Journal of Proteomics & EnzymologyISSN: 2470-1289

All submissions of the EM system will be redirected to Online Manuscript Submission System. Authors are requested to submit articles directly to Online Manuscript Submission System of respective journal.
antalya escort

Review Article, J Proteomics Enzymol Vol: 0 Issue: 0

Minor Differences in the Proteome of Bacillus subtilis and Bacillus mojavensis Based upon High Abundance/ Conserved Protein Mass Spectra; Implications for Rapid, Improved Identification of Two Pathogen Genetically Closely Related

Timothy Chambers1, Renata Culak1, Saheer E Gharbia2 and Haroun N Shah1*
1Department of Natural Sciences, Middlesex University,The Burroughs,Hendon, Middlesex, NW4 4BT
2Genomics Research Unit, Public Health England, 61 Colindale Avenue, London NW9 5EQ, UK
Corresponding author : Haroun N Shah
Department of Natural Sciences, Middlesex University, The Burroughs, Hendon, Middlesex, NW4 4BT
E-mail: [email protected] , [email protected]
Received: April 02, 2015 Accepted: April 30, 2015 Published: May 07, 2015
Citation: Chambers T, Culak R, Gharbia SE, Shah HN (2015) Minor Differences in the Proteome of Bacillus subtilis and Bacillus mojavensis Based upon High Abundance/Conserved Protein Mass Spectra; Implications for Rapid, Improved Identification of Two Genetically Closely Related Pathogens. J Proteomics Enzymol 4:1. doi:10.4172/2470-1289.1000119

Abstract

The genus Bacillus comprises a complex group of species, many of which cannot be readily differentiated by phenotypic and genotypic methods. Here, we utilised two species that are genetically very closely related, Bacillus subtilis and Bacillus mojavensis as a model to ascertain the potential of a linear MALDITOF MS to differentiate them against a background of their sporulation cycle and ultrastructural changes. Previous studies indicated that MALDI-TOF MS ionises the high abundance/conserved intracellular proteins but also produces additional lower mass ions that, with the appropriate software, may help to delineate such closely related taxa. Cells were grown in liquid culture over 24 hours and samples taken intermittently to monitor growth and the presence of spores. Harvested cells were studied by electron microscopy to detect potential changes in ultra structure and its potential effect on reliable and reproducible mass spectra. Sporulation was clearly evident by 24 hours and correlated inversely with reliable identification scores obtained using MALDI-TOF MS. Thus, cells cultured for 4, 6, 8 and 24 hours had identification scores of 2.157, 2.204, 2.295 and < 2 respectively. Contrary to previous findings, these results demonstrated unequivocally that Bacillus species may be reliably identified using this approach prior to sporulation. Furthermore, the software ClinProTools 3.0 was used to tease out minor components of the mass spectrum that enabled unambiguous separation of both groups of strains.

Keywords: MALDI-TOF MS; Sporulation; Bacillus mojavensis; Proteomics; Bacillus subtilis

Keywords
MALDI-TOF MS; Sporulation; Bacillus mojavensis; Proteomics; Bacillus subtilis
Introduction
The microbiological applications of ‘Matrix-assisted-laser desorption/ionization time-of-flight mass spectrometry’ (MALDITOF- MS) arose from early observations in 1996 [1-4] and over a period of some 12 years has arisen to the forefront of microbial diagnostics [5]. Today the technology is driven largely by Bruker Daltonics Ltd, Bio Mérieux-Shimadzu and to a lesser extent the MALDI-TOF MS Andromas system manufactured by SAI (Scientific Analysis Instruments, Manchester, UK). The method is now gaining universal appeal and few major laboratories are currently without a system. Commercial data bases,provided by these companies are continually updated and contain several thousand mass spectral profiles of relevant species. These databases often comprise extensive species diversity which is essential for comparative analysis of the mass spectrum of an unknown isolate. Poor identification is often assumed to be the result of absence of an appropriate matching mass spectral profile. Our work over the last seventeen years has shown that such defects cannot always be attributed to a deficiency in the database but rather the lack of clear demarcation between species that are genetically closely related species, often proposed on the basis of minute differences in characters, some of which are no longer extant. The present study focuses on such a problem involving members of the genus Bacillus which are known to exhibit many homogenous properties and the capacity of MALDI-TOF MS to delineate such closely related taxa. We selected Bacillus subtilis and Bacillus mojavensis as a model, as both are significant food pathogens and their accurate identification is critical.
It is widely recognised today that MALDI-TOF MS is capable of resolving most isolates to the species level and, in many instances to sub species. However, among members of the genus Bacillus, there are many species for which there is a paucity of reliable characters for accurate identification and many remain incertaesedis. A plethora of phenotypic and molecular methods [6-10] have been used to characterise this group that includes Bacillus amyloliquefaciens, Bacillus licheniformis, Bacillus mojavensis, Bacillus pumilis, and Bacillus valismortis but to date results can still be inconclusive. Years ago this was achieved using DNA-DNA hybridisation [9] but such methods are rarely used today and microbiologists have been searching for new approaches.
We have shown previously that the mass spectrum of Bacillus spp contains relatively few characteristic mass ions and several are suppressed by the presence of spores [11]. Consequently, the present study attempts to utilise the high abundance/conserved proteins that MALDI-TOF MS ionises against a background of the sporulation cycle of these species and aims to tease out minor spectral differences using newly available software as a test system to begin dissecting out such differences among the large number of taxa where this problem still prevails
Materials and Methods
Bacterial cultures and growth
For the growth and harvest time evaluation, four strains of B. subtilis were sourced from the National Collection of Type Cultures (NCTC 5398, NCTC 6431, NCTC 6432 and NCTC 10400). To create the species classification model, ten B. subtilis strains were used; 5 reference strains: NCTC 3610, NCTC 5398, NCTC 7861, NCTC 9933 and NCTC 10315 and 5 clinical isolates obtained from Public Health England. All 5 clinical strains were isolated from tissue. All strains were cultured on Colombia Blood Agar (CBA) at 37ºC under aerobic conditions.
To establish a growth curve, cells were was harvested from four overnight cultures of B. subtilis grown on Columbia Blood Agar and inoculated into tubes of 250 μl of distilled water until an optical density of 3 McFarlands standards was achieved. The suspensions were then pipetted into individual tubes containing 10 ml of nutrient broth and a 600 nm optical density reading of the cultures was taken at 0 minutes using a Vitek Densi Chek densitometer (bioMérieux). The cultures were incubated at 37ºC and readings taken every 30 minutes to an endpoint of 600 minutes. A mean from the readings was determined and plotted onto a graph to form a growth curve for B. subtilis.
Three isolates were cultured for 48 hours and harvested at 4 hours, 6 hours, 24 hours and at 48 hours for ultrastructural studies using electron microscopy (EM). The EM work was undertaken by the Electron Microscopy facility at Public Health England using a JEM-1400 Transmission Electron Microscope (JEOL USA). To fix the bacteria for EM, a 2 μl loopful of each of the bacterial strains was suspended in a 1:1 solution of 100μl of PBS and Carson’s Buffered Formalin solution [12].
Proteomic analysis using Matrix-assisted laser-desorption/ ionization time-of-flight mass spectrometry (MALDI-TOF MS): Mass spectrometry was undertaken using the Bruker Microflex MALDI-TOF instrument used in conjunction with MALDI Biotyper 3.1 classification software. The formic acid-acetonitrile extraction method was applied to all isolates and mass spectra were acquired as previously described [13]. Briefly, bacterial cells were suspended in 300 μl of water and mixed with 900 μl of absolute ethanol. Ethanol was completely removed after second centrifugation (14,000 g for 2 min) and the pellets were re-suspended in 50 μl of 70% formic acid and 50 μl of pure acetonitrile. Samples were vortexed and centrifuged at 14,000 × g for 2 min and 1 μl of each supernatant was placed in duplicate onto MSP 96 target polished steel (BrukeDaltonics) and air-dried. The spots were overlaid with 1 μL of α-cyano-4-hydroxycinnamic acid (CHCA) (saturated solution of CHCA in 50% acetonitrile/ 2.5 % trifluoroacetic acid) and air dried to co-crystalise the sample and matrix. The target plate was then processed using a Microflex LT mass spectrometer (BrukerDaltonics, UK, software version 3.4). Data collection was done in an automatic mode by collecting 240 laser shots from 6 different positions within the spot. MALDI measurements were recorded in a positive linear mode with mass range 2-20 kDa. Each isolate was analysed in triplicate to ensure reproducibility of the results. Bruker’s ClinProTools 3.0 software was used for data post-processing and biomarker detection to build classification models to differentiate between the two species as described below. The automated approach was performed using three ClinProTools functions: data preparation, model generation, and spectra classification. Data preparation involved baseline subtraction (tophat; 10% minimal baseline width), normalization (total ion current), recalibration (1,000 ppm maximal peak shift and 30% match to calibrant peaks, with exclusion of spectra that could not be recalibrated), average spectrum calculation (resolution; 800), average peak list calculation (signal-to-noise threshold; 5), peak calculation in the individual spectra, and normalisation of peak lists. Model generation using the genetic algorithm [13,14] was performed using the following settings: 15 peaks, automatic detection of initial number of peak combinations, 50 generations, 0.2 mutation rate, 0.5 crossover rate, no varying random seed, and 3 neighbours. Classification of unknown spectra was achieved by using the “Classify” function in ClinPro Tools. If 2 of 3 spectra per isolate were assigned to the same class, the identification was accepted.
Results and Discussion
In the evaluation of growth times on successful analysis, the first cohort of NCTC Bacillus subtilis strains were grown for 24 hours and growth monitored by optical density. Concurrently, spore staining was undertaken and ultrastructural changes in the cells were observed across time by electron microscopy. The growth curves demonstrated that exponential growth slowed after eight hours while spore staining revealed that sporulation became prominent at 24 hours, significantly earlier than expected (Figure 1).
Figure 1: The relationship between the life cycle of Bacillus subtilis and MALDI-TOF identification scores. MALDI-TOF MS analysis was undertaken for cells cultured for 4, 6, 8 and 24 hours which revealed identification scores of 2.157, 2.204, 2.295 and <2 respectively The results revealed clearly that for Bacillus species, MALDI-TOF MS needs to be carried out in the exponential growth phase and not at the stationary phase as is presently done.
Electron microscopy showed that between 4 and 8 hours, cultures of B. subtilis grew in an organised chain structure (Figure 2). Overnight, this uniformity was lost and the cells had degraded (Figure 3). Extracellular debris and spores were apparent. These changes correlated with a decrease in the quality of MALDI-TOF MS spectra and resulted in low confidence identification score (<2) or a failed result (Figure 4).
Figure 2: Electron micrograph of Bacillus subtilis cells grown for 4 (left) and 6 hours (right) Cells are filamentous and show no evidence of spores.
Figure 3: Typical mass spectral profiles resulting from MALDI-TOF MS analysis of Bacillus subtilis after culturing for 4 (A), 6 (B), 8 (C) hours and overnight (D) from top to bottom respectively. The changes in the density of mass ions resulting from the longer growth period are apparent. Many of the distinctive mass ions, eg. 4,305, 6,507, 7,713 and 15,083 Daltons were derived from the vegetative cells grown between 4 - 8 hours. With the onset of sporulation, these peaks disappeared and mass ions associated with the spores such as 3,460, 12,871, 16,022 Daltons were evident.
Figure 4: Typical mass spectral profiles resulting from MALDI-TOF MS analysis of Bacillus subtilis after culturing for 4 (A), 6 (B), 8 (C) hours and overnight (D) from top to bottom respectively. The changes in the density of mass ions resulting from the longer growth period are apparent. Many of the distinctive mass ions, eg. 4,305, 6,507, 7,713 and 15,083 Daltons were derived from the vegetative cells grown between 4 - 8 hours. With the onset of sporulation, these peaks disappeared and mass ions associated with the spores such as 3,460, 12,871, 16,022 Daltons were evident.
Consequently, a primary harvest time of biomass for successful identification was established at eight hours which correlated with a mean identification score of 2.293 (4 samples). This translates to an improvement of 44.7% on the identification score compared to an overnight analysis and therefore significantly more confidence inon the result. Several variable parameters such as culture in broth, alternative protein extraction procedures or direct spotting of cells onto the target plate did not improve identification scores.
To test the reproducibility of the results, the study was expanded to encompass a second cohort of B. subtilis strains including 5 clinical isolates and 5 new samples sourced from NCTC. MALDI-TOF analysis of the NCTC strains according to the new protocol successfully reproduced the higher confidence level (>2) of classification after 8 hours of growth, but the clinical isolates, ostensibly catalogued as B. subtilis were predominantly classified as the genetically very similar Bacillus mojavensis by the MALDI-TOF Biotyper’s 3.1 identification and classification software. This confounding development resulted in a decision to expand the study to encompass the creation of a classification model to delineate between the two species.
A typical mass spectrum of any bacterial isolate produces a mixture of major and minor mass ions. In the algorithms used in various software packages, these are sufficient to derive the identification of a species. In general the very small mass peaks have been ignored because they are often difficult to discern from background noise. However, it is becoming evident that these minor peaks are reproducible and new software is being introduced that utilises this information for typing particular groups of bacteria [13]. Here, we employed one such software package, ClinProTools 3.0 (Bruker Daltonics Ltd) to investigate its potential to delineate between the genetically closely related species, Bacillus subtilis and Bacillus mojavensis. Multivariate statistics is often used to analyse mass spectra of bacterial samples [4] and of these, principal component analysis (PCA) is the most commonly used multivariate data analysis tool and is part of the ClinProTools software. PCA is used to form new variables (i.e., principal components) which are linear combinations of the original variables (i.e., m/z values in mass spectra obtained here). The principal components account for the differences among spectra and showed the direct bacterial profiling of Bacillus subtilis and Bacillus mojavensis, using MALDI-MS. Cluster analysis grouped mass spectra into clusters by measuring similarities between spectra. The method successfully separated the mass spectra collected from the two groups that corresponded to the Bacillus subtilis and Bacillus mojavensis, isolates.
Even though this study is based upon a small dataset to investigate the potential of this approach, and an expansion of the work would require the validation of the Bacillus species by genomic methods before the creation of a final classification model, these results demonstrate unequivocally that the minor mass ions, that are an integral part of the high abundance/conserved protein mass spectrum of MALDI-TOF MS, are sufficiently robust to be used to distinguish genetically closely related species. These peaks appear to emanate from post-translational modifications but it is not possible to confirm using a linear MALDI-TOF mass spectrometer. Thus, studies involving top-down proteomics (unpublished) indicate that these are the likely sources of the mass ion peaks. Newly developing software tools (Figure 5) may be used to maximise the data produced by such a relatively simple instrument. In particular, new forms of tandem mass spectrometry are making such analyses more accessible and opening up new vistas in the field of microbial pathogenicity. This has been elegantly demonstrated for Neisseria meningitidis by Julia Chamot-Roote [15] or for studying the effects of nutritional depletion on cells as demonstrated for Salmonella [16]. However, these instruments are unlikely to be employed in a clinical laboratory in the near future and for the foreseeable future MALDI-TOF MS is likely to dominate the clinical laboratory where it will mostly likely be used with whole genome sequencing for charactering such closely related species. However, as more software such as ClinProTools become more manageable in the clinical laboratory, it is likely that microbial identification by MALDI-TOF will enter a new phase, with a significantly higher level of resolution than is currently achievable.
Figure 5: Principal component analysis using ClinProTools shows the unambiguous separation of the genetically closely related Bacillus subtilis (green cluster); top right and Bacillus mojavensis (red cluster); bottom left.The NCTC B. subtilis strains were analysed 12 times each from five wells on a target plate. This was repeated using the isolates identified as B. mojavensis. This data was imported into Clin Pro tools and it created two scatter populations for each species. Referred to as ‘Class 1, B. subtilis ONLY’ and ‘Class 2, B. mojavensis ONLY’. These data sets were used to create a model for the delineation of the two species using the classify option in Clin Pro Tools which reports the data as falling into class 1, 2 or neither.

References

Track Your Manuscript