An extensive identification and enumeration study of microfungi in a treated municipal water supply system in sub-tropical Australia over an 18 month period showed that microfungi are present throughout the system [1]. Analysis of the treated water leaving the single water treatment plant indicated that the treatment processes are very effective in removing or inactivating microfungal propagules since only occasional microfungal colony-forming units (CFU) were recovered from samples taken at the treatment plant outlet. However, microfungal CFU were routinely recovered from treated water storage reservoirs and distribution mains throughout the distribution system.

This finding led to a further investigation to attempt to identify the source(s) of such post-treatment microfungal population. During the collection of monthly water samples from one of the roofed storage reservoirs, designated R3, over an 18 month period it was noticed that numbers of the large Australian green tree frog, Litoria caerulea, were always resident within the reservoir, occupying the internal infrastructure of stairway, supporting columns and roofing beams as diurnal sleeping places. They particularly favoured the sloping C-section side rails of the stairway (Figure 1).

Frog excrement was also evident on these structures but it was clear that most faecal pellets are deposited directly into the water or are dislodged into the water by movement of the animals. During the study period a number of unexplained cases of low Escherichia coli counts were reported by a number of water authorities throughout Queensland. The regular observation of L. caerulea within water storage reservoir R3 in the water supply system under study led to the hypothesis that the source of those E. coli counts was the excrement of L. caerulea. Subsequent investigation demonstrated that the frog L. caerulea carries very large E. coli loads [2]. The water authority concerned has since taken steps to exclude L. caerulea from the reservoirs in its water distribution system.

Concurrently, it was noticed that old faecal pellets of L. caerulea which remained on the internal structures of reservoir R3 were showing evidence of extensive microfungal growth on their surfaces. This observation suggested three possibilities; (i) that frog excrement contained viable microfungal propagules which germinated post-voidance, (ii) that frog excrement acted as a suitable substrate for the germination and growth of airborne spores which landed on it, or (iii) both. This led to an investigation which aimed to determine whether or not the Australian green tree frog L. caerulea may be one of the sources of the microfungal propagules recovered from the mains and reservoir water during the identification and enumeration study mentioned earlier.

Samples were collected from the interior stairway and accessible structures of a roofed water service reservoir (R3) in a city in Central Queensland, Australia.

The large green tree frog L. caerulea is endemic across the northern half of Australia and parts of Papua New Guinea [8,9]. In its natural habitat this nocturnal, arboreal species hides during the day in hollow tree limbs, in dense foliage, and in similar protected places. It also has an affinity for human dwellings and other man-made structures including roofed water storage reservoirs as daytime hiding places, and seems to prefer these to its natural habitat. These structures provide a dark, warm, humid atmosphere, close proximity to water, and protection from predators, conditions which are ideal for this animal.

On every occasion that reservoir R3 was visited for water sampling purposes, mature specimens of L. caerulea were always found sleeping on the internal stairway, adjacent roof superstructure, and supporting columns within the reservoir. The suction pads on the toes of these animals enable them to easily climb vertical surfaces. At night this ambush predator comes to ground to prey on coleopterans and other insects as well as small vertebrates. The prey is engulfed, secured with its sticky tongue and vomerine teeth, and swallowed whole. As a consequence of this ground-feeding method the frog frequently ingests dried grass fragments and soil particles along with its prey, and these were commonly observed in the faecal pellets (Figure 2 a,c). These materials are natural substrates for many of the saprobic microfungi which were recovered from frog faecal pellets and also from water samples taken from the water body in the reservoirs [1]. Consequently it would be expected that such debris would introduce microfungal spores into the frog’s digestive system. Furthermore, many of the frog’s prey species are known to carry entomopathogenic fungi such as Paecilomyces spp. [10] which are also found in soil and on plants. The large number of fungal spores observed in needle-point samples of fresh frog faecal pellets also supports the theory that fungal spores are ingested with prey and extraneous debris, and pass through the animal’s digestive system unharmed.

It is clear that in areas of Australia where L. caerulea is endemic, the species is a potential source of the microfungal population of drinking water systems, and this could occur in two ways. Firstly by introduction of fungal spores from outside the reservoir in the frog’s faeces which are, in most cases, voided directly into the water body of the reservoir. The frog’s ground-feeding method ensures ingestion of extraneous material along with spore-laden prey and it is clear from the data that many of those spores survive the frog’s digestive system. This is supported by the variety of filamentous microfungal genera recovered from the frog pellets and the occasional large percentages of individual genera. For example, Table 1 shows that 97.7% of microfungal CFU recovered from frog pellet #2 was Paecilomyces, a known entomopathogenic genus. Samples #4 and #11 carried high percentages of Penicillium spp. while 45.1% of the filamentous fungi CFU recovered from sample #8 were Aspergillus spp. These genera are saprobes of plant debris and are also found in soil. The data in Table 3 also suggests that the frog’s digestive system naturally supports very high yeast populations. With the sole exception of sample #2, high numbers of yeasts were recovered from all of the pellets tested. It is notable that yeast and yeast-like colonies were frequently recovered from water samples collected from reservoir R3 during the separate identification and enumeration investigation [1]. Secondly, when excrement is deposited onto the internal superstructure above the maximum water level it provides a substrate where microfungi grow and sporulate thus adding further to the airborne spore load within the reservoir and subsequently in the water body.

The high prevalence and number of yeasts and yeast-like microfungi recovered from frog excrement and from skin swabs suggests that these microorganisms are resident commensals of the gut and skin of these animals. These high numbers, together with the knowledge that the animal carries high E. coli loads, suggests that L. caerulea should be excluded from reservoirs wherever possible and that adequate residual free chlorine levels should be maintained at all times. It is noteworthy that during the identification and enumeration study, 18 monthly water samples from reservoir R3, which was not rechlorinated during that study, produced an average of 147 yeast cells per litre of water. The water utility involved in this study has already taken appropriate action, and is constructing a rechlorination facility at reservoir R3 to manage the risk of recontamination.

The yeast genera Candida, Rhodotorula and Trichosporon which were recovered from excrement of L. caerulea collected from reservoir R3 contain species which can be opportunistic human pathogens [11]. Three of the ten representative yeast isolates from frog excrement were Candida species. It is known that metabolites of Candida are the aetiological agents of candidids, an allergic skin reaction [12]. The recovery of Candida glabrata is of some concern since it is a major cause of candidiasis in humans. This yeast, like C. albicans, is a commensal organism and is also an opportunistic pathogen causing superficial and systemic infections such as endocarditis, cystitis, osteomylitis, vaginitis and fungemia [13]. C. albicans was not found in this study but has been recovered from overseas water systems [14]. Thuraisingam and Denning [15] reported increasing frequency of systemic and vaginal infections with C. glabrata and stated that such infections are difficult to treat. Vaginitis is caused by both C. albicans and C. glabrata. However, increased prevalence of C. glabrata in diabetic women was reported by Ray et al. [16] who isolated C. glabrata from 61.3% and C. albicans from only 28.8% of 111 consecutive diabetic patients suffering from vulvovaginal candidiasis during their study into the prevalence of C. glabrata and its response to certain treatments. According to Gugic et al. [17], C. glabrata is the second most common species of Candida found in hospital patients in the U.S.A.

The other yeasts recovered from frog excrement and skin, Rhodotorula spp. and Trichosporon spp., are commonly occurring genera and while they have, on rare occasions, been associated with opportunistic human infections their involvement is of little importance [12].

It is not known if infections and allergies caused by C. glabrata or other yeasts can be acquired through contact with water containing them, but it is reasonable to expect that this may be so. In recent years it has been increasingly recognised that opportunistic, pathogenic microfungi, particularly filamentous genera such as Aspergillus, may be the waterborne aetiological agents of nosocomial mycoses but only a limited number of studies have been undertaken into this potential mode of infection.

The evidence presented in this study suggests that the Australian green tree frog L. caerulea, a diurnal resident of covered water reservoirs, is a source of yeast/yeast-like and filamentous microfungi found in water supply systems in northern Australia.

The excrement of this species introduces fungal spores directly into the water body of water storage reservoirs and, when deposited on internal reservoir architecture, it also acts as a substrate on which microfungi grow and sporulate thus increasing the spore load within the reservoir air and consequently in the water body. The results presented here suggest that the frog species L. caerulea is an important source of yeasts in water storage reservoirs in areas where the animal is endemic.

While this study was focused on a municipal water supply system where the water is treated and water quality is monitored closely, the findings would also be applicable to household water supply derived from rainwater tanks. In fact, if L. caerulea can gain access to domestic rainwater tanks to which exclusion devices are not fitted then a potential health problem could very well exist. Domestic rainwater tanks in Australia are rarely, if ever, chlorinated and the frog to water volume ratio is likely to be high because of the relatively small water-holding capacity of such tanks.