Molecular basis and evolutionary cost of a novel macrolides/lincosamides resistance phenotype in Staphylococcus haemolyticus

ABSTRACT Staphylococcus haemolyticus (S. haemolyticus) is a coagulase-negative Staphylococcus that has become one of the primary causes of nosocomial infection. After a long period of antibiotic use, S. haemolyticus has developed multiple resistance phenotypes for macrolides and lincosamides. Herein, we evaluated four S. haemolyticus clinical isolates, of which three had antibiotic resistance patterns reported previously. The fourth isolate was resistant to both erythromycin and clindamycin in the absence of erythromycin induction. This novel phenotype, known as constitutive macrolides-lincosamides-streptogramins resistance, has been reported in other bacteria but has not been previously reported in S. haemolyticus. Investigation of the isolate demonstrated a deletion in the methyltransferase gene ermC, upstream leader peptide. This deletion resulted in constitutive MLS resistance based on whole-genome sequencing and experimental verification. Continuous expression of ermC was shown to inhibit the growth of S. haemolyticus, which turned out to be the fitness cost with no MLS pressure. In summary, this study is the first to report constitutive MLS resistance in S. haemolyticus, which provides a better understanding of MLS resistance in clinical medicine. Importance This study identified a novel phenotype of macrolides/lincosamides resistance in Staphylococcus haemolyticus which improved a better guidance for clinical treatment. It also clarified the mechanistic basis for this form of antibiotic resistance that supplemented the drug resistance mechanism of Staphylococcus. In addition, this study elaborated on a possibility that continuous expression of some resistance genes was shown to inhibit the growth of bacteria themselves, which turned out to be the fitness cost in the absence of antibiotic pressure.

resistance mechanisms: (i) pumping antibiotics out by membrane transporter proteins that are encoded by an ATP binding cassette (ABC) transport superfamily, including vga(A) LC , and msrA (8,9); (ii) inactivation of antibiotics, such as by mphC, which encodes a macrolide phosphotransferase that modifies antibiotics by phosphorylation (10); and (iii) an alteration of the antibiotic's target, preventing the antibiotic from binding to its target.For example, the ribosomal RNA methyltransferase Erm family adds a methyl modification to bacterial 23S rRNA, which renders the antibiotic unable to bind to the 50S ribosomal subunit (11).
With excessive use of antibiotics, many bacteria have developed drug resistance.These emerging antibiotic resistance phenotypes have become a serious clinical issue that has resulted in fatality for an increasing number of patients.Herein, we explored emerging bacterial phenotypes that may be capable of reducing the spread of drugresistant pathogens.The results of this study provided a better understanding of drug resistance and improved guidance for clinical treatment.

Bacterial strains and media
S. haemolyticus strains A, B, C, and D used in this study were initially isolated from patients with invasive clinical infections at the Second Hospital of Soochow University (Jiangsu, China).Plasmid pBT2 (Staphylococcus-Escherichia shuttle vector) was used as a backbone (12).Since pBT2 is a temperature-sensitive plasmid, it can be lost when incubated at higher temperatures.Therefore, all strains with pBT2-related plasmid were cultured at 30°C.

Genome sequencing and assembly
We performed next-generation sequencing (Illumina, Hiseq2500, PE100) of these four isolates of S. haemolyticus.Raw reads were trimmed by FastQC (13), KAT (14), and BBTools (15).To achieve better assembly results, we integrated five tools to assemble clean reads: SuperReads from Masurca (16), BCALM 2 (17), Tadpole from BBTools, SPAdes (18), and Megahit (19) (Fig. S1).Assemblies of these tools were imported into Sequencher (Gene Codes, MI, USA), and the consensus sequences were manually verified and selected to form final assembly files according to the coverage depth of reads.BUSCO (20) was used to evaluate the accuracy and completeness of the final assembly results.The raw sequencing data and final assembly files were uploaded to NCBI database (accession numbers SRR14777770-SRR14777773).

Antibiotic resistance gene analysis
Prokka was used to annotate the assembled sequences (21).The potential resistance genes were analyzed by Resistance Gene Identifier (RGI) (22).

Sequencing reads depth calculation
We aligned sequencing reads to genome assembly files with the Burrows-Wheeler-Align ment tool (23) and then used Mosdepth (24) to calculate the read depth from two levels --contigs and genes.

Plasmid construction and transformation
The ermC gene was amplified from the genomic DNA of S. haemolyticus strain C and inserted into the pBT2 backbone to form pBT2-ermC.The upstream leader peptide sequence (short for LP) of ermC was obtained from the S. haemolyticus strain PK-01 chromosome (NCBI Accession: CP035541, region: 2,573,055-2,573,114) and synthesized by the GenScript Biotech Co.Plasmids were transformed into S. aureus strain ATCC25923 by electro-transformation at 2.0 kV, 25 µF, and 100 Ω.All primers for plasmid construction were provided in Table S2.

Isolation of RNA and RT-PCR
Total RNA was extracted by the TRIzol method (Invitrogen, Carlsbad, CA, USA).RNA purity and integrity were examined by RNA electrophoresis in 1.0% (wt/vol) agarose gel.One microgram of RNA was treated with DNase I prior to cDNA synthesis using a Revert-Aid First Strand cDNA Synthesis Kit (Thermo Scientific, MA, USA).RT-PCR was performed with specific primers for ermC and Staphylococcus 16S rRNA.All RT-PCR primers were provided in Table S2.

Antimicrobial susceptibility test
The turbidity of S. aureus transformed with recombinant plasmids was adjusted to a 0.5 McFarland turbidity (1.5 × 10 8 CFU/mL) and then evenly coated onto Mueller-Hinton agar medium.After 2 min, an erythromycin susceptibility test disc (15 µg) and a clindamycin susceptibility test disc (2 µg) were placed onto the middle of the plate.The distance between the disc and the edge of the Petri dish was more than 15 mm, while the two discs were 15-26 mm apart.The plate was incubated at 30°C for 20-24 h before the results were assessed.The current CLSI M100 (Performance Standards for Antimicrobial Susceptibility Testing) was used to determine whether the bacteria were resistant to the antibiotics.

Growth curve measurement
S. aureus transformed with different plasmids was cultured as above (30°C, trypticase soy broth, chloramphenicol, 200 rpm/min shaker).Bacterial count was measured every 2 h for 20 h.

Identification of a novel phenotype for macrolides/lincosamides resistance in S. haemolyticus
In the Second Hospital of Soochow University, we isolated four strains of S. haemolyti cus, identified as A, B, C, and D. During routine drug sensitivity testing, strains A and B were resistant to clindamycin and sensitive to erythromycin.Strain D was resistant to erythromycin and sensitive to clindamycin.Strain C was resistant to both erythro mycin and clindamycin.And it could tolerate clindamycin directly without erythromy cin induction (constitutive MLS resistance).These phenotypes for strains A, B, and D are common in clinical practice, and their mechanisms of drug resistance have been previously described.However, the phenotype of strain C has not been reported previously.

Genome assembly and antibiotic resistance gene analysis
The genomes of the four S. haemolyticus clinical isolates were sequenced and assembled.The assembly results are listed in Table 1, with all four assemblies demonstrated to be of high quality and high coverage.Potential resistance genes were analyzed by RGI (Table S1).Based on previous studies, several genes were found to be related to erythromycin and clindamycin resistance, vga(A) LC , msrA, mphC, and ermC.As expected, antibiotic resistance genes found in strains A, B, and D corresponded to their respective phenotypes (Fig. 1).Since the phenotype of strain C was unique, we focused analysis on that strain.By comparison of the resistance genes of strain C with their homologs in NCBI, we found that a short leader peptide was missing upstream of ermC (Fig. 2).This leader peptide was found to be essential for ermC expression during erythromy cin induction.Ribosome binding by erythromycin permitted translation of the leader peptide, which rearranged hairpin structures composed of four inverted repeats and exposed the initiation motifs of ermC for translation (Fig. 3A) (30,31).While in S. aureus, the variation of the leader peptide rendered the downstream ermC gene to switch from inducible expression to constitutive expression, resulting in constitutive MLS resistance (Fig. 3B) (32,33).Deletion of the leader peptide resulted in inverted repeat 1, allowing the others to pair as hairpin structures 2:3 or 3:4.Through RNAfold WebServer, we found that the secondary structure upstream of ermC was more likely to form as hairpin structure 2:3, which would expose an initiation sequence for the methylase translation (Fig. 3C).As such, we hypothesized that constitutive MLS resistance in S. haemolyticus strain C was due to deletion of the leader peptide upstream of ermC.

The ermC gene found in S. haemolyticus strain C was located on a plasmid
We aligned raw sequencing data with plasmid-specific elements in Staphylococcus spp.and found the specific elements in sequencing reads, indicating that genome sequenc ing results contained bacterial plasmid reads.We performed read depth calculation for S. haemolyticus strain C genome assembly files (the read depth calculation results for S. haemolyticus strains A, B, and D are shown in Fig. S2).Results (Fig. 4) showed read depth for both the contig containing ermC or the ermC gene itself were significantly higher, suggesting that ermC was most likely located on a plasmid.Based on in silico studies, we performed experimental validation and obtained the complete plasmid which closely resembled other known drug-resistant plasmids reported in S. aureus.

Deletion of the leader peptide upstream of ermC resulted in MLS constitutive resistance of S. haemolyticus
To validate the hypothesis that the constitutive MLS resistance of S. haemolyticus strain C was caused by deletion of the leader peptide upstream of ermC, we amplified the intact ermC gene from strain C and constructed a recombinant plasmid, pBT2-ermC.A synthesized upstream leader peptide (short for LP) sequence was inserted upstream of ermC in plasmid pBT2-ermC to form pBT2-LP-ermC.Due to the absence of a widely used standard strain of S. haemolyticus for drug sensitivity tests, we chose S. aureus for analysis.S. aureus ATCC25923 is usually used as a negative control strain for antimicrobial susceptibility testing (34).We performed a double disk diffusion test to determine MLS inducible/constitutive resistance (Fig. 5).S. aureus carrying pBT2-ermC was unaffected by erythromycin or clindamycin, forming a phenotype of constitutive MLS resistance.Whereas the strain harboring pBT2-LP-ermC showed clindamycin resistance only on the side of the disc close to the erythromycin susceptibility test disc (inhibition zone radius ≈ 7 mm).The side of the disc away from the erythromycin disc remained clindamycin sensitive (inhibition zone radius ≈ 15 mm) resulting in the formation of a D-Circle, which indicated that erythromycin induction was required for resistance to clindamycin.Therefore, constitutive MLS resistance by S. haemolyticus strain C was due to leader peptide deletion upstream of ermC.

The 23S rRNA methylation, modified by ermC, may have adverse effects on bacterial growth
Multiple studies have demonstrated constitutive MLS resistance phenotypes for common pathogenic Staphylococcus strains (35,36).Then since bacteria could consti tutively resist antibiotics, why have they evolved inducible resistance.An interesting observation in this study suggested a possibility.After different plasmids were trans formed into S. aureus, growth rates were found to vary considerably.To confirm these results, we performed a growth curve analysis, as shown in Fig. 6. Results showed that the growth of S. aureus carrying pBT2-ermC was significantly slower than the others.We speculated that this observation might be due to the adverse effects of ermC-mediated 23S rRNA methylation.Compared with unmethylated ribosomes, the translation speed of methylated ribosomes decreased, directly affecting bacterial metabolism and survival, while imposing an enormous fitness cost in the absence of antibiotics.

DISCUSSION
In this study, we analyzed four S. haemolyticus strains isolated from clinical patients.One strain exhibited a surprising drug-resistant phenotype, constitutive MLS resist ance.Although constitutive MLS resistance has been reported in Staphylococcus spp., Streptococcus spp., and Enterococcus spp., this was a novel observation for S. haemoly ticus.We found strain C contained a methyltransferase gene, ermC, with a deletion in the upstream leader peptide, which resulted in constitutive MLS resistance.In this manner we identified, for the first time, a variation in the ermC gene of S. haemolyticus.The plasmids that mimicked the wild-type and varied genotype were constructed and transformed into S. aureus for susceptibility testing.Based on mRNA expression and drug resistance, it was clear that this small leader peptide was critical to the regulation of MLS resistance.
Constitutive MLS resistance by clinical pathogenic bacteria has become an increas ingly important medical issue due to the selective pressure of antibiotic use.Based on the principle of fitness cost, restricting antibiotic usage could be beneficial for the elimination of constitutive drug-resistant bacteria.Such bacteria expressed nonessential genes that might aggravate extra strain burden, competing with other strains when antibiotic selection pressure decreased or disappeared.
It has been reported that multidrug resistance was exhibited by 75% of S. haemoly ticus clinical isolates (37).Intraspecies transfer of staphylococcal cassette chromosome mec (SCCmec) indicated that S. haemolyticus could be a reservoir of resistance genes (38,39).The genome plasticity of S. haemolyticus is characterized by many insertion sequences and SNPs, contributing to the acquisition of antibiotic resistance (40).Various Staphylococcus spp., including S. haemolyticus, may constitute a vast reservoir of resistance genes as well as multi-resistant strains, with frequent intraspecific or interspecies gene transfer, which shortens the usefulness of antibiotics.
Our study has identified a novel phenotype of macrolides/lincosamides resistance in S. haemolyticus.However, this resistance phenotype and corresponding mechanism have been found in S. aureus (32,33), so our discovery was not the first in Staphylococcus spp.In addition, we found that sustained expression of ermC may affect bacterial metabolism and survival, and we have provided some explanations for this phenomenon.But we have not proven our hypothesis through actual biological experiments.This led to our viewpoint that bacteria with inducible resistance had a competitive advantage in survival compared to bacteria with constitutive resistance.This point needs to be proven through more experiments in the future.

Conclusions
In summary, this study identified a novel phenotype of macrolides/lincosamides resistance in Staphylococcus haemolyticus and clarified the mechanistic basis for this form of antibiotic resistance.Furthermore, we demonstrated the expression of the 23S rRNA methylase gene, ermC, to attenuate the metabolism and growth of S. haemolyticus, suggesting it could be a fitness cost in the absence of antibiotic pressure.

FIG 1
FIG 1 Resistance genes related to erythromycin and clindamycin were detected in four S. haemolyticus strains A, B, C, and D. Plus or minus signs represent strains resistant or sensitive to antibiotics, respectively.Resistance mechanisms for these genes were shown on the right.Diamonds represent various antibiotics.

FIG 2
FIG 2 Identification of the target gene, ermC.Comparison of the ermC sequence identified from strain C with its homolog in NCBI.A short leader peptide was missing upstream.

FIG 3
FIG 3 Mechanism of inducible (A) and constitutive (B) MLS resistance.(C) Secondary structure prediction for a region upstream of ermC in S. haemolyticus strain C. (A) ermC mRNA was in an inactive conformation (upper portion of the panel) due to the structure of its 5ʹ end, which had a set of four inverted repeats paired as hairpin structure 1:2 and 3:4 that stalled the ribosomes and sequestered the initiation sequences for ermC.Thus, the methylase could not be synthesized because the initiation motifs for translation were not accessible to ribosomes.In the presence of inducer macrolides, the antibiotic (white diamond) first bound to the ribosome, dramatically slowing down the ribosome's speed and enabling translation of the leader peptide upstream of ermC.This induced the destabilization of the two hairpin structures and favored the association between inverted repeats 2 and 3 to form a new hairpin structure 2:3, which uncovered the initiation sequences that increase the translation efficiency of ermC.(B) The variations in the leader peptide decreased the stability of the hairpin structure and render ermC available for translation, resulting in constitutive MLS resistance.(C) In S. haemolyticus strain C, the deletion of the leader peptide caused the other inverted repeats to pair as hairpin structure 2:3 or 3:4.The minimum free energy of hairpin structure 2:3 was lower than 3:4, indicating that the secondary structure upstream of ermC was more likely to form as hairpin structure 2:3, which would expose initiation sequences for translation.

FIG 4 FIG 5
FIG 4 Read depth of S. haemolyticus strain C. (A) Black dots represented contigs from genome assembly files.The read depth of most contigs was consistent, while several were much higher than others, such as the one containing ermC.(B) Small black dots represented each annotated gene, with the read depth of ermC far above the average.

FIG 6
FIG6 Growth curves of S. aureus ATCC25923 transformed with different plasmids.ANOVA was used for statistical analysis, ***P < 0.001.