PE is a functional domain responsible for ... - Wiley Online Library

Molecular Microbiology (2007) 66(6), 1536–1547

doi:10.1111/j.1365-2958.2007.06023.x First published online 19 November 2007

PE is a functional domain responsible for protein translocation and localization on mycobacterial cell wall

Alessandro Cascioferro,1 Giovanni Delogu,2 Marisa Colone,3 Michela Sali,2 Annarita Stringaro,3 Giuseppe Arancia,3 Giovanni Fadda,2 Giorgio Palù1 and Riccardo Manganelli1* 1 Department of Histology, Microbiology and Medical Biotechnologies, University of Padua, Padua, Italy. 2 Institute of Microbiology, Catholic University of Sacred Heart, Rome, Italy. 3 Department of Technology and Health, Istituto Superiore di Sanità, Rome, Italy. Summary The PE family of Mycobacterium tuberculosis includes 98 proteins which share a highly homologous N-terminus sequence of about 110 amino acids (PE domain). Depending on the C-terminal domain, the PE family can be divided in three subfamilies, the largest of which is the PE_PGRS with 61 members. In this study, we determined the cellular localization of three PE proteins by cell fractionation and immunoelectron microscopy by expressing chimeric epitope-tagged recombinant proteins in Mycobacterium smegmatis. We demonstrate that the PE domain of PE_PGRS33 and PE11 (a protein constituted by the only PE domain) contains the information necessary for cell wall localization, and that they can be used as N-terminal fusion partners to deliver a sufficiently long C-terminuslinked protein domain on the mycobacterial cell surface. Indeed, we demonstrate that PE_PGRS33 and Rv3097c (a lipase belonging to the PE family) are surface exposed and localize in the mycobacterial cell wall. Moreover, we found that PE_PGRS33 is easily extractable by detergents suggesting its localization in the mycobacterial outer membrane. Beyond defining the cellular localization of these proteins, and a function for their PE domains, these data open the interesting possibility to construct recombinant mycobacteria expressing heterologous antigens on their surface for vaccine purposes. Accepted 19 October, 2007. *For correspondence. E-mail [email protected]; Tel. (+39) 049 8272366; Fax (+39) 0498272355.

© 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd

Introduction One of the most interesting characteristics of the Mycobacterium tuberculosis genome is the presence of two multigene families that together represent almost 10% of the chromosomal coding potential. These two protein families, named PE and PPE, are characterized by conserved N-terminal domains of about 100 and 180 amino acids, respectively (Cole et al., 1998). The name PE is derived from the signature motif Pro–Glu, while PPE represents the Pro–Pro–Glu motif usually present within the first 10 amino acids of these proteins. PE represents the largest of these two families. One hundred PE genes have been annotated in the genome of M. tuberculosis H37Rv strain (Cole et al., 1998), although some of these genes appear to be frame-shifted or not functional (http://genolist.pasteur.fr/ TubercuList/). Based on their structure, PE proteins can be divided into three main classes: (i) in the first class (17 members) the PE domain is followed by a C-terminal region with a unique sequence that can be as large as 400 amino acids (PE-unique sequence); (ii) the second class (20 members) is composed by proteins constituted by the PE domain (PE-only), whose structural gene is usually followed by a gene encoding a PPE protein (Strong et al., 2006; Tundup et al., 2006); and (iii) the third class is the most abundant (61 members) and contains proteins where the PE domain is linked at C-terminus with a highly variable Gly–Ala-rich sequence (PE_PGRS) (Cole et al., 1998; Brennan and Delogu, 2002). Some members of the PE and PPE structural genes are located within the ESAT-6 (esx) gene cluster regions, which encode novel secretion systems. In a recent work Gey van Pittius et al. (2006) showed that the PE–PPE structural gene couples associated to the ESAT-6 clusters represent the ancestral members of these multigene families and that their expansion is linked to ESAT-6 clusters duplication. Interestingly, while PE–PPE couples associated to ESAT-6 clusters are widespread among actynomicetes, PE_PGRS seems to be the result of a recent evolutionary event, and are present in multiple copies only in members of the M. tuberculosis complex and some of their closest

PE domains in mycobacteria 1537 relatives such as Mycobacterium ulcerans and Mycobaterium marinum (Gey van Pittius et al., 2006). Despite the great interest raised by PE proteins in the scientific community, their function remains largely unknown and the hypothesis that have been proposed in the last few years awaits experimental supports. It has been suggested that PE_PGRS may be implicated in virulence (Ramakrishnan et al., 2000) and in the immune evasion strategies deployed by M. tuberculosis to persist in host tissues, either by generating antigenic variations or by inhibiting the class I antigenic processing machinery (Cole et al., 1998; Brennan and Delogu, 2002). Most of the studies have been carried out on PE_PGRS33, which has been used as a model for this subfamily. It has been demonstrated that this protein is implicated in the interaction of the bacilli with macrophages (Brennan et al., 2001). More recently it has been shown that expression of PE_PGRS33 in Mycobacterium smegmatis, which does not naturally express any PE_PGRS, enhances the virulence properties of these bacteria by triggering cell death and bacterial persistence (Dheenadhayalan et al., 2006). Basu et al. (2007) recently demonstrated that PE_PGRS33 triggers apoptosis in macrophages by directly interacting with Toll-like receptor 2 (TLR2) and inducing secretion of tumour necrosis factor (TNF). Interestingly, these authors showed that interaction is mediated by the PGRS domain, that this property is specific for PE_PGRS33 (as two other PE_PGRS proteins could not induce apoptosis), and that polymorphism in the PGRS domain affects the ability of the protein to trigger TNF secretion (Basu et al., 2007). These results are in agreement with the hypothesis suggested by us and other authors that PE_PGRS proteins localize in the mycobacterial cell wall and are available on the mycobacterial cell surface, despite the fact that no known signal peptide or transmembrane region are predicted in their sequence (Brennan et al., 2001; Banu et al., 2002; Delogu et al., 2004). Despite the paramount importance of a better characterization of these proteins, final proof of their localization is still awaited due to the complex structure of the mycobacterial cell wall and to the lack of specific antibodies for the PE proteins. Moreover, while some function of the PGRS domain has been characterized (at least for PE_PGRS33), nothing is known about the function of the PE domain. PE-only proteins have been shown to strongly interact with the coexpressed PPE proteins, and this interaction was shown to be required for the proper protein folding at least when overexpressed in Escherichia coli (Strong et al., 2006; Tundup et al., 2006). Recently, Abdallah et al. (2006) showed in an elegant paper that PPE41 is surface exposed and partially secreted by the ESAT-6-like secretion system ESX-5, and that the PE protein whose

structural gene is upstream of that encoding PPE41 is essential for its secretion, opening the interesting hypothesis that PE domain could represent a functional domain involved in protein translocation. In this article, we study the cellular localization of PE_PGRS33, its PE domain when expressed in the absence of the PGRS domain (PE1818c), and Rv3097c, which has been annotated as a PE_PGRS by Cole et al. (1998) (PE_PGRS63), but should instead be included in the PE-unique subfamily. We could demonstrate that the three proteins localize in the mycobacterial cell wall and are available on the surface. Furthermore, we demonstrated that at least PE_PGRS33 localizes in the mycobacterial outer membrane, that the PE1818c domain, as well as the PE-only PE11 protein, contains the information for the cell wall localization, and that they can be used as fusion partners to localize chimeric proteins on the cell wall of mycobacteria, opening the interesting possibility to construct recombinant mycobacteria for vaccine purposes. Results To investigate the cellular localization of the proteins object of this study, we constructed three expression vectors where the chimeric proteins PE_PGRS33 (Rv1818c), its PE domain (PE1818c) and Rv3097c were fused to the HA epitope (Fig. 1) and expressed under the control of the strong mycobacterial promoter Phsp60. The expression plasmids were then introduced in M. smegmatis mc2155. PE_PGRS33–HA and Rv3097c–HA, but not the PE1818c–HA are exposed on the mycobacterial surface To determine if these proteins were exposed on the mycobacterial surface, M. smegmatis strains expressing the chimeras were subjected to proteinase K treatment. In this assay bacterial cells are washed and shortly exposed to proteinase K. After treatment, the protease is blocked with an inhibitor and washed away. This treatment is able to degrade surface-exposed proteins leaving the cells intact. Cells are then lysed by boiling and the same amount of each cell extract is analysed by Western blot. A strains expressing the wild-type (intracellular) GFP was used as negative control. As shown in Fig. 2, while the intracellular GFP was totally unaffected by the treatment, PE_PGRS33–HA and Rv3097c–HA were degraded (55% and 78%, respectively, as assessed by densitometric analysis) by the proteinase K, showing their availability on the bacterial surface. However, PE1818c–HA was not degraded, suggesting that this protein is not directly exposed on the bacterial surface. To rule out the possibility that PE1818c–HA resistance to proteinase K was due to an intrinsic resistance to this protease and not to its

© 2007 The Authors Journal compilation © 2007 Blackwell Publishing Ltd, Molecular Microbiology, 66, 1536–1547

1538 A. Cascioferro et al. Fig. 1. Schematic representation of the mycobacterial constructs used in this study. The predicted molecular weight for each chimeric protein is indicated on the right.

cellular localization, we repeated the experiment on bacterial cells lysed by sonication: in this case PE1818c–HA was efficiently degraded (data not shown). Subcellular localization of PE chimeric proteins To better characterize the localization of the chimeric proteins, we developed a cell subfractionation protocol. GFP was used as a marker for cytoplasmic proteins, while a fusion between the 5′-region of htrA, and the sequence encoding the HA epitope was used to obtain a cytoplasmic membrane marker. The same strategy was used to tag the secreted protein Mpt64 with the HA epitope (Fig. 1). A strain expressing the PE1818c–GFP chimera, previously shown to be associated to the insoluble fraction and hypothesized to localize into the cell wall (Delogu et al., 2004), was included in the experiment. M. smegmatis strains expressing the chi-

meric proteins were lysed by sonication and cell lysates were subjected to cell subfractionation by fractionated centrifugation as described in Experimental procedures and analysed by Western blot. The same amount of ‘cells equivalents’ was loaded in each well in order to be able to compare the amount of proteins in each cell fraction. As expected, wild-type GFP was only found in the soluble fraction representing the cytoplasm; Mpt64–HA was found mainly in cytoplasm and supernatant, suggesting that this protein accumulates in the cytoplasm before being secreted; HtrA–HA was found mainly in the cytoplasmic membrane fraction (as expected) and in the cellular debris fraction, while the PE1818c–GFP chimera was mainly associated to the cellular debris fraction (Fig. 3). These data suggest that the cellular debris fraction obtained with this protocol contains almost all the cell wall-associated proteins and part of the cytoplasmic Fig. 2. Proteinase K sensitivity assay performed on M. smegmatis recombinant strains expressing GFP, PE1818c–HA, PE_PGRS33–HA or Rv3097c–HA. Live bacteria were incubated with (+) or without (-) proteinase K. Whole-cell lysates obtained by the recombinant bacteria were separated on SDS-PAGE. Proteins were detected using anti-GFP or anti-HA monoclonal antibody.


PE domains in mycobacteria 1539

Fig. 3. Cell subfractionation analysis carried out on different cellular fractions of M. smegmatis strains expressing GFP, HtrA–HA, PE1818c–GFP or Mpt64–HA. Lane 1: cytoplasmic fraction; lane 2: membrane fraction; lane 3: cell wall fraction; lane 4: cellular debris fraction; lane 5: culture supernatant. Proteins were detected using anti-GFP or anti-HA monoclonal antibody.

membrane-associated proteins. HtrA–HA, GFP and PE1818c–GFP were not detected in the supernatant fractions (data not shown). Cellular localization of the PE chimeric proteins (Fig. 4) indicated that they localize mainly in the cellular debris and the cytoplasmic fraction. This was less evident for PE_PGRS33–HA which was present in all fraction almost in the same amount. In order to rule out the possibility that localization in the cell debris fraction was due to the formation of inclusion bodies as a result of overexpression, the PE1818c–HA chimera was also expressed by its

physiologic (weaker) promoter. As shown in Fig. 4A expression of a lower amount of the chimera resulted in an higher accumulation of the PE1818c–HA protein in the insoluble fractions (82% versus 57%), the opposite that would be expected in case of precipitation in inclusion bodies. These data strongly suggest that the PE1818c–HA chimera localizes in the cell wall. It is interesting to note the presence of bands migrating at molecular weights multiple of that expected for the PE1818c–HA monomer, suggesting that this protein may form multimers. None of the chimeras was found in culture supernatant.

Fig. 4. Cell subfractionation analysis carried out on different cellular fractions of M. smegmatis strains expressing PE1818c–HA from its physiologic promoter pP (A), or from the strong mycobacterial promoter Phsp60 (sP) (B), Rv3097c–HA (C) and PE_PGRS33–HA (D). Lane 1: cytoplasmic fraction; lane 2: membrane fraction; lane 3: cell wall fraction; lane 4: cellular debris fraction. Proteins were detected using anti-HA monoclonal antibody.


1540 A. Cascioferro et al.

Fig. 5. Immunoelectron microscopy performed on ultrathin cryosections of recombinant M. smegmatis strain expressing PE1818c–HA from its physiologic promoter (A), or from the strong mycobacterial promoter Phsp60 (B), Rv3097c–HA (C), PE_PGRS33–HA (D). The negative control (E) is represented by wild-type M. smegmatis MC2155. Bacteria were labelled with an anti-HA monoclonal antibody.

In support to the data obtained by cell fractionation, ultrathin cryosections of the recombinant M. smegmatis strains expressing Rv3097c–HA, PE_PGRS33–HA and the PE1818c–HA fusion from two the different promoters were subjected to immunoelectron microscopy. The surface of all the recombinant strains was labelled by the anti-HA antibody, showing the localization of the chimeric proteins in the envelope (Fig. 5A–D), while the wild-type strain did not show any labelling (Fig. 5E). In agreement with the cell fractionation data, the M. smegmatis strain expressing PE1818c–HA from its physiologic (weak) promoter showed a higher percentage of surface-associated protein (Fig. 5A) compared with that of the strain expressing the same protein from the Phsp60 (strong) promoter (Fig. 5B) (51.8 ⫾ 13.9% versus 17.0 ⫾ 8.2% respectively; P < 0.01), suggesting that overexpression leads to the cytoplasmatic accumulation of a soluble form of PE1818c–HA. Extraction of surface-lipid associated proteins: PE_PGRS33 is an outer membrane protein The outer part of mycobacterial cell wall is made of complex lipids that form an asymmetric lipid bilayer that

has been proposed to functionally resemble the outer membrane of Gram-negative bacteria (Hong and Hopfinger, 2004). The only protein that has been shown to be associated to this outer membrane is MspA, a porin identified in M. smegmatis and not found in slowly growing mycobacteria. MspA can be extracted from the mycobacterial cell wall using mild detergents such as Genapol (Heinz and Niederweis, 2000; Niederweis, 2003; Mahfoud et al., 2006). In order to investigate if PE protein could be localized in the outer membrane, the M. smegmatis strains expressing the protein chimeras were re-suspended in 0.5% Genapol. Samples were incubated for 30 min at different temperatures, centrifuged, and the supernatants were analysed by Western blot with anti-HA and anti-MspA specific antibodies. In agreement with previous findings, extraction of MspA was temperature dependent, reaching its maximum at 90°C (Fig. 6A). When the M. smegmatis strains expressing the proteins chimeras (PE1818c–HA, PE_PGRS33–HA, HtrA–HA and GFP) were subjected to this protocol, only PE_PGRS33–HA was extracted (Fig. 6B and data not shown). Interestingly, extraction of PE_PGRS33–HA was already at its maximum at the lowest temperature used in the experiment (30°C). The fact that neither GFP nor


PE domains in mycobacteria 1541 Fig. 6. Western blot analysis performed on samples obtained extracting live cells of M. smegmatis MC2155 wt (A) or MC2155 expressing PE_PGRS33–HA (B) at different temperature with the mild detergent Genapol. Anti-MspA (A) or anti-HA monoclonal antibody (B) was used for protein detection.

HtrA–HA was extracted confirmed that cells remained intact during the experiment, validating our results. These data suggest that PE_PGRS33 is localized in the outer membrane and that its association with it is weaker than that of the porin MspA. Surface localization of PE-based chimeric proteins To test the hypothesis that the PE domain is sufficient to drive cellular localization in the mycobacterial cell wall, we used a plasmid expressing PE1818c–GFP (Delogu et al., 2004) and a chimera between the PE-only protein PE11 (Rv1169c) and GFP (PE11–GFP) (Fig. 1). When M. smegmatis strains carrying these plasmids were used in a proteinase K assay, both chimeras were degraded (66% and 98% respectively) indicating their surface exposure (Fig. 7). In order to evaluate if PE needs to be located at the N-terminus of the protein to exploit its function, we constructed another chimeric protein, where the PE1818c domain was fused at the GFP C-terminal region (GFP–PE). When the strain expressing this protein was exposed to proteinase K, GFP–PE1818c was not degraded (Fig. S1). In order to test the possibility to use the PE domain of PE_PGRS33 as a fusion partner to express heterologous proteins on the mycobacterial surface, we constructed a mycobacterial plasmid expression vector designed to express PE1818c-fused proteins under the control of the hsp60 promoter: pSTE1 (Fig. S2). Subsequently, the sequence encoding Mpt64–HA without its signal peptide (205 amino acids) was cloned in pSTE1 in frame with the sequence encoding the PE1818c domain (PE1818c–DMpt64–HA). A construct to express Mpt64–HA without signal peptide (DMpt64–HA) was also constructed as a cytoplasmic control. Mycobacterium smegmatis strains expressing the chimeras were subjected to proteinase K assay. As shown in Fig. 8, 58% of the PE1818c–DMpt64–HA fusion was degraded by the protease, while the corresponding cytoplasmic control was not degraded, showing that the chimera was indeed surface exposed. In order to determine if the size of the sequence fused to the PE domain

could influence its own availability on the surface, pSTE1 was used to produce other two chimeric proteins where smaller portions of Mpt64 (60 and 125 amino acids, respectively) were fused to the PE domain and to the HA epitope. Interestingly, when expressed in M. smegmatis none of the two shorter chimeric proteins were available to the proteinase K degradation (Fig. 8). PE1818c– DMpt64–HA (205) was also expressed in Mycobacterium bovis BCG and the proteinase K experiment was repeated in this bacterium obtaining exactly the same results (Fig. S1). The M. smegmatis strain expressing the PE1818c– DMpt64–HA (205) chimera was finally analysed by immunoelectron microscopy. As shown in Fig. 9, the PE chimeric protein, but not the cytoplasmic control (DMpt64–HA), was clearly associated to the bacterial surface (47.3 ⫾ 16.0% versus 8.0 ⫾ 5.3%, respectively; P < 0.01).

Fig. 7. Proteinase K sensitivity assay performed on M. smegmatis recombinant strains expressing PE11–GFP, PE1818c–GFP or GFP. Live bacteria were incubated with (+) or without (-) proteinase K. Whole-cell lysates obtained by the recombinant bacteria were separated on SDS-PAGE. Proteins were detected using anti-GFP monoclonal antibody.


1542 A. Cascioferro et al. suggesting that PE_PGRS is surface exposed (Brennan et al., 2001; Banu et al., 2002; Delogu et al., 2004). The fact that a chimeric protein containing a PE domain is sensitive to degradation when the fusion partner is GFP (240 amino acids) (Fig. 7), or the PGRS domain (357 amino acids) (Fig. 2), but not when the fusion partner is the HA epitope (only 9 amino acids) (Fig. 2), suggests that only the distal C-terminal region of the chimeric proteins is available on the surface, while the PE domain itself is embedded into the mycobacterial cell wall. This hypothesis is also supported by the results obtained with the PE1818c–DMpt64–HA fusions (Fig. 8). Interestingly, when

Fig. 8. Proteinase K sensitivity assay performed on M. smegmatis recombinant strains expressing PE1818c–DMpt64–HA(60), DMpt64–HA(205), PE1818c–DMpt64–HA(125) or PE1818c–DMpt64–HA(205). Live bacteria were incubated with (+) or without (-) proteinase K. Whole-cell lysates obtained by the recombinant bacteria were separated on SDS-PAGE. Proteins were detected using anti-HA monoclonal antibody.

Discussion PE proteins are divided into three classes: PE_PGRS, PE-only and PE-unique sequence. In this article we study three PE proteins, one for each class: PE_PGRS33, PE11 and Rv3097c respectively. PE_PGRS33 is one of the few PE proteins for which experimental evidences have been produced and that has been used as a model for the PE_PGRS subfamily (Brennan et al., 2001; Dheenadhayalan et al., 2006; Basu et al., 2007). PE11 is a 100-aminoacid protein belonging to a PE–PPE couple (Adindla and Guruprasad, 2003; Strong et al., 2006; Tundup et al., 2006). Rv3097c encodes a PE protein that was annotated as a PE_PGRS by Cole et al. (1998) (PE_PGRS63), although a closer inspection of its predicted amino acid sequence indicates that it is not a PE_PGRS. In fact, the domain downstream PE lacks the typical Gly–Gly–Alarich sequences and a typical 35- to 41-amino-acid-long transition domain containing a consensus sequence that usually links the PE and the PGRS domains. Rv3097c encodes for a protein that has been recently demonstrated to have a lipase activity (Deb et al., 2006) and for this reason is the only one of the PE proteins for which a function could be clearly attributed. These proteins, or portion of them, were fused to easily detectable antigens (GFP and HA), expressed in M. smegmatis, and their availability on the bacterial surface was monitored by proteinase K assays. All but one of the chimeras (PE1818c–HA) were accessible to the proteinase showing their surface exposure (Figs 2 and 7). These results are in agreement with previous findings

Fig. 9. Immunoelectron microscopy performed on ultrathin cryosections of recombinant M. smegmatis strain expressing PE1818c–DMpt64–HA (A) or DMpt64–HA (B). Bacteria were labelled with an anti-HA monoclonal antibody.


PE domains in mycobacteria 1543 the GFP was fused at the N-terminus of the PE1818c domain we could not observe enzymatic degradation. Taken together these results suggest that the PE1818c domain may deliver on the mycobacterial cell surface a sufficiently long C-terminus-linked protein domain, but the PE domain itself is not available on the mycobacterial surface remaining embedded in the lipidic moiety of the mycobacterial outer membrane or in its periplasmic space. As PE domains are predicted to be hydrophilic (Strong et al., 2006), we favour the hypothesis that the PE1818c domain localizes in the mycobacterial periplasm where it could interact with the peptidoglycanarabinogalactan polymer and the hydrophilic portion of mycolic acids. When the PE proteins subject of this study were analysed using this cell subfractionation protocol, it was clear that they were mainly associated to the cytoplasmic and the cellular debris fractions, similar to what obtained with PE1818c–GFP (Fig. 3). Similar data were also observed for a PE11–GFP fusion (data not shown). The detection of small amount of these proteins in the cytoplasmic membrane and the cell wall fractions confirms that with this protocol we could not completely separate these two compartments which were mainly contained in the cell debris fraction. The only exception was PE_PGRS33–HA, which was found in all fractions in similar amounts. As the same result was obtained also in other laboratories (M. Brennan, pers. comm.), we can hypothesize that this protein has features that make it difficult to characterize its cellular location based on cell fractionation. One possibility is that, being mildly associated with the surface (as suggested by extraction with mild detergents), a significant amount of the protein could be released from the cell wall during sonication and/or during ultracentrifugation, resulting in subfractions contamination. In both cell subfractionation and proteinase K experiments two different forms of PE_PGRS33 were detected. The smaller form might represent a truncated protein lacking the PE domain. The fact that this form is detected also in the insoluble fractions and is sensitive to the proteinase K treatment suggests the presence of a fragile point (probably between the PE and the PGRS domain) causing the cleavage of the protein during cell lysis, as a result of a physical fracture or chemical cleavage by enzymes of cytoplasmic origin. However, we cannot exclude that the PE domain is only required for protein transport and not for anchoring to the surface and that it can be processed after the protein reaches its final location. To exclude that the localization of PE1818c–HA in the insoluble fractions was an artefact (e.g. precipitation in inclusion bodies) due to overexpression, this chimeric protein was also expressed from its physiologic (weaker) promoter. In these conditions most of the protein still

localized in the insoluble fractions while the relative amount of soluble protein decreased (Fig. 4A), suggesting that when the protein is produced from a strong promoter its soluble form accumulates in the cytoplasm probably because of a saturation of the secretion machinery. These data were confirmed by immunoelectron microscopy, which showed a clear association of PE1818c–HA with the mycobacterial surface both when expressed by the physiological promoter and when expressed by the stronger promoter (Fig. 5A–B). In the latter case, some of the protein was also detected in the cytoplasm. Thus, we conclude that the PE1818c–HA chimera accumulates in the cytoplasm before being translocated in the cell wall (similarly to DMpt64–HA that accumulate in the cytoplasm before being secreted in the culture supernatant). Interestingly, we could also detect multimeric forms of PE1818c–HA both in the soluble and in the insoluble fractions. As these multimers were detected on denaturating gels and after boiling with a reducing agent, the interaction among the PE1818c–HA monomers must be very strong. It is possible that also the other PE chimeric proteins could form multimers, but that they were resolved under conditions used for the analysis. The PE-only PE25 was previously shown to interact with the PPE domain of PPE41 (Strong et al., 2006; Tundup et al., 2006), and interaction of PE25 with PPE41 proteins was suggested to be essential for PPE41 secretion (Abdallah et al., 2006; Strong et al., 2006). It remains to be determined whether or not interaction among PE domains or with one of the two PPE encoded by M. smegmatis genome is important for the cellular localization of the PE proteins object of this study. Experiments to investigate this issue are in progress in our lab. Immunoelectron microscopy was used also to analyse PE_PGRS33–HA and Rv3097c–HA cellular localization. Also in this case, both proteins were detected on the bacterial surface, further supporting the hypothesis of their cell wall localization (Fig. 5C–D). The mycobacterial cell wall has a complex structure made of giant macromolecules formed by peptidoglican, arabinogalactan, mycolic acids and other complex sugars and lipids. It has been proposed that mycolic acids form the inner leaflet of an unique outer membrane whose outer leaflet is composed by a variety of extractable lipids, some of which play an important role in the immunopathogenesis of tuberculosis, like for instance the trehalose dimycolate (Minnik, 1982; Liu et al., 1995; Niederweis, 2003). Between the cytoplasmic membrane and the outer membrane is a periplasmic space, where the peptidoglycan is covalently linked to arabinogalactan and lipoarabinogalactan molecules. The protein composition of the outer membrane and periplasmic space is poorly understood. The only characterized outer membrane mycobacterial protein is the porine MspA (Mahfoud et al., 2006). It is


1544 A. Cascioferro et al. usually extracted from the cells after treatment with mild detergents as Genapol at high temperature (Heinz and Niederweis, 2000). Using the same extraction technique, we showed that PE_PGRS33–HA (but not the other PE chimeras) could be extracted from M. smegmatis cells (Fig. 6). This is an extremely interesting finding as PE_PGRS33 is the first M. tuberculosis protein shown to be extractable with mild detergents and hypothesized to be localized in the outer membrane (MspA is not present in slowly growing mycobacteria). Interestingly, while extraction of MspA could be achieved only at the highest temperature (90°C), complete extraction of PE_PGRS33–HA was obtained at 30°C, as increasing the temperature of incubation with the detergent did not result in a higher protein yield (Fig. 6). This finding suggests that this protein is loosely associated to the outer membrane and therefore could be easily released from the bacterium. Nevertheless, we could not detect PE_PGRS33–HA in the culture supernatants (data not shown). It is possible that even mild stresses or perturbations of the mycobacterial outer membrane, like for instance those occurring following interactions with alveolar surfactants, exposure to antibacterial peptides or other antibacterial defences encountered during the infectious process will result in a release of the PE_PGRS33, which could more easily interact with host components (Brennan et al., 2001; Dheenadhayalan et al., 2006; Basu et al., 2007). Our results, showing that PE_PGRS33 is loosely associated with the cell wall support recent findings showing that this protein is released into the extracellular environment via exosomes secreted by dendritic cells and macrophages infected by M. tuberculosis, and that the protein can be found in T cells where it can induce apoptotic signals (Balaji et al., 2007). Interestingly, PE1818c–HA and Rv3097c–HA could not be extracted by Genapol, suggesting that they are embedded more tightly or associated more strongly to the mycobacterial cell wall. We already hypothesized that the PE domain itself is not directly surface exposed and could be deeply embedded in the outer membrane or localize in the periplasmic space; the finding that it was resistant to detergent extraction reinforces this hypothesis. However, Rv3097c–HA is surface exposed and its predicted hydrophobicity pattern for Rv3097c is very similar to that of PE_PGRS33 (data not shown). The different behaviour showed by these two proteins during detergent extraction could depend on the peculiar structure of the PGRS domain that may contribute to the localization of PE_PGRS33. It would be interesting to determine if also other PE_PGRS proteins colocalize with PE_PGRS33 and are easily extractable with mild detergents. The finding that Rv3097c is surface exposed is particularly interesting. Rv3097c encodes for a true lipase

protein, which hydrolyses long-chain triacylglycerol with high specific activity (Deb et al., 2006). It has been proposed that it hydrolyses triacylglycerol accumulated in the cytoplasm and may be upregulated during nutrient starvation stages such as dormancy (Deb et al., 2006), although it remains to be determined whether Rv3097c activity is confined to the mycobacterial cytoplasm. Our results indicating that this protein is (at least in part) associated to the cell wall suggest that this enzyme may be active also outside the mycobacterial cytoplasm. An extracytoplasmatic lipase may be involved in the lipid turnover of the complex outer membrane or in lipids uptake. It has been previously demonstrated that heterologous antigens expressed on the M. bovis BCG surface fused with a lipoprotein are able to simulate the immune response more powerfully than when secreted or expressed in the cytoplasm (Grode et al., 2002). For this reason, after demonstrating that the PE domains can drive the localization of fused protein on the mycobacterial surface, we explored the possibility to use the PE1818c domain as a fusion partner to express heterologous antigens on the mycobacterial cell wall. To develop this antigen delivery system, we constructed an expression shuttle plasmid to facilitate the fusion of sequences encoding heterologous antigen to the PE1818c domain. The sequence encoding the highly immunogenic secreted M. tuberculosis antigen Mpt64, which is absent in M. bovis BCG, lacking its signal sequence, was cloned in this expression vector and introduced in both M. smegmatis and M. bovis BCG. Accordingly with previous results described in this article, the chimeric protein localized in the cell wall and was accessible to proteinase K in both hosts demonstrating that it was indeed surface exposed (Figs 8 and 9). In this article we demonstrate that PE_PGRS33, PE11 and Rv3097c localize into the mycobacterial cell wall and that the PE domains of PE_PGRS33 and PE11 contain the information sufficient to deliver a heterologous antigen on the mycobacterial cell wall, as long as it is found at its C-terminus. Moreover, we showed that PE_PGRS33 localizes on the outer membrane and we predict that our finding that it is extractable with mild detergents will represent an important tool for the future study of this interesting class of still almost unknown proteins. Interestingly, while the PE domains of PE_PGRS33 and Rv3097c belong to the same phylogenetical group (sublineage V, including 77 PE domains and all those found in PGRS proteins) and consequently are phylogenetically very close, the PE domain of PE11 belongs to another phylogenetical group (sublineage IV, including 10 PE domains) (Gey van Pittius et al., 2006). The fact that PE domains not phylogenetically close share the same function suggests that the results presented in this study could be generalized. However, we must take into consideration


PE domains in mycobacteria 1545 that the proteins object of this study represent only 3% of the total members of the PE family suggesting that it is still possible that not all PE proteins have the same cellular localization and the PE domains the same function. Finally, the finding described in this article will open the interesting possibility to use PE domains as fusion partners to express heterologous antigens on the mycobacterial surface. This will allow the design of new recombinant M. bovis BCG strains with increased immunogenic potential against M. tuberculosis, or against other intracellular pathogens.

Experimental procedures

Electroporation of M. smegmatis Bacteria were allowed to reach mid-exponential phase in liquid medium. Culture (50 ml) was centrifuged at 3000 g for 10 min at room temperature and washed twice with 20 ml of 10% glycerol. The pellet was then re-suspended in 1 ml of 10% glycerol, centrifuged again at room temperature at 3000 g and finally re-suspended in 10% glycerol to a final volume of 400 ml. Aliquots (50 ml) of cells were mixed with 500 ng of transforming DNA and loaded into a disposable Cuvette (0.2 cm electrode gap; Eppendorf). The sample was subjected to a single pulse using the Electroporator Gene Pulser Transfection Apparatus (Bio-Rad; capacitance 25 mF; voltage 12.5 kV cm-1; resistance 200 W). After the pulse, the cells were diluted in 900 ml of liquid medium, incubated for 3 h at 37°C and then plated on selective solid medium.

Bacterial strains, media and growth condition Mycobacterium smegmatis mc2155 (Snapper et al., 1990) was grown at 37°C in Middlebrook 7H9 (liquid medium) or 7H10 (solid medium; Difco), both supplemented with 0.05% v/v Tween 80 (Sigma-Aldrich) and 0.2% v/v glycerol (SigmaAldrich). Strains processed for proteinase K assay, cell subfractioning and immunoelectron microscopy were grown in Sauton (Difco). For cloning procedures E. coli strain HB101 was grown in Luria–Bertani medium (LB) (Sambrook et al., 1989). Hygromycin (Roche) was used at a final concentration of 100 mg ml-1 (solid media) or 50 mg ml-1 (liquid media) for M. smegmatis and at a final concentration of 200 mg ml-1 for E. coli.

DNA manipulation All genes expressed in this work were amplified with Pfu DNA polymerase (Stratagene) using the H37Rv genome as template. For the HA-tagged genes upper primers were designed to contain an NheI site immediately before the start codon, while lower primers were designed to contain the HA-coding sequence in frame to the coding sequence of the gene of interest, a stop codon and a BamHI site. All the HA-tagged genes were inserted in the mycobacterial expression vectors pMV10-25 and pMV4-36 (Delogu et al., 2004) digested with NheI and BamHI. To create the 5′ fusion to the GFP coding sequence, upper primers containing an XbaI site before the start codon and lower primers containing XbaI site after the last codon of genes were used. The sequence of the primer is available upon request. For the construction of pSTE1, the 453 bp sequence encoding the PE1818c domain plus the PE/PGRS transition domain was amplified with primers RP86 and RP87. The lower primer was designed to contain an XbaI site immediately before the start codon of the PE coding sequence, while the upper primer was designed to have a polylinker containing BamHI, PacI and NcoI restriction sites before a stop codon and a KpnI site. This fragment was inserted downstream into the Phsp60 promoter of the shuttle vector pMV10-25 (Delogu et al., 2004) after cut with NheI and KpnI, RP86: 5′-GCTCTAGAATGTCATTTGTGGTCACGATCC-3′; RP87: 5′-AAGGTACCTACCCATGGTTAATTAAGGATCCAT TGCCGATCAAGATTCCG-3′.

Proteinase K sensitivity assay Selected strains were grown in 20 ml of medium for 14 h starting from an OD600 of 0.1. Cells were washed once in TBS buffer (Tris HCl pH 7.5, NaCl 150 mM, KCl 3 mM) and re-suspended in 1 ml of the same buffer. Each sample was divided in two identical aliquots, one of which was added with proteinase K (Sigma-Aldrich) up to a concentration of 100 mg ml-1. Both aliquots were incubated for 30 min at 4°C. The reaction was stopped adding complete EDTA-free inhibitor (Roche) 1¥. Subsequently, samples were washed once in TBS and re-suspended in TBS plus loading buffer 5¥ (Sucrose 50% w/v, SDS 10% w/v, Tris HCl pH 6.8312 mM, bromophenol blue 0.05% w/v, b-mercaptoethanol 5% v/v). Finally, samples were boiled for 10 min to allow bacterial lysis and loaded on a polyacrylamide gel in equal amounts. Treated and untreated samples were processed in parallel using the same procedure to allow their comparison. Each experiment was performed at least twice with different biological samples

Cell fractionation Selected strains were grown in 20 ml of medium as previously described. Cells were washed once in 1 mM PBS/ phenyl methane sulphonyl fluoride (Sigma-Aldrich, PMSF) and re-suspended in 1 ml of the same buffer. Each sample was subjected to sonication (power 4; Sonifier B-12 Branson Sonic Power Company) for 30 s at room temperature, in 10 s cycles followed by chilling in an ice bath. The lysates were centrifuged at 3000 g at 4°C to precipitate cellular debris and unlysed cells. Supernatants were transferred to fresh tubes and sedimented at 27000 g for 30 min at 4°C in a BeckmanCoulter ultracentrifuge (L-90K Optima, SW55Ti rotor) in order to allow cell wall precipitation. Once again, the supernatant was precipitated at 100000 g for 2 h to separate cytoplasmic membrane from cytosolic fraction. Cytosolic proteins were subsequently precipitated with trichloroacetic acid 10% v/v, incubated on ice for 30 min, centrifuged for 10 min in a microcentrifuge at 16000 g at 4°C, and finally washed with acetone 80% v/v. All samples were washed once after each step of centrifugation in PBS/PMSF 1 mM and finally re-suspended in an


1546 A. Cascioferro et al. appropriate volume of PBS plus Loading buffer 5¥. Samples were boiled for 5 min before being separated on polyacrylamide gel and subjected to Western blotting as described below. Protein degradation was calculated by densitometric analisis using the Versadoc Imaging System (Bio-Rad) and Quantity one 4.2.3 software (Bio-Rad).

Genapol extraction Selected strains were grown in 20 ml of medium as previously described. Cells were washed once in PBS and their wet weight was determined. For each strain, four aliquots containing 10 mg of cells were re-suspended in 150 ml each of buffer PG05 (0.5% v/v of Genapol x-80; Sigma-Aldrich) (Heinz and Niederweis, 2000) and incubated for 30 min at 30°C, 50°C, 70°C, 90°C. All samples were centrifuged for 10 min at 4°C at 6000 g in a microcentrifuge. Supernatants were transferred to fresh tubes, and added with loading buffer 5¥. In this case samples were not boiled before being loaded on polyacrylamide gel.

SDS-PAGE and immunoblot Proteins were separated on 8% polyacrilamide gels (Sambrook et al., 1989) and than transferred to a polyvinylidene fluoride membrane (PVDF; Bio-Rad) using standard protocol (Sambrook et al., 1989). Proteins were visualized by immunoblotting using either monoclonal antibodies against the HA epitope (Anti-HA.11; Covance, dilution 1:2000) and GFP (Chemicon; dilution 1:2500) or antisera against MspA (Heinz and Niederweis, 2000) (dilution 1:2000). A secondary goat anti-mouse horseradish peroxidase conjugated (Santa Cruz Biotechnology; dilution 1:2000) and a secondary goat antirabbit horseradish peroxidase conjugated (Pierce; dilution 1:2000) were used to detect proteins on polyvinylidene fluoride (Bio-Rad) membrane. The anti-MspA antiserum was kindly provided by Michael Niederweis. The West Dura Signal Kit (Pierce) was used to develop the chemiluminescent signal. Images acquisitions and quantifications were performed using a Versadoc Imaging System (Bio-Rad) and Quantity one 4.2.3 software (Bio-Rad).

Immunoelectron microscopy Mycobacterium smegmatis samples were fixed with 4% paraformaldehyde plus 0.5% glutaraldehyde in PBS, pH 7.4, for 2 h at room temperature, washed in PBS and then infiltrated with 2.3 M sucrose in PBS overnight at 4°C, frozen in liquid nitrogen and cryosectioned following the method by Tokuyasu (1973). Ultrathin cryosections, obtained by Leica Ultracut UCT device (Leica Microsystem), were collected using sucrose and methylcellulose and incubated for 1 h at room temperature with the specific monoclonal antibody HA.11 (Covance; 1:20 diluted) and then with a goat antimouse 10 nm gold conjugates (1:10 diluted, Sigma-Aldrich). Control samples were obtained by omitting the incubation with the specific primary antibody. Finally, ultrathin cryosections were stained with a 2% methylcellulose and 0.4% uranyl acetate solution. Samples were examined with a Philips 208 transmission electron microscope (FEI Company). For the

quantitative evaluation of the PE protein localization in the various strains, at least five micrographs containing 5–10 sectioned cells have been selected for each strain. Then, the gold particles located on the cell wall and those located inside the cell were manually counted (independently by three different operators) and the relative percentages were calculated for each single micrograph. In such a way, a direct comparison between surface and cytoplasmic expression could be made, avoiding experimental variations. Then, the mean value and the relative standard deviation were calculated for each strain.

Acknowledgements This work was supported by EC-VI Framework Contract No. LSHP-CT-2006-036871 (awarded to R.M. and G.D.); by MIUR-COFIN 2006, Grant No. 2006064583 (awarded to R.M.); by the University of Padova, Progetti di Ateneo, Grant No. CPDA047993 (awarded to R.M.); and by the Istituto Superiore di Sanità, Progetto Nazionale AIDS, Grant No. 50G.13 (awarded to G.F.). The authors wish to thank Michael Niederweis for invaluable advice and for the gift of anti-MspA antibody, Michael J. Brennan and Roberta Provvedi for critically reading the manuscript, Stefania Zoncato and Paola Patuzzo for constructing some of the plasmids used in this study.

References Abdallah, A.M., Verboom, T., Hannes, F., Safi, M., Strong, M., Eisenberg, D., et al. (2006) A specific secretion system mediates PPE41 transport in pathogenic mycobacteria. Mol Microbiol 62: 667–679. Adindla, S., and Guruprasad, L. (2003) Sequence analysis corresponding to the PPE and PE proteins in Mycobacterium tuberculosis and other genomes. J Biosci 28: 169–179. Balaji, K.N., Goyal, G., Narayana, Y., Srinivas, M., Chaturvedi, R., and Mohammad, S. (2007) Apoptosis triggered by Rv1818c, a PE family gene from Mycobacterium tuberculosis is regulated by mitochondrial intermediates in T cells. Microbes Infect 9: 271–281. Banu, S., Honore, N., Saint-Joanis, B., Philpott, D., Prevost, M.C., and Cole, S.T. (2002) Are the PE-PGRS proteins of Mycobacterium tuberculosis variable surface antigens? Mol Microbiol 44: 9–19. Basu, S., Pathak, S.K., Banerjee, A., Pathak, S., Bhattacharyya, A., Yang, Z., et al. (2007) Execution of macrophage apoptosis by PE_PGRS33 of Mycobacterium tuberculosis is mediated by Toll-like receptor 2-dependent release of tumor necrosis factor-alpha. J Biol Chem 282: 1039–1050. Brennan, M.J., and Delogu, G. (2002) The PE multigene family: a ‘molecular mantra’ for mycobacteria. Trends Microbiol 10: 246–249. Brennan, M.J., Delogu, G., Chen, Y., Bardarov, S., Kriakov, J., Alavi, M., and Jacobs, W.R., Jr (2001) Evidence that mycobacterial PE_PGRS proteins are cell surface constituents that influence interactions with other cells. Infect Immun 69: 7326–7333.


PE domains in mycobacteria 1547 Cole, S.T., Brosch, R., Parkhill, J., Garnier, T., Churcher, C., Harris, D., et al. (1998) Deciphering the biology of Mycobacterium tuberculosis from the complete genome sequence. Nature 393: 537–544. Deb, C., Daniel, J., Sirakova, T.D., Abomoelak, B., Dubey, V.S., and Kolattukudy, P.E. (2006) A novel lipase belonging to the hormone-sensitive lipase family induced under starvation to utilize stored triacylglycerol in Mycobacterium tuberculosis. J Biol Chem 281: 3866–3875. Delogu, G., Pusceddu, C., Bua, A., Fadda, G., Brennan, M.J., and Zanetti, S. (2004) Rv1818c-encoded PE_PGRS protein of Mycobacterium tuberculosis is surface exposed and influences bacterial cell structure. Mol Microbiol 52: 725–733. Dheenadhayalan, V., Delogu, G., and Brennan, M.J. (2006) Expression of the PE_PGRS 33 protein in Mycobacterium smegmatis triggers necrosis in macrophages and enhanced mycobacterial survival. Microbes Infect 8: 262– 272. Gey van Pittius, N.C., Sampson, S.L., Lee, H., Kim, Y., van Helden, P.D., and Warren, R.M. (2006) Evolution and expansion of the Mycobacterium tuberculosis PE and PPE multigene families and their association with the duplication of the ESAT-6 (esx) gene cluster regions. BMC Evol Biol 6: 95. Grode, L., Kursar, M., Fensterle, J., Kaufmann, S.H., and Hess, J. (2002) Cell-mediated immunity induced by recombinant Mycobacterium bovis Bacille Calmette-Guerin strains against an intracellular bacterial pathogen: importance of antigen secretion or membrane-targeted antigen display as lipoprotein for vaccine efficacy. J Immunol 168: 1869–1876. Heinz, C., and Niederweis, M. (2000) Selective extraction and purification of a mycobacterial outer membrane protein. Anal Biochem 285: 113–120. Hong, X., and Hopfinger, A.J. (2004) Molecular modeling and simulation of Mycobacterium tuberculosis cell wall permeability. Biomacromolecules 5: 1066–1077. Liu, J., Rosenberg, E.Y., and Nikaido, H. (1995) Fluidity of the lipid domain of cell wall from Mycobacterium chelonae. Proc Natl Acad Sci USA 92: 11254–11258. Mahfoud, M., Sukumaran, S., Hulsmann, P., Grieger, K., and Niederweis, M. (2006) Topology of the porin MspA in the outer membrane of Mycobacterium smegmatis. J Biol Chem 281: 5908–5915.

Minnik, D.E. (1982) Lipids: complex lipids, their chemistry, biosynthesis and roles. In The Biology of the Mycobacteria: Physiology, Identification and Classification. Ratledge, C., and Stanford, J. (eds). London: Academic Press, pp. 95–184. Niederweis, M. (2003) Mycobacterial porins: new channel proteins in unique outer membranes. Mol Microbiol 49: 1167–1177. Ramakrishnan, L., Federspiel, N.A., and Falkow, S. (2000) Granuloma-specific expression of Mycobacterium virulence proteins from the glycine-rich PE-PGRS family. Science 288: 1436–1439. Sambrook, J., Fritsch, E.F., and Maniatis, T. (1989) Molecular Cloning: A Laboratory Manual. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press. Snapper, S.B., Melton, R.E., Mustafa, S., Kieser, T., and Jacobs, W.R., Jr (1990) Isolation and characterization of efficient plasmid transformation mutants of Mycobacterium smegmatis. Mol Microbiol 4: 1911–1919. Strong, M., Sawaya, M.R., Wang, S., Phillips, M., Cascio, D., and Eisenberg, D. (2006) Toward the structural genomics of complexes: crystal structure of a PE/PPE protein complex from Mycobacterium tuberculosis. Proc Natl Acad Sci USA 103: 8060–8065. Tokuyasu, K.T. (1973) A technique for ultracryotomy of cell suspensions and tissues. J Cell Biol 57: 551–565. Tundup, S., Akhter, Y., Thiagarajan, D., and Hasnain, S.E. (2006) Clusters of PE and PPE genes of Mycobacterium tuberculosis are organized in operons: evidence that PE Rv2431c is co-transcribed with PPE Rv2430c and their gene products interact with each other. FEBS Lett 580: 1285–1293.

Supplementary material This material is available as part of the online article from: http://www.blackwell-synergy.com/doi/abs/10.1111/ j.1365-2958.2007.06023.x (This link will take you to the article abstract). Please note: Blackwell Publishing is not responsible for the content or functionality of any supplementary materials supplied by the authors. Any queries (other than missing material) should be directed to the corresponding author for the article.
