Influence of the age and sex of human hosts on the distribution of Escherichia coli ECOR groups and virulence traits

David M. Gordon1, Steven E. Stern2 and Peter J. Collignon3

1 School of Botany and Zoology, Australian National University, Canberra ACT 0200, Australia
2 School of Finance and Applied Statistics, Australian National University, Canberra ACT 0200, Australia
3 Department of Microbiology and Infectious Diseases, The Canberra Hospital, PO Box 11 Woden, ACT 2606 Australia

Correspondence
David Gordon
David.Gordon{at}anu.edu.au


   ABSTRACT
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Escherichia coli were isolated from the faeces of 266 individuals living in the Canberra region of Australia. The isolates were characterized for their ECOR group membership (A, B1, B2 or D) and for the presence of 29 virulence-associated traits. Overall, 19·5 % of the strains were members of group A, 12·4 % B1, 45·1 % B2 and 22·9 % D. The frequency with which strains belonging to the four ECOR groups were observed varied with the age and sex of the hosts from which they were isolated. In males, the probability of isolating A or D strains increased with host age, whilst the probability of detecting a group B2 strain declined. In females, the probability of recovering A or B2 strains increased with increasing host age and there was a concomitant decline in the likelihood of isolating B1 or D strains. Of the 29 virulence-associated traits examined, 24 were detected in more than one strain. The likelihood of detecting most traits varied with a strain's ECOR membership, with the exception of afa/draBC, astA, cvaC, eaeA, iss and iutA, for which there was no statistically significant evidence of an association with ECOR group. The frequency with which fimH, iha, eaeA, iroN, hlyD, iss, ompT and K1 were detected in a strain depended on the age or sex of the host from which the strain was isolated. In group B2 strains many of the virulence traits were non-randomly associated, with some co-occurring in a strain less often than expected by chance, whilst others were co-associated. In 17 cases, the extent to which two virulence traits were co-associated was found to depend on host sex and age. The results of this study suggest that the morphological, physiological and dietary differences that occur among human individuals of different sex or age may influence the distribution of E. coli genotypes.


   INTRODUCTION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
A number of mathematical models have been developed with the purpose of exploring the dynamics of bacterial populations in the intestinal tract and have been formulated in the context of either chemostat or plug-flow reactor systems (Topiwala & Hamer, 1971; Baltzis & Fredrickson, 1983; Freter et al., 1986; Smith & Waltman, 1995; Ballyk & Smith, 1999; Stemmons & Smith, 2000; Ballyk et al., 2001; Jones et al., 2002). The chemostat models best describe intestinal tracts incorporating a microbial fermentation chamber, whilst the plug flow reactor models are more appropriate for intestinal tracts lacking a fermentation chamber. The predictions of these models have shown that the establishment of a bacterial population is more likely when the cells are capable of adhering to the lining of the intestinal tract than when they are not. Nutrient concentration is a factor and there is a threshold concentration of nutrients entering the model gut below which the bacterial population cannot establish and above which the population will establish and persist. Further, the time required for material to move through the gut can also determine establishment success, where decreasing transit times lead to populations being less likely to establish.

In a survey of vertebrates living in Australia, Gordon & Cowling (2003) demonstrated that the probability of detecting Escherichia coli in mammalian host was dependent on the body mass and diet of the host. E. coli was more likely to be recovered from hosts with intestinal tracts incorporating a microbial fermentation chamber (omnivores and herbivores) and less likely to be isolated from hosts with simple tube-like intestines (carnivores). The probability of detecting E. coli in a host increased with the body mass of the host. In most mammals, gut transit times increase with increasing body mass (Hume, 1999).

Faecal isolates of E. coli can be divided into four main genetic groups (ECOR groups) designated A, B1, B2 and D (Ochman & Selander, 1984; Herzer et al., 1990). These four groups differ in their phenotypic and genotypic characteristics (Selander et al., 1987; Johnson et al., 2001; Bergthorsson & Ochman, 1998). Gordon & Cowling (2003) demonstrated that the diet and body mass of the mammalian host from which an E. coli strain was isolated predicted the strain's E. coli group membership. For example, in omnivorous hosts, larger hosts were more likely than small-bodied hosts to harbour a group B2 strain. In carnivorous mammals, the probability of isolating a group A strain increased as host body mass increased.

Thus, the results of theoretical and empirical studies indicate that the morphology and dynamics of the intestinal tract influence the probability that E. coli will establish a population in the gut. Further, these same factors, in part, determine the genotype of the strain that successfully establishes. In adult humans, there is significant variation among individuals in the dynamics and morphology of the intestinal tract. The length of the intestine varies between men and woman and between people of different ages, with the intestine being longer in men and in young people (Hounnou et al., 2002). A number of studies have demonstrated that gut transit time differs between males and females (Graft et al., 2001). Transit times are, on average, significantly longer in females compared to males and this is true of transit times in the small intestine and colon. The majority of E. coli strains inhabit the colon (Hartl & Dykhuizen, 1984), and in this region, transit times are typically about 50 % longer in females compared to males. Gut transit times also change with age in adult humans, although the direction of change depends on the gut region under consideration. Transit times in the small intestine generally decrease with age, but colon transit times increase.

Mathematical theory shows that the quality and quantity of nutrients available to bacteria in the intestine will affect the growth rate of the bacteria. The magnitude of the population growth rate will determine whether the population is capable of overcoming the loss of cells from the intestine due to the movement of material through the gut. Energy intake is known to change with host age and can decline by 1000–1200 kcal (4200–5000 kJ) day–1 in men and by 600–800 kcal (2500–3300 kJ) day–1 in women (Wakimoto & Block, 2001). Diet choice also changes with age, with older people including a greater fraction of grains, vegetables and fruits in their diet compared to younger individuals (Drewnowski & Shultz, 2001). Nutrient availability in the colon will depend on events occurring in the small intestine. The ability of the gastro-intestinal tract to digest and absorb macro-nutrients appears to be unaffected by age, but it becomes less efficient at absorbing some micro-nutrients (Russell, 2000).

The purpose of this study was to determine whether the age or sex of a human host explained any of the genetic variation observed in E. coli strains isolated from human faeces. To this end, 266 E. coli strains were characterized with respect to their E. coli group membership and for the presence of 29 putative virulence genes associated with intestinal and extra-intestinal disease.


   METHODS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Strain isolation and identification.
During the first 6 months of 2002, faecal samples were collected from people living in the city of Canberra and its surrounds (Australian Capitol Territory, Australia). Samples were acquired from the microbiology laboratory of the Canberra Hospital and from the academic community of the Australian National University. A single faecal sample was collected per person and a single E. coli clone isolated per faecal sample. The Canberra Hospital isolates were collected from 189 in- and outpatients (95 females and 94 males) exhibiting symptoms suggestive of gastro-intestinal disease. There were 77 isolates from people in the age category of 0–19 years; 42 isolates, 20–39 years; 37 isolates, 40–59 years; 22 isolates, 60–79 years; and 11 isolates from people 80+ years of age. There were 77 isolates collected from the Australian National University community (36 females and 41 males); these individuals were asymptomatic. Individual host age data were not available for these isolates, but the isolates were taken from people between 21 and 65 years of age. There was no information available on the recent history of antibiotic use by any of the people sampled.

A subsample of the fresh faecal material was dilution-streaked onto a MacConkey agar plate. After incubation, a single colony with the appropriate colour and morphology was dilution-streaked onto a MacConkey agar plate and, after a further round of incubation, dilution-streaked onto another MacConkey agar plate and then onto a Luria broth agar plate. All lactose-positive isolates were tested for growth on a Simmons citrate agar plate and for indole production. Freezer cultures were produced by incubating the isolate overnight in Luria broth, after which 1 ml of the overnight culture was mixed with 30 µl glycerol and stored at –80 °C. All incubations were carried out at 35 °C.

ECOR group assignment.
All strains with a phenotype consistent with that of E. coli (Lac+ Cit Ind+) were genotyped. Template DNA for ECOR group assignment and virulence factor profiling was prepared using DNAzol (Invitrogen) according to the manufacturer's protocol. The recently developed method of Clermont et al. (2000) was used to assign the E. coli isolates to one of the four main groups of E. coli identified in the ECOR collection (Ochman & Selander, 1984; Herzer et al., 1990). The groups are designated A, B1, B2 and D. The Clermont method is based on a multiplex PCR protocol that determines the presence or absence of two genes (chuA and yjaA) and an anonymous DNA fragment (TSPE4.C2). The presence/absence of the three PCR products is used in the manner of a dichotomous key to assign an unknown isolate to one of the ECOR groups. A fraction of group A strains are negative for the three DNA products (Clermont et al., 2000). Strains failing to produce any PCR products were repeated twice more using newly prepared template DNA. After three attempts, any strain negative for all three PCR products was identified using the BBL Enteric/Nonfermenter ID kit in conjunction with the BBL Crystal System Electronic codebook. If the strain was identified as E. coli, it was considered to be an ECOR A strain. The presence of one or more PCR products of the correct size was considered further evidence of the identity of an isolate as E. coli.

Virulence factor profiling.
The 266 E. coli strains were screened for the presence of 29 virulence factors associated with intestinal and extra-intestinal disease using a modification of the multiplex PCR protocol described by Johnson & Stell (2000). There were five primer pools: pool 1, malX, papAH, fimH, focG, traT, ibeA; pool 2, gafD, fyuA, iroN, ompT, sfa/focDE, kpsMT.II; pool 3, hylD, iha, cvaC, afa/draBC, iss, ireA, kpsMT.K1; pool 4, H7, cnf1, eaeA, iutA, stx2, stx1; pool 5, she, eaag, 13fb, astA. Amplification was done in a 25 µl reaction mixture containing 2 µl template DNA, 4 mM MgCl2 (pools 1–4) or 2 mM MgCl2 (pool 5), and 0·5 units Platinum Taq in 1x reaction buffer [67 mM Tris/HCl (pH 8·8), 16·6 mM (NH4)2SO4, 0·45 % Triton X-100, 0·2 mg gelatin ml–1 and 0·2 mM dNTPs]. Amplification conditions were: 1 cycle at 95 °C for 12 min; 25 cycles of 94 °C, 30 s; 63 °C, 30 s (pools 1–4) or 55 °C, 30 s (pool 5); 68 °C, 3 min; and 1 cycle at 72 °C for 3 min. Electrophoresis was carried out using 2 % agarose gels. Gels were stained with ethidium bromide, visualized using an ultraviolet trans-illuminator, and photographed.

Statistical analysis.
The distribution of ECOR groups according to host age and sex was investigated by means of multinomial logistic regression analysis using a linear model on the logit scale and incorporating interactive effects for host age and sex. Observations for which either host age or sex was not available were excluded from the analysis. Classification tree analysis was used to verify that the linear logistic link structure was appropriate. Individual gene associations with host age and sex, after adjusting for any ECOR group effect, were investigated using standard logistic regression. Care was taken to assess the degree of influence associated with each observation, and any spurious significance caused by overly influential observations was discounted (which typically occurred in cases where the gene under investigation was very rare, either overall or within a particular host/sex category). Moreover, care was also taken in the determination of the final significant models for presentation, so that final models which show a host age effect in only a single sex were only considered when the overall interaction effect of host age and sex was found to be significant. Initially, observations for which either the host age or sex was not available were excluded from the analysis. However, in cases where it was determined that either host age or sex was not significantly related to the distribution of a gene, the corresponding predictor was removed from the model and any observations which had been excluded on the basis of a missing value for this factor were returned to the analysis. The effect of ECOR group on individual gene distributions was itself assessed using logistic regression analysis, and the results of these analyses were used as the basis for any ECOR group adjustments that were required for the analysis of host age and sex effects on gene distributions. Specifically, ECOR group effects were included in the overall logistic model for each individual gene only if the ECOR group analysis suggested that the distribution of a gene was significantly different within a particular ECOR group. In this way, the overall relationship between ECOR group and host age and sex was controlled for within the analysis of host age and sex relationships with the individual genes. Contingency table analysis was used to determine the extent to which virulence traits were non-randomly associated in a strain and the association was quantified using the estimate kappa={(O–+O+)–(E–+E+)}/(NE–+E+), where O– is the number of cases where both traits are absent in a strain and O+ is the number of cases where both traits are present. E– and E+ are the expected number of cases where the traits are either both present or absent assuming that the traits are randomly associating and N is the total number of samples. To determine if the degree to which two traits were co-associated varied as a function of host sex or age, nominal logistic regression was used, where one of the traits served as an independent variable and the other trait as the response variable.


   RESULTS
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Distribution of ECOR groups
Of the 266 strains characterized for their ECOR group membership, 19·5 % were group A, 12·4 % B1, 45·1 % B2 and 22·9 % D strains. The frequencies of the different ECOR groups varied with respect to the age and sex of the hosts from which the strains were isolated (likelihood ratio tests: age, {chi}2=8·036, P=0·045; sex, {chi}2=3·168, P=0·366; sexxage, {chi}2=21·000, P=0·0001). In males, the probability of isolating A or D strains increased with increasing host age, whilst the probability of isolating B2 strains declined with increasing host age (Fig. 1a). In females, the probability of isolating A or B2 strains increased with increasing host age, whilst the probability of isolating B1 or D strains declined with increasing host age (Fig. 1b). The appropriateness of the linear logistic structure was investigated visually by constructing mosaic plots analogous to Figs 1(a) and 1(b), based on standard classification tree (CART) analysis. The resultant figures (not presented) confirmed the adequacy of the chosen multinomial logistic model structure.



View larger version (41K):
[in this window]
[in a new window]
 
Fig. 1. Predicted frequency of E. coli ECOR groups A, B1, B2 and D isolated from the faeces of human hosts of different sex and ages based on the results of a multi-nominal logistic regression model.

 
Distribution of virulence factors
The virulence-associated traits eaag, gafD, stx1 and stx2 were not detected and 13fb was detected in a single group A strain. The probability of detecting the traits afa/draBC, astA, cvaC, eaeA, iss and iutA in a strain did not depend on the strain's ECOR group membership (Table 1). The probability of detecting all other traits varied with a strain's ECOR group membership, and most of the traits were more likely to be detected in B2 strains and less likely to be detected in A or B1 strains (Table 1).


View this table:
[in this window]
[in a new window]
 
Table 1. Frequency (%) with which virulence-associated traits were detected in E. coli strains isolated from human faeces with respect to the ECOR group membership of the strains

 
As has been demonstrated, the probability of isolating a strain of a particular ECOR group depends on the sex and age of the host from which the strain was recovered. Further, the probability of detecting most of these virulence-associated traits in a strain depends on the strain's ECOR group membership. Therefore, any attempt to determine if the probability of detecting a trait in a strain depends on the sex or the age of the host from which the strain was recovered must take into account the fact that the probability isolating a strain of a particular ECOR group varies as a function of host age and sex. For example, cnf1 is unlikely to be detected in any strain other than a group B2 strain. Furthermore, the probability of recovering a B2 strain in a host increases with host age in females but declines with host age in males. Consequently, it would be false to conclude that the probability of detecting cnf1 in a strain depended on the age and sex of the host, rather than concluding that it is the probability of isolating a B2 strain that depends on the age and sex of the host.

Of the 29 traits examined in this study, 24 were detected in more than one strain. Analysis of the data adjusting for ECOR group effects revealed that, for eight of the 24 traits, a statistically significant host sex or age effect on the probability of detecting the trait in a strain was observed (Table 2). For the virulence-associated traits fimH, iha and eaeA, the probability of detecting these factors in a strain declined as the age of the host from which the strain was recovered increased. fimH was detected in virtually all B1, B2 and D strains but in only 81 % of group A strains. However, the probability of detecting fimH in a group A strain declined with host age. The gene iha was detected in about a third of B2 strains, a quarter of group A and D strains and only 10 % of group B1 strains. The probability of detecting iha in B2 and D strains declined with increasing host age, but the frequency of iha in group A strains was independent of host age. The trait eaeA was relatively uncommon and the probability of detecting this gene was independent of a strain's ECOR group membership. However, eaeA was more likely to be detected in strains from young hosts compared to strains from older hosts. The probability of detecting iroN, hylD and iss in a strain was found to increase with the age of the host when the host was a male, but not if the host was a female.


View this table:
[in this window]
[in a new window]
 
Table 2. Influence of human host age and sex on the frequency with which virulence-associated traits were detected in E. coli strains isolated from human faeces

P values presented are for the model described in the Effect column.

 
The sex of the host from which the strain was recovered was found to influence the probability of detecting ompT and K1 in a strain. Groups A, B1 and D strains from males were more likely to be ompT positive than were these strains from females. ompT was detected in most (88 %) of group B2 strains and this may explain the lack of a host sex effect on the frequency of ompT in group B2 strains. Regardless of the ECOR group membership of a strain, K1 was less likely to be detected in strains isolated from males compared to strains recovered from females.

Many of the virulence-associated traits examined in this study are known to co-occur in strains significantly more frequently than would be expected by chance (Johnson & Stell, 2000). The co-association of virulence traits was examined for group B2 strains. B2 strains were selected as they were the most common group of strains and because B2 strains have the greatest diversity of virulence traits. The analyses were restricted to traits that were neither too common (>80 %) nor too rare (<20 %). Of the 77 possible pair-wise comparisons between traits, 31 were significant ({alpha}<0·01) (Table 3). In most cases traits were co-associated; for example papAH and hylD co-occurred more often than would be expected by chance (Table 3). Some traits, such as iutA and ibeA, co-occurred in a strain significantly less frequently than would be expected (Table 3).


View this table:
[in this window]
[in a new window]
 
Table 3. Co-occurrence of virulence-associated traits in ECOR group B2 E. coli strains isolated from human faeces

Numbers below the diagonal present the degree to which two traits are associated. Numbers above the diagonal denote the statistical significance of the association (probability of obtaining a greater likelihood ratio {chi}2 value by chance). Only those associations significant at the {alpha}=0·01 level are presented and all other associations are assumed to be 0. Entries in bold indicate those associations influenced by host sex or age (see Table 4). NS, Not significant.

 
The association analyses were extended by posing the following question: Does the extent to which two traits co-occur in a strain vary with respect to the age or sex of the hosts from which the strains were isolated? There were 17 cases where the nature of the association between virulence factors traits depended on host sex or age (Table 4) and the majority of these cases (13) occurred for traits where the simple association analysis indicated that the traits were un-associated (Table 3). For example, no significant association of papAH and iha was detected (Table 3), but an analysis including the effect of host sex and age revealed that papAH and iha were co-associated in strains isolated from females but not in strains from males (Table 4).


View this table:
[in this window]
[in a new window]
 
Table 4. The nature of host age and sex effects on the co-occurrence of virulence factor traits detected in ECOR group B2 E. coli strains isolated from human faeces

P values are for the model described in the Description of relationship column. OR, odds ratio.

 

   DISCUSSION
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Previous studies have shown that the E. coli clonal community of a person is numerically dominated by one strain, or at most a few strains (Caugant et al., 1981, 1983; Gordon et al., 2002; Schlager et al., 2002). Consequently, the sampling regime used in the present study was most likely to have recovered the dominant strain present in an individual.

In interpreting the results of this study two important caveats must be recognized. The first of these concerns the extent to which the hospital isolates constitute a non-representative sample of faecal E. coli. Although these isolates were from people with gastro-intestinal symptoms, the tests performed by the microbiology laboratory of the Canberra Hospital did not suggest that bacterial infections were the cause of the symptoms. Further, none of the E. coli isolates had virulence factor profiles that would suggest these strains might have been responsible for the symptoms, given that the genes eaag, 13fb, stx1 and stx2 were virtually undetected. Overall, the relative frequencies of strains of the four ECOR groups were virtually identical in the two samples, and very similar distributions were observed when the comparison of the hospital and community isolates was restricted to hosts between the ages of 20 and 65 years. All isolates were screened for resistance to 16 antibiotics using a disk diffusion assay (BBL Sensi-Disk, Bacto) and no statistically significant differences between the two samples were detected in frequency of resistance to any of the antibiotics tested (unpublished data). Finally, additional statistical analyses (not presented) did not suggest that the hospital and community isolates represent distinct populations.

The second caveat relates to the exploratory nature of the statistical analyses that were undertaken. A great many statistical tests were performed, and as the number of tests increases so does the risk of committing a type 1 error [rejecting the null hypothesis (no effect) when it is in fact true]. Consequently, some of the host age and sex effects may have arisen by chance, and the results of those statistical tests with P values >0·01 must be treated with caution. Thus, the outcomes of the analyses that were performed are best viewed as hypothesis generation rather than hypothesis testing.

The results clearly demonstrate that the genotype of the dominant E. coli strain present in a host is, in part, determined by the sex and age of that host. Host age and sex influence the probability of isolating a strain of a particular E. coli group (Fig. 1). Further, the analyses revealed that the frequency of a third of the virulence-associated traits examined varied in some manner with respect to the age or sex of the hosts from which the strains were isolated (Table 2). Finally, the extent to which two traits were associated in a strain and the nature of the association also varied with the age and sex of the hosts from which the strains were recovered (Table 4). There appears to be no pattern as regards the putative function of the genes whose frequencies or associations were influenced by host factors. The distribution of genes involved in adhesion, iron uptake, toxin production and cell surface attributes varied with respect to host age or sex. Significant host effects were detected for plasmid-encoded traits such as iroN and iss, chromosomally determined traits, such as K1 and eaeA, as well as chromosomally determined traits usually associated with pathogenicity islands, for example papAH and hlyD.

The results of this study suggest that intestinal tracts of people of unlike sex and ages represent different habitats to E. coli. However, the exact nature of the differences between the intestinal environment of males and females and the changes that occur in the intestine with age are unknown. Also unknown is the adaptive significance of the distribution of the virulence traits observed in this study. Clearly much more research is required if we are to understand the nature of the interactions that occur between E. coli and the intestinal environment. The results of this study potentially have significant implications for clinical microbiology, particularly in relation to disease prevention and the nature of pathogenicity in E. coli. Some of the clinically relevant implications of this study are discussed below.

Are there host sex- or age-specific strains of E. coli? At one level, this is unlikely to be true, largely because of the great genotypic diversity that exists in E. coli. For example, in a study of E. coli isolated from Australian mammals, the probability of isolating the same genotype from two hosts was <1 %. Similar results have been found for studies examining the clonal diversity to be found in human hosts (Caugant et al., 1983; Gordon et al., 2002). Consequently, it is highly unlikely that identical genotypes will ever be recovered sufficiently frequently from multiple hosts to allow one to conclude that a particular genotype is significantly more likely to be recovered from a host of a particular sex or age. However, it may well be true that there are gene combinations that are more likely to be isolated from a host of a particular sex or age. The results of the analyses addressing the question of whether there are differences in the extent to which two genes are associated in strains isolated from people of different sex or age provide support for this idea. However, any rigorous statistical investigation examining 3-, 4- or n-way gene associations will need to use a much larger sample of strains than was available for this study.

Human females are at a greater risk of urinary tract infection compared to males, largely because of the difference in their uro-genital morphology. The intestinal tract of males and females appears to represent different environments for E. coli. Is it possible that the nature of the adaptations required for E. coli to successfully invade and establish in the female intestine predisposes females to urinary tract infection? Many of the traits implicated in extra-intestinal infection are most prevalent in B2 strains (Table 1), and B2 strains are more common in females compared to males. The nature of the co-association between 17 pairs of virulence traits was found to vary with host sex or age (Table 4). For 13 of these cases, the two traits were more likely to co-occur in strains isolated from females compared to those isolated from males.

Duriez et al. (2001) demonstrated that, in faecal samples taken from humans, the relative frequencies of strains of the four ECOR groups varied among geographical localities (Table 5). In the Australian sample, the distribution of strains of the four ECOR groups is very different from those reported by Duriez et al. (2001). Group B2 strains were far more frequent in the Australian sample compared to the other populations (Table 5). The results of the present study demonstrate that when making comparisons such as those shown in Table 5 it would be prudent to adjust for the age and sex structure of the host populations from which the strains were isolated. Duriez et al. (2001) provided no data on the age and sex of the hosts from which their samples were collected. However, given the magnitude of the variation, it is unlikely that disparities in the age and sex structure of the populations could account for all of the among-population variation that is observed. Gordon & Cowling (2003) reported that the climate of the locality from which E. coli were isolated, as well as the diet of the host, influenced the likelihood of isolating a strain of a particular ECOR group. Consequently, as discussed by Duriez et al. (2001), it is unknown whether the variation among human populations in the distribution of strains of the four ECOR groups is a consequence of factors related to the external environment or due to cultural differences in diet or food processing and preparation practices.


View this table:
[in this window]
[in a new window]
 
Table 5. Frequency (%) of ECOR group strains isolated from human faeces at different localities

Data from Duriez et al. (2001).

 
The role played by many of the virulence-associated genes examined in this study in causing disease is largely unknown and this is particularly true of the genes implicated in extra-intestinal infection. Many of the genes confer an obvious in vitro phenotype, for example, haemolysin production (hlyD) and increased serum survival (iss); however, the mechanisms by which they cause disease in vivo are not precisely known. Many of these traits are thought to be associated with virulence because they are detected at higher frequencies in clinical compared to faecal isolates of E. coli. However, the results of the present study indicate that such comparisons are fraught with difficulties. It is well known that the majority of extra-intestinal E. coli infections in humans are caused by group B2 strains and to a lesser extent group D strains (Picard et al., 1999; Johnson et al., 2001). It is also well established that the frequency with which many of these genes are detected in a strain depends on the strain's E. coli group membership and that they are more frequent in B2 and D strains compared to A or B1 strains (Table 1; Johnson et al., 2001; Manges et al., 2001; Duriez et al., 2001). Finally, in many human populations B2 and D strains represent a minority of the strains isolated from faeces (Picard et al., 1993; Duriez et al., 2001). Therefore, given that the frequency of a gene may depend on a strain's ECOR group membership, is it appropriate to compare the frequency of a gene in clinical isolates (B2 and D strains) with the frequency of the gene in faecal isolates (A and B1 strains)? The results of the present study also indicate that such comparisons should not only account for a strain's ECOR group membership, but also should adjust for the age and sex structure of the host population from which the isolates are obtained.

E. coli causes a significant fraction of urinary tract infections in humans, and strains responsible for an infection are thought to originate from the intestinal E. coli community of the infected host. Two hypotheses have been suggested to explain why urinary tract infections are caused by strains residing in the intestine (Mobley & Warren, 1996). The first hypothesis predicts it is the numerically dominant faecal strain that is most likely to invade the urinary tract. The second hypothesis predicts it is the subset of the E. coli community in a host that possesses a particular suite of virulence factors that will infect the host. It has been suggested that if an E. coli strain possesses two or more of the traits papAH, afa/draBC, sfa/focDE, kpsMT.II or iutA it is capable of causing a urinary tract infection (Johnson et al., 2003). Excepting afa/draBC, all of these traits are most prevalent among group B2 strains (Table 1). Of the 266 strains examined, 44 % encoded two or more of these five virulence factors: 19 % of group A, 0 % of B1, 73 % of B2 and 31 % of D strains. If it were the numerically dominant strain that is most likely to cause urinary tract infections, then the expectation would be that the incidence of urinary tract infections would vary substantially among people living in different regions (Table 5). This assumes that the distribution of these traits is similar across different localities. Conversely, a similar incidence of urinary tract infection among people from different regions could suggest that the suite of traits that enables an E. coli strain to cause extra-intestinal infections needs to be redefined. However, there is little information concerning the incidence of urinary tract infections in different regions.

The results of this study and that by Gordon & Cowling (2003) suggest that the morphological, physiological and dietary differences that occur among different animal species, or among individuals of the same species, but of different sex or age, influence the distribution of E. coli genotypes. Further studies are required to better describe these patterns of variation and to understand the nature of the adaptations that enable strains of E. coli to exploit different intestinal environments.


   ACKNOWLEDGEMENTS
 
The technical assistance of Joanne Allison was invaluable. This study was funded in part by the Australian Research Council and the Faculty Research Fund of the Australian National University.


   REFERENCES
TOP
ABSTRACT
INTRODUCTION
METHODS
RESULTS
DISCUSSION
REFERENCES
 
Ballyk, M. & Smith, H. L. (1999). A model of microbial growth in a plug-flow reactor with wall attachment. Math Biosci 158, 95–126.[CrossRef][Medline]

Ballyk, M., Le, D., Jones, D. A. & Smith, H. L. (2001). Microbial competition in reactors with wall attachment: a comparison of chemostat and plug flow models. Microb Ecol 41, 210–221.[Medline]

Baltzis, B. C. & Fredrickson, A. G. (1983). Competition of two microbial populations for a single resource in a chemostat when one of them exhibits wall attachment. Biotechnol Bioeng 25, 2419–2439.

Bergthorsson, U. & Ochman, H. (1998). Distribution of chromosome length variation in natural isolates of Escherichia coli. Mol Biol Evol 15, 6–16.[Abstract]

Caugant, D. A., Levin, B. R. & Selander, R. K. (1981). Genetic diversity and temporal variation in the E. coli populations of a human host. Genetics 98, 467–490.[Abstract/Free Full Text]

Caugant, D. A., Levin, B. R. & Selander, R. K. (1983). Distribution of multilocus genotypes of Escherichia coli within and between host families. J Hyg 92, 377–384.

Clermont, O., Bonacorsi, S. & Bingen, E. (2000). Rapid and simple determination of Escherichia coli phylogenetic group. Appl Environ Microbiol 66, 4555–4558.[Abstract/Free Full Text]

Drewnowski, A. & Shultz, J. M. (2001). Impact of aging on eating behaviors, food choices, nutrition, and health status. J Nutr Health Aging 5, 75–79.[Medline]

Duriez, P., Clermont, O., Bonacorsi, S., Bingen, E., Chaventre, A., Elion, J., Picard, R. & Denamur, E. (2001). Commensal Escherichia coli are phylogentically distributed among geographically distinct human populations. Microbiology 147, 1671–1676.[Medline]

Freter, R., Brickner, S. & Temme, S. (1986). An understanding of colonization resistance of the mammalian large intestine requires mathematical analysis. Microecol Ther 16, 147–155.

Gordon, D. M. & Cowling, A. (2003). The distribution and genetic structure of Esherichia coli in Australian vertebrates: host and geographic effects. Microbiology 149, 3575–3586.[CrossRef][Medline]

Gordon, D. M., Bauer, S. & Johnson, J. R. (2002). The genetic structure of Escherichia coli populations in primary and secondary habitats. Microbiology 148, 1513–1522.[Medline]

Graft, J., Brinch, K. & Madsen, J. L. (2001). Gastrointestinal mean transit times in young and middle-aged healthy subjects. Clin Physiol 21, 253–259.[CrossRef][Medline]

Hartl, D. L. & Dykhuizen, D. E. (1984). The population genetics of Escherichia coli. Annu Rev Genet 18, 31–68.[CrossRef][Medline]

Herzer, P. J., Inouye, S., Inouye, M. & Whittam, T. S. (1990). Phylogenetic distribution of branched RNA-linked multicopy single-stranded DNA among natural isolates of Escherichia coli. J Bacteriol 172, 6175–6181.[Medline]

Hounnou, G., Destrieux, C., Desme, J., Bertrand, P. & Velut, S. (2002). Anatomical study of the length of the human intestine. Surg Radiol Anat 24, 290–294.[CrossRef][Medline]

Hume, I. D. (1999). Marsupial Nutrition. Cambridge: Cambridge University Press.

Johnson, J. R. & Stell, A. L. (2000). Extended virulence genotypes of Escherichia coli strains from patients with urosepsis in relation to phylogeny and host compromise. J Infect Dis 181, 261–272.[CrossRef][Medline]

Johnson, J. R., Delavari, P., Kuskowski, M. & Stell, A. L. (2001). Phylogenetic distribution of extraintestinal virulence-associated traits in Escherichia coli. J Infect Dis 183, 78–88.[CrossRef][Medline]

Johnson, J. R., Murray, A. C., Gajewski, A., Sullivan, M., Snippes, P., Kuskowski, M. A. & Smith, K. E. (2003). Isolation and molecular characterization of nalidixic acid-resistant extraintestinal pathogenic Escherichia coli from retail chicken products. Antimicrob Agents Chemother 47, 2161–2168.[Abstract/Free Full Text]

Jones, D., Kojouharov, H. V., Le, D. & Smith, H. (2002). Bacterial wall attachment in a flow reactor. SIAM J Appl Math 62, 1728–1771.[CrossRef]

Manges, A. R., Johnson, J. R., Foxman, B., O'Bryan, T. T., Fullerton, K. E. & Riley, L. W. (2001). Widespread distribution of urinary tract infections caused by a multidrug-resistant Escherichia coli clonal group. N Engl J Med 345, 1007–1013.[Abstract/Free Full Text]

Mobley, H. L. T. & Warren, J. W. (1996). Molecular pathogenesis and clinical management. Urinary Tract Infect 3, 67–94.

Ochman, H. & Selander, R. K. (1984). Standard reference strains of Escherichia coli from natural populations. J Bacteriol 157, 690–693.[Medline]

Picard, B., Journet-Mancy, C., Picard-Pasquier, N. & Goullet, P. (1993). Genetic structure of B2 and B1 Escherichia coli strains responsible for extra-intestinal infection. J Gen Microbiol 139, 3079–3088.[Medline]

Picard, B., Sevali Garcia, J., Gouriou, S., Duriez, P., Brahimi, N., Bingen, E., Elion, J. & Denamur, E. (1999). The link between phylogeny and virulence in Escherichia coli extra-intestinal infection. Infect Immun 67, 546–553.[Abstract/Free Full Text]

Russell, R. M. (2000). The aging process as a modifier of metabolism. Am J Clin Nutr 72, 529S–532S.[Abstract/Free Full Text]

Schlager, T. A., Hendley, J. O., Bell, A. L. & Whittam, T. S. (2002). Clonal diversity of Escherichia coli colonizing stools and urinary tracts of young girls. Infect Immun 70, 1225–1229.[Abstract/Free Full Text]

Selander, R. K., Caugant, D. A. & Whittam, T. S. (1987). Genetic structure and variation in natural populations of Escherichia coli. In Escherichia coli and Salmonella typhimurium, Cellular and Molecular Biology, pp. 1625–1648. Edited by F. C. Neidhardt and others. Washington, DC: American Society for Microbiology.

Smith, H. L. & Waltman, P. (1995). The Theory of the Chemostat. Cambridge: Cambridge University Press.

Stemmons, E. D. & Smith, H. L. (2000). Competition in a chemostat with wall attachment. SIAM J Appl Math 61, 567–595.[CrossRef]

Topiwala, H. & Hamer, G. (1971). Effect of wall growth in steady state continuous culture. Biotechnol Bioeng 13, 919–922.

Wakimoto, P. & Block, G. (2001). Dietary intake, dietary patterns, and changes with age: an epidemiological perspective. J Gerontol A Biol Sci Med Sci 56, 65–80.[Abstract/Free Full Text]

Received 23 June 2004; revised 18 August 2004; accepted 20 September 2004.



This Article
Abstract
Full Text (PDF)
Alert me when this article is cited
Alert me if a correction is posted
Citation Map
Services
Email this article to a friend
Similar articles in this journal
Similar articles in PubMed
Alert me to new issues of the journal
Download to citation manager
Google Scholar
Articles by Gordon, D. M.
Articles by Collignon, P. J.
Articles citing this Article
PubMed
PubMed Citation
Articles by Gordon, D. M.
Articles by Collignon, P. J.
Agricola
Articles by Gordon, D. M.
Articles by Collignon, P. J.


HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
INT J SYST EVOL MICROBIOL MICROBIOLOGY J GEN VIROL
J MED MICROBIOL ALL SGM JOURNALS
Copyright © 2005 Society for General Microbiology.