Multiple-Locus Variable-Number Tandem-Repeat Analysis of Mycoplasma pneumoniae Clinical Specimens and Proposal for Amendment of MLVA Nomenclature

Mycoplasma pneumoniae is one of the major respiratory bacterial pathogens that cause pneumonia in humans. Multiple-locus variable-number tandem-repeat analysis (MLVA) is currently the most discriminative method for typing M. pneumoniae strains. To better understand the epidemic of M. pneumoniae-related pneumonia in pediatric patients in Beijing, China, we performed MLVA analysis on 118 specimens collected during an epidemic from 2010–2012. Eleven distinct MLVA types were identified, including four novel types. There was no obvious association of macrolide resistance with any of the genotypes. Considering the instability of VNTR locus Mpn1, we propose an amended MLVA nomenclature system based on the remaining four VNTR loci.


Introduction
Mycoplasma pneumoniae is a common respiratory bacterial pathogen that causes pneumonia in humans, especially in children and young adults [1,2]. The infection is transmitted through close contact with an infected patient, and leads to epidemics in the immediate family and community. Worldwide epidemics of M. pneumoniae infection occur every 3 to 7 years [3,4]. Studies show that an epidemic has been spreading in many countries since 2010 [5,6,7,8,9,10]. To study an epidemic and identify the source of an outbreak, it is important to type the pathogenic strains rapidly and accurately.
Molecular typing of the M. pneumoniae P1 gene has been the most common genotyping method in the past 30 years. Two major types of P1 genes (types 1 and 2) [11,12,13], three subtypes, and three variants of these subtypes have been described [14,15,16,17,18]. Recently, a multiple-locus variable-number tandem-repeat (VNTR) analysis (MLVA) for molecular typing of M. pneumoniae was developed [19]. Twenty-six MLVA types (A-Z) were first introduced by typing 265 M. pneumoniae strains using five VNTR loci (Mpn1, . This highly-discriminatory method was quickly adapted by other research groups, and at least 18 novel types have been reported [20,5,6,21,22,8,23]. However, locus Mpn1 is unstable in both clinical strains and in laboratory passages, and most of the novel types came from variations in Mpn1 [20,8,23]. With more MLVA genotypes identified, the current M. pneumoniae MLVA nomenclature system needs to be amended.
In this study, we typed 118 clinical specimens collected from pediatric patients in Beijing, from 2010-2012, by MLVA and P1-RFLP. We also analyzed all reported MLVA types in the literature and propose a modified nomenclature for the M. pneumoniae MLVA types.

Ethics Statement
The present project was performed in compliance with the Helsinki Declaration (Ethical Principles for Medical Research Involving Human Subjects) and was approved by the research board of the Ethics Committee of the Capital Institute of Pediatrics, Beijing, China. All patient data were anonymously reported. Based on the guidelines of the Ethics Committee of the Capital Institute of Pediatrics this study did not need to be examined by the ethical committee, and therefore informed consent was not sought from patients.

Clinical Specimens, DNA Extraction, and Real-time PCR Detection
A collection of 681 clinical specimens was obtained from pediatric patients diagnosed with pneumonia or with respiratory tract infections in the Affiliated Children's Hospital of the Capital Institute of Pediatrics in Beijing. The specimens consisted of sputum samples, throat swabs, bronchoalveolar lavage fluid, and puncture fluid, of which 178 specimens were collected in 2010, 159 in 2011, and 344 in 2012. Each specimen was collected from a single patient. DNA samples were extracted using the TIANcombi DNA Lyse & Amp PCR Kit (Tiangen, Hangzhou, China) according to the manufacturer's instructions. Briefly, 0.4 mL of the specimen was centrifuged at 13,8006g for 15 min and the pellet was resuspended in 0.5 mL of lysis buffer (150 mM NaCl, 20 mM sodium citrate, 10 mM Tris-HCl, 1 mM EDTA, 1 mg/ mL proteinase K, 20 mg/mL RNase A, 1% (w/v) sodium dodecyl sulfate). Following incubation at 55uC for 60 min and 100uC for 10 min, lysates were centrifuged at 13,8006g for 3 min. The supernatants were transferred to a purification column and the DNA was eluted in a volume of 50 mL. The DNA was immediately used for real-time PCR detection of M. pneumoniae [15] or stored at -20uC.

P1 Gene Typing
To improve assay sensitivity, a nested P1 gene PCR-RFLP method was performed directly on the M. pneumoniae-positive samples,. To detect the RepMP4 region of the P1gene, the ADH1/ADH2 primer pair [13] was used for the first PCR, and the ADH1in/ADH1M and ADH2in/ADH2M primer pairs were designed and used for the second PCR (Table 1). To detect the RepMp2/3 region, the ADH3/ADH4 primer pair [13] was used for the first PCR, and the ADH3in/ADH3M and ADH4in/ ADH4M primer pairs were designed and used for the second PCR ( Table 1). The nested PCR conditions were as follow: 94uC for 10 min, 30 cycles of 94uC for 1 min, 55uC for 1 min, and 72uC for 2 min, followed 72uC for 10 min. PCRs were performed on a Thermal cycle PX2 apparatus (Becton Dickinson, Franklin Lakes, NJ). The nested PCR products of the RepMP4 and RepMp2/3 regions were digested using the restriction enzyme HaeIII and were run on a 2% agarose gel. The restriction fragment length polymorphism (RFLP) was analyzed as described [13].

MLVA Typing
MLVA was performed based on a previously described method, with modifications [19]. First, a nested PCR that had been adapted for direct clinical specimens was carried out using outer and inner primers targeting the five selected loci containing tandem repeats (TRs, Mpn1, Mpn13, Mpn14, Mpn15, Mpn16) [21]. Briefly, a 25 mL reaction mixture containing 3 mL extracted DNA, 0.5 mL each of forward and reverse primer (10 mmol/L), and 12.5 mL 26 Qiagen HotStarTaq polymerase mix was run under cycling conditions of 95uC for 10 min, followed by 35 cycles of 95uC for 30 s, 55uC for 30 s, and 72uC for 1 min, and a final extension at 72uC for 10 min. The PCR products were sequenced and the VNTR copy numbers were counted from the sequence results (Supporting Information).
In this study, to fully describe a strain using the molecular signatures from both typing methods, the MLVA type was designated by a prefix of M plus the number of repeats at each locus in the order of Mpn1-Mpn13-Mpn14-Mpn15-Mpn16 (Mnn-n-n-n) [24]. The combination of the MLVA and P1 typing was presented as Mn-n-n-n-nPn (M-P).

Detection of Macrolide Resistance Gene
The point mutations at positions 2063, 2064, 2616, and 2617 of the 23S rRNA gene of M. pneumoniae, which are responsible for macrolide resistance, were detected by a nested PCR-linked capillary electrophoresis and single-strand conformation polymorphism analysis (nPCR-CE-SSCP) based on a previous method [25].    The PCR products were purified and sequenced to verify mutations.

Data Analysis
The numbers of repeats identified at the five loci in the MLVA typing were recorded and imported into the Bionumerics (version 5.0) software package (Applied Maths). A minimum-spanning tree (MST) was generated based on categorical coefficient, and the priority rule was set as the highest number of single locus variants (SLVs). The polymorphism indexes of P1 gene typing, MLVA typing, and MLVA plus P1 (M-P) typing were calculated using the Hunter-Gaston diversity index (HGDI) [26].
Because the number of the patients with each of the genotypes was small, the associations between the genotype and sex, age, the average days of using macrolide and hospitalization, and the clinical severity of the patients could not be analyzed.
Examining the originally reported 26 MLVA types, we found that locus Mpn1 is highly variable, while the other four loci are more stable. If locus Mpn1 is omitted and only the other four loci are considered, the 26 MLVA types (A-Z) become nine types (Table 3). Using this simplified system, the MLVA types reported in the literature could be classified into 15 genotypes from type 2-6-6-2 to 4-6-7-2 (Table 4) [1,2,3,4,6,8,18,19,36]. The majority of the samples from this study (111/116, 95.7%) were type 4-5-7-2 (Table 4). This might indicate an association of a certain strain with this epidemic in pediatric patients in Beijing.

Discussion
The most recent epidemic outbreak of M. pneumoniae infection was reported in Denmark in late 2010 [10], and then a similar increase in M. pneumoniae infection was also noted in England and Wales over the same period [5]. In 2011, M. pneumoniae infection was reported in many other European and Asian countries [27, 28, 29, 3, 7, 9 and 30]. This study suggests that an epidemic outbreak also occurred in Beijing, China, during this period.    P1 gene PCR-RFLP has been the most common method for molecular typing of M. pneumoniae. The method only divided the strains into two types: types 1 and 2, while type 2 strains could be further subtyped into 2a, 2b, and 2c [31,14,32,8,33]. A type shift phenomenon was reported during different epidemic periods: type 2 strains were predominant between 1995 and 2001 [32], while type 1 strains were prevalent between 2002 and 2005 [32]. This study indicated that type 1 was the predominant genotype in the Beijing epidemic between 2010 and 2012. This was consistent with reports from China and France in 2011 [18,34].
The discriminatory power of MLVA is much higher than the P1 typing method. Combining MLVA typing with P1 typing provided a limited increase in discriminatory power, as shown in this study. While P1 typing is the gold standard of M. pneumoniae typing and has been used for over 30 years, it might be useful to compare the current results with the historical data. However, an open discussion of whether to continue using P1 typing in the future is encouraged.
Of particular concern is the uniformity of the definition of every tandem repeat number. As most reports used capillary electrophoresis (CE) and then GeneScan to count the number of TRs, and few reports used sequencing, there may be some difference of the TRs number between the two methods. Take M129 as an example, the genotype inferred from the CE result is 4-4-5-7-2, while the gene sequencing result gives a genotype of 4-3.2-5-6.2-2, or possibly 4-3-5-6-2. These discrepancies must be resolved for better comparison of the result of MLVA.
In this study, 90.8% specimens were shown to have macrolide resistance mutation in the 23S rRNA gene and this resistance was found to be scattered in each genotype, no association between any MLVA type and macrolide resistance was identified. This was consistent with a study from Shanghai, China [22], but differed from the report from France that found MLVA type Z was associated with resistance [8]. The macrolide-resistance prevalence is higher in China compared to those found in many other countries, a phenomenon that may be associated to the excessive use of antibiotics.
In summary, we reported an epidemic of M. pneumoniae in Beijing, China, during 2011-2012, with a predominant strain harboring MLVA type 4-5-7-2. An amended four-locus MLVA typing scheme was proposed for future M. pneumoniae molecular typing.  Supporting Information Figure S1 Different VNTR copy numbers of Mpn1 counted from the sequence results. (TIF)