Benjamin Lewin

Units of transcription and translation: sequence components of heterogeneous nuclear RNA and messenger RNA.

Defining the units in which the eucaryotic genome is transcribed and translated is central to any analysis of eucaryotic gene expression. The relationship between heterogeneous nuclear RNA and messenger RNA raises the question of whether the primary transcript may be more complex than the sequence which is translated; as I concluded last month in the first part of this review, kinetic analyses of these two RNA populations provide some suggestive indications but cannot prove whether the nuclear population includes messenger precursors that are much longer than mature cytoplasmic messengers (Lewin, 1975). Here I discuss recent analyses of the sequence components present in hnRNA and mRNA and how they may be related to each other and to the organization of the genome.

Oncogenic conversion by regulatory changes in transcription factors

In this review ,I shall adress the relationship between the retroviral-transforming v-onc genes and the c-onc genes coding for transcription factors to ask whether changes in the control of transcription may account for the generation of the oncogenic phenotype .As with other classes of oncogenes,we may enquire whether the expression patterns of the cellular proto-oncogenes offer any insights into the specifity of their oncogenic derivatives.

CHROMATIN AND GENE EXPRESSION : CONSTANT QUESTIONS, BUT CHANGING ANSWERS

Benjamin Lewin Cell It is now 20 years since the discovery of the nucleosome (Kornberg, 1974) and an appropriate time to take stock of the chromatin field, which was the objective of a meeting on chromatin structure and gene expression held at the Fundacion Juan March in Madrid from September 26-28, 1994. The common theme was the view that chromatin is a dynamic structure, that changes occur in the arrange- ment of histones and DNA, and that eukaryotic gene ex- pression can be understood only in the context of these changes. This review focuses on one aspect of the meet- ing: interactions between the transcription apparatus and chromatin and the preceding or consequent changes in structure. Nucleosome Structure and Mobility Data on the nature of transcribed chromatin have been conflicting. Two extreme situations have created a para- dox that remains at the heart of the issue. Intensively tran- scribed chromatin such as rDNA is organized in an ex- tended form that lacks nucleosomes. Yet the average transcribed gene is covered in nucleosomes, as can be seen from the generation of the standard ladder upon treatment with micrococcal nuclease. The problem with analysis in the latter situation has been the difficulty of determining whether any particular copy or region of the gene is both transcribed and nucleosomal, since the nu- cleasomes might be present on copies of the gene, or on parts of the gene, that are not actually being transcribed. The crucial question has been what happens to the histone octamer when an RNA polymerase traverses DNA. Recent work suggests that histone octamers are dis- placed by RNA polymerase, but that they retain contact with the DNA and reattach to the same template behind the polymerase (Studitsky et al., 1994). Figure 1 shows the rhodel for polymerase progression. DNA is displaced as the polymerase enters the nucleosome, but the poly- merase reaches a point at which the DNA loops back and reattaches, forming a closed region. As polymerase ad- vances further, it creates positive supercoils in this loop; the effect could be dramatic, because each base pair through which the polymerase advances makes a signifi- cant addition to the supercoiling in a closed loop of - 80 bp. Felsenfeld (NIDDK, NIH) reported that measurements of RNA polymerase progression, to analyze the addition of bases to the transcript with time, show that polymerase progresses easily for the first 30 bp into the nucleosome. Thenlit pauses with each additional base, as though en- countering increasing difficulty in progressing. When it reaches the midpoint of the nucleosome (the next bases to be added are essentially at the axis of dyad symmetry), pausing ceases, and the polymerase advances rapidly. This suggests that the midpoint of the nucleosome marks the point at which the octamer is displaced (possibly be- cause positive supercoiling has reached some critical level that expels the octamer from DNA). This releases tension ahead of the polymerase and allows it to proceed. The octamer then binds to the DNA behind the polymerase and no longer presents an obstacle to progress. These results were obtained with Sp6 phage RNA polymerase (which of course is much smaller than a eukaryotic RNA polymerase). Is the octamer released as an intact unit? Sradbury (UC Davis and Los Alamos National Laboratory) showed that cross-linking the proteins of the octamer does not create an obstacle to transcription (by T7 RNA polymerase). Tran- scription can continue even when cross-linkiing is exten- sive enough to ensure that the central regions of core histones have been linked. This implies that transcription does not require dissociation of the octamer into its compo- nent histones, nor is it likely to require any major unfolding of the central structure (C’ Neill et al., 1993). hlowever, ad- dition of histone Hl to this system causes a r,apid decline in transcription. The work suggests two conclusions: the histone octamer (whether remaining present or displaced) functions as an intact unit, and it may be necessary to remove HI from active chromatin or modify its interactions in some way.

Genes for SMA: Multum in parvo

In this issue of Ceil, we take the unusual step of publishing two articles each of which reports the identification of a different gene in the appropriate chromosomal region associated with the human disease of spinal muscular atrophy (SMA). Each paper individually meets the standard for publication in providing adequate evidence to identify systematic changes in the gene in patients with the disease but not in normal people, but the two groups have identified unrelated genes, called SMN and NAIP. SMA is a relatively common recessive autosomal disease, affecting 1 in 6000 births, and is one of the most common genetic causes of death in childhood. Three clinical types of the disease (I, II, and III) are distinguished by the (decreasing) severity of the symptoms. The common cause is depletion of motor neurons in the spinal cord, resulting in muscular atrophy with consequent paralysis of limb and trunk. All three types of SMA map to chromosome region 5q11.2-q13.3, and several research groups have been engaged for some years in constructing molecular maps of this region in order to identify the causal locus. The search for SMA has been complicated by the instability of this chromosomal region; it has many repetitive sequence elements that often cause instability in cloned sequences. It has several classes of expressed pseudogenes, which have an interrupted organization and generate transcripts that are spliced to give mRNA-like molecules that do not give functional products. This complicates attempts to identify the gene(s) for SMA, because

Alternatives for splicing: Recognizing the ends of introns

Since the discovery that eucaryotic genes may constitute mosaic structures in which the exon regions represented in mRNA may be interrupted by intron sequences that are removed from the primary transcript, a critical question has been how the correct sequences in RNA are selected for splicing together. In most cellular genes the splicing pathway is invariant; the same primary transcript always gives rise to the same mature messenger by removal of defined introns. But there are exceptions in which alternative processing patterns are available to a single gene; these may be used under different cellular conditions (as with the immunoglobulins) or simultaneously (as with SV40 and polyoma). This generates proteins whose sequences are in part common, in part different: here one gene’s intron can be another’s exon. Whatever system is responsible for recognizing and precisely splicing together the exact ends of each intron in RNA must therefore be flexible enough to accomodate both transitions in a given pathway and the simultaneous existence of alternative pathways. All the known sequences at intron-exon junctions in the nuclei of higher eucaryotes can be aligned around hypothetical common splice points by maximizing the brief nucleotide sequence homology that is found at the junctions. [These are hypothetical because in many cases the ends cannot be distinguished unambiguously due to the repetition of a short sequence (from 1-4 bases) at both junctions.] This suggests the “consensus” sequence:

The Mystique of Epigenetics

position effect variegation (in which the probability of Epigenetic effects have often enough been viewed as inactivation of a gene translocated to a position near verging on the mystical. It is a paradox of conventional heterochromatin decreases with its distance from hetgenetics that two alleles can have the same genetic erochromatin). One of the characteristics of a state that sequence but show different states of inheritance. This depends upon an alternative protein structure is that the can be resolved by supposing that the ability to inherit extent of the affected region is unlikely to have defined a nongenetic state reflects the existence of templating. limits. Because the determined condition propagates There must be a transition between two (discontinuous) (perhaps analogous to a crystal), its extent may be limstates (equilibrium situations are excluded). Each of the ited by the (variable) supply of components (unless a alternative states is stable. One of these states can be discrete boundary is encountered, which is not usually regarded as the naı̈ve state—what is achieved simply the case). by synthesis of the relevant components. The other state What conditions must be fulfilled to create an epigecan be regarded as a determined state, in which some netic effect? A discrete event must generate a difference special property has been conferred that distinguishes in structure, either by de novo methylation of DNA or the components (or their macromolecular assembly) by modifying proteins or nucleating a protein structure. from the naı̈ve state. In this issue of Cell, we dispel The structure must be perpetuated, in the case of methmysticism by considering the molecular bases for a variylation because of the existence of an extrinsic enzyme ety of situations involving epigenetic effects. Their system that acts constitutively on hemimethylated DNA, causes fall into two general classes, depending on in the case of a protein structure because the assembly whether DNA or protein is the target for conversion from is intrinsically self-templating (with the extrinsic condi-

Interaction of regulator proteins with recognition sequences of DNA

Benjamin Lewin The MIT Press 28 Carleton Street Cambridge, Mass 02142 Although models for the control of gene expression have for some time focused on the concept that regulator proteins act by recognizing specific nu- cleotide sequences, only now is it becoming possi- ble to define the molecular basis of such interac- tions. The existence of two sites at each control locus, the operator to which repressor protein binds and the promoter to which RNA polymerase binds, is well established in bacterial and phage systems. By preventing RNA polymerase from binding to and/or proceeding from the promoter, the binding of repressor protein at the operator blocks tran- scription of the adjacent structural genes. But because the control sites form only a small part of each system, experiments with DNA tem- plates in vitro have been unable to answer critical molecular questions: how long are the nucleotide sequences of the recognition sites; what is the rela- tionship between operator and promoter; and what features of their sequences are recognized? The first step in answering these questions, determi- nation of the recognition sequence, can now be achieved through the isolation of small fragments containing the control sites; these experiments con- firm earlier suspicions that symmetry may be impor- tant in recognition sequences. Isolation of the Lactose Control Region Genetic analysis has identified two control sites in the lactose operon. Mutations which abolish the usual response to repressor, causing continual (constitutive) transcription of the operon, define the operator and map immediately to the left of the three coordinately controlled structural genes (Beckwith, 1970). To the left of the operator, muta- tions which reduce the level of transcription of the operon identify a site essential for RNA synthesis. Although at first thought to inhibit the binding of RNA polymerase, these mutations are now known to influence the response to the catabolite activator protein, which is activated by cyclic AMP and whose mediation is required for the initiation of transcrip- tion (de Crombrugghe et al., 1971). The site at which RNA polymerase is presumed to bind lies be- tween these two groups of mutations, giving the order: