OUP user menu

Brucella pathogenesis, genes identified from random large-scale screens

Rose-May Delrue, Pascal Lestrate, Anne Tibor, Jean-Jacques Letesson, Xavier De Bolle
DOI: http://dx.doi.org/10.1016/S0378-1097(03)00963-7 1-12 First published online: 1 February 2004


Pathogenicity islands, specialized secretion systems, virulence plasmids, fimbriae, pili, adhesins, and toxins are all classical bacterial virulence factors. However, many of these factors, though widespread among bacterial pathogens, are not necessarily found among bacteria that colonize eukaryotic cells in a pathogenic/symbiotic relationship. Bacteria that form these relationships have developed other strategies to infect and grow in their hosts. This is particularly true for Brucella and other members of the class Proteobacteria. Thus far the identification of virulence factors for Brucella has been largely dependent on large-scale screens and testing in model systems. The genomes of the facultative intracellular pathogens Brucella melitensis and Brucella suis were sequenced recently. This has identified several more potential virulence factors for Brucella that were not found in large screens. Here, we present an overall view of Brucella virulence by compiling virulence data from the study of 184 attenuated mutants.

  • Brucella
  • Virulence
  • Intracellular survival

1 Introduction

The annotated genomic sequences of Brucella melitensis and Brucella suis were recently released [1,2]. These data were long awaited by the Brucella research community with the hope that the secrets of Brucella virulence would be revealed and clearly readable in their sequences. Instead, the sequences revealed that Brucella is surprisingly poor in classical virulence genes, earning it the name of ‘furtive nasty bug’ (for review [3,4]). This highlights the fact that though genomic sequences are of great value, they remain just starting points in the biological understanding of the nature of an organism and are not the end of the story. Therefore, while genome data open an easy door to a priori research, current research on Brucella pathogenesis will benefit from continued use of large-scale random approaches using relevant infection models and genetic tools, such as transposon mutagenesis, promoter-trap systems and proteome analysis.

Brucella pathogenesis is mainly based on its ability to survive and multiply in host cells [3]. Cellular models of infection using professional and non-professional phagocytic cell lines have been developed [5,6]. Once inside either of these cell types, Brucella replicates within a membrane-bound compartment, isolated from the classical destructive endocytic pathway [7,8]. This compartment possesses some characteristics of the endoplasmic reticulum [7,8].

Though cellular models are invaluable in studying Brucella pathogenesis, these models are simplistic compared to using an animal host model. It is well recognized that in vivo approaches are required for full understanding of bacterial pathogenesis [9]. Mouse models of Brucella infection have been developed. The mouse is a model for persistence in the reticuloendothelial system (which occurs in human disease as well as in domestic animals) and it is genetically well-defined, as compared to goats, cattle or swine, the natural hosts [10].

Various infectious models have been used by several research groups to screen for virulence genes using a priori or random approaches. Random screens to identify virulence factors have been performed with three different methods: transposon mutagenesis [1118,44,64], signature-tagged mutagenesis [1922] and differential fluorescence induction (DFI) [2325]. DFI identifies genes specifically induced intracellularly. In addition, DFI gives information about the intracellular environment encountered by the pathogen. Whether or not these genes are required for full virulence needs further analysis. Recently, Köhler et al. used the term virulome to describe the set of genes needed for survival of Brucella in macrophages [12]. We prepared a comprehensive list of genes that have been studied to date to determine their contribution to the virulence of Brucella along with the virulence model used in the virulence determination studies [7187]. Some additional data from our laboratory that have not been published are included as well. This collection consists of 184 unique genes divided into 12 functional classes based on sequence homology (Table 1).

View this table:
Table 1

This table compiles all the mutants ever published as attenuated in one of the three usual infectious models and some unpublished additional mutants from our laboratory

Embedded Image
  • All together this represents a collection of more than 184 unique genes divided into 12 functional classes based on sequence homology. For each mutant, all the infectious models tested are given, when one of them is not attenuated in a given infectious model, the model is written in red.

  • aThis region contains an authentic frameshift.

  • bThis region contains an authentic point mutation causing a premature stop.

  • cORFs encoding proteins underexpressed in B. melitensis strain Rev1 compared to strain 16M growing in rich medium [50].

  • *An uncountable number of intact intracellular mutants in these genes were observed at 48 h post infection in infected HeLa cells as with the parental strain. These mutants were classified in group A in the manuscript of Delrue and co-workers [16].

2 Classical virulence factors

2.1.1 Envelope molecule

Initial contact between Brucella and the host cell occurs obviously between the bacterial cell surface and the cellular plasma membrane. The large number of attenuated mutants with a structural defect in their lipopolysaccharide (LPS) confirms the importance of Brucella outer membrane in virulence. Recently, the LPS O-polysaccharide (O-PS) was shown to be involved in inhibition of phagocytosis, protection against bacterial killing inside the phagolysosome and inhibition of host cell apoptosis [26,27]. Porte et al. [27] suggest, also, that the O-PS could interact with lipid rafts during cell invasion contributing to diverting Brucella-containing phagosomes from the lysosomal pathway [28]. The O-PS is the first molecule clearly shown to be involved in Brucella intracellular entry. However, its cellular receptor has not been identified. It is possible, also, that the O-PS mutant phenotypes are not directly due to the absence of this structure but rather are the result of structural modification of bacterial cell surface ligands [3]. The O-PS contributes along with lipid A and the core of the LPS molecule in protection of Brucella against cellular defenses such as bactericidal cationic peptides and polycations, and humoral defenses such as complement-mediated lysis (for review see [29]). The outer membrane proteins (OMPs), another major constituent of the outer membrane, might also play a role in the protective properties of this membrane as suggested by the sensitivity of an omp19 mutant to cationic peptides [30]. It was also established that the two-component BvrR/BvrS system regulates the lipid A acylation pattern and expression of several group 3 OMPs including Omp3a, a protein known to be involved in Brucella virulence [3133].

It is also clear that Brucella LPS is not only involved in the molecular dialog between bacteria and host cells but acts also as an immunomodulator for protein antigen presentation of MHC class II molecules [34]. Similarly, Omp3a is an immunomodulator, inhibiting production of tumor necrosis factor α during Brucella's infection of human macrophages [35].

2.1.2 Secretion system

In addition to LPS, the type IV secretion system (T4SS) of Brucella encoded by the virB operon is a major virulence factor. The T4SS delivers macromolecules between bacteria and eukaryotic cells, crossing taxonomic kingdom boundaries, by a cell-contact-dependent mechanism [36]. The T4SS of Brucella is involved in the recruitment of lipid rafts for entry of Brucella into macrophages [37] and is required for Brucella to reach its proper niche and to replicate within host cells [16,17,38]. The probable involvement of T4SS at two steps of host–Brucella interaction suggests that more than one effector could be secreted. Recently, it was shown by the use of VirB5- and VirB8-specific antisera that the production of the T4SS differs among Brucella species and in response to environmental stress [39]. These observations suggest that regulation of virB expression is complex and probably that of the secreted T4SS effectors as well.

2.1.3 Flagella

Although Brucella are described as non-motile bacteria, genomic analysis showed that all the structural genes for building a flagellum are present [40]. However, as no chemotactic genes have been detected, the flagellar genes might be cryptic. Nevertheless, transient expression of fliF, flgE and fliC has been demonstrated in rich growth medium as well as intracellular induction of the fliF promoter (D. Fretin and S. Köhler, personal communication). Moreover, it has been shown in our laboratory that Brucella mutants in fliF, flhA, flgI, flgE, fliC and motB are attenuated in BALB/c mice ([20]; unpublished data). In flagellated bacterial species, these genes encode respectively the MS ring of the basal body, the export apparatus, the P-ring, the hook, the flagellin and the motor of the flagellum. Search for a favorable niche is usually the obvious function of flagellum-based mobility but, in pathogenic bacteria, this structure could also play a role in colonization or adhesion. Furthermore, the flagellar apparatus of Yersinia enterocolitica, in addition to having a dedicated role in flagellum biogenesis, is also involved in the transport of YplA bacterial protein into eukaryotic cells that interacts with host cells independently of mobility per se [41].

2.2 Regulation

To survive under various conditions ranging from the open environment to the intracellular milieu of eukaryotic cells, pathogenic bacteria coordinate the expression of an intricate network of factors as an adaptive response. For example, invasion and vacuolar jacking of eukaryotic cells is actively directed by Brucella (for review see [4]). This implies that there is a specific bacterial response to every change in the environment, adapting the bacterium at each step of the infectious cycle. Usually, the adaptive response is mediated by two-component regulatory systems (TCSs) and by transcriptional regulators (TRs). The detection of 21 predicted response regulators (RRs) and 20 predicted histidine kinases (HKs) was performed in the genome of B. melitensis using the PFAM HATPase_C domain and the PFAM response_reg and trans_reg_C domains, respectively. On the basis of the proximity of hk and rr genes in the genome, it is predictable that these RRs and HKs can form 13 HK/RR pairs. The remaining HKs and RRs may belong to phosphorelays, such as the DivK/CtrA pathway. Analysis of polypeptides for the helix-turn-helix motif, the main signature for DNA binding transcription factors of prokaryotes [42], identified 148 TRs in the genome of B. melitensis.

Six TCSs have been shown to affect the virulence of Brucella[12,20,22,43,44]. One of the TCSs, the Brucella BvrR/BvrS system, has an important effect on HeLa cell–pathogen interactions. At the step of penetration, this TCS is involved in the recruitment of small GTPases that are required for actin-dependent cell penetration [45]. After entry into the cell, it inhibits phagosome progress towards lysosomes [7]. No bacterial ligand or eukaryotic receptor has been identified as yet for any step of the infectious cycle. As some envelope components are regulated by BvrR/BvrS [31], these are good candidates for direct ligands of eukaryotic vacuolar molecules.

Among the 148 transcriptional regulators, six have been shown to be involved in Brucella pathogenesis so far [12,15,20,22] (Table 1). Most of them were identified in our lab but have not been published on as yet. One of the regulators we identified and studied in our laboratory belongs to the LuxR family of the quorum-sensing (QS) proteins. This regulator was designated vjbR (vacuolar jacking Brucellaregulator) and is involved in phagosome trafficking and activates the virB operon [15]. As the QS phenomenon links transcriptional regulation to population density, involvement of vjbR at the step of intracellular trafficking suggests that intracellular Brucella coordinates virulence gene expression in response to the accumulation of a pheromone inside a vacuole during trafficking towards the replicative niche. In support of this hypothesis, it has been shown that Brucella produces dodeca-acyl-homoserine lactone (Cl2-HSL) [46]. However as this QS pheromone represses the transcription of the virB operon [46], it seems unlikely that Cl2-HSL can be the activator of transcription by VjbR.

The spoT, ptsP and glnL mutants suggest that two global regulation systems, the stringent response (for review see [47]) and a phosphoenol pyruvate-dependent phosphotransfer transduction pathway (for review see [48]), are involved in the control of Brucella virulence [11,12,15].

2.3 Metal acquisition

Metal ions are known to play at least two major functions in bacterial virulence. They signal bacteria as to their cellular location and are required as co-factors for a wide variety of enzymes. For these reasons, the bacteria must tightly regulate their intracellular levels of metal ions. Mg2+, Zn2+, and Fe3+ transporter genes are involved in virulence, suggesting that these ions are essential for Brucella homeostasis. The role of iron metabolism in virulence of Brucella is still poorly understood. Mutants unable to synthesize DHBA siderophore are unable to cause abortion in the goat model and are not attenuated in tissue culture or mouse models of infection [49]. However, the Rev1 vaccine strain has been shown by proteomics to have altered expression of proteins associated with iron utilization [50].

2.4 Amino acid metabolism

This category is divided into three subclasses (synthesis, transport and unknown function) and contains 26 genes, 19 of which are in the synthesis subclass. From its ‘Brucella virulome’, Köhler et al. suggest that the intracellular replication site of Brucella is poor in amino acids [12]. However, all the mutants in amino acid metabolism that were identified as attenuated in cellular models by Delrue et al. [16] are able to replicate intracellularly (genes annotated by an asterisk in Table 1). This suggests that Brucella encounters an environment poor in nutrients before reaching the replicative compartment. As some amino acid/peptide transporters such as artL and dppA seem to be required during intracellular infection, it suggests that amino acids/peptides are available. It was also shown that minimal medium induces transcription of virB which is necessary for Brucella to reach its replicative niche [51]. In summary, Brucella may activate some critical virulence genes in response to starvation occurring at an early stage during its intracellular trafficking. Failure to resist this sudden lack of nutrients would result in attenuation. This hypothesis remains to be documented by future investigations.

Even though lysine, valine, leucine, isoleucine, serine, threonine, histidine, and cysteine could be synthesized de novo by the bacterium during infection, some of the genes involved in their biosynthesis might be involved in other metabolic processes as well. For example, the carAB mutant might be affected in both pyrimidine and glutamate synthesis. Likewise, a datA mutant might be impaired in both phenylalanine and peptidoglycan biosynthesis [19]. Thus, mutants at branch points in metabolic pathways must be considered carefully and will need further studies in order to decide which metabolic pathway affects virulence.

2.5 Sugar metabolism and transporter

Old data from the 1950s and the information coming from the genomic sequence about sugar metabolism in Brucella have been compiled nicely in a recent review [52]. Maltose, ribose, arabinose, galactose, glucose, glycerol, erythritol and rhizopine (inositol) degradation pathways appear to be essential for Brucella intracellular survival. Among the 25 attenuated mutants, genes encoding sugar transport systems were disrupted in eight, consistent with the replicative niche containing several carbon sources. However, it is not known whether their role is limited to carbon or energy supply or if they play a signal role in affecting gene regulation during the infectious process. Five mutants might be impaired in their pentose phosphate cycle (ppc) (mutants: cbbE, pgi, rbsK, araG, rbsA). Because Brucella lacks phosphofructokinase, an essential enzyme of the glycolysis [52], the ppc pathway is crucial for sugar degradation in Brucella. This pathway also furnishes ribose for nucleic acid synthesis, and it has been demonstrated that the ability to produce de novo purines and pyrimidines is essential for Brucella intracellular replication [12,53].

Virulence of Brucella is also dependent on certain synthetic abilities. For example, production of cyclic β-(1-2) glucan is necessary for Brucella's intracellular replication [4,54]. This compound might be involved in cholesterol sequestration during phagosome trafficking as suggested by Moreno and Moriyon [3]. Recent data suggest that the bacterium might also synthesize and/or catabolize rhizopine [20,23]. Rhizopine (l-3-O-methyl-scyllo-inosamine) is produced by Rhizobium only in the plant nodule [55,56]. This inositol-derived compound is a growth substrate for strains of Rhizobium carrying the moc operon and seems to give an advantage to moc+ strains in the rhizosphere [57].


Though most of the mutants in this category have deficiencies related to purine/pyrimidine synthesis, some of the mutants are more puzzling. A strain generated by Köhler is mutated in miaA[12], which encodes a tRNA δ(2)-isopentenylpyrophosphate transferase responsible for the specific modification of the A-37 residue in the UNN codon tRNA species. It has been shown that a mutation of the miaA gene of Agrobacterium tumefaciens results in reduced virB gene expression [58]. As Agrobacterium and Brucella are phylogenetically related, it would be interesting to study virB regulation in the Brucella miaA mutant. DNA repair systems are likely to play an important role in intracellular persistence, possibly by preventing DNA damage that might be induced by reactive oxygen intermediates. Mutants affecting RNA helicase or DNA gyrase activities suggest that Brucella might also use other strategies as well to regulate virulence genes. For example, tldD encodes a modulator of the DNA gyrase [59], and the modifications in DNA topology caused by the DNA gyrase are known to affect transcriptional regulation of virulence genes in some bacterial species [60].

2.7 Stress

Genes encoding stress proteins have been an obvious choice for directed mutagenesis for virulence studies and vaccine generation [61,8285]. In addition, they have been identified in different screens for attenuated strains. However, the role of htrA in virulence is unclear as contradictory results have been found [12,84,85]. Studies on the DnaK-DnaJ chaperones (BMEI2002-2001) have been initiated by Köhler et al. [61]. The results showed that though DnaK plays a role in intracellular survival, DnaJ does not. While the dnaJ mutant was not attenuated in macrophages, it should be noted that Brucella has four genes encoding proteins having a DnaJ domain (BMEI0047-I0564-I1513 and I2001). One of these genes, BMEI1513, was shown to affect virulence of Brucella in mice, macrophages and HeLa cells [19]. This suggests that DnaK might interact with different DnaJ domain proteins in various environments. Despite the fact that ClpA, ClpB and ClpAB play an important role in stress resistance, they are not involved in virulence [62,63].

2.8 Oxidoredox

Most of the genes in this category are probably involved in reactive oxygen intermediate detoxification, as demonstrated in the study of cydB by Endley et al. [64]. With the sequencing of the genome, it is suspected that Brucella might use nitrate as an electron acceptor, the attenuation of the narG and norE mutants suggests that the Brucella vacuole is deprived of oxygen and that the bacterium switches to anaerobic respiration and uses nitrate as an electron acceptor if available. Preliminary studies showed that Brucella is capable of anaerobic growth in the presence of nitrate [65]. Another possibility would be that these genes are involved in nitric oxide detoxification inside activated macrophages [66].

Proteins of the DsbA family are periplasmic proteins and function as soluble thiol:disulfide oxidoreductases. In a catalytic cascade pathway, the activity of DsbA is maintained by DsbB. DsbA proteins catalyze the oxidative folding and assembly of many secreted proteins, such as cholera toxin and pertussis toxin [67]. The implication of proteins of the DsbAB system in Brucella's effectors secretion remains to be investigated.

2.9 Coding sequences with unknown function

Among all the mutants presented here, 20 are disrupted in sequences for which no function could be assigned. These sequences are of particular interest as some might correspond to secreted effectors involved in vacuolar jacking. For example, BMEI1229 belongs to a cluster of six predicted coding sequences whose low GC content is less than that of the genome. All six sequences are specific for the Brucella genus and the class of rhizobia. As this cluster is bordered on one side by an integrase gene and on the other by a disrupted tRNA gene, it may have been acquired by horizontal gene transfer [68]. The protein encoded by BMEI1227 is a likely candidate for secretion by the T4SS as its carboxy-terminus contains an RPR motif. It has been suggested that this motif forms part of a transport signal for A. tumefaciens T4SS effectors [69,70].

3 Conclusion

In this review, we present a list of all the genes identified currently that affect the virulence of Brucella, regardless of the virulence model used. Because the function or functional group of most of these genes could be determined by homology searches, this list extends our understanding of the genetic basis of Brucella virulence. Of course, further work will be needed to confirm these homology-based predictions. Some sequences that potentially encode virulence factors were discovered by genome sequence analysis that had not been identified by random screens. Analysis of null mutants in these sequences which encoded putative hemolysins, invasins, and adhesins is awaited [2,3].

The next step is to understand the biological functions of the identified virulence factors during an infection. The role of flagella and of rhizopine metabolism in Brucella virulence is particularly intriguing. To adapt to a changing environment during the infectious cycle, pathogens express virulence factors in an orchestrated manner. The challenge now is also to identify all the partners involved in the regulation of expression of virulence factors as well as to draw regulation networks. Global proteomic and transcriptome studies done under conditions encountered by Brucella during its infectious cycle hand in hand with a mutagenesis approach will be critical for a detailed understanding of the genetic basis of virulence of Brucella. When we learn more about the host–pathogen interaction, we will have a better understanding of the brucellosis.


We thank Shirley Halling for critical reading of the manuscript. This work was supported by the Commission of the European Communities, Contracts QLK2-CT-1999-00014 and QLRT-2001-00918.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
  30. [30].
  31. [31].
  32. [32].
  33. [33].
  34. [34].
  35. [35].
  36. [36].
  37. [37].
  38. [38].
  39. [39].
  40. [40].
  41. [41].
  42. [42].
  43. [43].
  44. [44].
  45. [45].
  46. [46].
  47. [47].
  48. [48].
  49. [49].
  50. [50].
  51. [51].
  52. [52].
  53. [53].
  54. [54].
  55. [55].
  56. [56].
  57. [57].
  58. [58].
  59. [59].
  60. [60].
  61. [61].
  62. [62].
  63. [63].
  64. [64].
  65. [65].
  66. [66].
  67. [67].
  68. [68].
  69. [69].
  70. [70].
  71. [71].
  72. [72].
  73. [73].
  74. [74].
  75. [75].
  76. [76].
  77. [77].
  78. [78].
  79. [79].
  80. [80].
  81. [81].
  82. [82].
  83. [83].
  84. [84].
  85. [85].
  86. [86].
  87. [87].
View Abstract