JOURNAL OF BACTERIOLOGY, Dec. 1967, p. 1811-1816 Copyright © 1967 American Society for Microbiology

Vol. 94, No. 6 Printed in U.S.A.

Localization of Indigenous Yeast in the Murine Stomach DWAYNE C. SAVAGE' AmND RENE J. DUBOS The Rockefeller University, New York, New York 10021

Received for publication 15 September 1967

Certain strains of yeast were cultured frequently from the feces of adult CFW mice and Long-Evans and Sprague-Dawley rats, but not from infants of those murine strains, or from adults or infants of NCS or NCS-D mice. When the yeasts could be cultured from the feces, they could also be grown from all areas of the digestive tracts of the animals, but especially from the stomachs, where they formed layers on the epithelium of the glandular mucosa. Three of the yeast isolates, one each from the three murine colonies, were provisionally classified in the genus Torulopsis of the asporogenous yeasts. These yeast strains failed to colonize the digestive tubes of suckling infant mice of either the CFW or NCS-D colonies. In contrast, they colonized the guts of adult NCS-D mice and formed layers in their stomachs; tests with the yeast from CFW mice revealed that this strain colonized the guts and formed layers in the stomachs of germ-free CFW mice. When established in NCS-D mice, the yeast strains did not affect qualitatively or quantitatively the growth of the animals or the composition of the bacterial flora in the gastrointestinal tracts. Moreover, they did not elicit an unusual inflammatory response in the digestive tracts; nor were they pathogenic for NCS mice when injected by the intraperitoneal or intravenous routes. The yeasts thus appear to be harmless saprophytes that are able to flourish in the environment of the surface of the secreting epithelium of the murine stomach. The findings conform with the view that some types of microorganisms of the gastrointestinal tract are not just mixed randomly but rather occupy microenvironments in almost pure culture. This concept is important to the understanding of the ecology of the gut microflora. Recent reports from this laboratory have documented the existence of intimate associations between certain strains of bacteria and particular areas of the epithelium of the gastrointestinal tract of mice. In studies with mice from several colonies, it was found that lactobacilli and anaerobic streptococci (Group N) formed layers on the stratified squamous epithelium of the esophagi and nonsecreting portions of the stomachs (3, 7). In addition, dense layers of fusiform bacteria were found in the mucus on the epithelia of the cecums and colons (7). The bacteria in these layers appeared to occupy their microenvironments to the almost complete exclusion of other bacterial types of the normal gut flora. The layers formed very early in life, when the animals were suckling infants, and remained intact as long as the nutritional and environmental conditions of the animals remained optimal. These findings have made clear that the ecology 1 Present address: Department of Microbiology, University of Texas, Austin 78712.

of the normal flora cannot be fully understood from studies of mixed microbial populations. Studies of the normal microflora should probably include examinations for and studies of microorganisms as they occur in their microenvironments (7). To support this hypothesis, we have examined another association of microorganisms with gastrointestinal epithelium. In special histological sections, layers of yeast cells have been seen on the secreting epithelium of the glandular portion of the stomachs of certain strains of mice and rats. These layers of yeasts differ in certain important respects from the bacterial layers. In particular, they do not form until the animals are weaned, and may be influenced by stress phenomena in mice. The yeast strains observed in this study do not appear to harm their hosts in any way. MATERALS AND METHODS Animals. Strains of yeast were isolated from mice of the CFW colony (Carworth Farms, New City, N.Y.),

1811

1812

SAVAGE AND DUBOS

as well as from Long-Evans (LE) and Sprague-Dawley (SD) rats from the colonies maintained at Blue Spruce Farms (Altemont, N.Y.). Studies of the establishment of yeasts in the gastrointestinal tract were done with germ-free CFW mice (Carworth Farms), and NCS (2) and NCS-D (5) mice from the special colonies maintained under protected conditions at the Rockefeller University. During the experiments, the germ-free mice were maintained in disposable plastic isolators (ISOLAB, from Bioquest, Hackensack, N.J.), and the ordinary animals were housed in metal or plastic cages with wood shavings for bedding. All animals were given tap water and commercial pellets; the water and pellets for the germ-free mice were sterilized by auto-

claving. Preparation of specimens for detection of viable yeasts or bacteria. Fecal samples were collected directly from the animals and homogenized by mechanical shaking in 5 ml of sterile charcoal water (8). For examination of the microorganisms in various areas of the digestive tube, animals were sacrificed under chloroform anesthesia. Either the entire tract from the stomach to rectum or selected segments of it were homogenized in Teflon grinders in 5 ml of sterile charcoal water. Primary isolations of the yeasts from the feces or stomachs of the mice and rats were always made within 1 or 2 hr after the animals arrived at the laboratory. Care was taken that the CFW mice and LE and SD rats had no contact with each other or with other animals before the primary isolations; they were not placed in the animal rooms at the Rockefeller University until after the first samples were taken. Techniques for culturing yeast and bacteria. Homogenates of feces or digestive tracts were diluted in charcoal water in 10-fold steps. Calibrated loopfuls of each dilution were then spread on the surface of various selective agar media. The selective media and conditions of incubation used for the recovery and enumeration of gastrointestinal bacteria have been described elsewhere (8, 9). The medium used for isolation of the yeasts was Sabouraud's Dextrose Agar (Difco) containing 40,000 units of penicillin and 0.04 g of streptomycin per liter. This medium was incubated aerobically at 37 C, unless otherwise stated. Histological techniques. These have been described elsewhere (7). Segments of the intestines or the whole stomach, always with their contents intact, were frozen in a 2% solution of methylcellulose (15 centipoise) in a 0.15 M saline solution, either on the freezing shelf of a microtome-cryostat or in parafilm bags in a dry ice-alcohol mixture. Sections of the frozen tissues were fixed in absolute methyl alcohol and were stained with either hematoxylin and eosin or with a modified gram stain for tissues. Determination of pathogenicity of the yeasts. Micro Inoculum (Difco) broth cultures (24 hr) of each yeast isolate from the CFW mice and the two strains of rat were diluted in 10-fold steps in a sterile, 0.15 M saline solution. A portion (0.5 ml) of each dilution, or of the undiluted cultures, was injected either intraper-

J. BACTERIOL.

itoneally or intravenously into each of at least 10 five- to six-week-old male NCS mice. Determination of the taxonomy of the yeast strains. The method of Lodder and Kreger-Van Rij (4) was used to determine the genera of the yeast isolates. The mode of sexual reproduction was sought by examining for sexual spores in colonies of the yeast strains on Gorodkowa agar, acetate agar, and yeast autolysateglucose agar (4, 6). The morphology of the cells and their mode of vegetative reproduction were examined in cultures grown in malt extract broth and on malt extract agar (4). Ability of the strains to produce starch-like compounds was tested on Sabouraud's Dextrose Agar and on malt extract agar. The microorganisms were grown on the surface of the media, and iodine solution was poured over the colonies. A blue color in the medium surrounding the colonies was taken to indicate starch production (4). RESULTS

Isolation of yeastsfrom thefeces and gastrointestinal tracts of mice and rats. Isolations of yeast were made frequently from the feces of adult CFW mice and LE and SD rats (Table 1). In mice, the microorganisms were recovered in large numbers only from old animals of either sex or gravid or lactating females, but they were consistently recovered from young and old adult rats of either sex. They were never cultured from infants from the colonies of CFW mice, LE rats or SD rats, or from infants or adults from the colonies of NCS or NCS-D mice. When yeasts were present in the feces of an animal, they could also be isolated from all areas of its gastrointestinal tract but especially from its stomach (Table 2). In histological preparations, layers of yeasts were seen on the epithelium of the mucosa of the glandular portions of the stomachs of both the CFW mice and the two strains of rat. The yeast layers are shown for the mouse in Fig. 1 and 2, and for the rat in Fig. 3 and 4. The layering did not occur on any other area of the digestive tube. No inflammatory reaction to the yeasts was observed in histological preparations of the stomachs of either type of rodent. Whenever the yeasts were isolated from the feces and stomachs of the rodents, the layers were observed in the stomachs (Table 1). Since only one type of yeast could be seen in the histological sections of the stomachs, and these were predominantly restricted to the secreting epithelium, it is believed that the strains of yeast in the layers were the same as those that were recovered on culture media. Moreover, all of the isolates from any one source, for example, the CFW mice, had the same cellular and colonial morphology and method of vegetative reproduction on the isolation media

VOL. 94, 1967

1813

YEAST IN THE MURINE STOMACH

TABLE 1. Recovery of yeasts from the gastrointestinal tracts of conventional mice and rats and SPF mice Animal | Animal

CFW mice

NCS or NCS-D mice LE rats

SD rats

Age

~~~~~~~No.of Lanimals

2-21 days 3-10 weeks 8-12 monthsb 3-5 monthsc 1 day15 monthsb 2-21 days 4-5 weeks 10-12 monthsb 4-6 monthse 1 day-3 weeks 5-6 weeks 12 monthsb

No. with yeast No. with positive | No. with positive layers in stomach/ fecal cultures stomach cultu5re| crs no. tested

30 40 48 40 100

0 10 7 30 0

0 10 7 30 0

0/30 10/40 7/48 15/15" 0/100

30 20 6 5 31 30 10

1

19 6 5 0 30 10

0 19 6 5

0/30 10/10d 6/6 5/5 0/21 10/10"

0

30 10

10/10"

Counts ranged from 104 to 106 viable yeasts per g of fresh feces and from 10' to 107 viable yeasts per g of whole stomach. b Including postpartum females. e Gravid or lactating. d Only animals yielding yeast in feces were tested. a

used. Therefore, it is believed that for each rodent colony, the numerous isolations over the 2 years of the study probably involved the same yeast strain. One of the yeast isolates was selected from each of the three murine strains for further study and identification. Taxonomy of the yeast strains. Attempts were made to determine the genera of the three yeast strains, but not their specific epithets. The findings were the same for the three strains. No sexual spores were detected, the colonies on agar media were colorless, the cells were round to oval, vegetative reproduction was by multilateral budding, and no starch-like compounds were produced. The search for sexual spores was extensive and of long duration. It is still possible, however, that more examinations of the yeast strains may reveal a mode of sexual reproduction. However, on the basis of the results available thus far, we have provisionally classified all three of the yeast strains in the genus Torulopsis of the asporogenous yeasts. Colonization of the digestive tracts of NCS-D and germ-free CFW mice by the strains of yeast. Attempts to establish the yeast strains in the guts of infant mice by contaminating the animals per os were always unsuccessful, but the gastrointestinal tracts of adult NCS-D mice were colonized readily by either of the strains isolated from mice or LE rats. The digestive tracts of adult germ-free CFW mice were also readily colonized by the yeast strain isolated from CFW mice. No attempt

TABLE 2. Isolation of yeasts from various areas of the alimentary tracts of mice and rats Logso no. of yeast isolated6/ Animal

g

No. of

of fresh tissue

animals Small Ceu.Large FcsStomFeces Cecum intestine ach intestine

CFW mice LE rats SD rats a

10 10 10

6 6 6

7 7 7

6 6 6

6 6

6 6

6

6

Approximate averages.

was made to associate germ-free mice with the yeast strains isolated from rats (Table 3). The yeasts could be recovered from all parts of the digestive tubes, but formed layers only on the glandular mucosa of the stomachs. After contamination, it took 48 to 72 hr before the stomach layers were fully formed (Table 4). There is as yet no evidence that the yeast strains exerted harmful effects on the colonized mice. For example, they did not cause inflammation in the stomachs. Moreover, they did not affect the expected gain in body weight over a period of many weeks, nor influence in any detectable way the indigenous bacterial flora of the digestive tracts of the colonized NCS-D mice (Table 5). Large doses of cultures of the three yeast strains were injected intraperitoneally or intravenously into NCS mice; no deaths occurred over a period of several months, and no lesions could be observed upon autopsy, even in those animals which received as many as 107 viable yeast cells.

1814

SAVAGE AND DUBOS

4 ;.

.-

J. BACTERIOL.

...

.*.

..

.U

9.%

FIG. 1-4. Gram-stained tissue sections showing yeast layers on the epithelium of the glandular mucosa of the stomachs of mice and rats. FIG. 1. CFW mouse; X 500. FiG. 2. Same as Fig. 1; X 1,000. FiG. 3. LE rat; X 500. FIG. 4. Same as Fig. 3; X 1,000.

1815

YEAST IN THE MURINE STOMACH

VOL. 94, 1967

TABLE 3. Colonization of the gastrointestinal tracts of mice by yeasts isolated from CFW mice or LE rats Animal colonized Animal | Age at conclonizedtaminationa

CFW mice (ordinary)

2-21 days

CFW mice (germ-free) NCS-D mice

5-6 weeks 2-21 days 5-6 weeks

es Yeas

Md R M M None M R

No. of mice

20 20 6 20 35 50 15

No. of positive fecal culturesb No. of posiatomh tiv tiesoach culturesc ter 1-6 Afer weksAfmonths 2 weks Afte

0 0 6 0 0 50 15

No. with

stomach layers/no. tse

0/10

0 0 6 0 0 50 15

6 0 50 15

0/10

6/6 0/20 0/10 35/35 10/10

a Adults were contaminated by adding small amounts of an undiluted broth culture of yeast to their drinking water and food for 24 hr; infants were contaminated by placing one drop of undiluted broth culture into their mouths and pouring small amounts of culture into their nests. b Culture results as described in Table 1. c At termination, 1 to 6 months after contamination. d M, yeast strain isolated from CFW mice; R, yeast strain isolated from LE rats.

TABLE 4. Course of establishment of yeast isolated from CFW mice in guts of NCS-D mice After contamination

(hr)

Viable counta

TABLE 5. Fecalflora of NCS-D mice colonized more than 6 weeks with yeasts isolated from CFW mice or LE rats

Stomach layerb

No. of microorganismsa Origin of yeast

0 12 18 36 72 130

None