6122
Biochemistry 1989, 28, 6 122-61 27
Predicted Structures of the cGMP Binding Domains of the cGMP-Dependent Protein Kinase: A Key Alanine/Threonine Difference in Evolutionary Divergence of cAMP and cGMP Binding Sitest Irene T. Weber,*,$ John B. Shabb,§ and Jackie D. Corbins Crystallography Laboratory, NCI-Frederick Cancer Research Facility, BRI-Basic Research Program, P.O. Box B, Frederick, Maryland 21 701, and Howard Hughes Medical Institute and Department of Molecular Physiology and Biophysics, Vanderbilt University, Nashville, Tennessee 37232 Received January 20, 1989: Revised Manuscript Received April 3, 1989
Mammalian cGMP- and CAMP-dependent protein kinases show considerable similarity in amino acid sequence, although they specifically bind different cyclic nucleotides. Results of cGMP analogue binding experiments, combined with modeling of the cGMP binding sites by analogy to the structure of the homologous catabolite gene activator protein, suggest that a threonine residue forms a hydrogen bond with the 2-NH2 of cGMP. This threonine is invariant in all c G M P binding domains, but the corresponding residue in 23 out of 24 cAMP binding sites of protein kinases is alanine, which cannot form the same hydrogen bond. This alanine/threonine difference has the potential for discriminating between cAMP and cGMP and may be important in the evolutionary divergence of cyclic nucleotide binding sites. ABSTRACT:
x e two prominent cyclic nucleotides in eucaryotic cells, cGMP and CAMP, are believed to be second messengers for different physiological functions. Although other receptors may exist, many effects of cGMP and cAMP are mediated by the cGMP- and CAMP-dependent protein kinases (cAPK and cGPK), which presumably recognize different cellular protein substrates (Beebe & Corbin, 1986). The cGPK is composed of two identical subunits of 76 33 1 daltons, each of which contains a regulatory and a catalytic component (Lincoln & Corbin, 1983), and binds two molecules of cGMP (Corbin & Doskeland, 1983; MacKenzie, 1982). The cAPK also binds two molecules of cyclic nucleotide on each of two regulatory subunits (Beebe & Corbin, 1986). For both cyclic nucleotide dependent protein kinases, the two sites differ in cyclic nucleotide dissociation rate and analogue specificity (Rannels & Corbin, 1980; Corbin et al., 1986). Other than possible differences in conformation or electron distribution, there are two structural features that distinguish cAMP and cGMP. The cAMP molecule has an NH2 group at the 6-position and a hydrogen at the 2-position of the purine ring, while cGMP has a keto group at the 6-position and an NH2 group at the 2-position. The difference in relative specificity of cAMP and cGMP for the cAPK and cGPK approaches a factor of 100 (Lincoln & Corbin, 1983), despite the similarities in amino acid sequence of the two protein kinases. It might therefore be expected that differences in the cyclic nucleotide specificities of the respective kinases would be due to different amino acid contacts at the 2- or 6-position of the purine rings. Thus, a search has been carried out to find a structural difference that could discriminate between cAMP and cGMP binding. Each of the cyclic nucleotide binding sites of cGPK and cAPK shares amino acid sequence homology with the cAMP binding domain of the bacterial catabolite gene activator protein (CAP) (Weber et al., 1982; Takio et al., 1984b). CAP This research was sponsored by the National Cancer Institute, DHHS, under Contract N01-74101.
* NCI-Frederick 5 Vanderbilt
Cancer Research Facility. University.
senses the level of cAMP and regulates transcription from several operons in Escherichia coli (Zubay et al., 1970; Anderson et al., 1972), including lactose, galactose, and ara C [for reviews, see decrombrugghe et al. (1984) and deCrombrugghe and Pastan (1978)l. The amino-terminal domain of CAP binds CAMP, while the smaller carboxy-terminal domain forms the DNA binding site (Weber & Steitz, 1984, 1987). The crystal structure of the CAP dimer with two bound molecules of cAMP has been determined (McKay & Steitz, 1981; McKay et al., 1982), which permits modeling of cyclic nucleotide binding sites, as was done previously for cAPK (Weber et al., 1987). The models of the cAMP binding sites have been useful in studies of the cAPK, such as for predictions of specific amino acid residues that could be modified by site-specific mutagenesis (Bubis et al., 1988). It is expected that the models of the cGMP binding sites of cGPK will also be useful for these purposes. EXPERIMENTAL PROCEDURES The amino acid sequence of bovine cGPK was aligned with the sequences of CAP and the cAMP binding domains of bovine RI and RII, as shown in the structural alignment of Weber et al. (1987). In addition, amino acid sequences of 11 other R subunits and 1 other cGPK, that have recently been deduced by using molecular cloning techniques, have been aligned maintaining essentially the same gap locations. The starting coordinates for modeling the cGPK domains were taken from the RI model structure that had been built by using the CAP coordinates from the crystal structure refined to 2.5-A resolution (Weber & Steitz, 1987). The RI model structure was taken as a starting point since this protein is more similar to cGPK than is RII in amino acid sequence and kinetic properties (Doskeland et al., 1987). The deletions and insertions relative to the CAP structure had previously been determined (Weber et al., 1987). The amino acid side chains of the cGMP binding domains of cGPK were substituted for those of RI in the positions where the sequences differed. The positions of the amino acid side chains were examined with the PS300 computer graphics system using the program FRODO (Jones, 1978). Occasionally, a small adjustment in the position
0006-2960189 10428-6122$01.50/0 0 1989 American Chemical Society
Protein Kinases: Divergence of cAMP and cGMP Binding Sites CAMP BINDING DOMAIN
CAP
1
CAMP BINDING DOMAIN A cAPK
DNA BlNDlNQ DOMAIN
cAMP BINDING DOMAIN B
I
I
t
C O U P BINDING DOMAIN A cQPK
i
CATALYTIC SUBUNIT
REGULATORY SURUNll
cGMP BINDING DOMAIN E
---_
C
REGULATORY COMPONENT
CATALYTIC D O M A I N
FIGURE 1 : Summary of the domain structures of CAP, cAPK, and cGPK. Each polypeptide chain is indicated as a linear structure from the amino- to the carboxy-terminal direction. Homologous cyclic nucleotide binding domains are indicated by hatching; the catalytic portion is stippled.
of the amino acid side chain was required in order to avoid collisions with neighboring atoms. Cyclic GMP was substituted for CAMP, and both syn and anti conformations were examined in the two cGMP binding sites. RESULTSA N D DISCUSSION Conserved Domain Structure. General structural features of CAP, cAPK, and cGPK are summarized in Figure 1. All of the published amino acid sequences of the cyclic nucleotide
'LEW LSHCM I N K V D S K S ' L I U ~ K A E T L V VI V Y k V A V L I K D E
CAP
I?
A COWIa
2?
1 4 3 EDS31C3AUcPVSF
3c
43
50
~ A G E n V l O ~ E G ~ ~ V I 3 0 ~ U 3 V V V h
IAGE~VICOG3EGD'~FYVIDCE~~3VVVk E Q S 3 1 ~ 3 I U c S V S Fn A G E ' V l C W 3 E G 5 ' , F Y V I 3 C Q ' T 3 V V V N U3cR l a 1 4 5 E Q S 3 1 c 3 A U c P V S F I A G E l V l C O G 3 E G D ' , F Y V I D C O f U 3 V V V N R A P R l a I45 E R S 3 1 C D A U c P V S F I A G E l V I C P 0 3 E G D ' , F Y V I D C 6 ~ U 3 V V V N UOURIP I45 E Q S D l c D A U : P V T H I G G E ' V l C W ' ~ E G D ' . F Y V I D C O : V I ) V V V N C O W 1 la 1 4 3 O L S O V L D A U C E R T V K V D E ~ V I C O G 3 D G D ' ~ F Y V I E ^ G ' ~ 3 1 ! V T K 3
A OlGRla
144 EQS3IC3AU'PVSF
A "UURICI
'45
Biochemistry, Vol. 28, No. 14, 1989 6123
binding domains of CAP, bovine cGPK, 13 different regulatory subunits (RI and RII), and Drosophila cGPK are compared in Figure 2. The alignment of these 30 cyclic nucleotide binding domains shows a pattern of conserved residues, which can be analyzed with respect to the structure of the cAMP binding domain of CAP. There are only four positions where there are insertions or deletions relative to the CAP sequence, and these lie between elements of secondary structure. A region of variable length is observed between 04 and 05, single amino acid deletions occur at the positions corresponding to CAP 70 in all A domains and at CAP 79 in both A and B domains. Single amino acids are inserted in the Drosophila cGPK A and B domains between CAP 88 and 89. There are conserved differences between the A and B domains. For example, Tyr-77 is invariant in all A domains, but no B domain contains Tyr at residue 77; residue 70 which is phenylalanine in the B domains is deleted in all A domains; and residue 95 is tryptophan in the A domains and valine or leucine in the B domains. The region corresponding to the helices aB and crC differs in the A and B domains. Glycines-45 and -71 are invariant in all of the sequences, and Gly-33 is conserved in 29 out of 30 of the domains. These glycines also are conserved in two other bacterial proteins, the
EGKEMl LSVLhOS3F 'mLG(:EEGOE#Sw9A
6@
-*
e?
KTACEVAE ISYKKFROLIOVNPD ILURLSAOMARRLOVTSEKVGNL
9@
100
'IO
120
130
Y E W 4 T S V G E G S S F e L l l Y G 7 m V Y A K T N V K L W G I D R D S Y R ~ I L U G S LTR K R K U Y E E F L S K V S I L E S L D K W 2 6 0 YEW4'SVGEGSSF
M L I I L I V G 'PRAATVKA
K T N V K L ~ N D R D S V R ~ I L Y G LSRTK R K M V E E F L S K V S I L E S L D K W 2 6 1
2 COWIn
&€ L A L l V G ' P R A A l V I ( A K T N V K L V r l D R D S V F R I L U G S TL R K R K U V E E F L S K V S I L E S L D K W 2 6 2 Y E W 4 ' S V S E G S S F OL L I I L I V G ' P R A A T V Y A Y T N V K L ~ l D R D S Y F l I L U G S T L R K R K M Y E E F L S K V SI L E S L D K W 2 6 2 Y E W I Y S V G E G G S F QE L A L l Y G ' P R I I A T V K A Y T N V K L W G I D R D S V F R I L U G S T L R K R K M Y E E F L S K V S I L E S L D K W 2 6 2 G E W T k I S E G S S F O E L A L I Y G 'PRAA'VYA KTDLKLWGIDRDSYFAILUGSTLRKRKMYEEFLSKVSILESLEKW 262 NOT R S V G 3 V 3 4 H G S F E€ L ALWY 'P R A A T I VA T S E G S LWGLDRVT F R I I V K N Y A K K R K U F E S F IE S V P L L K S L E VS 2 6 4 OLSOVLDAUcERTVKVDEMVI?Q D O G ~ . F Y V I E F G ' W D I L V T K D N O T R S V G O V 3 h H S S F DL L A L U V Y 'PR O : S O V L O A U ~ ~ R ~ V K T D E ~ V l ~ ~ D D G D ~ ~ F Y V I E ~ ~ ~ V DN OI LT RVS~V ~G PD A N Q G S F M L I I L U V Y ' P R A A l I V A T S Q G S L W G L D R V T F R R I I V K N Y A K K R K U F E S F I E S V P L F K S L E M S 2 3 4 0.SOVLDAM~EY'\'K'DEMVIP~3DGD'~FYVIEFQ'VDILVTY3 N O T R S V G 3 v 5 h ~ S S F Q E L A L V V Y ' P R A A I I IA T S E G S L W G L D R V T F P R I I V K N Y A K K R K W E S F I E S V P L F K S L E M S 2 6 5 QVS3VLDAU~EKLUKEGEMVI~W3DGD'~FYVID~O'FDIVVKCI) GVGRCVGhv3hQSSf OELALUVY ' P R A A T I T A TSPGALWGLDRVTFF(I1lVKNYAKKRKUVESFlESLPFLKSLEVS 280 E P h V V ~ L A U \ ' E \ ' L V K A G 3 II I K P 0 3 E G D F Y V I D S G C D l V V C O k GSSP'LVUEVCEGSSF D E L A L I V G SPRAATVIA RTQVRLWALNGATVFRlLUDOT IKKRKLYEEFLEKVS I L R H I D K Y I96 S r R L V NCLEEKSVDKGA'I I K W 3 C G D . F Y V V E ~ O ' V 3 ~ V V H DNKVNSSGQGSSF 5 E L A L u V Y SPRAATVVA TSOCLLWALDRLTFF* ILLGSSFKKRLUYDDLLKSYPVLKSLTTY 3 0 9 I',~:MVC.EvCW~SCI Ib G3,C YVMECGV:fV'KE Ci'nK.C'VC95K.F 8E-K L V V C R L ' V Y - L V 4 V K L K 4 I C R Q C F C - I M V 9 ' G L K - ' E v U E c L I ( S V P T F O S L P E E 22' OVQEL\':SWSKS I A A G E C V I F : Q E . G : " . Y V ; 4 : O ' F A V V 3 0 G Y V L 3 K V S A G K A F BE L A L Y Y C ' R - I S l Q V L S E A A R V W V L D R R V F O O I M ~ ~ C T GI LE 0Y ~S V Y C L R S V P L L U N L S E E I40 Ea.-Vr_?LEC,3:E"j3KIV,~IGE~"JII-CF I ILCG.AAV.3aDSE h E E :\ E + S 2 C C 5 Z Y :@E I L UV F mrA- 1' V A 0 C: 3 K C * K L C 9 9 F E I V L C: 3 C S 3 L K R Y I OOY N S VS L S I' 3 9
9
E Q L ' V A 3 A L E P ~ ' O ~ E 3 G O K I V : O C I E F O D ~ F1 L~ El G j A A V . O Q Q S E
A
A A A
A D I G R I la A R 4 ' R I In
88 ''3
A U 3 L R I la ' 4 4 A R4'RI A
S,
91%
A VEASTR A
'59
aUER
C:\',?CY
-4
"2 ' I C
22 25' D I C R l n 262 MUM4 l a 2 6 3 UC)VRla 2 6 3 RA'Rla 263 U9URlP 4 6 3
9 ELVCGK
9 B 9 B 9 C O W I la 2 6 5
YEW4TSVGEGSSF
-
EQL'VA3ALEPVOCE3GOK I V \ 0 G E P G D 7 F C
I I LEGSAAVLORQSE
N E E ~ V E V G ~ L S P S 3 V : Q E I A L L U Y F P R A A T V V A ~ G P L K C V K L D R ~ Q F E R V L G ~ C S D l L K R N l O O Y N S F V S l3S8V0 NEEFVEVGQLGPS3YcGE IALLUY FPRAATVVA
RGDLKCVKLDRPRFElVLGDCSDlLKRNlOOYNSFVSLSV
381
E R L ' V A 3 4 L E P v O F E D G O K l V \ O G E P G D ~ F ~I LI E G - A A V L O R R S E
N E E ~ V E V G Q L G P S 3 V c B E I ~ L L U FYP R A A T V V A R G P L K C V K L D R P R F E R V l G ~ C S D l L K R Y l O O Y N S F V S L S V381
E ~ L ' V A 3 A L E P V O F E 3 G O K I V \ W E P G D ~ F c l ILEG'AAV.ORQSE
NEEFVEVSDLGPS3VcQEIALLUY FPRAATVVA RGPLKCVKlDRPRFERVlGPCSDlLKRYlOOYNSFVSLSV 3 8 1
EQLTVA3AlEPVOFEDGEK I V \ O G E P G D ^ F Y I I'EG'ASVLORQSD
NEEVVEVGQLGPS3VcQE IIILLLY FPRAATVVA RGPLKCVKLDRPRFERVLGPCSEILKRYIORYNSF I S L T V 3 8 1 I IESG:VS I: I K S K T K V N < 3 S E N 3 E \'E ; A Q C * Y G 3 Y c M L A L V ' Y K P R A A S A V A VGDVKClVYDVOAFERLlGPCVDlUKRYlSHYEEOLVKMFGSSMDL 3 9 5 9 R A ' R I la 2 3 5 EQUKI V 3 V I G E K L V K O G E R l I'OGEKAD.FY I IESG:VS I L I R S K T K ' K < N G G Y O E VE ~ A H C ~ K G 3 V ~ Q E L A L V 'KYP R A A S A V A VGDVKCLVUDVOAFERLLGPCWIUKRNISHYEEOLVKUFGSNLDL 365 9 U37R I 'a 2 6 6 EQUK I V 3 V I G E K ' V K D G E R I I A O G E K A D - F Y I IESG:VS I L I R S K T K S h Y N G G # 3 E V E AHCwYG3Y:BELALV'V K P R A A S A V G V G D V K C L V U D V O A F E R L L G P C U DI U K R N I S H Y E E O L V K M F G S N L D L 3 9 6 9 R A " R I I D 281 E Q L K V V 3 V I G T K V V N 3 G E Q I I A ( K 1 3 S A D ~ F C I V E S G ~ V R I T U Y P K G Y S 3 I EEN S A VE ' A Q C L n S 3 Y c G E L I L V " Y K P M A S A H A I G P V K C L A M D V O A F E R L L G P C U E I U K R N l A T Y E E O l V A L F G T N M D t 4 1 0 D U S T S * V V S E L - P S 3 Y c P E I L L L' D F PRAA'VT S I G Y T K C V E LDROQF h R L C G P ID O M L R R N M E T YNOF L N R P P S S P N L T 3 2 3 B S: ' M E R I 9 E Q V S L A 3 A L E P V N c 0 3 G E V I V F O G 3 P G D z F Y I I VE GK V V V T O E T VOG 9 V E A S T R 3 1 0 DQAKLA3ALDTK IVODGET I I F Q30G ' n F Y L I E V G 9 V 3 V S K K GOGV k Y L K D * 3 Y c G E V A L L k D L P A I T V T A T K R T K V A T L G K S G F C R L L G P A V D V L K L N D P T R H 4 1 6 3 C ! X G Y 2 2 5 .SKLA!!\vLEE'*vE4SEY I l ^ & ' ~ G ~ - IF ISdG.:',V'cE?SJ ~ 4 E C 3 .iL"'.Pz
