Adenylyl cyclase Rv0386 from Mycobacterium tuberculosis H37Rv uses a novel mode for substrate selection Lucila I. Castro, Corinna Hermsen, Joachim E. Schultz and Ju ¨ rgen U. Linder Abteilung Pharmazeutische Biochemie, Fakulta ¨ tfu ¨ r Chemie und Pharmazie, Universita ¨ tTu ¨ bingen, Germany The second messenger cAMP is synthesized by a large variety of adenylyl cyclases (ACs) which are separated into five classes that are not related in their protein sequences [1–3]. The vast majority of ACs belongs to class III which recently has been subdivided into clas- ses IIIa to IIId [4]. The catalytic domain of class III ACs has been termed cyclase homology domain (CHD) and appears to be linked with different protein domains which in several instances have been shown to impart peculiar regulatory features [5] (reviewed in [4]). So far, all CHDs operate as dimers with the cata- lytic centre positioned at the dimer interface [6–8]. Based on mutational and structural data catalysis is thought to rest on six highly conserved residues which are spaced in register in the CHDs. Two aspartate resi- dues coordinate two metal ions (Mg 2+ or Mn 2+ ), an asparagine and an arginine stabilize the transition-state and a lysine-aspartate couple specifies ATP as a sub- strate [9–11]. A common variant to this canon is the exchange of the usual substrate specifying aspartate for a threonine or serine in class IIIb ACs [4]. The hydroxyl group specifically serves as a hydrogen-bond acceptor and in this respect has the same function as the canonical aspartate [12,13]. However variations in all six canonical catalytic residues do occur as evident from whole genome sequencing projects [4] and the functional consequences of such changes are just beginning to be understood [14]. The genome of the human pathogen Mycobacterium tuberculosis H37Rv contains 15 ORFs that code for CHDs [15]. Two belong to class IIIa, four to class IIIb and nine to class IIIc [4]. One predicted class IIIa and six predicted class IIIc mycobacterial cyclase genes contain variations at canonical positions of the Keywords Adenylyl cyclase; cyclic nucleotide; guanylyl cyclase; Mycobacterium tuberculosis; substrate specificity Correspondence J. U. Linder, Abteilung Pharmazeutische Biochemie, 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 (Received 27 March 2005, revised 13 April 2005, accepted 18 April 2005) doi:10.1111/j.1742-4658.2005.04722.x Class III adenylyl cyclases usually possess six highly conserved catalytic residues. Deviations in these canonical amino acids are observed in several putative adenylyl cyclase genes as apparent in several bacterial genomes. This suggests that a variety of catalytic mechanisms may actually exist. The gene Rv0386 from Mycobacterium tuberculosis codes for an adenylyl cyclase catalytic domain fused to an AAA-ATPase and a helix-turn-helix DNA-binding domain. In Rv0386, the standard substrate, adenine-defining lysine-aspartate couple is replaced by glutamine-asparagine. The recombin- ant adenylyl cyclase domain was active with a V max of 8 nmol cAMPÆ mg )1 Æmin )1 . Unusual for adenylyl cyclases, Rv0386 displayed 20% guanylyl cyclase side-activity with GTP as a substrate. Mutation of the glutamine- asparagine pair either to alanine residues or to the canonical lysine-aspar- tate consensus abolished activity. This argues for a novel mechanism of substrate selection which depends on two noncanonical residues. Data from individual and coordinated point mutations suggest a model for purine definition based on an amide switch related to that previously identified in cyclic nucleotide phosphodiesterases. Abbreviations AC, adenylyl cyclase; CHD, cyclase homology domain; GC, guanylyl cylase; HTH, helix-turn-helix; PDE, cyclic nucleotide phosphodiesterase. FEBS Journal 272 (2005) 3085–3092 ª 2005 FEBS 3085 catalytic centre [4]. The four class IIIb cyclases contain the threonine variant mentioned above. To date almost all canonical and all class IIIb mycobacterial ACs have been investigated (class IIIa: AC Rv1625c; class IIIb: Rv1318c, Rv1319c, Rv1320c, Rv3645; class IIIc: Rv1264, Rv1647) [8,16–19]. Of the noncanonical CHDs only Rv1900c (class IIIc) has been examined in detail [14]. Here, the substrate-specifying lysine-aspar- tate pair is replaced by asparagine-aspartate and the catalytic asparagine is altered to histidine. Structure determination and mutagenesis experiments demon- strated that Rv1900c adopts a mode of catalysis in which these three otherwise canonical residues are dis- pensable. We investigated the mycobacterial Rv0386 gene product as a class IIIc AC which has a glutamine- asparagine pair at the positions defining ATP as a sub- strate instead of the lysine-aspartate consensus. We show that the purified catalytic domain of Rv0386 is active as an AC which has an unusually high GC side- activity. Mutational analysis of Rv0386 demonstrated that the catalytic activity specifically depends on the noncanonical glutamine-asparagine couple. This strongly indicates that an alternative substrate-binding mechanism evolved in Rv0386, distinct from that in canonical ACs. A model of purine-binding in Rv0386 is proposed. Results Sequence analysis The M. tuberculosis gene Rv0386 codes for a multido- main protein of 1085 amino acids (117 kDa, Fig. 1A). An AC catalytic domain is located at the N-terminus (amino acids 1–167), which is characterized as a class IIIc CHD because of a shortened ‘arm’ region [4]. Strikingly the canonical substrate-defining residues, lysine-aspartate, correspond to Gln57 and Asn106 in Rv0386, respectively (Fig. 1B, [4]). Therefore it was not at all a forgone conclusion whether the CHD of Rv0386 would in fact display AC activity. Further, sequence analysis by SMART and INTER- PRO-scan showed that the CHD is linked via 12 amino-acid residues to an AAA-ATPase domain (NB-ARC type [20], amino acids 180–477), a tetratrico- peptide repeat (TPR)-like region (amino acids 646–968) and a C-terminal helix-turn-helix (HTH) DNA-binding domain (amino acids 1024–1081, luxR family [21]). An identical domain composition, i.e. a CHD linked in this order to these three domains is present in the putative AC genes Rv1358 and Rv2488c from M. tuberculosis. Moreover, the AAA-ATPase ⁄ NB-ARC domain is sim- ilar to the respective domains of several bacterial tran- scriptional regulators (e.g. 40% similarity to afsR of Streptomyces coelicolor [22]). Therefore the presence of the HTH DNA-binding domain strongly suggests that in Rv0386 an AC may be functionally linked with a transcriptional regulator. AC activity of the Rv0386 CHD We expressed the CHD of Rv0386 (amino acids 1–175) in Escehrichia coli as a soluble protein and purified it to homogeneity (Fig. 2). At 4.9 lm recom- binant Rv0386 (1)175) displayed an AC activity of 5.0 nmol cAMPÆmg )1 Æmin )1 with Mn 2+ as a metal cofactor (Table 1). Activity with Mg 2+ was below the DHC /esaPTA-AAA C RA-BN 774081 5 761 6830vR )sdicaonima 5801 ( ekil-RPT 646 869 18014201 H T H )Rxul( A B Fig. 1. Sequence analysis. (A) Domain com- position of Mycobacterium tuberculosis Rv0386. (B) Local alignment of Rv0386 with the noncanonical class IIIc AC Rv1900c, the canonical class IIIc AC Rv1264 and the canonical class IIIa AC Rv1625c from M.tuberculosis. The six residues implicated in catalysis by canonical ACs are boxed. a, adenine-specifying; m, metal-coordinating; c, catalytic transition-state stabilizing. Solid arrowheads mark the deviations from the consensus in Rv0386. M. tuberculosis adenylyl cyclase Rv0386 L. I. Castro et al. 3086 FEBS Journal 272 (2005) 3085–3092 ª 2005 FEBS detection limit (Table 1). Rv0386 (1)175) had a substan- tial GC activity of 1.0 nmol cGMPÆmg )1 Æmin )1 , i.e. 20% of the AC activity. This is unusual because all canonical class III ACs investigated to date possess a stringent ATP specificity [23]. On the other hand it is reminiscent of the noncanonical AC Rv1900c which also possesses significant GC side-activity [14]. The temperature optimum of Rv0386 (1)175) was 30 °C, the activation energy 76 ± 3 kJÆmol )1 (SEM, n ¼ 2) and the pH optimum was at pH 7.5–8.0. V max was 7.5 ± 0.8 nmol cAMPÆmg )1 Æmin )1 (SEM, n ¼ 4) and the apparent K m for ATP was 0.6 ± 0.2 mm . A Hill coefficient of 1.0 ± 0.1 indicated no cooperativity for ATP with respect to the predicted two catalytic cen- tres. The V max was at the lower end of bacterial class III ACs which may reflect an unstimulated state of the isolated CHD (Discussion). For GC activity V max was 2.2 ± 0.1 nmol cGMPÆmg )1 Æmin )1 (n ¼ 3) and the apparent K m for GTP was 0.5 ± 0.03 mm with a Hill coefficient of 1.0 ± 0.1. Thus Rv0386 (1)175) had a lower turnover and a slightly higher substrate affinity to GTP compared to ATP. Mutational analysis of Rv0386 (1)175) What, if any, are the functions of those two putative substrate-binding amino acids, glutamine and aspara- gine which take the position of the canonical lysine-as- partate pair? First we removed the amide side-chains creating Rv0386 (1)175) Q57A and Rv0386 (1)175) N106A to determine whether the two residues actually are necessary for catalysis. Both mutants were expressed as soluble proteins and purified to homogeneity (Fig. 2). They were essentially inactive. This strongly implicated that Gln57 and Asn106 interact with the substrate. Next we asked whether the canonical lysine-aspartate pair would operate in Rv0386. The mutants Rv0386 (1)175) Q57K, Rv0386 (1)175) N106D and the double mutant Rv0386 (1)175) Q57K ⁄ N106D were generated, expressed and purified. Rv0386 (1)175) N106D was inactive. Rv0386 (1)175) Q57K had an AC activity of less than 5% of wild-type activity (Table 1). GC activity was below the detection limit (Table 1). Similarly, the purified double mutant protein Rv0386 (1)175) Q57K ⁄ N106D retained some AC activity (Table 1) while the GC side-activity was undetectable. Thus implementation of the canonical lysine-aspartate ensemble actually was incompatible with catalytic activity. This highlighted the importance of the gluta- mine-asparagine pair for substrate binding. The low activity of the mutants precluded a meaningful kinetic analysis. At this point, one may ask, whether Gln57 and Asn106 are indeed participating in substrate- binding or whether they are crucial for maintaining the fold of the protein. In the latter case the mutants would have been inactive due to misfolding. However, we regard this as unlikely, because all mutants were soluble, purified and stable in the absence of protease inhibitors. Actually, none of the previously reported mutants of the canonical lysine-aspartate couple in mammalian and mycobacterial ACs appeared to be misfolded [8,17,24]. The inactivity of Rv0386 (1)175) N106D was partic- ularly remarkable, because asparagine can act as both, a hydrogen-bond donor via its amide group and as a hydrogen-bond acceptor via its carbonyl oxygen atom. In contrast aspartate can only serve as a hydrogen- bond acceptor. Therefore, we reasoned that Rv0386 uses a novel substrate-defining and -binding mechan- ism which requires a precisely positioned hydrogen- bond donor at the position of the canonical aspartate. To test this hypothesis we replaced Asn106 by a serine R v 0 3 8 6 ( 1 - 1 7 5 ) W T Q 5 7 K Q 5 7 A N 1 0 6 D N 1 0 6 A Q 5 7 K / N 1 0 6 D N 1 0 6 S a D k 5 4 5 2 53 8 1 41 Q 5 7 E Fig. 2. Purification of recombinant proteins SDS ⁄ PAGE analysis of purified wild-type and mutant Rv0386 (1)175) ,1–3lg per lane, visual- ized by Coomassie stain. Some point mutants display a slightly altered electrophoretic mobility. Table 1. Activities of Rv0386 (1)175) and mutants. Assays were con- ducted with 850 l M substrate and 5 mM MnCl 2 at pH 7.5 and 30 °C. Standard errors of the mean are included (number of experi- ments in brackets). AC and GC activities of the mutants Rv0386 (1)175) Q57A, Rv0386 (1)175) N106A and Rv0386 (1)175) N106D were below the detection limits of 0.1 and 0.2 nmolÆmg )1 Æmin )1 , respectively. ND, not detectable. Enzyme Adenylyl cyclase (nmol cAMPÆ mg )1 Æmin )1 ) Guanylyl cyclase (nmol cGMPÆ mg )1 Æmin )1 ) Rv0386 (1)175) 5.0 ± 0.6 (10) 1.0 ± 0.2 (10) Rv0386 (1)175) Q57K 0.2 ± 0.05 (4) ND Rv0386 (1)175) N106S 1.8 ± 0.1 (4) ND Rv0386 (1)175) Q57K ⁄ N106D 0.2 ± 0.06 (4) ND Rv0386 (1)175) Q57E 0.1 ± 0.05 (4) ND L. I. Castro et al. M. tuberculosis adenylyl cyclase Rv0386 FEBS Journal 272 (2005) 3085–3092 ª 2005 FEBS 3087 as a potential hydrogen-bond donor and constructed Rv0386 (1)175) N106S (Fig. 2). The AC activity of purified Rv0386 (1)175) N106S was 1.8 nmol cAMPÆ mg )1 Æmin )1 , i.e. 36% of wild-type AC activity, whereas GC-activity was below the detection limit (Table 1). Thus serine was not only compatible with AC cata- lysis, but actually shifted substrate discrimination in favour of ATP. The kinetic analysis yielded a V max of 2.4 ± 0.4 nmol cAMPÆmg )1 Æmin )1 (SEM, n ¼ 4) and a K m of 0.4 ± 0.05 mm ATP with a Hill coefficient of 1.1 ± 0.1. Obviously, with the change from asparagine to serine ATP-binding affinity was retained while cata- lytic efficiency was attenuated. In analogy to the Rv0386 (1)175) N106D mutant we also generated Rv0386 (1)175) Q57E. On the one hand the mutation eliminated the hydrogen-bond donor property of the resident Q57; on the other hand a glu- tamate at this position is highly conserved in GCs where it may hydrogen-bond to the N1 amide and 2- amino groups of the guanine moiety [11,24]. Purified Rv0386 (1)175) Q57E displayed less than 5% of the AC activity of wild-type (Table 1) and no detectable GC activity. This confirmed that cyclase activity of Rv0386 relies specifically on the Gln57 ⁄ Asn106 couple. These unexpected findings cannot possibly be recon- ciled with and discussed on the basis of the available structural data of canonical mammalian class IIIa and mycobacterial class IIIc catalytic domains [5,7,9,25] nor do they parallel the findings on the noncanonical class IIIc AC Rv1900c [14]. Another novel substrate- specifying mechanism must exist in the CHD of Rv0386 probably brought about by peculiar structural elements yet to be recognized. Enzymatic activity of the Rv0386 holoenzyme To reveal a possible regulatory role of the C-terminal putative transcription factor domain we expressed the Rv0386 holoenzyme in E. coli. The majority of the expression product ended up in inclusion bodies. Yet it was possible to solubilize a few micrograms of enzyme with 2% CHAPS as a detergent (Fig. 3). Purification of the holoenzyme was impossible, because it was rapidly degraded upon incubation with the metal-affinity resin, a process which we were unable to stop in spite of the addition of an assort- ment of protease inhibitors (data not shown). The specific activity of the holoenzyme was estimated to be 3 nmol cAMPÆmg )1 Æmin )1 based on comparative protein quantification of the western blot signal indi- cating that in the absence of effector signals the C-terminal domains had no noticeable intrinsic regu- latory input on the catalyst. Discussion We characterized the unorthodox class IIIc AC Rv0386 from M. tuberculosis. AC activity of Rv0386 was surprising because the canonical amino acids which define substrate specificity are replaced in a non- conservative manner, glutamine-asparagine instead of lysine-aspartate. All mammalian membrane-bound ACs possess a strictly conserved and spaced hexad of catalytic residues. Emerging from mostly bacterial gen- ome sequencing projects deviations from this rule occur in a large number of putative AC genes. Actu- ally, predicted open reading frames for ACs exist where all six canonical amino acids are replaced non- conservatively [4]. Viewed from the structures of mammalian ACs those predicted proteins do not look like they could possibly have any AC activity unless alternate mecha- nisms of catalysis or substrate-binding exist for the conversion of ATP to cAMP [4]. However, the first structures of a variant AC were recently obtained with Rv1900c [14]. There the histidine residue which substi- tutes the canonical transition-state stabilizing aspara- gine does not contact the substrate and mutagenesis shows that it appears not to be involved in catalysis. Furthermore the asparagine-aspartate couple which replaces the usual substrate-specifying lysine-aspartate pair does not bind to the purine moiety and is dispen- sable for catalysis. This implies that the preference of Rv1900c for ATP over GTP is governed by other determinants, e.g. general steric constraints of the purine-binding pocket. aDk 611 53 54 5 2 Fig. 3. Rv0386 holoenzyme. Solubilized Rv0386 holoenzyme (calcu- lated molecular mass, 118 kDa) was analyzed by western blot with a commercial anti-RGSH 4 Ig. The signal of the holoenzyme corres- ponds to 80 ng protein. M. tuberculosis adenylyl cyclase Rv0386 L. I. Castro et al. 3088 FEBS Journal 272 (2005) 3085–3092 ª 2005 FEBS In contrast to Rv1900c, AC Rv0386 shows a differ- ent mechanism, because the glutamine-asparagine couple of Rv0386 is specifically needed for catalysis. Removal of either amide side-chain as in Q57A or N106A mutants abrogated cyclase activity. Thus mul- tiple variants of catalytic pockets seem indeed to exist in class III ACs. The inability of Rv0386 to operate with a consensus lysine-aspartate pair, as demonstrated by the catalytic incompetence of the Q57K, N106D, and Q57K ⁄ N106D mutants, suggests that Gln57 ⁄ Asn106 do bind the purine moiety of the substrate, but in a different mode compared to canonical ACs. The high GC side-activity of Rv0386 indicates that both, adenine and guanine can be accommodated in the substrate-binding pocket via Gln57 ⁄ Asn106. How can the purine be bound by the two amide side- chains? In all structures of canonical ACs, i.e. mam- malian AC, trypanosomal AC and mycobacterial AC Rv1264 the lysine-aspartate couple forms a salt bridge [5,7,26]. Even in Rv1900c the asparagine-aspartate pair is connected by a hydrogen bond when the sub- strate-binding pocket is unoccupied [14]. It is there- fore plausible to assume that Gln57 and Asn106 are similarly bonded in Rv0386. We propose that Gln57 and Asn106 are arranged in positions that could accommodate either a guanine or an adenine moiety (Fig. 4A,B). Mutation of either one would therefore be expected to abolish all cyclase activity, as has been observed experimentally. The results with the N106S point mutation, i.e. maintaining cyclase activity and enhancing ATP substrate specificity, are compatible with the proposed mechanism, because Ser106 could pair with Gln57 in the ATP-binding conformation (Fig. 4C). It should be noted that a related ‘amide switch’ mechanism of purine binding and specificity has previously been identified in mammalian cyclic- nucleotide phosphodieserases based on crystal struc- tures [27–29]. The specific activity of Rv0386 was robust and easily measurable with precision, yet, it represents a low activity CHD when compared to other bacterial class III ACs. Nevertheless this does not mean that cAMP production by Rv0386 is physiologically irrelevant. Both, high activity and low activity CHDs have been described previously in M. tuberculosis. High activity CHDs are Rv1625c (2 lmol cAMPÆmg )1 Æmin )1 ) [8], Rv1264 (1 lmol cAMPÆmg )1 Æmin )1 ) [17], Rv1900c (1 lmol cAMPÆmg )1 Æmin )1 ) and Rv1647 (3 lmol cAMPÆ mg )1 Æmin )1 ) [19]. Low activity CHDs are present in Rv1318c (0.3 nmol cAMPÆmg )1 Æmin )1 ), Rv1319c (7 nmol cAMPÆmg )1 Æmin )1 ), Rv1320c (0.2 nmol cAMPÆ mg )1 Æmin )1 ) and Rv3645 (9 nmol cAMPÆmg )1 Æmin )1 ) [18]. The low activity CHD of Rv3645 can be stimula- ted by almost two orders of maginitude via the adjoin- ing HAMP domain [18]. Thus a regulatory input can greatly enhance catalytic efficiency of a low activity CHD in the background of a holoenzyme. We envisage that the Rv0386 holoenzyme, which has a low AC activity comparable to the isolated CHD, will be stimu- lated by an as yet unknown effector. A further argu- ment in favour of a physiological relevance of the AC activity of Rv0386 lies in the evolution of the protein. The glutamine-asparagine couple is not a degenerate mutation, but specifically required for catalysis. Thus AC activity appears to have been retained by a pressure of selection. However, the biological function of Rv0386 in M. tuberculosis is unclear at this point. Recently, in a transposon-mutagenesis screen of M. tuberculosis a knock out mutant of Rv0386 (or all other presumed A PTA-6830vR B PTG-6830vR 601nsA 75 nlG P P Pbi R N N N N N H H O N H H O N H H 6 01ns A 7 5nl G P P PbiR N N N N O HN 2 H N O H H N O H H C PTA-S601N-6830vR 6 0 1reS 75nlG P P Pb iR N N N N N H H O N H H O H Fig. 4. Proposed mode of purine binding. Adenine and guanine binding in Rv0386 by paired Gln57 and Asn106 residues is based on a pos- sible amide switch (compare A and B). (C) Increased specificity for ATP in the Rv0386 (1)175) N106S mutant are the consequence of the hydrogen bond donor property of the serine. The ribose-5¢-triphosphate moiety is abbreviated by rib-P-P-P. L. I. Castro et al. M. tuberculosis adenylyl cyclase Rv0386 FEBS Journal 272 (2005) 3085–3092 ª 2005 FEBS 3089 mycobacterial AC genes) was viable under cell culture conditions [30]. Considering the elusive pathogenic pathways of mycobacteria in its host this does not exclude a vital function of Rv0386 under pathophysio- logical survival conditions. This suggestion is partic- ularly justified because a comparative genomic analysis of several mycobacterial strains identified Rv0386 as one of a few genes which are specifically retained in the M. tuberculosis complex while being lost in other strains, e.g. it is absent in M. smegmatis [31]. The substrate specificity of Rv0386 may be deter- mined by the concentrations of available ATP and GTP at the cellular location of the enzyme, yet the intracellular concentrations of ATP and GTP in M. tuberculosis are unknown to date. In fact, with the exception of Synechocystis [32] meaningful and unequi- vocal cellular cGMP levels have not been reported to date in M. tuberculosis nor in any other bacteria. In the finished genome of M. tuberculosis 10 putative pro- teins were identified which contain a cyclic nucleotide- binding domain [15]. However, no cGMP specificity has been predicted for any of these proteins. We are aware, of course, that this does not exclude the exist- ence of novel, hitherto unknown cGMP binding pro- teins in the pathogen. In conclusion the characterization of AC Rv0386 in this study reveals novel aspects in several respects. It has a completely novel mechanism of substrate binding so far not observed in other class III ACs. It has a rather striking new domain composition comprising an AAA-ATPase and transcription factor module with broad physiological implications to be elucidated. Experimental procedures Materials Radiolabelled nucleotides were from Hartmann Analytik (Braunschweig, Germany). Genomic DNA from M. tuber- culosis was from Dr Boettger (University of Zu ¨ rich Medical School, Switzerland). pBluescriptII SK(–) (Stratagene, Hei- delberg, Germany) was used for general cloning and pQE30 (Qiagen, Hilden, Germany) for expression in. Ni 2+ -nitrilo- triacetic acid-agarose slurry was from Qiagen. The anti- RGSH 4 antibody was obtained from Qiagen, the secondary antibody from Dianova, (Hamburg, Germany). Peroxidase detection was carried out with the ECL-Plus kit (Amer- sham-Life Sciences). Plasmid Construction The open reading frame of gene Rv0386 (GenBank Acces- sion Number BX842573) was amplified by PCR using specific primers and genomic DNA as a template and a BamHI and a HindIII site were added to the 5¢- and 3¢- ends, respectively. To remove the internal BamHI site a silent AfiT mutation was introduced at nucleotide 57. The PCR product was inserted into pQE30, adding an N-terminal MRGSH 6 GS tag. Similarly, the catalytic domain (Rv0386 (1)175) ) was fitted with a 5¢ Bam HI and a 3¢ HindIII site and inserted into pQE30. Point mutations were introduced by PCR using the expression cassette as a template and standard molecular biology techniques. The correctness of all DNA inserts was checked by double- stranded DNA sequencing. Primer sequences are available on request. Expression and purification of proteins Plasmids containing Rv0386 (1)175) or its mutants were transformed into E. coli BL21(DE3)[pRep4]. Protein expression was induced by 60 lm isopropyl-thio-b-D-gal- actoside for 3–5 h at 22 °C. Bacteria were washed once with buffer (50 mm Tris ⁄ HCl, 1 mm EDTA, pH 8) and stored at )80 °C. For purification cells from 200 to 600 mL culture were suspended in 25 mL of lysis buffer (50 mm Tris ⁄ HCl, pH 8, 50 mm NaCl, 10 mm 2-mercaptoethanol), lysed by sonication for 30 s and treated for 30 min with 0.2 mgÆmL )1 lysozyme on ice. Subsequently 5 mm MgCl 2 and 20 lgÆmL )1 DNAseI were added for 30 min. After cen- trifugation (31 000 g, 30 min) 15 mm imidazole pH 8 and 250 mm NaCl (final concentrations) were added to the supernatant. Protein was equilibrated for a minimum of 60 min with 250 lLNi 2+ ⁄ nitrilotriacetic acid agarose on ice, then transferred to a column and successively washed with 3 mL each of buffer A (lysis buffer containing 5 mm imidazole, 400 mm NaCl and 2 mm MgCl 2 ), buffer B (lysis buffer containing 15 mm imidazole, 400 mm NaCl and 2mm MgCl 2 ) and buffer C (lysis buffer containing 15 mm imidazole, 10 mm NaCl and 2 mm MgCl 2 ). The protein was eluted with 0.4 mL of buffer D (lysis buffer containing 150 mm imidazole, 10 mm NaCl and 2 mm MgCl 2 ). Puri- fied proteins were dialyzed against buffer E (50 mm Tris ⁄ HCl, pH 8, 10 mm NaCl, 2 mm 2-mercaptoethanol, 20% glycerol) and stored at )20 °C. The enzyme was stabile for several weeks at least. Cyclase assays AC activity was determined at 30 °C for 20 min in 100 lL [33]. The reactions contained 50 mm 3-(N-morpholino)- propanesulfonic acid pH 7.5, 22% glycerol, 5 mm MnCl 2 , 850 lm [ 32 P]ATP[aP] and 2 mm [2,8- 3 H]cAMP. The kinetic analysis was conducted from 10 lm to 2.3 mm ATP and kinetic constants were derived from a Hanes-Woolfe plot. GC activity was determined identically by using guanine nucleotides instead of the respective adenine nucleotides [34]. M. tuberculosis adenylyl cyclase Rv0386 L. I. Castro et al. 3090 FEBS Journal 272 (2005) 3085–3092 ª 2005 FEBS Western blot analysis Protein was mixed with sample buffer and subjected to SDS ⁄ PAGE (15%). The gel was blotted onto poly(vinylid- ene difluoride) membranes and probed sequentially with a commercial anti-RGSH 4 Ig and with a 1 : 5000 dilution of peroxidase conjugated goat anti-(mouse IgG) Ig as a secon- dary antibody. Sequence analyses INTERPRO-scans (http://www.ebi.ac.uk/InterProScan/index. html) and smart analysis (simple modular architecture research tool; http://smart.embl-heidelberg.de/) were per- formed. Acknowledgements This work was supported by the Deutsche Forschungs- gemeinschaft. References 1 Barzu O & Danchin A (1994) Adenylyl cyclases: a het- erogeneous class of ATP-utilizing enzymes. Prog Nucl Acid Res Mol Biol 49, 241–283. 2 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 hyperthermo- philic archaebacteria. J Bacteriol 180, 3339–3344. 3 Cotta MA, Whitehead TR & Wheeler MB (1998) Iden- tification of a novel adenylate cyclase in the ruminal anaerobe, Prevotella ruminicola D31d. FEMS Microbiol Lett 164, 257–260. 4 Linder JU & Schultz JE (2003) The class III adenylyl cyclases: multi-purpose signalling modules. Cell Signal 15, 1081–1089. 5 Tews I, Findeisen F, Sinning I, Schultz A, Schultz JE & Linder JU (2005) The structure of a pH-sensing myco- bacterial adenylyl cylase holoenzyme. Science 308, in press. 6 Whisnant RE, Gilman AG & Dessauer CW (1996) Interaction of the two cytosolic domains of mammalian adenylyl cyclase. Proc Natl Acad Sci USA 93, 6621– 6625. 7 Tesmer JJ, Sunahara RK, Gilman AG & Sprang SR (1997) Crystal structure of the catalytic domains of ade- nylyl cyclase in a complex with G s a. GTPcS. Science 278, 1907–1916. 8 Guo YL, Seebacher T, Kurz U, Linder JU & Schultz JE (2001) Adenylyl cyclase Rv1625c of Mycobacterium tuberculosis: a progenitor of mammalian adenylyl cyclases. EMBO J 20, 3667–3675. 9 Tesmer JJ, Sunahara RK, Johnson RA, Gosselin G, Gilman AG & Sprang SR (1999) Two-metal-Ion cata- lysis in adenylyl cyclase. Science 285 , 756–760. 10 Yan SZ, Huang ZH, Shaw RS & Tang WJ (1997) The conserved asparagine and arginine are essential for cata- lysis of mammalian adenylyl cyclase. J Biol Chem 272, 12342–12349. 11 Tucker CL, Hurley JH, Miller TR & Hurley JB (1998) Two amino acid substitutions convert a guanylyl cyclase, RetGC-1, into an adenylyl cyclase. Proc Natl Acad Sci USA 95, 5993–5997. 12 Kanacher T, Schultz A, Linder JU & Schultz JE (2002) A GAF-domain-regulated adenylyl cyclase from Ana- baena is a self- activating cAMP switch. EMBO J 21, 3672–3680. 13 Steegborn C, Litvin TN, Levin LR, Buck J & Wu H (2005) Bicarbonate activation of adenylyl cyclase via promotion of catalytic active site closure and metal recruitment. Nat Struct Mol Biol 12, 32–37. 14 Sinha SC, Wetterer M, Sprang SR, Schultz JE & Linder JU (2005) Origin of asymmetry in adenylyl cyclases: structures of Mycobacterium tuberculosis Rv1900c. EMBO J 24, 663–673. 15 McCue LA, McDonough KA & Lawrence CE (2000) Functional classification of cNMP-binding proteins and nucleotide cyclases with implications for novel regula- tory pathways in Mycobacterium tuberculosis. Genome Res 10, 204–219. 16 Reddy SK, Kamireddi M, Dhanireddy K, Young L, Davis A & Reddy PT (2001) Eukaryotic-like adenylyl cyclases in Mycobacterium tuberculosis H37Rv: cloning and characterization. J Biol Chem 276, 35141–35149. 17 Linder JU, Schultz A & Schultz JE (2002) Adenylyl cyclase Rv1264 from Mycobacterium tuberculosis has an autoinhibitory N-terminal domain. J Biol Chem 277, 15271–15276. 18 Linder JU, Hammer A & Schultz JE (2004) The effect of HAMP domains on class IIIb adenylyl cyclases from Mycobacterium tuberculosis. Eur J Biochem 271, 2446– 2451. 19 Shenoy AR, Sreenath NP, Mahalingam M & Viswes- wariah SS (2005) Characterization of phylogenetically distant members of the adenylyl cyclase family from mycobacteria: Rv1647 from M. tuberculosis and its ortholog ML1399 from M. leprae. Biochem J 387, 541–551. 20 van der Biezen EA & Jones JD (1998) The NB-ARC domain: a novel signalling motif shared by plant resis- tance gene products and regulators of cell death in animals. Curr Biol 8, R226–R227. 21 Showalter RE, Martin MO & Silverman MR (1990) Cloning and nucleotide sequence of luxR, a regulatory gene controlling bioluminescence in Vibrio harveyi. J Bacteriol 172, 2946–2954. L. I. Castro et al. M. tuberculosis adenylyl cyclase Rv0386 FEBS Journal 272 (2005) 3085–3092 ª 2005 FEBS 3091 22 Horinouchi S, Kito M, Nishiyama M, Furuya K, Hong SK, Miyake K & Beppu T (1990) Primary struc- ture of AfsR, a global regulatory protein for secondary metabolite formation in Streptomyces coelicolor A3 (2). Gene 95, 49–56. 23 Liu Y, Ruoho AE, Rao VD & Hurley JH (1997) Cata- lytic mechanism of the adenylyl and guanylyl cyclases: modeling and mutational analysis. Proc Natl Acad Sci USA 94, 13414–13419. 24 Sunahara RK, Beuve A, Tesmer JJ, Sprang SR, Gar- bers DL & Gilman AG (1998) Exchange of substrate and inhibitor specificities between adenylyl and guanylyl cyclases. J Biol Chem 273, 16332–16338. 25 Zhang G, Liu Y, Ruoho AE & Hurley JH (1997) Struc- ture of the adenylyl cyclase catalytic core. Nature 386, 247–253. 26 Bieger B & Essen LO (2001) Structural analysis of ade- nylate cyclases from Trypanosoma brucei in their mono- meric state. EMBO J 20, 433–445. 27 Xu RX, Hassell AM, Vanderwall D, Lambert MH, Holmes WD, Luther MA, Rocque WJ, Milburn MV, Zhao Y, Ke H & Nolte RT (2000) Atomic structure of PDE4: insights into phosphodiesterase mechanism and specificity. Science 288, 1822–1825. 28 Sung BJ, Yeon Hwang K, Ho Jeon Y, Lee JI, Heo YS, Hwan Kim J, Moon J, Min Yoon J, Hyun YL, Kim E et al. (2003) Structure of the catalytic domain of human phosphodiesterase 5 with bound drug molecules. Nature 425, 98–102. 29 Zhang KY, Card GL, Suzuki Y, Artis DR, Fong D, Gillette S, Hsieh D, Neiman J, West BL, Zhang C et al. (2004) A glutamine switch mechanism for nucleo- tide selectivity by phosphodiesterases. Mol Cell 15, 279–286. 30 Sassetti CM, Boyd DH & Rubin EJ (2003) Genes required for mycobacterial growth defined by high den- sity mutagenesis. Mol Microbiol 48, 77–84. 31 Gamieldien J, Ptitsyn A & Hide W (2002) Eukaryotic genes in Mycobacterium tuberculosis could have a role in pathogenesis and immunomodulation. Trends Genet 18, 5–8. 32 Ochoa De Alda JA, Ajlani G & Houmard J (2000) Synechocystis strain PCC 6803 cya2, a prokaryotic gene that encodes a guanylyl cyclase. J Bacteriol 182, 3839– 3842. 33 Salomon Y, Londos C & Rodbell M (1974) A highly sensitive adenylate cyclase assay. Anal Biochem 58, 541–548. 34 Klumpp S & Schultz JE (1982) Characterization of a Ca 2+ -dependent guanylate cyclase in the excitable cili- ary membrane from Paramecium. Eur J Biochem 124, 317–324. M. tuberculosis adenylyl cyclase Rv0386 L. I. Castro et al. 3092 FEBS Journal 272 (2005) 3085–3092 ª 2005 FEBS . that a variety of catalytic mechanisms may actually exist. The gene Rv0386 from Mycobacterium tuberculosis codes for an adenylyl cyclase catalytic domain. Adenylyl cyclase Rv0386 from Mycobacterium tuberculosis H37Rv uses a novel mode for substrate selection Lucila I. Castro, Corinna Hermsen, Joachim