Review article

European Journal of Microbiology and Immunology 5 (2015) 1, pp. 44–61 DOI: 10.1556/EuJMI-D-14-00038

REVIEW AND PHYLOGENETIC ANALYSIS OF qac GENES THAT REDUCE SUSCEPTIBILITY TO QUATERNARY AMMONIUM COMPOUNDS IN STAPHYLOCOCCUS SPECIES Trudy M. Wassenaar1,*, David Ussery2,4, Lene N. Nielsen3 and Hanne Ingmer3 1

Molecular Microbiology and Genomics Consultants, Zotzenheim, Germany Center for Biological Sequence Analysis, Department of Systems Biology, Technical University of Denmark, Lyngby, Denmark 3 Department of Veterinary Disease Biology, Faculty of Health and Medical Sciences, University of Copenhagen, Copenhagen, Denmark 4 Present address: Comparative Genomics Group, Biosciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA 2

Received: December 17, 2014; Accepted: December 23, 2014 The qac genes of Staphylococcus species encode multidrug efflux pumps: membrane proteins that export toxic molecules and thus increase tolerance to a variety of compounds such as disinfecting agents, including quaternary ammonium compounds (for which they are named), intercalating dyes and some antibiotics. In Stapylococcus species, six different plasmid-encoded Qac efflux pumps have been described, and they belong to two major protein families. QacA and QacB are members of the Major Facilitator Superfamily, while QacC, QacG, QacH, and QacJ all belong to the Small Multidrug Resistance (SMR) family. Not all SMR proteins are called Qac and the reverse is also true, which has caused confusion in the literature and in gene annotations. The discovery of qac genes and their presence in various staphylococcal populations is briefly reviewed. A sequence comparison revealed that some of the PCR primers described in the literature for qac detection may miss particular qac genes due to lack of DNA conservation. Despite their resemblance in substrate specificity, the Qac proteins belonging to the two protein families have little in common. QacA and QacB are highly conserved in Staphylococcus species, while qacA was also detected in Enterococcus faecalis, suggesting that these plasmid-born genes have spread across bacterial genera. Nevertheless, these qacA and qacB genes are quite dissimilar to their closest homologues in other organisms. In contrast, SMR-type Qac proteins display considerable sequence variation, despite their short length, even within the Staphylococcus genus. Phylogenetic analysis of these genes identified similarity to a large number of other SMR members, found in staphylococci as well as in other genera. A number of phylogenetic trees of SMR Qac proteins are presented here, starting with genes present in S. aureus and S. epidermidis, and extending this to related genes found in other species of this genus, and finally to genes found in other genera. Keywords: biocide resistance, qac, MFS, smr, MRSA, S. aureus, phylogeny

Introduction Multidrug efflux pumps are membrane proteins, typically containing multiple (four to fourteen) transmembrane domains that form a channel to transport toxic compounds out of the cell. Their substrates can be cations, quaternary ammonium compounds (qac), phosphonium derivatives and intercalating dyes [1–6]. These efflux pumps in part protect bacteria to a variety of toxic substances, including chemicals such as biocides and disinfectants, as well as fluoroquinolones and β-lactams [7, 8]. Repeated sub-lethal exposure can select for presence or active expression of efflux pumps, resulting in a population with reduced

susceptibility. A number of multidrug efflux pumps have been identified in Staphylococcus aureus and in other members of the Staphylococcus genus. S. aureus is an opportunistic but potentially serious human pathogen. It colonizes over 30% of individuals and although endogenous strains are often the source of infection, S. aureus can also be acquired from external sources. Pathogenic strains are circulating in hospitals that can cause nosocomial infections in high-risk patient populations, but community-acquired strains can also cause aggressive infections, even in healthy individuals [9, 10]. Adding to the problem is the ability of S. aureus to acquire resistance to antibiotics. In recent years methicillin-

* Corresponding author: Trudy M. Wassenaar; Molecular Microbiology and Genomics Consultants, Tannestrasse 7, 55576 Zotzenheim, Germany; Phone: +49 6701 8531; Fax: +49 6701 901803; E-mail: [email protected] ISSN 2062-509X / $ 20.00 © 2015 Akadémiai Kiadó, Budapest

Qac genes in Staphylococcus

resistant S. aureus (MRSA) strains have attributed to both hospital- and community-acquired infections [10, 11]. Exchange of MRSA strains between animal and human populations has also been observed, with the concerning example of sequence type ST398, which colonizes pigs in countries with a large pig-producing industry like Denmark and The Netherlands, and was shown to be transmitted from these animals to humans [12]. Another medically important staphylococcal species is S. epidermidis. While previously only considered to be a harmless colonizer of the skin, it is now recognized as an important opportunistic pathogen that gives rise to considerable numbers of nosocomial infections and infections associated with medical indwelling devices [13]. In S. aureus and S. epidermidis, genes encoding efflux pumps that are plasmid-borne can be variably present, and these are clinically relevant since they confer reduced susceptibility to commonly used antiseptics and disinfectants, such as cetrimide, benzalkonium chloride and chlorhexidine [6]. Their broad substrate specificity suggests they evolved to provide generic protection against toxic compounds. In clinically relevant populations their presence may have been selected for by human application of disinfectants [14, 15], although this selection could not always be experimentally demonstrated [16]. A number of these proteins are known by the name of Qac, for quaternary ammonium compound resistance / transporter. Generally, multidrug efflux pumps can be divided into five protein families, depending on their energy requirements and structure [3, 6, 17]. The Qac efflux proteins of Staphylococcus species belong to two of these protein families: the Major Facilitator Superfamily (MFS) and the Small Multidrug Resistance (SMR) family [3]. The MFS is the largest known superfamily of secondary transporters, containing 74 protein families (known to date), transporting a variety of substrates [18]. Both the MFS and SMR protein families contain a variety of membrane transporters, all involved in symport, antiport or uniport of a range of substrates. The proton motive force of the cellular transmembrane proton gradient provides energy for such transport [3]. The scope of this review is to briefly summarize the historic literature on qac genes and their products from Staphylococcus species, and to compare the publically available genes by phylogenetic analysis using either amino acid or DNA sequences. The results of these analyses were used to identify conserved domains in the genes encoding Qac that are found across various taxonomic groups. Phenotypic characteristics are not covered here.

The Qac alphabet: from QacA to QacZ Qac is the abbreviation for quaternary ammonium compounds, but not all genes conferring reduced susceptibility to these compounds are called qac, for historical reasons. Early observations of S. aureus resistance as a result of presence of plasmid-encoded genes were obtained from

45

studies using efflux of ethidium bromide as an assay system [19]. Twenty years later, in 1987, the first gene characterized from S. aureus responsible for this ethidium bromide efflux was therefore named ebr by Sasatsu and coworkers [20]. In that same year, Lyon and Skurray introduced the names QacA, QacB and QacC, in an influential review article that summarized the then current knowledge on antimicrobial resistance in S. aureus [1]. Based on observations published at that time, Lyon and Skurray coined the term QacA for a protein coded on pSK1 family plasmids that rendered resistance to multiple substrates: quaternary ammonium compounds, the intercalating dyes acriflavine and ethidium bromide, diamidines such as propamidine isothionate as well as diamidinodiphenylamine dihydroxichloride or chlorhexidine. The protein that produced resistance to quaternary ammoniums and intercalating dyes but not to diamidines or chlorhexidine was called QacB, while the name QacC was reserved for the protein that only produced resistance to quaternary ammoniums and ethidium bromide but not to acriflavine [1]. It was recognized at that time that these phenotypes were all due to efflux proteins and, while a genetic resemblance between QacA and QacB activity had been recognized, the gene coding for QacC was known to be different [1]. The gene encoding QacA was sequenced from plasmid pSK1 of S. aureus three years later, and its similarity to transporter genes involved in sugar transport systems was recognized [21]. QacA is 514 amino acids long, though meanwhile some qacA genes are annotated to start from alternative start codons, producing shorter open reading frames of 504 or 486 amino acids, respectively. The protein contains 14 transmembrane domains. Although an early observation suggested that antiseptic resistance could be detected on the chromosome of S. aureus [22], most of the literature describes qacA as plasmid-encoded. The qacA gene requires a transcriptional regulator for correct expression, which is coded by qacR [23]. This regulator gene is typically located in the upstream region of qacA, and the two genes are divergently transcribed. QacR functions as a repressor by binding to the DNA upstream of qacA, thus inhibiting expression. When cationic lipophilic compounds enter the cell, these bind to the repressor, resulting in its dissociation from the DNA, which in turn enables expression of qacA [23]. When the gene encoding QacB was sequenced, from plasmid pSK23, it was found to be very similar to qacA, despite the observed difference in substrate specificity that had been the basis of their distinction [24]. QacA and QacB only differ at six amino acid positions, of which the non-conserved amino acid change at position 323 (Asp in QacA, Ala in QacB) was demonstrated to be responsible for the difference in specificity [24]. Similar to qacA, expression of the qacB gene is regulated by the transcriptional repressor QacR. The proteins QacA and QacB belong to the MFS superfamily of transporters. Members of this superfamily have a variable number of transmembrane domains (also known as transmembrane segments, TMS), and it is believed they European Journal of Microbiology and Immunology

46

T. M. Wassenaar et al.

all evolved from a simple protein that contained a hairpin structure with two TMS [18]. This prototype seems to have triplicated into a protein containing six TMS, after which duplication formed a 12-TMS protein. The MFS proteins with 14 TMS, like QacA and QacB, have two extra, centrally positioned transmembrane domains, that are believed to be the result of an intragenic duplication [18]. The gene encoding QacC was first recognized in pSK89, a small rolling-circle replication plasmid of S. aureus [1]. QacC is only 107 amino acids long, though in other bacterial genera a length of 108 or 109 amino acids has been observed. These proteins contain four transmembrane domains and form dimers in the bacterial membrane. Expression of qacC does not require a transcriptional regulator. Littlejohn and coworkers recognized that QacC was identical to the ebr gene originally discovered by Sasatsu and coworkers [2]. The same gene was also isolated from a different S. aureus plasmid and published under the alternative name smr, for staphylococcal multidrug resistance [25, 26]. The identity in acronym may have been a coincidence, but qacC/smr happens to be a member of the SMR protein family of small multidrug resistance proteins [3, 17]. The name smr is nowadays used as synonym for qacC and seems to be favoured lately; here we refer to the gene as qacC/smr. The gene can be found on both conjugative and non-conjugative (rolling-circle) plasmids and has been described for both S. aureus and coagulase-negative Staphylococcus species [27]. Grinius and coworkers pointed out that QacC was similar to an efflux protein EmrE in E. coli [25]. An early characterization identified a glutamate residue at position 13 as essential for functionality [26]. Similar proteins are found in Gram-positive and Gram-negative bacteria, and most work performed to reveal structural requirements of SMR transporters has been carried out with E. coli EmrE. Although it was established that this transporter functions as a homo-dimer, the orientation of the two subunits remains unclear; some publications reported that these are oriented in parallel in the membrane, while other studies indicated an inverted dimer to be functional (reviewed in [28]). For Staphylococcus QacC, the exact position of the four alpha-helices, which form the transmembrane domains, was determined by nuclear magnetic resonance spectroscopy [29]. The name qacD was introduced in 1991 by Littlejohn and coworkers for a gene identical to qacC but with different promoter sequences [2]. These authors reserved the name QacD for proteins residing on large conjugative plasmids, but this name is no longer in use and the gene in question is now renamed as qacC. The next addition to the Qac family was QacE, a protein of 115 amino acids first described in E. coli, of which functionally active deletion derivatives (known as QacEΔ) are found in other bacterial genera. Their genes were sequenced from broad-host plasmids replicating in Gram-negative bacteria [30]. In one publication, qacE genes were also described in Gram-positives, including in Staphylococcus [31], though to our knowledge this observation has not been confirmed since. European Journal of Microbiology and Immunology

In the early 1990s it was recognized that qac genes were often contained on mobile elements, notably integrons, and that they could spread from one plasmid to another by site-specific recombination. The finding of a novel integron in a plasmid of Enterobacter aerogenes resulted in the next member the Qac family, QacF [32]. Although QacF (typically 110 amino acids, with a range of 108–112) has been detected in a number of Gram-positives [31], homologues of QacF have not been described for Staphylococcus. QacH was published before QacG, though the same research group coined their names. In 1998 the gene encoding QacH was published, after it was identified on plasmid p2H6 of S. saprophyticus [33]. The qacH gene encodes a protein of 107 amino acids, with strong resemblance to other SMR members, and it confers high-level ethidium bromide resistance but low-level proflavine resistance [33]. The same group identified QacG as the protein encoding resistance to benzalkonium chloride and ethidium bromide from plasmid pST94 of Staphylococcus; the species from which this plasmid was isolated was not specified [34]. QacG of Staphylococcus spp. is 107 amino acids long. The next Qac protein described was QacJ (107 amino acids long). Its gene was discovered from a newly identified rolling-circle replicating plasmid and subsequently identified in several equine isolates of S. aureus, S. simulans and S. intermedius [35]. Finally, QacZ was the name given for a new SMR-type Qac gene identified in Enterococcus faecalis [36]. This gene has so far not been described for Staphylococcus spp. In conclusion, although their names suggest a similarity between all Qac proteins, there is a clear division between on the one hand QacA and QacB (both members of the MFS superfamily and collectively called QacA/B here), and on the other hand QacC – J plus QacZ, which are all members of the SMR protein family. Genetically, these two protein families evolved separately; their similarity in function is an example of parallel evolution. The name QacD is no longer in use, and QacE, QacF, and QacZ are absent from Staphylococcus species.

Detection and distribution of qac genes in Staphylococcus populations Since resistance to biocide and disinfectants can severely hamper disinfection measures and current hygiene strategies to reduce nosocomial infections, the presence of qac genes has been investigated intensively in Staphylococcus populations from clinical settings as well as in the community. Most studies were carried out by PCR detection of the genes, sometimes backed up with phenotypic characterization to determine the nature and level of resistance. The primers used for studies employing classical PCR are summarized in Table 1. Based on multiple alignment of currently available qac sequences, three of the primers that have been in use can now be considered suboptimal:

[39]

295 bp

ATAGTCAGTGAAGTAATAG

qacH (3)

[15]

295 bp

CAATAGTCAGTGAAGTAATAGGCAGTG

[5]

qacH (3)

[39] 237 bp

qacJ (3)

306 bp

CCAACATTAGGCACACTTATTTC

qacJ (3)

[15]

GATCCAAAAACGTTAAGA

CTTATATTTAGTAATAGCG

232 bp (2)

TGACTTGATCCAAAAACGTTAAGA

GGCCAACATTAGGCACACTTA

qacJ (3)

[39]

275 bp

TACATTTAAGAGCACTACA (1)

CAACAGAAATAATCGGAACT

qacG

[15]

213 bp

AATGGCTTTCTCCAAATACA

TTTCGTTTGGAATTTGCTTT

qacG

[5]

139 bp (2)

CTGAAGTTATTGGAAGTGCATT

[14]

qacC/smr

[37]

286 bp

ATAAGTACTGAAGTTATTGGAAGT

qacC/smr

157 bp

AAACAATGCAACACCTACCACT

qacC/smr

[5]

215 bp (2)

ATTCCATTGAGTGCCTTTGC

qacA/B

[14]

[14], [47], [51]

[47], [51]

[5], [14], [47], [51]

[15], [38], [51]

[15], [41], [42], [45], [51]

Also used by

(3) Since only two sequences of qacH and only one of qacJ were publically available in GenBank at the time of analysis, conservation of the primer sequences could not be established

(2) The predicted amplicon length differs from that reported in the literature

(1) Nucleotides in bold represent positions where polymorphisms have been found in sequenced genes

*Amplicon length was prediction based on classical PCR using qac sequences available from public databases. Other PCR-dependent methods (realtime PCR, nested PCR, RFLP-PCR) are not included the table

AGTGTGATGATCCGAATGT

TGTGATGATCCGAATGTGTTT

CAATTAATGACTTGATCCAAAAACG

GTTATATTTAGTGGTAGGTGTTGC

TTCCGAAAATGTTTAACGAAACTA

AACGAAACTACGCCGACTATG

GCCCTTTCTTTAGGGTTTTC

AATCCCACCTACTAAAGCAG

630 bp

GCTGCATTTATGACAATGTTTG

qacA/B

[37]

Described by [38]

CCACTACAGATTCTTCAGCTACATG

CTATGGCAATAGGAGATATGGTGT (1)

qacA/B

361 bp

Amplicon length* 417 bp (2)

CCAGTCCAATCATGCCTG (1)

GCAGAAAGTGCAGAGTTCG

qacA/B

Reverse primer (5’ to 3’)

Forward primer (5’ to 3’)

Gene

Table 1. Primers used for detection of specific qac genes by classical PCR as described in the literature

Qac genes in Staphylococcus

47

European Journal of Microbiology and Immunology

48

T. M. Wassenaar et al.

the reverse primer for detection of qacA/B designed by Noguchi et al. [37], and used by a number of other researchers, is unlikely to amplify qacA of plasmid SAP070B due to a mismatch in its last 3’-nucleotide (Table 1). Moreover, the forward primer used by Mayer et al. [38] includes the polymorphic nucleotide that distinguishes qacA from qacB, and although the primer was reported to hybridize to both types, sequence analysis of the amplicon would not reveal whether the tell-tale qacA or qacB nucleotide had been present. Lastly, two mismatches in the reversed primer for qacG amplification designed by Bjorland et al. [39] might hamper amplification of some qacG genes that have since been sequenced, due to non-conserved nucleotides. Not all qac genes have been found in all Staphylococcus populations that are frequently monitored for reduced susceptibility to quaternary ammonium compounds. QacA and qacB are mainly found in S. aureus and S. epidermidis. QacC is also mostly found in these two species, though it can also be present in S. warneri. The other Qac names were introduced for proteins of the SMR family that were discovered in Staphylococcus spp. other than S. aureus or S. epidermidis, but with the exception of QacH, all of these have been described for the two medically relevant species S. aureus and S. epidermidis. A brief overview of the prevalence of the different qac genes that were determined in various bacterial populations is given here. Prevalence in clinical settings Worldwide differences in qac presence of bacterial populations can be observed, but these genes have been reported from most continents. The presence of various qac genes in a clinical setting obviously concentrated on S. aureus, whereby in some studies MRSA and methicillin-susceptible MSSA were analysed separately. The prevalence of qac genes seems to be particularly high in Asia. For instance, the genes qacA/B and qacC/ smr were found in 44% and 31%, respectively, in MRSA isolated between 1998 and 1999 in various Asian countries [37]. In human clinical isolates originating in Japan, qacA/B was found in 32% of MRSA, though at lower frequency (7.5%) in MSSA [40]. An alarmingly high prevalence of qacA/B (83%) in clinical MRSA isolates was recently reported from Malaysia, although qacC/smr was rarely found present [41]. Compared to the Asian situation, low prevalences were reported from the North-American continent. Investigations in Toronto, Canada, identified qacA/B in 2% of clinical MRSA isolates from an intensive care unit, with 7% positive for qacC/smr [42]. A lower incidence still was found in MRSA isolated from new admissions to nursing homes in the US, where only 0.6% of over 800 samples were positive for qacA/B [43]. Likewise, only 0.9% of clinical MRSA samples that had been submitted to a US surveillance network were positive for qacA/B by real-time PCR [44]. An international study comparing clinical and environmental S. aureus isolates from hospitals in three US states plus European Journal of Microbiology and Immunology

Freiburg in Germany found variable prevalences, whereby qacC/smr was more often detected than qacA/B [45]. Incidences of qacA/B in Europe are generally higher than reported for the US. A large European study that investigated S. aureus isolates from 1997–1999 reported 42% positive for qacA/B, again with higher frequency in MRSA (63%) than in MSSA (12%); qacC/smr was detected in 6% of the isolates with no difference between MRSA and MSSA [38]. A study from Scotland demonstrated that the minimal bactericidal concentration (MBC) of chlorhexidine gluconate and triclosan was 10 to 1000-fold lower than the concentration recommended for disinfection, and that this concentration was increased for Staphylococcus isolates that contained qac genes [15]. The authors demonstrated the presence of qacA and of qacC/smr (sometimes in combination) and reported absence of qacG, qacH or qacJ [15]. In contrast, Furi and coworkers observed effects on minimal inhibitory concentrations (MIC) but not on MBC for chlorhexidine when qacA/B was present [5]. High prevalence of qacA/B (55%) was also recently observed in a Danish study [46], where strains of S. epidermidis were recovered from scrub nurses who used chlorhexidine heavily, and this hardly differed for non-users (50%). The MIC/ MBC of these isolates did not correlate to presence or absence of qacA/B [46]. A lack of standard in experimental performance can explain these conflicting results [4]. Presence of genes other than qacA/B and qacC/smr in clinical isolates has not been extensively studied, but a few studies are available. In a study from the UK of clinical MRSA isolates, apart from qacA/B (8%) and qacC/smr, which was detected at a high frequency of 44%, qacH was also found in 3% of the strains, while qacG was absent and qacJ was not determined [14]. The gene qacG was identified, however, in 11 of 21 human clinical isolates of S. haemolyticus from Argentina, as were qacH and qacJ (in 10 and 4 isolates, respectively) [47]. In a Swedish study at a university hospital, qacA/B gene was present in 62/143 isolates (43%) and qacC/smr in 6%, while qacH was found once [48]. In historical isolates of S. epidermidis collected in the 1960s before the introduction of chlorhexidine, however, qacA/qacB were completely absent; the isolates in question were all from human patients in Denmark [46]. Thus, although presence of qacA/B does not always affect measurable MIC or MBC, long-term use of chlorhexidine or related compounds seems to have selected for presence of these genes in current clinically relevant populations. Prevalence in community, food-producing animals and food Since the importance of community-acquired Staphylococcus infection is drawing attention, presence of qac genes in community populations has also been investigated. A study from China identified few S. aureus but more coagulase-negative Staphylococcus (CNS) on an automated teller machine (ATM), as a sampling site representing

49

Qac genes in Staphylococcus

the bacterial population residing in the community. Fifteen per cent of the sampled ATMs contained Staphylococcus species, of which less than 1% were MRSA. Of the obtained S. aureus, 11% carried qacA/B and less than 2% had qacC/smr, while qacC/smr was found in 14% and qacA/B in 26% of the CNS isolates [49]. Another interesting approach to investigate community isolates was to sample the air in a Shanghai metro station but, although the primers used in this study were presented as specific for qac, unfortunately they detected a plasmid replication gene, so conclusions about presence of qac cannot be drawn [50]. Porcine Staphylococcus has been implied as a key population connecting community, food and clinical populations. In a study from Hong Kong, porcine MRSA contained qacG and qacC/smr but not qacA/B, qacH or qacJ [51]. In a recent study from Denmark that included 79 porcine S. aureus isolates, qacG was also identified. Strains containing qacG, as determined by PCR, were positive in an ethidium bromide efflux assay (Seier-Petersen and coworkers, unpublished data). Coagulase-negative Staphylococcus, which can cause opportunistic infections, are of concern in the food-industry, in particular strains that produce enterotoxin (a virulence factor once thought to be exclusive for coagulasepositive S. aureus). Disinfectants are widely used in food production, so it is no surprise that qac genes have been found in CNS isolated from food sources. The genes qacA/B, qacC/smr, qacG, qacH and qacJ have all been reported from S. epidermidis, S. saprophyticus, S. cohnii and S. hominis [39, 52].

A comparison of S. aureus qac genes extracted from public sequence databases In the studies described in the previous section, qac genes were mostly detected by PCR or real-time PCR but not

by sequencing; their presence and identity was inferred from the length of fragments generated by gene-specific primers. In contrast, genes that have been sequenced and submitted to public databases can be used for comparative analysis. In order to assess whether gene names have been used consistently, a name search was performed in the ‘Gene’ database at NCBI. The search revealed 6349 entries retrievable from that database with the query ‘QacA’, of which 198 were from Staphylococcus species. Only four entries were retrieved when queried with ‘QacB’, of which two were from Staphylococcus. This and other findings are summarized in Table 2. The non-redundant protein or nucleotide databases at NCBI contain far more qac genes from Staphylococcus spp. than those identified in the ‘Gene’ database, but many are filed under alternative names. Nowadays, most proteins are deduced from (complete) genome sequences, and their annotation is automated, based on detected homologies to existing gene families. Depending on the program used for gene annotation, genes with homology to qac can receive various alternative names and annotations: MFS as synonym for QacA and Smr or EmrE for QacC were most frequently encountered (Table 2). When the nonredundant database of NCBI was searched with QacB from pSK23 as the BlastP query, all hits to staphylococcal genes were close to 100% conserved. Some variation was observed in the start of the protein, which will be described in more detail in the next section. A hit with 99% identity was obtained to a Listeria monocytogenes protein and another to a protein from Propionibacterium acnes (though the first 124 amino acids were missing here). The most closely related protein from a different genus that was not 99% conserved over the complete aligned region (an indication that this protein evolved differently) was an ‘antiseptic resistance protein’ QacA from Kocuria palustris, with only 51% identity of the whole length of the protein.

Table 2. Search findings using the Gene database at NCBI Gene database queried at NCBI, searched by name Search term

Nr. of entries found / in Staphylococcus

qacA

6349 / 198

qacB

4/2

qacC

25 / 14

qacD

1/1

smr

Non-redundant protein database at NCBI, searched by sequence homology using PBLAST Query protein sequence

Protein names of hits

QacA

QacA, QacB, MFS (99–100% identity)

QacC

QacC, Smr, EmrE, QacG, QacH, SugE

5697 / 69

qacG

3/0

QacG

QacC, Smr, EmrE, QacG, QacH, SugE

qacH

29 / 0

QacH

QacC, Smr, EmrE, QacG, QacH, SugE

qacJ

1/1

QacJ

QacC, Smr, EmrE, QacG, QacH, SugE

ebr

4474 / 0

QacZ was not included in this search European Journal of Microbiology and Immunology

50

T. M. Wassenaar et al.

A BlastP search with either QacC or QacG protein sequences as the query resulted in very different findings compared to the QacB Blast, as they retrieved many hits that were less strongly conserved. The hits obtained with either QacC or QacG were mostly overlapping and carried various names (Table 2); some were annotated as hypothetical. A considerable number of close relatives to QacC, QacJ or QacG were detected in genera other than Staphylococcus. In combination, these findings suggest that 1) the sequences of QacA/B are all highly conserved and the proteins seem to have evolved little since they spread in Staphylococcus species; 2) the other Qac proteins are all related to each other though they show a higher degree of sequence variation than QacA/B; and 3) QacC, J and G homologues are found in a number of other bacterial genera as well as in staphylococci, while QacA/B has only been detected in two other species outside the Staphylococcus genus so far.

as described in the previous section were aligned using ClustalW. Gene fragments missing more than 30 amino acids were discarded. This resulted in 25 different proteins, a mixture of QacA and QacB. Collectively they contained variation in 19 amino acid positions, ignoring five proteins that lacked the first few amino acids due to alternative, downstream start codons; in addition, the L. monocytogenes gene that was included lacked the last three residues (it was annotated as ‘partial’). A phylogenetic tree of these 25 different variants of QacA/B is shown in Fig. 1. As can be seen in the figure, all proteins with predicted QacB substrate specificity (based on their Ala at position 323), are separated from the presumable QacA proteins. However, the genes from S. epidermidis and S. aureus are not found on separate clusters, and one protein variant was also detected in multiple S. hominis strains. This mixing of genes from different species within the observed clusters is to be expected, as the plasmids on which these genes reside are shared between Staphylococcus species. The membrane spanning domains (TMS) of QacA/B and the loops facing outwards or inwards have been deduced from structural analysis, as summarized by Wu and coworkers [53]. Application of this structural model suggests that none of the variable amino acid positions are found in the cytoplasmatic loops. Four are found within outward loops (E50D, E169D, Q170R, and D446A) but

Comparative analysis of QacA and QacB proteins and their genes In order to compare their phylogenetic relationship, the protein sequences with >99% homology to QacB of S. aureus plasmid pSK23 that were identified by BlastP

N=4 S. aureus

59%

(YP_001715985)

N=2 S. epidermidis

4%

N=1 S. epidermidis

23%

(WP_002511327) (AAK38457)

N=1 S. aureus

9%

(AAC38780)

N=4, S. aureus

3%

(YP_006938674)

N=1, S. aureus

76% 49%

N=1, S. aureus

68%

QacB

N=1, S. aureus

(AAQ10694)

N=5, S. aureus

(WP_000622773)

27%

QacA

N=5, S. aureus, S. epidermidis

5%

3% N=9, S. aureus, S. epidermidis, S. hominis

81%

19%

N=2, S. epidermidis 6% 11% 3%

N=1, L. monocytogenes

4% 3%

59%

(NP_115321)

N=2 S. epidermidis

(WP_002484475)

N=2 S. epidermidis

(WP_002504142)

N=1 S. aureus

2%

N=2 S. aureus

(WP_002455980)

(BAJ09383) (WP_000622775)

N=1 S. aureus

67%

N=11 S. aureus N=3, S. epidermidis

(WP_002476234)

(e.g. YP_006938271)

N=2 S. epidermidis

12%

(WP_002505088)

(WP_002496203)

(AGU67291)

N=56, S. aureus, S. epidermidis N=17 S. aureus

(WP_000622777)

(YP_006937800)

N=5, S. epidermidis, S. aureus N=3, S. epidermidis

(AAQ10697)

(BAJ09385)

(ACZ69235)

(YP_005759115)

(YP_006939164)

Fig. 1. Maximum likelihood phylogenetic tree of 25 different QacA/B proteins extracted from GenBank, together representing 142 submissions. Sequences shorter than 484 amino acids were removed. The arrow indicates a protein from L. monocytogenes, which lacks the last three amino acids. The number of entries per branch is given as N and for each branch; one accession number is given as an example in brackets. Cases where multiple entries covered proteins from different species are indicated. All proteins with the Ala at position 323, indicative of QacB substrate specificity, are above the dotted line. The tree was produced with PhyML at www.ATGCmontpellier.fr using default settings and visualized with TreeView (www.trex.uqam.ca) European Journal of Microbiology and Immunology

51

Qac genes in Staphylococcus

S. aureus pSK23 (GQ9004919) S. aureus pSK21 (GQ900490) 0.796

S. aureus pSK53 (GQ915270)

0.751

S. aureus pSK23 (AF053772) S. aureus pSK156 (AF053771)

0.749

0.725

S. aureus pSK156 1951 historical isolate (GQ900448) S. aureus pTZ2162 (AB304512)

0.925 S. epidermidis pST6 (AY028779) S. aureus SAP027A (GQ900388)

qacR-qacB

S. epidermidis SAP110A (GQ900465) 0.938

S. epidermidis pST75 (FR821778) S. epidermidis pSAP108A (GQ900458) S. aureus ST228 pI8T7 (HE579074)

0.925

qacR-qacA

S. aureus ST228 pI7S6 (HE579072) S. aureus ST228 pI6T6 (HE579070) S. aureus ST228 pI4T8 (HE579066)

0.437

S. aureus ST228 pI3T3 (HE579064) S. aureus ST228 pI1T1 (HE579060) 0.979

S. aureus ST228 pI2T2 (HE579062) S. aureus pSAP064A (GQ900419)

S. aureus pSK1 (GU565967) S. aureus pSAP070B (GQ900484)

0.844

S. aureus pTW20 (FN433597) S. aureus pSA1379 (AB255366) S. aureus JKD6008 (CP002120) S. aureus pSAP104C (GQ900498) S. aureus pSAP101A (GQ900495) S. aureus pSK77 (GQ900494) S. aureus pSK57 (GQ900493) S. aureus pSAP066A (GQ900482) S. epidermidis pSK105 (GQ900462) S. aureus pSK17 (GQ900514) S. aureus pVRSA (AP003367) S. haemolyticus (plasmid) (AJ400722) S. aureus pSAP107B (GQ900457) 0.01

Fig. 2. Maximum likelihood phylogenetic tree of DNA sequences of the qacA/B-qacR locus, spanning both open reading frames and their intervening sequence. Sequences were identified by BlastN with 2289 nt from pSK156, a historical isolate from 1951 (shadowed box). Hits were included only if they covered the complete query length and their accession number is shown in brackets. The arrow indicates a locus that is located on the chromosome of S. aureus strain JKD6008. The first node of the tree separates the qacA locus from qacB. The tree was produced with PhyML at www.trex.uqam.ca using default settings and visualized with TreeView European Journal of Microbiology and Immunology

52

T. M. Wassenaar et al.

most variation occurs within the TMSs, in particular in TMS-5 (three variable positions) and TMS-6 and 12 (2 positions each). Finally, three variable positions are found at the border of a TMS, two of which are positioned at the border of a cytoplasmatic loop and a TMS (V137L for TMS-5 and V498F for TMS-14) and A290T is located where TMS-9 changes to an outside loop. A number of these variants have been described before by Nakaminami et al. [8]. These authors stated that QacB was more heterogeneous in sequence than QacA. This trend was weakly observed in our dataset: in total 36 mutations were observed in 9 QacB sequences, compared to 30 mutations in 16 QacA sequences. Most of the variation described by Nakaminami and coworkers did not have consequences to functionality or specificity of the protein, with the exception of A320E in TMS-10 of QacB, which resulted in reduced susceptibility to fluoroquinolones [8]. Several site-directed mutagenesis studies have been performed for QacA, but other than TMS-10, which seems essential for substrate specificity (this is where D323A is located that distinguishes QacA from QacB), few amino acid positions could be identified as essential for functionality [24, 53–57]. This suggests that the protein is relatively robust. For evolutionary investigations, a DNA sequence comparison is more informative than protein analysis, as it also captures synonymous mutations. An analysis based on

A

DNA sequences was extended for the complete qacA/BqacR locus, as QacA/B cannot be functional without its regulator, which means that the two genes have evolved together. DNA sequences covering the open reading frames of qacR, qacA/B, and their intervening sequence were collected by BlastN, using 2289 nucleotides of pSK156, a historical isolate from 1951, as the query [58]. The resulting sequences were aligned and a phylogenetic tree was constructed for hits covering the complete query (Fig. 2). The historical isolate pSK156 is not positioned at the root of the tree shown in Fig. 2, indicating it does not resemble an ancestor of the genes captured here. The tree clearly splits qacA-bearing loci from those with qacB. Apparently, these two genetic lineages are distant enough to be detected as two separate clades when qacR and qacA/B DNA sequences are analyzed together. Moreover, another pSK156 sequence that is present in the database is not identical to the historical isolate (for the region analysed, eight mismatches and one single-nucleotide gap were detected). Whether these differences are due to sequencing errors or represent true biological variation could not be assessed. The qac locus from SequenceType ST228 isolates are all identical, as can be expected, since these are all representatives of a clone that recently spread globally. A chromosomal qac locus, present on an integrated mobile element, was indistinguishable from loci on plas-

Ala at pos 323 determines QacB

pTZ2162 pSAP027A pST6 pSK156

sin >

qacB >

< qacR

pST6 pSAP027A pSK156

pST6

pSAP027A 155 bp unique to pTZ2162

pSK156, pTZ2162

200 bp

B pVRSA pSAP110A

pSK105 pSAP110A, pVRSA

534 bp insert

pSAP110A

Asp at pos 323 determines QacA

498 bp

pST75

pSAP108A pI1T1*, pST75, pI3T3

pSK1 sin >

pTW20, pVRSA, pI1T1, pSA1379, pSAP070B, pI3T3, pSAP107B, pSK105, pSAP108A

first 10 aa missing in pI1T1, pI3T3 first 28 aa missing in pSAP070B

pSA1379**

pSK105 pSAP108A

qacA >

< qacR

pSK1, pSK105 pSAP070B*, pTW20, pSAP107B

pSAP107B

698 bp insert

1331 bp insert

pSK1, pTW20, pSAP070B, pSAP107B, pI1T1, pI3T3

pI3T3

*pSAP070B and pI1T1 merge further upstream **pSA1379 merges into Sin introducing a start codon further upstream

pTW20 pVRSA pI1T1 pSA1379 pSAP070B pI3T3

200 bp

Fig. 3. Compilation of the extended qac locus with flanking regions of 16 completely sequenced plasmids. In panel A, the relevant part of four plasmids with qacB are shown in colour for clarity. Downstream of qacR, to the left in the figure, a recombinase sin gene is conserved. Beyond this, homology is lost. Directly downstream of qacB, pTZ2162 diverges from the other three plasmids, while pSK156 merges with this after 155 nucleotides. Further downstream still, pST6 diverges. Black represents sequences also found in QacA plasmids. In panel B, twelve plasmids with qacA are included. The Sin recombinase is absent in pVRSA and pSAP110A, which share 498 bp before they diverge. Open reading frames other than sin, qacR, and qacA/B are not shown. The figure is drawn to scale European Journal of Microbiology and Immunology

53

Qac genes in Staphylococcus

mids from either S. aureus, S. epidermidis or S. haemolyticus. The sequences flanking the qacA/B-qacR locus were also analysed, as these could potentially reveal information on how the locus spread from one plasmid to another. For this, the DNA sequences of 16 completely sequenced plasmids (selected on the basis that they were not identical) were compared for extended regions of both flanks, until homology was lost. A compilation of the results is shown in Fig. 3. Even comparing only four extended qacB loci, shown in panel A of Fig. 3, already produces a complicated picture. Of note is a conserved gene coding Sin recombinase flanking qacR; the presence of this recombinase has been observed before [59], and a role of this recombinase in transfer of the qac locus can be postulated. However, no evidence of a conserved recombination site was identified downstream of qacB, where sequences diverged (pTZ2162) and converged again (pSK156, pTZ2162). An even more complicated picture resulted when 12 qacA-bearing plasmids were compiled, as shown in Fig. 3B. Two different events have taken place in a number of plasmids between sin and qacR that resulted in insertions, while two plasmids have lost the sin gene completely (pSAP110A and pVRSA). As was observed for qacBbearing plasmids, the plasmids shown in Fig. 3B diverge upstream of sin. The downstream flank of qacA, however, shows more conformity, and at least one qacB plasmid (pSAP027A) shares sequences with a number of qacA plasmids. Nevertheless, the evolutionary history of these plasmids can no longer be deduced, as too many recom-

bination events have occurred over time, one overprinting another, to recognize the order of events.

Comparative analysis of qac genes belonging to the SMR family For a similar phylogenetic comparison of QacC, homologues to QacC from Staphylococcus species were obtained by BlastP with the gene from pSK89 as the query. Using a cut-off of identity over at least 60% of the query length, this search identified not only proteins annotated as QacC and its synonym Smr, but also QacG, QacH, and proteins annotated with alternative names (Table 2). Therefore, the search was extended with queries of the other qac genes, and all hits above the cutoff-value, for proteins with a minimum length of 103 and a maximum of 113 amino acids were collected. Records specified as partial sequences were ignored. A phylogenetic tree with the genes annotated as specific qac genes, plus their closest relatives, is shown at the top of Fig. 4. Again, most genes were represented by multiple submissions. QacC was by far the most often recorded: a total of 76 Genbank entries were identified, with one variant being conserved in 70 cases. This predominant form of QacC, whose sequence is shown in panel B of Fig. 4 with the positions of the TMSs indicated as determined by Poget et al. [29], has so far been found in S. aureus, S. epidermidis, S. warneri and in Enterococcus faecalis. Two variants of QacG protein were identified, either from

(4) S. equorum, Staphylococcus spp. (WP_002506528) (2) S. epidermidis (WP_002495341) (70) S. aureus, S. epidermidis, S. warneri, Enterococcus faecalis (NP_647561) (4) S. warneri, Staphylococcus spp. (YP_007443144) (1) Staphylococcus spp. (WP_019467894) 0.827 QacH (1) S. saprophyticus (O87868) (1) S. aureus (AAB47993) 0.89 QacJ (4) S. aureus (NP_783299) 0.74 0.999 (5) S. aureus, Staphylococcus spp. (YP_006938120) QacG (2) S. epidermidis (WP_002477903) (2) S. pettenkoferi (WP_002470909) 0.999 (2) S. arlettae (WP_002509672) (1) Staphylococcus spp. (ERF48678) (3) S. arlettae, Staphylococcus spp. (WP_002509671)

A

0.488

0.673

QacC

B MPY I YL I I A I S TE V I GS AF L K S S E GF S K F I P S L G T I I S F G I C F Y F L S K T MQHLPL N I T Y AT W A GL GLVL TT VV S I I I FKE Q I NL I T I V S I VL I I V G VV S L N I F G T S H

C

MPY I YL I I A I S TE V I GS AF L K S S E GF S K F I P S L G T I I S F G I C F Y F L S K T MQHLPL N I T Y AT W A GL GLVL TT VV S I I I FKE Q I NL I T I V S I VL I I V G VV S L N I F G T S H 1

5

N-term

10

15

20

25

30

35

40

45

50

55

60

65

70

75

80

85

90

95

100

105

C-term

Fig. 4. Comparison of 14 different QacC, QacG, QacJ and QacH proteins, together with their closest homologues belonging to the SMR family from various Staphylococcus species. Panel A: Phylogenetic Neighbour Joining tree including the closest homologues of QacC, QacG, QacH and QacJ. Numbers between brackets indicate the number of entries in GenBank for a particular amino acid sequence, followed by the species in which these proteins were reported, with one accession number as an example. Bootstrap values are indicated. The tree was produced with PhyML at http://phylogeny.lirmm.fr using default settings. Panel B: Amino acid sequence for the most commonly recorded QacC protein (represented 70 times in GenBank). The boxes indicate alpha-helix transmembrane segments. The arrow points out Glutamate at position 13 that is essential for functionality. Panel C: Conserved amino acid positions based on a comparison of all proteins except for the two outliers from panel A. A sequence logo plot for these proteins is available as Supplementary Fig. S1 European Journal of Microbiology and Immunology

54

T. M. Wassenaar et al.

S. epidermidis or from S. aureus (both variants would be amplified with the qacG primers described by Smith et al. [15], shown in Table 1). Of the two variants of QacH, the one isolated from ‘Staphylococcus spp.’ was not annotated as such, but based on its close resemblance it is likely to represent QacH (note, however, that neither of the qacHspecific primer pairs shown in Table 1 would amplify this gene). The same applies to a variant of what is likely QacJ, which would again not be amplified by any of the qacJspecific primer pairs shown in Table 1. A comparison of the amino acid sequences from the main cluster containing all genes annotated as Qac plus the gene of S. pettenkoferi (the top 11 proteins in the tree) identified the conserved positions as shown in panel C of Fig. 4. A logo plot of this comparison is available as supplementary Fig. S1. The four TMSs are relatively strongly conserved, in particular the third TMS, while the loops display more variation. Whether this conversation is related to functionality of the protein has not yet been established. Even though the genes from the two bottom branches of the tree (isolated from S. arlettae and Staphylococcus spp.) are more distant, they are also likely to confer resistance to a broad variety of quaternary ammonium compounds. The proteins represented in Fig. 4 are not the only members of the SMR family present in Staphylococcus species: the searches described above identified many more homologues that were more distantly related and variably annotated as quaternary ammonium compound transporters, SugE, Smr or with alternative descriptors. A tree was composed of all identified members of the SMR family from Staphylococcus species (Fig. 5, a version with each branch annotated is available as Supplementary Fig. S2). Four clusters were recognized, of which Cluster 4 contained the QacC proteins that were already shown in Fig. 4. The fact that all QacC proteins were clustered together and separated from all other SMR genes is in accordance with their specialized phenotype. SMR proteins from pathogenic S. aureus were only found in this Cluster 4, while protein genes from S. epidermidis were

found in Clusters 1, 3 and 4. Clusters 2 and 3 mainly contained proteins derived from coagulase-negative species, although S. delphini and S. intermedius were also represented. Interestingly, the only species containing proteins from all four clusters was S. arlettae. In a review article by Bay and coworkers, the SMR family was subdivided into three subclasses, based on sequence comparisons across bacterial species [17]. These authors defined the subclass of small multidrug pumps (SMP) to which QacC-J of Staphylococcus and EmrE of E. coli belong, the SUG family with SugE of E. coli as the prototype (the name stands for suppressor of groEL mutation) and the group of paired small multidrug resistance proteins (PSMR) which produce a heterodimer of two proteins in the membrane [17]. In an attempt to identify if any of the four clusters we had identified coincided with these subclasses, representatives for each subclass from E. coli (EmrE, SugE, and the pair YdgE and YdgF for PSMR) were included in the analysis (results shown in Fig. S2). This identified that Cluster 3 was most similar to SugE but not EmrE of E. coli, while Cluster 4 with QacC was distantly related to E. coli EmrE, YdgE and YdgF. Only three genes (two from S. arlettae and one from Staphylococcus spp.) bore resemblance to the PSMR subclass of YdgE/F. That proteins belonging to the SugE-like Cluster 2 and the QacC Cluster 4 are transporter proteins is obvious, but they do not share phylogenetic characteristics to identify the three subclasses as defined by the E. coli genes. Whether the substrate specificity of each member within these two clusters is conserved remains unclear and would require phenotypic characterization. The function of Clusters 1 and 2 protein genes is mostly annotated as SugE or EmrE, for those cases where a functional annotation is specified, but phenotypic analysis would be required to justify these annotations. For each of the four clusters of the tree shown in Fig. 5, the protein alignment was analysed to identify conserved domains. This identified telltale amino acids that were typically conserved in members of each cluster, summa-

Table 3. Conserved domains within clusters of the Staphylococcal SMR genes TMS1 (starting from pos. 9)

Extended TMS3 (including parts of surrounding loops)

Cluster 1

F I V AGx E xGV xLNE L V I

A II I K IPMGTAYAIWTGIGT GGT GM FYxES G LV L T

M I AGxFE LG xxNxY V V

I

Cluster 2

Cluster 3

I AxxxE xWVxGL V

V I A TV LPVGT YAxFVG G xG xxxxxFF S L T VT

A

T T V LPLNx YA WAGLGLVLTTxxS xxFKE A S I

Cluster 4

I TA S S IxxE IG FLKx EGF K S V SS A T

S V AV AI K PM TAYA WTGIG GG LGMxxYxES L G I TA TL T

The clusters are numbered according to Fig. 5. Positions showing variation with 2 dominant amino acid residues are indicated. Positions with variation in 3 or more amino acids are shown as x. Amino acids conserved in all four clusters are shown in bold European Journal of Microbiology and Immunology

55

Qac genes in Staphylococcus

Cluster 1a 3 S. capitis 2 S. caprae 90 S. epidermidis 2 S. simiae

Cluster 1 S. arlettae S. camosus S. capitis S. caprae S. epidermidis S. equorum S. haemolyticus S. hominis S. lugdunensis S. massiliensis S. pettenkoferi S. saprophyticus S. simiae S. simulans S. warneri S. xylosus

Cluster 1b 2 S. arlettae 4 S. equorum 9 S. lugdunensis 2 S. pettenkoferi 5 S. saprophyticus 1 S. xylosus 2 Staphylococcus spp. Cluster 1c 3 S. camosus 1 S. epidermidis 3 S. haemolyticus 5 S. hominis 2 S. massiliensis 2 S. simulans 6 S. warneri 2 Staphylococcus

Cluster 2 2 S. arlettae 1 S. delphini 3 S. equorum 7 S. intermedius 1 S. lentus 2 S. saprophyticus 1 S. vitulinus 1 S. xylosus 6 Staphylococcus spp.

Cluster 3a 5 S. arlettae 2 S. caprae 85 S. epidermidis 4 S. equorum 5 S. saprophyticus 6 S. warneri 1 S. xylosus 1 Staphylococcus spp. Cluster 3b 3 S. arlettae 4 S. equorum 1 S. lentus 1 S. vitulinus Cluster 3c 1 S. xylosus 1 S. delphini 1 Staphylococcus spp.. 7 S. intermedius 2 Staphylococcus spp. SugE E. coli QacC QacH

Cluster 4

QacG

QacJ YdgE E. coli

Cluster 2 S. arlettae S. delphini S. equorum S. intermedius S. lentus S. saprophyticus S. vitulinus S. xylosus Staphylococcus spp.

Cluster 3 S. arlettae S. caprae S. delphini S. epidermidis S. equorum S. intermedius S. lentus S. saprophyticus S. vitulinus S. warneri S. xylosus

Cluster 4 S. aureus S. epidermidis S. equorum S. pasteuri S. saprophyticus S. simulans S. warneri Staphylococcus spp.

EmrE E. coli YdgF E. coli

Fig. 5. Phylogenetic tree of all homologues belonging to the SMR protein family identified from Staphylococcus species. Four clusters, numbered 1 to 4, are indicated and the species represented in each cluster are listed alphabetically to their right. Staphylococcal Qac proteins all belong to Cluster 4. The four SMR genes of E. coli were added for comparison, shown in red. The dotted branches at the bottom identify S. arlettae proteins that possibly belong to the paired small multidrug resistance proteins. The tree was produced with PhyML at http://phylogeny.lirmm.fr using default settings. A complete version with annotated branches is available as Supplementary Fig. S2 European Journal of Microbiology and Immunology

56

T. M. Wassenaar et al.

rized in Table 3. The results are also available as logoplots in Supplementary Fig. S3. The two domains that are specifically conserved per cluster coincide with the predicted TMS1 and TMS3, the latter extended with parts of its surrounding loops (Table 3). As indicated in Table 3, a few amino acids are conserved in all four clusters, and these include the glutamate (E) at position 13, which was shown to be essential for QacC functionality. Based on its conservation it is likely to be essential for all SMR proteins, though not specific for QacC function. Based on the findings summarized in Fig 5 and Table 3, it is questionable whether the nominal distinction of four different protein types within one cluster (QacC, QacG, QacJ and QacH) and the indiscriminate annotation of all other proteins as EmrE/SugE is reflective of the genetic variation and conservation within this large family. Genetically speaking, all genes of Cluster 4 are close enough to be given the same name, as they are more closely related to each other than to the other clusters. The current nomenclature therefore seems to be over-discriminatory, and is not a true reflection of genetic relationship.

Detection of qac genes in other bacterial genera The QacA/B locus has so far been detected in S. aureus and S. epidermidis mainly, and in a few cases qacA was found in S. hominis. Of concern is, however, that qacA was also detected in a L. monocytogenes sequence, submitted directly to GenBank (Accession number AGU57291). The information provided was that the source of the L. mono-

cytogenes strain was frozen food from China. Whether the sequence was obtained from chromosomal or plasmid DNA was not recorded. Unfortunately, the sequence only partially covers qacA, and no flanking sequences are currently available to assess if the gene resides on a mobile element; it is even possible that the sequenced sample was contaminated with Staphylococcal DNA. However, the recent detection (by PCR) of qacA/B in two E. faecalis isolates, one from cattle and one from human blood [60], suggest that QacA/B is being exchanged between bacterial genera. The direction of such transfer is not known, but since these genes have so far only been described in Staphylococcus species, it is tempting to assume the transfer was to, and not from E. faecalis. A similar situation is found for the SMR family members of Staphylococcus species, where qacC/smr has been found in E. faecalis (shown in Fig. 5). In this case, the gene was identified from a complete genome sequence but it was found on a short contig that has not (yet) been assembled into context. Bischoff and coworkers also produced a qacC amplicon from E. faecalis from human blood, and another from cheese [59]. None of the other qac members of the SMR family have so far been recorded in other genera: thus far, most of the SMR genes described for Staphylococcus species seem to be distinct from those in other bacteria. Figure 6 shows a phylogenetic tree of the closest relatives of staphylococcal SMR. These proteins share over 55% identity to either staphylococcal QacC or QacH and were identified by BlastP. An SMR member of each cluster of staphylococcal proteins was included for comparison. Although no further proteins identical to a Staphylococcus gene were identified, the similarity of some of these

Listeria monocytogenes Qac Listeria monocytogenes QacH Listeria monocytogenes QacH Sporosarcina newyorkensis Qac Paenibacillis polymyxa Qac Paenibacillis polymyxa Qac S. aureus, Enterococcus faecalis QacC S. warneri for Cluster 3 Enterococcus faecalis Mdr S. saprophyticus QacH Staphylococcus for Cluster 4 S. warneri for Cluster 1 S. arlettae for Cluster 2

Fig. 6. Closest relatives to SMR proteins of Staphylococcus species identified in other genera. The three branches at the bottom relate to Fig. 4. The tree was produced with PhyML at http://phylogeny.lirmm.fr using default settings European Journal of Microbiology and Immunology

Qac genes in Staphylococcus

proteins from other genera is strong enough to predict they could be potentially shared, in particular between staphylococci and E. faecalis, as these species are known to exchange genes.

Conclusions The nomenclature of efflux pumps resulting in reduced susceptibility to quaternary ammonium compounds and their genes has historical roots and unfortunately can be confusing and ambiguous. The universal name ‘Qac’ misses the fact that these proteins represent two major families, separating QacA and QacB (MFS superfamily members) from all other Qac proteins that belong to the SMR superfamily. These two major protein families have little in common, other than a similarity in efflux function with some shared substrates. The qacA and qacB genes show limited sequence variation and are mainly found in S. aureus and S. epidermidis. Apart from an identical gene found in E. faecalis, their closest homologues in different genera are quite distinct. The MFS-type qac genes, on the other hand, are less conserved within the Staphylococcus genus, and similar genes are present in other bacterial genera as well. Although QacC, QacG, QacH and QacJ reflect different branches of a phylogenetic tree, their nomenclatural distinction may be over-precise, in view of all other SMR-type efflux pump genes. The name ‘Smr’ has recently been favoured over QacC in Staphylococcus species, but this potentially leads to confusion, as ‘SMR’ is also the acronym for the larger protein family to which QacC belongs. Compared to all staphylococcal proteins of this SMR family, the amino acid sequences of QacC-J all cluster together, while a wide variety of other proteins of this family are represented in Staphylococcus species that are more distantly related to QacC. Many of these other SMR members do not alter susceptibility to quaternary ammonium compounds. The confusing nomenclature will not change soon, but the phylogenetic analyses presented here may at least assist in recognition of key sequence features of the SMR members, and to a better identification of these important genes.

Acknowledgements The authors thank all scientists who submitted their valuable sequences to public databases, without which this study could not have been performed. This study was supported by the Danish Research Council for Strategic Research (2101-08-0030).

References 1. Lyon BR, Skurray R: Antimicrobial resistance of Staphylococcus aureus: genetic basis. Microbiol Rev 51, 88–134 (1987)

57

2. Littlejohn TG, DiBerardino D, Messerotti LJ, Spiers SJ, Skurray RA: Structure and evolution of a family of genes encoding antiseptic and disinfectant resistance in Staphylococcus aureus. Gene 101, 59–66 (1991) 3. Putman M, van Veen HW, Konings WN: Molecular properties of bacterial multidrug transporters. Microbiol Mol Biol Rev 64, 672–693 (2000) 4. Buffet-Bataillon S, Tattevin P, Bonnaure-Mallet M, Jolivet-Gougeon A: Emergence of resistance to antibacterial agents: the role of quaternary ammonium compounds – a critical review. Int J Antimicrob Agents 39, 381–389 (2012) 5. Furi L, Ciusa ML, Knight D, Di Lorenzo V, Tocci N, Cirasola D, Aragones L, Coelho JR, Freitas AT, Marchi E, Moce L, Visa P, Northwood JB, Viti C, Borghi E, Orefici G; BIOHYPO Consortium, Morrissey I, Oggioni MR: Evaluation of reduced susceptibility to quaternary ammonium compounds and bisbiguanides in clinical isolates and laboratory-generated mutants of Staphylococcus aureus. Antimicrob Agents Chemother 57, 3488–3497 (2013) 6. Costa SS, Viveiros M, Amaral L, Couto I: Multidrug Efflux Pumps in Staphylococcus aureus: an Update. Open Microbiol J 7, 59–71 (2013) 7. Fuentes DE, Navarro CA, Tantaleán JC, Araya MA, Saavedra CP, Pérez JM, Calderón IL, Youderian PA, Mora GC, Vásquez CC: The product of the qacC gene of Staphylococcus epidermidis CH mediates resistance to beta-lactam antibiotics in gram-positive and gram-negative bacteria. Res Microbiol 156, 472–477 (2005) 8. Nakaminami H, Noguchi N, Sasatsu M: Fluoroquinolone efflux by the plasmid-mediated multidrug efflux pump QacB variant QacBIII in Staphylococcus aureus. Antimicrob. Agents Chemother 54, 4107–4111 (2010) 9. Lowy FD: Staphylococcus aureus infections. N Engl J Med 339, 520–532 (1998) 10. Taylor AR: Methicillin-resistant Staphylococcus aureus infections. Prim Care 40, 637–654 (2013) 11. Uhlemann A-C, Otto M, Lowy FD, DeLeo FR: Evolution of community- and healthcare-associated methicillin-resistant Staphylococcus aureus. Infect Genet Evol 21, 563– 574 (2014) 12. Fluit AC: Livestock-associated Staphylococcus aureus. Clin Microbiol Infect 18, 735–744 (2012) 13. Otto M: Staphylococcus epidermidis – the “accidental” pathogen. Nat Rev Microbiol 7, 555–567 (2009) 14. Vali L, Davies SE, Lai LLG, Dave J, Amyes SGB: Frequency of biocide resistance genes, antibiotic resistance and the effect of chlorhexidine exposure on clinical methicillin-resistant Staphylococcus aureus isolates. J Antimicrob Chemother 61(3), 524–532 (2008) 15. Smith K, Gemmell CG, Hunter IS: The association between biocide tolerance and the presence or absence of qac genes among hospital-acquired and community-acquired MRSA isolates. J Antimicrob Chemother 61, 78–84 (2008) 16. Cervinkova D, Babak V, Marosevic D, Kubikova I, Jaglic Z: The role of the qacA gene in mediating resistance to quaternary ammonium compounds. Microb Drug Resist 19, 160–167 (2013) 17. Bay DC, Turner RJ: Diversity and evolution of the small multidrug resistance protein family. BMC Evol Biol 9, 140 (2009) 18. Reddy VS, Shlykov MA, Castillo R, Sun EI, Saier MH: The major facilitator superfamily (MFS) revisited. FEBS J 279, 2022–2035 (2012) European Journal of Microbiology and Immunology

58

T. M. Wassenaar et al.

19. Johnston LH, Dyke KG: Ethidium bromide resistance, a new marker on the staphylococcal penicillinase plasmid. J Bacteriol 100, 1413–1414 (1969) 20. Sasatsu M, Shima K, Shibata Y, Kono M: Nucleotide sequence of a gene that encodes resistance to ethidium bromide from a transferable plasmid in Staphylococcus aureus. Nucleic Acids Res 17, 10103 (1989) 21. Rouch DA, Cram DS, DiBerardino D, Littlejohn TG, Skurray RA: Efflux-mediated antiseptic resistance gene qacA from Staphylococcus aureus: common ancestry with tetracycline- and sugar-transport proteins. Mol Microbiol 4, 2051–62 (1990) 22. Gillespie MT, Lyon BR, Skurray RA: Gentamicin and antiseptic resistance in epidemic methicillin-resistant Staphylococcus aureus. Lancet 1(8636), 503 (1989) 23. Peters KM, Sharbeen G, Theis T, Skurray RA, Brown MH: Biochemical characterization of the multidrug regulator QacR distinguishes residues that are crucial to multidrug binding and induction of qacA transcription. Biochem 48, 9794–9800 (2009) 24. Paulsen IT, Brown MH, Littlejohn TG, Mitchell BA, Skurray RA: Multidrug resistance proteins QacA and QacB from Staphylococcus aureus: membrane topology and identification of residues involved in substrate specificity. Proc Natl Acad Sci USA 93, 3630–3635 (1996) 25. Grinius L, Dreguniene G, Goldberg EB, Liao CH, Projan SJ: A staphylococcal multidrug resistance gene product is a member of a new protein family. Plasmid 27, 119–129 (1992) 26. Grinius LL, Goldberg EB: Bacterial multidrug resistance is due to a single membrane protein which functions as a drug pump. J Biol Chem 269, 29998–30004 (1994) 27. Leelaporn A, Firth N, Paulsen IT, Hettiaratchi A, Skurray RA: Multidrug resistance plasmid pSK108 from coagulase-negative staphylococci; relationships to Staphylococcus aureus qacC plasmids. Plasmid 34, 62–67 (1995) 28. Schuldiner S: EmrE, a model for studying evolution and mechanism of ion-coupled transporters. Biochim Biophys Acta 1794, 748–762 (2009) 29. Poget SF, Harris R, Cahill SM, Girvin ME: 1H, 13C, 15N backbone NMR assignments of the Staphylococcus aureus small multidrug-resistance pump (Smr) in a functionally active conformation. Biomol NMR Assign 4, 139–142 (2010) 30. Paulsen IT, Littlejohn TG, Rådström P, Sundström L, Sköld O, Swedberg G, Skurray RA:. The 3’ conserved segment of integrons contains a gene associated with multidrug resistance to antiseptics and disinfectants. Antimicrob Agents Chemother 37, 761–768 (1993) 31. Kazama H, Hamashima H, Sasatsu M, Arai T: Distribution of the antiseptic-resistance gene qacE delta 1 in gram-positive bacteria. FEMS Microbiol Lett. 165, 295–299 (1998) 32. Ploy MC, Courvalin P, Lambert T: Characterization of In40 of Enterobacter aerogenes BM2688, a class 1 integron with two new gene cassettes, cmlA2 and qacF. Antimicrob Agents Chemother 42, 2557–2563 (1998) 33. Heir E, Sundheim G, Holck AL: The Staphylococcus qacH gene product: a new member of the SMR family encoding multidrug resistance. FEMS Microbiol Lett 163, 49–56 (1998) 34. Heir E, Sundheim G, Holck AL: The qacG gene on plasmid pST94 confers resistance to quaternary ammonium com-

European Journal of Microbiology and Immunology

35.

36.

37.

38.

39.

40.

41.

42.

43.

44.

45.

46.

pounds in staphylococci isolated from the food industry. J Appl Microbiol 86, 378–88 (1999) Bjorland J, Steinum T, Sunde M, Waage S, Heir E: Novel plasmid-borne gene qacJ mediates resistance to quaternary ammonium compounds in equine Staphylococcus aureus, Staphylococcus simulans, and Staphylococcus intermedius. Antimicrob Agents Chemother 47, 3046–3052 (2003) Braga TM, Marujo PE, Pomba C, Lopes MFS: Involvement, and dissemination, of the enterococcal small multidrug resistance transporter QacZ in resistance to quaternary ammonium compounds. J Antimicrob Chemother 66, 283–286 (2011) Noguchi N, Suwa J, Narui K, Sasatsu M, Ito T, Hiramatsu K, Song JH: Susceptibilities to antiseptic agents and distribution of antiseptic-resistance genes qacA/B and smr of methicillin-resistant Staphylococcus aureus isolated in Asia during 1998 and 1999. J Med Microbiol 54, 557–565 (2005) Mayer S, Boos M, Beyer A, Fluit AC, Schmitz FJ: Distribution of the antiseptic resistance genes qacA, qacB and qacC in 497 methicillin-resistant and -susceptible European isolates of Staphylococcus aureus. J Antimicrob Chemother 47, 896–897 (2001) Bjorland J, Steinum T, Kvitle B, Waage S, Sunde M, Heir E: Widespread distribution of disinfectant resistance genes among Staphylococci of bovine and caprine origin in Norway. J Clin Microbiol 43, 4363–4368 (2005) Alam MM, Kobayashi N, Uehara N, Watanabe N: Analysis on distribution and genomic diversity of high-level antiseptic resistance genes qacA and qacB in human clinical isolates of Staphylococcus aureus. Microb Drug Resist 9, 109–121 (2003) Shamsudin MN, Alreshidi MA, Hamat RA, Alshrari AS, Atshan SS, Neela V: High prevalence of qacA/B carriage among clinical isolates of meticillin-resistant Staphylococcus aureus in Malaysia. J Hosp Infect 81, 206–208 (2012) Longtin J, Seah C, Siebert K, McGeer A, Simor A, Longtin Y, Low DE, Melano RG: Distribution of antiseptic resistance genes qacA, qacB, and smr in methicillin-resistant Staphylococcus aureus isolated in Toronto, Canada, from 2005 to 2009. Antimicrob Agents Chemother 55, 2999– 3001 (2011) McDanel JS, Murphy CR, Diekema DJ, Quan V, Kim DS, Peterson EM, Evans KD, Tan GL, Hayden MK, Huang SS: Chlorhexidine and mupirocin susceptibilities of methicillin-resistant Staphylococcus aureus from colonized nursing home residents. Antimicrob Agents Chemother 57, 552–558 (2013) Mc Gann P, Milillo M, Kwak YI, Quintero R, Waterman PE, Lesho E: Rapid and simultaneous detection of the chlorhexidine and mupirocin resistance genes qacA/B and mupA in clinical isolates of methicillin-resistant Staphylococcus aureus. Diagn Microbiol Infect Dis 77, 270–272 (2013) Kosmidis C, Schindler BD, Jacinto PL, Patel D, Bains K, Seo SM, Kaatz GW: Expression of multidrug resistance efflux pump genes in clinical and environmental isolates of Staphylococcus aureus. Int J Antimicrob Agents 40, 204– 209 (2012) Skovgaard S, Larsen MH, Nielsen LN, Skov RL, Wong C, Westh H, Ingmer H: Recently introduced qacA/B genes in Staphylococcus epidermidis do not increase chlorhexidine

59

Qac genes in Staphylococcus

47.

48.

49.

50.

51.

52.

53.

54. Xu Z, O’Rourke BA, Skurray RA, Brown MH: Role of transmembrane segment 10 in efflux mediated by the staphylococcal multidrug transport protein QacA. J Biol Chem 281, 792–799 (2006) 55. Hassan KA, Galea M, Wu J, Mitchell BA, Skurray RA, Brown MH: Functional effects of intramembranous proline substitutions in the staphylococcal multidrug transporter QacA. FEMS Microbiol. Lett 263, 76–85 (2006) 56. Hassan KA, Skurray RA, Brown MH: Transmembrane helix 12 of the Staphylococcus aureus multidrug transporter QacA lines the bivalent cationic drug binding pocket. J Bacteriol 189, 9131–9134 (2007) 57. Hassan KA, Souhani T, Skurray RA, Brown MH: Analysis of tryptophan residues in the staphylococcal multidrug transporter QacA reveals long-distance functional associations of residues on opposite sides of the membrane. J Bacteriol. 190, 2441–2449 (2008) 58. Paulsen IT, Brown MH, Skurray RA: Characterization of the earliest known Staphylococcus aureus plasmid encoding a multidrug efflux system. J Bacteriol 180, 3477–3479 (1998) 59. Sidhu MS, Heir E, Sørum H, Holck A: Genetic linkage between resistance to quaternary ammonium compounds and beta-lactam antibiotics in food-related Staphylococcus spp. Microb Drug Resist 7, 363–371 (2001) 60. Bischoff M, Bauer J, Preikschat P, Schwaiger K, Mölle G, Hölzel C: First detection of the antiseptic resistance gene qacA/B in Enterococcus faecalis. Microb Drug Resist 18, 7–12 (2012)

MIC/MBC. J Antimicrob Chemother 68, 2226–2233 (2013) Correa JE, De Paulis A, Predari S, Sordelli DO, Jeric PE: First report of qacG, qacH and qacJ genes in Staphylococcus haemolyticus human clinical isolates. J Antimicrob. Chemother 62, 956–960 (2008) Prag G, Falk-Brynhildsen K, Jacobsson S, Hellmark B, Unemo M, Söderquist B. Decreased susceptibility to chlorhexidine and prevalence of disinfectant resistance genes among clinical isolates of Staphylococcus epidermidis. APMIS 122, 961–967 (2014) Zhang M, O’Dononghue M, Boost M V: Characterization of staphylococci contaminating automated teller machines in Hong Kong. Epidemiol Infect 140, 1366–1371 (2012) Zhou F, Wang Y: Characteristics of antibiotic resistance of airborne Staphylococcus isolated from metro stations. Int J Environ Res Public Health 10, 2412–2426 (2013) Wong TZ, Zhang M, O’Donoghue M, Boost M: Presence of antiseptic resistance genes in porcine methicillin-resistant Staphylococcus aureus. Vet Microbiol 162, 977–979 (2013) Heir E, Sundheim G, Holck AL: Identification and characterization of quaternary ammonium compound resistant staphylococci from the food industry. Int J Food Microbiol 48, 211–219 (1999) Wu J, Hassan KA, Skurray RA, Brown MH: Functional analyses reveal an important role for tyrosine residues in the staphylococcal multidrug efflux protein QacA. BMC Microbiol 8, 147 (2008)

Supplement (4) S. equorum, Staphylococcus spp. (WP_002506528) (2) S. epidermidis (WP_002495341) (70) S. aureus, S. epidermidis, S. warneri, Enterococcus faecalis (NP_647561) (4) S. warneri, Staphylococcus spp. (YP_007443144) (1) Staphylococcus spp. (WP_019467894) 0.827 QacH (1) S. saprophyticus (O87868) (1) S. aureus (AAB47993) 0.89 QacJ (4) S. aureus (NP_783299) 0.74 0.999 (5) S. aureus, Staphylococcus spp. (YP_006938120) QacG (2) S. epidermidis (WP_002477903) (2) S. pettenkoferi (WP_002470909) 0.999 (2) S. arlettae (WP_002509672) (1) Staphylococcus spp. (ERF48678) (3) S. arlettae, Staphylococcus spp. (WP_002509671)

A

0.488

0.673

QacC

B MPY I YL I I A I S TE V I GS AFLK S S E GF S K F I P S L G T I I S F G I C F YFL S K TMQHLPLN I TY AT WA GL GL VL TT VV S I I I FKE Q I NL I T I V S I VL I I V G VV S LN I F G T S H

bits

N-term

S W T G II G TF S L L Y I

T

A

I S MT

L

ATVT

I

A FS

I

K

KH

T LF GV

LRD

L A I QFY

T

VS

GL A

I

S

I

C

S

VI

LL A

I

QI

II

LV F VL A

N

NV SL

ST

M T VG

I

V

I

F VF V V AL G I T

I V

LF

L

I

ST

V

SI

M

I T

L

I

S

G

K EN TS V

105

S

100

AD

95

A L

S

90

ST

L

85

V

80

L TA

75

I SS

70

V TA

65

AS

60

G VM

55

S

50

S

45

I

L

F VA

40

I

L

L

35

I

30

A H

25

P

20

0

M YLYL I A I TE I G FLK SEGF KL P L T SF I CFYFLS TM LPLNI YATWAGLGLVLT VSV FKE NL I I S I L I I GV LN FG HH 15

1

10

2

5

3

1

C

C-term

Fig. S1. Comparison of 14 different QacC, QacG, QacJ and QacH proteins, together with their closest homologues belonging to the SMR family from various Staphylococcus species. Panel A: Phylogenetic Neighbour Joining tree including the closest homologues of QacC, QacG, QacH and QacJ. Numbers between brackets indicate the number of entries in GenBank for a particular amino acid sequence, after the species in which these genes were reported are given, followed by an example accession number. Bootstrap values are indicated. The tree was produced with PhyML at http://phylogeny.lirmm.fr using default settings. Panel B: Amino acid sequence for the most commonly recorded QacC protein (represented 70 times in GenBank). The boxes indicate alpha-helix transmembrane segments. The arrow points out Glutamate at position 13 that is essential for functionality. Panel C: Sequence logo plot based on a comparison of all proteins except for the two outliers from panel A European Journal of Microbiology and Immunology

60

T. M. Wassenaar et al.

Cluster 1 S. arlettae S. camosus S. capitis S. caprae S. epidermidis S. equorum S. haemolyticus S. hominis S. lugdunensis S. massiliensis S. pettenkoferi S. saprophyticus S. simiae S. simulans S. warneri S. xylosus

Cluster 2 S. arlettae S. delphini S. equorum S. intermedius S. lentus S. saprophyticus S. vitulinus S. xylosus

0.94

Cluster 3 S. arlettae S. caprae S. delphini S. epidermidis S. equorum S. intermedius S. lentus S. saprophyticus S. vitulinus S. warneri S. xylosus

Cluster 4 S. aureus S. arlettae S. epidermidis S. equorum S. pasteuri S. saprophyticus S. simulans S. warneri 0.94

13 S. epidermidis WP020362966 0.849 2 S. epidermidis WP002501182 0.821 SugE 13 S. epidermidis WP002447282 3 S. epidermidis WP002493531 0.972 16 S. epidermidis NP765433 0.699 SugE 43 S. epidermidis YP189448 2 S. simiae WP002464054 0.673 ErmE 2 S. caprae WP002453828 0.977 3 S. capitis WP002432642 2 S. epidermidis WP002443859 3 S. saprophyticus YP300697 0.812 2 S. saprophyticus WP002482557 0.587 0 SugE 4 S. equorum WP002506908 0.647 0.957 S. xylosus WP017722869 0.867 EmrE Staphylococcus WP019468324 Staphylococcus ERF49491 0.918 0.742 2 S. arlettae WP002510079 0.865 2 S. pettenkoferi WP002471254 0.845 3 S. lugdunensis WP002460119 0.989 6 S. lugdunensis YP003471000 2 S. hominis WP002448630 0.979 3 S. hominis WP002454990 0.828 ErmE Staphylococcus WP017174884 0.796 2 S. warneri WP002450267 0.922 0.866 SugE 4 S. warneri, S. epidermidis YP007384373 0.883 Staphylococcus WP017636447 3 S. carnosus YP002634886 0.977 2 S. simulans WP002479639 2 S. massiliensis WP009382865 3 S. haemolyticus YP252668 2 Staphylococcus WP002506069 0.935 3 S. equorum WP002511983 0.675 Staphylococcus WP019469884 0.705 0.903 2 S. arlettae WP002509598 0.996 Staphylococcus ERF48598 0.746 0.95 S. xylosus WP017723366 0.98 2 S. saprophyticus WP002484281 S. vitulinus WP016912747 0.928 S. lentus WP016999239 0.844 S. intermedius WP020219772 0.784 5 S. intermedius YP004150128 0.897 S. delphini WP019165361 0.85 0.943 2 Staphylococcus WP016426296 S. intermedius WP019169642 3 S. saprophyticus YP300698 0.944 0.779 2 S. saprophyticus WP002482558 0.173 S. xylosus WP017722870 0.863 4 S. equorum WP002506909 0.306 Staphylococcus WP019468323 3 S. arlettae WP002510078 3 S. epidermidis WP002493530 0.906 SugE 66 S. epidermidis YP189447 0.904 15 S. epidermidis WP002438471 0.89 2 S. arlettae WP002443857 0.798 0.63 2 S. caprae WP002453827 4 S. warneri, S. epidermidis YP007384374 0.756 0.936 2 S. warneri WP002450268 Staphylococcus WP019469883 0.589 0.775 S. xylosus WP017723367 4 S. equorum WP002506070 0.987 S. vitulinus WP016912746 0.951 0.966 S. lentus WP016999238 0.659 3 S. arlettae WP002509599 3 S. intermedius YP006014693 0.072 3 S. intermedius YP004150127 0.797 0.867 0.126 S. delphini WP019165362 0.924 S. intermedius WP019169643 2 Staphylococcus WP016426297 SugE E.coli QacC 2 S. epidermidis WP002495341 QacC 5 S. warneri YP007443144 0.95 QacC 70 S. aureus, S. epidermidis, S. warneri, S. simulans, S. pasteuri, E. faecalis NP647561 0.983 QacH Staphylococcus WP019467894 0.745 0.745 QacH 2 S. saprophyticus O87868 QacG 5 S. aureus YP006938120 0.981 QacG 2 S. epidermidis WP002477903 QacJ S. aureus AAB47993 0.777 0.572 QacJ 4 S. aureus NP783299 0.906 4 S. equorum WP002470909 4 S. equorum WP002506528 0.874 YdgE E. coli 0.809 3 S. arlettae WP002509671 0.409 0.779 EmrE E. coli YdgF E. coli 2 S. arlettae WP002509672 0.998 Staphylococcus ERF48678

Fig. S2. Phylogenetic tree of all homologues belonging to the SMR protein family identified from Staphylococcus species. Each branch gives the species for which this gene was found and their numbers when >1, followed by an example accession number. Bootstrap values are indicated. Four clusters, numbered 1 to 4, are indicated and the species represented in each cluster are listed alphabetically to the left. Staphylococcal Qac proteins all belong to Cluster 4. The four SMR genes of E. coli were added for comparison, shown in red. The dotted branches at the bottom identify S. arlettae proteins that possibly belong to the paired small multidrug resistance proteins. The tree was produced with PhyML at http://phylogeny.lirmm.fr using default settings

European Journal of Microbiology and Immunology

61

Qac genes in Staphylococcus

logoplot for Cluster 1

bits

3 2 1 0

AM W L IAG E GV LNE RT K F SFSTLKLAM IPMGTAY I WTGI TAG T IGMLIFY ESK R I C I L I I SAVVIGLKRL I S

Y

LFY L C S II

F V

L

G W

MV

F I

V

N

bits

3 2 1 0

G

V

N

I

V

Y

T

FW M Q W A KK

T

Y

I

W

H

Q L S R WF S

S GH

Q

NQ

Q

Y

A

FV L L F GI

VF

V

C

MT Y M

I

S

I

T

LF

F

ND VT

FS

H Q

S V A

T I

M

E

G

T V

S

S

M

G LV A

C

F V

A

M

N

I

K

W

R

H

FL

S

T

G

Q

R

N L LA

HFI S R AK

K

S

I LM PL K W V

I

FI M L V V F G V

A

YLF V

G

S

V

T

I

L

I

F

A

G

S

IV

A

ML

AG-FE I -GVV -LNE L V I

M

C

A II I K IPMGTAYAIWTGIGT GGT GM FY-ES G LV L T

logoplot for Cluster 2

M W L A G F E L G L N Y L F S L A M P M T A Y A W T G I G G G M Y E S R I I I S GL K L A FI LA LI LI S L

V

L S

I

V

I

V

F M

VL

C

V

MV VILTFW A AH

LL G A I F R IV T R R K E T HQK

H Q

I

I

A

NA

I

I

Q

FV S

I

TK T KR R F

N Q

S

L

ND N

S

V

G

Y

F

K DA

SSV

I

G

LM

IG LL

Q Y A MR

F

F T MV

L

ASL LF A S LI SG VI ATVA AT I L FV KH KD I KI FC LV LI A ATVI LVSV I M GV

LS A I V

F

C

M

T

I W

I

LV I I V

QD

AQ TN S

M Y

S

K

L

L

E

M T

T

NY

S

A

A W

V

I F Q F L T V

FG

L

I

E A NY L

N

A L F

G

C

T

V

V

I

G

I

RM I

V

T

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104

4

L

L M

F I LI VL LI S KKN I YV I L GLI A I LC I AT LV I

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103

4

N

C

I PMS TAYAVWTGIGAVGGAILGM--Y-ESK L G I TA TL T

AG-FEM LG I --LN-Y V V

logoplot for Cluster 3 4

2 1 0

M W K E W V G L A W T I S F A L P V G Y A V G G F K L I L G I GL K T K

H

N

VI A

L L I K TV VF LI AT TVLI VDM F NGQPFV

QN L LI GL I I I TI NTH DHSLLFS I L LI I F FY LV S CS I V S M FF G A M V VF GS I V L

F T

A I

V

V

F

N

A

D

V

E LF

T

L

V Q T Y ST N T

Q

T

F I

H

A

Q

I

G

I

I

A

AA A FL F

TG V A

F

V I

MS

G

FL MS

VV

V

S I VL V C

I W MV D

N

E

V

A T

H

R

TN

S

V Y

S

Q

V T

S FI Y

L

A I

A

M

I

L V T GF F

A A MC TL

E

I

V A

I

T

G E S NS FI TI V I D

E L A H V L DN T

L

M

M

V

D

M

I

P S

T

V Y

A V

L

L

F L

V

M

T

FV

IL VT I

I LI LL A M T

F

M

A

V

I

M A

VV T

V

VS

S

C

S

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101

bits

3

A TV V LPVGT YA-FVG I G -G -----FF-S L T VT

A---E I -WV-GL V

... C

logoplot for Cluster 4 4

2 1 0

YM Y L I A I TEV I GT FLK SEGFSKL PTLGT I SFG I CFYFLS TM LPLNI TYATWAGLGLVLT VSV IFKE NL I S I S I L I I GV LLN FG P LI A H

I

L

L

I

L

F VA

S

S

G VM

AS

I

L TA

A S T

SS

L

VS

A L

W

A

D S

TF S Y I

T

A

I L

I S MT A T

L VT

L I

K KH

A FS

I

L I QFY LF G A T V

LRD

VS GL

A

S

I I

C

I

S

VI L

I

LV F VL LA N A

QI

NV SL

I

I

T V FV A L GV FI V

M T VG

T

I V

LF

V

I

SI

M

I TI

L

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104

bits

3

N

S A S I--E I IGTAFLK- EGF K T S V SS A

V T T LPLN- YA WAGLGLVLTT--S I --FKE A S

C

Fig. S3. Logoplot for the four clusters formed by SMR genes of Staphylococcus spp. The qacC/Smr genes belong to Cluster 4. Consensus sequences for two domains that are conserved within each cluster are indicated. The plots were produced at weblogo.berkeley.edu

European Journal of Microbiology and Immunology