Analysis of Sequence Elements Important for Expression and Regulationof the Adenylate Cyclase Gene ( cya ) of Salmonella typhimurium ABSTRACT THE nucleotide adenosine-3 ' ,5 ' - cyclic-monophosphate ( CAMP ) has long been known to be the central controlling molecule in the regulation of the large family of genes involved in theuptakeand utilization of carbon sources in enteribcacteria ( GOTTESMAN 1984 ) .
Cyclic AMP binds to the cAMP receptorprotein ( CRP ) inducingaconformational change in the protein thatallows it to bind specifically to DNA sequences proximal to sensitive promoters .
T h e best understood consequence of this binding is stimulation of the transcription of genes necessary for utilization of carbon sourcesother thanglucose .
While the physiological consequences of the increases and decreases in intracellular cAMP concentration in response to quantityand quality of the available carbon sources are well documented , the factors responsible for regulation of CAMP levels are incompletely understoo ( dreviewed in MAGASANIKand NEIDHARDT 1987 ) .
Study of the regulation of cya , which encodes adenylate cyclase ( the enzyme that synthesizes CAMP ) , provides an opportunity to examinethe regulation of a gene central to aglobal regulatory response .
T h e literaturecontains many conflicting reports regarding the regulation of activity of adenylate cyclase , the regulation of cya expression , and interpretation of the cya nucleotidesequence.Although it We determined the nucleotide sequence of the regulatory region of the cya gene of Salmonella typhimurium .
A setofnestedBAL-31deletions originatingupstream of the promoter/regulatory region and extending into the cya structural gene was constructed in M13mp : : cya phages and was tested for complementationof a chromosomalcya deletion mutationB .
AL-31deletion mutants unable to complement cya localized the major cya promoter .
The synthetic tac promoter was inserted upstream of the BAL-31 deletions so that expression of cya was dependent on transcription from tac .
Those tac derivative phages unableto complement cya localized the translation initiation region .
The cya DNA sequence revealed at least three potential promoters capableof transcribing cya , with a CRP binding site straddling the -10 hexamer of the promoter proximal to the structural gene .
The leader RNA sequence initiatedat the latter promoter is approximately 140 bases long and includes a region that may form a stable secondarystructure ( AG = -23.8 kcal ) .
There exist two possible in-frame translation start points , one of which is T T G and the other of which is ATG .
The sequence of the S. typhimurium regulatory regionwas compared with that reported for Escherichia coli .
' Present address : Divisionof Biochemistry and Molecular Biology , Department of Molecular and Cell Biology , University of California , Berkeley , California 94720 .
To wholn correspondence should be addressed .
appears that theenzymes of the bacterial phosphoen-o1pyruvate : sugar phosphotransferase system ( PTS ) may be involved in regulating adenylatecyclase activity , the mechanism is unclear ( DOBROGOSetZal .
1983 ; DANIEL1984 ; BLAAUWEaNnd POSTMA1985 ; REDDY et al. 1985 ) .
Mutants of cya with as much as 60 % of the carboxy-terminal end of adenylate cyclase deleted retain theability to synthesize cAMP ( LEIBand GERLT 1983 ; WANGC , LEGGand KOSHLAN1D981 ) .
T h e function of the C-terminal domain of adenylate cyclase is not known and contradictory resultshavebeenobtained concerning its role in regulating cAMP levels ( ROYet al. 1983 ; KOOP , HARTLEYandBOURGEOIS 1984 ) .
SeveralexperimentsdemonstratedCRP-de-pendent repression of cya expression by cAMP ( AIBA 1985 ; JOVANOVICH 1985 ; KAWAMUKAIet al. 1985 ; FANDLT , HORNEanRd ARTZ1990 ) , butthe results disagreed with those of other experiments ( BANKAITIS and BASSFORD1982 ; ROYH , AZIZAand DANCHIN 1983 ) .
There also exists some uncertainty regarding the number and location of cya specific promoters , the number and location of CRP binding sites , and the location of the translationstartpoint ( AIBA , KAWAMUKAI and ISHIHAMA1983 ; AIBAet al. 1984 ; DAN-CHIN et al. 1984 ; AIBA1985 ) .
In this paper we report the nucleotide sequence of the cya regulatory region of Salmonella typhimurium .
We report the identificationof three candidate cya promoters and the start of translation of adenylate cyclase based ondeletion analysis of M13mp : : cya phages.In the accompanyingpaper we study the regulation of cya and confirm the physiological signif icance of the transcription and translation sequences we identify here , and prove directly theCRP-depend-ent repressionof cya by CAMP .
Bacterial strains , plasmids and bacteriophages : S. typhi-murium strain AZ3409 was constructed asdescribed in Table 1 .
E. coli strain JM 103 has been described previously ( MESSING 1983 ) .
Plasmid pDR540 was purchased from P-L Biochemicals , Inc. , Milwaukee , Wisconsin .
Plasmid pCKlO2 ( WANG , CLEGGand KOSHLAND1981 ) was a gift of D. E. KOSHLANaDnd the 4.2 kb SmaI-EcoRI DNA fragment encoding most of the gene ofS .
typhimurium ( Figure 1 ) was cya subcloned from this plasmid intoM13mp8tomakethe phageMI 3mp : : cya .
PhagesM13mp8 and M13mp9 have been described ( MESSINGand VIEIRA 1982T ) h. e nucleotide sequences of both strands of the cya regulatory region and first part of the cya structural gene were determined with a series of phage M 13mp8 : : cya BAL-3 1 deletionmutants and with phage M 13mp9 : : cyaA33R .
Phage M 13mp9 : : cyaA33R was made by subcloning a 507-bp PstI-AvaII fragment of BAL-31 deletionphage M13mp8 : : cyaA33 in the reverse orientation .
The AvaII site used in the subcloning is at position 228 with respect to the startpoint of transcription of the major cya promoter ( P2 ) and thePstI site is from the multiple cloning sequenceof phage M13mp8 : : cyaA33 .
The AvaII end was filled in with DNApolymerase I ( Klenow fragment ) and the fragmentwas ligated into M13mp9 rep-licative form ( RF ) DNA that had been cut with PstI and HincII .
Media , chemicals and enzymes : T h e 2X Y T and C-minimal-medium have beendescribedelsewhere ( ALPER and AMES1978 ; MILLER1972 ) .
N ' , N ' - Methylene bisacryl-amide , acrylamide andammonium persulfatewere from Bio-Rad Laboratories , Richmond , California.Restriction enzymes were from New England Biolabs , Beverly , Massa-chusetts .
Agarose , nuclease BAL-31 and DNA polymerase I ( Klenow fragment ) were from Bethesda Research Laboratories , Gaithersburg , Maryland .
Low gelling temperature agarose , type VI1 was from Sigma Chemical Co. , St. Louis , Missouri .
The phage M 1si3ngle strand `` universal '' primer was from P-L Biochemicals Inc. , Milwaukee , Wisconsin and was used to obtain the cya DNA sequence of the nontemplate strand with M13mp : : cya BAL-31 deletion mutants , and partof the tion phages and their tac derivatives .
Recombination was prevented by the presence of the recA mutation in strain AZ3409 .
The recipient was grown overnight in 2 ml2X Y T at 37 '' , pelleted , resuspended in 2 ml sterile 0.85 % NaCI , and 0.1 ml was spread on C-citrate ( 0.4 % W/V ) agaprlates supplemented with 2 mM proline .
The phage lysates ( 2 pl of a 1:10 dilution ) were spotted on the plates which were then incubated for72 h at37 '' .
Phages M13mp8 : : cya and M13mp8 were included as positive and negative controls , respectively .
Lysates were spotted onto plates without recipient cells to test for contamination .
Positive complementation was scored as confluentgrowth in the spot ; lack of complementation gave no growth in the spot .
The M 13mp donorphage lysates were grown in 2 ml2X Y T at 37 '' as described by MESING ( 1983 ) .
Lysates were cleared of bacteria by centrifugingthem for 5 minin a microfuge ; remaining bacteria were killed by incubating the lysates at 65 '' for 20 min .
This heating had no significant effect on phage viability and lysates were approximately 10 '' plaque forming units per mI .
RESULTS Localization of the cya promoter region : T o facilitate localization of the cya promoter region we constructed a set of nested BAL-31-generated deletions extending into cya from the 5 ' end of the gene ( Figures 1 and 3 ) .
The sizesof the deletions were estimated by agarose gel electrophoresis of the recombinant phageDNAs and precise deletion endpointswere determined by DNA sequencing .
Phage lysates were tested for their ability to complement the extended cya deletion in strain AZ3409 ( Table1 ) .
Those phages carrying BAL-31 deletions extending into sequences required for expression of cya should fail to complement strain AZ3409to growth on citrate .
The results of these experimentsare summarized in Figure 3 and show that the ability to complement the cya mutation in AZ3409 was lost between the endpointsof deletions-32 and 62 .
Therefore , a sequence required for expressionof cya is located in the39-bpregiondefined by thesedeletionendpoints.Insertion of a95-bp fragmentcontaining the tac promoter upstream of the BAL-31 deletion ( Figure 2 ) restored the ability of phage M13mp8 : : cyaA62 to complemens t train AZ3409 ( Figure 3 ) .
This indicates that the deletionin phage M13mp8 : : cyaA62 removes a cya promoter .
Localization of the cya translation initiation region : T o localize the start of translation of the cya structural gene , we extended the approachdescribed above used to locate the promoter region .
BAL-31 deletion phages containing the tac promoter fragment were tested fortheir ability to complementstrain AZ3409 to growthoncitrate.Starting with A62 , expression of cya was dependent on transcription from the tac promoter but translation was dependent on cya sequences since the tac fragment lacksa translation initiation sequence .
Figure 3 shows that several of the deletion phages that had been unable to complement cya in the absence of tac were able to do so when transcription originated from the tac promoter , indicating that cya sequences required for translation initiation remained intact.T h e 75-bp region defined by the endpoints of deletions 125 and 36 must contain sequences requiredfor initiation of cya translation since the ability to complement cya was lost between these endpoints Nucleotide sequence of the cya regulatory region and the beginningof the cya structural gene : Based on the results of the complementation analysis with BAL-31 deletions of phage M13mp8 : : cya , we were able to locate that part of the DNA sequence importantfor cya expression andgeneregulation in S. typhimurium .
The DNA sequence of thepromoter/regulatory region and the first part of the cya structural gene is shown in Figure 4 .
Within the 39-bp region defined by the endpoints ofBAL-31 deletions-32 and62 , and indicated by complementation analysis to contain a cya promoter ( Figure3 ) occursasequence5 ' - TTTACG-17 bp-TAAATT-3 ' showing homology to the Ea '' consen-suspromoter5 ' - TTGACA-17bp-TATAAT-3 ' ( HAWLEYandMCCLURE1983 ) .
We designated this sequence the P2 promoter ( Figure 4 ) by analogy to AIBA ( 1985 ) who observed the identical sequence for Escherichia coli cya immediately upstream of a major transcriptionstartpoint .
( The numbering of the S. typhimurium cya DNA sequence corresponds to the E. coli cya sequence with the P2 startpointdefined as + -1 .
) Overlapping the 10 hexamer sequence ( `` Pribnow box '' ) of cya P2 is the sequence 5 '' AG-TGTTA - '' `` - TCACG-TT3 ' which is closely homologous to the consensus CRP binding site 5 '' AA-TGTGA -- -- TCACA-TT-3 ' ( EBRIGHTet al. 1987 ; BERGand VONHIPPEL1988 ) .
In addition , threeother Ea '' promoter-like sequences , designated P1 , P1 ' and Px , were identified upstream of the P2 promoter .
The PI and PI ' promoter sequences are oriented toward cya DISCUSSION CCGTTAAGAATTTTGTAACA C CTC TAT ATT GAG ACT CTG AAA CAG AGA CTG GAT GCACTA AAT CAA CTG CGT G GAG ATA TAA CTC TGA GAC TTT GTC TCT GAC CTA CTGAGT TTA GTT GAC GCA r Leu Tyr Ile Glu Thr Leu Lys Gln Arg Leu AspAla Ile Asn Gln LeuA r g GTG GAT CGC GCG CTT GCT GCC ATGGGA CC CAC CTA GCG CGCGAA CGA CGG TAC CCTGG Val Asp Arg Ala Leu Ala Ala Met Gly region .
We localized a cya promoter to a39-bp region and the starotf cya translation to a 75-bp regiownithin the initial 4.2-kb cloned fragment .
The same set of BAL-31-generated deletion phages allowed us to determine the nucleotide sequenceof the cya regulatory region.Inthe accompanying paper , we use allele replacementtechniquesdevelopedfor the M13mp vector system ( ARTZet al. 1983 ; BLUMet al. 1989 ) to study the effects of regulatory mutations constructed in-vitro .
Identification of multiple cya promoters : The nu ¬ C-G +73 + eo c-6 * ' * +15 Q +14 ?
A. G. C. A. A. C. ~ .
`` Secondary structure of the cya leader RNA .
Numbers above theRNA sequence correspond to the numberingof the cya DNA sequence i n Figure 4 .
The AUGandUUG initiation codonsareboxedandthesequenceshomologoustothe Shine-Dalgarnosequencefor eachinitiation codonaredesignated by double lines .
The AG of the secondary structure was calculated to be -23.8 kcal ( CECH1983 ) .
the sequences around the transcription startpoint of the P 1 promoter and saingle base-pair substitution in the 17-bp spacer between the -10 and -35 hexamer sequences of the P2 promoter .
The Px promoter is less homologous with differences aroundthetranscription startpoint , in the spacer between the -10 and -35 hexamer sequences , and a 1-bp substitution at position -173 which is a highly conserved position for Ea7 ' promoters in the-35hexamer sequence .
This latter difference would correspond to a strong promoter-downmutation in the S. typhimurium Px promoter ( HAWLEY anMdCCLURE1983 ) .
The striking conservation of the proposedpromoter sequences in s. typhimurium and E. coli provides evidence fortheirfunctionalimportance ( with the possible exception of the divergent Px promoter ) .
The sequences between the P2 promoter and the upstream P1 and P 1 ` promoters , andbetween P2 and the start of translation , are poorly conserved as would be expected for sequences of less important function .
The existence of multiple promoters implies that transcriptional regulation of cya may be complex as might be expected for a gene that , itself , is involved in global regulation .
We did notfind a S. typhimurium promoter sequence corresponding tothe E. coli `` upstreampromoter '' proposed by ROYet al. ( 1983 ) ( positions -162 to-131 in Figure 6 ) .
Because of the very poor homology between the S. typhimurium and E. coli sequences in this region and because AIBA ( 1 985 ) failed to find the transcription startpoint , the functional significance of this `` upstream promoter '' in E. coli remains problematic .
Potential cya promoters were identified in E. coli by inspection of the DNA sequence ( ROY , HAZIZAand DANCHIN 198a3n ) d by S1-nuclease-mapping to determine transcription startpoints ( AIBA1985 ) .
We identified cya promoters of S. typhimurium by deletion mutant analysis ( Figure3 ) ( FANDL , THORNEanRd ARTZ 1990 ) andby DNA sequence comparison with the E. coli cya sequence ( Figure 6 ) .
Identification of a CRP binding site : Analysis of the nucleotide sequence revealed one potential CRP binding site which straddles the -10 hexamer of the P2promoter ( Figure 4 ) .
The location of this CRP binding site suggests a role for the CAMP-CRPcom-plex as a repressor of cya expression .
This site corresponds to one of the CRP binding sites proposed by .
C. G. U. C ~ ~ ~ .
~ AIBA , KAWAMUKAaI nd ISHIHAMA ( 1983 ) for E. coli cya but the S. typhimurium sequence does not reveal a second CRP bindingsite proposed forE .
coli cya ( AIBA , KAWAMUKAaInd ISHIHAMA1983 ; ROY , HAZIZA and DANCHIN 1983T ) .
he location of the proposed second site ( centered around position +115 in Figure 6 ) is in a region of poor homology between the nucleotide sequences of S. typhimurium and E. coli suggesting lack of a functional importanceof the sequence .
T h e contacts made between Eo7 ' RNA polymerase and DNA , and between CAMP-CRPand DNA , have been analyzed in detail ( SIEBENLISTS , IMPSONand GILBERT1980 ; SHANBLATaTnd REVZIN 1986 ) .
This information can be used to infer how RNA polymer-ase and CAMP-CRPmay interact at cya P2 .
Figure 7 ( top ) shows the positions of ethylated phosphates in the cya P2sequence that are predicted to interfere with the binding of RNA polymerase and CAMP-CRP .
Figure 7 ( bottom ) shows aplanarrepresentation of the contactsprojectedon the surface of the DNA .
According to this analysis , CAMP-CRPwould bind to the same surface of the DNA as RNA polymerase at cya P2 and interfere directly with the accessof the polymerase to the-10 hexamer and theregion of the cya P2 promoter predicted to be melted in forming the open complex .
Since it has beenproposedthat activation of transcription by CAMP-CRPinvolves di-rectprotein-proteincontact with RNA polymerase ( SIEBENLISTSI , MPSON and GILBERT1980 ; IRWIN and PTASHNE1987 ) , it may be speculated that repression of transcription by CAMP-CRPinvolves protein-pro-tein contact as well .
Evidence that lac repressor forms a ternary complex with RNA polymerase and DNA at thelac operon promoterhas recently been obtained ( STRANLEaYnd CROTHERS1987 ) .
A mechanism in which RNA polymerase remains poised at cya P2 under repressing conditions would be consistent with interpretation of the physiological analysis of cya regulation ( FANDLT , HORNaEnRd ARTZ 1990 ) .
We have constructed mutations that alter the CRP binding site which overlaps the P2 promoter .
These mutations resultin the loss of repression of cya expression by CAMP-CRP , thus confirming the importance of this CRPbinding site in the regulation of cya expression ( FANDLT , HORNaEnRd ARTZ 1990 ) Initiation of cyu translation : AIBA , KAWAMUKAI and ISHIHAMA ( 1983 ) proposed that the E. coli cya structural gene startswith the ATG atposition +225 ( using the numbering in Figures 4 and 6 ) .
ROY et al. 1983 ) proposed an unusual T T G initiation codon at position +147 andthis was subsequently supportedby N-terminal sequence analysis of a cya-lac hybrid protein ( DANCHIN et al. 1984 ) .
Our results with S. typhi-murium supportthe conclusion that cya translation initiates at the TTG.Deletion analysis with tac deriv-ativephages clearly places the translation initiation site forS .
typhimurium cya between positions +120 and +195 ( Figure 3 ) .
This region , which is perfectly homologous to the correspondingE .
coli sequence , does not contain a potentiaA l TG initiation codon followed by an open reading fram ( eFigures 4 and 6 ) .
The only likely initiatorcodon associated with a good Shine ¬ Dalgarno sequence is the TTG atposition +147 ( Figure 4 ) .
AIBAet al. ( 1984 ) also identified an additional translation reading framein E. coli beginning with the ATG atposition +88 ( Figure 6 ) and extending for 30 amino-acids .
This peptide is unlikely to be of important function since the homology between E. coli and S. typhimurium is poor and thSe .
typhimurium sequence has an in-frame TGA stop codon at position +106 .
Although rare , T T G initiator codons occuirn other wild-type genes or have been made by mutation ( Ko-ZAK 1983 ) .
REDDY , PETERKOFSKY and MCKENNEY ( 1985 ) changedthe T T G codon at position +147 ( Figure 4 ) of the E. coli cya gene to ATG and found that translation of cya was six times more efficient with theATG codon when transcriptioncouldbe repressed fromthe XPLpromoter .
When transcription initiated atthe cya promoters , however , theAT codon was lethal in a multicopy plasmid suggesting that the T T G codon may provide a mechanism for limiting translation of the wild-type gene .
An alternative possibility is that the T T G initiator codon could have been selected as the result of a down mutation during the initial cloning of E. coli cya on a multicopy plasmid .
However , this possibility is unlikely since the T T G initiator codon has been found in three independent , multicopy cya clones , two from E. coli ( ROY , HAZIZA anDdANCHIN1983 ; AIBAet al. 1984 ) and one from S. typhimurium ( this work ) and selection of the same down mutation three times would be highly fortuitous .
Upstream of the T T G initiatorcodon in the S. typhimurium cya sequence is an in-frame ATG codon at position +117 with an associated Shine-Dalgarno sequence that , if functional as a ribosome-binding site , would add 10 amino-acids to the N terminus of the protein ( Figure 4 ) .
The corresponding ATG in the E. coli sequence could notbe utilized toinitiate cya translation because of the one bpdeletion at position +13 1 which changes the reading frame ( Figure 6 ) .
+ The potential ATG startpoint atposition 1 17is not required for translation of S. typhimurium cya mRNA as determinedby complementation analysis ( Figure 3 ) and it is interesting to speculate that function of the ATGstartpoint could bedeleterious to adenylate cyclaseactivity .
If this is true , then thReNA secondary structure which includes the ATG codon and its associated Shine-Dalgarno sequence ( Figure 5 ) may function to prevent use of this ribosome-binding site in S. typhimurium .
The importance of the possible regulatory role of the RNAsecondary structure in proper initiation of translation of S. typhimurium cya might account fotrhe divergence of the S. typhimurium and E. coli sequences in the region between positions +90 and +1 17of the S. typhimurium sequence ( Figure menu RNAinterveningsequence : structuralhomology with fungal mitochondrial intervening sequences .
DANCHINA , N , , G. urso , A. ROY and A. ULLMANN1 ,984 Identification of thEescherichia coli cya gene product as authen tic adenylate cyclase .
DANIEL , J. , 1984 Enzyme 111 stimulation of cyclic AMP synthesis in an Escherichia coli crp mutant .
, G. W. HALL , D. K. SHERBAD .
G. HARMANand T. MELTON1 ,98 R3 egulatoriynteractions among the cya , crp and pts gene products in Salmonella typhi-murium .
EBRIGHT , R. H. , A. KOLB , H. Buc , T. H. KUNKEL , JS .
KRAKOW and J. BECKWITH1,987Roleofglutamicacid-181inDNA-sequence recognition by the catabolic gene activator protein ( CAP ) of Escherichia coli : Altered DNA-sequence-recognition properties of [ Val '' ' ] CAP and [ Leu `` I ] CAP .
FANDL , J.P. , L. K. THORNaEnRd S. W.ARTZ ,1990 Mutations that affect transcription ancdyclic AMP-CRP regulation of the adenylatecyclase gene ( cya ) ofSalmonella typhimuriumG .
, 1984 Bacterial regulation : global regulatory networks .
, 1983 Studies on transformation of Escherichia coli with plasmids.J .
, and W. R. MCCLURE , 1983 Compilation andanalysis of Escherichia coli promoterDNAsequences.Nucleic Acids Res .
IRWIN , N. , and M. PTASHNE ,1987 Mutantsofthecatabolite activator proteinofEscherichiacoli that arespecifically deficient in the gene-activation function .
JOVANOVICH , S. B. , 1985Regulationofa cya-lac fusion by cyclic AMP in Salmonella typhimurium .
KAWAMUKAI , M. , J. KISHIMOTO , R. UTSUMI , M. HIMENO , T. KOMGNOand H. AIBA , 1985 Negative regulation of adenylate cyclase gene ( cya ) expression by cyclic AMP-cyclic-AMP-receptor-protein in Escherichia coli : studies with cya-lac protein and operon fusion plasmids.J .
KOOP , A. H. , M. HARTLEYandS .
BOURGEOIS1 ,984 Analysis of the cya locus of Escherichia coli .
KOZAK , M. , 1983 Comparison of initiation of protein synthesis in procaryotes , eucaryotes and organelles .
LEIB , T. K. , and J. A. GERLT , 1983 Evidence for a small catalytic domain in the adenylatecyclase from Salmonella typhimurium .