Abstract
Meticillin-resistant Staphylococcus aureus (MRSA) remains an important pathogen in the nosocomial setting and is being seen with increasing frequency within the community on a global basis (Fluit et al., 2001; Chambers, 2001). In recent years, there have been a considerable number of studies attempting to address the requirement for a rapid and economical method for determining strain relatedness in both local outbreak investigations and global epidemiological studies. Methods which span the entire genome, such as amplified fragment length polymorphism and PFGE, offer very high resolution and are used for distinguishing strains isolated over short periods of time (Melles et al., 2007), but do not lend themselves to the study of genetic relatedness or evolution. PFGE (the current ‘gold standard’) is criticized for being technically demanding and time-consuming, with difficulty in making interlaboratory comparisons (Sabat et al., 2006). Multilocus sequence typing (MLST), which assigns a strain to a sequence type (ST) based on the sequences of seven housekeeping genes, is often used for global studies (Robinson & Enright, 2004). However, the workload and expense associated with MLST mean that it is often limited to well-resourced facilities.
Robertson et al. (2004) describe an alternative approach involving the detection of a specific set of single nucleotide polymorphisms (SNPs) based on the MLST database using a real-time allele-specific PCR format. The SNP sets were optimized through maximization of Simpson's index of diversity (D) and a set of seven SNPs was identified that provides a D of 0.95 with respect to the S. aureus MLST database (Robertson et al., 2004). Stephens et al. (2006) validated the seven SNP set using 32 MRSA isolates (characterized by MLST) covering 11 STs, which was completely successful in identifying the bases at the SNPs. In addition, the 11 STs were resolved into nine SNP profiles. The two unresolved STs represented single locus variants (SLVs) of clonal complex (CC) progenitors. This is expected in a clonal organism such as S. aureus where there are low levels of diversity and the majority of SLVs have arisen by mutation. For a given SLV, the SNP that discriminates it from the CC progenitor is not polymorphic elsewhere in the species, and so will not be readily discriminated by SNP detection. Stephens et al. (2006) also performed SNP typing using a further 76 isolates and demonstrated that the genotypes defined by the seven high-D SNP set were concordant with the population structure of S. aureus enabling the resolution of CCs corresponding to the major lineages.
We show that the SNP typing described by Robertson et al. (2004) is amenable for use in a standard allele-specific PCR format as well the real-time format already described. In order to overcome the main problem associated with allele-specific PCR, namely the extension from mismatched 3′ ends of oligonucleotides by Taq polymerase lacking 3′ to 5′ exonuclease activity, we used oligonucleotides containing a locked nucleic acid (LNA) at the terminal 3′ position. LNA is a nucleic acid analogue with a 2′-O, 4′-C methylene bridge. This structure locks the ribose moiety into a C3′-endo conformation, which results in an increase in melting temperature when an oligonucleotide containing LNA is hybridized to its complementary sequence (Petersen & Wengel, 2003). Latorra et al. (2003) successfully demonstrated increased specificity using 3′ LNA oligonucleotides in comparison to DNA oligonucleotides for SNP detection.
A small number of S. aureus bacteraemia isolates collected from Imperial College Healthcare NHS Trust covering 14 S. aureus MLST STs (determined according to the method of Enright et al., 2000) were interrogated using a seven SNP set based on those of Robertson et al. (2004). The reverse oligonucleotides described by Robertson et al. (2004) were redesigned to provide a product ∼200 bp in length that could be detected by agarose gel electrophoresis (Table 1⇓). Each PCR was run in a 25 μl volume containing 1× PCR buffer, 200 μM dNTPs, 1.5 mM MgCl2 and 1 U Taq polymerase (Fermentas Life Sciences). Desalted LNA oligonucleotides were obtained from Proligo (Sigma-Genosys) and were used in the concentrations listed in Table 1⇓. PCR was performed using a thermal cycler (MJ Research) using the following parameters: 94 °C for 4 min, followed by 30 cycles of 94 °C for 30 s, 61 °C for 30 s and 72 °C for 30 s, followed by a final extension at 72 °C for 4 min. Products were visualized using a 2 % agarose gel containing ethidium bromide.
Meticillin-resistant Staphylococcus aureus (MRSA) remains an important pathogen in the nosocomial setting and is being seen with increasing frequency within the community on a global basis (Fluit et al., 2001; Chambers, 2001). In recent years, there have been a considerable number of studies attempting to address the requirement for a rapid and economical method for determining strain relatedness in both local outbreak investigations and global epidemiological studies. Methods which span the entire genome, such as amplified fragment length polymorphism and PFGE, offer very high resolution and are used for distinguishing strains isolated over short periods of time (Melles et al., 2007), but do not lend themselves to the study of genetic relatedness or evolution. PFGE (the current ‘gold standard’) is criticized for being technically demanding and time-consuming, with difficulty in making interlaboratory comparisons (Sabat et al., 2006). Multilocus sequence typing (MLST), which assigns a strain to a sequence type (ST) based on the sequences of seven housekeeping genes, is often used for global studies (Robinson & Enright, 2004). However, the workload and expense associated with MLST mean that it is often limited to well-resourced facilities.
Robertson et al. (2004) describe an alternative approach involving the detection of a specific set of single nucleotide polymorphisms (SNPs) based on the MLST database using a real-time allele-specific PCR format. The SNP sets were optimized through maximization of Simpson's index of diversity (D) and a set of seven SNPs was identified that provides a D of 0.95 with respect to the S. aureus MLST database (Robertson et al., 2004). Stephens et al. (2006) validated the seven SNP set using 32 MRSA isolates (characterized by MLST) covering 11 STs, which was completely successful in identifying the bases at the SNPs. In addition, the 11 STs were resolved into nine SNP profiles. The two unresolved STs represented single locus variants (SLVs) of clonal complex (CC) progenitors. This is expected in a clonal organism such as S. aureus where there are low levels of diversity and the majority of SLVs have arisen by mutation. For a given SLV, the SNP that discriminates it from the CC progenitor is not polymorphic elsewhere in the species, and so will not be readily discriminated by SNP detection. Stephens et al. (2006) also performed SNP typing using a further 76 isolates and demonstrated that the genotypes defined by the seven high-D SNP set were concordant with the population structure of S. aureus enabling the resolution of CCs corresponding to the major lineages.
We show that the SNP typing described by Robertson et al. (2004) is amenable for use in a standard allele-specific PCR format as well the real-time format already described. In order to overcome the main problem associated with allele-specific PCR, namely the extension from mismatched 3′ ends of oligonucleotides by Taq polymerase lacking 3′ to 5′ exonuclease activity, we used oligonucleotides containing a locked nucleic acid (LNA) at the terminal 3′ position. LNA is a nucleic acid analogue with a 2′-O, 4′-C methylene bridge. This structure locks the ribose moiety into a C3′-endo conformation, which results in an increase in melting temperature when an oligonucleotide containing LNA is hybridized to its complementary sequence (Petersen & Wengel, 2003). Latorra et al. (2003) successfully demonstrated increased specificity using 3′ LNA oligonucleotides in comparison to DNA oligonucleotides for SNP detection.
A small number of S. aureus bacteraemia isolates collected from Imperial College Healthcare NHS Trust covering 14 S. aureus MLST STs (determined according to the method of Enright et al., 2000) were interrogated using a seven SNP set based on those of Robertson et al. (2004). The reverse oligonucleotides described by Robertson et al. (2004) were redesigned to provide a product ∼200 bp in length that could be detected by agarose gel electrophoresis (Table 1⇓). Each PCR was run in a 25 μl volume containing 1× PCR buffer, 200 μM dNTPs, 1.5 mM MgCl2 and 1 U Taq polymerase (Fermentas Life Sciences). Desalted LNA oligonucleotides were obtained from Proligo (Sigma-Genosys) and were used in the concentrations listed in Table 1⇓. PCR was performed using a thermal cycler (MJ Research) using the following parameters: 94 °C for 4 min, followed by 30 cycles of 94 °C for 30 s, 61 °C for 30 s and 72 °C for 30 s, followed by a final extension at 72 °C for 4 min. Products were visualized using a 2 % agarose gel containing ethidium bromide.
Oligonucleotides used for SNP detection
The results obtained using the LNA oligonucleotides for SNP typing using the seven SNP set (Table 2⇓ and Fig. 1⇓) are concordant with those expected after analysis of the MLST database, demonstrating that LNA oligonucleotides are capable of the reliable and unambiguous detection of specific SNPs.
Agarose gel electrophoresis of products of arcC162. (a) arcC162 (T). Lanes: 1, 100 bp ladder; 2, ST1; 3, ST5; 4, ST7; 5, ST8; 6, ST12; 7, ST22; 8, ST30; 9, ST36; 10, ST45; 11, ST188; 12, ST199; 13, ST221; 14, ST278; 15, ST361; 16, PCR-negative; 17, 100 bp ladder. (b) arcC162 (A). Lane order as for (a).
Summary of MLST and SNP typing results
We conclude that LNA oligonucleotides are a useful tool for the detection of high D SNP sets that can be used for the designation of S. aureus isolates into CCs. As a result, this method may be used as an initial screen in the event of a local outbreak investigation to identify whether isolates are likely to be related, belong to a CC of interest or belong to an emerging lineage in a manner more timely and less costly than a full MLST profile, whilst still using a common language enabling comparison with other local and global studies. For example, the vast majority of MRSA isolates from hospitals in England and Wales belong to EMRSA-15 (ST22, CC22) and EMRSA-16 (ST36, CC30); this method would immediately identify any MRSA isolates not belonging to either of these two lineages.