Regulation of Kinase Activity

SCIENTIFIC CORRESPONDENCE

negative control of kinase activity among Regulation of kinase activity the src family resides within substrate like sequences outside the catalytic SIR-In a recent News and Views article, interpretation of these results is that in the domain. Hardie' described the emerging evidence off state the catalytic carboxy-terminal CARLA GRANDORI that protein kinases can be inhibited by domain interacts with the amino-terminal The Rockefeller University, interacting with substrate-like sequences. substrate-like sequences. Phosphoryla Box 102, 1230 York Ave, These sequences are present either within tion of the amino-terminal site would only New York, the same molecule, as in kinase C, or in a occur when the kinase is activated, New York 10021, USA different molecule, as in the regulatory presumably by an induced conformational subunits of cyclic AMP-dependent kinases. change. Phosphorylation at these sites 1. Hardie. G. Nature 335, 592-593 (1988) . 2. Brugge, J.S. & Darrow, D. J. bioi. Chern. 259, 1550-1557 I would like to point out that an inhibitory may allow maintainance of the on state. (1981). role of amino-terminal sequences over the A similar effect is observed with certain 3. Pawson, T. Oncogene 3, 491-495 (1988). 4. Yonemoto, W., Jarvis·Morar, M., Brugge, J.S., Bolen. J.B. carboxy-terminal catalytic domain has amino-acid substitutions of p60'."' . These & Israel. M. Proc. natn. Acad. Sci. U.S.A. 82, 4568--4572 also been shown for p60'."' , a member of are Gly 6HAsp, Arg 9~Trp , and Thr (1985). 5. Bolen. J.B., Rosen, N. & Israel , M. Proc. natn. Acad. Sci. the non-receptor family of tyrosine 9~Ile , which are present in the viral U.S.A. 82, 7275-7279 (1985). 2 kinases • The mechanism whereby this protein p60'.'", and are sufficient to 6 . Kato. J.Y. eta!. Malec. cell. Bioi. 6, 4155-4160 (1986) . increase the specific activity of the kinase". 7. Potts, W.M., Reynolds, A.B., Lansing, T.J . & Parsons, T.J. negative regulation occurs has not yet Oncogene Res. 3, 343-355 (1989). been addressed, although studies per The recombinant chimaeras between v formed on p60'."', the cellular (c) homol src and c-src used in these studies' are ogue of the viral (v) protein, support the shown schematically in the figure . Among idea of substrate-like sequences being these three mutations, the crucial one is Tale of two serines involved. likely to be Arg 95~Trp, because Gly SIR-In a recent paper', Brenner used the Amino-acid positions 90 and 92 of 63~Asp when present alone in p60'.'" is fact that serine is encoded by two non p60'."' are tyrosine residues, one of which silent, and Thr 9~Ile is not conserved in linked codon types, UCN and AGY, in is surrounded by acidic amino acids as arc other strains of Rous sarcoma virus. conjunction with his observation that many tyrosine-kinase target sequences Arg 95 is three residues away from the within several enzyme families catalytic (see figure). These sequences are located tyrosine-phosphorylation site and its serine residues have different codon in the amino-terminal modulatory domain presence could weaken the interaction representations, to propose that these of the molecule within a stretch of about with the kinase domain, thereby releasing serines evolved convergently by single 50 amino acids that are conserved among the inhibition and allowing constitutive substitutions in cysteine or threonine all the members of the src family. This activation of the kinase. codons (UGY and ACN, respectively), region has been called the SH3 or A box' A more rigorous proof that Arg 95 lies the latter being catalytic residues in and it is not present in the receptor class of within a kinase regulatory domain has ancestral enzymes of each class . This pro tyrosine kinases. been obtained by Potts et at. ', who find posal, however, fails to account for the Activated p60'-"' molecules, such as that the Arg 95~ Trp mutation alone is serine codon representations of two those bound to middle-T antigen of sufficient to activate p60''". In addition, groups of proteins. polyomavirus, or those present in certain this group and also L. Fox, K. Frost and First, the chymotrypsin-like proteases neuroblastoma cell lines, are phosphory J .S. Brugge (personal communication), from Streptomyces griseus, SGPA and lated on tyrosine at their amino termini'"' . have shown that mutations at either tyro SGPB (ref. 2): in these enzymes, the Basoo on their neighbouring sequences, sines 90 or 92, or the deletion of amino catalytic serine residues are encoded by tyrosine residues 90 and 92 represent the acids 92-95, activate the transforming AGU and UCC, respectively, but the 61 most likely targets of phosphorylation potential and the kinase activity of p60'.'". per cent identity of amino acids between within this amino-terminal region. An Thus, it is possible that an intramolecular SGPA and SGPB, and their similar sizes, make it highly improbable that they w belong to two different phylogenetic line 95 ages, as predicted by Brenner's hypo * * thesis. Second, the viral replicative pro peoc-arc(amino acid s1-101) GGVTTFVAL YD YESRTETDLSFKKGE teins containing the widespread nucleoside-5' -triphosphate (NTP)-binding Relative motif (refs 3-5) Gly-X-X-X-X-Gly-Lys 63 95 96 260 516 I I , catalytic domain. • kinaee activity Ser/Thr (GXXXXGK SIT) where X is any I I amino acid: as shown in the figure and v-arc (SRA-RSV-NY) j{ II 01 fl table, the lineages of the GKS-containing proteins of picornavirus, comovirus and v-arc (SRA-RSV-SF) nepovirus derived from serine codon types are incompatible with the phylogeny vfc-orc (NY851) 0.5-0.11 arising from sequence comparisons (com vfc-arc (NYHB5) 0.0-1.-0.05 pare a and b in the figure). Moreover, the conspicuous absence in this family of a c-ore (NY5H) 0.0-1.-0.05 GKT-containing protein is at odds with Brenner's additional suggestion that, in Amino acids 81 to 106 of p60'·"c W, substitution of Arg-. Trp as in p60'."', asterisks are above the case of the NTP-binding motif, threo tyrosines 90 and 92, underlines, acidic amino acids surrounding the tyrosine-phosphorylation nine might be an evolutionary inter sites. Lower panel, schematic diagram of various src proteins from: Schmidt Ruppin subgroup A mediate between the two kinds of serine, Rous sarcoma virus, New York strain (SR.A.-RSV-NY); San Francisco strain (SRA-RSV-SF) ; rather than their predecessor. It seems to chimaeric viral and cellular src recombinant viruses (NY851 and NYHB5) ; a non-mutated cellular srcrecombinantvirus (NY5H). Mutations 63, 95 and 96 are indicated; vertical lines show amino• be the case, in the proteins of this family, acid substitutions in other regions of the viral protein as compared with the cellular src protein. that threonine is not acceptable in the The relative kinase activity was measured by in vitro phosphorylation of enolase as previously NTP-binding motif and that evolutionary 5 described . transitions between UCN and AGY NATURE · VOL 338 · 6 APRIL 1989 467

Phylogeny of putative NTPases CODON UCN AGY UCN AGY UCN AGY UCG AGU GKSS sequence with the two serines of picornavirus. comovirus and ENTERO- encoded by codons of different series) nepovirus. a, Phylogenetic VIRUS RHINO- FMDV followed by deletion of the original codon. scheme derived from the data CARDIO- HAV ENTERO- FMDV HAV TBRV CPMV RHINO- CARDIO- TBRV CPMV of the table according to Bren• This mechanism is equally feasible with ner's hypothesis; b, phylo• DNA and RNA genomes and might oper genetic tree generated by ate wherever there is no strict constraint comparison of amino-acid on the residue(s) next to the functional sequences of evolutionary serine. Generally, understanding the conserved segments of puta• evolutionary history of serine codons in tive NTPases using a rate• catalytic centres of each enzyme class independent distance matrix 7 8 requires knowledge of its phylogeny method · Only the branching a derived from independent data. order is shown; the branch EUGENEV. KOONIN lengths were chosen arbitrarily. The branching order of this tree is identical to that generated for 9 10 viral RNA polymerases and for capsid proteins , and presumably reflects the phylogeny of viral ALEXANDER E. GORBALENYA genomes as a whole. Codon representations of serine in the GKS consensus are shown. Where, Institute of Poliomyelitis and in a group of viruses, only one codon series is utilized, the branching order was not further Viral Encephalitides, specified. USSR Academy of Medical Sciences, codons occur without GKT intermediates. bility. Nevertheless, it seems that if this 142782 Moscow Region, In two other families of viral GKS/T occurred then it did so only rarely. USSR containing proteins serine is encoded Although our survey of serine codon 1. Brenner. S. Nature 334, 528--530 (1988). almost exclusively by UCN, with AGY usage in putative viral NTPases does not 2 Henderson. G. et al. I Bact. 169. 3778- 3784 occurring only once (see table). Again, it support Brenner's hypothesis, it exposes (1987). is unreasonable to suppose that the potex some intriguing variations in the evolu 3. Walker. J.E.. Saraste, M .J.. Runswick, M.J . & Gay, N.J. EMBOI 1, 945-951 (1982) virus gene containing AGU originated tionary mechanisms of different phylo 4 . Gorbalenya . A.E.. Blinov. V.M .. Donchen~o . A.P . & from a separate line of descent. Both these genetic lineages. At least two mechanisms Koonin. E.V. I molec. Evol. 28 (in the press) . 5. Evans . R.K .. Haley. B. E. & Roth. D.A. J. bioi. Chern. 260, families, however, include G KT may be invoked to explain how the UCN 7800- 7804 (1985). containing proteins seemingly making the to AGY transition occurs without loss of 6 . Steinhauer. D.A. & Holland, J.J. A. Rev. Mrcrobiol. 41, 409-433 (1988) evolution of differentially encoded serines serine at the enzyme active site. The first 7. Yushmanov, S.V. & Chumakov. K.M. Malek. Genetika 3, through threonine intermediates a possi- of these requires the simultaneous change 9-15 (1988). of two adjacent bases. Given the high 8. Chumakov. K.M . & Yushmanov, S.V. Malec. bioi. Evol. (In Serine codon representations in the nucleotide- error rate of RNA replication', this mech the press) . binding motif of positive strand RNA viruses 9. Koonin. E.V .. Chumakov. K.M .. Yushmanov, S.V. & anism is more feasible for RNA viral Gorbalenya. A. E. Molek. Genetika3,16-19 (1988). 10. Pal men berg, A.C. in Molecular Biology of Picomaviruses Family of viral Consensus Serine genomes than for DNA genomes. An (ICN- UCI. in the press). 'NTPases' sequence (threonine) alternative mechanism involves the gener 11. Gort>alenya. A.E.. Koonin. E.V., Donchenko, A.P. & Blinov. codons V.M. FEBS Lett. 23516-24 (1988). ation of a new serine codon next to the 12. Gorbalenya. A.E .. Koonin. E.V .. Donchenko. A.P . & Family I functionally important one (yielding a Blinov, V.M . (submitted).

Alphaviruses GKS UCN Coronavirus GKS ucc Telomere formation in yeast Furovirus GKS UCN Hordeivirus GKS UCA SIR-We recently demonstrated that dur-, ing telomeric repeats at each end; (2) Tobravirus GKS UCG ing formation of new telomeres in the asymmetrical resolution of the circles thus Potexviruses yeast Saccharomyces cerevisiae, telomeric formed; and (3) telomere formation on an WC1MVp147 GKS ucu sequences are often transferred between end with the telomeric repeats in the p26 GKS ucu PVX p165 GKS AGU DNA termini'. We argued that the most 'wrong'u orientation. Not only have none p26 GKS ucc reasonable explanation for this transfer is of these processes been demonstrated in Tricornaviruses GKT ACN recombination between DNA termini. yeast, but even symmetrical resolution is Tobamovirus GKT ACC In a recent News and Views article', inefficient (- 1 per cent per cell division)' Tymovirus GKT ACA however, Szostak suggested that the telo compared with the sequence transfer we Family II mere resolution reaction' (the cleavage observe. between two blocks of telomeric Because the resolution reaction is Enteroviruses GKS UCN sequences that are oriented as a head-to excluded unequivocally by DNA seq Rhinoviruses GKS UCN Cardioviruses GKS UCN head inverted repeat'') can provide an uence data, telomere-telomere recomb Aphtoviruses (FMDV) GKS AGY alternative explanation for our data, a ination remains the only reasonable expla Hepatitis A GKS AGY possibility that can be addressed defini nation for the transfer of telomeric virus (HAV) tively by DNA sequencing. It is not clear sequences that we have observed. Comovirus (CPMV) GKS AGU Nepovirus (TBRV) GKS UCG to us why this possibility was raised Whether or not yeast exploits telomere because we stated' that our unpublished telomere recombination in the replication Family Ill sequence data support the interpretation or maintenance of telomeres remains to presented in the article; that is, the orien be determined. Potyviruses GKS UCN Flaviviruses GKT ACN tation of the transferred repeats is the VIRGINIAA. ZAKIAN Pestivirus GKT ACA same as that of the test termini (S.-S. ANN F. PLUTA Wang and V.A.Z ., in preparation). Fred Hutchison Cancer Research Center, For sources of sequence data see refs 4, 11 and 12. Although the sequence data eliminated 1124 Columbia Street, N, any nucleotide; Y, pyrimidine. The grouping of viral proteins containing the NTP-binding motif is the resolution model as an explanation for Seattle, according to refs. 4 and 12. Potexviruses (as well as the telomeric transfer, we did not discuss Washington 98104, USA furoviruses and probably hordeiviruses) have two these data specificially in terms of this 4 1. Pluta. A.F . & Zakian. V.A. Nature 337, 429-433 (1989). putative NTPases each . Different species of entero• model'. The resolution reaction never virus and rhinovirus. and different strains of foot and 2. Szostak. J.W. Nature 337. 303-304 (1989). provided a likely explanation for our 3. Szostak. J.W. Cold Spring Harbor Symp quant. Bioi. 47. mouth disease virus (FMDV) and HAV have either Cor 1187- 1194 (1982) . U in the third position; hence, N or Y is indicated, results because it requires three events: 4. Murray. A.W .. Claus. T.E . & Szostak. J.W. Malec. cell. Bioi. respectively. (1) circularization of linear plasm ids bear- 8, 4642-4650 (1988). 468 NATURE · VOL 338 · 6 APRIL 1989