How do transcription factors work? - Genes & Development

MEETING REVIEW

How do transcription factors work? A r n o l d J. B e r k a n d M a r t i n C. S c h m i d t

Department of Microbiology and Molecular Biology Institute, University of California, Los Angeles, California 90025-1570 USA

A casual look about us demonstrates the complexity of biological form that likely depends on intricate mecha- nisms of transcriptional regulation. Yet, initial studies of transcriptional regulation in eukaryotes presented a relatively simple picture. DNA sequences of 6 -20 bp within several kilobases of transcription initiation sites were shown to be binding sites for transcription factors which in most cases activate and in some cases repress initiation. It seemed that for any gene, if one determined which transcription factor binding sites were present and the cell types in which the relevant transcription factors were expressed, the developmental regulation of expression could be understood. The recent Cold Spring Harbor Cancer Cells Meeting on Regulation of Eukary- otic mRNA Transcription (September 6-10) made it clear that the situation is much more complex than this simple picture. Control sequences can interact with multiple different proteins, transcription factors may be multimers of distinct polypeptide chains in which dif- ferent combinations of subunits yield different func- tions, and combinations of control sequence commonly produce expression patterns that differ markedly from the summed effects of the individual control sequences. Combinatorial diversity of transcription factor polypep- tides and control sequences was a major theme of the meeting. Limited space permits review of only a fraction of the interesting work presented.

Control of factor activity

A talk by Weintraub (Hutchinson Cancer Center, Seattle) exemplified the principle that different combi- nations of polypeptide chains in dimeric transcription factors may have profound effects on factor activity. Pre- vious work by Davis, Lassar, and Weintraub had shown that expression of the MyoD gene activates myogenic differentiation of 10TV2 cells and the expression of striated muscle specific genes, such as muscle creatine kinase (MCK), in many types of cultured cells. MyoD protein binds to the MCK control region both as a ho- modimer and as a heterodimer with El2, a ubiquitously expressed nuclear protein, whose gene was originally cloned by Baltimore's group (MIT). Paradoxically, both myoD and El2 are expressed in replicating myoblasts before differentiation and their levels remain the same when myoblasts are stimulated to develop into myo- tubes by removing serum from the culture medium. Why don't MyoD and El2 activate myogenesis in repli- cating myoblasts?

Benegra, Davis, and Weintraub used the "he l ix - loop- helix" (HLH) homology region of MyoD and El2, which is responsible for dimerization, as a probe to search for

additional cDNAs encoding proteins in this class. A cDNA encoding a protein called Id was identified which contains the HLH dimerization domain but lacks the basic region just amino terminal to it that is required for DNA binding. Id was shown to complex with both El2 and MyoD, and to repress the ability of MyoD to acti- vate the MCK promoter, presumably because Id forms heterodimers with MyoD that cannot bind DNA. Most interesting of all, Id message is expressed in myoblasts and falls in concentration as they differentiate into myo- tubes. Thus, differentiation may result when MyoD and El2 are released from inactive heterodimers with Id, al- lowing them to form activate homo- and heterodimers, which then bind to muscle-specific gene control regions and activate transcription.

The El2 gene figured in another surprising report at the meeting. Mellentin and co-workers (Stanford) in col- laboration with the Baltimore group mapped the El2 gene to human chromosome 1, which is involved in the tl :19 translocation observed in -40% of acute lympho- cytic leukemias (ALL). The group found that the translo- cation falls within the El2 gene in 95% of all ALL cell lines with this translocation. As a result, the 5' end of the El2 gene which is expressed in developing lympho- cytes is fused to a gene in chromosome 19 not normally expressed in these cells. Cloning and sequencing of the cDNA from the gene created by the translocation in one cell line revealed that the translocation removed the he l ix- loop-he l ix and DNA-binding domain of El2 and substituted a sequence from chromosome 19 including a homology to the homeo box class of DNA-binding do- mains. The amino-terminal portion of E 12 remaining in the recombinant protein contains a transcription-acti- vating domain mapped by Kadesch (University of Penn- sylvania) using a yeast expression and assay system. As a consequence, an activation domain may be fused to a new DNA-binding specificity not normally expressed in developing lymphocytes. The resulting protein could be a novel transcriptional activator important in the patho- genesis of this class of ALL.

In the introduction to one session, Baltimore empha- sized that nature is not likely to have missed any pos- sible mechanisms to regulate gene expression. This prin- ciple was dramatically demonstrated in a talk by Levine (Columbia University) on the regulated nuclear trans- port of the dorsal protein of Drosophila, required to es- tablish the dorsal-ventral axis of the early embryo. dorsal protein is distributed throughout the cytoplasm of the early embryo; however, following the ninth nu- clear cleavage, dorsal protein is transported to the nuclei in the ventral region of the embryo specifically, while it

GENES & DEVELOPMENT 4:151-155 © 1990 by Cold Spring Harbor Laboratory Press ISSN 0890-9369/90 $1.00 151

remains in the cytoplasm in the dorsal region. Muta- tions in any of the 11 genes known to regulate dorsal function prevent transport of dorsal to the nuclei in the ventral embryo, indicating that this regulation of trans- port is extraordinarily complex.

In collaboration with Manley (Columbia University), Levine showed that dorsal protein expressed in Droso- phila tissue culture cells following transient transfec- tion of the gene is principally cytoplasmic. However, when the carboxy-terminal 6 or 8 residues are deleted, the protein is transported into the nucleus. Two of the 11 genes involved in regulating dorsal function have been cloned and their sequences show homology to serine proteases, indicating a possible mechanism for regulation of dorsal subcellular localization.

Evidence is accumulating that the activity of several transcription factors can be regulated by changes in their state of phosphorylation. Gonzalez and Montminy (Salk Institute, La Jolla) presented evidence along these lines with CREB, a factor that binds to the cAMP response element (CRE) and stimulates transcription in vitro. They showed that co-transfection of expression vectors for CREB and the catalytic (C) subunit of protein kinase A (PICA) resulted in strong stimulation of a co-trans- fected CRE-CAT reporter gene in F9 teratocarcinoma cells. Co-transfection of the CRE-CAT reporter gene with the CREB expression plasmid alone or with the ex- pression plasmid for the C-subunit of PICA alone did not result in activation. Moreover, mutation to alanine of the CREB serine phosphorylated by PICA in vitro pre- vented activation. These results indicate CREB activity is directly induced by PICA phosphorylation.

In contrast, Karin (University of California, San Diego) and Hunter (Salk Institute) presented a poster arguing that activity of AP-1, a heterodimer of c-Jun and c-Fos, is induced by dephosphorylation. They presented a model in which c-Jun in unstimulated cells is phosphorylated by a glycogen synthase kinase 3 (GSK-3)-like kinase, decreasing the affinity of AP-1 for TPA-responsive ele- ments (TREs). In response to TPA treatment a specific phosphatase is activated, removing phosphates from these sites and resulting in increased TRE binding by AP-1 and transcriptional activation.

Another provocative model for the control of tran- scription factor activity came from Tan and Richmond (Institute for Molecular Biology and Biophysics, Zurich) who argued that activation by yeast PRTF (also known as GRM and MCM-1) requires a conformational change in the protein induced by binding to some specific sites and not others. PRTF is thought to be required for tran- scription from both a- and a-mating-type-specific genes. However, binding of PRTF alone is thought to activate transcription of a-specific genes, whereas another se- quence specific binding protein, MATch,, is required for activation of a-specific genes where it binds a site neigh- boring the PRTF binding site. Through analysis of PRTF protease-sensitive sites, Tan and Richmond presented evidence that binding to an a site causes a conforma- tional change in PRTF that exposes an activation do- main on the protein surface. At c~-sites, they propose

that an interaction with MATc,1 is required to force the conformational change.

Combinatorial effects of control dements

The significance of the combinatorial effects of control elements was apparent in a number of studies. Ptashne (Harvard University) showed that in yeast, multiple binding sites for the Gal4 transcription factor result in much greater than additive increases in transcription compared to the level observed with a promoter having a single binding site. This synergism was observed even when the level of Gal4 protein was saturating so that single and multiple binding sites were fully occupied by Gal4 in vivo. Consequently, the synergism could not be due to cooperativity of Gal4 DNA binding. Ptashne sug- gested that the synergism of activation may be the result of multiple cooperative contacts between the activator proteins and the general transcription apparatus assem- bled at the TATA box and initiation site. Driever (Max Planck Institut, Tubingen) discussed the importance of cooperativity of transcription factor function in defining the sharp boundaries of gene expression in Drosophila development. A gradient of bicoid protein across the early embryo from anterior to posterior results in a re- gion of hunchback expression in the anterior portion of the embryo sharply demarcated from a region where hunchback is not expressed. When a single binding site for bicoid was placed upstream of a [3-galactosidase re- porter gene, the gene was turned on by bicoid in the an- terior, but f3-galactosidase expression fell off slowly in the posterior direction, following the bicoid gradient. When multiple bicoid binding sites were place upstream of the reporter gene, there was a sharp demarcation be- tween the anterior region of f~-galactosidase expression and the posterior region of no expression, as seen for hunchback expression. Thus, the sharp boundary of hunchback expression required for normal development is mediated by strong cooperativity of bicoid function which in turn depends on multiple bicoid binding sites. A similar sharp on-off switch of the bacteriophage h PL and PR promoters is mediated by cooperative binding of h repressor to adjacent binding sites (Ptashne, Harvard University). But as yet, there is no indication as to whether the cooperativity observed with bicoid function is due to cooperativity in DNA binding or to coopera- tivity in the activation mechanism as observed for Gal4.

An example of dramatic differences in the effects of combinations of control proteins compared to the effects of the individual factors came in the presentation by Manley (Columbia University) concerning Drosophila homeo box proteins. Manley discussed studies on tran- scriptional activation by the zen and ftz homeo box pro- teins using transient transfection assays in Drosophila tissue culture cells. Co-transfection of either a zen or a ftz expression plasmid with a reporter gene containing upstream binding sites for both proteins resulted in a low level of activation compared to transfection of the reporter gene alone. However, a triple transfection of the reporter gene with both the ftz and zen expression

152 GENES & DEVELOPMENT

plasmids resulted in a dramatic level of activation much greater than that seen with either ftz or zen alone. Fur- ther results suggested that this synergism is not due to cooperative DNA binding of the homeo box proteins. A zen- f t z fusion protein containing only the zen homeo box DNA-binding domain produced a high level of acti- vation similar to co-transfection of the wild-type zen and ftz genes, arguing that the synergism is due to an interaction between the homeo box proteins that does not involve DNA binding and is reproduced in the fu- sion protein.

Multiple proteins bind to one control element

Further complexity of control elements was made ap- parent by reports that multiple proteins can be found that bind to the same promoter element. For example, Green (Harvard University) isolated eight distinct cDNA clones of proteins that all bind the ATF control element found in several adenovirus promoter regions. Only one of three of these thus far tested activated tran- scription in a transfection assay. Maniatis (Harvard Uni- versity) reviewed the interferon-B (IFN-[3) control region where three positive elements are flanked by two nega- tive elements as identified in transient transfection assays with mutated control regions. One of the positive elements known as PRD-I binds several factors in nu- clear extracts. Three of these have been cloned by the groups of Maniatis and Fujita (Osaka University). When a plasmid expressing the factor cloned by the Maniatis group was co-transfected with a reporter gene bearing the PRD-I control region, transcription of the reporter gene was repressed. A distinct factor cloned by the Fujita group also repressed transcription, while another of Fu- jita's factors activated transcription. A second positive control region, PRD-II, also binds several factors in nu- clear extracts, including H2TF1 and NF-KB. Maniatis commented that determining which of these factors reg- ulates IFN-f~ expression in vivo may require the applica- tion of new approaches to knock-out the genes encoding them.

Another case in point is the octamer element required for both B-lymphocyte-specific expression of immuno- globulin genes as well as the ubiquitous expression of U snRNA and histone H2b genes. Analysis of the proteins that bind to this sequence identified the Oct-1 factor, found in all mammalian cell types examined, and the Oct-2 factor, found in B-lymphocytes. It was not clear how these proteins, which bind to the same DNA se- quence, activate transcription of U snRNA and histone H2b genes in all cell types and of immunoglobulin genes in B lymphocytes alone. An explanation for this ap- parent paradox was suggested by co-transfection studies reported by Herr (Cold Spring Harbor). His group used Oct-1 and Oct-2 expression plasmids and a B-globin re- porter gene with copies of the octamer inserted near the TATA box. Oct-2 expression activated the reporter gene in HeLa cells, whereas Oct-1 expression did not. These results suggest that Oct-2 carries an activation domain, whereas Oct-1 does not include one that can act inde-

pendently at a typical TATA-box containing promoter. It was proposed that for Oct-1 to stimulate transcription, it must function in specific promoter contexts found in the U snRNA and histone H2b genes. In contrast, the isolated octamer sequence acts as a strong tissue-specific promoter element in B lymphocytes because Oct-2 is specifically expressed in these cells and contains a gen- eral activation domain.

Although the model that only Oct-2 has an indepen- dent activation domain helps to explain why an isolated octamer acts as a promoter element in B lymphocytes, other results showed that B-lymphocyte-specific en- hancer activity of the octamer is still more complex. Schaffner (University of Zurich) reported that when multiple octamer sequences were inserted several kilo- bases from the B-globin promoter, the octamer repeat acted as an enhancer in B lymphocytes, but not in HeLa cells co-transfected with the Oct-2 expression plasmid. Perhaps other B-lymphocyte-specific factors are required to mediate long-range activation by Oct-2.

The influence of chromatin structure

Three talks exemplified studies on the effect of chro- matin structure on transcription. Grunstein (University of California, Los Angelesl presented experiments showing that removal of nucleosomes in vivo in yeast derepresses transcription and suggested that one of the functions of upstream transcription factors may be to re- move nucleosomes from the TATA-box/transcription initiation site region, allowing general transcription factors to interact with the promoter. Grunstein also showed that histone H4 is involved in the shut-off of the yeast silent mating-type loci. Deletion or substitution of the basic region at the amino-terminus of H4 results in failure to suppress the silent loci. This effect can be sup- pressed by three specific alleles of SIR 3, a protein re- quired to repress the silent loci along with Sir 1, 2, and 4. The results suggest that the SIR 3 protein silences through an interaction with histone H4.

Kellum and Schedl {Princeton Universityl studied two regions on each side of the 87A7 heat-shock locus of Drosophila, called scs left and right, which may delimit higher-order chromatin structures. These scs regions ex- hibit extreme protection from micrococcal nuclease di- gestion over a 350-bp region. In several types of in vivo experiments, Kellum and Schedl showed that the scs re- gions block the effects of enhancers when placed be- tween an enhancer and promoter. They suggested that these scs DNA segments may form nucleoprotein struc- tures that act as boundaries between higher order chro- matin domains, allowing independent gene expression in neighboring regions of the genome.

Workman, Kingston (Massachusetts General Hos- pital}, and Roeder {Rockefeller University) reported that the assembly of nucleosomes onto template DNA dra- matically increased activation by an upstream transcrip- tion factor in vitro. They added purified transcription factors to a plasmid containing the adenovirus major late promoter at the same time as histones and a nucleo-

GENES & DEVELOPMENT 153

some assembly extract from Xenopus oocytes. When TFIID (the TATA-box factor) alone was added during nucleosome assembly, followed by the remaining gen- eral factors after assembly, transcription was inhibited progressively with increasing numbers of nucleosomes per template DNA molecule. Addition of the upstream factor USF with TFIID at the time of nucleosome as- sembly resulted in much less inhibition. As a result, the ratio of transcription with USF compared to that without USF was much greater following nucleosome assembly compared to the effect of USF on transcription from a free DNA template. Apparently, the effect of USF on TFIID binding is augmented during a period of com- petition between nucleosomes and transcription factors for DNA binding.

RNA polymerase H and general transcription factors

The details of transcription initiation by pol II and the general transcription factors are poorly understood. Yet it is probably these processes that are regulated by gene- specific transcription factors. Considerable progress was reported in the study of the basic factors and pol II. Sev- eral talks focused on the carboxy-terminal heptapeptide repeat (CTD) of the largest pol II subunit, which is highly phosphorylated in vivo. Corden (Johns Hopkins School of Medicine) reported on the isolation of a mu- rine protein kinase which can phosphorylate this region in vitro. The purified kinase has subunits of 58 and 34 kD, and the 34-kD subunit was shown to be similar or identical to the murine homolog of the cdc 2 protein ki- nase of Schizosaccharomyces pombe involved in cell cycle control. However, Greenleaf (Duke Medical Center) purified a similar CTD kinase from Saccharo- myces cerevisiae and found it to be composed of three polypeptides of 58, 38, and 32 kD, none of which had the properties of the cdc 28 protein kinase, the S. cerevisiae equivalent of S. pombe cdc 2. Young (MIT) also reported that incubation of the cdc 28 mutant at the nonpermis- sire temperature did not alter phosphorylation of the large pol II subunit. Greenleaf cloned the gene encoding the 58-kD polypeptide of the CTD kinase. Sequencing showed that it contains a homology to the region of cdc 28 conserved between various protein kinases. Bura- towski and Sharp (MIT) showed that the CTD is not nec- essary for pol II to enter fully assembled transcription complexes identified in native polyacrylamide gels. Both Greenleaf and Buratowski reported that the CTD is not required to observe stimulation of transcription by up- stream factors in vitro. Thus, while the CTD is essential in vivo, its function remains unclear.

Dahmus (University of California, Davis) showed that when pol II with an unphosphorylated CTD is incubated with the adenovirus major late promoter, general tran- scription factors, and ATP, GTP, or dATP, the pol II en- tering transcriptionally competent complexes becomes phosphorylated on the CTD. Dahmus suggested that the CTD kinase may be a component of the general tran- scription factors and that CTD phosphorylation may be required to release pol II from the initiation complex,

154 GENES & DEVELOPMENT

allowing elongation. This model has implications for transcriptional regulation since CTD phosphorylation could be controlled through the action of gene-specific factors.

The groups of Hahn (Fred Hutchinson Cancer Center, Seattle), Buratowski, Sharp, and Guarente (MIT); Win- ston (Harvard Medical School); Horikoshi and Roeder (Rockefeller University); Egly and Chambon (University of Strasbourg); and Schmidt and Berk (University of Califomia, Los Angeles) all reported on cloning the gene for S. cerevisiae TFIID, the protein that binds to the TATA box, initiating a cascade of assembly of general transcription factors TFIIA, B, E, F, and pol II. Winston showed that the gene is essential for viability and that a mutant, sptlS, alters transcription initiation in vivo. Only a single gene could be found in S. cerevisiae, and Hahn et al. showed that the isolated protein bound to multiple TATA boxes of different sequence. Struhl (Har- vard Medical School) showed that multiple single and double point mutations in the cerevisiae his3 TATAAA box had similar effects on in vivo transcription in yeast and in vitro transcription using a HeLa nuclear extract, indicating that yeast and HeLa TFIID have nearly iden- tical DNA-binding specificities. Horikoshi and Roeder presented an extensive analysis of in phase deletions which showed that the amino-terminal 60 residues of the 240-residue protein are dispensable for DNA binding and in vitro transcription. However, all mutations in the carboxy-terminal 180 residues eliminated both tran- scriptional and DNA-binding activities. This finding contrasts with gene-specific transcription factors that generally have much smaller DNA-binding domains and often have distinct activation domains. Schmidt and Berk reported that yeast TFIID does not bind DNA at 0°C, in contrast to most DNA-binding proteins, binds with slower kinetics, and untwists the DNA helix as it binds. Taken together these results raise the possibility that TFIID binds DNA by an unusual mechanism, which may be an essential feature of its function in transcription initiation.

Greenblatt (University of Toronto) presented the iso- lation from HeLa cells of a general transcription factor called RAP 30/74 composed of two tightly associated polypeptides of 30 and 74 kD, and the isolation of a cDNA clone for RAP 30. RAP 30/74, isolated using a pol II affinity column, is probably equivalent to TFIIF which Reinberg (University of Medicine and Dentistry of New Jersey) isolated as a complex of polypeptides of similar size. TFIIF together with TFIIE are the last general tran- scription factors to bind to the adenovirus major late promoter, following TFIID, A, B, and pol II. The RAP 30 sequence shows homology with a conserved region shared by all eubacterial sigma factors that is thought to interact with core RNA polymerase. Most significantly, Greenblatt showed that RAP 30/74 has DNA helicase activity, leading him to suggest that this factor melts the DNA strands, allowing access to the coding strand by pol II. Consistent with this, Buratowski and Sharp showed that binding of TFIIE/F to the initiation complex results in hypersensitivity to cutting by orthophen-

anthroline-Cu +2 at - 6 to - 4 . Hypersensitivity to cut- ting by this reagent is also seen in the single-stranded region generated when E. coli RNA polymerase melts the DNA strands to form an open complex during tran- scription initiation. Greenblatt pointed out that the transition from the closed to the open polymerase-pro- moter complex is a target for regulation in E. coli. By analogy, regulation of the RAP 30/74 helicase activity is another potential site of transcription control by up- stream transcription factors.

In the introduction to one session, Sigler (Yale Univer- sity) drew an interesting analogy between the activation domains of transcription factors and signal sequences that target proteins to particular cellular compartments. Different classes of signal sequences that target proteins to the endoplasmic reticulum, mitochondria, or nucleus, each display a wide variety of sequences with only gen- eral properties in common. Similarly, activation do- mains have little sequence homology, but often have general properties in common such as high concentra- tion of acidic residues, glutamine, or proline. Sigler in- terpreted this to indicate that these domains may partic- ipate in low-affinity interactions that may suffice for specificity of function. Weak interactions may be suffi- cient because of the high local concentrations of inter- acting proteins brought together through DNA binding near a promoter region. Indeed weak interactions may be required so that transcription factors do not interact until they bind to a promoter region, preventing the "squelching" type of phenomenon that can be observed when they are expressed at high level (Ptashne, Harvard University). Furthermore, weak interactions are more readily reversible and better suited to control than high- affinity interactions.

Obviously, a great many significant advances were presented at this meeting, including many important studies that can not be summarized here. It is equally obvious that a great many interesting questions remain to be answered before we have a clear understanding of the regulation of eukaryotic mRNA transcription. How do the general transcription factors function in tran- scription initiation? How do gene-specific factors regu- late the activity of the general factors? How do cells dif- ferentiate between multiple factors that bind a single control element? How do different factors bound at mul- tiple sites in complex control regions interact to deter- mine the final level of transcription observed? The ad- vances presented at this stimulating meeting under- scored the significance of these basic questions.

GENES & DEVELOPMENT 155

INFORMATION FOR CONTRIBUTORS

Aims and Scope

Genes & Development welcomes high-quality research papers that address questions concerning the regulation and expres- sion of genes of any organism or virus. Publication t ime from acceptance of manuscript is between two and three months. For papers accepted subject to revision, only one revised version will be considered; it must be submitted within 2 months of the provisional acceptance.

Submission of Papers

The journal accepts papers which present original research that has not previously been published. Submission to the Journal implies that a paper is not currently being considered for an- other journal or book. It is also understood that researchers who submit papers to this journal are prepared to make available to qualified academic researchers materials needed to duplicate their research results (probes, plasmids, clones, sequences, and the like).

Contributors from North and South America, and Australia should submit their papers to:

Judy Cuddihy, Managing Editor Cold Spring Harbor Laboratory P.O. Box 100, Bungtown Road Cold Spring Harbor, New York 11724, USA Phone (516) 367-8492

Contributors from Europe, Asia, and Africa should submit their papers to:

Grahame Bulfield, Editor AFRC Institute of Animal Physiology & Genetics Research Edinburgh Research Station Roslin, Midlothian EH25 9PS, UK Phone (031) 440-2726 or 0977

Questions regarding papers should be directed to Judy Cuddihy, Managing Editor, at Cold Spring Harbor Laboratory (516-367- 8492).

Manuscript Preparation

1. General. Papers submi t t ed to Genes & Development should be concisely written. The entire paper (including tables, figure legends, references, footnotes) should be typed double- spaced on standard-sized European or American bond paper with at least 1-in (2.5 cm) margins on all four sides. Computer printouts should be of letter quality. Each page should be la- beled with the first author's name and a page number. Five copies should be submitted; at least four of these copies should have original artl A cover letter should include: (a) name, ad- dress, and telephone number of author responsible for corre- spondence regarding the manuscript; (b) s ta tement that the manuscript has been seen and approved by all listed authors; (c) any specific requirements for reproduction of art; and (d) status of any permissions needed.

2. Submitting Papers on Computer Discs. Publication can be speeded up if accepted papers are supplied on 51/4-inch floppy discs. We can accept IBM PC, AT and XT (1.2 meg or 360K floppy), or compatible, formats. Please supply the manuscript on the disc as a " text" or ASCII file, if possible. If a word-pro-

cessing file is being sent, do not use any underscoring, italic, or boldface; spell out special characters (Greek, math); use two carriage returns at the end of each paragraph, subheads, and list items. Indicate on the disc: computer brand name, whether the disc contains a text or word-processing file (name of software), and the disc format. Five hardcopy versions should also be sub- mit ted for use by referees and editors.

3. Form. The following order is preferred: Title page, Ab- stract, Introduction, Results, Discussion, Methods, Acknowl- edgments, References, Tables, Figure legends. The Title page should include: (a) title; (b) all authors' full names; (c) all affilia- tions clearly indicated; (d) a shortened version of the title for use as a running head (maximum 45 characters); and (e) key words (up to 6) for use in indexing. The Abstract should be about 200 words long and should summarize the aim of the report, the methodological approach, and the significance of the results. Methods will appear at the end of the paper and should be detailed enough to allow any qualified researcher to dupli- cate the results.

4. Figures and Legends. Five sets of figures should be sup- plied as high-quality glossy prints. Half-tones should be high- contrast, particularly in the case of gels, for the best reproduc- tion. If accurate reproduction is crucial, please describe your requirements in your cover letter so that the best possible re- production me thod can be selected. Line drawings, graphs, charts, and chemical formulae should be professionally pre- pared and labeled.

The back of each figure should be labeled with the first au- thor's name, figure number, and an indication of "top." The figures should be numbered consecutively in the order to which they are referred in the text. The size of the figures will be ad- justed to fit the Journal format, therefore please try to keep labels, symbols, and other call-out devices in proportion to the figure size and detail.

Figure legends should be brief and should not conta in methods. Symbols indicated in the figure should be identified in the legend text. If figures are reprinted from another source, permission to reprint is required.

5. Tables. Tabular data should be presented concisely and logically. Tables should be numbered consecutively according to the order cited in the text and each should have a title. Use only horizontal rules and make sure column headings are un- ambiguous in indicating columns to which they refer. Table legends and footnotes should be included where needed. If tables are reprinted from another source or if data included are from another source, permission to reprint is required.

6. References. References are name/date citations in text; please do not cite by number. Undated citations (unpublished, in preparation, personal communication) should include first initials and last names of authors. Bibliographic information should be supplied in the following order. For Journal articles: Author(s), year, article title, journal title, volume, inclusive page numbers. For books: Author(s), year, chapter title, book title, editors ' names, volume, inclusive page numbers , pub- lisher, city of publication.

7. Proofs. Proofs are considered the final form of the paper and correction can be made only in the case of factual errors. If additional information must be added at this stage, it should be in the form of "Note added in proof," subject to the approval of the editors.

8. Reprints. A reprint order form will be included with the proofs.

To help defray the cost of publication, a charge of $20 per page will be made for publication in Genes & Development. Authors unable to meet these charges should include a letter of explanation upon acceptance for publication; inability to mee t these charges will have no effect on acceptance and publication of submitted papers.

UCLA Symposia on Molecular & Cellular Biology 1990 Conferences STRUCTURE AND ASSEMBLY OF THE MYOFIBRIL Organizers: Richard Lymn, Donald Fischman and Henry Epstein January 13-16, 1990 ° Frisco, Colorado

IONIZING RADIATION DAMAGE TO DNA: MOLECULAR APPROACHES Organizers: Susan Wallace and Robert Painter January 16-21, 1990 • Lake Tahoe, California

MOLECULAR BIOLOGY OF STREPTOMYCETES Organizers: Mervyn Bibb, Richard Losick and Dean Taylor January 16-21,1990 • Frisco, Colorado

MOLECULAR BASIS OF CELL ADHESION Organizers: Richard Hynes and Erkki Ruoslahti January 20-26,1990. Steamboat Springs, Colorado

CYTOSKELETON AND CELL REGULATION Organizers: Kermit Carraway, Elias Lazarides and Ron Vale January 20-26,1990 ° Steamboat Springs, Colorado

NEW STRATEGIES IN BONE MARROW TRANSPLANTATION Organizers: Richard Champlin and Robert P. Gale January 20-27,1990 ° Keystone, Colorado

MANIPULATING THE MAMMALIAN GENOME Organizers: Alan Bernstein, C. Thomas Caskey and Dusty Miller January 21-28,1990. Tamarron, Colorado

MOLECULAR PATHWAYS OF CYTOKINE ACTION Organizers: Stanley Cohen and Charles Evans January 27-February 3,1990 ° Park City, Utah

CELLULAR IMMUNITY AND THE IMMUNOTHERAPY OF CANCER Organizers: Olivera Finn and Michael Lotze January 27-February 3, 1990, Park City, Utah

TRANSCRIPTIONAL CONTROL OF CELL GROWTH Organizers: Kathryn Calame and Leonard Guarente January 27-February 3,1990 ° Keystone, Colorado

RECEPTOR MEDIATED SECOND MESSENGER PATHWAYS Organizers: Harvey Herschman and Lewis T. Williams January 27-February 3, 1990. Keystone, Colorado

MOLECULAR, BIOCHEMICAL AND CELLULAR BIOLOGY OF HUMAN BREAST CANCER Organizers: Marc Lippman, Robert Dickson and Dennis Slamon February 3-8, t990 • Tamarron, Colorado

BIOCHEMISTRY AND MOLECULAR BIOLOGY OF BIOSENSORS AND BIOPROBES Organizers: Joseph Andrade, Daphne Kamely and James Lear February 3-8,1990. Frisco, Colorado

GENETIC AND IN VITRO ANALYSIS OF CELL COMPARTMENTALIZATION Organizer: David Meyer February 3-9,1990 • Taos, New Mexico

POLYAMINES IN CELL PROLIFERATION AND DIFFERENTIATION Organizers: Laurence Marton, David Morris and Anthony Pegg February 9-14, 1990 • Taos, New Mexico

SIGNAL TRANSDUCTION BY G PROTEINS Organizers: Lutz Birnbaumer and Ravi lyengar February 9-15, 1990 • Tamarron, Colorado

MOLECULAR BIOLOGY OF YERSINIAE Organizer: Saumya Bhaduri February 22-27, 1990 • Frisco, Colorado

THE INORGANIC CHEMISTRY/MOLECULAR BIOLOGY INTERFACE Organizers: Jeremy Berg, Thomas O'Halloran and Thomas Tullius February 24-March 1, 1990 • Taos, New Mexico

SYNTHETIC PEPTIDES: APPROACHES TO BIOLOGICAL PROBLEMS Organizer: James Tam February 27-March 4, 1990. Frisco, Colorado

NEGATIVE CONTROLS ON CELL GROWTH Organizers: Harold Moses and Robert Weinberg March 3-9, 1990 • Taos, New Mexico

AIRWAYS INFLAMMATION Organizers: Stephen Rennard March 5-11, 1990 • Tamarron, Colorado

BIOLOGY OF SARCOMAS Organizers: Beverly Emanuel, Joseph Madri and Richard Womer March 11-16, 1990 • Lake Tahoe, California

PROTEIN PURIFICATION AND BIOCHEMICAL ENGINEERING Organizers: Edward Cussler, William Drohan and William Velander March 19-25, 1990 • Lake Tahoe, California

ANIMAL MODELS OF HUMAN VIRAL DISEASES: RELEVANCE TO DEVELOPMENTAL THERAPEUTICS Organizers: John Blasecki, Catherine Laughlin and John McGowan March 31-April 5, 1990 • Keystone, Colorado

HIV AND AIDS: PATHOGENESIS, THERAPY AND VACCINE Organizers: Samuel Broder and Flossie Wong-Staal March 31-April 6,1990. Keystone, Colorado

B LYMPHOCYTE DEVELOPMENT Organizers: Max Cooper, Roger Perlmutter and Irving Weissman March 31-April 6, 1990 • Park City, Utah

RECEPTOR-MODULATED TRANSPORT SYSTEMS Organizer: Michael Czech March 31-April 6,1990. Steamboat Springs, Colorado

GROWTH AND DIFFERENTIATION FACTORS IN DEVELOPMENT Organizers: Rik Derynck and Brigicl Hogan March 31-April 7, 1990 • Steamboat Springs, Colorado

SIGNAL TRANSDUCTION AND GENE ACTIVATION IN DEVELOPMENT Organizers: Richard Firtel, Judith Kimble and M. Geoffrey Rosenfeld March 31-April 7,1990. Steamboat Springs, Colorado

THE ENDOTHELIAL CELL Organizers: Peter Lelkes and Thomas Maciag April 6-12, 1990. Keystone, Colorado

TISSUE ENGINEERING Organizers: Richard Skalak, Randall Swartz and C. Fred Fox April 6-12, 1990 • Keystone, Colorado

MOLECULAR STRATEGIES FOR CROP IMPROVEMENT Organizers: Charles Amtzen, James Peacock and Marc Van Montagu April 16-22, 1990 • Keystone, Colorado

CELL AND MOLECULAR ASPECTS OF THE DEVELOPING NERVOUS SYSTEM Organizers: Theodore Stotkin and lan Zagon April t7-23,1990 • S. Padre Island, Texas

MOLECULAR BIOLOGY OF NEUROTRANSMITTERS AND THEIR RECEPTORS Organizers: Robert Steiner and Stanley Watson April 17-23,1990 • S. Padre Island, Texas

MOLECULAR NEUROBIOLOGY Organizers: Tom Curran and James Morgan April 17-23,1990. S. Padre Island, Texas

NEUROTROPHIC FACTORS Organizer: Ralph Bradshaw and Dennis Cunningham April 17-23, 1990 • S. Padre Island, Texas

. . . . . . ~ m . . . . . . . . . . . . m m m u ~ m ~ m

F - - (Please Detach and Return) UCLA Symposia on Molecular & Cellular Biology

INFORMATION REQUEST FORM Please send me Program and Application Information on the following meeting(s)

Name (Last, First, Middle Initial): Department: Address (Street/P.O. Box):

Institution:

City: Previous Address:

State: Zip Code: Country:

Business Telephone: Telefax: Please mail to: UCLA Symposia, 2032 Armacost Avenue, Los Angeles, CA 90025

Telephone: (213)207-5042 Telefax: (213)207-2397 Telex: 3716012

Reader Serv i ce No. 338

*Cancer