Whole genome sequence of pan drug-resistant clinical isolate of Acinetobacter baumannii ST1890

Acinetobacter baumannii is an opportunistic gram-negative bacteria typically attributed to hospital-associated infection. It could also become multidrug-resistant (MDR), extensively drug-resistant (XDR), and pan drug-resistant (PDR) during a short period. Although A. baumannii has been documented extensively, complete knowledge on the antibiotic-resistant mechanisms and virulence factors responsible for pathogenesis has not been entirely elucidated. This study investigated the drug resistance pattern and characterized the genomic sequence by de novo assembly of PDR A. baumannii strain VJR422, which was isolated from a catheter-sputum specimen. The results showed that the VJR422 strain was resistant to any existing antibiotics. Based on de novo assembly, whole-genome sequences showed a total genome size of 3,924,675-bp. In silico and conventional MLST analysis of sequence type (ST) of this strain was new ST by Oxford MLST scheme and designated as ST1890. Moreover, we found 10,915 genes that could be classified into 45 categories by Gene Ontology (GO) analysis. There were 1,687 genes mapped to 34 Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. The statistics from Clusters of Orthologous Genes (COG) annotation identified 3,189 genes of the VJR422 strain. Regarding the existence of virulence factors, a total of 59 virulence factors were identified in the genome of the VJR422 strain by virulence factors of pathogenic bacteria databases (VFDB). The drug-resistant genes were investigated by searching in the Comprehensive Antibiotic Resistance Database (CARD). The strain harbored antibiotic-resistant genes responsible for aminoglycoside, β-lactam-ring-containing drugs, erythromycin, and streptogramin resistance. We also identified resistance-nodulation-cell division (RND) and the major facilitator superfamily (MFS) associated with the antibiotic efflux pump. Overall, this study focused on A. baumannii strain VJR422 at the genomic level data, i.e., GO, COG, and KEGG. The antibiotic-resistant genotype and phenotype as well as the presence of potential virulence associated factors were investigated.

Introduction Acinetobacter baumannii is an opportunistic, gram-negative coccobacilli [1] commonly associated with hospital-acquired nosocomial infections that can cause bacteremia, pneumonia, meningitis, and urinary tract infections [2,3]. It is also considered the most common organism in the intensive care unit (ICU) and has been recognized as an emerging cause of nosocomial outbreaks globally [3]. The Infectious Diseases Society of America (IDSA) reported that A. baumannii ranked in the six top for priority dangerous microorganisms [4].
With serious concern for a multidrug-resistant (MDR) crisis, multidrug-resistant A. baumannii is one of the most alarming strains in terms of treatment and control. MDR has been increased all over the world that is considered a public health threat. Several recent investigations reported the emergence of multidrug-resistant bacterial pathogens from different origins including humans, birds, cattle, and fish that increase the need for routine application of the antimicrobial susceptibility testing to detect the antibiotic of choice as well as the screening of the emerging MDR strains [5][6][7][8][9][10][11][12][13][14][15]. MDR exhibited antibiotic resistance to different antibiotic including β-lactams, fluoroquinolones, tetracyclines, and aminoglycosides [16]. One of the major explanations for multidrug-resistant A. baumannii is the expression of resistance-nodulation-division (RND) transporters or outer membrane proteins, which actively pump drugs out of the cells [17,18]. Treatment with colistin and tigecycline has become the only remaining active antibiotics treatment and the last resort in terms of treatment for MDR-A. baumannii [19][20][21]. Therefore, MDR-A. baumannii was also suggested as being extensively drug-resistant (XDR), which refers to resistance to all antibiotics except colistin and tigecycline, and an XDR-A. baumannii is a common cause of severe healthcare-associated infections worldwide [22,23]. Recently, pan drug-resistant (PDR) A. baumannii strains have also been reported to resist colistin and tigecycline [24][25][26][27]. The emergence of PDR-A. baumannii has increased mortality rates and limited treatment management because monotherapy treatment is insufficient for curing [28][29][30]. Several virulence factors responsible for the pathogenicity of A. baumannii have been identified, including pilus, outer membrane porins, phospholipases, proteases, lipopolysaccharides, capsular polysaccharides, protein secretion systems, and iron-chelating systems. Some strains share genes related to adherence, invasion and survival, and form biofilms on the surface [31,32].
Acinetobacter baumannii is one of the most successful pathogens responsible for hospitalacquired nosocomial infections because of the high prevalence of infections and scarcity of effective antibiotics for treatment. To overcome this problem, knowledge of the antibioticsresistant mechanism and virulence factors responsible for pathogenesis is necessary.
Advances in whole-genome sequencing technology have facilitated bacterial whole-genome characterization, enhancing the ability to elucidate the antibiotic-resistant mechanism and pathogenesis in Acinetobacter baumannii [33,34]. However, data concerning genome analysis on colistin resistance of Acinetobacter baumannii isolated from Thailand is currently limited in the literature [35]. To understand the antibiotics-resistant mechanism and virulence factors in A. baumannii, we described the whole genome of PDR-A. baumannii strain VJR422 by using de novo assembly with Illumina technology. The prediction of gene annotation and functional annotation employed a public database. Genome studies were also applied to predict potential antibiotics-resistant genes and virulence factors in this strain. The identification of genes involved in antibiotics resistance as well as virulence factors could be a potentially rewarding step towards a better understanding of the mechanism for antibiotics resistance in A. baumannii and could also provide foundational information for developing potential clinical management and treatment in the future.

Ethics statement
This study was approved by the Ethics Committee of the Faculty of Medicine Vajira Hospital, Navamindradhiraj University, Bangkok, Thailand (COA 009/2561). Informed consent was waived because the study and the analysis used anonymous clinical data.

Isolation and identification of Acinetobacter baumannii
In this study, we obtained A. baumannii strain VJR422 isolated from the single catheter-aspirated sputum of a patient who received care at our hospital in 2017, at the Faculty of Medicine Vajira Hospital, Navamindradhiraj University, Bangkok, Thailand. The bacterial strain was cultured on 5% sheep blood agar, chocolate agar, and MacConkey agar (commercially prepared by Clinical Diagnostics Ltd., Part, Bangkok, Thailand), followed by incubation for 18-24 hours at 35˚C. Subsequently, the strain was identified using the matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) (Bruker Microflex, Bremen, Germany).

Antimicrobial susceptibility testing
Antimicrobial susceptibility testing was also performed by using a BD Phoenix NMIC-203 commercial kit (Becton Dickinson Diagnostic Systems, Sparks, Maryland, USA) on the BD Phoenix Automated Microbiology System (Becton Dickinson Diagnostic Systems, Spark, Maryland, USA). The manufacturer's instructions were followed. Determining the minimum inhibitory concentration (MIC) breakpoint was done according to Clinical and Laboratory Standards Institute (CLSI) guidelines (M100, 27 th Ed.) [36]. The antimicrobials tested were cefoxitin, ceftazidime, ceftriaxone, imipenem, meropenem, aztreonam, ciprofloxacin, gentamicin, piperacillin/tazobactam, trimethoprim/sulfamethoxazole, tigecycline, and colistin. The control strains were Escherichia coli ATCC 25922 and Pseudomonas aeruginosa ATCC 27853 as recommended by the CLSI.

Whole-genome DNA sequencing and analysis
Genomic DNA was extracted using a QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany) following the manufacturer's recommended protocol. The extracted DNA was visualized for quality on 0.8% w/v agarose gel electrophoresis, quantified with a NanoDrop 2000 spectrophotometer (Thermo Fisher, Wilmington, DE, USA), and stored at −30˚C until further use. The poly-A-tailed DNA was ligated to paired-end adaptors and PCR amplified with a 500-bp insert. A mate-pair library was constructed with an insert size of 5 kb at Beijing Novogene Bioinformatics Technology Co., Ltd., Beijing, China. Whole-genome sequencing was performed on the Illumina platform with MPS (massively parallel sequencing) technology. Paired-end low-quality reads, mate-pair library, and PCR adaptor read were filtered by the quality control step using a Beijing Novogene Bioinformatics Technology pipeline. All good-quality paired reads were assembled using SOAP de novo (http://soap.genomics.org.cn/soapdenovo.html) into several scaffolds [37,38]. In the next step, the filter reads were processed by gap closing.

In silico Multilocus Sequence Typing (MLST)
In silico Multilocus Sequence Typing (MLST) and sequence types (STs) from whole-genome sequence data was performed using the MLST

Conventional MLST
Genomic DNA was extracted by using the QIAamp DNA Mini Kit (QIAGEN, Hilden, Germany). MLST was carried out on extracted DNA using the methodology described by Bartual et al. In brief, fragments of seven housekeeping genes (gltA, gyrB, gdhB, recA, cpn60, gpi, and rpoD) were amplified by the polymerase chain reaction (PCR). The amplified PCR products were then purified using polyethylene glycol-sodium chloride (PEG-NaCl) precipitation (20% w/v of PEG, 2.5 M NaCl). Both strands of all PCR products were fully sequenced by A T G C Co; Ltd. (Pathum Thani, Thailand). The obtained sequences were assigned allele numbers by using the MLST website (https://pubmlst.org/organisms/acinetobacter-baumannii). The ST code was generated based on the combination of detected alleles for gltA, gyrB, gdhB, recA, cpn60, gpi, and rpoD with the Oxford scheme [40].

Functional gene annotation
Three databases were used to predict gene function: 1. GO-The Gene Ontology database [54], 2. KEGG-The Kyoto Encyclopedia of Genes and Genomes [55-57], and 3. COG-Protein sequences can be classified into Clusters of Orthologous Groups of proteins [58] Virulence factors were analyzed using the VFDB (Virulence Factors of Pathogenic Bacteria) [59] for a virulence factor search. The abundance of resistant genes was detected by using the RGI program (version 4.2.2) to identify drug resistant genes by comparison with the reference genome in the Comprehensive Antibiotic Resistance Database (https://card.mcmaster.ca/) (submitted file to the online database and analyzed on 4 th January 2020).

Expression levels of efflux pump genes
Total RNA was extracted using conventional hot phenol RNA extraction and converted into cDNA using a cDNA synthesis kit (iScript Reverse Transcription Supermix, Bio-Rad, Hercules, CA, United States). The quality and purity of the RNA were evaluated using a Nanodrop-100 spectrophotometer (Nanodrop Technology Inc., Wilmington, DE, USA). Real-time quantification of cDNA was carried out on a CFX96 Touch TM real-time PCR detection system (Biorad, California, USA) using the iScript SYBR green PCR master mix (Bio-Rad, Hercules, CA, United States). The amplification cycle included initial denaturation at 95˚C for 1 minute and 40 cycles of denaturation at 95˚C for 15 seconds followed by annealing and extension at 62˚C for 30 seconds. The primers used for amplification included adeB_RT_F: GGATTA TGGCGACTGAAGGA and adeB_RT_R: AATACTGCCGCCAATACCAG for adeB [60], adeG_RT_F: CGTAACTATGCGGTGCTCAA and adeG_RT_R: ATCGCGTAGTCACCAG AACC for adeG [60], and adeJ_RT_F: CATCGGCTGAAACAGTTGAA and adeJ_RT_R: GCCTGACCATTACCAGCACT for adeJ [60]. Relative expression values were determined using the 2 -ΔΔCt method. A. baumannii ATCC 19606 was used as a standard strain for normalization of relative mRNA levels.

Bacterial isolate, identification, and antimicrobial susceptibility test
A. baumannii strain VJR422 was isolated from the catheter-aspirated sputum of a patient who admitted to Vajira hospital. Colonies on blood agar and chocolate agar were smooth, raised, and opaque with non-lactose fermenter colonies on MacConkey agar. The antimicrobial susceptibility profile of this strain was resistant to all antibiotics tested, as shown in Table 1.
The COG database is divided into 25 parts according to function. The results in Fig 4. were obtained from the statistics of COG annotation for the VJR422 strain. A total of 3,189 genes were annotated and classified into 22 functional groups. No genes were allocated to the chromatin structure and dynamics, nuclear structure, and cytoskeleton functional domains. Among the COG functional classifications, "General function prediction only" comprised the largest group (314 genes), followed by "Amino acid transport and metabolism" (294 genes), and "Transcription" (273 genes). Moreover, 179 genes were classified as "Function unknown" (Fig 4.).
The COG categories in the genome of VJR422 were divided into four main domains comprising 1) Information storage and processing, 2) Cellular process and signaling, 3) Metabolism, and 4) Poorly characterized [62,67]. All were compared with another COG toward the genome of A. baumannii DMS06669 [68] and A. baumannii XH386 [69] in Table 3. The predominance of VJR422 includes extracellular structures, annotated genes, secondary metabolite biosynthesis, transport, and catabolism-annotated genes.
The abundance of resistant genes in the VJR422 strain was assessed by searching the Comprehensive Antibiotic Resistance Database (CARD). The CARD includes BLAST and the Resistance Gene Identifier (RGI) software tools for the analysis of molecular sequences, the prediction of the resistome based on homology and single nucleotide polymorphism (SNP) models. The distribution of antibiotic resistance genes in the genome of VJR422 is also shown in S4 Table. A total of 25 genes that respond to different mechanisms of drug-resistance in Acinetobacter were identified by the CARD tool. The percentage identity of the matching region with the reference sequence in the program was in a range from 98.98% to 100%. A. baumannii VJR422 was resistant to aminoglycosides (gentamycin and ciprofloxacin), and genes responsible for aminoglycoside resistance (armA, APH(6)-Id, APH(3'')-Ib, and ANT(3'')-IIc) were found. The 16S rRNA methylase, which confers high-level resistance on all aminoglycosides encoded by the armA gene, was initially identified in Citrobacter freundii in Poland in 2002 and has now been identified worldwide among gram-negative bacteria [70][71][72][73]. Commonly, resistance to aminoglycoside is conferred by aminoglycoside-modifying enzymes (acetyltransferases, nucleotidyltransferases, and phosphotransferases) [74]. The VJR422 was found to possess variants of phosphotransferases, i.e. APH(6)-Id and APH(3'')-Ib, and ANT(3'')-IIc, a variant of nucleotidyltransferase. Four β-lactamase-encoding genes that can hydrolyze antimicrobials containing a β-lactam ring were predicted, i.e. TEM-1, ADC-73, OXA-23, and OXA- 66. The VJR422 strain was resistant to all β-lactam-ring-containing drugs (ceftazidime, imipenem, meropenem, cefoxitin, ceftriaxone, and aztreonam.) TEM-1 (Temoneira-1) β-lactamase is one of the best-known drug-resistant enzymes able to hydrolyze penicillin and the first generation of cephalosporin [75]. The ADC-73 (Acinetobacter-derived cephalosporinase-73) βlactamase is regarded as a chromosomally encoded class C β-lactamase that confers resistance to penicillin, cephalosporin, and β-lactam/β-lactamase inhibitor combinations [76]. Therefore, the VJR422 strain is also resistant to piperacillin/tazobactam. The class D carbapenem-hydrolyzing oxacillinases (OXA type), OXA-23 and OXA-66 were predicted in WGS of VJR422. OXA-23 is one of the largest groups of OXA-type carbapenemases in A. baumannii, and OXA-66 is a variant of OXA-51 classified as an OXA-51-like group of enzymes [77,78]. Regarding the Ade pump, resistance-nodulation-cell division (RND) transporter genes (adeABC, adeFGK, adeN, and adeRS) were identified in the VJR422 genome analysis. Investigation of the tigecycline resistance mechanism in the VJR422 isolate was challenging. Several point mutations in the regulatory gene adeRS were observed, resulting in overexpression of the AdeABC efflux pump system. AdeIJK and AdeFGH showed high expression compared with the susceptible strain (ATCC 19606) (Fig 5). The presence of numbers and polyspecificities of RND transporters correlate with high intrinsic and clinical resistance in gram-negative bacilli [79]. Moreover, Acinetobacter baumannii AbaF, AbaQ, and AmvA, the major facilitator superfamily (MFS) antibiotic efflux pump, were also identified in the VJR422 genome. MFS transporters are involved in drug efflux systems and lead to antibiotic resistance in both grampositive and gram-negative bacteria [80]. AbaF was identified as an efflux pump associated with fosfomycin resistance in A. baumannii [81,82]. AbaQ is mainly involved in the extrusion of quinolone-type drugs in A. baumannii [83]. AmvA contributes resistance to erythromycin, acriflavine, benzalkonium chloride, and methyl viologen [81,84]. Multidrug efflux pumps of the small multidrug resistance (SMR) type are made of a transport protein located in the inner membrane. AbeS, an efflux pump of the SMR type, was identified in the VJR422 genome. AbeS could decrease susceptibility in the various class of antibiotics, disinfectants, dyes, and detergents [85]. The mphE and msrE genes associated with erythromycin resistance and streptogramin resistance were identified as well [68,86,87].

Conclusions
In conclusion, this study identified and characterized the MDR A. baumannii strain VJR422 from a clinical specimen using the WGS analysis tool at the Faculty of Medicine Vajira Hospital, Navamindradhiraj University. Knowledge of this bacterial pathogen at the genomic level has not been reported previously at our hospital. We reported and updated the new ST1890 in the PubMlst Database and characterized the VJR422 in the genomic level data, i.e. GO, COG, and KEGG. The antibiotic resistance genotype and phenotype as well as the presence of potential virulence associated factors were investigated.
Supporting information S1 Table. GO classification into three main parts in the genome of VJR422.