(7.5.2)--2.Archaelintronssplicing,interce.pdf

POINT TISS 22 - SEPTEMBER 1997 simultaneously. And even seven inter- connected reactions are capable of sur- prisingly sophisticated behaviour, such as spike generation, oscillations and chaotic fluctuations. So, in this respect, there are clear limits to reductionism fur bio- chemists. Victims of our own success, we have probed so deeply into our sur- roundings that we are no longer able to see the wood for the trees. What can we do about it? There are, I suggest, two possible paths to salvation. The first is to loosen up. Abandon slavish reliance on molecules and look for ex- planations at a higher level. This is of course a normal process, because our ac- counts of the natural world are frequently multi-level, hierarchical, onion skins of explanation. Who would describe a cilium or a chromosome by listing all the mol- ecules they contain together with the properties of these molecules and how they interact? Cilia and chromosomes are clearly delineated structures with defined location and function in the cell. We represent them mentally as distinct entities that operate as units within the ceil. Surely, then, this is the way to tackle other parts of the cell, even those that presently seem impenetrable. We might, for example, try to resolve the plethora of interconnected reactions involved in cell signalling into discrete computational modules. Perhaps the signalling com- plexes that are being found increasingly in many systems will provide us with the elements we need to construct a satisfy- ing reductionist explanation of intracel- lular communication? The second road to salvation is through the keyboard of a computer. Although we poor mortals have difficulty manipulat- ing seven things in our head at the same time, our silicon prot6g6s do not suffer this limitation. Computers have a gargan- tuan appetite for numbers, digesting them at lightning speed and regurgitat- ing them in any desired form - even one that we can understand. Suppose you are interested in the structure of a pro- tein molecule and have been given the precise spatial coordinates of all of the atoms it contains. How much would you learn by studying this long list of num- bers? I suggest ve little. But now feed these numbers inl j a molecular graphics package running on a workstation and. voile! Now you have a three-dimensional image of your protein that you can in- vert, rotate and survey for different residues. You can manipulate this graphical image in ways impossible with the real molecule it stands for, and in so doing achieve an understanding. The dramatic success of this ap- proach encourages me to ask: why not do the same for living cells? If we can use computer-based graphical elements to understand the world of protein structure, why should we not do the same for the universe of cells? The data are accumulating and the computers are humming. What we lack are the words, the grammar and the syntax of the new language. Reference I The Limits of Reductionism in Biology, CIBA Sympesium (Vol. 213) (in press) DENNIS BRAY Department of Zoology, Downing Street, Cambridge, UK CB2 3EJ. Email d.brayzoo.cam.ac.uk At h eal sDr g a |n ons: ICm intercellular mobility and evolution Jens Lykke-Andersen, Claus Aagaard, Mikhail Semionenkov and Roger A. Garrett Until recently, it appeared that archaeal introns were spliced by a process specific to the archaeal domain in which an endoribonuclease cuts a bulge-helix-bulge motif that forms at exon-intron junctions. Recent re- sults, however, have shown that the endoribonuclease involved in archaeal intron splicing is a homologue of two subunits of the enzyme complex that excises eukaryotic nuclear tRNA introns. Moreover, some archaeal introns encode homing enzymes that are also encoded by group I introns. ARCHAEA do not appear to carry either group ! introns, group It introns or nu- clear mRNA.type introns that are found in eukaryotes and/or bacterial Instead, M, ,mlommkov and R, A. Gairett are at the RNA Regulation Centre, Institute of Molecular Biolo, University of Copenhagen, Sglvgade 83H, DK-1307 Copenhagen K, Denmark. Email: garrettmermaid.molbio.ku.dk they carry introns in their tRNA and rRNA genes that are spliced by an apparently archaeal-specific mechanism. However, despite the distinctiveness of the archa,ai introns, there are strong indications that they share some structural and func- tional characteristics with other intron types. Here, the structural characteris- tics of archaeal introns, their splicing mechanism and the mode of action of their encoded homing endonucleases are 326 summarized together with their possible evolutionary relationships to other bac- terial and eukaryotic introns. Structure and splicing mechanism o archaea ntrons Archaeal introns have been detected in some tRNA genes of euryarchaeotes and in tRNA and rRNA genes of crenarchae- otes. in pre-tRNAs, the introns are gen- erally located one nucleotide 3 to the anticodon, at the same location as all eukaryotic nuclear tRNA introns. There are exceptions, however, where archaeai introns are positioned at other tRNA sites, including one within the variable arm z3 (Fig. la). Archaeal introns are also found at diverse sites within the large pre-rRNAs, all of which appear to corre- spond to important functional centres on the ribosomal surface 4. Some of these rRNA introns contain open reading frames (ORFs) of about 600 nucleotides, which are low in G+C content relative to the co-transcribed rRNA and intron core 54. All archaeal intron transcripts generate a bulge-helix-bulge motif at the exon- intron junction aq2 (Fig. lb). The higher order structure of this splicing motif remains unknown, although the three- nucleotide bulges exhibit limited acces- sibility to chemical and enzymatic probes in the free pre-RNA |2, and gel-mobility Copyright 1997, Elsevier Science Ltd. All rights reserved. 0968-0004/97/$17.00 Pil: S0968-0004(97)01113-4 TIBS 22 - SEPTEMBER 1997 TALKING POINT studies indicate that the bulges produce bending of the motif (3. Z. Dagaard and R. A. Garrett, unpublished). The bulge- helix-buge motif is recognized by a splic- ing endoribonucease that cuts at symmet- rica positions within fi e three-nucleotide buges, producing . -cychc phosphates and 5-OH ends u.:) (Fig. lc). Subse- quently, the ends of the exons are lig- ated and the introns circularize (Fig. lc), although the latter process has only been demonstrated or three o the larger introns that carry ORFs 6,3. So far, no archaeal RNA ligase has been isolated. The splicing endoribonuclease has been detected exclusively in archaeal cell extracts 9u. Moreover, it has a broad substrate specificity and it seems likely that the endoribonuclease of any ar- chaeon can cleave any archaeal exon- intron junction 3. It is also likely that this enzyme has a more general processing function, because the bulge-helix-bulge motif is found in other archaeal cellular RNA structures, including the long pro- cessing stems of the large rRNAs 4. Given the wide variation of the primary se- quence of the substrates, it is probable that substrate recognition occurs mainly at a tertiary structural level. Insight into the; process is minimal, however, be- cause although the enzyme has been purified to different levels from cell ex- tracts 9J4, only recently has a gene from the haloarchaeon Haloferax volcanii been cloned and sequence&. A common evolutionary origin The following lines of evidence sup- port a common origin for archaeal and eukaryotic nuclear tRNA introns. (l) Both splicing mechanisms require an endoribo- nuclease that generates a 2-3-cyclic phosphate and 5-OH. However, whereas the archaeal enzyme recognizes and cuts a bulge-helix-bulge motif (Fig. Ic), the nuclear enzyme primarily recog- nizes the tRNA structure and employs some kind of ruler mechanism to rec- ognize the intron-exon junction . (2) Sequence-comparison studies indicate that the archaeal enzyme and the Sen2p and Sen34p subunits of the yeast hetero- tetrameric tRNA-splicing endoribonu- clease are homologues s78. (3) Most archaeal tRNA introns and all nuclear tRNA introns are ocated one nucleotide 3 to the anticodon n6 (Fig. la). Their putative similar origin raises the question as to how these introns and their splicing apparatus evolved. Searches within databases of bacterial genome sequences have failed to reveal homo- logues of the splicing endoribonuclease, 3 5 3 go/ e-g c-gA e C - e -C o - o 5 o G-c / og o-o R Y A a o- e o Aa- C - C-G A 7 - o U-R g c o-o C-g g-c g- e .-o oo g g-c oe - e o R oG- C G m 0 / o o - o e o * o e ! I I I I o o ) O0 o - iO o | I I I Q o e o o e g- C e I o , o e o U o g-C e i e o e g- C o-,) o L ) g-c 0)- - O) -o o_. (- : :-(16) pre-tRNA O) oR pre-rRNA () pre-RNA 5 3 5 3 %, / I _ o - o 5 3 - Ligation - o _ o / . OH - - - ttructural o o o - o %v. -.i j rearrangement) i Slicino . endonuclease ) _ o e o - /4 . o o - a (-.) 6 o o o Q o i HO . (Ligation) - 0-o o - o o - o - - Figure 1 (a) Secondary structural model of tRNA where the locations and numbers (in brackets) of known archaeal introns are indicated; the anticodon is boxed. (b) Conserved bulge-helix- bulge motifs at the exon-intron junctions of archaeal tRNA introns (left) and rRNA introns (right) where cleavage positions are indicated by arrowheads. Exon and intron nucleotides are sllown in brown and red, respectively: upper-case letters indicate nucleotide conser- vation of 85%, and lower-case letters 60-85% conservation of the known sequences; less- conserved positions are presented as dots. R, purine and Y, pyrimidine. Base pairs, includ- ing G-U pairs, are indicated when present in =,85% of the structures. The internal loop in the core structure of the rRNA intron, indicated by open circles, is variable in size. ( Splicing mechanism of archaeal introns. The splicing endoribonuclease cleaves at the exon-intron junctions (indicated by arrows) and generates 5-OH and 2,3.cyclic phos- phates. Subsequently, the exons are ligated by an unknown mechanism, and sometimes a structural rearrangement occurs post-ligation n. The large rRNA intronc circularize. which is consistent with it having evolved after the bacteria separated from the putative archaeal/eukaryotic bran,.h. It is also likely that the more-complex splicing enzyme developed within the eukaryotic branch. Similarily, failure to detect archaeal- or nuclear-tRNA-type in- trons in bacterial genomes suggests that the introas also arose after this putative split. However, for the small tRNA in- trons and rRNA core introns, it remains likely that they derive from more ancient introns that may have been lost by at least some bacteria. Moreover, such a loss (or gain) of tRNA introns seems to be a fairly common event judging by their irregular distribution in the same tRNAs of different organisms. Some archaeal rRNA introns encode homing endonucleases Some of the rRNA introns contain ORFs positioned in the terminal loop of the in- tron core s-7 (Fig. l b). Such introns have only been found in the crenarchaeotal kingdom in the genera Desulfurococcus and Pyrobaculum, despite the sequenc- ing of many archaeal 16S and 23S rRNA genes 4 and of the whole genome of Methanococcus jannaschii 19. The encoded proteins carry two copies of a partly de- generate LAGLIDADG motif, which is 327 TIBS 22 - SEPTEMBER 1997 (a) Intron homlig-slto I-Dmol: 5-ATGCCTTGCCGGGTTCCGGCGcGC-3 3 -TACGGAACGGCC2-CAAGGCCGCgCG- 5 I-Porl: Intron homiig-site 5-GCgAGCCCGtAAGGGtGtACGGGGGC-3 3-CGcTCGGGCaTcaCaTGCCCCCG-5 (c (b) Domain I Domain II LAGLI- LAGLI- DADG DNA DNA DADG DNA DNA b+v b+v M2* M 2* () Endonuclease monomer M m M m s.% . = f-.% . , f %-,= = , a= la 5 Domain I Domain II I l / N / I ! / r Figure 2 (a) Homing sites for introns encoding the archaeal endonucleases I.Dmol and I.Pod are indicated by arrowheads and the cutting positions are marked by red lines. The approxi- mate limits of the recognition site of the homing enzymes are underlined 2z,25. Base pairs conserved among all known crenarchaeotal rRNA genes are indicated by upper-case letters. (b) Diagram of archaeal homing endonucleases showing the linear organization of the two do- mains and the sites implicated in DNA and divalent metal ion binding (M +) that were deduced from protein footprlntlng experiments performed on I-Dmol and I-PorI-DNA complexes 2.6. LAGLIDADG motifs and putative DNA.contact regions are represented as dark and light blue boxes, respectively. Regions of protein backbone positioned close to the divalent metal ions are indicated by orange boxes. The catalytic metal ions, coordinated by the conserved acidic residues (shown in yellow) of the LAGLIDADG boxes, are given as red circles 26. () A two= dimensional projection of a homing endonuclease-DNA-metal ion complex, determineO from the results of both DNA- and protein400tprinting experiments 22.25o28. A probable distor- tion of the DNA double helix 2s is not illustrated. The sugar-phospate backbone is shown as lines, where dots indicate the positions of sugar residues. Base pairs are depicted as ver- tical lines and the major and minor grooves are labelled M and m, respectively. The scissiie phosphates are indicated with asterisks. The two putative domains of the monomeric endo- nuclease are presented as yellow ellipses and the two catalytic metal ions are shown as reu circles. (d) Crystal structure of l-Crel, a homing endonuclease related to I-Drool and I.Porl. The DNA substrate is indicateo by a dashed circle (upper) or dashed lines (lower). In the lower panel, the DNA is located above the protein. Protein-footprinting data on I-Drool and I-Pod are superimposed on the I.Cre! structure, using the same colour code as in (b). The putative positions of the catalytic metal ions (red circles) were deduced from Fe2+-hydroxyl radical and mutagenesis experiments on the archaeal homing enzyme-DNA complexes 26. also found in some proteins encoded by group 1 introns of mitochondria and chloroplasts, and occurs in inteins t, At least two of the archaeal intron-encoded proteins are homing endonucleases that specifically cleave intron- alleles of the rRNA genes that encode them 2.z (Fig. 2a). The DNA interactions of two of the ar- chaeal homing enzymes, l-Drool and l-Porl, have been examined in some detail. Both recognize large regions of double-stranded DNA (15.20 bp) and, like all characterized LAGLIDADG endonucleases, they gener- ate four nucleotide 3-extensions and 5-phosphates on cleavage z,-I (Fig. 2a). Protein footprinting studies on l-Drool- and I-Porl-substrate complexes indicated that both endonucleases exhibit a two- domain structure with one conserved LAGUDADG motif and two DNA-contacting sites within each domain 2 (Fig. 2b,c). Re- cent analyses of the crystal structures of monomeric PI-Sce123 and dimeric l-Cre124, two eukaryotic homing endonucleases, distantly related to monomeric l-Drool and I-Porl, confirmed these predictions. Each DNA-binding region, one within each protein half (Fi