Two dimensional gel electrophoresis (2-DE) for high-throughput proteome analyses of Mycoplasma bovis

1Department of Animal Health, The University of Agriculture Peshawar, Khyber Pakhtunkhwa, 25120, Pakistan; 2The State Key Laboratory of Agricultural Microbiology, Huazhong Agricultural University, Wuhan 430070, China; 3College of Veterinary Medicine, Huazhong Agricultural University, Wuhan 430070, China; 4Institute of Basic Medical Sciences, Khyber Medical University, Peshawar, Pakistan; 5Department of Livestock Management, Breeding and Genetics, The University of Agriculture Peshawar, Khyber Pakhtunkhwa, 25120, Pakistan


INTRODUCTION
Mycoplasma bovis (M. bovis), a causative agent of mastitis, was documented as a pathogen of cattle pneumonia in 1976 (Caswell & Archambault, 2007). It is still a serious threat to the beef and dairy industry, due to its evasion from the host immunity, variability in the surface antigen, resistance to antibiotics (Robino et al., 2005) and invasion of peripheral-blood mononuclear cells and erythrocytes (Van der Merwe et al., 2010).
Although M. bovis has been discovered nearly six decades ago, its pathogenesis remains obscure. Accessible information regarding its proteome is very scant for inferring its pathogenesis. Certain proteins, such as p48 (Sachse et al., 1993), pMB67 (Behrens et al., 1996), variable surface proteins (VSPs) (Scherm et al., 2002), p26 (Thomas et al., 2003), GAPDH (Prysliak et al., 2013), TrmFO , NADH oxidase (Gang et al., 2017) and P27  have been reported to be involved in its adhesion to the host cells and are immunogenic. However, the immune response to these antigens proved to be non-protective. Fluctuations in the expression, conformation, and antigenicity of the VSPs (Nicholas & Ayling, 2003) and pMB67 proteins might be one way in which M. bovis continuously evades the host immune response (Caswell & Archambault, 2007).
Comprehensive investigation of pathogenesis of M. bovis was sought and began in 2008 in China with the identification and isolation of M. bovis strain HB0801 from calf lung lesions in the Hubei province (Lei et al., 2008). Consequently, complete genome sequencing of the HB0801 strain and its comparison with the genomes of other strains and Mycoplasmas species laid down a foundation for the clarification of pathogenesis of M. bovis (Qi et al., 2012;Rasheed et al., 2017).
Full genome sequence of the HB0801 strain basically paved the way for proteomics in order to analyse protein coding genes for understanding of the biological value of the sequenced genome. Proteomics based on MS represent a potentially important tool for incorporating protein-level information into the genome annotation process. Combination of 2-DE with MS is currently the workhorse for proteomics (Jungblut et al., 2008;Khan et al., 2017), and particularly for immunoproteomics (Thomas et al., 2005;Zhang et al., 2008;Jores et al., 2009;Sun et al., 2012;Zhao et al., 2012;Khan et al., 2016;Khan et al., 2017). M. bovis proteins were usually analyzed employing conventional SDS-PAGE (1D), ELISA and immunoblotting. Utilization of 2-DE for the first time in the comparative proteomic analysis with detergent phase proteins of virulent and non-virulent strains of M. bovis identified a novel 24 kDa membrane associated protein (Thomas et al., 2005). More recently, 2-DE and immunoblotting were performed with whole-cell proteins of M. bovis PD revealing 19 immunogenic proteins. Moreover, an iELISA was established for diagnosis exploiting the E1 beta subunit of the pyruvate dehydrogenase complex (Sun et al., 2014). Although infrequent, successful application of 2-DE revealed important proteins of M. bovis as mentioned. These findings provide an insight for further improvement and application of 2-DE for comprehensive analyses of total proteins, membrane proteins, and immunogenic proteins of M. bovis.
The study presented here had efficiently calibrated each technical link of 2-DE for the analyses of total proteome, and more importantly immunoproteome of M. bovis. The 2-DE conditions standardized in this study would be likely very helpful in the future research on the elucidation of important factors of M. bovis related to its virulence in cattle.

Strain and culture conditions
M. bovis strain HB0801 (CCTCC # M2010040), isolated from pneumonia calf lung lesions (Lei et al., 2008), was utilized in the subsequent analysis. This strain was grown in a stationary CO 2 incubator as reported previously (Han et al., 2015), with minor modification, at 37 o C in the presence of 5% CO 2 in PPLO (Pleuropneumonia like organism) broth for 36 hrs, supplemented with 0.1% (w/v) sodium pyruvate and 0.5% (w/v) yeast extract, 0.001% (w/v) phenol red (pH indicator of the medium, color of PPLO broth changes with the growth of M. bovis due to acidic pH), 20% (v/v) heat inactivated (56 o C, 30 min) donor equine serum (Thermo Fisher Scientific, Waltham, MA, USA) and 40 000 U of penicillin G/100ml. Stock of this isolate was prepared by freezing broth culture in aliquots (1 ml and 5 ml) with 15% (v/v) glycerol as preservative at -80°C. For further analysis, the frozen stock was thawed and inoculated at a ratio of 1:100 in PPLO broth, in triplicate.

Confirmation of M. bovis HB0801
The bacterial strain used was confirmed to be M. bovis HB0801 by physical examination of biofilm formation over the PPLO broth (characteristic of M. bovis growth) and by PCR after 36 hrs of incubation in a log phase with 10 9 CFU, as described elsewhere (Qi et al., 2012). Briefly, two sets of primers were used for targeting the uvrC gene and Mbov-0732 gene for confirmation and differentiation from attenuated strains of HB0801. PCR was conducted using 2 µL of genomic DNA as a template in a 25 µL reaction mixture with 1 U of LA Taq DNA polymerase (TaKaRa, Tokyo, Japan) in 1X buffer supplied by the manufacturer, 200 µM dNTPs and 50 pmol of each primer. The amplification was programmed over 30 cycles, each consisting of 35 s at 95ºC, 40 s at 55ºC, and 42s at 72ºC. Initial denaturation step was programmed at 95ºC for 4 min, and final extension at 72ºC for 5 min.

Bacterial cell lysis for the extraction of proteins
Two methods, i.e. sonication and high pressure homogenization, commonly used for bacterial cell lysis and extraction of whole cell proteins were employed and their extraction efficiency and quality was compared.
Sonication. Total proteins of HB0801 were extracted after 36 hrs of incubation in log phase with 10 9 CFU as reported previously (Jungblut et al., 2008), with slight alteration. Briefly, PPLO broth cultured cells were centrifuged at 15400 × g, for 20 min, at 4 o C. Cell pellets (100-200 mg), in triplicate, were washed three times with cold PBS (150 mM NaCl, 1.5 mM KH 2 PO 4 , 9 mM Na 2 H-PO 4 -12H 2 O, 2.5 mM KCl, pH 7.4) to remove interfering substances and then resuspended in the lysis buffer (Table 1) containing a protease inhibitor (Sigma-Aldrich, USA). M. bovis cells were lysed by sonication with a 3 mm diameter probe, at 60-80% amplitude, 60% cycle duty (0.6 sec pulse rate) for 30 min at 4 o C in ice-water, incubated for 3 h at room temperature and centrifuged 30 min at 3000 × g to remove debris and unlysed cells. Protein concentration was estimated with 2D Quant Kit (GE healthcare, Fisher Scientific, Thermo Fisher Scientific, USA). Solubilized proteins were immediately used or stored in aliquots at -80 o C until further use.

High pressure homogenization
Slightly modified high pressure lysis conditions were used as reported previously (Parraga-nino et al., 2012;Zhang et al., 2008). In brief, cells grown in PPLO broth were centrifuged at 15400 × g, for 20 min, at 4 o C. Pellet (100-200 mg), in triplicate, was washed with cold PBS (1X) three times. Cells were resuspended in 30 ml PBS (1X) containing a protease inhibitor (Sigma-Aldrich, USA) and lysed by three passages through a French pressure cell (Thermo, USA) at 20 000 lbf/in. (poundforce per square inch); unlysed cells were removed by centrifugation at 15400 × g, for 20 min, at 4 o C. Supernatant was collected in another tube and treated with 15% TCA (Sigma). After 1h incubation at -20 o C, the sample was centrifuged at 15400 × g, for 20 min, at 4 o C. Protein pellet was then washed three times with ice-cold acetone. After drying, protein pellet was resuspended in the lysis buffer (Table 2) and incubated for 3h at room temperature. Protein concentration was estimated with 2D Quant Kit (GE healthcare, Fisher Scientific, Thermo Fisher Scientific, USA).

Two dimensional gel electrophoresis (2-DE)
Various technical steps of 2-DE given below were compared and analyzed.
Immobilized pH gradient (IPG) gel strips. IPG gel strips (Bio-Rad, USA) of various pH ranges and lengths (Table 3), in triplicate, were thoroughly and separately evaluated in this study to generate a high-throughput resolution and reproducible proteome map of M. bovis HB0801.
Rehydration of IPG strips. To figure out promising rehydration conditions for IPG strips, each strip (in triplicate) of different length (7 cm/17 cm) and pH range (pH 3-10 NL/pH 4-7) was rehydrated passively (Sun et al., 2014) for 1 h without overlaying mineral oil (Bio-Rad IPG strip manual). This was followed by active rehydration (50 V) of each 7 cm strip for 12 h and 17 cm for 16 h in IEF tray (Bio-Rad) at 20 o C, layered with mineral oil (Parraga-Nino et al., 2012). Moreover, the effect of passive rehydration in disposable rehydration tray for 12 h (7 cm strips) and 16 h (17 cm strips) was also examined. For re-swelling of dry IPG strips, the rehydration solutions, as described elsewhere (Thomas et al., 2005;Zhang et al., 2008;Jores et al., 2009;Sun et al., 2012;Zhao et al., 2012) were compared, evaluated, and standardized (Table 3).
First dimension -iso-electric focusing (1D-IEF). Using the Protean IEF cell (Bio-Rad), IEF was carried out on 7 cm IPG strips (Table 1) according to a program reported elsewhere (Zhang et al., 2008) under slightly modified conditions, i.e. 200 V for 30 min, 300 V for 30 min, 1000 V for 3 h, and 4000 V for 10 000 vhr. Whereas the IEF program, as described previously (Sun et al., 2014), for 17 cm was modified as followed: 150 V for 3 h, 300 V for 3 h, 1000 V for 6 h, 10 000 V for 3 h, and 10 000 V for 60 000 vhr. IEF on each strip (7 cm/17 cm) was performed in triplicate.
Gel staining. All 2D gels of each strip (Table 1) were stained with 0.15% CBB R-250 for about 7 h with freshly prepared stain and overnight with a used stain. Furthermore, 2-DE gels were twice de-stained with de-staining solution (40% ethanol, 10% acetic acid) for 30 min each, as previously reported (Sun et al., 2014). Then, the gels were washed three times with distilled water and   Gel staining was performed in triplicate for each strip.

Confirmation of Strain
The strain under investigation was confirmed to be M. bovis strain both, physically by biofilm formation and color change of broth, and by molecular methods by obtaining a 238 bp amplicon of the targeted uvrC gene and 146 bp of Mbov-0732 gene with PCR ( Fig.  1). Presence of Mbov-0732 gene differentiated the virulent HB0801 from its attenuated strains (not shown).

Total protein concentration of M. bovis HB0801
Two methods widely used for the extraction of total bacterial proteins (sonication and high pressure lysis) were compared for the whole cell protein extract of M. bovis HB0801 and were subsequently evaluated. Sonication at 4 o C extracted a measurable protein concentration of about 6.2 mg/ml, whereas a higher protein concentration (8.3 mg/ml) with high purity was successfully acquired with French Pressure Cell coupled with TCA precipitation and acetone washing. Moreover, quality of the resultant 2-DE map of protein sample extracted by sonication was found to be lower than that of the high pressure lysis and TCA treated sample (Fig. 2 and 3).

IEF, Equilibration, and SDS PAGE
Two-dimensional electrophoresis was carried out analyzing different conditions. After extensive assessments, electrophoresis with IEF of proteins on 17 cm IPG strip compared to 7 cm IPG strips, gel side down, pH 3-10 NL using program: 150 V for 3 h, 300 V for 3 h, 1000 V for 6 h, 10 000 V for 3 h, and 10000 V for 60 000 vh, followed by equilibration as mentioned above were determined as good conditions, when compared to those reported elsewhere Sun et al., 2014).

Gel running conditions, staining, and detection of protein spots
Focused IPG strip (17 cm), in triplicate, run simultaneously on freshly prepared gel (10%) for 3 h at 50 V and at 100 V for 15 h, in the same fresh electrode buffer and stained with fresh CBB-R250, were found to generate relatively better 2-DE images than 8% and 12% gels ran for 12 h and 18 h respectively. About 639-680 highly resolved protein spots were detected (Fig. 2), representing 83.85-89.23% of the total coding sequences of M. bovis HB0801. This 2DE process was found to be reproducible by running different batches of protein samples in triplicate.

DISCUSSION
M. bovis is one of the leading threats to the dairy and beef industry. It has been identified nearly six decades ago. However, its presence in China has been reported since 2008. Availability of very little information regarding virulence factors and pathogenesis makes its control very challenging. Recent complete genome sequences of M. bovis typed strain PG45 (Wise et al., 2011), M. bovis Hubei-1 (Li et al., 2011) and M. bovis HB0801 (Sun et al., 2012) paved the way for elucidation of its pathogenesis exploiting proteomics approaches. Comprehensive proteomics approaches enable us to understand vital processes of any organism (Le . Proteomic studies are developing quickly and use of 2-DE is not uncommon for proteome analysis. Hence, obtaining high resolution 2DE maps with good reproducibility is the bottleneck of proteomics. Although the advent of IPG gel strips has enhanced the reproducibility of 2-DE results (Gorg et al., 2007), there are several other factors which have to be streamlined in ordered to establish a 2DE for each particular subject, as documented previously (Ansong et al., 2008). Furthermore, 2-DE with high resolution power has an ability to separate proteins with distinct post-translational modifications, which makes it an invaluable method for proteomic analyses.
Existences of significant evolutionary changes in the molecular machinery of M. bovis (Qi et al., 2012) urge optimization of 2-DE for the in-depth exploration of common virulence factors in various strains of M. bovis. This may lead to the development of potentially more comprehensive control strategy. Although 2-DE used over the past decade for the proteome and immunoproteome analysis of M. bovis (Thomas et al., 2005;Sun et al., 2014), the study presented here is presenting more efficient 2-DE conditions for the detail proteome analysis of M. bovis strain HB0801, after extensive calibration of each technical step of 2-DE.
Extraction and preparation of the protein sample are the benchmarks in the 2-DE. Their quality may directly affect the process of 2-DE and determine the accuracy of final results. An appropriate method and buffer have to be utilized to delicately remove and solubilize proteins with minimal losses. In this study, the HB0801 strain cells were lysed by sonication and high pressure homogenization. The same lysis buffer as described previously (Parraga-Nino et al., 2012) was used for solubilization of proteins extracted with both methods. However, 2-DE images obtained with French pressure cell and TCA treated sample were of good quality (Fig. 2), which might be due to TCA, as described previously (Gorg et al., 2007), which is effective in the removal of interfering substances (salts, nucleic acids, fats, and carbohydrates)  Proteins were separated on 17 cm, pH 4-7 IPG strip and 10% SDS-PAGE, and stained with CBB R-250. Protein samples of different batches were run in triplicates. Quality 2-DE image was obtained, but certain proteins at the basic end were missing as compared to Fig. 1. pI values are shown on top. and eventually prevents vertical and horizontal streaking in 2-DE gels.
Reproducibility of 2-DE map has increased enormously with the advent of immobilized pH gradient strips (Gorg et al., 2007). Still, certain 2-DE conditions still need extensive work to optimize for specific protein sample. Therefore, various IPG strips were evaluated in this project. Strip (17 cm) of pH ranging from 3-10 NL was proved to provide good separation distance for the proteins of HB0801. Additionally, narrow pH range strips (Table 1) were also evaluated, as described elsewhere (Gorg et al., 2007), since they provide a large separation distance to proteins. Although 17 cm strips of pH ranges 4-7 and 5-8 (Fig. 4), were found good in resolution of respective proteins but some basic proteins were found missing in the gel strip of pH 4-7 when compared to pH 3-10 NL (Fig. 6). Moreover, 7 cm strips (pH 3-10 NL, 4-7, 5-8) were found not to be feasible in resolving M. bovis proteins, which might be due to their short length.
Appropriate rehydration solution is also very important for re-swelling of a dry IPG strip to their inherent size of 0.5 mm, according to manufacturer's instructions (Bio-Rad). Therefore, rehydration was performed both, actively at 50 V and passively for 16 h, respectively, as described elsewhere (Gorg et al., 2007). Active rehydration with optimized rehydration solution (Table 3) excluding destreak reagents (Parraga-Nino et al., 2012), and addition of 65 mM DTT (Li et al., 2009), was observed to work well for resolution. DTT is known to be a good reducing agent, and in combination with other reagents in a given buffer it helps in solubilization of proteins by reducing the disulfide bonds of proteins. Furthermore, protein loading concentration also affects the quality of subsequent 2-DE map (Biorad Bulletin 3110). In this study, various protein loading concentrations were analyzed. Proteins concentration range of 330-380 μg for 17 cm strip, as reported (Li et al., 2009), was found to be more promising than 1-4 mg reported elsewhere (Zhao et al., 2012), which might be due to the differences in behavior of various microbial proteins in different 2DE environments.
Isoelectric focusing was performed with a slightly modified setup (Jores et al., 2009;Pang et al., 2010;Sun et al., 2014), using Protean IEF cell at 50 μA/strip. Samples that were extracted by sonication were found to have ionic contaminants, which was initially indicated in the run by very early achievement of the target current for IEF at a very low voltage, and later on by poor quality 2DE images. On the other hand, the TCA/acetone treated samples (Zhang et al., 2008), were found to be almost free of contaminants, which was indicated by 6 μA/strip at low voltage (150 V, 300 V) and 33 μA/ strip at high voltage (10 000 V) till the end of IEF, and later on by an effective 2DE map. Equilibration of strips needs to be done delicately for better transfer of focused protein and movement during SDS-PAGE. Equilibration of the focused proteins was performed in a reducing and alkylating solution (Zhang et al., 2008) for a bit longer (20 min each).
Furthermore, SDS-PAGE was performed using both, a relatively old and fresh 30% acrylamide. Fresh 30% acrylamide (Bio-Rad) with C=3.3 (29:1) was found to be relatively better in separation of proteins according to their respective molecular mass (M r ) than an old one (more than 6 months old).
Previously, certain efforts recognized some proteins of M. bovis by SDS-PAGE (1D), and 2-DE. However; potentially important virulent factors for understanding the pathogenesis of M. bovis are yet to be explored. 2-DE was utilized infrequently in the past decade for getting high resolution proteome map of M. bovis. Therefore, the effort presented here had standardized more precise 2-DE conditions, after extensive calibration of each step in the proteomics procedure, for a self-explanatory 2-DE map (Fig. 2). Exploiting these 2-DE conditions would be possibly useful in future efforts of elucidation of virulent factors and pathogenesis of M. bovis HB0801 (Chinese strain).