Cloning, expression and characterization of immunogenic aminopeptidase N from Brucella melitensis

Abstract

A 97-kDa purified aminopeptidase N (PepN) of Brucella melitensis was previously identified to be immunogenic in humans. The B. melitensis pepN gene was cloned, expressed in Escherichia coli and purified by affinity chromatography. The recombinant PepN (rPepN) exhibited the same biochemical properties, specificity and susceptibility to inhibitors as the native PepN. rPepN was evaluated as a diagnostic antigen in an indirect enzyme-linked immunosorbent assay (ELISA) using sera from patients with acute and chronic brucellosis. The specificity of the ELISA was determined with sera from healthy donors. The ELISA had a cutoff value of 0.156 with 100% specificity and 100% sensitivity. Higher sensitivity was obtained using rPepN compared with crude extract from B. melitensis. Anti-PepN sera did not exhibit serological cross-reaction to crude extracts from Rhizobium tropici, Ochrobactrum anthropi, Yersinia enterocolitica 09 or E. coli O157H7.

PepN, aminopeptidase N, Brucella, recombinant, ELISA

Introduction

Brucellae are small, gram-negative, nonspore-forming cocobacilli and are able to replicate in the phagosomes of macrophages (Liautard et al., 1996). Brucellosis is a widespread zoonotic disease that affects a variety of livestock and wildlife, and results in decreased reproductive efficiency and abortion. The disease remains endemic in many Latin American countries where it causes important economic losses. The infection in humans is commonly acquired by consumption of raw milk and cheese made from unpasteurized milk (López-Merino, 1989; Luna-Martinez & Mejia-Teran, 2002). The bacteria cause a severe and debilitating disease that can involve many organs and tissues, and may lead to death in some untreated cases (López-Merino, 1989; Luna-Martinez & Mejia-Teran, 2002).

The lipopolysaccharide O-side chain of the smooth Brucella species seems to be the immunodominant antigen that elicits a long-lasting serological response (Baldi et al., 1996). Nevertheless, it is well known that diagnostic tests based on lipopolysaccharide O-side chain antibodies can generate false positives based on cross-reaction with antigens from other gram-negative bacteria (Weynants et al., 1996). The identification of antigens that could potentially be useful for the specific serological diagnosis is an important goal in brucellosis research. Several such immunoreactive proteins have been identified, e.g. BP26 (Rossetti et al., 1996), superoxide dismutase (Tabatabai & Hennager, 1994), Brucella lumazine synthase, an 18-kDa cytoplasmic protein (Goldbaum et al., 1993), the brucellar homologue of the ribosome recycling factor, named CP24 (Cloeckaert et al., 1996), and a 14-kDa protein that possesses lectin-like properties (Chirhart-Gilleland et al., 1998; Vemulapalli et al., 2006). Letesson et al., (1997) used a set of different recombinant proteins toward the development of multiprotein diagnostic reagent for brucellosis using an enzyme-linked immunosorbent assay (ELISA). Cassataro et al., (2002) reported increased detection in subacute and chronic brucellosis cases using lumazine and CP24 together in an ELISA (Cassataro et al., 2002).

We have previously reported the purification of an aminopeptidase N (PepN) from Brucella melitensis that was recognized by sera from human patients with acute and chronic brucellosis using western blot analysis (Contreras-Rodriguez et al., 2003). The native B. melitensis PepN was found predominantly in a cell-free extract (soluble extract). Furthermore, protein sequence analysis did not identify hydrophobic segments or a signal sequence related to membrane attachment (Contreras-Rodriguez et al., 2003). In Lactococcus lactis, immunogold labelling using antibodies against PepN strongly supported an intracellular location of this aminopeptidase (van Alen-Boerrigter et al., 1991).

In the present study we cloned and purified recombinant PepN (rPepN) of B. melitensis 16M produced in Escherichia coli. Biochemical characterization of recombinant enzyme was performed. The usefulness of rPepN was evaluated as a diagnostic reagent in an ELISA using sera from cases of acute and chronic human brucellosis.

Materials and methods

Bacterial strains, media, plasmids and growth conditions

Brucella melitensis 16M (ATCC 23456) and Ochrobactrum anthropi 95-5 were kindly donated by the Central Veterinary Laboratory (New Haw, Weybridge, UK). The strains were grown at 37°C on trypticase soy agar supplemented with 0.5% (w/v) yeast extract. Rhizobium tropici CFN 299 was kindly provided by E. Martinez-Romero, Centro de Ciencias Genomicas, Cuernavaca, Mexico. Rhizobium tropici was grown at 28°C on peptone broth supplemented with 0.5% yeast extract for 48 h. Yersinia enterocolitica 09 and Escherichia coli O157:H7 from our strain collection were grown at 37°C in trypticase soy broth.

Escherichia coli TOPO10F′ and E. coli BL21(DE3)pLysS were used as hosts (Invitrogen, Inc.). pCR2.1 vector containing the TA cloning sites and pPRSET A containing the inducible T7 RNA polymerase promoter were also obtained from Invitrogen. All E. coli strains were grown at 37°C in Luria Bertani broth (LB), or LB plates solidified with 1.5% agar. Ampicillin was used at 50 µg mL⁻¹.

DNA isolation and manipulation

DNA manipulations were performed according to standard procedures (Sambrook & Russell, 2001).

PCR amplification

A primer pair consisting of one forward primer (5′-GGATC-CATGCGTACTGAAAC-3′) and one reverse primer (5′-GAATTCTCAGGCGAGCGTTCG-3′) were designed based on the nucleotide sequence of the pepN gene from B. melitensis 16M (GenBank accession number NP_540241). A restriction site (underlined) was included into each primer (BamHI in the forward primer, and EcoRI in the reverse primer).

Expression of PepN in E. coli

The amplified pepN gene from B. melitensis 16M was cloned into the pCR2.1 vector. The PepN gene was excised from the pCR2.1 plasmid and subcloned into the pRSET A plasmid. The resulting plasmid was transformed into E. coli BL21(DE3) and expression was induced according to the manufacturer's instructions (Invitrogen, Inc.). The rPepN was isolated by affinity chromatography on His-Bind resin (Amersham-Pharmacia Biotech).

Enzyme characterization

Aminopeptidase activity was assayed by measuring the hydrolysis of Ala-pNA as previously described (Contreras-Rodriguez et al., 2003). The effect of pH, temperature, substrate specificity, inhibitors and determination of molecular mass measurement upon rPepN was examined as previously described (Contreras-Rodriguez et al., 2003).

Human sera

Sera from 25 patients with brucellosis at different stages of the disease belonging to our sera collection were used. The samples were selected according to the following criteria: 15 had acute brucellosis as determined by a clinical diagnosis of brucellosis, an agglutination titre equal to or higher than 1 : 80, and a positive 2-mercaptoethanol (2-ME) agglutination titre. Eight of these patients were positive for Brucella isolation from blood cultures. The other 10 patients were considered to have chronic brucellosis as determined by a history of persisting symptoms or relapses and persistent agglutination titres over the course of more than 1 year. Sera from 53 volunteers without any history of brucellosis, and negative to Rose Bengal, standard tube and 2-ME agglutination tests, were included as negative controls.

Preparation of antigens, cross-reaction analysis and western blot using anti-PepN serum

We obtained extracts from B. melitensis 16M, R. tropici CFN 299, O. anthropi 95-5, Y. enterocolitica 09 and E. coli O157H7 as previously described (Contreras-Rodriguez et al., 2003). Immunoblotting was performed using anti-PepN serum as previously described (Contreras-Rodriguez et al., 2003).

ELISA

Microtitre plates were coated with the corresponding crude extract from B. melitensis 16 M or rPepN at 0.5 µg per well diluted in carbonate bicarbonate buffer, pH 9.6, and then blocked with 2% albumin. Human sera were tested at a dilution of 1 : 100 in phosphate-buffered saline-0.05% Tween 20. After incubation for 1 h at 37°C, binding of antibodies was detected by a further incubation for 1 h at 37°C with horseradish peroxidase-labelled antihuman immunoglobulin G (Cappel). The substrate used to reveal binding was ortho-phenylenediamine. To discount nonspecific reactivity, each sample was also tested in wells not coated with the antigen, and the specific optical density (OD_{492 nm}) was calculated as OD antigen−OD no antigen. To establish the cutoff value of the assay, serum samples from healthy blood donors were tested under the same conditions described above. The cutoff value of the ELISA system was calculated as the mean specific OD_{492 nm} of control sera plus three standard deviations.

Statistical analysis

Absorbance values observed in different groups were compared by anova analysis using Analyse-it software (Analyse-it for Microsoft Excel, General+Clinical Laboratory statistics v.1.71, Analyse-it Software, Ltd).

Results and discussion

The rPepN expressed in E. coli was isolated by Ni²⁺ affinity resin to more than 95% purity. The specific activity (150 U mg⁻¹) was very similar to that obtained for the purified enzyme from B. melitensis VTRM1 (159 U mg⁻¹) and the yield was 25 mg from 2 L of E. coli grown in LB (Table 1).

Table 1

Purification of recombinant Brucella melitensis PepN

Purification step	Total protein (mg)	Total activity (U)	Specific activity (U mg−1)	Yield (%)	Purification (fold)
Sonication	350	5130	14.6	100	1
Metal affinity resin	25	3750	150	70	10.2

Protein concentrations were determined using a Pierce assay (Pierce Biotechnology).

The enzyme activities were assayed with Ala-pNA for the substrate and are expressed as micromoles of pNA released per minute.

Sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis of purified rPepN revealed the presence of a single protein band with an apparent molecular mass of 98 kDa (Fig. 1); the same molecular mass was observed following gel filtration chromatography (data not shown). The enzyme was stable in the pH range 6.0 8.5 and had no activity at a pH below 6.0 or above 9.0. The activity of the rPepN was optimal at 40°C. The activity of rPepN was almost completely inhibited in the presence of the chelating agents 1,10-phenanthroline (10 µM) or EDTA (2.5 mM). Bestatine inhibited almost 50% of the activity. rPepN was insensitive to leupeptin and slight inhibition was observed in the presence of phenylmethylsulphonylfluoride (PMSF), E64 and pepstatin. The presence of Zn²⁺ and Hg²⁺ caused complete inhibition at 1.0 mM, while Ca²⁺, Mg²⁺ and Co²⁺ caused 3 15% inhibition (Table 2). rPepN was able to release amino acids Ala, Leu, Met, Gly and Lys from their corresponding aminoacyl-pNA substrates. Ala-pNA was most actively hydrolysed, whereas (in order of susceptibility) Lys-pNA, Leu-pNA, Met-pNA and Gly-pNA were hydrolysed at lower rates. rPepN showed no detectable reactivity against Glu-pNA. Weak activity (2%) was detected against Ala-Pro-pNA, whereas no activity was detected against Gly-Phe-pNA, Gly-Pro-pNA, Ala-Ala-pNA and Ala-Phe-pNA. Native PepN from Brucella and rPepN expressed in E. coli exhibited very similar biochemical properties, substrate specificity and susceptibility to inhibitors (Contreras-Rodriguez et al., 2003).

Figure 1

(a) Denaturing gel electrophoresis analysis of total cell lysate of Escherichia coli producing recombinant PepN enzyme. Lane 1, ladder (kDa); lane 2, culture grown at 37°C (uninduced); lane 3, culture 4 h postinduction (37°C); lane 4, purified recombinant Brucella melitensis PepN. Calibration of the gels is indicated in kilodaltons. Protein detection was by Coomasie blue staining. (b) Western blot analysis of purified recombinant PepN. Lane 1, molecular mass markers; lane 2, purified recombinant PepN using sera antinative PepN.

Table 2

Effect of proteinase inhibitors and divalent cations on recombinant PepN

Inhibitor	Concentration	Relative activity (%)
None		100
Zn2+	1 mM	0
Co2+	1 mM	91
Ca2+	1 mM	85
Hg2+	1 mM	0
Mg2+	1 mM	97
EDTA	2.5 mM	0
Pefabloc	5 µM	64
Bestatin	250 µM	51
Pepstatin	250 µM	92
Leupeptin	50 µM	100
PMSF	5 µM	95
E-64	50 µM	93
1,10-phenanthroline	10 µM	15

The detection of antibodies against metalloproteases as a diagnostic test has been reported for other infectious agents. A 52-kDa metalloenzyme of Candida albicans was used for the diagnosis of systemic candidiasis (El-Moudni et al., 1998). In the same context, a metalloprotease (110 kDa) purified from Toxoplasma gondii reacted with sera from patients who experienced toxoplasmosis (Berthonneau et al., 2000).

In this ELISA, the sera examined was from patients with acute (n=15) and chronic (n=10) brucellosis, and from healthy donors (n=53). The median reactivities of sera from patients with acute and chronic brucellosis were 0.293 and 0.296, respectively. By contrast, the mean absorbance value for the healthy donors was 0.065 (Fig. 2).

Figure 2

Reactivities of human sera against recombinant Brucella melitensis PepN determined by ELISA. Groups: 1, healthy blood donors (n=53); 2, patients with acute brucellosis (n=15); 3, patients with chronic brucellosis (n=10). Significant differences with healthy donors were found (P=0.0001).

Statistical analysis using anova showed significant differences between absorbance values obtained with sera from acute and chronic brucellosis relative to those from healthy donors (P<0.0001). In addition, there was no significant difference between absorbance values obtained with sera from acute and chronic brucellosis patients (P=0.816). We expected that the sera from acute and chronic patients would recognize rPepN as we had observed reactivity previously using native PepN in a western blot (Contreras-Rodriguez et al., 2003).

The rPepN ELISA had a diagnostic value at a cutoff point of 0.156 with 100% specificity and 100% sensitivity. It was very clear that sensitivity of the ELISA was lower when a B. melitensis crude extract was used (see Table 3).

Table 3

Comparison of diagnostic values of the recombinant PepN and crude extract from Brucella melitensis using an ELISA

Antigen	Cutoff	Sensitivity (%)	Specificity (%)	TP	TN	FP	FN
Recombinant PepN	0.156	100	100	25	53	0	0
Crude extract	0.333	84	100	21	53	0	4

TP, true positives (acute and chronic patients, n=25); TN, true negatives (n=53); FP, false positives; FN, false negatives; cutoff, absorbance at 492 nm.

Previously, we had performed an ELISA test with crude B. melitensis antigens, but using a rPepN we have increased the specificity of the test, avoided the cross-reaction due to lipopolysaccharide, and excluded the possible interference of other components from B. melitensis or E. coli present in soluble protein extracts.

Infections by Y. enterocolitica 09 and a few gram-negative bacteria cause false-positive reactions in Brucella diagnostic assays; other cross-reactivities between antigens of Brucella and Alphaproteobacteria (Delpino et al., 2004) have been reported. We investigated the serological cross-reaction between rPepN and two related Alphaproteobacteria, R. tropici and O. anthropi, and two unrelated organisms, Y. enterocolitica and E. coli O157:H7. The results indicated no reaction of anti-PepN serum against extracts from this range of bacteria (data not shown).

In summary, this study shows that rPepN was successfully expressed and purified from E. coli. Moreover, biochemical analysis demonstrated that the properties of rPepN were in agreement with those reported for native B. melitensis PepN. rPepN could be useful for diagnosis of acute and chronic human brucellosis, as it can be easily and safely prepared from E. coli as opposed to B. melitensis. Further analysis should be performed in order to confirm the value of this enzyme as a target antigen in the diagnosis of brucellosis.

Acknowledgements

A.C.-R. was supported by a fellowship from CONACYT and PIFI. This work was supported in part by grants for the CONACYT 1631PM and CGPI-IPN 32.1 project.

Author notes

Editor: Alex van Belkum