Genomic Insights into Staphylococcus aureus Isolates Exhibiting Diminished Daptomycin Susceptibility

Daptomycin is one of the last therapeutic resources for multidrug-resistant gram-positive bacteria. Despite its structural similarities with glycopeptides, its mechanisms of action and resistance are different and in some respects are not completely understood. Mutations in several genes have been associated with daptomycin resistance, especially in mprF, walkR, rpoB and rpoC, but their role and importance remain to be elucidated. We have studied mutations in 11 genes, which have been previously associated with daptomycin non-susceptibility, in nine daptomycin-non-susceptible Staphylococcus aureus clinical isolates (daptomycin MIC: >1 mg/L). Susceptibility to daptomycin, vancomycin, linezolid, oxacillin, telavancin and dalbavancin was studied. walkR, agrA, cls1, cls2, fakA, pnpA, clpP, prs, rpoB, rpoC and mprF were amplified by PCR and sequenced. The sequences were compared with the S. aureus ATCC 25923 complete genome (GenBank gi: 685631213) by using BLAST® software. We did not find any changes in walkR, pnpA, prs and clpP. All isolates excepting isolate MSa5 showed a high number of significant mutations (between 13 and 25 amino acid changes) in mprF. Most isolates also showed mutations in the rpo genes, the cls genes and fakA. Daptomycin non-susceptibility in S. aureus clinical isolates seems to be reached through different mutation combinations when compared to S. aureus ATCC 25293. Especially mprF and cls1 showed very high polymorphism in most isolates. Meanwhile, one isolate, MSa5, showed only single mutation in mprF (P314T).


Introduction
Staphylococcus aureus has shown a high capacity for developing resistance to each newer antistaphylococcal antibiotic developed.Antibiotic resistance in S. aureus may result in higher morbidity, mortality, lengths of hospitalization and health expenditures [1].The cyclic anionic lipopeptide antibiotic daptomycin (DAP) [2] is becoming a main resource in therapy against multidrug-resistant S. aureus, especially methicillin-resistant S. aureus (MRSA), due to its activity against glycopeptide and linezolid-resistant MRSA [3,4].The mechanism of action of DAP differs from conventional glycopeptides (vancomycin, teicoplanin) and is more like the cationic antimicrobial peptides produced by the immune system.Vancomycin and teicoplanin impair bacterial cell wall synthesis.They bind to the terminal dipeptide D-alanyl-D-alanine (D-Ala-D-Ala) within the disaccharide pentapeptide attached to lipid II and the forming PGN chain.This way, transglycosylation and transpeptidation are prevented, thus hindering proper wall synthesis [5,6].In addition, vancomycin has other secondary actions, probably less important in the lysis of the microorganism, such as inhibition of RNA synthesis and alteration of wall permeability.
The action of DAP is based on binding to the bacterial cell membrane through its hydrophobic end in the presence of calcium ions [7][8][9].This binding occurs both in the exponential growth phase and the stationary phase, causing membrane depolarization due to the loss of potassium ions from the cytoplasm [10].This process leads to the disruption of multiple functions of the bacterial cell membrane, such as the synthesis of proteins, DNA and RNA without the need for the antimicrobial to penetrate the cytoplasm.Unlike other antimicrobials, the mechanism of action does not involve the lysis of bacterial cells [11] but rather bacterial cell death due to alterations in cellular homeostasis [7,12].
Intrinsic resistance to DAP by the hydrolytic cleavage of ester bonds inside the DAP molecule has been described, in absence of contact with this or other antibiotics" among gram-positive bacteria with a low percentage of GC, such as staphylococci [13].
Streptomyces strains highly resistant to daptomycin (MIC ≥ 256 µg/mL) have been found to inactivate the antibiotic through two primary modes of daptomycin inactivation: ring hydrolysis, leading to linearization of the cyclic compound, and deacylation of the lipid tail [14].Additionally, Paenibacillus lautus has been shown to inactivate daptomycin.The production of the inactivating activity in P. lautus is inducible by exposure to daptomycin, and some data suggest it might be caused by metalloesterases [14].
This "natural" resistance mechanism has been suggested to be associated with the natural production of DAP-like molecules by different microorganisms.Though this type of resistance has not been shown in clinically important bacteria yet, resistance genes coding for these DAP-inactivating enzymes might be a potential source of resistance if they could be captured by pathogenic bacteria, as has happened before for other resistance determinants, such as CTX-M enzymes, which originated in Kluyvera spp.and are now present in most clinically important enterobacteria [15].
Nevertheless, DAP non-susceptibility observed in isolates obtained both in vitro and from therapeutic failures in S. aureus seems to be mainly associated with electrostatic repulsion of the DAP-calcium complex from the cell surface due to an increase in the positive charge of the bacterial surface [16].
DAP-non-susceptible isolates obtained both in vitro and from therapeutic failures usually harbor mutations in genes associated with cytoplasmic membrane (mprF) and cell wall homeostasis (yycFG, also known as walK) and mutations in RNA polymerase subunits rpoB/rpoC [7,8].DAP-non-susceptible isolates have been described, both in vitro and in vivo, lacking mutations in some of these genes but not in all of them [7,[9][10][11].Other authors have shown that DAP non-susceptibility may also involve the upregulation of genes associated with cell wall synthesis and turnover, such as the two-component regulator vraSR [12].Otherwise, whole-genome-sequencing studies have shown that DAPnon-susceptible isolates may harbor mutations in other genes, such as agrA, pnpA, pgsA, prs, cls2 and clpP, some of them associated with the cytoplasmic membrane and cell wall metabolism [10,11].
Therefore, the present study had the following aims: the identification of S. aureus strains (methicillin-resistant Staphylococcus aureus (MRSA) and methicillin-susceptible S. aureus (MSSA)) non-susceptible to DAP, isolated from three different hospitals in Spain; testing sensitivity to different drugs such as oxacillin, vancomycin, linezolid, dalbavancin and telavancin; and the amplification and sequencing of the main genes associated with DAP resistance.

Selection of Bacterial Strains
Nine S. aureus clinical isolates identified as DAP-non-susceptible S. aureus (Minimum Inhibitory Concentration, MIC > 1 mg/L) by microdilution conventional susceptibility systems (WalkAway, Perkin Elmer, Shelton, CT, USA) were included in this study.Seven of them had been isolated in hospitals in the province of Badajoz (Southwest of Spain) and two in the Hospital Universitario de Salamanca (Middle West of Spain).All isolates except MSa9 were from patients who had received daptomycin treatment, although data on duration of treatment and dosage were not accessible.To our knowledge, the patient from whom the MSa9 isolate was obtained did not receive treatment with daptomycin at any time.
Microorganisms were identified by MALDI-TOF mass spectrometry (MicroflexTM, Bruker Daltonics, Bremen, Germany).All the isolates used were preserved and stored at −80 • C in glycerol broth with 10% skimmed milk.

Antibiotic Susceptibility
Susceptibility to DAP, vancomycin (VAN), linezolid (LZD) and oxacillin (OXA) were tested by using a broth microdilution method (MicroScan WalkAway, Perkin Elmer, Waltham, MA, USA), according to manufacturer instructions.Telavancin (TLV) and dalbavancin (DLV) were studied by E-test (Liofilchem, Roseto degli Abruzzi, Italy).The microdilution sensitivity study was carried out using an automated commercial method (MicroScan Walk-Away, Perkin Elmer, USA) in which antimicrobials are lyophilized so that a predetermined concentration is reached by adding a set amount of culture medium.The microorganism is inoculated in cation-adjusted Mueller Hinton broth supplemented with 2% NaCl for oxacillin sensitivity and 50 microg/mL calcium for DAP.An inoculum of the microorganism equivalent to 0.5 of the McFarland (McF) standard is inoculated and incubated at 37 • C for 20-24 h.The reading is carried out in an automated way in MicroScan equipment.
For the E-test study, a Mueller Hinton agar plate was inoculated with an inoculum equivalent to 0.5 McF; strips of the different antimicrobials were deposited on the inoculated agar, incubated at 37 • C for 24 h and read visually.For the DAP E-tests, it was not necessary to supplement the agar with Ca 2+ , as the DAP strips had it incorporated.

DNA Extraction and Amplification of Genes Associated with DAP Non-Susceptibility in S. aureus
A suspension with turbidity equivalent to 1 McF was obtained from colonies grown for 24 h at 37 • C on blood agar.Then, 500 µL of the suspension was transferred to the automated nucleic acid extraction system NucliSENS ® easyMAG (bioMérieux, Marcyl'Étoile, France).Nucleic acid extraction was performed according to manufacturer in-structions.DNA was preserved at −80 • C. A series of genes previously associated with DAP non-susceptibility [7,14] was amplified and sequenced.Amplification was carried out in an Eppendorf Master Cycler ® thermal cycler (Eppendorf, Hamburg, Germany).All primers used in this study were designed specifically (Table 1) based on the S. aureus ATCC 25,923 complete genome sequence deposited in GenBank (gi: 685631213) using the Primer Blast ® application (NCBI, Bethesda, MA, USA).The primers were manufactured by Sigma-Genosys (The Woodlands, TX, USA), and the MasterMix ® PCR amplification mix (Promega, Madison, WI, USA), a premixed 2X solution of Taq DNA Polymerase, dNTPs and reaction buffer, was used for PCR.PCR conditions were as follows: 95 • C × 5 min (denaturation); 35 cycles of: 95 • C × 30 s, 1 min at the temperature reported in the supplementary section (Table S1) specific to each primer (hybridization), 72 • C × 1 min (elongation), and finally, 72 • C × 5 min (elongation).

Visualization of Amplicons and Purification of PCR Product
The products were subjected to electrophoresis in 2% agarose gel together with a size control (DNA Ladder 100 bp, Promega, Madison, WI, USA).Blue/Orange 6x Loading Dye (Promega, Madison, WI, USA) was used as a loading buffer, and RedSafe ® Nucleic Acid Staining Solution (Invitrogen, Carlsbad, CA, USA) was used as a fluorescent dye for DNA detection.
As a general rule, amplified DNA was cleaned for sequencing by using Exo-SAP It ® (Chang Biosciences, Tianjin, China).When DNA was purified from the gel, the MEGAquick-Spin Total Fragment DNA Purification kit (Intron Biotechnology, Gyeonggi-do, Republic of Korea) was used.
Amplicons were sequenced by the Sanger method in a 3500 Genetic Analyzer sequencer (Applied Biosystems, San Francisco, CA, USA) and visualized by using the software Chromas 2.5.1.0.Sequences were compared with the S. aureus ATCC 25,923 complete genome sequence deposited in GenBank (gi: 685631213).S. aureus ATCC 25,923 is commonly used as a control in standard laboratory tests.Comparison and analysis were carried out by using the BLAST ® software freely available on the NIH website.

Mutation Profiles Associated with Non-Susceptibility
All the isolates, excepting MSa5, showed significant mutations in the genes rpoB, mprF, rpoC, cls1, cls2 and fakA, while MSa5 showed significant mutations only in the rpoB and mprF genes (Figure 1) (Tables 2-4).We did not detect any significant mutations in the walk, agrA, pnpA, clpP and prS genes.The genes with the highest number of mutations were mprF (31-35 amino acid changes) and cls1 (6-16 amino acid changes) (Figure 1).
The mprF gene presented a high number of mutations in all isolates (Table 2), excepting MSa5, which only had one mutation in this region: P314T.Mutations in this position also appeared in other clinical isolates, such as MSa1 (P314T) and MSa6 (P314L).Fortythree different mutations were identified in all in the mprF genes.Most isolates showed different combinations of these mutations.Most changes appeared in the central area of the genes, although we found changes throughout the genes.MSa6 had the highest number of mutations in mprF, exactly 35, followed by the strains MSa1, MSa4 and MSa8, with 34 changes.The mprF gene presented a high number of mutations in all isolates (Tab excepting MSa5, which only had one mutation in this region: P314T.Mutations in position also appeared in other clinical isolates, such as MSa1 (P314T) and MSa6 (P3 Forty-three different mutations were identified in all in the mprF genes.Most iso showed different combinations of these mutations.Most changes appeared in the ce area of the genes, although we found changes throughout the genes.MSa6 had the hig number of mutations in mprF, exactly 35, followed by the strains MSa1, MSa4 and M with 34 changes.   1 Amino acid numbering corresponds to their positions within the mprF gene.* Base numbering corresponds to the complete genome of S. aureus subsp.aureus strain ATCC 25923 (GenBank: CP009361.1). 2 Friedman [7]; 3 Friedman [7] and Peleg [11] S295L; 4 Mishra [8]; 5 Friedman [7], Mishra [8], Murthy [17] and Cameron [16]: T345A.As in mprF, we found a high number of changes in the genes that encoded for the cardiolipin synthase, especially in cls1.Only MSa5 showed no mutations in cls1 (Table 3).Though a number of mutations appeared in more than one isolate, no isolates showed the same mutation profile.Only one mutation (I238V) appeared in all the isolates showing mutations in this gene.MSa6 was the isolate with the highest number of mutations in this gene (16 mutations).cls2 only showed four different mutations (Table 3), half of which were shared by all the isolates, excepting MSa5, which did not show any mutation in this gene.
All the isolates, including MSa5, showed mutations in rpoB; however, all the isolates excepting MSa5 had mutations in rpoC (Table 4).
fakA, a protein also referred to as mw1109 in previous studies, is a fatty acid kinase that has been associated with unsensitivity to DAP in mutants selected in vitro [10].All the isolates tested in this study, excepting the clinical isolate MSa5, showed mutations in this gene (Table 4).Mutation profiles in fakA were more homogeneous than in mprF and cls1.Only MSa6 showed a mutation profile that was neatly different.
We did not identify changes in walkR in any of our nine clinical isolates.Other genes also associated occasionally with reduced susceptibility to DAP, such as agrA, clpP, pnpA and prs, did not show any significant mutations in this study.
In summary, we found two different mutation profiles.A group of eight isolates had changes in six genes (mprF, rpoB, rpoC, cls1, cls2 and fakA), while one isolate had changes in only two genes (mprF and rpoB).mprF and cls1 showed high numbers of mutations/isolate in most isolates, while for the other genes, the number of mutations/isolate was much shorter.In the absence of the Multilocus Sequence Typing (MLST) analysis, we could not rule out that the identified substitutions could potentially be associated with a specific genetic lineage rather than daptomycin resistance.

Discussion
DAP resistance seems to be based on complex mechanisms and has been associated with the presence of mutations in different genes, most of them involved in the metabolism and homeostasis of the cell membrane and, in some cases, of the bacterial wall [7,14,15].A cause-effect relationship between specific mutations in specific genes and a specific increase in the MICs of DAP has hardly been described, and the repulsion hypothesis does not explain the resistance to DAP in all S. aureus isolates.Only some studies have shown that the replacement of the mutated mprF gene with a wild one reverses, at least partially, the MICs of DAP [16,18].In the present study, the MIC interval of DAP ranged between 2 and 4 mg/L in all cases, except MSa2, which had an MIC of 1 mg/L.Although the non-sensitivity to daptomycin in S. aureus is MIC > 1 mg/L, due to its proximity to the range, MSa2 was not ruled out from this study.
Mutations in the rpoB and mprF genes have been reported in most staphylococcal clinical isolates showing DAP MIC increases [7,10,11].mprF codes for the bifunctional enzyme MprF, which contributes to the positive charge of the cell membrane surface through the lysinylation of phosphatidylglycerol (PG) and the translocation of lysinylated PG (L-PG) from the inner to the outer leaflet of the cell membrane.The positively charged cell membrane surface resulting from this process helps repel the DAP molecule from the surface [19][20][21].The association between mprF mutations and DAP non-susceptibility is supported by the finding that the inactivation of mutated mprF can reverse, at least partially, DAP non-susceptibility [22].
The mprF protein is composed of 14 transmembrane domains and a cytosolic Cterminal domain [11].The first eight N-terminal transmembrane domains are involved in the translocation of L-PG to the outer CM.The next four transmembrane "central" domains are involved both in L-PG synthesis and flipping, while the cytosolic C-terminal domain is involved only in L-PG synthesis.Single-nucleotide polymorphism (SNP) in mprF genes has been reported in most staphylococcal clinical isolates showing DAP MIC increases [7,10,11] and is associated with gain of function.Nevertheless, some authors have shown that SNPs can also be found in 30% of DAP-susceptible isolates and that some DAP-susceptible isolates with MICs of 1 mg/L can show up to 30 amino acid changes in mprF [23], suggesting a high level of heterogeneity within mprF in some staphylococcal isolates.Nevertheless, the DAP-non-susceptible isolates studied by these authors [23] showed only one amino acid substitution in the mprF (L341S or L826F).Changes in these or in very near positions have been previously reported to be associated with DAP non-susceptibility (S829L, S295L, P314L, S337L, T345I, T345A, I420N) [7,8,11,17,24].
Unlike results reported by other authors, we found one SNP in mprF only in one isolate (MSa5).Mutations at the 314 position had been reported to be previously associated with DAP non-susceptibility, but as P314L [7] instead of P314T, as appears in MSa5.P314L was found in another strain (MSa6, Table 2).The other isolates had high numbers of mutations affecting the three areas.They had between six and eight mutations in the N-terminal "flippase" domains (mode: six mutations), seventeen and twenty mutations (mode: eighteen mutations) in the central, bifunctional domains and eight and nine mutations (mode: eight) in the phosphatidylglycerol lysinylation domain.Six isolates (MSa1, MSa4, MSa5, MSa6, MSa7 and MSa8) included in this wide group of mutations showed some amino acid changes in positions previously associated with DAP non-susceptibility (S295, P314, T345) [7,11,16], but other isolates (MSa2, MSa3 and MSa9) showed mutations previously not reported in association with DAP non-susceptibility.Moreover, they showed no mutations in the area comprising between S295 and I420, where most DAP non-susceptibility-associated mutations reported are located, though they showed more than 30 amino acid changes throughout the gene.This study reflects a large genetic variability between our clinical isolates and the strain used as a reference strain (S. aureus ATCC 25923), especially in some of the genes studied.Some of the mutations observed in the different genes have been previously described to be associated with daptomycin resistance in previous studies [7,8,11,16,17], but others have not.As already mentioned in the results section, the absence of MLST forced us to consider the possibility that at least some of these previously undescribed changes are not associated with daptomycin resistance but with the genetic variability existing between the different MLST types.
All the isolates showed rpoB mutations.The rpoB and rpoC genes encode for the bacterial RNA polymerase β and β' subunits [7,8].The mutations described in rpoB and rpoC, associated with DAP non-susceptibility, are different from those affected by other antibiotics, such as rifampicin [25].Unlike mprF mutations that, at least in vitro, usually emerge at relatively early times during the selection process, rpoB and rpoC mutations emerge later [7].Some studies have reported rpoB changes in mutants selected in vitro but not in clinical isolates [7], while other studies [11] did not detect rpoB changes in strains selected in vitro and detected them only in clinical isolates.Some authors [25,26] have shown that single-point mutations in rpoB, such as A477D or A621E, can reduce susceptibility to both daptomycin and vancomycin.Such mutations have been shown to cause cell wall thickening and reduction of the negative charge of the outer layer in S. aureus.All the isolates tested in this study showed a rpoB mutation (F737Y) that had not been previously described in non-susceptible isolates, whether clinical or obtained in vitro.Changes in the rpoC gene were also detected in all isolates, excepting MSa5.Single rpoC mutations (N341D) have also been associated with treatment failures in staphylococcal infections [27].Mutations in the walkR sequence, a regulator of cell wall metabolism and virulence in S. aureus, B. subtilis and S. pneumoniae [28], have been associated with vancomycin resistance [29] and also with reduced DAP susceptibility, in association or not with mprF and agrA mutations [10,29].
fakA has only been associated with DAP non-susceptibility in one study, and only in mutants selected in vitro [10].In that study, the authors found a L133H mutation associated with DAP non-susceptibility.In our study, this protein showed different changes in eight out of nine clinical isolates.Since the function of this protein is unknown, it is not possible to know its role in DAP susceptibility or even if there is really a cause-effect relationship.
cls genes encode for a cardiolipin synthase, and mutations in these genes have been associated with low DAP susceptibility in staphylococci [11].Cls proteins are membranebound enzymes that synthesize cardiolipin, an important anionic membrane phospholipid, from the phosphatidyl moiety of two PG molecules [30].While in logarithmic growth, PG is the main membrane phospholipid, in stationary growth conditions and under conditions of stress, such as unfavorable growth conditions or cell wall acting antibiotics, cardiolipin can accumulate up to 25-30% of the membrane phospholipid.cls2 has been the most frequently reported cls gene in this aspect.In this study, all isolates excepting MSa5 showed mutations in this gene.Most isolates shared two mutations (V135I, H205R), and two of them added a third mutation (A459L, A471E).None of these mutations had been described previously.The mutations previously reported appeared between amino acids 20 and 60 (A23V, T33N, L52F, F60S) [11] in the two transmembrane domains described in this protein, while our mutations appeared outside these domains and the cardiolipin synthase active domains.
Mutations in cls1 have been described with a much lower frequency in association to reduced DAP susceptibility.Nevertheless, we found mutations in this gene in the same isolates in which we found mutations in cls2, and the number of mutations in this gene/isolate ranged from six to sixteen.Most mutations found in the isolates tested in this study appeared outside the transmembrane and cardiolipin synthase active domains, but V18A (in cls1), which appeared in all isolates excepting MSa5 and G20A and appeared in MSa6 and MSa9 (Table 3), was in the first transmembrane domain, and I238V, which also appeared in all isolates excepting MSa5, and I421M (isolates MSa1, MSa4 and MSa9) appeared in the cardiolipin synthase domains.Results concerning cls1 and cls2 suggest that, though DAP non-susceptibility-associated mutations are more frequently described in cls2, in the isolates tested in this study, it seems more likely that the most transcendent mutations are those observed in cls1, since, unlike cls2, in almost all isolates, we found mutations that specifically affect the transmembrane and cardiolipin synthase functional domains.Mutations in these areas may have a summation effect with those found in mprF, since the negative charge provided by cardiolipin would be reduced or even suppressed, reducing the attraction between DAP and the bacterial membrane [31].

Conclusions
The present study confirms that the acquisition of daptomycin resistance may be associated with different mutational profiles.Therefore, various combinations of mutations in different genes may end up leading to the same result of low susceptibility to daptomycin.However, resistance is mainly associated with mutations in the mprF gene involved in the amount of and situation on the membrane of lysyl phosphatidyl glycerol molecules; rpoB, which encodes an RNA polymerase; and the cls genes, which encode a cardiolipin synthase involved in the metabolism of membrane phospholipids.In the isolates analyzed in this study, mprF, cls1 and cls2 showed much higher numbers of mutations than reported in other studies, many of which had not been described so far.There is no relationship between the number of changes in the genes studied and the degree of sensitivity (MICs).That is, we could find strains with the same MICs and different genetic profiles.Future studies should clarify the real role of each of these mutations.

Institutional Review Board Statement:
This study was conducted in accordance with the Declaration of Helsinki.Since the study was conducted exclusively with microorganisms, and no human subjects or human tissues were involved, the Hospital Ethics Committee considered that his approval was not necessary.
Informed Consent Statement: Not applicable.

Figure 1 .
Figure 1.Number of mutations of each clinical isolate for each gene studied.

Figure 1 .
Figure 1.Number of mutations of each clinical isolate for each gene studied.

Author
Contributions: Conceptualization, J.L.M.-B.and N.G.-C.Study direction: J.L.M.-B.Study development: N.G.-C., M.N.G.-Z.and A.M.B.d.C. Conclusions and writing: N.G.-C.and J.L.M.-B.All authors have read and agreed to the published version of the manuscript.
N • of Isolate (
1 Amino acid numbering corresponds to their positions within the cls1 and cls2 genes.* Base numbering corresponds to the complete genome of S. aureus subsp.aureus strain ATCC 25923 (GenBank: CP009361.1).NC: no change.