Development of a One-Step Multiplex PCR Assay for Differential Detection of Four species (Enterobacter cloacae, Enterobacter hormaechei, Enterobacter roggenkampii, and Enterobacter kobei) Belonging to Enterobacter cloacae Complex With Clinical Significance

Enterobacter cloacae complex (ECC) is composed of multiple species and the taxonomic status is consecutively updated. In last decades ECC is frequently associated with multidrug resistance and become an important nosocomial pathogen. Currently, rapid and accurate identification of ECC to the species level remains a technical challenge, thus impedes our understanding of the population at the species level. Here, we aimed to develop a simple, reliable, and economical method to distinguish four epidemiologically prevalent species of ECC with clinical significance, i.e., E. cloacae, E. hormaechei, E. roggenkampii, and E. kobei. A total of 977 ECC genomes were retrieved from the GenBank, and unique gene for each species was obtained by core-genome comparisons. Four pairs of species-specific primers were designed based on the unique genes. A total of 231 ECC clinical strains were typed both by hsp60 typing and by species-specific PCRs. The specificity and sensitivity of the four species-specific PCRs ranged between 96.56% and 100% and between 76.47% and 100%, respectively. The PCR for E. cloacae showed the highest specificity and sensitivity. A one-step multiplex PCR was subsequently established by combining the species-specific primers. Additional 53 hsp60-typed ECC and 20 non-ECC isolates belonging to six species obtained from samples of patients, sewage water and feces of feeding animals were tested by the multiplex PCR. The identification results of both techniques were concordant. The multiplex PCR established in this study provides an accurate, expeditious, and cost-effective way for routine diagnosis and molecular surveillance of ECC strains at species level.


INTRODUCTION
Enterobacter spp. are natural commensals of the animal and human gut microbiota (Sanders and Sanders, 1997;Mesa et al., 2006). Over the past decades, some members of Enterobacter spp. have become important nosocomial pathogens thus are classified as one of the "ESKAPE" bugs (Enterococcus faecium, Staphylococcus aureus, Klebsiella pneumoniae, Acinetobacter baumannii, Pseudomonas aeruginosa, and Enterobacter species) (Rice, 2008). Among Enterobacter spp., the E. cloacae complex (ECC) is of major importance, accounting for 65%-75% of infections in clinical settings (Sanders and Sanders, 1997). The ECC is frequently associated with a multidrug resistance (MDR) phenotype, mainly due to the ability to easily acquire numerous genetic mobile elements carrying resistance genes. Moreover, the chromosome of ECC intrinsically encodes ampC genes, resulting in a constitutive production of the AmpC b-lactamase and is thus intrinsically resistant to ampicillin, amoxicillin, amoxicillin/ clavulanate, first-generation cephalosporins, and cefoxitin (Mezzatesta et al., 2012). Available epidemiological data show that the ECC has become the third major drug-resistant Enterobacteriaceae species involved in nosocomial infections after Escherichia coli and K. pneumoniae (Zhou et al., 2018). Hence, it is important to control the dissemination of MDR-ECC for public health.
According to physiology and biochemistry phenotypes, as well as genotypes, more than 20 species and clades (without taxonomy terms) are classified into ECC till now (Sutton et al., 2018), and the taxonomic status is consecutively updated. Of note, increasing studies show that the clinical importance and drug resistance of different ECC members is significantly different. A number of epidemiological studies reveal that E. hormaechei is the most prevalent species in the clinical setting (Morand et al., 2009;Chavda et al., 2016;Moradigaravand et al., 2016), and this species is commonly considered a causative pathogen responsible for various infections (Davin-Regli et al., 1997;Paauw et al., 2009;Mezzatesta et al., 2012). Another important member of ECC is E. kobei. Clinical strains of E. kobei have been isolated from various clinical samples, including blood, sputum, urine, bronchoalveolar lavage, abscesses, and especially intra-abdominal samples (Zhou et al., 2017;Davin-Regli et al., 2019). A recent study showed that E. hormaechei and E. kobei were the predominant ESBL-positive species accounting for more than 70% of community-acquired ECC strains collected across China (Zhou et al., 2018). E. cloacae is very often isolated in samples of clinical origin and is found frequently during systematic sampling of neonates who have been colonized early. The species is involved in 10% of postsurgical peritonitis cases, 5% of nosocomial sepsis and pneumonia cases, and 4% of urinary tract infections (Davin-Regli et al., 2019). A large genomic study by collecting 316 MDR-ECC isolates from bloodstream infections between 2001 and 2011 across hospitals in the UK and Ireland shows that E. hormaechei, E. kobei, and E. cloacae account for more than 80% of the studied population, suggesting the epidemiological and clinical significance for the three species (Moradigaravand et al., 2016). E. roggenkampii is a recently described species based on a computational analysis of sequenced Enterobacter genomes, and can cause multi-site infection (e.g. blood, urine, feces, and body fluids). The species has widely distributed worldwide, including Australia, the United States, Germany, and China, etc. (Xu et al., 2021). Collectively, the data highly support the necessity of studying ECC on species level.
Currently, the accurate identification of ECC species and subspecies remains a challenge. In clinical laboratories, the routine identification of ECC is mainly dependent on phenotypic characteristics by using commercialized systems, like Vitek-II (bioMeŕieux) and the Matrix-assisted laser desorption/ionization-time-of-flight (MALDI-TOF) mass spectrometry (MS) technology. Although these systems can identify the ECC, they fail to differentiate the species within the group (Davin-Regli et al., 2019). Molecular methods are more suitable for precisely identification of the ECC on species level, and currently hsp60 typing is the most widely used method for this purpose. By sequencing and phylogenetically analyzing the hsp60 gene, Hoffman and Roggenkamp grouped the ECC into 12 genetic clusters (I-XII) and one unstable sequence cluster (XIII) (Hoffmann and Roggenkamp, 2003). Additionally, the technique of microarray comparative genomic hybridization (microarray-CGH) (Paauw et al., 2008), multi-locus sequence analysis (MLSA) (Singh et al., 2018), and whole-genome sequencing (Chavda et al., 2016) show the power on the species-level identification of ECC. However, these methods are time-consuming, expensive, and therefore difficult to implement routinely. Recently, a qPCR-based method targeting the hsp60 gene was developed and showed high accuracy to detect E. hormaechei (Ohad et al., 2014). The limited detection spectrum cannot fulfill the needs of clinical diagnosis yet.
Given that current methods have significant disadvantages, there is a need to develop an efficient and accurate method to identify the ECC population on the species level for routine diagnosis and research. The primary goal of our study is to setup and optimize a one-step multiplex PCR assay capable of differentiating the major ECC species (i.e. E. cloacae, E. hormaechei, E. roggenkampii, and E. kobei) with clinical significance. Our optimized multiplex PCR assay is able to provide a reliable and rapid detection method for laboratory diagnosis to employ in the clinical setting.

Bacterial Collection
A total of 284 ECC and 20 non-ECC strains were used in this study, including 294 clinical isolates, eight environmental isolates recovered from hospital sewage water and two livestockassociated isolates recovered from feces of farmed animals. Species identification was performed by using VITEK MS (bioMerieux, Marcy-l'Étoile, French). The bacteria were cultured in Luria-Bertani (LB) liquid (Tryptone 10 g, NaCl 10 g, Yeast extract 5 g, H 2 O 1L, pH 7.0) or agar medium. The bacterial genomic DNA were extracted using the conventional boiling lysis method as previously described (Chen et al., 2011).  (Pritchard et al., 2016) to accurately identify the species of each genome, with a cut-off of 95%. The genomes of E. hormaechei, E. kobei, E. roggenkampii, and E. cloacae identified were re-annotated using Prokka (Seemann, 2014). The annotated genomic sequences were used as the input to Roary (Page et al., 2015) to calculate the core and accessory genes of each species. The identity threshold was set at 90% without split paralogs. The unique genes of each species were defined as the core genes present in more than 95% genomes of the species but less than 5% genomes of the other species. The pair-wise blast was further performed for the unique genes to exclude sequences with high homology (30% of amino acid sequence identity and 10% of sequence coverage). The resulting unique genes were selected as the candidates of species markers. PCR primers ( Table 1) were designed using Primer Premier 6.0 (GraphPad Software, San Diego, CA, USA) based on the sequence of each candidate of species markers.

Establishment of Single and Multiplex PCR Systems
Species-specific primers were used to establish the single and multiplex PCR systems. Type strains (E. cloacae ssp. cloacae ATCC 13047, E. hormaechei ssp. hormaechei ATCC 49162, E. roggenkampii DSM 16690 and E. kobei DSM 13645) were used as positive controls. The mixture of a single PCR reaction (25 ml) contained 2×Taq Plus Master Mix II (Vazyme, Nanjing, China) (12.5 ml), 10 mM of each primer (1 ml), the DNA template (0.5 ml), and nuclease-free ultra-pure water (10 ml). The multiplex PCR mixture (50 ml) contained 25 ml of Vazyme 2×Taq Plus Master Mix II, 2 ml of primer mix containing 0.25 ml of each primer in 10 mM, 2 ml of DNA template, and 21 ml of nuclease-free ultra-pure water in a final volume of 50 ml. The reaction condition of single and multiplex PCR is identical, comprising an initial denaturation step at 95°C for 3 min, followed by 30 cycles of amplification (95°C for 15 s, 55°C for 20 s, and at 72°C for 90 s), and a final elongation step at 72°C for 5 min. PCR amplification was performed by using a Bio-Rad T100 ™ Thermal cycler (Bio-Rad, Hercules, USA). The PCR products were separated by using electrophoresis with 2.0% agarose gel for 40 min at 110 V in 1×Tris-Acetate-EDTA (TAE) buffer, and gels were visualized by using the gel imaging system instrument (Bio-Rad, Gel Doc ™ XR + with Image Lab ™ Software). The total time cost of the PCR reactions and electrophoresis is approximately 150 min. kobei DSM 13645) were used as positive controls and E. coli ATCC25922 was used as the negative control.

Species Identification
Species identification for the ECC members were performed by using hsp60 typing as previously described (Hoffmann and Roggenkamp, 2003). Briefly, the fragments of hsp60 gene of 284 ECC strains were amplified with using primers Hsp60-F (5′-GGTAGAAGAAGGCGTGGTTGC-3′) and Hsp60-R (5′-ATGCATTCGGTGGTGATCATCAG-3′) (Hoffmann and Roggenkamp, 2003) followed by Sanger sequencing. The 20 non-ECC isolated were identified by Vitek-II (bioMeŕieux) mass spectrometer and 16S rRNA gene sequencing. The results of ECC sequences were aligned and the phylogenetic tree was calculated by using MEGA X (Kumar et al., 2018 The untypeable strains by hsp60 typing method were further sent to WGS to confirm the species. Genomic DNA was extracted using a Gentra Puregene Yeat/Bact. Kit (Qiagen, San Francisco/Bay area, CA, USA), and subjected to WGS by using Illumina novaseq 6000 system (Illumina, San Diego, United States), using 2×150-bp pairend libraries. Illumina reads were trimmed using Trimmomatic (Bolger et al., 2014) (REF or Website) and assembled with SPAdes v3.12.0 (Bankevich et al., 2012). ParSNP v1.2 (Treangen et al., 2014) was used to align the core genome to calculate the phylogenetic tree with settings: min LCB size 25, maximal diagonal difference 12% and enabling filtering of SNPs located in PhiPack identified regions of recombination.

Evaluation of the Multiplex PCR Applicability for DNA Amount
Type strains of four species were used to estimate the detection limit of the multiplex PCR established here. The analytical sensitivity of the multiplex PCR using DNA directly extracted from tested strains which serially diluted ranged from 1000 to 1 pg. After amplification, 5 ml of each PCR product was loaded onto 2.0% agarose gels. The detection limit was determined to be the smallest DNA quantity that generated the correct band pattern for the targeted gene.

Statistical Analyses
Specificity and sensitivity were used to quantify the identification accuracy of species-specific primers designed in this study. The specificity is the proportion of true negatives that are correctly identified by the primer tests, and sensitivity is the proportion of true positives that are correctly identified by the primer tests. The results of species-specific PCRs were compared with that of hsp60 typing. Specificity and sensitivity were measured with a 95% confidence interval (95% CI) based on true positives (TP), true negatives (TN), false positives (FP), and false negatives (FN) (Lalkhen and McCluskey, 2008). The specificity as TN/(TN + FP), while sensitivity was calculated as TP/(TP + FN). The sensitivity and specificity values were calculated using MedCalc Software 1 .

RESULTS
Design of Species-Specific Primers for E. cloacae, E. hormaechei, E. roggenkampii, and E. kobei A collection of 977 ECC genomes were retrieved from the NCBI RefSeq database. According to the ANI results, 808 genomes were assigned to E. hormaechei (n=714), E. cloacae (n=50), E. kobei (n=26), and E. roggenkampii (n=18). Pan-genome construction of each species was conducted to query the species-specific core genes. To avoid high sequence homology between all of the candidate unique genes which might be a consequence of the high identity threshold chosen for pan-genome construction, and to obtain species-specific primers with high possibility, the pair-wise blast was further performed. Four species-specific genes were obtained for the four species, respectively, including a gene (ECL_01098) encoding an aminoacylase (Curley and van Sinderen, 2000;Dettori et al., 2018) in the genome of E. cloacae ssp. cloacae strain ATCC 13047 (accession no. CP001918.1); the pqqB (LI64_09975) gene encoding a small redox active molecule (Shen et al., 2012) in the genome of E. hormaechei ssp. hormaechei strain 34983 (accession no. CP010377.1); a gene (EGY04_15785) encoding an esterase family protein (Bornscheuer, 2002) in the genome of E. roggenkampii strain FDAARGOS_523 (accession no. CP033800.1); and the secretion ATPase gene gspE (BFV64_21570) (Lu et al., 2013) in the genome of E. kobei strain DSM 13645 (accession no. CP017181.1).
The full length of each species-unique gene was extracted from genomes of each species and aligned to obtain the conserved fragments for the design of species-specific primers. This resulted in four pairs of primers as listed in Table 1. In silico PCR showed that the primers generated products with different sizes, i.e. 397 bp for E. cloacae, 487 bp for E. hormaechei, 1132 bp for E. roggenkampii, and 1403 bp for E. kobei (Table 1 and Figure 1).

Evaluation of the Specificity and Sensitivity for Designed Species-Specific Primers
A total of 231 ECC strains were initially typed by using hsp60 typing. All of the strains were tested by using the primers EC-F/ R, EH-F/R, ER-F/R, and EK-F/R. PCR results showed that 12 strains were positive and 219 were negative to the primers EC-F/ R, which were fully consistent with that of hsp60 typing (specificity: 100.00%; sensitivity: 100.00%). PCR results of the  typing results showed that seven were false positives and two were false negatives (specificity, 96.59%; sensitivity, 92.31%). The results were detailed in Table 2 and Table S1.

Development and Validation of Multiplex PCR Assay
Following the validated single species-specific PCRs, a multiplex PCR system using the four pairs of primers was set up. The product of four type strains using the multiplex PCR system was separated well by the electrophoresis with 2.0% agarose gel, resulting in four bands ( Figure 2). We further evaluated the multiplex PCR system by testing 53 ECC strains which were isolated from clinical samples, sewage water, and animal feces and 20 non-ECC strains (four Acinetobacter baumannii, five K. pneumoniae, two Serratia marcescens, four Escherichia coli, two Pseudomonas aeruginosa, three Flavobacterium meningosepticum). The established multiplex PCR identified 1 isolate as E. cloacae, 17 as E. hormaechei, 16 as E. roggenkampii, 9 as E. kobei, and the remaining 10 ECC strains and 20 non-ECC strains were negative to the PCR reactions ( Table 3).
The hsp60 sequences of six ECC strains (1339, 1340, 1342, 1383, 6069, and 12740) clustered together without any of type strains, thus failed being typed. We further performed WGS for two of them (1339 and 1383) to identify the species, and the phylogenetic analysis showed that they belonged to E. asburiae (data not shown). Collectively, the results of the multiplex PCR are fully consistent with those of hsp60 typing and WGS ( Table 3). The multiplex PCR assay showed 100% sensitivity (95% CI 91.78-100%) and 100% specificity (95% CI 88.43-100%) for the identification of E. cloacae, E. hormaechei, E. roggenkampii, and E. kobei ( Table 2).

Evaluation of the Detection Limit for Multiplex PCR
We further examined the detection limit for the multiplex PCR employed in the clinical setting by using serially diluted DNA of isolates of the targeted ECC. The detectable DNA amount was 1 pg for E. cloacae ssp. cloacae ATCC 13047 ( Figure 3A), 50 pg for E. hormaechei ssp. hormaechei ATCC 49162 ( Figure 3B), 5 pg for E. roggenkampii DSM 16690 ( Figure 3C) and 50 pg for E. kobei DSM 13645 ( Figure 3D). The data prove that the multiplex PCR established in this study is with very high analytical sensitivity.

DISCUSSION
According to the ANI results of 977 ECC genomes retrieved from the NCBI RefSeq database, the four species (E. cloacae, E. hormaechei, E. roggenkampii, and E. kobei) accounted for 82.7% of ECC population studied in this work, and most genomes were clinically associated, suggesting that the four species were predominant in the clinical setting. This is consistent with the previous studies (Zhou et al., 2017;Davin-Regli et al., 2019;Wang et al., 2020). It is meaningful to develop a method which can identify ECC organisms efficiently and differentially, and has the potential to aid the control of four major ECC infections greatly. In the present study, we tried to develop a simple and effective multiplex PCR assay that enables the differential identification of the four ECC species. The species-specific core genes of the four ECC species detected by comparative genomics provide meaningful markers for their identification. All primers were newly designed and optimized for development of the multiplex PCR assay in this study.
In clinical laboratories, various methods for identifying ECC species have been introduced over the past decades, and molecular techniques and MALDI-TOF MS are frequently used to identify ECC species (Hoffmann and Roggenkamp, 2003;Paauw et al., 2008;Ohad et al., 2014;Chavda et al., 2016;Singh et al., 2018;Davin-Regli et al., 2019). These techniques reduce the error rates for the results, thereby enabling the acceleration of treatment decisions according to the accurate identifications. MALDI-TOF MS, microarray-CGH, MLSA, WGS, and qPCR technologies have been surpassed by hsp60 typing as the methods of choice for rapidly discriminating the species of this genus. The one-step multiplex PCR platform is a kind of common PCR, and it does not require additional steps. Researchers had developed one-step multiplex PCR methods to solve the clinical problems for the accurate identification of mycobacteria at the species and/or subspecies level, and it has demonstrated 100% specificity for the targeted Mycobacterium species (Chae et al., 2017). Similar as what we did here, a multiplex PCR method has been developed to identify three species (K. pneumoniae, K. variicola, and K. quasipneumoniae) of Klebsiella genus, and has been verified that it is feasible to carry out Klebsiella identification in the clinical routine work (Fonseca et al., 2017). Although several multiplex PCR systems of some species have already been commercialized and introduced into  laboratories to do strain classifications (Brito et al., 2003;Chae et al., 2017;Fonseca et al., 2017), to our knowledge, there were no multiplex PCR assays capable of simultaneously discriminating various ECC species with clinically importance. 1 We initially performed single PCRs to evaluate the specificity and sensitivity for each pair of species-specific primers designed in this study. A unique band with expected size presented in each single PCR detection by using primers EC-F/R for E. cloacae ATCC 13047 ( Figure 1A), EH-F/R for E. hormaechei ATCC 49162 ( Figure 1B), ER-F/R for E. roggenkampii DSM 16690 ( Figure 1C), and EK-F/R for E. kobei DSM 13645 ( Figure 1D). A large set of clinical strains were further tested for the primers to estimate their specificity and sensitivity. Our tests showed that the specificity and sensitivity of our primers were reliable with a 95% confidence interval. The sensitivity of ER-F/R and EK-F/R is lower than that of EC-F/R and EH-F/R. We suppose that the reason could be a limited number of genomes available in the public database for the two species (26 genomes of E. kobei, and 17 of E. roggenkampii), resulting in a relatively low sensitivity. As the accumulation of genomes in the public database, the primers ER-F/R and EK-F/R could be updated in the near future. Of note, the multiplex PCR designed in our study currently is culture dependent. To accelerate microbial identification, direct detection from clinical specimens is a promising strategy. Therefore, it is worth evaluating the capacity of our multiplex PCR for culture-independent detections in the future.
In summary, we here developed and validated a novel multiplex PCR amplification method that can simultaneously identify four major members of ECC (i.e. E. cloacae, E. hormaechei, E. roggenkampii, and E. kobei) with clinical significance. In comparison to available molecular methods, our method is rapid, accurate, and cost-efficient which is suitable for being used in routine diagnostic tests. Additionally, the application of the established multiplex PCR method would highly facilitate us to understand the ECC in the aspect of clinical significance, epidemiology, and drug resistance at the species level.

DATA AVAILABILITY STATEMENT
The data sets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ncbi. nlm.nih.gov/, JAEKKA000000000 https://www.ncbi.nlm.nih. gov/, JAEKKB000000000.

AUTHOR CONTRIBUTIONS
YJ: strain collection, run all tests, and writing-original draft. PW: strain genome data collection and comparison, primer design, and genome data analysis. TX: experimental design, writing-review and editing, and statistical analysis. YZ: project administration. RC: statistical analysis. HZ and KZ: conceptualization, writing-review and editing, project A B D C FIGURE 3 | Analytical sensitivity of the multiplex PCR. Analytical sensitivity was evaluated for the type strains of four species using 1 to 1,000pg DNA per reaction mixture. The analytical sensitivities were 1 pg for E. cloacae ssp. cloacae ATCC 13047 (A); 50 pg for E. hormaechei ssp. hormaechei ATCC 49162 (B); 5 pg for E. roggenkampii DSM 16690 (C); 50 pg for E. kobei DSM 13645 (D). administration, resources, and supervision. All authors contributed to the article and approved the submitted version.