3

Protein± interactions Editorial overview Aneel K Aggarwal and Jennifer A Doudna

Current Opinion in Structural 2003, 13:3±5

0959-440X/03/$± see front matter ß 2003 Elsevier Science Ltd. All rights reserved.

DOI 10.1016/S0959-440X(03)00004-6

Aneel K Aggarwal Abbreviations w pseudouridine Department of Physiology and , AFM atomic force Mount Sinai School of Medicine, New York, NCP nucleosome core particle NY 10029, USA RNAP RNA polymerase e-mail: [email protected] SRP signal recognition particle TGT tRNA guanine transglycosylase The Aggarwal group studies the structures of enzymes and An extraordinary range of ±nucleic acid interactions negotiates the transcriptional/translational ¯ow of genomic information between DNA, RNA and protein. The reviews regulators, and their interactions with in this section of Current Opinion in cover the spectrum of nucleic acids. expression from DNA packaging, unwinding and repair, mechanisms and regulation of RNA synthesis and post-transcriptional modi®cation, the Jennifer A Doudna initiation of protein synthesis to the targeting of nascent polypeptides for secretion or membrane insertion. A recurrent theme in these reviews is the Department of Molecular and Biology, interplay between structure and the dynamics of macromolecular assembly, Department of Chemistry, Howard Hughes catalysis and regulation. Medical Institute, University of California, Berkeley, CA 94720, USA e-mail: [email protected] Akey and Luger, in the ®rst review, highlight advances in our structural understanding of the packaging of vast linear sequences of DNA into The Doudna laboratory studies the compact units: the nucleosomes. The nucleosome core particle (NCP), molecular structure and of consisting of 146 base pairs of DNA wound twice around an octameric core RNA, focusing on ribozymes and the of four histone , is the fundamental building blockof packaged large involved in protein DNA (or chromatin) in eukaryotic cells. The 1997 atomic structure of synthesis initiation and protein NCP, a milestone in the study of chromatin structure and function, traf®cking. provided the starting point for more recent high-resolution structures, including the NCP containing histone variants that differ subtly in amino acid sequence. Akey and Luger raise the intriguing possibility that these histone variants may modulate the accessibility of the genome to tran- scription factors. These NCP structures have also set the stage for inves- tigating the roles of chaperones and chromatin remodeling complexes in organizing chromatin. In this context, the authors highlight the recent crystal structure(s) of oligomeric nucleoplasmin, a histone chaperone. The structure suggests a model for histone storage during oogenesis and foreshadows future workon other chaperones that regulate nucleosome assembly.

In the second review, we are aptly reminded that ``DNA is not a static and unchanging molecule'', but one that is continually modi®ed by specialized enzymes such as recombinases, helicases and topoisomerases, among many others. MondragoÂn and colleagues offer a tantalizing lookat the mechanisms of some of these enzymes, with an emphasis on intermediates that catch these DNA `manipulators' at different stages of their catalytic pathways. Multiple structures of Cre±LoxP, for instance, now provide a fairly complete picture of the sequence of events underlying site-speci®c recombination, including the order in which the DNA strands are cleaved, exchanged and www.current-opinion.com Current Opinion in Structural Biology 2003, 13:3±5 4 Protein±nucleic acid interactions

reconnected. Structures of transposase Tn5 in both pre- À10 and À35 elements of the promoter. Together, the synaptic and postsynaptic DNA complexes are now avail- structures help to map the progression of bacterial tran- able, and support a two-metal-ion mechanism of catalysis. scription, from initial promoter recognition, melting of The recent structure of Escherichia coli DNA topoisome- the DNA, the beginning and end of abortive initiation, rase III bound to single-stranded DNA frames an impor- promoter escape, to the ®nal release of the s subunit. In tant intermediate during the topological rearrangement of addition, despite the structural complexity of the multi- DNA. The authors point out similarities between the subunit RNAPs, the authors note intriguing similarities to different enzymes, best exempli®ed by the recent struc- the simpler single-subunit phage RNAPs, including chan- ture of Archaeoglobus fulgidus reverse gyrase, which con- nels for nucleotide entry and protein elements that block tains both helicase and topoisomerase domains. These the elongating RNA product. DNA manipulations are the stuff of , preparing DNA for all manner of cellular processes, including transcrip- The recruitment of RNAP to a promoter is regulated by a tion and replication. mix of transcription factors commonly bound to upstream DNA elements. The stereospeci®c assembly of these The discovery of E. coli DNA polymerase I (PolI) in 1956 factors in response to developmental/extracellular cues was followed over the years by the discovery of PolII and is central to survival. Ogata, Sato and Tahirov review PolIII in prokaryotes, and Pola±e in eukaryotes. Remark- recent structures of Runx1±CBFb±DNA and c-Myb±C/ ably, in just the past four years, the number of known EBPb±DNA complexes that offer insights into cases in DNA polymerases has more than doubled with the dis- which one transcription factor is DNA bound (Runx1) and covery of the Y-family. These newly discovered poly- the other is not (CBFb), and in which two transcription merases allow cells to cope with unrepaired DNA damage factors are bound to widely separated DNA sites. We by promoting replication through lesions that would learn that the binding of CBFb does not alter the struc- otherwise stall the replication fork. In her review, Yang ture of Runx1 per se, but stabilizes its conformation for outlines the main structural features of Y-family DNA optimal DNA recognition. c-Myb and C/EBPb interact polymerases, as derived from recent crystal structures of via their DNA-binding domains attached to distant sites, archaeal Dbh and Dpo4, and eukaryotic PolZ. The struc- with the intervening DNA looped out, as visualized by tures reveal an architecture consisting of the palm, ®ngers atomic force microscopy (AFM). A striking AFM image is and thumb domains common to all DNA polymerases, one in which oncogenic v-Myb Ð known not to cooperate and a unique C-terminal domain described variously as with C/EBPb Ð does not loop out the intervening DNA. PAD (polymerase associated domain) in PolZ, a little The authors discuss some of the factors in¯uencing DNA ®nger domain in Dpo4 and a wrist domain in Dbh. looping in both eukaryotic and prokaryotic transcriptional The thumb and ®ngers domains are unusually small systems, the connotations of which extend to genetic and stubby when compared to those of replicative events ranging from nucleosome assembly to DNA trans- DNA polymerases, resulting in a more `open' active site position. that can accommodate various DNA-distorting lesions. However, despite this coterie of new structures Ð pub- A critical post-transcriptional event is the chemical mod- lished within four months of each other Ð the author i®cation of RNA in from to man. The notes that we still have much to learn about how each extensively modi®ed bases in tRNAs, for instance, are a member of the Y-family bypasses a speci®c lesion, and reminder of the importance of these modi®cations to how replication is coordinated between replicative and cellular life. In his review, FerreÂ-D'Amare presents repair polymerases in cellular organisms. advances in our structural understanding of two types of RNA-modifying enzymes: pseudouridine synthases, A multisubunit RNA polymerase (RNAP) synthesizes which catalyze the isomerization of a speci®c uridine to mRNA in all cellular organisms. In the past few years, a pseudouridine, and tRNA guanine transglycosylases an extraordinary set of multisubunit RNAP structures has (TGTs), which replace a guanine base with a 7-deaza- begun to uncover the details of how mRNA is transcribed guanine derivative. Pseudouridine (c) is the most fre- in both prokaryotes and eukaryotes. The past year quent base modi®cation in RNA and is familiar as the yielded another crop of exciting RNAP structures, which conserved middle base in the TcC loop of tRNAs. includes, for the ®rst time, bacterial holoenzymes with Structures of several c synthases from E. coli have been the promoter speci®city factor, the s subunit. In their determined recently, including `free' TruA, RsuA in review, Murakami and Darst discuss the rapid develop- complex with uracil/UMP and TruB bound to a 22- ments in our structural understanding of how transcrip- nucleotide portion of a tRNA. In spite of a lackof tion is initiated, derived from both `free' (from Thermus sequence similarity, the structures reveal a common aquaticus [Taq] and Thermus thermophilus [Tth]) and pro- catalytic domain with an invariant catalytic aspartic acid moter-DNA-bound (Taq) holoenzymes. The authors residue. The TruB±RNA complex shows that the enzyme highlight the essential features of the s subunit, includ- gains access to the substrate by using a base-¯ipping ing those that permit the recognition of the conserved mechanism that is likely to be shared by other c

Current Opinion in Structural Biology 2003, 13:3±5 www.current-opinion.com Editorial overview Aggarwal and Doudna 5

synthases, as well as by ribonucleoproteins responsible for increasingly on multisubunit factors and complexes, as producing c residues in rRNAs. However, as the author well as the interactions of these proteins with the ribo- points out, the mechanism by which the substrate base is some and the translating mRNA. actually detached from the sugar, rotated and then reat- tached to the sugar is a puzzle at present. FerreÂ-D'Amare A secretory or membrane protein is escorted to the also draws parallels between c synthases and the recently membrane as soon as its nascent polypeptide emerges determined structure of an archaeal TGT. Like c from the . The escort is the signal recognition synthases, the TGT contains a catalytic aspartic acid particle (SRP), a ribonucleoprotein that binds the emer- residue and a structurally homologous domain for inter- ging polypeptide and delivers it to the endoplasmic actions with RNA. Whether this kind of sharing of reticulum in eukaryotes or to the plasma membrane in domains among unrelated RNA-modifying enzymes is prokaryotes. The mammalian SRP consists of a 300 widespread will become clearer with further structural nucleotide RNA that can be divided into S and Alu work. domains, and six proteins: SRP9, SRP14, SRP19, SRP54, SRP68 and SRP72. In their review, Sauer-Eriks- The initiation of protein synthesis is a complex process son and Hainzl stress recent crystallographic studies that involving the assembly of a -competent ribo- elucidate the roles of SRP19 and SRP54 in S-domain some at the start codon of an mRNA. In several coordi- assembly. These new structures, together with earlier nated steps, initiator tRNA binds to the small ribosomal workon SRPs (including bacterial SRP), offer compelling subunit, this complex associates with the mRNA and comparisons between species, between binary and tern- locates the initiation codon, and the large ribosomal ary complexes, and between cognate and noncognate subunit joins to produce a translationally active ribosome. complexes. From these comparisons, the authors point As reviewed by Sonenberg and Dever, structural studies out intriguing commonalities and differences in RNA± of the initiation factors that mediate this assembly process RNA and RNA±protein interactions, as well as in the have revealed several folds common to both prokaryotes sequence of conformational transitions in S-domain and eukaryotes, suggesting evolutionary conservation of a assembly in all three kingdoms of life. core set of proteins essential for translation. Factors that recruit to mRNA, unique to eukaryotes, form Together, these reviews truly capture the pace and exci- large multisubunit complexes that bind to the ribosome tement of structural biology in elucidating genomic inter- and both ends of the mRNA. The authors discuss recent actions. The remarkable snapshots of the assemblages progress in understanding the molecular basis of these and `molecular machines' provide unprecedented detail interactions from both NMR and X-ray crystallographic on the processing of genomic information. These, com- data for proteins that bind the 50 and 30 ends, respectively, bined with the challenge of looking at even larger assem- of mRNA Ð eIF4E and polyA-binding protein (PABP) Ð blies and their underlying dynamics, presage an era of as well as for parts of eIF4G involved in multiple recogni- increasing synergism between X-ray crystallography, tion events. Whereas current workhas focused on single NMR, cryo-electron microscopy and single-molecule polypeptide factors, future structural studies will focus techniques.

www.current-opinion.com Current Opinion in Structural Biology 2003, 13:3±5