaerobically oxidize ammonium and their in situ dominance in biofilms of trickling filters and contacting disc reactors with unexpected N-losses (Schmid et al., in press). A detailed understanding of the in situ physiology of these microorganisms will almost certainly allow the design of novel treatment strategies for nitrogen removal from sewage with high ammonium and low organic carbon contents.

In conclusion, the battery of molecular tools which has been developed does today allow a detailed census of the microbes present within sewage treatment plants. Now, we have to precisely define and evaluate the links between microbial community composition and function. Those microorganisms responsible for nitrogen- and phosphorous-removal have to be identified and their in situ physiology has to be characterized. The obtained results can directly be funneled into mathematical activated sludge and biofilm models and will, in addition, promote a coming of age of bioaugmentation strategies. The set of probes available for rapid and reliable in situ identification of filamentous microorganisms (Wagner et al., 1994b), which are responsible for sludge bulking and foaming, has to be extended and the ecology of these microorganisms has to be studied in detailed in order to develop efficient control strategies for these harmful microorganisms. Furthermore, factors influencing the stability of those microbial populations responsible for nutrient removal against perturbations have to be identified. Intuitively, one would assume that the diversity of the respective population has a pronounced influence on the stability of the process. If this holds true we might be able to increase the diversity of microbial key populations by smart treatment strategies and thus protect the microbial consortium against process failure.

References

Bond P.L., Erhart R., Wagner M., Keller J., Blackall L.L. 1999. Identification of some of the major groups of bacteria in efficient and nonefficient biological phosphorus removal activated sludge systems. Appl. Environ. Microbiol. 65:4077-4084.

Daims, H., Brühl, A., Amann, R., Schleifer, K.H., and Wagner, M. 1999. The domain specific probe EUB338 is insufficient for the detection of all Bacteria: Development and evaluation of a more comprehensive probe set. Syst. Appl. Microbiol. 22:434-444.

Hesselmann R.P., Werlen C., Hahn D., van der Meer J.R., Zehnder A.J. 1999. Enrichment, phylogenetic analysis and detection of a bacterium that performs enhanced biological phosphate removal in activated sludge. Syst. Appl. Microbiol. 22:454-465.

Juretschko, S., Timmermann, G., Schmid, M., Schleifer, K.-H., Pommerening-Röser, A., Koops, H.-P., and Wagner, M. 1998. Combined molecular and conventional analyses of nitrifying bacterium diversity in activated sludge: Nitrosococcus mobilis and Nitrospira-like bacteria as dominant populations. Appl. Environ. Microbiol., 64:3042-3051.

Lee, N., Nielsen, P.H., Andreasen, K.H., Juretschko, S., Nielsen, J.L., Schleifer, K.-H., and Wagner, M. (1999). Combination of fluorescent in situ hybridization and microautoradiography - a new tool for structure-function analyses in microbial ecology. Appl. Environ. Microbiol. 65:1289-1297.

Mobarry, B.K., M. Wagner, V. Urbain, B.E. Rittmann, and D.A. Stahl. 1996. Phylogenetic probes for analyzing abundance and spatial organization of nitrifying bacteria. Appl. Environ. Microbiol. 62:2156-2162.

Strous M., Fuerst J.A., Kramer E.H., Logemann S., Muyzer G., van de Pas-Schoonen K.T., Webb R., Kuenen J.G., Jetten M.S. 1999. Missing lithotroph identified as new planctomycete. Nature 400:446-449.

Wagner, M., R. Amann, H. Lemmer, and K.H. Schleifer. 1993. Probing activated sludge with oligonucleotides specific for proteobacteria: inadequacy of culture-dependent methods for describing microbial community structure. Appl. Environ. Microbiol. 59:1520-1525.

Schmid, M., Jetten, M.S.M., Klein, M., Juretschko, S., Twachtmann, U., Metzger, J.W., Schleifer, K.H., and Wagner, M. Molecular evidence for genus-level diversity of bacteria capable to catalyze anaerobic ammonium oxidation. Syst. Appl.. Microbiol., in press.

Wagner, M., R. Erhart, W. Manz, R. Amann, H. Lemmer, D. Wedi, and K.H. Schleifer. 1994a. Development of an rRNA-targeted oligonucleotide probe specific for the genus Acinetobacter and its application for in situ monitoring in activated sludge. Appl. Environ. Microbiol. 60:792-800.

Wagner, M., R. Amann, P. Kämpfer, B. Assmus, A. Hartmann, P. Hutzler, N. Springer, and K.H. Schleifer. 1994b. Identification and in situ detection of gram-negative filamentous bacteria in activated sludge. System. Appl. Microbiol. 17:405-417.

Wagner, M., G. Rath, R. Amann, H.-P. Koops, and K.-H. Schleifer. 1995. In situ identification of ammonia-oxidizing bacteria. System. Appl. Microbiol. 17:251-264.

Wagner, M., G. Rath, H.-P. Koops, J. Flood, and R. Amann. 1996. In situ analysis of nitrifying bacteria in sewage treatment plants. Wat. Sci. Tech. 34 (1-2):237-244.

Wagner, M., P Hutzler, and R. Amann. 1998. 3-D analysis of complex microbial communities by combining confocal laser scanning microscopy and fluorescence in situ hybridization (FISH). In: Digital Image Analysis of Microbes (Wilkinson, M.H.F.; Schut, F., eds.); pp. 467-486. John Wiley & Sons, Chichester, England.

1. Introduction. Since the mid-80s it has become increasingly evident that recombinant DNA technology could be successfuly used for constructing bacterial strains with novel phenotypes targeted to biodegradation of environmental pollutants. Thus, many expectations have been raised on the possibility to use genetically modified microorganisms (GMOs) to address otherwise intractable contamination problems caused by aromatic compounds and other chemical wastes and side-products. However, it is likely that the tremendous power of modern molecular biology applied to this field cannot be fully realized until some technical gaps are filled, both in side of genetics and that of process engineering. Many perceive that molecular genetics has failed so far to produce authentically useful GMOs with the degreee of predictability in their performance and ecological behaviour which would be required for applications outdoors. For nearly one decade, a number of strains, belonging mostly to the genus Pseudomonas, have been developed with enhanced biodegradation capabilities. However, one issue is biodegradation of one particular chemical in a Petri dish under the well-defined conditions of the Laboratory, and a very different one is to achieve the proper expression of the degradative phenotype under the conditions prevailing in the field, on which we have little or no control. This is a major difference with other biotechnological processes (i.e. based on bioreactors) in which the working conditions can be fixed externally.

Fortunately, the last few years have witnessed rapid advances in the design of genetic assets specifically tailored for construction of GMOs which are more predictable in terms of performance and ecological behaviour. In this communication I will briefly disclose some problems that the molecular geneticist has to face -and the potential solutions available, for designing recombinant bacteria fit and reliable for applications in the field. Such predictability is a prerequisite for the development of innoculation techniques and large-scale bioprocess engineering which are essential for bioremediation applications.

2. Imitating Nature: Insertion and deletion of DNA segments in the chromosome of bacteria. Adaptation of bacteria to novel carbon sources or to changing environmental conditions is frequently accompanied by the acquisition or loss of single or multiple DNA segments which encode the functions required for survival. One archtypical case of adaptation is the sequence of molecular events which lead to the assembly of metabolic pathways for the degradation of xenobiotic compounds. Analysis of the DNA sequences of the corresponding catabolic genes and operons suggests that they have resulted from joining DNA segments recruited from preexisting pathways in which a small number of changes in their DNA sequence have been effected. Modularity of the genetic complement of bacteria implies the existence of natural genetic engineering systems to assemble the various sequence components of the different elements. Catabolic operons are frequently included within transposons which, at least in the case of the Tn3 /Tn21 family, have the capacity to exchange DNA segments through transposon-determined site-specific recombination systems.

What practical lessons should we learn from these molecular processes? In an effort to imitate in the Laboratory some of the events which lead naturally to the creation of new degradative phenotypes, a series of transposon-vectors derived from Tn5 have become available since 1990 which permit the insertion and stable inheritance of heterologous DNA segments into the chromosome of target bacteria through a process which resembles considerably the existing mechanisms of acquisition or loss of adaptation-related phenotypes. Transposon vectors permit to engineer stable recombinant phenotypes with very few manipulations in the Laboratory. Their simple design has made them increasingly popular for construction of GMOs for uncontained applications. These vectors have found broad utilization for manipulations of a variety of bacteria including E. coli, Klebsiella, Salmonella, Proteus, Vibrio, Bordetella, Actinobacillus, Rhizobium, Rhodobacter, Agrobacterium, Alcaligenes and several pseudomonads. Besides allowing insertion of heterologous genes into the chromosome of various Gram-negative strains, transposon-vectors permit also to employ native promoters for expresion of recombinant genes even in the absence of any information on promoter structure or regulation in the native host.

In a further step to develop tools for metabolic engineering, i. e., degradation of toxic compounds, inspired by natural processes of DNA insertion/excision, the capability of the multimer resolution system (mrs) of broad host range plasmid RP4 has been exploited to generate predetermined deletions within DNA segments inserted in the chromosome of Gram-negative bacteria. The mrs system is one of the functions present within the par region, that determines the major plasmid stabilisation system of RP4. The product of the parA gene is a site-specific resolvase that catalyzes intramolecular recombination between two directly oriented res sites flanking any supercoiled DNA. The result of this process is the excision of the intervening nucleotide sequence. Transient expression of parA suffices to effect differential deletions within DNA segments inserted in the chromosome of Pseudomonas putida and other Gram-negative strains, thus permitting eventual addition of heterologous DNA segments devoid of any phenotypic marker to the genome of the target bacterium. At the end, the resulting microorganism endowed with a novel phenotype determined by the DNA segment of interest does not carry any of the genetic determinants used for its construction. Therefore, utilization of this new type of mrs-vectors leaves little room for the somewhat arbitrary distinction between recombinant and natural bacteria, since they mimic faithfully the natural processes of DNA shuffling between distant locations of the bacterial genome. Should such an absence of phenotypic markers be undesirable, a whole collection of alternative non-antibiotic selection determinants have become available also, combined with transposon-vectors to produce equally effective mobile elements devoid of antibiotic resistances. Such alternatives include the use of resistances to herbicides such as bialaphos and glyphosate, nutritional markers like growth on lactose as the only carbon source and resistances to heavy metal ions like mercury, arsenite or tellurite.

All these recent developments have virtually solved the problem of genetic stability of the recombinant genes as well as that of the use of antibiotic gene markers as selection determinants, the later being one of the features of recombinant organisms which, to this day, cause more concern in public perception.

3. Gene expression in the field: Can we force bacteria to do what they normally do not do? Our possibilities to imitate Nature also in the way catabolic operons are expressed are still hampered by the very little information available on promoters of biodegradative pathways. Only the regulators of some toluene (TOL), naphtalene (NAH) and catechol/Cl-catechol (cat, clc, tcb) degradation systems have been studied to a greater molecular detail. Even with this very scarce information, the comparison of these few cases of regulation of some biodegradative pathways of Pseudomonas puts forward very intriguing questions on how regulatory circuits evolve. Different transcriptional control systems seem to gain the same ability to regulate expression of similar phenotypes which are essential for adaptation to a hostile environment. Understanding how such controls develop and eventually work is of essence in cases where signals present in the external medium are desired to drive transcription of a recombinant phenotype in the field or when the genetic manipulation engineered in the recombinant strain is expected to lead to overexpression of a certain gene or gene cluster.

A key question to be addressed is whether the energy balance of the cells can support efficient expression of an energy-costly phenotype from which there is not much metabolic return. It is a widespread observation that biodegradative pathways are subjected to some type of catabolite control which prevents transcription of the corresponding genes as long as other carbon sources are available. It is therefore highly desirable to uncouple growth from expression of catabolic genes, what will require a substantial (and certainly rewarding) research effort to identify the corresponding genetic and biochemical connections and bottlenecks.

In spite of these limitations, there is already available a variety of genetic tools to engineer gene expression under uncontained conditions. Regulated promoters of the few studied biodegradative pathways of Pseudomonas have been particularly appealing to design heterologous expression sistems in the field. Unlike schemes used in the laboratory (thermosensitive promoters, lac-derived promoters), catabolic promoters can be induced with low concentrations of generally cheap inducers, thus making possible extensive induction operations which would be just unrealistic for other expression systems. Along these lines, a number of recombinant transposon vectors have been constructed which contain various catabolic promoters (along with their cognate regulatory genes) assembled in such a way that any gene of interest can be expressed upon exposure of the cells to specific aromatic inducers such as toluenes, xylenes, benzoates and salicylates. Furthermore, some of these promoters have been combined with the gene for the T7 phage polymerase so that regulatory cascades can be designed to express different genes or gene clusters through the addition of a single inducer. This is particularly interesting in cases when various DNA segments containing catabolic genes are engineered for simultaneous expression triggered by a single chemical signal, i.e., a common metabolite for all pathways.

In some cases, starvation might be considered an universal signal potentially useful to drive gene expression when no other know signal can be used for the purpose. Promoters responsive to carbon, nitrogen, iron and phosphate starvation, are in principle, adequate building blocks for expression of heterologous genes in the field. Furthermore, some specialized genetic probes are available to identify promoters which are preferentially active wen cells have ceased to grow. Such type of promoters may have interesting applications in GMOs destined to environmental release which need to perform under nutrient starvation and/or very low growth rates.

4. Outlook. Legal troubles notwithstanding, an authentic bottleneck for the use of GMOs in uncontained applications has been, so far, that of ecological predictability. However, the recent development of a whole collection of genetic tools specifically tailored for metabolic engineering of novel catabolic pathways has reached such a sophistication that genetic unstability, presence of antibiotic markers or expression of heterologous genes in the field do not represent anymore a serious tecnical problem. Bioremediation of sites polluted with recalcitrant chemicals is one of the cases where the use of recombinant bacteria may prove to be extremely useful in the not-so distant future, as soon as matching technologies evolve in parallel with the design of novel bacterial phenotypes.

5. Recent references.

Panke, S., V. de Lorenzo, A. Kaiser, B. Witholt, and M. Wubbolts (1999) Engineering of a Stable Whole-Cell Biocatalyst Capable of (S)-Styrene Oxide Formation for Continuous Two-Liquid-Phase Applications. Appl. Envir. Microbiol. 65: 5619-5623.

de Lorenzo, V. and Kuenen, G. (1999). Scientific basis for the remediation of the toxic spill of the Aznalcóllar mine : combining bacteria and plants to address an intractable kind of pollution. Enviromental Microbiology 1: 275-278.

Lau, P., and de Lorenzo, V. (1999) Genetic engineering: the frontier of bioremediation. Environmental Science & Technology (News & Reseach Notes) 4: 124A-128A.

Sánchez-Romero, J. M. and V. de Lorenzo. 1999. Genetic engineering of non-pathogenic Pseudomonas strains as biocatalysts for industrial and environmental processes. In Manual of Industrial Biotechnology (2nd ed). J. Davies, G. Cohen (ed). American Association of Microbiology Press, Washington DC. pp. 460-474.

de Lorenzo, V., Herrero, M., Sánchez, J.M. and Timmis, K. N. 1998 Mini-transposons in microbial ecology and environmental biotechnology. FEMS Microbiol. Ecology 27: 211-224.

Cases, I. and de Lorenzo V. 1998. Expression systems and physiological control of promoter activity in bacteria. Curr. Op. Microbiol. 1 : 303-306.

Panke, S., Sánchez-Romero, J. M. and de Lorenzo, V. 1998. Engineering quasi-natural Pseudomonas putida strains for metabolism of toluene through an ortho-cleavage degradation pathway. Appl. Environ. Microbiol. 64 : 748-751.

Gallardo, M. E., Fernández, A., de Lorenzo, V., García, J.L. and Díaz, E. 1997. Designing recombinant Pseudomonas strains to enhance biodesulfurization. J. Bacteriology 179: 7156-7160.

de Lorenzo, V. (1997) Bioremediation : more than cloning and pasting DNA. Trends in Biotechnology 15 : 235-236.

Sousa, C., Cebolla, A. and de Lorenzo, V. 1996. Enhanced metallo-adsorption and metallotropism of bacterial cells displaying poly-His peptides on their outer membrane. Nature Biotechnology 14, 1017-1020.

Introduction

Waterborne diseases are at a relatively low level in most industrialized countries but waterborne diseases have killed innumerable people since man has been living on earth. Waterborne diseases still kill millions of children and adults every year, especially in developing countries that are struggling for water itself.

In the course of the 20th century, water treatment has become a science. A science that relies on scientific knowledge to better understand what chemicals and microorganisms are present in water and how to remove them it in order to minimize risks to the population. The link between disease and water has been known for centuries, but it only through the works of early microbiologists such as Pasteur that the link between the newly identified microorganisms and specific diseases was demonstrated. The epidemiological demonstration by John Snow that cholera was transmitted by water from a well, resulted in the first water treatment: removing the handle of the pump on the contaminated well controlled the outbreak of cholera that was ongoing in London.

This paper will review briefly the principles that have guided public health authorities in their struggle against waterborne diseases, the methods used for water treatment and their effect on the level of illness. It will also attempt to provide some insight on the developing methodologies for treating water in an overall scheme of risk management of waterborne diseases. The messages that will be presented are:

The microorganisms and the diseases

When the link between disease and specific pathogenic microorganisms was demonstrated, epidemiological studies rapidly evolved in establishing their effect on populations. Early studies centered mainly on bacterial or parasitic pathogens that could be grown on artificial media or seen through the microscope. Viruses that had been suspected for decades were finally seen with the advent of the electron microscope.

The nature of the waterborne pathogens and their principal source, the feces, require that the control of these enteric diseases is approached on two fronts: sanitation and water supply. Both will limit the amount of fecal material reaching us, and thus the ensuing risk of being infected by the pathogens it contains as most pathogens found in fecal material can survive in the environment for a certain period awaiting a susceptible host. The microbial content of sewage provides a good example of what lurks around in our environment (Table 1).

A colleague once reflected that we live in a world covered by a film of fecal material and that the difference between countries is its thickness. This thickness can be reduced by sanitation, wastewater facilities, personal hygiene, and appropriate food processing and conservation. Low levels of fecal bacteria are unavoidable as they are part of our daily lives including contact with other individuals and inanimate surfaces and objects. The objective is thus to reduce their level and prevent the pathogens from reaching our mouth and nose, their portal of entry.

In terms of disease-causing waterborne pathogens, those that have been targeted as being the most important are the human enteric viruses and two protozoan parasites, Giardia lamblia and Cryptosporidium parvum. The reason for their importance lies in their physical nature: they are small and very resistant to treatment. Protozoan parasite cysts and human enteric viruses can survive at pH 3 to 4 for prolonged periods. In the aquatic environment they can survive for several months especially when water is cold. Viruses are so small that they are difficult to remove by filtration and some are quite resistant to disinfection by chlorine. Protozoans cysts can be partly removed by filtration if it is applied efficiently, but they are very resistant to disinfection. Their presence in tapwater resulted in large outbreaks. Bacterial pathogens are in lower numbers in contaminated water, most of them do not survive well in the environment and they are quite sensitive to water treatments, especially disinfection.

At the beginning of the 20th century, the number of outbreaks and cases of waterborne disease was high in most industrialized countries, but it was rapidly reduced through the development of water supplies, filtration and disinfection. At the same time, pasteurization, better food preparation and conservation in refrigerators and freezers also reduced considerably the level of enteric diseases. Cholera and typhoid fever are good examples of waterborne diseases controlled by sanitation and water treatment during the 20th century. The number of cases and the number of reported waterborne outbreaks were at their lowest between 1950 and 1960. In the United-States, where a good reporting system was in place, this changed in the second part of the century and the number of reported outbreaks increased slightly and culminated recently by several large outbreaks due to the parasite Cryptosporidium. Similar outbreaks have been reported in many countries and reflect the ubiquitous nature of these pathogens.

The low level of endemic gastrointestinal illness in industrialized countries is deceiving in terms of waterborne disease: the proportion of illnesses attributable to water could be the same as demonstrated by our studies in Canada. The rate of gastrointestinal diseases in developing countries can be as high as 5 to 10 episodes per person per year, with an important rate of mortality especially in children. The proportion of illnesses due to water had not been measured until recently. Two epidemiological studies conducted in Canada by our group have suggested that up to half of the enteric illnesses reported were due to tap water meeting North American water quality standards. Part of the problem in establishing the contribution of water to the disease burden is the variability of the rate of gastrointestinal illnesses that can vary from low in summer to high in winter in the Northern hemisphere.

Lessons from the 20th Century

Several lessons have been learned during the last century and I have summarized them under 5 categories under which intervention is possible to reduce or eliminate the microbial risks associated with drinking water:

Level 1: Find a clean and protected source of drinking water.

Level 2: Remove as many contaminants as possible.

Level 3: Kill the remaining microbes.

Level 4: Maintain water quality in the distribution system.

Level 5: Use appropriate quality control measures.

Level 1: Use a safe source of drinking water and protect it. Pathogenic microorganisms in water originate from 2 sources: man or animals. While some non-enteric microorganisms can be transmitted through water (i.e., bilharziosis), the main focus of water treatment is to protect from enteric pathogens found in the feces of warm-blooded animals. Water unpolluted by sewage or animal feces should be a much safer source of drinking water. The first rule for selecting a water source is to find such an unpolluted stream or groundwater and to put in place a watershed protection program. Both animal and human presence should be limited and controlled. This certainly creates problems for watersheds that span over several countries.

Level 2: Remove as many contaminants as possible. Sand filtration has been the main method used for the purification of water. It is relatively inexpensive, but it requires large infrastructures with expert management and continuous monitoring. Maintaining the level of removal to peak performance is difficult and even then, when treating poor quality raw water, the results are often far from perfect. Our laboratory has detected viruses in finished drinking water on several occasions and others have found protozoan cysts in many filtered waters. Slow sand filtration and bank infiltration have provided better results in pathogen removal where they have been used in conjunction with watershed protection. Even when turbidity levels are quite low (0.3 to 0.5 NTU), increases in turbidity have been suggested as indicative of an increase in the rate of gastrointestinal disease. Links between gastrointestinal disease in the population and turbidity have recently been reported in the literature. While this remains to be further substantiated, it is quite plausible as waterborne outbreaks have often been traced to poor filtration and turbidity levels approaching 1 NTU. Membrane technology has developed considerably over the last decades and has moved from point-of-use devices to large-scale plants that can provide water to large populations. Using membrane technologies, safe drinking water can be produced from heavily contaminated water as well as from water with high salt content.

Level 3: Kill the remaining microbes. This has been the approach taken by most countries since the beginning of the 20th century. Chlorine disinfection provided an easy approach to microbe-killing. Chlorine disinfection in individual dwellings or as part of a water supply often reduces dramatically the level of gastrointestinal disease in countries with poor sanitation. Various forms of chlorine have been used and their effects on pathogens can be quite different. Ozone can be a very powerful disinfectant: its production is costly and it does not leave a residual (good and bad). The problems associated with disinfectants have been touched briefly earlier. Some viruses such as coxsackieviruses require long contact time with chlorine before their number is sufficiently reduced to protect public health. Most cysts of protozoans are almost impervious to treatment with chlorine requiring extremely long contact times at the usual doses of chlorine (0.5 to 2 mg/L). The potential cancer-causing effect of disinfectant by-products has raised serious concerns on the use of chlorine disinfectants in water that is rich in organic material: this remains to be resolved by health authorities. Disinfection is governed by concentration, time and temperature: temperature plays a critical role especially in colder climates where water temperature can be close to 0°C in winter. The killing efficacy of chlorine can be several orders of magnitude lower in cold water.

Level 4: Maintain water quality in the distribution system. High quality drinking water produced at a centralized water treatment plant can be degraded very rapidly in a poor distribution system. In many developing countries where the infrastructure is difficult to maintain this is quite a challenge. In industrialized countries, a distribution system hundreds of kilometers in length and often near the sewerage system, is probably infiltrated daily. While it was believed by many that the pressure maintained in the distribution system was sufficient to prevent the entry of contaminated material, this is now seriously challenged both by epidemiological and engineering studies. This also brings forward the control of water in the distribution system. Many countries still rely heavily on a chlorine residual to protect public health from the entry of contaminated water hoping that the residual concentration will be sufficient to kill the pathogens. This is controlled by the measurement of the residual chlorine concentration and by sampling for coliforms (total, fecal or E. coli) at various points in the system. The efficacy of these parameters is challenged by many scientists. The main reason lies in the intrinsic resistance of pathogens to chlorine. While coliforms and some bacterial pathogens are easily killed by even low levels of this disinfectant, none of the viral and protozoan pathogens are seriously affected. The absence of coliform results in a false sense of safety that has been challenged by recent epidemiological studies linking gastrointestinal disease to water free of these indicators.

Level 5: Use appropriate quality control measures. For all of the 20th century, coliforms have remained the principal indicator of water quality. From total coliform counts, it became the thermotolerant (fecal) coliform count that best represented contamination by feces from warm-blooded animals. With the development of new culture media and fluorochromes the culture and detection of Escherichia coli has become routine but it use has been challenged. It is too sensitive to reflect correctly water treatment and it does not correlate with health effects. Thermotolerant coliforms and E. coli are good indicators of fecal pollution but they appear to have little value for other purposes. Various parameters, microbial or physicochemical have been proposed to better reflect virus and pathogen removal as well as disinfection efficiency. The most promising parameters for drinking water are turbidity and particle counting because they can provide in-line continuous measurement before the water is distributed. Microbial parameters (culture-based or using molecular methods) will remain quality control tools for the final validation of the treatment processes and for the sanitary evaluation of watersheds.

What lies in the future: the coming century

The facts that I have been presented reflect some of the current knowledge on drinking water as it was accumulated during the 20th century. We can now look in the crystal ball and attempt to see what is coming for both industrialized and developing countries.

In developing countries, the task still remains at the level of sanitation and education. The simple fact that electricity, and thus pressure in the distribution system, cannot be maintained continuously suggests that efforts should be put at other levels. It has been shown that putting in place a disinfected water distribution system in an unhygienic environment can result in large waterborne outbreaks. The distribution system becomes the vehicle for poorly treated water that is consumed by individuals that are more susceptible to the pathogens because they have been less exposed. Treating drinking water to the highest standards in an environment that is still highly contaminated and where a distribution system is difficult to maintain is not cost effective in socioeconomic and public health terms. Sewerage, wastewater treatment, education in hygienic practices, food protection and disinfection of drinking water at the family level should provide the best results. In these countries, the control of waterborne diseases will also rely heavily on microbial testing of the environment. Low cost methods for the detection and enumeration of microbial indicators of fecal pollution should be developed for poorer countries as well as for poorer or remote areas of industrialized countries.

In developing countries that have attained a higher level of industrialization, when treated water can be provided under pressure continuously to households that have a relatively good level of hygiene, wastewater treatment and watershed protection should become the critical points to insure minimal contamination of the water to be treated. This is a place where better indicators of water quality and treatment will also be needed.

In industrialized countries, where a very high level of disease control has been achieved, the obligation of providing drinking water continuously with minimal health risk will compel water utilities to:

Conclusion

The principles for the control of waterborne diseases in the 3rd millenium will not be very different from what has been done in the 20th century. The main objective is to prevent pathogens in fecal material from reaching us. The control of fecal pollution through sewerage and full wastewater treatment remains the basic step. This will insure protection of the watersheds thus providing clean beaches and unpolluted drinking water sources. In developing countries, in addition to the fecal pollution control through sanitation, the level of hygiene will need to be enhanced for a better control all enteric diseases in a holistic approach. Only then will it be possible to apply the most modern methods of pathogen removal, disinfection of drinking water and quality control using the best available indicators to all countries.

The benefits of improved water and sanitation include both health and non-health effects. The direct health benefits are related to two contrasting roles of water: that of disease vector when it carries pathogens; and that of health mediator through its use in personal and domestic hygiene. Indirect effects related to health include for example improved quality of life and decreased expenditure on medical expenses. Non-health effects include time savings for productive activity or education.

One of the most fundamental developments in ‘sector wisdom’ regarding the interactions between water and health was the development of a categorisation of disease based in terms of how an intervention would change the situation (Bradley, 1974; Bradley and Emurwon, 1968). Four fundamental categories of disease were identified, all of which concerned microbial hazards of diverse types as follows:

Water-borne disease transmission occurs when a pathogen (or noxious chemical) is contained in water which is consumed whether as drinking-water or otherwise. This includes all of the faecal-oral diseases and therefore many causes of diarrhoea, cholera, typhoid and hepatitis A.

Water scarcity or water-washed diseases arise from transmission which can be reduced by increased water use for personal and domestic hygiene. Since faecal-oral disease transmission may be interrupted by personal and domestic hygiene this group includes many of the infectious water borne diseases in addition to eye infections and infections and infestations of the skin.

Water-based disease occurs through transmission of a pathogen with an obligatory aquatic intermediate host or hosts. Examples include the schistosomes and Guinea Worm. The latter is also water borne but is not water-washed.

Water-related insect vector diseases involve transmission by insects which breed in water or which live and bite near water. This group includes malaria since its insect vector breeds in water and the arthropods which carry the viruses which cause dengue and yellow fever.

These mechanisms are not necessarily mutually exclusive. Most infectious water-borne disease for example is also water washed (being transmitted by the faecal-oral route) (Bartram and Ballance, 1996).

There is also a series of secondary and indirect health benefits which may arise from improved water supply services and dis-benefits associated with inadequate water supply. They are of three inter-linked types:

• the effects of ill health (from both the water supply per se and the act of securing water from alternative sources);
• time and energy saved or lost (arising from both ill health and from time spent in water collection and household management); and
• financial (through cost of care for ill health and sometimes from purchase of water).

Both direct and indirect health effects are closely inter-related. Thus for example time demands for water collection increase pressure to utilise more accessible (but potentially more contaminated) nearby sources of water; and the time and energy costs of fetching water govern women’s perceptions of the importance of hygiene in disease prevention and take priority over them in practice (Kendie, 1992).

Diverse social benefits may also accrue which are largely independent of the quality of the water supply service per se. Thus for example Krishna (1990) reports breakdown in patterns of class-based segregation and decreased importance being given to the veiling of women in some areas in addition to time savings. Others have indicated accelerated political maturity, gender equality, community cohesion and national identity (IRC, 1992b). Time, energy and other savings derived from improved water supply accrue particularly to women and children and children may be more likely to attend school (Falkenmark, 1982).

The disease burden associated with inadequate access to safe drinking water and the health impact of improved water supply are distinct, although the difference is sometimes overlooked. There are numerous identified problems in undertaking assessments (World Bank, 1976; Blum & Feachem; 1983; Esrey et al., 1985; Esrey et al., 1991) and as a result a limited body of evidence is available.

The most recent estimates of the global burden of disease (GBD) suggest that around six per cent of the global disease burden is linked to ‘basic hygiene’ (water, sanitation food, hygiene behaviours). However these estimates do not take account of more severe infectious outcomes that can be transmitted or prevented through safe drinking water supply (such as infectious hepatitis or typhoid); neither do they take account of water-washed disease, nor water contact, nor vector borne disease (schistosomiasis, malaria) nor non-infectious diseases (arsenicosis, fluorosis, (Murray and Lopez, 1996) suggest a declining relative importance for some of the basic hygiene diseases. This trend may not seem to be in line with other authoritative sources that suggest an ‘unequivocal increase in infectious disease attributable to microbiologically unsafe drinking water’ (Ford and Colwell, 1996). Such GBD estimates relate to present-day disease burden and not to the gains already achieved and ‘held in check’ by good water and sanitation management.

The reviews of health impact studies by Esrey et al. (1985; 1991) like that of Cairncross (1990a) suggest that water supply and sanitation improvements can reduce the overall incidence of infant and child diarrhoea substantially, in the range of 15 - 36 per cent for single or combined interventions, and total infant and child mortality substantially. Thus, through an exhaustive critical appraisal of published health impact studies, Esrey and co-workers (Esrey et al., 1985; 1991) reported the results summarised in Table 1.

Generalised analyses of this type may undervalue the importance of water quality, because of the ability of piped water supplies to propagate outbreaks of disease. Large-scale outbreaks such as those caused by Cryptosporidium in Milwaukee, USA (Mackenzie et al., 1994) and hepatitis, India (Ramalingaswami and Purcell, 1988) and the outbreaks of cholera in Latin America (Craun et al., 1991) attest their importance and are unaccounted for in such studies. The effect of quality during extreme disruptions, whether seasonal or related to natural disasters may also be underestimated (Ramalingaswami and Purcell, 1988).

The development of standards for drinking-water quality centres upon a risk management process which normally consists of three steps: research, risk assessment and risk management. This sequential process as a linear flow of activity is characterised by a hazard-specific outlook and is frequently also highly compartmentalised, such that hazards arising in drinking-water for example are considered in isolation of other hazards. The result is a failure to relate associated hazards with one another; and problems in relating the degree of health protection secured against one hazard with another. Water-related hazards may have some commonality. For example inadequate treatment of sewage to inactivate human pathogens may lead to exposures through drinking-water, recreational water use or foodstuffs through irrigation, aquaculture or shellfish farming or harvesting. Approaches to risk management which are excessively compartmentalised may prioritise interventions which relate to the route of exposure, such as drinking-water treatment to the detriment of more general protective measures such as water resource and source protection.

Alternatively, risk management could be viewed as a complete circle, showing the feedback of policy evaluation into hazard identification and prioritization. Thus the effectiveness of policy implementation could be validated and available, usually limited, resources used in a optimal manner for health protection.

WHO supports the development of rational, health based standards through normative 'Guidelines'. The three principal normative Guidelines are the Guidelines for Drinking-water Quality, the Guidelines for Safe Use of Wastewater and Excreta in Agriculture and Aquaculture and Guidelines for Safe Recreational-water Environments (the latter in two volumes, one addressing coastal and freshwaters and the other addressing swimming pools, spas and similar environments. These Guidelines are based upon critical review of the available evidence and upon scientific consensus.

WHO Guidelines are not international standards. They are intended to assist competent national and other authorities to establish standards protective of public health. In doing so such authorities will take account of social, economic and environmental factors that may result in Standards significantly different to the Guidelines themselves. Because resources may be very inadequate when confronted by the full range of hazards to be addressed it will often also be beneficial to adopt strategies towards risk-benefit and progressive implementation.

Recommendations made in the Guidelines for Drinking-water Quality are usually presented as numerical values termed guideline values. In the context of this paper, guideline values and criteria are synonymous: both are science-based and both are of an advisory nature.

The development of guideline values entails three major steps:

• selection of priority microbial and chemical contaminants to be evaluated;
• quantitative assessment of the risk to human health from exposure to these microbial and chemical contaminants;
• using certain exposure assumptions, derivation of health-based guideline values that are protective of public health and for which practical methods of analysis are available.

Chemical versus Microbiological Risks

Prevailing standards for microbiological quality of drinking water originate from the early twentieth century when they were developed to monitor the treatment processes (slow sand filtration, chlorination) that had proven successful in controlling bacterial pathogens causing diseases such as cholera and typhoid. These problems are still real and important in large parts of the globe and such standards and their application for this purpose remain valid.

Whilst the development of quantified risk assessment in relation to chemicals in drinking-water has been entirely logical, as has its expression in terms of end-product quality targets, this relates to features of chemical quality which are not mirrored by microbes. For the great majority of chemicals occurring in drinking-water, health effects arise from long-term exposures and acute exposure is of limited health significance. Exceptions such as nitrate do of course exist, but concentrations of most chemicals sufficient to cause health effects from short-term exposure are relatively rare and when they occur it is often in association with conditions such as spills that render the water unpalatable.

However in the case of microbes, health effects arise from acute exposure - a single glass of water may contain an infectious dose and lead to disease. Furthermore, a limited number may reasonably easily be identified as presenting the greatest challenges to provision of safe drinking water - whether through treatment or through protection of sources and it is therefore feasible to identify critical challenges. These challenges may also be related relatively readily to knowledge regarding source water quality and to an increasing body of knowledge regarding removal and attenuation. Furthermore, the provision of microbiologically safe drinking water hinges upon placing barriers between the hazard and the human population. Regulating the microbiological safety of drinking water may therefore be undertaken both through end product standards and through requirements of these barriers - as was so successfully done during the first sanitary revolution. Such barriers relate not only to drinking-water treatment, but extend through source protection to resource management more generally. The development of WHO Guidelines for the microbiological quality of drinking-water therefore relates water quality targets, expressed in terms of ‘challenges’ by reference pathogens - whether generalised or related to specific knowledge about challenges likely to arise from a given water resource/source; and relating this to process requirements (WHO,1998b). Verification of success is then feasible using a two prong approach based upon inspection or audit of processes and procedures and end-of-product quality testing.

WHO has recently embarked upon a process of re-assessment and development of the microbiological component of the GDWQ with the intention of having a revised scheme in place in a third edition for publication in around 2003 (WHO, 1998). Whilst still in development, the outline of the scheme has been prepared and may be readily conceived in two compartments. One of risk assessment leading to specification of water quality targets and the second relating these targets to measures for their achievement and verification of success in this.

The first compartment (Figure 3) closely resembles the process of risk assessment as applied to, for example chemical contaminants of drinking water.

Available evidence suggests that the major hazards are known and much information concerning pathogens and human response to exposures to them exists. Whilst some of this remains descriptive, there is an increasingly quantitative literature. The overall process is also well characterised since it closely reflects approaches taken to risk assessment for chemicals in drinking water - a well-established and well-characterised process (WHO, 1993).

This situation may be compared with that for many chemicals. The profusion of recognised pathogens is analogous to the profusion of chemicals. And the recognition that there are inevitably as-yet-unrecognised pathogens remains analogous to the continuing synthesis of new compounds and the unknown toxicity of many synthesized compounds, of their by-products of degradation and of their by-products in water treatment, especially disinfection, processes. However in the case of microbes, health effects arise from acute exposure - a single glass of water may contain an infectious dose and lead to disease. Furthermore, a limited number of different pathogens may reasonably easily be identified as presenting the greatest challenges to provision of safe drinking water - whether through treatment or through protection of sources and it is therefore feasible to identify "critical challenges". These challenges may also be related relatively readily to knowledge regarding source water quality and to an increasing body of knowledge regarding removal and attenuation.

The provision of microbiologically safe drinking water hinges upon placing barriers between the hazard and the human population. Such barriers relate not only to drinking water treatment and management of distribution systems, but extend through source protection to resource management more generally. The second compartment for development the microbiological aspects of the WHO GDWQ therefore relates water quality targets, expressed in terms of ‘challenges’ by reference pathogens (whether generalised or related to specific knowledge about challenges likely to arise from a given water resource/source); and relating this to process requirements. Verification of success is then feasible using a two prong approach based upon inspection (audit) of processes and procedures and of end-product quality testing.

This process builds upon the approach advocated by WHO for many years, that water quality depends upon source protection, treatment determined by source water quality and should be assured by a combination of inspection/audit and microbiological verification (WHO, 1976; WHO, 1983). It also reflects the extensive and positive experience with the management and regulation of control measures rather than the management and regulation of end-product quality as described above.

Discussion of this outline process will inevitably be refined as further discussions take place and new evidence emerges.

Les citoyens des sociétés modernes et surtout occidentales, très matérialistes, expriment un besoin de savoir. Ils exigent de connaître la qualité de ce qu'ils consomment, ils expriment une demande de garantie du service et des biens qu'ils utilisent. Ils rejettent l'arbitraire et réclament des certitudes objectives. Ces requêtes sont très souvent relayées par les autorités et sont à l'origine de réglementations. Ces dernières demandent une vérification, un contrôle de l'état du service rendu ou du produit livré. Ces vérifications sont exprimées le plus souvent sous forme de résultats de mesures ou d'essais de conformité, de performance, de sécurité etc. Le non-respect des réglementations ou les litiges sur les résultats finissent en conflits entre les individus, les groupes sociaux, avec les entités juridiques ou étatiques. La protection de l'environnement, la sécurité et la protection des consommateurs sont des domaines typiques où fleurissent des lois et décrets "techniques" nationaux ou européens.

Le contrôle de la qualité de l'eau est un exemple typique de réglementation technique destinée à protéger le consommateur ou l'utilisateur. L'approche prise par le législateur pour rédiger ces textes a longtemps été rigide et peu adaptée à la complexité du problème à résoudre et surtout à l'évolution rapide des connaissances et des techniques. Les textes de loi et leur propension à vouloir préciser jusqu'au méthodes de contrôle ont souvent pour conséquence de fossiliser la technologie mise en œuvre pour les contrôles au jour de la conception du règlement. L'approche "contrôle du produit" utilisée mésestime l'importance de l'échantillonnage et du prélèvement. Une conséquence de ces réglementations rigides, pour la science et la technologie, est la difficulté de faire bénéficier le contrôle de l'eau du progrès et entre autre des méthodes modernes de mesures. Sans modifications lourdes de la Loi elle-même, aucune modification n'est possible. La Directive européenne concernant la qualité des eaux de baignades et en particulier pour le volet qui touche au contrôle de la qualité microbiologique des eaux est un exemple typique de situation bloquée.

Il existe des moyens par lesquels le scientifique peut conseiller le législateur afin de sortir des impasses créées par la loi sans diminuer, bien au contraire, la qualité du contrôle, et en favorisant le progrès technique et scientifique.

La Directive "eaux de baignades" (76/160/CEE, JO L 31 du 05-02-1976) est d'une approche du type "contrôle produit" : elle fixe les limites à ne pas dépasser en terme de contamination microbienne, liste les méthodes de mesures et les fréquences d'échantillonnage et certains critères pour le prélèvement. Elle définit des critères d'acceptabilité des eaux de baignade. Elle est basée sur de la science des années 1970, son but originel était de vérifier, de façon indirecte l'état des infrastructures d'un pays pour l'assainissement des eaux usées et leur rejet en mer ou dans les eaux intérieures. Typique de l'approche juridique du législateur, elle recherche la preuve du "délit", donc de la pollution déjà survenue, mais ignore la prévention des désordres. On applique à peu de chose prêt à l'environnement un principe juridique mis au point pour les crimes et délits.

Depuis diverses Directives ont précisé les obligations des Etats pour l'assainissement des eaux usées. La Directive "eaux de baignade" n'a donc plus ce rôle, et redeviendra vraiment utile si elle apporte une vraie information aux baigneurs.

Divers amendements ont été introduits depuis l'adoption de la Directive par ex. entérovirus, mais n'ont pas changé l'approche.

ALTERNATIVES PROPOSEES POUR AMENDER LA DIRECTIVE DE 1976

La stratégie du programme cadre de recherche (programme normes mesures et essais) pour aider les juristes à réviser la Directive "eau de baignade" au niveau de ses demandes microbiologiques a consisté à améliorer à la fois les méthodes de mesures (contrôle de qualité pour une meilleure comparabilité des résultats), redéfinir les paramètres à mesurer (diminuer l'effet méthode et améliorer la traçabilité des mesures), définir une approche établissant des performances à atteindre par une méthode et le moyen de démontrer ces performances (donner la préférence à des normes de performances et non des méthodes figées dans l'annexe), sortir de l'impasse posée par l'échantillonnage en examinant la faisabilité d'une approche du type prévention du comportement d'un système (contrôle a priori des sources de contamination et établissement de modèles de risques). Tous ces travaux étant toujours menés dans l'esprit de développer des outils adéquats pour un système préventif des risques.

Améliorer les outils et affiner les paramètres mesurés

Il n'est pas possible de progresser dans le contrôle d'un système si on ne dispose pas de méthodes de mesures fiables. La Directive mentionne des paramètres totaux (coliformes, streptocoques fécaux) dont la mesure est très dépendante de la méthode. Un premier projet (Coordonné par l'Institut Pasteur de Lille avec 34 autres laboratoires européens) a fait le tri des méthodes et a défini quels micro-organismes pouvaient être suivis [1]. Le groupe est arrivé à la conclusion que E. coli remplaçait avantageusement les coliformes totaux ou fécaux et que les streptocoques étaient à remplacer par quatre entérocoques. Les performances atteintes par certaines méthodes (NPP 3 et 5 tubes) étaient si faibles qu'elles ne permettaient aucune étude quantitative. Elles sont pourtant permises par la Directive.

Cette étude menée sur des échantillons dopés et des échantillons naturels de diverses origines a prouvé qu'aucune méthode seule ne pouvait produire dans tous les cas des résultats 100% justes et reproductibles mais que seule une combinaison de deux méthodes permettait d'éliminer tous faux positifs et de détecter tous les organismes (faux négatifs). Certaines par contre n'ont échoué que pour des échantillons de composition très particulière (contamination d'origine rare). La Figure 1 montre les méthodes de référence pour E. coli et les entérocoques. La Figure 2 compare cette référence à d'autres méthodes pour E. coli.

Développer une approche normalisée pour comparer les méthodes

Une deuxième étude également conduite par l'Institut Pasteur de Lille a utilisé cette méthode de référence pour préciser ses performances réelles (r et R) et développer un protocole pour comparer une méthode à cette méthode de référence [2]. Ces résultats sont transmis au CEN et à l'ISO. En parallèle, plusieurs projets pour aider les laboratoires à monter des systèmes d'assurance de qualité et de contrôle de qualité internes et externes (tests interlaboratoires, matériaux de références) ont été soutenus [3,4]. Un guide des bonnes pratiques a été publié. Ces projets ont concerné plus 300 laboratoires de contrôle des eaux en Europe [5].

Le cas des entérovirus

Ce paramètre est listé comme obligatoire en cas de dégradation de la qualité des eaux (?) pose des problèmes méthodologiques et financiers. Leur détermination est en effet assez peu fiable, très chère et la technologie de mesure disponible dans de rares laboratoires. Un projet de recherche a été soutenu pour tenter de remplacer les entérovirus par des bactériophages [6]. Il a conclu que ceux-ci étaient de très bons indicateurs de pollution fécale et de surcroît les méthodes étaient très fiables, simples et n'étaient pas plus chères que pour déterminer des bactéries (Figure 3).

L'échantillonnage et le prélèvement

Le prélèvement et les précautions à prendre pour le transport des échantillons a été abordé dans le Guide pour l'assurance de qualité [5]. La stratégie d'échantillonnage n'a jamais pu être abordée sérieusement dans le cadre d'une approche "contrôle produit". Suivre un système aussi vaste et complexe qu'un site de baignade (plage, rivière ou lac) pose inévitablement la question de la représentativité de l'échantillon dans l'espace et dans le temps. Ceci est particulièrement vrai pour des eaux de baignades situées dans des zones à risque (activité économique et population importante, sources de contaminations diverses et nombreuses).

Dans ces zones à risque, les sources de contamination des eaux de baignades sont représentées par les eaux usées d'origine humaine, industrielles et agricoles, aux sols contaminés, les causes sont en général liées aux conditions d'utilisations inadéquates des installations de traitement des eaux ou à des conditions climatiques défavorables (débordements d'émissaires ou de retenues, lessivage de sols industriels ou agricoles etc.). Dans de nombreuses régions ou agglomérations les inventaires des sources et des causes sont en partie réalisés pour des raisons de gestion des risques liés aux catastrophes naturelles. Les modèles de gestion doivent donc être simplement ouverts à la gestion des eaux de baignades et entrer dans des modèles mathématiques adaptés. Dans le cadre du développement et de l'affinement, de la surveillance de ces modèles il est nécessaire de réaliser des mesures microbiologiques ciblées. Il peut s'avérer nécessaire d'équiper des sites stratégiques de systèmes de mesure in-situ et automatisés. L'échantillonnage, dans cette optique, est lui-même optimisé en fonction de modèles statistiques qui quantifient les risques de contamination. On peut aussi étendre ces modèles à la protection de sites d'aqua- ou de conchyliculture.

De tels systèmes préventifs ont l'avantage de prévenir le baigneur du risque plutôt que de lui annoncer à posteriori qu'il s'est baigné dans des eaux contaminées. L'efficacité de tels systèmes réside dans la fiabilité des méthodes de mesure ou de détection, dans leur capacité à donner des résultats rapides et le faible coût de leur mise en œuvre, en particulier pour assurer une réponse rapide aux gestionnaires des eaux de baignade pour le retour à la "normale". La faisabilité d'une telle approche a été discutée dans un projet de la Commission européenne en 1997 [7] et devrait mener prochainement a un projet d'étude.

CONCLUSIONS

Les projets de recherche en microbiologie, menés par la Direction générale de la recherche dans le cadre de la révision de la Directive "eaux de baignade" ont permis de prouver que les analystes peuvent proposer de nouvelles approches permettant au progrès scientifique de venir en support à une meilleure mise en œuvre de la réglementation :

• en ciblant mieux les paramètres (E coli, entérocoques, bactériophages)
• en développant une méthode normalisée pour comparer les performances des méthodes analytiques, donc en autorisant les méthodes nouvelles (rapides) d'être utilisées ;
• en donnant aux microbiologistes les moyens de développer des systèmes d'assurance qualité dans leurs laboratoires ;
• en proposant à partir de ces outils une approche plus prédictive des risques de contamination.

Toutes ces informations ont été données aux juristes et au législateur afin de nourrir ses réflexions dans le cadre de la révision de la Directive de 1976. Si pour les modifications de paramètres et de méthodes les discussions sont assez faciles il en va autrement de l'approche préventive par modèles qui nécessitent une nouvelle façon de penser. Cette nouvelle approche pour les sites de baignade est très nouvelle et demandera encore de la part du législateur un certain temps d'adaptation. Ce changement de la réglementation a déjà été accepté pour le contrôle des chaînes de production dans l'industrie agro-alimentaire (HACCP : Hazard Analysis Critical Control Points). Il faut ajouter que de tels changements ne pourront être acceptés que si on met en place pour les mesures en microbiologie de l'eau une approche "qualité totale" comme l'a fait l'industrie agro-alimentaire. Ceci présuppose une adhésion totale des laboratoires aux systèmes de certification et d'accréditation.

Enfin l'approche alternative proposée pour les eaux de baignade est transférable à beaucoup de Directives sur l'eau qui réclament une analyse microbiologique. Si ces approches pouvaient être généralisées, elles assureraient aux industriels des méthodes et de l'instrumentation d'analyse microbiologique une entrée plus facile sur le marché des analyses réglementaires et favoriserait donc le progrès et en final réduirait les coûts tout en améliorant le service rendu au consommateur.

[1]J.F. Hernandez, J.M. Delattre and E.A. Maier. Sea water microbiology. Performance of methods for the microbiological examination of bathing water. Part 1. EUR Report 16601 EN, Luxembourg (1994)

C. Demarquilly, B. Boniface and E.A. Maier. Sea water microbiology. Performance of methods for the microbiological examination of bathing water. Part 2. Statistical analysis, EUR Report 16613 EN, Luxembourg (1994)

[2]T. Simonart, C. Demarquilly and E.A. Maier. The microbath project. EUR Report (to be published), Luxembourg (2000)

[3]N.F. Lightfoot and D.A. Ramaekers. The EQUASE project, Final report (to be published), Luxembourg (2000)

[6]J.Jofre, Bacteriophages in bathing waters. Final report (to be published), Luxembourg (2000)