Edinburgh Research Explorer Susceptibility to Aminoglycosides and Distribution of aph and aac(3)-XI Genes among Corynebacterium striatum Clinical

Corynebacterium striatum is an opportunistic pathogen, often multidrug-resistant, which has been associated with serious infections in humans. Aminoglycosides are second-line or complementary antibiotics used for the treatment of Corynebacterium infections. We investigated the susceptibility to six aminoglycosides and the molecular mechanisms involved in aminoglycoside resistance in a collection of 64 Corynebacterium striatum isolated in our laboratory during the period 2005–2009. Antimicrobial susceptibility was determined using E-test. The mechanisms of aminoglycoside resistance were investigated by PCR and sequencing. The 64 C . striatum were assessed for the possibility of clonal spreading by Pulsed-field Gel Electrophoresis (PFGE). Netilmicin and amikacin were active against the 64 C . striatum isolates (MICs 90 = 0.38 and 0.5 mg/L, respectively). Twenty-seven of the 64 C . striatum strains showed a MIC 90 for kanamycin > 256 mg/L, and 26 out the 27 were positive for the aph(3’)-Ic gene. Thirty-six out of our 64 C . striatum were streptomycin resistant, and 23 out of the 36 carried both the aph(3”)-Ib and aph(6)-Id genes. The gene aac(3)-XI encoding a new aminoglycoside 3-N acetyl transferase from C . striatum was present in 44 out of the 64 isolates, all of them showing MICs of gentamicin and tobramycin > 1 mg/L. CS4933, a C . striatum showing very low susceptibility to kanamycin and streptomycin, con-tains an aminoglycoside resistance region that includes the aph(3’)-Ic gene, and the tandem of genes aph(3”)-Ib and aph(6)-Id . Forty-six major PFGE types were identified among the 64 C . striatum isolates, indicating that they were mainly not clonal. Our results showed that the 64 clinical C . striatum were highly resistant to aminoglycosides and mostly unrelated. . . against our . isolates mechanism involved in resistance to these two compounds. and showed poor activity against our strains (MICs 90 > 256 and > 1024 mg/L, respectively).


Introduction
The genus Corynebacterium consists of more than 100 species (https://www.dsmz.de/bacterialdiversity/prokaryotic-nomenclature-up-to-date, accessed November 18, 2016). Corynebacterium spp. are widely distributed in the environment. Pathogenic Corynebacterium species include Corynebacterium diphtheriae and nondiphtheroid Corynebacterium. Nondiphtheroid corynebacteria are part of the normal skin and mucous membranes microbiota of humans. They have traditionally been considered as sample contaminants when isolated from clinical stuffs. However, nondiphtheroid Corynebacterium are actually recognized as causing opportunistic disease [1], particularly in specific circumstances, such as repeated exposure to broadspectrum antibiotics, after the use of invasive medical procedures or in patients who have suffered long-term hospitalization [2,3,4]. Corynebacterium striatum has been implicated sporadically as etiological agent of a wide variety of infections, such as bacteraemia associated with central venous catheters, arthritis, vertebral osteomyelitis, septicaemia, meningitis, endocarditis, breast abscess, peritonitis, wound infections, and prosthetic joint infections [5], in both immunocompetent and immunocompromised patients [3]. In two C. striatum outbreaks, patient-to-patient transmission of a single C. striatum strain was reported [6,7].
Hospitals are "antibiotic-rich" environments, creating a selective pressure on the survival of resistant bacteria by which resistance genes can be transferred to susceptible bacteria, spreading multi-resistance. It has been suggested that this selective pressure favours the overgrowth of C. striatum as a secondary colonizer in immunocompromised hosts [6], so that C. striatum is considered a multi-resistant nosocomial pathogen [5]. In this context, the emergence of multidrug resistant strains is of particular concern, because the therapeutic options for the management of severely ill patients can be limited, increasing the reliance on vancomycin, which is still active against practically all clinical isolates of the C. striatum. Therapeutic efficacy of vancomycin can be improved if it is combined with a second antibiotic, most often an aminoglycoside. Thus, several cases of endocarditis due to C. striatum were successfully treated with gentamicin combined with vancomycin [8,9]. Therefore, aminoglycosides are considered as complementary antibiotics for the treatment of C. striatum infections.
There is a need for more data from clinical isolates to evaluate the risk of emergence of antimicrobial-resistant C. striatum strains and to track the transfer of resistance genes. In this respect, we aimed to investigate the susceptibility to six aminoglycosides of 64 clinical isolates of C. striatum from our sanitary area and characterized the molecular mechanisms of resistance. The 64 C. striatum were typed by means of PFGE to inquire about their clonal relationships and possible spread mechanisms.

Bacterial strains and growth conditions
Sixty-four C. striatum obtained from clinical samples of 63 patients collected during the period 2005-2009 at the Clinical Microbiology laboratory of the Marqués de Valdecilla University Hospital, Santander, Spain, were included in this study. Only in one case, two isolates from the same patient were included (C. striatum CS4108 and CS4489), since they showed different PFGE patterns and MICs. The 64 C. striatum were isolated from surgical wound exudates (32), non-surgical wound exudates (22, of which 7 were from diabetic foot lesions), blood (3), abdominal drainage (1), peritoneal ascitic fluid (2), sputum (2), gluteus abscess (1), and bronchoalveolar lavage fluid (1). They were grown on blood agar plates at 37˚C. All isolates were identified as putative C. striatum using API CORYNE V2 strips (bioMérieux, Marcy l'Etoile, France). C. striatum was differentiated from C. amycolatum using additional phenotypic tests (tyrosine hydrolysis, N-acetylglucosamine assimilation and phenylacetic acid assimilation). Identification was confirmed by MALDI-TOF, using the Vitek-MSTM (bioMérieux) system, according to manufacturer's instructions.

Susceptibility to aminoglycosides
The aminoglycosides used in this study were chosen because either they are frequently used in clinical practice (amikacin, gentamicin, netilmicin, and tobramycin) or are good markers of resistance mechanisms (kanamycin and streptomycin). MICs were determined using Etest1 (bioMérieux) on Mueller-Hinton agar (Pronadisa, Madrid, Spain), following the instructions of the manufacturer. Clinical categories were established according to breakpoints defined by the European Committee on Antimicrobial Susceptibility Testing (EUCAST) [10]. Since EUCAST has only defined breakpoints for gentamicin, MIC values of amikacin, kanamycin, netilmicin and tobramycin were interpreted according to EUCAST criteria for Staphylococcus spp. The EUCAST lacks breakpoints of streptomycin for staphylococci and those for enterococci are specified for detecting high-level resistance, which was not the primary objective of this study. For this reason we have considered the MIC breakpoints values proposed by the French Society of Microbiology (http://www.sfmmicrobiologie.org/UserFiles/files/casfm/ CASFM2013vjuin.pdf) to define the number of streptomycin sensitive, intermediate and resistant strains. Staphylococcus aureus ATCC 29213 and Enterococcus faecalis ATCC 29212 were used as control strains for susceptibility testing assays.
Pulsed-field gel electrophoresis and dendrogram analysis PFGE was performed with a CHEF-DRIII system (Bio-Rad, Hercules, California, USA). Bacteria were grown in Brain Heart Infusion agar (Pronadisa) with shaking at 37˚C for 48-72h. Cultures were adjusted to A 600 = 2.0, cells from 250 μl were pelleted and re-suspended in 300 μl of TE buffer (10 mM Tris, 1 mM EDTA) containing 2 mg/ml lysozyme. This suspension was incubated at 37˚C for 1 h, inverting the tubes every ten minutes. An equal volume of 2% LM agarose (Pronadisa) in TE buffer containing 1% SDS and 0.2 mg/ml proteinase K was added, and plugs were cast with a standard casting tray. After the plugs solidified, they were incubated overnight at 55˚C with shaking in 4 ml of TE buffer containing 1% sarcosyl and 0.15mg/ml proteinase K. The plugs were washed eight times with pre-warmed TE buffer and then digested with 30 U of XbaI at 37˚C overnight. Electrophoresis was performed in a 1% agarose gel at 6 V/cm and 14˚C with 0.5xTBE buffer. Pulse times ramped from 0.1 to 5 seconds for 18h. Low range PFGE marker (New England Biolabs, Beverly, USA) was used as the molecular size marker. Cluster analysis was performed with Fingerprinting II v4.5 software (Bio-Rad, Madrid, Spain) by using the Dice similarity coefficient and the unweighted pair Group method with arithmetic means (UPGMA), with 1% of optimization and tolerance. Isolates were classified as indistinguishable if they showed 100% similarity, as closely related subtypes if they showed 95-99% similarity, and as different strains if they showed <95% similarity.

Search of aminoglycoside resistance genes by PCR
The 64 C. striatum were screened for the presence of aminoglycoside modifying enzyme (AME) genes common in gram-positive bacteria [aac (3) and aac(6')-Ib] by PCR using primers specific for each gene (S1 Table) and by sequencing of the PCR products. As a control, the 24 isolates susceptible to aminoglycosides were also used in the PCR analysis. DNA was extracted using InstaGene matrix (Bio-Rad) according to manufacturer's instructions. Then, 10 μl of DNA template (10 pg/μl) was added to a reaction mixture containing 4 U of Taq DNA polymerase (Bioline, London, United Kingdom), 1.5 mM MgCl 2 , 5 μl of 10× PCR amplification buffer [160 mM (NH 4 ) 2 SO 4 , 670 mM Tris-HCl (pH = 8.8), 0.1% Tween 20], 40 pmol of each primer, 0.2 mM each deoxynucleoside triphosphate (Bioline), and double-distilled water to a final volume of 50 μl. DNA was first denatured at 95˚C for 5 min and then subjected to 30 cycles of amplification using a Perkin Elmer model 9600 thermal cycler under the following conditions: denaturation at 95˚C for 1 min, annealing for 1 min, and extension at 72˚C for 1 min. After the final cycle, the reactions were terminated by an extra run at 72˚C for 10 min, and the reaction products were then kept at 4˚C until analysis. Amplification products were characterized by electrophoresis in 1.5% agarose using 0.5x TBE running buffer. PCR products were cleaned of amplification primer using the QIAquick PCR Purification kit (Qiagen, Madrid, Spain) following the manufacturer's instructions. Purified DNA was sequenced with the primers outlined in S1 Table. Genome sequencing and analysis C. striatum CS4933 whole DNA was prepared from cultures grown on blood agar plates overnight. Bacterial cells were mixed with 2 ml of distilled water poured on the plate, transferred to a 2-ml Eppendorff tube, collected by centrifugation at 10,000g for 4 min, re-suspended in 0.25 ml of Tris-EDTA buffer containing 20 mg of lysozyme/ml and 50 mg of proteinase K/ml, and incubated at 37˚C for 1 h. Bacterial cells were then lysed by the addition of 0.25 ml of 0.1 M Tris containing 1% sodium dodecyl sulfate (SDS) and 400 μg/ml of proteinase K and incubated at 55˚C for 1 h. The lysate was mixed with 0.1 ml of 5 M NaCl and 100 μl of cetyltrimethylammonium bromide-NaCl and incubated at 65˚C for 10 min. DNA was then extracted with chloroform-isoamyl alcohol and phenol-chloroform, precipitated with isopropanol, and re-suspended gently in distilled water. For genomic sequencing genomic libraries were prepared using the Illumina TruSeq DNA Sample Prep Kit and sizeselected to a 400 bp mean insert size using Sage Pippin Blue. Final libraries were sequenced on a MiSeq platform using 250 base paired-end reads (v2 chemistry). Contigs were assembled with the MacVector tool (http://macvector.com), and analysed using the NCBI ORF finder (www.ncbi.nlm.nih.gov/gorf) and BLAST (www.ncbi.nlm.nih.gov/BLAST) tools. The sequence of the C. striatum CS4933 region including aph(3')-Ic, aph(3")-Ib and aph(6)-Id genes has been deposited in GenBank under accession number KP119857.

Susceptibility of C. striatum to aminoglycosides
According to the breakpoints previously indicated, 52 of the 64 C. striatum included in this study were resistant to at least one of the six aminoglycosides, whereas the remaining 12 isolates were aminoglycoside-susceptible. The MIC 50 and MIC 90 distributions of the different aminoglycosides for the 64 C. striatum are presented in Table 1.
Considering the MIC 90 , netilmicin was the more active aminoglycoside against the C. striatum we tested (MIC 90 = 0.38 mg/L), followed by amikacin (MIC 90 = 0.5 mg/L). All the isolates were susceptible to these two compounds. Martínez-Martínez et al [11] have previously reported that 27 out of 31 C. striatum were sensitive to amikacin. In this study, the activity of gentamicin was lower than that of netilmicin and amikacin, with the MIC 90 of gentamicin (8 mg/L) being 16 times higher than that of amikacin. Forty-five out of 64 isolates were resistant to gentamicin. Other studies reported variable activity of gentamicin against C. striatum: Campanile et al. [12] found a MIC 90 of 32 mg/L for 36 C. striatum isolated in Italy whereas Gómez-Garcés et al. [13] reported a MIC 90 of 2 mg/L for a collection of 30 Spanish C. striatum. Tobramycin was scarcely active against our C. striatum (MIC 90 = 16 mg/L). The 45 isolates tobramycin-resistant were resistant to gentamicin as well, suggesting a common mechanism involved in resistance to these two compounds. Kanamycin and streptomycin showed poor activity against our strains (MICs 90 >256 and >1024 mg/L, respectively).
According to EUCAST criteria, six resistance profiles were found among the 52 C. striatum isolates resistant to at least one of the tested aminoglycosides ( Table 2). Forty-four of the 64 isolates were resistant to more than one of the tested aminoglycosides. The most frequently encountered multi resistance phenotype was Kanamycin/Gentamicin/Tobramycin/Streptomycin (n = 24), followed by Gentamicin/Tobramycin (n = 13). Furthermore, eight isolates were resistant to just one of the considered aminoglycosides: seven to streptomycin and one to kanamycin.
The data obtained in this study may support the consideration of netilmicin and amikacin in combination with vancomycin for empiric therapy against infections caused by C. striatum, although more studies are necessary to confirm these results. In any case appropriate antimicrobial therapy against C. striatum infections must be prescribed on the basis of the data provided by antibiotic susceptibility tests for each strain.

Detection and mapping of aminoglycoside resistance genes
Inactivation by enzymatic modification is the most prevalent mechanism of resistance of corynebacteria to aminoglycosides in the clinical setting. The aph(3')-Ic gene, encoding an aminoglycoside-O-phosphotransferase conferring resistance to kanamycin, neomycin, paromomycin, ribostamycin and lividomycin, is part of plasmids and transposons and its wide distribution includes Corynebacterium spp. [14]. Twenty-six out of the 27 C. striatum resistant to kanamycin rendered the expected 0.48-kb amplicon when amplified with aph(3')-Ic specific primers. The isolate C. striatum CS3253 (MIC of kanamycin > 256 mg/L) was not amplified, suggesting that resistance is mediated by allelic variants of the gene or by another mechanism. The 37 C. striatum susceptible to kanamycin were negative when they were analysed by PCR- aph(3')-Ic. Therefore, in our C. striatum, there is a correlation between resistance to kanamycin and the presence of the aph(3')-Ic gene. As kanamycin is not prescribed in Spain, the high prevalence of this gene in C. striatum may be due to its genetic linkage with other resistance genes (see below). In Corynebacterium spp., the main mechanism of streptomycin resistance is related to the presence of the tandem of genes aph(3")-Ib and aph(6)-Id, encoding for aminoglycoside-3"phosphotransferase [APH (3")-Ib] and aminoglycoside-6-phosphotransferase [APH (6)-Id], respectively [14]. Thirty-six out of our 64 C. striatum were streptomycin resistant. In 23 streptomycin resistant C. striatum PCR amplification resulted in the expected aph(3")-Ib (0.55 Kb) and aph(6)-Id (0.51 Kb) amplicons. However, twelve out of the 24 C. striatum streptomycin resistant isolates did not give aph(3")-Ib neither aph(6)-Id amplicons, suggesting that in these strains streptomycin resistance is mediated by allelic variants of these genes or by another mechanism. One C. striatum (CS6865) was susceptible to streptomycin (MIC = 1 mg/L) but gave the aph(3")-Ib and aph(6)-Id amplicons, which might be due to mutations affecting the aph(3")-Ib-aph(6)-Id genes or their promoter region.
Forty-four out of 64 C. striatum were positive for the gene aac(3)-XI, encoding AAC(3)-XI, a new aminoglycoside 3-N acetyl transferase from C. striatum [15]. AAC (3)-XI acetyl transferase confers resistance to gentamicin, tobramycin, sisomicin, dibekacin and fortimicin. The 44 C. striatum carrying the aac(3)-XI gene showed MICs of both gentamicin and tobramycin > 1 mg/L. Nineteen out the 20 isolates negative for the aac(3)-XI gene were susceptible to both gentamicin and tobramycin (MICs < 0.064 mg/L) whereas one isolate showed a MIC = 1.5 mg/L for both compounds. Therefore, there is a good correlation between resistance to these two compounds and the presence of the aac(3)-XI gene.
The 44 isolates positive for the gene aac(3)-XI have a gentamicin MIC range of 1-16 mg/L, supporting the EUCAST clinical breakpoint of > 1 mg/L. If the gentamicin breakpoint recommended by the CLSI was used (>16 mg/L) [16], most of the isolates positive for the aac(3)-XI gene are not considered as resistant. This difference in interpretation between the European and the US recommendations could result in an underestimate of gentamicin resistant C. striatum if the CLSI breakpoint instead of the EUCAST breakpoint was used, with possible serious consequences for patients.
In order to know the genomic context of the aminoglycoside resistance genes and inquiry about their transfer mechanisms we sequenced the genome of the isolate C. striatum CS4933, which was chosen for its clinical relevance (it was isolated from soft tissue abscess) and its resistance profile (it was resistant to kanamycin and streptomycin). Annotation of the genome sequence of C. striatum CS4933 revealed that the aph(3')-Ic gene is flanked by two divergently oriented IS26-like elements (Fig 1), forming a new transposon similar to Tn5715 but with the aph(3')-Ic gene and the IS26 modules inversely oriented [17]. The aph(3')-Ic gene is part of a larger DNA region containing the aph(3")-Ib-aph(6)-Id tandem pair of resistance genes conferring streptomycin resistance. The aph(3")-Ib-aph(6)-Id coding regions of C. striatum CS4933 are located downstream of the resolvase gene from a Tn5393-like transposon that is disrupted by the Tn5715 element containing the aph(3')-Ic gene. Essentially identical aminoglycoside resistance regions were found in the plasmid pTP10 from C. striatum [17], in the chromosome of C. urealyticum DSM1709 [18] and in the plasmid pJA144188 of C. resistens DSM 45100 [19] (Fig 1).

Molecular epidemiology of the C. striatum isolates
The PFGE method displayed a high tipability (100% of the isolates were clearly assigned to one pulsotype), and discriminatory power. PFGE delineated the 64 C. striatum isolates into forty-six distinct PFGE types (Fig 2). On these, PFGE types 3, 8, 34 and 46 could be further classified into 2 subtypes. Most PFGE patterns corresponded to single isolates, whereas one PFGE pattern was observed for 7 isolates, 3 patterns contained three isolates each, and 6 patterns two isolates each. The dendrogram showed a Dice similarity coefficient ranging from 63 to 97% (Fig 2).
Most of the pulsotypes include strains showing the same resistance properties. The 3 strains of the PFGE type 4 were resistant to streptomycin and positive for the genes aph(3")-Ib and aph(6)-Id, whereas the 3 strains of the type 23 were sensitive to all aminoglycosides. The 3 strains of the pulsotype 28 were resistant to four aminoglycosides (kanamycin, streptomycin, gentamicin and tobramycin) and contain the aph(3')-Ic, aph(3")-Ib, aph(6)-Id and aac(3)-XI genes.