Anal. Chem. 2008, 80, 7097–7104
Detection and Identification of Coxiella burnetii Based on the Mass Spectrometric Analyses of the Extracted Proteins Lenka Hernychova,† Rudolf Toman,‡ Fedor Ciampor,‡ Martin Hubalek,† Jana Vackova,† Ales Macela,† and Ludovit Skultety*,‡ Institute of Molecular Pathology, Faculty of Military Health Sciences, University of Defense, 500 01 Hradec Kralove, Czech Republic, and Laboratory for Diagnosis and Prevention of Rickettsial and Chlamydial Infections, Institute of Virology, Slovak Academy of Sciences, 845 05 Bratislava, Slovak Republic Rapid and reliable detection, identification, and typing of bacterial species are necessary in response to natural or terrorist-caused outbreaks of infectious diseases and play crucial roles in diagnosis and efficient treatment. We report here two proteomic approaches with a high potential in the detection and identification of Coxiella burnetii, the causative agent of Q fever. The first of them starts with the acetonitrile (ACN) and trichloroacetic acid extractions of inactivated C. burnetii cells followed by the detection of extracted molecules and ions derived from the inactivated cells by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. In the second approach, identification of the proteins extracted by ACN is accomplished after enzymatic digestion by electrospray tandem mass spectrometry coupled to a nanoscale ultraperformance liquid chromatography (LC-MS/MS). In order to observe morphological differences on the surface structures upon extraction, the inactivated and treated cells of the bacterium were examined by electron microscopy. The LC-MS/MS approach has allowed identification of 20 proteins in the ACN extracts of C. burnetii strain RSA 493 that were observed in more than 3 out of 10 experiments. Various mass spectrometric approaches became popular in the last several years for rapid identification of intact microorganisms1-4 or viruses.5 Matrix-assisted laser desorption/ionization time-offlight (MALDI-TOF) mass spectrometry (MS) of intact bacterial cells provides characteristic and reproducible mass spectral * To whom correspondence should be addressed. Phone: (421) 259 302406. Fax: (421) 254 774284. E-mail:
[email protected]. † Institute of Molecular Pathology. ‡ Institute of Virology. (1) Claydon, M. A.; Davey, S. N.; Edward-Jones, V.; Gordon, D. B. J. Microbiol. Methods 1996, 14, 1584–1586. (2) Holland, R. D.; Wilkes, J. G.; Rafii, F.; Sutherland, J. B.; Persons, C. C.; Voorhees, K. J.; Lay, J. O. Rapid Commun. Mass Spectrom. 1996, 10, 1227– 1232. (3) Krishnamurthy, T.; Ross, P. L.; Rajamani, U. Rapid Commun. Mass Spectrom. 1996, 10, 883–888. (4) Krishnamurthy, T.; Ross, P. L. Rapid Commun. Mass Spectrom. 1996, 10, 1992–1996. (5) Yao, Z.-P.; Demirev, P. A.; Fenselau, C. Anal. Chem. 2002, 74, 2529–2534. 10.1021/ac800788k CCC: $40.75 2008 American Chemical Society Published on Web 08/16/2008
fingerprints containing unique biomarker profiles6-8 that might be exploited to improve identification and typing of bacterial strains. In order to characterize these proteins in a more detail, tandem mass spectrometry coupled to nanoscale ultraperformance liquid chromatography (LC-MS/MS) has been employed.9-14 Many different strategies for analyses of intact bacterial cells were published.7,9,15 Most of them are based on bacterial mass fingerprints obtained by linear mode MALDI-TOF MS accompanied by the different biostatistical analyses7,17-23 or phylogenetic grouping.24 For security reasons, a cell free acetonitrile (ACN) extraction procedure for MS analysis of biological warfare (BW) agents has been developed.6,8,16,17 In addition, countless numbers of various software products and databases were built up for rapid screening and characterization of bacteria. They may employ an (6) Hathout, Y.; Demirev, P. A.; Ho, Y.-P.; Bundy, J. L.; Ryzhov, V.; Sapp, L.; Stutler, J.; Jackman, J.; Fenselau, C. Appl. Environ. Microbiol. 1999, 65, 4313–4319. (7) Bright, J. J.; Claydon, M. A.; Soufian, M.; Gordon, D. B. J. Microbiol. Methods 2002, 48, 127–138. (8) Shaw, E. I.; Moura, H.; Woolfitt, A. R.; Ospina, M.; Thompson, H. A.; Barr, J. R. Anal. Chem. 2004, 76, 4017–4022. (9) Goodacre, R.; Heald, J. K.; Kell, D. B. FEMS Microbiol. Lett. 1999, 176, 17–24. (10) Stapels, M. D.; Cho, J.Ch.; Giovannoni, S. J.; Barofsky, D. F. J. Biomol. Tech. 2004, 15, 191–198. (11) Demirev, P. A.; Ramirez, J.; Fenselau, C. Anal. Chem. 2001, 73, 5725– 5731. (12) Lo, A. A.; Hu, A.; Ho, Y. P. J. Mass Spectrom. 2006, 41, 1049–1060. (13) Garza, S.; Moini, M. Anal. Chem. 2006, 78, 7309–7316. (14) Marrie, T. J.; Raoult, D. Int. J. Antimicrob. Agents 1997, 8, 145–161. (15) Demirev, P. A.; Feldman, A. B.; Kowalski, P.; Lin, J. S. Anal. Chem. 2005, 77, 7455–7461. (16) Fenselau, C.; Demirev, P. A. Mass Spectrom. Rev. 2001, 20, 157–171. (17) Demirev, P. A.; Ho, Y. P.; Ryzhov, V.; Fenselau, C. Anal. Chem. 1999, 71, 2732–2738. (18) Arnold, R. J.; Reilly, J. P. Rapid Commun. Mass Spectrom. 1998, 12, 630– 636. (19) Bright, J. J.; Claydon, M. A.; Soufian, M.; Gordon, D. B. J. Microbiol. Methods 2002, 48, 127–138. (20) Keys, C. J.; Dare, D. J.; Sutton, H.; Wells, G.; Lunt, M.; McKenna, T.; McDowall, M.; Shah, H. N. Infect. Genet. Evol. 2004, 4, 221–242. (21) Hettick, J. M.; Kashon, M. L.; Slaven, J. E.; Ma, Y.; Simpson, J. P.; Siegel, P. D.; Mazurek, G. N.; Weissman, D. N. Proteomics 2006, 6, 6416–6425. (22) Jarman, K. H.; Cebula, S. T.; Saenz, A. J.; Petersen, C. E.; Valentine, N. B.; Kingsley, M. T.; Wahl, K. L. Anal. Chem. 2000, 72, 1217–1223. (23) Pineda, F. J.; Lin, J. S.; Fenselau, C.; Demirev, P. A. Anal. Chem. 2000, 72, 3739–3744. (24) Stackebrandt, E.; Pauker, O.; Erhard, M. Curr. Microbiol. 2005, 50, 71– 77.
Analytical Chemistry, Vol. 80, No. 18, September 15, 2008
7097
improved data analysis25 that generates a database of microorganisms which contain biomarker masses derived from the ribosomal protein sequences and N-terminal Met losses.25 Coxiella burnetii, a short pleomorphic, intracellular parasite of eukaryotic cells, is the etiological agent of Q fever (Query fever). The disease is a widespread zoonosis and causes several outbreaks within the world each year.26 The bacterium is extremely resistant to harsh environmental conditions due to spore formation, it readily becomes airborne, and it is highly infectious for humans.27,28 For these properties, it is on the list of BW agents in “Category B”.29 Human infections arise as a result of contact with infected farm animals such as cattle, sheep, and goats, although pet animals especially cats have also been implicated particularly in urban outbreaks.26,27 In most cases, the infection follows inhalation of aerosols derived from the excretions and secretions of infected animals.30 Ingestion of C. burnetii (consumption of contaminated dairy products) is considered a rare alternative for acquiring infection.31 The acute form of Q fever is characterized as a flulike illness or atypical pneumonia, or less frequently as granulomatous hepatitis with a significant incidence of neurological complications.26 Persistent infections may lead to a chronic form of the disease.32 Q fever endocarditis is the most frequent clinical manifestation of this form of the disease. It is well-known that the proteins located at the surface of C. burnetii represent the major immunoreactive antigens in the serological diagnosis of Q fever. In addition, it has been found that an LPS-protein complex extracted from the bacterial cells with trichloroacetic acid (TCA) should contain immunoprotective substances as it represents an effective experimental vaccine against human Q fever.33,34 In spite of these facts, there is a lack of information on both composition and function of the proteins located at the surface of C. burnetii. Our investigations have focused on reliable detection and identification of C. burnetii employing MS analyses. In order to obtain specific fingerprints preferentially of the surface-exposed molecules of C. burnetii, we prepared two cell-free extracts of the bacterium and subjected them together with the inactivated bacterium to the MALDI-TOF-based MS analysis. The fingerprints of the linear MS spectra revealed some characteristic peaks that were used for initiation of an integrated database as an approach for the tentative recognition of C. burnetii strains/isolates. Moreover, the proteins extracted from the bacterium were identified in the tryptic digest by LC-MS/MS technique. MATERIALS AND METHODS Materials. Sinapinic acid, 2, 5-dihydroxybenzoic acid, Protein calibration mix2, and Peptide calibration mix1 were purchased (25) Pineda, F. J.; Antoine, M. D.; Demirev, P. A.; Feldman, A. B.; Jackman, J.; Longenecker, M.; Lin, J. S. Anal. Chem. 2003, 75, 3817–3822. (26) Kazar, J. Ann. N. Y. Acad. Sci. 2005, 1063, 105–114. (27) Marrie, T. J.; Raoult, D. Int. J. Antimicrob. Agents 1997, 8, 145–161. (28) Fournier, P. E.; Marrie, T. J.; Raoult, D. J. Clin. Microbiol. 1998, 36, 1823– 1834. (29) Madariaga, M. G.; Rezai, K.; Trenholme, G.; Weinstein, R. A. Lancet 2003, 3, 709–721. (30) Babudieri, B. Adv. Vet. Sci. 1959, 5, 81. (31) Fishbein, D. B.; Raoult, D. Am. J. Trop. Med. Hyg. 1992, 47, 35–40. (32) Maurin, M.; Raoult, D. Clin. Microbiol. Rev. 1999, 12, 518–553. (33) Brezina, R.; Urvolgyi, J. Acta Virol. 1961, 5, 193–195. (34) Brezina, R.; Schramek, S.; Kazar, J.; Urvolgyi, J. Acta Virol. 1974, 18, 269– 271.
7098
Analytical Chemistry, Vol. 80, No. 18, September 15, 2008
from LaserBio Laboratories (Sophia-Antipolis Cedex, France). Trifluoroacetic acid (TFA), formic acid (FA), TCA, and Tris-HCl were obtained from Sigma-Aldrich (St. Louis, MO). Trypsin was purchased from Promega (Madison, WI), and Microcon centrifugal filter device YM-10 was obtained from Millipore (Bedford, MA). ACN (LiChrosolv quality), water (LiChrosolv quality), thiomersal, and dithiotreithol (DTT) were furnished by Merck (Darmstadt, Germany). Guanidine hydrochloride and EDTA were purchased from Janssen (Geel, Belgium). Sodium 2-iodoacetate and NH4HCO3 were obtained from Fluka (Buchs, Switzerland). Phenol, NaCl, Na2HPO4, and diethyl ether were received from the local distributor, Slavus. Microorganism. C. burnetii strain RSA 493 (Nine Mile), serologically in the virulent phase I, was obtained from the WHO Collaborating Centre for Rickettsial Reference and Research at the Institute of Virology, Slovak Academy of Sciences, Bratislava, Slovak Republic. Embryonated, antibiotic-free, and pathogen-free eggs were inoculated with C. burnetii and incubated for 9 days at 37 °C and 98% humidity. The eggs were continuously rocked. When ∼50% embryo died off, the yolk sac of each egg was harvested. The yolk sac mass was homogenized in 2 M aqueous solution of NaCl containing 0.45% phenol and allowed to stand at 5 °C for 7-10 days. These conditions destroyed C. burnetii viability. This was confirmed by the fact that, after this treatment, no propagation of the bacterium occurred after its inoculation and cultivation. The C. burnetii cells were purified by series of differential centrifugation and ether treatments.35 The cells were stored in phosphate-buffered saline with thiomersal as a preservative at 4 °C. All these steps were performed in a biological safety level 3 facility following protocols wherein all manipulations are carried out in biological safety cabinets. Sample Preparation. The C. burnetii cells were extracted either by ACN or TCA solutions. Briefly, the cellular material (5 mg) was washed two times by deionized water (1 mL) and centrifuged (12 000 rpm, 20 min, 4 °C). The pellets were resuspended in 300 µL of deionized water containing 0.5% (final concentration) of TFA. ACN (700 µL) was continually added to the suspension. After extraction by vigorous vortexing (5 min), the cells were sedimented by centrifugation (12 000 rpm, 15 min, 10 °C) and the supernatant was used for further analyses. A cell-free TCA extract (TCA extract) according to Lukacova et al.36 was prepared by the following modification. The pellets of the washed cellular material (5 mg) were resuspended in 1 mL of deionized water, and an aliquot of 20% TCA was added. The mixture was stirred at 0 °C for 45 min and neutralized with 5 M NaOH. The cells were sedimented by centrifugation (12 000 rpm, 30 min, 4 °C), and the supernatant was dialyzed against deionized water at 4 °C for 2 days. Both cell-free extracts were concentrated on a Speed Vac (Eppendorf) until the volume was reduced to 5 µL. Purification of the ACN Extract. The cell-free ACN extract (ACN extract) was purified according to Wessel and Fluge37 with the following modification. Briefly, the freshly prepared ACN extract was evaporated with a stream of nitrogen to 0.2 mL, and (35) Skultety, L.; Toman, R.; Patoprsty, V. Carbohydr. Polym. 1998, 35, 189– 194. (36) Lukacova, M.; Brezina, R.; Schramek, S.; Pastorek, J. Acta Virol. 1989, 33, 75–80. (37) Wessel, D.; Flu ¨ gge, U. I. Anal. Biochem. 1984, 138, 141–143.
Figure 1. Proteomic workflow.
an aliquot (0.8 mL) of methanol was added. The mixture was vortexed and mixed with chloroform (0.2 mL) and phased by centrifugation (1 min, 5000 rpm) after addition of 0.6 mL of deionized water. The upper phase was carefully removed and discarded. Methanol (0.6 mL) was added to the remaining material and mixed carefully. The precipitated proteins were centrifuged for 1 min at 5000 rpm. The supernatant was discarded, and the protein pellet was resuspended in 50 µL of 70% ACN. Electron Microscopy. The C. burnetii cells and specimens after extraction of the surface antigens either by ACN or TCA were absorbed on electron microscopic grids coated with a Formvar-carbon film. The samples were negatively stained with 2% phosphotungstic acid, pH 7.5 in distilled water, and examined in JEOL JEM 1200Ex electron microscope at 80 kV. Proteomics Workflow. After propagation and purification of C. burnetii strain RSA 493 according to the standard protocol,35 the cell-free extracts of the pure culture were prepared either with a solution of TCA or ACN containing TFA with the aim to avoid working with the BSL3 agent in the proteomics laboratory and to make the sample not as complex. The inactivated bacteria and the residual cells obtained after extraction were examined by electron microscopy in order to observe morphological differences at the surface structures upon extraction. All samples were analyzed by MALDI-TOF MS in a linear mode for rapid screening and characterization of the bacterial surface. The extract that gave homogeneous and reproducible MALDI-TOF MS profiles was chosen for further MS/MS studies. The extracted, surfaceexposed proteins were identified after enzymatic digestion by LC-MS/MS (Figure 1). Enzymatic Digestion. The C. burnetii ACN extract was reduced, alkylated, and digested on a Microcon centrifugal device (YM-10, MWCO 10000; Millipore) as follows. The concentrated cell-free extract (5 µL) was diluted by denaturing freshly prepared buffer (300 µL) containing 6 M guanidine hydrochloride, 100 mM Tris-HCl, and 5 mM EDTA. The volume of the mixture was reduced on a Microcon filter (12 000 rpm, 4 °C, 1 h) to ∼50 µL. A DTT (100 µL of 100 mM) solution in 100 mM ammonium hydrogen carbonate was added, mixed, and incubated for 1 h at 56 °C. The volume was reduced again to ∼50 µL and the mixture alkylated with iodoacetamide (100 µL of 300 mM) in the same solvent for 30 min at 25 °C in dark. The sample was concentrated and washed twice with 200 µL of cleavage buffer (50 mM ammonium hydrogen carbonate, 5% ACN). Sequence grade trypsin (0.2 µg) in cleavage buffer (100 µL) was added, and the proteins were digested in a thermomixer at 37 °C overnight. The
peptides were recovered by spinning at 12 000 rpm for 1 h at 15 °C into clean vials followed by washing the filter unit with 50 µL of 0.1% FA containing 30% ACN and 50 µL of 60% ACN. The peptides were vaccuum-dried and reconstituted in 30 µL of sample buffer containing 0.1% FA and 2% ACN. MALDI-TOF MS. A concentrated ACN extract (1 µL) was spotted on a MALDI target sample plate with a hydrophobic surface (2 × 96 well positions) and allowed to air-dry at room temperature. Matrix solution (1 µL) containing SA (10 mg/mL) in aqueous 30% ACN with 0.5% TFA was dropped onto each sample spot. The sample plate was inserted into the MALDI-TOF instrument after solvent evaporation. Mass spectra were acquired using a Voyager DE STR MALDI-TOF (Applied Biosystems, Framingham, MA) mass spectrometer equipped with a delayed extraction and UV nitrogen laser (337 nm, 3-ns pulse width). Analyses were performed in a linear positive ion mode at accelerating voltage 25 kV, 93% grid voltage, 0.15% guide wire, extraction delay time 320 ns, and low mass gate 1000 m/z. Mass range was set from 1000 to 25 000 m/z. The instrument was calibrated before each analysis with a mixture of both Protein calibration mix1 and Protein calibration mix2. The mass accuracy for each standard was within 500 ppm of the corresponding average molecular weight. Each mass spectrum was obtained by averaging 200 laser shots. The Data Explorer program (Matrix Science) was used to view and process data files from the instrument. Processing parameters were as follows: advanced baseline correction 32, 0.5, 0.1; noise filter/smooth, Gaussian smooth at filter width 15. LC-MS/MS. Protein identification in the ACN extract was performed on a tandem mass spectrometer coupled to a nanoscale ultraperformance liquid chromatography (UPLC). The extracted peptide mixture was separated on a nanoAcquity UPLC system (Waters) using precolumn concentration (Atlantis dC18, 5-µm NanoEase Trap column, Waters) and gradient elution on an analytical nanocolumn (Atlantis dC18 75 µm × 150 mm, 3-µm NanoEase) at a flow rate 250 nL/min. The gradient consisted of a linear increase from 3 to 50% ACN containing an aqueous solution of 0.1% FA in 50 min. The column was connected to the PicoTip emitters (New Objective) mounted into the nanospray of the Q-TOF Premier (Waters). The data acquisition was performed in data-dependent manner for the time of the separation collecting up to three MSMS events at the same time. Data were processed by ProteinLynx Global Server v. 2.2 (Waters) that provided background subtraction (polynomial order 5 and threshold 35%), smoothing (Savitzky-Golay, twice, over three channels), and Analytical Chemistry, Vol. 80, No. 18, September 15, 2008
7099
Figure 2. Electron microscopy images of the inactivated C. burnetii (A) and residual cells obtained after extraction either by 70% acetonitrile (B) or 10% trichloroacetic acid (C). Magnification, 20000×.
centroiding (top, 80%, minimal peak width at half-height 4). Resulting data were searched against C. burnetii (National Center for Biotechnology Information) and all species nonredundant databases (Swiss-Prot, Swiss Institute of Bioinformatics) with the following criteria: fixed carbamidomethylation of Cys, variable Met oxidation, tryptic fragments with 1 miss cleavage, peptide mass tolerance 100 ppm, and fragment mass tolerance 0.1 Da. The results were validated by the identification of three or more consecutive fragment ions from the same series. RESULTS AND DISCUSSION Electron Microscopy Study. The observations (Figure 2) revealed no substantial morphological differences of the surface structures of both inactivated and extracted C. burnetii cells. Disintegration into the small vesicular structures was not observed. Nevertheless, some very distinct differences with regard to the electronoptic density character of the cells were noticed. The intact C. burnetii cells were more compact compared with the extracted ones. The cellular material of extracted cells was concentrated into the central area of the cells; their periplasmic space became larger and less dense, which might be due to the extraction of surface complexes (Figure 2B and C) MALDI-TOF MS Analysis. The linear MALDI-TOF mass spectra of pure inactivated bacterial cells and the ACN and TCA cell-free extracts from quintuple preparations were acquired in a positive ion mode. Mass spectral data were normalized, baselinecorrected, and denoised. The extraction procedures were optimized with respect to the signal-to-noise ratios, reproducibility, and number of proteins extracted. Table 1 lists the m/z values and the corresponding intensities of mass spectral peaks clearly distinct in the spectra. Only those values of peaks are shown in which an average intensity of quintuple preparations was higher than 12 and 10% of the base peak in the ranges of 3000-8000 and 8000-21 000 m/z, respectively. Peaks that unfit these parameters are displayed only if equivalent peaks of at least one method of preparation accommodate the given criteria. The data presented in Table 1 are in a good agreement with the work of Shaw et al.8 as the spectral markers they reported for C. burnetii strain Nine Mile in phase I at m/z of 5475, 7685, 7964, 8175, 10 006, 10 502, 11 133, 15 273, and 16 382 have been recognized and confirmed. It is well-known that the reproducibility of MALDI-TOF MS spectra, particularly variations in signal intensity, are an important concern that can adversely affect mass spectral fingerprinting approaches to microbial identification. For these reasons, Table 1 details also the variability encountered in the C. burnetii MALDITOF MS data following base peak normalization, and the mean intensity differences observed between the sets of samples by displaying their corresponding standard deviation. 7100
Analytical Chemistry, Vol. 80, No. 18, September 15, 2008
From the results obtained we concluded that the extraction procedure using ACN gave a spectral profile similar to that of the inactivated C. burnetii cells, and thus, it represents the optimum balance between maximizing the extraction of proteins and maintaining both homogeneity and reproducibility of the MALDI-TOF MS spectra. Therefore, the ACN extract was chosen for further MS/MS studies. LC-MS/MS Analysis. Since the genome of C. burnetii strain RSA 493 is completely sequenced,38 there was no obstruction for the subsequent proteomic studies toward identifying the proteins that were extracted from C. burnetii. Thus, proteins of the ACN extract of the bacterium were cleaved with trypsin, and the resulting peptides were separated by reversed-phase chromatography and analyzed by MS/MS. The experiment was repeated with 10 freshly prepared ACN extracts. Based on the MS/MS data, 20 distinct proteins were identified in a more than three experiments (Table 2) including 4 low-mass (
