Comparative Proteome Analysis of Brucella melitensis Vaccine Strain Rev 1 and a Virulent Strain, 16M

Abstract

The genus Brucella consists of bacterial pathogens that cause brucellosis, a major zoonotic disease characterized by undulant fever and neurological disorders in humans. Among the different Brucella species, Brucella melitensis is considered the most virulent. Despite successful use in animals, the vaccine strains remain infectious for humans. To understand the mechanism of virulence in B. melitensis, the proteome of vaccine strain Rev 1 was analyzed by two-dimensional gel electrophoresis and compared to that of virulent strain 16M. The two strains were grown under identical laboratory conditions. Computer-assisted analysis of the two B. melitensis proteomes revealed proteins expressed in either 16M or Rev 1, as well as up- or down-regulation of proteins specific for each of these strains. These proteins were identified by peptide mass fingerprinting. It was found that certain metabolic pathways may be deregulated in Rev 1. Expression of an immunogenic 31-kDa outer membrane protein, proteins utilized for iron acquisition, and those that play a role in sugar binding, lipid degradation, and amino acid binding was altered in Rev 1.

Brucellosis is a major infectious disease afflicting humans and a wide range of domesticated animals worldwide. The disease is caused by several Brucella species, which are aerobic, nonmotile, gram-negative, facultative intracellular coccobacilli. The genus Brucella belongs to the α-2 subgroup of the class Proteobacteria. It is subdivided, on the basis of its pathogenicity and host preference, into six nomen species: Brucella abortus, B. canis, B. melitensis, B. neotomae, B. ovis, and B. suis (12). In addition, a new strain affecting marine mammals was recently isolated and tentatively named B. maris (32). On the basis of DNA hybridization data, it was suggested that all of these organisms be placed into a single species, B. melitensis (39). Among the various nomen species, B. abortus, B. canis, B. suis, B. maris, and B. melitensis have been reported to infect humans (21, 35, 13). B. melitensis is a pathogen of goats and sheep and is considered the most virulent species for humans. Human infection can result from either occupational contact or ingestion of contaminated food.

Vaccination and eradication of infected hosts have been key factors in the control of brucellosis. Rev 1, an attenuated strain of virulent B. melitensis, was developed in 1957 (19). It is considered the most effective vaccine for the control of brucellosis in small ruminants and was used in comprehensive vaccination programs in many countries, including Saudi Arabia, Kuwait, Mongolia, Spain, and Turkey (8).

Our laboratory has been involved in a comprehensive analysis of the B. melitensis 16M proteome, and initial results have been published recently (40). Previous proteomics studies using B. melitensis cells grown under different conditions have been reported (36, 37), and initial work on the B. abortus proteome has been described (26, 29). A comparative study was conducted with B. abortus vaccine strains S19 and RB51 and virulent strain 2308 (34). In the present investigation, we compared the proteome of laboratory-grown strain Rev 1 to that of strain 16M by using two-dimensional (2-D) gel electrophoresis and matrix-assisted laser desorption/ionization (MALDI)-mass spectrometry (MS) to elucidate differences between the protein expression patterns of the two strains. Differentially expressed proteins were identified and grouped into three major classes: (i) protein spots unique to either 16M or Rev 1, (ii) proteins overexpressed in Rev 1, and (iii) proteins underexpressed in Rev 1.

MATERIALS AND METHODS

Materials.

The reagents and equipment used in this investigation have been described in a recent publication (40) and also at our website (http://www.proteome.scranton.edu).

Bacterial strains.

B. melitensis 16M is the virulent reference strain. Vaccine strain B. melitensis Rev 1 was obtained in 1957 by successively growing virulent strain B. melitensis 6056 (4) on streptomycin-containing media until streptomycin-resistant clones developed. Strain Rev 1 was one of the antibiotic-resistant clones isolated. This commercially available vaccine strain was produced in accordance with specific requirements issued by the World Health Organization (43). Rev 1 resembles other B. melitensis strains metabolically but differs from its virulent 16M counterpart in its slower growth, dye sensitivity, and growth on antibiotic-containing media (3).

Bacterial cultures.

B. melitensis strains 16M and Rev 1 were grown from frozen laboratory stocks on Schaedler blood agar in petri dishes at 37°C in an atmosphere of 5% CO₂. After 3 days of growth, cells were collected and killed with chloroform as previously described (40). The cells were then separated from the chloroform phase and stored at −20°C.

Protein extraction.

Proteins were extracted by the protocol of Rafie-Kolpin et al. (29) with modifications (40). Essentially, chloroform-killed cells (40 μl containing 40 μg of protein) were suspended with an equal volume of 10% trichloroacetic acid (TCA), incubated for 5 min, and centrifuged. The pellet was washed with 5% TCA, followed by acetone; centrifuged; and resuspended in 40 μl of sample buffer 1 (0.3% sodium dodecyl sulfate [SDS], 0.2 M dithiothreitol [DTT], 50 mM Tris-HCl, pH 8.0) and 4 μl of sample buffer 2 (50 mM MgCl₂, 8 U of DNase I, 3 U of RNase A, 0.5 M Tris-HCl, pH 8.0). The mixture was incubated for 10 min, after which 160 μl of loading buffer {8 M urea, 4% 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (CHAPS), 40 mM Tris base, 65 mM DTT, 0.01% bromophenol blue} and 200 μl of rehydration buffer (8 M urea, 2% CHAPS, 10 mM DTT, 2% carrier ampholytes, 0.01% bromophenol blue) were added.

Total protein determination.

Protein concentrations were determined as described by Bradford (9), by using the Bio-Rad protein stain with bovine serum albumin as the standard. All samples and bovine serum albumin standards contained 10 μl of a 10:1 (vol/vol) mixture of sample buffers 1 and 2.

IEF.

Eighteen-centimeter immobilized pH gradient (IPG) strips (pH ranges, 4 to 7, 4 to 5, 4.5 to 5.5, and 5 to 6) were rehydrated overnight at room temperature with 400 μl of sample containing 40 μg of total protein. Isoelectric focusing (IEF) was conducted at 20°C for 24 h (maximum voltage of 5,000 V, maximum current of 80 μA/gel, 80,000 Vh, and an end-of-run hold at 125 V).

SDS-polyacrylamide gel electrophoresis.

After IEF, each IPG strip was washed for 15 min in 10 ml of equilibration buffer 1 (6 M urea, 133 mM DTT, 30% glycerol, 50 mM Tris-acetate, pH 7.0) and then for 15 min in 10 ml of equilibration buffer 2 (6 M urea, 2.5% iodoacetamide, 30% glycerol, 50 mM Tris-acetate, pH 7.0). The IPG strips were loaded onto 10% precast Duracryl gels (22 cm by 23 cm by 1 mm; Tris/Tricine/SDS chemistry). Electrophoresis was carried out for either 5 to 6 h (500 V; 14,000 to 20,000 mW/gel) or 18 to 19 h (500 V; 1,600 mW/gel) at 4°C. All subsequent operations were carried out in an automated staining apparatus. After polyacrylamide gel electrophoresis, the gels were fixed for 30 min in 40% methanol with 10% acetic acid and then rinsed for 5 min with deionized water to remove excess fixative. Gels were stained with SYPRO Ruby for 12 to 15 h in the dark with gentle rocking. After staining, the gels were submerged in deionized water for 5 min, destained for 30 min with 10% methanol and 6% acetic acid, rinsed with deionized water, and stored in 2% glycerol in the dark at 4°C. Gel images were captured under 470-nm light with a Fujifilm LAS-1000 plus imager.

Gel analysis.

Each gel was analyzed with the Investigator HT Analyzer program developed by Nonlinear Dynamics (Newcastle upon Tyne, United Kingdom). For each pH range, gels were run in triplicate. Protein spots were detected and matched on each of these three subgels, and an average gel was generated for each pH range. Spots present in at least two subgels were included in the average gel. The volume of each spot was normalized by selecting one base spot in each gel whose shape and intensity were reproducibly observed in all three subgels to compensate for possible variation in staining quality. Each spot was normalized by dividing its volume by that of the base spot and then multiplying the result by 100. The average standard deviation of the spot volumes ranged from 8 to 44% for strain 16M and from 8 to 50% for strain Rev 1. Standardized average gels for 16M and Rev 1 were compared. Unique spots were defined as those exclusive to either 16M or Rev 1 under the laboratory conditions described above. Spots were characterized as unique on the basis of their distinct electrophoretic mobility. These findings were validated by inspection of all three subgels. To be considered differentially expressed, the protein spot had to show a reproducible pattern (overexpressed or underexpressed) in at least two gels for each strain.

Spot excision.

Protein spots (1.2-mm diameter) were excised from the SYPRO Ruby-stained 2-D gels by a UV box-equipped ProPic robot. Plugs were frozen at −80°C and thawed prior to trypsin digestion.

Automated in-gel trypsin digestion with o-MU modification.

Digestion with trypsin was performed in accordance with the ProGest default long trypsin digestion protocol, which was altered to include modification of the tryptic peptides with o-methylisourea (o-MU) (40). Tryptic peptides were modified for 1 h at 37°C with 1 M o-MU in 100 mM ammonium bicarbonate at pH 10.0 as previously described (24). Modified tryptic peptides were recovered in 7 parts acetonitrile and 2 parts 10% (vol/vol) formic acid (31).

Preparation of tryptic peptides for MALDI analysis.

Tryptic peptides generated with the ProGest station were dried under a vacuum and resuspended in 100 μl of 10% (vol/vol) formic acid. They were then desalted and concentrated with Zip Tips from Millipore (Bedford, Mass.) in the ProMS workstation. Finally, a 1-μl mixture of peptide and matrix (α-cyanohydroxycinnamic acid, 10 mg/ml) was spotted onto the stainless steel 384-well Kratos MALDI target plate.

Mass spectral analysis.

All spectra were obtained with the Kratos Axima-CFR running in reflectron mode. The spectra were collected and analyzed as described previously (40), with the Kompact software package (Kratos Analytical Ltd., Manchester, United Kingdom). After manual selection of the monoisotopic peaks in each spectrum, peptide mass fingerprints were searched in house against all 3,198 open reading frames (ORFs) of the B. melitensis 16M genome with the Mascot and Mascot Daemon software packages from Matrix Science (London, United Kingdom) (28). The search parameters were a maximum of one missed cleavage by trypsin, fixed modifications of oxidized methionine and carbamidomethylated cysteine, variable modification of acetylated lysine (this modification is of the same mass as guanidination [42.02 Da]), a charge state of +1, and a mass tolerance of ±0.2 Da.

RESULTS

General observations.

Annotated proteome maps obtained by using one type of IPG strip (pH range, 4.0 to 7.0) of strains 16M and Rev 1 grown under similar laboratory conditions are presented in Fig. 1. A total of 513 and 522 protein spots were detected for strains 16M and Rev 1, respectively. With IPG strips with various narrow pH ranges (i.e., 4.0 to 5.0, 4.5 to 5.5, and 5.0 to 6.0), the total number of detected protein spots increased to 591 from pH 4 to pH 6. A composite Rev 1 proteome map (presented in Fig. 2) was compared to the recently published 16M proteome map (40). Other IPG strips (i.e., pH ranges of 5.5 to 6.7, 6.0 to 9.0, and 6.0 to 11.0) were also tested with Rev 1 samples, but the unexplained poor protein resolution obtained was not adequate for analysis (data not shown).

FIG. 1.

Proteomes of laboratory-grown B. melitensis strains 16M and Rev 1 in the pH range of 4.0 to 7.0. Protein extracts (40 μg) of each strain were focused with IPG strips and run on SDS-10% Duracryl gels. The gels were stained with SYPRO Ruby and imaged at 470 nm. Underexpressed and overexpressed proteins were labeled in green and red, respectively. Differentially expressed proteins not identified were circled. Blue numbers indicate putative identification.

FIG. 2.

Composite 2-D gel image of the strain Rev 1 proteome. Overlapping images from three different one-unit pH ranges (4 to 5, 4.5 to 5.5, and 5 to 6) were assembled. Underexpressed and overexpressed proteins were labeled in green and red, respectively. Differentially expressed proteins not identified were circled. Blue numbers indicate putative identification. MW, molecular mass.

Protein spots found exclusively in 16M or Rev 1 are listed in Table 1. Those that were underexpressed by Rev 1 are listed in Table 2; those that were overexpressed by Rev 1 are presented in Table 3. The products of a single ORF can have different 2-D electrophoretic mobility because of possible posttranslational modifications. As a consequence, the same function would be assigned to a unique spot and to one or more differentially expressed spots encoded by the same gene. The products of ORF BMEI0755, listed as unique in 16M (spot 7, Table 1) while underexpressed (spots 22 and 23, Table 2) and overexpressed (spot 43, Table 3) in Rev 1, are typical examples. Unique protein spots were often present in very small amounts, making their identification by MALDI-MS difficult. Thus, only 35% of these proteins were identified. Approximately, 87% of underexpressed and 88% of overexpressed proteins were assigned functions. These included outer membrane and periplasmic immunogenic proteins and those involved in sugar and amino acid binding, iron metabolism, and lipid degradation.

TABLE 1.

Unique protein spots in strains 16M and Rev 1 in different pH ranges

Strain and pH range	Spot no.a	Protein name	ORF	pI		Molecular mass (kDa)
				Exptl	Theoretical	Exptl	Theoretical
16M
4.0-7.0	1	NDb		5.69		54.9
	2	Peptidyl-prolyl cis-trans isomerasec	BMEI0123	4.94	4.94	41.2	36.1
4.0-5.0	3	ND		4.39		16.9
	4	Hypothetical protein	BMEI1843	4.38	7.67	15.5	25.8
4.5-5.5	2	Peptidyl-prolyl cis-trans isomerasec	BMEI0123	4.88	4.94	42.1	36.1
5.0-6.0	5	Aconitate hydratase	BMEI1855	5.52	5.55	109.1	97.9
	6	ND		5.51		51.6
	7	EF-Tud	BMEI0755	5.34	5.29	44.9	42.9
			BMEI0742		5.60		44.7
	8	ND		5.64		27.2
Rev 1
4.0-7.0	9	ND		6.09		39.3
	10	ND		5.47		34.0
	11	ND		4.73		16.5
	12	ND		5.23		16.0
4.0-5.0	11	ND		4.68		17.1
4.5-5.5	13	ND		5.31		66.3
	14	Sugar-binding proteinc	BMEII0590	4.83	4.97	40.2	43.5
	15	ND		5.34		34.2
	16	ND		5.11		34.2
	17	ND		5.32		32.1
	11	ND		4.69		16.8
5.0-6.0	18	Chromosome partitioning protein PARB	BMEI0010	5.75	5.81	37.5	32.7
	19	Chromosome partitioning protein PARB	BMEI0010	5.88	5.81	37.3	32.7
	20	Nitrate reductasec	BMEII0305	5.56	8.44	31.6	35.6
	21	ND		5.75		25.1
	22	ND		5.54		20.1
	23	ND		5.33		51.0

^aSpots with the same number were repeatedly observed in different pH ranges.

^bND, not determined.

^cTentative identification.

^dSame protein function assigned to two different ORFs with similar sequences in the genome.

TABLE 2.

Proteins underexpressed in B. melitensis strain Rev 1, compared to strain 16M, and also illustrated in Fig. 1 (pH range 4 to 7) and 2 (pH range 4 to 5, 4.5 to 5.5, and 5 to 6)

pH range and spot no.a	Protein name	ORF	pI		Molecular mass (kDa)		V16M/ VRev1c
			Exptl	Theoretical	Exptl	Theoretical
4.0-7.0
1	31-kDa outer membrane immunogenic protein precursor	BMEII0844	4.88	5.21	34.5	23.3	11.9
2	NDb		5.96		66.5		6.4
3	31-kDa outer membrane immunogenic protein precursor	BMEII0844	5.16	5.21	33.7	23.3	3.9
4	Tetratricopeptide repeat family proteind	BMEI1531	6.10	5.75	60.4	68.1	3.8
5	ND		6.27		19.3		3.6
6	31-kDa outer membrane immunogenic protein precursor	BMEII0844	5.02	5.21	33.0	23.3	3.5
4.0-5.0
7	Arylesterase precursord	BMEI1861	4.56	4.74	22.5	23.5	3.5
4.5-5.5
8	ATP synthase beta chain	BMEI0251	5.16	5.48	57.5	55.0	7.7
9	Glutamate N-acetyltransferase/amino acid acetyltransferased	BMEI0124	5.39	6.17	30.9	43.7	6.0
10	Periplasmic immunogenic protein	BMEI0536	5.28	6.40	32.7	26.7	6.0
3	31-kDa outer membrane immunogenic protein precursor	BMEII0844	5.11	5.21	33.8	23.3	6.0
11	EF-Tu	BMEI0755	5.34	5.29	47.1	42.9	5.4
12	d-Galactose-binding periplasmic protein precursor	BMEII0983	4.98	5.37	40.8	36.4	5.1
13	d-Ribose-binding periplasmic protein precursor	BMEII0435	4.77	5.60	32.0	31.1	4.9
14	Isovaleryl-CoA dehydrogenase	BMEI1923	5.41	5.36	44.2	42.1	4.7
15	Isovaleryl-CoA dehydrogenase	BMEI1923	5.31	5.36	44.7	42.1	4.6
16	Ferric anguibactin-binding protein	BMEII0607	5.40	5.36	37.5	31.3	4.3
17	2-hydroxyhepta-2,4-diene-1,7-dioate isomerase/5-carboxymethyl-2- oxo-hex-3-ene-1,7-dioate decarboxylase	BMEI1708	4.97	6.45	35.4	37.0	4.1
18	60-kDa chaperonin GroEL	BMEII1048	5.09	5.04	66.9	57.8	3.5
6	31-kDa outer membrane immunogenic protein precursor	BMEII0844	4.96	5.21	33.8	23.3	3.4
19	DNA protection during starvation protein	BMEI1980	5.42	5.43	23.6	20.0	3.0
20	4-Hydroxybutyrate dehydrogenase	BMEII1094	5.32	5.23	39.5	34.2	3.0
5.0-6.0
21	DNA protection during starvation protein	BMEI1980	5.19	5.43	23.1	20.0	19.1
22	EF-Tue	BMEI0755	5.20	5.29	45.5	42.9	13.1
		BMEI0742		5.60		44.7
23	EF-Tue	BMEI0755	5.23	5.29	44.9	42.9	10.0
		BMEI0742		5.60		44.7
8	ATP synthase beta chain	BMEI0251	5.10	5.48	54.5	55.0	6.6
19	DNA protection during starvation protein	BMEI1980	5.32	5.43	22.7	20.0	6.3
24	Nitrogen assimilation regulatory protein NTRXd	BMEI0868	5.55	5.53	58.4	50.3	6.2
25	Succinyl-diaminopimelate desuccinylased	BMEII0268	5.52	5.52	44.9	43.1	6.1
26	ATP-dependent CLP protease, ATP-binding subunit CLPB	BMEI0195	5.19	5.41	110.8	103.8	6.1
27	ND		5.61		13.8		6.0
28	N-Acyl-l-amino acid amidohydrolase	BMEI1827	5.37	5.37	57.9	52.7	5.4
29	ATP synthase beta chain	BMEI0251	5.44	5.48	57.5	55.0	5.1
30	DNA protection during starvation proteind	BMEI1980	5.07	5.43	22.8	20.0	4.9
31	ND		5.40		30.8		4.6
32	Chromosome partitioning protein PARB	BMEI0010	5.68	5.81	38.0	32.7	4.2
33	Nickel-binding periplasmic protein precursor	BMEII0487	5.24	5.44	56.7	58.5	4.1
34	Gamma-d-glutamyl-l-diamino acid endopeptidase IId	BMEI1950	5.68	5.64	37.0	32.3	4.1
9	Glutamate N-acetyltransferase/amino acid acetyltransferased	BMEI0124	5.30	6.17	30.0	43.7	3.9
35	Nitrogen assimilation regulatory protein NTRX	BMEI0868	5.65	5.58	57.9	50.3	3.8
10	Periplasmic immunogenic protein	BMEI0536	5.21	6.40	31.2	26.7	3.8
36	Hydroxypyruvate isomerase	BMEII1092	5.11	5.11	29.3	29.7	3.7
37	Acetylglutamate kinase	BMEII0273	5.43	5.46	32.8	34.3	3.3
38	ND		5.42		49.5		3.1
39	ND		5.26		31.9		3.1

^aSpots with the same number were repeatedly observed in different pH ranges.

^bND, not determined.

^cEach value is the ratio of the volumes of matched spots between 16M and Rev 1.

^dTentative identification.

^eSame protein function assigned to two different ORFs with similar sequences in the genome.

TABLE 3.

Proteins overexpressed in B. melitensis strain Rev 1, compared to strain 16M, and also illustrated in Fig. 1 (pH range, 4 to 7) and 2 (pH ranges 4 to 5, 4.5 to 5.5, and 5 to 6)

pH range and spot no.a	Protein name	ORF	pI		Molecular mass (kDa)		VRev 1/ V16Mc
			Exptl	Theoretical	Exptl	Theoretical
4.0-7.0
40	Enoyl-CoA hydratase	BMEI1945	5.65	6.05	30.0	32.8	18.6
41	Bacterioferritin	BMEII0704	4.74	4.74	21.2	20.2	12.7
42	Acyl-CoA dehydrogenase	BMEI1521	5.42	5.30	65.5	63.8	12.5
43	EF-Tud	BMEI0755	5.35	5.29	55.0	42.9	6.6
44	Sugar-binding protein	BMEII0590	4.85	4.97	43.2	43.5	6.2
45	General l-amino-acid-binding periplasmic protein AapJ precursor	BMEI1211	5.06	5.30	36.9	37.4	5.6
46	Branched-chain amino acid ABC transporter, periplasmic amino acid-binding protein	BMEII0344	5.34	5.31	43.0	43.5	5.4
47	Iron-regulated outer membrane protein FRPB	BMEII0105	5.63	5.71	72.1	73.1	5.4
48	Glycine betaine/l-proline-binding protein proX	BMEII0550	5.30	5.57	33.9	32.1	4.8
49	NDb		5.71		55.9		4.7
50	2-Oxoisovalerate dehydrogenase beta subunit	BMEII0747	5.84	5.50	41.4	37.3	4.4
51	Sensory transduction regulatory protein	BMEI0372	4.92	4.76	32.7	29.2	4.3
52	Iron-regulated outer membrane protein FRPB	BMEII0105	5.72	5.71	71.5	73.1	4.1
53	ND		6.29		14.8		4.0
54	ND		5.98		15.3		3.9
55	Leu-, Ile-, and Val-binding protein precursor	BMEII0103	5.18	5.66	43.6	48.4	3.8
56	Protein translation elongation factor G	BMEI0754	5.10	5.05	91.4	76.8	3.8
57	d-Ribose-binding periplasmic protein precursor	BMEI1390	4.47	4.61	35.8	33.3	3.7
58	Peptidoglycan-associated lipoprotein	BMEI0340	5.77	9.92	18.7	18.4	3.7
59	Alcohol dehydrogenase class III	BMEI1819	5.79	5.93	43.5	39.7	3.7
60	Leu-, Ile-, Val-, Thr-, and Ala-binding protein precursor	BMEII0633	5.60	5.54	43.7	42.7	3.5
61	ND		5.30		43.0		3.5
62	Alcohol dehydrogenase	BMEI1746	5.86	6.07	41.7	39.8	3.5
63	Phage-related DNA-binding protein	BMEI0899	4.69	6.92	21.2	19.4	3.2
64	UTP-glucose-1-phosphate uridylyltransferase	BMEII0023	6.12	5.73	37.2	33.0	3.1
65	ND		5.36		43.4		3.0
66	Periplasmic dipeptide transport protein precursor	BMEII0284	5.38	5.83	57.7	55.4	3.0
67	Iron(III)-binding periplasmic protein precursor	BMEII1120	4.79	4.88	42.2	39.2	3.0
4.0-5.0
44	Sugar-binding protein	BMEII0590	4.85	4.97	42.1	43.5	5.7
68	Hypothetical membrane-associated protein	BMEI1551	4.63	4.96	29.4	23.7	4.9
57	d-Ribose-binding periplasmic protein precursor	BMEI1390	4.48	4.61	35.6	33.3	4.9
69	Sugar-binding protein	BMEII0590	4.80	4.97	42.3	43.5	4.7
67	Iron(III)-binding periplasmic protein precursor	BMEII1120	4.77	4.88	41.6	39.2	3.6
70	Hypothetical membrane-associated protein	BMEI1551	4.85	4.96	31.3	23.7	3.4
71	2-Oxoisovalerate dehydrogenase beta subunit	BMEII0747	4.83	5.50	16.5	37.3	3.1
4.5-5.5
72	Isovaleryl-CoA dehydrogenase	BMEI1923	4.93	5.36	42.4	42.1	3.7
73	Putative binding protein YDDS precursor	BMEII0203	4.70	4.92	58.6	57.3	3.1
5.0-6.0
74	Transaldolase	BMEI0244	5.25	5.69	27.5	23.7	5.1
75	Lipoamide acyltransferase component of branched-chain α-keto acid dehydrogenase complexd	BMEII0746	5.72	5.77	53.8	46.5	4.5
76	Nitrogen assimilation regulatory protein NTRX	BMEI0868	5.18	5.58	53.2	50.3	3.8

^aSpots with the same number were repeatedly observed in different pH ranges.

^bND, not determined.

^cEach value is the ratio of the volumes of matched spots between 16M and Rev 1.

^dTentative identification.

Unique protein spots.

Nine of the 23 unique protein spots (excluding spots repeatedly detected in different pH ranges, e.g., spots 2 and 11) detected in both strains were assigned functions (Table 1). Strain-specific proteins were observed in various pH ranges in strains 16M and Rev 1. In the pH range of 4 to 7, two unique protein spots were detected in 16M but only one was identified. In Rev 1, none of the four proteins detected was identified. In the pH range of 4 to 5, two proteins were detected in 16M and one was detected in Rev 1, whereas in the pH range of 4.5 to 5.5 one protein was detected in 16M and six were detected in Rev 1. Four unique protein spots were detected in 16M and six were detected in Rev 1 in the pH range of 5 to 6. A representative example of a unique spot in strain 16M, which was tentatively identified as a peptidyl-prolyl cis-trans isomerase (ORF BMEI0123, Table 1), is shown in Fig. 3. This protein was consistently observed with IPG strips with pH ranges of 4.0 to 7.0 and 4.5 to 5.5.

FIG. 3.

A strain 16M-specific peptidyl-prolyl cis-trans isomerase encoded by ORF BMEI0123. The use of IPG strips with narrower pH ranges (4.0 to 7.0 and 4.5 to 5.5) improved the protein resolution and revealed that the 41.5-kDa protein is absent from strain Rev 1.

Underexpressed proteins in strain Rev 1.

Several proteins in Rev 1 were expressed at a significantly lower level (3.0- to 19.1-fold) and thus were classified as underexpressed in comparison to their expression in 16M (Table 2). A total of 45 Rev 1 proteins (including spots repeatedly detected in different pH ranges, e.g., spots 3, 6, 8, 9, 10, and 19) were noted with apparently low expression levels. They were distributed among the different pH ranges as follows: six proteins in the pH range of 4.0 to 7.0, one protein in the pH range of 4.0 to 5.0, 15 proteins in the pH range of 4.5 to 5.5, and 23 proteins in the pH range of 5.0 to 6.0. The greatest differential expression was found for the DNA protection during starvation protein (19.1-fold), followed by protein translation elongation factor EF-Tu (10- to 13.1-fold), the immunogenic 31-kDa outer membrane protein precursor (11.9-fold), and the ATP synthase β chain (7.7-fold) (Table 2). Some sugar-binding proteins and those involved in protein synthesis were also underexpressed (i.e., products of ORFs BMEII0983, BMEII0435, BMEI0755, and BMEI0742). Of the 45 underexpressed proteins, 31 were identified with high certainty by peptide mass fingerprinting whereas eight spots were given tentative identification because of their poor MS spectra. The remaining six spots were not identified.

Proteins overexpressed in Rev 1.

Rev 1 proteins whose amounts increased 3.0- to 18.6-fold compared to their amounts in 16M were categorized as overexpressed (Table 3). A total of 40 proteins (including spots repeatedly detected in different pH ranges, e.g., spots 44, 57, and 67) in this group were distributed as follows: 28 proteins in the pH range of 4.0 to 7.0, seven proteins in the pH range of 4.0 to 5.0, two proteins in the pH range of 4.5 to 5.5, and three proteins in the pH range of 5.0 to 6.0. Thirty-three of these proteins were assigned functions, two were tentatively identified, and five were not assigned a function. Overexpressed proteins included those involved in iron metabolism, such as bacterioferritin (12.7-fold), the iron-regulated outer membrane protein FRPB (5.4-fold), and the iron(III)-binding periplasmic protein precursor (3.0- to 3.6-fold). Likewise, proteins involved in sugar binding, protein synthesis, and lipid metabolism were overexpressed.

DISCUSSION

The metabolic status of an organism can be evaluated by transcriptome analysis; however, the amount of mRNA often does not correlate with protein expression (23). The proteome is a more accurate and complete description and is often referred to as a snapshot of the physiological state of the cell at any time point. The availability of a completely sequenced and annotated B. melitensis genome (16) (http://www.genome.scranton.edu/Brucella; GenBank accession numbers AE008917 [chromosome I] and AE008918 [chromosome II]) has greatly facilitated an investigation of the proteome of this organism (40).

Strains 16M and Rev 1 grown under defined laboratory conditions had proteomes highly similar to each other in terms of the total number of spots and the general protein patterns on 2-D gels (Fig. 1). This is not surprising since strains 16M and Rev 1 belong to the same species and thus could be expected to have a very high level of genomic sequence homology. One striking difference between 16M and Rev 1 is the expression level of certain classes of proteins. Whether this variation is a consequence of transcriptional, posttranscriptional, translational, or posttranslational modifications cannot be stated. Protein expression levels may reflect differences in metabolic properties that may have resulted from ancestral lineage or attenuation of the vaccine strain.

Altered iron metabolism in Rev 1.

One striking difference between the two strains of B. melitensis is the number of differentially expressed proteins involved in iron metabolism. Most are categorized as overexpressed in Rev 1 (Table 3), such as bacterioferritin (12.7-fold), iron regulated outer membrane protein B (5.4- and 4.1-fold), and the iron(III)-binding periplasmic protein precursor (3-fold). The above proteins are components of different iron acquisition systems, suggesting that Rev 1 may have an increased capacity to assimilate iron via multiple transport systems.

Iron plays a salient role in the survival of pathogens once inside host cells (30). This element is essential in the synthesis of iron- and heme-containing enzymes. During infection, a macrophage reduces iron availability by producing chelating agents and actively exporting iron from the phagosome, where the pathogen multiplies (22). In response, the pathogen synthesizes a battery of proteins designed to compete with the host for iron (30). One system utilizes siderophores, which capture iron from the host's intracellular environment, together with a transmembrane transport system that then returns the ferrisiderophore into the bacterial cell. A bacterial outer membrane receptor binds specifically to the ferrisiderophore, translocating it toward the periplasm. A periplasmic protein then transfers the ferrisiderophore to the cytoplasmic membrane, where it is placed into the cytosol by a membrane ATP-binding cassette (ABC) transporter. Synthesis of all of these components is derepressed during iron deprivation (30).

Upregulation of bacterioferritin synthesis in Rev 1 grown in nutrient-rich media is expected since this protein functions in cytoplasmic iron storage (7). We found that bacterioferritin accumulation is significantly greater in strain Rev 1 than in strain 16M. It may be that Rev 1 lost the ability to regulate bacterioferritin synthesis and degradation. Furthermore, proteins such as the outer membrane protein and an iron(III)-binding periplasmic protein, which are normally derepressed during low iron availability, were found to be simultaneously overexpressed in Rev 1. This may indicate a misregulation of both the iron capture system and iron metabolism.

The annotated B. melitensis genome has three different ORFs for the ferric uptake regulation repressor protein (Fur), which controls the expression of genes essential for extracellular iron uptake and transfer toward the cytosol (20, 38). In Helicobacter pylori, replacement of the fur gene with a kanamycin cassette resulted in derepression of the genes for ferritin and iron-regulated outer membrane protein B (FrpB) (15). Both proteins of the H. pylori mutant were expressed at high levels. This was also the case for Rev 1. A comparative analysis of fur gene sequences in strains Rev 1 and 16M may provide some insights into the molecular regulatory mechanism of Fur-regulated genes. We have not been able to detect Fur proteins on our 2-D gels. This is not surprising since these proteins are likely present in small amounts, which is usually the case for regulatory proteins.

The effects of various stress conditions on the B. melitensis proteome have been investigated (37). One of the proteins significantly affected was bacterioferritin, whose level of expression was reduced in response to heat shock. A bacterioferritin deletion mutant of B. melitensis 16M generated by gene replacement was recently described (17). The survival and growth of the mutant were similar to those of its parental strain in human monocyte-derived macrophages, suggesting that bacterioferritin is not essential for intracellular survival in monocyte-derived macrophages. However, this experiment was conducted with an in vitro system. It would be interesting to test this mutant strain in live animals. Recently, a DNA vaccine encoding bacterioferritin (2) and the bacterioferritin antigens of B. melitensis 16M (1) have been reported to induce cellular and humoral immune responses but did not protect BALB/c mice against a challenge with B. abortus.

In Neisseria gonorrhoeae, the protein FrpB (also referred to as FetA) functions as an enterobactin receptor (11). It may operate along with periplasmic binding protein FetB in transporting the ferrisiderophore into the cell. The iron(III)-binding periplasmic protein is not involved in transport of the siderophore through the membrane. Thus, it appears that the enhanced ability of Rev 1 to import iron via multiple transport systems results in increased iron uptake, which consequently may lead to higher oxygen toxicity. This could occur by Fenton chemistry, resulting in increased accumulation of reactive oxygen species. If this were the case in Rev 1, it could partially explain its slower growth compared to that of 16M (5).

The oxidative killing pathway is one of the major defense systems utilized by the macrophage against infectious microorganisms (27). It involves the production of various oxidants such as superoxide radical (^·O₂⁻), which, with its relatively low reactivity, diffuses into the targeted cell before inactivating bacterial enzymes. Another important component is hydrogen peroxide (H₂O₂), which destabilizes cellular membranes, oxidizes enzymes, inhibits membrane transport systems, and damages DNA. The hydroxyl radical (^·OH) is most reactive and requires a metal catalyst (i.e., Fe³⁺) for its formation during the reaction of H₂O₂ with ^·O₂⁻. This compound is highly toxic since it can oxidize a large variety of biomolecules, including proteins, DNA, and lipids. It can also generate free-radical chain reactions, inducing further host cell injury. Thus, it is important for a cell to strictly regulate the storage of iron so that its availability for the production of free radicals is reduced.

Sugar-binding proteins.

Sugars are used as carbon sources and as precursors of other complex cell constituents such as lipopolysaccharides. Several sugar-binding proteins were differentially expressed in Rev 1. Proteins that were up-regulated included the sugar-binding proteins specific to glucose (BMEII0590) and a d-ribose-binding periplasmic protein (BMEI1390) (Table 3). Likewise, two underexpressed proteins were the d-galactose-binding periplasmic protein precursor BMEII0983 and the d-ribose-binding periplasmic protein precursor BMEII0435 (Table 2). Despite the differential expression of these proteins, strains 16M and Rev 1 appear to have similar physical properties on the basis of available biotyping tests (http://www.moag.gov.il/brunet). Rev 1 and other B. melitensis strains do not differ significantly in the ability to oxidize various sugar substrates, including erythritol (14). Since galactose can be interconverted to glucose-6-phosphate and enter the glycolytic pathway, we could speculate that a decrease in the intracellular level of galactose would induce enhanced glucose uptake. The increase in glucose-binding protein in Rev 1 may be a compensatory response to a defective system for galactose uptake or metabolism. The low level of galactose-binding protein in Rev 1 supports this hypothesis. Sugar transport toward the cytosol is mediated by different transport systems, including permeases and ABC transporters composed of membrane-associated and periplasmic proteins. Previous studies of periplasmic sugar-binding protein ChvE from Agrobacterium tumefaciens, a plant pathogen closely related to Brucella, indicated that inducible ChvE may be involved in the expression of virulence genes (18). However, a ChvE homologue in B. suis is constitutively expressed and may not be essential for survival inside the macrophage (6).

Lipid metabolism and protein synthesis.

The increased accumulation of enoyl coenzyme A (CoA) hydratase (18.6-fold) and acyl-CoA dehydrogenase (12.5-fold) in Rev 1 indicated up-regulation of fatty acid metabolism. Both enzymes are involved in the dehydrogenation and hydration reactions during the β oxidation of fatty acids to generate acetyl-CoA. Consequently, acetyl-CoA enters the TCA cycle, in which reducing equivalents are liberated as NADH and FADH₂. As in the TCA cycle, β oxidation generates reduced electron carriers whose reoxidation generates ATP via oxidative phosphorylation from ADP.

A number of amino acid-binding proteins were overexpressed in Rev 1. These included the general l-amino acid-binding periplasmic protein AapJ precursor (5.6-fold); a branched-chain amino acid ABC transporter/periplasmic amino acid-binding protein (5.4-fold); glycine betaine/l-proline-binding protein proX (4.8-fold); the Leu-, Ile-, and Val-binding protein precursor (3.8-fold); and the Leu-, Ile-, Val-, Thr-, and Ala-binding protein precursors (3.5-fold). All of these proteins were characterized as periplasmic in different bacteria (41, 42, 10, 33, 25). It is possible that the up-regulation of lipid metabolism and the overexpression of various amino acid-binding proteins are essential survival mechanisms that drive the generation of more ATP to compensate for the energy wasted on the inefficient metabolism of iron.

Conclusion.

Comparative proteome analysis of vaccine strain Rev 1 and virulent strain 16M of B. melitensis indicates that the two strains have significant metabolic differences. Differentially expressed proteins involved in iron metabolism, sugar transport, lipid metabolism, and protein synthesis were identified. The expression of proteins essential for both low and high iron availability suggests a misregulated system for iron metabolism and capture, leading to possible unnecessary expenditure of energy. This may be a consequence of successive in vitro passages of B. melitensis in the presence of streptomycin. It is difficult to state what changes were directly or indirectly effected by this stressful growth condition. However, one plausible theory is that to compensate for these changes in gene expression, Rev 1 may have up-regulated other pathways, such as those involved in the β oxidation of fatty acids and protein synthesis, to generate more reducing equivalents, ultimately for use in the production of ATP. These alterations would compensate for the energy loss due to misregulation of iron metabolism.