Two FHA domains on an ABC transporter, Rv1747, mediate its phosphorylation by PknF, a Ser/Thr protein kinase from Mycobacterium tuberculosis

Abstract

Bacterial genomics have revealed the widespread occurrence of eukaryotic-like protein kinases in prokaryotes, but little is known about their regulation, endogenous substrates, and physiological role. The present study concerns one of these enzymes, the serine/threonine protein kinase PknF from Mycobacterium tuberculosis. It is shown that, in addition to its autokinase activity, PknF is able to phosphorylate Rv1747, a newly described ABC transporter. This reaction appears to involve two FHA domains of Rv1747. It is suggested that recruitment and phosphorylation of Rv1747 depend on the interaction between its two non-redundant FHA domains and the autophosphorylated form of PknF.

Protein phosphorylation, Ser/Thr protein kinase, FHA domain, ABC transporter, Mycobacterium tuberculosis

• STPK

  serine/threonine protein kinase

• FHA

  forkhead associated domain

• ABC

  ATP-binding cassette

• SBP

  substrate-binding protein

• NBD

  nucleotide-binding domain

• LB

  Luria–Bertani medium

• GST

  glutathione-S-transferase

• IPTG

  isopropyl-1-thio-β-galactopyranoside

• Tris

  tris(hydroxymethyl)aminomethane

• EDTA

  ethylene diamine tetraacetic acid disodium salt

• DTT

  1,4-dithio-d,l-threitol

• SDS—PAGE

  sodium dodecyl sulfate–polyacrylamide gel electrophoresis.

1 Introduction

For many years after the discovery of protein phosphorylation catalyzed by serine/threonine protein kinases (STPKs), the prevailing view was that these enzymes were present only in eukaryotes [1]. However, since the identification of bacterial homologues a few years ago [2, 3], genomics has now demonstrated that serine/threonine protein kinases and phosphatases are also widespread in prokaryotes [4, 5]. In the case of Mycobacterium tuberculosis, analysis of the genome sequence has predicted the presence of 11 different STPKs [6,7]. Eight of these appear to be transmembrane proteins, with a putative extracellular signal sensor domain and an intracellular kinase domain. Among them, PknA, PknB, PknD, PknE, PknF, and PknG have been shown, so far, to catalyze autophosphorylation [8–14] but none has been yet demonstrated to phosphorylate endogenous protein(s). To date, only PknH has been shown to phosphorylate an endogenous substrate, EmbR, which is a putative transcriptional activator that belongs to the OmpR-like family [15]. Moreover, it has been demonstrated that EmbR phosphorylation by PknH is mediated by an FHA (f orkh ead a ssociated) domain [15].

An FHA domain is a phosphopeptide recognition domain with a 55- to 75-amino-acid homology region forming an 11-stranded β-sandwich that mediates specific phosphorylation-dependent protein–protein interaction [16–19]. Although FHA containing proteins are present in all phyla from bacteria to mammals, the specific interaction partners of individual FHA domains are poorly understood. Indeed, till now, phosphorylation of EmbR by PknH constitutes the unique experimental characterization of such phosphorylation-dependent protein–protein interaction in bacteria, even though 64 bacterial FHA proteins are currently listed in the SMART database (http://smart.ox.ac.uk). Since regulatory phosphorylation by STPKs is an emerging theme in prokaryotic signaling, the identification of protein substrates specifically phosphorylated by prokaryotic STPKs is needed to decipher the bacterial signaling pathways. Analysis of the M. tuberculosis genome database has revealed that only six genes (rv0019c, rv0020c, rv1747, embR, Rv1827, and Rv1860) encode putative FHA-containing proteins [20]. Interestingly, among these genes, rv1747 encodes a putative ATP-binding cassette (ABC) transporter [21], and is located in the same operon as pknF, which has been demonstrated to encode a Ser/Thr protein kinase [14]. ABC transporters act as mechanical pumps that couple ATP hydrolysis to either the uptake or export of a wide diversity of substrates across biological membranes including ions, sugars, lipids, peptides, and complex organic molecules [22]. These transporters have a similar topology, with two transmembrane domains and two nucleotide-binding domains, that can be found on either the same polypeptide or as many as four different polypeptides. In the case of ABC importers, an additional subunit, termed SBP for “substrate-binding protein”, is involved in both the capture and the docking of the extracellular substrate onto the transporter [22].

Considering that both Rv1747 and EmbR contain a FHA domain and that the corresponding genes, pknF and rv1747, are adjacent along the chromosome, it seemed of interest to study the possible cross-regulation of PknF and Rv1747. In this work, we describe the overproduction and purification of the nucleotide-binding domain (NBD) of Rv1747 and the assessment of its ATPase activity. Also, the cytosolic kinase domain of PknF has been overproduced and purified and the capacity of this kinase, besides its autophosphorylation activity, to phosphorylate the NBD of Rv1747 has been tested. In addition, the effects of the two FHA domains of Rv1747 on the level of its trans-phosphorylation by PknF have been determined.

2 Materials and methods

2.1 Bacterial strains, plasmids, and growth conditions

Strains, plasmids, and primers used in this study are listed in Tables 1 and 2. Overproduction of NBD_1–559 and PknF_1–292 was performed, respectively, in E. coli strain BL21 (pRep4-groESL) [23] and BL21(DE3) [24]. E. coli DH5 strain [25] was used to propagate plasmids in cloning experiments. All strains were grown and maintained in LB medium at 37 °C. When required, media were supplemented with either 50 μg/ml ampicillin or 25 μg/ml kanamycin.

1

Bacterial strains and plasmids used in this study

Strain or plasmid	Genotype or description	Reference or source
DH5	supE44 lacU169(80lacZM15) hsdR17 recA1 endA1 gyrA96 thi-I relA1	[25]
BL21λ (DE3)	B F−dcm ompT hsdS(rB−mB−) gal (DE3)	[24]
BL21(pRep4-groESL)	F−dcm ompT hsdS(r−Bm−B) gal λ[pRep4GroESL]	[23]
pET44b	E. coli expression vector generating His6 and Nus fusion proteins for overexpression	Novagen
pGEX(M)	pGEX with an 321-bp EcoRI/BamHI fragment from pET19b introducing an HindIII site in the pGEX polylinker	[15]
pGEX(M)-NBD	pGEX(M) with an 1677-bp BamHI/HindIII fragment encoding the nucleotide-binding domain of Rv1747	This study
pET44b-PknF1-876	pET44b with an 876-bp EcoRI/PstI fragment encoding encoding the cytoplasmic domain of PknF1–292	This study
pGEX-NBDR33A	pGEX(M) with an 1677-bp BamHI/HindIII fragment encoding the entire NBD fragment mutated on R33	This study
pGEX-NBDS47A	pGEX(M) with an 1677-bp BamHI/HindIII fragment encoding the entire NBD fragment mutated on S47	This study
pGEX-NBDN69A	pGEX(M) with an 1677-bp BamHI/HindIII fragment encoding the entire NBD fragment mutated on N69	This study
pGEX-NBDR234A	pGEX(M) with an 1677-bp BamHI/HindIII fragment encoding the entire NBD fragment mutated on R234	This study
pGEX-NBDS248A	pGEX(M) with an 1677-bp BamHI/HindIII fragment encoding the entire NBD fragment mutated on S248	This study
pGEX-NBDN270A	pGEX(M) with an 1677-bp BamHI/HindIII fragment encoding the entire NBD fragment mutated on N270	This study
pGEX-NBDE479Q	pGEX(M) with an 1677-bp BamHI/HindIII fragment encoding the entire NBD fragment mutated on E479	This study

2

Primers used in this study

Primer a	5′ to 3′ Sequence b,c
135 (+)	TATGGATCCGTGCCGATGAGCCAACCAGCC
99 (−)	TATAAGCTTCAGTCGTCGGCGACGGTGCTGAAG
94 (+)	TATGAATTCCATGCCGCTCGCGGAAGGTTCG
95 (−)	TATCTGCAGTCACGGTTGCGACACCCGCGT
162 (−)	CGCGCGCAGGTCAGCCCCGACGACGACATC
163 (−)	CAGGTGTGCCCGGGCGATCAGGGGG
164 (−)	GAGGTAGAGCCCAGCGAGGCTGCCATTG
118 (−)	GTTGTCGTTCGCCGCGCCGATCCTGAC
155 (−)	GGCGTGGTGACGTGCGGCCAACACCTCG
156 (−)	GACGAAGGTGCCAGCGATGCTGCGGTTG
160 (+)	GATCCTCGACCAGCCGACATCCG

a

Forward and reverse primers are represented by plus (+) or minus (−), respectively.

b

Restriction sites are italicized.

c

The bases mutated from those present in the wild-type are underlined.

2.2 Construction of GST-tagged NBD domain and Nus-tagged PknF expression plasmids

The 1677-bp Rv1747 gene fragment, with appropriate restriction sites at both ends, was synthesized by PCR amplification using M. tuberculosis H37Rv genomic DNA as a template and primer pair 135/99 (Table 2). This DNA fragment was ligated into vector pGEX(M), thus yielding pGEX-NBD (Table 1). The 876-bp pknF gene fragment corresponding to the cytosolic domain of PknF, with appropriate restriction sites at both ends, was synthesized by PCR amplification using M. tuberculosis H37Rv genomic DNA as a template and primer pair 94/95 (Table 2). This DNA fragment was ligated into vector pET44b, thus yielding pET44b-PknF_1–876 (Table 1).

2.3 Site-directed mutagenesis

Site-directed mutagenesis was carried out based on PCR amplification with the primer presented in Table 1, as previously described [15].

2.4 Overproduction and purification of the Nus-PknF_1–292 domain

PknF was purified on Ni–NTA agarose columns as previously described [15].

2.5 Overproduction of the GST-tagged NBD_1–559 domain and related mutant proteins

E. coli BL21(pRep4-groESL) cells were transformed with the pGEX-NBD vector and derivatives expressing the wild-type or mutated NBD_1–559 domain (Table 1). Recombinant strains harboring the pGEX-NBD vector and derivatives were used to inoculate 100 ml of LB medium supplemented with ampicillin, and were incubated at 37 °C under shaking until A₆₀₀ reached 0.5. IPTG was then added at a final concentration of 1 mM and growth was continued for an additional 3 h at 37 °C under shaking. E. coli BL21(DE3) cells were transformed with the pET44b-PknF_1–876 vector and derivatives expressing the wild-type or mutated PknF cytoplasmic domain (Table 1). Recombinant E. coli strains harboring the pET44b-PknF_1–876 vector and derivatives were used to inoculate 50 ml of LB medium supplemented with ampicillin, and were incubated at 37 °C under shaking until A₆₀₀ reached 0.5. IPTG was then added at a final concentration of 1 mM and growth was continued for an additional 3 h at 37 °C under shaking.

2.6 Purification of the NBD_1–559 domain and related mutant proteins

Cells were harvested by centrifugation at 6000g for 10 min, washed in 10 ml of phosphate buffer saline (PBS), and centrifuged again in the same conditions. The cell pellet was resuspended in PBS. Cells were disrupted by sonication (Branson Sonifier 250). The resulting suspension was centrifuged at 4 °C for 30 min at 30,000g. The supernatant was incubated for 1 h with glutathione–Sepharose 4B matrix (Pharmacia Biotech) at 4 °C. The protein–resin complex was packed into a column and washed with 50 ml PBS. Protein elution was carried out with elution buffer (50 mM Tris–HCl, pH 8.0, 120 mM NaCl) containing 15 mM glutathione. Eluted fractions were analyzed by SDS–PAGE [26]. After dialysis against buffer C (50 mM Tris–HCl, pH 8.0, 120 mM NaCl, 0.2 mM DTT, 0.2 mM EDTA, and 10% glycerol), pure fractions were pooled and tested for in vitro phosphorylation in the presence of PknF. All mutant proteins were purified in the same conditions.

2.7 In vitro kinase assay

In vitro phosphorylation of about 1 μg of PknF_1–292 was carried out for 15 min at 37 °C in a reaction mixture (20 μl) containing buffer P (25 mM Tris–HCl, pH 7.0, 1 mM DTT, 5 mM MgCl₂, and 1 mM EDTA) with 200 μCi/ml [γ-³²P]ATP. Phosphorylation of NBD_1–559 by PknF_1–292 was performed with 5 μg of NBD_1–559 in 20 μg of buffer P with 200 μCi/ml [γ-³²P]ATP and 1 μg of PknF_1–292 for 15 min at 37 °C. The reaction was stopped by addition of an equal volume of 2× sample buffer and the mixture was heated at 100 °C for 5 min. One-dimensional gel electrophoresis was performed as described by Laemmli [26]. After electrophoresis, gels were soaked in 16% TCA for 10 min at 90 °C, and dried. Radioactive proteins were visualized by autoradiography using direct exposure films. When needed, radioactivity was measured with a Molecular Dynamics Typhoon phosphoimager.

2.8 ATPase assay

The ATPase activity of NBD_1–559 was tested in a reaction mixture containing 50 ng protein, 1 μCi [γ-³²P]ATP (20 μCi/mmol), 5 μM ATP, 5 mM MgCl₂, 5 mM MnCl₂, 1 mM β-mercaptoethanol, and 50 mM Tris–HCl, pH 7.5. After 30 min incubation at 37 °C, the products were separated by thin layer chromatography on polyethyleneimine cellulose sheets (Macherey-Nagel) using 0.3 M potassium phosphate buffer, pH 7.4, as solvent. Radioactive signals on the dried sheets were visualized with a STORM phosphoimager.

3 Results

3.1 Overproduction of the nucleotide-binding domain of Rv1747

Analysis of the protein primary structure deduced from M. tuberculosis genome indicated that Rv1747 would belong to the ABC transporter family [21] (Fig. 1). It was predicted to contain 865 amino acids, with a putative nucleotide-binding domain (NBD) characteristic of the ABC transporter family: this includes the highly conserved Walker_A (corresponding to Gly352 to Ser359) and Walker_B (corresponding to Leu474 to Asp478) motifs which form an ATP-binding pocket, and the typical signature of the ABC family, the C-motif, a stretch of ∼12 residues usually starting with the sequence LSGGQ [27] (corresponding to Leu454 to Ala465, see Fig. 1A). Rv1747 was also predicted to contain six putative transmembrane α-helices in its C-terminal region, likely allowing the translocation of an unknown substrate through the membrane (Fig. 1A). Moreover, previous sequence analysis had revealed that the N-terminal region of Rv1747 possesses a putative FHA domain highly conserved among orthologues of Rv1747 [20], spanning from residue Glu224 to Phe299. Interestingly, in a more detailed analysis of the primary sequence of Rv1747, we identified a second region, spanning from residue Phe23 to Ala98, and sharing strong homology with a FHA domain (Fig. 1A and B). From this observation, the hypothesis was made that Rv1747 would possess in fact two FHA domains, FHA1 and FHA2 (Fig. 1A), as previously reported in the case of the Rad53 yeast kinase [28]. Considering the possibility that Rv1747 might be a substrate for the serine threonine/kinase PknF, an attempt was made to overproduce and purify the NBD of Rv1747. For that purpose, the truncated Rv1747 gene encoding the NBD (residues 1–559) was synthesized by PCR amplification using genomic DNA from M. tuberculosis H37Rv. The amplified product was then cloned into plasmid pGEX(M) [15] to yield pGEX(M)-NBD and used to transform E. coli cells. The analysis of the GST-chimeric protein by SDS–PAGE revealed that the protein was expressed in a soluble form and was migrating as a unique band with the predicted molecular mass of 85.5 kDa.

1

Structural domains of Rv1747. (A) Schematic presentation of the Rv1747 domains. The M. tuberculosis Rv1747 gene encodes a predicted ABC transporter which includes the highly conserved Walker_A and Walker_B motifs, a C-motif, and six putative transmembrane α-helices (TM) in the C-terminal region. Rv1747 is also predicted to contain two FHA domains in the N-terminal region. (B) Sequence alignment of FHA domain-containing proteins. Sequences were aligned using ClustalW program. Alignments are shaded using the BOXSHADE server. Sequence designations and NCBI GI numbers are as follows: Rad53/FHA1: FHA domain from S. cerevisiae Rad53, GI 134835; EmbR from M. tuberculosis, GI 15608407; and Rv1747 from M. tuberculosis, GI 15608885. Black boxes represent identical residues and grey boxes indicate similar residues.

3.2 NBD_1–559 exhibits ATPase activity

In order to check whether the recombinant protein was able to hydrolyze ATP, NBD_1–559 was incubated in a reaction mixture containing [γ-³²P]ATP. As shown in Fig. 2A (lane 1 vs. lane 2), free radioactive inorganic-phosphate (Pi) was released from [γ-³²P]ATP in the presence of NBD_1–559, suggesting that this domain exhibits an intrinsic ATPase activity. To examine the possibility that the ATPase activity thus measured might be due to a contaminant enzyme activity, the catalytic Glu of this ABC transporter [29,30] (corresponding to Glu479 in Rv1747, see Fig. 1A) was mutated to a Gln residue. The mutant domain thus prepared did not exhibit any ATPase activity. In addition, this mutant showed a strong tendency to bind ATP [30], resulting in a intense radioactivity spot observed in the loading zone, (Fig. 2A, lane 3).

2

ATPase activity of NBD_1–559. (A) ATP hydrolysis by NBD_1–559 was carried out in a reaction mixture containing 50 ng protein, 1 μCi [γ-³²P]ATP (20 μCi/mmol), 5 mM MgCl₂, 5 mM MnCl₂, 1 mM β-mercaptoethanol, 5 μM ATP, and 50 mM Tris–HCl, pH 7.5. After 30 min at 37 °C, the products were separated by thin layer chromatography on polyethyleneimine cellulose sheets (Macherey-Nagel) using 0.3 M potassium phosphate buffer, pH 7.4, as solvent. (B) Radioactivity was measured with a Molecular Dynamics Typhoon phosphoimager and plotted.

3.3 Rv1747 is a substrate for in vitro phosphorylation by PknF

To assess the ability of PknF_1–292 to phosphorylate NBD_1–559, an in vitro phosphorylation assay was performed by incubating purified PknF_1–292 in the presence of [γ-³²P] ATP and purified NBD_1–559. The reaction products were then separated by SDS–PAGE and the labeled proteins were identified by autoradiography. As shown in Fig. 3B (lane 2), PknF could phosphorylate NBD_1–559, whereas NBD_1–559 alone was unable to incorporate [γ-³²P] (Fig. 3B, lane 3), thus confirming that NBD_1–559 was a substrate for PknF. As previously observed for the myelin basic protein [14], the autophosphorylation of PknF was markedly reduced in the presence of NBD_1–559.

3

Phosphorylation of Rv1747 by PknF. (A) Genetic organization of the 7-kb region containing genes pknF and Rv1747. Black arrows represent pknF and Rv1747, and grey arrows represent hypothetical orf genes. (B) In vitro phosphorylation of Rv1747 by PknF_1–292. The following protein fragments were incubated with [γ-³²P]ATP: PknF_1–292 alone (lane 1), NBD_1–559 with PknF_1–292 (lane 2), and NBD_1–559 alone (lane 3). Proteins were separated by SDS—PAGE, and radioactive bands were revealed by autoradiography.

3.4 Both FHA domains are required for NBD_1–559 phosphorylation

The predicted presence of two FHA domains in the NBD of Rv1747 strongly suggested that Rv1747 could interact with a phosphoprotein, namely carrying phosphothreonine residues [31]. Since PknF is known to autophosphorylate on threonine residue(s) one the one hand, and is able to phosphorylate NBD_1–559 on the other hand, one could suspect that the FHA domains (both or either one of them) of Rv1747 would mediate a specific interaction required for its phosphorylation by the kinase. To check whether these phosphopeptide recognition motifs were involved in this phophorylation process, each FHA domain was mutagenized at a crucial position. The concerned punctual mutations consisted of Ala substitutions of the conserved Arg33, Ser47, and Asn69 residues of FHA1 and the Arg234, Ser248, and Asn270 residues of FHA2 corresponding, respectively, to Arg70, Ser85, and Asn107 in Rad53 (Fig. 1A). In the Rad53 FHA1 domain [18, 31], these three highly conserved residues have been shown to play a key role in the direct binding of the phosphopeptide, either by contacting the phosphopeptide backbone or via a phosphothreonine residue through a network of hydrogen bonding. Moreover, punctual mutations of these three conserved residues in EmbR (respectively, Arg312, Ser326, and Asn348) have been demonstrated to totally abolish phosphorylation of EmbR by PknH, as well as to prevent a specific interaction between these two proteins [15]. To test whether the FHA1 and/or the FHA2 domain of Rv1747 could mediate the interaction with PknF, in vitro phosphorylation assays were performed with either wild-type or mutated NBD_1–559. In each mutant of either FHA domain, the level of phosphorylation was strongly reduced as compared to the wild-type NBD_1–559 (decreased to 2–17% of the value measured for wild-type NBD_1–559; Fig. 4A, lanes 1–7; Fig. 4B, columns 1–7). Therefore, these data suggested that NBD_1–559 phosphorylation by PknF involved both FHA domains of Rv1747. To confirm this finding, a double mutant was prepared in which both Arg33 and Arg234 were mutated to Ala, in FHA1 and FHA2, respectively. As expected, such double mutation abolished almost completely the incorporation of radioactivity in NBD (2% of the wild-type value: Fig. 4A, lane 8; Fig. 4B, column 8).

4

Phosphorylation of NBD_1—559 mutants by PknF^1–292 (A) In vitro phosphorylation of NBD_1–559 mutants by PknF^1–292. Incubation of PknF^1–292 was performed in the presence of [γ-³²P]ATP for 15 min with either NBD_1–559 (lane 1), NBD_R33A (lane 2), NBD_S47A (lane 3), NBD_N69A (lane 4), NBD_R234A (lane 5), NBD_S248A (lane 6), NBD_N270A (lane 7), and NBD_R33A/R234A(lane 8). Proteins were analyzed by SDS–PAGE, and radioactive bands were revealed by autoradiography. (B) Radioactivity was measured with a Molecular Dynamics Typhoon phosphoimager. Phosphorylation of NBD_1–559 and its related mutants by PknF was plotted. Error bars indicate SDs.

4 Discussion

Although the regulatory networks involving STPKs and phosphatases represent an emerging theme in M. tuberculosis signaling cascades [10,13,15,32], little is known about the different protein partners, kinases and substrates, involved in this process. The main result of this study is the demonstration of a relationship between the PknF autokinase activity and its phosphorylating activity towards protein Rv1747. The latter reaction involves two FHA domains of Rv1747, one previously detected by in silico analysis [20] and another newly identified in this report.

On the basis of amino acid sequence analysis, it was predicted that Rv1747 would belong to the ABC transporter family of M. tuberculosis[21]. Protein Rv1747 contains not only the classical signature of an ABC transporter consisting of an NBD followed by a transmembrane domain, but also two N-terminal FHA domains. Given that (i) an ABC transporter requires the presence of at least two transmembrane domains and two NBDs to function and (ii) no additional putative half ABC transporter close to the rv1747 gene could be detected in the genome of M. tuberculosis, this suggests that Rv1747 would function in vivo as a homodimer rather than a heterodimer [21]. Based on sequence analysis and on the fact that no putative substrate-binding protein can be predicted to be associated with Rv1747, this transporter can be considered as a putative exporter even though its substrate specificity remains to be determined [21]. Interestingly, Rv1747 displays an unusual topology with its NBD domain preceding its transmembrane domain in the primary structure, whereas a reverse topology is found in the vast majority of ABC transporters. This is the only example of an ABC transporter with such topology in M. tuberculosis[21]. In comparison, no transporter with a similar organization has been found in Bacillus subtilis[33]. In human, a transporter with a similar topology, BCRP (also called MXR or ABCG2), has been found to be involved in multidrug resistance in certain cancer cell lines [34]. Further work will be needed to determine the functional role of rv1747 in M. tuberculosis. The phosphorylation of the NBD of Rv1747 by PknF points to a possible mechanism of regulation of the transport mediated by Rv1747. Indeed, protein phosphorylation has been shown to regulate the activity of several ABC transporters including CFTR [35,36], multidrug resistance 1 (MDR1) [37], and ABCA1, a cholesterol and phospholipid transporter [38]. Attempts to analyze the effect of phosphorylation on the ATPase activity of the NBD of Rv1747 have been so far unsuccessful, possibly because of the very low ATPase activity of the isolated NBD. In fact, phosphorylation can regulate the functioning of ABC transporters in a very subtle way. For instance, phosphorylation of MDR1 seems to modulate the drug transport by abolishing negative cooperativity between the drug-binding sites, while it does not affect the rate of ATPase activity [39]. Further investigation of the role of phosphorylation on the ATPase activity of the whole transporter Rv1747 might help to clarify this point.

Referring to eukaryotic systems, a large number of STPKs are known to catalyze autophosphorylation in an intramolecular process generally modulated by regulatory ligands, which allow rapid switching of numerous cellular functions. A similar situation can be envisaged for prokaryotes, here for PknF, which would behave like an “eukaryotic-like” STPK receptor, capable of modulating the activity of Rv1747 through phosphorylation. In eukaryotic organisms, the formation of such complex networks of interacting proteins involves conserved protein modules or domains that regulate signal transduction by mediating protein–protein interaction. Among these, the FHA domains mediate protein–protein interaction via a modular phosphopeptide recognition domain with a striking specificity for a phosphothreonine-containing epitope [18,31]. Since a similar situation has been demonstrated to occur in M. tuberculosis through the FHA-dependent phosphorylation of EmbR by PknH [15], the presence of two FHA domains in Rv1747, as well as the adjacent chromosomal position of the pknF and Rv1747 genes, suggests that these modules mediate similar cross-regulation between these two proteins.

The observation that phosphorylation of NBD_1–559 by PknF is abolished by several individual mutations present in both modules, shows that the two FHA domains of Rv1747 are required for the PknF/Rv1747 interaction. On this basis, it can be proposed that recruitment and phosphorylation of Rv1747 depend on interaction between its two non-redundant FHA domains and the autophosphorylated form of PknF. This suggests that PknF is autophosphorylated on at least two different threonine residues in order to interact simultaneously with the two FHA domains of Rv1747. Such mode of phosphorylation between an STPK and an ABC transporter provides a new framework for future investigation of possible phosphorylation of other ABC transporters in bacteria, such as Streptomyces coelicolor. Indeed, in this bacterial species, the sco1806 gene encodes a putative ABC transporter which seems to possess, as well as Rv1747, two FHA domains. Moreover, since some ABC transporters are able to export a wide diversity of substrates across biological membranes including sugars and drugs, the results obtained with M. tuberculosis open new perspectives to decipher the molecular mechanisms involved in mycobacterial physiology and pathogenicity. Currently, experiments are in progress to overproduce the entire Rv1747 protein and to identify the precise role of phosphorylation in transport activity.

Acknowledgements

This work was supported by grants from the Fondation pour la Recherche Médicale, the Société Ezus-Lyon 1 (contract 482.022), and the Institut Universitaire de France.