2 chromosome mec 3 accepted -...
TRANSCRIPT
1
Systematic derivation of marker sets for Staphylococcal Cassette 1
Chromosome mec typing. 2
3
4
Alex J. Stephens, Flavia Huygens and Philip M. Giffard* 5
6
7
8
9
10
11
12
13
*Corresponding Author. 14
Philip M. Giffard 15
Cooperative Research Centre for Diagnostics 16
Institute of Health and Biomedical Innovation 17
Queensland University of Technology 18
60 Musk Ave 19
Kelvin Grove QLD 4059 20
Australia 21
Email [email protected] 22
Tel 61 7 3138 6194 23
Fax 61 7 3138 6030 24
25
Running title: Identifying markers for SCCmec genotyping 26
27
Keywords: Methicillin Resistant Staphylococcus aureus, MRSA, Staphylococcal 28
Cassette Chromosome mec, SCCmec, genotyping, Minimum SNPs, Bioinformatics. 29
ACCEPTED
Copyright © 2007, American Society for Microbiology and/or the Listed Authors/Institutions. All Rights Reserved.Antimicrob. Agents Chemother. doi:10.1128/AAC.01323-06 AAC Accepts, published online ahead of print on 21 May 2007
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
2
Abstract 1
2
The aim of this study was to identify optimised sets of genotyping targets for the 3
Staphylococcal Cassette Chromosome mec (SCCmec). We analysed the gene contents 4
of 46 SCCmec variants in order to identify minimal subsets of targets that provide 5
useful resolution. This was achieved by firstly identifying and characterising each 6
available SCCmec element based on the presence or absence of 34 binary targets. This 7
information was used as input for the software “Minimum SNPs”, which identifies the 8
minimum number of targets required to differentiate a set of genotypes up to a 9
predefined Simpson’s index of diversity (D) value. It was determined that 22 of the 34 10
targets were required to genotype the 46 SCCmec variants to a D of 1. The first 6, 9, 11
12 and 15 targets were found to define 21, 29, 35 and 39 SCCmec variants, 12
respectively. The genotypes defined by these marker subsets were largely consistent 13
with the relationships between SCCmec variants and the accepted nomenclature. 14
Consistency was made virtually complete by forcing the computer program to include 15
ccr1 and ccr5 in the target set. An alternative target set biased towards discriminating 16
abundant SCCmec variants was derived by analysing an input file in which common 17
SCCmec variants were repeated, thus ensuring that markers that discriminate 18
abundant variants have a large effect on D. Finally, it was determined that mecA 19
single nucleotide polymorphisms (SNPs) can increase the overall genotyping 20
resolution, as different mecA alleles were found in otherwise identical SCCmec 21
variants. 22
23
24
25
26
27
28
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
3
Introduction 1 2
Comparative bacterial genomics is revealing numerous hypervariable regions in 3
bacterial chromosomes. Interrogation of such regions can efficiently provide an 4
epidemiological fingerprint, or insight into pathogenic capability, antimicrobial 5
resistance phenotype, or vaccine susceptibility (7, 9, 11). However, the often complex 6
nature of the variation can make it difficult to devise standardised protocols for sub-7
typing such regions. 8
9
An excellent example of a clinically relevant hypervariable region in a major bacterial 10
pathogen is the Staphylococcal cassette chromosome mec (SCCmec) (16, 19). This 11
element defines methicillin-resistant Staphylococcus aureus (MRSA) through carriage 12
of the beta-lactam resistance gene mecA. A combination of multi-locus sequence type 13
(MLST) and SCCmec type is frequently used as an identifier for MRSA clones (8, 31, 14
36). As a consequence, in recent years there has been considerable interest in 15
developing SCCmec sub-typing methods. 16
17
A number of broadly similar but distinct SCCmec typing methods have been 18
described. These methods include; PCR-based detection of type specific targets (ccr 19
variants, mec gene complexes and junkyard regions (10, 29, 30, 46), restriction digests 20
(43, 45) and full sequence analysis (15-17). In general, SCCmec is divided into five 21
major types based on the combination of mec class and recombinase-encoding (ccr) 22
genes present (15, 17, 25). The mec classes are themselves defined by the 23
arrangement of genes adjacent to mecA (38). Subtypes are defined by binary variation 24
(presence or absence) of sequence blocks in the “junkyard” regions (4). The various 25
typing methods and associated terminology have evolved in a somewhat ad hoc 26
fashion as more SCCmec variants have been discovered. Recently, a proposal for a 27
rationalised SCCmec typing nomenclature was published (4). This proposal has merit 28
as it incorporates the core structural features and the variable junkyard regions. 29
However, there is currently no standardised SCCmec genotyping method that 30
systematically and efficiently utilises all known variation. 31
32
Our research group has previously developed the computer program “Minimum 33
SNPs” (35). This was designed to derive sets of single nucleotide polymorphisms 34
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
4
(SNPs) from MLST databases on the basis of maximisation of resolving power. 1
Resolving power may be measured either on the basis of the power to discriminate a 2
user-defined sequence type (ST) from all other STs, or, on the basis of maximisation 3
of the Simpson’s Index of Diversity (D) (13). More recently, this approach was 4
adapted to the identification of resolution optimised sets of binary markers. Price and 5
co-workers derived a set of such markers from Campylobacter jejuni comparative 6
genome hybridization data on the basis of D maximisation, and these markers were 7
shown to have considerable utility as genotyping targets (33). SCCmec diversity can 8
also be considered to be a database of binary gene variation, and is therefore amenable 9
to a similar analysis. Accordingly, we have carried this out, with the central 10
hypothesis of the study being that the sets of markers derived by our systematic 11
approach would efficiently provide resolving power. 12
13
14
Materials and Methods 15
16
Identification of SCCmec variants 17
18
A literature and NCBI database search was undertaken to identify SCCmec variants 19
for this study. Variants were named according to the proposed SCCmec nomenclature 20
system (4). This procedure was complicated by the fact that described SCCmec 21
variants differ in the detail to which they have been analysed, some are completely 22
sequenced, while others are classified only though PCR amplification. In a small 23
number of instances, SCCmec variants reported in the literature lacked sufficient 24
structural information to be included in this study. 25
26
Identification of resolution optimised sets of binary markers. 27
28
Sets of binary markers were identified using the computer program ‘Minimum SNPs” 29
by a strategy illustrated in Fig. 1 (33, 35). This program extracts resolution optimised 30
sets of SNPs from DNA sequence alignments. It does this by identifying the single 31
SNP with the highest resolving power, labelling this as SNP 1, then identifying the 32
SNP that in combination with SNP 1 gives the highest resolving power, and labelling 33
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
5
that as SNP 2 etc. A valuable feature of the software is that the user can force the 1
program to include or exclude any SNP in/from the SNP set. 2
3
Minimum SNPs can measure resolving power in more than one way, but the most 4
generally applicable method is by calculation of D with respect to the sequence 5
alignment. This algorithm was used throughout this study. Resolution optimised sets 6
of binary markers were identified by first converting the binary marker data for the 7
SCCmec variants into a string of “A’s” and “T’s”, with “A” denoting binary maker 8
absence and “T” denoting binary marker presence. In this way the binary data for each 9
SCCmec variant becomes a pseudo-DNA sequence that can be aligned with other 10
pseudo-DNA sequences representing other SCCmec variants. This alignment was then 11
mined by Minimum SNPs in order to identify sets of binary markers that give a high 12
D value with respect to that alignment. The alignment input file reflecting estimated 13
SCCmec abundance included 49 extra copies of variants: 1B.1.1 (I), 2A.1.1 (II), 14
3A1.1.1 (III), 3A.1.2 (IIIA), 3A.1.3 (IIIB), 2B.1 (IVa), 2B.2.1 (IVb), 2B.3.1 (IVc) and 15
5C.1 (V) 16
17
mecA nucleotide sequence determination and SNP analysis 18
The mecA genes from 19 diverse Australian MRSA isolates from nine MLST types 19
were amplified using primers mecAF1 and mecAR3 (Sigma-Proligo, Lismore, 20
Australia) (39, 41). The amplicons were purified using Exo-SapIt (Amersham 21
Biosciences, Castle Hill, Australia) for 15 min at 37oC and 15 min at 80
oC, then 22
sequenced for 1400bp from the 3` end using primers mecAF1 and mecAR2. The 23
sequence traces were viewed and analysed in SeqMan II version 4.06 (DNAstar, 24
Madison, USA). 25
26
The literature and NCBI databases were searched for Staphylococcus aureus mecA 27
gene sequences whose corresponding SCCmec type was known. Twenty-five mecA 28
sequences were identified, cropped to 1440bp and added to a sequence alignment with 29
the 19 partial mecA sequences identified from this study. Clustal X (v1.64b) was used 30
to align the partial mecA gene sequences to identify SNPs. In order to investigate the 31
possibility of combinatorial genotyping methods based on SNPs and binary genes, the 32
SNP data were added to the pseudo-DNA sequences derived from the binary marker 33
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
6
genotypes, so as to construct an input file for Minimum SNPs that would allow the 1
identification of resolution optimised SNP-binary marker combinations. 2
3
Full details of the mecA sequences used in this analysis and the isolates from which 4
they were derived are available on line as supplementary data at 5
http://www.ihbi.qut.edu.au/research/cells_tissue/phil_giffard/. All new mecA 6
sequences have been submitted to Genbank. 7
8
Results 9
10
Collation of currently available data concerning SCCmec binary diversity 11
12
The overall aim of this study was to apply a systematic approach for the identification 13
of genetic targets to efficiently genotype SCCmec. In order to do this effectively, it 14
was necessary to extract all available SCCmec data from the literature and publicly 15
available databases. Diversity can be in the form of binary variability (gene 16
presence/absence) or single nucleotide polymorphisms (SNPs). As complete 17
sequences are known for only a subset of the genes that make up all the known 18
SCCmec variants, binary diversity was collated and analysed first. 19
20
Analysis of the literature and online databases resulted in 34 SCCmec-associated 21
binary targets being defined (Table 1). So as to facilitate the design of genotyping 22
methods, unique sequences associated with the targets together with explanatory notes 23
are provided as supplementary data at 24
http://www.ihbi.qut.edu.au/research/cells_tissue/phil_giffard/. The 34 binary targets 25
in turn defined 46 SCCmec variants (Table 2). Each SCCmec element differing at one 26
or more of the 34 binary targets (Table 1) was considered a separate variant. This 27
extensive compilation of SCCmec variants reveals the extensive rearrangement and 28
mutation the five major types have undergone during their evolutionary histories, with 29
diversity in the junkyard regions defining the majority of the variants 30
31
Identification of a resolution-optimised set of binary targets, without constraints 32
33
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
7
The database of binary gene variation was converted into a pseudo-sequence 1
alignment (see Methods) and analysed using Minimum SNPs for sets of markers that 2
provide a high D. For the first experiment, a set of binary markers was derived from 3
the data in Table 2 simply on the basis of maximisation of D, with no attempt to make 4
the results consistent with previously described typing methods. It was determined 5
that all 46 SCCmec variants can be completely resolved by interrogating 22 of the 34 6
binary targets. Table 3 shows the resolving powers of the best 6, 9, 12 and 15 target 7
sets, which differentiate 21, 29, 35 and 39 of the 46 variants, respectively. 8
9
These targets were then compared with those used in currently extant SCCmec typing 10
methods, with respect to their identity, resolving power, and whether the variants they 11
define correspond with current SCCmec classification schemes. Current PCR-based 12
SCCmec typing generally make use of 10 targets (ccr 1, 2, 3, 5, mec classes A, B, C 13
and three junkyard regions from type 2B) to assign a SCCmec element to one of the 14
five major types, and also provide some sub-typing information. When tested against 15
the dataset containing 46 variants, these 10 targets discriminated 18 genotypes. In 16
comparison, the first 10 targets derived on the basis of D maximisation discriminated 17
32 genotypes. This supports our conjecture that binary target identification on the 18
basis of computerised D maximisation will provide a superior result to an ad hoc 19
approach. In order to further test this, the resolving powers of randomly selected sets 20
of markers were determined. This always gave a much lower resolving power than 21
marker sets selected on the basis of D maximisation (data not shown), thus supporting 22
the utility of our method for target selection. 23
24
A resolution-optimised set of binary targets nucleated by ccr1 and ccr5 25
26
It would be desirable for a SCCmec genotyping method to define genotypes consistent 27
with accepted terminology, and current models concerning the degrees of relatedness 28
of the different SCCmec variants. In other words, a genotyping method that fails to 29
discriminate some pairs of distantly related SCCmec variant may be regarded as 30
problematic. The genotypes defined by the D maximised marker set largely met this 31
requirement. However, the 6 and 9 target sets were unable to discriminate the 32
unrelated 4B and truncated SCCmec from 1B (I) variants, and also the 5/2G.1 and 33
2C.1 from 2B (IV) variants (Table 3). Accordingly, a second set of resolution 34
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
8
optimised targets were derived using the “Include” function in Minimum SNPs. This 1
forces the program to include a chosen marker(s) in the set. In this instance, ccr1 and 2
ccr5 were forced into the set, because these discriminate the variants that are not 3
discriminated with the original marker set. In this analysis, Minimum SNPs builds 4
markers sets using the set of ccr1 plus ccr5 as a starting point, so in effect the derived 5
marker sets are nucleated by ccr1 and ccr5. 6
7
The derived marker sets are very similar to the unconstrained set and provided almost 8
identical resolution. However, the occurrence of incorrectly grouped variants was 9
largely rectified. The only instances of disparate variants being grouped together were 10
2C.1 with 2B variants and 2B.N.2 with 2A variants when using the six target set 11
(Table 4). 12
13
A resolution optimised target set derived from an input file reflecting SCCmec 14
variant abundances 15
16
An assumption inherent in our approach to identifying resolution-optimised sets of 17
genotyping targets is that the database is a useful surrogate of the population structure. 18
If this is the case, then the resolving power of the targets with respect to the database 19
is a useful measure of their resolving power on actual collections of isolates. This 20
assumption, although unlikely to be completely wrong, may be simplistic. This is 21
because all the SCCmec variants are not similarly abundant. This could result in a 22
difference between the D value calculated from the database, and the D value 23
obtained from actual collections of isolates. A corollary of this is that markers that 24
discriminate between the abundant genotypes should be preferentially included in the 25
marker set if the D for actual collections of isolates is to be maximised. 26
27
We addressed this issue by creating an input file for “Minimum SNPs” that contains 28
multiple copies of abundant SCCmec variants. A comprehensive and accurate 29
determination of relative abundances in nature was not practical, so this exercise was 30
carried out somewhat crudely; the published literature was used to make a judgement 31
as to which SCCmec variants are abundant, and these variants were repeated 50 times 32
in the Minimum SNPs input file. It was hypothesised that the derived marker sets 33
would provide a very high performance at discriminating the abundant variants 34
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
9
because a marker that discriminated between abundant variants would have a big 1
impact on the D value. It was also predicted that the derived markers would still be 2
effective for discriminating the less abundant variants from the abundant variants and 3
from each other. 4
5
The resolution optimised marker sets from this exercise are shown in Table 5. The 6
target sets identified are significantly different to the unconstrained and ccr1/ccr5-7
nucleated target sets, and interestingly, do not discriminate as many genotypes. 8
However, as expected, this marker set provided a superior performance to the 9
unconstrained and ccr1/ccr5-nucleated marker sets in resolving the nine SCCmec 10
variants identified as being abundant. There were, however, a small number of 11
instances where the six target set failed to discriminate rare SCCmec variants from 12
other unrelated SCCmec variants. This was partially rectified by manually promoting 13
the thirteenth identified target, ccr1, into the six target set. This discriminates 4B and 14
the aberrant variant we have termed “truncated” from the abundant unrelated 15
variant1B.1.1. The resulting seven target set is what in shown in column 1 of Table 5. 16
This target set does not discriminate 5/2G.1 and 2C.1 from the abundant unrelated 17
variant 2B3.1. However, the 11 target set does discriminate all these unrelated 18
variants. 19
20
To further test the efficacy of this approach, a separate dataset containing only the 21
nine abundant variants was assembled and analysed. “Minimum SNPs” calculated that 22
six of the 34 targets were required to attain a D of 1. These six targets were very 23
different from the marker set in Table 5, provided poor resolving power with the 24
entire data set, and there were numerous instances of unrelated variants failing to be 25
discriminated (data not shown). Therefore, the approach of using the entire data set, 26
with multiple copies of abundant genotypes was superior. 27
28
29
mecA SNPs can increase resolving power 30
31
It has previously been reported that the MRSA mecA gene contains several SNPs (37, 32
39, 44). These are potential genotyping markers that could be used either as 33
replacements for the binary targets selected, or to define more SCCmec subtypes. The 34
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
10
literature and sequence databases were searched for mecA sequences from 1
characterised SCCmec elements. In total, 25 sequences were retrieved from the NCBI 2
database, and in addition, partial mecA sequence analysis was undertaken on 19 3
selected isolates from our collection of previously characterised SCCmec elements 4
(41). From this set of 44 partial mecA sequences, eight SNPs in total were identified, 5
three of which are novel (Table 6). 6
7
To determine whether mecA SNPs add resolution to the binary genes, they were 8
included in a Minimum SNPs analysis as additional data points. Because no precise 9
mecA sequence data could be assigned to many SCCmec variants, a new SCCmec 10
data set was created containing only variants with corresponding mecA sequences. In 11
this case, the data used as input into Minimum SNPs was in effect an alignment that 12
contained a region of pseudo-DNA sequence derived from binary gene variation, and 13
a region of actual DNA sequence, derived from mecA sequences. This dataset 14
consisted of 18 of the original 46 SCCmec variants (Table 6). As different mecA SNP 15
profiles were found within identical SCCmec variants, the 34 binary markers in 16
combination with the eight SNPs defined 24 variants. These expanded mixed 17
SNP/binary target profiles were analysed using “Minimum SNPs”, which revealed 18
that a D of 1.0 was achieved with 9 binary targets and five mecA SNPs. Overall, of 19
these five SNPs only SNP 737 could be considered as a possible replacement for the 20
binary targets, as it was selected at the third position in the Minimum SNPs output. 21
The remaining SNPs were selected at positions 9 and 12-14, which demonstrates 22
minimal D-value contribution. The entire marker set is ccr2, Tn554 MLS, mecA737, 23
mec class A, dcs, pT181, ccrC, Jyb, mecA438, ccrC-VT, Tn4001, mecA75, mecA415, 24
mecA448 25
26
27
Discussion 28
29
It is now generally accepted that a powerful strategy for bacterial genotyping is to 30
interrogate the genome backbone plus one or more hypervariable regions. This 31
generalised approach has been termed “Phylogenetic Hierarchical Assays using 32
Nucleic Acids” (20). The practice of identifying MRSA clones by a combination of 33
MLST and SCCmec type is an example of this. The widespread adoption of this 34
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
11
MRSA genotyping method has proven an impetus to the accumulation of considerable 1
information regarding SCCmec diversity, and the development of several specific 2
SCCmec sub-typing schemes. 3
4
SCCmec has been classified into five major “types”, composed of mec classes and ccr 5
gene identity, and several SCCmec genotyping methods classify the element to this 6
level only (10, 45). The results of our comprehensive search of the literature and 7
databases emphasised the high diversity of SCCmec and revealed that much of the 8
diversity is invisible to previously published SCCmec typing methods. Additional to 9
the initial characterisation of types 1B, 2A and 3A (I, II, and III) (15), the community 10
acquired type 2B (IV) (25) and the most recent type 5C (V) (17), more variable and 11
unusual variants have been identified. Of particular interest are variants carrying 12
multiple copies of ccr and instances of new combinations of mec and ccr classes (2, 13
28). 14
15
Our approach to identifying sets of binary markers for genotyping SCCmec differs 16
greatly from previously published methods (21). Genotyping approaches reported to 17
date yield approximately one mec class or ccr gene per marker interrogated. This is 18
needlessly inefficient as the extensive recombination in SCCmec means that resolving 19
power can potentially increase logarithmically rather than arithmetically as more 20
targets are interrogated. Accordingly, rather than identifying binary markers 21
diagnostic for particular classes/types/subtypes, we used a computerised approach that 22
identifies sets of markers that maximise D. What this algorithm does is to attempt to 23
identify a marker that splits the known variants into two equal halves, and then 24
attempts to find a single marker that splits each of the groups defined by marker 1 into 25
two equal halves, and so forth. In effect, it is a search for markers that are maximally 26
unlinked. This approach proved to be valuable. It provided sets of markers with 27
greater resolving power than equivalently sized sets of markers identified by the 28
traditional approach. Also, the analyses clearly defined a range of options regarding 29
the numbers of markers interrogated and resolution obtained. These results bore out 30
the predication that resolution could increase exponentially with the number of targets 31
interrogated; when the numbers of targets was graphed against log(1-D), the points 32
formed a straight line (data not shown). 33
34
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
12
The markers identified have obvious potential to inform the design of specific 1
SCCmec genotyping methods around the multiplexing capacity of the technology to 2
be used and/or the resolution required. One intriguing result was the strong 3
consistency between the SCCmec classes and the groups defined by small numbers of 4
resolution optimised marker sets, even when there were no constraints on marker 5
selection. The fact that maximally unlinked markers define these groups supports the 6
notion that the SCCmec classes indeed represent distinct phylogenetic clusters of this 7
element. The strategy to increase the consistency between genotypes and the 8
relationships between the SCCmec variants by nucleating the marker set with ccr1 9
and ccr5 proved successful, as was the analysis to identify markers especially efficient 10
at discriminating the abundant SCCmec variants. Overall, our approach to marker 11
selection proved effective and flexible. The marker sets identified provide a wide 12
choice of well-understood options for the design of SCCmec genotyping procedures. 13
This general approach could be applied to any genome region displaying a high 14
degree of binary variability e.g. the loci encoding the enzymes that synthesise 15
complex antigenically active cell-wall associated polysaccharides. 16
17
Rapid interpretation of the results of a genotyping procedure based on the marker sets 18
described here is potentially problematic. It can be done by manually correlating the 19
results with the information in Table 2, but this is quite laborious. However, it can 20
also be done more rapidly using Minimum SNPs, which is able to run in reverse and 21
return all sequences in an alignment that correspond to a user-defined SNP profile. 22
23
The utility of including mecA SNPs in the marker set was explored, and it was found 24
that such SNPs define additional SCCmec variants. However, the SNPs are not 25
effective substitutes for binary markers. Of interest were the observations that 26
identical SNP allelic combinations were found in different SCCmec variants while 27
conversely, different mecA alleles were found in identical SCCmec variants. For 28
example, the first four SCCmec classes (1B, 2A, 3A and 2B) each carry the most 29
abundant mecA sequence, while seven of the nine mecA sequences are found in type 30
3A. This suggests that some SCCmec variants are very much older than others, or that 31
there has been recombination between different SCCmec variants. Other observations 32
of interest include variant 2B.1 (IVa) exclusively carrying the most numerous mecA 33
sequence, which is different to the mecA sequence in the closely related variant 2B.3.1 34
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
13
(IVc), and the complete linkage found between ccr5 and a particular mecA sequence 1
(Table 6). It was also observed that two SNP profiles were primarily associated with 2
mec class B, while seven SNP profiles were associated with mec classes A and C. It 3
was concluded that mecA SNPs constitute a possible means of increasing resolving 4
power, with inclusion of the SNP data increasing the number of genotypes. However, 5
the mecA SNPs have only a limited potential use as replacements for the binary 6
markers, especially considering that it is more technically straightforward to assay for 7
the presence of a binary marker, than to interrogate a SNP. 8
9
10
Recently, our research group used Minimum SNPs to derive sets of seven (35, 41) or 11
eight (14, 26) resolution optimised SNPs from the S. aureus MLST database. The 12
genotypes defined by these SNP sets display a high level of consistency with the S. 13
aureus clonal complex structure as revealed by eBURST analysis of the MLST 14
database. A combinatorial MRSA typing method that interrogated these SNPs plus 15
e.g. the first seven targets in Table 5 (ccr1, dcs, Tn554 MLS, ccr2, mec class A, J1b 16
and pT181) would be a method of MRSA clone identification that would provide 17
sufficient resolution for many purposes, and would be easily adaptable to standardised 18
real-time PCR, medium density array, or “lab on a chip” technologies. Additional 19
resolution for exploring small scale epidemiological questions (e.g. in the practice of 20
infection control) could be provided by interrogating one or more variable number 21
tandem repeat loci. This also could be done in a real time PCR device by Tm 22
measurement. 23
24
In conclusion, this study has demonstrated the effectiveness and flexibility of a 25
systematic computer assisted approach to deriving genotyping targets from loci 26
displaying high levels of complex binary gene variation. 27
28
Acknowledgements 29
The authors thank E.P Price for scientific advice. This work was supported by the 30
Cooperative Research Centres Program of the Australian Federal Government. 31
32
References 33
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
14
1
1. Aires de Sousa, M., and H. de Lencastre. 2003. Evolution of sporadic 2
isolates of methicillin-resistant Staphylococcus aureus (MRSA) in hospitals 3
and their similarities to isolates of community-acquired MRSA. J Clin 4
Microbiol 41:3806-15. 5
2. Boyle-Vavra, S., B. Ereshefsky, C. C. Wang, and R. S. Daum. 2005. 6
Successful multiresistant community-associated methicillin-resistant 7
Staphylococcus aureus lineage from Taipei, Taiwan, that carries either the 8
novel Staphylococcal chromosome cassette mec (SCCmec) type VT or 9
SCCmec type IV. J Clin Microbiol 43:4719-30. 10
3. Cha, H. Y., D. C. Moon, C. H. Choi, J. Y. Oh, Y. S. Jeong, Y. C. Lee, S. Y. 11
Seol, D. T. Cho, H. H. Chang, S. W. Kim, and J. C. Lee. 2005. Prevalence 12
of the ST239 clone of methicillin-resistant Staphylococcus aureus and 13
differences in antimicrobial susceptibilities of ST239 and ST5 clones 14
identified in a Korean hospital. J Clin Microbiol 43:3610-4. 15
4. Chongtrakool, P., T. Ito, X. X. Ma, Y. Kondo, S. Trakulsomboon, C. 16
Tiensasitorn, M. Jamklang, T. Chavalit, J. H. Song, and K. Hiramatsu. 17
2006. Staphylococcal cassette chromosome mec (SCCmec) typing of 18
methicillin-resistant Staphylococcus aureus strains isolated in 11 Asian 19
countries: a proposal for a new nomenclature for SCCmec elements. 20
Antimicrob Agents Chemother 50:1001-12. 21
5. Deurenberg, R. H., C. Vink, G. J. Oudhuis, J. E. Mooij, C. Driessen, G. 22
Coppens, J. Craeghs, E. De Brauwer, S. Lemmen, H. Wagenvoort, A. W. 23
Friedrich, J. Scheres, and E. E. Stobberingh. 2005. Different clonal 24
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
15
complexes of methicillin-resistant Staphylococcus aureus are disseminated in 1
the Euregio Meuse-Rhine region. Antimicrob Agents Chemother 49:4263-71. 2
6. Donnio, P. Y., D. C. Oliveira, N. A. Faria, N. Wilhelm, A. Le Coustumier, 3
and H. de Lencastre. 2005. Partial excision of the chromosomal cassette 4
containing the methicillin resistance determinant results in methicillin-5
susceptible Staphylococcus aureus. J Clin Microbiol 43:4191-3. 6
7. Dorrell, N., J. A. Mangan, K. G. Laing, J. Hinds, D. Linton, H. Al-7
Ghusein, B. G. Barrell, J. Parkhill, N. G. Stoker, A. V. Karlyshev, P. D. 8
Butcher, and B. W. Wren. 2001. Whole genome comparison of 9
Campylobacter jejuni human isolates using a low-cost microarray reveals 10
extensive genetic diversity. Genome Res 11:1706-15. 11
8. Enright, M. C., D. A. Robinson, G. Randle, E. J. Feil, H. Grundmann, and 12
B. G. Spratt. 2002. The evolutionary history of methicillin-resistant 13
Staphylococcus aureus (MRSA). Proc Natl Acad Sci U S A 99:7687-92. 14
9. Fitzgerald, J. R., D. E. Sturdevant, S. M. Mackie, S. R. Gill, and J. M. 15
Musser. 2001. Evolutionary genomics of Staphylococcus aureus: insights into 16
the origin of methicillin-resistant strains and the toxic shock syndrome 17
epidemic. Proc Natl Acad Sci U S A 98:8821-6. 18
10. Francois, P., G. Renzi, D. Pittet, M. Bento, D. Lew, S. Harbarth, P. 19
Vaudaux, and J. Schrenzel. 2004. A novel multiplex real-time PCR assay for 20
rapid typing of major staphylococcal cassette chromosome mec elements. J 21
Clin Microbiol 42:3309-12. 22
11. Hakenbeck, R., N. Balmelle, B. Weber, C. Gardes, W. Keck, and A. de 23
Saizieu. 2001. Mosaic genes and mosaic chromosomes: intra- and interspecies 24
genomic variation of Streptococcus pneumoniae. Infect Immun 69:2477-86. 25
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
16
12. Hisata, K., K. Kuwahara-Arai, M. Yamanoto, T. Ito, Y. Nakatomi, L. Cui, 1
T. Baba, M. Terasawa, C. Sotozono, S. Kinoshita, Y. Yamashiro, and K. 2
Hiramatsu. 2005. Dissemination of methicillin-resistant staphylococci among 3
healthy Japanese children. J Clin Microbiol 43:3364-72. 4
13. Hunter, P. R., and M. A. Gaston. 1988. Numerical index of the 5
discriminatory ability of typing systems: an application of Simpson's index of 6
diversity. J Clin Microbiol 26:2465-6. 7
14. Huygens, F., J. Inman-Bamber, G. R. Nimmo, W. Munckhof, J. 8
Schooneveldt, B. Harrison, J. A. McMahon, and P. M. Giffard. 2006. 9
Staphylococcus aureus Genotyping Using Novel Real-Time PCR Formats. J 10
Clin Microbiol 44:3712-9. 11
15. Ito, T., Y. Katayama, K. Asada, N. Mori, K. Tsutsumimoto, C. 12
Tiensasitorn, and K. Hiramatsu. 2001. Structural comparison of three types 13
of staphylococcal cassette chromosome mec integrated in the chromosome in 14
methicillin-resistant Staphylococcus aureus. Antimicrob Agents Chemother 15
45:1323-36. 16
16. Ito, T., Y. Katayama, and K. Hiramatsu. 1999. Cloning and nucleotide 17
sequence determination of the entire mec DNA of pre-methicillin-resistant 18
Staphylococcus aureus N315. Antimicrob Agents Chemother 43:1449-58. 19
17. Ito, T., X. X. Ma, F. Takeuchi, K. Okuma, H. Yuzawa, and K. Hiramatsu. 20
2004. Novel type V staphylococcal cassette chromosome mec driven by a 21
novel cassette chromosome recombinase, ccrC. Antimicrob Agents Chemother 22
48:2637-51. 23
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
17
18. Ito, T., K. Okuma, X. X. Ma, H. Yuzawa, and K. Hiramatsu. 2003. 1
Insights on antibiotic resistance of Staphylococcus aureus from its whole 2
genome: genomic island SCC. Drug Resist Updat 6:41-52. 3
19. Katayama, Y., T. Ito, and K. Hiramatsu. 2000. A new class of genetic 4
element, staphylococcus cassette chromosome mec, encodes methicillin 5
resistance in Staphylococcus aureus. Antimicrob Agents Chemother 44:1549-6
55. 7
20. Keim, P., M. N. Van Ert, T. Pearson, A. J. Vogler, L. Y. Huynh, and D. 8
M. Wagner. 2004. Anthrax molecular epidemiology and forensics: using the 9
appropriate marker for different evolutionary scales. Infect Genet Evol 4:205-10
13. 11
21. Kondo, Y., T. Ito, X. X. Ma, S. Watanabe, B. N. Kreiswirth, J. Etienne, 12
and K. Hiramatsu. 2007. Combination of Multiplex PCRs for Staphylococcal 13
Cassette Chromosome mec Type Assignment: Rapid Identification System for 14
mec, ccr, and Major Differences in Junkyard Regions. Antimicrob Agents 15
Chemother 51:264-74. 16
22. Kwon, N. H., K. T. Park, J. S. Moon, W. K. Jung, S. H. Kim, J. M. Kim, 17
S. K. Hong, H. C. Koo, Y. S. Joo, and Y. H. Park. 2005. Staphylococcal 18
cassette chromosome mec (SCCmec) characterization and molecular analysis 19
for methicillin-resistant Staphylococcus aureus and novel SCCmec subtype 20
IVg isolated from bovine milk in Korea. J Antimicrob Chemother 56:624-32. 21
23. Lim, T. T., F. N. Chong, F. G. O'Brien, and W. B. Grubb. 2003. Are all 22
community methicillin-resistant Staphylococcus aureus related? A comparison 23
of their mec regions. Pathology 35:336-43. 24
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
18
24. Lim, T. T., G. W. Coombs, and W. B. Grubb. 2002. Genetic organization of 1
mecA and mecA-regulatory genes in epidemic methicillin-resistant 2
Staphylococcus aureus from Australia and England. J Antimicrob Chemother 3
50:819-24. 4
25. Ma, X. X., T. Ito, C. Tiensasitorn, M. Jamklang, P. Chongtrakool, S. 5
Boyle-Vavra, R. S. Daum, and K. Hiramatsu. 2002. Novel type of 6
staphylococcal cassette chromosome mec identified in community-acquired 7
methicillin-resistant Staphylococcus aureus strains. Antimicrob Agents 8
Chemother 46:1147-52. 9
26. McDonald, M., A. Dougall, D. Holt, F. Huygens, F. Oppedisano, P. M. 10
Giffard, J. Inman-Bamber, A. J. Stephens, R. Towers, J. R. Carapetis, 11
and B. J. Currie. 2006. Use of a Single-Nucleotide Polymorphism 12
Genotyping System To Demonstrate the Unique Epidemiology of Methicillin-13
Resistant Staphylococcus aureus in Remote Aboriginal Communities. J Clin 14
Microbiol 44:3720-7. 15
27. Mongkolrattanothai, K., S. Boyle, T. V. Murphy, and R. S. Daum. 2004. 16
Novel non-mecA-containing staphylococcal chromosomal cassette composite 17
island containing pbp4 and tagF genes in a commensal staphylococcal species: 18
a possible reservoir for antibiotic resistance islands in Staphylococcus aureus. 19
Antimicrob Agents Chemother 48:1823-36. 20
28. O'Brien, F. G., G. W. Coombs, J. C. Pearson, K. J. Christiansen, and W. 21
B. Grubb. 2005. Type V staphylococcal cassette chromosome mec in 22
community staphylococci from Australia. Antimicrob Agents Chemother 23
49:5129-32. 24
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
19
29. Okuma, K., K. Iwakawa, J. D. Turnidge, W. B. Grubb, J. M. Bell, F. G. 1
O'Brien, G. W. Coombs, J. W. Pearman, F. C. Tenover, M. Kapi, C. 2
Tiensasitorn, T. Ito, and K. Hiramatsu. 2002. Dissemination of new 3
methicillin-resistant Staphylococcus aureus clones in the community. J Clin 4
Microbiol 40:4289-94. 5
30. Oliveira, D. C., and H. de Lencastre. 2002. Multiplex PCR strategy for rapid 6
identification of structural types and variants of the mec element in 7
methicillin-resistant Staphylococcus aureus. Antimicrob Agents Chemother 8
46:2155-61. 9
31. Oliveira, D. C., A. Tomasz, and H. de Lencastre. 2001. The evolution of 10
pandemic clones of methicillin-resistant Staphylococcus aureus: identification 11
of two ancestral genetic backgrounds and the associated mec elements. Microb 12
Drug Resist 7:349-61. 13
32. Oliveira, D. C., S. W. Wu, and H. de Lencastre. 2000. Genetic organization 14
of the downstream region of the mecA element in methicillin-resistant 15
Staphylococcus aureus isolates carrying different polymorphisms of this 16
region. Antimicrob Agents Chemother 44:1906-10. 17
33. Price, E. P., F. Huygens, and P. M. Giffard. 2006. Fingerprinting of 18
Campylobacter jejuni by Using Resolution-Optimized Binary Gene Targets 19
Derived from Comparative Genome Hybridization Studies. Appl Environ 20
Microbiol 72:7793-803. 21
34. Roberts, R. B., A. de Lencastre, W. Eisner, E. P. Severina, B. Shopsin, B. 22
N. Kreiswirth, and A. Tomasz. 1998. Molecular epidemiology of 23
methicillin-resistant Staphylococcus aureus in 12 New York hospitals. MRSA 24
Collaborative Study Group. J Infect Dis 178:164-71. 25
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
20
35. Robertson, G. A., V. Thiruvenkataswamy, H. Shilling, E. P. Price, F. 1
Huygens, F. A. Henskens, and P. M. Giffard. 2004. Identification and 2
interrogation of highly informative single nucleotide polymorphism sets 3
defined by bacterial multilocus sequence typing databases. J Med Microbiol 4
53:35-45. 5
36. Robinson, D. A., and M. C. Enright. 2003. Evolutionary models of the 6
emergence of methicillin-resistant Staphylococcus aureus. Antimicrob Agents 7
Chemother 47:3926-34. 8
37. Ryffel, C., F. H. Kayser, and B. Berger-Bachi. 1992. Correlation between 9
regulation of mecA transcription and expression of methicillin resistance in 10
staphylococci. Antimicrob Agents Chemother 36:25-31. 11
38. Shore, A., A. S. Rossney, C. T. Keane, M. C. Enright, and D. C. Coleman. 12
2005. Seven novel variants of the staphylococcal chromosomal cassette mec in 13
methicillin-resistant Staphylococcus aureus isolates from Ireland. Antimicrob 14
Agents Chemother 49:2070-83. 15
39. Shukla, S. K., S. V. Ramaswamy, J. Conradt, M. E. Stemper, R. Reich, K. 16
D. Reed, and E. A. Graviss. 2004. Novel polymorphisms in mec genes and a 17
new mec complex type in methicillin-resistant Staphylococcus aureus isolates 18
obtained in rural Wisconsin. Antimicrob Agents Chemother 48:3080-5. 19
40. Shukla, S. K., M. E. Stemper, S. V. Ramaswamy, J. M. Conradt, R. Reich, 20
E. A. Graviss, and K. D. Reed. 2004. Molecular characteristics of 21
nosocomial and Native American community-associated methicillin-resistant 22
Staphylococcus aureus clones from rural Wisconsin. J Clin Microbiol 23
42:3752-7. 24
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
21
41. Stephens, A. J., F. Huygens, J. Inman-Bamber, E. P. Price, G. R. Nimmo, 1
J. Schooneveldt, W. Munckhof, and P. M. Giffard. 2006. Methicillin-2
resistant Staphylococcus aureus genotyping using a small set of 3
polymorphisms. J Med Microbiol 55:43-51. 4
42. Taneike, I., T. Otsuka, S. Dohmae, K. Saito, K. Ozaki, M. Takano, W. 5
Higuchi, T. Takano, and T. Yamamoto. 2006. Molecular nature of 6
methicillin-resistant Staphylococcus aureus derived from explosive 7
nosocomial outbreaks of the 1980s in Japan. FEBS Lett 580:2323-34. 8
43. van der Zee, A., M. Heck, M. Sterks, A. Harpal, E. Spalburg, L. 9
Kazobagora, and W. Wannet. 2005. Recognition of SCCmec types 10
according to typing pattern determined by multienzyme multiplex PCR-11
amplified fragment length polymorphism analysis of methicillin-resistant 12
Staphylococcus aureus. J Clin Microbiol 43:6042-7. 13
44. Wu, C. Y., J. Hoskins, L. C. Blaszczak, D. A. Preston, and P. L. Skatrud. 14
1992. Construction of a water-soluble form of penicillin-binding protein 2a 15
from a methicillin-resistant Staphylococcus aureus isolate. Antimicrob Agents 16
Chemother 36:533-9. 17
45. Yang, J. A., D. W. Park, J. W. Sohn, and M. J. Kim. 2006. Novel PCR-18
restriction fragment length polymorphism analysis for rapid typing of 19
staphylococcal cassette chromosome mec elements. J Clin Microbiol 44:236-20
8. 21
46. Zhang, K., J. A. McClure, S. Elsayed, T. Louie, and J. M. Conly. 2005. 22
Novel multiplex PCR assay for characterization and concomitant subtyping of 23
staphylococcal cassette chromosome mec types I to V in methicillin-resistant 24
Staphylococcus aureus. J Clin Microbiol 43:5026-33. 25
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
22
Figure legend 1
2
Fig. 1 3
4
5
Illustration of the strategy used to identify sets of resolution-optimised binary 6
markers. Each genotype of interest is characterised based on the T (presence) (T) or 7
absence (A) of the total set of binary markers (5 are used in this example). Thus the 8
binary gene configuration is converted into a pseudo-DNA sequence composed of A’s 9
and T’s. The alignment of pseudo-DNA sequences is then analysed using Minimum 10
SNPs for combinations of binary markers that maximise the Simpson’s index of 11
Diversity (D) (33). In the example shown above, binary genes 2 and 4 provide a D of 12
1 i.e. they completely resolve the three genotypes. In the present study, 46 SCCmec 13
types were defined by the presence or absence of 34 binary markers. Even with this 14
relatively small data set, manual identification of resolution optimised marker sets is 15
extremely difficult. 16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
1
1
Table 1. The 34 binary targets used for SCCmec characterisation. 2
3
Core Junkyard
ccr
mec
class
J1 J2/J3
ccr1 A a pT181 dcs
ccr2 A1a b pI258 pls
ccr3 A.3 c pUB110 kdp
ccr4 A.4 d Tn4001
3A.2.1
unique
ccr5 B g Tn554 (MLS)
ccrC-VT B1 b
ΨTn554 (cad)
C2 IS256-dcs
E IS256-Tn4001
F b
IS256-mecI
G c
4
a mec class A1 is characterised by a 166bp deletion in mecR1 (15, 24, 32). 5
b Two separate mec classes have been termed B1. For this analysis, the mec class described by Lim et al remains B1 and the mec configuration described by 6
Shukla et al has been renamed mec class F (23, 39). 7
c The novel mec class featured in SCCmecZH47 at the time of writing was unnamed, therefore the tentative name of mec class G was used (GenBank Accession 8
No. AM292304). 9
10
11
12
ACCEPTED on June 1, 2018 by guest
http://aac.asm.org/
Dow
nloaded from
2
1
2
Table 2. Forty-six SCCmec variants collated for this study a 3
4
5
ACCEPTED on June 1, 2018 by guest
http://aac.asm.org/
Dow
nloaded from
3
Defining characteristics: SCCmec
Class
SCCmec
(n = 46)
Uniform
Nomenclature b ccr/mec class J1 J2 and J3
PCR
genotype c
Strain/isolate and/or Sequence Reference
I (1B) I 1B.1.1 ccr1, mec class B Pls I NCTC10442 (AB033763), COL (CP000046) (15)
I variant (1B.2.1) ccr1, mec class B No pls I PhII (38)
IA 1B.1.2 ccr1, mec class B Pls pUB110 I PER34 (31)
IA variant (1B.1.3) ccr1, mec class B Pls IS256 I PER184 (31)
IA variant (1B.1.4) ccr1, mec class B Pls IS256 and pUB110 I PER88 (31)
II (2A) IIa 2A.1.1 ccr2, mec class A kdp operon pUB110 II N315 (D86934), MRSA252 (BX571856) (16)
II variant 2A.1.2 ccr2, mec class A kdp operon II Not described (3)
II variant (2A.1.3) ccr2, mec class A kdp operon pUB110 and no dcs II Not described (3)
IIA 2A.3.1 ccr2, mec class A.4 (IS 1182) J1 of IVb (2B.2.1) Tn554 and pUB110 II NV Not described (38)
IIB 2A.3.2 ccr2, mec class A J1 of IVb (2B.2.1) pUB110 II AJ810123 (38)
IIC 2A.3.3 ccr2, mec class A.3 (IS 1182) J1 of IVb (2B.2.1) Tn554 and pUB110 II NV Not described (38)
IID 2A.3.4 ccr2, mec class A.4 (IS 1182) J1 of IVb (2B.2.1) Tn554 II NV Not described (38)
IIE 2A.3.5 ccr2, mec class A.3 (IS 1182) J1 of IVb (2B.2.1) Tn554 II NV AR13.1/330.2 (AJ810120) (38)
IIb 2A.2 ccr2, mec class A Unique sequence for 2A.2 IS256, Tn554 II JCSC3063 (AB127982) (12)
IIb variant (2F.1.1)d ccr2 mec class F kdp operon dcs and no pUB110 NV Not described (40)
III (3A) III (3A.1.1) ccr3, mec class A Tn554 MLS, pT181 and pI258 III HUSA304 (23, 31)
III (3A1.1.1)e ccr3, mec class A1 (∆mecR1) Tn554 MLS, pT181 and pI258 III 85/2082 (AB037671), ANS46 (15, 31)
IIIA 3A.1.2 ccr3, mec class A No pT181 or ips III HU25 (AF422651-AF422696) (31)
IIIB 3A.1.3 ccr3, mec class A No pT181, pI258 or Tn554 III HDG2 (31)
III variant (3A.1.5) ccr3, mec class A No Tn554 III R35 (31)
III variant (3A.1.6) ccr3, mec class A pUB110 III 15814-9852 (41)
III variant (3A.2.1) ccr3, mec class A Tn554 cad with tnpA III 85/3907 (18)
III variant (3A.3.1) ccr3, mec class A Tn554 (MLS) III 85/961 (18)
III variant (3A.4.1) ccr3, mec class A pls dcs III DOM068 (5)
III variant (3A.1.6) ccr3, mec class A dcs III HSA10 (1)
ACCEPTED on June 1, 2018 by guest
http://aac.asm.org/
Dow
nloaded from
4
1
a Table modified from Chongtrakool et al (4). 2
b Uniform nomenclature values in parentheses denote tentative naming of unnamed variants as defined by Chongtrakool et al (4). 3
c Hypothetical SCCmec genotype using standard PCR assays for ccr, mec class and the 2B J1 regions; a, b, c and d. NV and NT indicate a new variant and a non typable variant, respectively. 4
d Shukla et al., described this isolate as a SCCmec 2A variant (40). 5
e The mec class of isolate 85/2082 features a 166bp deletion of the mecR1 membrane spanning domain, which is characteristic for mec class A1 (15, 24, 32). 6
IV (2B) IVa 2B.1 ccr2, mec class B Specific for 2B.1 (IVa) IVa MW2 ( BA000033), CA05 (AB063172) (25)
IVb 2B.2.1 ccr2, mec class B Specific for 2B.2.1 (IVb) IVb 8/6-3P (AB063173) (25)
IVc 2B.3.1 ccr2, mec class B Specific for 2B.3.1 (IVc) Tn4001 IVc MR108 (AB096217) (18)
IVc variant 2B.3.2 ccr2, mec class B Specific for 2B.3.1 (IVc) No Tn4001 IVc 2314 (AY271717) (27)
IVd 2B.4 ccr2, mec class B Specific for 2B.4 (IVd) IVd JCSC4469 (AB097677) (12)
IVE 2B.3.3 ccr2, mec class B Identical to 2B.3.1 (IVc) Unique left extremity sequence IVc AR43/3330.2 (AJ810121) (38)
IVF 2B.2.2 ccr2, mec class B Identical to 2B.2.1 (IVb) Unique left extremity sequence IVb AR43 (38)
IVg 2B.5 ccr2, mec class B Specific for 2B.5 (IVg) IVb M03-68 (DQ106887) (22)
IV variant (2B.6) ccr2, mec class B Different to IVa,b,c or d IV NV SD179-1 (12)
IV variant 4B ccr4, mec class B NT HDE288 (AF411935) (31)
IV variant (2B.1.2) ccr2, mec class B IS256 IVa BARGII17 (31, 34)
IVA variant 2B.N.2 ccr2, mec class B pUB110 IVa PER2 (3, 30)
V (5C) V 5C.1 ccr5, mec class C2 hsd V WIS (AB121219) (17)
V variant (5B.1) ccr5, mec class B No details described NV 04-17489 (28)
V variant (5B1.1) ccr5, mec class B1 No details described NV WBG8404 (28)
V variant (5E.1) ccr5, mec class E No details described NV WBG10198 (28)
VT (5C2.1) ccrC2, mec class C2 No details described V TSGH 17 (2)
V variant (5/2G.1) ccr5 and ccr2, mec class C2 hsd NV ZH 47 ( AM292304) Not published
V variant (5/1C.1.1) ccr5 and ccr1, mec class C2 No details described pT181 NV B827549 This study, (41)
Other NV (2C.1) ccr2, mec class C2 No details described NV 04-16419 (28)
Truncated - - - IS431, pUB110 and dcs NT 479968 (6)
ACCEPTED on June 1, 2018 by guest
http://aac.asm.org/
Dow
nloaded from
5
1
2
3
4
5
6
7
8
9
10
11
12
13
14
ACCEPTED on June 1, 2018 by guest
http://aac.asm.org/
Dow
nloaded from
6
Table 3. Resolution-optimised sets of binary targets, selected without constraints. Solid horizontal lines 1
represent discrimination by the binary targets listed in the row that is second from the top.. The 2
discriminatory power increases as additional targets are added, so the number of horizontal lines 3
increases from left to right.. SCCmec variants in parentheses were named in this study. 4
6 targets 9 targets 12 targets 15 targets
SCCmec
class
ccr2, Tn554 (MLS),
dcs, pUB110,
mec class A, J1b
6 + mec class C2,
IS256-dcs, pT181
9 + mec class B,
ccr1, mec class A.3
12+ J1a, J1c, pls
1B (I) 1B.1.1 1B.1.1 1B.1.1 1B.1.1
(1B.2.1) (1B.2.1) (1B.2.1) (1B.2.1)
4B* 4B
* 4B 4B
(1B.1.3) (1B.1.3) (1B.1.3) (1B.1.3)
1B.1.2 1B.1.2 1B.1.2 1B.1.2
Truncated* Truncated
* Truncated Truncated
(1B.1.4) (1B.1.4) (1B.1.4) (1B.1.4)
2A (II) 2A.3.1 2A.3.1 2A.3.1 2A.3.1
2A.3.3 2A.3.3 2A.3.3 2A.3.3
2A.3.4 2A.3.4 2A.3.4 2A.3.4
2A.3.5 2A.3.5 2A.3.5 2A.3.5
2A.1.1 2A.1.1 2A.1.1 2A.1.1
2A.3.2 2A.3.2 2A.3.2 2A.3.2
2A.2 2A.2 2A.2 2A.2
2A.1.2 2A.1.2 2A.1.2 2A.1.2
(2A.1.3) (2A.1.3) (2A.1.3) (2A.1.3)
(2F.1.1) (2F.1.1) (2F.1.1) (2F.1.1)
3A (III) 3A.1.1 3A.1.1 3A.1.1 3A.1.1
(3A.2.1) (3A.2.1) (3A.2.1) (3A.2.1)
(3A.3.1) (3A.3.1) (3A.3.1) (3A.3.1)
3A.1.2 3A.1.2 3A.1.2 3A.1.2
3A.1.3 3A.1.3 3A.1.3 3A.1.3
(3A.1.5) (3A.1.5) (3A.1.5) (3A.1.5)
(3A.1.6) (3A.1.6) (3A.1.6) (3A.1.6)
(3A.4.1) (3A.4.1) (3A.4.1) (3A.4.1)
(3A1.1.1) (3A1.1.1) (3A1.1.1) (3A1.1.1)
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
* Variants grouped with unrelated variants. 20
(3A.1.4) (3A.1.4) (3A.1.4) (3A.1.4)
2B (IV) 2B.4 2B.4 2B.4 2B.4
(2B.6) (2B.6) (2B.6) (2B.6)
2B.1 2B.1 2B.1 2B.1
2B.3.2 2B.3.2 2B.3.2 2B.3.2
(2B.1.2) (2B.1.2) (2B.1.2) (2B.1.2)
2B.3.1 2B.3.1 2B.3.1 2B.3.1
2B.3.3 2B.3.3 2B.3.3 2B.3.3
(5/2G.1)* (5/2G.1)
* (5/2G.1) (5/2G.1)
(2C.1)* (2C.1) (2C.1) (2C.1)
2B.2.1 2B.2.1 2B.2.1 2B.2.1
2B.5 2B.5 2B.5 2B.5
2B.N.2 2B.N.2 2B.N.2 2B.N.2
2B.2.2 2B.2.2 2B.2.2 2B.2.2
5C (V) 5C.1 5C.1 5C.1 5C.1
(5C2.1) (5C2.1) (5C2.1) (5C2.1)
(5B1.1) (5B1.1) (5B1.1) (5B1.1)
(5E.1) (5E.1) (5E.1) (5E.1)
(5B.1) (5B.1) (5B.1) (5B.1)
(5/1C.1.1) (5/1C.1.1) (5/1C.1.1) (5/1C.1.1)
No.
variants
resolved
21/46
D = 0.9507
29/46
D = 0.9768
35/46
D = 0.9855
39/46
D = 0.9923 ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
8
Table 4. Resolution-optimised sets of binary targets nucleated by ccr1 and ccr5. Solid horizontal lines 1
represent discrimination by the binary targets listed in row that is second from the top. SCCmec variants in 2
parentheses were named in this study. 3
4
5
6 targets 9 targets 12 targets 15 targets
SCCmec
class
ccr1, ccr5,
Tn554 (MLS), dcs,
pUB110, ccr2
6 + J1b,
mec class C2,
IS256-dcs
9 + mec class A,
pT181,
mec class A.3
12 + mec class B,
J1a, J1c
1B (I) 1B.1.1 1B.1.1 1B.1.1 1B.1.1
(1B.2.1) (1B.2.1) (1B.2.1) (1B.2.1)
(1B.1.3) (1B.1.3) (1B.1.3) (1B.1.3)
1B.1.2 1B.1.2 1B.1.2 1B.1.2
(1B.1.4) (1B.1.4) (1B.1.4) (1B.1.4)
2A (II) 2A.1.1 2A.1.1 2A.1.1 2A.1.1
2A.3.1 2A.3.1 2A.3.1 2A.3.1
2A.3.3 2A.3.3 2A.3.3 2A.3.3
2A.3.4 2A.3.4 2A.3.4 2A.3.4
2A.3.5 2A.3.5 2A.3.5 2A.3.5
2A.1.2 2A.1.2 2A.1.2 2A.1.2
(2F.1.1) (2F.1.1) (2F.1.1) (2F.1.1)
2A.3.2 2A.3.2 2A.3.2 2A.3.2
2B.N.2*
2B.N.2 2B.N.2 2B.N.2
2A.2 2A.2 2A.2 2A.2
(2A.1.3) (2A.1.3) (2A.1.3) (2A.1.3)
3A (III) 3A.1.1 3A.1.1 3A.1.1 3A.1.1
(3A.2.1) (3A.2.1) (3A.2.1) (3A.2.1)
(3A.3.1) (3A.3.1) (3A.3.1) (3A.3.1)
3A.1.2 3A.1.2 3A.1.2 3A.1.2
(3A1.1.1) (3A1.1.1) (3A1.1.1) (3A1.1.1)
3A.1.3 3A.1.3 3A.1.3 3A.1.3
(3A.1.5) (3A.1.5) (3A.1.5) (3A.1.5)
(3A.1.4) (3A.1.4) (3A.1.4) (3A.1.4)
(3A.1.6) (3A.1.6) (3A.1.6) (3A.1.6)
(3A.4.1) (3A.4.1) (3A.4.1) (3A.4.1)
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
9
1
* Variants grouped with unrelated variants. 2
2B (IV) 2B.1 2B.1 2B.1 2B.1
2B.3.2 2B.3.2 2B.3.2 2B.3.2
2B.4 2B.4 2B.4 2B.4
(2B.6) (2B.6) (2B.6) (2B.6)
2B.2.1 2B.2.1 2B.2.1 2B.2.1
2B.5 2B.5 2B.5 2B.5
(2B.1.2) (2B.1.2) (2B.1.2) (2B.1.2)
2B.3.1 2B.3.1 2B.3.1 2B.3.1
2B.3.3 2B.3.3 2B.3.3 2B.3.3
2B.2.2 2B.2.2 2B.2.2 2B.2.2
(2C.1)*
(2C.1) (2C.1) (2C.1)
5C (V) 5C.1 5C.1 5C.1 5C.1
(5C2.1) (5C2.1) (5C2.1) (5C2.1)
(5B.1) (5B.1) (5B.1) (5B.1)
(5B1.1) (5B1.1) (5B1.1) (5B1.1)
(5E.1) (5E.1) (5E.1) (5E.1)
(5/2G.1) (5/2G.1) (5/2G.1) (5/2G.1)
(5/1C.1.1) (5/1C.1.1) (5/1C.1.1) (5/1C.1.1)
Other 4B 4B 4B 4B
Truncated Truncated Truncated Truncated
No.
variants
resolved
18/46
D = 0.9391
28/46
D = 0.9729
34/46
D = 0.9836
37/46
D = 0.9903 ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
10
Table 5. Resolution-optimised binary target sets derived from an input file reflecting SCCmec variant 1
abundance. Solid horizontal lines represent discrimination by the binary targets listed in the row that is 2
second from the top. SCCmec variants in parentheses were named in this study. 3
7 targets 9 targets 11 targets 13 targets c
SCCmec
(ccr1)d, dcs,
Tn554 (MLS),
ccr2, mec class A,
J1b, pT181
7 + pUB110,
mec class B
9 + mec class C2,
J1a
11 + pls,
IS256-dcs
Abundant
variant
1B (I) 1B.1.1 1B.1.1 1B.1.1 1B.1.1 Yes
(1B.1.3) (1B.1.3) (1B.1.3) (1B.1.3) No
(1B.2.1) (1B.2.1) (1B.2.1) (1B.2.1) No
1B.1.2 1B.1.2 1B.1.2 1B.1.2 No
(1B.1.4) (1B.1.4) (1B.1.4) (1B.1.4) No
2A (II) 2A.1.1 2A.1.1 2A.1.1 2A.1.1 Yes
2A.1.2 2A.1.2 2A.1.2 2A.1.2 No
2A.2 2A.2 2A.2 2A.2 No
(2A.1.3) (2A.1.3) (2A.1.3) (2A.1.3) No
2A.3.2 2A.3.2 2A.3.2 2A.3.2 No
(2F.1.1) (2F.1.1) (2F.1.1) (2F.1.1) No
2A.3.1 2A.3.1 2A.3.1 2A.3.1 No
2A.3.3 2A.3.3 2A.3.3 2A.3.3 No
2A.3.4 2A.3.4 2A.3.4 2A.3.4 No
2A.3.5 2A.3.5 2A.3.5 2A.3.5 No
3A (III) 3A.1.1 3A.1.1 3A.1.1 3A.1.1 Yes
(3A.2.1) (3A.2.1) (3A.2.1) (3A.2.1) No
(3A.3.1) (3A.3.1) (3A.3.1) (3A.3.1) No
(3A1.1.1) (3A1.1.1) (3A1.1.1) (3A1.1.1) No
3A.1.2 3A.1.2 3A.1.2 3A.1.2 Yes
(3A.1.4) (3A.1.4) (3A.1.4) (3A.1.4) No
3A.1.3 3A.1.3 3A.1.3 3A.1.3 Yes
(3A.1.5) (3A.1.5) (3A.1.5) (3A.1.5) No
(3A.1.6) (3A.1.6) (3A.1.6) (3A.1.6) No
(3A.4.1) (3A.4.1) (3A.4.1) (3A.4.1) No
2B (IV) 2B.1 2B.1 2B.1 2B.1 Yes
(2B.1.2) (2B.1.2) (2B.1.2) (2B.1.2) No
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
11
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
* Variants grouped with unrelated SCCmec variants. 20
a Due to the input dataset containing repeats of identical sequences, a D value close to 1 cannot be 21
obtained 22
2B.3.2 2B.3.2 2B.3.2 2B.3.2 No
2B.4 2B.4 2B.4 2B.4 No
(2B.6) (2B.6) (2B.6) (2B.6) No
2B.N.2 2B.N.2 2B.N.2 2B.N.2 No
2B.2.1 2B.2.1 2B.2.1 2B.2.1 Yes
2B.5 2B.5 2B.5 2B.5 No
2B.3.1 2B.3.1 2B.3.1 2B.3.1 Yes
2B.3.3 2B.3.3 2B.3.3 2B.3.3 No
(5/2G.1) * (5/2G.1)
*
(5/2G.1) (5/2G.1) No
(2C.1)* (2C.1)
*
(2C.1) (2C.1) No
2B.2.2 2B.2.2 2B.2.2 2B.2.2 No
5C (V) 5C.1 5C.1 5C.1 5C.1 Yes
(5C2.1) (5C2.1) (5C2.1) (5C2.1) No
(5B1.1) (5B1.1) (5B1.1) (5B1.1) No
(5E.1) (5E.1) (5E.1) (5E.1) No
(5B.1) (5B.1) (5B.1) (5B.1) No
(5/1C.1.1) (5/1C.1.1) (5/1C.1.1) (5/1C.1.1) No
Other 4B* 4B 4B 4B No
Truncated*
Truncated Truncated Truncated No
No.
variants
resolved
19/46
D = 0.8976
28/46
D = 0.9012
31/46
D = 0.9034
36/46
D = a0.9047
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
12
Table 6. SCCmec variants and associated mecA SNP profiles a. 1
2
mecA SNP position SCCmec
Class
SCCmec
Variant 75 312* 415* 438* 448* 612* 675* 737
1B (I) 1B.1.1 A C G T G T T A/G
2A (II) 2A.1.1 A/C C G T G T T G
2A.3.5 A C G T G T T G
2B (IV) 2B.1 A C G T G T T G
2B.2.1 A C G T G T T G
2B.3.1 A C G T G T T A/G
2B.3.2 A C G T G T T A
2B.3.3 A C G T G T T G
2B.5 A C G T G T T A
3A (III) 3A.1.1 A C G T/A G/A G T A
3A.1.2 A A G T G G T A
3A.1.3 A C G T G T T G
3A.1.4 A C G/A T G T T G
3A.2.1 A C G T G T T A
5C (V) 5/1C.1.1 A C G T G T A G
5C.1 A C G T G T A G
5C2.1 A C G T G T A G
5/2G.1 A C G T G T A G
3
a The most abundant SNP profile, ACGTGTTG, was found 22 times. 4
* Non synonymous codon change. 5
mecA SNP references: 75 (16), 312 (This Study) , 415 (This Study), 438 6
(41), 448 (This Study), 612(39), 675 (17) and 737 (39). 7
8
9
10
11
12
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
13
1
2
3
4
5
6
7
ACCEPTED
on June 1, 2018 by guesthttp://aac.asm
.org/D
ownloaded from
Type 1
1 2 3 4 5
T T T T T
Type 3
Type 2
T A T T A
T T T A T
Binary marker numbers
Pseudo-DNA sequence
Fig. 1
ACCEPTED on June 1, 2018 by guest
http://aac.asm.org/
Dow
nloaded from