The plo gene, encoding the Arcanobacterium pyogenes cholesterol-dependent cytolysin, pyolysin (PLO), was localized to a 2.7-kb genomic islet of reduced %G+C content and alternate codon usage frequency. This islet, conserved among isolates from diverse hosts and geographical locations, separated the housekeeping genes smc and ftsY, which are found adjacent in many prokaryotes. The ftsY and ffh genes, located downstream of the plo islet, encode components of the signal recognition particle. Mutational analysis suggested that these genes were essential for viability in A. pyogenes. The A. pyogenes ffh gene was unable to complement a conditional ffh mutant of Escherichia coli and its overexpression was toxic in E. coli. Mutagenesis of the islet-encoded orf121 did not affect plo expression, indicating that it may not be involved directly in the regulation of plo expression. Regardless, the presence of the plo gene as part of a genomic islet inserted between genes essential for normal growth may provide selective pressure for the retention of this important virulence factor.
Arcanobacterium pyogenes is a Gram-positive, facultative anaerobe that can exist in the host both as a commensal of the mucous membranes and as an opportunistic pathogen, invading after physical or microbial induced trauma . As an opportunistic pathogen, A. pyogenes has been identified in a number of animal species, including cattle, where it has been isolated from liver abscesses  and mastitis , which have a significant economic impact on the beef and dairy industries.
A. pyogenes secretes a cholesterol-dependent cytolysin (CDC), pyolysin (PLO) , which is both a host-protective antigen and virulence factor . An A. pyogenes plo mutant was attenuated for virulence in a mouse model, and immunization with recombinant PLO protected mice from infection with wild-type A. pyogenes. PLO is cytolytic for host cells including erythrocytes, murine peritoneal macrophages, and bovine and ovine polymorphonuclear leukocytes , through the formation of large pores in the host cell membrane, characteristic of the CDC family of toxins .
The genetic regulation of CDC expression is not well understood, although some are regulated by genes closely associated with the CDC structural gene. The Listeria monocytogenes CDC, listeriolysin O (LLO) is encoded, along with its positive regulator, PrfA, on the pathogenicity island LIPI-1 . Similarly, Clostridium perfringens perfringolysin O (PFO) is encoded immediately downstream of its putative regulator, PfoR . In an effort to identify factors involved in the expression of PLO, the nucleotide sequence of regions flanking the A. pyogenes plo gene was determined. Here we report the localization of plo and an upstream open reading frame (ORF), orf121, on a genomic islet flanked by conserved housekeeping genes.
Materials and methods
Bacterial strains and growth conditions
A. pyogenes strains were grown on brain heart infusion (BHI) agar supplemented with 5% bovine blood at 37°C with 5% CO2 for 48–72 h, or in BHI supplemented with 10% bovine calf serum with shaking at 37°C. Escherichia coli was grown on Luria–Bertani (LB) agar at 37°C, or in LB broth with shaking at 37°C. The E. coli ffh conditional mutant WAM113 , in which ffh was under the control of the araB promoter, was grown on LB agar with 0.2% arabinose, for maintenance of wild-type levels of ffh expression. Media were supplemented with antibiotics at the following concentrations: for A. pyogenes: erythromycin (Erm) 15 µg ml−1, kanamycin (Kan) 30 µg ml−1, and streptomycin (Str) 200 µg ml−1; and for E. coli: ampicillin (Amp) 100 µg ml−1, chloramphenicol (Cam) 30 µg ml−1, Erm 200 µg ml−1, Kan 50 µg ml−1, Str 50 µg ml−1, and tetracycline 10 µg ml−1.
Recombinant DNA techniques
A. pyogenes genomic DNA was isolated by the method of Pospiech and Neumann ; plasmid DNA isolation and electroporation protocols for A. pyogenes were performed as previously described . E. coli plasmid DNA preparations and transformations, agarose gel electrophoresis, restriction endonuclease digestions, and ligation reactions were performed as previously described . Polymerase chain reactions (PCRs) were performed on template derived from isolated colonies of A. pyogenes using Taq DNA polymerase (Fisher Scientific) and its supplied buffer in a standard protocol with an initial hot start for 5 min at 94°C, followed by 35 cycles consisting of 1 min at 94°C (DNA denaturation), 30 s at 55°C (primer annealing) and 1 min kb−1 at 72°C (DNA synthesis).
Nucleotide sequencing and computer analysis
The plo gene was originally identified on the cosmid ApH1 and sequenced by Billington et al. . Sequencing of the plo flanking regions was performed on subclones of ApH1 using either vector or insert specific primers (Sigma-Genosys). Sequencing was performed on both strands on a 377A DNA sequencer (Applied Biosystems Inc.) at the University of Arizona's Genomic Analysis Technology Core Facility. Sequences were assembled using Sequencher™ 3.1 (GeneCodes), and database searches were performed using the BlastX and BlastP programs . Further analyses, including the construction of %G+C and codon usage tables, analysis of patterns and possible structural motifs, and multiple sequence alignments were conducted using the GCG suite of programs (Accelyrs).
The nucleotide sequence data reported in the paper have been deposited in the GenBank nucleotide sequence database under the accession number U84782.2.
Allelic exchange mutagenesis
To construct an A. pyogenes strain carrying a mutation in orf121, a 2.4-kb PCR product containing orf121, amplified using primers SmcF1 (5′-GACGAGGTCGAGGCACAC-3′) and FtsYR3 (5′-GCAAGTTCCTCGTGTCCG-3′), was cloned into the vector pHSS19  to generate pJGS436. An Erm resistance cassette derived from pNG2  was cloned into the unique KpnI site within orf121, which was previously blunt-ended with T4 DNA polymerase (New England Biolabs), to construct pJGS438, which was then introduced into BBR1 by electroporation. In strain JGS534, the wild-type copy of orf121 was replaced by homologous recombination with the insertionally inactivated copy, as confirmed by PCR. To assess the effect of overexpression of orf121, a 1.3-kb orf121-containing fragment was cloned directly downstream of Plac in pJGS181, a Str-resistant derivative of the multicopy shuttle vector pEP2, to generate pJGS440.
The construction of A. pyogenes ftsY and ffh mutants was carried out using both allelic exchange and vector integration strategies. A 1.4-kb PCR product containing ftsY was amplified using primers FtsYF1 (5′-GGTGCGCTCTGATCGCGAAGGC-3′) and FtsYR1 (5′-AAGCTGATTGACGGCGTTGGGG-3′), cloned into pHSS19 and disrupted at the unique NruI site with a 2.0-kb Str resistance cassette derived from pKRP13 , to construct pJGS222. Similarly, a 1.7-kb PCR product containing ffh was amplified using primers FfhF1 (5′-CTACCTGAGGGCATACATGTTTA-3′) and FfhR1 (5′-CGTCAGTGCCCGCAAAGACCC-3′), cloned into pHSS19 and disrupted at the unique EcoRV site with the pKRP13 2.0-kb Str resistance cassette, to construct pJGS225. Each of these constructs was introduced independently into A. pyogenes BBR1 by electroporation, selecting for Str resistance, indicating allelic exchange. Alternatively, internal fragments of ftsY or ffh amplified with primers FtsYF4 (5′-GGACGTCTGGCCGCTTCTGGTG-3′) and FtsYR5 (5′-CCCACGTGGAGAGCTGATCGGC-3′), or FfhF2 (5′-AACGCCGTCAATCAGCTTCAGG-3′) and FfhR2 (5′-CTCCTCGCTCCAGGTCTTTTCG-3′) were cloned into pHSS19 to construct pJGS313 and pJGS315, respectively. These plasmids were introduced into BBR1, selecting for the vector-encoded Kan resistance and, therefore, disruption of either the wild-type ftsY or ffh, resulting in two incomplete copies of the gene.
Hemolytic assays were performed on A. pyogenes culture supernatant essentially as described . Hemoglobin release was measured at A410, with one hemolytic unit (HU) representing the amount of hemolysin required to release 50% of the hemoglobin from 200 µl of 0.25% ovine erythrocytes in 1 h.
Complementation of an E. coli ffh conditional mutant
A 1.6-kb NotI fragment carrying the A. pyogenes ffh gene was blunt-ended with DNA polymerase I Klenow fragment (Promega) and cloned into SmaI-digested pDK7  to construct pJGS369. In this vector, ffh is under the control of the isopropyl-β-d-thiogalactopyranoside (IPTG)-inducible Ptac. pJGS369 and pDK7 were introduced into WAM100, and its conditional ffh mutant WAM113 by electroporation. The resulting strains were grown on LB agar plates supplemented with 1 µM–10 mM IPTG to induce expression from the ffh gene on pJGS369, in the presence of decreasing arabinose (0.2–0.002%) to limit transcription from the endogenous ffh gene in WAM113. In a separate experiment, E. coli DH5α derivatives carrying pDK7 or pJGS369 were inoculated into LB broth with or without 100 µM IPTG to determine the effect of the A. pyogenes ffh gene on the growth of E. coli.
Results and discussion
The A. pyogenes plo gene and orf121 are located on a genomic islet
In an attempt to identify genes physically linked to plo involved in the expression of PLO activity, the nucleotide sequence 2160 bp upstream and 4121 bp downstream of plo was determined. The organization of ORFs in the plo region is shown in Fig. 1. Identified in the sequence upstream of plo was one complete ORF, orf121, which encoded a 13.4-kDa protein with no homologs in the GenBank database, and the 3′-end of a partial ORF, which encoded a protein with similarity to Smc proteins from a number of bacterial species, and most similarity to Streptomyces coelicolor Smc (67.9% identity, 78.5% similarity across the amino acids examined) (GenBank accession number T35661). Smc proteins are responsible for the structural maintenance of chromosomes and are essential for chromosome partitioning in Bacillus subtilis and E. coli. Downstream of the plo transcriptional terminator  were two ORFs which encoded proteins with similarity to the protein components of the signal recognition particle (SRP), FtsY and Ffh, from a number of bacterial species. Interestingly, the A. pyogenes FtsY and Ffh proteins shared the highest similarity with the S. coelicolor FtsY (58.2% identity, 66.8% similarity) (GenBank accession number CAA22423) and Ffh (74.7% identity, 65.4% similarity) (GenBank accession number CAA19378) proteins, respectively. Immediately downstream of ffh was orf353, whose translated product shared similarity to a S. coelicolor conserved hypothetical protein (37.2% identity, 48.7% similarity) .
The conserved smc and ftsY genes are genetically linked in many prokaryotes, particularly Gram-positive bacteria (Fig. 2). In A. pyogenes, they are separated by orf121 and plo, suggesting the insertion of a small genomic islet. Further analysis of this region identified a lower %G+C content for orf121 (51.5%) and plo (54.5%) than the average %G+C content of A. pyogenes housekeeping genes (62.5%) (S.J. Billington, S.T. Rudnick and B.H. Jost, unpublished data) and genes flanking the proposed islet such as smc (62.3%), ftsY (63.5%), ffh (62.3%) and orf353 (62.1%) (Fig. 1). Furthermore, the codon usage of plo and orf121 was different compared to that of smc, ftsY, ffh, and orf353, with a significant AT bias in the wobble position of the codons for Ala, Gln, Asn, Thr, and His within the plo islet. For example, 62% of His-encoding codons in non-islet genes are CAC, while in islet genes, 73% of these codons are represented by the alternate codon CAT.
The identification of plo on a genomic islet of reduced %G+C content and alternate codon usage is suggestive of horizontal transfer of this virulence gene. The acquisition of plo may have converted the commensal A. pyogenes into a potential pathogen, similar to the acquisition of the LIPI-1 pathogenicity island of L. monocytogenes. The mechanism by which the plo islet was incorporated into the A. pyogenes chromosome is unknown. Insertion via transposition, or bacteriophage integration, seems unlikely, as neither integrase, transposon, insertion sequence, nor repeat sequences, often found associated with pathogenicity islands , were identified within or flanking the plo islet. However, it is possible that this islet may have inserted into the intergenic region between smc and ftsY via homologous recombination, as was hypothesized for sly on the Streptococcus suis chromosome . The integration of the plo islet between smc and ftsY may provide a selective advantage for retention of the islet since both smc and ftsY are required for normal bacterial growth and deletion of the islet may affect expression of either of these genes.
The plo islet is conserved across geographically diverse isolates
To examine the distribution of the plo islet among different isolates, several bovine and porcine strains from geographically diverse locations were analyzed by PCR for conservation of this region. Primers to the 3′-end of smc, SmcF1 (5′-GACGAGGTCGAGGCAGCAC-3′), and the 5′-end of ftsY, FtsYR3 (5′-GCAAGTTCCTCGTGTCCG-3′), were used to amplify a 3.2-kb product containing the entire plo islet. All isolates tested were positive for this amplification product (Fig. 3). Furthermore, PCRs using primers internal to plo, 2125 (5′-GGCCCGAATGTCACCGC-3′) and 2127 (5′-AACTCCGCCTCTAGCGC-3′), in conjunction with SmcF1 or FtsYR3, confirmed the location of plo between smc and ftsY in each of the isolates (data not shown). These results indicate that the location of the plo islet is invariant, and that if it was horizontally acquired, its introduction occurred prior to isolate diversification and may represent an ancestrally derived virulence islet.
orf121 does not regulate plo expression
Two homologs of PLO, LLO and PFO, are both regulated by proteins encoded by linked ORFs [7,8]. As orf121 was identified immediately upstream of plo, and within the plo islet (Fig. 1), this ORF was considered a candidate for a regulator of PLO expression. To analyze the role of orf121 on plo expression, an allelic exchange mutant of orf121, JGS534, was constructed. Hemolytic activities of culture supernatants collected at points during a growth curve were compared for JGS534 and BBR1. Growth rates and hemolytic activities in log and early stationary phase were similar between the two strains (Table 1), indicating that inactivation of orf121 has no observable effect on plo expression under these conditions. However, the effects of orf121 may be subtle, as has been observed with pfoR and sloR, putative regulatory genes of CDCs from C. perfringens and Streptococcus pyogenes, respectively [8,24]. To determine the effect of orf121 overexpression on plo production, orf121 was cloned downstream of Plac, on pJGS440. BBR1 derivatives carrying either pJGS181 or pJGS440 were grown in broth culture with Str selection to retain the plasmid. Both strains demonstrated similar growth rates in log phase and similar hemolytic activities in both log and stationary phase (Table 1). While it is possible that orf121, or some trans-acting factor necessary for its effect on plo expression, is not expressed in either of these systems in vitro, the combined results of these experiments, including the fact that computer analyses did not identify Orf121 DNA binding motifs, suggest that orf121 does not have a direct role in the expression of plo in vitro.
|Strain||Growth rate (h)||Hemolytic activity (HU ml−1)|
|Mid-log phase||Early stationary phase|
|Strain||Growth rate (h)||Hemolytic activity (HU ml−1)|
|Mid-log phase||Early stationary phase|
Doubling time in log phase.
Each value represents the average of duplicate assays from one representative experiment, and duplicate experiments demonstrated the same trends.
The A. pyogenes SRP protein component genes are physically linked and essential for growth
The A. pyogenes ftsY, ffh and orf353 genes were oriented in an operon-like arrangement and may be co-transcribed (Fig. 1), as ftsY and ffh were separated by 85 bp and ffh and orf353 by only 10 bp, and no obvious transcriptional terminators were identified in the intergenic sequences. This is an unusual arrangement for ftsY and ffh when compared to other bacteria. Analysis of the annotated prokaryotic genomes currently present in the GenBank database indicated that in only C. perfringens, L. monocytogenes, Staphylococcus aureus, Bacillus halodurans, B. subtilis, and Mycobacterium tuberculosis were ftsY and ffh found in close proximity, but in each case separated by at least one ORF (Fig. 2), while in most prokaryotic species, ftsY and ffh were separated by over 100 kb. The association of these genes in A. pyogenes could be indicative of an ancestral organization, which has been dispersed through evolution.
In E. coli, SRP is primarily a targeting mechanism for inner membrane proteins (IMPs) , although in Gram-positive organisms, SRP may also function in the targeting of secreted products . The identification of ftsY and ffh in close association with plo lead to the hypothesis that SRP may be involved in the targeting of PLO to the membrane for secretion. To address this hypothesis, strategies were devised to disrupt the A. pyogenes ftsY and ffh genes. Plasmid constructs carrying either ftsY or ffh disrupted by a Str resistance cassette, or constructs containing internal fragments of ftsY or ffh were introduced into BBR1. The former strategy attempted to disrupt the wild-type genes by allelic exchange, and the latter through a single crossover recombination event resulting in two defective copies of the target gene. Neither of these strategies resulted in viable recombinants with inactivated ftsY or ffh genes, despite the fact that plo mutants of BBR1 could be isolated at a frequency of 2.5×102 mutants µg−1 plasmid DNA using a previously constructed plo mutagenic plasmid, pJGS79 . While these experiments do not directly confirm that ftsY and ffh are essential in A. pyogenes, the number of failed attempts to construct stable mutants strongly suggests that they are essential under the conditions tested. Furthermore, these results are consistent with the essential nature of SRP genes in other bacteria .
The A. pyogenes ffh gene can not complement an E. coli conditional ffh mutant and has a dominant negative effect on normal E. coli growth
Since allelic exchange mutations in ftsY or ffh were not obtained in A. pyogenes, the ability of the A. pyogenes ffh gene to complement an E. coli ffh conditional mutant was investigated. E. coli WAM113, a conditional ffh mutant of WAM100, contains ffh under tight control of the araB promoter, such that depletion of arabinose results in severe growth defects . The A. pyogenes ffh gene was placed under the control of Ptac, on pJGS369. To determine if the A. pyogenes ffh could circumvent the growth defects observed in the absence of arabinose in WAM113, a range of IPTG concentrations (1 µM–10 mM) was used to induce expression of ffh on pJGS369 in the presence of decreasing arabinose concentrations (0.2–0.002%) on LB agar. Analysis of WAM113(pJGS369) showed that while decreasing arabinose alone, as expected, negatively affected the growth of this strain, addition of IPTG did not rescue growth (data not shown). Furthermore, induction of the A. pyogenes ffh from pJGS369, with IPTG concentrations ≥100 µM, inhibited growth of WAM100(pJGS369). These results not only suggest that the A. pyogenes ffh does not complement the E. coli ffh deficiency, but that its expression is toxic to E. coli. The lack of complementation of WAM113 by the A. pyogenes ffh gene was surprising as the E. coli and A. pyogenes ffh genes share 50.2% identity and 60.7% similarity. The Streptococcus mutans ffh gene has previously been demonstrated to complement the growth defect in WAM113 . However, expression of the mammalian SRP 54-kDa homolog of Ffh in WAM113 was unable to complement pleiotropic defects characterized by cell elongation, and increased accumulation of precursor proteins .
To confirm the toxic nature of A. pyogenes ffh in E. coli, growth curves of DH5α(pDK7) and DH5α(pJGS369) were performed in the presence and absence of 100 µM IPTG, levels that were found to be detrimental, but not completely inhibitory to WAM100(pJGS369). While the growth of DH5α(pDK7) was not affected by the addition of IPTG, DH5α(pJGS369) exhibited a significant decrease in growth in the presence of IPTG (Fig. 4).
These results suggest that A. pyogenes ffh encodes an active protein that affects normal growth of E. coli, but does not complement the loss of E. coli Ffh production. Therefore, expression of A. pyogenes ffh had a dominant negative effect on the normal functioning of the E. coli SRP system, probably resulting from the inability of A. pyogenes Ffh to interact with some components of the E. coli SRP, e.g. the E. coli FtsY, as is the case with the mammalian SRP 54 kDa . In toto, these results suggest that the E. coli system cannot currently be used as a heterologous system to study the role of the A. pyogenes SRP on targeting of A. pyogenes proteins. Future analyses may require the construction of conditional expression strains of A. pyogenes.
The A. pyogenes plo gene was localized to a genomic islet conserved across geographically diverse isolates. The identification of this and other CDCs on genomic islets suggests that the introduction of these genomic regions may have increased the virulence of previously non-pathogenic bacteria. Furthermore, the presence of plo on a genomic insertion between two essential genes may provide selective pressure for the maintenance of this important virulence gene during periods when A. pyogenes is a commensal on the mucosal membrane of its host. While we report here that orf121, located on the plo pathogenicity islet, does not play a direct role in the expression of plo in vitro, its role in vivo has yet to be determined.
The authors would like to thank Dr. G.J. Phillips for strains WAM100 and WAM113. We would also like to thank Dr. Paul Rudnick for helpful discussions and critical reading of the manuscript, and Dawn Bueschel, Dr. Kevin Keel and Hien Trinh for technical assistance. This work was supported by an NRICGP/USDA award (99-35204-7818).