PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or betore date due. DATE DUE DATE DUE DATE DUE MSU Is An Affirmative Action/Equal Opportunity Institution ANALYSIS OF THE CHICKEN ERYTHROID- SPECIFIC H5 HISTONE GENE BY Paul Llewellyn Boyer A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Biochemistry 1988 5454/K' ABSTRACT ANALYSIS OF THE CHICKEN ERYTHROID-SPECIFIC H5 HISTONE GENE BY Paul Llewellyn Boyer The chicken erythroid-specific H5 histone gene has been isolated from a phage library and analyzed. It was found that the chicken H5 histone gene is not closely linked (within 20-30 kb) to any other known histone gene and that only one H5 histone gene exists per haploid chicken genome. The H5 histone gene was further studied by in vitrg muta- genesis. The linker scanner mutagenesis procedure of Mc- Knight and Kingsbury was utilized to construct both linker scanner mutants and deletion mutants within the putative chicken H5 histone promoter region. The deletion mutants were studied by transfection into QT6 quail fibroblasts. RNA was isolated from the transfected cells 36-48 hours after the transfection was completed and subjected to 81 analysis. Deletions within the promoter region have iden- tified a number of possible H5 promoter elements. The H5 transcript steady state levels for each mutant was normal- ized to the steady state levels of the wild type H5 gene by utilizing an internal control mutant H5 gene. As a number of putative promoter elements were deleted, the H5 tran- script steady state levels decreased, relative to the steady state levels for the normal as gene. A mutant which deletes the 3' flanking region of the H5 gene was also found to have an effect on the steady state levels of the H5 transcript. Since this mutant has an intact promoter region, it is felt that this mutant has altered the processing of the H5 transcript. To Mom and Dad iv W I would like to thank "the Boss", Jerry Dodgson, who put up with such minor details as rose bushes on the window- sill, ship models on my desk, and a grad student who only worked at night. Thanks to Horiko Ito and Sue Kalvonjian for being friends without equal, and who let me ramble on about my hobbies without telling me to shut up. Thanks to Jiing Dwan Lee, truly someone stranger than I Thanks to Dr. Richard Schwartz who proofread this thesis in a record time. And finally, thanks to JoEllyn McGrath who made the last few months here much more enjoyable. I will miss you more than you know. TABLE OF CONTENTS LIST OF TABLES............................... ..... .... Vii LIST OF FIGURES....................................... viii LIST OF ABBREVIATIONS................................. ix INTRODUCTION.......................................... 1 CHAPTER 1 Literature Review Histones......................................... 3 Organization of Histone Genes.................... 11 Eukaryotic Gene Regulatory Elements.............. 16 Erythropoeisis................................... 21 H5 Histone Protein............................... 24 H5 at the mRNA Level................ ..... ........ 29 H5 at the DNA Level.............................. 32 Bibliography............... ..... . ................ 41 CHAPTER 2 materials andMethOdSOOOOOIOOOOO0.0.0.000...00...... 56 CHAPTER 3 Results Isolation of H5 Lambda Phage Clones....... ...... . 66 H5 Plasmid Subclones............................. 72 Construction of Linker Scanner Mutants........... 75 References....................................... 94 CHAPTER 4 Results Mutational Analysis of the H5 Histone Promoter... 95 Discussion....................................... 122 References....................................... 127 CHAPTER 5 Results Methylation of the HS gene....................... 129 Summary.......................................... 139 H5 Gene Restriction Enzyme Site Polymorphisms.... 139 Discussion....................................... 144 References....................................... 148 V1 CHAPTER 3 CHAPTER 4 LIST OF TABLES EAQE Regulatory elements deleted 103 by various mutants Areas under densitometer scan 116 curves Recognition sequences for CfoI, 130 HpaII, and MspI Predicted bands from restric- 136 tion enzyme digestion Lysozyme dissimilarity 145 Albumin and transferrin dis- 145 similarity CHAPTER 1 scauaw CHAPTER 2 0‘0! 13 14 15 CHAPTER 3 16 17 18 19 20 21 22 CHAPTER 4 23 24 25 26 LIST OF FIGURES Structure of the Nucleosome.......... Histone Gene Arrangments............. Chicken Erythropoiesis............... Sequence of the H5 Gene.............. H5 Phage Clone Maps.................. Southern Blot of HindIII............. Digested Chicken DNA H5 plasmid Subclone Maps............. Construction of 5’ to 3’ mutants..... Construction of 3’ to 5' mutants..... Maxam-Gilbert Sequencing............. HindIII Linker Position for "S"...... Series clones, -130 to -40 HindIII Linker Position for "S"...... Series clones, -50 to +40 HindIII Linker Position for "C"...... Series Clones, -150 to —50 HindIII Linker Position for "C"...... Series Clones, -60 to +30 Scheme for Constructing Linker. ...... Scanner Mutants Duck H5 vs. ChiCken HSOOOOOOOOOOOOOIO Mapping the H5 CAP Site.............. Location of Deletion Mutants......... Location of Linker Scanner Mutants... Expected Fragment Sizes From 81...... Analysis 81 Analysis of Mutant H5 mRNA........ Sample Densitometer Tracing.......... Methylation of H5, Southern Blot..... Methylation of H5, Southern Blot..... Mg Histone, SacI Polymorphisms....... a Globin, SacI Polymorphisms......... viii 5 13 22 35 69 71 73 77 78 82 85 87 89 91 93 97 99 105 108 111 115 118 132 134 142 143 LIST OF ABBREVIATIONS bp.................................. kb.................................. LTR................................. A................................... G................................... TOOOOOOOOOOOOOOOO 000000000000 0...... ix base pair kilobase Long Terminal Repeat Adenine Cytosine Guanine Thymidine Introduction The chicken erythroid-specific H5 histone gene has a num- ber of unusual characteristics. The H5 histone gene is not closely linked to any other known histone gene, and the H5 transcript is polyadenylated. While the H5 histone gene is considered to be a member of the H1 class of histones, many of the H1 histone promoter elements have been replaced in the H5 histone promoter. What regulatory elements are responsible for the expression of the H5 histone gene are uncertain. The isolation of phage clones containing the chicken gen- omic H5 gene is described in Chapter 3. Study of the iso- lated phage clones indicated that all were isolates of the same H5 gene, and that the H5 histone gene was unlinked to other known histone genes. The H5 gene was subcloned into the plasmid pBR322. The promoter region of the H5 gene was studied by in_yi§zg mutagenesis. Several hundred deletion clones were obtained from the Bal 31 digestions. The del- etions’ boundaries were roughly determined by restriction enzyme digestion. The exact deletion boundaries for 106 mu- tants have been determined by sequencing of the clones. The sequencing data allowed the construction of deletion mutants whose boundaries were known to one base pair. The various mutants that were constructed were analyzed by transfection into QT6 quail fibroblasts (Chapter 4). RNA was isolated from the cells 36-48 hours after 1 2 transfection. The level of H5 transcripts within the assay was determined by $1 analysis. A mutant gene, which has a small deletion 3’ of the transcriptional start sites, was used as an internal control. The H5 transcript level of the various mutants were normalized to the expression level of this internal control gene. The effects of the various mutants are described in this chapter. The last chapter (Chapter 5) includes two other stud- ies of the H5 histone gene. The first part of the chapter studies the degree of DNA methylation at various stages of erythropoeisis. DNA from normal and RAV-l infected eryth- roblast and erythrocyte cells were digested with the re- striction enzymes CfoI, MspI, and HpaII. These enzymes differ in their sensitivity to DNA methylation in the form cmec . The second portion of Chapter 4 covers SacI restriction enzyme site polymorphisms for both the H5 histone gene and the aA-globin gene. DNA from 14 inbred chicken lines as well as DNA from the domesticated turkey, the Japanese Quail, and the ring-necked pheasant were tested. Chapter 1 mm 81532132: Histones are a group of small basic proteins that inter- act with DNA within the nucleus to form the elemental sub- unit of chromatin structure, the nucleosome (97). There are five major classes of histones based upon their electro- phoretic mobility: the core histones (H2A, H28, H3, and H4), and the linker histone, H1 (73). An unusual histone variant, H5, is considered to be a member of the H1 class of histones and will be discussed in more detail later. All of the core histones show a similar organization in their protein sequence. Most of the charged amino acids are present in the amino-terminal end of the protein while the carboxy-terminal end is hydrophobic in nature (73). This may indicate that the core histones share a common ancestral gene (55). The H1 histones have charged amino acids (mostly lysine) in both the amino- and carboxyl-ter- minals while the central region is apolar (73). By protein sequence analysis, the histones are evolution- arily well conserved across species barriers (42,118,147). H2A and HZB (the slightly lysine rich histones) show the most variability among the core histones while H3 and H4 (the arginine rich histones) show little variation. The changes that do occur in the core histones are usually seen 3 4 in the hydrophilic amino-terminus rather than in the hydro- phobic carboxyl-terminus (73). This is probably due to the function of the core histones within the nucleosome, and will be discussed below. H1 histones (including H5) are very lysine rich and show the most variation of all the histones. Even so, H1 his- tones are well conserved evolutionarily. H1 histones have charged amino acids in both the amino- and carboxyl-terminal ends and have an apolar central region. Most changes in the protein sequence occur in one of the two hydrophilic termini rather than in the central region (73). The histones were known to complex with DNA but the ex- act structural role of the histones was unclear until the discovery of the nucleosome (83,97), which is the basic subunit of chromatin structure. By various techniques (58, 89,105), the nucleosome was found to consist of a histone octamer core, containing an H3:H4 tetramer and two H2A:H2B dimers (Fig. 1). The basic amino-terminal ends of the core histones form the external surface of the nucleosome while the apolar carboxyl-terminal ends interact with each other inside the core (97). Wrapped twice around this histone core is 140 base pairs (bp) of DNA. Between two adjoining nucleosomes is anywhere from twenty to eighty base pairs of DNA. H1 (and H5) histones appear to bind in this "linker" region. With their two basic ends, they are believed to m €19 DNA ".nks'..0_NA Subiect to rapid Length 140 be; Length 60 hp; end trimming by attacked at may vary in units oi 10 bp: micrococcal intervals at about half is susceptible nuclease; this 10 by DNAue l to first cleavage by remaining H1 Figure 1 microcoml nucieese Structure of the nucleosome. The histone octamer consists of an H3:H4 tetramer and two H2A:HZB dimers. 140 base pairs of DNA is wrapped twice around this core. The H1 histone family (includ- ing histone H5) binds to the linker region between two adjoining nucleosomes. Figure from Lewin (93) 6 "seal" the nucleosome ends where the DNA enters and leaves. A second function assigned to the H1 histones is in the formation of higher order chromatin structure. The H1 his- tone protein appears to be essential for forming the 30nm fiber, which consists of six nucleosomes arranged in a coil (73). This fiber itself can be further compacted to form even more complex structures whose molecular details are as yet unclear. As discussed below, the H5 histone also is responsible for the assembly of higher order chromatin structures. The replacement of H1 histone with the H5 his- tone protein within the erythroid chromatin condenses the chromatin, and may play a role in rendering the mature ery- throcyte transcriptionally inactive. Though the histones are usually grouped within the five major classes (H1, H2A, H2B, H3, H4), there exist variant subtypes of histone proteins within these classes. One type of histone variant is that of the modified histones. Modified histones are histones that have undergone post- translational chemical modifications. The most common such modifications are listed below. ,Methylation is usually found on certain lysine groups for both H3 and H4 and appears to be reversible (73). Increased levels of methylation may be detected during late S and G2 phases of the cell cycle (165). The function of this modification is unknown. Acetylation may occur both on the N-terminal serine resi- due of the histones H1, H2A, and H4 and on lysine residues within the core histones (H2A, H2B, H3, and H4). The ter- minal N-acetyl-serine group appears to be a stable, irre- versible modification but its function is unclear. In con- trast, the internal acetylations are reversible through the action of the enzyme deacetylase (18). H2B, H3, and H4 have four modification sites while H2A has only one site (73). The core histones are acetylated soon after they are syn- thesized but not all possible sites are modified. In trout H4 for example, only two of the four possible sites are modified on the newly synthesized protein. The addition of the third and fourth acetyl groups procedes more slowly. Shortly after the trout H4 protein has been fully acetyl- ated, removal of these groups begins by the action of the enzyme deacetylase until the protein is completely unmodi- fied or has only one acetyl group remaining. This cycle takes around one day to complete (24,95). A somewhat sim- ilar series of acetylation/deacetylation reactions occurs for the other core histones (25,26). This cycle of acetyl group addition and removal creates heterogeneity in the in- ternal extent of modification. Calf thymus histone H4 for example, has one of its modification sites (lysine 16) ace- tylated only 60% of the time and HeLa cell H4 histones are 8 acetylated only 40% of the time (93). There is evidence that hyperacetylated histones congre- gate in certain discrete regions of the genome (32). A number of these regions are DNase I sensitive (114,148). This is of interest since DNase I sensitivity has been shown to occur in those areas of the genome that are trans- criptionally active or potentially active (152,155). It is thought that the hyperacetylation of the histones affects their interactions with the DNA and allows the chromatin to relax, but this is conjecture (2). Phosphorylation may occur for all the histones and usu- ally is found on specific threonine groups as well as on histidine (H4) and lysine (H1 and H5). This modification is found mostly in the S phase of the cell cycle but may also occur under hormonal induction. The modification is usually lost after the anaphase stage (73). The phosphor- ylation of H5 does not appear to be cell cycle related, but is related to the stage of maturation of the erythroid cell (140,141). Addition of a poly(ADP-ribose) moiety to the glutamate or aspartate amino acids of H2A, H2B, H3, and H1 is thought to affect the interactions between the histones but this is uncertain (73). Attachment of the protein ubiquitin to lysine #119 of 9 H2A has also been found. Ubiquitin, as the name implies, is found both in prokaryotes and all eukaryotes (61). The ubiquitin:H2A complex is designated A24 and the cellular levels of this complex is thought to be related to the cell’s mitotic activity (55). Histone variants may also be grouped according to the timing of their synthesis (162,163). The major class (for which most histone genes isolated to date code) is that of the replication-dependent histones. The transcripts coding for these histones are most abundant in rapidly dividing tissues. These histone gene transcripts are absolutely de- pendent on DNA synthesis for their production and they de- grade rapidly once DNA synthesis has halted. These histones are produced only during the S phase of the eukaryotic cell cycle. A second group of histone variants is that of the par- tially replication-dependent histones. These were discov- ered during studies on regenerating mouse liver. These histones are produced during the initial stages of S phase but unlike the first class of histones, these continue to be synthesized after DNA replication has ceased. A third group of histone variants is that of the repli- cation-independent histones, also known as the replacement histones. These histones gradually accumulate during cell 10 maturation and replace a portion of the replication depen- dent histones within the chromatin of non-replicating cells such as liver, kidney, and erythrocytes (162,163). The best known examples are the H3.3 variants (23,53), histone H5, HZAF (70), and H1° from mammals (137). Messenger RNA coding for these histones is synthesized constitutively at a low level regardless of the replication state of the cell. Histones may also be classified as to the stage at which they appear in the organism's maturation or development. This classification would include such stages as embryogen- esis, spermatogenesis/oogenesis, and tissue specific matur- ation. One example is the development of the sea urchin embryo which undergoes rapid cell division. The first se- veral cell divisions use histones translated from stored maternal mRNA (19,136), but after this the histones are synthesised from the early histone genes (128,142). As the embryo matures, expression from the early genes is repress- ed and instead, the late histone genes are used (33,116). There is no evidence however, for such a developmental use of different histone gene sets in avians or mammals. 11 MW Since sea urchins produce large quantities of histone mRNAs during their maturation, it was likely that there were many histone genes within their genome. This made them a likely target for isolating the first histone genes. Sea urchin species that had evolved separately for millions of years had very similar organizational patterns for their histone genes (81). It was originally believed that most other organisms would share this arrangement or something similar. As more species have been studied however, it has become apparent that the organizational pattern of the sea urchin histone genes is unique. It reflects their need for large amounts of histone synthesis in a short period of dev- elopment rather than a standard motif (Fig. 2). Sea urchins have organized the five major early histone genes into a quintet that is tandemly repeated several hun- dred times (81). The order of the histones within the quin- tet (5' to 3’ relative to transcription) is H4-H2B-H3-H2A- H1. They are all transcribed from the same strand of DNA, but there is no evidence that a polycistronic message is made. Each early histone gene appears to have its own pro- moter elements (76). Each gene within the quintet is sep- arated from the flanking genes by a stretch of DNA that is AT-rich which appears to contain the needed regulatory ele- ments. Figure 2 12 Histone gene arrangements in selected organisms. The identity of each histone gene is shown, as well as the direction of transcription (if known). The map for the chicken histone genes identifies 35 out of the 42 histone genes mapped to date. The histone genes were isolated from two separate chromosome regions. The scale for the chicken histone gene map is half that of sea urchin and Drosophila histone gene maps. The histone gene quintet of the Drosophila is roughly 5.0 kilo base pairs in length and is tandemly repeated. The histone gene quintet of the sea urchin covers a span roughly 6-7 kb in length and is also tandemly repeated. Figures from S. Dalton, Ph.D. thesis, University of Adelaide, Adelaide, S. Australia 13 z. .3 :8 :u. :m» 633626 1. :u .3») g 1.. :3 E :u :8 z. z... :u .3 :~» 1 > :6 r. IN) I. 2&0 El: 13%.. 19:3 3.18.3.3»: as» 5» :u 2.2. 1. Figure 14 The sea urchin late histone genes have been found to be present in only 5-12 copies per haploid genome, and they are organized in clusters of irregular pattern similar to that of the vertebrate histone genes discussed below (96). It was also discovered that a third class of sea urchin histone genes existed. These genes, termed orphons, appear to be solitary genes unlinked to any of the other histone genes. It is uncertain if some of these are transcription- ally active or if they are all psuedogenes (30). Using probes made from the sea urchin early histone genes, the histone genes for Drosophila melanogaster were isolated (80). The organizational pattern of these genes shows similarities to that of the sea urchin but with dis- tinct differences (Fig. 2). The order of the Drosophila histone genes is H1-H3-H4-H2A-HZB, and the genes are not all transcribed from the same DNA strand. This quintet is, like the sea urchin quintet, tandemly repeated but Droso- phila has two distinct repeat units instead of one (94). Flies have been bred that lack one or the other repeat unit with no apparent deleterious effects (108,156). It appears as though either type of repeat unit is sufficient to main- tain the required level of histone. Why two sets are main- tained and whether they actually are completely equivalent 15 is uncertain. In higher organisms that have a reduced number of his- tone genes (compared to either the sea urchin or Droso- phila), the regularity of the organization pattern de- creases. Both birds (39,51,69,139) and mammals (28,134, 133,163) have irregular (i.e.no tandem repeat unit) clus- ters of histone genes. In the chicken there are two such clusters which vary in both content and in organization (51,139). There is no common order to the closely clus- tered chicken histone genes, although several examples of a few genes linked in an inverted repeat fashion can be noted. Furthermore an H2A gene is often but not always paired with an H2B gene in such a way that they are trans- cribed in opposite directions. It also appears that H3 genes tend to pair with H4 genes though this is not as com- mon. The two clusters in chicken contain replication de- pendent and partially replication dependent histone genes. The replication-independent histone genes (mentioned above) appear to be unlinked to any other histone gene. This may be due to the differences in the way the replication-inde- pendent histone genes are expressed. 15 WW Gene expression in eukaryotic cells may be modulated at a number of points. Transcriptional control is the most common form of regulation but modulation may also occur during the course of mRNA processing (removal of interven- ing sequences and addition of the CAP structure and the polyadenylate tail), in the transport of the mRNA from the nucleus to the cytoplasm, in the stability levels of the mRNA, during the translation of the mRNA into the protein product, and during post-translational modification of the protein (40,115). Transcriptional regulation controls both the frequency and the timing (i.e., tissue specificity, hormone inducibil- ity, etc.) of transcription from the gene. This regulation results from the interactions between specific promoter ele- ments on the DNA with cellular factors. One of the first promoter elements identified was the "TATA" box. This sequence (consensus sequence 5’-TATA§A§- 3’) is highly conserved and is usually found 25 to 30 base pairs upstream from the transcriptional start site (36). This sequence appears to regulate the location of the mRNA start site since removing or mutating the "TATA" sequence reduces the number of correctly initiated transcripts (65, 149). Deletion mutations have indicated that the spacing 17 between the ”TATA" box and the transcriptional start site is important (20). The "TATA" sequence also appears to be needed for maximal mRNA transcription (65,99). A second common promoter element is the "CCAAT" box. This sequence (consensus 5’-GGPyCAATCT-3') is generally lo- cated in the region between 40 and 90 base pairs upstream of the mRNA start site (12) and has been shown to interact with other promoter elements (15,66). Unlike the "TATA" box, the "CCAAT" sequence seems to be involved only in re- gulating the levels of transcription, not in the selection of the transcriptional start site (44,100,106). A third common promoter sequence is the G/C box (consen- sus 5'-GGGGCGGGGG-3’) (79). This sequence is usually found in the region between 40 and 200 base pairs upstream from the transcriptional start site and is often present in mul- tiple copies. This element does not have any obvious se- quence symmetry but nevertheless, this element can function in either orientation (47,59). Like the "CCAAT" sequence, the G/C box is involved in the efficient transcription of the gene (12,29,56). Proteins from cellular extracts have been found that bind to the "TATA" box (127,131), the "CCAAT" box (74,75), and the G/C box (22,47). While little is known about the mechanisms involved, it is known that interactions between 18 the ”TATA" box and its factor(s) (41), between the "CCAAT" sequence and its factor(s) (e.g., "CCAAT” transcription factor or CTF) (75,101), and between the G/C sequence and its factor (designated SP1) (101) are required for either accurate initiation of the transcript or for efficient transcription. Along with these general promoter elements are gene specific regulatory sequences and enhancers. Gene specific elements are usually detected by mutational analysis, gene fusions, or simply by comparing the DNA sequences of re- lated genes. Examples of this would be the detection of an H1 specific sequence found in all H1 genes studied to date (34), and regulatory elements common to cAMP-regulated genes (35). These gene specific elements combine with the general promoter elements already described to provide cor- rect regulation of gene expression. Each class or type of histone gene studied to date, usually has highly conserved, type-specific elements in the regulatory region of the gene. Along with these other elements are enhancers. An en- hancer was first discovered in the SV40 virus (13,67). En- hancers have several unusual features that differ from most other control elements. Enhancers are usually orientation and position independent (109) and may exert their influ- ence over long distances (91). Enhancers may activate genes with heterologous promoters and as the name suggests, 19 expression of a gene is higher in the presence of an enhan- cer than in its absence (8,27). Enhancers were found in many other eukaryotic viruses besides SV40. While large stretches of sequence homology have not been detected (72), the enhancers are often functionally equivalent (92). A common feature of viral enhancers is the presence of tan- demly repeated elements (43,120,150), though not all virus- es, notably polyoma, have these (72). Viral enhancers also tend to show host cell specificity (43,88). For example, replacing the natural enhancer of polyoma (which infects mouse cells) with the SV40 enhancer (which infects primate cells) changes the host cell prefer- ence of polyoma from mouse cells to primate cells (43). Similar cell type preferences have been noted for other viruses (110,120,132). Since enhancers are widely used in viruses that infect eukaryotic cells, cellular genes were examined to determine if eukaryotic cells also use enhancers as a means gene reg- ulation. A number of cellular enhancers have indeed been identified. The first cellular enhancer detected was the enhancer from the mouse immunoglobulin heavy chain locus. This enhancer has structural and functional similarities to the viral enhancers (9,60,102), and stimulates trans- cription when fused to heterologous genes (60). Interest- 20 ingly, this stimulation is strictly lymphoid-specific (9, 60). Several more enhancer sequences have since been discov- ered and many more probably exist. Examples include the pancreas-specific enhancer from the rat elastase I gene (117), the pancreas specific enhancer from the rat insulin II gene (68), the rat prolactin and growth hormone enhancer (113), the chicken p-globin enhancer (which is located in the 3' flanking region of the gene) (31,78), and the chic- ken histone H5 enhancer which also is present in the 3' flanking region of the gene (146). 21 W Erythropoiesis is the process by which an erythroid cell differentiates from a pluripotent stem cell to a non- dividing, transcriptionally inactive erythrocyte (Fig. 3). During this process, various tissue specific genes must be activated so that the erythrocyte can function as the oxy- gen carrying cell of the organism. These include those genes that encode enzymes for the manufacture of the heme ring (62,159), the genes for the a- and fi-globin protein subunits of hemoglobin (45,50,52), the carbonic anhydrase II gene (57,161), various erythroid antigens and membrane proteins (82,112,125,153), cytoskeletal proteins (54,90, 160), iron metabolism proteins, and the erythroid specific histone H5. After the precursor stem cell, the first identifiable erythroid cell type are the colony forming unit-marrow cells (CFU-M). These cells are self-renewing and can col- onize the bone marrow of irradiated chickens, hence the name (124,125). Cells of this type continue to to repli- cate indefinitely in the self-renewal mode or they may ir- reversibly commit to terminal hematopoietic differentia- tion. After the CFU-M stage, the cell enters the burst forming unit-erythroid (BFU-E), which is then followed by the Pluripotent Stem Cell CPU—M BFU-E OFU-£ lrythroblast Reticulocyte Erythrocyte Figure 3 Pathway of chicken erythropoiesis. The CPU-M stage may either continue to self-replicate it— self or irreversibly commit to further differen- tiation. The "X" between the CPU-E and erythro- blast stages indicates the differentiation step blocked by the Avian Erythroblastosis Virus (AEV). Abbreviations: CFU-M-QColony forming unit-Marrow. BFU-E-9 Burst forming unit-erythroid. CFU-E-D Colony forming unit-erythroid (124.125.126.164) 23 colony forming unit-erythroid (CFU-E) stage (126). These cells are defined by their potential to produce different numbers of cells when cultured in_yit;g. Bursts may be distinquished from colonies in that bursts have a large number of cell groups clustered together (usually 3-20 groups) with each group containing 8-60 erythroid cells. Colonies on the other hand are a single group with 8-150 cells within the group. The two cell types may also be identified by their antigen expression and their response to growth factors. CFU-E cells show an absolute require- ment for erythropoietin before they will continue to dif- ferentiate while BFU-E cells do not have this requirement. After the CFU-E stage is the erythroblast stage. These cells are the last erythroid cells capable of cell division, and hemoglobin production begins at this stage. The cell size begins to decrease and the nucleus begins to condense as the erythroblast matures through the reticulocyte stage and finally into the erythrocyte stage. Reticulocytes re- tain mRNA and protein synthetic activities while little or no synthetic activity occurs in the mature erythrocytes. As described below, it has been hypothezised that H5 his- tone may play a role in the differentiation process. 24 W31 H5 histone (previously designated V or f2c) was origin- ally detected as an additional band during gel electrophor- esis of histone proteins. It was only seen in histone ex- tracts from mature avian erythrocytes and not from mammal- ian histone extracts. H5 was found to be in all nucleated erythrocytes tested, i.e., reptiles, fish, amphibians, and birds (103,111). The presence of H5 can be detected in the earliest testable precursors including both adult and embryonic erythroid cells arrested at the CPU-E stage by a temperature-sensitive avian erythroblastosis virus (14,un- published data). It is also the only histone produced by the nearly mature, non-dividing reticulocyte (5). The levels of H5 are low in the precursor cells, lower than the levels for H1, but H5 gradually accumulates as the cell ma- tures. As the concentration of H5 increases, H5 tends to displace H1 from the chromatin (6,121), until the concen- tration ratio of H5 to H1 in the chromatin is roughly 2:1 (10). This replacement correlates with the gradual compac- tion of the erythroid chromatin (143,144) and probably plays a role in the decreased levels of transcription seen in the reticulocyte and erythrocyte (151). In this respect it resembles the mammalian H1 histone variant H10. The mammalian histone however, is not tissue specific like H5 25 and is found in most adult non-dividing cells (137). H5 histone is thought to function in rendering the nuc- leus transcriptionally inactive but uncertainty remains as to why it can be detected so early in the differentiation process when such a function would not seem to be needed until the late reticulocyte stage. It is thought that the concentration of the H5 protein in the earlier dividing cells remains below a certain threshold level which allows transcription to proceed. It appears as though the concen- tration of H5 protein within the erythroid cell does not markedly increase until the late erythroblast stage (1). The level of post-translational modification may also play a role. Five sites have been found in the protein that may be phosphorylated. Two of these sites are in the amino terminal end, while the remaining three are in the car- boxyl-terminal end. All predicted phosphopeptides may be detected in a tryptic digest of monophosphorylated H5 pro- tein, indicating that phosphorylation at any given site is a random event. The fact that the phosphorylation sites are present in the basic amino- and carboxyl-terminal ends suggests that the phosphorylation event may interfere with the DNA binding potential of the H5 histone (141). It has been determined that the phosphorylation levels of the H5 histone decreases as the erythroid cell matures. Early ery- 26 throid precursor cells isolated from chicken bone marrow have 70% of their H5 histone protein phosphorylated while erythroblasts and reticulocytes isolated from anemic chic- ken blood have only 50% of their H5 histone proteins modi- fied. In comparison, mature erythrocytes isolated from nor- mal chicken blood have little or no detectable phosphoryla- tion. It has been hypothesized that the H5 phosphorylation present in the early precursor cells prevents the H5 his- tones from binding to the chromatin and displacing the H1 histones. As the cells mature, this modification is grad- ually removed and the H5 histone may bind to the chromatin (140). From the protein sequence, H5 is usually considered to be a member of the H1 family of histones. The H5 protein has partial sequence homology with the H1 protein and is similar in its primary structure (i.e. charged amino acids in the amino- and carboxyl-terminal regions and an apolar central region). Like H1, H5 binds to the linker region between adjacent nucleosomes (7,145,157). When the protein sequences for the chicken H5 histone (21) and the goose H5 histone (158) are compared to the protein sequences of var- ious H1 proteins (rabbit, trout, calf and sea urchin), the similarities suggests that these two proteins were derived from a common ancestor. It is apparent though, that these 27 two proteins have been evolving separately for a lengthy period of time. It is uncertain when the evolutionary di- vergence began, but since fish, reptiles, amphibians, and birds all have been shown to contain H5 histone proteins, it is probable that the separation occured relatively early in animal evolution (3,138,158). H5 consists of two variant forms that differ only at amino acid #15. This amino acid may either be glutamine or arginine (21). Since one of the codons for glutamine is CAG and one of the codons for arginine is CGG, it was felt that a single base pair change was responsible for the two forms of H5 and that the two H5 proteins are different al- leles of the same H5 gene. This was shown to be the case when the H5 gene was sequenced (discussed below). H5 is not as well conserved between species as most other histone proteins. There is only 84% homology between the amino acid sequence of chicken H5 histone and the se- quence of either the duck H5 histone or the goose HS his- tone (21, unpublished observation). Besides these three complete H5 protein sequences, fragments of the pigeon and quail H5 proteins have also been sequenced (130,157). Com- parisons between these various H5 proteins indicate that most changes occur in the amino-terminal end with relative- ly few changes in the carboxyl-terminal or central regions 28 (21). Unfortunately, only avian H5 proteins have been se- quenced to date. Analysis of H5 histone proteins from fish and frogs however, indicates that these proteins have amino acid compositions similar to that of the chicken H5 protein (48,103). 29 H§_a§_§hs_mBHA_Lexsl All chicken replication dependent and partially replica- tion dependent histone genes lack introns and are not poly- adenylated (17,87,119). These genes are transcribed into a precursor RNA which is then processed to generate the cor- rect 3' terminal end. This process utilizes a conserved stem loop motif (77) and a U7 snRNP complex (129) to remove the unneeded 3’ portion of the RNA. The H5 histone gene does not have introns, and the H5 transcript was found to undergo the 3' terminal polyadenylation common to most other cellular mRNAs (107). Other non-replication dependent chicken histone transcripts, including transcripts from the two known H3.3 variant genes (23,53), and from the H2AF gene (70), are also polyadenylated rather than contain the stem loop sequence. H5, as described above, is one of the replication inde- pendent histones. When the cell enters the S phase of the cell cycle, the mRNA levels of replication dependent his- tones increase 10-20 fold. As the cell leaves the S phase and enters the G2 phase, these mRNA levels rapidly drop off. The increase in mRNA levels can be blocked by use of DNA synthesis inhibitors, such as cytosine arabinoside, hydrox- yurea, and aphidicolin (11,63). The increase in mRNA levels is due both to an increase in the mRNA stability 30 and to an increase in the rate of transcription (4,71,135). The H5 histone gene appears to be constitutively expressed. How independent H5 is from the cell cycle is still being debated. Two groups used aphidicolin synchronized cells to arrive at different conclusions. One group (37) used both Northern blotting and in_yi§;g pulse labeling. Erythroid cells were blocked in the cell cycle by use of aphidicolin. Once the block was released, the cells entered the cell cycle synchronously. At various time points both nuclei and RNA were isolated. By Northern blot analysis of the RNA, the steady state levels of the H5 mRNA appeared to remain constant while the other histone messages went up 15 fold at S phase and faded out at the G2 phase. By in_yit;g pulse labeling of the isolated nuclei, it also appeared that the H5 message was being transcribed at all points in the cell cycle. Their conclusion was that H5 is completely cell cycle independent. Another group (1) also used aphidicolin blocked eryth- roid cells. They report that while H5 mRNA is transcribed at all points of the cell cycle, the levels fluctuate. In their studies, the rate of transcription during the G1 phase is only 65% of the levels found in the S or 62 phases. The differences between these two studies cannot be easily resolved but may be related to the respective sensitivity levels of the measurements. 31 The level of transcription of the H5 gene is low in ery- throid precursor cells. Before the cells lose the ability to replicate (the erythroblast stage), the rate of trans- cription increases 6-fold compared to that of the other histone genes. As the cells enter the reticulocyte stage, the H5 gene is the only histone gene being transcribed. The level of H5 mRNA starts to decline as the cells near the end of their differentiation process even though the trans- cription rate is still high at that time. Transcription of the H5 gene probably ceases entirely as the cells enter the inactive erythrocyte stage (1). 32 W As the first step towards isolating the H5 histone gene, cDNA clones were constructed and identified. Two slightly different strategies were employed to isolate H5 cDNA clones. One group (85) utilized the known chicken H5 pro- tein sequence (21) to construct a synthetic DNA primer com- plimentary to a portion of the H5 mRNA. The synthetic DNA primer was hybridized to chicken reticulocyte poly A+ mRNA which was enriched for H5 mRNA. A cDNA library was con- structed from this hybridization and screened with the syn- thetic DNA primer. Ten clones containing H5 histone se- quences were isolated from the library. From this experi- ment, it was estimated that H5 mRNA represents only 0.2% of the chicken reticulocyte poly A+ population. One of the cDNA clones was found to encode the arginine H5 protein variant, while two other cDNA clones encoded the glutamine variant. As predicted, the only DNA sequence difference be- tween the two variants is the single base pair change re- quired to change the arginine codon CGG to the glutamine codon CAG. The second group (122) constructed a cDNA library from reticulocyte poly AI mRNA that was enriched for H5 his- tone sequences. The cDNA library was screened with an anti- 5 H5 antibody:12 I-protein A complex, which would detect 33 any bacterial colonies that were producing the H5 protein. Three clones, including p541 which will be discussed later, were isolated. From this data, it was calculated that the H5 mRNA represented 0.1-0.2% of the total reticulocyte mRNA population, a figure in agreement with the figure mentioned above. Using the cDNA clones as probes, the H5 gene has been isolated from a chicken phage library by several groups in- cluding our own (86,123,Chapter 3). It has been determined that the H5 gene is present only once per chicken haploid genome, compared to six known, unique H1 genes. Unlike most other known histone genes, the H5 gene is not closely link- ed to any other known histone gene. The DNA sequence of the H5 gene (Fig. 4) reveals many differences from the H1 genes. The H5 gene was found to be missing the H1 gene- specific 5’ element 5’-AAACACA-3’ which is present in all H1 genes sequenced to date (123). Instead, the H5 gene has replaced this A-rich sequence with the C-rich element 5’-CCGCCC-3' (34). The H5 gene does not have a "CCAAT” box in its promoter region nor does it contain a consensus sequence "TATA" box. Instead of the canonical sequence "TATATAT", it has the related sequence "TTAAAT", which does not contain the "ATA" motif. This sequence ("TTAAAT") does occur in the region expected of a "TATA" box (see Figure 4 34 DNA sequence of the chicken histone H5 gene. Sequence is from the group of Renaud (123). The gene shown here encodes for the arginine var- iant of the H5 protein. Putative control elements are underlined or enclosed by boxes. The two open arrows indicate the two transcription start sites, while the solid arrow at +726 indicates the site of polyadenylation. Direct and inverted repeat elements are underlined by horizontal arrows. Differences between this sequence and the H5 DNA sequence reported by Krieg et al (86) are shown above the sequence. Characters on top of the se- quence and between nucleotides indicate base dif- replacements and insertions in the gene sequenced by Krieg et al while hyphens represent deletions. The large open triangle indicates the position where the H5 cDNA clone p541 has a 9 base pair insertion relative to the H5 genomic clone shown here. 35 -850 WWWCWATG. g CTDCG 1' 6T LCCAAAGCC'I'C‘I'LCT W666 ALMJZCGGGACTW. nanmumcmwmmcmnamrficccnmccm &. ---.. -- -- .. GCMCACCACCCCNCCATAURACCACCUGLTAAUEGGAGCCACCACAMCCCCAEGGTGACACCA I'AGC'I’GCTCACGGCA ' 650 — 6 0 0 GACAGMGAGTGAGACCCCCGGTGTCTGTCCMTGTCCCCCCCTCAUMTCCCTACATTC666me ATCCAGTGTTCCCTAACACCCI‘TCTGACTCTCCTT GCCACCTC'ITZCCT 6CTCCCI'CCC‘1‘6TCCCCAGCAC6A6CTCCTCCTTGCC mnmmcmttcrmmcccmCCAGCACCTATCCfiGTCCTACCCAc‘rcTCAGTCCMAcTAcmrmc— -uoo WWW—LCCUWLCWCCSCUWURWWCWCW Accmarccccacccnmccicrcurccmumdficmagggqmcggawré cameo Accrccrtccmcrcccrcccccacccchmmccuucccomuccucrccrccccccccrc -100 cémomcac__c<fi@bccrccrccrcccgtccccficcwmcccwc'cccccccccmucmcuccccmrcccm TTGCIGGCGGCTCCTTFHTAAGCTCCCTAACCCCAGTGCCCTGCCGTGGGGTGAAGCGGCGGCC ATG AC6 GAG AGC CTG Thr Glu Sor Lou to 20 GTCCTATCCCCAGCCCCAGCCAAGCCCAAGCGGG‘I'GAAGGCATCGCGGCGC‘I‘CGGCATCGCAC Vol Lou Sor Pro Alo Pro Mo Lyo Pro Lyo Arc Vol Lyo Alo Sor Ar. Ar. Sor Alo Sor His 30 no CCCACCTACWGGAGATGATCGCGGCOGCCATCCGTGCGGAAAAGAQ:CGCGGCGGC‘I'CCTCG Pro Thr Tyr Sor Olu Hot Ilo Me Me Me Ilo Arg Ale Glu Lyo Sor Arc Gly Giy Sor Sor 50 ' 60 CGGCAGTCCATCCAGAAGTACATCAAGAGCCACTACAAGGTCGGCCACAACGCCGATCTGCAG Ar. em Sor Ito Gin Lyo Tyr Ilo Lyo Sor Hio Tyr Lyo Vol Giy His Aon Ato Aoo Lou Bin 70 80 ATCAAGCTCTCCATCCGACGI'C‘I‘CCTGGCI‘GCCGGCGTCC‘I'CAAGCAGACCMAGGGGTCGGG Ilo Lyo Lou Sor Iio Ar. Arc Lou Lou Aio Alo Oly Vol Lou Lyo Bin Thr Lyo Oly Vol Qty 90 100 GCCTCCGGCTCCTTCCGCTTGGCCAAGAGCGACAAGGCCAACAGGTCCCCCCGGAAGAAGAAC Ale Sor Ole Sor Pho Are Lou Alo Lyo Soc A» Lyo Alo Lyo Arc Sor Pro Gly Lye Lyo Lye 110 120 130 AAGCCCGTCAGGAGGTCCACGTCTCCCAAGAAGGCAGCGAGGCCCAGGAAGGCCAGGTCACCG Lyo Aio Vol Arg Arc Sor Thr Sor Pro Lyo Lyo Alo Me Are Pro Arg Lye Me Are Sor Pro 1'00 150 GCCAAGAAGCCCAMGCCACCGCCAGGAAGGCCAGGAAGAAGTCGCGGGCAAGCCCCAAGAAG Alo Lyo Lye Pro Ln Mo Thr Me Are Lyo Alo Arg Lyo Lyo Sor Arg Alo- Sor Pro Ln Ln 60 170 GCCAAGAAGCCAAACACI'GITAAG QZMG'I'CGCGGAAGCCCTCCAAGGCCAAGAAGGTGAAG Alo Lyo Lyo Pro Lyo Thr Vol Lyo Alo Lyo Sor Ar. Lyo Alo Sor Lyo Alo Lyo Lyo Vol Lyo 180 109 . CGGTCGAAACCCAGAGCCAAGK‘I’GGCGCCCGGAAATCGCCCAAGAAGAAGTGAW Ar. Sor Lyo Pro0 Ar. Alo Lyo Sor Oly Alo Arc Lyo Sor Pro Lyo Lyo Ln 650 WUWAMTAWAWACWAWMATMW AMAWMLWWGAAWMAWTAW ccnmucncmmmnmmcmmmmw WAWG Figuro 4 36 Promoter elements section), in this case 20 base pairs 5' of the transcriptional start site. There are no other AT- rich areas near the transcription start site that could act as a ”TATA” box except for this sequence. It is interest- ing to note that the duck H5 gene also has the "TTAAAT" se- quence at the same location as that of the chicken H5 "TTAAAT" sequence and that perfect homology exists between the two genes in the region surrounding this element, indi— cating selective pressure to keep this sequence (46). Other genes besides these two H5 genes have been found to contain noncanonical "TATA" boxes, including the "TTAAAA" sequence of the chicken lysozyme gene (64), the "ATTTAAA" element for the human intestinal alkaline phosphatase gene (104), and the "TTTAAAA" sequence of the Qigtygstgligm actin gene (98). The body of the H5 histone gene is similar to that of most other histone genes in that it does not contain in- trons. Despite being a member of the H1 family of histones, the coding sequence for H5 has less homology with the cod- ing sequence of H1 then might be expected (123). Computer analysis of the two coding sequences indicates that there is only 40% homology between an H1 histone gene and the H5 histone gene (unpublished data). As mentioned before, most other histones utilize a stem 37 loop structure to produce the 3' mRNA terminal end. The H5 gene does not contain this dyad repeat but may retain a remnant of it (34,123). Since the H5 mRNA is polyadenyl- ated, the expected signal sequence, "AATAAA", was searched for but not found. A similar sequence, "TATAAA", was de- tected in the sequence of the cDNA clones (85), but it is only four base pairs away from the polyadenylation site. This appears to be too close; usually the distance between the signal sequence and the site of polyadenylation is 11- 30 base pairs. The sequence, "TATAAA", may be present only by chance since this area is very AT-rich (84). It is known that the duck H5 gene does not have either the "AATAAA" or the "TATAAA" sequence present in the equivalent position. The 3’ untranslated region and the 3' flanking region for both the duck and the chicken H5 histone genes contain two inverted repeats that may be involved in the processing of the mRNA 3' terminal end. One possible stem loop structure is present near the 3’ end of the coding region while the second possible stem loop is near the polyadenylation site (46). Whether these stem loop elements are actually in- volved in the mRNA 3' terminal processing is uncertain, but as far as is known, the H5 histone gene does not use the standard polyadenylation signal sequences. The sequence analysis then, confirms that the H5 and H1 38 histones are related but only distantly. The H5 histone gene lacks many of the consensus 5' and 3’ elements of the H1 histone genes and contains other regulatory elements in their place. One of these elements is an erythroid-speci- fic enhancer (146). This H5 gene enhancer has structural and positional similarities to the chicken fi-globin enhanc- er (31,78). Like the fi-globin enhancer, the H5 gene enhan- cer is located in the 3' flanking region of the gene. The enhancer position was identified by use of an ”enhancer trap" plasmid (150) which allows the production of the SV40 large T antigen only when an enhancer sequence is inserted into the plasmid. Enhancer activity was detected in a restriction enzyme fragment isolated from the gene’s 3’ flanking region. This activity, however, was detected only when the plasmid was introduced into erythroid-lineage chicken HD3 cells and not when introduced into non-eryth- roid chicken embryo fibroblasts. In addition, the level of the large T antigen production increased when the HD3 cells (which are arrested at the CPU-E stage by a temperature sensitive avian erythroblastosis virus) were induced to re- enter the erythroid differentiation pathway. These results indicate that the H5 enhancer is erythroid-specific, and that the enhancer may mediate the temporal induction of H5 gene transcription during erythropoiesis. 39 Smaller restriction enzyme fragments were tested in the ”enhancer trap" plasmid to locate the position of the enhan- cer. An an I- Xma III fragment near the 3’ terminal end of the H5 gene still retained enhancer activity. Within this fragment, a sequence was identified that had a large degree of homology with the p-globin enhancer sequence. When the two sequences were aligned, it was determined that there were 25 out of 34 correct matches. Such a high de- gree of similarity between the two sequences is unlikely to have arisen by chance. This may represent a ‘core’ enhan- cer sequence, but this is uncertain. H5 enhancer 5'-GGAGGAGAGGGGACTCCTTCTTGTCCATAGGAGT-3’ *** ******** * * * **** **** *** fi-globin 5'-GGAAGAGAGGGGGTTAATCC-TGTCAATAGTAGT-3' enhancer A region that exhibits H5 gene-specific trans-activa- tion has been detected on the 5' side of the H5 gene (154). Xenopus oocytes were either injected with chicken histone genes or coinjected with chicken histone genes and chroma- tin salt wash fractions (CSWFs) isolated from chicken ery- throid cells. Oocytes coinjected with H1 and H2B histone genes plus the erythroid CSWFs showed the same level of transcription from the histone genes as oocytes injected with the histone genes alone. Coinjection of the H5 gene 40 and erythroid CSWFs however, showed a 10-fold increase in the level of transcription from the H5 gene relative to that observed in oocytes injected only with the H5 gene. If the H2B gene was physically linked to the H5 gene, the H2B gene would also show an increase in its levels of tran- scription when coinjected with the erythroid CSWFs. There- fore, there appears to be a sequence (or sequences) within the H5 DNA that recognizes a factor (or factors) within the erythroid chromatin extracts. The binding of this factor(s) increases the level of H5 transcription in this system and may activate genes that are linked near it. The H5 histone gene is not associated with the nuclear matrix in the same fashion as other histone genes. While the other histone genes appear to be associated with the nuclear matrix regardless of their transcriptional activi- ty, the H5 gene appears to be associated with the nuclear matrix only in erythroid cells where it is being transcrib- ed. It was not associated with the matrix in non-erythroid‘ T-cell line. The region of association to the nuclear mat- rix falls within a 780 base pair area, spanning part of the coding region and all of the 5’ untranslated region. This apparently erythroid-specific association with the nuclear matrix suggests that the H5 gene is activated much more se- lectively than are the H1 histone genes (38). 41 EIBLIQQBARHI 1) Affloter, M., Cote, J., Renaud, J., Ruiz-Carrillo, A., MOI. Cell. Biol. 1 pp. 3663-3672 (1987) 2) Allfrey, V., Faulkner, R., Mirsky, A. PNAS 51 pp. 786-794 (1964) 3) Allan, X., Hartman, P., Crane-Robinson, C., Aviles, F. Nature (London) 288 pp. 675-679 (1980) 4) Alterman, R., Ganguly, S., Schulze, D., Marzluff, W., Schildkraut, L., Skoultchi, A. Mol. Cell. Biol. 4 pp. 123-132 (1984) 5) Appels, R., Wells, J.R.E. J. Mol. Biol. 19 pp. 425-434 (1972) 6) Appels, R., Wells, J.R.E., Williams, A. J. Cell. Sci. 19 pp. 47-59 (1972) 7) Aviles, F., Chapman, G., Kneale, G., Crane-Robinson, C., Bradbury, C. Eur. J. Bioch. 88 pp. 363-371 (1971) 8) Banerji, J., Rusconi, S., Schaffner, W. Cell 21 pp. 299-308, (1981) 9) Banerji, J., Olson, L., Schaffner, W. Cell 3; pp. 729-740 (1983) 10) Bates, D., Thomas, J. NAR 2 pp. 5883-5894 (1981) 11) Baumbach, L., Harashi, F., Plumb, M., Stein, G., Stein, J. Biochemistry 23 pp. 1618-1625 (1984) 12) 13) 14) 15) 16) 17) 18) 19) 20) 21) 22) 23) 42 Benoist, C., O'Hare, K., Breathnach, R., Chambon, P. NAR 8 pp. 127-142 (1980) Benoist, C., Chambon, P. Nature 290 pp. 304-310 (1981) Beug, H., Palmieri, S., Zentgraf, C., Graf, T. Cell 23 pp. 907-919 (1982) Bienz, M. Cell 46 pp. 1037-1042 (1986) Birchmeier, C., Folk, W., Birnstiel, M. Cell 35 pp. 433-440 (1983) Birchmeier, C., Schfimperli, D., Sconzo, G., Birnstiel, M. PNAS 31 pp. 1057-1061 (1984) Boffa, L., Vidali, G., Mann, P., Allfrey, V. J. Bio. Chem. 25; pp. 3364-3366 (1978) Brandhorst, B. Dev. Bio. 12 pp. 139-148 (1980) Breathnach, R., Chambon, P. Ann. Rev. Bioch. 5g pp. 349-383 (1981) Briand, G., Kmiecek, D., Sautiere, P., Wouters, D., Borie-Loy, O., Biserte, G., Mazen, A., Champagne, M. FEBS 11; pp. 147-151 (1980) Briggs, M., Kadonga, J., Bell, 8., Tjian, R. Science 234 pp. 47-52 (1986) Brush, D., Dodgson, J., Choi, 0., Stevens, P., Engel, J. Mol. Cell. Bio. 5 pp. 1307-1317 (1985) 24) 25) 26) 27) 28) 29) 30) 31) 32) 33) 34) 35) 43 Candido, E., Dixon, G. J. Biol. Chem. 24_ pp. 3182-3188 (1971) Candido, E., Dixon, G. J. Biol. Chem. 241 pp. 3868-3873 (1972) Candido, E., Dixon, G. PNAS 62 pp. 2015-2019 (1972) Capecci, M. Cell 22 pp. 479-488 (1980) Carozzi, N., Marashi, F., Plumb, M., Zimmerman, S., Zimmerman, A., Coles, L., Wells, J.R.E., Stein, G., Stein, J. Science 224 pp. 1115-1117 (1984) Charnay, P., Mellon, P., Maniatis, T. MOI. Cell. Biol. 5 pp. 3261-3269 (1985) Childs, G., Maxson, R., Cohen, R., Kedes, L. Cell 23 pp. 651-663 (1981) Choi, O., Engel, J. Nature 3 3 pp. 731-734 (1986) Christensen, M., Rattner, J., Dixon, G. NAR 12 pp. 4575-4592 (1984) Cohen, 1., Lowry, J., Kedes, L. Science 129 pp. 994-997 (1975) Coles, L., Wells, J.R.E. NAR ;; pp. 585-594 (1985) Comb, M., Birnberg, N., Seasholtz, A., Herbert, E., Goodman, H., Nature 323 pp. 353-356 (1986) 36) 37) 38) _39) 40) 41) 42) 43) 44) 45) 46) 47) 44 Corden, J., Wasylyk, B., Buchwalder, A., Chambon, P. Science 292 pp. 1406-1414 (1980) Dalton, 8., Coleman, J., Wells, J.R.E. Mol. Cell. BiOl. 6 pp. 601-606 (1986) Dalton, S., Youngblood, H., Wells, J.R.E. NAR 15 pp. 6507-6523 (1986) D'Andrea, R., Coles, L., Wells, J.R.E. Mol. Cell. Biol. 5 pp. 3108-3115 (1985) Darnell, J. Nature 221 pp. 365-371 (1982) Davison, B., Egly, J., Mulvihill, E., Chambon, P. Nature 301 pp. 680-686 (1983) DeLange, R., Fambrough, D., Smith, E., Bonner, J. J. Biol. Chem. 21$ pp. 5669-5679 (1969) deVilliers, J., Schaffner, W. Nature 112 pp. 242- 246 (1984) Dierks, P., VanOoyen, A., Cochran, M., Dobkin, C., Reiser, J., Weissman, C. Cell _1 pp. 695-706 (1983) Dodgson, J., Strommer, J., Engel, J. Cell 11 pp. 879-887 (1979) Doenecke, D., Ténjes, P. J. Mol. Biol. 178 pp. 121-135 (1984) Dynan, W., Tjian, R. Nature 3 6 pp. 774-778 (1985) 48) 50) 51) 52) 53) 54) 55) 56) 57) 58) 59) 60) 45 Edwards, L., Hnilca, L. Experimentia 22 pp. 228- 229 (1968) Engel, J., Dodgson, J.B. PNAS 11 PP. 2596- 2600 (1980) Engel, J., Dodgson, J.B. PNAS 12 pp. 2856- 2860 (1981) Engel, J., Dodgson, J.B. J. Biol. Chem. 222 pp. 8239-8246 (1981) Engel, J., Sugarman, B., Dodgson, J.B. Nature 221 pp. 434-436 (1982) Elgsalter, A., Stokke, B., Mikkelson, A., Branton, D. Science 225 pp. 1217-1223 (1986) Elgin, S., Weintraub, H. Ann. Rev. Bioch. 22 pp. 725-774 (1975) Everett, R. NAR 11 pp. 6647-6666 (1983) Ferrell, R., Stroup, S., Tanis, R., Tashian, R. J. Biol. Chem. 222 pp. 4459-4463 (1978) Finch, J., Brown, R., Rhodes, D., Richmond, T., Rushton, B., Lutter, L., Klug, A. J. Mol. Biol. .155 pp. 757-769 (1981) Gidoni, D., Dynan, W., Tjian, R. Nature .212 pp. 409-413 (1984) Gillies, S., Morrison, 8., Oi, V., Tonegawa, S. Cell 22 pp. 717-728 (1983) 61) 62) 63) 64) 65) 66) 67) 68) 69) 70) 71) 72) 46 Goldstein, G., Scheid, H., Hammerling, U., Boyse, E., Schlesinger, D., Nioll, H. PNAS 12 pp. 11-15 (1975) Granick, S., Beale, S. Advan. Embryol. 22 pp. 33-203 (1978) Graves, R., Marzluff, W. Mol. Cell. Biol. 5 pp. 351-355 (1984) Grez, M., Land, H., Giesecke, K., Schutz, G., Jung, A., Sippel, A. Cell 22 pp. 743-752 (1981) Grosschedl, R., Birnstiel, M. PNAS 11 pp. 1432-1436 (1980) Grosveld, G., deBoer,, E., Shewmaker, C., Flavell, R. Nature 222 pp. 120-126 (1982) Gruss, P., Dhar, R., Khoury, G. PNAS 12 pp. 943-947 (1981) Hanahan, D. Nature 212 pp. 115-122 (1985) Harvey, R., Krieg, P., Robins, A., Coles, L., Wells, J.R.E. Nature 221 pp. 49-51 (1981) Harvey, R., Whiting, J., Coles, L., Krieg, P., Wells, J.R.E. PNAS 29 pp. 2819-2823 (1983) Heintz, N., Sive, H., Roeder, R. Mol. Cell. Biol. 2 pp. 539-550 (1983) Hen, R., Borrelli, E., Chambon, P. NAR 11 pp. 8747-8760 (1983) 73) 74) 75) 76) 77) 78) 79) 80) 81) 82) 83) 84) 47 Isenberg, L. Ann. Rev. Bioch. 22 pp. 159-191 (1979) Jones, H., Yamamoto, K., Tjian, R. Cell 22 pp. 559-572 (1985) Jones, R., Kadonoga, J., Rosenfeld, P., Kelly, T., Tjian, R. Cell 22 pp. 79-89 (1987) Hentschel, C., Birnstiel, M. Cell 22 pp. 301-313 (1981) Hentschel, C., Irminger, J.-C., Bucher, P., Birnstiel, M. Nature 222 pp. 147-151 (1981) Hesse, J., Nikol, J., Lieber, M., Felsenfeld, G. PNAS 22 pp. 4312-4316 (1986) Kadonaga, J., Tjian, R. 'PNAS 22 pp. 5885- 5893 (1986) Karp, R., Hogness, D. Fed. Proc. 22 pp. 1623- 1635 (1976) Kedes, L. Ann. Rev. Bioch. 22 pp. 837-870 (1979) Kline, K., Allison, J., Sanders, B. Develop. Biol. 12 pp. 389-396 (1982) Kornberg, R. Ann. Rev. Bioch. 22 pp. 931-954 (1977) Krieg, P., Robins, A., Colman, A., Wells, J.R.E. NAR 12, pp. 6777-6785 (1982) 85) 86) 87) 88) 89) 90) 91) 92) 93) 94) 95) 96) 48 Krieg, P., Robins, A., Gait, M., Titmas, R., Wells, J.R.E. NAR 12 pp. 1495-1502 (1982) Krieg, P., Robins, A., D’Andrea, R., Wells, J.R.E. NAR 11 pp. 619-627 (1983) Krieg, P., Melton, D. Nature 222 pp. 203-206 (1984) Laimins, L., Gruss, P. PNAS 12 pp. 6453-6457 (1982) Laskey, R., Earnshaw, W. Nature 222 pp. 763- 767 (1980) Lazarido, E., Nelson, W. Cell 21 pp. 505-508 (1982) Lee, F., Mulligen, R., Berg, P., Ringold, G., Nature 222 pp. 228-232 (1981) Levinson, B., Khoury, G., Van deWoulde, G., Gruss, P. Nature 222 pp. 568-572 (1982) Lewin, B. Ge 5 'o (Wiley Interscience, N.Y.) pp. 299-321 (1980) Lifton, R., Goldberg, M., Karp, R., Hogness, D. CSH Symp. Quant. Biol. 22 pp. 1047-1051 (1977) Louie, A., Dixon, 6. PNAS 22 pp. 1975-1979 (1972) Maxson, R., Mohun, T., Gormerzano, G., Childs, G., Kedes, L. Nature 301 pp. 120-125 (1983) 97) 98) 99) 100) 101) 102) 103) 104) 105) 106) 107) 108) 49 MCGhee, J., Felsenfeld, G. Ann. Rev. Bioch. 22 pp. 1115-1153 (1980) McKeown, M., Firtel, R. Cell 22 pp. 797-807 (1981) McKnight, S., Kingsbury, R. Science 217 pp. 316-324 (1982) McKnight, 8. Cell 21 pp. 403-408 (1984) McKnight, S., Tjian, R. Cell _2 pp. 795-805 (1986) Mercola, M., Wang, X., Olsen,J., Calame,K. Science 221 pp. 663-665 (1983) Miki, B., Neelin, J. Can. J. Bioch. 22 pp. 1158-1169 (1975) Millan, J. NAR 12 pp. 1059-1065 (1987) Mirzabekov, A., Belyovsky, A., Bauykin, S., Shick, V., Biosystems 12 pp. 265-271 (1980) Miyamoto, N., Moncollin, V., Wintzerith, M., Hen, R., Egly, J., Chambon, P. NAR 12 pp. 8779-8799 (1984) Molgaard, H., Perucho, M., Ruiz-Carrillo, A. Nature 222 pp. 502-504 (1980) Moore, G., Procunier, J., Cross, D., Grigliatti, A. Nature pp. 312-314 (1979) 109) 110) 111) 112) 113) 114) 115) 116) 117) 118) 119) 120) 50 Moreau, P., Hen, R., Everett, R., Gaub, M., Chambon, P. NAR 2 6047-6068 (1981) PP- Mosthof, L., Pawlito, M., Gruss, P. Nature 597-600 (1985) PP- Neelin, J., Callahan, P., Lamb, D., Murray, K. Can. J. Bioch. 1743-1752 (1964) 52 PP- Nelson, C., Allison, J., Kline, K., Sanders, B. Can. Res. 22 pp. 4625-4630 (1982) Nelson, C., Crenshaw, E., Franco, R., Rosenfeld, M. Nature 222 pp. 557-562 (1986) Nelson, D., Ferris, P., Zhang, D., Fereny, C. NAR 12 pp. 1667-1682 (1986) Nevins, J. Ann. Rev. Bioch. 22 pp. 441-446 (1983) Newrock, R., Cohen, C., Hendricks, M., Donnelly, R., Weinburg, E. Cell 12 pp. 324-336 (1978) Palmiter, R., Brinster, R. Ann. Rev. Genet. 22 pp. 465-499 (1986) Patthy, L., Smith, E., Johnson, J. J. Biol. Chem. 222 pp. 6834-6840 (1973) Price, D., Parker, C. Cell 22 pp. 423-429 (1984) Rogers, B., Saunders, G. BioEssays 2 pp. 62-65 (1985) 121) 122) 123) 124) 125) 126) 127) 128) 129) 130) 131) 51 Ruiz-Carrillo, A., Wangh, L., Littou, V., Allfrey, V. J. Biol. Chem. 222 pp. 7358-7368 (1974) Ruiz-Vazquez, R., Ruiz-Carrillo, A. NAR 12 pp. 2093-2108 (1982) Ruiz-Carrillo, A., Affolter, M., Renaud, J. J. Mol. Biol. 112 pp. 843-859 (1983) Samarut, J., Jurdic, J., Nigon, V. J. Embryol. Exp. Morph. 22 pp. 1-20 (1979) Samarut, J., Blanchet, J., Nigon, V. Develop. Biol. 12 pp. 155-166 (1979) Samarut, J., Bouabdelli, M. J. Cell. Physio. 122 pp. 553-563 (1980) Sawadago, M., Roeder, R. Cell 22 pp. 165-175 (1985) Schaffer, W., Kunz, G., Daetwyler, H., Telford, J., Cell Smith, H., Birnstiel, L. 655-671 13 PPo (1978) Schaufele, F., Gilmartin, G., Bannworth, W., Birnstiel, M. Nature 222 pp. 777-781 (1986) Seligy, V., Roy, C., Dove, M., Yaguchi, M. Bioch. Bioph. Res. Comm. 11 pp. 196-202 (1976) Selleck, S., Majors, J. Nature 222, pp. 173-177 (1987) 132) 133) 134) 135) 136) 137) 138) 139) 140) 141) 142) 52 Shaul, Y., Rutter, W., Laub, O. EMBO 2 pp. 427- 430 (1985) Sierra, F., Lichtler, A., Marashi, F., Rickles,R., VanDyke, T., Clark, 8., Wells, J., Stein, G., Stein, J. PNAS 12 pp. 1795-1799 (1982) Sittman, D., Chiu, 1., Pan, C., Cohn, R., Marzluff, W. 4078-4082 (1981) PNAS 12 pp. Sittman, D., Graves, R., Marzluff, W. PNAS 22 pp. 1849-1853 (1983) Skoulttchi, A., Gross, P. PNAS 12 pp. 2840- 2844 (1973) Smith, B., Harris, M., Sigourney, L., Mayes, E., 122 pp. 309-317 Bustin, M. Eur. J. Bioch. (1984) Strickland, W., Strickland, M., Brandt, W., VonHolt, C., Lehmann, A., Whittman, B. Eur. J. Bioch. 122 pp. 567-578 (1980) Sugarman, B., Dodgson, J.B., Engel, J. J. Biol. Chem. 222 pp. 9005-9016 (1983) Sung, M. (Biochemistry 12 pp. 286-290 (1977) Sung, M., Freedlender, E. Biochemistry 11 pp. 1884-1890 (1978) Sures,I., Lowry, J., Kedes, L. Cell 12 pp. 1033- 1044 (1979) 143) 144) 145) 146) 147) 148) 149) 150) 151) 152) 153) 53 Thoma, F., Koller, T., Klug, A. J. Cell. Biol. 22 403-427 (1979) PP- Thomas, J., Rees, E. Eur. J. Bioch 1&1 Torres-Martinez, S., Ruiz-Carrillo, A. C., Pearson, 143-151 (1985) PP- NAR 2323-2335 (1982) PP- Trainor, C., Stamler, S., Engel, J. Nature 228 827-830 (1987). PP- Van Helden, P., Strickland, W., Strickland, M., Von Holt, C. Bioch. Bioph. Acta 12; pp. 17-20 (1982) Vidali, G., Boffa, L., Bradbury, E., Allfrey, v. 15 Wasylyk, B., Derbyshire, R., Guy, A., Molko, D., PNAS ~ 2239-2243 (1978) PP0 PNAS 7 Roget, A., Teule, R., Chambon, P. 7024-7028 (1980) PP- Weber, F., deVilliars, J., Schaffner, W. Cell 22 983-992 (1984) PP- Weintraub, H., Cambell, G., Holtzer, H. 59 PP- Weintraub, H., Groudine, M. J. Cell. Biol. 652-668 (1971) Science 848- 122 PP- 856 (1976) Weise, M., Chan, L. J. Biol. Chem. 251 1892-1897 (1978) PP- 154) 155) 156) 157) 158) 159) 160) 161) 162) 163) 164) 54 Wigley, P., Wells, J.R.E. Mol. Cell. Biol. 1 pp. 3853-3856 (1987) Wood, W., Felsenfeld, G., J. Biol. Chem. 227 pp. 7730-7736 (1982) Wright, T., Hodgetts, R., Sherald, A. Genetics 22 267-285 (1976) PP- Yaguchi, M., Roy, C., Dove, M., Seligy, V. Bioch. Bioph.Res. Comm. 12 pp. 100-106 (1977) Yaguchi, M., Roy, C., Seligy, V. Bioch. Bioph. Res. Comm. 22 1400-1406 (1979) PP- Yamamoto, M., Yew, N., Dodgson, J.B., Hayashi, N., Engel, J. PNAS 22 pp. 3702-3706 (1985) Yew, N., Choi, H., Gallarda, J., Engel, J. £1. PP- Yoshihara, C., Dodgson, J.B. PNAS 1035-1039 (1987) Annal§_cf_thc_nlxl .. ' o . e garbcaic_Aanx§re§e§ 112 Bioch. 332-334 (1983) pp- 15 PP6 Hiet2nc_Qene_Structnrcl_9rsenizatien and_Begulatign eds. Zweidler, A. Dev. 47-56 (1980) Zweidler, A. Stein, G., Stein, J., Marzluff, W., J. Wiley and Sons, New York. 339-375 PP- (1984) Gazzolo, L., Samurut, J., Boubdelli, M., Blanchet, J. 22 683-691 (1980) Cell pp. 55 165) Borun, T., Pearson, D., Paik, W. J. Biol. Chem. 227 pp. 4288-4298 (1973) sham: NW Material: Restriction enzymes, calf alkaline phosphatase, T4 DNA ligase, T4 polynucleotide kinase, and E. coli DNA polymer- ase I were obtained from the following sources: Bethesda Research Laboratories, IBI, Promega Biotec, New England Biolabs, or Boeringer Mannheim. The nuClease Bal 31 (mix- ed form) was obtained from Bethesda Research Laboratories. The plasmids pHS-BR 2.4, pHS-HR 1.5 and pHS-BR delta Cla I were constructed as described in the Results section. The H5 cDNA clone, p541 was obtained from Dr. Ruiz-Carrillo (University of Toronto) QT6 and HD3 cells were both obtained from Dr. Hsing-Jien Kung (Case Western Reserve University). m Most of the cloning procedures listed below generally follow the protocols outlined by the Cloning Manual of Maniatis, Fritsch, and Sambrook (1). E . E !l :1. 1 El 1.! The library of Dodgson, Strommer and Engel (2) was screened for the presence of phage containing the H5 his- tone gene. The phage library was plated at a density of 56 57 50,000-100,000 plaque forming units (pfu) per 150 mm petri dish, using the E. coli strain K803 SupF as the host. The phage plaques were transferred to nitrocellulose fil- ters and were treated as described by Maniatis et al. (1). The H5 cDNA clone, p541, was used to probe the resulting filters. To decrease the possibility of background prob- lems, the p541 plasmid was digested with Pst I and the cDNA insert was isolated from the pBR322 vector sequences. Any bacteriophage plaques from the initial screening which appeared to have hybridized to the p541 probe were rescreened at a lower density (100-200 pfu/lOOmm petri dish). Those bacteriophage plaques which tested positive in this second screening were isolated and a stock culture of the bacteriophage was obtained. Bacteriophage DNA was isolated from bacteriophage puri- fied on a CsCl step gradient (1). The DNA was then di- gested with various restriction enzymes and subjected to agarose gel electophoresis. The DNA was transferred to a nitrocellulose filter by Southern blotting. The filter was probed with the H5 cDNA clone, p541. 58 EJEJD. !. The Bal 31 exonuclease digestions were generally per- formed as described by Maniatis 22_21 (1). Five micro- grams of the target plasmid were linearized with either Sac I (for pH5-BR 2.4) or Cla I (for pHS-BR delta Cla I). The DNA was then digested with 1-5 units of Bal 31 at 30° for the lengths of time indicated (3 to 5 minutes). The digestion was halted by the addition of EGTA (which che- lates the necessary calcium ions) and cooling of the mix- ture to 4°. The samples were deproteinized by extraction with an equal volume of 1:1 phenol to chloroform. The DNA was ethanol precipitated, then made blunt-ended by treatment with DNA polymerase Klenow fragment. HindIII linkers were ligated to the DNA using T4 DNA ligase. The linkers were then digested with Hind III, and the excess Hind III link- ers removed by ethanol precipitation. The plasmids were afterwards recircularized by ligation under dilute condi- tions and transformed into the E. coli HB101. Individual colonies were picked and analyzed. DNA was isolated from the bacteria (3) and digested with either Pst I + Hind III (for the pH5-BR 2.4 derived clones, the "S" series) or Sac I + Hind III (for the pH5-BR delta Cla I derived clones, or the "C" series). The digestions were 59 subjected to gel electrophoresis on a 5% polyacrylamide gel, then stained with ethidium bromide. Those bacterial clones which were identified as having a HindIII linker within the target region were stored as a glycerol stock. W Clones were sequenced by the chemical degradation meth- od of Maxam and Gilbert (4) as modified by Smith and Calvo (5). Individual clones were digested with Hind III to linearize the plasmid at the linker site. The DNA was treated with the enzyme calf alkaline phosphatase to remove the end terminal phosphate groups. The DNA was then radio- tively labelled by incubating the DNA with T4 polynucleo- tide kinase and 1-32P ATP. The unneeded radioactively lab- elled end was removed by digesting the DNA with either BamHI or Sal I. Since the sequence of the H5 gene is known, only three of the five standard Maxam-Gilbert reactions were used. These three reactions (C+T, A+G, G) were sufficient to determine the portion of the 5’ flanking region being studied. Only the terminal 20-30 base pairs of the clone’s sequence needed to be determined. The chemical reactions 60 therefore, were allowed to proceed for twice the usual time. The reaction products were run on a 20% denaturing polyacrylamide gel and exposed to X-Ray film. Selim QT6 cells (6) are chemically transformed quail fibro- blasts. These cells are cultured in Dulbecco Modified Eagle's Medium (Grand Island Biological Co., Grand Island, NY) supplemented with 4% fetal calf serum, 1% chicken se- rum, and 1% v/v DMSO. HD3 cells (7,8) are chicken CFU-E erythroid cells differentially arrested by a temperature sensitive Avian Erythroblastosis Virus (ts-AEV). These cells are cultured in Dulbecco Medified Eagle’s Medium sup- plemented with 8% fetal calf serum, 2% chicken serum, and 10 mM HEPES pH 7.4. Both cell types are grown at 37° and 5% C02. W _The transformation protocol is that of Wigler et al. (9). Twenty four hours before transformation, QT6 cells 6 were plated to a density of 1-2 x 10 cells per 100mm tissue culture plate. Four to five hours prior to the 61 transformation, the old media was removed from the cells and replaced with 7 mls of fresh media. 25 to 50 micrograms of each individual clone’s DNA was ethanol precipitated. If more than one clone was to be transformed into the same plate of cells, the DNA for both clones were co-precipitated. The DNA was resuspended in 0.62 mls of ddH20 and adjusted to a final concentration of 250 mM CaCl2 by addition of 0.08 mls of 2.5 M CaClz. The DNA/CaCl2 mixture was rapidly added to an equal volume of 2x Hepes buffered saline (280 mM NaCl, 50 mM Hepes, 1.5 mM Na HPO4, pH 7.10 t 0.05). The DNA-calcium phosphate pre- 2 cipitate was allowed to form for 15-30 minutes at room tem- perature. The DNA-calcium phosphate mixture was then added to the QT6 cells. After 8-12 hours, this mixture was re- moved and replaced with fresh QT6 media. RNA was isolated from the cells 36-40 hours after the DNA-calcium phosphate mixture was removed. BNA_I§Ql§§iQn For total cellular RNA, two 100 mm plates of QT6 cells were lysed with 5 mls of guanidinium isothiocyanate buffer (4 M guanidinium isothiocyanate, 50 mM Tris pH 7.5, 0.14 p- mercaptoethanol, 2% Sarcosyl, and 10mM EDTA). The cellular 62 homogenate was then passed five times through a 22 gauge needle to shear the high molecular weight DNA. Two grams of RNase free CsCl was added to every 5 mls of homogenate. This mixture was then layered over 4 mls of RNase free 5.7 M CsCl (in 0.03 M NaOAc, pH 5.2) in a polyallomer SW 41 tube. The samples were spun at 28,000 rpm for 16 hours in a SW 41 rotor. After the run was complete, the supernatant was carefully removed with an RNase free pipet. The RNA pellet was resuspended in 300 microliters of RNase free ddHZO and extracted with an equal volume of 4:1 Chloro- form to n-Butanol solution. NaOAc was added to a final concentration of 300 mM and the RNA precipitated with 2.5 volumes of ethanol. For Poly A+ RNA, two 100 mm plates of QT6 cells were lysed with 3 mls of Proteinase K buffer (0.5 M NaCl, 10 mM Tris pH 7.5, lmM EDTA, 1.0% SDS, 200 micrograms/ml protein- ase K) and the high molecular weight DNA released by the lysis sheared by passage through a 22 gauge needle. Fresh proteinase K (100 micrograms/ml) was added to the homogen- ate which was then incubated at 37a for one hour. Oligo dT cellulose was swelled in RNase free ddH 0, then 2 transferred to proteinase K buffer and proteinase K treated to remove any residual RNases. The oligo dT cellulose was added to the cell homogenates and the homogenates were 63 rocked at room temperature for one hour. The oligo dT cellulose was spun out of the solution and was washed twice with a high salt buffer (0.5 M NaCl, 10 mM Tris, pH 7.5, 1mM EDTA), then twice with a low salt buffer (0.1 M NaCl, 10 mM Tris pH 7.5, 1 mM EDTA) . The poly A+ RNA was removed from the cellulose by addition of two mls of RNase free ddHZO. 50 micrograms of RNase free tRNA was added to the supernatant and the RNA precipitated with 2.5 volumes of ethanol. W Two DNA probes were used to analyze the RNA obtained from the QT6 transformations. Early studies utilized a BstEII restriction enzyme site present at +500 while later studies utilized a EcoOlO9 site present at +380. The pro- cedure used to label this DNA is similar to that of Mania- tis 22_21. Briefly, the plasmid pHS-BR 2.4 was digested with either BstEII or Eco0109, then the terminal phos- phates were removed by the enzyme calf alkaline phospha- tase. The DNA was separated on a 6% agarose gel, and the desired fragment isolated. This fragment was radioactively 2 labelled by treatment with 1-3 P ATP and T4 polynucleotide kinase. The DNA was then digested with the restriction en- 64 zyme Hian to remove the unneeded labelled end. The DNA was separated on a 6% agarose gel and the desired fragment was isolated. The labelled DNA fragment was mixed with the RNA being studied and both were ethanol precipitated. The pellet was resuspended in 10 microliters of hybridization buffer (80% formamide, 0.4 M NaCl, 0.04 M Pipes, pH 7.25). The sample was heated at 90° for 4 minutes to denature both the RNA and the DNA. The sample was allowed to hybridize at 550 for 12 hours. After the hybridization was completed, 200 microliters of 1X 81 buffer (0.03 M NaOAc, pH 4.5, 0.25 M NaCl, 4mM ZnOAc, 100 micrograms/ml denatured, sheared sal- mon sperm DNA) and 200 units of $1 nuclease was added to the sample. The reaction was allowed to proceed for 15 minutes at room temperature, then was stopped by deprotein- ation by extraction with an equal volume of a 1:1 phenol: chloroform mixture. The supernatant was ethanol precipi- tated. The reaction products were analyzed on a 6% dena- turing polyacrylamide gel. 65 References 1) 2) 3) 4) 5) 6) 7) 3) 9) 'Maniatis, T., Fritsch, E., Sambrook, J. NQIEQHIAI Qlcningl_A_Laberatcr2_nanual (Cold Spring Harbor Laboratory, NY, 1982) Dodgson, J., Strommer, J., Engel, J. Cell 11 pp. 879-887 (1979) Ish-Horowicz, D., Burke, C. NAR 2 pp. 2989- 2998 (1981) Maxam, A., Gilbert, W. Methods Enzymol. 22 pp. 499-560 (1980) Smith, D., Calvo, J. NAR 3 PP. 2255-2274 (1980) Moscovici, C., Moscovici, G., Jimenez, H., Lai, M., Hayman, M., Vogt, P. Cell 11 pp. 95-103 (1977) Beug, H., Palmieri, 8., Freudenstein, C., Zentgraf, H., Graf, T. Cell 22 pp. 907-919 (1982) Weintraub, H., Beug, H., Groudine, M., Graf, T. Cell 22 pp. 931-940 (1982) Wigler, M., Pellicer, A., Silverstein, 8., Axel, R. Cell 12 pp. 725-731 (1978) Results W The first attempt at isolating lambda phage clones containing the chicken H5 histone gene made use of the fact that both the chicken replication-dependent and partially replication-dependent histone genes are loosely clustered together within the chicken genome. It was hoped that the H5 gene would be present within one of these histone gene clusters, even though it is a replication-independent his- tone gene. Fifty phage clones which were known to contain various histone genes were screened for the presence of an H5 gene with an H5 cDNA clone, p541 (1) which was obtained from Dr. Ruiz-Carrillo. The cDNA insert of p541 is 250 base pairs in length, and is cloned into the Pst I site of pBR322. None of the phage clones hybridized to the p541 probe, indicating that the H5 gene is not closely linked to any other known histone gene. Two other groups (4,5) have also determined that the H5 gene is a solitary histone gene. This is possibly due to the fact that the H5 gene is a replication-independent histone as well as a tissue-spec- ific histone. Interestingly, the H5 probe did not cross- hybidize to any histone H1 genes, even under low stringency conditions. This is further evidence that the H5 gene has 66 67 evolved separately from the H1 genes for a lengthy period of time. The chicken phage library of Dodgson, Strommer, and En- gel (2) was then screened for lambda phage clones contain- ing all or part of the H5 histone gene. Screening the phage library with the p541 probe identified two different phage clones, designated cH5-1 and cH5-2, which hybridized strongly to the probe. The restriction enzyme maps of the two phage clones and the p541 hybridizing region are shown in Fig. 5. It is not apparent from the maps whether these two clones are overlapping each other, and therefore con- tain the same H5 gene. While there is a 350 base pair SacI fragment in both clones within the region of H5 hybridiz- ation, this alone was not enough to prove that both phage clones contained the same H5 gene. At the time this work was being done, it had not been established that only one H5 gene/haploid genome was present in the chicken. 0n the contrary, an early study involving hybridization kinetics (3) indicated that the H5 gene might be present at up to 10 copies/haploid genome. While this has since been shown to be incorrect, it was not possible at the time these clones were being studied to state with certainty that both clones contained the same H5 gene. If several H5 genes did exist, more phage clones containing an H5 gene should have 68 Figure 5 Restriction enzyme maps of the four phage clones containing the H5 gene. The region which hybrid- ized to the p541 H5 cDNA clone is shown as a thick line. The transcriptional direction of the H5 gene is shown above this region. oousosoon 2.1:... D ocotj Locoo> (v soLosO 2: £2.23 Elle: .Iouu \—1 o...- o: 2: . llv 2:233»: ~me I seen . .0 ~.nAll.m~ do /—\ so_uou=o...O .5... q Loo-c... .33.” 3 L 6:22...» T234282. deer-Hm I.» HE .3 41 4 _ m M. nuns» Iii. ._, NE, DD D EL: 8...... infill??? q a e.» be m 4 w :1" «luau Jar D. AV DAN DAM Av.My % . fl Figuro 5 70 been present in the phage library. The 2.4 kilobase EcoRI- BamHI fragment from cH5-2 which hybridized to the p541 probe was selected as a probe for rescreening the phage library. Two additional phage clones, designated cH5-3 and cH5-4 (Fig. 5), were isolated and analyzed. Comparisons between the four restriction enzyme maps indicate that not only are cH5-3 and cH5-4 similar to one another, but that they overlap both cHS-l and cH5-2. All four phage clones appear to span the same region of the genome and contain the same H5 gene. This would indicate that there is only a single H5 gene/haploid genome rather than multiple copies. To test this further, DNA was isolated from chicken reticu- locytes and digested with the restriction enzyme HindIII. After Southern blotting, the blot was probed with the 2.4 kb EcoRI-BamHI fragment from cH5-2. The H5 gene analyzed in our lab lies within a 7 kb HindIII fragment (cH5-4, Fig. 5). If there were other H5 histone genes, it was likely that they would be surrounded by different flanking regions and would appear as different sized bands on the Southern blot. As shown in Fig. 6, only a single hybridizing band, roughly 7 kb in size as predicted, was detected. This blot however, would not rule out a reiterated H5 gene. Two other laboratories (4,5) have also isolated H5 his- tone phage clones. Several of these clones are similar to Figure 6 Chicken reticulocyte DNA was digested with HindIII and transferred to nitrocellulose paper by Southern blotting. 20 and 50 micrograms of DNA was tested in this gel. The blot was probed with the pHS-BR 2.4 subclone described in Figure 6. Predicted size of the hybridizing band is 7 kb. 72 clones isolated in our laboratory, and the restriction en- zyme maps of the region surrounding their H5 genes are very similar or identical to the cloned region described above. That all three groups have isolated identical H5 genes is further evidence that only one H5 gene is present in the chicken genome. W A number of restriction enzyme fragments have been iso- lated from the phage clones and subcloned into the plasmid pBR322. The two subclones that have been analyzed in de- tail are pHS-HR 1.5 and pH5-BR 2.4 (Fig. 7). pHS-HR 1.5 contains the 1.5 kb HindIII-EcoRI fragment from cH5-1 which hybridizes to the p541 probe, while pH5-BR 2.4 contains the 2.4 kb BamHI-EcoRI fragment from cH5-2 which likewise hy- bridizes to the p541 probe. It should be noted that the EcoRI sites in both cases are not present in the normal H5 genomic sequence (compare the maps of cH5-3, and cH5-4 ver- sus the maps of cH5-1 and cH5-4), but were introduced by use of synthetic EcoRI linkers during the construction of the chicken phage library (2). Both subclones were mapped by restriction enzymes (Fig. 7) and compared to the known restriction enzyme map of the H5 gene (4,5). This allowed both the position of the gene Figure 7 73 Restriction enzyme maps of the H5 gene sub- clones described in the text. The transcrip- tional direction of the H5 gene is in all cases left to right. The location of the transcrip- tion start site and the polyadenylation site are shown. 74 ezm «Loam _ .uel so:s_uuese..h e .emz 4.6 ”magma NH. "866...... H so no so Icuhouflm: l u ecu « 33. H . low :5: m ooze-ase—oohzm ~ _ eEmF =. 1:...- D : cum .1 -_u «on « uses—c... 3.25 .3 _uemw .ern e :0 3.6.73... _ A. 4 . A z :1... I .3 J l vx w ilk H *m #1” _ 2 szlmzennllluflfwa :3 ffi Figure 7 75 within the clone as well as the orientation of the gene to be determined. As shown in Fig. 7, pHS-BR 2.4 contains the entire H5 gene as well as 1.2 kb of the 5’ flanking region and 0.7 kb of the 3' flanking region. The pH5-HR 1.5 sub- clone contains a truncated H5 gene. Only about 100 bases of the H5 coding region is present within the insert, which also includes 1.4 kb of the 5' flanking region. WWW Linker scanner mutagenesis (6) uses synthetic oligonuc- leotide linkers to generate a cluster of point mutations within the region of DNA being studied. The region being studied for the H5 histone gene is the region extending from the CAP site to roughly -200, which appears to con- tain most of the putative promoter elements. Before the linker scanner mutants can be constructed, two sets of deletions must be made. One set deletes prog- ressively larger amounts of the region being analyzed in a 5'-3' manner, while the second set progressively deletes over the same region in a 3'-5' manner. The plasmid pH5- BR 2.4 was used to construct the first set of deletion mu- tants (Fig. 8). The plasmid was first digested with the restriction enzyme SacI. There are only two SacI sites, 76 5' of the promoter region (Fig. 7). After the plasmid was digested with SacI, the DNA was incubated with the double- stranded exonuclease, Bal-31. At various time points dur- ing the Bal-31 digestion, aliquots were taken and the Bal- 31 reaction stopped. This procedure allowed a series of progressively larger deletions to be obtained. Synthetic HindIII linkers were ligated to the Bal-31 treated pHS-BR 2.4. HindIII linkers were chosen because no HindIII sites are present within the pH5-BR 2.4 plasmid. The link- ers are ten nucleotides in length, with the sequence GCAAG CTTGC. After the ligation was completed, the linkers were digested with HindIII and the plasmid recircularized and ligated. The result of this series of reactions is a set of deletions which removes progressively larger amounts of the putative promoter region in a 5’ to 3' direction. This set of deletions has been designated the "8" series. A similar set of reactions was done on the 3' side of the putative promoter region (Fig. 9), using the plasmid pHS-BR delta ClaI. pHS-BR delta ClaI (Fig. 7) is a deri- vative of pHS-BR 2.4, and differs from the parent plasmid in two respects. The first difference is that the 350 base pair SacI fragment has been deleted and only one SacI site remains, instead of two. The second difference is that a synthetic ClaI linker was inserted into one of the 77 " Sac I (2 sites) ‘/ EcoRI \ Pst I BamHI 1. Digest plasmid with Sac I 2. Incubate for various lengths of time with Bal 31 3. End fill with Klenow polymerase and ligate to synthetic Hind III linkers 4. Digest linkers with Hind III and recircularize plasmids by ligation in a dilute solution 5. Transform E. coli and screen clones by Hind III + Pst I digestion Figure 8 Scheme for construction of the 5' to 3' set of deletions. The location and transcriptional direction of the H5 gene is shown by the arrow within the circle. The area deleted is shown by the arrows on the outside of the circle. The pBR322 vector sequences are shown as a thick line while the HS insert is shown as a thin line. BamHI pHS-delta Cla I 1. Digest plasmid with Cla I 2. Incubate with Bal 31 for various lengths of time 3. End fill with Klenow polymerase and ligate to synthetic Hind III linkers 4. Digest linkers with Hind III and recircularize plasmids by ligation in a dilute solution 5. Transform E. coli and screen clones by Hind III + Sac I digestion Figure 9 Scheme for construction of the 3' to 5' set of deletions. Position and transcriptional direction of the H5 gene is shown by the internal arrow while the area of deletion is shown by the external arrow. The thin line indicates the HS insert while the thick line represents the pBR322 vector sequences. 79 SacII recognition sites. This ClaI linker is within the coding region of the H5 gene and is 3’ of the putative pro- moter region. The pH5-BR delta ClaI plasmid was linear- ized by digestion with ClaI, and afterwards treated as described above and in Figure 9. The deletion mutants that resulted from this series of reactions progressively delete the putative promoter region in a 3’ to 5' manner. This set of deletions has been designated the "C" series. After the plasmids were recircularized, the DNA was transformed into the E. coli strain, HB101. Plasmid DNA was isolated from the resulting colonies, and analyzed by digestion with restriction enzymes. pH5-BR 2.4 deletion mutants ("8” series) were digested with both HindIII and PstI, while the pH5-BR delta ClaI derived mutants ("C" series) were digested with HindIII and SacI. The purpose of this digestion is twofold: first, it determines whether a HindIII linker is present within the mutant, and second, if a HindIII linker is present, its location within the promoter region may be established within a roughly 50 base pair area. Over 1,000 bacterial colonies have been tested in this manner. Not all of the mutants tested had a Hind III linker present within their DNA. Efficiencies varied from 80% Hind III+ to a low of only 40% Hind III+. The reasons behind these different efficiencies is unclear. 80 The majority of the clones which have the HindIII linker, have it within the target area, i.e. between -200 and the CAP site. Measuring the position of the linker by restriction en- zyme digestion gives only a rough estimate of the linker’s location. To make the linker scanner mutants, the exact location of the linker must be determined by DNA sequenc- ing. A modified Maxam and Gilbert sequencing protocol (7) was used to sequence selected mutants. Unfortunately, no usable restriction enzyme sites were present near the re- gion to be sequenced. Therefore, the mutants were digest- ed with HindIII and opened directly at the linker. The end 32P ATP as described in of the DNA was labeled with gamma the Materials and Methods section. Because the sequence of the H5 gene was known, only three out of the standard five reactions were done. These three reactions ("C+T", "A+G", and "6”), were sufficient to determine the location of the linker. Another modification to the standard protocol was in the length of time the reactions were allowed to pro- cede. Only the terminal 20-30 base pairs of the sequence needed to be determined. For this reason, the chemical reactions were allowed to proceed twice as long as normal- ly required. Due to the small size of the fragments, the reaction products were electrophoresed on a 20% denaturing 81 Figure 10 Maxam-Gilbert sequencing of the ”8" series clone 8'4-6. This particular clone was studied with five chemical reactions (C+T, C, A+G, G, and T>G). Most clones were se- quenced with only three of the reactions (C+T, A+G,and G). The bottom five nucleotides (below the arrow) are from the artificial Hind III linker. The first nucleotide above the arrow, represents the beginning of the chicken H5 histone DNA sequence. Roughly 25 to 30 base pairs of se- quence may be determined from the gel. Since the sequence of the H5 flanking region is known, the exact location of the artificial HindIII linker may then be determined. Figure 10 M 6+1 .5: l. as w r .. . £35113 3-2.4». $3.9 .. , s.AC.....CA....As sec 6 c c 83 polyacrylamide gel. As shown in Figs. 11 through 14, the exact location of the HindIII linker has been determined for 106 mutants: 53 from the "8" series and 53 from the "C" series. The construction of a linker scanner mutant requires an exact match between the location of the linker in the "8" series clone and in the "C" series clone. When these two clones are digested with HindIII and BamHI and are ligated together (Fig. 15), a mutant is created that maintains the original spacing of the H5 gene promoter, but has replaced 10 base pairs of the original promoter sequence with 10 base pairs of synthetic HindIII linker sequence. By using the same technique with "8" series and "C" series clones that are not an exact match, deletions or duplications of various sizes within the H5 histone promoter may be obtain- ed. The various mutants that were constructed for studying the H5 histone promoter will be discussed in more detail in the next chapter. Figure 11 84 Location of the Hind III linker for "8" series clones between -130 and -40. The sequence of the H5 flanking region is at the top. The thin line to the right of the Hind III linker se- quence indicates unchanged H5 sequences. The designation for each mutant is on the far right. 85 m UOHHUU< Location and transcriptional direction of the H5 gene Fig. 15 94 Referensea 1) Ruiz-Vazquez, R., Ruiz-Carrillo, A. NAR 12 pp. 2093-2108 (1982) 2) Dodgson, J., Strommer, J., Engel, J. Cell 11 pp. 879-887 (1979) 3) 4) 5) 6) 7) Scott, A., Wells, J.R.E. Nature 222 pp. 635- 638 (1976) Ruiz-Carrillo, A., Affloter, M., Renaud, J. J. Mol. Biol. 112 pp. 843-859 (1983) Krieg, P., Robins, A., D’Andrea, R., Wells, J.R.E. NAR 11 pp. 619-627 (1983) McKnight, 8., Kingsbury, R. Science 211 pp. 316-324 (1982) Maxam, A., Gilbert, W. Methods Enzymol. 22 pp. 499-560 (1980) 9.112% W As described in the literature review, the putative chicken H5 histone promoter region lacks many of the chick- en Hl histone promoter elements. The putative H5 promoter region lacks a ”CCAAT" sequence and unlike all H1 histone promoter regions sequenced to date, the H5 promoter region does not include an "H1 box" (5'-AAACACA-3’) (1). There are a number of sequences within the H5 promoter region, however, which may act as promoter elements for the H5 his- tone gene (Figure 16). The first of these sequences is an element eight nucleo- tides in length located near the transcriptional initiation site. This element is also present at a similar location in the duck H5 promoter region. While the purpose of this element is unclear, it may serve as a recognition sequence for determining the location of transcription initiation (6). The chicken H5 histone gene has two transcriptional start sites, which are three nucleotides apart and located immediately 5’ of the octanucleotide described above (7,8, Figure 17). The two transcriptional start sites appear to be used at an equal frequency. Another probable promoter element is the noncanonical "TATA" box (5'-TTAAAT-3’) located at -20. As shown in Fig- ure 16, the duck H5 histone gene has the identical sequence in this area of the promoter. It is interesting to note 95 Figure 16 96 Comparison of the chicken and duck H5 puta- tive promoter regions. The two sequences have been aligned at the "TTAAAT" element. Asterisks indicate differences between the two sequences (only shown for the region be- tween +10 and -60). The numbering system for the chicken H5 promoter starts from the 5’ transcriptional start site. The second start site is at +4. The duck H5 promoter is numbered from the single transcriptional start site. Putative promoter elements (described in the text) are underlined. 97 Chicken 'TO I '70 I -i° I +i Ti 71° GTGCQQQAQQQIIAAAIGCGTGCTGGTGGCGACGCGCGQQQQQAQACGCA * * * *** * * ******* GCACQAQAQQQIIAAAITCGGGGCAGCGCCGGGTGCQQQQQAQAGGCGGC l | I l | I I l | I -30 -20 -10 +1 +10 Duck Chicken -90 -80 -70 -60 -50 -40 | I | l I I TCCTQQQQQQQQQQQQAGAGGGGGGACAQQQQQACAGGCAGTCCTCCCCGCGGTCC * QQQQQGGGACGGGACGGGGGGGGGGACAQQQQQACAGGCAGTCCTCCCCGCCGTCC -90 -80 -70 -60 -50 -40 Duck Chicken -l40 -130 -120 -110 -100 l l | l | l CCCCATCACATCCCTTCTGGTCCCAACCTCQIQQQIQQQIQQQIQQQQQAQGCATG CTCCTGTCCCC?CGGT%I§Q§$Q§§I$IQT§fCCAGTGCAQ%§§A§TAT§Q%§§§Q -150 -140 -130 -120 -110 -100 Duck Figure 16 Figure 17 98 poly A+ RNA from Rat 3A fibroblasts (lane 1), QT6 fibroblasts (lane 2), and HD3 eryth- roid cells (lane 3) were hybridized to a 27 nucleotide H5 primer. The primer covers the first 9 codons of the H5 coding region. The primer was then extended using reverse tran- scriptase. The two bands in the HD3 erythroid RNA lane indicates that there are two tran- scriptional start sites, 3 base pairs apart. Predicted band sizes are 141 bp and 144 bp. 2 2 2 Fig. 17 100 that the duck and chicken H5 promoter regions do not share extensive areas of homology in the region between the "TTAAAT" box and the octanucleotide around the transcrip- tional initiation site. Starting at the "TTAAAT" sequence, however, a large stretch of sequence homology extending to -70 is found. Apparently, the noncanonical "TTAAAT" box is under selective pressure. Immediately upstream of the "TTAAAT" element is a hexa- nucleotide, which is designated the proximal H1 homology element (3). This element (5’-CGCACC-3') is located at -30 in the chicken H5 gene and a related sequence (5'-CAC ACC-3') is located at an identical site in the duck H5 pro- moter. The 5'-CGCACC-3' sequence is also present in two chicken H1 genes (2,9) but in the two chicken H1 genes, this element is located at roughly -60 and is separated from the "TATA" element by approximately 10 nucleotides instead of being adjacent as in the H5 genes. The signif- icance of this difference, if any, is unknown. The proximal H1 homology hexanucleotide sequence almost completely overlaps a second sequence (5’-CACCC-3'). The 5’-CACCC-3' pentanucleotide has been shown to be required for high levels of rabbit and chicken p-globin gene expres- sion after transfection of these genes into fibroblasts (4, 5). While the location of this sequence varies slightly species to species (4), the pentanucleotide is usually found in the -100 region. In the chicken adult p-globin 101 gene, this pentanucleotide is present at roughly position -130. In the chicken H5 histone promoter, however, this pentanucleotide is adjacent to the "TATA" element at posi- tion -30. Unfortunately, the mutagenesis assay being des- cribed here will only determine if this area of the promo- ter is important for the expression of the H5 histone gene. It will not determine which element (if either) is respon- sible for any changes detected when this area is mutated. Further 5’ of these elements is another H1 homology hexanucleotide (5’-CGGGGA-3’) which is designated the dis- tal H1 homology element. This element is located at -60 in both the chicken and duck H5 histone promoter. One chicken H1 histone gene (2) has this element at position -100 while another chicken H1 gene (9) has a related hexanucleotide (5'-CGGGGC-3’) at roughly position -130. The function of this element, if any, is unknown. After the distal H1 homology element, the spacing of putative promoter elements is different between the duck and the chicken H5 genes. An Spl protein binding sequence, 5’-GGGGCGGGG-3', is located at position -80 in the chicken H5 promoter region. The duck H5 promoter region, however, has this sequence displaced ten nucleotides upstream rela- tive to the chicken H5 gene. Upstream of the Spl binding element in the chicken H5 promoter, between -95 and -115, is a sequence designated 102 the duck H5 homology element. In the duck H5 promoter, this element consists of two isolated sequences, 14 and 6 nucleotides in length respectively. In the chicken H5 pro- moter, however, these two sequences have been joined into a 20 base pair element. The function of this element, if any, is unknown. In summary, six possible promoter elements have been identified to date. They include the octanucleotide near the H5 transcriptional start site, the noncanonical "TATA" element at -20, the proximal H1 homology sequence/"CACCC" element at -30 to -25, the distal H1 homology sequence at -60, the Spl binding element at -80, and the duck H5 homo- logy sequences at -115 to -95. Other regulatory sequences may be present in the region of homology (-30 to -70) that exists between the duck and chicken H5 histone promoter regions. There may also exist regulatory elements further upstream (> -150) or downstream (> +20) that have not been detected by homology searches. ' The chicken H5 promoter was initially analyzed by mak- ing gross deletions in the region between -145 and +20. These deletions were made in the same manner as the linker scanner mutants described in the previous chapter (Figure 15). In this case, however, a single "C" series clone was matched to a series of different "8" series clones. These mutants and the region deleted in each mutant are shown in 103 Table_1 "CACCC"/ duck distal prox. "TATA" CAP hgmol. Spl H1 H11 box .2122 -145/-75 - - + + + + -145/-42 - - - + + + ' * -145/-28 - - - - +/- + -145/+20 - - - - - - *: The -145/-28 mutant deletes the first two thymidines in the "TTAAAT" sequence. It is uncertain if this dele- tion would effect the function of the sequence. Figure 18 and Table 1. This procedure yields a series of mutants which have a common boundary on one side (at posi- tion -145) but which progressively delete further 3' to- wards the H5 histone gene. Also shown in figure 18 is an internal control (+5/+20) which will be described in more detail below. A number of linker scanner mutants have also been con- structed within this region (-145 to +20). A linker scanner mutant requires an exact match between the location Figure 18 104 Size and location of the deletion mutants described in the text. Thin line repre- sents chicken H5 sequence. Putative pro- moter elements are indicated at the bottem of the page. 105 3L db 1r- .J. Fig. 18 oofiuooaoscmuoo ouwm m