Previous Article | Next Article ![]()
Applied and Environmental Microbiology, January 2003, p. 18-23, Vol. 69, No. 1
0099-2240/03/$08.00+0 DOI: 10.1128/AEM.69.1.18-23.2003
Copyright © 2003, American Society for Microbiology. All Rights Reserved.
Centre National de Référence des Staphylocoques, INSERM E-0230,1 Clinical Pharmacology Unit, EA 643 & EZUS-APRET, IFR 62, Faculté de Médecine Laennec, Lyon, France2
Received 12 April 2002/ Accepted 30 September 2002
|
|
|---|
|
|
|---|
Longitudinal studies of healthy individuals have shown three patterns of S. aureus carriage: (i) noncarriage, (ii) intermittent carriage, and (iii) persistent carriage of the same or different strains (23, 24). The differences could be due to host factors and/or to antagonism between members of the nasal flora. Indeed, a lower incidence of S. aureus colonization is observed in individuals heavily colonized by Corynebacterium spp. (22), and interaction between these two species was confirmed by in vivo experiments showing that experimental colonization by Corynebacterium spp. inhibits colonization by S. aureus (22). Inconsistent results have been obtained with other species, including non-aureus staphylococci (18, 22).
Expression of cell wall-associated and extracellular proteins in staphylococci is controlled by the agr locus, which encodes a two-component signaling pathway whose activating ligand is a bacterial-density-sensing peptide (autoinducing peptide [AIP]) which is also encoded by agr (10). A polymorphism in the AIP amino acid sequence and in that of its corresponding receptor has been described in staphylococci (4, 7, 9). S. aureus strains can be divided into four major groups (designated agr-1Sa to agr-4Sa), such that within each group, all strains produce a peptide that can activate the agr response in the other members of the same group whereas autoinducing peptides are usually mutually inhibitory between members of different groups (7, 9). Functional agr loci are present in other staphylococcal species, including S. epidermidis (agr-1Se to agr-3Se) (3, 4, 26), which are different from each other and from agrSa. The agr-1Se AIP inhibits the activity of agr-1Sa to agr-3Sa but not agr-4Sa, while among S. aureus AIPs, only type 4 (weakly) inhibits agr-1Se activity (20). It has been proposed that agr-2Sa S. aureus strains hinder umbilical stump colonization by agr-1Sa strains (19). The biological mechanism of this interference is unknown but might be caused by molecular cross-interference between agr alleles.
The aim of the present investigation was to determine the qualitative and quantitative composition of the nasal flora of healthy individuals, focusing on S. aureus, coagulase-negative staphylococci, and corynebacteria, and to identify potential interactions between these bacteria. Staphylococcal isolates were analyzed at the species and agr allele level, and a mathematical model of bacterial nasal interference was constructed.
|
|
|---|
Estimation of the nasal vestibule flora.
The standard cotton swabbing technique was used to sample the nasal vestibule. Swabs were streaked on sheep blood agar and incubated at 37°C in an aerobic atmosphere for 48 h. Bacterial density was estimated by counting CFU in logarithmic graduations. The representative colonies were subcultured and identified using standard methods, as described below. Twenty randomly selected S. aureus-positive swabs were inoculated in brain heart broth (bioMérieux) and cultured for 24 h at 37°C. After centrifugation, pellets were harvested and stored at -20°C until used for DNA extraction.
Identification of isolates.
Staphylococcus species were identified on the basis of conventional phenotypic characteristics, namely, Gram staining, cell morphology and cell arrangement, colony morphology and pigmentation on P agar and Trypticase soy agar (bioMérieux) supplemented with horse blood, catalase activity, coagulase production in rabbit plasma (bioMérieux), and production of clumping factor (Pastorex Staph Plus; bioMérieux). For species identification of coagulase-negative staphylococci, we used individual tests (susceptibility to furazolindone [300 µg], bacitracin [0.02 U], desferrioxamine [250 µg], and novobiocin) and the ID32 Staph gallery (bioMérieux). Corynebacterium spp. were identified on the basis of colony morphology and pigmentation on Trypticase soy agar supplemented with horse blood and also on the basis of cell morphology and cell arrangement after Gram staining; they were not identified to the species level.
agr typing by multiplex PCR.
Genomic DNA was extracted from staphylococci grown on agar plates or in brain heart infusion broth (13) and used as an amplification template with primers (Table 1) designed from the agr-1Sa to agr-4Sa and agr-1Se to agr-3Se sequences (GenBank accession numbers X52543, AF001782, AF001783, AF288215, Z49220, AF346724, and AF346725, respectively) to amplify specific agr alleles. For multiplex PCR, two primer sets were prepared: one to amplify agrSa alleles and another to amplify agrSe alleles. Amplification was carried out under the following conditions: an initial 5-min denaturation step at 95°C followed by 25 stringent cycles (1 min of denaturation at 94°C, 1 min of annealing at 55°C, and 1 min of extension at 72°C) and a final extension step at 72°C for 10 min. The quality of the DNA extracts and the absence of PCR inhibitors were confirmed by amplification of gyrA (S. aureus) or the 16S-23S intergenic region of the ribosomal DNA operon (S. epidermidis) (13). PCR products were analyzed by electrophoresis through 0.8% agarose gels (Sigma). The following strains were used to control the specificity of PCR amplification: (i) S. aureus RN6390 (agr-1Sa), RN6923 (agr-2Sa), RN8462 (agr-3Sa), and A880740 (agr-4Sa) (7); (ii) S. epidermidis CCM2124 (agr-1Se), N910160 (agr-2Se), and N910191 (agr-3Se) (4).
|
View this table: [in a new window] |
TABLE 1. Nucleotide sequences of agr type-specific oligonucleotide primers used in this study, and anticipated sizes of PCR products
|
2 test. To explain the probability that S. aureus is present in terms of the number of colonies of the other species, standard linear regression does not apply. Hence, a multiple logistic regression model was used to analyze simultaneously the influence of Corynebacterium spp. and S. epidermidis agr alleles on the probability of the presence of S. aureus agr-1Sa or agr-2Sa. Age, sex, and school of origin were used as candidate covariates for adjustment. Those with a P value below 0.05 were finally retained in the final model. In this model, the colony counts were categorized into four groups for corynebacteria (<102 [reference group], 102 to 103, 103 to 104, and
104) and into two groups for agrSe-1 to agrSe-3 (<102 [reference group] and
102). Statistical analyses were done with software from SAS Inc. |
|
|---|
Determination of the agr type of staphylococcal isolates.
The agr type of all isolates belonging to the two most prevalent staphylococcal species, S. aureus and S. epidermidis, was determined by using specific multiplex PCR. To reduce possible bias of strain cloning, a minimum of 10 colonies of each morphological type were randomly selected for multiplex PCR. As shown in Fig. 1, strong specific signals of the expected sizes were obtained with the reference strains. All 65 S. aureus isolates fell into one of the four previously described agrSa groups (alleles). Only one agrSa allele was detected in each carrier: 34 isolates belonged to agr-1Sa, 19 belonged to agr-2Sa, 7 belonged to agr-3Sa, and 5 belonged to agr-4Sa. To verify that each carrier harbored only one agrSa allele, PCR was also performed on total DNA extracted from brain heart infusion broth cultures of samples from 20 of the S. aureus carriers. In all cases, only one agrSa allele (identical to one of the four previously identified alleles) was detected in each culture by multiplex PCR.
![]() View larger version (34K): [in a new window] |
FIG. 1. Multiplex PCR detection of S. aureus (A) and S. epidermidis (B) agr alleles. Lanes 1 to 4, PCR amplicons from primer set agrSa alleles using DNA from S. aureus RN6390 (agr-1Sa), RN6923 (agr-2Sa), RN8462 (agr-3Sa), and A880740 (agr-4Sa), respectively; lanes 5 to 7, PCR amplicons from set agrSe alleles using DNA from S. epidermidis CCM2124 (agr-1Se), N910160 (agr-2Se), and N910191 (agr-3Se), respectively.
|
Interactions between staphylococci and corynebacteria.
To detect possible interference between Corynebacterium and staphylococcal species in carriers, we examined the number of culture-positive samples for each staphylococcal species as a function of the number of Corynebacterium CFU. As shown in Table 2, within the range of 0 to 103 Corynebacterium CFU, the proportion of samples positive for S. aureus was quite similar (34 to 46%), whereas it fell significantly at higher Corynebacterium CFU levels (16 and 11% of samples were S. aureus positive at 104 and >104 Corynebacterium CFU, respectively; P = 0.08 and 0.001, respectively). The multiple-regression model (Table 3) confirmed that the probability of S. aureus isolation was reduced by about 3-fold when the Corynebacterium CFU level rose from <102 to 104 (odds ratio = 0.36; P = 0.16) and by more than 10-fold when it rose to
105 (odds ratio = 0.08; P = 0.0001). In contrast, the rate of non-aureus staphylococcal colonization (including S. epidermidis) was not affected by the degree of Corynebacterium colonization (Table 2). These results confirm those of previous studies indicating that Corynebacterium spp. specifically inhibit colonization by S. aureus but not by non-aureus staphylococci (22).
|
View this table: [in a new window] |
TABLE 2. Incidence of nasal carriage of S. aureus, non-aureus staphylococci, and S. epidermidis according to the degree of colonization by Corynebacterium spp.
|
|
View this table: [in a new window] |
TABLE 3. Influence of the number of Corynebacterium or S. epidermidis colonies on the probability of S. aureus isolation
|
Relationships between staphylococcal species.
To determine if non-aureus staphylococci, especially S. epidermidis, inhibited S. aureus nasal colonization, we examined the number of S. aureus culture-positive samples as a function of the non-aureus staphylococci (including S. epidermidis) and S. epidermidis sensu stricto CFU values. The number of S. aureus-positive samples fell markedly as the non-aureus staphylococcal (not shown) and S. epidermidis CFU values increased (from 78 to 23%, P = 0.001; and from 66 to 30%, P = 0.004, respectively) (Table 4), in particular in the volunter with <103 UFC of Corynebacterium spp. (from 88 to 31%, P = 0.001). Multiple-regression analysis (Table 5) showed that the probability of S. aureus isolation fell by about 6- to 10-fold when the S. epidermidis CFU value rose from <102-103 to 104 and
105 (odds ratios, 0.1 to 0.16; P, 0.0001 to 0.005). This model also showed that the presence of S. epidermidis and Corynebacterium spp. was independently associated with the presence or absence of S. aureus (P = 0.0001). Colonization by >103 non-aureus staphylococci, and especially S. epidermidis, was clearly protective against S. aureus colonization. The rate of non-aureus staphylococcal and S. epidermidis colonization fell from 100 to 81% and from 99 to 81% as the number of S. aureus CFU increased (not shown); however, these results were not statistically significant, probably owing to the small number of volunteers strongly colonized by S. aureus. Lastly, some rarely detected staphylococcal species such as S. warneri, S. lugdunensis, S. haemolyticus, and S. cohnii (but not S. capitis or S. hominis) were never isolated simultaneously with S. aureus.
|
View this table: [in a new window] |
TABLE 4. Incidence of nasal carriage of S. aureus according to the degree of colonization by S. epidermidis
|
|
View this table: [in a new window] |
TABLE 5. Influence of the number of colonies of corynebacteria on the probability of isolating S. aureus strains bearing agr-1Sa and agr-2Sa
|
102 versus <102 (odds ratio, 0.39; P = 0.004), while no statistical differences were observed with other agrSe alleles (P > 0.35 or higher). The probability of isolating agr-1Sa S. aureus was reduced by a factor of about 10 when the number of Corynebacterium CFU was
104 versus <102 (odds ratio, 0.10; P = 0.004), as observed above. In contrast, the probability of isolating agr-2Sa S. aureus was reduced by a factor of about 3 when the CFU count of agr-1Se, agr-2Se, or agr-3Se S. epidermidis was
102 versus <102. The results were statistically significant for agr-1Se and agr-3Se but not for agr-2Se, probably because of the small number of isolates (odds ratio, 0.31, P = 0.06; odds ratio, 0.37, P = 0.12; and odds ratio, 0.33; P = 0.05, respectively). Surprisingly, the probability that the agr-2Sa S. aureus colonization rate fell as the Corynebacterium colonization rate rose was not statistically significant, despite the large number of isolates (odds ratio minimal 0.33, P = 0.17) (Table 5; Fig. 2). agr-3Se S. epidermidis and Corynebacterium isolation was independently associated with the presence of agr-1Sa S. aureus (P = 0.027 and P = 0.0002, respectively), while agr-1Se and agr-2Se S. epidermidis isolation was independently associated with the presence of agr-2Sa S. aureus (P = 0.046 and P = 0.044, respectively).
![]() View larger version (16K): [in a new window] |
FIG. 2. Model of agr-dependent staphylococcal interference. Based on semiquantitative analysis of the aerobic nasal flora, a multiple logistic regression model was used to analyze simultaneously the effect of S. aureus agrSa alleles, Corynebacterium spp., and S. epidermidis agrSe alleles on S. aureus colonization. OR, odds ratio; threshold of 102 CFU for S. epidermidis and 104 CFU for Corynebacterium spp.
|
|
View this table: [in a new window] |
TABLE 6. Influence of the number of colonies of agr-1Se to agr-3Se allele-bearing S. epidermidis on the probability of isolating S. aureus strains bearing agr-1Sa and agr-2Sa
|
|
|
|---|
Most of the S. aureus and S. epidermidis isolates were agr typed by PCR. Only one agrSa allele was detected in each individual's nasal flora, as previously described (25). Moreover, multiple logistic regression analysis with the two most prevalent agrSa alleles (agr-1Sa and agr-2Sa) and the three agrSe alleles confidently showed agr-specific interaction in the nasal vestibule (Fig. 2). Our initial hypothesis was that staphylococcal interaction in the nose would reflect the heteroallelic inhibitory activity previously detected in vitro. In these in vitro experiments, it was shown that the autoinducing peptides produced by the different groups are usually mutually inhibitory on the expression of RNAIII, the effector of the agr system, and subsequently on the expression of exoproteins and toxins of S. aureus. Hence, agrSa alleles, with the exception of agr-1Sa and agr-4Sa, are all mutually inhibitory in vitro on the expression of RNAIII (7, 9), and agr-1Se is also inhibitory for agrSa but not for agr-4Sa (20). If these in vitro interferences were relevant in nasal colonization, we would have observed an inhibitory effect of agr-1Se on agr-1Sa and agr-2Sa and frequent simultaneous detection of agr-4Sa and agr-1Sa or agr-4Sa and agr-1Se. In fact, we found no correlation between in vitro and in vivo data on agr alleles. Our results were not in support of a predominant role of the agr system in staphylococcal interaction in the human nasal vestibule, as previously observed in several staphylococcal diseases (8). Jarraud et al. (8) examined the possible relationship between agr groups and human S. aureus disease by studying 198 S. aureus strains isolated from patients with suppurative infections and acute toxemia. A relationship between the genetic background, agr group, and disease type was observed in most cases of toxin-mediated disease and in several suppurative infections such as infective endocarditis. Jarraud et al concluded that the agr type had no direct responsibility for disease initiation and speculated that the preferential association between certain agr alleles, certain toxin genes, and a particular genetic background may reflect an ancient evolutionary division of S. aureus in terms of this fundamental biology of the species (8).
Hence, the apparent agr-dependent in vivo interactions observed in the present study may have been due to other mechanisms. One possibility is the synthesis of antagonists such as bacteriocins, bacteriolytic enzymes, hydrogen peroxide, lactic acids or fatty acids, and ammonia (2). Bactericidal exoproteins have been already detected in staphylococcal and Corynebacterium spp. Among the staphylococci, previous studies have identified S. aureus bactericiocin against some Corynebacterium species (17, 21) and S. epidermidis bacteriocin against S. aureus (6, 16). However, no bacteriocin-like activity produced by corynebacteria against S. aureus or S. epidermidis or by one S. aureus strain against another has been described. Another possible mechanism of bacterial interference is competition among Corynebacterium spp., S. aureus, and S. epidermidis for specific attachment to epithelial cells (1). Uehara et al. suggested that binding competition might involve the carbohydrate portion of the human nasal mucin support and showed that Corynebacterium spp. had higher affinity for mucus than did S. aureus and that S. aureus had higher affinity than did S. epidermidis (22). In our study, Corynebacterium spp. inhibited S. aureus colonization, but only strains harboring the agr-1Sa allele (Table 5). We did not find other concordant results between those reported by Uehara and our in vivo data. The physiological role of mucus is to bind and remove bacteria, not to promote bacterial adhesion to the epithelium. In adhesion experiments with human airway epithelial cells, Mongodin et al. recently showed that S. aureus did not adhere in vivo to intact mucus-producing airway epithelium but did adhere to the basolateral plasma membrane of columnar cells, to basal cells, and to the basement membrane (15).
Finally, while we did not identify the precise mechanism of the observed bacterial interference in nasal colonization, our mathematical analysis of ecological data produced a working model of bacterial interactions in the nasal vestibule. Our subsequent experiments aimed at determining these mechanisms will focus on specific interactions identified by the model. Importantly, our results show that the likelihood of nasal colonization by S. aureus in healthy subjects varies with the composition of the local flora. The relevance of our model to patients with underlying diseases remains to be tested, but it is noteworthy that most methicillin-resistant S. aureus (MRSA) strains harbor agr-1Sa (reference 25 and unpublished personal data) and that colonization by agr-1Sa strains was specifically associated with a low rate of colonization by Corynebacterium spp. and agr-3Se S. epidermidis. Indeed, our model predicted that the probability of agr-1Sa S. aureus (probably MRSA) colonization in such cases would be increased by a factor of 33. Larger cross-sectional and longitudinal studies are required to understand how one S. aureus strain can displace another, especially in the case of MRSA colonization.
This work was supported by a grant from Recherche en Microbiologie et Maladies Infectieuses et Parasitaires of the Ministère de l'Education Nationale.
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»