The type IV secretion system VirB
Perhaps one of the most studied factors required for Brucella trafficking diversion is the T4SS encoded by the VirB operon (de Jong & Tsolis, 2012). The VirB genes are homologous to those of type IV DNA transfer systems known from other Gram-negative bacteria such as A. tumefaciens or Legionella pneumophila (Alvarez-Martinez & Christie, 2009). T4SS are classified in two subgroups, IVA and IVB. The latter is closely related to conjugation systems and includes those of pathogens such as L. pneumophila and Coxiella burnetii. The Brucella VirB T4SS apparatus, which belongs to the subgroup IVA, was first identified in B. suis and is encoded by 12 open reading frames (virB1 to virB12) on chromosome II (O'Callaghan et al., 1999). It is essential to the virulence of all Brucella strains investigated in their respective models including the more recently described species B. microti (Hanna et al., 2011). As mentioned earlier, the VirB apparatus is crucial for the intracellular trafficking of Brucella within professional phagocytes and nonprofessional phagocytes including macrophages, DCs and epithelial cells (Fig.1, O'Callaghan et al., 1999; Foulongne et al., 2000; Sieira et al., 2000; Comerci et al., 2001; Delrue et al., 2001; Celli et al., 2003; Kim et al., 2003; Billard et al., 2005; Rajashekara et al., 2006; den Hartigh et al., 2008; Salcedo et al., 2008). BCVs of VirB deficient strains are unable to sustain interaction with the ER. The vacuoles of most of these mutants eventually fuse with lysosomes where they are degraded (Sieira et al., 2000; Comerci et al., 2001; Delrue et al., 2001; Celli et al., 2003; Salcedo et al., 2008; Starr et al., 2008). Interestingly, a polar mutant of virB10 that lacks transcription of downstream genes is degraded in phagolysosomes, whereas a nonpolar mutant of virB10 is recycled to the cell surface (Comerci et al., 2001), suggesting that the escape from the degradative pathway and establishment of the ER-derived niche for intracellular replication are two distinct events mediated by different mechanisms (Comerci et al., 2001).
The inability of virB mutants to establish an intracellular replicative niche reflects their attenuation in the mouse infection model (Hong et al., 2000; Lestrate et al., 2000; Sieira et al., 2000; Kahl-McDonagh & Ficht, 2006; Rajashekara et al., 2006) and in goats (Kahl-McDonagh et al., 2006; Zygmunt et al., 2006). After infection by gavage, virB-deficient B. melitensis present lower bacterial numbers in liver, spleen and intestinal tissues (Paixao et al., 2009). Bypassing the digestive tract using intraperitoneal injection, virB-deficient strains are able to disseminate to lymph nodes, liver and spleen (Rajashekara et al., 2005; Rolan & Tsolis, 2007; Paixao et al., 2009). However, after the initial phase, Brucella virB mutant strains are cleared much faster than wild-type bacteria, suggesting defects in persistence to maintain chronic infection. The difference in survival of virB mutants within infected cells in vitro (eliminated in 24 h) and in vivo (persistence equivalent to wild type for the first few days) could reflect different bacterial localization during mouse infections (e.g. intra- vs. extracellular) or that cells that are able to kill in vitro cannot do so in vivo because, for example, they are regulated by other cells or a particular cytokine environment.
In contrast to wild-type brucellae, virB mutants with a deficient T4SS induce essentially no transcriptional changes in spleen cells. They do not induce inflammation-related genes and the infection remains quiescent, suggesting that T4SS function triggers innate immune responses (Roux et al., 2007). However, the level of response is much lower than that induced by pathogens such as Salmonella (Barquero-Calvo et al., 2007), underlining the immune evasive strategies of Brucella.
The regulation of the virB operon expression correlates to its virulence functions. The B. suis virB operon is expressed maximally in minimal medium (rather than rich medium) at early exponential phase, at temperature of 37 °C and inside of the host cell (Boschiroli et al., 2002). In all Brucella species investigated, an induction of VirB protein expression is observed in response to an acidic environment, and this environmental stimulus seems to account for the major part of induction observed intracellularly (Boschiroli et al., 2002; Rouot et al., 2003). The initial acidification of phagosomes which is essential for B. suis intramacrophage replication (Porte et al., 1999) is required to induce VirB expression. Interestingly, there seem to be some differences in the regulation of VirB expression among Brucella strains: whereas B. abortus, B. melitensis and B. ovis express VirB at neutral pH in a rich medium, B. suis, B. canis and the vaccine strains S19, RB51 (both derived from B. abortus) and Rev1 (derived from B. melitensis) show little to no VirB protein expression in these conditions (Rouot et al., 2003).
Transcriptional regulators of VirB synthesis are the quorum-sensing regulators VjbR (Delrue et al., 2005) and BlxR (Rambow-Larsen et al., 2008), the histidine utilization regulator HutC (Sieira et al., 2010), the transcription factors DeoR, AraC8, AraC2, GntR4 and NolR (Haine et al., 2005), the two-component system (TCS) BvrR/BvrS (Martinez-Nunez et al., 2010), the stringent response regulator encoded by Rsh (Dozot et al., 2006) (see below) and the integration host factor that specifically interacts with the virB promoter during intracellular and vegetative growth (Sieira et al., 2004).
Recent findings suggest that apart from being subject to regulation, mutation of VirB may itself affect expression of other genes. Relative transcription levels of several genes including the quorum-sensing regulator VjbR are downregulated in a B. melitensis VirB mutant during growth in culture and inside macrophages (Wang et al., 2009). Differential expression at transcriptional and post-transcriptional levels can also be observed with several outer membrane proteins, which may explain higher sensitivity of virB mutants to polymyxin B and several environmental stresses (Wang et al., 2010).
Substrates of the T4SS
Bacterial proteins (‘effectors’) translocated into the host cell are important elements of T4SS in the intracellular survival of pathogens. Although the VirB Brucella T4SS was first described more than 10 years ago, few potential effector proteins have been identified and only one for which a function has been ascribed (Table 1). Two proteins that are translocated into the macrophage cytoplasm in a T4SS-dependent manner are VceA and VceC (de Jong et al., 2008). Translocation of VceA and VceC fused with the TEM1 β-lactamase (N-terminus) into mouse macrophage-like J774 cells can be detected from 7 h onwards and is dependent on the last C-terminal 20 amino acids, which contain a motif reminiscent of the secretion signal identified in Agrobacterium T4SS substrates. The open reading frames of VceA and VceC are conserved in all Brucella species sequenced, including B. suis, canis, ovis, abortus and melitensis (de Jong et al., 2008). VceA and VceC are co-regulated with the T4SS as they belong to the VjbR regulon (de Jong et al., 2008), which controls expression of several virulence-associated factors including VirB. The cellular targets and the functions of VceA and VceC remain unknown.
Table 1. Brucella proteins translocated into host cells during infection
|Name||ORFa||Signals required for secretion/translocation||T4SS translocation||Tag used||Function||Reference|
C terminus required for translocation
N-terminus Sec signal present
|Yes||TEM1||Unknown||de Jong et al. (2008)|
|VceC||BAB1_1058/BMEI0948||C terminus required for translocation||Yes||TEM1||Unknown|| |
|RicA||BAB1_1279/BMEI0736||nd||Yes||TEM1||Interacts with Rab2||de Barsy et al. (2011)|
|BPE123||BAB2_0123/BMEII1111||N-terminus required for translocation (Sec signal)||Yes||CyA and 3FLAG||Unknown||Marchesini et al. (2011)|
The third Brucella protein found to be translocated into the cytoplasm of RAW264.7 macrophages is RicA (de Barsy et al., 2011). It has been identified because of its interaction with the guanosine diphosphate (GDP)-bound form of the eucaryotic protein Rab2, a small GTPase that is crucial for Brucella intracellular replication (Fugier et al., 2009). A B. abortus ricA deletion mutant replicates faster in HeLa cells and shows an accelerated loss of LAMP1 from its BCVs than wild-type bacteria. The ricA mutant resides in BCVs, which recruit less GTP-locked Rab2 (de Barsy et al., 2011) suggesting that RicA plays a role in control of BCV maturation. However, the fact that inhibition of Rab2 blocks intracellular replication but the ricA mutant is not attenuated in virulence suggests other effector proteins are involved in the control of Rab2 function. At this stage, it is not clear what is the activity of RicA on Rab2. Results indicate that RicA does not act as a guanine nucleotide exchange factor catalysing the replacement of GDP by GTP on Rab GTPases. Because RicA interacts with the GDP-bound form of Rab2, it is unlikely to be a GTPase-activating protein (GAP) that stimulates GTP hydrolysis by converting the GTPase back to its GDP-bound form. It is possible that RicA is functioning as a guanine nucleotide dissociation inhibitor (GDI) that stabilizes the inactive Rab form by preventing GDP dissociation or GDI displacement factor that removes GDI and allows membrane insertion of Rab2 by its geranylgeranyl anchor. Alternatively, RicA could be interacting with Rab2 without having a direct control on its activity.
Translocation of TEM1-RicA is observed in the wild-type strain but not in the virB mutant and only 24 h after infection of RAW264.7 macrophages. At this time point, there is strong attenuation of virB mutants, which are in very different compartments than the virulent strain, as they are degraded in lysosomes. This could account for the lack of translocation of TEM1-RicA in the virB mutant. Interestingly, there is no obvious C-terminal motif like for VceA and VceC, and secretion of RicA into the culture media was independent of the T4SS. It is possible that the in vitro secretion assay used in this study induces an alternative pathway for effector proteins to cross the bacterial membranes. Another explanation could be that the secretion across the bacterial membranes is independent of VirB and uncoupled from the translocation across the vacuolar membrane. This has recently been shown to occur in the case of T3SS (Akopyan et al., 2011). Alternatively, RicA may not be a T4SS effector. Additional work is now necessary to test these hypotheses.
An in silico screen has recently identified four additional substrates of the VirB T4SS (Marchesini et al., 2011). Furthermore, it identified two Brucella proteins that are translocated into macrophages in a VirB-independent manner suggesting that Brucella has another secretion system yet to be identified. Because of the structural similarities of bacterial export machineries and flagella, it has been speculated that in Brucella (normally considered nonmotile), the flagella genes may serve as a secretion apparatus rather than as an organelle mediating bacterial movement (Lestrate et al., 2003). Moreover, the co-regulation of flagellum components with the VirB secretion system and the attenuation of a flagellum mutant during persistence of B. melitensis in the mouse model of infection suggest a role for this organelle in Brucella pathogenicity (Delrue et al., 2005; Fretin et al., 2005). However, it remains to be demonstrated if these proteins are translocated via the flagella system.
Translocation of all proteins identified in the screen was assayed 5 h after infection of murine macrophages, using C-terminal fusions with the adenylate cyclase reporter CyA from Bordetella pertussis. In the case of the T4SS substrate BPE123-CyA, a protein with no predicted function, the concentration of cAMP peaked between 2 and 5 h postinfection consistent with maximal activation of VirB (Sieira et al., 2004). In murine bone marrow-derived macrophages, a 3xFLAG-tagged BPE123 was found in the proximity of BCVs at 4 h after infection. In contrast with VceA and VceC, translocation of BPE123 was dependent on the N-terminal 25 amino acids, which contain a Sec secretion signal. Although no obvious C-terminal motif was identified, BPE123 contains several positively charged amino acids that could constitute a signal for the T4SS. It is curious that for some substrates (e.g. BPE123), translocation by the VirB T4SS would be coupled to Sec-dependent secretion but not for all effectors (e.g. VceA and VceC). T4SS substrates of other bacteria have been shown to depend on a two-step translocation, including the pertussis toxin which contains a sec-dependent signal to cross the inner membrane (Gauthier et al., 2003) and some substrates in Agrobacterium that lack a signal peptide but form a soluble complex with VirJ in the periplasm, which then enables further interaction with T4SS components and translocation across the bacterial outer membrane and host cell membrane (Pantoja et al., 2002).
Other potential translocated effectors
Although Brucella induces a minimal inflammatory response in the host, it has acquired proteins that help modulate host innate and adaptive immune response mechanisms and which may therefore be important in the development of a chronic infection. One such interesting protein contains a Toll/interleukin-1 receptor domain (TIR), which is essential in TLR signalling. It has been designated Btp1 for B. abortus (Salcedo et al., 2008) and TcpB for B. melitensis (Cirl et al., 2008) but for clarity we will refer to it as Btp1 (Brucella TIR-containing protein 1). This nomenclature is in our opinion more appropriate because of the presence of a second TIR-containing protein in the genome of Brucella (Btp2; Salcedo and Gorvel, unpublished results). Btp1 does not have a role in trafficking but is seems to contribute to infection by manipulating intracellular host pathways. Btp1 has been shown to interfere with TLR4- and TLR2-mediated NF-κB activation as well as cytokine secretion (Cirl et al., 2008; Salcedo et al., 2008; Radhakrishnan et al., 2009). Although one study suggested that Btp1 interacted with Myd88 (Cirl et al., 2008), there is also good evidence that it targets the adaptor protein TIRAP/MAL required for both TLR2 and TLR4 signalling. A more recent study describes a stronger interaction between Btp1 and Myd88 when compared with TIRAP and that Btp1 specifically targets the death domain of Myd88 (Chaudhary et al., 2011). Differences observed by various research groups regarding the eucaryotic target of Btp1 may simply reflect the different methodology used, and it is now essential to determine what is the ‘real’ target of Btp1 during infection. Btp1 binds phosphoinositides similarly to TIRAP, so it may be targeted to the plasma membrane and interfere with the Myd88-TIRAP complex (Radhakrishnan et al., 2009) where it can induce ubiquitination and degradation of TIRAP (Sengupta et al., 2010). More recently, Btp1 was shown to modulate microtubule dynamics. Btp1 stabilizes polymerized microtubules and it enhances their rate of nucleation and polymerization (Radhakrishnan et al., 2011). However, most of the work regarding Btp1 has been performed in vitro, by ectopically expressing the protein in host cells. It is now important to determine the cellular localization of translocated Btp1 and its molecular interactions during infection to confirm these hypotheses. During Brucella infection, Btp1 has been shown to contribute to the modulation of TLR2-dependent activation of murine bone marrow-derived DCs (Salcedo et al., 2008) and to decrease TIRAP degradation in macrophage-like J774 cells (Sengupta et al., 2010). Although the btp1 mutant is attenuated neither in any of the cell systems tested so far (cultured macrophages, epithelial cells and DCs) nor in immune competent intraperitoneally inoculated mice, it plays a role in control of the immune response during infection. In intranasally infected mice, specific subsets of lung DCs showed higher maturation levels when infected with the btp1 mutant compared with the wild-type strain (C. Archambaud, S.P. Salcedo and J.P. Gorvel, unpublished results). It will be interesting to determine whether Btp1 is a substrate for the VirB T4SS and whether it has a role in control of microtubules or microtubule-dependent vesicular trafficking during infection.
One poorly characterized candidate effector is BvfA, a small protein of 11 kDa which is unique to the genus Brucella (Lavigne et al., 2005). It has been identified in a random screen using the Yersinia YopP as a reporter system, which induces apoptosis when the fusion protein is translocated into the host cytosol. Although the function of BvfA is unknown, it is necessary for intracellular survival within both human and murine macrophages, and bvfA mutants are highly attenuated in mice. Interestingly, the promoter of BvfA is induced intracellularly upon acidification but it is unclear whether this small protein predicted to be periplasmic is translocated into host cells. It would be worth re-analysing the translocation of BvfA into host cells using the TEM1 or CyA reporter system as it may be another substrate for the VirB T4SS. Its specific role during intracellular trafficking needs to be investigated.