Changes in the agr Locus Affect Enteritis Caused by Methicillin-Resistant Staphylococcus aureus

Abstract

We studied the characteristics of methicillin (meticillin)-resistant Staphylococcus aureus (MRSA) strains that caused enteritis. In a previous report, we demonstrated that both phenotypic and genotypic changes were associated with MRSA enteritis; and we hypothesized that the accessory gene regulator (agr), which is a global regulator of staphylococcal virulence and upregulates several exoproteins, is the key factor associated with the development of MRSA enteritis. In this study, we examined 12 MRSA isolates associated with enteritis from stool samples and 17 MRSA isolates not associated with enteritis that had the following characteristics: the strains associated with enteritis had the same genotype (genotype A), as detected by pulsed-field gel electrophoresis, or the strains were isolated from stools. The differences between strains that caused enteritis and those that did not cause enteritis strains were examined by quantitative reverse transcription-PCR to assess RNAII, agrA, RNAIII, and tst expression and by sequencing of the agr locus. The levels of expression of agrA, RNAIII, and tst were higher by the MRSA isolates associated with enteritis than by the MRSA isolates not associated with enteritis, whether or not they were of the same genotype. The levels of expression of RNAII by almost all the clinical isolates were similar. Sequencing of the agr locus showed that all MRSA isolates that caused enteritis have agr mutations, whereas the MRSA isolates that did not cause enteritis, with three exceptions, did not. Many of the isolates associated with enteritis had the same mutation, especially at the C-terminal end of agrA. These results suggest a trend in which mutations in the agr locus modify the expression of agrA and RNAIII and the production of toxin, all of which may increase the virulence and influence the occurrence of MRSA enteritis.

Severe postsurgical enteritis caused by methicillin (meticillin)-resistant Staphylococcus aureus (MRSA) (MRSA enteritis) was prevalent in Japan in the early 1990s. In the late 1990s, most patients with MRSA enteritis were successfully treated with vancomycin, and the number of cases with MRSA enteritis gradually decreased, but whenever it occurred, it was associated with a severe clinical course and, occasionally, death (12, 28).

Several studies have discussed the characteristics of isolates causing MRSA enteritis. Staphylococcal enterotoxin typing, coagulase typing, and toxic shock syndrome toxin 1 (TSST-1) titers are used as markers of MRSA enteritis (10, 30, 31). Many studies have described the phenotypes of isolates that cause enteritis, but there have been only a few reports on the genomic analysis of MRSA strains that cause enteritis (36). Our laboratory has studied the molecular epidemiology of MRSA enteritis strains by using pulsed-field gel electrophoresis (PFGE) and other methodologies for more than three decades. Our work has provided evidence that the disappearance of MRSA enteritis may have resulted from the decreased incidence of enteritis-causing clones (21). We also found that among isolates with the same genotype, as determined by PFGE, some caused enteritis, whereas others did not. This indicates that not only differences in the genotypes determined by PFGE but also differences in phenotypes are important factors associated with the development of MRSA enteritis.

The accessory gene regulator (agr) is a global regulator of staphylococcal virulence and other accessory gene functions and is important in coordinating the expression of many gene products required for invasive infection (5). It suppresses the expression of surface proteins and upregulates the expression of several exoproteins (16, 24, 25). The agr locus encodes two divergently transcribed transcripts, RNAII and RNAIII. RNAII is a polycistronic transcript that encodes agrB, agrD, agrC, and agrA. The protein encoded by agrD is a propeptide that matures into an autoinducing peptide. This process is carried out with the aid of the protein encoded by agrB. The protein encoded by agrC is the receptor of the autoinducing peptide and the sensor of the system, whereas the protein encoded by agrA is a response regulator. These two proteins form a two-component signal transduction pathway that upregulates the synthesis of RNAIII (17). RNAIII mediates the downregulation of the synthesis of cell-wall-associated proteins and the upregulation of exoproteins, including enterotoxins, TSST-1, and Panton-Valentine leukocidin (1, 25, 37).

agr promoter activation is proportional to the level of active protein encoded by agrA and the ability of the protein encoded by agrA to bind to the promoters (33). Because the transcription of RNAIII is increased in proportion to the concentration of the protein encoded by agrA, the level of active protein encoded by agrA influences the activation of RNAIII and the subsequent virulence of MRSA (11).

Polymorphism of the agr locus also results in specific differences in the sequences encoding agrB, agrD, and agrC; and this polymorphism causes bacterial interference (8). Variations in the protein encoded by agrA at the amino acid level may provide for variation in the activity of the protein encoded by agrA beyond that of the consensus activities of four interference groups, and small phenotypic differences encoded by different agrA alleles might be selectable into larger evolutionary differences (27).

MRSA enteritis has been associated with gastrectomy and the administration of antibiotic agents, and it has been postulated that the use of broad-spectrum antibiotics following gastrectomy allows the colonization and the subsequent overgrowth of MRSA isolates that may produce enteritis-causing toxins (29). Other triggers causing enteritis are not well understood. TSST-1 and staphylococcal enterotoxins are known to be bacterial superantigens that strongly activate T cells and that produce cytokines, such as interleukin-2, gamma interferon, and tumor necrosis factor.

Thus, macrophages play a critical role in a mouse model of postoperative MRSA enteritis (32), and interleukin-15 may be associated with the pathogenesis of postoperative MRSA enteritis (14). The production of TSST-1 and staphylococcal enterotoxin is thought to be associated with the occurrence of MRSA enteritis (10). The mechanism of their production is complicated, but it seems clear that the production of these enterotoxins and the virulence of the MRSA strains are controlled by the agr locus (23, 25, 37).

The aim of the study described here was to assess the differences in the levels of expression of agrA-associated proteins and to examine the role of the agr locus in the pathogenesis of MRSA enteritis. For this purpose, we compared the levels of expression of RNAII, RNAIII, agrA, and tst using quantitative reverse transcription-PCR (RT-PCR) and analyzed MRSA isolates causing enteritis and those not causing enteritis for the production of TSST-1. In addition, we sequenced the agr locus to study the relationship between mutations in the agr locus, the levels of expression of agrA, and MRSA enteritis.

MATERIALS AND METHODS

Bacterial strains and culture conditions.

A total of 198 MRSA strains were obtained from the surgical ward at Hiroshima University Hospital. Twelve strains were isolated from stools from 12 different patients with MRSA enteritis between 1990 and 1993, when MRSA enteritis was prevalent. A total of 186 strains were isolated from stools, sputum samples, drains, pus, and pleural fluid between 1998 and 2002 from patients who did not develop MRSA enteritis (nonenteritis strains). For this study, we examined the 12 strains associated with MRSA enteritis and 17 nonenteritis isolates that had the following characteristics: the strains associated with enteritis had the same genotype (genotype A), as determined by PFGE, or the strains were isolated from stools. A total of 29 isolates were used in this study. They had been stored at −80°C until use. Stock cultures were passaged on brain heart infusion (BHI) agar (Nissui, Tokyo, Japan), and selected single colonies were grown in BHI broth at 37°C for 24 h. The concentration of each culture was standardized to 1 × 10⁸ CFU/ml by using the absorbance correlated with standard colony counts. The bacterial cultures were stored at −80°C until use.

Assay for TSST-1.

All isolates were cultured as described above and were centrifuged at 3,000 × g for 30 min, and the supernatant was assayed for TSST-1 as described previously (21). A reverse passive latex agglutination (RPLA) test kit was used to detect TSST-1, according to the manufacturer's instructions (Denka Seiken Co., Tokyo, Japan). To determine the TSST-1 titers, supernatants were tested in twofold serial dilutions ranging from 1:2 to 1:1,024.

RNA isolation and cDNA synthesis.

Total RNA samples were prepared by using an RNeasy midikit (Qiagen GmbH, Hilden, Germany), according to the manufacturer's instructions. Extracted total RNA samples were treated with DNase I on RNeasy columns and were finally dissolved in 60 μl of RNase-free water. The extracted RNA samples were analyzed with an RNA 6000 Nano Chip kit (Agilent Technologies, Santa Clara, CA). cDNA was synthesized with a high-capacity cDNA Archive kit (Applied Biosystems, Foster City, CA), according to the manufacturer's instructions. The reaction mixtures were incubated first at 25°C for 10 min and then at 37°C for 120 min. The products were stored at −20°C until use.

Quantitative real-time PCR conditions.

Gene quantification was performed with an ABI Prism 7900 HT instrument (Applied Biosystems). cDNA was diluted 10-fold with water for use for real-time PCR. With the exception of agrA, quantitative PCR was performed with 1 μl of cDNA, 10 μl of Sybr green PCR master mixture (Applied Biosystems), and 5 μM each of the forward and the reverse primers (Table 1) (2, 3) in a final volume of 20 μl. For the quantification of agrA, PCR was performed with 2 μl of cDNA, 10 μl of TaqMan universal master mix (Applied Biosystems), 10 μM each of the forward and the reverse primers, and Universal ProbeLibrary probes (Roche Diagnostics, Mannheim, Germany) in a final volume of 20 μl. The probe and the primer specific for agrA (Table 1) were selected by using the system of the Universal ProbeLibrary (Roche Diagnostics). The thermal cycling programs consisted of 2 min at 50°C and 10 min at 95°C, followed by 40 cycles of 15 s at 95°C and 1 min at 60°C. The specificity of the PCR was verified by ethidium bromide staining on 2% agarose gels.

TABLE 1.

Primers used in this study

Primer purpose and primer	Sequence (5′-3′)	Product size (bp)	Reference or source
Primers for quantitative RT-PCR
RNAII
RNA II F	TATGAATAAATGCGCTGATGATATACCACG	73	2
RNA II R	TTTTAAAGTTGATAGACCTAAACCACGACC
RNAIII
RNA III F	TAAGAAAATACATAGCACTGAG	259	3
RNA III R	GAAGGAGTGTTTCAATGG
agrA
agrA F	AAAGTTGCAGCGATGGATTT	96	This study
agrA R	AAACACAGATGGCAGCAT
tst
tst F	AAACACAGATGGCAGCAT	129	This study
tst R	GTTCCTTCGCTAGTATGT
16S rRNA gene
16S F	CGGTCCAGACTCCTACGGGAGGCAGCA	455	2
16S R	GCGTGGACTACCAGGGTATCTAATCC
gmk
gmk F	TCGTTTTATCAGGACCATCTGGAGTAGGTA	153	2
gmk F	CATCTTTAATTAAAGCTTCAAACGCATCCC
Primers for sequencing
agr-Nested F	TGCCACGTATCTTCAAA	1,421	16
agr-Nested R	ATAATCATGACGGAACTT
agrB1	TATGCTCCTGCAGCAACTAA	1,072	32
agrC2	CTTGCGCATTTCGTTGTTGA
agrB2	GATGAACCTAAAGTACCCGC		This study
agrA F	ATTAACAACTAGCCATAAGG	836	31
agrA R	GCCAGCTATACAGTGCATTTGC		This study

Quantitative real-time PCR data analysis.

Fluorescence emission was detected with an ABI Prism 7900 sequence detection system (Applied Biosystems) and was analyzed with Sequence Detector software (version 1.7; Applied Biosystems). The levels of expression of the target genes in the RNA samples extracted from the different strains were normalized on the basis of internal standards (16S rRNA and the guanylate kinase gene [gmk]). Standard curves for the sequences were generated by using 10-fold serial dilutions of the specific standard, strain N315. A negative control (distilled water) and a standard dilution that permitted gene quantification with the supplied software, according to the manufacturer's instructions, were included in each run. For each RNA sample, the measurement of gene expression with the entire experiment was performed twice, and the results for biological replicates were compared. The mean of these values was used for further analysis.

Bacterial DNA lysates and amplification of the agr locus.

Genomic DNA was extracted by using a lysis buffer and 20 U lysostaphin (Wako Pure Chemical Industries, Ltd., Osaka, Japan). After incubation at 37°C for 30 min, 0.1% sodium dodecyl sulfate and 0.3 μg of proteinase K were added and the samples were then incubated at 56°C for 60 min. Proteinase K was denatured by heating it at 95°C for 10 min. Samples were extracted with an equal volume of a mixture of phenol-chloroform-isoamyl alcohol (25:24:1), and the aqueous phase was diluted with distilled water. The DNA template was quantified with a NanoDrop ND-1000 spectrophotometer (Thermo Fishers Scientific, Wilmington, DE) before use in the PCR.

In order to amplify the variable region of the agr gene by PCR, we used a nested PCR (18). The nucleotide coordinates of the primers were derived from the sequence of the agr locus of S. aureus agr group II strain N315 (GenBank accession number AP003135). The first primer pair was 5′-ACCAGTTTGCCACGTATCTCA-3′ and 5′-AACCACGACCTTCACCTTTAGTAG-3′. The nested primer pair was 5′-TGCCACGTATCTTCAAA-3′ and 5′-ATAATCATGACGGAACTT-3′. The first PCR was performed with 50-μl volumes containing 1 μl of genomic DNA, KOD-Plus enzyme solution (Toyobo, Osaka, Japan), and 50 pmol each of the forward and the reverse primers. For the nested PCR amplification, 1 μl of the first PCR product was added as the template to the PCR mixture containing the same components described above, except that the nested primers were used instead of the first set of primers. Thermal cycling was performed in a PC-800 programmed temperature control system (Astec, Fukuoka, Japan) and consisted of 30 cycles of denaturation (94°C, 15 s), annealing (for the first PCR, 59°C for 30 s; for the nested PCR, 52°C for 30 s), and extension (68°C, 90 s). The PCR product was verified by ethidium bromide staining using a 1.15% agarose gel containing Synergel powder (Diversified Biotech, Boston, MA).

Cloning and sequence analysis of agr locus variants.

The products of the nested PCR and the agrA products were electrophoresed, and the DNA bands were excised from the agarose gel. The PCR products in the bands were purified with a Qiaex II gel extraction kit (Qiagen, Hilden, Germany). The purified PCR products were cloned into competent Escherichia coli JM109 cells with the pGEM-T vector system (Promega, Madison, WI). The recombinants were screened by PCR with primers T7 and SP6 to check the sizes of the inserts. The plasmid DNA was purified with the reagents provided with the QIAprep spin miniprep kit (Qiagen, Valencia, CA). The sequencing reactions were carried out with a BigDye Terminator (version 1.1) cycle sequencing kit (Applied Biosystems). The sequencing reaction mixture contained 8 μl of BigDye Terminator ready reaction mix, 3.2 pmol of sequencing primer (Table 1), and 8 μl of purified PCR product in a total volume of 20 μl. The PCR thermal cycling program was performed according to the manufacturer's instructions. The sequencing products were purified with CentriSep spin columns (Princeton, Adelphia, NJ) and were then prepared to be run on an ABI 3100 Avant genetic analyzer (Applied Biosystems), according to the manufacturer's instructions.

Statistical analysis.

The statistical significance of pairwise differences in the levels of expression of RNAII, agrA, RNAIII, tst, and TSST-1 was evaluated by the Mann-Whitney test; and the significance for isolates of the same genotype was evaluated by Student's t test. The frequency of an agr locus mutation in enteritis and nonenteritis strains was statistically evaluated by the Fisher exact test. Data were considered significant when P values were <0.05.

RESULTS

Quantitative RT-PCR analysis of RNAII, agrA, RNAIII, and tst. (i) RNAII expression.

All of the isolates were positive for RNAII expression, and there was no statistically significant difference in the levels of expression of RNAII between the enteritis and the nonenteritis isolates (Fig. 1). In addition to 16S rRNA, we also used gmk as an internal standard. The RNAII expression levels evaluated by using gmk as the internal standard were also not statistically significantly different (data not shown).

FIG. 1.

Levels of expression of RNAII (A), agrA (B), RNAIII (C), and tst (D) by enteritis isolates compared to those by nonenteritis isolates, as determined by real-time RT-PCR with an ABI Prism 7900 sequence detection system. Transcripts were quantified with reference to the level of transcription of 16S rRNA. The results are the means of two experiments and are presented as means ± standard errors of the means. *, not significant; **, P < 0.01; ***, P < 0.05.

(ii) agrA expression. A total of 27 isolates (93.1%) were positive for agrA mRNA expression. All enteritis isolates were positive for agrA expression, and there was a significant difference in the levels of expression of agrA between the enteritis and the nonenteritis isolates (Fig. 1). The lowest level of expression of agrA among the enteritis isolates was 4.1 (MRSA strain 3895), but only one nonenteritis isolate had a level of agrA expression of greater than 4.1. The agrA-negative isolates (n = 2) were negative for tst expression and showed no production of TSST-1.

(iii) RNAIII expression. A total of 28 isolates (96.6%) were positive for RNAIII expression. All of the enteritis isolates were RNAIII positive, and there was a significant difference in the levels of expression of RNAIII between the enteritis isolates and the nonenteritis isolates (Fig. 1). Among the nonenteritis isolates, only one isolate (MRSA isolate 0015) expressed RNAIII at a high level. However, this isolate was negative for agrA and tst expression and showed no production of TSST-1.

tst expression. A total of 26 isolates (89.7%) were positive for tst expression. The level of expression of tst was statistically significantly different between the enteritis isolates and the nonenteritis isolates (Fig. 1). Among the enteritis isolates, only one isolate (MRSA isolate 3992) was negative for tst expression. Among the nonenteritis isolates, two isolates were negative for tst expression, and these isolates were also negative for agrA expression.

Sequences and agr grouping.

We sequenced the most variable region of agrBDC and agrA for all 29 clinical isolates. To confirm the presence of mutations, we sequenced the agr locus and compared it with the published sequence for strain N315 (13). Among the enteritis isolates, all isolates had mutations in either one of the two genes, whereas only 3 of 17 (18%) nonenteritis isolates had mutations (χ² = 16.1; P < 0.001). The mutations in agrA detected among the enteritis isolates are summarized as follows: there was a run of seven adenines at the C-terminal end, but many of the enteritis isolates had nine adenines, owing to the substitution of an A for a T; MRSA strain 3980 demonstrated a point mutation that converted an amino acid from a serine to a tyrosine; and MRSA strain 3992 demonstrated a point mutation that converted an amino acid from a lysine to an arginine. The mutations in agrBDC detected among the enteritis isolates are summarized as follows: strain MRSA 3992 belonged to agr group I, which means that the sequence of its agr locus was very different from the sequence of the agr locus in strain N315; MRSA strain 3895 demonstrated a point mutation in agrD that converted an amino acid of the agrD protein from a phenylalanine to a leucine; MRSA strain 3936 demonstrated a frameshift mutation in agrC; MRSA strains 3980, 5070, and 5083 demonstrated mutations in agrB; MRSA strains 3980 and 5083 demonstrated frameshift mutations in agrB; MRSA strain 5070 demonstrated a point mutation that converted an amino acid of the protein encoded by agrB from a glycine to an arginine; MRSA strain 0641, the only nonenteritis isolate with mutations in agr, demonstrated many point mutations in agrB and agrC, with 4 amino acid changes in the protein encoded by agr. The characteristics of these mutations are shown in Fig. 2.

FIG. 2.

Partial sequences of agrBDC and agrA in strain N315 and enteritis isolates with a mutation in the agr locus. The amino acid sequence of strain N315 (GenBank accession number AP003135) is shown on a shaded background, and mutations are on a white background.

The agr groups of the isolates were examined, and the published sequence of N315 revealed that it belongs to agr group II. The agr sequences of almost all isolates examined in this study were identical to the agr sequence of N315 and to a previously published agr group II DNA sequence. Only two isolates (MRSA strain 3992 and one nonenteritis isolate) had sequences concordant with an agr group I DNA sequence (GenBank accession number AJ617708) (4).

The characteristics of all strains are summarized in Table 2. All enteritis isolates except strain 3992 showed high levels of tst expression and TSST-1 production, resulting in a significant difference in the levels of production of TSST-1 between enteritis and nonenteritis isolates (P = 0.037). Strain 3992, which belonged to agr group I, had high levels of expression of agrA and RNAIII, but it was negative for tst expression and TSST-1 production. All enteritis isolates except strains 3895 and 3992 had equivalent, high levels of agrA and tst expression; but the levels of RNAIII expression varied, and the levels of RNAIII expression correlated with TSST-1 production.

TABLE 2.

Characteristics of all clinical isolates (enteritis and nonenteritis)

Isolate	Sample	Enteritis isolate	PFGE type	agr group	agrBDC mutation	agrA mutation	Quantitative RT-PCRa				TSST-1 titer by RPLAb
							RNAII	agrA	RNAIII	tst
3895	Stool	+	A	II	+	−	9.6	4.1	113.9	2.1	1024
3936	Stool	+	A	II	+	−	4.3	5.4	13.4	4.6	128
3980	Stool	+	A	II	+	+	4.5	8.8	26.6	3.7	256
3992	Stool	+	B	I	+	+	27.2	89.4	3,716.4
5070	Stool	+	C	II	+	+ (9Ac)	8.4	12.0	27.4	3.6	64
5083	Stool	+	D	II	+	+ (9A)	4.5	5.5	14.7	2.7	128
6211	Stool	+	C	II	−	+ (9A)	40.3	65.1	358.0	26.4	1,024
6251	Stool	+	C	II	−	+ (9A)	8.4	12.0	198.0	3.3	512
6463	Stool	+	C	II	−	+ (9A)	3.5	4.9	3.3	2.5	256
6873	Stool	+	C	II	−	+ (9A)	18.8	31.0	5,862.5	14.8	128
6947	Stool	+	C	II	−	+ (9A)	1.8	9.8	3,655.7	2.3	512
7014	Stool	+	C	II	−	+ (9A)	12.1	12.4	27.3	7.2	512
0009	Sputum	−	A	II	−	−	7.4	6.2	9.1	3.2	1,024
0323	Drain	−	A	II	−	+	3.1	0.8	0.9	0.6	64
0641	Drain	−	A	II	+	−	3.7	2.5	10.6	1.7	2
1017	Stool	−	A	II	−	−	3.7	1.8	14.0	1.3	32
1124	Sputum	−	A	II	−	+	2.9	0.9	5.0	0.9	16
1137	Pleural fluid	−	A	II	−	−	3.2	1.2	5.4	1.4	32
1372	Pus	−	A	II	−	−	18.0	1.1	19.3	1.7	256
0015	Stool	−	Other	I	−	+	6.4		56.3
0084	Stool	−	Other	II	−	−	2.2	3.1	1.5	0.4	512
0215	Stool	−	Other	II	−	−	2.3	2.3	2.9	0.6	256
0356	Stool	−	Other	II	−	−	2.4	2.1	2.3	0.8	128
0392	Stool	−	Other	II	−	−	2.0	2.4	3.3	0.4	32
0594	Stool	−	Other	II	−	−	4.2	3.2	2.4	1.8	64
0638	Stool	−	Other	II	−	−	6.2	0.7	9.3	0.8	256
0726	Stool	−	Other	II	−	−	11.6	2.7	5.0	4.2	64
0824	Stool	−	Other	II	−	−	3.5	3.5	5.4	1.1	128
0925	Stool	−	Other	II	−		178.0

^aQuantities are normalized on the basis of 16S rRNA as the internal standard.

^bThe TSST-1 titers were tested in serial dilutions ranging from 1:2 to 1:1,024. The data represent the inverse dilution.

^c9A, nine adenine residues owing to the substitution of A for T at the C-terminal end of agrA.

Characteristics of PFGE type A isolates (same genotype).

In our previous study (21), we detected enteritis isolates of the same genotypes up to 8 years after the first outbreak of MRSA enteritis. These isolates belonged to PFGE type A (the one type that caused MRSA enteritis) but did not cause enteritis. In this study there were a total of 10 PFGE type A isolates. These comprised three enteritis isolates and seven nonenteritis isolates.

In this study we examined the PFGE type A isolates for their agr group, mutations in the agr gene, and gene expression using quantitative RT-PCR analysis and the RPLA test for TSST-1 expression. The results are displayed in Table 2. Sequence analysis revealed that all PFGE type A isolates belong to agr group II. All PFGE type A enteritis isolates had agrBDC mutations, whereas none of the nonenteritis isolates except MRSA strain 0641 did. There was no nine-adenine mutation in PFGE type A enteritis isolates, even though that mutation was present in most enteritis isolates.

The levels of expression of RNAIII, agrA, and tst by the enteritis isolates tended to be higher than those by the nonenteritis isolates, even isolates of the same genotype.

RPLA tests of these isolates revealed that they all produced TSST-1. The enteritis isolates produced higher levels of TSST-1 than the nonenteritis isolates except for MRSA strains 0009 and 1372. MRSA strain 0641, which was the only nonenteritis isolate with agrBDC mutations, produced the lowest level of TSST-1.

DISCUSSION

In the present study, we have shown that the levels of expression of agrA, RNAIII, and tst in MRSA enteritis isolates were higher than those in MRSA nonenteritis isolates; and these differences in the levels of expression persisted, even though the isolates had the same genotype. Surprisingly, there was little difference in the levels of RNAII expression among almost all the clinical isolates. The most likely explanation is that changes during the process of encoding agrB, agrD, agrC, and agrA caused differences in agrA expression levels. Indeed, our study revealed that all of the isolates causing MRSA enteritis had mutations in agr, but among the nonenteritis isolates, with the exception of three isolates, there were no agr mutations. Many of enteritis isolates had same mutation, especially at the C-terminal end of agrA. This result shows a trend in which mutations in the agr locus influence the levels of expression of agrA and correlate with the occurrence of MRSA enteritis.

There is no proof of a direct relationship between an agr group type and a specific infectious disease caused by S. aureus in humans, but some investigators have reported on such relationships (6, 7, 19). Some methicillin-sensitive Staphylococcus aureus strains causing specific syndromes have been linked to certain agr types: staphylococcal scalded skin syndrome and generalized exfoliative syndromes have been associated with agr group IV (6). The strains causing endocarditis belong to agr groups I and II (7). Some investigators have reported that agr group II strains are more commonly identified among health care facility-associated MRSA strains (19). Our study results are compatible with those presented in that report, but almost all the clinical isolates in our study belonged to agr group II, so agr group II does not seem to be exclusively associated with MRSA enteritis.

Many studies have suggested that RNAIII is the agr-specific effector of exoprotein gene regulation in Staphylococcus aureus (20). The increase in the concentration of agrA leads to the transcription of RNAIII and the subsequent upregulation of RNAIII-mediated virulence responses (11). Thus, we speculate that increased levels of expression of agrA are associated with toxin production, causing MRSA enteritis. The results of the present study indicate that the levels of expression of agrA correlate with the levels of expression of RNAIII and TSST-1 production, which were associated with MRSA enteritis. This tendency was observed even among isolates of the same genotype.

In addition, there is convincing evidence that the agr locus demonstrates polymorphism, and variation in the agr locus is related to differences in the activation of strains (8, 15, 18, 22, 34). Therefore, polymorphism in the agr locus may result in specific differences in the agrB-, agrD-, and agrC-encoding sequences, which influence the downstream region of the agr locus and change the level of expression of agrA. In this study, the level of expression of agrA was significantly different in enteritis isolates from that in nonenteritis isolates, but there was little difference in the RNAII expression levels between the two types of isolates. All of the enteritis isolates also had mutations in the agr locus, whereas the nonenteritis isolates did not. In this study, we found that many of enteritis isolates had the same mutation in agrA, whereas the nonenteritis isolates did not. This mutation was nine adenines, owing to the substitution of an A for a T at the C-terminal end and the same Mu3 and Mu50 sequences. These isolates have high levels of expression of agrA, so it may be presumed that this mutation influences the expression of active agrA. However, the same mutation in agrA was not found in PFGE type A isolates, which are genotypically similar to each other, but all of enteritis isolates had agrBDC mutations, whereas only one nonenteritis isolate had the agrBDC mutations. These results show a trend in which mutations in agrBDC influence the levels of agrA expression. Thus, in PFGE type A isolates, the agrBDC mutations appeared to be play an important role as a cause of agrA overexpression. To explain this fact, we thought it was important that some operon was expressed as monocistronic and polycistronic mRNA (35). We hypothesized that agrA may have both monocistronic and polycistronic mRNA. A decrease in the level of polycistronic mRNA (e.g., a frameshift mutation in agrBDC) may have influenced monocistronic agrA mRNA expression. Further studies are needed to verify our hypothesis.

Several important studies have described the genotypes of enteritis isolates (9, 36); therefore, PFGE is a useful surveillance method for the identification of genotypically different isolates. However, virulent isolates are not usually associated with a particular PFGE type. Actually, in our study, some strains with different base sequences that appeared to influence the levels of TSST-1 expression were genetically the same isolate, according to the results of PFGE analysis. Therefore, our results provide evidence that sequencing of the agr locus and quantification of agrA expression levels provide additional information in surveys of MRSA enteritis, and it is possible to assume that combining that technology with PFGE will allow further discrimination of MRSA enteritis isolates.

There is convincing evidence that some environmental conditions, including glucose levels and changes in pH, affect agr expression and toxin production (26). Surgery on the upper gastrointestinal tract, especially cancer-related surgery, and the administration of histamine H₂ receptor antagonists and broad-spectrum antimicrobial agents are considered clinically significant causes of MRSA enteritis. These factors cause environmental changes in the gut; therefore, it is possible that these environmental changes in the gut accumulate and effect mutations in the agr locus, resulting in agr overexpression, which leads to MRSA enteritis (Fig. 3). However, to verify our hypothesis, further studies are needed to evaluate the association between mutations in the agr locus and these factors and the mechanisms that cause MRSA enteritis.

FIG. 3.

Integration of data from our study and other investigations. A hypothetical pathway relating MRSA enteritis to environmental changes in the gut is indicated.

In conclusion, the agr locus is associated with the development of MRSA enteritis. Moreover, mutations in the agr locus have a potential association with the induction of the overexpression of agrA and increased toxin production, which influence the occurrence of MRSA enteritis. Therefore, sequencing of the agr locus and quantification of agrA expression provide additional information in surveys of MRSA enteritis.