Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 7-Day Trial for You or Your Team.

Learn More →

Repetitive zinc‐binding domains in the protein transcription factor IIIA from Xenopus oocytes.

Repetitive zinc‐binding domains in the protein transcription factor IIIA from Xenopus oocytes. The EMBO Journal vol.4 no.6 pp. 1609 - 1614, 1985 Repetitive zinc-binding domains in the protein transcription factor LiA from Xenopus oocytes A.D.McLachlan and A.Klug Results J.Miller, MRC Laboratory of Molecular Biology, Hills Road, Cambridge CB2 2QH, content of 7S particle Zinc UK methods for 7S purification involve ion ex- Because published Communicated by A.Klug high ionic strength buffers, which in our hands dis- change and The 7S particle of Xenopus laevis oocytes contains 5S RNA at least 30% of particle samples, we developed a new sociates and a 40-K protein which is required for 5S RNA transcrip- purification which causes no detectable dis- method for particle tion in vitro. Proteolytic digestion of the protein in the particle We found, in agreement with other workers (Denis sociation. yields periodic intermediates spaced at 3-K intervals and a le Maire, 1983), that the 7S complex dissociated at salt con- and limit digest containing 3-K fragments. The native particle is >0.2 M, and that the isolated protein precipitated centrations shown to contain 7-11 zinc atoms. These data suggest that We also observed that gel filtration of the complex in readily. the protein contains repetitive zinc-binding domains. Analysis of 0.1 mM dithiothreitol (DTT) invariably resulted the presence of the amino acid sequence reveals nine tandem similar units, elution of protein and 5S RNA. This dissociation was in separate each consisting of approximately 30 residues and containing seen when the particle incubated with 0.1 mM DTT was also two invariant pairs of cysteines and histidines, the most com- on agarose gels, suggesting that the 20 cysteine electrophoresed mon ligands for zinc. The linear arrangement of these re- per molecule (Picard and Wegnez, 1979) were somehow groups peated, independently folding domains, each centred on a zinc involved in particle stability. However, when we found that ion, comprises the major part of the protein. Such a struc- 25 mM NaBH4 did not disrupt the complex, we suspected, by ture explains how this small protein can bind to the long analogy with the results of Lewis and Laemmli (1982) on meta- internal control region of the 5S RNA gene, and stay bound phase chromosomes, that metal binding might be involved. When during the passage of an RNA polymerase molecule. the was incubated with 1, 10-phenanthroline, DTT, EDTA particle Key words: Xenopus laevis/transcription factor IA/zinc-binding other chelating agents and run on agarose gels, or a number of observed which could only be prevented by prior domains/7S particle dissociation was Mn2+, Ni2+, Fe2+, Zn2+, and not by Mg2+, Ca2+, addition of + by them- and Cd2+ induced dissociation Ca2+ or Co2 +. Cu2 Introduction bound Zn2+. selves, apparently by displacing RNA genes of Xenopus laevis, which are transcribed by preparation of particle by atomic The 5S Analysis of a 30-50% pure RNA III, have been the subject of intensive study revealed insignificant concentrations of polymerase absorption spectroscopy in the last decade by the research groups of D.D.Brown and R.G. or but a significant concentration of Zn, Cu, Ni, Co Fe, Cd, Roeder. The two types, oocyte and somatic, provide a system of at least 5 mol Zn per mol. particle. at a ratio for studying the differential regulation of gene expression as well were in progress, Hanas et al. (1983) While these experiments as mechanisms (Brown, 1984). In the course of these of Zn in the 7S RNP particle, and the re- transcription reported the presence it has been discovered that the correct initiation of tran- for Zn for the association of TFIIA with the internal studies quirement the binding of a 40-K protein factor, variously of the SS gene. They found a Zn/particle ratio scription requires control region called factor A or transcription factor HIA (TFIIA), which has two in the presence of 5 mM EDTA and three in its of about been purified from oocyte extracts (Engelke et al., 1980). By the latter fell to two in samples purified by absence, although deletion mapping it was found that this factor interacts with a filtration. gel region - 50 nucleotides long within the gene, called the internal We have now repeated the analysis with pure and undissociated control (Bogenhagen et al., 1980). This initiation com- To ensure that no contamination of the prep- region particle preparations. all and ware was washed in is stabilised by the sequential binding of two further pro- could glass plastic plex aration occur, of 10 mM EDTA for a time several times longer tein factors, called B and C (Segall et al., 1980; Bieker et al., several changes The buf- 1985). of the entire preparative procedure. than the duration and submitted with sample for Immature oocytes store 20 000 5S RNA molecules in the form was concentrated particle fer (Picard and Wegnez, 1979), 7S at 65 4M con- of 7S ribonucleoprotein particles spectrophotometry. particle atomic absorption which has been shown to be ident- a ratio of 7.0 + 0.5 mol Zn per each containing a single protein Zn at 460 giving tained iM, factor IHA (Pelham and Brown, 1980; buffer used was at most 60 nM in Zn, ical with transcription mol The original particle. TFiIA therefore binds both 5S RNA was to at most 1.5 litres of buffer during Honda and Roeder, 1980). and exposed particle DNA and it was therefore suggested that it may Since the of was 2 ml solution at and its cognate the yield particle purification. 920 nmol of Zn, at most 90/920 = 10% could mediate autoregulation of 5S gene transcription. Whether this 65 iM, containing time of frog adsorbed at any following homogenization occurs in vivo or not, the dual interaction pro- have been autoregulation structural problem which can be approached ovaries. vides an interesting Hanas et al. contained 0.5 mM of of the protein TFIIIA used (1983) because of the presence large quantities The buffers by 1 has a constant for Zn of In this we report some results of our which binding in oocytes. paper or mM DTT, large Xenopus - and so their value for the on TFiIA which reveal a remarkable repeat- and 1972), studies 1010 (Cornell Corviro, preliminary underestimate. Our buffer contained within the be an structure protein. Zn content may ing IRL Press Limited, Oxford, England. J.Miller, A.D.McLachlan and A.Klug 20 mM MES, a which weakly chelating buffer, nevertheless, .. because of the volumes used in the filtration large gel and dialysis steps in the preparative have reduced the procedure, might Zn content of the particles, suggesting that the value we have ob- tained of seven still be an underestimate. Zn/particle may An experiment in which the was in the particle prepared presence of 10 Zn and then separated from unbound Zn on a gel fil- itM 17K BPM tration column a value of 11-12 mol Zn/mol gave The particle. sequence described below that there analysis suggests be may 6K at least 9 mol Zn/mol particle. 4K BPTI _W 3K ^ The 7S particle is not unique in its for Zn. requirement The 42S particle of X. laevis, which Denis and le Maire (1983) have shown to contain, other a 5S RNA and among components, 5S RNA-binding protein distinct from also TFIIA, requires Zn for stability by agarose has a molar Zn gel assay, very large con- tent, and contains no amounts of other metals significant (Miller, unpublished results). d f Small domains in TFIIIA protein Smith et al. have shown on treatment Fig. Trypsin digestions. Particle samples (0.2 mg/ml) in 20 mM MES (1984) that, with proteo- 50 mM buffer, pH 6.0, KCl were dialyzed against 50 mM Tris-Cl, pH 8.1, lytic enzymes, the 40-K intact rise to a 30-K protein gives break- 50 mM 20 KCl. aliquots were digested with trypsin (20 for Al Ag/mn) down product, which is then converted to a 20-K product. These varying times and with 2 mM stopped benzamidine. Times were (a) 0; proteolytic fragments remain bound to the RNA 5S but Smith 17 24 39 64 h. (b) h; (c) h; (d) h; (e) Electrophoresis was as described in et al. (1984) have purified them and studied their interactions Materials and methods. with the 5S gene by DNase I From these footprinting. exper- iments, Smith et al. (1984) concluded that TFIIIA consists of three structural domains which bind to different of the parts internal control region of the 5S gene: a 20-K protein which binds to the 3' end of the coding strand, an 10-K domain which adjacent extends the to binding the 5' end of the control and a region, third domain of 10-K which does not bind to the DNA directly - - 39K but enables the intact protein to enhance transcription, presumably I3 TMV interaction with RNA through polymerase. We have also carried out proteolysis of the 7S particle using and also trypsin, chymotrypsin, elastase, papain (D.Rhodes, per- BPTI _i'~ sonal communication) to determine whether smaller fragments -#p of the protein bind the RNA might 5S and the 5S RNA gene. a V Our measurements of the breakdown tryptic products that suggest Smith's 30-K and 20-K be 40-K, may closer to 33-K and 39-K, 23-K (data not shown). (-I t-} We have also observed that the 23-K fragment may be reduced to a 17-K fragment which remains to bound the 5S RNA. Fur- C. r,. ther, the 17-K after fragment, prolonged can be proteolysis, Fig. 2. (a) TFIIIA used in these experiments; (b) markers: 12.3 K, 17.2 K, reduced to a limit digest consisting of a mixture of 4-K 6-K, and 25.7 K, 45 K, 66.3 78 K, K; (c) 17 h chymotrypsin (20 tg/ml) digest of 3-K fragments (Figure 1), which themselves ultimately disappear. sample as described in prepared legend to Figure 1. Doublet bands are Chymotrypsin produces higher of these multiples fragments as marked (see Results). Acrylamide concentrations for (a) and (b) were 15%. well. Doublets are seen at -11 and 9.5-K, 7.5 and 6-K 4-K and and a 3-K, spacing between them of 3-3.5-K 2c (Figure and has been deduced from a cDNA clone (Ginsberg et al., 1984), time in d). Early points tryptic and reveal chymotryptic digestion contains an unusually large number of Cys and His residues. At metastable bands between the production of the major fragments first sight these residues appeared to us to form roughly periodic described above not (data shown), which are also approximately groupings. We therefore made a systematic search for repeats periodically spaced. in both the amino acid sequence and the cDNA, using the diagonal The of finding periodic intermediates in the course of proteo- comparison matrix method and the damped Needleman and lytic digestion, and the persistence of small fragments even after Wunsch method (see Materials and methods). prolonged digestion, suggests a periodic arrangement of small, The protein comparison matrices (with short window lengths compact protein domains of size - 3-K. If such repetitive do- of 11, 22 and 30 residues) showed an exceptionally strong and mains existed they might account for the large number of cys- regular pattern of 30-residue repeats in the sequence, with four teine and histidine residues and the multiple Zn binding. repetitions evident in the first half of the molecule (residues 13- We therefore investigated the newly published amino acid 156) and two more clear repeats in the second half (residues 223 sequence (Ginsberg et al., 1984) to see if any structural period- -276). Further analysis of the protein by the damped Needleman icity was manifested in the sequence. and Wunsch method showed that the repeats were even more Analysis of sequence TFIIIA of extensive, being partly obscured by a few gaps in the middle Sequence repeats. The amino acid sequence of TFIIIA, of the sequence. The which final best alignment (Figure 3) shows that 1610 Zinc-binding domains in Xenopus transcription factor MILA 1 8 13 17 23 26 30 T G E K *P V® D G D K K K .. K *R *. R)T 1 ( M G E K A L P V V Y K R ) 1 I ®S F A D®G A A®N K N W K®Q * A®L C * K® 37 tI) 2 T G * E K E E G©E K S L H H®T * R®S L * T 67 P®)P®)K G(®)T 3 T G E K * S D L R®T T K A N M K * K®F N R F® 98 G©)D NOTED 4 N I K I C V@V©H F E N©G K A®K K H N V®Q F * S® 129 QMK 5 Q L * P H E G(D K RS L P S K LQ P®E®C * R®E * V® 159 RMK 6 A G - - * K D D G K T W T®Y L K®V A E C® 188 S®S(®V -(1)PO)K 7 - * A D V®N R H K D K * T® 214 Q D - L Y®R * K(®)R D®Q V(C- - 8 E K E R T R D R ST T A F N®R * Q S F® 246 GC©D 1 V(®)L®P S®2 E E Q R * P®V®E H A K M K K * V® 276 9 G©G C®A S®E * R®S D P E K R K L * K E K C P R P K R S L A S R L T G Y I P P K S K E K 311 [5IA S V S G T E K T D S L V K N K P S G T E T[8 G S L V L D K L T I Q ) 344 Fig. 3. Amino acid sequence of transcription factor from X. oocytes, aligned to show the repeated units. The sequence is in one-letter code IHA laevis (Dayhoff, 1978). The molecule contains: an amino end region (residues 1-12); a lysine-rich zone (277- 309) near the carboxyl end; a short tail region which may bind carbohydrate at Asn 310 and Asn 333, which are indicated by squares. The repeat units are numbered 1-9 on the left side of the diagram. The boxed-in consensus sequence at the top shows the characteristic features of a typical repeat unit, numbered as for a length of 30 residues. The end-point of each unit has been chosen arbitrarily after His-30. The best-conserved residues are ringed. Cys-8, Cys-13, His-26 and His-30 are believed to bind to a Zn2+ ion. Tyr-6, Phe-17 and Leu-23 are hydrophobic residues which may form an inner core for the proposed multiple-domain structure. Asterisks (*) mark positions where an insertion sometimes occurs in the normal pattern, and dots (.) mark variable positions in the sequence. In the main body of the repeats a dash indicates an The are (-) alignment gap. underlined regions those which show clear evidence of a relationship with at least one other unit (see Table I). Note that units 6 and 7 have diverged considerably from the usual pattern, and that residues 277-282 may form a fragmentary extension of repeat unit 9. or conversion et It by gene duplication gene (Hood al., 1975). Table I. Overall similarities between the nine repeat units of the is not yet possible to make a proper analysis evolutionary because the end points of the units are not known; relationships, Residues Unit Similarity values if the TFIIIA these could gene possesses intervening sequences 5 9 2 8 3 4 1 7 6 be Table I a estimate of the of revealing. gives rough degree 130- 159 5 * 23 20 15 18 14 6 8 2 of each of derived from the numbers of relatedness pair units, 247-276 9 23 * 23 16 18 12 10 1 0 in the matrix. The totals of high-scoring windows comparison 38 -67 2 20 23 * 11 17 16 6 1 8 columns in the table a measure of how each unit the give closely 215-246 8 15 16 11 * 17 13 11 5 5 resembles the consensus of the nine global repeats. 68-98 3 18 18 17 17 * 6 6 6 2 It can be seen that units 9 and in that are 5, 2, order, clearly 99-129 4 14 12 16 13 6 * 16 7 0 the most members of the and are like one typical family, very 13 - 37 1 6 10 6 11 6 16 * 00 another. Common features include and Pro-5, Gly-12, Lys-15 189-214 7 8 1 1 5 6 7 0 * 1 Arg-25 (here we refer to the numbering of the consensus sequence 160-188 6 2 0 8 5 2 0 0 1 * in Figure 3). A second group comprises units 8, 3 and 4. These Column totals 106 103 102 93 90 84 28 18 contain some insertions at 4A or 28A. single-residue positions Units 8 and 3 form a as shown the ident- recognisable pair by the number of 1 The value given to each pair is overlapping 1-residue Asn-22 and ities Asp-11, Gly-12, Asp-15, Thr-18, Thr-19, windows, in the correctly aligned comparison of the two units, for which Phe-29. units 7 and 6 have diverged considerably from Lastly, 1, the score exceeds the 0.1 % matching probability level. Since each unit is 7 are shortened at the normal in units 6 and -30 residues long, the maximum possible value is 30. The column totals pattern: particular, of give a measure of how close each unit is to the global consensus them We note their amino ends with (Cys-8)-(Cys-13) loops. irregular 1 of all. The underlinings below the sequence in Figure show the parts each Ala-16 and in that units 1 and 4 have Phe-10, Gly-14, Lys-19 unit that lie in the centre of a window, where the comparison score, with at while units 1 and 6 both have common, Trp-21. least one other unit, reaches the 0. I% level. Thus although all nine units belong to the same family they have diverged in detail and may have taken on specific individual residues 13-276 form a continuous run of a of evolved. repeated motif, DNA-binding functions as the 5S gene control system nine similar each of - 30 residues. Since the It if the same 30-residue units were units, sequence would also not be surprising repeats itself cyclically, the position of the boundary of the motif in numbers in other related later found to occur varying gene is uncertain, but the choice of an end-point shortly after His-30 control proteins. of the consensus allows the largest number of complete units to The of a are evolutionary advantages repeating design probably as be assigned. With this choice, most of the insertions and deletions much the same as those in other linear such many proteins, in a occur near the ends of the units rather than the central ovomucoid inhibitor et loop (Laskowski al., 1980). Probably single DNA was once (consensus position numbers 14-25). functional unit that binds to a half-tum of evolved, Our results that residues 13-276 of TFIIIA in much and suggest strongly and then became used more subtle specialised ways from a ancestral unit of 30 amino acids when a number of similar units were in series. have evolved primitive large joined 1611 and A.D.McLachlan J.Miller, A.Klug Characteristic structural features of the small domains The well-marked repeat pattern in TFLIA suggests strongly that each of the sequence repeats corresponds to a series of small structural elements, or domains, of -30 residues, arranged linearly. In large globular proteins with repeated sequences the units most often form compact dimeric or tetrameric pseudo- symmetrical structures (Rossman and Argos, 1981; McLachlan, 1980). But there are also well-known examples of proteins with compact independently active structural units strung together in line, which can be separated by enzyme cleavage: the 'kringles' in plasminogen (Sottrup-Jensen et al., 1978) and the Kazal-type protease domains in the ovoinhibitor of Japanese quail (Laskow- ski et al., 1980; Bolognesi et al., 1982). In these molecules the structural units have lengths of 80 and 63 residues. The TFIA unit is exceptional because of its small size and unique arrange- ment of Cys Cys His His residues. scheme for a linear of Fig. Folding arrangement each repeated domains, There is strong evidence, both that the repeat unit is a self- centred on a tetrahedral of zinc arrangement ligands. Ringed residues are sufficient folded domain, and that it is stabilised the conserved amino by zinc ions. acids which include the and Cys His zinc ligands, the The first conclusion is negatively charged Asp-l and the three suggested by the appearance of 1, hydrophobic groups that may form multiple a structural core. Black circles mark the most probable DNA-binding side small fragments on proteolysis, with quantized mol. wts. which chains (see Figure In the scheme drawn here 5). the metal ion draws the are often multiples of 3 kd (see above). The second is supported ends of each unit the together, leaving central residues 14-25 to form a by evidence that one TFIIIA molecule binds many zinc ions or An potential DNA-binding loop 'finger'. alternative but much less likely (7-11, see above) and is for inactive in the absence of position the zinc is between the His residues of one the metal unit and the Cys residues of an one. adjacent (Hanas et al., 1983). We therefore believe that most, if not all, of the nine units bind zinc. 13 17 6 8 23 26 30 Zinc is normally tetrahedrally coordinated in inorganic and metallo-organic compounds. The amino acids Cys and His are 63tD its most common ligands in enzymes (see Fersht, 1977), such as carbonic anhydrase, liver alcohol dehydrogenase, carboxy- peptidase A, Cu-Zn superoxide dismutase and thermolysin. Therefore we may picture each small domain as a compact unit formed round a central zinc ion to confer stability. Each zinc BASI I r nnIJ would be coordinated tetrahedrally to the two invariant pairs of Cys and His residues in each unit. The ends of the domain will then be pulled together round the zinc. In Figure 4 we have drawn ACIDICnl one possible arrangement of the zinc ligands which fits this type of scheme. It is important to remember that the Cys . . Cys and His . His loops probably cross over at an angle to form HYDROPHOBTC a tetrahedral box, and that the zinc may IFhI have more than four ligands. For example, a Cys side chain might be shared between ,,_r_ L Fh n L , F1 n FIr zincs in two adjacent units. Figure 5 shows how each domain contains three invariant Fig. 5. Histograms for the average distribution of amino acids along the hydrophobic groups (normally Tyr-6, Phe-17 and Leu-23) length of each 30-residue repeat unit. The height of each in the bar, range and how the acidic side chains are nearly all in or near the the number of 0-9, gives times that class of amino acid occurs at each (Cys-8) . . . (Cys-13) position in the nine loop. units. Positions are numbered as in Figure 3, with Cys-8, Cys-13, His-26 and His-30 treated as special The extended protein loop positions (zinc between residues 14 and ligands). 25 might Amino acids have been classed as follows. (a) DNA-binding: Lys, Arg, form a DNA-binding region. The three-dimensional structures Thr His, Asn, Gln, (Ohlendorf and Matthews, 1983). (b) Basic: Arg, Lys, of gene repressor proteins (Ohlendorf and Matthews, 1983) show His. (c) Acidic: Asp, Glu. (d) Large hydrophobic: Leu, Ile, Val, Phe, Tyr, that the side chains of Lys, His, Asn, Gln and Thr often Trp. The strongest potential interact DNA-binding sites, with five or more DNA- with binding amino acids in the nine the phosphate backbone units, are also marked in Figure of DNA, while Arg can form 4. hydro- gen bonds to the base pairs. Amino acids of this type are con- centrated in the region 14-25 of TFIIIA (Figure 4). instead built into a 'metal-sulphide' pseudocrystal bordered by A structural parallel for our proposed 30-residue metal-centered the polypeptide chain. domain might be the 40-residue calcium-binding unit of the cal- TFIIIA thus appears to have an exceptionally compact mol- modulin family (Moews and Kretsinger, 1975) or the 26-residue ecular architecture for a self-sufficient structure; however, it may ion-sulphur half-domain of bacterial ferredoxin (Adman et al., be that strong interactions between adjacent units are important 1973). However, neither of these structures is an independent for maintaining the structure. We have examined the two-helix folding unit, but forms part of a linked pair of fragments. The DNA-binding motif of the Cro and Lac repressors, but it does metallothionein family of proteins (Boulanger et al., 1983), which not seem possible to accommodate the TFHIA consensus se- contain repeated Cys residues, are also not a good analogy for quence into this model: the loop 14-25 may instead fold into TFIIIA, but a rather peculiar group. The reason is that they bind a long twisted ribbon of ,3-sheet which wraps in some way round metal ions in clusters, so that the the bridged by Cys residues, ions zinc-binding pocket, using the invariant hydrophobic groups are not separately packaged in independent protein cages, but as nuclei. The characteristic Cys-Cys-His-His consensus motif 1612 Zinc-binding domains in Xenopus transcription factor LiHU 5'Non- 2). As pointed out by them, the small fragment CB2 stains quite Coding differently with silver from the two other fragments. We there- RNA Pol Internal Control Other factors ? ,,, fore identify the fragments CB3, CB1 and CB2 with the resi- DNA Upstream dues 1-90, 91- 266 and 267- 344, respectively, since the latter is of different character from the first two (particularly in its high content of lysine). The lower mobility of the shorter peptide CB3 (78 residues) relative to CB3 (90 residues) could be accounted for if some of its residues were modified, as suggested above. The same The question of modifications is being investigated. found the amino-carboxyl orientation (Figure 6) has been by Carnegie group (D. D.Brown, personal communication). Carboxyl rich Amino Discussion Carbohydrate- binding above have The experimental and theoretical studies described lIIA consists led to a picture in which the transcription factor Fig. 6. An interpretation of the structural features of the protein TFIIIA and each mostly of a linear arrangement of nine repeated domains, its interactions with DNA. The DNA is drawn curved, as if resting on a on of zinc The centred a tetrahedral arrangement ligands. repeats long beaded surface of the protein. The internal control region of the 5S RNA gene (bases 40- 100) is drawn as six turns of DNA, with the 5' end are thought to be extended DNA-binding fingers, linked by (flex- of of the non-coding strand at the left. A separate piece upstream DNA may with the ible?) joints. This model is consistent highly asymmetric be from required to promote transcription. The protein sequence runs right of the molecule indicated its shape by physico-chemical proper- to left: the evidence for this orientation is in the text. The amino end given A structure of this kind would ties (Bieker and Roeder, 1984). is followed units 1 in contact with the by nine repeat (residues -276) how a small of 40 K can bind to a control These units form the 30-K of explain relatively protein long region. together proteolytic fragment Smith et al. The first six units 13- (1984). repeat (residues 188) probably of DNA: if each domain interacted with stretch double-helical constitute the 20-K in the proteolytic fragment: irregularities repeat pattern half a DNA for the nine about period, then, allowing end effects, of and 7 to a between units 6 may correspond susceptible cleavage region could cover - 50-60 the of the domains nucleotides, length the two The are to be extended fragments. repeats thought DNA-binding internal control et The DNA or RNA region (Sakonju al., 1980). linked each a zinc ion centre. Residues 'fingers' by flexible joints, having 277- 344 include a followed two of each domain could be and lysine-rich region, by potential binding strength modulated, specific near the end with filled carbohydrate-binding sequences carboxyl (marked recognition of short nucleic acid stretches established, by vari- and these form a the 10-K triangles), together parts separate domain, ations in the at the sequence finger tips. proteolytic fragment. The relationship between the grooves of the DNA and One of for the advantage a many-fingered design transcrip- the of the units is unknown and the must not be positions protein drawing of the taken but each unit binds to about one half-turn of DNA. The tion factor is that it binds to an internal control region gene literally, tip of a unit could fit into a groove. in a where a stable is formed system transcriptional complex et Lassar (Bogenhagen al., 1982; Gottesfeld and Bloomer, 1982; found in TFIHIA appears to be unique to this protein. We have et This al., 1983). complex can sustain many rounds of transcrip- searched for it without success in a large number of zinc enzymes, tion which the factor to during presumably remains bound the including those mentioned above, as well as in the metallothio- As the gene. polymerase passes through the gene, the many- neins. There is, therefore, no evidence yet that this motif, or fingered protein could release those fingers bound ahead of the indeed any other repetitive pattern, is a typical feature of zinc- but its processing polymerase, stay bound by remaining fingers, as a class. binding proteins whether to the DNA or intact double helix to the non-coding Relation oflarge domains to the protein sequence. The nine small strand. of domains containing repeated pairs cysteines and histidines com- We have already mentioned above that the transcription fac- prise residues -276 or 1-280 (depending where the boun- tor to be a evolved version of a appears highly small molecule at the dary is taken). The remaining 70 residues carboxyl end of the size of one of the 3-K domains contemporary stabilised units The se- have no homology with the repeating (Figure 3). a metal ion. The molecule could have assisted by primitive simply this terminal be of quence suggests that region might composed an form of Evolution to the elaborate tran- early transcription. - is rich in two parts. The first half of 30 residues very lysines found could have taken scription apparatus today place by gene in this a no and arginines, resembling respect histone, although with different extra functions. It duplication, repeats taking up The second half lacks is as remarked Hanas et al. that significant sequence homology emerges. noteworthy, by (1983) RNA as has been out to us Dr H-C. this enrichment and, pointed by m contains and it could be that the initial polymerase zinc, activity sites for addition Th0gersen, contains two potential carbohydrate of primitive TFIIIA promoted transcription in the absence of a characteristic is N X S for at 310 and 333. which evolved later aspargines [The sequence polymerase molecule, presumably only (Marshall, 1972).] greater efficiency. in character of the last 70 residues et al. a between a The marked difference sug- Ginsberg (1984) have noted homology region this be identified with the 10-K 'domain' of DNA and the of the TFIIA gest that region might 5S (or RNA) coding sequence one end of the the studies of have TFIIHA could interact with its own revealed at protein by proteolysis gene, and suggested that This 10-K domain is for efficient or the derived RNA to at the tran- Smith et al. (1984). required gene autoregulate expression RNA unlike the 30-K or translational levels. The of this transcription, but, remaining fragment (and scriptional evolutionary origin 20-K not for to the internal con- be as follows. RNA is be- its smaller subfragment), binding contemporary property may widely amino and orien- a small RNA trol of the RNA The lieved to have so that region 5S gene. carboxyl preceded DNA, primitive would then be as shown of the 3-K tation of the relative to the could have coded for a small protein gene protein (the precursor RNA in Figure 6. repeating unit) which bound back to the and so stimulated with the bromide This orientation is consistent its own cyanogen production. shown Smith et al. the deletions in units 6 and 7 of the of the It be that cleavage map protein by (1984, Figure may contemporary 1613 J.Miller, A.D.McLachlan and A.Klug with distance from the centre of each window. This newer method is more suitable are to enhance its and protein (Figure 3) necessary flexibility for with dealing gaps, but is less susceptible to statistical analysis. enable it to accommodate to the or structure secondary tertiary In the comparison matrices the highest observed score with a window of 30 of the 5S as well as to the DNA. RNA, to a double corresponded matching probability for the two peptide segments of The 10-K domain of Smith et al. does not bind to the (1984) 0.53 x 10-10 only (McLachlan, 1971) and was highly significant. We also calcu- lated the score for the of two random internal control of the and we have identified it with highest expected comparison 344-residue region gene the score which would achieved sequences, i.e., be on average just once with the 70 residues at the C-terminal end of the Smith et protein. our protein. This score was exceeded not but 359 times in the natural se- once, al. have that this domain (1984) suggested 'transcription' may quence and showed that there are many significant repetitions. interact with other factors to form a tran- required competent In the damped Needleman-Wunsch method we used an effective range of 30 RNA or with Ill. To this residues with rather low 2.0 to start each and 2.0 for scription complex directly polymerase gap penalties: gap each extension by one residue. we would add in view of its and only that, high lysine arginine it also be in involved to elements content, might binding upstream Acknowledgements of the DNA outside the of the coding region gene. We thank Dr G.L.Everett (Johnson-Matthey Ltd., Royson) for determinations of the content of zinc and other metals, and Professor N.Hales and Dr G.Maguire Materials and methods (Department of Clinical Biochemistry, Cambridge University) for zinc determin- 7S ations. We are grateful to Drs D.D.Brown and H.Pelham for advice, and our particle punfication colleagues Drs D.Rhodes and H-C.Th0gersen for their help. 7S RNP were from the ovaries of immature X. laevis particles prepared (South African Snake a method to be described elsewhere in Farm) by (Miller, prep- The method References does not the to such as aration). subject particle strong chelators, or to salt concentrations above 0.07 M. Protein is at least 95% DTT, pure by Adman,E.T., Sieker,L.C. and Jensen,L.H. (1973) J. Biol. Chem., 248, 3987- Coomassie or silver and >90% amino acid not stain, by analysis (data shown). The is no more than 5% dissociated as measured elec- particle by agarose gel Bieker,J.J., Martin,P.L. and Roeder,R.G. (1985) Cell, 40, 119-127. and contains no detectable free 5S RNA. trophoresis (see below) usually Bieker,J.J. and Roeder,R.G. (1984) J. Biol. Chem., 259, 6158-6164. For of metal and item to which quantitation content, glassware, tubing any Bogenhagen,D.F., Sakonju,S. and Brown,D.D. (1980) Cell, 19, 27-35. buffer was was treated for at least 48 h in 30% nitric washed exten- exposed acid, Bogenhagen,D.F., Wormington,W.M. and Brown,D.D. (1982) Cell, 28,413-421. with water and for h then treated at least 72 in 10 mM sively glass-distilled Na2 Bolognesi,M., Gatti,G., Menegatti,E., Guarneri,M., Marquart,M., Papamokos, EDTA the for EDTA (pH 4), optimum pH Zn2+ complexation by (West, 1969). E. and Huber,R. (1982) J. Mol. Biol., 162, 839-868. items not resistant to nitric column such as matrices or Any acid, dialysis tubing, Boswell,D.R. and McLachlan,A.D. (1984) Nucleic Acids Res., 12, 457-464. were rinsed in several of 10 mM EDTA 72 h. for at least changes Na2 (pH 4) Boulanger,Y., Goodman,C.M., Forte,C.P., Fesik,S.W. and Armitage,J.M. (1983) All materials were washed with water before use. Buffers were glass-distilled Proc. Natl. Acad. Sci. USA, 80, 1501-1505. a Chelex 100 column before use passed through ion-exchange (BioRad) directly 359-365. Brown,D.D. (1984) Cell, 37, to remove divalent cations. Protein concentration was measured amino acid by Cornell,N.W. and Crivaro,K.E. (1972) Anal. Biochem., 47, 203-208. analysis. Dayhoff,M.0. (1978) Atlas of Protein Sequence and Structure, Vol. 5, suppl. Particle dissociation 3, published by National Biomedical Research Foundation, Washington D.C. assay in Denis,H. and le Maire,M. (1983) Roodyn,D.B. (ed.), Subcellular Biochemistry, 7S at concentrations of 100 were loaded with particle preparations, usually ytg/ml, Vol. 9, Plenum Publishing Co., pp. 263-297. an volume of 50% 50 mM 0.1% equal glycerol, Tris-Cl, pH 7.5, xylene cyanol and Engelke,D.R., Ng,S.Y., Shastry,B.S. Roeder,R.G. (1980) Cell, 19, 717-728. onto 0.7% in 50 mM 90 mM boric acid 8.5 and run agarose gels Tris-Cl, pH Structure and Fersht,A. (1977) Enzyme Mechanism, published by W.H.Freeman for - 1 h at 8 V/cm. Gels were stained with 1 ethidium bromide and mg/ml & San Co., Francisco. visualised under u.v. have confirmed that the fraction of RNA in Experiments and Ginsberg,A.M., King,B.O. Roeder,R.G. (1984) Cell, 39, 479489. with free 5S RNA correlates with particle preparations co-migrating qualitatively and Gottesfeld,J. Bloomer,L.S. (1982) Cell, 28, 781-791. the fraction of 5S RNA in a from 7S in filtration eluting separate peak particle gel Hanas,J.S., Hazuda,D.J., Bogenhagen,D.F., Wu,F.Y.H. and Wu,C.W. (1983) that it be taken as a measure of (Miller, unpublished results), suggesting may J. Biol. 14 120-14 125. Chem., 258, free and bound 5S RNA in solution. and 119-126. Honda,B.M. Roeder,R.G. (1980) Cell, 22, Polyacrylamide gel electrophoresis and Annu. Rev. Hood,L.M., Campbell,J.H. Elgin,S.C.R. (1975) Genet., 9, 305- was as described Laemmli with the modifi- Electrophoresis by (1970) following cations. was 3% 0.15% 0.125 M Tris- 680-685. Stacking gel acrylamide, bis-acrylamide, Laemmli,U.K.(1970) Nature, 227, 0.1% 0.01% 0.1% ammonium and P04, pH 6.8, SDS, TEMED, persulphate. Running Laskowski,M., Kato,I., Kohr,W.J., March,C.J. Bodard,W.C. (1980) in was 22.5% 0.73% 0.75 Tris-CI 0.1% 0.01% Protides Vol. gel acrylamide, bis, pH 8.8, SDS, Peeters,H. (ed.), of Biological Fluids, 28, Pergamon Press, 0.1% ammonium were diluted with an 123-128. TEMED, persulphate. Samples equal Oxford, pp. volume of mM mM 100 NaPO4 2% 100 DTT and 50% and pH 6.8, SDS, glycerol Lassar,A.B., Martin,P.L. Roeder,R.G. (1983)Science, 222, 740-748. and boiled for 3 before mol. wts. were found min and directly loading. Approximate Lewis,C.D. Laemmli,U.K. (1982) Cell, 29, 171-181. from a of the of mol. wts. two tobacco logarithmic plot mobility against markers, J. Mol. McLachlan,A.D. (1971) Biol., 61, 409-421. mosaic virus and bovine inhibitor Gels in Protein protein (TMV) pancreatic trypsin (BPTI). McLachlan,A.D. (1980) Jaenicke,R. (ed.), Folding, Elsevier/North were stained with 0.5% PAGE blue 83 (BDH Biochemicals). 79-96. Holland, Amsterdam, pp. J. Mol. 15-30. McLachlan,A.D. (1983) Biol., 169, Proteolysis Annu. Rev. Marshall,R.D. (1972) Biochem., 41, 673-702. was at conducted concentrations of 200-500 Proteolysis always particle mg/ml and J. Mol. 201-225. Moews,P.C. Kretsinger,R.H. (1975) Biol., 91, in 50 mM Tris-CI 50 mM 0.5 mM with 20 pH 8.1, KCI, MgCI2, tryp- JAg/mIl and J. Mol. 443453. Needleman,S.B. Wunsch,C.D. (1970) Biol., 48, room Reactions were either with 2 mM sin, temperature. stopped benzamidine, and Annu. Rev. Ohlendorf,D.H. Matthews,B.W. (1983) Biophys. Bioeng., 12, or for 2 in buffer. min by boiling loading 259-284. methods Sequence repeats of analysis and Pelham,H.R.B. Brown,D.D. (1980) Proc. Natl. Acad. Sci. USA, 77, 4170- We used two well-established In methods. the the matrix 4174. first, diagonal comparison the is divided into all its and (McLachlan, 1971, 1983), sequence possible overlapping Picard,B. Wegnez,M. (1979) Proc. Natl. Acad. Sci. USA, 76, 241-245. or of a fixed without insertions or and segments, windows, given length, deletions, Rossmann,M.G. Argos,P.W. (1981) Annu. Rev. Biochem., 50, 497-532. and of is The score for every pair segments compared independently. comparing Sakonju,S., Bogenhagen,D.F. and Brown,D.D. (1980) Cell, 19, 13-25. each of amino acids in the window is derived from mutation likeli- and J. pair Dayhoffs Segall,J., Matsui,T. Roeder,R.G. (1980) Biol. Chem., 255, 11986-11991. hood tables and is assessed the exact and (Dayhoff, 1978) against calculated-prob- Smith,D.R., Jackson,T.J. Brown,D.D. (1984) Cell, 37, 645-652. distribution of the window scores for random with the same ability sequences Sottrup-Jensen,L., Claeys,H., Zajdel,M., Petersen,T.E. and Magnusson,S. (1978) as the whole The cDNA se- Prog. Chem. Fibrinolysis Thrombolysis, 3, 191-209. average composition protein (McLachlan, 1983). were scored identical bases. In the second the quences by counting method, damped West,T.S. (1969) Complexometry with EDTA and Related Reagents, published Needleman-Wunsch alignment with gaps (Boswell and McLachlan, 1984; Needle- by BDH Chemicals Ltd., Poole, UK. and are man with for insertions Wunsch, 1970), sequences aligned locally penalties and but the scores are which die Received on 4 deletions, 1985 given weights away exponentially April http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png The EMBO Journal Springer Journals

Repetitive zinc‐binding domains in the protein transcription factor IIIA from Xenopus oocytes.

Miller, J.; McLachlan, A.D.; Klug, A.
The EMBO Journal , Volume 4 (6) – Jun 1, 1985
Download PDF
6 pages

Loading next page...
 
/lp/springer-journals/repetitive-zinc-binding-domains-in-the-protein-transcription-factor-DGWSO0AVJc

References (2)

Publisher
Springer Journals
Copyright
Copyright © European Molecular Biology Organization 1985
ISSN
0261-4189
eISSN
1460-2075
DOI
10.1002/j.1460-2075.1985.tb03825.x
Publisher site
See Article on Publisher Site

Abstract

The EMBO Journal vol.4 no.6 pp. 1609 - 1614, 1985 Repetitive zinc-binding domains in the protein transcription factor LiA from Xenopus oocytes A.D.McLachlan and A.Klug Results J.Miller, MRC Laboratory of Molecular Biology, Hills Road, Cambridge CB2 2QH, content of 7S particle Zinc UK methods for 7S purification involve ion ex- Because published Communicated by A.Klug high ionic strength buffers, which in our hands dis- change and The 7S particle of Xenopus laevis oocytes contains 5S RNA at least 30% of particle samples, we developed a new sociates and a 40-K protein which is required for 5S RNA transcrip- purification which causes no detectable dis- method for particle tion in vitro. Proteolytic digestion of the protein in the particle We found, in agreement with other workers (Denis sociation. yields periodic intermediates spaced at 3-K intervals and a le Maire, 1983), that the 7S complex dissociated at salt con- and limit digest containing 3-K fragments. The native particle is >0.2 M, and that the isolated protein precipitated centrations shown to contain 7-11 zinc atoms. These data suggest that We also observed that gel filtration of the complex in readily. the protein contains repetitive zinc-binding domains. Analysis of 0.1 mM dithiothreitol (DTT) invariably resulted the presence of the amino acid sequence reveals nine tandem similar units, elution of protein and 5S RNA. This dissociation was in separate each consisting of approximately 30 residues and containing seen when the particle incubated with 0.1 mM DTT was also two invariant pairs of cysteines and histidines, the most com- on agarose gels, suggesting that the 20 cysteine electrophoresed mon ligands for zinc. The linear arrangement of these re- per molecule (Picard and Wegnez, 1979) were somehow groups peated, independently folding domains, each centred on a zinc involved in particle stability. However, when we found that ion, comprises the major part of the protein. Such a struc- 25 mM NaBH4 did not disrupt the complex, we suspected, by ture explains how this small protein can bind to the long analogy with the results of Lewis and Laemmli (1982) on meta- internal control region of the 5S RNA gene, and stay bound phase chromosomes, that metal binding might be involved. When during the passage of an RNA polymerase molecule. the was incubated with 1, 10-phenanthroline, DTT, EDTA particle Key words: Xenopus laevis/transcription factor IA/zinc-binding other chelating agents and run on agarose gels, or a number of observed which could only be prevented by prior domains/7S particle dissociation was Mn2+, Ni2+, Fe2+, Zn2+, and not by Mg2+, Ca2+, addition of + by them- and Cd2+ induced dissociation Ca2+ or Co2 +. Cu2 Introduction bound Zn2+. selves, apparently by displacing RNA genes of Xenopus laevis, which are transcribed by preparation of particle by atomic The 5S Analysis of a 30-50% pure RNA III, have been the subject of intensive study revealed insignificant concentrations of polymerase absorption spectroscopy in the last decade by the research groups of D.D.Brown and R.G. or but a significant concentration of Zn, Cu, Ni, Co Fe, Cd, Roeder. The two types, oocyte and somatic, provide a system of at least 5 mol Zn per mol. particle. at a ratio for studying the differential regulation of gene expression as well were in progress, Hanas et al. (1983) While these experiments as mechanisms (Brown, 1984). In the course of these of Zn in the 7S RNP particle, and the re- transcription reported the presence it has been discovered that the correct initiation of tran- for Zn for the association of TFIIA with the internal studies quirement the binding of a 40-K protein factor, variously of the SS gene. They found a Zn/particle ratio scription requires control region called factor A or transcription factor HIA (TFIIA), which has two in the presence of 5 mM EDTA and three in its of about been purified from oocyte extracts (Engelke et al., 1980). By the latter fell to two in samples purified by absence, although deletion mapping it was found that this factor interacts with a filtration. gel region - 50 nucleotides long within the gene, called the internal We have now repeated the analysis with pure and undissociated control (Bogenhagen et al., 1980). This initiation com- To ensure that no contamination of the prep- region particle preparations. all and ware was washed in is stabilised by the sequential binding of two further pro- could glass plastic plex aration occur, of 10 mM EDTA for a time several times longer tein factors, called B and C (Segall et al., 1980; Bieker et al., several changes The buf- 1985). of the entire preparative procedure. than the duration and submitted with sample for Immature oocytes store 20 000 5S RNA molecules in the form was concentrated particle fer (Picard and Wegnez, 1979), 7S at 65 4M con- of 7S ribonucleoprotein particles spectrophotometry. particle atomic absorption which has been shown to be ident- a ratio of 7.0 + 0.5 mol Zn per each containing a single protein Zn at 460 giving tained iM, factor IHA (Pelham and Brown, 1980; buffer used was at most 60 nM in Zn, ical with transcription mol The original particle. TFiIA therefore binds both 5S RNA was to at most 1.5 litres of buffer during Honda and Roeder, 1980). and exposed particle DNA and it was therefore suggested that it may Since the of was 2 ml solution at and its cognate the yield particle purification. 920 nmol of Zn, at most 90/920 = 10% could mediate autoregulation of 5S gene transcription. Whether this 65 iM, containing time of frog adsorbed at any following homogenization occurs in vivo or not, the dual interaction pro- have been autoregulation structural problem which can be approached ovaries. vides an interesting Hanas et al. contained 0.5 mM of of the protein TFIIIA used (1983) because of the presence large quantities The buffers by 1 has a constant for Zn of In this we report some results of our which binding in oocytes. paper or mM DTT, large Xenopus - and so their value for the on TFiIA which reveal a remarkable repeat- and 1972), studies 1010 (Cornell Corviro, preliminary underestimate. Our buffer contained within the be an structure protein. Zn content may ing IRL Press Limited, Oxford, England. J.Miller, A.D.McLachlan and A.Klug 20 mM MES, a which weakly chelating buffer, nevertheless, .. because of the volumes used in the filtration large gel and dialysis steps in the preparative have reduced the procedure, might Zn content of the particles, suggesting that the value we have ob- tained of seven still be an underestimate. Zn/particle may An experiment in which the was in the particle prepared presence of 10 Zn and then separated from unbound Zn on a gel fil- itM 17K BPM tration column a value of 11-12 mol Zn/mol gave The particle. sequence described below that there analysis suggests be may 6K at least 9 mol Zn/mol particle. 4K BPTI _W 3K ^ The 7S particle is not unique in its for Zn. requirement The 42S particle of X. laevis, which Denis and le Maire (1983) have shown to contain, other a 5S RNA and among components, 5S RNA-binding protein distinct from also TFIIA, requires Zn for stability by agarose has a molar Zn gel assay, very large con- tent, and contains no amounts of other metals significant (Miller, unpublished results). d f Small domains in TFIIIA protein Smith et al. have shown on treatment Fig. Trypsin digestions. Particle samples (0.2 mg/ml) in 20 mM MES (1984) that, with proteo- 50 mM buffer, pH 6.0, KCl were dialyzed against 50 mM Tris-Cl, pH 8.1, lytic enzymes, the 40-K intact rise to a 30-K protein gives break- 50 mM 20 KCl. aliquots were digested with trypsin (20 for Al Ag/mn) down product, which is then converted to a 20-K product. These varying times and with 2 mM stopped benzamidine. Times were (a) 0; proteolytic fragments remain bound to the RNA 5S but Smith 17 24 39 64 h. (b) h; (c) h; (d) h; (e) Electrophoresis was as described in et al. (1984) have purified them and studied their interactions Materials and methods. with the 5S gene by DNase I From these footprinting. exper- iments, Smith et al. (1984) concluded that TFIIIA consists of three structural domains which bind to different of the parts internal control region of the 5S gene: a 20-K protein which binds to the 3' end of the coding strand, an 10-K domain which adjacent extends the to binding the 5' end of the control and a region, third domain of 10-K which does not bind to the DNA directly - - 39K but enables the intact protein to enhance transcription, presumably I3 TMV interaction with RNA through polymerase. We have also carried out proteolysis of the 7S particle using and also trypsin, chymotrypsin, elastase, papain (D.Rhodes, per- BPTI _i'~ sonal communication) to determine whether smaller fragments -#p of the protein bind the RNA might 5S and the 5S RNA gene. a V Our measurements of the breakdown tryptic products that suggest Smith's 30-K and 20-K be 40-K, may closer to 33-K and 39-K, 23-K (data not shown). (-I t-} We have also observed that the 23-K fragment may be reduced to a 17-K fragment which remains to bound the 5S RNA. Fur- C. r,. ther, the 17-K after fragment, prolonged can be proteolysis, Fig. 2. (a) TFIIIA used in these experiments; (b) markers: 12.3 K, 17.2 K, reduced to a limit digest consisting of a mixture of 4-K 6-K, and 25.7 K, 45 K, 66.3 78 K, K; (c) 17 h chymotrypsin (20 tg/ml) digest of 3-K fragments (Figure 1), which themselves ultimately disappear. sample as described in prepared legend to Figure 1. Doublet bands are Chymotrypsin produces higher of these multiples fragments as marked (see Results). Acrylamide concentrations for (a) and (b) were 15%. well. Doublets are seen at -11 and 9.5-K, 7.5 and 6-K 4-K and and a 3-K, spacing between them of 3-3.5-K 2c (Figure and has been deduced from a cDNA clone (Ginsberg et al., 1984), time in d). Early points tryptic and reveal chymotryptic digestion contains an unusually large number of Cys and His residues. At metastable bands between the production of the major fragments first sight these residues appeared to us to form roughly periodic described above not (data shown), which are also approximately groupings. We therefore made a systematic search for repeats periodically spaced. in both the amino acid sequence and the cDNA, using the diagonal The of finding periodic intermediates in the course of proteo- comparison matrix method and the damped Needleman and lytic digestion, and the persistence of small fragments even after Wunsch method (see Materials and methods). prolonged digestion, suggests a periodic arrangement of small, The protein comparison matrices (with short window lengths compact protein domains of size - 3-K. If such repetitive do- of 11, 22 and 30 residues) showed an exceptionally strong and mains existed they might account for the large number of cys- regular pattern of 30-residue repeats in the sequence, with four teine and histidine residues and the multiple Zn binding. repetitions evident in the first half of the molecule (residues 13- We therefore investigated the newly published amino acid 156) and two more clear repeats in the second half (residues 223 sequence (Ginsberg et al., 1984) to see if any structural period- -276). Further analysis of the protein by the damped Needleman icity was manifested in the sequence. and Wunsch method showed that the repeats were even more Analysis of sequence TFIIIA of extensive, being partly obscured by a few gaps in the middle Sequence repeats. The amino acid sequence of TFIIIA, of the sequence. The which final best alignment (Figure 3) shows that 1610 Zinc-binding domains in Xenopus transcription factor MILA 1 8 13 17 23 26 30 T G E K *P V® D G D K K K .. K *R *. R)T 1 ( M G E K A L P V V Y K R ) 1 I ®S F A D®G A A®N K N W K®Q * A®L C * K® 37 tI) 2 T G * E K E E G©E K S L H H®T * R®S L * T 67 P®)P®)K G(®)T 3 T G E K * S D L R®T T K A N M K * K®F N R F® 98 G©)D NOTED 4 N I K I C V@V©H F E N©G K A®K K H N V®Q F * S® 129 QMK 5 Q L * P H E G(D K RS L P S K LQ P®E®C * R®E * V® 159 RMK 6 A G - - * K D D G K T W T®Y L K®V A E C® 188 S®S(®V -(1)PO)K 7 - * A D V®N R H K D K * T® 214 Q D - L Y®R * K(®)R D®Q V(C- - 8 E K E R T R D R ST T A F N®R * Q S F® 246 GC©D 1 V(®)L®P S®2 E E Q R * P®V®E H A K M K K * V® 276 9 G©G C®A S®E * R®S D P E K R K L * K E K C P R P K R S L A S R L T G Y I P P K S K E K 311 [5IA S V S G T E K T D S L V K N K P S G T E T[8 G S L V L D K L T I Q ) 344 Fig. 3. Amino acid sequence of transcription factor from X. oocytes, aligned to show the repeated units. The sequence is in one-letter code IHA laevis (Dayhoff, 1978). The molecule contains: an amino end region (residues 1-12); a lysine-rich zone (277- 309) near the carboxyl end; a short tail region which may bind carbohydrate at Asn 310 and Asn 333, which are indicated by squares. The repeat units are numbered 1-9 on the left side of the diagram. The boxed-in consensus sequence at the top shows the characteristic features of a typical repeat unit, numbered as for a length of 30 residues. The end-point of each unit has been chosen arbitrarily after His-30. The best-conserved residues are ringed. Cys-8, Cys-13, His-26 and His-30 are believed to bind to a Zn2+ ion. Tyr-6, Phe-17 and Leu-23 are hydrophobic residues which may form an inner core for the proposed multiple-domain structure. Asterisks (*) mark positions where an insertion sometimes occurs in the normal pattern, and dots (.) mark variable positions in the sequence. In the main body of the repeats a dash indicates an The are (-) alignment gap. underlined regions those which show clear evidence of a relationship with at least one other unit (see Table I). Note that units 6 and 7 have diverged considerably from the usual pattern, and that residues 277-282 may form a fragmentary extension of repeat unit 9. or conversion et It by gene duplication gene (Hood al., 1975). Table I. Overall similarities between the nine repeat units of the is not yet possible to make a proper analysis evolutionary because the end points of the units are not known; relationships, Residues Unit Similarity values if the TFIIIA these could gene possesses intervening sequences 5 9 2 8 3 4 1 7 6 be Table I a estimate of the of revealing. gives rough degree 130- 159 5 * 23 20 15 18 14 6 8 2 of each of derived from the numbers of relatedness pair units, 247-276 9 23 * 23 16 18 12 10 1 0 in the matrix. The totals of high-scoring windows comparison 38 -67 2 20 23 * 11 17 16 6 1 8 columns in the table a measure of how each unit the give closely 215-246 8 15 16 11 * 17 13 11 5 5 resembles the consensus of the nine global repeats. 68-98 3 18 18 17 17 * 6 6 6 2 It can be seen that units 9 and in that are 5, 2, order, clearly 99-129 4 14 12 16 13 6 * 16 7 0 the most members of the and are like one typical family, very 13 - 37 1 6 10 6 11 6 16 * 00 another. Common features include and Pro-5, Gly-12, Lys-15 189-214 7 8 1 1 5 6 7 0 * 1 Arg-25 (here we refer to the numbering of the consensus sequence 160-188 6 2 0 8 5 2 0 0 1 * in Figure 3). A second group comprises units 8, 3 and 4. These Column totals 106 103 102 93 90 84 28 18 contain some insertions at 4A or 28A. single-residue positions Units 8 and 3 form a as shown the ident- recognisable pair by the number of 1 The value given to each pair is overlapping 1-residue Asn-22 and ities Asp-11, Gly-12, Asp-15, Thr-18, Thr-19, windows, in the correctly aligned comparison of the two units, for which Phe-29. units 7 and 6 have diverged considerably from Lastly, 1, the score exceeds the 0.1 % matching probability level. Since each unit is 7 are shortened at the normal in units 6 and -30 residues long, the maximum possible value is 30. The column totals pattern: particular, of give a measure of how close each unit is to the global consensus them We note their amino ends with (Cys-8)-(Cys-13) loops. irregular 1 of all. The underlinings below the sequence in Figure show the parts each Ala-16 and in that units 1 and 4 have Phe-10, Gly-14, Lys-19 unit that lie in the centre of a window, where the comparison score, with at while units 1 and 6 both have common, Trp-21. least one other unit, reaches the 0. I% level. Thus although all nine units belong to the same family they have diverged in detail and may have taken on specific individual residues 13-276 form a continuous run of a of evolved. repeated motif, DNA-binding functions as the 5S gene control system nine similar each of - 30 residues. Since the It if the same 30-residue units were units, sequence would also not be surprising repeats itself cyclically, the position of the boundary of the motif in numbers in other related later found to occur varying gene is uncertain, but the choice of an end-point shortly after His-30 control proteins. of the consensus allows the largest number of complete units to The of a are evolutionary advantages repeating design probably as be assigned. With this choice, most of the insertions and deletions much the same as those in other linear such many proteins, in a occur near the ends of the units rather than the central ovomucoid inhibitor et loop (Laskowski al., 1980). Probably single DNA was once (consensus position numbers 14-25). functional unit that binds to a half-tum of evolved, Our results that residues 13-276 of TFIIIA in much and suggest strongly and then became used more subtle specialised ways from a ancestral unit of 30 amino acids when a number of similar units were in series. have evolved primitive large joined 1611 and A.D.McLachlan J.Miller, A.Klug Characteristic structural features of the small domains The well-marked repeat pattern in TFLIA suggests strongly that each of the sequence repeats corresponds to a series of small structural elements, or domains, of -30 residues, arranged linearly. In large globular proteins with repeated sequences the units most often form compact dimeric or tetrameric pseudo- symmetrical structures (Rossman and Argos, 1981; McLachlan, 1980). But there are also well-known examples of proteins with compact independently active structural units strung together in line, which can be separated by enzyme cleavage: the 'kringles' in plasminogen (Sottrup-Jensen et al., 1978) and the Kazal-type protease domains in the ovoinhibitor of Japanese quail (Laskow- ski et al., 1980; Bolognesi et al., 1982). In these molecules the structural units have lengths of 80 and 63 residues. The TFIA unit is exceptional because of its small size and unique arrange- ment of Cys Cys His His residues. scheme for a linear of Fig. Folding arrangement each repeated domains, There is strong evidence, both that the repeat unit is a self- centred on a tetrahedral of zinc arrangement ligands. Ringed residues are sufficient folded domain, and that it is stabilised the conserved amino by zinc ions. acids which include the and Cys His zinc ligands, the The first conclusion is negatively charged Asp-l and the three suggested by the appearance of 1, hydrophobic groups that may form multiple a structural core. Black circles mark the most probable DNA-binding side small fragments on proteolysis, with quantized mol. wts. which chains (see Figure In the scheme drawn here 5). the metal ion draws the are often multiples of 3 kd (see above). The second is supported ends of each unit the together, leaving central residues 14-25 to form a by evidence that one TFIIIA molecule binds many zinc ions or An potential DNA-binding loop 'finger'. alternative but much less likely (7-11, see above) and is for inactive in the absence of position the zinc is between the His residues of one the metal unit and the Cys residues of an one. adjacent (Hanas et al., 1983). We therefore believe that most, if not all, of the nine units bind zinc. 13 17 6 8 23 26 30 Zinc is normally tetrahedrally coordinated in inorganic and metallo-organic compounds. The amino acids Cys and His are 63tD its most common ligands in enzymes (see Fersht, 1977), such as carbonic anhydrase, liver alcohol dehydrogenase, carboxy- peptidase A, Cu-Zn superoxide dismutase and thermolysin. Therefore we may picture each small domain as a compact unit formed round a central zinc ion to confer stability. Each zinc BASI I r nnIJ would be coordinated tetrahedrally to the two invariant pairs of Cys and His residues in each unit. The ends of the domain will then be pulled together round the zinc. In Figure 4 we have drawn ACIDICnl one possible arrangement of the zinc ligands which fits this type of scheme. It is important to remember that the Cys . . Cys and His . His loops probably cross over at an angle to form HYDROPHOBTC a tetrahedral box, and that the zinc may IFhI have more than four ligands. For example, a Cys side chain might be shared between ,,_r_ L Fh n L , F1 n FIr zincs in two adjacent units. Figure 5 shows how each domain contains three invariant Fig. 5. Histograms for the average distribution of amino acids along the hydrophobic groups (normally Tyr-6, Phe-17 and Leu-23) length of each 30-residue repeat unit. The height of each in the bar, range and how the acidic side chains are nearly all in or near the the number of 0-9, gives times that class of amino acid occurs at each (Cys-8) . . . (Cys-13) position in the nine loop. units. Positions are numbered as in Figure 3, with Cys-8, Cys-13, His-26 and His-30 treated as special The extended protein loop positions (zinc between residues 14 and ligands). 25 might Amino acids have been classed as follows. (a) DNA-binding: Lys, Arg, form a DNA-binding region. The three-dimensional structures Thr His, Asn, Gln, (Ohlendorf and Matthews, 1983). (b) Basic: Arg, Lys, of gene repressor proteins (Ohlendorf and Matthews, 1983) show His. (c) Acidic: Asp, Glu. (d) Large hydrophobic: Leu, Ile, Val, Phe, Tyr, that the side chains of Lys, His, Asn, Gln and Thr often Trp. The strongest potential interact DNA-binding sites, with five or more DNA- with binding amino acids in the nine the phosphate backbone units, are also marked in Figure of DNA, while Arg can form 4. hydro- gen bonds to the base pairs. Amino acids of this type are con- centrated in the region 14-25 of TFIIIA (Figure 4). instead built into a 'metal-sulphide' pseudocrystal bordered by A structural parallel for our proposed 30-residue metal-centered the polypeptide chain. domain might be the 40-residue calcium-binding unit of the cal- TFIIIA thus appears to have an exceptionally compact mol- modulin family (Moews and Kretsinger, 1975) or the 26-residue ecular architecture for a self-sufficient structure; however, it may ion-sulphur half-domain of bacterial ferredoxin (Adman et al., be that strong interactions between adjacent units are important 1973). However, neither of these structures is an independent for maintaining the structure. We have examined the two-helix folding unit, but forms part of a linked pair of fragments. The DNA-binding motif of the Cro and Lac repressors, but it does metallothionein family of proteins (Boulanger et al., 1983), which not seem possible to accommodate the TFHIA consensus se- contain repeated Cys residues, are also not a good analogy for quence into this model: the loop 14-25 may instead fold into TFIIIA, but a rather peculiar group. The reason is that they bind a long twisted ribbon of ,3-sheet which wraps in some way round metal ions in clusters, so that the the bridged by Cys residues, ions zinc-binding pocket, using the invariant hydrophobic groups are not separately packaged in independent protein cages, but as nuclei. The characteristic Cys-Cys-His-His consensus motif 1612 Zinc-binding domains in Xenopus transcription factor LiHU 5'Non- 2). As pointed out by them, the small fragment CB2 stains quite Coding differently with silver from the two other fragments. We there- RNA Pol Internal Control Other factors ? ,,, fore identify the fragments CB3, CB1 and CB2 with the resi- DNA Upstream dues 1-90, 91- 266 and 267- 344, respectively, since the latter is of different character from the first two (particularly in its high content of lysine). The lower mobility of the shorter peptide CB3 (78 residues) relative to CB3 (90 residues) could be accounted for if some of its residues were modified, as suggested above. The same The question of modifications is being investigated. found the amino-carboxyl orientation (Figure 6) has been by Carnegie group (D. D.Brown, personal communication). Carboxyl rich Amino Discussion Carbohydrate- binding above have The experimental and theoretical studies described lIIA consists led to a picture in which the transcription factor Fig. 6. An interpretation of the structural features of the protein TFIIIA and each mostly of a linear arrangement of nine repeated domains, its interactions with DNA. The DNA is drawn curved, as if resting on a on of zinc The centred a tetrahedral arrangement ligands. repeats long beaded surface of the protein. The internal control region of the 5S RNA gene (bases 40- 100) is drawn as six turns of DNA, with the 5' end are thought to be extended DNA-binding fingers, linked by (flex- of of the non-coding strand at the left. A separate piece upstream DNA may with the ible?) joints. This model is consistent highly asymmetric be from required to promote transcription. The protein sequence runs right of the molecule indicated its shape by physico-chemical proper- to left: the evidence for this orientation is in the text. The amino end given A structure of this kind would ties (Bieker and Roeder, 1984). is followed units 1 in contact with the by nine repeat (residues -276) how a small of 40 K can bind to a control These units form the 30-K of explain relatively protein long region. together proteolytic fragment Smith et al. The first six units 13- (1984). repeat (residues 188) probably of DNA: if each domain interacted with stretch double-helical constitute the 20-K in the proteolytic fragment: irregularities repeat pattern half a DNA for the nine about period, then, allowing end effects, of and 7 to a between units 6 may correspond susceptible cleavage region could cover - 50-60 the of the domains nucleotides, length the two The are to be extended fragments. repeats thought DNA-binding internal control et The DNA or RNA region (Sakonju al., 1980). linked each a zinc ion centre. Residues 'fingers' by flexible joints, having 277- 344 include a followed two of each domain could be and lysine-rich region, by potential binding strength modulated, specific near the end with filled carbohydrate-binding sequences carboxyl (marked recognition of short nucleic acid stretches established, by vari- and these form a the 10-K triangles), together parts separate domain, ations in the at the sequence finger tips. proteolytic fragment. The relationship between the grooves of the DNA and One of for the advantage a many-fingered design transcrip- the of the units is unknown and the must not be positions protein drawing of the taken but each unit binds to about one half-turn of DNA. The tion factor is that it binds to an internal control region gene literally, tip of a unit could fit into a groove. in a where a stable is formed system transcriptional complex et Lassar (Bogenhagen al., 1982; Gottesfeld and Bloomer, 1982; found in TFIHIA appears to be unique to this protein. We have et This al., 1983). complex can sustain many rounds of transcrip- searched for it without success in a large number of zinc enzymes, tion which the factor to during presumably remains bound the including those mentioned above, as well as in the metallothio- As the gene. polymerase passes through the gene, the many- neins. There is, therefore, no evidence yet that this motif, or fingered protein could release those fingers bound ahead of the indeed any other repetitive pattern, is a typical feature of zinc- but its processing polymerase, stay bound by remaining fingers, as a class. binding proteins whether to the DNA or intact double helix to the non-coding Relation oflarge domains to the protein sequence. The nine small strand. of domains containing repeated pairs cysteines and histidines com- We have already mentioned above that the transcription fac- prise residues -276 or 1-280 (depending where the boun- tor to be a evolved version of a appears highly small molecule at the dary is taken). The remaining 70 residues carboxyl end of the size of one of the 3-K domains contemporary stabilised units The se- have no homology with the repeating (Figure 3). a metal ion. The molecule could have assisted by primitive simply this terminal be of quence suggests that region might composed an form of Evolution to the elaborate tran- early transcription. - is rich in two parts. The first half of 30 residues very lysines found could have taken scription apparatus today place by gene in this a no and arginines, resembling respect histone, although with different extra functions. It duplication, repeats taking up The second half lacks is as remarked Hanas et al. that significant sequence homology emerges. noteworthy, by (1983) RNA as has been out to us Dr H-C. this enrichment and, pointed by m contains and it could be that the initial polymerase zinc, activity sites for addition Th0gersen, contains two potential carbohydrate of primitive TFIIIA promoted transcription in the absence of a characteristic is N X S for at 310 and 333. which evolved later aspargines [The sequence polymerase molecule, presumably only (Marshall, 1972).] greater efficiency. in character of the last 70 residues et al. a between a The marked difference sug- Ginsberg (1984) have noted homology region this be identified with the 10-K 'domain' of DNA and the of the TFIIA gest that region might 5S (or RNA) coding sequence one end of the the studies of have TFIIHA could interact with its own revealed at protein by proteolysis gene, and suggested that This 10-K domain is for efficient or the derived RNA to at the tran- Smith et al. (1984). required gene autoregulate expression RNA unlike the 30-K or translational levels. The of this transcription, but, remaining fragment (and scriptional evolutionary origin 20-K not for to the internal con- be as follows. RNA is be- its smaller subfragment), binding contemporary property may widely amino and orien- a small RNA trol of the RNA The lieved to have so that region 5S gene. carboxyl preceded DNA, primitive would then be as shown of the 3-K tation of the relative to the could have coded for a small protein gene protein (the precursor RNA in Figure 6. repeating unit) which bound back to the and so stimulated with the bromide This orientation is consistent its own cyanogen production. shown Smith et al. the deletions in units 6 and 7 of the of the It be that cleavage map protein by (1984, Figure may contemporary 1613 J.Miller, A.D.McLachlan and A.Klug with distance from the centre of each window. This newer method is more suitable are to enhance its and protein (Figure 3) necessary flexibility for with dealing gaps, but is less susceptible to statistical analysis. enable it to accommodate to the or structure secondary tertiary In the comparison matrices the highest observed score with a window of 30 of the 5S as well as to the DNA. RNA, to a double corresponded matching probability for the two peptide segments of The 10-K domain of Smith et al. does not bind to the (1984) 0.53 x 10-10 only (McLachlan, 1971) and was highly significant. We also calcu- lated the score for the of two random internal control of the and we have identified it with highest expected comparison 344-residue region gene the score which would achieved sequences, i.e., be on average just once with the 70 residues at the C-terminal end of the Smith et protein. our protein. This score was exceeded not but 359 times in the natural se- once, al. have that this domain (1984) suggested 'transcription' may quence and showed that there are many significant repetitions. interact with other factors to form a tran- required competent In the damped Needleman-Wunsch method we used an effective range of 30 RNA or with Ill. To this residues with rather low 2.0 to start each and 2.0 for scription complex directly polymerase gap penalties: gap each extension by one residue. we would add in view of its and only that, high lysine arginine it also be in involved to elements content, might binding upstream Acknowledgements of the DNA outside the of the coding region gene. We thank Dr G.L.Everett (Johnson-Matthey Ltd., Royson) for determinations of the content of zinc and other metals, and Professor N.Hales and Dr G.Maguire Materials and methods (Department of Clinical Biochemistry, Cambridge University) for zinc determin- 7S ations. We are grateful to Drs D.D.Brown and H.Pelham for advice, and our particle punfication colleagues Drs D.Rhodes and H-C.Th0gersen for their help. 7S RNP were from the ovaries of immature X. laevis particles prepared (South African Snake a method to be described elsewhere in Farm) by (Miller, prep- The method References does not the to such as aration). subject particle strong chelators, or to salt concentrations above 0.07 M. Protein is at least 95% DTT, pure by Adman,E.T., Sieker,L.C. and Jensen,L.H. (1973) J. Biol. Chem., 248, 3987- Coomassie or silver and >90% amino acid not stain, by analysis (data shown). The is no more than 5% dissociated as measured elec- particle by agarose gel Bieker,J.J., Martin,P.L. and Roeder,R.G. (1985) Cell, 40, 119-127. and contains no detectable free 5S RNA. trophoresis (see below) usually Bieker,J.J. and Roeder,R.G. (1984) J. Biol. Chem., 259, 6158-6164. For of metal and item to which quantitation content, glassware, tubing any Bogenhagen,D.F., Sakonju,S. and Brown,D.D. (1980) Cell, 19, 27-35. buffer was was treated for at least 48 h in 30% nitric washed exten- exposed acid, Bogenhagen,D.F., Wormington,W.M. and Brown,D.D. (1982) Cell, 28,413-421. with water and for h then treated at least 72 in 10 mM sively glass-distilled Na2 Bolognesi,M., Gatti,G., Menegatti,E., Guarneri,M., Marquart,M., Papamokos, EDTA the for EDTA (pH 4), optimum pH Zn2+ complexation by (West, 1969). E. and Huber,R. (1982) J. Mol. Biol., 162, 839-868. items not resistant to nitric column such as matrices or Any acid, dialysis tubing, Boswell,D.R. and McLachlan,A.D. (1984) Nucleic Acids Res., 12, 457-464. were rinsed in several of 10 mM EDTA 72 h. for at least changes Na2 (pH 4) Boulanger,Y., Goodman,C.M., Forte,C.P., Fesik,S.W. and Armitage,J.M. (1983) All materials were washed with water before use. Buffers were glass-distilled Proc. Natl. Acad. Sci. USA, 80, 1501-1505. a Chelex 100 column before use passed through ion-exchange (BioRad) directly 359-365. Brown,D.D. (1984) Cell, 37, to remove divalent cations. Protein concentration was measured amino acid by Cornell,N.W. and Crivaro,K.E. (1972) Anal. Biochem., 47, 203-208. analysis. Dayhoff,M.0. (1978) Atlas of Protein Sequence and Structure, Vol. 5, suppl. Particle dissociation 3, published by National Biomedical Research Foundation, Washington D.C. assay in Denis,H. and le Maire,M. (1983) Roodyn,D.B. (ed.), Subcellular Biochemistry, 7S at concentrations of 100 were loaded with particle preparations, usually ytg/ml, Vol. 9, Plenum Publishing Co., pp. 263-297. an volume of 50% 50 mM 0.1% equal glycerol, Tris-Cl, pH 7.5, xylene cyanol and Engelke,D.R., Ng,S.Y., Shastry,B.S. Roeder,R.G. (1980) Cell, 19, 717-728. onto 0.7% in 50 mM 90 mM boric acid 8.5 and run agarose gels Tris-Cl, pH Structure and Fersht,A. (1977) Enzyme Mechanism, published by W.H.Freeman for - 1 h at 8 V/cm. Gels were stained with 1 ethidium bromide and mg/ml & San Co., Francisco. visualised under u.v. have confirmed that the fraction of RNA in Experiments and Ginsberg,A.M., King,B.O. Roeder,R.G. (1984) Cell, 39, 479489. with free 5S RNA correlates with particle preparations co-migrating qualitatively and Gottesfeld,J. Bloomer,L.S. (1982) Cell, 28, 781-791. the fraction of 5S RNA in a from 7S in filtration eluting separate peak particle gel Hanas,J.S., Hazuda,D.J., Bogenhagen,D.F., Wu,F.Y.H. and Wu,C.W. (1983) that it be taken as a measure of (Miller, unpublished results), suggesting may J. Biol. 14 120-14 125. Chem., 258, free and bound 5S RNA in solution. and 119-126. Honda,B.M. Roeder,R.G. (1980) Cell, 22, Polyacrylamide gel electrophoresis and Annu. Rev. Hood,L.M., Campbell,J.H. Elgin,S.C.R. (1975) Genet., 9, 305- was as described Laemmli with the modifi- Electrophoresis by (1970) following cations. was 3% 0.15% 0.125 M Tris- 680-685. Stacking gel acrylamide, bis-acrylamide, Laemmli,U.K.(1970) Nature, 227, 0.1% 0.01% 0.1% ammonium and P04, pH 6.8, SDS, TEMED, persulphate. Running Laskowski,M., Kato,I., Kohr,W.J., March,C.J. Bodard,W.C. (1980) in was 22.5% 0.73% 0.75 Tris-CI 0.1% 0.01% Protides Vol. gel acrylamide, bis, pH 8.8, SDS, Peeters,H. (ed.), of Biological Fluids, 28, Pergamon Press, 0.1% ammonium were diluted with an 123-128. TEMED, persulphate. Samples equal Oxford, pp. volume of mM mM 100 NaPO4 2% 100 DTT and 50% and pH 6.8, SDS, glycerol Lassar,A.B., Martin,P.L. Roeder,R.G. (1983)Science, 222, 740-748. and boiled for 3 before mol. wts. were found min and directly loading. Approximate Lewis,C.D. Laemmli,U.K. (1982) Cell, 29, 171-181. from a of the of mol. wts. two tobacco logarithmic plot mobility against markers, J. Mol. McLachlan,A.D. (1971) Biol., 61, 409-421. mosaic virus and bovine inhibitor Gels in Protein protein (TMV) pancreatic trypsin (BPTI). McLachlan,A.D. (1980) Jaenicke,R. (ed.), Folding, Elsevier/North were stained with 0.5% PAGE blue 83 (BDH Biochemicals). 79-96. Holland, Amsterdam, pp. J. Mol. 15-30. McLachlan,A.D. (1983) Biol., 169, Proteolysis Annu. Rev. Marshall,R.D. (1972) Biochem., 41, 673-702. was at conducted concentrations of 200-500 Proteolysis always particle mg/ml and J. Mol. 201-225. Moews,P.C. Kretsinger,R.H. (1975) Biol., 91, in 50 mM Tris-CI 50 mM 0.5 mM with 20 pH 8.1, KCI, MgCI2, tryp- JAg/mIl and J. Mol. 443453. Needleman,S.B. Wunsch,C.D. (1970) Biol., 48, room Reactions were either with 2 mM sin, temperature. stopped benzamidine, and Annu. Rev. Ohlendorf,D.H. Matthews,B.W. (1983) Biophys. Bioeng., 12, or for 2 in buffer. min by boiling loading 259-284. methods Sequence repeats of analysis and Pelham,H.R.B. Brown,D.D. (1980) Proc. Natl. Acad. Sci. USA, 77, 4170- We used two well-established In methods. the the matrix 4174. first, diagonal comparison the is divided into all its and (McLachlan, 1971, 1983), sequence possible overlapping Picard,B. Wegnez,M. (1979) Proc. Natl. Acad. Sci. USA, 76, 241-245. or of a fixed without insertions or and segments, windows, given length, deletions, Rossmann,M.G. Argos,P.W. (1981) Annu. Rev. Biochem., 50, 497-532. and of is The score for every pair segments compared independently. comparing Sakonju,S., Bogenhagen,D.F. and Brown,D.D. (1980) Cell, 19, 13-25. each of amino acids in the window is derived from mutation likeli- and J. pair Dayhoffs Segall,J., Matsui,T. Roeder,R.G. (1980) Biol. Chem., 255, 11986-11991. hood tables and is assessed the exact and (Dayhoff, 1978) against calculated-prob- Smith,D.R., Jackson,T.J. Brown,D.D. (1984) Cell, 37, 645-652. distribution of the window scores for random with the same ability sequences Sottrup-Jensen,L., Claeys,H., Zajdel,M., Petersen,T.E. and Magnusson,S. (1978) as the whole The cDNA se- Prog. Chem. Fibrinolysis Thrombolysis, 3, 191-209. average composition protein (McLachlan, 1983). were scored identical bases. In the second the quences by counting method, damped West,T.S. (1969) Complexometry with EDTA and Related Reagents, published Needleman-Wunsch alignment with gaps (Boswell and McLachlan, 1984; Needle- by BDH Chemicals Ltd., Poole, UK. and are man with for insertions Wunsch, 1970), sequences aligned locally penalties and but the scores are which die Received on 4 deletions, 1985 given weights away exponentially April

Journal

The EMBO Journal – Springer Journals

Published: Jun 1, 1985

There are no references for this article.