improvement in staphylococcus and bacillus strain differentiation by matrix-assisted laser...

Research Article

Received: 28 November 2013 Revised: 22 May 2014 Accepted: 18 June 2014 Published online in Wiley Online Library

Rapid Commun. Mass Spectrom. 2014, 28, 1855–1861

Improvement in Staphylococcus andBacillus strain differentiationbymatrix-assisted laser desorption/ionization time-of-flight massspectrometry profiling by using microwave-assisted enzymaticdigestion

Tereza Balážová1,2, Ondrej Šedo1,3*, Polonca Štefanić4, Ines Mandić-Mulec4, Michiel Vos5

and Zbyněk Zdráhal1,31Research Group Proteomics, CEITEC – Central European Institute of Technology, Masaryk University, Kamenice 5, 62500Brno, Czech Republic2Department of Microbiology, Faculty of Science, Masaryk University, Kamenice 5, 62500 Brno, Czech Republic3National Centre for Biomolecular Research, Faculty of Science, Masaryk University, Kamenice 5, 62500 Brno, Czech Republic4Department of Food Science and Technology, Biotechnical Faculty, University of Ljubljana, Vecna pot 111, 1000 Ljubljana,Slovenia5European Centre for Environment and Human Health, The University of Exeter Medical School, University of Exeter, TR10 9EZPenryn, United Kingdom

RATIONALE: Distinguishing between individual bacterial strains below the species level is a challenge to matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) bacterial profiling. We proposea quick method for improving strain differentiation of two Staphylococcus and one Bacillus species.METHODS: An alternative procedure to the extraction protocol recommended by Bruker Daltonics was developed.Ethanol-sterilized cells of six S. aureus and six S. haemolyticus strains were digested by trypsin using 2-min microwaveirradiation and were then analyzed. Twenty-eight strains belonging to two ecotypes of B. subtilis were subjected to thesame procedure to extend the scope of the method.RESULTS: S. aureus and S. haemolyticus strains, only partially distinguishable by the standard sample preparationprocedure, were subjected to microwave-assisted tryptic digestion. The repeatability of the procedure was checked inthree experiments accomplished at weekly intervals. Clear distinction of the strains was achieved by cluster analysis.The differentiation of B. subtilis ecotypes was also improved significantly by the digestion method. The discriminatorypower of the novel method was supported by an increase in the number of strain-specific peaks, as compared to thestandard method.CONCLUSIONS: The method modulates the discriminatory power of MALDI-TOF MS profiling. The differentiation of aset of S. aureus, S. haemolyticus and B. subtilis strains was improved significantly after microwave-accelerated trypticdigestion of the cellular material. Copyright © 2014 John Wiley & Sons, Ltd.

(wileyonlinelibrary.com) DOI: 10.1002/rcm.6966

Matrix-assisted laser desorption/ionization time-of-flight massspectrometry (MALDI-TOF MS) profiling is widely used inmicrobiology for identification of bacteria on the basis of theirpeptide/protein profiles. However, the discriminatory powerof MALDI-TOF MS is often not sufficient to distinguishbetween closely related strains.[1] Demands for rapid andreliable identification of pathogens have brought a number ofalternatives to the classical MALDI-TOF MS protein profilingsample preparation protocols.[2] Identification can be alsoaccomplished by the profiling of bacterial lipids using an onlinedroplet-deposited MALDI-TOF MS, as shown with Escherichia

* Correspondence to: O. Šedo, Research Group Proteomics,CEITEC –Central European Institute of Technology,MasarykUniversity, Kamenice 5, 62500 Brno, Czech Republic.E-mail: [email protected]

Rapid Commun. Mass Spectrom. 2014, 28, 1855–1861

18

coli, Salmonella, and Bacillus subtilis.[3] Extensive and advancedcomputational analysis may also contribute to improvementsin the discriminatory power of the method. Improvements inthe distinction between very similar mass spectra of closelyrelated Erwinia subspecies were achieved by the applicationof a weighted pattern matching algorithm.[4] Another solutionfor distinguishing between closely related strains wasintroduced by Warscheid and Fenselau, who proposed thecombination of in situ proteolytic digestion of a limited set ofproteins with MS-based microsequencing.[5] They were ableto distinguish between closely related Bacillus species, but theprocedure required a humidified chamber to prevent solventevaporation during digestion.

The digestion process can be accelerated withmicrowave irradiation, which has been used in organicsynthesis,[6] enzymatic digestion of proteins,[7–11] and acidhydrolysis with HCl[12,13] and TFA.[14] Microwave digestion of

Copyright © 2014 John Wiley & Sons, Ltd.

55

T. Balážová et al.

1856

bacterial cells, in conjunction with MALDI-TOF MS profiling,has mainly been tested in connection with the use ofmagnetic beads.[15]

We assessed a novel method for strain distinction based onrapid tryptic digestion of cellular material using microwaveirradiation and MALDI-TOF MS protein profiling. This newmethod was tested using strains of Staphylococcus aureus andStaphylococcus haemolyticus and also on a panel of closelyrelated B. subtilis strains (’ecotypes’).

EXPERIMENTAL

Strains and culture conditions

Six Staphylococcus aureus (SA) strains (CCM 885T, CCM 1484,CCM 2022, CCM 2107, CCM 2351, and CCM 2352) and sixStaphylococcus haemolyticus (SH) strains (CCM 2737T, CCM1798, CCM 2603, CCM 2574, CCM 2729, and CCM 4832)were obtained from the Czech Collection of Microorganisms(CCM, Brno, Czech Republic). The strains were aerobicallycultured in tryptone soya agar (TSA; Oxoid, Basingstoke,UK) at 37 °C for 24 h. Bacillus subtilis (BS) strains included13 representatives of ecotype PE10 (PS-11, PS-13, PS-14,PS-18, PS-30, PS-31, PS-51, PS-65, PS-68, PS-96, PS-168,PS-210, and PS-233) and 15 representatives of ecotypePE32 (PS-24, PS-25, PS-52, PS-53, PS-55, PS-93, PS-95, PS-108,PS-119, PS-130, PS-131, PS-149, PS-160, PS-218, and PS-263)isolated from a sandy bank soil of the Sava River inSlovenia.[16] The strains were inoculated from freezer stock(–80 °C) onto LB agar plates and were cultured for 24 hat 37 °C prior to sample preparation. Putative ecotypesof the listed river bank strains were demarcated byEcotype Simulation (ES) as described elsewhere.[17] Therates of ecotype formation and of periodic selectionper nucleotide substitution were based on threeconcatenated sequenced genes (gyrA, rpoB, dnaJ) and thedistinct ecotypes (PE10 and PE32) demarcated by ESrepresent two separate clades putatively adapted to adifferent niche.[18]

Figure 1. Comparison of mass spectra o(A) and digested (B) Staphylococcus aureus

wileyonlinelibrary.com/journal/rcm Copyright © 2014 John Wile

Sample preparation

Approximately 5–10 mg of the bacterial culture from TSA/LBagar plates was suspended in 1 mL ddH2O, vortexed andcentrifuged at 1.365g for 2 min after which the supernatantwas discarded. The pellet was suspended in 1.2 mL of 75%ethanol, vortexed and centrifuged at 12 100g for 2 min. Pelletedcells were resuspended in 5–10 μL of ddH2O depending on thesize of the pellet. A volume of 1 μL of each bacterial cellsuspension was mixed with 1.25 μL of trypsin (200 ng/μL),0.2 μL of ammonium bicarbonate (500 mM) and 7.55 μL ofddH2O in a 500 μL Eppendorf tube. This digestion mixturewas prepared in triplicate for each strain. The digestionexperimentswere primarily performed in a domesticmicrowaveoven (Eta 0205, Prague, Czech Republic); to confirm therobustness of the method, selected experiments were repeatedin two other microwave ovens (LG WAVEDOM 23 L, Seoul,South Korea and Samsung M171N, Seoul, South Korea). TheEppendorf tubes were placed on a plastic rack and heated in amicrowave oven for 2 min at a maximum power of 700W. Afterthe treatment, 0.2 μL of 5% trifluoroacetic acid (TFA) was addedto stop the digestion process, and themixture was centrifuged at12,100g for 2 min. The supernatant, in a volume of 1 μL, wasplaced onto a MALDI target plate (MSP 384 target, BrukerDaltonics, Germany) in one replicate (one well) for each of thethree digestion mixtures of respective samples. The sampleswere allowed to dry at room temperature and overlayed withthe saturated solution of α-cyano-4-hydroxycinnamic acid inACN/H2O/TFA mixture (50:47.5:2.5, v/v/v) in a volume of0.5μL and allowed to dry again. To test reproducibility, digestionand MALDI-TOF MS analysis were repeated two more times atweekly intervals for the six S. aureus strains.

In parallel, ethanol-sterilized bacterial pellets of the samestrains were processed by the standard extraction protocolrecommended by Bruker Daltonics (extraction using amixture of equal volumes of acetonitrile and 70% formicacid[19]) and the extracts were subsequently analyzed in thesame way as the digests.

α-Cyano-4-hydroxycinnamic acid was obtained fromBruker Daltonics (Leipzig, Germany), ethanol, acetonitrile,trifluoroacetic acid, formic acid and ammonium bicarbonate

btained from a standard preparationCCM 885T pellets.

y & Sons, Ltd. Rapid Commun. Mass Spectrom. 2014, 28, 1855–1861

MALDI-TOF MS differentiation of bacterial strains by microwave-assisted digestion

were obtained from Sigma-Aldrich (Steinheim, Germany),and trypsin was obtained from Promega (Mannheim,Germany). Water was prepared on a Milli-Q plus 185apparatus (Millipore, Billerica, MA, USA).

Figure 2. Dendrogram calculated on theTOF mass spectra of six Staphylococcus aurthe standard sample preparation methomarked as ’I’,’II’, and ’III’.

Figure 3. Dendrogram calculated on theTOF mass spectra of six Staphylococcus(marked as ’I’,’II’, and ’III’) at different(marked as ’D1’, ’D2’, and ’D3’) using mi

Copyright © 2014Rapid Commun. Mass Spectrom. 2014, 28, 1855–1861

MALDI-TOF MS analysis

All mass spectra were acquired by using an Ultraflex IIIinstrument (Bruker Daltonics, Germany) operated in linearpositive mode under FlexControl 3.3 software. External

basis of cluster analysis of MALDI-eus strains analyzed in triplicate usingd. Repeated analyses of strains are

basis of cluster analysis of MALDI-aureus strains analyzed in triplicatetimes separated by weekly intervalscrowave-assisted digestion.

wileyonlinelibrary.com/journal/rcmJohn Wiley & Sons, Ltd.

1857


1858

calibration of the mass spectra was performed using E. coliDH5 alpha standard. Three independent spectracomprising 1000 laser shots each were acquired from eachof the wells, where only peaks detected in all three massspectra from one well were taken into account for furtherevaluation. Mass spectra were processed using FlexAnalysis (version 3.4, Bruker Daltonics) and BioTypersoftware (version 3.0, Bruker Daltonics). The MALDI-TOFMS-based dendrogram was generated using the correlationdistance measure with the average linkage algorithm(Unweighted average distance – UPGMA).

Figure 4. Dendrogram calculated on theTOF mass spectra of six Staphylococcus haeusing microwave-assisted digestion. Repe’I’,’II’, and ’III’.

Figure 5. Comparison of mass spectra o(A) and sterilized (B) Staphylococcus aureu


RESULTS AND DISCUSSION

As seen in Fig. 1, digestion affected the nature of mass spectraobtained. In samples prepared by the standard method(undigested), signals were observed in the range from2000 to 11 000 Da, whereas, after digestion, signals werepresent across the mass range from 1000 to 8000 Da(mainly 2000–4000 Da). With the exception of a fewsignals of low intensity, entirely different signals werepresent in the mass spectra obtained from digested versusnon-digested pellets. While the mass spectra obtained

basis of cluster analysis of MALDI-molyticus strains analyzed in triplicateated analyses of strains are marked as

btained from digests of non-sterilizeds CCM 1484 pellets.



from the non-digested bacterial pellets allowed onlypartial strain differentiation (as seen on mixed clusters ofS. aureus strains shown in Fig. 2), spectra obtained byaccelerated microwave-assisted digestion enabled reliabledifferentiation between all S. aureus (Fig. 3) and S. haemolyticus(Fig. 4) strains used in our study. Differentiation wasenhanced by strain-specific signals in the mass spectra.These were found in two of the S. aureus strains – SACCM 2351 (3294 and 3597 Da) and SA CCM 2352 (2828,

Figure 6. Dendrogram calculated on theTOF mass spectra of three selected Staphyin triplicate using microwave-assisted diovens. Repeated analyses of strains are m

Figure 7. Dendrogram calculated on theTOF mass spectra of 28 Bacillus subtilisPE32) analyzed using standard sample p


3585, 6915, 7036 and 8470 Da). We also observedrepeatable differences in the relative intensity of individualsignals and the absence of some peaks that were presentin other strains. In total, seven strain-specific signals weredetected in digests of S. aureus, while only two strain-specific signals of lower intensity were found when thestandard sample preparation protocol was carried out.Increased discriminatory power in the case of S. haemolyticuswas also associated with an increase in the number of

basis of cluster analysis of MALDI-lococcus haemolyticus strains analyzedgestion in three domestic microwavearked as ’I’,’II’, and ’III’.

basis of cluster analysis of MALDI-strains of two ecotypes (PE10 and

reparation.


1859

Figure 8. Dendrogram calculated on the basis of cluster analysis of MALDI-TOF mass spectra of 28 Bacillus subtilis strains of two ecotypes (PE10 andPE32) analyzed using microwave-assisted digestion.


1860

strain-specific peaks, where 29 strain-specific signals weredetected from digested samples, whereas only six werefound in mass spectra from non-digested samples.Equivalent improvements in strain differentiation were

obtained after digestion was carried out at 37 °C for 2 hwithout the use of microwave irradiation (dendrogram notshown). We also evaluated the newly developed protocol oncells harvested directly from culture. However, ethanoltreatment prior to the enzymatic digestion was necessary toobtain suitable mass spectral quality since spectra obtainedfrom non-sterilized pellets contained a low number ofsignals above 4000 Da (Fig. 5) and did not enabledifferentiation between strains (dendrogram not shown).We also confirmed the versatility of our method on selectedstrains of S. haemolyticus (CCM 1798, CCM 2603 and CCM4832) using two other microwave ovens. Repeated analysesof these strains were divided into three main strain-specificsub-branches of the dendrogram, regardless of microwavemake and model (Fig. 6). We also checked the reliability ofthe method by blind analysis of 12 digested samples ofS. haemolyticus (two of each strain); all were correctly assignedto their respective strains with a Biotyper log(score) valuegreater than 2.5.Cluster analysis of mass spectra from the non-digested

pellets of Bacillus subtilis strains belonging to two differentecotypes did not result in reliable strain differentiation inthe dendrogram (Fig. 7). Differentiation of ecotypes PE10 andPE32 was significantly improved after microwave-assistedtrypsin digestion. Although six samples were not groupedaccording to ecotypes, most strainswere divided into twomainecotype-specific sub-branches (Fig. 8). A wide range of geneticpolymorphisms could potentially underlie the proteomicdifferences observed. For example, the comQXPA operon,which encodes the B. subtilis quorum sensing system, andregulates many genes during late growth, is polymorphic


within ecotypes.[18] Furthermore, it is highly encouraging thatour findings were supported by the detection of an ecotype-specific signal of 4087 Da found in mass spectra of most PE10ecotype strains (except samples PS-31, PS-65, and PS-210), butwas not detected in any of PE32 ecotype strains.

CONCLUSIONS

A combination of microwave-accelerated trypsin digestion,followed by MALDI-TOF MS profiling, is a methodical variantthat has the potential to improve strain differentiation, asdemonstrated on a set of Staphylococcus aureus andStaphylococcus haemolyticus strains and two Bacillus subtilisecotypes. Apart from the digestion step, all other steps in thesample preparation protocol, acquisition of mass spectra, andBiotyper software settings were equivalent to those used underthe standard extraction procedure. The advantages of thepresent approach are its rapidity and simplicity, making itaccessible to all MALDI-TOF MS profiling users.

AcknowledgementsThis work was supported by CEITEC open access project(LM2011020), funded by the Ministry of Education, Youth,and Sports of the Czech Republic under the activity ’Projectsof major infrastructures for research, development andinnovations’, by projects CEITEC (CZ.1.05/1.1.00/02.0068)and CEB (CZ.1.07/2.3.00/20.0183), by the European RegionalDevelopment Fund and the European Social FundConvergence Programme for Cornwall and the Isles of Scillyand by the research program P4-0116 financed by theSlovenian Research Agency. The English language was kindlyrevised by Prof. J. D. Brooker.



REFERENCES

[1] T. R. Sandrin, J. E. Goldstein, S. Schumaker. MALDI TOFMS profiling of bacteria at the strain level: A review. MassSpectrom. Rev. 2013, 32, 188.

[2] O. Šedo, I. Sedláček, Z. Zdráhal. Sample preparationmethods for MALDI-MS profiling of bacteria. MassSpectrom. Rev. 2011, 30, 417.

[3] X. Shu, Y. Li, M. Liang, B. Yang, C. Liu, Y. Wang, J. Shu.Rapid lipid profiling of bacteria by online MALDI-TOFmass spectrometry. Int. J. Mass. Spectrom. 2012, 321, 71.

[4] S. Sauer, A. Freiwald, T. Maier, M. Kube, R. Reinhardt,M. Kostrzewa, K. Geider. Classification and identificationof bacteria by mass spectrometry and computationalanalysis. PLoS ONE 2008, e2843.

[5] B.Warscheid, C. Fenselau. A targeted proteomics approach tothe rapid identification of bacterial cell mixtures by matrix-assisted laser desorption/ionization mass spectrometry.Proteomics 2004, 4, 2877

[6] P. Lidström, J. Tierney, B. Wathey, J. Westman. Microwaveassisted organic synthesis - a review.Tetrahedron 2001, 57, 9225.

[7] H. F. Juan, S. C. Chang, H. C. Huang, S. T. Chen. A newapplication of microwave technology to proteomics.Proteomics 2005, 5, 840.

[8] J. R. Lill, E. S. Ingle, P. S. Liu, V. Pham, W. N. Sandoval.Microwave-assisted proteomics. Mass Spectrom. Rev. 2007,26, 657.

[9] B.N. Pramanik, U. A.Mirza, Y.H. Ing, Y.H. Liu, P. L. Bartner,P. C. Weber, M. K. Bose. Microwave-enhanced enzymereaction for protein mapping by mass spectrometry: A newapproach to protein digestion in minutes. Protein Sci. 2002,11, 2676.

[10] W. Sun, S. Gao, L. Wang, Y. Chen, S. Wu, X. Wang,D. Zheng, Y. Gao. Microwave-assisted protein preparation


and enzymatic digestion in proteomics. Mol. Cell. Proteomics2006, 5, 769.

[11] S. Swatkoski, P. Gutierrez, C.Wynne, A. Petrov, J. D. Dinman,N. Edwards, C. Fenselau. Evaluation of microwave-accelerated residue-specific acid cleavage for proteomicapplications. Proteome Res. 2008, 7, 579.

[12] S. T. Chen, S. H. Chiou, K. T. Wang. Enhancement ofchemical reactions by microwave irradiation. J. Chin. Chem.Soc. 1991, 38, 85.

[13] H. Y. Zhong, Y. Zhang, Z. H. Wen, L. Li. Protein sequencingby mass analysis of polypeptide ladders after controlledprotein hydrolysis. Nat. Biotechnol. 2004, 22, 1291.

[14] H. Zhong, S. L. Marcus, L. Li. Microwave-assisted acidhydrolysis of proteins combined with liquid chromatographyMALDI MS/MS for protein identification. J. Am. Soc. MassSpectrom. 2005, 16, 471.

[15] W. J. Chen, P. J. Tsai, Y. C. Chen. Functional nanoparticle-based proteomic strategies for characterization of pathogenicbacteria. Anal. Chem. 2008, 80, 9612.

[16] P. Stefanic, I. Mandic-Mulec. Social interactions anddistribution of Bacillus subtilis pherotypes at microscale.J. Bacteriol. 2009, 191, 1756.

[17] A. Koeppel, E. B. Perry, J. Sikorski, D. Krizanc, A. Warner,D. M.Ward,A. P. Rooney, E. Brambilla,N.Connor, R.M. Ratcliff,E. Nevo, F. M. Cohan. Identifying the fundamental units ofbacterial diversity:Aparadigm shift to incorporate ecology intobacterial systematics. Proc. Natl. Acad. Sci. USA 2008, 105, 2504.

[18] P. Stefanic, F. Decorosi, C. Viti, J. Petito, F. M. Cohan,I. Mandic-Mulec. The quorum sensing diversity within andbetween ecotypes of Bacillus subtilis. Environ. Microbiol. 2012,14, 1378.

[19] A. Freiwald, S. Sauer. Phylogenetic classificationand identificationof bacteria by mass spectrometry. Nat. Protocols 2009,4, 733.


1861

improvement in staphylococcus and bacillus strain differentiation by matrix-assisted laser...

Documents