RAPD and ERIC-PCR coupled with HRM for species identication of non-dysenteriae Shigella species isolated from clinical specimens; As a potential of alternative method

Background: Species identication of Shigella isolates are so prominent for epidemiological studies and implementing infection prevention strategies. We developed and evaluated RAPD and ERIC-PCR assays coupled with HRM for differentiation of non-dysenteriae Shigella species as potential alternative methods. isolation of eighteen Shigella strains from faecal specimens collected from children under 2 years old with diarrhea symptom (n=143), species of the isolates were identied by slide agglutination assay as the golden standard method. Species of isolates were identied using developed RAPD-PCR-HRM and ERIC-PCR-HRM assays and differentiation of data sets obtained from difference plots of HRM was implemented by PCA as a dimension reduction method then sensitivity and specicity of the methods were evaluated using ROC curve. Results: Consequently, we found that RAPD-PCR-HRM with sensitivity and specicity of 100 and 85% respectively is signicantly more suitable for identication of non-dysenteriae Shigella species. However, sensitivity and specicity of ERIC-PCR-HRM were evaluated 33 and 46% respectively. Conclusion: Regardless of inherent poor reproducibility of DNA ngerprinting based methods, RAPD-PCR-HRM assay can be considered as a potential of alternative method for species identication. Also, HRM technique used in this study is more rapid, inexpensive, and sensitive than gel electrophoresis for characterization of PCR amplicons.


Background
Shigella is a facultative anaerobe, gram-negative bacilli belonging to the Enterobacteriaceae family with intestinal and extraintestinal pathogenicity in human (Shigellosis). Four species (with different serotypes) have been characterized for this genus including: S. sonnei, S. exneri, S. boydii, and S. dysenteriae. Shigella species cause mild to severe diarrhea in adults and children (up to 28% fatality rate in children with severe symptoms) as intestinal pathogens; however, S. dysenteriae type 1 lead to hemorrhagic uremic syndrome (HUS) because of releasing Shiga-toxin; it is regarded as the only extraintestinal disease caused by this genus [1]. There are signi cant differences between the severity and type of the gastrointestinal disorders triggered by Shigella species. Also, the geographical distributions of prevalence rate of Shigella species are considerably varied among developing and industrial countries [2]. Epidemiological studies revealed that identi cation of Shigella species plays a crucial role for providing acceptable clinical investigation and tracing of the outbreaks generated by this pathogen [3]. Therefore; to support the epidemiological investigation of shigellosis outbreaks especially among high-risk population as well as infant cases, different serological, molecular and immunological assays have been developed for identi cation of Shigella species. The most prominent challenge of Shigella species identi cation is differentiating all species from each other [4]. However; differentiation of Sh. dysenteriae is more practicable because of harboring stx gene as the speci c marker [5]. Some of these identi cation methods consist of DNA hybridization, multiplex PCR, whole genome sequencing, immunocapture PCR, MALDI-TOF mass spectrometry, rRNA gene restriction pattern, and serological methods as the gold standard. However, many of these methods, despite the optimization, are too expensive and complicated to be employed in clinical labs. Thus, new methods as well as more affordable and user-friendly assays with acceptable speci city and sensitivity are still needed to be developed and optimized [6].
DNA ngerprinting techniques are carried out for characterization and typing of microbial genomes. These methods are divided into two main categories: sequencing-based assays such as multi-locus sequence typing (MLST), repetitive element sequence-based (Rep-PCR), microarray, comparative hybridization, and whole genome sequencing; and non-sequencing-based techniques e.g. random ampli cation polymorphism DNA (RAPD-PCR), enterobacterial repetitive intergenic consensus (ERIC-PCR), pulsed eld gel electrophoresis (PFGE), and ampli ed fragment length polymorphism (AFLP) assays [7]. Obviously, sequence-based typing techniques are more expensive, time consuming, and complicated than other group of assays [8]. DNA PCR ngerprinting methods e.g. RAPD and ERIC have been widely used as genotyping assay for enterobacterial foodborne pathogens. However, these techniques have been developed for species identi cation of some bacterial and fungal strains regardless of species-speci c sequences [9]. Because of the complicated PCR products of RAPD and ERIC-PCR, statistical and mathematical analysis of the results are suggested to be carried out for further genomic characterization and species identi cation. Also; requiring characterization of the PCR products of RAPD and ERIC-PCR, usually implemented by electrophoresis on agarose gel, is considered as another challenge of these methods for species identi cation [10].
High resolution melting curve analysis (HRMA) provides characterization of PCR products as a rapid alternative assay of gel electrophoresis with higher precision and differentiation potential of PCR products with the same size [11]. This property of HRMA makes it considerable to be employed for analysis of the complicated PCR products as well as in genotyping and DNA ngerprinting protocols. PCR molecular typing methods such as RAPD coupled with HRMA has been developed successfully to characterize microbial strains in intra-serovar levels and for species identi cation [12]. However, use of PCR based DNA ngerprinting methods coupled with HRMA despite many advantages for serotype and species identi cation of microbial strains is limited to a few researches [13]. In the present study, we developed RAPD and ERIC-PCR ngerprinting assay coupled with HRMA following analysis by principal component analysis (PCA) for identi cation of Sh. sonnei, Sh. exneri and Sh. boydii strains after isolation from clinical specimens.

Results
So far, the most reliable method for identi cation of Shigella species is the gold standard assay by slide agglutination method using species-speci c polyvalent antisera [14]. Despite higher sensitivity and speci city, this method is very expensive so many researchers have made efforts to develop new alternatives. Because of high-cost, complicated, and sensitive production process of monoclonal antibodies in serology-based methods, molecular assays as well as DNA based techniques are preferred to be developed for species identi cation of bacterial pathogens [15]. However, identi cation of S. dysenteriae because of harboring shig-toxin encoded gene as an individual marker of this pathogen is not regarded as a real challenge in comparison with identi cation of other species [16]. At the present study, we isolated and identi ed Shigella species in eighteen specimens (totally from 143 children under 2 years old with gastrointestinal disorder; 12.58%) consisting of 11 S. sonnei, 3 S. boydii, and 4 S. exneri isolates by golden standard method using culture-based technique for Shigella isolation and serologic assay for species identi cation (Table 1). After isolation by culture-based methods, we evaluated and developed RAPD and ERIC-PCR coupled with HRM methods as sensitive and speci c alternative assays for Shigella species identi cation. Normalized and difference melting curves of isolates by RAPD-PCR-HRM are shown in Figs. 1 and 2, respectively. Because of extremely complicated PCR products ampli ed in RAPD, very complex melting curves were obtained after HRM step which made it impossible to categorize melting pro les by the instrument software algorithm [17]. Therefore; at the present study, we have employed a dimension reduction method, PCA as recommended by many researchers, to simplify analysis of this complex and huge data. PCA assay was performed on the data obtained from the output of HRM difference plot. PC 3D loading plot for difference melting curve data set of RAPD-PCR-HRM in species identi cation of Shigella isolates has been illustrated in Fig. 3. PCA plot showed that three distinctive groups as we considered three different species for Shigella isolates. Except one isolate of S. boydii, all species of isolated Shigella strains were identi ed correctly corresponded with the golden standard method for species identi cation by developed RAPD-PCR-HRM assay. Table 1.
Shigella isolates identified in this study by golden standard methods ERIC-PCR-HRM was evaluated as another assay carried out in this study for species identi cation of the same Shigella strains. Figures 4 and 5 showed normalized and difference plots of ERIC-PCR-HRM assay respectively for Shigella isolates. As the same as the uorescence plots obtained from the RAPD-PCR-HRM; because of extremely varied amplicon size and sequence, high complexity of melting pro les also does not allow normal analysis to be implemented for HRM data set in this procedure [17]. Difference plot output data of ERIC-PCR-HRM was analyzed for principal components determination with PCA. PC 3D loading plot of ERIC-PCR-HRM data set has been shown in Fig. 6. Seven species among eighteen isolates were recognized incorrectly showed that RAPD assay is more sensitive and speci c than ERIC-PCR to identify Shigella species while they are coupled with HRM and analyzed by PCA. Sensitivity and speci city of an assay as the accuracy characteristics of a screening or identi cation test relative to a golden standard method are commonly evaluated by ROC curve [18]. ROC curves of the assays developed in this study for Shigella species identi cation including RAPD and ERIC-PCR-HRM assays are illustrated in Figs. 7 and 8, respectively. Sensitivity and speci city of the RAPD-PCR-HRM assay were observed 100 and 85%, respectively. However, sensitivity and speci city of the ERIC-PCR-HRM technique were calculated 33 and 46% respectively regarding signi cant differences in comparison with the other assay. Nevertheless, these remarkable differences in methodological characteristics were predictable considering the results of PCA for each individual assay. of RAPD for characterization of Leptospira strains to serovar level. They were successful to develop RAPD-HRM assay as an inexpensive and rapid method to identify different serotypes of Leptospira. However, melting pro le analysis by Rotor-Gene 6000 series software package was used for clustering of the isolates and differentiating between the serotypes [19]. Miller et al. (2015) developed HRM-based assay for identi cation of Pasteurellaceae species by ampli cation of varied regions in 16 s rRNA sequence and they found this method sensitive and rapid for species identi cation [20]. Chen et al. (2019) also employed RAPD-HRM assay for differentiation of pathogenic and non-pathogenic Escherichia coli isolates from each other and they found it a sensitive, speci c, and practical assay. There are few studies about species identi cation of bacterial strains using DNA ngerprinting methods coupled with HRM; however, many researchers investigated these methods for differentiating nonmicrobial species such as raw meat, animals, and plant species especially by RAPD-PCR-HRM assay. Also, RAPD-HRM assay has been used for genotyping of bacterial pathogens previously [17].

Discussion
High resolution melting curve analysis of ERIC-PCR has not been developed and employed for genotyping and other molecular characterization of microbial genomes yet and the present study is the rst one evaluated this method coupled with HRM as a molecular technique for diagnostic purpose [21]. During HRM procedure, saturating dyes bound to DNA then emit uorescence because of dissociation of doublestrand DNA while the temperature is rising gradually then the emission is measured by a highly sensitive uorescence detector. Considering mechanism of this molecular assay, it can be found that any difference as well as a single nucleotide polymorphism in amplicon sequences signi cantly changes the melting pro le [22]. In DNA ngerprinting methods such as RAPD and ERIC-PCR assays, broad range of amplicons are produced throughout the PCR step. These complex PCR products are individually considered as a DNA marker for each organism and can be characterized by different methods [23]. Gel electrophoresis is commonly carried out for characterization of PCR products of DNA ngerprinting assay. HRM as a more rapid, inexpensive, sensitive and user-friendly method is used and more preferred for characterization of PCR products [24]. By HRM, RAPD and ERIC-PCR products are barcoded for better characterization and clustering of isolates in serotype and species identi cation. Because of the complex properties of these methods in PCR amplicon production, it is unclear that why one method is more practical, sensitive, and speci c than the other one in species identi cation in comparison with the golden standard method [25]. At the present study, we found that RAPD-PCR-HRM assay is more appropriate and acceptable than ERIC-PCR-HRM method for Shigella species identi cation. However, this result may not be obtained for species or serotype identi cation of other microorganisms in the future studies. ERIC are repetitive sequences located on some speci c genomic regions of Enterobacteriaceae family bacteria which are ampli ed during the PCR reaction [26]. In RAPD-PCR, randomly ampli cation is performed throughout the microbial genome on a wide range of different DNA regions which can be located on any probable position. Although both methods are based on random ampli cation, higher variation of ampli ed genomic regions in RAPD-PCR method may result in more e cient differentiation capability [27]. Limited genomic regions for DNA ngerprinting and barcoding classi cation makes ERIC-PCR ine cient to discriminate genomic differences among microbial strains and species. As previously investigated by many researchers, discriminatory index of RAPD was signi cantly higher than that of ERIC-PCR for genotyping of bacterial isolates. It is worthy to note that the most prominent drawback of DNA barcoding and ngerprinting methods based on random primer DNA typing such as RAPD and ERIC-PCR to identify microbial serotype and species is inherent poor reproducibility [28]. This main inherent problem has not been solved yet; also, we have not investigated on this challenge at the present study. However, it is more appreciated and suggested to improve and develop DNA ngerprinting methods with higher reproducibility in future studies. Nevertheless, we introduced RAPD-PCR-HRM assay as a potential alternative method for non-dysenteriae Shigella species identi cation; however, future studies with more sample size, and wide range of different molecular techniques are suggested to be implemented.

Conclusion
DNA ngerprinting methods usually are used for genotyping of microbial pathogens. These methods can be considered as the potential of alternative methods for differentiation of microbial serotypes and species. In this study, we developed and evaluated RAPD and ERIC-PCR methods coupled with HRM assay for identi cation of non-dysenteriae Shigella species isolated from clinical specimens. The authors attempted to analyze the data set obtained from the difference plots of HRM by PCA as a dimension reduction method to simplify the classi cation of melting pro les. We isolated eighteen Shigella strains in faecal specimens collected from children (n = 143) with gastrointestinal disorders and diarrhea symptom then species of the strains were identi ed by slide agglutination assay as the golden standard method. Also, species of the isolates were identi ed by developed RAPD-PCR-HRM and ERIC-PCR-HRM methods analyzed with PCA. We found RAPD-PCR-HRM assay more appropriate than ERIC-PCR-HRM as the potential of alternative method for non-dysenteriae Shigella species identi cation with sensitivity and speci city of 100 and 85%, respectively. Inherent poor reproducibility is the only disadvantage of all DNA ngerprinting based methods such as RAPD for genotyping or species and serotype identi cation which is suggested to be improved for the future studies.

Fecal samples collection and reference bacterial strains
Children under 2 years old with diarrhea symptom suspected to gastroenteritis (n = 163) in Qazvin, Iran, during December 2019 to February 2020 were referred to the central laboratory of Qods Children Hospital of Qazvin for microbiological investigation. As previously described by WHO Laboratory investigations manual of Enteric infections for faecal sample preparation, all suspected specimens were collected in disposable and clean containers then examined macroscopically and microscopically for color, mucus, consistency, and red blood cells [14]. All specimens were kept at 4 °C prior to bacterial isolation procedure. We also used S. sonnei ATCC 25931; S. exneri ATCC 12022; and S. boydii ATCC 12030, purchased from Pasteur institute of Iran (Pasteur institute, Tehran, Iran), as Shigella reference strains in this study. All reference strains were activated by inoculation in Bovine Heart Infusion (BHI) broth and incubation at 37 °C overnight.

Shigella isolation and species identi cation
All specimens were initially cultured on MacConkey agar and Salmonella-Shigella agar (SSA) media (Promedia, Spain) incubated aerobically for 24 h at 37 °C. Non-lactose-fermenting and H2S-negative colonies on MacConkey agar and SSA were isolated for biochemically identi cation. Biochemical tests consisted of Urease, Triple Sugar Iron (TSI), Motility, Siminous Citrate, and Indole tests. Biochemically con rmed colonies were further serologically analyzed for Shigella species identi cation. Serological identi cation was carried out using slide-agglutination test with DIFCO Shigella species speci c antisera (Difco Lab, Michigan, USA) as the gold standard assay [15].

Genomic DNA extraction
Bacterial isolates and the reference strains were inoculated into LB broth then incubated at 37 °C overnight. After centrifugation of the incubated LB broth tubes at 6000 RPM for 10 min, the cell palette in each centrifuge tube were subjected to genomic DNA extraction employing Sinaclon gram-negative DNA extraction Kit (Sinaclon, Iran) according to the instructions described by the manufacturer. The quantity and quality of the extracted genome were evaluated using NanoDrop Spectrophotometer (Thermo Scienti c, USA). Before PCR-HRM, concentration of all DNA templates was adjusted to 50 ng. µL-1.
RAPD and ERIC-PCR coupled with HRM assay In this study, we used the single random oligonucleotide primer 5-CGCGTGℂAG -3 for RAPD-PCR and species identi cation of isolated Shigella strains [16]. Each reaction tube contained 20 µL of nal volume, including 10 µL of 2X HotStart EvaGreen PCR-HRM master mix (Solis BioDyne Co, Estonia), 0.5 µL of RAPD primer (20 pM), 1 µL of DNA template (50 ng), and distilled deionized water to the nal reaction volume. RAPD-PCR coupled with HRM assay was performed using Rotor-Gene 6000 real-time PCR instrument (Corbett, Australia) as follows: initial denaturation at 95 °C for 5 min; 35 cycles of 95 °C for 1 min, 36 °C for 1 min, and 4 min at 72 °C. Then HRM was performed between 70 and 95 °C with continuous ramping 0.1 °C s-1. Fluorescence of melting pro le was normalized between 80 and 95 °C by Rotor-Gene 6000 series software version 1.7 (Corbett, Australia) then normalized and difference melting curves were obtained for each isolate.
For ERIC-PCR ampli cation two primers have been used including ERIC1R 5 -ATGT ∀ GCTℂTGGGGATTCAC -3 and ERIC2 5-∀ GT ∀ GTGACTGGGGTGAGCG -3 [16]. Both RAPD and ERIC primers used in this study previously were employed by researchers for DNA ngerprinting of enterobacterial pathogens. ERIC-PCR was performed in 20 µL nal reaction volume containing 10 µL of 2X HotStart EvaGreen PCR-HRM master mix (Solis BioDyne Co, Estonia), 1 µL of each primer (10 pM), 1 µL of DNA template, and distilled deionized water to 20 µL. Rotor-Gene 6000 real-time PCR instrument (Corbett, Australia) also was employed for ERIC-PCR-HRM procedure. Thermal cycling program was initial denaturation step at 95 °C for 5 min followed by 30 cycles of denaturation at 95 °C for 1 min, annealing at 52 °C for 1 min, and extension at 65 °C for 8 min. Afterwards, HRM was carried out from 70 to 95 °C with 0.1 °C s-1 ramping. Also, normalized and difference melting curves nally were achieved from the Rotor-Gene 6000 series software by uorescence normalization of melting pro le between 80 and 95 °C.

Data analysis
All difference curves obtained from Rotor-Gene 6000 series software were exported to Excel (Microsoft O ce Excel software, Microsoft Corp., Redmond, USA) les then they were used as the input data of analysis. Principal components analysis (PCA) was employed for analysis and data categorization as a dimension reduction method by SPSS software version 23.0 (SPSS, Inc., Chicago, USA). Also, sensitivity and speci city of the methods have been evaluated using receiver operating characteristic (ROC) curve analysis by SPSS. All measurements and analysis were carried out in triplicates.