Rapid and Accurate Identification of Human-Associated Staphylococci by Use of Multiplex PCR

Abstract

Although staphylococci are identified by phenotypic analysis in many clinical laboratories, these results are often incorrect because of phenotypic variation. Genetic analysis is necessary for definitive species identification. In the present study, we developed a simple multiplex-PCR (M-PCR) for species identification of human-associated staphylococci, which were as follows: Staphylococcus aureus, S. capitis, S. caprae, S. epidermidis, S. haemolyticus, S. hominis, S. lugdunensis, S. saprophyticus, and S. warneri. This method was designed on the basis of nucleotide sequences of the thermonuclease (nuc) genes that were universally conserved in staphylococci except the S. sciuri group and showed moderate sequence diversity. In order to validate this assay, 361 staphylococcal strains were studied, which had been identified at the species levels by sequence analysis of the hsp60 genes. In consequence, M-PCR demonstrated a sensitivity of 100% and a specificity of 100%. By virtue of simplicity and accuracy, this method will be useful in clinical research.

INTRODUCTION

Staphylococci are normal inhabitants of human skin and mucous membranes. Staphylococcus aureus is a common cause of invasive and life-threatening infections, but coagulase-negative staphylococci (CNS) cause mainly nosocomial infections, such as catheter-related bloodstream infections, prosthetic valve endocarditis, central nervous system shunt infections and prosthetic joint infections (17, 31, 42). Unlike other CNS, however, S. lugdunensis is highly pathogenic and can cause aggressive skin and soft tissue infections, bone and joint infections, and native valve endocarditis (13, 31, 42). In addition, the oxacillin MIC breakpoints to determine methicillin resistance differ among staphylococcal species. Therefore, it is important to identify staphylococci at the species levels.

A variety of methods have been reported for species identification of staphylococci. Although conventional biochemical assays are employed in many clinical laboratories, they are frequently imprecise due to phenotypic variation (5, 16, 19, 26). Real-time PCR has been developed, but some staphylococcal species are indistinguishable, and/or interpretation of results is intricate (12, 37). Sequence analyses of several target genes are the most reliable methods (10, 25, 28, 36). However, these are expensive, laborious, and time-consuming. Thus, a simple and reliable assay is needed for identification of staphylococcal species.

The purpose of the present study was to develop a rapid and accurate multiplex-PCR (M-PCR) for species identification of human-associated staphylococci, which contained the following: S. aureus, S. capitis, S. caprae, S. epidermidis, S. haemolyticus, S. hominis, S. lugdunensis, S. saprophyticus, and S. warneri. This method was based on nucleotide sequences of the thermonuclease (nuc) genes that had been used for staphylococcal identification (4, 35).

MATERIALS AND METHODS

Bacterial strains and species identification.

A total of 24 staphylococcal strains, which included 5 species with whole-genome sequences (14, 23, 24, 32, 40), were used for phylogenetic analysis of the nuc genes (Table 1). A total of 361 staphylococcal strains were studied to assess the utility of M-PCR for species identification of human-associated staphylococci, which were as follows: S. aureus (n = 60), S. epidermidis (n = 104), S. lugdunensis (n = 26), S. haemolyticus (n = 24), S. saprophyticus (n = 23), S. warneri (n = 23), S. hominis subsp. hominis (n = 8), S. hominis subsp. novobiosepticus (n = 14), S. capitis subsp. capitis (n = 7), S. capitis subsp. ureolyticus (n = 24), S. caprae (n = 16), S. auricularis (n = 1), S. arlettae (n = 1), S. carnosus (n = 1), S. chromogenes (n = 1), S. cohnii subsp. cohnii (n = 1), S. cohnii subsp. urealyticum (n = 1), S. croceolyticus (n = 1), S. delphini (n = 1), S. devriesei (n = 1), S. equorum (n = 1), S. felis (n = 1), S. fleurettii (n = 1), S. gallinarum (n = 1), S. hyicus (n = 1), S. intermedius (n = 1), S. kloosii (n = 1), S. lentus (n = 1), S. lutrae (n = 1), S. nepalensis (n = 1), S. muscae (n = 1), S. pasteuri (n = 1), S. pettenkoferi (n = 1), S. piscifermentans (n = 1), S. pseudintermedius (n = 1), S. saccharolyticus (n = 1), S. schleiferi (n = 1), S. sciuri (n = 1), S. simiae (n = 1), S. simulans (n = 1), S. succinus (n = 1), S. vitulinus (n = 1), S. xylosus (n = 1).

Table 1.

Strains used for sequence analysis of the nuc genes

Species	Strain	Source	Size of ORF (bp)	GenBank accession no. for sequence of nuc	Reference
S. aureus	N315	Human	534	BA000018	23
S. epidermidis	RP62A	Human	537	CP000029	14
S. haemolyticus	JCSC 1435	Human	537	AP006716	40
S. saprophyticus	ATCC 15305T	Human	531	AP008934	24
S. lugdunensis	ATCC 43809T	Human	543	AB598384	This study
S. warneri	ATCC 27836T	Human	531	AB598385	This study
S. caprae	CCM 3573T	Goat	537	AB598386	This study
S. capitis subsp. capitis	ATCC 27840T	Human	537	AB598387	This study
S. capitis subsp. ureolyticus	JCSC 4868	Human	537	AB598388	This study
S. hominis subsp. hominis	ATCC 27844T	Human	546	AB598389	This study
S. hominis subsp. novobiosepticus	JCSC 3392	Human	546	AB598390	This study
S. pseudintermedius	LMG 22219T	Cat	507	AB327164	34
S. delphini group A	LMG 22190T	Dolphin	516	AB327167	34
S. delphini group B	P-27B	Pigeon	507	AB327166	34
S. intermedius	ATCC 29663T	Pigeon	507	AB327165	34
S. schleiferi subsp. schleiferi	P-45C	Pigeon	504	None	35
S. schleiferi subsp. coagulans	JCM 7470T	Dog	504	AB465334	35
S. hyicus	JCM 2423T	Pig	510	AB465332	35
S. felis	JCM 7469T	Cat	507	AB465335	35
S. chromogenes	P-29B	Pigeon	507	AB465333	35
S. xylosus	h-2B	Horse	537	AB465329	This study
S. kloosii	P-40	Pigeon	543	AB465330	This study
S. simulans	JCM2424T	Human	534	AB465331	This study
S. carnosus	TM300	Dry sausage	543	AM295250	32

All strains used in this study were identified at the species levels by sequence analysis of the hsp60 genes (25), which were determined by high sequence similarity (>95%).

DNA extraction.

DNA was extracted by using the procedure described previously (35). A single colony was suspended to a 1.0 McFarland standard in 100 μl of TE buffer (10 mM Tris, 1 mM EDTA [pH 8.0]) with 10 U of achromopeptidase (Wako Chemical, Co. Ltd.). The suspension was heated at 55°C for at least 10 min until it became transparent. The supernatant was used as template DNA for PCR.

Sequence analysis of the nuc genes.

The nuc genes were sequenced by using the procedure described previously (35). Primers Nuc-alF1 (5′-CCNAAYACNCCNGTNCARCCN-3′) and Nuc-alR (5′-NADCCANACRTANGCNARNGT-3′) were used to amplify the conserved regions of the nuc genes. PCR products were cloned into plasmid pCR-4 I-TOPO (Invitrogen, Life Technologies, Carlsbad, CA) and were transformed into Escherichia coli TOP10 cells (Invitrogen). Insert DNA of the recombinant plasmid was sequenced by using a BigDye Terminator (version 3.1) cycle sequencing kit (Applied Biosystems, Foster City, CA) with an ABI Prism 3100 genetic analyzer (Applied Biosystems). The 5′ and 3′ regions were obtained by inverse PCR, and the complete sequences of the nuc genes were determined. For the species for which a degenerate PCR of the nuc genes was inadequate, degenerate primers Nuc-AsdF (5′-WRNCKRTTCATNARRTAYTT-3′) and AsdR (5′-ACNTAYMGNGARATGMGNGAR-3′) were used to amplify the conserved regions of the aspartate kinase genes, which were located about 2 to 8 kbp upstream of the nuc genes. Downstream sequences were analyzed by inverse PCR in order to determine the complete sequences of the nuc genes. Multiple alignments were carried out by using the CLUSTAL W program (41). Construction of the phylogenetic tree was performed by the neighbor-joining method (33).

M-PCR for species identification of human-associated staphylococci.

As previously described (35), the primer pairs for M-PCR were designed on the basis of nucleotide sequences of the nuc and adjacent genes, and they were specific for each species (Table 2). For M-PCR, the reaction mixture contained 2 μl template DNA, 0.2 μM each primer, 200 μM each deoxynucleotide triphosphate, Ex Taq buffer, and 2 U Ex Taq polymerase (Takara Co., Ltd., Kyoto, Japan), in a final volume of 50 μl. A Takara PCR thermal cycler was used for amplification, with an initial denaturation step (95°C, 5 min); 30 cycles of denaturation (95°C, 30 s), annealing (58°C, 30 s), and extension (72°C, 70 s); and a final elongation step at 72°C for 2 min. PCR products were visualized by electrophoresis in 1× Tris-acetate-EDTA on a 1% agarose gel stained with ethidium bromide.

Table 2.

Oligonucleotide primers for M-PCR for species identification of human-associated staphylococci

primer	Sequence (5′–3′)	Size of PCR product (bp)	Species	Reference
hom-F	TACAGGGCCATTTAAAGACG	177	S. hominis	This study
hom-R	GTTTCTGGTGTATCAACACC
epi-F	TTGTAAACCATTCTGGACCG	251	S. epidermidis	This study
epi-R	ATGCGTGAGATACTTCTTCG
aur-F	TCGCTTGCTATGATTGTGG	359	S. aureus	35
aur-R	GCCAATGTTCTACCATAGC
hae-F	TAGTGGTAGGCGTATTAGCC	434	S. haemolyticus	This study
hae-R	ACGATATTTGCCATTCGGTG
cap-F	ACTACGCCTATGATTATTGC	525	S. capitis	This study
cap-R	GAYGCTTCTTTACCATAGGG
lug-F	TCCAATGATGGTAACGAGGC	695	S. lugdunensis	This study
lug-R	TTTTGCGCCTCGTTTTGTGC
sap-F	TTTTGGATGCGATAGATTGG	843	S. saprophyticus	This study
sap-R	TCTTCAGACTTTTCAAAGGC
war-F	CGTTTGTAGCAAAACAGGGC	999	S. warneri	This study
war-R	GCAACGAGTAACCTTGCCAC
rae-F	TTGTTCTWGCACTYATTGCG	1,227	S. caprae	This study
rae-R	TTTTATAGAACAGGGTCGAC

Nucleotide sequence accession numbers.

GenBank accession numbers for the sequence of the nuc genes determined in this study are as follows: AB598384 for S. lugdunensis, AB598385 for S. warneri, AB598386 for S. caprae, AB598387 for S. capitis subsp. capitis, AB598388 for S. capitis subsp. ureolyticus, AB598389 for S. hominis subsp. hominis, AB598390 for S. hominis subsp. novobiosepticus, AB465329 for S. xylosus, AB465330 for S. kloosii, and AB465331 for S. simulans.

RESULTS

Development of M-PCR for species identification of human-associated staphylococci.

M-PCR successfully amplified the DNA fragments corresponding in size to each species as follows: S. hominis, 177 bp; S. epidermidis, 251 bp; S. aureus, 359 bp; S. haemolyticus, 434 bp; S. capitis, 525 bp; S. lugdunensis, 695 bp; S. saprophyticus, 843 bp; S. warneri, 999 bp; and S. caprae, 1,227 bp (Fig. 1). In order to validate this assay, 361 staphylococcal strains, which had been identified at the species levels by sequence analysis of the hsp60 genes, were studied. For nine human-associated staphylococcal species, the sizes of the amplified DNA fragments matched those predicted for each species, and others showed negative results. The 177-bp fragments contained S. hominis subsp. hominis and S. hominis subsp. novobiosepticus, the 525-bp fragments contained S. capitis subsp. capitis and S. capitis subsp. ureolyticus, and the 843-bp fragments contained S. saprophyticus subsp. saprophyticus and S. saprophyticus subsp. bovis, so this method could not identify these three species at the subspecies levels. Consequently, M-PCR had a sensitivity of 100% and a specificity of 100% in the identification of human-associated staphylococci at the species levels.

Fig. 1.

Electrophoresis after multiplex PCR for species identification of human-associated staphylococci on a 1.0% agarose gel. Lanes: 1, S. hominis; 2, S. epidermidis; 3, S. aureus; 4, S. haemolyticus; 5, S. capitis; 6, S. lugdunensis; 7, S. saprophyticus; 8, S. warneri; 9, S. caprae; 10, water (negative control).

Phylogenetic analysis of the nuc genes of staphylococci.

We constructed the phylogenetic tree based on amino acid sequences of the nuc genes of 24 staphylococcal strains (staphylococcal thermonuclease group), staphylococcal nuclease family proteins of Gram-positive bacteria (staphylococcal nuclease group), and thermonuclease family proteins of thermophilic bacteria (thermophilic bacterial thermonuclease group) (Fig. 2).

Fig. 2.

The phylogenetic tree based on amino acid sequences of the nuc genes of staphylococci (staphylococcal thermonuclease group), staphylococcal nuclease family proteins of Gram-positive bacteria (staphylococcal nuclease group), and thermonuclease family proteins of thermophilic bacteria (thermophilic bacterial thermonuclease group). The tree was constructed by the neighbor-joining method using CLUSTAL W. The genes on plasmids and a prophage are indicated by open circles and a filled circle, respectively.

There were the nuc genes at specific gene loci in 24 staphylococcal strains, which were located about 2 to 8 kbp downstream of the aspartate kinase genes (Table 1). We could not detect this gene in any S. sciuri group (S. sciuri, S. lentus, and S. vitulinus) strains. The staphylococcal thermonuclease group formed two main clusters (Fig. 2). One included S. aureus, and the other included S. intermedius. Phylogenetic relationships of the nuc genes among staphylococci were analogous mostly to those of the hsp60 genes (25). The sequence identity of the nuc genes among staphylococci ranged from 59.6 to 99.6% (mean, 67.2%). The sequence similarity of these genes was 99.6% between two subspecies of S. schleiferi, 98.4% between two subspecies of S. hominis, 89.9% between two subspecies of S. capitis, and 79.5% between S. delphini group A and S. delphini group B. It was 95.9% between S. delphini group B and S. pseudintermedius and 87.7% between S. delphini group B and S. intermedius, so the nuc gene sequence of S. delphini group B was more similar to that of S. pseudintermedius and S. intermedius than that of S. delphini group A (34, 35).

There were staphylococcal nuclease genes in Gram-positive bacteria (2, 21–23, 27, 29, 32, 40) and nuc genes in thermophilic bacteria (1, 9, 30, 38, 43) whose complete genome sequences had been determined. Each group formed a distinct cluster from the staphylococcal thermonuclease group, and some genes were on plasmids or a prophage (Fig. 2).

DISCUSSION

Staphylococci are identified by phenotypic analysis in many clinical laboratories. Although many phenotypic identification methods are commercially available, diagnostic accuracy has been reported to be 36.7 to 93.6% (5, 16, 19, 26). This inaccuracy is problematic in clinical practice. Because CNS are closely related to prosthetic device infections, CNS isolated from blood cultures of previously healthy individuals may be regarded as contaminants. However, S. lugdunensis can cause native valve endocarditis, and community-acquired S. lugdunensis bacteremia has been associated with endocarditis (7, 44). Failure to identify S. lugdunensis might lead to delayed or inadequate treatment with an increase in morbidity and mortality.

Currently, the breakpoint for oxacillin resistance is ≥4 μg/ml in S. aureus and S. lugdunensis, whereas it is ≥0.5 μg/ml in other staphylococci (8). If other staphylococcal strains are misidentified as S. aureus or S. lugdunensis, those showing oxacillin MICs of 0.5 to 2 μg/ml are incorrectly considered to be sensitive, which will result in usage of ineffective antimicrobial agents. If S. aureus or S. lugdunensis strains are misidentified as other staphylococci, those showing oxacillin MICs of 0.5 to 2 μg/ml are falsely considered to be resistant. This indicates that antibacterial agents for methicillin-resistant S. aureus, such as vancomycin, can be used for treatment of methicillin-sensitive S. aureus infection. This is also problematic since vancomycin was previously reported to be inferior to beta-lactam antibiotics in the treatment of methicillin-sensitive S. aureus bacteremia (6, 20, 39). In addition, it has been reported that isolates with oxacillin MICs of 0.5 to 2 μg/ml are mecA negative, particularly in S. saprophyticus and S. warneri (15, 18). This can lead to inappropriate use of antibiotics, such as vancomycin. For these species, the new oxacillin MIC breakpoints need to be determined for methicillin resistance. As a result, it is necessary to precisely identify staphylococci at the species levels.

Little is known about the ecology of CNS despite commensal bacteria. It is thought that S. capitis is seen mostly on the head and that S. saprophyticus is a member of the normal rectal or urogenital flora of some women (31, 42). However, isolates have been identified not genotypically but phenotypically. Although matrix-assisted laser desorption ionization–time of flight mass spectrometry, which is the novel phenotypic method, appears to be sensitive and specific (11), the results are unreliable without the protein extraction step (3). Genetic analysis is needed for definitive species identification. Our assay can also be helpful to elucidate the ecology of CNS in humans.

Sequence analyses of several target genes have been reported for staphylococcal identification, and the mean sequence similarity of 16S rRNA, rpoB, hsp60, sodA, and dnaJ genes was 97%, 86%, 82%, 81%, and 77%, respectively (10, 25, 28, 36). Since the nuc genes showed moderate sequence diversity (mean similarity, 67%) and were generally found in staphylococci except the S. sciuri group, we could develop an M-PCR for species identification of staphylococci based on nucleotide sequences of the nuc and adjacent genes.

The nuc genes were present at specific gene loci in 24 staphylococcal strains but absent in the S. sciuri group. In whole-genome analysis of a Macrococcus caseolyticus strain (2), which is a species closely related to the genus Staphylococcus (25), this strain also did not have the nuc gene. Additionally, there were nuc genes in thermophilic bacteria, and one was on a prophage. These results suggest that the nuc genes might have been derived from thermophilic bacteria and might have been acquired by the common ancestor of staphylococci after divergence from the S. sciuri group.

In summary, our M-PCR has been proved to be a rapid and accurate method for species identification of human-associated staphylococci. This is simple and easy to interpret. This will be useful mainly in clinical research and help to elucidate the ecology of human-associated staphylococci.