The effect of HAMP domains on class IIIb adenylyl cyclases from Mycobacterium tuberculosis Ju¨ rgen U. Linder, Arne Hammer and Joachim E. Schultz Abteilung Pharmazeutische, Biochemie Fakulta ¨ tfu ¨ r Chemie und Pharmazie, Universita ¨ tTu ¨ bingen Morgenstelle, Tu ¨ bingen, Germany The genes Rv1318c, Rv1319c, Rv1320c and Rv3645 1 of Mycobacterium tuberculosis are predicted to code for four out of 15 adenylyl cyclases in this pathogen. The proteins consist of a membrane anchor, a HAMP region and a class IIIb adenylyl cyclase catalytic domain. Expression and purification of the isolated catalytic domains yielded aden- ylyl cyclase activity for all four recombinant proteins. Expression of the HAMP region fused to the catalytic domain increased activity in Rv3645 21-fold and slightly reduced activity in Rv1319c by 70%, demonstrating iso- form-specific effects of the HAMP domains. Point muta- tions were generated to remove predicted hydrophobic protein surfaces in the HAMP domains. The mutations further stimulated activity in Rv3645 eight-fold, whereas the effect on Rv1319c was marginal. Thus HAMP domains can act directly as modulators of adenylyl cyclase activity. The modulatory properties of the HAMP domains were con- firmed by swapping them between Rv1319c and Rv3645. The data indicate that in the mycobacterial adenylyl cyclases the HAMP domains do not display a uniform regulatory input but instead each form a distinct signaling unit with its adjoining catalytic domain. Keywords: adenylyl cyclase; HAMP-domain; Mycobac- terium tuberculosis. Synthesis of the universal second messenger cAMP is accomplished by a plethora of adenylyl cyclases (ACs) which are currently arranged in five classes of unrelated primary structure [1–3]. The vast majority of ACs fall into class III which in turn has been subdivided recently into four subclasses (IIIa–d, [4]). The catalytic domain of these ACs, also designated as the cyclase homology domain (CHD), is often linked with further protein domains which in general appear to be regulators of cAMP production. For example, GAF, BLUF, histidine kinase, receiver, RAS-associating and cation channel domains have been found in conjunction with the class III AC catalytic domain [4]. A prominent illustration of AC diversity occurs in the human pathogen Mycobacterium tuberculosis with 15 predicted class III ACs, two of class IIIa, four of class IIIb and nine of class IIIc [4,5]. Fusion partners of the mycobacterial CHDs include membrane anchors, a novel autoinhibitory domain, AAA- ATPase domains, helix-turn-helix DNA-binding domains, an a/b-hydrolase domain and HAMP-domains. So far only two mycobacterial ACs have been reported to be enzymat- ically active, Rv1625c (class IIIa) and Rv1264 (class IIIc) [6–8]. Here we report on the four class IIIb ACs from M. tuberculosis H37Rv (Rv1318c, Rv1319c, Rv1320c, Rv3645) which contain a single HAMP domain as part of an 8 kDa region which connects the CHD to a 31 kDa membrane anchor with six predicted transmembrane spans (Fig. 1A). HAMP-domains [abbreviation originating from their primary occurance in histidine kinase, adenylyl cyclase, methyl accepting chemotaxis proteins (MCPs) and phos- phatases] are amphiphilic protein regions of about 50 amino acids which are predicted to fold into two amphipathic a-helices joined by a short linker [9]. The exact physical structure of HAMP still awaits elucidation [9]. The biochemical function of the HAMP domain has been investigated exclusively in receptor histidine kinases and in MCPs [9–13]. Mutagenesis studies with the Escherichia coli Aer protein, an aerotactic MCP, suggest an interaction of a HAMP domain with a flavin adenine dinucleotide-binding PAS-domain (acronym for period clock protein, aryl hydrocarbon receptor, single-minded protein) [13]. Further, it has been speculated that HAMP domains may function as an autonomous switch between two signaling states, or that they may regulate receptor histidine kinases by formation of four-helix bundles with a downstream dimerization domain [9,10]. Yet, respective experiments with the HAMP domains of the related sensor kinases NarX and NarQ resulted in completely different phenotypes, e.g. a deletion of seven amino acids yielded a constitutively active NarX while the same mutation did not affect the regulation of NarQ [11]. Thus, it must be acknowledged that minor variations among HAMP-domain primary structures may result in rather individual structure-function relationships including crosstalk between HAMP-domains and their adjoining effector modules [11]. Correspondence to J. U. Linder, Abteilung Pharmazeutische Bioche- mie, Fakulta ¨ tfu ¨ r Chemie und Pharmazie, Universita ¨ tTu ¨ bingen, Morgenstelle 8, 72076 Tu ¨ bingen, Germany, Fax: + 49 7071 295952, Tel.: + 49 7071 2974676, E-mail: juergen.linder@uni-tuebingen.de Abbreviations: AC, adenylyl cyclase; CHD, cyclase homology domain; MCP, methyl accepting chemotaxis protein. Enzyme: adenylyl cyclase (EC 4.6.1.1). (Received 25 February 2004, revised 8 April 2004, accepted 19 April 2004) Eur. J. Biochem. 271, 2446–2451 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04172.x Here we investigated the HAMP domain function in the context of mycobacterial class IIIb ACs. The isolated CHDs of all four isoforms were purified and had AC activity in vitro. Inclusion of the N-terminal HAMP-region enhanced the activity of the CHD of Rv3645 by more than an order of magnitude. Disruption of the predicted hydro- phobic epitopes of the HAMP-domain enhanced substrate affinity about eightfold, in principle demonstrating the ability of the HAMP domain region to regulate AC activity. The effect of the HAMP domain on the CHD of Rv1319c was much less pronounced, supporting the earlier notion of a distinct individuality of the interactions of HAMP- domains within a particular protein. Experimental procedures Materials Genomic DNA from M. tuberculosis was a gift from E. C. Boettger 2 ,UniversityofZu ¨ rich Medical School. Radio- chemicals were from Hartmann Analytik (Braunschweig, Germany). All enzymes were purchased from either Roche Diagnostics or New England Biolabs. pQE30 and Ni-nitrilotriacetic acid–agarose were from Qiagen. Fine chemicals were from Merck KGaA, Roche Diagnostics, Roth and Sigma (Germany). Plasmid construction The open reading frames of genes Rv1318c, Rv1319c, Rv1320c (GenBank Accession Number BX842576) and Rv3645 (GenBank Accession Number NC_000962) were amplified by PCR using specific primers and genomic DNA as a template. HindIII sites were added at the 3¢-ends,the 5¢-ends were fitted with BglII sites (Rv1318c, Rv3645)or BamHI sites (Rv1319c, Rv1320c). The open reading frames were cloned between the BamHI and HindIII sites of pQE30, adding an N-terminal MRGSH 6 GS-tag. Partial constructs comprised the following amino acids: Rv1318c- CHD, 355–541; Rv1319c-CHD, 356–535; Rv1320c-CHD, 355–567; Rv3645-CHD, 356–549; Rv1319c-HAMP-CHD, 279–535; Rv3645-HAMP-CHD, 279–549. BamHI and HindIII sites were added at the 5¢-and3¢-ends, respectively, and the fragments were cloned into pQE30. The following point mutations were introduced simultaneously by PCR using the expression cassettes as a template, a silent BglII site engineered at nucleotide 996 of both, Rv1319c and Rv3645, and standard molecular biology techniques: V284N, V294S, L311N, F318S, V322N in Rv1319c; I280T, L284N, V294S, L311N, F318S, V322N in Rv3645. The silent BglII site was also used for construction of chimeras in which the HAMP-domains (amino acids 279– 332) were swapped between Rv1319c and Rv3645.For constructs containing only the 23 amino acids long linker between HAMP and the CHD, i.e. amino acids 333–535 of Rv1319c and 333–567 of Rv3645, the respective BglII/ HindIII fragments were cloned into pQE31, adding an N-terminal MRGSH 6 T-tag (Fig. 1B for definition of sequence segments). The correctness of all DNA inserts was checked by double-stranded DNA sequencing. Primer sequences are available on request. Expression and purification of proteins Expression plasmids were transformed into E. coli BL21(DE3)[pRep4]. Expression was induced by 60 l M isopropyl thio-b- D -galactoside for 4–6 h at 22 °C. Bacteria were washed with buffer (50 m M Tris/HCl, 1 m M EDTA, pH 8), frozen in liquid nitrogen and stored at )80 °C. For purification, cells from 200 to 600 mL culture were suspen- ded in 20 mL of lysis buffer (50 m M Tris/HCl, 2 m M 3-thioglycerol, pH 8), lysed by sonication for 30 s and Fig. 1. Sequence analysis of mycobacterial Rv1318c, Rv1319c, Rv1320c and Rv3645 genes. (A) Domain composition of class IIIb ACs from M. tuberculosis. Hatched rectangles depict predicted transmembrane helices. Note the fusion of the last transmembrane helix to the HAMP domain. (B) Sequence alignment of the HAMP regions. Residues similar in all sequences are inverted. TM 6, transmembrane helix 6. Me, a metal- cofactor binding aspartate residue is marked for orientation purposes. Chim., cross-over site for the construction of chimeras. The predicted secondary structure of two a-helices is indicated by lightly shaded boxes on top. Ó FEBS 2004 M. tuberculosis HAMP-adenylyl cyclases (Eur. J. Biochem. 271) 2447 treatedfor30minwith0.2mgÆmL )1 lysozyme on ice. Subsequently, 5 m M MgCl 2and 10 lgÆmL )1 DNaseI were added for further 30 min on ice. After centrifugation (31 000 g,30min)15m M imidazole, pH 8, and 250 m M NaCl (final concentrations) were added to the supernatant. Protein was equilibrated for a minimum of 60 min with 250 lLNi 2+ -nitrilotriacetic acid–agarose 3 on ice, then transferred to a column and successively washed with 10 mL of buffer A (lysis buffer containing, 15 m M imidaz- ole, 250 m M NaCl and 5 m M MgCl 2 ) and 5 mL of buffer B (lysis buffer containing 15 m M imidazole and 5 m M MgCl 2 ). The protein was eluted with 0.4 mL of buffer C (37.5 m M Tris/HCl,pH8,250m M imidazole, 2 m M MgCl 2, 1.5 m M 3-thioglycerol). Purified proteins were stored at )20 °Cin buffer C after addition 20% of glycerol to the eluate. For membrane preparations of cells expressing the holoenzymes, lysis was performed by a French Press. Cell debris was removed at 3000 g for 30 min and membranes were sedimented at 100 000 g for 1 h at 0 °C. Membranes were suspended in buffer (40 m M Tris/HCl, pH 8.0, 1.6 m M 3-thioglycerol, 20% glycerol) and assayed for AC activity. Cells transformed with pQE30 served as a negative control. Adenylyl cyclase assays AC activity was determined at 37 °Cfor10minin100lL [14]. Standard reactions contained 50 m M Tris/HCl, pH 8.0, 22% glycerol, 3 m M MnCl 2 ,200l M [ 32 P]ATP[aP] and 2 m M [2,8- 3 H]cAMP. For kinetic analysis, variable amounts of MnATP were used in the presence of 3 m M free Mn 2+ . At least two independent purifications were performed for recombinant protein. All data are means of 4–11 measurements ± SD. Results Sequence features of mycobacterial class IIIb ACs The predicted gene products of Rv1318c, Rv1319c, Rv1320c and Rv3645 from M. tuberculosis H37Rv would be members of the class IIIb AC family. Characteristics of class IIIb ACs include the substitution of a threonine for the canonical substrate-specifying aspartate and an arm region extended by one residue compared to class IIIa ACs (e.g. mycobac- terial Rv1625c, mammalian membrane-bound ACs [4]). The N-terminal of all four mycobacterial ACs is predicted to constitute a membrane anchor with six transmembrane helices (Fig. 1A). The last transmembrane helix is fused directly to a HAMP-domain which is part of the region connecting the membrane anchor to the catalytic domain (CHD). Secondary structure prediction suggested that the second amphipathic helix of the HAMP domain is C-terminally extended by about 20 amino acids (Fig. 1B) which implies that HAMP is embedded in a larger structural module. The CHDs of the mycobacterial class IIIb ACs are predicted to form homodimers with two catalytic centers at the dimer interface, as has been demonstrated for several other bacterial ACs [4]. The overall amino acid identities within the cluster of Rv1318c–Rv1320c are 67–77%, the identities of these three cyclases with Rv3645 are 37–38%. Higher identities are observed among the CHDs (76–79% within the cluster, 53–57% to Rv3645) and the HAMP regions (87–91% within the cluster, 46–47% to Rv3645). Expression of the isolated catalytic domains All four catalytic domains which start 11 amino acids upstream of the first metal-binding aspartate ([8,15]) were overexpressed in E. coli and purified essentially to homo- geneity via an N-terminal His 6 metal-affinity tag (Fig. 2, lanes 1–4). All proteins possessed significant enzymatic activity in the presence of 3 m M Mn 2+ as a cofactor, demonstrating that the four genes code for functional ACs. All enzymes were specific for ATP as a substrate. Tiny guanylyl cyclase side-activities were detectable in Rv1319c- CHD ( 0.08 nmol cGMPÆmg )1 Æmin )1 ) and Rv3645-CHD ( 0.03 nmol cGMPÆmg )1 Æmin )1 ). For a further kinetic characterization, we selected the two most active enzymes, Rv1319c-CHD and Rv3645-CHD (Table 1). The maximal velocity of both enzymes was rather similar, yet Rv1319c-CHD had  20-fold higher affinity for the substrate, ATP than Rv3645-CHD (Table 1). The Hill coefficients of 1.0 indicated that the two predicted catalytic centers did not interact cooperatively in the recombinant CHDs. The large difference in substrate affinity between the isoforms and the known differences of metal-cofactor affinity of mammalian AC isoforms [16] prompted us to investigate the metal-cofactor dependence in detail. Nota- bly, Mg 2+ was ineffective as a cofactor. At 25 m M Mg 2+ Rv1319c-CHD had 3% of the activity with 3 m M Mn 2+ , Rv3645-CHD was inactive with Mg 2+ .Mn 2+ affinities (EC 50 values) were 0.48 ± 0.02 and 3.9 ± 0.3 m M for Rv1319c-CHD and Rv3645-CHD, respectively. Thus the Fig. 2. 15% SDS/PAGE of affinity-purified AC proteins. Enzyme (1.5– 2.1 lg per lane) were stained with Coomassie blue. AC activities ± SD at 2.2–3.5 l M enzyme under standard conditions are given below each lane. Lane 1, Rv1318c-CHD; lane 2, Rv1319c-CHD; lane 3, Rv1320c- CHD; lane 4, Rv3645-CHD; lane 5, Rv1319c-HAMP-CHD; lane 6, Rv1319c-HAMP mut -CHD; lane 7, Rv3645-HAMP-CHD; lane 8, Rv3645-HAMP mut -CHD; lane 9, Rv3645HAMP-1319cCHD; lane 10, 1319cHAMP-3645CHD. Note, that the apparent molecular mass of Rv1318c-CHD, Rv1320c-CHD Rv3645c-CHD appears about 3 kDa higher than calculated. We attribute this slight deviation to unusual electrophoretic mobility as often observed. Rv1319c-CHD runs canonically. An extended translation is highly unlikely, because all constructs were completely sequenced and contain two in-frame stop codons. 2448 J. U. Linder et al. (Eur. J. Biochem. 271) Ó FEBS 2004 differences in substrate affinity parallel those for Mn 2+ - affinity. Modulation of AC activity by the HAMP-domain To examine a possible regulatory input of the HAMP domain on AC activity, we expressed constructs comprising the CHD and the HAMP-region, i.e. Rv1319c-HAMP- CHD and Rv3645-HAMP-CHD. The purified proteins displayed altered cyclase activities compared to the respect- ive catalysts alone (Fig. 2, lanes 5, 7). In Rv1319c the attached HAMP region reduced activity by 70% whereas the activity of the CHD of Rv3645 was enhanced 21-fold. As a control we expressed and purified constructs in which only the 23 amino acid linker was present N-terminally of the respective CHD. The recombinant proteins had the same activity as the CHDs alone. This suspends the possibility that the slightly diverged linkers alone influence the activities of Rv1319c-HAMP-CHD and Rv3645- HAMP-CHD (data not shown). A kinetic analysis revealed that the reduced activity in Rv1319c-HAMP-CHD was due to a simultaneous reduction of V max and the substrate-affinity (Table 1) whereas the increased activity of Rv3645-HAMP-CHD was due to a higher V max with a slightly reduced substrate affinity (Table 1). In general the presence of the HAMP region actuated a certain extent of cooperativity between the two catalytic centers (Table 1). Taken together it appears that the HAMP regions modulate the AC activity of the CHDs in a rather distinct manner. Next we examined the role of the HAMP region by targeting its amphiphilic nature. Helical wheel representa- tions of the HAMP-domains indicate hydrophobic surfaces in both helices of the two isoforms (Fig. 3). To weaken hydrophobic interactions between these surfaces, we simul- taneously replaced hydrophobic residues in both helices of Rv1319c and Rv3645 by amino acids with hydrophilic uncharged side chains (Fig. 3). The recombinant proteins, Rv1319c-HAMP mut -CHD and Rv3645-HAMP mut -CHD, were purified and assayed. In Rv1319c, the mutations increased AC activity of the HAMP-CHD ensemble to 195% of the respective unaltered construct (Fig. 2, com- pare lanes 5 and 6). In Rv3645, removal of the hydro- phobic surface of the HAMP region caused a sevenfold increase of activity (Fig. 2, lanes 7, 8), rendering Table 1. Kinetic properties of protein constructs for Rv1319c and Rv3645. Concentrations of affinity-purified, homogenous proteins were 1–5 l M to limit substrate conversion to < 10%. SC 50 , substrate concentration at half-maximal velocity. Enzyme V max (nmol cAMPÆmg )1 Æmin )1 ) SC 50 (l M ) Hill-coefficient Rv1319c-CHD 6.6 ± 0.1 57 ± 1 1.0 ± 0.04 Rv1319c-HAMP-CHD 3.6 ± 0.1 150 ± 3 1.2 ± 0.1 Rv1319c-HAMP mut -CHD 6.1 ± 0.3 110 ± 8 1.0 ± 0.1 3645HAMP-1319cCHD 8.6 ± 0.1 180 ± 5 1.3 ± 0.03 Rv3645-CHD 8.2 ± 0.6 1200 ± 100 1.0 ± 0.03 Rv3645-HAMP-CHD 590 ± 10 2700 ± 100 1.4 ± 0.01 Rv3645-HAMP mut -CHD 470 ± 30 380 ± 50 1.3 ± 0.1 1319cHAMP-3645CHD 79 ± 2 2600 ± 200 1.3 ± 0.02 Fig. 3. Helical wheel models of the HAMP- domains. Hydrophobic residues (L, I, M, V, F, A, P) are inverted, nonhydrophobic ones are lightly shaded. Letters and arrows denote the mutations introduced to eliminate hydro- phobic surfaces. Ó FEBS 2004 M. tuberculosis HAMP-adenylyl cyclases (Eur. J. Biochem. 271) 2449 Rv3645-HAMP mut -CHD 100-fold more active than the isolated CHD (Fig. 2 compare lanes 4 and 8). Thus the hydrophobic epitopes of the HAMP-domain appear to act like a throttle on AC activity, their disruption enhances the catalytic efficiency. A kinetic analysis revealed that Rv3645- HAMP mut -CHD had an eightfold higher substrate-affinity compared to Rv3645-HAMP-CHD (Table 1). Do the differences in the effects of the HAMP domains reside in particular features of the respective ensemble or are they intrinsic properties of the respective HAMP domains? To answer this question we swapped the HAMP domains and generated 3645HAMP-1319cCHD and 1319cHAMP- 3645CHD, respectively (Fig. 1B). Under standard assay conditions, the activity of Rv1319c was repressed by the Rv3645 HAMP domain by 38%, i.e. slightly less than the 70% by the wild-type module (Rv1319c-HAMP-CHD, Fig. 2, compare lanes 2, 5 and 9). The Rv1319c HAMP domain stimulated Rv3645 only threefold compared to the 21-fold stimulation seen in Rv3645-HAMP-CHD (Fig. 2, compare lanes 4, 7 and 10). The kinetic parameters confirmed that the external HAMP domain affected catalysis of Rv3645 similarly as the intrinsic HAMP domain, i.e. large increase in V max , accompanied by a slight reduction in substrate-affinity (Table 1). However, in Rv1319c the Rv3645 HAMP domain reduced substrate affinity threefold and increased V max only marginally. With regard to the above questions this means that it is predominantly the kind of interaction between catalyst and regulatory domain which determines the regulatory output of HAMP domains and not an intrinsic feature of a peculiar HAMP domain. As each hydrophilic surface of the HAMP domains carries 4–5 charged residues (Fig. 3), we tested whether the cationic amphiphilic antibiotic Gramicidin S interferes with their function. Gramicidin S (10 l M ) had no effect on the constructs of Rv1319c, but it differentially affected Rv3645 constructs. While the activity of Rv3645-CHD remained unaffected, Gramicidin inhibited Rv3645-HAMP-CHD by 43 ± 4% and Rv3645-HAMP mut -CHD by 18 ± 2%. Thus Gramicidin appeared to impair slightly the interaction of the HAMP domain of Rv3645 with its catalytic domain. As Gramicidin also inhibited the Rv3645-HAMP mut -CHD construct somewhat, it may be intimated that it partially interacted with the hydrophilic surfaces of the HAMP helices. Finally, we wished to characterize the effect of the HAMP domains in the context of the membranous holoenzymes. Therefore we attempted to express all four mycobacterial class IIIb AC isoforms. However, significant AC activity was only seen upon expression of the Rv3645 holoenzyme (1 nmol cAMPÆmg )1 Æmin )1 )inisolatedcell membranes. Solubilization and purification of the holo- enzyme was impossible due to instability of the enzyme upon detergent treatment needed for solubilization. This stymied all attempts to reliably characterize a potential modulatory effect of the HAMP domain in the holoenzyme. Discussion For the first time we show here that the catalytic domains of all four class IIIb ACs of M. tuberculosis H37Rv (Rv1319c- 1320c, Rv3645) are actually enzymatically active. Although the multiplicity of these highly similar gene products may suggest a redundancy of functional class IIIb ACs in M. tuberculosis, the data show that each cyclase displays rather individual properties and thus may serve distinctive cellular demands. Striking differences are exemplified by a closer analysis of Rv1319c and Rv3645. Rv1319c has a 20-fold higher substrate affinity than Rv3645 while the effect of the HAMP region is much more dramatic in Rv3645. Even within the cluster of Rv1318c–Rv1320c AC activities vary by one order of magnitude, supporting the suggestion that their physiological role may be tailored to particular cellular states and needs during the survival of the pathogen under changing environmental conditions in the host. The AC activities of these class IIIb CHDs are two orders of magnitude lower than those published for the CHDs of the mycobacterial ACs Rv1625c (class IIIa) and Rv1264 (class IIIc) [7,8]. In our view, this does not imply that the class IIIb ACs are of minor importance or physiologically irrelevant, because catalysis of Rv3645 can be potentiated by the HAMP region, yielding V max values of comparable magnitude as Rv1625c and Rv1264 [7,8]. In all likelihood, the Rv1318c-1320c isoforms will also be regulated in such an individual manner although the mechanisms remain enigmatic. The four ACs require Mn 2+ for efficient catalysis. This has also been reported for the mycobacterial ACs Rv1625c [6,7] and Rv1264 [8] and we observe the same with four further mycobacterial ACs currently under investigation (Rv0386, Rv1647, lipJ, Rv2212). The Mn 2+ -dependence of the ACs suggests that millimolar concentrations of Mn 2+ are physiological in M. tuberculosis as it has been discussed previously [6]. The cytosolic concentration of Mn 2+ in Mycobacterium is unknown. Yet, the presence of a high- affinity Mn 2+ -transporter in M. tuberculosis supports this suggestion [17]. Furthermore Lactobacillus plantarum is known to contain 16–25 m M Mn 2+ in the cytosol [18] demonstrating that some bacteria indeed accumulate Mn 2+ to high concentrations. An interesting question concerns the role of the HAMP- region and the membrane anchors in AC regulation. So far, HAMP domains coupled in a similar manner with AC catalysts have been detected in the genomes of a variety of bacteria, such as Corynebacteria, Legionella, Leptospira and Treponema. However, a biochemical study has not been carried out with any of those potential gene products. In Rv1319c the HAMP region exerts a slightly inhibitory effect on the CHD and in Rv3645 it has a large stimulatory effect indicating that obviously no general rule can be deduced from our data for the effect of HAMP domains on AC catalysts. Even more striking, the disruption of the predicted hydrophobic surfaces in the Rv3645 HAMP domain itself caused a large increase in catalytic efficiency. Thus, HAMP domains can act as a regulatory module on the CHD independently of the presence of the N-terminal membrane anchor. A crucial involvement of the hydrophobic epitopes in signaling through HAMP has been predicted earlier [10]. Our experiments indicate that hydrophobic interactions in or with the HAMP domains affect the activity status. As for the effect of Gramicidin S on the HAMP domain function of Rv3645 we speculate that the AC is regulated in vivo by an as yet unknown factor that may directly and reversibly 2450 J. U. Linder et al. (Eur. J. Biochem. 271) Ó FEBS 2004 interact with HAMP domains. Such an interaction of a HAMP domain with another module has been demonstra- ted previously for the oxygen sensor protein Aer from E. coli, where an intramolecular interaction between the HAMP domain and the PAS domain is observed [13]. The role of the membrane anchors in signaling through HAMP domains remains enigmatic. The close attachment of the HAMP domains to the last transmembrane span is a structural feature shared by MCPs and receptor histidine kinases [19]. Therefore the membrane anchors may serve as external sensors of physical or chemical stimuli which are then transmitted via a sophisticated system for cAMP generation, possibly spatially and temporally fine-tuned by the multiplicity of AC isoforms. In fact, the membrane anchors are the most divergent parts of the four mycobac- terial HAMP-ACs studied here and in different M. tuber- culosis strains the number of HAMP-ACs varies. The cluster of three ACs (Rv1318c)1320c) in strain H37Rv corresponds to a cluster of four genes (MT1359–1362)in strain CDC1551 [20]. The MT1361 protein corresponds to Rv1319c (99% identity). MT1360 has a membrane anchor which is diverged by 43% compared to Rv1319c whereas the HAMP-CHD ensemble is identical to that of Rv1319c. This again suggests that the membrane anchor of each isoform has evolved towards a specific, yet elusive function. The physiological role of the four mycobacterial class IIIb ACs is as yet unknown. Our biochemical studies, which may be termed genomics-based biochemistry, reveal new and valuable aspects of the function of HAMP-domains in general. HAMP domains appear to have regulatory func- tions of their own in addition to a role as signal transducers. The HAMP-CHD ensembles of the mycobacterial class IIIb ACs reported here may be the first step toward an establishment of the mechanistic relationship between the HAMP domain and its attached AC catalysts. References 1. Barzu, O. & Danchin, A. (1994) Adenylyl cyclases: a hetero- geneous class of ATP-utilizing enzymes. Prog. Nucleic Acid. Res. Mol. Biol. 49, 241–283. 2. Cotta, M.A., Whitehead, T.R. & Wheeler, M.B. (1998) Identifi- cation of a novel adenylate cyclase in the ruminal anaerobe, Pre- votella ruminicola D31d. FEMS Microbiol. Lett. 164, 257–260. 3. Sismeiro, O., Trotot, P., Biville, F., Vivares, C. & Danchin, A. (1998) Aeromonas hydrophila adenylyl cyclase 2: a new class of adenylyl cyclases with thermophilic properties and sequence similarities to proteins from hyperthermophilic archaebacteria. J. Bacteriol. 180, 3339–3344. 4. Linder, J.U. & Schultz, J.E. (2003) The class III adenylyl cyclases: multi-purpose signalling modules. Cell. Signal. 15, 1081–1089. 5. McCue, L.A., McDonough, K.A. & Lawrence, C.E. (2000) Functional classification of cNMP-binding proteins and nucleo- tide cyclases with implications for novel regulatory pathways in Mycobacterium tuberculosis. Genome Res. 10, 204–219. 6. Reddy,S.K.,Kamireddi,M.,Dhanireddy,K.,Young,L.,Davis, A. & Reddy, P.T. (2001) Eukaryotic-like adenylyl cyclases in Mycobacterium tuberculosis H37Rv: cloning and characterization. J. Biol. Chem. 276, 35141–35149. 7. Guo, Y.L., Seebacher, T., Kurz, U., Linder, J.U. & Schultz, J.E. (2001) Adenylyl cyclase Rv1625c of Mycobacterium tuberculosis: a progenitor of mammalian adenylyl cyclases. EMBO J. 20, 3667–3675. 8. Linder, J.U., Schultz, A. & Schultz, J.E. (2002) Adenylyl cyclase Rv1264 from Mycobacterium tuberculosis has an autoinhibitory N-terminal domain. J. Biol. Chem. 277, 15271–15276. 9. Aravind, L. & Ponting, C.P. (1999) The cytoplasmic helical linker domain of receptor histidine kinase and methyl-accepting proteins is common to many prokaryotic signalling proteins. FEMS Microbiol. Lett. 176, 111–116. 10. Williams, S.B. & Stewart, V. (1999) Functional similarities among two-component sensors and methyl-accepting chemotaxis proteins suggest a role for linker region amphipathic helices in transmembrane signal transduction. Mol. Microbiol. 33, 1093– 1102. 11. Appleman, J.A., Chen, L.L. & Stewart, V. (2003) Probing con- servation of HAMP linker structure and signal transduction mechanism through analysis of hybrid sensor kinases. J. Bacteriol. 185, 4872–4882. 12. Appleman, J.A. & Stewart, V. (2003) Mutational analysis of a conserved signal-transducing element: the HAMP linker of the Escherichia coli nitrate sensor NarX. J. Bacteriol. 185, 89–97. 13. Bibikov, S.I., Barnes, L.A., Gitin, Y. & Parkinson, J.S. (2000) Domain organization and flavin adenine dinucleotide-binding determinants in the aerotaxis signal transducer Aer of Escherichia coli. Proc. Natl Acad. Sci. USA. 97, 5830–5835. 14. Salomon, Y., Londos, C. & Rodbell, M. (1974) A highly sensitive adenylate cyclase assay. Anal. Biochem. 58, 541–548. 15. Kanacher, T., Schultz, A., Linder, J.U. & Schultz, J.E. (2002) A GAF-domain-regulated adenylyl cyclase from Anabaena is a self-activating cAMP switch. EMBO J. 21, 3672–3680. 16. Pieroni, J.P., Harry, A., Chen, J., Jacobowitz, O., Magnusson, R.P. & Iyengar, R. (1995) Distinct characteristics of the basal activities of adenylyl cyclases 2 and 6. J. Biol. Chem. 270, 21368– 21373. 17. Vidal, S.M., Malo, D., Vogan, K., Skamene, E. & Gros, P. (1993) Natural resistance to infection with intracellular parasites: iso- lation of a candidate for Bcg. Cell 73, 469–485. 18. Archibald, F.S. & Fridovich, I. (1981) Manganese and defenses against oxygen toxicity in Lactobacillus plantarum. J. Bacteriol. 145, 442–451. 19. Galperin, M.Y., Nikolskaya, A.N. & Koonin, E.V. (2001) Novel domains of the prokaryotic two-component signal transduction systems. FEMS Microbiol. Lett. 203, 11–21. 20. Fleischmann, R.D., Alland, D., Eisen, J.A., Carpenter, L., White, O., Peterson, J., DeBoy, R., Dodson, R., Gwinn, M., Haft, D., et al. (2002) Whole-genome comparison of Mycobacterium tuberculosis clinical and laboratory strains. J. Bacteriol. 184, 5479–5490. Ó FEBS 2004 M. tuberculosis HAMP-adenylyl cyclases (Eur. J. Biochem. 271) 2451 . activity. The effect of the HAMP domain on the CHD of Rv1319c was much less pronounced, supporting the earlier notion of a distinct individuality of the interactions. partially interacted with the hydrophilic surfaces of the HAMP helices. Finally, we wished to characterize the effect of the HAMP domains in the context of the membranous holoenzymes.