High prevalence of carbapenem-hydrolysing oxacillinases in epidemiologically related and unrelated Acinetobacter baumannii clinical isolates in Spain

Authors


Corresponding author and reprint requests: J. Vila, Servei de Microbiologia, Centre de Diagnòstic Biomèdic, Hospital Clinic, Facultat de Medicina, Universitat de Barcelona, Barcelona, Spain
E-mail: jvila@ub.edu

Abstract

Carbapenem-hydrolysing oxacillinases are reported increasingly in Acinetobacter baumannii. This study investigated the role of these β-lactamases in causing resistance to carbapenems in 83 epidemiologically related and unrelated imipenem-resistant A. baumannii clinical isolates. The isolates were also analysed for the presence of ISAba1 in the promoter region of the blaOXA-51-like gene in order to investigate the role of ISAba1 in OXA-51 expression. All clinical isolates contained a blaOXA-51-like gene, 20% contained a blaOXA-58-like gene, and 42% contained a blaOXA-40-like gene; blaOXA-23-like, blaIMP and blaVIM genes were not detected in any of the isolates investigated. ISAba1 was found in 24 (82.7%) of 28 pulsetypes, and was located in the promoter region of the blaOXA-51-like gene in five (20.8%) of these pulsetypes. Expression of blaOXA-51 was detected in the five isolates with ISAba1 located in the promoter region, but was not detected in an isogenic imipenem-susceptible A. baumannii isolate that did not have ISAba1 located in the promoter region. It was concluded that there is a high prevalence of oxacillinases with activity against carbapenems among genetically unrelated A. baumannii clinical isolates from Spain, and that concomitant expression of two carbapenemases (OXA-51-like and either OXA-40-like or OXA-58-like) may take place. Insertion of an ISAba1-like element in the promoter of the blaOXA-51-like gene promotes the expression of this gene, although this did not seem to play a major role in carbapenem resistance.

Introduction

Acinetobacter baumannii is a nosocomial pathogen that is recognised as being responsible for a wide spectrum of infections, including bacteraemia, secondary meningitis, pneumonia and urinary tract infections [1]. It has been implicated increasingly in hospital-acquired infections, mostly affecting debilitated patients in intensive care units, in whom such infections are associated with high mortality rates [1]. Administration of appropriate antimicrobial therapy to these patients is therefore essential. Carbapenems usually have good potency against A. baumannii, with imipenem being the most active agent [1], but carbapenem resistance in A. baumannii has increasingly been reported worldwide during the last decade [2,3]. Several mechanisms responsible for resistance to carbapenems in A. baumannii have been described: (i) synthesis of carbapenemases [2]; (ii) decreased outer-membrane permeability caused by the loss or reduced expression of porins [4–8]; and (iii) alterations in penicillin-binding proteins [9,10].

Although IMP-type [11–14] and VIM-type [15] carbapenemases have been reported in A. baumannii, the carbapenemases found most frequently are those belonging to class D. To date, four groups of carbapenem-hydrolysing oxacillinases (Ambler class D β-lactamases) have been described in A. baumannii[16–20]. The first oxacillinase described in A. baumannii with activity against carbapenems was OXA-23 [21], and this enzyme is in a group that currently includes OXA-23, OXA-27 and OXA-49 [17,22]. A second group comprises OXA-24, OXA-25, OXA-26 and OXA-40 [16,22,23], and shares 60% amino-acid identity with the first group. The third and largest group comprises the OXA-51-like carbapenemases, which are encoded by chromosomally located genes that show 56% and <63% amino-acid identity with groups 1 and 2, respectively [24,25]. Finally, a new carbapenemase (OXA-58) has been characterised that shares <50% amino-acid identity with the other three groups.

Recently, a novel insertion sequence (IS), ISAba1 (http://www-is.biotoul.fr/is.html), which has 11-bp inverted repeat sequences flanked by 9-bp direct repeats of the target sequence, was identified in A. baumannii[26]. Many IS elements contain promoters that play a role in the expression of antibiotic resistance genes situated downstream from the site of insertion [27,28]. ISAba1 has been identified adjacent to a β-lactamase resistance gene (ampC) in A. baumannii[26,29], and primer extension studies showed that transcription of the ampC gene was dependent on promoter sequences within ISAba1[27]. ISAba1 has also been found upstream from blaOXA-51-like and, probably, blaOXA-23-like genes [30,31].

The aim of the present study was to investigate the distribution of the different OXA-type enzymes, as well as VIM and IMP enzymes, in a collection of A. baumannii isolates from various locations in Spain. The presence of ISAba1 upstream of blaOXA-51-like genes, together with its role in the expression of these genes, was also investigated.

Materials and methods

Bacterial isolates

Eighty-three isolates were selected from among 221 clinical isolates collected during November 2000 from 25 Spanish hospitals as part of a previous study of carbapenem resistance. All isolates were identified by amplified rDNA restriction analysis [32], and their epidemiological relationships were determined by pulsed-field gel electrophoresis, according to the method of Gautom [33].

Susceptibility testing

Microdilution assays according to CLSI guidelines [34] were used to determine MICs of the following antimicrobial agents: ampicillin, piperacillin, cephalothin, cefoxitin, gentamicin, amikacin, tobramycin, tetracycline, minocycline, doxycycline, rifampicin and colistin (Sigma, Madrid, Spain), ceftazidime (GlaxoSmithKline, Uxbridge, UK), cefepime (Bristol-Myers Squibb, Madrid, Spain), sulbactam and azithromycin (Pfizer, Sandwich, UK), imipenem (Merck, Hoddesdon, UK), meropenem (AstraZeneca, Macclesfield, UK), ciprofloxacin (Bayer, Leverkusen, Germany), and co-trimoxazole (Galloso, Madrid, Spain) [35]. Breakpoints used were those recommended by the CLSI for non-fermentative Gram-negative bacilli [34]; control strains used were those described previously [35].

PCR analysis

DNA was extracted by boiling a single colony in 25 µL of water for 10 min and then centrifuging in a microcentrifuge at maximum speed for 1 min. This was followed by the addition of 25 µL of a reaction mixture containing 0.5 µM of the relevant primers (Table 1), 200 µM dNTPs and 2.5 U of Taq DNA polymerase, to give a final volume of 50 µL. Initial denaturation (95°C for 5 min) was followed by 30 cycles of 95°C for 1 min, the optimal annealing temperature for each gene (Table 1) for 1 min, and 72°C for 1 min, with a final extension step at 72°C for 10 min. PCR products were resolved in agarose 2% w/v gels in TBE buffer (89 mM Tris base, 89 mM boric acid, 2 mM EDTA). Gels were stained with ethidium bromide and the DNA was visualised by UV light transillumination at 302 nm. When required, PCR products were recovered directly from the agarose gels and were purified with the Wizard SV gel and PCR system (Promega, Madison, WI, USA) according to the manufacturer's instructions. DNA sequencing was performed using a BigDye Terminator v.3.1 Cycle Sequencing Kit (Applied Biosystems, Warrington, UK) and an automated DNA sequencer 3100 Genetic Analyzer (Applied Biosystems).

Table 1.   Primers used in this study
PrimersGenes detectedSequence (5’→3’)Product size (bp)Annealing temp.
OXA51 Uoxa51, oxa69, oxa71, oxa75, oxa 78AACAAGCGCTATTTTTATTTCAG64153°C
OXA51 L CCCATCCCCAACCACTTTT  
OXA58 Uoxa58AGTATTGGGGCTTGTGCT45350°C
OXA58 L AACTTCCGTGCCTATTTG  
OXA24 Uoxa24, oxa25, oxa26, oxa33, oxa40, oxa72ATGAAAAAATTTATACTTCCTATATTCAGC82550°C
OXA24 L TTAAATGATTCCAAGATTTTCTAGC  
OXA23 Uoxa23, oxa27, oxa49GATGTGTCATAGTATTCGTCGT64152°C
OXA23 L TCACAACAACTAAAAGCACTGT  
ISaba1UISAba1CATTGGCATTAAACTGAGGAGAAA45153°C
ISaba1L TTGGAAATGGGGAAAACGAA  
VIM UVIM-typeATTGGTCTATTTGACCGCGTC78055°C
VIM L TGCTACTCAACGACTGCGCG  
IMP UIMP-typeCATGGTTTGGTGGTTCTTGT48855°C
IMP L ATAATTTGGCGGACTTTGGC  

Quantification of mRNA by RT-PCR

One milliliter of bacterial culture (OD600 0.6) was added rapidly to a solution comprising 125 µL of ethanol 95% v/v plus phenol 5% v/v, and was then centrifuged. The pellet was resuspended in 100 µL of lysozyme solution in water (0.1 mg/mL), vortexed, and then incubated for 30 min at room temperature, after which mRNA was extracted and purified using an RNAwiz kit (Ambion, Austin, TX, USA). RT-PCR was then performed using a SuperScrip One-step RT-PCR Kit with Platinum Taq (Invitrogen, Barcelona, Spain). Two sets of primers, OXA51 U and OXA51 L (Table 1), were used for the blaOXA-51-like gene, with primers for the 16S rRNA gene as an internal control. Reaction mixtures comprised 1× reaction mix (SuperScrip One-step RT-PCR Kit), 0.5 μM each primer, 1 U of RT/platinum Taq MIX (SuperScrip One-step RT-PCR Kit), 500 ng of RNA template, and distilled water to 50 µL. Each reaction was performed with two initial steps, the first at 50°C for 30 min (reverse transcription), and the second at 95°C for 2 min to activate the Taq polymerase, followed by 19 cycles of 95°C for 1 min, 53°C for 1 min and 72°C for 1 min.

For quantification, it is important to stop the reaction, usually between cycles 10 and 25, to compare the expression of a gene in different isolates. After several trials, 19 cycles were used for each of the two genes. This low cycle number produces amplicons that are difficult to see in an agarose gel stained with ethidium bromide; the RT-PCR products obtained were therefore analysed in acrylamide gels (Amersham Biosciences, Barcelona, Spain) using a GenePhor apparatus (Pharmacia Biotech, Barcelona, Spain). The gel was then stained using a DNA silver-staining kit (Amersham Biosciences).

Results

In total, 221 A. baumannii clinical isolates were collected from 25 Spanish hospitals [35]. All 83 imipenem-resistant A. baumannii isolates from this collection were chosen for the present study. The 83 selected isolates belonged to 28 different pulsetypes from 12 different hospitals. The number of isolates in each pulsetype ranged from one to 18. Table 2 summarises the presence of the genes encoding the main carbapenemases described to date in A. baumannii, including the four subgroups of oxacillinases having carbapenemase activity, as well as IMP- and VIM-type enzymes. All the isolates carried a gene encoding a β-lactamase belonging to the OXA-51-like group. In addition, 19% and 42% of the isolates carried a gene encoding an OXA-58-like or an OXA-40-like enzyme, respectively. The MICs of imipenem for isolates carrying an OXA-40-like or an OXA-58-like carbapenemase were ≥128 mg/L and 16–64 mg/L, respectively. No genes encoding β-lactamases of the OXA-23-like group, or VIM or IMP metallo-β-lactamases, were detected. Sixty-two (74.7%) of the 83 isolates contained the insertion sequence ISAba1, which was located in the promoter region of the blaOXA-51-like gene for 11 (17.7%) of the 62 isolates carrying this insertion element. Some degree of heterogeneity was observed within groups of isolates; thus, pulsetype 72 contained 18 isolates, but only one carried a blaOXA-40-like gene, and pulsetype 53 contained six isolates, five of which carried a blaOXA-58-like gene, with four of these overexpressing a blaOXA-51-like gene, while the remaining isolate did not carry a blaOXA-58-like gene and overexpressed a blaOXA-51-like gene.

Table 2.   Distribution of different carbapenemase-encoding genes among isolates of Acinetobacter baumannii from various locations in Spain
HospitalCityNo. of isolatesPulsetype P designationGenes detectedImipenem MIC (mg/L)
OXA-51-likeOXA-58-likeOXA-40-likeISAba1IS-OXA-51-likea
  1. City: BCN, Barcelona; MD, Madrid; SLM, Salamanca; SAN, Santander; TO, Toledo; CO, Córdoba; GR, Granada; SE, Sevilla.

  2. P, number assigned to each pulsetype.

  3. IS-OXA-51-like, isolates that contain ISAba1 inserted in the promoter region of the blaOXA-51-like gene.

H. BellvitgeBCN287140040128
72180120 64–128
732000064
741000064
7510010128
7810010128
7910110128
H. 12 OctubreMD 44440440 64–128
H. C. SalamancaSLM 1111000064
H. La PrincesaMD 24120220128
H. M. ValdecillaSAN 2322202016
H. ParapléjicosTO 31510110128
2220220128
H. Reina SofiaCO 64733030 16–64
4832030 32–64
H. Vall d’HebronBCN 14510011128
H. V. SaludTO 65050550128
5210110128
H. V. NievesGR 75365065 16–32
541001116
H. V. RocioSE205690990128
5720220128
5920220128
6133033 16–32
61.230330128
651100016
H. GetafeMD 33420220128
3510011128
Total83288316356211 
%  10019.342.274.713.3 

Expression of the blaOXA-51-like gene was investigated in five isolates belonging to pulsetypes 35, 45, 53, 54 and 61, respectively, in which ISAba1 was located in the promoter region of the blaOXA-51-like gene. In isolates from three pulsetypes (35, 45 and 54), the blaOXA-51-like gene was the only carbapenemase, whereas the remaining two isolates (pulsetypes 53 and 61) also carried a blaOXA-58-like carbapenemase. In addition, an isolate belonging to pulsetype 53, which also carried ISAba1, but not in the promoter region of the blaOXA-51-like gene, was included as a control. Fig. 1 shows that the blaOXA-51-like gene was clearly expressed in all five isolates with ISAba1 in the promoter region of the gene (lanes 2–6), but not in the control isolate (lane 1). The PCR product obtained from the amplification was verified by sequencing using the ISAba1 forward primer and the blaOXA-51-like gene reverse primer.

Figure 1.

 Expression of blaOXA-51-like genes (lanes 1–6) and, as controls, 16S rRNA genes (lanes 7–12). Lanes 1–6 and 7–12 show isolates belonging to pulsetypes 53, 35, 45, 53, 54 and 61, in that order. Lanes 1 and 7 contain an isolate belonging to pulsetype 53 that carries ISAba1, but not in the promoter region of the blaOXA-51-like gene. All other lanes contain isolates with ISAba1 located in the promoter region of the blaOXA-51-like gene.

Discussion

A. baumannii is associated frequently with nosocomial infections, with imipenem being the most active antimicrobial agent against these microorganisms. Acquisition of imipenem resistance means that colistin, and perhaps tigecycline, may be the only remaining therapeutic options for treating infections caused by multiresistant A. baumannii. The high level of resistance to imipenem in A. baumannii clinical isolates and the clinical risk-factors favouring the acquisition of imipenem-resistant A. baumannii have been reported previously [36]. Although carbapenem resistance may be caused, in part, by impaired permeability, resulting from decreased expression of porins, or by modifications in penicillin-binding proteins [4–10], most recent reports have indicated that carbapenem-hydrolysing β-lactamases play a more significant role [2]. In the present study, all isolates produced an OXA-51-like β-lactamase, with those that also produced an OXA-40-like β-lactamase having a higher imipenem MIC than isolates that also produced an OXA-58-like enzyme (Table 2). Other mechanisms of resistance to imipenem may be present, possibly including over-expression of the adeABC efflux pump [37].

Da Silva et al. [38] showed that a multiresistant epidemic clone of A. baumannii carrying the blaOXA-40 gene was disseminated widely in Portugal and Spain. In the present study, epidemiologically unrelated A. baumannii clinical isolates carrying blaOXA-40-like genes were more prevalent than isolates carrying blaOXA-58-like genes, suggesting that dissemination of a genetic element carrying a blaOXA-40-like gene may have taken place. However, the spread of an A. baumannii clone or a genetic element carrying this gene seems, at present, to be limited to Spain and Portugal [38].

The OXA-58 carbapenemase described by Poirel et al. [20] shares <50% amino-acid homology with the three remaining OXA groups. OXA-58 was described originally in a carbapenem-resistant isolate of A. baumannii from France, but this carbapenemase has been found subsequently in isolates from Austria, Greece, Romania, Spain, Turkey and the UK [39–42], Argentina and Kuwait [39], and Venezuela (E. Salazar, personal communication). In the present study, an OXA-58-like enzyme was found in six of 28 A. baumannii clinical isolates belonging to different pulsetypes, again suggesting that a genetic element carrying a blaOXA-58-like element may be disseminating.

The heterogeneity among the OXA-51-like group of enzymes is probably very high [24,41], and it is known that the OXA-51-like enzymes are intrinsic chromosomally-located β-lactamases in A. baumannii[43]. In the present study, eight PCR products from the amplified region of the blaOXA-51-like gene were randomly chosen and sequenced, and were found to show 100% homology with either the blaOXA-66/76 gene (PCR products from six different isolates) [25], the blaOXA-71 gene (one isolate) [25], or the blaOXA-51 gene (one isolate) [24]. Héritier et al. [43] showed that OXA-69, which is closely related to OXA-51, had only very weak catalytic efficiency for most β-lactam antibiotics, including carbapenems, and suggested that OXA-69 may not play a significant role in resistance to β-lactam antibiotics. The present study demonstrated over-expression of the blaOXA-51-like gene in all isolates with ISAba1 located in the promoter region of the blaOXA-51-like gene. Although another mechanism of resistance to carbapenems cannot be ruled out, the differing expression of the blaOXA-51-like gene observed in the isogenic sensitive and resistant isolates belonging to pulsetype 53 suggests that insertion of ISAba1 in the promoter region of the blaOXA-51-like gene may produce a slight increase in the MIC of imipenem. The end of the ISAba1 element found inserted in the promoter region of the blaOXA-51 gene was located seven bases from the start codon of the blaOXA-51-like gene, as described previously by Turton et al. [31].

In conclusion, although additional mechanisms of resistance to carbapenems cannot be ruled out, the present study demonstrated a high prevalence of oxacillinases with activity against carbapenems in genetically unrelated A. baumannii clinical isolates from Spain, and confirmed that insertion of an ISAba1-like element in the promoter region of blaOXA-51-like genes enhances the expression of such genes, and produces a small increase in the imipenem MIC.

Acknowledgements

We thank the Grupo de Estudio de la Infección Hospitalaria (GEIH) from the Sociedad Española de Enfermedades Infecciosas y Microbiología Clínica (SEIMC) for supporting this study. This study was also supported by Grant SGR050444 from the Departmanet d'Universitats, Recerca I Societat de la Informació de la Generalitat de Catalunya, Spain (to J.V.) and by the Ministerio de Sanidad y Consumo, Instituto de Salud Carlos III, Spanish Network for the Research in Infectious Diseases (REIPI C03/14) and Spanish Network for the Research in Infectious Diseases (REIPI RD06/0008). M.R. and S.M. are each in receipt of a fellowship from REIPI.

Ancillary