REVIEWS
24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40
Li, M. et al. (1997) J. Virol. 71, 1984–1991 Arvanitakis, L. et al. (1997) Nature 385, 347–350 Bais, C. et al. (1998) Nature 391, 86–89 Henderson, S. et al. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 8479–8483 Sarid, R. et al. (1997) Nat. Med. 3, 293–298 Sunil-Chandra, N.P., Efstathiou, S., Arno, J. and Nash, A.A. (1992) J. Gen. Virol. 73, 2347–2356 Cardin, R.D., Brooks, J.W., Sarawar, S.R. and Doherty, P.C. (1996) J. Exp. Med. 184, 863–871 Sunil-Chandra, N.P., Efstathiou, S. and Nash, A.A. (1992) J. Gen. Virol. 73, 3275–3279 Sunil-Chandra, N.P., Efstathiou, S. and Nash, A.A. (1993) Virology 193, 825–833 Usherwood, E.J. et al. (1996) J. Gen. Virol. 77, 2819–2825 Usherwood, E.J., Stewart, J.P. and Nash, A.A. (1996) J. Virol. 70, 6516–6518 Weck, K.E. et al. (1996) J. Virol. 70, 6775–6780 Nash, A.A. and Sunil Chandra, N.P. (1994) Curr. Opin. Immunol. 6, 560–563 Sunil-Chandra, N.P., Arno, J., Fazakerley, J. and Nash, A.A. (1994) Am. J. Pathol. 145, 818–826 Penn, I. (1986) Surg. Gynecol. Obstet. 162, 603–610 Doherty, P.C. et al. (1997) Curr. Opin. Immunol. 9, 477–483 Rickinson, A.B. and Moss, D.J. (1997) Annu. Rev. Immunol. 15,
405–431 41 Ehtisham, S., Sunil-Chandra, N.P. and Nash, A.A. (1993) J. Virol. 67, 5247–5252 42 Stevenson, P.G. and Doherty, P.C. (1998) J. Virol. 72, 943–949 43 Usherwood, E.J. et al. (1997) J. Gen. Virol. 78, 2025–2030 44 Sarawar, S.R. et al. (1996) J. Virol. 70, 3264–3268 45 Sarawar, S.R. et al. (1997) J. Virol. 71, 3916–3921 46 Dutia, B.M., Clarke, C.J., Allen, D.J. and Nash, A.A. (1997) J. Virol. 71, 4278–4283 47 Weck, K.E. et al. (1997) Nat. Med. 3, 1346–1353 48 Tripp, R.A. et al. (1997) J. Exp. Med. 185, 1641–1650 49 Usherwood, E.J., Ross, A.J., Allen, D.J. and Nash, A.A. (1996) J. Gen. Virol. 77, 627–630 50 Sunil-Chandra, N.P., Efstathiou, S. and Nash, A.A. (1994) Antiviral Chem. Chemother. 5, 290–296 51 Neyts, J. and De-Clercq, E. (1998) Antimicrob. Agents Chemother. 42, 170–172 52 Ensser, A., Pflanz, R. and Fleckenstein, B. (1997) J. Virol. 71, 6517–6525 53 Albrecht, J.C. et al. (1992) J. Virol. 66, 5047–5058 54 Russo, J.J. et al. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 14862–14867 55 Telford, E.A. et al. (1995) J. Mol. Biol. 249, 520–528 56 Baer, R. et al. (1984) Nature 310, 207–211 57 Nicholas, J. et al. (1997) J. Virol. 71, 1963–1974
Fimbrial expression in enteric bacteria: a critical step in intestinal pathogenesis Robert A. Edwards and José Luis Puente
A
prerequisite for the de- The ability of species of enteric bacteria to around the bacterial cell, with velopment of any bac- recognize and colonize unique niches along hundreds of fimbriae per bacterial disease is localizathe intestine is mainly based on receptor terium; however, in some tion of the bacteria to a niche distribution and interpretation of a systems, the fimbriae are exthat is suitable for growth and combination of environmental signals pressed in a polar orientation pathogenesis. In a mammalian leading to the expression of specific (Fig. 1). intestine, attachment is critical adherence factors. Such elaborate Two of the best-studied to avoid displacement from a orchestration of events is critical during enteric pathogens, Escherichia preferred site by the continuthe initial steps of pathogenesis. coli and Salmonella enterica, ous flow of the intestinal conare able to infect a wide range R.A. Edwards* is in the Dept of Microbiology, tents. Attachment is also hinof hosts, but particular strains University of Illinois Urbana–Champaign, dered by competition with the of either species can often only 601 S. Goodwin Ave, Urbana, IL 61801, USA; multitude of indigenous microcause infection in a small variJ.L. Puente is in the Molecular Microbiology Dept, flora for binding sites on the ety of potential hosts or coloInstituto de Biotecnología, Universidad Nacional intestinal epithelium1. Autónoma de México, Apartado Postal 510-3, Colonia nize a particular segment of Miraval, Cuernavaca, Morelos, 62250, Mexico. The initial step in bacterial the intestine. This host and *tel: ⫹1 217 333 2203, attachment to the host epithetissue-range specificity is often fax: ⫹1 217 244 6697, lium is usually mediated by mediated by different fimbriae e-mail:
[email protected] fimbriae (Box 1). Fimbriae are and receptors2,3. Fimbriae are proteinaceous appendages of critical to the pathogenic provarying lengths and diameters, consisting of a poly- cess; however, their importance is often overlooked mer of a single subunit tipped or interspersed with by many researchers in bacterial pathogenesis, who adhesive proteins, which protrude from the bacterial focus on the intracellular aspects of pathogenesis or cell. The fimbriae are usually arranged peritrichously the later stages of disease. Copyright © 1998 Elsevier Science Ltd. All rights reserved. 0966 842X/98/$19.00
TRENDS
IN
MICROBIOLOGY
282
PII: S0966-842X(98)01288-8
VOL. 6 NO. 7 JULY 1998
REVIEWS
Box 1. Fimbrial diversity and nomenclature49,50 The term fimbriae (meaning threads or fibers) was first coined in 1955 to describe the fiber-like protrusions of bacterial cells. Subsequently, the term pili (meaning hair) was used to describe the same appendages. These two terms have become used interchangeably; however, several authors have recommended that the term pili be reserved for specialized appendages such as sex pili, which are structurally and functionally distinct from adhesive fimbriae. Fimbriae were initially characterized into F groups, similar to the O and K antigens. For example, 987P fimbriae are sometimes called F6 fimbriae. Some fimbriae, such as K88 and K99, were initially wrongly identified as capsule antigens and were incorrectly given ‘K’ denominations. Type I fimbriae were the first fimbriae to be identified by their ability to agglutinate erythrocytes. However, this agglutination is inhibited by the presence of mannose [hence, it is often called mannose-sensitive hemagglutination (MSHA)]. Type I fimbriae are ubiquitous and are found in most Gram-negative bacteria, including Escherichia coli, Salmonella spp., Klebsiella spp. and Vibrio cholerae. A second fimbrial classification, mannose-resistant hemagglutination (MRHA), is used for type I-like fimbriae that can agglutinate erythrocytes but are not inhibited by the addition of mannose. At least five more fimbrial types have been defined. Type II fimbriae are similar to type I but are non-adhesive and are thought to be type I mutants. Type III fimbriae cause MRHA only with erythrocytes that have been pretreated with tannic acid. Two fimbrial types have been classified as type IV fimbriae. Initially, type IV fimbriae were characterized in Proteus as causing an MRHA but being serologically distinct from type I fimbriae. Subsequently, a different type of fimbriae, also called type IV fimbriae, was identified from Pseudomonas aeruginosa. Other members of this group have been identified based on amino acid sequence homologies and similar export mechanisms. Clearly, the characterization of fimbriae as either causing MRHA or MSHA is now redundant and, with the characterization of fimbriae at the molecular level, this nomenclature should be reconsidered. Moreover, the identification of different fimbrial morphologies has rendered the type system somewhat outdated and perhaps confusing. Recently, fimbriae have been given denominations based upon unique characteristics, such as their distribution [for example, the Salmonella enterica serovar Enteritidis fimbria (Sef)] or the electron-microscopic appearance of the fimbriae [for example, the bundle-forming pilus (BFP)]. Although these names are not ideal, as Sef fimbriae have subsequently been identified in other Salmonella species, the unique identification of a fimbria is preferable to the vague, and perhaps physiologically irrelevant, hemagglutination characteristics.
In this review, we take a journey through the mammalian intestinal tract, comparing the diverse fimbrial systems used by intestinal pathogens, such as E. coli, Salmonella spp. and Vibrio cholerae, to adhere to different regions of the intestine (summarized in Table 1) and the particular signals that might mediate their specific expression in a particular niche. Upper gastrointestinal tract The stomach is often considered as the first major barrier to intestinal colonization, because the acid environment (pH 1–2) kills the majority of ingested bacteria. How do bacteria overcome this barrier to ensure intestinal colonization? Some intriguing studies on type I fimbriae suggest that fimbriate strains are able to colonize the pharynx4 and replicate, thereby increasing the chance that a few bacteria will survive the transit through the stomach to colonize the lower intestine. In addition, the pH immediately adjacent to the gastric epithelium is neutralized to protect epithelial cells; this layer of less acidic mucus might provide a region where bacteria can survive during transit. Indeed, some species, for example Helicobacter pylori, colonize this mucus layer. Although adherence is essential for the pathogenesis of H. pylori, specific adhesins have not been identified, but several surface molecules, for example flagella, lipopolysaccharide, the urease enzyme and other outer membrane proteins, have been proposed as adhesins5.
out the lumen of the small intestine, the pH remains between 6.5 and 7.5 (Ref. 6). The epithelium of the proximal small intestine consists mainly of villous epithelial cells, interspersed with specialized cells, called M cells, which are clustered in Peyer’s patches. In contrast, in the epithelium of the distal small intestine, the M cells are much more numerous7. It has been hypothesized that a diverse range of pathogens preferentially invade the intestinal epithelium through the M cells8. Differences in the expression of glycoconjugates on the epithelial surfaces
Small intestine The stomach contents pass into the intestine, where they are neutralized by pancreatic secretions. Through-
TRENDS
IN
MICROBIOLOGY
283
Fig. 1. Electron micrographs of fimbriated bacteria. (a) 987P-fimbriated enterotoxigenic Escherichia coli. Photograph courtesy of Dieter Schifferli. (b) Bundle-forming pilus (BFP)-fimbriated enteropathogenic E. coli . Photograph courtesy of William Murray.
VOL. 6 NO. 7 JULY 1998
REVIEWS
Table 1. Fimbriae used by pathogenic bacteria to adhere to host intestinal epithelial cellsa Microorganism
Fimbriae
Gene cluster
Host
Initial bindingb
Carbon responsec
Ref.
Clostridium difficile Campylobacter jejuni Escherichia coli E. coli E. coli E. coli E. coli E. coli E. coli Salmonella typhimurium S. typhimurium Shigella flexneri Vibrio cholerae
Unknown Unknown AF/R1 AAF/I BFP CFA K88 K99 987P LPF PEF Unknown TCP
Unknown Unknown afr agg bfp cs family fae fan fas lpf pef Unknown tcp
Unknown Unknown Rabbit Human Human Human Porcine Porcine Porcine Various Various Unknown Human
LI LI DSI LI PSI PSI PSI DSI DSI DSI DSI LI PSI
Unknown Unknown Unknown Unknown Good Good None Poor Poor Unknown Unknown Unknown Good
39 38 30 35 14 23 21 47 25 10 32 37 18
a Abbreviations: AF/R1, aggregative fimbriae from rabbit E. coli; AAF/I, aggregative adherence fimbriae; BFP, bundle-forming pilus; CFA, colonization factor antigen; CRP, catabolite repressor protein; DSI, distal small intestine; LI, large intestine; LPF, long polar fimbriae; PEF, plasmid-encoded fimbriae; PSI, proximal small intestine; TCP, toxin co-regulated pilus. b Location of initial binding. Subsequent infection may occur throughout lower regions of the intestine. c Regulation of fimbrial gene expression in response to growth on different carbon sources. ‘Poor’ indicates that expression is activated by the cAMP–CRP complex; ‘good’ indicates that expression is repressed by the cAMP–CRP complex or is optimal in the presence of glucose.
of M cells in the mouse intestine offer potential for the differential targeting of microbial pathogens via their adhesins; thus, species- and site-related variations in the M-cell surface might contribute to microbial tropism9. For example, pathogenic bacteria, such as S. enterica serotype Typhimurium, appear to target M cells specifically for invasion by fimbrial adhesion. S. enterica serotype Typhimurium encodes long polar fimbriae (LPF) that specifically mediate attachment to M cells10,11. However, it is not yet clear whether this property is limited to S. enterica serotype Typhimurium or is a general mechanism used by bacteria that invade the intestinal epithelium via M cells. Proximal small intestine (duodenum/jejunum) Several pathogenic bacteria, such as enteropathogenic E. coli (EPEC), V. cholerae and enterotoxigenic E. coli (ETEC), can initiate intestinal colonization by adhering to the proximal small intestine, either in the duodenum or proximal jejunum, by means of specific fimbriae. Pathological symptoms of disease are initially localized to these regions of the intestinal epithelium. EPEC induce dramatic changes in the host cell cytoskeleton and the destruction of the microvilli, a phenotype known as the attaching and effacing (A/E) lesion12. Analysis of biopsies obtained from EPECinfected children, has shown that these bacteria preferentially attach to jejunal epithelial surfaces as discrete microcolonies13 in a pattern called localized adherence (LA). Expression of the bundle-forming pilus (BFP; Fig. 1b) correlates with the ability of EPEC strains to exhibit LA or ‘autoaggregation’ phenotypes14,15. The EAF-plasmid-encoded gene cluster (bfp), which consists of 14 tandemly arrayed genes, directs the synthesis of the BFP, even in a nonEPEC strain16. It is not clear if the BFP mediates direct adherence to an epithelial host cell receptor or merely
TRENDS
IN
MICROBIOLOGY
acts as an interlinking structure that mediates the formation of free bacterial aggregates. Such aggregates may, in turn, act as functional infectious units that promote direct contact with the epithelial surface and facilitate the interaction of other fimbrial ligands with particular receptors. However, strains carrying mutations in bfpA, the gene coding for the major structural subunit, are attenuated in vivo and fail to produce the LA or autoaggregation phenotypes in vitro, confirming the role of BFP in intestinal pathogenesis (D. Bieber and G.K. Schoolnik, pers. commun.). The attachment of V. cholerae to epithelial cells is achieved by the synthesis of the toxin co-regulated pilus (TCP), which has been shown to be an essential colonization factor17, although a role for other accessory factors in colonization has been suggested18. The TCP biosynthesis gene cluster is located in the chromosome of all virulent strains19. Furthermore, it has been shown that TCP is the receptor for the cholera toxin phage (CTX⌽), a lysogenic filamentous bacteriophage that encodes the two subunits of the cholera toxin. This toxin phage is transferred between V. cholerae strains in the intestine, as seen in the infant mouse cholera model20. Although BFP and TCP belong to the type IV fimbriae family (Box 1), adherence to the upper small intestine is not limited to bacteria expressing fimbriae from this family, as other fimbriae may also be involved (described below). ETEC host tropism and their ability to colonize different intestinal niches are attributable to the diversity of fimbriae produced by different strains. For example, ETEC strains producing K88 fimbriae initially adhere to the proximal region of the small intestine of piglets and calves21. There are three antigenic variants of the K88 fimbriae, and these mediate attachment to three different receptors on the piglet intestinal epithelium22. In human ETEC strains, a variety of fimbriae have been identified and
284
VOL. 6 NO. 7 JULY 1998
REVIEWS
Distal small intestine (jejunum/ileum) Porcine ETEC strains expressing either 987P fimbriae (Fig. 1a) or K99 fimbriae specifically adhere to the epithelium in the distal small intestine2. In contrast to the three K88 antigenic variants described above, identical 987P fimbriae can attach to two different receptors24. A 987P-fimbriated strain only binds to the distal epithelium, even when its receptors are found along the entire length of the piglet intestine, suggesting that a requirement for binding, probably the regulated expression of the fimbriae25, is only being fulfilled in this region. Interestingly, this strain mainly infects neonatal piglets under ~3 weeks of age, after which time the receptor on the intestinal membrane appears to be released so that 987P fimbriae can no longer bind as effectively26. Maturation of the piglet intestine also affects K88 receptors, resulting in altered adherence of strains expressing these fimbriae27,28. Rabbit enteropathogenic E. coli (REPEC) strains produce an A/E-like lesion characteristic of human EPEC infections, although they differ from the human EPEC strains in their host specificity and their capacity to express a variety of adherence factors that are not related to BFP (Ref. 29). A REPEC strain expresses the plasmid-encoded AF/R1 pilus that mediates adherence to M cells30. Other REPEC strains, for example B10 and 83/39, express fimbriae related to the K88 fimbriae described for porcine ETEC (Ref. 31). S. enterica are able to invade the epithelium and cause septicemia in many animals. Salmonella spp. contain several different fimbrial gene clusters, allowing multiple sites of infection. This redundancy complicates efforts to identify unique roles for each fimbrial system. Nevertheless, several studies on the role of fimbriae in the attachment of S. enterica serotype Typhimurium by Bäumler and co-workers11,32 have shown that at least two fimbrial systems, the LPF and the plasmid-encoded fimbriae (PEF), are involved in the binding of this species to the intestinal epithelium. Interestingly, the LPF specifically mediate attachment of S. enterica serotype Typhimurium to the Peyer’s patches, which presumably enables the species to initiate invasion at this site. Conversely, PEF appear to mediate attachment to the villous intestinal epithelium and probably allow S. enterica serotype Typhimurium to multiply in the intestine prior to, or instead of, invading the epithelium11,32.
suitable niches for opportunistic pathogens1; consequently, fewer pathogens have been characterized as causing disease in the large intestine. Enterohemorrhagic E. coli (EHEC) strains (particularly those belonging to the O157:H7 serotype) are the causative agent of diarrhea, hemorrhagic colitis and the hemolytic uremic syndrome in humans. EHEC colonize and produce characteristic A/E lesions in the large intestine of different animals or in in vitro-cultivated cell lines33. A fimbria encoded on the 60-MDa plasmid found in EHEC strains mediates adherence to intestinal cells in tissue culture34; however, further characterization of this adherence factor has not been pursued. Enteroaggregative E. coli (EAEC) strains have been shown to bind to human intestinal colonic mucosa, where they express a 38-amino acid enterotoxin, the EAEC heat-stable enterotoxin 1 (EAST1). EAEC isolates display a distinct pattern of adhesion, called aggregative adherence, which for most of the strains seems to be mediated by a fimbria termed aggregative adherence fimbria (AAF/I)35. Recently, a second aggregative adherence fimbria (AAF/II) has been described in a different EAEC strain36. Although the two AAF fimbrial subunits show some homology at the amino acid level, the organization of their accessory genes seems to be different. Furthermore, the two fimbriae have been shown, by electron microscopy, to be morphologically distinct and, by comparing their binding properties in vitro, to be phenotypically distinct36. Not all EAEC strains produce AAF/I or AAF/II, suggesting that other types of aggregative adherence fimbriae also exist. Shigella flexneri, a close relative of E. coli, specifically invades colonic epithelial cells. Although the mechanism of invasion has been well characterized, the specifics of attachment to the intestinal epithelium have received much less attention. S. flexneri is able to produce several different fimbriae37; however, it has yet to be shown whether any of these are directly involved in the attachment of Shigella to the intestinal epithelium prior to invasion. Similarly, it has been speculated that both Campylobacter jejuni and Clostridium difficile attach to the intestinal epithelium by fimbriae38,39, although no putative adhesive factors have yet been characterized. The normally nonpathogenic bacterium Bacteroides fragilis is also able to produce fimbriae40, which may be important for the colonization of the large intestine. Unfortunately, the paucity of information on the mechanisms of bacterial adhesion to the colonic epithelium prevents comparison with the betterstudied fimbriae used by bacteria to attach to the small intestine.
Large intestine As there is less fluid movement of the contents of the large intestine than those of the small intestine, rapid adhesion of bacteria to epithelial surfaces may be less critical in the former, as bacteria have more time to establish an infection. However, the large numbers of indigenous microorganisms limit the availability of
Fimbrial regulation Using this battery of fimbrial systems, different pathogens are able to adhere to the epithelium at different locations during their journey along the intestinal tract. Enteric bacteria are confronted by a complex combination of varying physicochemical signals, such as temperature, pH, osmolarity, nutrient
characterized as the human colonization factor antigen (CFA) family23. Members of the CFA family recognize different receptors on the human intestinal epithelium and do not mediate binding of human ETEC strains to piglets or other mammals.
TRENDS
IN
MICROBIOLOGY
285
VOL. 6 NO. 7 JULY 1998
REVIEWS
constitutive expression of CS1 or CFA/I fimbriae with no requirement for Rns or CfaR. For CFA Microorganism Fimbriae Initial bindingb Effect of glucose on regulationc Ref. fimbriae, H-NS appears to repress transcription, whereas the positive Escherichia coli BFP PSI Expressed 14 activator is required to overcome E. coli CFA PSI Expressed 23 this repression43,44. E. coli K88 PSI None 21 Expression of V. cholerae TCP E. coli K99 DSI Repressed 47 differs between the two clinically E. coli 987P DSI Repressed 25 important biotypes (classical and Vibrio cholerae TCP PSI Expressed 18 El Tor). Classical strains express TCP optimally in rich growth a Abbreviations: BFP, bundle-forming pilus; CFA, colonization factor antigen; CRP, catabolite medium with an initial pH of 6.5 repressor protein; DSI, distal small intestine; PSI, proximal small intestine; TCP, toxin co-regulated and a high salt concentration. Unpilus. b Location of initial binding. Subsequent infection may occur throughout lower regions of the usually, the optimal temperature intestine. for expression of TCP appears to c Expression of fimbrial genes in response to growth on glucose. ‘Expressed’ indicates expression be 30⬚C instead of the expected in is optimal in the presence of glucose (inhibited by the cAMP–CRP complex); ‘repressed’ vivo temperature (37⬚C), although indicates expression is minimal in the presence of glucose (activated by the cAMP–CRP complex). it is possible that the intestinal conditions that promote expresavailability and the presence of toxic substances. A sion of TCP have yet to be completely mimicked in combination of these environmental signals may pro- vitro. The in vitro conditions for optimal expression vide enteric bacteria with an intestinal ‘road map’ to of TCP in classical strains do not favor the expression identify their appropriate niche and stimulate expres- of TCP in El Tor strains, even though in both strains sion of fimbrial genes. This ensures attachment in the activation of the tcp operon depends on ToxT, whose appropriate host compartment at the correct time. expression is in turn dependent upon ToxR (Ref. 18). Table 2. Environmental regulation of fimbrial genesa
Temperature regulation The expression of several virulence factors is optimal at the temperature of their natural host. For example, E. coli strains that attach to rabbit intestines via REPEC fimbriae express virulence factors (including fimbriae) optimally at 39⬚C, the natural body temperature of the rabbit41. Expression of most fimbrial operons studied, including EPEC bfp and ETEC fas (encoding 987P fimbriae), fae (K88 fimbriae) and fan (K99 fimbriae), is also temperature regulated, with optimal expression occurring at 37⬚C (Refs 14,25). Temperature regulation of many fimbrial operons is mediated by the DNA-binding histone-like protein H-NS (Refs 25,42). Although not all members of the CFA family have been thoroughly characterized, many of them are regulated by similar mechanisms. A transcriptional activator (either Rns or CfaR) is required for gene expression. In an hns strain, there is
Questions for future research • Which environmental and nutritional signals regulate expression of fimbrial operons in vitro and in vivo? • How do the nutrient concentrations and environmental signals interrelate to regulate expression of the multitude of virulence genes required for pathogenesis? • Can nutrients act as signals for fimbriation and, if so, what are the concentration thresholds for different species of pathogenic bacteria. What are the concentrations of nutrients at different points in the intestine? • What fimbriae or adhesins are used by bacteria that attach to or invade the large intestine? • Is fimbrial expression coordinately regulated with the expression of other virulence factors in response to the intestinal signals?
TRENDS
IN
MICROBIOLOGY
Carbon source Nutrient gradients probably provide significant signals to intestinal bacteria. For example, dietary glucose is rapidly absorbed in the upper intestine by all mammals. Thus, changes in luminal glucose concentration offer a precise environmental signal to colonizing bacteria, as the glucose concentration drops from ~25 mM in the upper duodenum to ⬍1 mM in the mid-intestine and is barely detectable in the distal small intestine45. Many bacterial fimbrial operons are regulated in response to glucose, which confirms the notion that this nutrient could be used by colonizing microorganisms as a niche indicator (Table 1). For example, production of both EPEC BFP and V. cholerae TCP is enhanced by growth on good carbon sources, such as glucose14,46, and these bacteria adhere to the upper small intestine to cause disease. Conversely, expression of both ETEC 987P and K99 fimbriae is enhanced by growth on poor carbon sources, such as those that induce the cAMP–catabolite repressor protein (CRP) complex25,47. These fimbriae mediate attachment to the distal small intestine, which is the site of infection. The correlation between glucose gradients and binding of bacteria to different portions of the intestine are shown in Table 2. Nitrogen source Ammonia is excreted as a waste product from intestinal bacteria and influences the nitrogen content of the intestine. This probably varies from low levels of available nitrogen in the upper gastrointestinal tract to high levels in the distal ileum, providing another gradient signal for the regulation of fimbrial gene expression. Ammonia dramatically reduces bfpA expression in EPEC and, consequently, might prevent
286
VOL. 6 NO. 7 JULY 1998
REVIEWS
adherence of EPEC BFP to the distal small intestine, allowing its shedding and further propagation14. Expression of the genes encoding the 987P fimbriae (fas genes) is also regulated by the nitrogen source25 but, unlike bfp regulation, expression of 987P fimbriae is optimal in media containing ammonia. Low levels of ammonia in the proximal intestine trigger BFP expression required for binding in this region, whereas high levels of ammonia stimulate expression of 987P fimbriae required for binding in the distal small intestine. The mechanism by which BFP and 987P fimbriae are regulated in response to nitrogen status remains to be elucidated, although regulation might occur either directly or indirectly through the Ntr nitrogen-regulation system48. Conclusions Studies on the attachment of intestinal pathogens to epithelial cells have led to both an understanding of the importance of adhesion to disease and a better knowledge of how disease progresses; however, our knowledge is still rudimentary. Nevertheless, it appears that environmental cues, such as temperature, carbon or nitrogen sources and their gradients in the intestine, can act not only as regulatory signals for both fimbrial expression and bacterial pathogenesis but also as ‘street signs’, indicating the niches suitable for infection. Acknowledgements We thank Andreas Bäumler and Dave Bieber for sharing data prior to publication, Dieter Schifferli and William Murray for the electron micrographs used in Fig. 1, and Stanley Maloy, David Nunn, Paula Ostrovsky and Dieter Schifferli for critically reviewing the manuscript. This work was partly supported by grants from the NIH (PHS GM34715) (to R.A.E.), Universidad Nacional Autónoma de México (DGAPA IN208095) and Consejo Nacional de Ciencia y Tecnologia (CONACyT 1027P-N) (to J.L.P.). References 1 Batt, R.M., Rutgers, H.C. and Sancak, A.A. (1996) J. Small Anim. Pract. 37, 261–267 2 Isaacson, R.E., Nagy, B. and Moon, H.W. (1977) J. Infect. Dis. 135, 531–539 3 Bäumler, A.J., Tsolis, R.M. and Heffron, F. (1997) Adv. Exp. Med. Biol. 412, 149–158 4 Bloch, C., Stocker, B. and Orndorff, P. (1992) Mol. Microbiol. 6, 697–701 5 Nedrud, J.G. and Czinn, S.J. (1997) Curr. Opin. Gastroenterol. 13, 71–78 6 Evans, D.F. et al. (1988) Gut 29, 1035–1041 7 Siebers, A. and Finlay, B.B. (1996) Trends Microbiol. 4, 22–29 8 Neutra, M.R., Pringault, E. and Kraehenbuhl, J.P. (1996) Annu. Rev. Immunol. 14, 275–300 9 Jepson, M.A. et al. (1996) J. Anat. 189, 507–516 10 Bäumler, A.J., Tsolis, R.M. and Heffron, F. (1996) Proc. Natl. Acad. Sci. U. S. A. 93, 279–283 11 Norris, T.L., Kingsley, R.A. and Bäumler, A.J. Mol. Microbiol. (in press) 12 Donnenberg, M.S., Kaper, J.B. and Finlay, B.B. (1997) Trends Microbiol. 5, 109–114 13 Rothbaum, R.J. et al. (1983) Ultrastruct. Pathol. 4, 291–304 14 Puente, J.L. et al. (1996) Mol. Microbiol. 20, 87–100 15 Ramer, S.W., Bieber, D. and Schoolnik, G.K. (1996) J. Bacteriol. 178, 6555–6563
TRENDS
IN
MICROBIOLOGY
16 Sohel, I. et al. (1996) J. Bacteriol. 178, 2613–2628 17 Herrington, D.A. et al. (1988) J. Exp. Med. 168, 1487–1492 18 Skorupski, K. and Taylor, R.K. (1997) Mol. Microbiol. 25, 1003–1009 19 Manning, P.A. (1997) Gene 192, 63–70 20 Waldor, M. and Mekalanos, J. (1996) Science 272, 1910–1914 21 Moon, H.W., Isaacson, R.E. and Pohlenz, J. (1979) Am. J. Clin. Nutr. 32, 119–127 22 Mouricout, M. (1997) Adv. Exp. Med. Biol. 412, 109–123 23 Gaastra, W. and Svennerholm, A-M. (1996) Trends Microbiol. 4, 444–452 24 Khan, A.S. and Schifferli, D.M. (1994) Infect. Immun. 62, 4223–4243 25 Edwards, R.A. and Schifferli, D.M. (1997) Mol. Microbiol. 25, 797–809 26 Dean-Nystrom, E.A. and Samuel, J.E. (1994) Infect. Immun. 62, 4789–4794 27 Conway, P.L., Welin, A. and Cohen, P.S. (1990) Infect. Immun. 58, 3178–3182 28 Grange, P.A. et al. (1998) Infect. Immun. 66, 1613–1621 29 Robins-Browne, R.M. et al. (1994) Infect. Immun. 62, 1584–1592 30 Rafiee, P. et al. (1991) J. Cell Biol. 115, 1021–1029 31 Fiederling, F. et al. (1997) Infect. Immun. 65, 847–851 32 Bäumler, A.J. et al. (1996) Infect. Immun. 64, 61–68 33 Donnenberg, M.S. et al. (1993) J. Clin. Invest. 92, 1418–1424 34 Karch, H. et al. (1987) Infect. Immun. 55, 455–461 35 Savarino, S.J. et al. (1994) J. Bacteriol. 176, 4949–4957 36 Czeczulin, J.R. et al. (1997) Infect. Immun. 65, 4135–4145 37 Snellings, N.J., Tall, B.D. and Venkatesan, M.M. (1997) Infect. Immun. 65, 2462–2467 38 Dolg, P. et al. (1996) Mol. Microbiol. 20, 885–894 39 Borriello, S.P. et al. (1990) Rev. Infect. Dis. 12, S185–S191 40 Brook, I. and Myhal, M.L. (1991) Infect. Immun. 59, 742–744 41 Abe, A. et al. (1997) Infect. Immun. 65, 3547–3555 42 Atlung, T. and Ingmer, H. (1997) Mol. Microbiol. 24, 7–17 43 Murphree, D., Froehlich, B. and Scott, J.R. (1997) J. Bacteriol. 179, 5736–5743 44 Jordi, B.J.A.M. et al. (1992) EMBO J. 11, 2627–2632 45 Ferraris, R.P. et al. (1990) Am. J. Physiol. 259, G822–G837 46 Skorupski, K. and Taylor, R. (1997) Proc. Natl. Acad. Sci. U. S. A. 94, 265–270 47 Lo-Tseng, T., Lee, J. and Isaacson, R.E. (1997) Adv. Exp. Med. Biol. 412, 303–310 48 Merrick, M.J. and Edwards, R.A. (1995) Microbiol. Rev. 59, 604–622 49 Duguid, J. and Old, D. (1994) in Fimbriae. Adhesion, Genetics, Biogenesis, and Vaccines (Klemm, P., ed.), pp. 1–7, CRC Press 50 Low, D., Braaten, B. and van der Woude, M. (1996) in Escherichia coli and Salmonella typhimurium. Cellular and Molecular Biology (Neidhardt, F.C. et al., eds), pp. 146–157, ASM Press
287
Writing a review for TIM Most of the reviews published in Trends in Microbiology are specially commissioned by the Editor. However, other ideas for articles will also be considered. If you wish to write a review article for TIM, please contact the Editor first with an outline of your intended article. Note that all articles are subject to rigorous peer and editorial review: commissioning does not guarantee publication.
VOL. 6 NO. 7 JUNE 1998
