A multiplex loop-mediated isothermal amplification assay for rapid screening of Acinetobacter baumannii and D carbapenemase OXA-23 gene

Background: Acinetobacter baumannii is a health burden responsible for various nosocomial infections, and bacteremia in particular. The resistance of A. baumannii to most antibiotics including carbapenem has increased. OXA-23-producing A. baumannii is the chief source of nosocomial outbreaks with carbapenem-resistant A. baumannii. Successful antibiotic treatment relies on the accurate and rapid identification of infectious agents and drug resistance. Here, we describe a multiplex loop-mediated isothermal amplification (LAMP) assay for simultaneous and homogeneous identification for A. baumannii infection screening and drug-resistance gene detection. Methods: Four primer pairs were designed to amplify fragments of the recA gene of A. baumannii and the oxa-23 gene. The reaction with a 25 μl of final volume was performed at 63°C for 60 min. For comparative purposes, we used a traditional method of bacterial identification to evaluate assay efficacy. Results: The multiplex LAMP assay enables simultaneous and homogeneous detection of the recA gene of A. baumannii and the oxa-23 gene and requires less than 21 min with no pre-requisite for DNA purification prior to the amplification reaction. The detection is specific to A. baumannii, and the coincidence rate of the multiplex LAMP and the traditional method was 100%. Conclusions: Our data indicate that the multiplex LAMP assay is a rapid, sensitive, simultaneous and homogeneous method for screening of A. baumannii and its drug-resistance gene.


Introduction
Acinetobacter baumannii, an opportunistic pathogen, is a health burden responsible for various nosocomial infections, and bacteremia in particular [1][2][3]. Due to the wide application of broad-spectrum antibiotics, resistance of A. baumannii to most antibiotics has increased, and the emergence of carbapenem-resistant (CR) A. baumannii has been described as a sentinel event of clinically relevant antimicrobial resistance [4]. Recent research shows that bacteremia caused by CR A. baumannii typically leads to increased mortality and prolonged hospital stays [5][6][7].
The propensity of A. baumannii to be multidrug-resistant or extensively drug-resistant presents therapeutic and infection control challenges. Successful antibiotic treatment of infections relies on accurate and rapid identification of the infectious agents and drug resistance. Various culture-based methods and susceptibility testing are time-consuming and delay confirmation of the strain and the start of appropriate antibiotic therapy. Some recently developed molecular techniques (loop-mediated isothermal amplification (LAMP)) are rapid and relatively simple to use and have been described for identifying DNA directly from clinical specimens [11,12].
We describe the development of a multiplex LAMP assay for simultaneous and homogeneous identification for A. baumannii infection screening and drug-resistance gene detection. To the best of our knowledge, this is the first study that outlines this approach. The present study was performed at the Chinese PLA General Hospital from 2016 to 2017, and data show that the multiplex LAMP assay is a rapid, sensitive and direct method for screening of A. baumannii and its drug-resistance gene and requires less than 21 min with no pre-requisite for DNA purification prior to the amplification reaction.

Materials and methods
Definitions CR A. baumannii is defined if an A. baumannii isolate was resistant to both imipenem and meropenem [13].

Bacterial isolates
All consecutive and nonduplicate clinical isolates from hospitalized patient samples (such as gastric or drainage fluid, urine, blood and sputum) were collected including Candida albicans isolates, Candida tropicalis isolates, Candida parapsilosis isolates, Klebsiella pneumonia isolates, Enterobacter cloacae isolates, Escherichia coli isolates, Streptococcus viridans isolates, Micrococcus isolates, Enterococcus isolates, coagulase negative Staphylococcus isolates, A. baumannii isolates, Stenotrophomonas maltophilia isolates and Pseudomonas aeruginosa isolates. Standard microbiologic methods were used to isolate and identify bacteria from clinical specimens. Bacterial antimicrobial susceptibility testing was performed in accordance with the methodology of the Clinical and Laboratory Standards Institute 2014 [14].

Primer design and LAMP method
The primer pairs (outer and inner) used for identification of A. baumannii and its antimicrobial susceptibility were designed to amplify fragments of the recA gene of A. baumannii and the oxa-23 gene using LAMP Designer software (http://www.optigene.co.uk) (OptiGene Ltd, United Kingdom) ( Table 1). Bacterial genomic DNA was collected by boiling. Single colonies were collected in 500 μl of sterile purified water and then serially diluted. Diluted bacterial solutions were boiled for 30 min and centrifuged at 10,000×g for 8 min. Supernatants (total genomic DNA) were aliquoted for storage at −80 • C for later study.    When the recA and oxa-23 genes of A. baumannii were simultaneously and homogeneously amplified in a single reaction tube, primer pairs for fragment amplification were premixed (ratio 1.3:1) and added to the reaction tube as described. Amplification of the positive DNA yielded a sigmoidal amplification curve, and the negative control tube had no measurable fluorescence or turbidity as indicated by a flat line in the plot. The desired products were confirmed by agarose gel electrophoresis. A negative control was performed using sterile water instead of DNA template.

Baumannii and carbapenem-resistance
To evaluate specificity of fragment amplification of the recA gene for confirming A. baumannii, a total of 22 bacterial strains were used including two C. albicans isolates, a single C. tropicalis isolate, a single C. parapsilosis isolate, two K. pneumonia isolates, a single E. cloacae isolate, a single E. coli isolate, two S. viridans isolates, a single Micrococcus isolate, a single Enterococcus isolate, a single coagulase negative Staphylococcus isolate, three A. baumannii isolates, a single P. mirabilis isolate, two S. maltophilia isolates and three P. aeruginosa isolates. Bacterial colonies were boiled for 30 min to obtain bacterial genomic DNA. Amplifications were carried out using a Loopamp DNA Amplification Kit on Loopamp Realtime Turbidimeter LA-320c. The specificity of the LAMP assay for A. baumannii was evaluated by comparing results identified by standard methods. The detection limit was ascertained using serial dilutions of bacteria. These experiments were replicated to ensure reproducibility.

Multiplex LAMP assay
Two A. baumannii isolates with positive oxa-23 gene amplification were used to perform the multiplex LAMP assay that simultaneously and homogeneously amplified recA and oxa-23 genes in a single reaction tube with four pairs of primers. The multiplex LAMP assay was carried out using the commercially available Loopamp DNA Amplification Kit (Isothermal Master Mixes and DNA Polymerases) on a Genie II or a ROCHE LightCycler 480.

Results
We developed a specific multiplex LAMP assay for measuring A. baumannii and carbapenem-resistance. This assay facilitated amplification of the recA gene of A. baumannii within 36 min of initiation using the Loopamp Realtime Turbidimeter LA-320c (Figure 1), 25 min using a ROCHE LightCycler 480 (Figure 2A) and 20 min with a Genie II ( Figure 3A). Using serial dilutions, ∼300 copies μl −1 of specific nucleic acid targets were assayed after 25 min using the ROCHE LightCycler 480 (Figure 4). Of the 22 identified bacterial isolates, three were positive for recA gene amplification and identified as A. baumannii using classical culture-based methods. No amplification curve was observed with the 19 non-Acinetobacter baumannii bacterial DNA samples. An analysis of specificity associated with the LAMP assay confirmed that only target A. baumannii DNA generated a positive result. DNA extracted from C. albicans, C. tropicalis, C. parapsilosis, K. pneumoniae, E. cloacae, E. coli, S. viridans, Micrococcus, Enterococcus, coagulase negative Staphylococcus, P. aeruginosa, mucoid P. aeruginosa, S. aureus, Cryptococcus and S. maltophilia were negative with the LAMP assay. This confirmed specificity for A. baumannii using the LAMP assay amplifying the recA gene, and that A. baumannii could be identified by amplifying the recA gene.
All three isolates with positive recA gene amplification were resistant to imipenem and meropenem according to standard antibiotic susceptibility tests and two were oxa-23 gene-positive ( Figure 2B). When the multiplex amplification was conducted on the Genie II, a positive curve appeared within 18 min ( Figure 3A). The multiplex amplification products were confirmed by agarose gel electrophoresis and two DNA bands, a 136-bp fragment for the oxa-23 gene and a 210-bp fragment for the recA gene appeared. Conversely, only one gene was amplified, yielding an amplification product ( Figure 2D).
We observed that the annealing temperature of recA gene products was ∼0.7 • C higher than the oxa-23 gene (Figure 3B), and a greater ratio of recA to oxa-23 primer pairs, the greater the annealing temperature of mixed products of both genes in the multiplex amplification ( Figures 5 and 6). The annealing temperature of mixed products was not specific (85.3-85.8 • C) depending on the ratio of primer pairs for both genes. Additionally, when the multiplex amplification was carried out on a ROCHE LightCycler 480, a positive curve presented a specific superposition of two curves ( Figure 2C). Data show that identification of A. baumannii and carbapenem resistance can be done simultaneously and homogeneously in a short time.

Discussion
The recA gene is an important housekeeping gene in most bacteria. Its function mainly involves homologous recombination, DNA repair, the SOS response, combining single-stranded DNA, DNA unraveling and seeking homologous loci for homologous recombination [15][16]. The recA gene has a high discriminatory power for differentiating species and subspecies within various bacterial genera, such as Vibrio and Lactobacillus species [17][18][19][20][21]. In 2012, recA sequences were used to develop a real-time PCR assay for discrimination of S. pseudopneumoniae from closely related species with excellent specificity [22]. A multiplex PCR targeting the 16S-23S rRNA intergenic spacer gene sequence and the recA gene sequence were also established to identify A. baumannii rapidly, which was effective for identifying reference strains and clinical isolates of A. baumannii in environmental specimens [23]. Thus, the recA gene is a confirmatory alternative tool for molecular identification and epidemiological studies.
Carbapenemases produced by A baumannii are mostly class D oxacillinases, although metalloenzymes of the IMP, VIM and NDM families have been detected in some well-defined geographical regions [8, [24][25]. Carbapenemase activity is also an intrinsic property of many class D β-lactamases, which are referred to as CHDLs. The first and most common subgroup of CHDLs is made of OXA-23, OXA-27 and OXA-49 [26], and the blaOXA-23 gene is considered to be a significant cause of carbapenem resistance in A. baumannii worldwide [27][28].
Real-time PCR is sensitive and accurate for identifying carbapenemase genes [29][30]. LAMP could identify bacteria directly from clinical specimens (plasma) within 20 min without DNA purification [31]. Compared with PCR, the biggest advantage of LAMP is that nucleic acid amplification is carried out isothermally at a lower temperature and in clinical samples such as serum, swabs and whole blood with little or no sample preparation [31][32]. Some multiplex LAMP methods that involve homogeneous and heterogeneous amplification have been reported [33][34][35][36].
Here, we describe a multiplex LAMP assay to directly amplify recA and OXA-23 genes simultaneously and homogeneously in real-time at 63 • C, which enables detection within 18 min without DNA purification before the multiplex amplification reaction. Figure 4 shows that the detection limit of the LAMP assay for A. baumannii was 10 2 CFU μl −1 for the strain without loop primers for the LAMP reaction. All A. baumannii strains were amplified, and 19 non-baumannii Acinetobacter strains produced negative reactions. We also used the traditional method of bacterial identification to evaluate the newer assay, and the coincidence of both was 100%. These data indicated that the assay provides a sensitive, specific and simple diagnostic tool for A. baumannii infection. Tone and co-workers investigated the influences of T m values of LAMP products and found that inorganic pyrophosphatase (PPase) reduced the T m value variances and allowed differentiation of pathogenic agents using an annealing curve analysis for the LAMP method [37]. However, annealing curve analysis is difficult to apply to the multiplex LAMP assay because products of different target genes were similar in structure and length, all of which were dumbbell-like structures. Therefore, the T m (annealing temperature) values for each LAMP product do not differ significantly. We report that one peak appeared in the annealing curve of the mixed products of two different target genes, which indicates a common T m value for the mixed products generated from different target genes. However, the common T m value ranged over a narrow interval depending on the proportion of various products is not stable and specific ( Figures 3B, 5B and 6B). Our assay is robust enough to screen suspected A. baumannii infection and simultaneously detect the OXA-23 carbapenemase gene, which reduces the time to the diagnosis of infection and improves medical treatment. This assay could also be performed on various real-time nucleic acid amplification apparatuses with target detection using fluorescence (Genie II is an isothermal amplification instrument, and ROCHE LightCycler 480 is a universal PCR device), suggesting a significant advantage for on-site application and efficient screening and testing of clinical samples. More work is required to confirm the sensitivity and specificity of this assay and how to indicate multiplex detection being introduced into the LAMP method.