Development Of A Binary Gene Typing Scheme For Clostridium Difficile And Comparison With MLST

Objective: Clostridium difficile（CD）is an obligate anaerobic,Gram-positive,spore–forming rod bacillus.Early in the 1980 s,C.difficile was firstly found to be of the etiological agents of antibiotic-associated pseudomembranous colitis.Recently C.difficile infection（CDI）is commonly recognized the main cause of antibiotic-associated diarrhea（AAD）with clinical symptoms ranging from mild and self-limiting diarrhea to severe pseudomembranous colitis and even death during antibiotic treatment or shortly afterwards.The major cause of CDI is the use of antibiotics destroy the balance of normal intestinal flora and cause C.difficile to multiply producing toxins-TcdA and TcdB.Molecular typing for C.difficile is important for tracking epidemic strains of C.difficile associated diarrhea（CDAD）,surveillance of outbreaks within healthcare facilities,investigation of its propagation pathways across the globe and exploration of its pathogenic mechanism.As the prevalence of CDI increases globally as a consequence of high levels of antibiotic use,several techniques for investigating its epidemiology have been developed including PCR ribotyping（PCR RT）,pulsed-field gel electrophoresis（PFGE）and multilocus sequence typing（MLST）.However,these typing methods have different defects around discriminatory power,reproducibility,operability,cost and transportability.Binary gene typing,which relies on a code derived from presence or absence of genes widely distributed in bacterial genomes,is another typing technique that has been successfully used for Campylobacter jejuni and Staphylococcus aureus.This technique is based on the idea that every binary target（gene present in some strains but not in others）could divide strains into two different groups.Binary typing has showed the following advantages over the many commonly used typing techniques described above: short turnaround time,simplicity,cost-effectiveness,portability,and discriminatory ability.In this study,we aimed to establish the basis of a novel PCR binary typing system that is inexpensive,rapid,and highly portable for future routine surveillance and outbreak investigations of Clostridium difficile.Methods: The full length genome sequences of C.difficile R20291,QCD-63q42,QCD-76w55 and C.difficile 630 were downloaded from Genebank.Genomes of C.difficile R20291,QCD-63q42,QCD-76w55 were compared with C.difficile 630 repectively using genome alignment software Mauve 2.3.1.Binary loci were defined as genes,which only existed in these strains or C.difficile 630,with a length >500bp and <1000bp.Sequences of every potential binary gene were used to “fish” its whole gene sequence by Basic Local Alignment Search Tool（BLAST;www.NCBI.nlm.nih.gov/BLAST/）.Initially,a total of 50 potential binary genes were found.Then,50 whole genome sequences of C.difficile selected randomly from NCBI genome database were used as a sample library for intial selection.Each of the 50 potential binary genes was BLAST with the database of “Whole-genome shotgun contigs”.The positive detection of a binary gene in each genome sequence of the sample library was recorded as “1”,while the negative result of a binary gene in an isolate was recorded as “0” in an Excel file according to the results of BLAST query.The Simpson’s index（SI）of all possible combinations of a given number of loci from the 50 loci was calculated for the 50 genome sequences using the computer program Optimal Combination Finder（OCF）.Then,a subset of 28 loci was selected for subsequent local strains analysis which yielded the same discriminatory potential as the entire 50 loci scheme.Primers were designed for the subset of 28 binary genes and the presence and absence of these genes in 99 clinical C.difficile isolates and one reference strain ATCC BAA-1870 were determined using PCR amplification.The optimal binary gene combination of the 28 binary loci in typing was evaluated using OCF.The optimal binary typing combination was found to be a panel of 10 loci,which produced the same SI as that generated by all 28 loci for the tested strains.To evaluate the advantage of this typing technique,comparison of PCR ribotyping and binary typing was performed using the same test isolates group in this study.Finally,10 loci typing scheme defined as the finalized typing methods for C.difficile binary typing.We also conducted a comparison of binary typing and MLST to determine their utility in epidemiologic investigations of C.difficile using 100 C.difficile isolated from different geographic areas and population groups from 2010 to 2014.Results: The whole genome sequence alignment software Mauve version 2.3.1 was used to compare four fully sequenced C.difficile genomes of 630,R20291,QCD-63q42,and QCD-76w55.Fifty candidate binary loci were chosen.With 50 fully sequenced C.difficile genomes publically available in GenBank as an in silico testing population,the distribution of the selected 50 binary genes was assessed among the C.difficile isolates in the NCBI genome database.As a result,a subset of 28 loci,which showed promising typing potential,was determined for further analysis.Primers were designed for the subset of 28 binary genes and the presence and absence of these genes in all 100 test isolates were determined using PCR amplification.The optimal binary typing combination was found to be a panel of 10 loci,which produced the same SI as that generated by all 28 loci for the tested strains.The optimal 10 loci are listed in Table 1.The binary typing profile of the 10 loci was arranged in the order of CDBI1₀2275,tetM,20070-1,23S_rRNA,CDBI1₀2830,CDBI1₀9860,CD630₃1520,CDR20291₀581（tcdR）,CDBI1₀1190,and CDBI1₀3750.This binary typing panel produced a discriminatory ability of SI = 0.908.A comparison between PCR ribotyping and binary typing was performed using the same test isolates in this study.In total,16 PCR ribotypes were identified with an SI value of 0.847.The hypervirulent ribotype 027 was not identified in the clinical isolates tested.The most commonly identified PCR ribotype was S3,being 30 out of 100 isolates tested.The other frequent identified PCR ribotypes were S4,S1,and S2 with 16,14,and 13 isolates,respectively.By contrast,binary typing was relatively more discriminatory,producing 28 binary type profiles.The combination of these two typing methods can increase the differentiation ability.The combined SI of ribotyping and binary typing was 0.934 with 39 genotypes.For the 100 C.difficile isolated from different geographic areas and population groups,10 loci binary typing and MLST reviewed different discriminatory ability with SI=0.938 and 0.886,respectively.Conclusions: Binary gene typing is a useful method for C.difficile subtyping,offering confident information with advantages of speed,cost and practicability by providing high-resolution bacterial fingerprints of C.difficile and was sufficiently discriminatory for isolates at an epidemiologically useful level.In addition,some binary genes of the 10 loci binary gene panel are associated with virulence and antibiotic resistance.So the results of binary typing could also provide information that is relevant to the consequent clinical treatment.In conclusion,compared with other conventional typing methods,binary typing shows unique clinical significance.