598 Vol. 277 22 February 1979

Exons encode functional units from c. C. F. Blake AN important contribution to relation­ the / intron mosaic in the coding coded by a separate exon is most ship between protein structure and the region could be most illuminating. exciting. So clear cut is the corres­ newly discovered mosaic structure of The DNA was obtained from mouse pondence that one is tempted to con­ eukaryotic , in which coding MOPC 21 myeloma by complete clude that the problem of has segments of DNA (exons) alternate digestion with EcoRI nuclease and been solved along the lines proposed with non-coding segments (intervening the identification of a 6.8 kilobase by Gilbert (News and Views 271,501; sequences or introns), is reported by segment that contained the required 1978) in which the presence of exons Tonegawa and his colleagues in this sequence by hybridisation with cDNA was seen as allowing new to week's Nature (Sakano et aI. page prepared from MOPC 21 y l mRNA. be constructed from parts of old ones. 627). The protein chosen was the The segment was recombined with The results reported by Tonegawa are constant region of the immunoglo­ the EK2 vector, phage Agt WES, and exactly in line with this hypothesis, bulin "I heavy chain. Sequence studies cloned. After digestion of the clone and show that individual exons can and crystallography have already the position of the 11 in define even very small functional shown that the constant region is the 6.8 kb fragment was determined by units, like the 14-residue hinge region, composed of three homologous struc­ electron microscopy of the R-Ioops and not just domains or folding units tural domains: the CHI, CHz and formed between it and the "II-chain as I previously suggested (News and CH3 domains. Each domain is com­ mRNA. This revealed three smaIl R­ Views 273, 267; 1978). In view of the posed of about 110 amino acids, loops separated by two intervening precise correlation between exons and folded into almost identical tertiary double-stranded DNA segments, sug­ the functional units in the immuno­ structures made up from seven gesting the presence of three exons globulins, the presence of three exonS /1-strands. The CHI domain pairs with interrupted by two introns. A restric­ in the /1-globin , when the protein the CL domain of light chains to form, tion enzyme cleavage map showed cannot readily be divided into indi­ together with paired VH and VL that this interpretation was correct, vidual units of either structural or domains, the Fab parts of the "I-glo­ but that there was an additional very functional significance seems at first bulin molecule. The CHz and CR. short exon located between exons 1 very puzzling. However, closer exam­ domains pair with equivalent domains and 2, giving a final pattern of 4 exons ination shows that the central exon of a second heavy chain to form the separated by 3 introns. Sequence corresponds very closely to the protein Fe fragment, which is separated from analysis of the intron-exon boundaries fragment that defines the haem the Fab fragment by a hinge region. and comparison with the known crevice, and contains nearly all the Each of these structures has a well amino acid sequence of the MOPC amino acid residues that interact with defined function: the paired CH~ 21 y I chain, revealed that the exons the haem. Tonegawa's work implies domains act in surface inter­ represent, in sequence, the CHI that exons may reveal unsuspected actions, the paired CHz domains in domain, the 14-residue hinge region, functional protein units-perhaps like complement fixation; the hinge prob­ the CH ~ domain, and the CR! domain. the 'minimal globin' outlined above­ ably transfers information from the There seem to be no introns within even when the joins are not visible in Fab to the Fo. fragment; the CH, pair the individual sequences coding for the protein structure. The thought is the attachment point for the light these units. that the exon/intron mosaic may chain; the VII VL pair forms the anti­ The conclusion from this study, that carry the frozen evolutionary history gen binding site; and there is also an each of the six known functional of proteins makes one look forward N-terminal signal peptide. Thus the units in the "I-globulin heavy chain is to see whether further determinations immunoglobulin molecule is com­ of exon-intron patterns of other pro­ C. C. F. Blake is in the Laboratory of teins of known structure support this posed of six discrete structural/ Molecular Biophysics. Department of functional units whose relationship to Zoology, University of Oxford. view, or point in a different direction.

molecule whose active group is nic­ RNA splicing and polynucleotide evolution otinamide. T suggest that rhe first stJeps from D. Reanney on the road to precellular metabolism were taken when electron-donating and INFORMATION flow in cells, is defined biology has come from several lines of electron-accepting reactions were by the 'oentral dogma' 'as evidence whioh strongly reinforce the coupled by means of molecules whose view of Crick (1. molec. BioI. 38, 367; 'handles' were complementary in the DNA~RNA~Protein 1968) and others that RNA is the Watson-Crick sense. For example primary genetic material. It is worth But although the central dogma ac­ re-emphasising that most key features UDP-X (red)~UDP-X (ox) curately describes the flow of infor­ of the translation of genetic informa­ where V is paired with A mation in modern cells, it probably tion depend on Watson-Crick pairing NAD (ox)~NADH+H (red) fails to reflect evolutionary perspective. between (partly) complementary . Considered fwm the chronological Translation is mediated by interactions Here the 'gen,etic' monomers A and viewpoint a more accurate represen­ between different RNAs-mRNA: U are used to bring the two elements tation of the oentral dogma might be tRNA; mRNA:rRNA; rRNA:tRNA of the redox reaction into appropriate and rRNA:rRNA. The same was pro­ juxtaposition; this is effectively what /DNA bably true of primitive 'metabolism'. an enzyme does. The addition of RNA With respect to a metabolic fossil such 'Protein as the ribodinucleoticte NAD, for D. Reanney is Senior Lecturer in the example, one may ask why is a nucleo­ Department of Microbiology, La Trobe This rewriting of a basic tenet of tide 'handle' a necessary part of a University, Australia.

0028~836J79/0277-OS98S01.00 IC Macmillan Journals Ltd 1979