Sep 7, 1995 - between the catalytic domain structures of the biotin synthetase/ repressor ... there is no available information about a specific PheRS-DNA.
Protein Science (1995), 4:2429-2432. Cambridge University Press. Printed in the USA.
Copyright 0 1995 The Protein Society
Structural similarities in the noncatalytic domains of phenylalanyl-tRNA and biotin synthetases
MARK SAFRO
AND
LIDIA MOSYAK
Department of Structural Biology, Weizmann Institute of Science, Rehovot 76100, Israel
(RECEIVED July 12, 1995; ACCEPTED September 7, 1995)
Abstract: Detailed comparison between the structures of the Escherichia coli biotin synthetasehepressor protein (BirA) and the recently solved Thermus thermophilus phenylalanyl-tRNA synthetase (PheRS)reveals significant similarities outside their respective catalytic domains. These comprise a DNA-binding a+0 domain and an Src-homology 3 (SH3)-like domain that were observed in both enzymes. This similarity provides a novel example in which all domains of one multidomain protein appear to be constituents of the other multidomain protein and supports aconcept of a common ancestor for two different synthetase families. Keywords: aminoacyl-tRNA synthetases; biotin synthetasehepressor protein; DNA-binding motif; transcription regulation mechanism; X-ray crystallography ~~
~~
~~~~~~~
~
~~~
~
~~~~
Recently, Artymiuk et al. (1994) have reported a striking similarity between the catalytic domain structures of the biotin synthetase/ repressor protein (BirA) (Wilson et al., 1992) and of seryl-tRNA synthetase (SerRS) (Cusack et al., 1990). The X-ray structure of another representative of the class I1 aminoacyl-tRNA synthetases (aaRSs) (Erianiet ai., 1990), phenylalanine-tRNA synthetase (PheRS) fromThermus thermophilus, has recently been solved at 2.9A resolution (Mosyak et al.,1995). Presented here are theresults of a detailed comparison of the BirA and PheRS structures that reveals that the two enzymes, apart from close similarity of their catalyticdomains, sharealso two noncatalytic domains: a DNA-binding domain containing a helix-turn-helix (HTH) motif and abarrel-like domain resembling an Src-homology 3 (SH3) fold. PheRS is a tetramer of ( ( ~ 0type ) ~ with 350 amino acid residues per a-subunit and 785 residues per 0-subunit. Unlike the catalytic a-subunit (domains A1 and A2), the 0-subunit (domains B1-B8) has no invariant functional amino acids directly involved in the aminoacylation process. Nevertheless, it turned out that both the catalytic a-subunit and two domains of the ~~~~
~~
~
~~~
~
~~~
Reprint requests to: Mark Safro, Department of Structural Biology, Weizmann Institute of Science, Rehovot 76100, Israel; e-mail: cssafro@ weizmann.weizmann.ac.il.
0-subunit (B6 and B7) display the typical fold of class I1 aaRSs, observed in the structures of SerRS (Cusack etai., 1990), AspRS (Ruff et al., 1991), and LysRS (Onesti et al., 1995). Moreover, the PheRS @subunit shares, besides the catalytic-like, two more common folds with other class I1 aaRSs (Mosyak et al., 1995). This suggests that the 0-subunit, having no catalytic function, has evolved from the same ancestor asall class I1 synthetases. This, together with the finding of Artymiuket al. (1994), encouraged us to search for further structuralsimilarities between BirA and the multidomain 0-subunit of PheRS. We used the structural alignment program STRUCTAL-E (Subbiah et al., 1993) to compare thecoordinates of PheRS with those of BirA (PDB code lbia). The B5 domain of PheRS (residues 407-473) and the N-terminal domain of BirA (BirAI, residues 1-60) display similar architecture, with a common a+P fold formed by three a-helices packed against a three-stranded antiparallel 0-sheet (Fig. IA and Kinemage I). The differencein topological organizationof the domains is related to the pointsof entry and exit: helix D is at the C-terminus in B5 instead of being at the N-terminus (helix A) as in BirAI. For the superpositionof 43 equivalent C,? pairs, the RMS deviation is 2.26 A. Two nearly perpendicular a-helices (B and C) separatedby a short turn constitute the HTH motif, characteristic of many DNA-binding proteins (reviewed by Brennan, 1992). In PheRS, the HTH motifis exposed to the region located in the vicinity of the twofold axis, where the two heterodimers form a wide saddle. In fact, B5 is the sixth example of an a+p fold previously observed in the DNA-interacting proteins (McKay & Steitz, 1981; Wilson et al., 1992; Clark et al., 1993; Ramakrishnan et al., 1993; Lima et al., 1994), even though there is no available information about aspecific PheRS-DNA interaction. Seven strands andthree a-helices (Fig. 1 B andKinemage 3) constitute a common structuralmotif of the PheRS “catalytic-like” module (CLM, B6 and B7) and theBirA biotin-binding domain (BirAII). The superposition of CLM on BirAII is characterized by RMS deviation of 2.55 A for 83 equivalent C, atoms. Topologically, both CLM (482-679) and BirAII (68-269) are located downstream of their DNA-binding folds. B5 and CLM are connected by an extended polypeptide segment(1 5 residues) and separated in space by the catalytic module (CAM), whereas the
2429
2430
M. Safro and L . Mosyak
A DNA-binding fold ofPheRS
DNA-binding fold ofBirA
6 Catalytic-like module of PheRS
Biotin-binding domain of BirA
4
C
SH3-like foldof PheRS
SH3-like fold of BirA
Fig. 1. MOLSCRIPT (Kraulis, 1991) drawings of the equivalent folds observed in the structuresof PheRS (left) and BirA (right). The common secondary structure elements are shown as sequentially labeled ribbons. A: B5 and BirAl domains displaying DNAbinding a+@ topology. The C-terminal connection of helix D to the rest of the fold is also observed in the two cu+B motifs of DNA topoisomerase I structure (Limaet al., 1994). R: CLM and BirAll domain,both characterized by a seven-stranded &sheet surrounded by three cu-helices. The 12-residue segment connecting strand 3 andhelix C in CLM is omitted for clarity.C: B4 and BirAlll domains with SH3-like topology. Dashed lines emphasize different exit pathways of the domains.
243 1
Noncatal.vtic domains of phenylalanyl-tRNA and biotin synthetases
Fig. 2. Stereo view showing the superposition of BirAII (residues 68-21 I ) on CLM (480-679), together with the DNA-binding domains (residues 1-60 and 403-474, respectively). The domainsare in gray (BirA) and black (PheRS).BirAI and BS domains occupy roughly similar relative positions in their structures, but with different orientations (they can be superimposed after rotation of about 90" around the axis perpendicular to the plane of the drawing and passing between the two domains). Drawings generated by MOLSCRIPT (Kraulis, 1991).
association of BirAl and BirAll is tighter and the corresponding linker is five residues shorter. Theoverall topological relationship of the DNA-binding domain plus the biotin-binding domain of BirA together with the respective counterparts in the PheRS structure (B5 and CLM) is shown in Figure 2 and Kinemage 3. The C-terminal, SH3-like domain of BirA (BirAIII, 272-317) and the B4 domain of PheRS (271-320) display a similar barrellike structure, except that strand 6 is missing in PheRS (Fig. IC and Kinemage 2). Although strand 1 in B4 is broken by a loop insertion into two parts (strands0 and I), the domain fit is fairly good: 26 equivalent C,, pairs can be superimposed with an RMS deviation of 2.5 A. B4 is the insertion domain between the two flafl motifs of a larger structural fragment 8 3 (Mosyak et al., 1995). The architecture of B3 dictates the proximity of the Nand C-termini of B4, which in turn modifies the exit pathway of B4 (lack of strand 6) as compared to the classical SH3 fold. Structure-derived sequence alignment of BirA and PheRS shows noclear sequence homology. Nevertheless, the fact that all three domainsof BirA appeared to have counterpartsin the PheRS structure supportsa concept of a common ancestor for two different synthetase families. Structural relationships between the catalytic domainsof BirA and theclass l l aaRSs (Artymiuk et al., 1994) correlate with the well-known mechanistic similarities in their operation modes (Eisenberg, 1973). But what might be the functional reason of the conserved features related with noncatalytic domains? SH3 domains area common feature of a number of eukaryotic proteins involved in signal transduction pathways (Koch et al., 1991), whereas BirAlII and B4 are constituents of two prokaryotic synthetases. This may bea clue for understanding yet-uncertain functions of these domainsin the proteins of the prokaryotic kingdom. An attractive explanation for the DNA-binding domainsimilarities is to assume that PheRSserves as a transcriptional regulator by drawing analogy with the BirA repressor function (Eisenberg et al., 1982). In such a transcription regulation mechanism, aminoacylated-tRNAPhe might play a role similar to that of already-biotinylated biotin-accepting proteins. Depending on the concentration level of the uncharged tRNAPhe,
PheRS may regulate its own transcription by binding to DNA with twofold related helices of domain B5. In light of the regulatory rolesuggested for aaRSs, only one direct indication is available at thepresent time ofa similar regulation mechanism observed for AlaRS. As was shown by Putney and Schimmel (1981), in vitro, AlaRS binds to a specific DNA sequence and represses transcription of its own gene inthe presence of alanine. About the sametime, Springer et al. (1983) showed that the expression of the Escherichia coli PheRS operon has manyof the characteristic featuresof an attenuationcontrolled operon, whereas no attenuator-like sequences were found in the region of PheRS operon in 7: thermophilus(Kreutzer et al., 1992). Therefore, it remains to be seen whether the repression regulatory mechanism for 7: thermophilus PheRS can be considered as an alternative one. Further biochemical and crystallographic studies of PheRS andBirAfunctionalcomplexes will berequiredtoclarify whether or not the additional structural relationshipswe have found can be extended to their noncatalytic functions.
Acknowledgments
We thank B. Matthews for reading the manuscript. Thiswork was supported by grant from the Basic Research Foundation of Israel and by Kimmelman Center for Biomolecular Structure and Assemble. References Artymiuk PJ. Rice DW. Poirette AR, Willet P. 1994.A tail of two synthetases. Nature SIrucl Biol 1:758-760. Brennan RG.1992.DNA recognition by the helix-turn-helix motif. Curr Opin Struct Biol2 :1 0 0 - 108. Clark KL, Halay ED, Lai E, Burley SK. 1993.Co-crystal structure of the HNF-3/fork head DNA-recognition motif resembles histone H5.Nature 364:412-420. Cusack S , Berthet-ColominasG, Hartlein M, Nassar N, Leberman R. 1990. A second class of synthetase structure revealed by X-ray analysis of E. coli seryl-tRNA synthetase at 2.5 A. Nature 347:249-256. Eisenberg MA. 1973. Biotin: Biogenesis, transport. and their regulation. Adv Enzymol38:317-372. Eisenberg MA, Prakash 0,Hsiung SC. 1982. Purification and properties
2432 of the biotin repressor-A bifunctional protein. JBiolChem257:1516715173. Eriani G , Delarue M, Poch 0, Gangloff G, Moras D. 1990. Partition of aminoacyl-tRNA synthetases in two classeson the basis of mutually exclusive sets of sequence motifs. Nature 347:203-206. Koch CA, Anderson D, Moran MF, Ellis C, Pawson T. 1991. SH2 and SH3 domains: Elements that control interactions of cytoplasmic signaling proteins. Science 252:668-674. Kraulis P. 1991. MOLSCRIPT: A program to produce both detailed and schematic plots of protein structures. J Appl Crysrallogr 24:946-950. Kreutzer R, Kruft V, Bobkova EV, Lavrik 01, Sprinzl M. 1992. Structure of the phenylalanyl-tRNA synthetase genes from Thermus rhermophilus HB8 and their expression in Escherichia coli. Nucleic Acids Res20:41734178. Lima CD, Wang JC, Mondragon A. 1994. Three-dimensional structure of the 67K N-terminal fragment of E. coli DNA topoisomerase I . Nature 367:138-146. McKay DB, Steitz TA. 1981. Structure of catabolite gene activator protein at 2.9 Aresolutionsuggestsbinding t o left-handed B-DNA. Nature 250:744-749. Mosyak L, Reshetnikova L, Goldgur Y, Delarue M, Safro M. 1995. Structure of phenylalanyl-tRNA synthetase from Thermus therrnophilus. Nature Struct Biol2:537-547.
M . Safro and L . Mosyak Onesti S, Miller AD, Brick P. 1995. The crystal structure of the lysyl-tRNA synthetase (LysU) from Escherichia coli. Structure 3:163-176. Putney SD, Schimmel P. 1981. An aminoacyl tRNA synthetase binds to a Nature specific DNAsequenceandregulatesitsgenetranscription. 251:632-635. Ramakrishnan V, Finch JT, Graziano V, Lee PL, Sweet RM. 1993. Crystal structure of globular domainof histone H5 and its implications for nucleosome binding. Nufure 362:219-223. Ruff M, Krishnaswamy S, Boeglin M, Poterszman A, Mitschler A, Podjarny A, Rees B, Thierry JC, Morras D. 1991. Class I1 aminoacyl-tRNA synthetases: Crystal structure of yeast aspartyl-tRNA synthetase complexed with tRNAAyp.Science 252:1682-1689. Subbiah S, Laurents D, Levitt M. 1993. Structural similarity of DNA-binding domains of bacteriophage repressor and the globin core. Curr Biol 3 : 141-148. Springer M, Trudel M, Graffe M, Plumbridge JA, Fayat G, Mayaux JF, Sacerdot C , Blanquet S, Grunberg-Monago M. 1983. Escherichia coli phenylalanyl-tRNA synthetase operon is controlled by attenuation in vivo. J Mol Biol 17/:263-279. Wilson KP, Shewchuk LM, Brennan RG, Otsuka AJ, MatthewsBW. 1992. Escherichia coli biotin holoenzyme synthetase/bio repressorcrystal structure delineates the biotin- and DNA-binding domains.Proc Nut/ Acud Sci USA 85:9257-9261.
