An electron microscopical - Wiley Online Library

Journal of Applied Microbiology 2001, 90, 294±300

Epithelium-associated bacteria in the gastrointestinal tract of Arctic charr (Salvelinus alpinus L.). An electron microscopical study E. Ringù1, J.B. Lùdemel1,2, R. Myklebust3, T. Kaino1, T.M. Mayhew4 and R.E. Olsen5 1

Department of Arctic Veterinary Medicine, The Norwegian School of Veterinary Science, Tromsù, Norway, Department of Marine Biochemistry, Norwegian College of Fishery Science, University of Tromsù, Norway, 3 Department of Morphology, Faculty of Medicine, University of Tromsù, Norway, 4School of Biomedical Sciences, Queen's Medical Centre, University of Nottingham, UK, and 5Institute of Marine Research, Matre Aquaculture Research Station, Matredal, Norway 2

428/6/00: received 28 June 2000, revised 25 October 2000 and accepted 31 October 2000

E . R I N G é , J . B . L é D E M E L , R . M Y K L E B U S T , T . K A I N O , T . M . M A Y H E W A N D R . E . O L S E N . 2001.

Aims: The primary aim was to use transmission electron microscopy (TEM) and scanning electron microscopy (SEM) to de®ne the location of epithelium-associated bacteria in the digestive tract of the salmonid ®sh, Arctic charr (Salvelinus alpinus). Methods and Results: TEM and SEM examination of the gastrointestinal tract demonstrated substantial numbers of ovoid and rod-shaped bacterial cells associated with the microvillous brush borders of enterocytes. Bacteria were found at the tips of microvilli as well as between adjacent microvilli. Endocytosis of bacteria by epithelial cells was observed in two regions (pyloric caeca and midgut). Conclusions: Electron microscope examination of the gut is an important tool for evaluating the microbial ecology of the ®sh digestive tract ecosystem. Signi®cance and Impact of the Study: The results of the current study clearly demonstrate that the intestine is involved in bacterial endocytosis.

INTRODUCTION The epithelial cells of the digestive tract have three main functions: (i) digestion of food and absorption of nutrients; (ii) continuous cell renewal, and cell death and extrusion mechanisms, which facilitate the elimination of damaged cells; and (iii) protection against environmental pathogens. Light and electron microscope examinations of gut samples are important tools for investigating the microbial ecology of the vertebrate digestive tract ecosystem. Studies on chickens, pigs, rodents and humans have demonstrated the presence of epithelium-associated microbial populations (Knutton et al. 1987; Tannock 1987; Tannock et al. 1987). Although several recent reviews have described the composition of the intestinal microbiota of ®sh (Cahill 1990; Ringù et al. 1995; Hansen and Olafsen 1999; Ringù and Birkbeck Correspondence to: Dr E. Ringù, Department of Arctic Veterinary Medicine, The Norwegian School of Veterinary Science, NO-9292 Tromsù, Norway (e-mail: [email protected]).

1999), relatively few investigations have attempted to identify the location of micro-organisms in the gastrointestinal tract using transmission electron microscopy (TEM) and/or scanning electron microscopy (SEM) (LeÂsel and Pointel 1979; Austin and Al-Zahrani 1988; Hansen and Olafsen 1999; Ringù and Olsen 1999). In view of the paucity of information on whether bacterial cells are located at the tips of microvilli or between microvilli, the primary aim of the present study was to use TEM and SEM to de®ne the location of micro-organisms in three different regions (pyloric caeca, midgut and hindgut) of the gastrointestinal tract in the salmonid ®sh, the Arctic charr (Salvelinus alpinus). Endocytosis of bacterial cells has been reported in the gastrointestinal tract of ®sh larvae (Hansen and Olafsen 1990; Hansen et al. 1992; Olafsen and Hansen 1992; Grisez et al. 1996). However, when discussing endocytosis, the development of the digestive tract is an important factor to be considered. At the time of hatching, the digestive tract of ®sh is an undifferentiated straight tube which is morphoã 2001 The Society for Applied Microbiology

BACTERIA IN THE GASTROINTESTINAL TRACT OF ARCTIC CHARR

logically and physiologically less elaborate than that of adults (Govoni et al. 1986). Therefore, the gradual development of the digestive tract from larval stages to adult ®sh must be taken into account. Consequently, the secondary aim of the present investigation was to evaluate whether endocytosis of bacteria occurs in different regions (pyloric caeca, midgut and hindgut) of the gastrointestinal tract of adult salmonid ®sh using TEM. This is highly relevant as the digestive tract is a potential port of entry for pathogens (Chair et al. 1994; Olsson 1995; Grisez et al. 1996; Olsson et al. 1996). MATERIALS AND METHODS Fish and rearing conditions Arctic charr of the anadromous Hammerfest strain were reared in fresh water at KaÊrvika Research Station, northern Norway, at ambient light and temperature conditions until an average weight of 141 g was attained. In February 1999, ®sh were divided into three feeding groups each containing 98 ®sh in triplicate tanks (300 l PVC self-cleaning) and supplied with aerated fresh water at a constant temperature of 8 °C. Food was supplied in excess (2%) using 2 h (between 0900 and 1100 h) disc feeders. The duration of feeding trials was 4 weeks. The experimental protocols and procedures were approved by the Animal Use and Care Committee at KaÊrvika Research Station. Diet For present purposes, an experimental diet (4 mm) was prepared by Nor Aqua Innovation Ltd, Dirdal, Norway, based on their commercial recipe for salmonid ®sh. The pellets contained 4% marine fat and were coated with 18% soybean oil. Protein, carbohydrate, lipid and fatty acid composition of the diet was analysed according to Ringù et al. (2001). A detailed description of the chemical composition of the experimental diet is presented in Table 1. Electron microscope studies of Arctic charr intestines After random sampling (lottery method), three ®sh were killed by a sharp blow to the head. The intestine was immediately excised and strati®ed random samples from three areas of the intestineÐpyloric caeca, midgut (middle portion) and hindgut (middle portion)Ðwere taken. Tissue slices were immediately ®xed in McDowell's ®xative (McDowell and Trump 1976), and prepared for TEM and SEM as described elsewhere (Olsen et al. 2000). Intestinal specimens were examined using a Jeol 1010 1,2 transmission microscope and a Jeol 5300 scanning electron 1,2 microscope (Tokyo, Japan).

295

Table 1 Chemical composition and fatty acid composition of the experimental diet Chemical composition

(% of dry weight)

Carbohydrate Protein Lipid

11á3 + 0á4 48á2 + 1á2 20á8 + 0á2

Fatty acid composition 14 : 0 16 : 0 16 : 1 n-7 18 : 0 18 : 1 (n-7 + n-9) 18 : 2 n-6 18 : 3 n-3 20 : 1 n-9 20 : 5 n-3 22 : 1 n-11 22 : 6 n-3

(% of total) 1á4 12á6 1á1 3á4 21á5 40á6 7á8 1á9 1á9 1á9 3á2

RESULTS AND DISCUSSION In general, the population level of bacteria in ®sh intestines is substantially lower, and the microbiota simpler, than those reported for warm-blooded animals, including humans. While the digestive tract of ®sh is colonized mainly by aerobes and facultative anaerobes (Cahill 1990; Ringù et al. 1995), the predominant bacterial genera/species isolated from the gastrointestinal tract of endotherms are obligate anaerobes (Finegold et al. 1983). The relatively low population level of bacteria recovered in ®sh intestine (Cahill 1990; Ringù et al. 1995) may be due to: (i) destruction of ingested bacteria as a result of acid in the stomach and bile in the intestine; (ii) the media used being unable to support the growth of the major bacteria present; (iii) antibacterial activity of intestinal secretions, including mucus; (iv) temperature effects; and (v) the relatively high gastrointestinal evacuation rate. Given the low bacterial populations in the gastrointestinal tract of ®sh, electron microscope examination of gut samples by TEM and SEM may provide important information regarding the presence and localization of bacteria in the digestive tract ecosystem. Transmission electron microscopy of the intestine In a recent study, Hansen and Olafsen (1999) demonstrated bacteria associated with the microvillous layer in the posterior part of the hindgut in 14 day-old herring (Clupea harengus) larvae. In the present study, using TEM, bacterial cells were clearly demonstrated between microvilli in all three regions investigated (Figs 1, 2, 3).

ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 294±300

296

E. RINGé ET AL.

Fig. 1 Transmission electron micrograph of the apical regions of enterocytes in the pyloric caecum. Bacterial pro®les are seen scattered at different levels within the brush border from the tips to the bases of microvilli. In addition, one bacterial pro®le (arrowhead) is seen to be contained in an internalized, membranebound endocytic vacuole. ´15 000

Autochthonous microbiota in ®sh The microbiology of the intestinal tract of marine and freshwater ®sh has been widely investigated (for reviews see Cahill 1990; Ringù et al. 1995; Hansen and Olafsen 1999; Ringù and Birkbeck 1999), and the gut microbiota can be de®ned as either autochthonous (indigenous) or allochtonous (transient). To succeed in colonizing the gastrointestinal tract, adherent micro-organisms must be able to survive and multiply in the gastrointestinal tract (Cahill 1990; Ringù et al. 1995). There is evidence that the microbial populations occurring within the hindgut are greater than those of the surrounding water, but there is still uncertainty as to whether the microbiota of ®sh is autochthonous or allochtonous. To de®ne the presence of an autochthonous

microbiota in ®sh, electron microscope examination might be a useful tool. This method has been used successfully in studies of the microbial ecology of the digestive tract of pigs, rodents, chickens and humans (Knutton et al. 1987; Tannock 1987; Tannock et al. 1987). The present investigation con®rmed the presence of microbiota in these adult Arctic charr, and this was considered as evidence for autochthonous bacteria closely associated with microvilli in the pyloric caeca (Fig. 1), midgut (Fig. 2) and hindgut regions (Fig. 3). Scanning electron microscopy of the intestine Bacterial adhesion is a cell-surface interaction phenomenon which makes it ideal for examination by SEM (Knutton

Fig. 2 Higher-power transmission electron micrograph of the midgut. The opposed surfaces of two enterocytes are shown. Both cells have appreciable numbers of bacterial pro®les between their microvilli. Note the internalized bacterium in the subapical cytoplasm (arrowhead). L: intestinal lumen. ´15 000 ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 294±300


297

Fig. 3 Transmission electron micrographs showing bacteria associated with the microvilli of enterocytes in the hindgut. Enterocytes in this region show endocytic activity and are characterized by large numbers of intracytoplasmic vacuoles (V) with contents of varying electron density. An internalized bacterium is discernible (arrowhead). ´15 000

1995). Compared with TEM, SEM has the advantage that large areas of cell surface can be rapidly examined for adherent micro-organisms. In the present study, substantial associations of coccoid and rod-shaped bacterial cells with the tips, and between, enterocyte microvilli in pyloric caeca, midgut (Fig. 4) and hindgut (Fig. 5) were demonstrated by SEM. Sometimes, these bacteria had their luminal ends protruding above the level of the microvilli. Micrographs displayed clear differences, even within the same areas, as some enterocytes were heavily colonized while others did not have any associated bacteria. The reasons for these differences are not clear but may be due to enterocyte ageing, differentiation, or possessing speci®c receptors for receptor-mediated endocytosis of bacteria.

In contrast to the results of the present study (which clearly demonstrated that some enterocytes were heavily colonized by bacteria), a different situation was observed when ®sh were fed dietary linseed oil. In the latter situation, most bacteria associated with enterocytes were located at the apical brush border (Ringù et al. 2001). These differences in bacterial colonization of the enterocytes between ®sh fed different dietary oils will be the subject of further investigations. They might be related to the fact that different gut microbiota were observed in ®sh fed soybean oil or linseed oil (Ringù et al. 2001). The adhesion processes and mechanisms involved for different adherent bacterial species have been described in ®sh and seem to include adhesive factor(s) (adhesin,


298

E. RINGé ET AL.

Fig. 4 Scanning electron micrograph of the apical aspects of enterocytes in the midgut. The borders between adjacent cells are clearly visible, as are the microvilli (MV) which cover the cell apex. The luminal ends of bacteria located in the interstices between microvilli are also visible (arrows). ´7500

Krovacek et al. 1987), salinity and pH (Balebona et al. 1995), and cell surface hydrophobicity (Parker and Munn 1984; Bruno 1988). In addition, receptor-speci®c interactions such as pili-like structures, speci®c receptors on enterocytes such as sugar residues as documented from 3 mammalian models, may be involved (Knutton 1995). Furthermore, it is known that dietary lipids in¯uence intestinal membrane composition, function and ¯uidity in ®sh (Pelletier and Leray 1987; Behar et al. 1989). In a recent study, Ringù et al. (2001) demonstrated that the hindgut microbiota was affected by dietary lipids (soybean, linseed and marine oils). If brush border membranes are in¯uenced by dietary lipid, the fatty acid composition and, possibly, the

adherence mechanisms of the mucosa, could change with diet, favouring the establishment of certain bacterial species. Further studies regarding the interactions between dietary lipid and the adherence mechanisms of the intestinal microbiota should be given high priority. To date, methods for studying the gut microbiota of ®sh, whether attached or luminal populations, have been undertaken by homogenizing sections of gut and plating the homogenate onto a range of selective media. However, such methods detect only the micro-organisms capable of growing on speci®c, selective media. A proportion of the bacterial population will remain undetected. This underestimation of microbial numbers has been observed in natural environ-

Fig. 5 Scanning electron micrograph showing cell apices in the hindgut. Cell borders can be seen and all cells have associated bacteria (arrows), although numbers vary from cell to cell. Note the small spaces (arrowheads) between microvilli. These may represent the transit paths of more deeply embedded bacteria, or they may be created by bacterial loss. The latter may be an artefact of tissue preparation or a consequence of local bacterial cell division. ´5000 ã 2001 The Society for Applied Microbiology, Journal of Applied Microbiology, 90, 294±300


ments, where growth on agar plates only accounted for 2±4% of the total microbial population counted using microscopy (Olsen and Bakken 1987). In a recent study, Ringù et al. (2000) detected 3á3 ´ 105 bacteria associated with the hindgut region. If it is assumed that only 25% (1 ´ 105) of hindgut enterocytes are colonized, and that each bears approximately 100 bacteria, this corresponds to roughly 1 ´ 107 bacteria, and implies that only 3% of the total bacterial population associated with the hindgut region are culturable. This calculation suggests an underestimation of the gut microbiota in ®sh. Endocytosis of bacteria As the intestine seems to be involved in infection (Chair et al. 1994; Olsson 1995; Grisez et al. 1996; Olsson et al. 1996), a fundamental question is whether or not there are differences between the anterior and posterior part of the intestine with respect to bacterial endocytosis. The results of the present study clearly demonstrated endocytosis of bacterial cells in the pyloric caeca (Fig. 1) and midgut (Fig. 2), indicating that these two regions are also involved in bacterial endocytosis in adult ®sh. Endocytosis of bacteria by enterocytes has been shown to occur mainly in the epithelial border of the foregut region of larvae from herring (Hansen et al. 1992; Olafsen and Hansen 1992), herring and cod (Gadus morhua) (Olafsen and Hansen 1992), cod (Hansen and Olafsen 1990) and turbot (Scophthalmus maximus) (Grisez et al. 1996). On the other hand, Vigneulle and Laurencin (1991) and Tamura et al. (1993) showed that phagocytosis of formalin-killed Vibrio anguillarum occurred primarily in the posterior intestine of rainbow trout (Oncorhynchus mykiss), sea bass (Dicentrarchus labrax), turbot and eel (Anguilla anguilla). In vitro adhesion of Vibrio anguillarum to rainbow trout gut sections was reported by Horne and Baxendale (1983), who noted that adhesion was more predominant in the upper and midgut regions than in the oesophagus, stomach and hindgut. In vivo experimental challenge did not reveal any differences in adhesion to the various gut segments but adhesion of the pathogen to the midgut region of vaccinated ®sh was signi®cantly lower than in non-vaccinated ®sh. However, as the results of the present study clearly indicate that the intestine is involved in bacterial endocytosis, it is suggested that there are no longitudinal differences between regions of the intestine with respect to bacterial infection. This controversial hypothesis calls for further investigations which are currently underway. ACKNOWLEDGEMENTS Financial support from Norwegian Research Council (grant no. 122851/122) is gratefully acknowledged.

299

REFERENCES Austin, B. and Al-Zahrani, A.M.J. (1988) The effect of antimicrobial compounds on the gastrointestinal micro¯ora of rainbow trout, Salmo gairdneri Richardson. Journal of Fish Biology 33, 1±14. Balebona, M.C., Morinigo, M.A., Faris, A. et al. (1995) In¯uence of salinity and pH on adhesion of pathogenic Vibrio strains to Sparus aurata skin mucus. Aquaculture 132, 113±120. Behar, D., Cogan, U., Viola, S. and Mokady, S. (1989) Dietary ®sh oil augments the function and ¯uidity of the intestinal brushborder membrane of carp. Lipids 24, 737±742. Bruno, D.W. (1988) The relationship between auto-agglutination, cell surface hydrophobicity and virulence of the ®sh pathogen Renibacterium salmoninarum. FEMS Microbiology Letter 51, 135±140. Cahill, M.M. (1990) Bacterial ¯ora of ®shes: a review. Microbial Ecology 19, 21±41. Chair, M., Dehasque, M., van Poucke, S., Nelis, H., de Sorgeloos, P. and Leenher, A.P. (1994) An oral challenge for turbot larvae with Vibrio anguillarum. Annual Review of Fish Diseases 10, 5±36. Finegold, S.M., Sutter, V.L. and Mathisen, G.E. (1983) Normal indigenous intestinal ¯ora. In Human Intestinal Micro¯ora in Health and Disease ed. Hentgens, D.J. pp. 3±31. London: Academic Press. Govoni, J.J., Boehlert, G.W. and Watanabe, Y. (1986) The physiology of digestion in ®sh larvae. Environmental Biology of Fishes 16, 59±77. Grisez, L., Chair, M., Sorgeloos, P. and Ollevier, F. (1996) Mode of infection and spread of Vibrio anguillarum in turbot Scophthalmus maximus larvae after oral challenge through live feed. Diseases of Aquatic Organisms 26, 181±187. Hansen, G.H. and Olafsen, J.A. (1990) Endocytosis of bacteria in yolksac larvae of cod (Gadus morhua L.). Microbiology in Poecilotherms ed. LeÂsel, R. pp. 187±191. Amsterdam: Elsevier. Hansen, G.H. and Olafsen, J.A. (1999) Bacterial interactions in early life stages of marine cold water ®sh. Microbial Ecology 38, 1±26. Hansen, G.H., Strùm, E. and Olafsen, J.A. (1992) Effect of different holding regimes on the intestinal micro¯ora of herring (Clupea harengus) larvae. Applied and Environmental Microbiology 58, 461±470. Horne, M.T. and Baxendale, A. (1983) The adhesion of Vibrio anguillarum to host tissues and its role in pathogenesis. Journal of Fish Diseases 6, 461±471. Knutton, S. (1995) Electron-microscopic methods in adhesion. Methods in Enzymology 253, 145±158. Knutton, S., Lloyd, D.R. and McNeish, A.S. (1987) Adhesion of enteropathogenic Escherichia coli to human intestinal enterocytes and cultured human intestinal mucosa. Infection and Immunity 55, 69±77. Krovacek, K., Faris, A., Ahne, W. and MaÊnsson, I. (1987) Adhesion of Aeromonas hydrophila and Vibrio anguillarum to ®sh cells and to mucus-coated glass slides. FEMS Microbiology Letters 42, 85±89. LeÂsel, R. and Pointel, J.G. (1979) Settlement of bacterial ¯ora in the digestive tract of rainbow trout. Scanning electron microscope study. Annales de Zoologie, Ecologie Animale 11, 327±335. McDowell, E.M. and Trump, B.R. (1976) Histological ®xatives suitable for diagnostic light and electron microscopy. Archives of Pathology and Laboratory Medicine 100, 405±414. Olafsen, J.A. and Hansen, G.H. (1992) Intact antigen uptake in intestinal epithelial cells of marine ®sh larvae. Journal of Fish Biology 40, 141±156.


300

E. RINGé ET AL.

Olsen, R.A. and Bakken, L.R. (1987) Viability of soil bacteria: Optimisation of plate counting technique and comparison between total counts and plate counts within different size groups. Microbial Ecology 13, 59±74. Olsen, R.E., Myklebust, R., Ringù, E. and Mayhew, T.M. (2000) The in¯uences of dietary linseed oil and saturated fatty acids on caecal enterocytes in Arctic charr (Salvelinus alpinus L.): a quantitative ultrastructural study. Fish Physiology and Biochemistry 22, 207±216. Olsson, C. (1995) Bacteria with inhibitory activity and Vibrio 4 anguillarum in ®sh intestinal tract. PhD Dr. Thesis, GoÈteborg University, Sweden. Olsson, J.C., JoÈborn, A., Westerdahl, A., Blomberg, L., Kjelleberg, S. and Conway, P.L. (1996) Is the turbot, Scophthalmus maximus L., intestine a port of entry for the ®sh pathogen Vibrio anguillarum. Journal of Fish Diseases 19, 225±234. Parker, N.D. and Munn, C.B. (1984) Increased cell surface hydrophobicity associated with possession of additional surface protein by Aeromonas salmonicida. FEMS Microbiology Letters 21, 233±237. Pelletier, X. and Leray, C. (1987) Effects of eicosatetraynoic acid on membrane lipids of trout liver and intestine. Lipids 22, 1053±1056. Ringù, E. and Birkbeck, T.H. (1999) Intestinal micro¯ora of ®sh larvae and fry. Aquaculture Research 30, 73±93. Ringù, E., Lùdemel, J.B., Myklebust, R., Jensen, L., Lund, V., Mayhew, T.M. and Olsen, R.E. (2001) The effect of soybean oil,

linseed oil and marine oil on aerobic gut microbiota of Arctic charr (Salvelinus alpinus L.) prior and post challenge with Aeromonas 5 salmonicida subsp. salmonicida. submitted. Ringù, E. and Olsen, R.E. (1999) The effect of diet on aerobic bacterial ¯ora associated with intestine of Arctic charr (Salvelinus alpinus L.). Journal of Applied Microbiology 86, 22±28. Ringù, E., Strùm, E. and Tabachek, J.-A. (1995) Intestinal micro¯ora of salmonids: a review. Aquaculture Research 26, 773±789. Tamura, S., Shimizu, T. and Ikegami, S. (1993) Endocytosis in adult eel intestine: immunological detection of phagocytic cells in the surface epithelium. Biological Bulletin 184, 330±337. Tannock, G.W. (1987) Demonstration of mucosa-associated microbial populations in the colons of mice. Applied and Environmental Microbiology 53, 1965±1968. Tannock, G., Blumershine, R. and Archibald, R. (1987) Demonstration of epithelium-associated microbes in the oesophagus of pigs, cattle, rats and deer. FEMS Microbiology Ecology 45, 199±203. Vigneulle, M. and Laurencin, F.B. (1991) Uptake of Vibrio anguillarum bacterin in the posterior intestine of rainbow trout (Oncorhynchus mykiss), sea bass (Dicentrarchus labrax) and turbot (Scophthalmus maximus) after oral administration or anal intubation. Diseases of Aquatic Organisms 11, 85±92.
