Brucella spp. causes brucellosis, a zoonotic disease still endemic in many countries of the world. This infectious disease affects different animal species and is transmitted to humans in several ways, the most common through ingestion of raw milk or other unpasteurized dairy products. The preferred ecological niche for the brucellae is within phagosomal compartments of host macrophages; the capacity of this bacterial pathogen to establish and maintain chronic infections is dependent upon its ability to replicate within these phagocytic cells 1. Brucella belongs to the alpha-2 subdivision of the Proteobacteria and they are therefore phylogenetically related to the plant cell-associated species of the genera Rhizobium and Agrobacterium 2.

A wide variety of environmentally and medically important bacteria produce the enzyme urease (urea amidohydrolase; EC 3.5.1.5), which catalyzes the hydrolysis of urea, leading to the production of carbon dioxide and ammonia 3. This enzyme allows many soil bacteria to use urea as a nitrogen source. Urease is also an important virulence factor that improves survival of pathogenic bacteria under acidic conditions within the host and can also cause direct damage to the host tissue due to ammonia, CO₂ or alkali production 45. Interestingly, some species of Rhizobiales, such as Brucellae and Bradyrhizobium BTAi1 show multiple urease clusters 6. This reiteration so far is found in other bacterial species, namely Streptomyces coelicolor, S. avermitilis, Pseudomonas syringae and Escherichia coli 7.

Several bacterial ureases have already been purified and characterized 5. One of the most studied ureases is from Helicobacter pylori, a bacterium that is able to persist in the stomach where the pH is very acidic 8. Urease activity is an important colonization factor by generating ammonia in the immediate bacterial microenvironment, thus protecting H. pylori from the deleterious effects of gastric acid 9. Furthermore, urease activity appears to be responsible for the acid resistance of the invasive enteric pathogen Yersinia enterocolitica 10. Interestingly, some humans have a genetic predisposition to develop reactive arthritis following a Y. enterocolitica infection, which correlates to their serum reactivity with the UreB subunit of Yersinia urease 11.

The present work describes the purification as well as the enzymatic and phylogenomic characterization of urease from Brucella suis strain 1330; in addition, the urease reactivity of sera from patients diagnosed with brucellosis was examined. To our knowledge, no Brucella urease has been previously purified and characterized. Since Brucella is a human pathogen, it is important to obtain as much information as possible to complement databases, such as BRENDA 121314. These results can then be used by experimentalists and modelers to understand, from a systems biology point of view 15, the mode of action of enzymes in the pathogen and how that can be modified in order to develop new disease treatments.

Urease was purified from the soluble extract of B. suis 1330 using two consecutive chromatography steps that resulted in the final isolation of a homogeneous enzyme (Figure 1). As previously reported in the case of Mycobacterium tuberculosis 16, hydrophobic-interaction chromatography was consistently the most useful method to purify B. suis urease. The urease was purified 51-fold, with a recovery of 12% of total urease activity and 0.24% of the total protein (Table 1).

Urease activity was always eluted as a single symmetrical peak at 450 mM NaCl during anion exchange chromatography, and eluted at 180 mM of NH₄SO₄ in hydrophobic interaction chromatography. During the chromatographic purification, all fractions were measured for urease activity and consistently produced one activity at the same salt concentrations. The purified protein showed a specific activity of 540 U/mg. This is in the range of urease specific activities described for other organisms: from very low – 8 U/mg for Rhodobacter capsulatus 17 – to as high as 3500 U/mg in Brevibacterium ammoniagenes 18.

An SDS-PAGE analysis of the purified enzyme revealed the presence of three protein bands with apparent molecular masses of 66,000, 15,500 and 14,000 Da; the low molecular weight bands were poorly stained in contrast to the large subunit (Figure 2). The stoichiometry of the subunits in bacterial ureases has been the subject of considerable investigation. Separation of individual subunits by SDS-PAGE and integration of the staining intensities for the multiple bands reveal near-integer values for the subunits stoichiometries, but these ratios vary for the enzymes from different sources. It is likely that all ureases possess equal numbers of each or their distinct subunits polypeptides 5. The molecular weight of the native enzyme obtained by gel filtration on Superose was 130–135 kDa, while molecular weight obtained by nondenaturing zymograms gels, appeared as one prominent band with a molecular weight of 157 ± 7 kDa (Figure 3). Bacterial native ureases show a range of molecular weights from 140 kDa in S. salivarius 19 to 300 – 600 kDa in H. pylori 5. The isoelectric point was 5.0, very similar to the 5.1 value of urease from Proteus penneri 20. The predicted molecular weight for the three subunits of B. suis urease operon 1: URE-A1 (11,129 Da), URE-B1 (11,377 Da) and URE-C1 (61,054 Da), the total molecular weight expected is 83,520 Da, this result is very closed with the molecular weight obtained in the polyacrylamide gradient non-denaturing gel if B. suis urease could be in a dimer form (2). Molecular weights higher than expected for the three subunits in the denaturing polyacrylmide gel could be a product of aberrant running. Differences between results in gel filtration chromatography compared with polyacrylamide gradient non-denaturing gel could be a product of different conformation of the standard proteins and the possible multimeric conformation of the B. suis urease.

B. suis urease activity was present between pH 6–8 but activity decreased sharply below pH 6. Recently the optimal pH of B. abortus urease was determined as 7.3 21. Similarly, the ureases from Sthaphylococcus saprophyticus, S. salivarium, Ureaplasma urealyticum, and M. tuberculosis have a neutral optimum pH between 7 and 7.5 16192223. In contrast, the ureases from L. reuteri, L. fermentum, L. ruminis, and S. mitior exhibit acidic optimal pH values of 2, 2, 3, and 4.5 respectively 19, while H. pylori and K. aerogenes have a slightly alkaline pH of 8.0 824. We assayed urease activity in Tris-HCl buffer at pH 8.0, as suggested by Dunn and co-workers 8. Although Tris has been described as an inhibitor of urease, a simple control experiment comparing the activity in Tris-HCl and the synthetic buffer HEPES, showed no significant difference (results not shown).

In terms of temperature stability, ureases from different species of Lactobacillus have a very high optimal temperature between 55 to 65°C 19, and M. tuberculosis is stable between 22 to 60°C but is inactivated above 60°C 16. B. suis urease was active between 10–45°C, and it was almost completely inactivated above 45°C. The optimal activity was observed between 28–35°C, temperatures that are close to the corporal temperature of humans and other animals who are hosts for this pathogen.

The kinetic study of B. suis urease showed a simple Michaelis-Menten-type kinetic behavior (Figure 4), in accordance with what has been described for ureases from other organisms 5. The K_m obtained by hyperbolic regression was 5.60 ± 0.69 mM, and a very similar value (5.24 mM) was obtained with the still widely used Lineweaver-Burk or double reciprocal plot (Figure 4). Hence, B. suis urease shows a higher affinity for its substrate than the enzymes from B. abortus (13 mM) 21, P. mirabilis (39 mM) or Providencia stuartii (12 mM) 25, but less affinity than H. pylori whose urease has one of the lowest K_m values (0.3 mM) reported 8. A limited correlation may exist between the value of this kinetic constant and the ecological niche of the host organism. For example, H. pylori inhibits the gastric mucosal lining, where low concentrations of urea are present (1.7 to 3.4 mM); such a low kinetic constant would allow this urease to function under close to saturation conditions despite the low substrate concentration. In contrast, microorganisms that are exposed to large amounts of urea such as are found in the urinary tract or in the soil typically possess ureases with higher K_m values, such as 13, 60, and 40 to 130 mM observed for the enzymes from P. mirabilis, Sporosarcina ureae, and Bacillus pasteurii, respectively 5.

Several compounds have been described as inhibitors of urease (for a review see 526). These are potentially important to control urease related pathogenesis of bacterial infections and in enhancing the efficiency of urea fertilizers 5. We tested the inhibitory effect of three of these compounds, hydroxyurea and thiourea, substrate analogues, and AHA. Both hydroxyurea and thiourea show a competitive type of inhibition (Figure 5), with K_i of 1.04 ± 0.31 mM for hydroxyurea and 26.12 ± 2.30 mM for thiourea. These results agree well with what is described for ureases from other bacteria: the enzyme from K. aerogenes is inhibited by thiourea with a K_i > 25 mM, and hydroxyurea is a competitive inhibitor of the Brevibacterium ammoniagenes urease with a K_i of 0.23 mM 26. For AHA, we used a similar approach and obtained a competitive inhibition with K_i of 0.77 ± 0.06 mM. AHA has been described as a potent inhibitor of ureases from bacteria and other microorganisms 5 and the study of urease from jack bean and K. aerogenes revealed that AHA is a slow-binding competitive inhibitor with a very low K_i, 2.6 – 4 μM 2627. Our results confirm that AHA is an inhibitor of B. suis urease, with a lower K_i than thiourea and hydroxyurea. To test AHA as a slow binding inhibitor is beyond the scope of this study.

Due to the low protein concentration of the low molecular weight bands (Figure 2), it was not possible to sequence them. Since the active urease apoenzyme consists of UreA, UreB and UreC, it is likely that the smaller bands correspond to UreA and UreB. In the case of 66 kDa protein, we obtained the amino acid sequence MPARISRATYAQM; this sequence matched with B. suis UreC1 subunit (BR0270).

During the last years, several immunoreactive Brucella proteins have been identified and evaluated against animal and human serum samples 282930. It is well known that urease is the most abundant protein of H. pylori and as a result it serves as a significant immunogen and is useful in diagnostic tests. Patients with active gastritis due to H. pylori show significantly elevated immunoglobulin G and A sera titers to the urease 5. In the current study, B. suis urease was tested for reactivity with human sera from patients with acute and chronic brucellosis by western blot analysis. The IgG antibodies in sera from patients with acute and chronic brucellosis reacted against the large subunit while the lower molecular weight subunits were not recognized. It is possible that the smaller subunits (UreA and UreB) did not react because their concentration was too low or that they are not immunogenic. As expected, healthy individuals were seronegative for reactivity to the UreC1 (data not shown). This is somewhat analogous to the serum reactivity of patients recovering from Y. entercolitica infections in which the UreB subunit is recognized 11.

Although the sequenced genomes of the Brucella spp. have two urease operons (ure-1 and ure-2), it appears that only one is responsible for encoding an active urease. In the case of B. abortus, Tn mutagenesis produced mutations of ure-1, but not ure-2, resulting in decreased urease activity 31. Similarly, in the case of B. suis, the disruption of the ureC2 or ureB2, genes does not affect measurable urease activity in B. suis 16. Thus, it is perhaps not unexpected that the urease purified in this study was identified as that encoded by ureC1 based on N terminal amino acid sequencing. Moreover, the serum from humans diagnosed with brucellosis also reacted with UreC1 and supports the mutational studies that only the ure-1 operon is used to produce urease during a Brucella infection. It is also noteworthy that the serum from patients infected with Y. enterocolitica also shows reactivity to the UreB subunit and is a candidate antigen involved in the induction of reactive arthritis in the same patients 32. This ReA syndrome is also a typical side effect of a significant proportion of patients who have been infected with Brucella 3334 or with a number of other pathogens acquired through the gastro-intestinal tract 35. The question remains: why are the bacterial ureases so immunodominant and are they responsible, at least in part, for the induction of ReA as suggested by the arthritogenic peptide theory 36.

PSI-BLAST comparisons showed that UreC1 from B. suis has 99% of similarity with either the putative UreC1 from B. melitensis or UreC1 from B. abortus (encoded by BMEI0647 and BruAb1_0296 genes, respectively). Also, high similarity values were obtained by comparison with the alpha subunits from M. loti (81%), R. leguminosarum (78%), S. meliloti (79%) and A. tumefaciens (78%). In contrast, the comparison with UreC2 from B. suis, B. melitensis and B. abortus (BR1358, BMEI1652, and Bru Ab1_1355 genes, respectively), yielded lower values (around 55%, in each case). To find the relationships of the multiple urease clusters of B. suis, we carried out a phylogenetic and structure analysis with ureases from bacterial species with unique or multiple clusters.

The phylogenetic tree deduced for B. suis UreC1 (BR0270, denoted as B. suis 270) grouped it within the Rhizobiales order (Figure 6, left). The closely related clades included the typical UreC from other alpha (not shown), beta (Burkholderia thailandensis), and gamma (P. syringae and P. syringae tomato) proteobacterial species. However, in the other main clade, the UreC2 (BR1358, denoted as B. suis 1358) was grouped with Yersinia and Photorhabdus species, both from the gamma proteobacteria subdivision [bootstrap value = 100]. The unexpected phylogenetic relationship of cluster 2 possibly reflects a horizontal gene transfer event, because both Brucella and Yersinia share the niche (the mammalian cell), are classified as facultative intracellular pathogens and can replicate inside macrophages 137. In a closely related clade appeared two UreC from Bradyrhizobium BTAi1 (denoted as 1962 and 4442) and one from Pseudomonas syringae strain B728a (PSYR2197). A similar explanation of horizontal gene transfer could be applied, because these species are common inhabitants of the soil and the plant rhizosphere. Also in this group were the UreC subunits of the distant Actinobacteria species Streptomyces coelicolor and S. avermitilis.

Interestingly, the main group relationships observed in the phylogenetic tree for UreC were also revealed in the structures of the urease clusters. In the alpha proteobacteria subdivision (including Silicibacter pomeroyi, Rhodobacter sphaeroides, Rhodopseudomonas palustris, not shown), the Rhizobiales and B. suis, the cluster 1 structure has the order ureDABCEFG (Figure 6, right). A slight variation, with the structure ureDABCFG, is present in B. japonicum, in one of the clusters of Bradyrhizobium strain BTAi1 (denoted as 7009), in some gamma proteobacteria such as Pseudomonas aeruginosa (not shown) and in one of the two clusters of P. syringae strain B728a; in two last species, the ureFG genes are located several genes downstream. A second main structure, with the order ureABCEFGD, was shared by B. suis cluster 2 and the Yersinia pestis, Y. pseudotuberculosis and Photorabdus luminiscens clusters and with one of the clusters of Streptomyces coelicolor and S. avermitilis (Actinobacteria). Lastly, a third structure was found with the order ureEF(AB)CGD. This structure was present in two of the clusters of Bradyrhizobium strain BTAi1 (genes 1962 and 4442) and in one cluster of P. syringae strain B728a. In this case (AB) symbolizes the fusion of ureA and ureB genes. In Streptomyces coelicolor and S. avermitilis the urease clusters consist of only ure(AB)C genes. The conservation of the cluster and the coherence with the phylogenetic tree among the analyzed species suggests a coordinated evolution of genes and clusters and, perhaps, a relationship between origin and function.

We have for the first time purified and characterized an active urease from B. suis. The enzyme was characterized at the kinetics, immnunological and phylogenetic levels. Our results confirm that the active urease of B. suis is product of ure-1 operon.

Acetohydroxamic acid: AHA; HEPES: 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid; 2-mercaptoethanol: 2-ME; sodium dodecylsulfate – polyacrylamide gel electrophoresis: SDS-PAGE.

AC-R carried out the laboratory work and wrote the draft of the manuscript. JQ-L and EA-C carried out part of the purification and biochemical characterization. AMM performed the determination of kinetic parameters and inhibition studies. HP performed the phylogenetic and gene cluster analyses. AL-M, NS and SMB conceived of the study and participated in its design, coordination and helped to write the final manuscript. All authors read and approved the manuscript.