Genome Sequence of Staphylococcus aureus Strain Newman and Comparative Analysis of Staphylococcal Genomes: Polymorphism and Evolution of Two Major Pathogenicity Islands

Abstract

Strains of Staphylococcus aureus, an important human pathogen, display up to 20% variability in their genome sequence, and most sequence information is available for human clinical isolates that have not been subjected to genetic analysis of virulence attributes. S. aureus strain Newman, which was also isolated from a human infection, displays robust virulence properties in animal models of disease and has already been extensively analyzed for its molecular traits of staphylococcal pathogenesis. We report here the complete genome sequence of S. aureus Newman, which carries four integrated prophages, as well as two large pathogenicity islands. In agreement with the view that S. aureus Newman prophages contribute important properties to pathogenesis, fewer virulence factors are found outside of the prophages than for the highly virulent strain MW2. The absence of drug resistance genes reflects the general antibiotic-susceptible phenotype of S. aureus Newman. Phylogenetic analyses reveal clonal relationships between the staphylococcal strains Newman, COL, NCTC8325, and USA300 and a greater evolutionary distance to strains MRSA252, MW2, MSSA476, N315, Mu50, JH1, JH9, and RF122. However, polymorphism analysis of two large pathogenicity islands distributed among these strains shows that the two islands were acquired independently from the evolutionary pathway of the chromosomal backbones of staphylococcal genomes. Prophages and pathogenicity islands play central roles in S. aureus virulence and evolution.

Staphylococcus aureus is a human pathogen that causes both nosocomial and community-acquired infections. The emergence of strains resistant to many antibiotics (methicillin-resistant S. aureus [MRSA]) and of highly virulent community-acquired MRSA that can cause fatal infections such as necrotizing pneumonia is of considerable concern even in countries with well-developed health surveillance systems (24, 30). In order to study mechanisms of staphylococcal antibiotic resistance and virulence, whole genome sequences of several different S. aureus strains have been determined. MRSA strains N315 and Mu50 were the first staphylococcal genomes to be sequenced (18), which were followed by nine additional strains (1, 5, 9, 10, 13, 14). All staphylococcal genomes are approximately 2.8 Mbp in size with a relatively low G+C content. Comparative analysis revealed that most regions of the staphylococcal genome are well conserved, whereas several large sequence blocks display high variability. S. aureus strains likely acquired these genomic islands horizontally and, at least initially, their integration into the genome must have required dedicated DNA recombination (integrase) genes. Furthermore, variable blocks of genome sequence frequently carry virulence and antibiotic resistance determinants that aid in the development of staphylococcal diseases. Variable regions can be classified as prophages, pathogenicity islands, or staphylococcal cassette chromosomes. The overall combination of variable sequence elements and the encoded spectrum of virulence properties varies from strain to strain and appears to be reflective of the overall large spectrum of clinical disease manifestations in humans (1, 2).

S. aureus strain Newman was isolated in 1952 from a human infection (6) and has been used extensively in animal models of staphylococcal disease due to its robust virulence phenotypes. Thirty genes that are required for staphylococcal pathogenesis were identified in S. aureus Newman after a screen of 1,736 bursa aurealis mutants with transposon insertions in different genes. Both well-characterized virulence genes and genes with unknown function were shown to be involved in the pathogenesis of staphylococcal infections (4). Additional benefits of systematic insertional mutagenesis are the identification of genes that are dispensable for staphylococcal growth under laboratory conditions. Subsequent work identified four prophages, φNM1 to φNM4, in the genome of strain Newman genome. Indeed, six paralogous groups of virulence determinants that were identified via bursa aurealis mutagenesis are encoded by these prophages (3). S. aureus Newman variants that lacked either φNM3 or φNM1, φNM2, and φNM4, or all four prophages (φNM1 to φNM4) displayed dramatic reductions in their ability to form organ specific abscesses after intravenous infection of mice, suggesting that the prophages φNM1 to φNM4 play important roles during the pathogenesis staphylococcal infections.

To further unravel molecular mechanisms of the physiology and pathogenesis of disease caused by S. aureus Newman, all of its genes must be known. This experimental goal was achieved, as we report here the complete genome sequence of S. aureus Newman. In contrast to other staphylococcal strains, which carry some virulence genes in mobile pathogenicity islands or genomic islets, virulence determinants of S. aureus Newman strain are conspicuous in prophages (2). Strain Newman carries a similar combination of major pathogenicity islands, νSaα and νSaβ, as S. aureus strains COL, NCTC8325, and USA300. In contrast to hospital-acquired MRSA, S. aureus Newman harbors only a small number of insertion sequences (IS) and lacks known antibiotic resistance determinants.

MATERIALS AND METHODS

Shotgun sequencing and contig assembly.

The whole-genome sequence was determined as described previously (1, 18). Shotgun sequencing was carried out by Hitachi High-Tech Fielding Co. (Tokyo, Japan) and Takara Bio, Inc. (Otsu, Japan). Sequences were also assembled as described previously. We have entered the whole-genome sequence of Newman in the DNA Database of Japan, with accession number AP009351.

Determination of open reading frames and structural RNA.

Determination of open reading frames, structural RNAs and annotations were performed as described previously (1, 18). Briefly, open reading frames were initially extracted with Genome GAMBLER program (Xanagen, Kawasaki, Japan) based on GLIMMER and rbsfinder software. The predicted open reading frames were then individually reviewed with GAMBLER. We searched a nonredundant protein database with the determined open reading frames using BLAST software for annotation. tRNA and tmRNA genes were identified by tRNAscan-SE (22) and with web-based software (http://www.indiana.edu/∼tmrna/), respectively. Illustration of G+C contents on a genome map was drawn by using Insilico molecular cloning software (Insilico Biology, Yokohama, Japan). This software was also used for comparative genome analysis among strains.

S. aureus genomes used for comparative analysis.

Sequences of S. aureus strains MW2 and N315 (accession numbers are BA000033 and BA000018, respectively) were used for whole-genome comparative analysis with strain Newman. The genome sequences of strains Mu50 (BA000017) (18), NCTC8325 (CP000253) (10), COL (CP000046) (9), MRSA252 (BX571856) (14), MSSA476 (BX571857) (14), RF122 (AJ938182) (13), USA300 (CP000255) (5), JH1 (CP000736) (A. Copeland et al., unpublished data), and JH9 (CP000703) (Copeland et al., unpublished) were also used for comparison among pathogenicity islands νSaα and νSaβ.

Calculation of phylogenic relationship among strains.

From genomes sequences of 10 different S. aureus strains nucleotide sequences of carbamate kinase, shikimate dehydrogenase, glycerol kinase, guanylate kinase, phosphate aceryltransferase, triosephosphate isomerase and acetyl coenzyme A acetyltransferase were used for multilocus sequence typing (MLST) analysis (7). Nucleotide sequences were combined and then aligned with CLUSTAL X (26). For phylogenic tree display, results from the CLUSTAL X calculation were visualized with TreeView (25).

RESULTS

Overview and genomic islands of the S. aureus Newman genome.

The whole-genome sequence of S. aureus Newman was determined as described previously (1, 17). Briefly, shotgun cloning of S. aureus strain Newman genomic DNA allowed for DNA sequencing of random fragments and data assembly into contiguous genome segments (contigs). Sequence gaps between assembled contigs were read by series of PCRs, using primers based on predetermined sequences. We encountered assembly difficulties for genome segments surrounding prophages due to the high degree of sequence homology between phages, in particular for φNM1 and φNM2 (a nearly 40-kb identical sequence). This obstacle was overcome by deriving DNA sequences from isolated φNM1 or φNM2 phage particles that had been isolated distinctively by mitomycin treatment of strain Newman (3). Indeed, sequences of φNM1 and φNM2 showed clear differences only in their attachment core sequences and integrases that recognize the attachment sequence and had little uniqueness in other domains. The clear differences in the attachment sites and the integrase sequences strongly supported that φNM1 and φNM2 were distinct prophages from each other, and this was confirmed by identifying unique chromosomal locus where insertion of each phage occurred in contiguous sequences upon shotgun assembly along with the different integrase sequences of φNM1 and φNM2. In conjunction with sequences for phage integration sites, this eventually permitted assembly of a circular chromosomal DNA sequence. The length of the S. aureus Newman chromosome is 2,878,897 bp, and it encodes 2,614 open reading frames (Table 1) . Plasmids sequences were not identified in our experiments involving S. aureus Newman.

TABLE 1.

Overview of S. aureus strain Newman genome in comparison with other strains

Parameter	Strain
	Newman	MW2	N315	Mu50	NCTC8325	MRSA252	MSSA476	COL	RF122	USA300	JH1	JH9
Length of sequence (bp)	2,878,897	2,826,402	2,814,816	2,878,529	2,821,361	2,902,619	2,799,802	2,809,422	2,742,531	2,872,769	2,906,507	2,906,700
G+C content (%)	32.9	32.8	32.8	32.9	32.9	32.8	32.9	32.8	32.8	32.8	33.0	33.0
Open reading framesa
No. of protein coding regions	2,614	2,632	2,593	2,714	2,892	2,744	2,619	2,673	2,589	2,560	2,747	2,697
% Coding	83.4	83.5	83.4	83.8	85.1	90.0	85.6	82.9	83.9	82.1	83.7	83.4
rRNAs
16S	5	6	5	5	5	5	6	6	5	5	6	6
23S	5	6	5	5	5	5	6	6	5	5	6	6
5S	6	7	6	6	6	6	7	7	6	6	7	7
tRNAs	56	61	62	60	61	60	60	53	60	53	60	59
tmRNA	1	1	1	1	1	1	1	1	1	1	1	1
No. of insertion sequencesb
IS1181	2	0	8	10	2	0	0	3	0	2	8	8
IS431	0	1	2	2	0	2	1	1	2	2	1	1
IS1272	0	1	0	0	1	9	0	0	0	1	1	1
Others (including remnant)	10	4	10	11	5	19	4	6	9	7	7	7
Transposon Tn554	0	0	5	2	0	3	0	0	0	0	2	2
Genomic islands
Prophages	4	2	1	2	3	2	2	1	1	2	4	4
SCCmec type	None	IV	II	II	None	II	Nonec	I	None	IV	II	II
νSaα typed	I	II	I	I	I	III	II	I	IV	I	I	I
νSaβ typed	II	II	I	I	II	III	II	II	II	II	I	I
Other islandse	2	3	2	3	2	2	2	3	3	4	1	1

^aBased on published records and therefore the criteria for open reading frame adoption differs from one genome to another.

^bBased on published records or BLAST searches using IS Finder (http://www-is.biotoul.fr/is.html), where genetic elements that e values were 0.0 to known IS were adopted.

^cSCChsd is present instead of SCCmec.

^dSee Fig. 3 and 4.

^eThat is, islands without pathogenic features are also included.

We compared the Newman genome with other S. aureus chromosomes thus far sequenced (Table 1). Although the strains are categorized as a single species, S. aureus, the chromosomes from different strains had unique features. The G+C contents did not vary drastically; however, the lengths of chromosomes differed by more than 5% when the longest chromosome from strain JH9 was compared to the shortest RF122, with lengths ranging 2.74 to 2.91 Mbp. Chromosomes were classified into two groups according to the number of rRNA genes. Importantly, the ribosomal gene numbers did not correlate with genomic island subtypes, as seen in SCCmec and νSa islands (see below). A number of genomic islands also had large varieties among different strains. At least one prophage was found in each genome, and Newman has as many as four, which is maximum number thus far identified in a genome. There was also wide variability among strains possessing other classes of islands and IS, indicating that these genetic elements play key roles in conferring chromosomal diversity to S. aureus strains.

Figure 1 displays a circular map of S. aureus Newman chromosomal DNA. Genomic islands including prophages and pathogenicity islands are shown as green lines. The integration of four different prophages is unique to strain Newman; for example, S. aureus MW2 and N315 harbor either two or only one prophage, respectively (Fig. 2). Two IS1181 insertion sequences and ten remnants of IS were found in strain Newman. One of the IS1181 was inserted into the corresponding site of IS1181-6 in strain N315 (17). Major transposons such as Tn554 were not found in the S. aureus Newman chromosome, in contrast to strain N315 with many IS, as well as five copies of Tn554 (Table 1). SCCmec (16) was not found in the chromosome of strain Newman, a finding in agreement with the observation that this strain is susceptible to methicillin and other β-lactam antibiotics (data not shown).

FIG. 1.

Circular display of S. aureus strain Newman chromosome. Green bars inside the first (outer) scale circle indicate the positions of genomic islands. The second circle shows open reading frames oriented in the forward direction, whereas the third circle indicates those oriented in the reverse direction. The fourth and fifth circles show genes for rRNAs and transfer RNAs, respectively. The sixth circle represents G+C content values. Purple indicates domains with G+C contents higher than 50%. The seventh circle shows G+C skew, in which purple indicates positive values.

FIG. 2.

Homology alignment of whole genomes between strains Newman and MW2 (A) and strains Newman and N315 (B). Gaps are highlighted with circles in green (unique elements to Newman), purple (unique to MW2 or N315), or yellow (the same elements but with differences). The νSaβ in panel A and νSaα in panel B are homologous each other, but only the pathogenicity island names are indicated in parentheses.

Open reading frames are indicated in second (found in the forward strand) and third (found in the reverse strand) circles as either red (virulence determinants) or blue (others) bars. The open reading frame orientation showed clear contrast according to movement of replication fork, which is in agreement with a tendency seen in other strains previously sequenced (1, 18). The G+C skew value distribution was also asymmetrical across the axle of replication origin termination sites. G+C contents tended to be relatively high not only in the loci where structural RNA genes were concentrated but also where genomic islands were located, a finding in agreement with the general hypothesis that horizontal gene transfer causes acquisition of the genomic islands.

Genomic island νSa and the nomenclature.

The term νSa refers to nonphage and non-SCC genomic islands that are exclusively present in S. aureus, often (but not always) encode for virulence determinants, are inserted at a specific locus in chromosome, and are associated with either intact or remnant DNA recombinase (1). The feature of a νSa genomic island possessing DNA recombinase also supports the hypothesis that staphylococcal pathogenicity islands are acquired by horizontal gene transfer. Due to the allelicity of the islands among different strains, the designation of the islands based on their structures or genetic content may create multiple island names that differ from strain to strain, regardless of the fact that they are inserted in identical loci of S. aureus chromosomes. We therefore propose that the term νSa does not designate an island with specific structure or a particular genetic content in a strain, but rather a locus where the island is inserted in the S. aureus chromosome. Hence, a term indicating a specific pathogenicity island such as “SaPI” should be used as well as “νSa.”

Among all of the S. aureus strains sequenced thus far, two major pathogenicity islands νSaα and νSaβ are present in strain Newman, and these islands appear to be allelic among different strains (2) (see also Fig. 3 and 4). The previously identified νSaγ, which has been shown to be present in all sequenced S. aureus strains, encoding exfoliative toxin and exotoxins (9), was also found in strain Newman. A fourth class of νSa island, νSa4, was located downstream of φNM3. This island was inserted at the corresponding site into the chromosome of strain N315, where a νSa4 island coding for 20 open reading frames, including three superantigen genes, is present in N315. Unlike the νSa4 in N315, the one found in strain Newman lacks known virulence determinants and encodes for only integrase and three functionally unknown proteins. Therefore, the νSa4 island in strain Newman probably lacks the features of a pathogenicity island and is structurally similar to the νSa4 island in strain MW2. The difference in νSa4 in strain Newman or strain MW2 compared to that in N315 also indicates that νSa4 shows polymorphism among strains as νSaα and νSaβ.

FIG. 3.

Comparison of major pathogenicity islands νSaα (A) and νSaβ (B) in Newman with other S. aureus strains. Arrows represent open reading frames and their orientations. The two islands for strain Newman are shown on the top of each panel, whereas each type of νSaα and νSaβ is shown with representative strains below that for Newman: νSaα of N315, MW2, MRSA252, and RF122 are shown as types I through IV (A), whereas νSaβ of N315, MW2, and MRSA252 are shown as types I through III (B), respectively. Refer to Fig. 4 for the classes of the islands carried by other strains. Note that HsdS sequences for strains are polymorphic and that a common type of the pathogenicity island always harbors identical HsdS sequences regardless of the strains. Another pathogenicity island, SaPIbov (8), is inserted downstream of guaA gene in type IV νSaα of strain RF122, and only the insertion position is indicated.

FIG. 4.

Tree view showing the phylogenic relationships among 12 S. aureus strains calculated based on the sequence diversity of seven housekeeping genes and the classification of the genomic islands (Fig. 3) νSaα (A) and νSaβ (B) using ellipses. The genomic island classification is based on its structural differences and HsdS sequences (1): when the HsdS sequences in either νSaα or νSaβ island from different strains are identical, the islands are categorized into a common class, whose genetic content is strongly associated with the HsdS allotype in the island, as shown in Fig. 3. If strains harbor a common class of νSaα (A) or νSaβ (Β), their names are covered by an ellipse with a unique color. Both HsdS in νSaα and νSaβ of RF122 are truncated, although νSaβ of the strain is classified into type II. The scale indicates the relative distance on the phylogenic tree view.

Comparative analysis of staphylococcal genomes.

Figure 2 shows a comparison of the S. aureus Newman genome with those of strains MW2 and N315. Depiction of homologous regions as dots or lines revealed that the overall size of the chromosome and the order of its genes are conserved among all three strains. Major homology gaps are caused by the insertion of four prophages into the genome of strain Newman. Similar to the insertion of φNM3 in strain Newman, S. aureus MW2 and N315 also carry prophages in the hlb gene (encoding beta-hemolysin); however, sequence homology between these phages is low, and differences are readily detectable in the plots in Fig. 2. SCCmec elements are found in strains N315 and MW2; however, this element is absent in S. aureus Newman. The observed gap size is larger in S. aureus N315 than in strain MW2, since SCCmec in strain N315 carries not only β-lactam resistance but also determinants for resistance to other antibiotics, whereas MW2 possesses only the β-lactam resistance gene mecA (23). Additional sequence gaps between staphylococcal strains are mainly due to differences in pathogenicity islands. The gap at 0.45 Mbp (Fig. 2A) is due to differences in pathogenicity island νSaα between Newman and MW2, whereas the gap at 1.9 Mbp in Fig. 2B reflects the differences in the νSaβ pathogenicity island between strains Newman and N315. These data suggest that S. aureus Newman νSaα is similar to νSaα in N315 but not to νSaα in MW2. Further, S. aureus Newman νSaβ is similar to νSaβ in MW2 but not to νSaβ in N315.

In S. aureus Newman, prophage φNM3 is inserted into hlb, and similar insertions have been observed for hlb-converting phages of S. aureus strains N315, MW2, Mu50, NCTC8325, MSSA476, MRSA252, USA300, JH1, and JH9. The genetic content of hlb-converting phages is, however, variable, especially with regard to virulence determinants. For example, hlb-converting phages of S. aureus N315 and Newman carry genes for staphylococcal complement inhibitor and chemotaxis inhibitory protein (31); the latter is absent in the S. aureus MW2 prophage. In contrast, the MW2 hlb-converting phage carries genes for enterotoxins K2 and Q that are not found in phages of strains N315 and Newman. Despite these differences, the integrase gene of φNM3 and those of related phages are virtually identical, suggesting that all hlb-converting phages evolved from a common ancestor. Other prophages of strain Newman, φNM1, φNM2, and φNM4, are absent in S. aureus MW2 and N315 (Fig. 2). However, φNM1 is inserted at the same integration site as φ11 in S. aureus NCTC8325 (15), and φ11 and φNM1 harbor the same integrase gene. The integrase of φNM4 is identical to that of φL54a in S. aureus COL and, similarly, φL54a and φNM4 insert into the same locus (geh). Thus, it seems highly likely that site-specific integration of phages into the staphylococcal genome is associated with different classes of integrase genes.

Virulence determinants.

Table 2 summarizes major virulence-related genes found in strain Newman compared to strains MW2 and N315. Twelve additional virulence-related genes that belong to six paralogous groups identified in previous studies (3, 4) were encoded by the four Newman prophages, in addition to genes in other loci of the chromosome. The functionally unknown phage genes were conspicuously present in Newman, whereas most of them were absent in prophages of strains MW2 and N315, suggesting that the phage genes play an important role in the virulence of strain Newman.

TABLE 2.

Major virulence-related genes in strains Newman, MW2, and N315 and their locations^a

Type	Strain
	Newman			MW2			N315
	Locus	Gene	Island	Locus	Gene	Island	Locus	Gene	Island
Toxins
Enterotoxin	NWMN1883	sea	φNM3	MW1889	sea	φSa3mw	SA1761	sep	φSa3n
				MW1938	sek2	φSa3mw	SA1817	sec3	νSa4
				MW1937	seq	φSa3mw	SA1816	sel	νSa4
				MW0759	sec4	νSa3	SA1642	seg	νSaβ
				MW0760	sel2	νSa3	SA1643	sen	νSaβ
				MW0051	seh		SA1646	sei	νSaβ
							SA1647	sem	νSaβ
							SA1648	seo	νSaβ
Exotoxin	NWMN0388	set1nm	νSaα	MW0382	set16	νSaα	SA0382	set6	νSaα
	NWMN0389	set2nm	νSaα	MW0383	set17	νSaα	SA0383	set7	νSaα
	NWMN0390	set3nm	νSaα	MW0384	set18	νSaα	SA0384	set8	νSaα
	NWMN0391	set4nm	νSaα	MW0385	set19	νSaα	SA0385	set9	νSaα
	NWMN0392	set5nm	νSaα	MW0386	set20	νSaα	SA0386	set10	νSaα
	NWMN0393	set6nm	νSaα	MW0387	set21	νSaα	SA0387	set11	νSaα
	NWMN0394	set7nm	νSaα	MW0388	set22	νSaα	SA0388	set12	νSaα
	NWMN0395	set8nm	νSaα	MW0389	set23	νSaα	SA0389	set13	νSaα
	NWMN0396	set9nm	νSaα	MW0390	set24	νSaα	SA0390	set14	νSaα
	NWMN0397	set10nm	νSaα	MW0391	set25	νSaα	SA0393	set15	νSaα
	NWMN0400	set11nm	νSaα	MW0394	set26	νSaα
Toxic shock syndrome toxin							SA1819	tst	νSa4
Exfoliative toxin	NWMN1082	eta	νSaγ	MW1054	eta	νSaγ	SA1016	eta	νSaγ
Alpha-hemolysin	NWMN1073	hly	νSaγ	MW0955	hly	νSaγ	SA1007	hly	νSaγ
Beta-hemolysin	Truncated	hlb	φNM3	Truncated	hlb	φSa3mw	Truncated	hlb	φSa3n
Delta-hemolysin	NWMN1942	hld		MW1959	hld		SAS065	hld
Gamma-hemolysin	NWMN2318	hlgA		MW2342	hlgA		SA2207	hlgA
component	NWMN2319	hlgC		MW2343	hlgC		SA2208	hlgC
	NWMN2320	hlgB		MW2344	hlgB		SA2209	hlgB
LukDE leukocidin	NWMN1717	lukD	νSaβ	MW1767	lukD	νSaβ	SA1637	lukD	νSaβ
component	NWMN1718	lukE	νSaβ	MW1768	lukE	νSaβ	SA1638	lukE	νSaβ
Panton-Valentine leukcidin				MW1379	lukS-PV	φSa2mw
component				MW1378	lukF-PV	φSa2mw
Exoenzymes
Serine protease	NWMN1706	splA	νSaβ	MW1755	splA	νSaβ	SA1631	splA	νSaβ
	NWMN1705	splB	νSaβ	MW1754	splB	νSaβ	SA1630	splB	νSaβ
	NWMN1704	splC	νSaβ	MW1753	splC	νSaβ	SA1629	splC	νSaβ
	NWMN1703	splD	νSaβ	MW1752	splF	νSaβ	SA1628	splD	νSaβ
	NWMN1702	splE	νSaβ				SA1627	splF	νSaβ
	NWMN1701	splF	νSaβ
	NWMN0892	htrA		MW0903	htrA		SA0879	htrA
Serine V8 protease	NWMN0918	sspA		MW0932	sspA		SA0901	sspA
Cysteine protease	NWMN0917	sspB		MW0931	sspB		SA0900	sspB
	NWMN0916	sspC		MW0930	sspC		SA0899	sspC
Lipase	Truncatedg	geh	φNM4	MW0297	geh		SA0309	geh
	NWMN2569	lip		MW2590	lip		SA2463	lip
Lipase/esterase	NWMN0624	lipA		MW0617	lipA		SA0610	lipA
Hyaluronate lyase	NWMN2106	hysA		MW2129	hysA		SA2003	hysA
Thermonuclease	NWMN1236	nuc		MW1211	nuc		SA1160	nuc
Immunomodulators
Staphylokinase	NWMN1880	sak	φNM3	MW1885	sak	φSa3mw	SA1758	sak	φSa3n
Chemotaxis inhibiting protein	NWMN1877	chp	φNM3				SA1755	chp	φSa3n
Complement inhibitor	NWMN1876	scn	φNM3	MW1884	scn	φSa3mw	SA1754	scn	φSa3n
Immunoglobulin G binding protein A	NWMN0055	spa		MW0084	spa		SA0107	spa
Immunoglobulin G binding protein sbi	NWMN2317	sbi		MW2341	sbi		SA2206	sbi
Adhesins
Collagen-binding protein				MW2612	cna
Fibronectin-binding protein	NWMN2399	fnbAb		MW2421	fnbA		SA2291	fnbA
	NWMN2397	fnbBb		MW2420	fnbB		SA2290	fnbB
SD-rich fibrinogen-binding	NWMN0756	clfA		MW0764	clfA		SA0742	clfA
protein	NWMN2529	clfB		MW2551	clfB		SA2423	clfB
	NWMN0523	sdrC		MW0516	sdrC		SA0519	sdrC
	NWMN0524	sdrD		MW0517	sdrD		SA0520	sdrD
	NWMN0525	sdrE		MW0518	sdrE		SA0521	sdrE
	NWMN1940	sdrH		MW1956	sdrH		SA1839	sdrH
Elastin-binding protein	NWMN1389	ebpS		MW1369	ebpS		SA1312	ebpS
Other virulence-related phage proteinsc
Functionally unknown proteinA	NWMN0270d		φNM4
Functionally unknown proteinB	NWMN0273		φNM4
Functionally unknown proteinC	NWMN0280		φNM4	MW1420		φSa2mw	SA1786		φSa3n
Functionally unknown protein	NWMN0284		φNM4
Functionally unknown protein	NWMN0995		φNM2
Functionally unknown proteinB	NWMN1001		φNM2
Functionally unknown proteinC	NWMN1008		φNM2
Functionally unknown proteinC	NWMN1800		φNM1
Functionally unknown proteinB	NWMN1807		φNM1
Functionally unknown proteinA	NWMN1809e		φNM1
Functionally unknown proteinC	NWMN1905f		φNM3
Functionally unknown protein	NWMN1912		φNM3	MW1924		φSa3mw	SA1795		φSa3n

^aAmong virulence-related genes shown as red bars in Fig. 1, those involved in capsular polysaccharide biosynthesis and iron uptake are not listed.

^bThe C terminus, including cell wall anchoring signal, is truncated, and therefore the product is nonfunctional (11).

^cIdentified by bursa aurealis mutagenesis of strain Newman followed by nematode killing assay employing the mutants (4). Unknown proteins with the same superscript capital letter are paralogs.

^dLocated between NWMN0270 and NWMN0271.

^eLocated between NWMN1809 and NWMN1810.

^fLocated between NWMN1905 and NWMN1906.

^gTruncated due to insertion of the indicated prophage.

Most of the exoenzymes and adhesins were commonly present in the three strains shown in Table 2, although a lipase encoded in geh gene was truncated due to the insertion of φNM4, and extra genes were present in the spl serine protease cluster in νSaβ in strain Newman. It has also been noted that fibronectin-binding proteins encoded by fnbA and fnbB genes were present in strain Newman, but they lack C-terminal cell wall sorting signals (27), as reported previously (11). The three strains in Table 2 share the same sets of hemolysin components and LukDE leukocidin genes. However, unlike S. aureus N315 and MW2, genes for staphylococcal superantigens were not found in the genome in strain Newman with the single exception of enterotoxin A (sea), which is encoded by φNM3. Since other sequenced S. aureus strains normally harbor at least two enterotoxin genes in their chromosomes (data not shown), strain Newman is characteristic in its possession of a smaller number of enterotoxin genes. In S. aureus MW2, genes for enterotoxin H (seh) and collagen-binding protein (cna) are located within genome islets (1) and Panton-Valentine leukocidin component genes, whose product is responsible for fatal outcomes in humans such as necrotizing pneumonia (20) in a prophage; however, these virulence genes are not present in S. aureus Newman.

Major pathogenicity islands of S. aureus Newman.

Prophages, staphylococcal cassette chromosome, and pathogenicity islands (8, 21) are categorized as genomic islands and carry characteristic integrase genes (DNA recombinases) (1). Unlike most other genomic islands, the pathogenicity islands νSaα and νSaβ are present in all S. aureus genomes sequenced thus far; however, νSaα and νSaβ harbor only remnants of their integrase genes. It therefore seems reasonable to assume that νSaα and νSaβ are no longer mobile and that these pathogenicity islands must have played a major role in the evolution of this pathogen (2). For example, even though the genome of Staphylococcus epidermidis is highly homologous to that of S. aureus, the genomic islands νSaα and νSaβ are absent from the genome of this or any other coagulase-negative staphylococci (9, 19, 29, 33). These findings are in agreement with a general hypothesis that acquisition of νSaα and νSaβ into a primordial staphylococcal genome may have been associated with subsequent evolution of S. aureus as a major human pathogen.

νSaα and νSaβ display polymorphisms among strains and can be classified into three to four groups based on their structural differences and HsdS subtypes generated in each of the two islands. Tandem arrays of exotoxin and lipoprotein genes are characteristic features of νSaα in Newman (Fig. 3). νSaβ carries genes responsible for lantibiotic biosynthesis in addition to genes for leukocidin components (lukD and lukE) and a serine protease gene (spl) cluster. Compared to other strains, νSaα for Newman belongs to the same type as νSaα in strains N315, Mu50, NCTC8325, COL, JH1, JH9, and USA300 (type I in Fig. 3A), whereas νSaβ belongs to the same type as νSaβ in strains NCTC8325, COL, MW2, MSSA476, and RF122 (type II in Fig. 3B). Therefore, the combination of νSaα and νSaβ in S. aureus Newman (blue and brown ellipses in Fig. 4A and B, respectively) is the same as in strains NCTC8325, COL, and USA300 (see also Table 1). It is noteworthy that different types of νSaα and νSaβ correlate with sequence variation in hsdS, whose product determines the sequence specificity of DNA methylation and restriction via staphylococcal restriction-modification systems (17). Figure 4, however, also shows that the type distribution of the two major pathogenicity islands in the strains does not correlate with phylogenic relationships among strains when a phylogenic tree image is drawn based on allelic distribution of seven essential housekeeping genes used for MLST analysis (7). In addition, the distribution of νSaα types among strains differs from that of νSaβ, suggesting that the two pathogenicity islands have been acquired by the strains independently of each other and differently from housekeeping genes that have presumably evolved in a vertical fashion.

Due to their general distribution in S. aureus strains and absence in other staphylococci, the two major pathogenicity islands are considered to play important roles in virulence for their human hosts. The molecular mechanisms that implement such putative strategies are, however, still unknown, and future work will need to unravel how pathogenicity islands are involved in staphylococcal virulence during host infection.

DISCUSSION

Following the first sequencing of S. aureus N315 (18), 11 additional S. aureus genomes have been determined and deposited into the databases. Here we add the whole genome sequence of S. aureus Newman to this rapidly growing list. Genome sequencing projects for multiple isolates of a bacterial pathogen are of considerable scientific value because the generated data reveal not only gene content but also conservation and variability between different strains and their associated human or animal diseases. Staphylococcal diversity is mainly due to polymorphisms that occur in genomic islands, which also carry many virulence and antibiotic resistance determinants. Nevertheless, some genes, such as the staphylocoagulase gene, are located outside of genomic islands and are known to be polymorphic (32). One can add to this list certain combinations of virulence genes, for example, seh (enterotoxin H) and cna (collagen-binding protein), which are present only in certain types of S. aureus strains. A hallmark of the S. aureus classification is the ability of these microbes to ferment mannitol and to produce characteristic proteins such as DNase, coagulase, and protein A. S. aureus strains differ from one another in virulence and drug resistance features that are carried in or outside of genomic islands.

Previous works (3, 4) revealed virulence genes or candidate virulence genes within four prophages that have integrated into the genome of S. aureus Newman. Our determination of the whole genome sequence for strain Newman showed that many virulence-related genes are encoded by prophages. One superantigen, staphylococcal enterotoxin A (sea), is located in φNM3; however, unlike other staphylococcal strains, additional superantigen genes were not found. Furthermore, S. aureus Newman carries a small pathogenicity island but lacks known virulence genes. We also failed to identify the collagen adhesin gene that is present in strains MW2, MRSA252, and MSSA476. Therefore, it is likely that virulence caused by strain Newman largely relies on prophages, in addition to the contribution by other virulence determinants present in all S. aureus strains, and the nonprophage regions of strain Newman genome seem to form the basic backbone of pathogenic S. aureus. While en bloc transfer of virulence genes via prophages and pathogenicity islands appears to be important for S. aureus acquisition of virulence properties, stepwise incorporation of additional genes and/or mutations may play an additional role in the evolution of clones with similar, yet discretely different strategies for the pathogenesis of human disease.

As shown in Fig. 4, analysis of two major pathogenicity islands in 12 different S. aureus genome sequences revealed that these strains do not always share the same combinations of νSaα and νSaβ classes. Moreover, the classes do not correlate with phylogenic relationship based on the allelic distribution of seven housekeeping genes upon MLST analysis (7). This clearly shows that these two pathogenicity islands were horizontally acquired and must have evolved independently of S. aureus genomes, whereas housekeeping genes are considered to evolve in a vertical fashion. Interestingly, sequences of hsdS gene products that determine the site specificity of methylation and restriction in restriction-modification systems vary depending on the type of pathogenicity islands that encodes them. The reasons why modification subunits of the R-M system are present in νSaα and νSaβ and have sequence variations remain unknown. One possible explanation is that sequence diversity in pathogenicity islands requires its distinct restriction modification site determined by HsdS: since self-DNA protection by modification system is promoted by sequence-specific methylation on DNA, sequence diversity in genomic islands should coincide with the methylation site determined by HsdS. DNA methylation of pathogenicity islands may further influence expression of the virulence gene and thereby affect the pathogenesis of infectious diseases caused by this organism. Recent studies have revealed that type I RM system activity and modification site specificity are related to changes in the surface antigenic protein in Mycoplasma pulmonis, depending on the organism's infection sites (12, 28). This suggests that the RM system in S. aureus also plays a direct role in virulence.

Some of the genes located within the major pathogenicity islands, νSaα and νSaβ, are presumed to be involved in virulence. However, their molecular contributions to pathogenicity are still unclear. It should also be noted that the presence of any one gene does not result in its expression. In order to reveal the mechanisms of virulence further, microarray experiments could be used to reveal their expression.

The overall spectrum and individual combinations of virulence genes, as they are diversely encoded by different genomic islands, appears to be the major factor in determining clinical symptoms after S. aureus infection and may even dictate the severity of diseases caused by this pathogen. Together with an analysis of transposon insertion mutants (4), our work here may provide experimental strategies for better understanding the pathogenicity and physiology of S. aureus.