JOURNAL OF BACTERIOLOGY, May 1990, p. 2614-2621 Vol. 172, No. 5 0021-9193/90/052614-08$02.00/0 Copyright X 1990, American Society for Microbiology Rhizobium meliloti Adenylate Cyclase Is Related to Eucaryotic Adenylate and Guanylate Cyclases A. BEUVE,1 B. BOESTEN,2 M. CRASNIER,l A. DANCHIN,l* AND F. O'GARA2 Institut Pasteur, 28, rue du Dr Roux, 75724 Paris Cedex 15, France,' and Microbiology Department and National Food Biotechnology Centre, University College, Cork, Ireland2 Received 25 October 1989/Accepted 8 February 1990
A gene from Rhizobium meliloti coding for an adenylate cyclase was sequenced, and the deduced protein sequence was compared with those of other known adenylate cyclases. No similarity could be detected with the procaryotic counterparts. However, striking similarity was found with the catalytic region of Saccharomyces cerevisiae adenylate cyclase, the cytoplasmic domains of bovine adenylate cyclase, and two mammalian guanylate cyclases. The gene was fused to the enteric j-galactosidase, and the chimeric protein was purified by affinity chromatography. This fusion protein was found to direct the synthesis of cyclic AMP in vitro. This activity was strongly inhibited by the presence of GTP, but no cyclic GMP synthesis could be detected in conditions permitting cyclic AMP synthesis.
Rhizobia are gram-negative bacteria that are able to enter MATERIALS AND METHODS symbiotic association with legume plants, leading to the Bacterial strains and plasmnids. E. coli strains and plasmid formation of N2-fixing root nodules. This process involves vectors are described in Table 1. Strain TP610A was selected differential gene expression between the host plant and as a derivative of strain TP610 on maltose plates in the bacteria, but the signals involved in the modulation of gene presence of 50 ,uM cAMP. Plasmid pCU5 (18), which can expression are not well understood. Cyclic AMP (cAMP) complement E. coli strains that are defective in cAMP might play a role in regulating metabolism of rhizobia. synthesis, was used as a source of the R. meliloti cya gene Indeed, in Bradyrhizobium japonicum, it has been shown for DNA sequencing and for creating a gene fusion to the that s-ynthesis of glutamate synthetase and glutamate dehy- lacZ gene of E. coli. Plasmid pMC1403 (3) provided a drogenase is repressed by cAMP (40). Genes presumably cartridge containing the enteric ,-galactosidase gene trun- encoding an adenylate cyclase were isolated from both B. cated at its 5' end. Plasmid pDIA7010 was created during our japonicum (9) and Rhizobium meliloti (12). The gene cloned work by cloning the P-galactosidase gene from pMC1403 from R. meliloti only weakly complemented an Escherichia into pCU5. coli strain that was defective in cAMP production (12). It Growth media. M63 minimal medium was prepared as was therefore necessary to identify the putative cya gene described by Miller (25) and supplemented with the appro- product as an authentic adenylate cyclase. This prompted us priate carbon source (0.4%) and with amino acids (40 ,ug/ml) to determine the nucleotide sequence of the gene and to as required. MacConkey medium supplemented with the purify the corresponding gene product for enzymological appropriate carbon source (1%) was used as an indicator assessment. In particular, it was important to assay whether medium for sugar utilization. Ampicillin was used at 50 the protein could synthesize cyclic GMP (cGMP) in addition ,ug/ml (final concentration). to cAMP. Genetic procedures. Plasmid preparations, restriction en- Previous experiments with the maxicell technique had zyme digestions, and ligations were performed as described suggested that the region encompassing the R. meliloti cya by Maniatis et al. (21). Competent bacterial cells were gene could direct synthesis of two proteins (18). To identify prepared by the technique of Chung and Miller (5). the translational product that was endowed with adenylate Nucleotide sequencing. Standard methods for DNA se- cyclase activity, a gene fusion with the E. coli lacZ gene was quencing by using dideoxy sequencing (32) and M13 single- performed, and the corresponding P-galactosidase hybrid stranded phage (24) were employed. Sequential deletions of protein was purified. the A new class of phylogenetically related eucaryotic en- M13 clones were generated by using DNase I as de- zymes directing synthesis of cAMP and cGMP was recently scribed by Lin et al. (19). The sequences of both strands discovered (4, 11, 13, 14, 35, 37). It was therefore of interest were determined independently in our two laboratories. The to investigate whether the R. meliloti enzyme had any DNA sequence information was analyzed in Cork by using relationship with this new class. As demonstrated below, the Microgenie sequence analysis program (Beckman Instru- this was indeed the case; the R. meliloti enzyme, although ments, Inc., Fullerton, Calif.). The data bank was the able to recognize GTP, was not able to synthesize cGMP but Genbank compilation of DNA and protein sequences (Na- could synthesize cAMP at the expense of ATP. tional Institutes of Health, Bethesda, Md.). In Paris, se- This indicates either horizontal gene exchange or, more quences were analyzed by using the Institut Pasteur soft- probably, that synthesis of cyclic nucleotides was already ware designed by the Unitd d'Informatique Scientifique. present very early in evolution before the separation be- Hybrid adenylate cyclase- -galactosidase protein purifica- tween tion and (-galactosidase assays. Bacteria were disrupted by eucaryotes and procaryotes. sonication in the presence of 2 M NaCl added to a 20 mM Tris hydrochloride-10 mM MgC12 buffer (pH 7.4) (6). The purification of the hybrid protein by affinity chromatography * Corresponding author. on TPEG (thiophenylethyl-,-D-galactoside)-Sepharose was 2614 VOL. 172, 1990 RHIZOBIUM MELILOTI ADENYLATE CYCLASE 2615
TABLE 1. Bacteria and plasmids 5.4-kilobase (kb) BglII fragment, was previously subcloned a DNA Strain or Genotype or Origin or into various high-copy-number vectors, and 0.8-kb plasmid relevant features reference segment was found to be able to weakly complement cya- defective E. coli strains (18). The nucleotide sequence of the E. coli 1.5-kb BamHI-EcoRI fragment cloned in pBR322 (pCU5) TP610 F- thi-J thr-J leuB6 pro lacY1 tonA21 10 and containing the 0.8-kb PstI-EcoRI core fragment of the supE44 hsdR hsdM recBC lop-il is in 1. lig' cya-610 cya gene presented Fig. TP610A As TP610 but carrying a mutation This work Two long open reading frames (ORFi and ORF2) that permitting growth on maltose in the overlap, as well as a third one (ORF3) in the opposite strand presence of 50 ,um cAMP and coinciding with ORFi, are prominent. ORFi spans TP2010 F- xyl Acya argH1 AlacX74 recA 28 almost the entire fragment and stops 39 base pairs (bp) TP2339 F- xyl Acya Acrp-39 AlacX74 argH1 31 upstream from the EcoRI site. At its 5' end it is in frame with MC1060 AlacX74 galU galK rpsL hsdR 3 the tet gene of pBR322 present in pCU5 and pDIA7010, as Y1083 A(lacIPOZYA) U169 proA+ Alon 41 well as with the cat gene of pACYC184 present in pCUll, araDJ39 rpsL thi and used by Lathigra et al. (18). are enriched in codons with a Plasmids Translated ORFs starting pMC1403 Contains the lac'ZYA gene truncated 3 purine and ending with a pyrimidine (34). Analysis of the at its 5' end, Ampr RNY triplets of ORFi and its third base G+C content pCU5 pBR322 containing the cya gene of R. 18 strongly suggested that it might code for a protein (34). Since meliloti cloned at the BamHI most of the known adenylate cyclase proteins comprise a restriction site catalytic domain coupled to a regulatory domain (11, 16, 22, pDIA7010 Constructed by cloning the lac'ZYA This work 30), it seemed important to identify the catalytic domain of gene of pMC1403 downstream from the R. meliloti protein. The DNA sequence revealed, inter- the TGA stop codon of the cya nal to ORFi and downstream from the second PstI restric- gene of pCU5 tion a consensus site at pDIA7010A Derivative of pDIA7010 containing an This work site, perfect ribosome-binding posi- in-phase cya-lacZ gene fusion tion 875 followed by an ATG start codon 7 nucleotides obtained by a spontaneous deletion downstream. As discussed below, it appeared that in E. coli pDIA7020 Derivative of pDIA7010A with a This work this start codon serves as an initiator of translation and deletion of the 0.15-kb XhoI directs synthesis of a protein endowed with adenylate cy- restriction fragment clase activity. ORF2 extended from the TGA at position 134 pDIA7030 Derivative of pDIA7010A with a This work to the TGA at position 1502. There was no obvious signal for deletion of the 0.298-kb NotI initiation of translation, and this ORF did not score highly restriction fragment with the RNY analysis. Furthermore, it featured an unlikely high number of arginine residues. Because of the high G+C content in the third base of ORFi and the fact that this performed as described by Ullmann (39). The eluted protein third-base position of ORFi coincides with the first base of was dialyzed against 50 mM phosphate buffer (pH 7.4). ORF2 (66% in the present case), the probability of finding a 3-Galactosidase activities were measured as described by termination codon (A+T rich) in ORF2 is low, thus explain- Pardee et al. (27). The amount of protein was determined as ing the presence of such a long noncoding ORF. Similarly, described by Bradford (2). ORF3 on the opposite strand, extending over the entire Protein electrophoresis and immunoblotting. Polyacryl- 1.5-kb fragment, also shared its first base of the triplet code amide gel electrophoresis of proteins in sodium dodecyl with the third base of ORF1. Although fitting with the RNY sulfate and electrophoretic blotting of these proteins onto rule (because it is in frame with ORF1), ORF3 did not reveal nitrocellulose were carried out by the procedures of Laem- a high third-base G+C content as expected. It is therefore mli (17) and Towbin et al. (38), respectively. The fusion unlikely that it codes for a protein. protein was visualized by Western immunoblotting with Upstream from the ribosome-binding site and the ATG anti-p-galactosidase antibodies (diluted 1:200) by using a start codon located at position 875 in ORF1, several G+C- reaction with peroxidase-conjugated antibodies provided by rich palindromic sequences and inverted repeats were Diagnostic Pasteur. present, located at positions 325 through 345, 360 through Adenylate cyclase and guanylate cyclase assays. In vitro 430, and 791 through 912. This suggested that a correspond- cyclase reactions were performed at 28°C in 10 mM MgCl2- ing transcript might be folded in stable hairpin structures, 2.5 mM dithiothreitol-60 mM Tris hydrochloride buffer at possibly preventing translation of the corresponding region various pHs and ATP and GTP concentrations. ATP and (15). This puzzling organization, together with the two GTP were obtained from Sigma Chemical Co. (St. Louis, translational products of 65 and 28 kilodaltons (kDa) previ- Mo.). Reactions were stopped with 50 mM sodium acetate ously found in maxicells (18), prompted us to identify the buffer (pH 6.2) and incubated at 65°C for 5 min. Measure- corresponding protein. ments of adenylate cyclase activity were performed by Construction of cyaA-kacZ fusion. To purify the R. meliloti monitoring cAMP synthesis with a radioimmunological as- adenylate cyclase protein, a fusion gene that expressed a say with 1251I-labeled cAMP from Institut Pasteur Production hybrid protein with both adenylate cyclase and ,B-galactosi- and anti-cAMP antibodies kindly provided by A. Ullmann. A dase activities was constructed. First, the SmaI-SalI frag- cGMP radioimmunological assay kit from Amersham was ment of pMC1403 was inserted in pCU5 between the SmaI used to assay for guanylate cyclase activity (36). and Sall sites (Fig. 2). The plasmid pDIA7010 thus obtained was still able to complement the cya-deficient strain E. coli RESULTS TP610A, but no p-galactosidase activity could be detected. Nucleotide sequence of the R. meliloti cyaA gene. The R. This was as expected, because the stop codon of the putative meliloti adenylate cyclase gene, originally located on a cya gene was located 79 bp upstream from the SmaI site 2616 BEUVE ET AL. J. BACTERIOL.
il I f I d a s g I q f a e r h a a s n g f d p r t r p w y r a a v n g k a p v a GGATCCTGTTCCTGGACGCGAGCGGCCTGCAGMGCCGAACGTCACGCTGCCTCCAACGGCTTCGATpGAGGACGCGGCCCTGGTACCGCGCGGCCGTCMCGGCAGGCGCCGGTGG 4uiIla 10 20 i g p y e m a t t g n I g m t il s q a h r g n p q i v i g a d v v I d t i t d f CCATCGGTCCTATGAGATGGCCACCACAGGCAATCTCGGATGACCATATCGCMGCGCACCGCrGG CCCCMATCnTCATCGGCGCCGATGTCGTTCTCGATACGATCACGGATT 130 140 150 160 170 180 190 200 210 220 230 240 1 s r e r I t d d s v s f v I d a v g r p 1 i h s d s t m m r r il m a s k g r d GCCGC6AGCGGCTGACcGACGACTCGGmCCTTCGTGCTcGATGcGGTGGAcGACCGATcATccAcTCCGAccCAcCATGATGCGGCGCATCATGCATCGAAGGGCCGGG 250 260 270 280 290 300 310 320 330 340 350 360 r p v a t p q e d g I i e s i r r n p p p a g k a t I v e v g n r t y I v t v a ACCGrCCGGTGGCCACGCCGCG TGGACTGATCGAGAGCAT GGCGCAACCCCCACCGGCCGAMGGC ACTCTCGTCG AGTCGGAACCGCACCTATCTCGTCACGGTGG 370 380 390 400 410 420 430 440 450 460 470 480 p I e s a 1 I I s g h r v v v a a p I d e I I a a a n e t I v q g I a v s g a v CGCCGCTCGAATCGGCATTGCTTCTGTCCGGGCACCGGGTGGTCGTCGCCGCCCCTCTCGACGAGCTGCTGGCGGCCGCMAMCGAGACGCTCGTTCAGGGACTTGCCGTCTCGGGCGCCG 490 500 510 520 530 540 550 < Notl > 570 580 590 600 v v v a v I I a I v I a h I i t k s I n q I t d s a n r I q d I d f a t p i d v TGGTGGTaGTCGCCGatCTCCTGGCCCTCGTGCTTGCGCPATCTGATCACGACGCTCACCAGCTCACCGACAGCGCCACCGCCTGCAiVCTGGAMCGCCACTCCTATCGACG 610 620 630 640 650 660 670 680 s s h v a e il s t I n g a ml n r a r d a il f t f a I y v p k e I v r k a i e s_g IvCGCATGTGGCGGAMTCTCGACGCTCAACGGCGCAATGAACAGGGCTCGCGACGCGATCTCACCTTCCGCTCTATGTTCCGAAGGAGCTGGTGCGGCA ATCATCCG 730 740 750 760 770 780 790 800 810 820 830 840 h f a a r a a w r a e v t a M F T D I Y D F T T I S E G R S P E E V V A H L S E GCCAMCGGCGGCCGCGCCGCATGGCGGA TACGGCA TGTTCA DFGACATCTACGATCACCACCATCAGCGAGGGCCGGTCGCCGGMEGMGTGGTCGCGATGCTCTCGG 850< Notl > 870 880 890 900 910 920 930 940 950 960 Y F D L F S E V V A A H D G T I I Q F H G D S V F A M W N A P V A D T R H A E H AGTAMCGACCTGTTCAGCAGTCGTCGCCGCCCACGACGGMACCATCATCCMTTCCATG6GATCGGTCMTGCCAT6TG6MCGCGCCGGTCGCCGATACCAGGCATGCCGAGC 970 980 990 1000 1010 1020 1030 1040 1050 1060 1070 1080 A C R C A L A V E E R L E A F N S A Q R A S G L P E F R T R F G I H T G T A V V ATGCCTGTCGATGCGCACTCGCGGTCGAGGAGAGGCTCGAQGCCTTCMTTCTGCGCAACGCCAGCGGATTGCCLP AGTTCCGCACCCGCTTCGGCATCCACACCGGAACGGCCGTCG 1090 1100 1110 4Xhol 1130 1140 1150 1160 1170 1180 1190 1200 G S V G A K E R L Q Y T A M G D T V N V A S R L E G H N K D Y G T S V L A S G A TCGGCAGCGTCGGCGCCMGGAAGCTGCMTATACGGCGATGGGCGACACGTGMCGTCGCCTCGCGGCTCGAGGGCATM CMGGATTACGGCACGAGCGTTCTTTMAGCGGCG 1210 1220 1230 1240 1250 1260 1270 4Xhol 1290 1300 1310 1320 V V A Q C K D M V K F R P L G T A K A K G R S T A L D I Y E V V G V V R A V N T CGGTGGTCGCCCAATGCMAGCTGGTGAAGTTCCGCCCGCTCGGCASCCAGGCOGMAGGGCCGTTCGACGGCGCTC6ACAMACGAAGTCGTGGGCGTCGTCCGCGCGGTGAACA 1330 1340 1350 1360 1370 1380 1390 1400 1410 1420 1430 1440 T E A G T A A * CTACCGAAGCCGGMACGGCCGCCTGAGGMAGGCAGATGCCGCGGCGAACGGCGGCCCCGCTGAATTCGCTTCGAAACTCTGAAAGAMMAGCCCGGMACCCGGGCTTIlTTGACT 1450 1460 1470 1480 1490 1500 corresponding to the truncated lacZ gene start. Strain was chosen for further study. Strain TP2010 transformed TP2339 (AlacX74) was then transformed with plasmid with pDIA7010A was able to grow on minimal medium pDIA7010. To obtain a spontaneous cya-lacZ fusion, mu- supplemented with 0.4% lactose or 0.4% maltose. The tants that restored growth on minimal medium supplemented 0.28-kb XhoI-BamHI fragment of pDIA7010A that included with 0.4% lactose were then selected, and it was determined the junction of the gene fusion was sequenced. It showed a that the mutation permitting growth on lactose was carried deletion of 34 bp that had removed the TGA stop codon and by the plasmids thus obtained. Strain TP2010 was subse- resulted in an in-phase fusion of the cya and lacZ reading quently transformed with the mutant plasmids. One such frames. The resulting gene fusion also confirmed that all or plasmid (pDIA7010A), shown to be cya and lacZ positive, part of ORF1 was expressed in E. coli under the present VOL. 172, 1990 RHIZOBIUM MELILOTI ADENYLATE CYCLASE 2617 1 kb -- I cru E 0 hJ L C ul) LUJ I I pMC 1403 ...... -...... / r \1ac ZYA ...... Amp \ ~~/ \ ~~/ _ -I_ ~~/ \=-rF a: O E O E= 0 L) LUI I I 11u1j1 LO I Ii 'IIII II I pCU5 ...... I...... ,...... Il ...... _1_ 1- ~~~~~~~~~~~.~ r ptet cyaA Amp o E o 0 F- C-) LU c LULLO I uC) LUJIr tIII pDIA7010 :-:---:::::::::: ....:-:--:-:-:::::::::::::::::: r p tet cya A l1ac ZYA Amp FIG. 2. Construction of plasmid pDIA7010. Plasmid pDIA7010 was constructed by cloning the SmaI-SalI fragment, containing the enteric P-galactosidase genes truncated at the NH2 terminus, into pCU5, downstream from the R. meliloti cya gene. IacZYA, E. coli lactose operon genes; ptet, tetracycline resistance promoter; AmpR, ampicillin resistance gene; cyaA, R. meliloti adenylate cyclase gene. conditions. This deletion appeared to be created by a cross- Two main bands were observed (Fig. 4, lane A). The higher over event involving an 11-bp direct repeat (Fig. 3). To band of 135 kDa corresponded to the adenylate cyclase- substantiate the hypothesis that it was generated by homol- P-galactosidase hybrid protein. The second band of 116 kDa, ogous recombination, the experiment designed to provide corresponding to the size of the P-galactosidase, was pre- Lac' colonies in minimal plates was repeated in a recA- sumably a degraded form of the first protein. This latter form deficient background (strain TP2010); no Lac' clones were was predominant under our growth conditions. Other con- recovered after strain TP2010 harboring plasmid pDIA7010 ditions of purification (cells disrupted with alumina, elution was plated under conditions similar to those of the original at lower pH) were tried to increase the amount of the hybrid experiment (data not shown). protein; they did not permit significant improvement of the Biochemical analysis of the adenylate cyclase-s-galactosi- hybrid protein yield. Similar experiments were also carried dase hybrid protein. Bacteria harboring plasmid pDIA7010A out with a protease-deficient strain (Alon) (41) with no were grown to the stationary phase in minimal medium further improvement (data not shown). supplemented with lactose, and their P-galactosidase con- The Mr of the hybrid protein corresponded to that ex- tent was purified as described in Materials and Methods. The pected from the synthesis of a protein translated from the eluted protein was submitted to sodium dodecyl sulfate- ATG at position 885 and was in agreement with the smaller polyacrylamide gel electrophoresis, and Western blotting product previously found in maxicells (18). The protein was analysis was performed with anti-i-galactosidase antibodies. blotted on a polyvinylidene difluoride membrane, which Smoa 1150 1500 I L I RCCGRRGCCGGRRCGGC GCCtGApGRRRGGCRGRTGCCGCIGGCGRRCGGC GGCCCCGCTGRRTTCGCTTCGRRRCTCTGRRAGCRRRRRRGCCCG nRRCCCCGGGGRTCCCGTCGTT ThrGluRlAoGlyThrRI aRloProLouRsnSsrLouRrgRsnSerGluSerLysLysRlaRrgLysProGlyRspProUoalIUal FIG. 3. Nucleotide sequence of the spontaneous deletion in pDIA7010 that resulted in a gene fusion of the R. meliloti cya gene and the lac gene. The deletion is boxed, and the two direct repeats involved in homologous recombination are underlined. The TGA stop codon is boxed. The lacZ gene starts from the SmaI restriction site. 2618 BEUVE ET AL. J. BACTERIOL. B A lIW.(kDa) 180 135 ka.__ E 1 16 k.--am am 1 16 0) 5E 84 C( 58 48 8 10 GTP (mM) FIG. 6. GTP inhibition of the adenylate cyclase reaction. Reac- tions were carried out at 4 mM ATP in the presence of different concentrations of GTP. To check that the hybrid protein of 135 kDa was the main FIG. 4. Western blotting analysis of the hybrid prc encoded significant translational product endowed with an adenylate by the plasmids pDIA7010A and pDIA7030. The atein a small deletion was introduced tosidase antibodies were used at a dilution of 1:200, anId the blotting cyclase activity, upstream was performed as described by Towbin et al. (38). Lane-s: A, purified from the ATG codon present at position 885 (see below). In E. coli extract from cells harboring plasmid pDIA7010O A; B, purified this latter case, Western blotting analyses indicated synthe- E. coli extract from cells harboring plasmid pDIA70330. Molecular sis of a hybrid protein with an Mr similar to that of the size markers were from Sigma. protein synthesized from the native plasmid pDIA7010A (Fig. 4, lane B). could be directly used in a gas-phase sequencer (23). Deter- Genetic analysis of plasmid pDIA7010A. Deletion of the mination of the NH2-terminal residues of the hybrid protein 0.15-kb XhoI fragment preserving the translation frame (Fig. was then attempted but failed to yield signific,ant results, 1) in plasmid pDIA7010A led to the construction of plasmid suggesting that under our conditions the N terminus was pDIA7020. Such a deletion maintained P-galactosidase ac- blocked (data not shown). The purified preparati onrwasused tivity, as expected, but abolished the adenylate cyclase to assay adenylate cyclase activity as described in Materials activity. This confirmed that the catalytic center of the and Methods. The reactions were carried out at pH7e5 protein was, at least in part, coded for by this region of the which was found to be the optimum pH for the adenylate ORF. cyclase activity. The hybrid protein synthesized cAMP, and To demonstrate that the translation of the fusion protein is the Km for ATP was determined and estimated to be nmM initiated at the ribosome-binding site at position 875, a (i.e., the usual concentration of ATP inside livini cells) (Fig deletion was created that modified the reading frame up- 5). The enzyme activity was inhibited by a concgentration of stream from this translational start. Plasmid pDIA7010A was substrate of 10 and 20 mM. GTP was a strong inhtibitor of the digested by the restriction enzyme NotI (Fig. 1), generating reaction (Fig. 6). However, no guanylate cycl,ase activity plasmid pDIA7030, which harbored a deletion of a 298-bp was detected under the assay conditions. Fiurthermore, fragment. This resulted in alteration of the ORF upstream measurements on cultures overexpressing the uInfused pro- from the ribosome-binding site and ATG located at positions tein showed no synthesis of cGMP (data not shown). 875 and 885, respectively. Strains TP2339 and TP610A were transformed with plasmid pDIA7030. They displayed both 3-galactosidase and adenylate cyclase activities. This con- 0,4 - firmed that initiation of translation of the fusion protein could occur at the more downstream ATG codon (located at position 885) and that the corresponding protein could 0,3- synthesize cAMP. Analysis of the protein sequence of the cya gene. Analysis of C) E the polypeptide sequence translated from the cya gene (from >n 0,2 - the ATG at (D position 885 to the TGA at position 1464) 0 revealed no significant similarity with other procaryotic E adenylate cyclases either from enterobacteria (1) or from Bordetella pertussis (8) or any other protein translated from the DNA sequences present in the EMBL data bank (release 0, 18) when the algorithm of Lipman and Pearson (20) was used. However, a more thorough comparison of the Rhizo- /t 0 1 2 3 4 5 6 bium cyclase sequence with all other cyclases revealed -1 blocks of identical amino acid residues (Fig. 7). This ap- -1/ Km 1/S (mM) peared to be highly significant, since the regions of identity FIG. 5. In vitro activity of the R. meliloti aden3ylate cyclase were located in regions whose integrity is necessary for protein. Initial velocities were measured as describecd in the text. cAMP synthesis (Saccharomyces cerevisiae [11, 22], R. Lineweaver-Burk plots allowed us to estimate the Km f the reaction meliloti, XhoI deletion). A further substantiation of this to be 4 mM at pH 7.5 and 28°C. interpretation stems from comparison with guanylate and VOL. 172, 1990 RHIZOBIUM MELILOTI ADENYLATE CYCLASE 2619 KLDNKRILFNLLPAHVAQHFLMSNPRNMDLYY-QSYSQVGVMFASIPNFNDFYIELDGNNM 893 BTCYA +L*** * +Q* + *4+R +++++++FTP*+*K * ** 53* CTENITILCLALYENIQQQN4RfTLLRRLQPEISPPTGNLAMVFTDIKSS-TFLWELFPN-- 530 SCCYA +** * * + MFTDIYDMIS-E--G--R 15 RMCYA *+.** **++4* * + KAE--ALLYQILPHS-----VAEQLKRGETVQAEAFDSVTIYFSDIVGFTALSAESTP--- 862 RNCYG *44 +Y*** +*** **++VT*+++* * +A**** ** **** * + 44 KTD--TLLYSVLPPS---VANELRHKRPVPAKRYDNVTILFSGIVGFNAFCSKHASGEG 442 BTCYG GVEC-LRLLNEIIADFDELMDKDFYKDLEKIKTIGSTYMAAVGLAPTAG-T----KAKKCI 947 BTCYA 44 ++ *+.*+ +* * *4+** *++* *4+ +**+.** *4+ AMRTAIKTHNDINRR--QLRIYGGY---E-VKTEGDAFNVAFP-TPTSGLTWCLSVQLKLL 584 SCCYA + + + + + **+. + +.*4+ SPEEVVANLSE-Y--FDLFSEVVAAHDGTIIQFHGDSVFAMWN-APVADTR----HA---- 64 RMCYA + + + * * ** + * 44 **44+**+ *+ --NQVVTLLNDLYTCFDAVID---NFDVYKVETIGDAYNVWSGL-PVRNGQ--L-HAR--- 911 RNCYG * +L******** **++ * * ******+I**** **** * + *** A-MKIVNLLNDLYTRFDTLTDSRKNPFVYKVETVGDKYMTVSGL-PEP----CIHHAR--- 494 BTCYG -SSHLSTADFAIEWFDVLDEINYQSYNDFVLRVGINVGPVWAGV-IGARRPQYDIWGNTVN 1007 BTCYA + + +.** *+. 4.*4** * *+ + + + * * * ** DAQWPEEITSVQDGCQVTDRNGNIIYQGLSVRMGIHWGCPVPELDLVTQRMDY--LGPNVN 643 SCCYA + 4.4. 4. + * ****+*+I*++* +G* ** EHACRCALAVEERLEAFNSAQRASGLPEFRTRFGIHTGTAVVGSVGAKERLQYTAMGDTVN 125 RMCYA *****+ +* *+ 4*+ *******+ * *** * ***** EVA-RHALALLDAVRSFRIRHRPQE--QLRLRIGIHTGPVCAGVWGLKMP-RYCLFGDTVN 968 RNCYG + ++** 4+44 + ++ + I+I ******* +T**+R*+Y********+G*** SI-CHLALDIH{EIAG--QVQVDGE---SVQITIGIHTGEWVTGVIGQRMP-RYCLFGNTVN 548 BTCYG VASRMDSTGVQ-GRIQVTEEVH----RLL-RRGSYRFVCRGKVSVKGKG---EMLTYFLEG 1059 BTCYA *+.*+++.+.+.* *4++4+ 4+4 *+. * * **** *+.4+ .* KMRVQGVA-DGGQIAMSSDFYSEFNKIM-----KYHE--R--V-VKGKESLKEVYGEEIIG 694 SCCYA *+.*++*+ * * ++4* * * *++ **+. + + *++4* VASRLEGMNKDYGTSVLASGAWAQCKDM ---VKFRP-L-GTAKAKGRSTA--LDIYEWG 179 RNCYA ***+-*+L*4+*+.*4+ * 4. * ** **++*KR +* TASRMES-NGEALKIHLSSET-KAVLEEF--DG-FELELRGDVEMK(GGKVR---TYWLLG 1021 RNCYG 4.** *+ ** **4+* *4.+.++ *+* ** ** ***** * *+ LTSRTET-TGEKGKINVSEYTYRCLMTPENSDPQFHLEHRGPVSWKGKKEP--MQ-VWFLS 605 BTCYG FIG. 7. Alignment of catalytic domains from S. cerevisiae adenylate cyclase (SCCYA), R. meliloti adenylate cyclase (RMCYA), Rattus norvegicus guanylate cyclase (RNCYG), Bos taurus guanylate cyclase (BTCYG), and bovine brain adenylate cyclase (BTCYA). Identical residues are marked by asterisks, residues of the same class (33) are indicated by plus signs, and gaps are indicated by dashes. adenylate cyclases from higher eucaryotes (4, 13, 14, 37): the The status of this cya gene, which could be expressed in E. alignments performed by Chinkers et al. (4) and by Krupin- coli, was not clear in R. meliloti until recently, when it was ski et al. (14) highlight the same blocks of identity (Fig. 7). It possible to inactivate it in R. meliloti F34 (26). It could be can be therefore inferred that the procaryotic Rhizobium demonstrated that in these latter mutants the level of cAMP adenylate cyclase shares a common ancestor with adenylate was decreased as compared with the level found in the cyclases from S. cerevisiae and bovine brain and also with wild-type parental strain. However, a significant level of guanylate cyclases from higher eucaryotes. cAMP was still present in the mutant strain, leading O'Regan et al. to propose that two cya genes exist in R. meliloti (26). DISCUSSION The presence of two adenylate cyclases might permit fine Determination of the nucleotide sequence of a gene from tuning of cAMP synthesis according to growth or symbiotic Rhizobium meliloti complementing cya strains of E. coli conditions. It was therefore important to characterize as allowed characterization of a Rhizobium adenylate cyclase. precisely as possible the protein coded for by the isolated 2620 BEUVE ET AL. J. BACTERIOL. cya gene. It seems important at this stage to emphasize that cAMP or cGMP in archaebacteria would be of unusual the Mr of the protein as expressed in E. coli was about interest. 20,000, a very low value when compared with those of other adenylate cyclases (7-9, 11, 14, 22, 29), and therefore ACKNOWLEDGMENTS presumably constituted nothing more than a catalytic do- We thank G. Lenzen and S. Fanning for their support in sequenc- main. In this respect it is worth noting that plasmid ing the cya gene, A. Ullmann for her help at the affinity chromatog- pDIA7030 directed synthesis of a protein that was able to raphy step, Pat Higgins for technical assistance, and B. George for synthesize cAMP. Indeed, in this construction (NotI dele- expert typing. tion), the only start codon is ATG88. encoding a 193-residue Financial support came from the Centre National de la Recherche protein. Analysis of the DNA sequence located upstream Scientifique (UA1129), Institut Pasteur (13845), EOLAS (SC/BLG/ from the translation start site revealed the presence of a long 002/87), EC Stimulation [ST2J-0478-C (EDB)], and EC BEP ORF. However, this ORF encompasses a region rich in (GB1.5.030 ETR). potential secondary structures. This suggests that a complex LITERATURE CITED regulatory pattern involving translation of the upstream 1. Aiba, H., K. Mori, M. Tanaka, T. Ooi, A. Roy, and A. Danchin. region and coupling to transcription might be operating in R. 1984. The complete sequence of the adenylate cyclase gene of meliloti. This is reminiscent of the situation in S. cerevisiae, Escherichia coli. Nucleic Acids Res. 12:9427-9439. in which adenylate cyclase seems to be produced from 2. Bradford, M. 1976. A rapid and sensitive method for the transcripts of different lengths, leading to different polypep- quantitation of microgram quantities of protein utilizing the tides depending on growth conditions (11, 22). It cannot be principle of protein-dye binding. Anal. Biochem. 72:248-254. ruled out at that 3. Casadaban, M. J., J. Chou, and S. N. Cohen. 1980. In vitro gene completely present, however, other, longer fusions that join an enzymatically active ,-galactosidase seg- translation products can be synthesized from this ORF. In ment to amino-terminal fragments of exogenous proteins: Esch- particular a protein with 41 extra NH2-terminal residues erichia coli plasmid vectors for the detection and cloning of might have been synthesized, starting at ATG762. In this translational initiation signals. J. Bacteriol. 143:971-980. respect it seems worth noting that Lathigra et al. found two 4. Chinkers, M., D. L. Garbers, M. S. Chang, D. G. Lowe, H. polypeptides of 26 and 64 kDa in maxicells when the cya Chin, D. V. Goeddel, and S. Schulz. 1989. A membrane form of gene was placed under the control of the chloramphenicol guanylate cyclase is an atrial natriuretic peptide receptor. Na- In our case the ture (London) 338:78-83. acetyltransferase promoter (18). cya-lacZ 5. Chung, C. T., and R. H. Miller. 1988. A rapid and convenient gene fusion was cloned downstream from the Tet promoter, method for the preparation and storage of competent bacterial and the resulting 135-kDa protein correspond to the fusion of cells. Nucleic Acids Res. 16:3580. the 26-kDa polypeptide with P-galactosidase. The 64-kDa 6. Crenon, I., D. Ladant, N. Guiso, A. M. Gilies, and 0. Barzu. product might derive from translation initiated at an ATG 1986. Characterization of a P-galactosidase hybrid protein car- codon located upstream from the BamHI cloning site and rying the catalytic domain ofEscherichia coli adenylate cyclase. going through the secondary structures present in the region Eur. J. Biochem. 159:605-609. at positions 791 through 912, to be fused to the catalytic 7. Escuyer, V., E. Duflot, 0. Sezer, A. Danchin, and M. Mock. 1988. Structural homology between virulence-associated bacte- center of the protein. Thus, adenylate cyclase might be (as rial adenylate cyclases. Gene 71:293-298. seems to be the case in S. cerevisiae [11]) present in two 8. Glaser, P., D. Ladant, 0. Sezer, F. Pichot, A. Ullmann, and A. forms, one composed of a catalytic center alone and another Danchin. 1988. The calmodulin-sensitive adenylate cyclase of fused to a regulatory domain. Bordetella pertussis: cloning and expression in Escherichia coli. The Km of the enzyme for ATP was 4 mM, which is Mol. Microbiol. 2:19-30. somewhat higher than that of its E. coli counterpart (ca 0.6 9. Guerinot, M. L., and B. K. Chelm. 1984. Isolation and expres- mM) (6). Another important feature of the adenylate cyclase sion of the Bradyrhizobium japonicum adenylate cyclase gene was its inhibition GTP. measurements of (cya) in Escherichia coli. J. Bacteriol. 159:1068-1071. activity by Activity 10. Hedegaard, L., and A. Danchin. 1985. The cya gene region of cGMP were performed on purified preparations of a cyclase- Erwinia chrysanthemi B374: organisation and gene products. ,-galactosidase fusion protein and on cultures overexpress- Mol. Gen. Genet. 201:38-42. ing the unfused protein. In both cases we failed to detect 11. Kataoka, T., D. Broek, and M. Wigler. 1985. DNA sequence and cGMP. characterization of the S. cerevisiae gene encoding adenylate The polypeptide sequence comparison indicated signifi- cyclase. Cell 43:493-505. cant identities between R. meliloti and S. cerevisiae catalytic 12. Kiely, B., and F. O'Gara. 1983. Cyclic 3'5'-adenosine mono- domains. This strongly suggests that the corresponding phosphate synthesis in Rhizobium: identification of a cloned domain in bovine adenylate cyclase (14) is the catalytic sequence from Rhizobium meliloti coding for adenyl cyclase. domain of the Since two are Mol. Gen. Genet. 192:230-234. protein. homologous domains 13. Koesling, D., J. Herz, H. Gausepohl, F. Niroomand, K.-D. prominent in the latter, one may wonder whether the en- Hinsch, A. Mulsch, E. Bohme, G. Schultz, and R. Frank. 1988. zyme from the higher eucaryote does not comprise two The primary structure of the 70 kDa subunit of bovine soluble catalytic domains. A further consequence of these observa- guanylate cyclase. FEBS Lett. 239:29-34. tions is that they indicate a common descent between 14. Krupinski, J., F. Coussen, H. A. Bakalyar, W.-J. Tang, P. G. eucaryotes and procaryotes. Another possibility, although it Feinstein, K. Orth, C. Slaughter, R. R. Reed, and A. G. Gilman. is unlikely because the overall base composition of the gene 1989. Adenylyl cyclase amino acid sequence: possible channel- is typical of R. meliloti, would be horizontal transfer of the or transporter-like structure. Science 244:1558-1564. gene. Thus cAMP synthesis may have been present at the 15. Kubo, M., and T. Imanaka. 1989. mRNA Secondary structure in of the between the two In an open reading frame reduces translation efficiency in Bacillus origin separation kingdoms. subtilis. J. Bacteriol. 171:4080-4082. addition, similarity with guanylate cyclase proteins from 16. Ladant, D. 1988. Interaction of Bordetella pertussis adenylate higher eucaryotes, still visible in R. meliloti through the cyclase with calmodulin: identification of two separated calm- inhibitory potential of GTP, indicates that a nucleotide- odulin-binding domains. J. Biol. Chem. 263:2612-2618. cyclizing enzyme ofbroad specificity was present very early. 17. Laemmli, U. K. 1970. Cleavage of structural proteins during the In this respect, unequivocal confirmation of the presence of assembly of the head of bacteriophage T4. Nature (London) VOL. 172, 1990 RHIZOBIUM MELILOTI ADENYLATE CYCLASE 2621 227:680-685. 31. Roy, A., C. Haziza, and A. Danchin. 1983. Regulation of 18. Lathigra, R., M. O'Regan, B. Kiely, B. Boesten, and F. O'Gara. adenylate cyclase synthesis in Escherichia coli: nucleotide 1986. Organization of the adenyl cyclase (cya) locus of Rhizo- sequence of the control region. EMBO J. 2:791-797. bium meliloti. Gene 1460:89-96. 32. Sanger, F., S. Nicklens, and A. R. Coulson. 1977. DNA sequenc- 19. Lin, H. C., S.-P. Lei, and G. Wilcox. 1985. An improved DNA ing with chain terminating inhibitors. Proc. Natl. Acad. Sci. sequencing strategy. Anal. Biochem. 147:114-149. USA 74:5463-5467. 20. Lipman, D. J., and W. R. Pearson. 1985. Rapid and sensitive 33. Schwartz, R. M., and M. 0. Dayhoff. 1978. Matrices for detect- protein similarity searches. Science 227:1435-1441. ing distant relationships, p. 353-358. In M. 0. Dayhoff (ed.), 21. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecular Atlas of protein sequence and structure, vol. 5. National Bio- cloning: a laboratory manual. Cold Spring Harbor Laboratory, medical Research Foundation, Washington, D.C. Cold Spring Harbor, N.Y. 34. Shepherd, J. C. W. 1981. Method to determine the reading 22. Masson, P., G. Lenzen, J. M. Jacquemin, and A. Danchin. 1986. frame of a protein from the purine/pyrimidine genome sequence Yeast adenylate cyclase catalytic domain is carboxy terminal. and its possible evolutionary justification. Proc. Natl. Acad. Curr. Genet. 10:343-352. Sci. USA 78:1596-1600. 23. Matsudaira, P. 1987. Sequence freom picomole quantities of 35. Singh, S., D. G. Lowe, D. S. Thorpe, H. Rodriguez, W.-J. proteins electroblotted onto polyvinylidene difluoride mem- Kuang, L. J. Dangott, M. Chinkers, D. V. Goeddel, and D. L. branes. J. Biol. Chem. 262:10035-10038. Garbers. 1988. Membrane guanylate cyclase is a cell-surface 24. Messing, J., and J. Vieira. 1982. A new pair of M13 vectors for receptor with homology to protein kinases. Nature (London) selecting cither DNA strand of double-digest restriction frag- 334:708-712. ments. Gene 19:269-276. 36. Steiner, A. L., R. E. Wehmann, C. W. Parker, and D. M. Kipnis. 25. Miller, J. H. 1972. Experiments in molecular genetics. Cold 1972. Radioimmunoassay for the measurement of cyclic nucle- Spring Harbor Laboratory, Cold Spring Harbor, N.Y. otides. Adv. Cyclic Nucleotide Res. 2:51-61. 26. O'Regan, M., B. Kiely, and F. O'Gara. 1989. Expression of the 37. Thorpe, D. S., and D. L. Garbers. 1989. The membrane form of adenyl cyclase-encoding gene (cya) of Rhizobium meliloti F34: guanylate cyclase: homology with a subunit of the cytoplasmic existence of two cya genes? Gene 83:243-249. form of the enzyme. J. Biol. Chem. 264:6545-6549. 27. Pardee, A. B., F. Jacob, and J. Monod. 1959. The genetic control 38. Towbin, U., T. Staehelin, and J. Gordon. 1979. Electrophoretic and cytoplasmic expression of inducibility in the synthesis of transfer of proteins from polyacrylamide gels to nitrocellulose 3-galactosidase of Escherichia coli. J. Mol. Biol. 1:165-168. sheets: procedure and some applications. Proc. Natl. Acad. Sci. 28. Roy, A., and A. Danchin. 1981. Restriction map of the cya USA 76:4350-4354. region of the Escherichia- coli K12 chromosome. Biochimie 39. Ulimann, A. 1984. One-step purification of hybrid proteins 63:719-722. which have ,B-galactosidase activity. Gene 29:27-31. 29. Roy, A., and A. Danchin. 1982. The cya locus ofEscherichia coli 40. Upchurch, R. G., and G. H. Eklan. 1978. The role of ammonia, K12: organization and gene products. Mol. Gen. Genet. 188: L-glutamate, and cyclic adenosine 3',5'-monophosphate in the 465-471. regulation of ammonia assimilation in Rhizobium japonicum. 30. Roy, A., A. Danchin, E. Joseph, and A. Ullmann. 1983. Two Biochim. Biophys. Acta 538:244-248. functional domains in adenylate cyclase of Escherichia coli. J. 41. Young, R. A., and R. W. Davis. 1983. Efficient isolation of genes Mol. Biol. 165:197-202. by using antibody probes. Proc. Natl. Acad. Sci. 80:1194-1198.