Whole-Genome Sequencing of Staphylococcus aureus Strain RN4220, a Key Laboratory Strain Used in Virulence Research, Identifies Mutations That Affect Not Only Virulence Factors but Also the Fitness of the Strain^,

Abstract

Staphylococcus aureus RN4220, a cloning intermediate, is sometimes used in virulence, resistance, and metabolic studies. Using whole-genome sequencing, we showed that RN4220 differs from NCTC8325 and contains a number of genetic polymorphisms that affect both virulence and general fitness, implying a need for caution in using this strain for such studies.

TEXT

Staphylococcus aureus is a versatile pathogen that is responsible for the majority of nosocomial and community-acquired infections. One of the biggest challenges in treating staphylococcal infections is that many S. aureus strains have developed resistance against various antibiotics. In contrast to clinical isolates, RN4220 is a commonly used laboratory strain that is characterized by a mutation in the sau1 hsdR gene, making it restriction deficient and hence an ideal intermediate cloning host. RN4220 was originally derived from NCTC8325-4 using UV and chemical mutagenesis (10). NCTC8325-4, in turn, was derived from an early clinical isolate, NCTC8325 (also known as PS47 or RN1), cured of three prophages (16). Indeed, RN4220 is not a suitable candidate for the study of antibiotic resistance, because newer methicillin-resistant S. aureus (MRSA) lineages have evolved [e.g., ST239 (hospital), ST80, and ST59 (community)], presumably due to recombination events between lineages. RN4220 is also known to harbor a small deletion in rsbU, a gene within the stress-induced sigB operon, which renders it deficient in σ^B expression. Additionally, RN4220 shows a Δagr mutant phenotype and does not produce α-hemolysin despite producing small amounts of RNAIII in late log phase (24). Recent studies revealed that a deletion of cvfB (encoding conserved virulence factor B) in RN4220 resulted in diminished agr expression, with a reduction in hla expression, protease production, and virulence, in a silkworm model of systemic infection (14), but the loss of hemolytic activity in the cvfB mutant of RN4220 was found to be due to a defective agr locus and not attributable to the cvfB mutation (14). In another study, SrrAB, a two-component regulatory system, was found to repress the transcription of RNAIII of the agr locus in RN4220 (20, 27). However, the interpretation of virulence and of the associated regulatory data in these studies with RN4220 is suspect due to an inherent agr defect in this strain (24). Finally, O'Neill showed by comparative sequencing that NCTC8325-4, which was thought to be identical to its parent NCTC8325 except for the deletion of three prophages, possesses previously undescribed polymorphisms that may influence the virulence and pathogenicity of NCTC8325-4 (18). As a result of these issues, it is extremely important to delineate an accurate picture of the mutations in RN4220, given the polymorphisms in this strain which can have an impact upon virulence and resistance phenotype.

Using Illumina Solexa-based whole-genome sequencing (paired end) (P. Mayer, L. Farinelli, and E. Kawashima, U.S. patent application WO98/44151), we obtained the whole-genome sequence of strain RN4220 and subsequently identified the polymorphisms in RN4220 compared with the released NCTC8325 genome. Briefly, RN4220 was grown with aeration at 37°C in tryptic soy broth to log phase (optical density at 620 nm [OD₆₂₀], 0.7). Genomic DNA, isolated with a phenol-chloroform extraction method (6), was sent to Ambry Genetics (California) for library preparation and sequencing. The library preparation was carried out by shearing genomic DNA and blunting it, followed by the addition of adenine at the 3′ end. A specific adapter with bar coding was then ligated to these DNA fragments, followed by PCR amplification (Illumina). Fragments of ∼76 bases were generated and assembled.

Genome assembly, single-nucleotide polymorphism (SNP) calling, and annotation were done as follows. The obtained 3.5 million paired reads, each of which is 68 nt, were de novo assembled using an Edena assembler (9), development version 3.0. The assembly has been slightly refined using the Minimus assembler (23). Final assembly resulted in 179 contigs (sum = 2.67 Mb, N₅₀ = 80.5 kb [N₅₀ is the contig size such that 50% of the entire assembly is contained in contigs equal to or larger than that size], max = 148 kb). Contigs were annotated using the RAST server (2). Comparison, SNP calling, SNP annotation, and graphical mapping were performed using the MUMmer software package (12), the CIRCOS visualization engine (11), and applications developed in-house. A map of the genome showing the polymorphisms in RN4220 is shown in Fig. 1. The entire genomic sequence of RN4220 can be found in the supplemental material.

Fig. 1.

Graphical mapping of the strain RN4220 contigs onto the strain NCTC8325. The RN4220 contigs are shown on the left-hand side of the circle, while the complete genome sequence of the strain NCTC8325 is shown on the right-hand side. Homologous sequences between the two strains are linked by colored ribbons. The black dots along the NCTC8325 genome sequence indicate the SNP positions, as described in Table S2 in the supplemental material. The three open circles indicate the large regions in NCTC8325 that are deleted in RN4220.

Compared with the published NCTC8325 genome, we identified 121 SNPs and 4 large-scale deletions (see Tables S1 and S2 in the supplemental material). Among the SNPs, 14 were synonymous. The remaining SNPs involve 80 nonhomologous substitutions in coding regions and 27 substitutions in the intergenic regions. As anticipated, three of the four large deletions were associated with the absence of Φ11, Φ12, and Φ13 (Table S1). We also confirmed a subset of 29 nonsynonymous mutations (those that might affect virulence or metabolism) by PCR amplification followed by DNA sequencing (Table 1). This subset included nine nonsynonymous mutations that were also identified by O'Neill (18).

Table 1.

Nonsynonymous, PCR-verified SNPs identified in RN4220 relative to NCTC8325

Genome position	Putative gene product and function	Nucleotide change		Amino acid changea	Locus tag
		NCTC8325	RN4220
174867	HsdR family type I site-specific DNase	G	A	W197*	SAOUHSC_00162
281533	EssC (DNA segregation FtsK/SpoIIIE, S-DNA-T family)	G	A	W52*	SAOUHSC_00262
292106	Hypothetical protein	G	K		SAOUHSC_00274
292107		C	Y
292179		C	Y
292199		A	R
292328		G	R
388693	Conserved hypothetical protein	C	T	P134S	SAOUHSC_00383
590402	Conserved hypothetical protein	G		Frameshift	SAOUHSC_00591
751285	SecA preprotein translocase subunitb	A	T	E449V	SAOUHSC_00769
795429	Clumping factor ClfA	C	T	S815L	SAOUHSC_00812
827849	Hypothetical proteinb	A	T	T164S	SAOUHSC_00859
939304	Competence transcription factor, putativeb	C	T	E53K	SAOUHSC_00961
1016979	Spermidine/putrescine ABC transporter, putativeb	G	A	E220K	SAOUHSC_01048
1020577	Manganese transport protein MntHb	G	T	S286*	SAOUHSC_01053
1063555	UvrC, excinuclease ABC subunit C	C	T	P331S	SAOUHSC_01102
1123048	Orotate phosphoribosyltransferaseb	G	A	G42S	SAOUHSC_01
1160531	RimM, 16S rRNA-processing protein	G	A	A106T	SAOUHSC_01209
		A	G
1358230	Kgd, alpha-ketoglutarate decarboxylaseb	C	T	D590N	SAOUHSC_01418
1632629	(5-Methylaminomethyl-2-thiouridylate)methyltransferase		A	Frameshift	SAOUHSC_01726
1733572	Septation ring formation regulator EzrA	G	T	T73N	SAOUHSC_01827
		A	G	F54S
2087725	GroEL chaperone	A	T	F218I	SAOUHSC_02254
2096628	AgrA		A	Frameshift	SAOUHSC_02265
2106539	ABC transporter, ATP-binding protein, putative	A	T	L602F	SAOUHSC_02274
2446162	Phosphotransferase system sucrose-specific IIBC component, putative	C	T	A211T	SAOUHSC_02661
		C	G
		C	A

^a*, truncation.

^bMutation identified by O'Neill (18).

Besides the deletions of the three phages, there were also deletions of two hypothetical proteins and an 1,195-bp region that codes for the B subunit of excision endonuclease (also called excinuclease) ABC, which catalyzes the processing of DNA lesions by the UvrABC excinuclease complex for DNA repair (25). Mutations in hsdR (SAOUHSC_00162), essC (SAOUHSC_00262), mntH (SAOUHSC_01053), and a hypothetical gene (SAOUHSC_02790) result in a premature stop codon (Table 1). More specifically, the mutation in hsdR, resulting in a lack of restriction, confirmed previous data of Waldron and Lindsay (26). A rhomboid family protein (SAOUHSC_01649) shows an N-terminal deletion. There are 20 indels (insertion/deletions) in coding regions, causing frame-shift mutations in SAOUHSC_00269, SAOUHSC_00270, SAOUHSC_00274, SAOUHSC_00275, and SAOUHSC_00276, all hypothetical proteins. Among the indels, one of the insertions is in agrA and another is in the gene encoding (5-methylaminomethyl-2-thiouridylate)methyltransferase, while the remaining two are in noncoding regions. Of the 27 mutations in noncoding regions, 5 are within 100 bp of the coding regions of putative regulatory loci. Most of these intergenic SNPs were located in untranslated regions or within putative operons, possibly impacting multiple factors. We found some intergenic SNPs in predicted regulatory noncoding RNA such as Sau-25 (1) and within regions potentially affecting the structure of the highly stable small RNA Teg27/RsaX18 (4, 21).

Eleven of the 121 mutations were also found in NCTC8325-4 (18) (Table S2 in the supplemental material). The remaining mutations, unique to RN4220, can be grouped into two categories: (i) those affecting the survival and fitness of the bacteria, which include uvrC (part of the uvrABC system) (25), ezrA, rimM, and genes encoding ribosomal protein S2 and GroEL, and (ii) those affecting virulence factors, namely, agrA, essC, clfA, and a gene encoding superantigen-like protein (Table 1).

UvrC, part of the system involved in DNA repair, is specifically involved in the incision of the 5′ and 3′ sides of the lesion. The P331S mutation in UvrC in RN4220 may result in disruption of the secondary structure of UvrC, thus making RN4220 more prone to spontaneous mutations. The ezrA gene, likely essential to S. aureus, is thought to regulate septum ring formation (7). RimM (ribosome maturation factor), ribosomal protein S2, and GroEL are important proteins collectively involved in protein synthesis and protein folding. More specifically, RimM is an accessory factor required for 30S maturation and assembly in Escherichia coli (13), and deletion of rimM has been known to decrease the growth rate and reduce translational efficiency at 37°C. In S. aureus, treatment with heat or a cell wall-active antibiotic results in increased transcription of groEL, indicating a role for GroEL in protein folding under heat and antibiotic stress (17, 22). Mutations in these three proteins presumably would thus have an effect on the general fitness (slower growth and translational deficiency) of the strain, especially under stress. The mutation of these genes might result in an inadequate response to antibiotics, which would actually be a false-positive effect of the antibiotic in question. With regard to the virulence factors, the agrA mutation results in the insertion of an extra adenine residue at the 3′ end of agrA, leading to a run of eight adenines and a frameshift that adds three amino acids to the C terminus of AgrA. This finding with agrA in RN4220, resulting in delayed activation of agr and a failure to synthesize delta and alpha hemolysins, has been described by Traber and Novick (24). The membrane protein EssC is one product of the eight-gene cluster of the ESAT-6-like secretion system (Ess), which is essential for the secretion of EsxA-EsxB (5). The essC mutation in RN4220 results in a truncated EssC. S. aureus mutants that fail to secrete EsxA and EsxB display significantly reduced virulence, dissemination, and colonization in mice (5). The S. aureus clumping factor ClfA, a surface protein belonging to the MSCRAMM family, binds the γ-chain of fibrinogen (15) and induces platelet aggregation (3); it also mediates adherence of S. aureus to fibrinogen-coated surfaces and contributes to protection against phagocytosis by neutrophils (19). The S815L mutation in ClfA in RN4220 results in the substitution of a polar hydrophilic amino acid with a nonpolar hydrophobic residue. This may conceivably disrupt the protein fold and reduce its affinity for fibrinogen. The superantigen-like protein is similar to exotoxin Set6, which is a virulence-associated protein (8).

Based on the above analysis, investigators using RN4220 in virulence studies should proceed with caution, since mutations identified in this paper show that the virulence genes as well as those involved in fitness and numerous stress-associated putative regulators are altered in RN4220.