CHIGAN STATE UNIVERSITY LIBRARIE /‘ "’ IIIIIII IIIII III IIIIIII a 3 1293 00623 2593 ”exam L Michigan State University This is to certify that the dissertation entitled A. The Search for a Chicken Major Histocompatibility Complex II Alpha Gene B. Transformation-Related Viral Transcripts in Marek's Disease Virus-Transformed Cell Lines presented by Joanne Kivela-Tillotson has been accepted towards fulfillment of the requirements for Ph . D . degree in Biochemistry 4144/44 d “jor professij Date 8/3 /8? MSU is an Affirmatiw Action /Equal Opportunity Institution 0 712771 ,4 "\ II C". I J r K PLACE II RETURN BOX to runove this checkout from your record. TO AVOID FINES return on or before due due DATE DUE DATE DUE DATE DUE MSU Is An Affirmative Action/Equal Opportunity lndltution A. THE SEARCH FOR A CHICKEN MAJOR HISTOCOMPATIBILITY COMPLEX CLASS II ALPHA GENE B. TRANSFORMATION-RELATED VIRAL TRANSCRIPTS IN MAREK'S DISEASE VIRUS- TRANSFORMED CELL LINES BY Joanne Kivela Tillotson A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Biochemistry 1987 ‘u’ ”I ‘11” \I‘) 4} ABSTRACT A. THE SEARCH FOR A CHICKEN MAJOR HISTOCOMPATIBILITY COMPLEX CLASS II ALPHA GENE. B. TRANSFORMATION-RELATED VIRAL TRANSCRIPTS IN MAREK'S DISEASE VIRUS- TRANSFORMED CELL LINES. BY Joanne Kivela Tillotson A. Because the chicken major histocompatibility complex (Ml-1C) has been implicated in disease resistance, studies of the genes were initiated using molecular cloning techniques. Since the RP-9 B—lymphoblastoid cell line is positive for surface MHC Ia antigens, it was used a source for the isolation of cDNA clones which hybridized at low stringency with a human DQ alpha cDNA. Sequence analysis of these cDNA clones, however, revealed that they did not code for MHC proteins, but did represent multiple, tandemly-repeated copies of a 1% base pair sequence of chicken highly repetitive DNA. It was concluded that there is no no alpha gene expressed in the chicken which contains regions that are even 70% identical to the human 00 alpha gene. B. Historically, the avian disease of most economic impartance to the poultry industry has been Marek's disease, a T—lymphoblastoid tumor induced by a herpesvirus, Marek's disease virus (MDV). Although the disease has been well characterized and vaccines developed, little is known about the mechanism of viral transformation. Toward this end, established tumor cell lines were studied to determine their transcription of RNAs from various regions of the viral genane. A ii region of the viral genome from the repeats flanking the long unique sequences was identified which is transcribed in all tumors and tumor cell lines tested, and at higher levels than in productively infected cells. These RNA transcripts appear to be spliced to a region of DNA which may be altered in the tumor cells, coupared to its configuration in DNA obtained from productively infected cells. Further data will be required to define the relationship between the transformed state, the viral transcripts, and the DNA configuration. iii This dissertation is dedicated to Mark and Robin; their patience and love have supported me throughout these studies. iv TABLE OF CONTENTS Chapter A. The search for a chicken major histocompatibility complex Class II alpha gene. INTRODUCTION Proteins of the chicken MHC Functions of the chicken MHC Cloning of mammalian MHC genes MATERIALS AND METHODS CEIl line RNA preparation Plasmid DNA isolation Genanic DNA isolation Bacteriophage DNA isolation RNA electrophoresis and Northern blotting DNA electrophoresis and Southern blotting Probe radiolabelling Hybridizations cDNA library construction DNA sequencing RESULTS DISCUSSION REFERENCES 11 13 18 18 18 19 20 21 22 22 23 24 25 27 29 69 78 Chapter B: Transformation-related transcripts in Marek's disease virus-transformed cell lines INTRODUCTION Biology of Marek's disease vaccines Genetic resistance to MD Other herpesviruses Molecular biology studies of MDV MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES vi 88 89 9B 95 96 97 108 185 107 136 145 LIST OF TABLES Table B1: MDV genomic clones used in transcription survey. 111 Table 82: MDV transcripts identified in survey. 119 Table B3: Transcript sizes detected with probes specific to the long repeat regions of MDV. 124 vii LIST OF FIGURES Al: Genetic maps of avian, murine, and human MHC. A2: Characterization of MHC subloci and molecular composition. A3: Southern blot of genomic DNA from congenic chickens. A4: Northern blot of RP-9 total RNA hybridized with nick-translated pDSlZ insert DNA (DO). A5: Secondary screening filters, after picking positive plaque # 18A on initial screening of the RP-9 cDNA library. A6: Figure A6: Characteristics of recombinant clones from RP—9 cDNA. A7: Southern blots of DNA from recombinant clones. A8a: Mapping of the p38 homology region within pDSlZ cDNA. A8b: Mapping of the p38 homology region within pDSlZ cDNA. A9: Northern blot of RP-9 total RNA hybridized with eitherpDSlZ or p38 cDNA. viii 38 31 34 35 37 40 41 42 A10: A11: A12: A13: A14: A15: A16: Bl: B2: Partial restriction map of the recombinant subclones. Partial sequences of p16 and p38 deduced using the dideoxynucleotide method. Region of homology between p16 and pDSlZ. Consensus sequence of the 10 bp repeats found in p38 and p16 cDNAs. Sequences that were shown in Figure A11, aligned to maximize their identity with the repeat consensus sequence shown in Figure A13. Translation in three reading frames for each of the sequenced regions. Genomic Southern blots of four different haplotype DNAs probed with nick-translated p38. Restriction map of Marek's disease virus. Photograph of ethidium bromide stained gel showing Bamfll or EggRl digested plasmid clones. MDV tumor cell lines used in transcription survey. Northern blots of RNA fran cell lines and infected cells, probed with nick-translated DNA from the short region of the genane. Northern blots probed with DNA from the long unique region. Northern blots probed with DNA from the repeats flanking the long unique region. ix 44 45 52 54 55 62 68 108 109 110 114 116 118 B7: 88: 810: 811: 812: 813: 814: 815: Northern blots of RNA from various tumor cell lines probed with cloned DNA which maps to the inverted repeats flanking the long unique region of the MDV genome. Northern blots of RNA from DEF infected with MDV virus, probed with the same probes as in Figure 87. a. Map of MDV region a right end of long unique segment, including most of inverted repeat sequences. b. DNA fragments used for probes. Identity of EcoRl-X fragment found in different genomic clones. Northern blots of RPl and M881 RNA probed with subclone DNA fragments Tumor cell line Southern blots, probed with BamHI-IZ. Genomic blot of cell line DNA digested with either EggRl or Bgmhl and probed with nick- translated EcoRl-Q DNA. Southern blot of MDV-infected cell DNA, digested with either _E_c_<_>_Rl or 81mm, and probed with nick-translated EcoRl-Q DNA. Northern blots of RNA extracted from in vivo tumors induced by MDV, and probed with either EcoRl-Q or EcoRl-F. 122 123 126 127 129 131 132 133 135 Chapter A: The search for a chicken major histocompatibility complex Class II alpha gene. The major histocompatibility complex (MHC) is a cluster of genes which are involved in controlling a number of critical inmune functions of vertebrates. The majority of the known genes within the MHC are included in the groups referred to as Class I and Class II. These classes code for highly polymorphic, cell-surface proteins, and are members of the inmunoglobulin supergene family (Steinmetz and Hood, 1983.) The MHC antigens consist both of domains which are characteristically "variable" or polymorphic, and also constant danains; it is the constant danains which have strong sequence homology to the inmunoglobul ins (Hood et a1, 1985.) Each domain is usually encoded by a single exon, and follows the general size and structural characteristics of the inmunoglobulin family, including the conserved locations of the disulfide bonds (Flavell et a1, 1986.) The Class I and II antigens are transmembrane proteins which must be recognized in conjunction with foreign antigens in order for the T cell receptor to recognize the antigen and mount an effective inmune response (Malissen, 1986.) Cytotoxic T lymphocytes (CI'L) normally lyse cells expressing a "foreign" antigen on the cell surface, in an attempt to destroy tumor or virus-infected cells, but they can only recognize that antigen in the context of the appropriate me Class I molecule (Zinkernagel and Doherty, 1979.) Apparently, some CI‘Ls can also recognize a "foreign" MHC Class I antigen and attack any cell expressing that antigen. This strong reaction is responsible for the rejection of tissue transplants (Zaleski et a1, 1983,) and Class I antigens are also 3 the traditional transplantation antigens which are responsible for the general name "major histocanpatibility complex.” 0n the other hand, helper T lymphocytes can stimulate 8 cells to produce antibody to a foreign antigen only if that antigen is recognized by the helper T cell in the context of the correct Class II MHC antigen (Flavell et a1, 1986.) These antigens are normally expressed only by antigen-presenting cells such as monocytes and macrophages, and by 8 cells (Kaufman et a1, 1984,) but their presence can be induced on other cell types by many factors, including interferon and probably some growth factors (Flavell et a1, 1986.) Both the human (HLA) and the mouse (Ii-2) major histocompatibility complexes have been extensively studied. In both, there are at least three Class I loci expressed, and two or three distinct Class 11 protein products (each heterodimers encoded by two MHC genes.) All of these antigens exhibit considerable polymorphism—as many as 50-60 alleles being known for some Class I loci, and 25-40 alleles for sane of the Class II loci (Zaleski et a1, 1983.) This polymorphism is one of the most unique characteristics of the MHC. The specific arrangement of alleles at each locus is collectively known as the haplotype; with multiple loci and many alleles available at each locus, the number of haplotypes possible within a species is huge. This large number of haplotypes apparently confers selective advantage upon species in their ability to recognize foreign antigens and effectively resist disease- causing organisms (Longenecker and Mosmann, 1981.) This polymorphism is apparantly maintained by a combination of mechanisms (Quddus et a1, 4 1986): pseudogenes exist in the MHC which have a high rate of mutation and appear to be a reservoir of genetic material used in gene conversion mechanism, which is widely recognized to occur in the MHC (Arden and Klein, 1982; Weiss et a1, 1983; Klein, 1984; Miyada et a1, 1985; Jaulin et a1, 1985), recombinational hot spots occur at various places in the MHC (Steinmetz et a1, 1986; Kobori et a1, 1986), and unequal crossing over has apparently occured in some species, but its role is still not well understood (Steinmetz and Hood, 1983.) The major histocompatibility complex of the chicken, the B—complex, like the manmalian MHC, is composed of a series of genes which control certain inmune functions. It is located on the microchromosome which also contains the ribosomal RNA gene cluster and the nucleolar organizer region (Bloom and Bacon, 1985; Muscarella et a1, 1985.) The 8-comp1ex codes for at least three antigens, B-F, B—L, and B—G, which have been characterized and compared with both the structure and function of the well-studied mammalian MHC gene products (Figure Al; Hala et al, 1981a; Toivanen and Toivanen, 1983.) The B—G antigen does not have a manmalian homolog (Hala et al, 1981b): its unique nature may reflect its presence on erythrocytes, which are nucleated in avian species. The B-F and B—L antigens appear to represent the chicken MHC Class I and II molecules, not only in structure (Ziegler and Pink, 1975,) but also in tissue distribution and in function (Hala, 1981a). Like the manlnalian MHC, the genes are clustered, but appear to undergo less frequent reconbination in the chicken (Koch et a1, 1983; Skjodt et a1, 1985.) Thus, the B L F G chicken < > < > ? 0.039cM .5. I s D E K A,E C4181p D,L mouse < ----- ><--->< > < > 0.02 0.1 0.11 0.19cM D _c2___C4 .2. .2. .9. fl 01534012 aJB£,c4 B c A human < > < ------- > <-->< > 0.7 0.3 0.1 0.7cM Figure A1: Genetic maps of avian, marine, and human MHC. Recombinational distances between loci are shown; multiple antigens coded within each locus are shown for mouse and human (adapted from Hala et al, 1981a.) Class Chicken Mouse Human Cellular Molecular distribution composition I F K,D,L A,B,C All nucleated about 40kD, cells associates with beta-2+microglobulin II L A,E DR,DQ, 8 cells, heterodimer, each DP monocytes about 30kd macrophages III ? Ss,Slp 8f,C2,C4 secreted complement components IV G ? ? erythrocytes heterodimer, Figure 31 and 42 kD A2: Characterization of MHC subloci and molecular composition. 7 complex genetic maps constructed for the murine and human MHCs have not been duplicated in the chicken (see Figure A2,) and assignment of the MHC functions to subregions or specific antigens has assumed parallelism between manmalian and avian MHC structure-function relationships. Hala, 1981a.) More precise understanding of both the individual antigens and their functions would be beneficial in studies of iumune function in various avian diseases (Bacon, 1987.) For example, obese strain (OS) chickens provide a useful model for studying human autoimmne thyroiditis, and an MHC influence has been shown (Bacon et a1, 1974; Bacon et a1, 1977.) Whether information obtained from this model will also be applicable to other autoinmune disorders has yet to be shown. Connections between the MHC and susceptibility to coccidiosis (Ruff and Bacon, 1984; Clare et al, 1985) and fowl cholera (Lamont et a1, 1987) have been postulated. Sane of the most striking disease correlations have investigated the role of the mo in virus—induced neoplasms. Marek's disease is an economically important T-cell lymphana caused by a herpesvirus. Resistance to lymphoma developnent (but not virus infection) is associated with certain B-haplotypes (8ri1es et a1, 1977; Bacon et a1, 1981, 8ri1es et a1, 1983; Bacon et a1, 1983.) Rous sarcoma virus is a laboratory retrovirus which contains the src oncogene and causes tumor development at the site of injection within two weeks, regardless of B- type. However, the B-haplotype governs the fate of the tumor, with complete regression occurring in some haplotypes, but continued tumor growth in other B-types until host death occurs ( Bacon et a1, 1981; 8 Plachy and Benda, 1981, Bacon et a1, 1983, Plachy et a1, 1984.) The long latency tumors developing as a result of ALV infection and subsequent oncogene activation by promotor insertion vary in type depending upon the cell type and the identity of the activated oncogene. There is a B-haplotype influence upon the virus spread (Bacon et a1, 1987), and perhaps also upon the tumor developnent or progression (Bacon et a1, 1981; Bacon et a1, 1983.) None of these disease correlation studies has been able to fully understand mechanisms by which the MHC influence is mediated, or to attribute disease resistance functions to either the Class I or II subregions of the B—complex because of the lack of knowledge about the molecules in the chicken mic and the scarcity of recombinant haplotypes. Strong linkage disequilibrium of the chicken MHC genes is indicated by the existence in field strains of chickens of haplotypes which consistently pair the same sets of alleles at each locus within the haplotype; that is, L2 is always seen with F2 and 62, etc. This is confirmed by numerous studies which have attempted to identify MHC recanbinants within large groups of birds which have been typed serologically. A few recombinants have been identified separating B-G fran the other two loci, but the B-F and B—L have not been seen to recombine in these attempts. Thus, an estimated genetic distance between B—G and B-F/L is 0.039-0.048 centimorgan, and the distance between B—F and B—L is less than that (Hala et al, 1981b; Koch et a1, 1983; Skjodt et a1, 1985.) Thus, the frequency of crossing over between Class I and Class II regions in the chicken is at least one and probably 9 two orders of magnitude lower than in the human and mouse MHC (Skjodt et a1, 1985.) Therefore, a molecular approach to learn more about the chicken MHC would not only contribute to understanding of evolutionary relationships of MHC molecules, but also would lay the foundations for continued study of the inmmological canponents of these diseases, which are either of economic importance to the poultry industry, or are animal models which may prove extremely valuable in the understanding of specific human di seases . PROTEINS OF THE Q-lICKEN MHC In 1947, 8ri1es et a1 (1948) described a group of erythrocyte antigens in the chicken. With the subsequent description of the mouse H—Z complex as the major histocompatibility complex (Gorer et a1, 1948,) the chicken Ea-B antigens were identified as the chicken MHC (Briles, 1950) because of their similar characteristics. A standardized nomenclature for the B-canplex has been adopted (Briles and Briles, 1982; Briles et a1, 1982.) B-F antigens are found on virtually all somatic cells, are highly polymorphic, and consist of a 40-42 kD MHC-encoded transmembrane alpha- chain which associates non-covalently on the cell surface with beta-2- microglobulin (Zeigler and Pink, 1975.) These characteristics identify it as the hanolog of the manmalian Class I antigens, such as the murine H-Z-K, -D, and -L and the human HLA-A, -B, and -C proteins. Although three different, but related, polymorphic Class I antigens occur in both 10 of these manmalian MHCs, the chicken may have only one Class I locus. Inmune precipitation of antigens and subsequent two dimensional gel analysis have shown a 40-42 kD peptide, non-covalently joined to beta-2- microglobulin, but have not provided clear evidence for more than one locus (Kubo et a1, 1977.) Limited N-terminal microsequencing of inmunoprecipitated protein shows that B-F is clearly homologous to the manmal ian Class I antigens; the sequence homogeneity suggests that a single locus is expressed in each line (Vittetta et a1, 1977; Huser et a1, 1978.) It may well be that the chicken has only a single Class I gene, and therefore is considerably less canplex than the mamnalian species in this major histocompatibility complex locus. However, in one case, sequential inmunoprecipitation by separate antisera have indicated the possibility of two separate Class I antigens (Bisati and Brogen, 1980 .) B-L antigens are the chicken Class 11 proteins. They are polymorphic antigens found on sane cells of the monocyte/macrophage lineage, as well as on B-cells and activated T—cells (Ewert and Cooper, 1978; Ewert et a1, 1984; Hala et a1, 1984.) The proteins expressed on the cell surface consist of a 33-34 kD non—polymorphic alpha-chain and a variable beta chain; this dimer is bound noncovalently with an invariant chain intracellularly, before its expression on the cell surface (Crone et al, 1981a; Guillemot et a1, 1986.) The size of the beta chain has been reported at 27 kD (Ewert et a1, 1984) or 32-36 kD (Guillemot et a1, 1986); No amino acid sequence data is available, but antisera of varying reactivities have indicated the possibility that two distinct 8- L molecules may exist on some cells (Crone et al, 1981b.) This is in 11 agreement with the two—dimensional gel data of Guillemot et a1 (1986), who found a single alpha chain, but two distinct beta chains. Therefore, although it appears that the Class II region of the chicken is less canplex than the hanologous human region, it may be more similar to the mouse in complexity. B—G is the third antigen mapped to the chicken MHC. It is found only on erythrocytes, consists of two chains of 31 and 42 kD, does not associate with beta-Z-microglobulin, and is highly polymorphic (Hala, 1981a.) There is no mamalian homolog which has been described. FUNCTIONS OF THE CHICKEN MHC Because recombinants have been identified which separate the B—G region from the remainder of the MHC, functional testing of the role of the B-G antigen has been possible. Such experiments have revealed that it is not involved in the well-known MHC characteristics such as allograft rejection, GVH reaction, inmune response to specific antigens, resistance to Marek's disease, or regression of RSV induced tumors (Briles et a1, 1983; Vainio et a1, 1984; Plachy et a1, 1984.) B-G does appear to have an adjuvant effect, however, in that injection of erythrocytes expressing both new F and G antigens produces high titers of both anti-F and anti-G antibodies. But if only the F antigen differs, few or no chickens produce detectable anti-F antisera; if only 12 G differs, all chickens seem capable of responding with antibody production to G (Hala et al, 1981b). The B—F region has been shown to regulate allograft rejection (Hala et al, 1981a) and GVH reaction, which can be blocked by specific anti-F antisera (Simonsen et al, 1977, Vilhelmova et a1 1977.) These funcions are hanologous to the functions determined by manmalian Class I antigens, and help to confirm the identity of B-F as the true Class I region of the chicken MHC. The 8—L region is assumed to play a role in inmune response similar to the control exerted by the manmalian Class II antigens (Hala et al, 1981a.) In one study, B-L control of T-B cell interactions was studied using H.819 chickens which have been serologically typed as B-F19, B— 619, and B—L12. Only the B-L antigen identity was significant in controlling the adoptive transfer of bursal cells (vainio et a1, 1984.) It has been assumed that the high/low response to the synthetic polypeptide GAT as studied by Pevzner et a1 (1978) would be controlled by this region, and preliminary work with antisera produced by reciprocal innunizations showed that the antigens targeted by the antisera were of the correct size for B—L antigens (Birkemeyer and Nordskog, 1982; Pevzner et a1, 1978,) but to date this has not been confirmed by other investigators (S. Lamont, personal conmunication.) Whether this Ir-GAT gene actually codes for an MHC antigen is still uncertain, but there does appear to be a linkage between Ir-GAT and Rous l3 regression. (Gebriel et a1, 1979; Lee and Nordskog, 1981; Gebriel and Nordskog, 1983.) Thus, the major functions attributed to the manmalian MHC are also present in the chicken MHC. There is sane evidence that certain of these functions are associated with the homologous proteins, and there is no evidence to date that the remaining functions are not correlated with the hanologous proteins. Therefore, it is a reasonable assumption that the chicken MHC is highly analagous in both structure and function with the manlnalian MHC, although its apparent decreased complexity makes it a useful model for studying correlations of the MHC with various questions of disease resistance and inmune function. Therefore, further molecular and functional characterization of the MHC is highly desirable. A first step in this characterization could be the cloning of various regions of the MHC to facilitate carparison of the genes at a molecular level by restriction fragment length polymorphism analysis, and to deduce the protein sequences for further carparison with known manmalian proteins. CLONING OF MAMLIAN MHC GENES The decade of the 80's has seen considerable progress in the elucidation of manmalian major histocarpatibility genes, with a variety of laboratories using a variety of methodologies for isolation and identification of the MHC genes. Early efforts at understanding the structure of the Class I antigens included the laborious task of protein 14 sequencing (Coligan et a1, 1981, Maloy et a1, 1981, Nathenson et al, 1981, Ezquerra et a1, 1985), and more recently have included X-ray crystallography and determination of three-dimensional structure (Bjorkman et al, 1985.) The information gained fran sequencing has proven immensely valuable in subsequent cloning efforts, both in confirmation of the gene identity after DNA sequencing and in synthesis of oligonucleotides to be used in the cloning strategies. For example, Sood et al (1981) used a synthetic mixture of ll-mers, derived from a reported conserved BLA 3'-peptide sequence, in a primer extension experiment to produce a specific 30mer using reverse transcriptase, ddATP, and enriched mRNA. After sequencing it, they were able to use this specific 30—mer as a probe to select specific HLA clones fran a cDNA library (Sood et a1, 1981.) These clones were then used by others at lower stringency to select clones from mouse cDNA and genomic libraries (Steinmetz et al, 1981a, Steinmetz et al, 1981b, Steinmetz et a1, 1982.) Early cloning and characterization also depended heavily upon the use of highly specific antisera. For example, Ploegh et a1 (1980) and Kvist et a1 (1981) constructed human and mouse cDNA libraries respectively, both fran poly A+ RNA which had been size fractionated, then selected fractions which could be translated in vitro to produce inmunoprecipitable products of the correct size. After constructing libraries, the plasmid DNAs were used to hybrid select mRNA which was tested in in vitro translation and inmunoprecipitation. The BIA-B7 clone (Ploegh et a1, 1980) provided sequence information which later proved useful in further cloning attempts (Sood et a1, 1981.) Kvist's 15 pH-2d-1 clone was used by others to clone additional mouse (Bregegere et a1, 1981) and human Class I sequences (Jordan et a1, 1981; Malissen et al, 1982a; Malissen et al, 1982b.) A nImIber of groups, using various procedures, reported the cloning of class II antigen genes beginning in 1982. Again, the procedures depended heavily upon good, specific antibodies. In the approach taken by Wake et al (1982) and Long et a1 (1982), a cDNA clone bank was constructed fran B-lymphoblastoid mRNA enriched for specific DR mRNA as assayed by injection into Xenopus laevis oocytes , and inmunoprecipitation of the translated products. Individual clones were then identified which were able to hybrid select mRNA which could be translated to give the correct, inmunoprecipitable products. These DR alpha and DR beta clones were later used to identify additional human genomic genes (Gorski et a1, 1984; Tonnelle et a1, 1985) as well as in cloning the entire mouse Ia region (Steimnetz et al, 1982.) Another approach requiring specific antibodies was taken by Korman et al (1982a) in cloning DR alpha cDNA by innunoprecipitation of polysomes to yield the specific mRNA used for library construction and for 32P-cDNA probe preparation. The identity of clones was confirmed by hybrid selection of mRNA, in vitro translation and inmunoprecipitation of products. This clone was later used to select genanic clones, identify the exon/intron organization, and confirm the protein sequence (Korman et al, 1982b; Lee et a1, 1982,; Schamboeck et a1, 1983.) It was also used by Auffray (1982) in differential 5'- and 3'-hybridization to detect related, but not hanologous clones. A third approach to cloning the DR alpha chain used a combination of methods including mRNA enrichment by selecting 16 manbrane-bound polysomes to prepare RNA for the cDNA cloning, and used a synthetic oligonucleotide probe prepared after primer extension experiments identified the correct nucleotide sequence out of the possibilities deduced fran the available amino acid sequence (Stetler et a1, 1982.) Stetler's human DR alpha cDNA clone was then used to identify mouse E and A alpha cDNA and genomic clones (Benoist et a1, 1983; Mathis et a1, 1983.) A similar primer extension/oligonucleotide probe strategy was employed to identify both DR alpha and DR beta clones (Das et al, 1983a; Das et al, 1983b; Bell et a1, 1985.) A final methodology used by Davis et al (1984) was successful in locating the mouse A alpha gene. Cosmids coding for the entire Ia region (Steinmetz et al, 1982) were probed with B—cell 32P—cDNA after it had been hybridized with T—cell mRNA to remove conmon sequences. Similar subtracted cDNA was also cloned. In most of the examples given here, the identity of the Class 11 genes was confirmed by inmunoprecipitation of the protein coded for by the clones. However, particularly for some of the Class I genes, identity of the genes has been established by transfecting L cells, or 8—ce11 lines, and detecting a new specificity with antibodies (Goodenow et al, 1982a; Goodenow et al, 1982b; Goodenow et a1, 1983; Zuniga et a1, 1983.) The cloning and transfection techniques have enabled researchers to dissect functions of MHC gene products in new and productive ways 17 (Wbodward et al, 1982; Reiss et a1, 1983; Malissen et a1, 1984; Shimizu et a1, 1986; VOgel et a1, 1986.) Even more elegant functional expression of both Class I and Class II antigens have been studied in transgenic mice (BieberiCh et al, 1986; LeMeur et al, 1985; Yamamura et a1, 1985.) The technology is available and the desire is certainly present for similar functional assays of genes related to disease resistance, including MHC genes, in the transgenic chicken (Crittenden, 1986; Salter et al, 1987.) 18 MATERIALS AND METHODS Cell line: LSCC-RP-9 is a B-Iymphoblastoid cell line transformed with RAV-Z virus (Okazaki et al, 1980). Cells were maintained in Leibowitz- McCoy medium with 10% bovine fetal serum, 20% chicken serum, 5% tryptose phosphate broth, and 0.1% fungizone, at 41 C with 5% C02' RNA preparation: RP-9 cells were washed with sterile PBS and collected by centrifugation at 100 x g for 5 minutes. A cell pellet containing 108 cells was lysed by pipetting up and down in 8 m1 6M guanidinium isothiocyanate, containing SmM sodium citrate (pH 7.0), 0.1M beta- mercaptoethanol, and 0.5% sodium lauryl sarcosinate. Cesium chloride was added to 2.4M and the solution was layered over a 3 m1 cushion of 5.7M CsCl in an SW41 tube. This was centrifuged at 30,000 rpn at 20 C for 24 hr. The supernatant was carefully poured off and the walls of the tube were wiped to remove as much liquid as possible. The RNA pellet was resuspended in 400 ul water, NaCl added to 100m, and 2,5 volumes of absolute ethanol added. The RNA was precipitated at ~80C overnight or longer, and collected by centrifugation for 15 min in an Eppendorf microcentrifuge. The RNA was pelleted, washed carefully with 70% ethanol, dried, and resuspended in water inmediately prior to use. Polyadenylated RNA was selected on a column of oligo d(T) cellulose. Pelleted and dried RNA was dissolved in water containing 1mm vanadyl riboncleoside complexes and heated to 65C for 5 minutes. An equal volume of 2x RNA loading buffer was added, and the cooled mixture was applied to a 1 ml column of oligo d(T) cellulose (Collaborative 19 Research, Inc., Lexington, MA) which had been washed thoroughly in succession with 0.1M NaOH, 0.005M EDTA, water, and 1x loading buffer. The colunm was washed with 1x loading buffer until the 00260 of the effluent was 0 (about 10 ml). The poly (A)+ RNA was eluted with several m1 of 101m Tris (pH 7.5), 11m EDTA, 0.05% SDS, and the concentration was estimated by 00260 absorbance. Sodium chloride was added to 0.1M and 2.5 volumes of absolute ethanol was added before incubation overnight at -8gCo Plasmid DNA isolation: pDSlZ was the generous gift of J. Silver (Chang et a1, 1983,) and includes the entire coding sequence of a human Class II (00) alpha chain. pRAV-Z DNA, used as a control plasmid, was the gift of E.J. Smith. Plasmids were grown in E. coli strains H8101 or bus at 37C with shaking in LB broth, plus the appropriate drugs. Plasmids using the vectors pBR322 or pBR328 were grown to (D of 0.5-0.6, 600 chloramphenicol was added to 150ug/ml to amplify yield, and incubation continued for a further 12-16 hr. Plasmids using the vectors pGEMB or pGEM4 were allowed to grow under the same conditions for 20 hr, with no ampl if ication step . The cells were pelleted at 2500xg for 5 min and the media removed. Cells were resuspended in 1% original volume of GET buffer (usually 5 m1, 5011M glucose, 251m Tris, pH 8.0, 10mM EDTA), and placed in a clean tube on ice. Ten m1 of 1% SDS, 0.2M NaOH, was added, mixed gently, and left on ice for 5 min. 7.5 m1 of cold potassium acetate (5M acetate, 3M potassium) was added with a few quick inversions of the tube, and incubated on ice for 5-10 min. After centrifugation for 10 min at 20 15000xg at 4C, the supernatant was removed to a new tube an mixed with 0.6 volume iSOpropanol. After 10 min at room temperature, the DNA was pelleted by centrifugation at 15000xg for 10 min. The pellet was drained very well and resuspended in 5 ml TE. 50u1 heat-treated RNase (10mg/ml RNaseA, l0kU/ml RNaseTl) was added, and the sample was incubated at 37C for 30 min. After addition of NaCl to 0.5M, the sample was thoroughly extracted with an equal volume of phenol :chloroform (1:1), and the aqueous phase was precipitated with two volumes of ethanol for at least 1 hr at -20C. The pelleted INA was recovered after centrifugation at 15000xg for 10 min. The DNA was dissolved in 5 m1 TE, NACl was added to 1.5M and 0.25 volumes of 30% PEGB000 in 1.5M NaCl was added. After mixing, the solution was incubated on ice for 1 hr, and centrifuged at 4C at 15000xg for 15 min. After draining thoroughly, the pellet was rinsed gently with 70% ethanol and dried. The pellet was resuspended in TE and the concentration estimated from absorbance readings at 260 nm. Genomic DNA isolation: Red blood cells (RBC) were obtained from 1515 congenic lines of White Leghorn chickens developed and maintained at the Regional Poultry Research Laboratory in East Lansing (Bacon et al, 1987b). Approximately 0.2 ml packed RBCs were resuspended in 10 ml NET buffer (100nm NaCl, 1w EDTA, 50 nM Tris, pH 7.5). Pronase was added to 0.5mg/ml and SDS added to 0.5%, and the mixture was incubated at 37C overnight. 10ml phenol:chloroform (1:1) was added and mixed gently by inversion. After centrifugation to separate the phases, the aqueous phase was removed to a new tube using a large bore pipet to minimize 21 shearing of the DNA. This extraction was repeated several times until the interface was clear. The sample was then extracted one time with chloroform. RNaseA was added to 50ug/ml and the sample was incubated for 1 hr at 37C. The sample was extracted again with phenol:chloroform (1:1), and then with chloroform. The aqueous phase was then dialyzed exhaustively against TE, and the DNA concentration was determined by measurement of absorbance at 260 and 280 nm. Bacteriophage DNA isolation: Minipreps were grown by infecting 100u1 of overnight E. coli culture with 1/20 of a resuspended plaque from a plaque purified preparation. This was incubated 12-15 hr at 37C after plating in 3 ml 0.75% agarose on L8 agarose plates. The top agarose was scraped off into 5 m1 SM (100w NaCl, 10mM M9804, SmM Tris pH 7.5) plus l00u1 chloroform, and incubated at 4C overnight to elute the phage. Debris was pelleted by centrifugation at 12000xg for 15 min, and the supernatant containing the phage was placed in a new tube. NaCl was added to 1M, and polyethylene glycol (MW=6000) was added to 10%. After all solids were dissolved, the solution was cooled to 4C on ice, and incubated on ice for at least 1 hr.The precipitated phage particles were collected by centrifugation at 15000xg for 30 min, and the tube was drained well. The pellet was resuspended in 400ul, transferred to an Eppendorf tube, and extracted once with chloroform. EDTA was added to 20m, proteinase K to S0ug/m1, and SDS to 0.5%, and the mixture was incubated at 65C for 1 hr. The mixture was extracted 2 times with phenol and 2 times with chloroform, and 2 volumes ethanol was added to precipitate the DNA. 25% of the preparation was used for a restriction 22 digestion: if the concentration was too high or low, another sample was run with the corrected amount. RNA electrophoresis and Northern blotting: Gels were prepared consisting of 1% agarose, 2.2M formaldehyde, and 1x gel buffer (200:!!! MP8, 50mM sodiumacetatelpH 7.0], and 1011M EDTA.) RNA samples containing 20ug total RNA, 50% formamide, 2.2M formaldehyde, and 1x gel buffer were heated to 60C for 5 min and cooled on ice. Before loading on the gel, 1/4 volume of 50% glycerol containing l w .TA, 0.4% bromophenol blue and 0.4% xylene cyanol was added to each sample. RNA was subjected to electrophoresis at 40 volts, 80 milliamps, for 16 hr. Lanes containing RNA markers were cut off, stained with l ug/ml ethidium bromide in 0.1M beta-mercaptoethanol, washed in distilled water, and photographed. The remaining lanes were soaked in 20xSSC for 30-60 min and transferred to nitrocellulose with 20xSSC by passive transfer overnight. The dried filter was baked at 80C for 2 hr before hybridization. DNA electrophoresis and Southern blotting: DNA was digested in the enzyme buffer reconlnended by the vendor, in a volume appropriate for the size of the electrophoresis well, usually 20-100ul. Samples included 15ug genomic DNA, l-2ug phage DNA or 0.5-1ug plasmid INA. Digestion was carried out for 2-18 hr at the recannended tauperature, using 1-20 enzyme/ug DNA. One tenth volume of 20% Ficoll, 101m EDTA, 0.5% 23 bromophenol blue, 0.5% xylene cyanol was added to stop the reaction and prepare the sample for loading on the gel. Agarose gel electrophoresis was carried out in TBE buffer (0.089M Tris, 0.089 M boric acid, 1mM EDTA), with the gel concentration, voltage, and time varying depending upon the size of fragnents to be separated. Genanic DNA was electrophoresed through a 18x15cm gel containing 0.8% agarose at 35 volts for 16 hr. After electrophoresis, the gel was removed to a shallow dish containing 0.1ug/ml ethidium bromide, and soaked at roan tanperature for 30 min with occasional shaking. The gel was photographed under ultraviolet illmnination to identify the location of bands. The DNA was denatured by soaking the gel in 0.5M NaOH, 1.5M NaCl for 45-60 min, and then neutralized by soaking for 45-60 min in 0.5M Tris, pH 7.5, 1.5M NaCl. The DNA was passively transferred to nitrocellulose filters (S88, Keene, NH) using l0xSSC, for 16 hr (cloned DNA) or 24-48 hr (genanic DNA.) The filters were baked at 80C for 2 hr. Probe radiolabellirg: All DNA to be used for low stringency hybridization probes was first digested with restriction enzymes and the specific DNA insert was gel purified. After staining with ethidium bromide and visualizing bands under UV light, the appropriate band was cut out with a razor blade, placed in a small piece of dialysis tubing with 2-5 ml TBE, and electroeluted at 50-80 volts for 2-4 hours, until the DNA was all in solution. The DNA was then purified on an Elutip—d 24 column (888, Keene, NH), according to manufacturer's instructions, and concentrated by ethanol precipitation. For nick translation, 100-500 ng DNA was added to a tube containing 50- l00uCi 32 P-alpha-dCI‘P (NEN, >800uCI/mMole) in 50Mris, pH 7.2, l0mM M9504, 0.11m DTl‘, 50ug/ml bovine serlml albumin (BSA), 20uM dATP, 20uM dGTP, and 20uM dTTP. To this mixture was added 5U E. coli DNA polymerase I and 0.5u1 of 10-4mg/ml DNaseI. After mixing, the reaction was incubated at 14C for 1.5 hr. Separation of probe fran unincorporated nucleotides was accomplished by either multiple ethanol precipitations or passage over Sephadex G-50. Incorporation was usually in the range of 25-65%, yielding probes labelled to a specific activity of 107 to 108dpn/ug. Hybridizations: All low stringency hybridizations were prepared in 20% formamide, 5xSSPE (1xSSPE=0.18M NaCl,l0mM sodium phosphate, pH 7.4, 11m .TA), 2x Denhardt's (lx=0.02% each Ficoll, polyvinylpyrrolidone, and BSA), and 100ug/ml denatured, sheared herring sperm DNA. Filters were pre-hybridized in this solution for 1-15 hours at 25C, and then incubated in the same solution plus 5x106-5xl07 dpn nick-translated probe for 20-48 hours at 25C. Filters were then washed twice with 2xSSC at 25C for 15-30 min each, and then 2 times in 2xSSC at 56C for 15 min each. "Very" low stringency was the same conditions except the final washes were done at 50C. "Moderate" stringency was the same conditions except the final washes were done at 65C. "High" stringency was the 25 same conditions except that the final washes were done in 0.2xSSC at 65C. cDNA library construction: Poly A+ RNA was used as a template for ch synthesis using reverse transcriptase and RNaseH, using the reagent purchased fran Amersham Corp. Briefly, 5ug RNA was mixed with 1000 AMV reverse transcriptase, in a 50 ul volume which included 10mm Tris, pH 8.3, 10!!!! MgC12,140 mM KCl, 100 ug/ml oligo dT 0.SU.ul RNasin, 1mM 32 18' dATP, 1mM dCI‘P, 1mM dGTP, 111M dTTP, and 20uCi P-alpha—dCI'P. The reaction was allowed to continue at 42C for 45-90 minutes. Second strand synthesis was then initiated by addition of 50 RNaseH and 115 U E. coli polymerase I, in a total volume of 250ul, containing 0.1M HEPES, pH 6.9, 10 mM MgC12, 2.5mM D'I'I‘, 70mM KCl, and 1mM of each dNTP. This reaction was incubated at 14C for 2 hours; then the reaction was heated to 70C for 10 min to stop further enzyme activity, and cooled on ice. T4-DNA polymerase (10 U) was added to repair single-stranded ends, and the reaction was incubated at 37C for 15min. Further enzyme activity was stopped by addition of EDTA and SDS to 1011M and 0.1%. The cDNA was then extracted with phenol:chloroform (1:1) and then chloroform, and concentrated by ethanol precipitation. In order to block any internal Echl sites from cleavage during the cloning process, the cDNA was treated with 100 §c_o_R1 methylase in 100w Tris, pH 8.0, 1mm EDTA, 400ug/ml BSA, and 80uM S-adenosyl methionine for 1 hour at 37C. After ethanol precipitation, the cDNA was mixed with lug phosphorylated 8-mer EgRl linkers (GGAA'I'I'CID and 10 T4-DNA ligase in 70mM Tris, pH 7.6, 10 um MgClZ, SmM MT, and 50um ATP. The ligation 26 was allowed to proceed for 16 hours at 12C; the cDNA was digested with 50U EggRl for 2 hours at 37C, and excess linkers were removed by mul tiple spermine prec ipi tat ions , after extract ions with phenol:chloroform, and chloroform. The spermine precipitations involved adding 360ul TE, 40ul DMSO and l0ul 100m spermine-HCl, pH 6.8, and then freezing briefly in dry ice, followed by slow thawing on ice and centrifugation for 30 min at 4C. The pellet was washed 3 times with 75% ethanol and dried. The precipitated cDNA was taken up in 1x ligation buffer (301m Tris, pH 7.5, 30 mM NaCl, 4m MgClZ, 0.5mM ATP, 21m DTT, 100ug/ml EA), and 50 T4-DNA ligase and 0.5ug dephosphorylated lambda gtl0 arms were added. Ligation was allowed to continue at 12C for 16 hours. Phage DNA was then packaged in an in vitro packaging extract (Pranega Biotec) by incubating in the freshly thawed mixture for 2 hours at 22C. To titer the recombinant lambda phage produced in this library, serial l0-fold dilutions were made and 100ul of each dilution was added to 200u1 overnight cultures of C600 or C600hf1 strains of E. coli. After allowing 15 min at 37C for the phage to attach to the bacteria, 3 m1 molten top (0.75%) agar was added to each sample. The samples were mixed gently and poured onto 100nm plates containing 25 ml 1.5% agar in LB broth. After the top agar had hardened, the plates were inverted and incubated for 12-16 hours at 37C. Plaques were counted on both C600 and C600hf1; gtl0 wild—type will grow lytically on C600, but forms a lysogen on C600hf1. When recombinants are produced at the EcoRI site, 27 insertional mutagenesis prevents lysogeny on 0600hf1, and the recombinant phage grow lytically on this strain as well. DNA sequencing: DNA sequencing was perforned by the dideoxynucleotide nethod, directly on the double-stranded plasmids, using primers which were canplementary to the SP6 and T7 pranotor regions flanking the polylinker cloning site of pGEM plasmids (Pronega Biotec.) The purified plasmid template DNA was linearized using _;Pv_u_II, which cuts only in the vector. After phenol:chloroform extraction and ethanol precipitation, lug of recombinant plasmid and 30ng of appropriate promotor primer were mixed in a l0ul volume, containing 1011M Tris, pH 7.5 and 5m NaCl. The mixture was heated in a boiling water bath for 3 min, then chilled quickly in ice water for 5-15 min. To the annealed tanplate mix, 4ul [32P]-alpha-dATP (DuPont, 800Ci/nlnole, l0mCi/ml) was added and mixed; 3u1 of this mixture was added to tubes for each of the four reactions, containing 4ul including 1.25 U Klenow enzyme in 34 mM Tris,pH 8.3, 611M MgClz, 511M DTT, 5mm NaCl, 250 nM each dGI‘P, d'I'l‘P, and dCI‘P, and one of the following: 100uM ddCI‘P, 3.6uM ddATP, 200uM ddTTP, or 50uM mm. The tubes were mixed and incubated at 37C for 15 min; then 1 ul chase solution was added (containing 2mM concentration of each dATP, dCI‘P, dGI‘P, and dTTP in the sane buffer) and the tubes were incubated at 37C an additional 15 min. The reactions were stopped by adding 5ul of a solution containing 98% formamide, 101m EDTA, 0.3% xylene cyanole FF and 0.3% branophenol blue. The reactions were heated to 70C for 3 min 28 innediately before loading 2.5 ul of each reaction onto adjacent wells of a sequencing gel. Sequencing gels were 6% polyacrylamide (Bis:acrylamide = 1:20), and 8M urea in leBE buffer; the gel size was 30x40x0.04 an. Gels were pre-run in an 181 sequencing apparatus at 70-75 watts for 30-60 min before loading samples, and were run at the same power settings for about 2 hr (until the bromophenol blue reached the bottom of the gel.) Usually a second loading of samples was applied to the gel and electrophoresis was carried out for an additional two hours. Occasionally, a third application was also done. When the branophenol blue of the final loading neared the bottom.of the gel, the power was stopped and the gel removed from the apparatus. The gel was fixed by soaking in 5% acetic acid/5% nethanol, transferred to Whatman 3m paper, dried, and exposed to X-ray film. 29 RESULTS In order to test whether manmalian cDNAs would hybridize to homologous regions of the chicken genone, digests of DNA from White Leghorn 15I B— 5 congenic chickens were Southern blotted and probed with a variety of cloned cDNAs. Figure A3 is an autoradiogram of DSlZ-probed genomic blots, and shows the hybridization of 0812 to chicken genomic DNA digested with E1. A major band of 4.55 kb is seen in all lanes, with both larger and smaller minor bands also seen in several lanes. Northern blots of total RNA isolated from the Ia+ cell line, RP-9, are shown in Figure A4. A single RNA of about 1.95 kb, which hybridized to the 0812 probe is seen in lane 2, and marked with an asterisk. Lane 1 shows an autoradiogram of the same RNA blot after hybridization with pRl.9, a clone consisting of single copy chicken sequences which are not transcribed in RP-9 cells (unpublished observations); the background due to the low stringency hybridization conditions showed up in the rRNA bands of both lanes, and was clearly different from the specific hybridization of 0812 to the 1.95 kb RNA. In order to clone cDNA of this 1.95 kb RNA in RP-9 cells, we isolated poly A+ RNA fran RP-9 cells, used reverse transcriptase to make double stranded cDNA, added linkers, and cloned into the E3931 site of lambda gtll. After in vitro packaging, the numbers of total and recombinant phage were titered on C600 and C600hfl strains of E. coli, and the library was amplified before further screening. There were 25000 reconbinant phage, which was 53% of the total library. 30 B2 812 815 313 319 35 _— —_—— 4.55- ,. ‘,-....._.......-.—.—- Figure A3: Southern blot of genomic DNA fran congenic chickens. DNA from chickens of the labelled haplotypes, was digested with PstI and hybridized with nick translated pDSlZ insert DNA at low stringency. The major band seen in all lanes is 4.55 kb. 31 Figure A4: Northern blot of RP-9 total RNA hybridized with nick- translated pDSlZ insert DNA (DQ). Position of the 28S and 18S ribosomal RNA is shown, determined from low stringency hybridization with a control plasmid (C).. 32 The amplified library was plated on six 150 nm plates, with 50000 total plaques per plate. Duplicate filters were pulled fran each plate and processed according to the protocol of Benton and Davis (1977.) These filters were hybridized at low stringency with nick-translated DSlZ insert DNA. The 32 regions which included any plaques whose autoradiogram signals were clearly above background on both filters were picked and eluted into 1 ml SM + 50ul CHCl at 4C overnight. Each of 3 these phage stocks was diluted and aliquots of each dilution plated on 100 um plates. Triplicate filters were processed as before, with two hybrized to 0812 and the third to pRAV-Z. A sample of filters processed from one plate (10A) is shown in Figure A5; panel A shows very low stringency washing, with all plaques exhibiting similar high signals; panel B shows the same filter after "normal" low stringency washes; panel C. shows the duplicate to B, with arrowheads pointing out sane of the positive duplicated signals; panel D shows a third filter from the sane plate hybridized with pRAV-Z, and not showing positive signals with the same plaques . Out of the 32 plaques picked from the primary screening, 20 remained positive through the secondary screening and plaque purification procedures. These isolates are listed in Figure A6 along with details of characterization of their cDNA inserts. All 20 clones were grown as plate lysates and minipreps of the gtll DNA were prepared, digested with EcoRl, and electrophoresed to analyze the cDNA inserts. Only 15 of the clones resulted in preparations of 33 Figure A5: Secondary screening filters, after picking positive plaque # 10A on initial screening of the RP-9 cDNA library. Filter A was washed at very low stringency; Panel B shows the same filter after rewashing at normal low stringency; Filter C is the duplicate filter, washed as in Panel B: All three were hybridized with nick-translated pDSlZ insert DNA. Filter D was hybridized with the control pRAV-Z plasmid probe. Arrows in panels 8 and C point to plaques which are positive on the duplicate filters. Figure A5: 35 Clone secondary screen insert DNA Subclone DS pRAV2 size DS p38 obtained? hybridization hybridization le + + 2800bp ++ - - 3A + - ? - + — 4A + - 450 - - - 7A + - ND 9A + - 375 + + - 10A + - 1500 ++ ++ p38 11A + - 800 ++ ++ - 12A + - 400 + + - 13A + - ND 14A + - ND 17A + - 400 - - - 18A + — 375 ++ ++ p16 21A + - 400 + + - 22A + - 400 + + - 23A + - 400 + + - 25A + - ND 26A + - 400 + + — 27A + - ND 28A + - 175 - - - 29A + - ? - - - Figure A6: Characteristics of recombinant clones from RP9 cDNA. 36 Figure A7: Southern blots of DNA from recombinant clones. Panel A shows the results of a Southern blot of the lambda DNA isolated from eight of the recombinant clones isolated, after probing with the nick- translated pDSlZ insert DNA.. Panel B shows the same filter after removing the probe and rehybridization with nick-translated p38 plasmid DNA, and washing at high stringency. 3 4 91011121718 p38 Figure A7: 38 adequate DNA for restriction analysis. Two of these produced vector bands, but no visible insert bands. This could have been due to the small quantity of DNA digested, the very small size of the insert, or one of the Echl sites could be missing at the ends of the insert. The visible inserts ranged from 175 to 2800 hp, with nine of them being around 400 bp. These fragments were transferred to nitrocellulose and hybridized with D812 to ensure that the positive signal detected in the phage screening was due to hybridization of the probe to the insert DNA (Figure A7a). Ten of the 15 clones had detectable inserts hybridizing with 0812. Three of the inserts detectable with ethidium branide were not recognized by hybridization with D612; there was no hybridization seen with either of the clones which had no detectable inserts by ethidium bromide staining. The DNAs from several of the clones which showed the most positive signals were digested with EggRl, mixed with EcoRl-cut and phosphatased pGEM-4 DNA, ligated, transformed into canpetent DHS bacteria, and selected on L8 plates containing 25ug/ml ampicillin. For two of the clones, # 10A and 18A, subclones of the cDNA were obtained, designated p38 and p16. Since 10A has the longest cDNA insert hybridizing specifically with 0812, additional subcloning was not deemed necessary at this time. In order to examine the relationships between the various clones, nick- translated p38 was used to hybridize at high stringency to the same filter seen in figure A7a, after the D812 had been removed. All of the inserts previously positive with the D512 probe were also positive with 39 p38 (Figure A7b.) In addition, one of the inserts (3A) which had not been detected by either ethidium bromide or 0812 hybridization, was positive using p38; we assune that the quantity of DNA in this fragment was too small to be seen with the first two techniques, but was adequate for high stringency hybridization with an homologous probe. Thus, there was no evidence that we had isolated any D812-hybridizing clones which did not hybridize to p38; p38 was therefore assuned to include most of the sequence of the mRNA hybridizing to 0812. In order to determine which portion or portions of D812 were homologous to the cDNA clones which had been isolated, the D812 plasmid was digested with various enzymes and the separated fragments were Southern blotted and probed with p38. The enzymes used did not cut within the vector, and the pertinent restriction sites with the D812 insert are shown in Figure A8a, along with sizes of the major fragments. Figure A8b shows the autoradiograms after probing these fragments with p38. The large AEI fragment ,which includes the entire coding region for the mature polypeptide chain, does hybridize to p38, indicating that the region of homology is within the D812 coding region. The _S_t_P_I and £91 sites, when double digested with 531 remove the cytoplasmic and transmembrane regions, respectively. The AEI-flul fragment retained the ability to hybridize, but this was entirely removed in the ApaI-Sagl fragment. This indicates that the entire low stringency hybridization involves this 90 bp 8331-8131 fragment, which encodes primarily the transmembrane reg ion. 40 .moEmNco @3835 on» :33 cofiumwmwo Houwo mucmsmmum no woman owuomaxo on» mongoose“ can .393 .830 5,50 c355 on» you ace cofiuowuummu Luciana o mzonm o amend .mzoo N309 :3»; cofimmu b.6328: 93 may no means: "we 39mg A I 1' Ono v Ofik ' mun .AI v mko 119:! 1'" [NS [009 106V 113:! Se :2 5:8— uu _ B .5.» 41 APO/Stu Ape/Soc o o. < Figure A8: Mapping of the homology region within pDSlZ cDNA. Panel b shows the results of a Southern blot of fragments from pDSlZ, probed with the p38 chicken cDNA clone. 42 DQ p38 3 f id. Figure A9: Northern blot of RP-9 total RNA hybridized with either pD812 or p38 cDNA. Filters were washed at low or high stringency, as designated. 43 In order to ascertain whether the p38 cDNA represented the expected 1.95 kb RNA originally seen with the D812, we ran Northern blots of RP-9 RNA as in Figure A4, and probed adjacent lanes with nick-translated p38 and D812. The autoradiogram is shown in Figure A9. At low stringency, both probes hybridized to the 1.95 kb RNA, although p38 also recognized a small amount of larger molecular weight RNA. After high stringency washes, the 0812 probe washed off, while the p38 ranaind hybridized, indicating extensive homology with the RNA. Thus, p38 represents most of the 1.95 kb RNA which hybridized with D812. At this time sequence analysis was begun on both the p16 and p38 clones. The sequencing strategy is shown in Figure A10; this was simplified by the availability of primers homologous to the SP6 and T7 promotors which flank the polylinker in the pGEM4 plasmid. Therefore, dideoxy sequencing could be carried out directly on the double-stranded plasmid DNA. To facilitate sequencing of the internal portion of p38, two subclones were constructed in pGEM4, utilizing the £191 site in the middle of the insert. The six areas sequenced are shown in Figure All(a-f), with an arrow indicating the priner, strand, and direction of sequencing . These sequences were compared with the Genbank data files, using the Bionet system program Align. The most significant correlation found is shown in Figure A12; it aligns a region of p16 with the human 00 chain transmembrane region. This introduces one gap of two nucleotides and matches 29 of 34 nucleotides, for approximately 85% homology over this region. This homology would be adequate to account for the 44 .omwu mo3 cocoovwm may noes: mom mwpfiuowaooc mo wanes: onu noes meoHo .mzouuo on» an owumoflpcfi m“ amoumuuu mofiocmovwm .mwcoHoQom Homewnaoowu ecu mo ans cofiuofiuumwu Hofiuumm "mam muomfim _ fiO—h . . MmMerlllllllllllllllllflmww mtafiflu" hm?— AflWWlIIIIIIIIIIIIIIIIIIIIIIIIIIIMVI muflxwmxn mm bk . amu— IloIJm AflumVIIIIIIIIIIIIIIIIIIIhnVIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIMV mzuu Em C mop OMN fl oum‘ s" x 0mm t. 45 Figure All (a-f): Partial sequences of p16 and p38 deduced using the dideoxynucleotide method. Sequencing strategy and location of promotor primer regions are shown in Figure A10. Arrows show direction and origin of sequencing reactions. The p38 internal Sagl site and the polyadenylation signal are underlined. 46 CG CG AT GC CG GC AT CG AT CG CG CG GC CG GC AT CG AT CG CG CG CG GC CG GC GC TA TA AT CG CG CG AT AT CG GC AT TA AT CG CG CG CG GC AT GC AT CG AT CG CG TA CG AT CG TA AT TA AT CG CG TA TA GC AT GC GC TA AT CG CG CG CG GC AT CG CG AT CG AT CG CG CG GC CG CG AT TA AT CG CG CG CG GC CG CG AT GC AT CG CG CG GC AT GC TA. TA AT CG CG CG GC GC CG GC AT TA AT CG CG CG CG GC CG CG AT GC AT CG AT CG CG CG AT AT GC TA AT CG CG CG CG GC TA GC AT CG AT CG CG AT CG CG CG GC TA GC TA TA AT AT CG AT CG GC AT TA AT CG CG CG TA AT CG TA AT CG AT CG GC AT GC TA AT CG CG CG GC CG CG AT CG AT CG CG GC CG GC AT TA Figure All(a):p16—T7 sequence 47 CG GC AT TA AT CG CG CG GC TA GC TA AT CG AT CG CG AT GC TA AT AT CG CG 44 53 CG CG GC CG GC AT TA AT CG CG CG CG GC CG CG CG CG AT CG AT CG CG CG CG GC CCACT GAGCC CCACA GTGCC CCACT GGTGA CTCGG GGTGT CACGG GGTGA TA CG GC AT TA AT CG CG CG CG AT CG CG AT GC AT CG AT CG CG CG GC AT AT GC AT CG CG CG CG GC CG GC AT TA AT CG CG CG CG GC TA GC TA TA AT CG AT CG CG GC CG GC GC GC GC CG CG GC AT AT AT AT TA CG CG CG CG CG TA TA AT TA TA TA Figure All(b): p16-SP6 sequence 48 AT GC AT GC TA CG AT CG CG CG AT GC TA GC AT CG AT CG CG CG CG GC CG GC AT AT CG CG CG CG GC AT GC GC CG TA AT CG CG CG AT GC CG GC AT TA AT CG CG CG p38-T7 sequence Figure All(c): 49 5'-A G C C T A T G A G C C A C A N A G T 3'-T C G G A T A C T C G G T G T N T C A GC AT GC GC TA GC AT CG CG CG AT GC CG GC AT TA AT CG CG CG AT GC CG GC AT C C G A C A C A G C G A C A C A T G G A G G C T G T G T C G C T G T G T A C C T A C T T G A G C C C C A T A G C G A C C C A C T G T G A A C T C G G G G T A T C G C T G G G T G A C CG CG CG GC AT GC GC TA AT CG CG CG AT GC CG GC AT TA AT CG CG CG CG GC AT C A T T G A G C C C C A C G G A G C T C-3' G T A A C T C G G G G T G C C T C G A G-5' p38/25-SP6 sequence Figure All(d): 50 5'-GAGCT CACAC CACGC CCCAT 3'-CTCGA GTGTG GTGCG GGGTA CG AT CG CG CG AT GC CG GC AT CG AT CG CG CG TA GC AT GC GC CG AT CG CG CG ACCAA CGCGA CCCAT GGAGC CCCAT TGGTT GCGCT GGGTACCTCG GGGTA CTGAG CCCCA CAGAA GTGCA CCCAT GACTCGGGGT GTCTTCACGT GGGTA ACCTA CCCAT GGA TGGATGGGTA CCT CCCAC TATAG AGCAC GGGTGATATCTCGTG AT GC TA CG CG AT AT CG CG CG GC AT GC GC TA GC AT CG CG CG AT CG CG GC AT p38/8-SP6 sequence GCCCA TAGCA GCCCA TT CGGGT ATCGT CGGGT AA Figure All(e): 51 5'-C A C A G T G C C C C A C T G A A G C C 3'-G T G T C A C G G G G T G A C T T C G G CG CG CG CG CG TA GC TA AT CG AT CG CG TA CG CG AT TA AT CG CG CG CG AT CG TA TA TA CG GC GC CG CG CG CG GC CG CG AT TA AT CG CG CG CG GC CG AT TA AT p38-SP6 sequence Figure All(f): 52 pDSlZ--GAGCTCACAGAGACTGTGGTCTGTGCIZCTGGGG'I'I‘GTCTGTGGGCCTC * i i i * * * ** ** *** t * *** *** *** ** p16SP6-GGGCGCTATGwGCACAATGTGGAGCTATGCK9GTGGTCTGTGGGC'I‘TC GTGGGCATTGTGGTGGGGACCGTCTTGATCATCCGAGGCCI‘G “A * *** * * *** * * t * t * * ** ** AGTGECXSCACTGTGGCISCTCAGTGEGGCGCTATGGCISCC-IXSGTG Figure A12: Region of homology between p16 and pDSlZ. The nucleotide sequence of the pDSlZ iagI-St_uI fragment is taken from Chang et al, 1983. The p16 sequence is aligned according to the best fit found by the BioNet Align program (between the open arrows.) The restriction sites are underlined; asterisks represent identical nucleotides; " represent insertions to maximize the homology. 53 hybridization seen between these two cDNAs, especially since 65% of the matching nucleotides are Gs and Cs. Further significant regions were not found by comparing these sequences with MHC genes, or with any sequences present in the Genbank files. An interesting aspect of the sequence data, however is obvious from inspection of the sequences determined. Nearly all regions sequenced are quite conserved repeats of a 10bp 'consensus' sequence, shown in Figure A13 along with the percentage of repeats having that nucleotide in that position. Figure A14 (a-e) shows all of the sequenced regions aligned to show the repeat homology. It is evident that there are deviations fran the consensus, including variations in the number of nucleotides as well as their arrangements. However, the origin of this DNA as duplications of a 10bp repeat seems obvious. In order to determine whether these sequenced portions could code for an unidentified protein, all forward reading frames were translated, as shown in Figure A15. There is no open reading frame in any region sequenced . A final blotting experiment was carried out to determine the genomic structure of the DNA transcribed into the RNA represented by p38. Figure A16 shows the autoradiograms of B—congenic DNAs cut with 132931 (a) or WI (b). A large number of bands was seen in each blot, indicating multiple DNA sequences in the genome. 54 AC/TA GA/CG c c C C 90% C-38% 54% 81% A-36% 76% 66% 88% 91% 96% T—44% C-48% FIGURE A13: Consensus sequence of the 10bp repeats found in p38 and p16 cDNAs. Percentage of all repeats which contain consensus base at the given location is indicated on the second line. 55 Figure A14 (a-f): Sequences that were shown in Figure All, aligned to maximize their identity with the repeat consensus sequence shown in Figure A13. 56 C A C T A G C G C C C C A C T G C G C C C C A.C T (3(Z(3 C3C3C C A C A G C G C C C A C A G C G A C C C A C A G A G C C C C A T A G C A A C C C A T T G A G C C C C A T G G A G T T C C A T A T C A C T C C A G A G C G C C C C A T A G C G C C C A C A C C C C G C C C C A T A G C G G C C C A T T G A G C C C C A C A G T G C C C C A T G A A C C C A C A G A T C A T C C C A T A G C A C A A T T G T G C C C A T A G C G C C A C A C C G C C C A T G A G C A C A T G G C C Figure Al4(a): p16-T7 sequence 57 C C C A A T G A C C A C A T G T G C C C A T A G C G C C C C A C A C C C C G C C C C A T A G C G C C C C A C T G A G C C C C A C A G T G C C C C A C T G A A G C C C A C A GIA C C A C C C C A T A G C T C C A C A T T G T G C C C C A T A G C G C C C C A T T T A T T G C C C C T A A A A G C C G G G G C G C T A T G G C G G Figure Al4b: p16-SP6 sequence 58 GTA GCG ACCC ATG GAG CCCC AC ACA GCG CCCC ACA GTGACCC ACT GAG ACCC ATA GCGACCC ATCGGAG CCCC ACT GAG CCCC ACA GCG ACCC Figure A14 (0): p38-SPG sequence 59 AGCC TAT GAG CCAC ANA GTG CGC AGA GCG ACCC ATA GCGACCC AGT GGA GCCG ACA CAG CGAC ACA TG GAG CCCC ACTT GAG CCCC ATA GCG ACCC ACT GAG CCCC ATA GCG ACCC ATG GAG CCCC ATT GAG CCCC ACG GAG CTC Figure A14 (d): p38/25-SP6 sequence ACA ATA ACG ACA ACA ACT AT :8 5 1'3 O G) n 3’ Figure Al4(e): 60 GA GA 0 C) 6) 3’ GC CCT GAG GCA GAG NN z GC GA GAA GT GC CTG GCA GAG CG CA CA ease CCCC ACCC TCCC ACCC ACCC CCCC GCCC CCCC CACC ACCC CCCC CCCC CCC ACCC CCC GCCC GCCC CCC p38/8-SP6 sequence GGC T 61 C GTG CCCC GAA GCCA CCA CCCC GCT CCAC TG CCCC CG CCCC GCG CCCC TAG GGGC AAA TGGG G GCA15 Figure Al4(f): A C CGG TCCC GTG CATC p38-SP6 sequence 62 com Hm> use one now and cum haw now was cum ohm use mam was now ma: was now we: cum 9H0 oHH cum and . cum mac one mam cum and . com com mmU cHU cum mam . mam was mad oHH one one cum use cum mac mHH com 000 009 oeu euo 000 ewe aeu 000 woo cam 990 090 090 com 000 woo can 090 mma mma cum mam mum now was was com Ho> haw mg: and saw can cum cam . cum mH< cam now cam and mom cam new now one cum mac . ma: use cam mHH cum now and saw cum use wHH m“: mom 5H0 am: cum mom and cum one now mam cum saw 000 Gem euo Bum use emu Use wee mac QQU 00¢ use moo Odd awe emu 000 Gem mad am one ma: one nae use can cad use one mu< one one o~< who m“: can odd sou cHo cum one How new mac now we: cum mam use cum one own cum cam now use one can one cao cum mum CH0 one mad mud n“: cum mam use cum one . ma: mum UGO coo oeu e00 000 Geo moo 000 009 ocu 900 000 90¢ 000 090 was oeu am pm coaumamcMHL asumam “lav nae magmas .mwfiufiomflnen mcfiocmovom mowuomaooc cu moo nonwauoump on uoccoo cofiumamcmuu on» muons moons mumowocfi mxcoam .mcoooo noun wwwcmflm moon .chMOmu poocmovom may no some new weapon mowooou mouse m“ newumHmcmus ”Animv mam muomwm 63 use emu see em: we: ees use use eeu em> mes mes eds euz mes ees wee Hm> em> we: mes use use see ees mes mes sea eeu use use use we: mes mes use ees em: wee mes use em> mes mes use use see u use eau ess ese use uss sea ess uss use ues uus uss use «es uss mme ees use ces mes ees mes we: eds em> emu sea use use mes eds use use seu ees em> uwu eeu one use see mes emu use ees em> emu see ees em: mes eeu ees emu ass ees em> use mes see ces sou mes mes em> em> ees use emu sue uus esa sec uss ses e64 uus use sac ees uus ese ses ses uus eue usu use see em Hm> one we: was em: was mew mes who cum mew one one see oem see now we: mes :es mes was use use mes em: . ceu mes . was ees one use . one em: one ees see ass sea eeu emu see was mes use sou use one was wee mes sea use uss use mes uos uss emu sea ass uus see ees uss uus sou sea uus em e~ coeuoamcmuu oQMImHa ”Any mam mesmem 64 and How we: can new use one new mum me: one oH< oem mus sew sum was . sex com new mom cum awe How one see one one emu use one see mes wee cum mus . u use sss use sou use ses ass sos ass sea use «so use as see sum was one cum and now me: one was use see cum 9H0 em: sum was . we: use new use one . new one see now me: one see has me: use was Hm> cum amc . we: use Hm> use can me< new use one now Que cum and now one ego use use oeu 00¢ 090 eu< 990 000 Geo sue 000 00m woe e00 0&0 00¢ 900 am eN coeumemamue eeImms "luv mes messes 65 uww uom mum cum cam . we: cum 5H0 em: cum mum own new use oum umm sou cum mam has men use mam mum ees we: ous ses see cum umu sue cue sum usu 0 e00 moo omo ooo omo eem ooo oom woe moo omo oom Nee mme . cum cam . we: use mam see cum uom owe me: sum 9H0 em: we: mum new see cum umm and cum mum uwm we: cum new use cum cam ado we: use cam use we: cum new use cum mum . cum cam . use cum uwm Que use mum uow we: EEO QQO 0mm 90¢ DUO mwO Umfi £00 000 mwe 90¢ QQU 00¢ 609 com Umw 00¢ UEU use us use cum new Hm> cum mum uom we: une cam mum mum Hm> we: uww em: cum sum mum euu uuu ous sum . oue sum umu nus mum umu cue sus use sum mum umu sue nus use mum sue oue mum nus mam mes use mum . ass uwu mus sum use sms sum sus mem uus mus sms mus use smz mum uus use mes usm em em useumemcouu I coeuuoa =m .socmoosm mma Hocuwuce ”on mam mesmem 66 :Ho wee cum cHo . me: mam mam sou cum uom uoo 0m 0ee mUO 00¢ 004 Hmm . cum uwo sou cum cum oHo Ho> cum mum uom we: use who seq omo cum cam see cum oeo cum ceo cam eHo uom oua mum . cum we: Ho> oeo use cum use sex cam . use cum uom Que cHo use cam wee cum mam How mum me: see oHo emo ooo moe oom moo oom ooe omo oom ooo mem ooo moo eom mom omo ooo omo mNm 5mm . see cum uom Que cum mum mum use we: cum cam mum nee we: use we: cum 9H0 em: cum mum use we: cum use uom . ooo cum uom cum cam eeo sea use mam cum use oum we: ado wee use one eoe moo ooo moo emo oom ooo omz moo moz zzz mom ooo omo omo mem eom ooo new mom mam mam cum cum uom Que Cum uee use we: use cam use cum uow mum cum mum uom who ceo cum oHo em: sum one sex oum mum uoo we: uom seo use cum mum cao em: cum mam eeo we: use cum use sum mum cao cum Hm> ado we: use oHo omo oem moo ooo moo emo oom eoo mom ooo moo omo moo oeo moo omo oom omo mam awe mum men sum cum use uow uom . cum mam use we: use oHo wee cum mum sex cum mam mem ooo ooo zmo omo moo eoo mo Noe coeuuoa .m .oocwooom mma accumuCe "Asa mam seamen 67 who who who who who mum who who who sea :Ho umm men men men men cam wee mm mmm mmm mmm mmm moo oem mme we: Ho> mum cum eeo who ado cum wee cam Mao zoo zoo mum cum cam wee cum cam mum mum uom mum Ho> Que sea cum eeo mum one mam cum mam uwm we: cum mes use oue em> mum mes em: . nus eus . use eus ous mum . oue «em ooe oom ooo oeo ooo eoo oem mme mmo ooo ome eee ooo ooo ooo ome moo ooo use as sum wee cue mes es: we: ass . sum use «em use mam ues use oue mes nes use we: sum Hm> we: cum cam see cum we: uom we: uom . uem cum Hm> use uee oum mam who use ooo uom new oua oua :Ho oum men own cum mam uwm we: ome moo ooo eoe mom ooe oom emo ooo moo omo moo omm oeo moo ooo eom omo mm em couumumcmuu ueuummu "luv mam museum 68 Figure A16: Genomic Southern blots of four different haplotype DNAs probed with nick translated p38. Panel A shows 13le digestion; panel 8 shows 8amHI digestion. 69 DISCUSSION Once the decision was made to begin cloning genes of the MHC of the chicken, a number of possible approaches were evaluated. 1) Use of synthetic oligonucleotide probes was not possible due to lack of protein amino acid sequence data. In addition, the difficulty in obtaining adequate quantities of purified protein and lack of expertise in this area precluded an attempt on our part to obtain this sequence data prior to the cloning attempts. 2) Several approaches require the use of highly specific antibodies to screen expression libraries or to recognize in vitro translation products from hybridization-selected mRNAs. Although we have access to sone excellent serologic typing reagents (Bacon et al, 1987b), these allo—antisera recognize the polymorphic determinants on the native cell-surface molecules which are glycosylated hetero-dimers. The antisera titers against the individual polypeptide chains produced in bacteria or in vitro is too low to be very useful for cloning purposes. 3) Since the structure and function of the chicken MHC proteins appears highly homologous to the manlnalian proteins, it is possible that low stringency hybridization using the mamnalian genes would be able to identify the homologous chicken genes. This approach has proven successful in cloning human Class I genes (using mouse Class I cDNA) and in cloning mouse Class II genes (using human Class II cDNA) . Although it is very likely that chicken MHC genes would be more divergent from either mouse or human than the two manmalian species would be from each other, several lines of evidence led us to believe this approach might work. First, highly conserved genes, such as the precursors of oncogenes, had been detected in a 70 variety of species using low stringency hybridization conditions. Second, the degree of similarity between conserved MHC gene segments in mouse and human was as high as 68 - 76% (Auffray et al, 1984,) leading us to believe that even the greater divergence of the avian species might still allow for low stringency hybridization. Third, other chicken genes which had been cloned had diverged from their mamnalian counterparts to a small enough degree that low stringency detection between them might be possible. For example, the chicken lambda light chain inmunoglobulin gene had been cloned and was 73% and 54% identical to the human and mouse homologs, while the identity between mouse and human was only 41-51% (Reynaud et al, 1983.) This encouraged us to believe that chicken MHC genes might be picked up at low stringency using a manmalian homologous probe. Consequently, we obtained several manmalian cDNA clones and tested their hybridization at low stringency to restriction enzyme-digested chicken genomic DNA (Figure A3). A single major band appeared in each lane, with sane variations between haplotypes in the minor bands. However, we were concerned about the possibility of not being able to identify the location of less conserved or more polymorphic exons, even if we were successful in cloning the most conserved exon. We therefore felt that obtaining a cDNA clone would not only enhance our possibility for successful cloning, but also would yield additional information concerning the divergence of less conserved regions. Since cDNA clones would need to be obtained from cells which were actively synthesizing the protein and we wanted a reproducible source 71 for the mRNA, we chose to look at RNA from RP-9 cells. RP-9 is a chicken B—lynphoblastoid cell line cultured from a transplantable tumor 1 x 72 chicken infected with RAV-Z. It has been shown to be positive for cell surface "Ia," using a specific anti- which originated in a 15 chicken monoclonal antibody (C-L. Chen, personal conmunication.) Since this is the equivalent of the human Class II MHC proteins, we examined the ability of the DS‘12 cDNA clone to hybridize with RNA from RP-9 cells. As shown in Figure A4, a 1.95 kb RNA was detected under low stringency conditions. We therefore proceeded with construction of an RP-9 cDNA library in lambda gtl0. Initially, 200,000 lambda gtl0 reconbinant clones were screened with both DS‘12 and pRAV-2 as probes. At high stringency, approximately 0.14% of the clones hybridized with the pRAV-2 probe, which was within the expected range, since these cells release infectious virus and therefore are producing messages for the structural proteins of the virus. Using very low stringency conditions for the DS‘12 probe, twenty plaques were considered to be positive on duplicate filters and were picked for further plaque purification. 75% of these were confirmed to be positive during the secondary screenings. However, it is estimated that those plaques picked after the low stringency primary screening represented only 10-20% of the similarly positive plaques. This low percentage arose because of the extremely low stringency used in screening the plaques; many regions of the filters contained sufficient background to prevent ability to discern positive plaques on the filter. Only those plaques detected on duplicate filters were picked for analysis. If the estimate of 10-20% is correct, then there were 72 actually 100-200 positives out of the original 200 ,000 plaques screened, or 0.05-0.l% of the mRNA is represented in the DS‘lZ-positive fraction. This appears to be similar to the 0.3% abundance found in a human lymphocyte cDNA library by Erlich et a1 (1984) . Secondary screening and plaque purification yielded fifteen clones which remained positive throughout the procedure. Phage DNA was isolated from these clones, the E_CORI fragments were transferred to nitrocellulose, and the filters were hybridized with the D8‘12 probe. Thirteen of the fifteen clones exhibited ethidium branide-detectable inserts ranging in size from 175 to 1500 bp. Nine of the inserts were detectable on autoradiograms after the hybridization with DS‘12 probe, as shown in Figure A7a. Subcloning of two of these inserts (10A and 18A) into the plasmid vector pGEM4 gave the recombinants designated p38 and p16, which allowed further mapping, sequencing, and hybridization analyses. After 32P-labelled probe from the filter shown in Figure A7a, removing the this filter was rehybridized with the nick-translated insert fran p38. The results shown in Figure A7b indicate that all of the gtl0 inserts which could be detected by D812 at low stringency could also be detected by p38 at high stringency, indicating that this set of clones consists largely (if not exclusively) of various cDNA copies of mRNA transcribed from a single gene or members of a closely related gene family. Since the initial subcloning attempts yielded a plasmid containing the largest of these closely related inserts, further subcloning did not appear to be necessary at this tine. All further studies described here were conducted using the p38 and p16 subclones. 73 In order to determine which portion of the human cDNA was responsible for the hybridization to these clones, D8‘12 was restricted with various enzyttes and the nitrocellulose containing the fragments was hybridized with nick-translated p38. As shown in Figure A8b, all fragments which hybridized contain the small §a£I--§_t_u1 fragment which spans the transmembrane portion of the D8 alpha polypeptide. Hybridization using the p16 insert or either §c_c_>RI--§agl fragment of p38 as the probe showed identical hybridization patterns to the pattern shown in Figure A8b (data not shown) . When Northern blots of RP-9 RNA hybridized with the nick-translated p38 clone were conpared with an identical blot hybridized with the DS‘12 probe, it was determined that both probes identified a band of identical size, although the D812 band could only be seen under conditions of low stringency. Also, p38 picked up several less abundant bands of higher molecular weight, which may be partially processed precursors of the 1.95 kb band present in the total cellular RNA. Thus, the 1500 bases of p38 represent the major portion of the only RNA sequence which hybridizes to D812 . Sequence analysis was begun on both p38 and p16, according to the strategy shown in Figure A10, and the deduced sequences are shown in Figure All(a-f). Analysis of these sequences and their conparison with other published sequences leads to several conclusions. First, a 35 bp portion of p16 is >80% identical with the transmanbrane portion of D8‘12 (Figure A12.) This probably is adequate to account for the hybridization seen at low stringency. Second, p38 contains a 74 polyadenylation signal (AATAAA) 25 bp upstream of the 15 bp poly (A) stretch at one end of the clone. Assuming that no recombination events have occurred during cloning, this identifies the ”coding" strand for the entire length of the clone. Third, all areas sequenced appear to consist primarily of highly conserved repeats of a 10 bp sequence Figures A13-A14.) Assmning that this repeating structure is the same on both clones, the portion of the p16 sequence which is homologous to the DS‘12 transmembrane coding sequence is found on the p16 "non-coding” strand, indicating that it is insignificant to any protein structure which may be coded by this RP-9 RNA. Fourth, no long open reading frames have been located on the coding strand within those portions sequenced (Figure A15,) indicating that if a protein is coded for by this mRNA, it is not as large as an MC Class II peptide chain appears to be. Fifth, no significant homology exists between the sequenced portions of this cDNA and D812 or any other published manmalian MHC sequence, aside fran the previously-mentioned transmembrane portion. This leads to the question of the source of this unusual, repeat- containing, poly A+ RNA. Our first attanpt to answer this was to look at its hybridization to genomic blots of normal chickens. As shown in Figure A16, when used as a probe, the p38 plasmid detected a large number of genomic fragnents using several enzymes, such as B_a_mH1 or H_i_r_i_dIII, indicating that it represents a repetitive sequence present many tines in the genone. However, it was noticed that E_coRI digestion yielded many fewer bands, perhaps indicating that many of the repeats are clustered whithin these large EcoRI fragnents. Thus, p38 apparently 75 represents a short, moderately repetitive sequence which is transcribed in RP-9 cells into a specific, large, polyadenylated RNA. We have considered the possibility that p38 represents a cloning artifact, and that it may be misleading to base conclusions on the structure and sequence of p38. However, the size of p38 is reasonable to have cane from the 1.95 kb RNA detected in Northern blots at low stringency with D812 and at high stringency with p38. Although the polyadenylation signal lies within the only sequence which does not clearly fit the repeat structure, it probably represents the true 3' end of the RNA; since the cDNA was made fran oligo dT cellulose-selected RNA, and was made using an oligo dT priner, the original cDNA most likely was polyadenylated, in order to be represented multiple times in the l ibrary. If this is truely a polyadenylated RNA transcribed from repetitive INA, why is it present in RP-9 cells? Transcriptional activation due to promotor insertion is known to occur as a result of retroviral integration. We do not know whether this has occurred in this cell line, but it could be tested by cloning the junction fragments of the inserted provirus(es) and looking for adjacent repeats of the type found in p38. It is also possible that the transcription of this RNA is cannon to sure or all transforned cells. Numerous reports have docuttented the presence of highly repetitive sequences in poly A+ RNA of transforned cells or during embryonic developnent (Scott et al, 1983; Murphy et al, 1983, Kramerov et al, 1982; Miyahara et a1, 1985; Yamamoto et al, 1983,) although the presence of repetitive elements is 76 frequently restricted to the 3'-untranslated portion, where it may affect mRNA stability (Croce, 1987.) Of course, it is possible that the 5'-400 nucleotides which are missing from this clone may contain coding sequences and that the entire 1500 bp of the cDNA is contained within the 3'-untranslated region of the 1950 bp mRNA. However, regardless of whether there are 150 or 190 copies of the repeats, the reason for this very large copy number of the repeat in this transcript is unknown. Most of the very short, tandem repeats in highly repetitive portions of DNA have been shown to be present at telonere and centroneres of chromosanes (Jelenik and Schmid, 1982) where their function is unknown. In the newt, one such repetitive sequence is also present interspersed among histone genes, and is transcribed in lampbrush chromosates when the polymerase fails to recognize the correct termination signal for the histone gene, and continues transcription through the adjacent repeat sequences (Diaz et al, 1981.) Now that we know the 1.95 kb message does not code for an MHC Class II alpha chain, what conclusions can be reached about the chicken MHC fran this study? This hybridization strategy should have picked up a D0 alpha homolog expressed in 8 cells of the chicken, if that gene maintained at least 65-70% homology over a substantial portion of the region coding for a single danain. This estimate is based upon the actual G-C content of pDSlZ, and wash stringency conditions of 2xSSC at 56C, which allows at least 30% mismatch in hybridization. The actual homology detectable in the Northern blots is lower than this, due to increased stability of RNA-DNA hybrids. 77 This means that either the D0 alpha homolog lacks adequate similarity to be detected by hybridization, or that there is no DQ alpha hanolog expressed in the chicken. The latter explanation would fit with the data of Guillemot et al (1986) , who found only a single, non-polymorphic alpha chain expressed on splenic lymphocytes of several B-haplotypes and on RP-9 cells. Thus, the single alpha chain expressed on these cells is likely to be the homolog of the human DR alpha or the mouse E alpha, which are both non-polymorphic. The data described here lead to the following conclusions: 1) there may not be a DO alpha homolog in the chicken, recognizing the possibility of fewer genetic loci and expressed proteins means less complexity in the chicken Class II MHC than the mattrnalian Ia regions, and 2) if there is a D0 alpha homolog in the chicken, it may not be expressed (Guillemot et al, 1986), and certainly diverges sufficiently to interfere with the success of hybridization analyses such as that described here.. 7 8 REFERENCES Arden, 8. and Klein, J. (1982) Biochemical conparison of major histocanpatibility complex molecules from different subspecies of Mus musculis: Evidence for trans-specific evolution of alleles. Proc. Natl. Acad. Sci. USA _7_9_:2342-2346. Auffray, C., Korman, A.J., Roux-Dosseto, M., Bono, R., and Strominger, J.L. (1982) cDNA clone for the heavy chain of the human B cell alloantigen DCl: Strong sequence homology to the HLA-DR heavy chain. Proc. Natl. Acad. Sci. USA 3:6337-6341. Auffray, C., Lillie, J.W., Arnot, D., Grossberger, D., Kappes, D., and Strominger, J.L. (1984) Isotypic and allotypic variation of human class II histocompatibility antigen alpha-chain genes. Nature 308:327-333. Bacon, L.D. (1987b) Influence of the MHC on disease resistance and productivity. Poultry Sci., in press. Bacon, L5D., Crifgenden, L.B., Witter, R.L., Fadly, A., and Motta, J. (1983) _B_ and _8 Associated with Progressive Marek's Disease, Rous Sarcoma, and Avian Leukosis Virus-Induced Tumors in Inbred lSI4 Chickens. Poultry Sci. 2:573-578. Bacon, L.D., Ismail, N., and Motta, J. (1987) Allograft and antibody responses of 151 -_l_3_-congenic chickens. In: Avian Inmunology II. Webber, W.T., and 5Ewert, D., editors, Alan R. Liss and Co., New York, 219-233. Bacon, L.D., Kite, J.H., and Rose, N.R. (1974) Relation between the major histocatpatibility (B) locus and autointnune thyroiditis in Obese chickens. Science 186:274. Bacon, L.D., Sundick, R.S., and Rose, N.R. (1977) Genetic and cellular control of spontaneous autoitrmune thyroiditis in OS chickens. Adv. Exp. Med. Biol. @5309. Bacon, L.D., Witter, R.L., Crittenden, L.B., Fadly, A., and Motta, J. (1981) _B_-Haplotype Influence on Marek's Disease, Rous Sarcana, and Lymphoid Leukosis Virus-Induced Tumors in Chickens. Poultry Sci. 60:1132-1139. Bell, J.I., Estess, P., St. John, T., Saiki, R., Watling, D.L., Erlich, H.A., and McDevitt, H.O. (1985) DNA sequence and characterization of human class II major histocompatibility complex beta chains from the E haplotype. Proc. Natl. Acad. Sci. USA 82:3405-3409. Benoist, C.O., Mathis, D.J., Kanter, M.R., Williams, V.E., and McDevitt, H.O. (1983) The murine Ia alpha chains, E alpha and A alpha, show a surprising degree of sequence homology. Proc. Natl. Acad. Sci. USA 82:534-538. Benton, W.D. and Davis, R.W. (1977) Screening lambda gt reconbinant clones by hybridization to single plaques in situ. Science 196:180-182. 79 Bieberich, C. , Scangos, G., Tanaka, K. , and Jay, G- (1986) Regulated expression of a murine Class I gene in transgenic mice. Mol. Cell. Biol. _6_:l339-1342. Birktteyer, R.C. and Nordskog, A.W. (1982) Proc. 18th Intl Conf. Anim. Blood Groups Biochan. Polymorphisms. p. 71. (Cited by Toivanen and Toivanen, 1983.) Bisati, S. and Brogren, C.H. (1980) Identification and inmunochemical characterization of chicken MHC alloantigens in lymphocytes and serum. Scand. J. Inmunol. l_2:532. Bjorkman, P.J., Strominger, J.L., and Wiley, D.C. (1985) Crystallization and x-ray diffraction studies on the histocompatibility antigens HLA-AZ and HLA-A28 from human cell menbranes. J. Mol. Biol. 18_6_:205-210. Bloom, 8.8. and Bacon, L.D. (1985) Linkage of the Major Histocatpatibility (_B_) Canplex and the Nucleolar Organizer in the Chicken: Assignnent to a Microchromosone. J. Hered. £:146-154. Bregegere, F., Abastado, J.P., Kvist, 8., Rask, L., Lalanne, J.L., Garoff, H., Cami, 8., Wiman, K., Larhanmar, D., Peterson, P.A., Gachelin, G., Kourilsky, P., and Dobberstein, 8. (1981) Structure of C- terminal half of two H-2 antigens from cloned mRNA. Nature 3_9_Z_:78-81. Briles, W.E. and 8ri1es, R.W. (1982) Identification of Haplotypes of the Chicken Major Histocompatibility Complex (8). Inmunogenetics 15:449-459. Briles, W.E., Briles, R.W., Taffs, R.E. and Stone, H.A. (1983) Resistance to Malignant Lymphana in Chickens Is Mapped to a Subregion of Major Histocompatibility (B) Complex. Science 219:977-979. Briles, W.E., Bumstead, N., Ewert, D.L., Gilmour, D.G., Gogusev, J., Hala, K., Koch, C., Longenecker, 8.M., Nordskog, A.W., Pink, J.R.L., Schierman, L.W., Simonsen, M., Toivanen, A., Toivanen, P., Vainio, 0., and Wick, G. (1982) Nanenclature for Chicken Major Histocatpatibility (B) Conplex. Inmunogenetics gum-447. Briles, W.E., McGibbon, W.H., and Irwin, M.R. (1948) Studies of the time of development of cellular antigens in the chicken. Genetics 33:97. Briles, W.E., McGibbon, W.H., and Irwin, M.R. (1950) On multiple alleles effecting cellular antigens in the chicken. Genetics 3_S_:633- 652. Briles, W.E., Stone, H.A., and Cole, R.K. (1977) Marek's Disease: Effects of 8 Histocatpatibility Alloalleles in Resistant and Susceptible Chicken Lines. Science 195:193—195. Chang, H-C., Moriuchi, T., and Silver, J. (1983) The heavy chain of human B cell alloantigen HLA-DS has a variable N-terminal region and a constant inlnunoglobulin-like region. Nature 305:813-815. Clare, R.A., Strout, R.G., Taylor, R.L., Collins, W.M., and 8ri1es, W.E. (1985) Major Histocompatibility (g) Complex Effects on Acquired Innunity to Cecal Coccidiosis. Ittmunogenetics 23:593-599. Clenens, M.J. (1987) A potential role for RNA transcribed fran 82 repeats in the regulation of mRNA stability. Cell 49:157-158. Coligan, J.E., Kindt, T.J., Uehara, H., Martinko, J., and Nathenson, S.G. (1981) Primary structure of a murine transplantation antigen. Nature 291:35-39. Crittenden, L.B. (1986) Identification and cloning of genes for insertion. Poultry Sci. 65:1468-1473. Crone, M., Jensenius, J.Chr., and Koch, C. (1981a) B-L Antigens (Ia- like Antigens) of the Chicken Major Histocatpatibility Canplex. Scand. J. Inmunol. E:591-597. Crone, M., Jensenius, J. and Koch, C. (1981b) Evidence for Two Populations of B—L (Ia-Like) Molecules Encoded by the Chicken MHC. Inmunogenetics l_3:381-39l. Das, H.K., Biro, P.A., Cohen, 8.N., Erlich, H.A., von Gabain, A., Lawrance, S.K., Lanaux, P.G., McDevitt, H.O., Peterlin, B.M., Schulz, M- F., Sood, A.K., and Weissman, S.M. (1983a) Use of synthetic oligonucleotide probes canplementary to genes for human HLA-DR alpha and beta as extenson priners for the isolation of 5'-specific genomic clones. Proc. Natl. Acad. Sci. USA 80:1531—1535. Das, H.R., Lawrance, S.K., and Weissman, S.M. (1983b) Structure and nucleotide sequence of the heavy chain gene of HLA-DR. Proc. Natl. Davis, M.M., Cohen, D.I., Nielsen, E.A., Steinnetz, M., Paul, W.E., and Hood, L. (1984) Cell-type-specific cD§A probes and the murine I region: The localization and orientation of A alpha. Proc. Natl. Acad. Sci. USA 81:2194—2198. Diaz, M.O., Barsacchi-Pilone, G., Mahon, R.A., and Gall, J.G. (1981) Transcripts from both strands of a satellite DNA occur on lampbrush chromosone loops of the newt Notophthalmus. Cell 21:649-659. Erlich, H., Stetler, D., Sheng-Dong, R., and Saiki, R. (1984) Analysis by molecular cloning of the human class II genes. Federation Proceedings £:3025—3030. Ewert, D.L. and Cooper, M.D. (1978) Ia—Like Alloantigens in the Chicken: Serologic Characterization and Ontogeny of Cellular Expression. Inlnunogenetics 1:521-535. 81 Ewert, D.L., Munchus, M.S., Chen, C-L., and Cooper, M.D. (1984) Analysis of Structural Properties and Cellular Distribution of Avian Ia Antigen by Using Monoclonal Antibody to Monomorphic Determinants. J. Inmunol. 133325244530. Ezquerra, A., Bragado, R., Vega, H.A., Strominger, J.L., Woody, J., and Lopez de Castro, J.A. (1985) Primary structure of papain-solubilized human histocompatibility antigen HLA-827. Biochemistry 24:1733-1741. Flavell, R.A., Allen, H., Burkly, L.C., Sherman, D.H., Waneck, G.L., and Widera, G. (1986) Molecular biology of the H-2 histocompatibility canplex. Science 233:437-443. Folsom, V., Gay, D., and Tonegawa, S. (1985) The beta-1 domain of the mouse E beta chain is important for restricted antigen presentation to helper T-cell hybridomas. Proc. Natl. Acad. Sci. USA 8_2_:1678-1682. Gebriel, G.M. and Nordskog, A.W. (1983) Genetic Linkage of Subgroup C Rous Sarcoma Virus-Induced Tumour Expression in Chickens to the Ir-GAT Locus of the _B Complex. J. Inmunogenetics 10:231-235. Gebriel, G.M., Pevzner, I.Y., and Nordskog, A.W. (1979) Genetic Linkage between Inmune Response to CAT and the Fate of RSV-Induced Tumors in Chickens. Innunogenetics 9:327-334. Goodenow, R.S., McMillan, M., Nicolson, M., Sher, B.T., Eakle, K., Davidson, N., and Hood, L. (1982a) Identification of the class I genes of the mouse major histocompatibility canplex by DNA-mediated gene transfer . Nature, 301: 231-237 . Goodenow, R.S., McMillan, M., Orn, A., Nicolson, M., Davidson, N.a Frelinger, J.A., and Hood, L. (1982b) Identification of a BALB/c H-2L gene by DNA-mediated gene transfer. Science 215:677-679. Goodenow, R.S., Stroynowski, 1., McMillan, M., Nicolson, M., Eakle, Sher, B.T., Davidson, N., and Hood, L. (1983) Expression of conplete transplantation antigens by manmalian cells transformed with truncated class I genes. Nature fl:388-394. Gorer, P.A., Lyman, S., and Snell, G.D. (1948) Proc. R. Soc. London Ser. 8. 135:499. Gorski, J., Rollini, P., Long, 8., and Mach, 8. (1984) Molecular organization of the HLA-SB region of the human major histoconpatibility canplex and evidence for two 88 beta-chain genes. Proc. Natl. Acad. Sci. USA _8_l_:3934-3938. Guillemot, F., Turuel, P., Charron, D., Le Douarin, N., and Auffray, C. (1986) Structure, Biosynthesis, and Polymorphism of Chicken MHC Class II (B-L) Antigens and Associated Molecules. J. Inmunol. 137:1251-1257. Hala, K., Boyd, R., and Wick, G. (1981a) Chicken Major Histocompatibility Complex and Disease. Scand. J. Intnunol. _l_4_:607-616. 82 Hala, K., Plachy, J ., and Schulmannova, J. (1981b) Role of theB-G- Region Antigen in the Humoral Inmune Response to the B-F-Region Antigen of Chicken MHC. Inmunogenetics 14:393-401. Hala, K., Wick, G., Boyd, R.L., Wolf, H., Bock, G., and Ewert, D.L. (1984) The B-L (Ia-Like) Antigens of the Chicken. Lymphocyte Plasma Membrane Distribution and Tissue Localization. Develop. Compar. Inlmnol. 8:673—682. Hood, L., Kronenberg, M, and Hunkapiller, t. (1985) T cell antigen receptors and the inmmoglobulin gene family. Cell 40:225. Huser, H., Ziegler, A., Knecht, R., and Pink, J.R.L. (1978) Partial Aminoterminal Amino Acid Sequences of Chicken Major Histocompatibility Antigens. Inmunogenetics _6_:301-307. Jaulin, C., Perrin, A., Abastado, J-P., Dumas, 8., Papamatheakis, J., and Kourilsky, P. (1985) Polymorphism in mouse and human Class I 5:2 and fl genes is not the result of random independent point mutations. Inlnunogenetics 2_2:453-470. Jelinek, N.R. and Schmid, C.H. (1982) Repetitive sequences in eukaryotic DNA and their expression. Ann. Rev. Biochem. 51:813-844. Jordan, B.R., Bregegere, F., and Kourilsky, P. (1981) Human HLA gene segnent isoloated by hybridization with mouse H-2 cDNA probes. Nature 290 :521—523 . Kaufman, J.E., Auffray, C., Korman, A.J., Shackelford, D.A., and Straninger, J. (1984) The Class II molecules of the human and murine major histocompatibility canplex. Oell _3_6_:1-13. Klein, J. (1984) Gene conversion in MHC genes. Transplantation fi:327- 329. Kobari, J.A., Strauss, E., Minard, K., Hood, L. (1986) Molecular analysis of the hotspot of recombination in the murine major histocompatibility complex. Science 234:173-179. Koch, C., Skjodt, K., Toivanen, A., and Toivanen, P. (1983) New Recombinants within the MHC (B—Complex) of the Chicken. Tissue Antigens 21:129-137. Korman, A.J., Knudsen, P.J., Kaufman, J.E., and Strominger, J.L. (1982a) cDNA clones for the heavy chain of HLA-DR antigens obtained after inmunopurification of polysones by monoclonal antibody. Proc. Natl. Acad. Sci. USA 19:1844-1848. Korman, A.J., Auffray, C., Schamboeck, A., and Strominger, J.L. (1982b) The amino acid sequence and gene organization of the heavy chain of the HLA-DR antigen: Homology to inmunoglobulins. Proc. Natl. Acad., Sci. USA _7_9_:6013-60l7. 83 Kramerov, D.A., Lekakh, I.V., and Ryskov, A.P. (1982) The sequences homologous to major interspersed repeats 81 and 82 of mouse genome are present in mRNA and small cytoplasmic poly A+ RNA. Nucleic Acids Res. _l_0:7477-7491. Kubo, R.T., Yamaga, K., and Abplanalp, H.A. (1977) Isolation and partial characterization of the major histocompatibility antigen in the chicken. Adv. Exp. Med. Biol. 8_8_:209. Kvist, 8., Bregegere, F., Rask, L., Cami, 8., Garoff, H., Daniel, F., Wiman, K., Larhamnar, D., Abastado, J.P., Gachelin, G., Peterson, P.A., Dobberstein 8., and Kourilsky, P. (1981) cDNA clone coding for part of a mouse H-2 major histocompatibility antigen. Proc. Natl. Acad. Sci. USA 18: 2772-2776 . Lamont, 8.J., Bolin, C., and Cheville, N. (1987) Genetic resistance to fowl cholera is linked to the major histocatpatibility canplex. Inmunogenetics 25:284-289. Lee, J.S., Trowsdale, J., Travers, P.J., Carey, J., Grosveld, F., Jenkins, J ., and Bodtner, W.E. (1982) Sequence of an HLA-DR alpha-chain cDNA clone and intron-exon organization of the corresponding gene. Nature 29_9_:750-752. Lee, R.W.H. and Nordskog, A.W. (1981) Role of the inmune-response region of the 8 complex in the control of Graft-vs-Host reaction in chickens . Intnunogenetics _l_3_I: 85 . Le Meur, M., Gerlinger, P., Benoist, C., and Mathis, D. (1985) Correcting an inmune—response deficiency by creating E alpha transgenic mice. Nature 316:38-42. Long, E.O., Wake, C.T., Strubin, M., Gross, N., Accolla, R.S., Carrel, 8., and Mach, 8. (1982) Isolation of distinct cDNA clones encoding HLA-DR beta chains by use of an expression assay. Proc. Natl. Acad. Sci. USA 19:7465-7469. Longenecker, B.M. and Mosmann, T.R. (1981) Structure and Properties of the Major Histoconpatibility Complex of the Chicken. Speculations on the Advantages and Evolution of Polymorphism. Inmunogenetics E :1-23. Malissen, 8. (1986) Transfer and expression of MHC genes. Innunology Today 1:106-112. Malissen, 8., Price, M.P., Goverman, J.M., McMillan, M., White, J., Kappler, J., Marrack, P., Pierres, A., Pierres, M., and Hood, L. (1984) Gene transfer of H-2 class 11 genes: Antigen presentation by mouse fibroblast and hamster B—cell lines. Cell _3_6_:319-327. Malissen, M., Damotte, M., Birnbaum, D., Trucy, and Jordan, B.R. (1982a) HLA cosmid clones show conplete, widely spaced human class I genes with occasional clusters. Gene 2_0_:485-489. 84 Malissen, M., Malissen, 8., and Jordan, B.R. (1982b) Exon/intron organization and canplete nucleotide sequence of an HLA gene. Proc. Natl. Acad. Sci. USA _7_9_:893-897. Maloy, W.L., Nathenson, S.G., and Coligan, J.E. (1981) Primary structure of murine major histocompatibility complex alloantigens. J. Biol. Chan. 256:2863-2872. Mathis, D.J., Benoist, C.O., Williams, V.E., Kanter, M.R., and McDevitt, H.O. (1983) The murine B alpha innune response gene. Cell 32:745-754. Miyada, D.G., Klofelt, C., Reyes, A.A., McLaughlin-Taylor, E., and Wallace, R.8. (1985) Evidence that polymorphism isn the major histoconpatibility canplex may be generated by the assortment of subgene sequences. Proc. Natl. Acad. Sci. USA _8_2_:2890-2894. Miyahara, M., Sumiyoshi, H., Yamamoto, M., and Endo, H. Strand specific transcription of satellite DNA I in rat ascites hepatoma cells. Biochem. Biophys. Res. Contn. 130:897-903. Murphy, D., Brickell, P.M., Latchman, D.S., Willison, K., and Rigby, P.W.J. (1983) Transcripts regulated during normal embryonic develogtent and oncogenic transformation share a repetitive elenent. Cell £:865—871. Muscarella, D.E., VOgt, V.M., and Bloom, S.E., (1985) The Ribosomal RNA Gene Cluster in Aneuploid Chickens: Evidence for Increased Gene Dosage and Regulation of Gene Expression. J. Cell Biol. 101:1749-1756. Nathenson, S.G., Uehara, H., Ewenstein, B.M., Kindt, T.J., and Coligan, J.E. (1981) Primary structural analysis of the transplantation antigens of the murine H—Z major histocompatibility canplex. Ann. Rev. Biochem. 50:1025-1052. Okazaki, W., Witter, R.L., Romero, C., Nazerian, K., Sharma, J.M., Fadly, A. , and Ewert, D. (1980) Induction of lymphoid leukosis transplantable tumors and the extablishnent of lymphoblastoid cell lines. Avian Pathol. _9_:3ll. Pevzner, I.Y., Trowbridge, C.L., and Nordskog, A.W. (1978) Recombination between genes coding for inmune response and serologically determined antigens in the chicken 8 system. Intnunogenetics _7_: 25. Plachy, J. and Benda, V. (1981) Location of the Gene Responsible for Rous Sarcoma Regression in the B-F Region of the 8 Conplex (MHC) of the Chicken. Folia Biologica (Praha) 27:363-367. Plachy, J., Jurajda, V., and Benda, V. (1984) Resistance to Marek's Disease is Controlled by a Gene within the B—F Region of the Chicken Major Histoconpatibility Complex in Rous Sarcoma Regressor or Progressor Inbred Lines of Chickens. Folia Biologica (Praha) _3_0_:251—258. Ploegh, H.L., Orr, H.T., and Straninger, J.L. (1980) Molecular cloning of a human histoconpatibility antigen cDNA fragment. _7_7_:6081-6085. 85 Quddus, J., Prakash, 8., Bahn, R., Banerjee, 8., and David, C. (1986) Genetics of imnune response and susceptibility to disease. Proceedings of the Third World Congress on Animal Genetics. Lincoln, N8, 1986. Reiss, C.S., Evans, G.A., Margulies, D.H., Seidnan, J.G., and Burakoff, S.J. (1983) Allospecific and virus-specific cytolytic T lymphocytes are restricted to the N or Cl danain of H-2 antigens expressed on L cells after DNA-mediated gene transfer. Proc. Natl. Acad. Sci. USA 8_0:2709- 2712. Reynaud,C-A., Dahan, A., and Weill, J-C. (1983) Complete sequence of a chicken lambda light chain itttnunoglobulin derived from the nucleotide sequence of its mRNA. Proc. Natl. Acad. Sci. USA _80_:4099-4103. Ruff, M.D. and Bacon, L.D. (1984) Coccidiosis in 15.8-Congenic Chicks. Poultry Sci. 93172-173 (abstract). Salter, D.W., Smith, E.J., Hughes, S.H., Wright, S.E., and Crittenden, L.B. (1987) Virology 157:236-240. Schamboeck, A., Korman, A.J., Kamb, A., and Strominger, J.L. (1983) Organization of the transcriptional unit of a human histocanpatibility antigen: HLA-DR heavy chain. Nucleic Acids Res. _l_l_:8663-8675. Scott, M.D.R., Westphal, K-H., and Rigby, P.W.J. (1983) Activation of mouse genes in transforned cells. Cell 34:557-567. Shimizu, Y., Koller, 8., Geraghty, D., Orr, H., Shaw, 8., Kavathas, P., and DeMars, R. (1986) Transfer of cloned human Class I major histoconpatibility canplex genes into HLA mutant htmtan lymphoblastoid cells. Mol. Cell. Biol. 6:1074-1087. Simonsen, M., Hala, K., and Vilhelmova, M. (1977) Alloaggression in chickens: analysis of the B-canplex by means of GVH splenategaly and by inhibitory antibodies. adv. Exp. Med. Biol. 8_8_:267. Skjodt, K., Koch, C., Crone, M., and Simonsen, M. (1985) Analysis of Chickens for Recombination within the mo (B-complex) . Tissue Antigens 25:278-282. Sood, A.K., Pereira, D., and Weissman, S.M. (1981) Isolation a partial nucleotide sequence of a cDNA clone for human histoconpatibility antigen HLA-8 by use of an oligodeoxynucleotide primer. Proc. Natl. Acad. Sci. USA 18:616-620. Steinnetz, M., Grelinger, J.G., Fisher, D., Hunkapiller, T., Pereira, D., Weissman, S.M., Uehara, H., Nathenson, S., and Hood, L. (1981) Three cDNA clones encoding mouse transplantation antigens: Homology to inmunoglobulin genes. Cell 24:125-134. Steinnetz, M. and Hood, L. (1983) Genes of the major histocompatibility canplex in mouse and man. Science 222:727-733. 86 Steinmetz, M., Minard, K., Horvatb, 8., MxNicholas, J., Srelinger, J., Wake, C., Long, E., Mach, 8., and Hood, L. (1982) Molecular map of the imnune response region from the major histocompatibility coupler of the mouse. Nature 400335-42. Steinmetz, M., Moore, K.W., Frelinger, J.G., Sher, B.T., Shen, F-W., Boyse, E.A., and Hood, L. (1981) A pseudogene hanologous to mouse transplantation antigens: Transplantation antigens are encoded by eight exons that correlate with protein domains. Cell 2_5:683-692. Steinmetz, M., Stephan, D., and Lindahl, K.F. (1986) Gene organization and reconbinational hotspots in the murine major histocompatibility canplex. Cell fi:895-904. Stetler, D., Das, H., Nunberg, J.H., Saiki, R., Sheng-Dong, R., Mullis, K.B., Weissman, S.M., and Erlich, H.A. (1982) Isolation of a cDNA clone for the human HLA-DR antigen alpha chain by using a synthetic oligonucleotide as a hybridization probe. Proc. Natl. Acad. Sci. USA _7_9_: 5966-5970 . Toivanen, P. and Toivanen, A. (1983) Avian Major Histoconpatibility Canplex. The International Union of Immunological Societies Proceedings No. _6_6:107-114. Tonnelle, C., DeMars, R., and Long, 8.0. (1985) DO beta: a new beta chain in HLA-D with a distinct regulation of expression. EMBO J. 4: 2839—2847 . Vainio, 0., Koch, C., and Toivanen, A. (1984) B—L Antigens (Class II) of the Chicken Major Histocompatibility Canplex Control T-B Cell Interaction. Inmunogenetics 42:131-140. Vilhelmova, M., Miggiano, V.C., Pink, J.R.L., Hala, K., and Hartmanova, J. (1977) Analysis of the alloinlnune properties of a reconbinant genotype in the major histocompatibility complex of the chicken. Eur. J. Inmunol. 1:674. Vittetta, E.S., Uhr, J.W., Klein, J., Pazderka, F., Moticka, E.J., Ruth, R.E., and Capra, J.D. Homology of (murine) H-2 and (human) HLA with a chicken histocompatibility antigen. Nature 270:535. Vogel, J., Kress, M., Khoury, G., and Jay, G. (1986) A Transcriptional Enhancer and an Interferon-Responsive Sequence in Major Histocatpatibility Canplex Class I Genes. Mol. Cell. Biol. _6_:3550- 3554. ‘ Wake, C.T., Long, E.O., Strubin, M., Gross, N., Accolla, R., Carrel, S., and Mach, 8. (1982) Isolation of cDNA clones encoding HLA-DR alpha chains. JProc. Natl. Acad. Sci. USA 72:6979-6983. Weiss, E.H., Mellor, A., Golden, L., Fahrner, K., Simpson, E., Hurst, J., and Flavell, R.A. (1983) The structure of a mutant _B_-:3 gene suggests that the generation of a polymorphism in L2 genes may occur by gene conversion-like events. Nature 104:671-674. 87 Woodward, J.G., Orn, A., Harmon, R.C., Goodenow, R.S., Hood, L., and Frelinger, J.A. (1982) Specific recognition of the product of a transferred major histocompatibility complex gene by cytotoxic T lymphocytes. Proc. Natl. Acad. Sci. USA 19:3613-3617. Yamamoto, M., Maehara, Y., Takahashi, K., and Endo, H. (1983) Cloning of sequences expressed specifically in tumors of rat. Proc. Natl. Acad. Yamamura, K., Kikutani, H., Folsom, V., Clayton, L.K., Kimoto, M., Akira, S., Kashiwamura, S., Tonegawa, S., d Kishimoto, T. (1985) Functional expression of a microinjected E alpha gene in C57BL/6 transgenic mice. Nature _3_l_6_:67-69. Zaleski, M.B., Dubiski, S., Niles, 8.8., and Cunningham, (1983) Chapter 16: The murine major histocompatibility H-2 complex. Innunogenetics, Pitman Publ. Inc., Boston. p. 307-430. Ziegler, A. and Pink, J.R.L. (1975) Characterization of Major Histocanpatibility (8) Antigens of the Chicken. Transplantation 40:523- 527. Zinkernagel, R.M. and Doherty, P.C. (1974) Nature 251:547. Zuniga, M.C., malissen, 8., McMillan, M., Brayton, P.R., Clark, S.S., Forman, J., and Hood, L. (1983) Expression and function of transplantation antigens with altered or deleted domains. Cell 3:535- 544. Chapter 8: Transformation-related viral transcripts in Marek's Disease Virus-transfommed cell lines. 88 89 Marek's disease (MD) is a malignant lymphoma of chickens, first described by Josef Marek (1907) as a polyneuritis causing paralysis. It was recognized as a viral disease, differentiated from the retroviral- induced lymphanas, as the result of work by Biggs (1961), Campbell (1956), Biggs and Payne (1963.) It has been among the most economically important diseases in contnercial chicken flocks, especially in the late 1950s and 1960s, when its losses in the United States reached $0.5 million per day (Calnek and Witter, 1984.) Marek's disease is a highly contagious disease, caused by a herpes virus (Marek's disease virus, MDV) which is usually strictly cell-associated (Nazerian, 1980). However, latent infection persists for the life of the bird and periodically erupts into productive infection in the feather follicle epithelium, releasing mature infectious virions into the envirorment when feathers are molted or dead cells are shed (Calnek, 1986.) Thus, the virus persists in all flocks. Losses have decreased tremendously, however, as a result of the introduction in 1969 and 1970 of vaccines composed of either attenuated MDV strains (Cnurchhill et al, 1969) or the antigenically—related, nonpathogenic herpesvirus of turkeys (HVT), (Okazaki et al, 1970.) Thus, MDV is the first tumor virus for which a safe and effective vaccine has been developed, and is in routine use 0 Although economic losses have declined due to the vaccine, MDV and its lymphana continue to be studied in several areas. The mechanians of viral transformation, genetic resistance, and vaccinal inmunity are all being studied, both in an effort to better understand Marek's disease, 90 and as a model for other herpesvirus oncology (Nonoyama, 1982; Calnek, 1986.) MD is an excellent model systetn because it is a naturally occuring disease for which the appropriate genetic strains of both host and virus are available and well characterized. In addition, vaccine development work continues. Very virulent forms of the virus have emerged in the field, against which the vaccine was less effective (Witter, 1983; Witter, 1985), and new approaches, such as developnent of bivalent vaccines (Witter, 1982), have been required in the past, and may be necessary again in the future. Thus, the virus, its pathogenicity, the host's iminity to it, and vaccines have all been studied extensively, making Marek's disease one of the best understood of all the herpes virus-induced neoplastic diseases (Calnek, 1986.) However, even with the well—understood descriptive biology of the tumors, the molecular mechanisms responsible for the transformation remain elusive. BIOLCBY OF MAREK'S DISEASE MDV usually infects the host via the respiratory tract when the bird occupies a contaminated environment. Little or no virus infection can be detected in the cells of the respiratory passages, and means of travel to the primary lymphoid organs has not been established. Early cytolytic infection occurs in the spleen, bursa of Fabricius, and thymus. The viral internal antigens (VIA) are first detectable in a few cells at 1-2 days post infection. Maximum infection occurs between 3 and 5 days post infection, and the cytolytic infection is largely over 91 by 7 days post infection (Calnek et al, 1984a.) This stage can be largely prevented by embryonic bursectomy (Schat et al, 1980); the cells which are cytolytically infected at this stage have been shown by staining with appropriate antisera to be primarily 8 cells (Shek et a1, 1983). However, a small proportion (usually less than 10%) of these cells are T cells (Calnek et al, 1984a.) As a result of the cytolytic infection, both the bursa and thymus show significant atrophy. As the number of cytolytically infected cells drops off during days 6-7 post infection, latently infected cells appear in the spleen and peripheral blood. These cells appear to be primarily activated T cells, bearing both T cell markers and Ia antigen, although about 2-4% of latently infected cells are 8 cells (Shek et al, 1983, Calnek et al, 1984a.) The viral genone is present in only a few copies per cell (Calnek et al, 1984a) and VIA are not detected (Calnek et al, 1981b.) However, VIA expression can be induced in these cells by 24-48 hour cultivation in vitro (Calnek et al, 1981a.) The amount of transcription or protein expression from other viral genes is unknown. These lymphocytes remain latently infected for the life of the bird (Witter et al, 1971.) It is not known whether other cell types may also harbor latent infections. However, after the second or third week post infection, foci of cytolytic infection appear in epithelial tissues, such as kidney, adrenal gland, and feather follicle epithelium. These infections result in necrosis and local inflattmatory response with infiltration of mononuclear cells. The feather follicle epithelium is 92 the only tissue known to support a fully productive infection, releasing enveloped, infectious virus to the environment. During this sane period, beginning in the second or third week post infection, another wave of cytolytic infection begins in lymphocytes in the central lymphoid organs. It is not known whether it is primarily 8 or T cells which undergo cytolytic infection at this titne. A prominent feature of Marek's disease is the induction of innunosuppression (Lee, 1984.) A transient inmnosuppression follows the initial wave of cytolytic infection in the central lymphoid organs. Although this may be partly due to the atrophy of both bursa and thymus following the cytolytic infection, both humoral and cell-mediated inlnune response to MDV are detectable by about 7 days post infection (Sharma and Coulson, 1977; Sharma et al, 1978; Confer et al, 1980). Therefore, the atrophy is not sufficiently canplete to account for the total suppression seen. At this time, about 6—7 days post infection, a transient decrease in in vitro mitogen responsiveness (Lee et al, 1978; Liu and Lee, 1983b) has been attributed to the presence of suppressor macrophages (Theis, 1977.) However, by about three weeks post infection, a "permanant" inmunosuppression develops at the sane time as the second wave of cytolytic infection in genetically susceptible chickens. This inmunosuppression includes depressed antibody responses (Jokowski et al, 1973), delayed skin graft rejection (Purchase et a1, 1968) and impaired PHA response (Lee et al, 1978.) 93 The early transient inmunosuppression, early cytolytic infection, and persistence of latently infected T cells occur in all chickens infected by MDV, regardless of the genetic resistance of the host chicken or the degree of pathogenicity attributed to the virus strain (Confer and Adldinger, 1980.) However, beginning at about three weeks post infection, the "permanent" inmunosuppression and the second wave of cytolytically infected lymphocytes occur only in those chickens undergoing full-blown disease. It may be that there is a cause-effect relationship between the later cytolytic infection and the permanent inmunosuppression (Calnek, 1986) but this has not been proven. It is also possible that the inmunosuppression is due to the effect of soluble suppressor factors, which have been associated with the spleens of inmunosuppressed chickens (Theis, 1977.) A final (and probably the most characteristic) facet of Marek's disease is the neoplastic transformation of lymphocytes and the developnent of gross lymphanas. These tumors contain not only the transformed cells, but also a variety of other lymphoid cells, including T and 8 cells and macrophages (Payne and Roszkowski, 1973.) The exact time of cellular transformation is unknown, but microscopic lesions can be detected as early as two weeks post infection, and gross lymphanas of organs such as gonads, spleen, and nerves are seen after about three to four weeks (Payne, 1982.) Tumor cell lines have been established which can form tumors when injected into susceptible birds (Akiyama and Kato, 1974; Nazerian et al, 1977; Calnek et al, 1978;) and these cells are characterized as activated T cells, possessing both T cell markers and Ia (Nazerian and Sharma, 1975; Ross et a1 1977; Schat et al, 1982.) 94 Numerous other antigens have been detected on the ID cell lines, but the most carefully studied of these has been MATSA, Marek's disease tumor- associated surface antigen (Witter et al, 1975.) Distribution of MATSA has been investigated using a variety of antisera—it is usually present on a majority of cells within the cultured MD cell lines, on a minority of cells from an MD tumor, and on a small percent of normal spleen cells (Witter et al, 1975.) It is not present on latently infected lymphocytes (Calnek et al, 1981b.) MATSA is probably host-gene encoded, as it varies among species or strains of host (Elmubarak et al, 1981; Ross et al, 1982; Powell et al, 1984.) It is also present on sane cells infected with non-transforming strains of MDV (Schat and Calnek, 1978.) MATSA-bearing cells have been found in the lymphoid organs as early as 5-7 days post infection (Murthy and Calnek, 1978), and in the gonad at 16 days post infection (Powell and Rennie, 1985.) This indicates that either transforned cells, or "pre—tumor" cells are present in the organs at these times. Although most tumor cells support a small amount of partially productive infection, very limited viral gene transcription and expression has been detected. Silver et al (1979) found only 12-14% of the viral genone was transcribed in a non-producer MD cell line, mar-1, canpared with transcription from nearly 50% of the genone in productively infected cells. Calnek et al (1981) found a very low level of expression of VIA and viral manbrane antigens in a survey of 31 MD cell lines. 9 5 VACCINES MDV strains fall into three serological groups; serotype I includes all pathogenic strains, and the tissue-culture attenuated strains derived from them. Serotype II strains are naturally occuring, non-pathogenic strains which infect chickens in the field. Serotype III is the herpesvirus of turkeys (HVT), which is normally found non-pathogenically infecting turkey flocks, but also is apathogenic when it infects chickens. Although there is sone antigenic cross reactivity between these serotypes, they can easily be distinguished using various serologic reagents, including monoclonal antibodies (Lee et al, 1983.) Vaccines have been developed using all three serotypes; attenuated serotype I or naturally non-pathogenic serotypes II and III. Most vaccines rapidly induce a cell-associated viremia, but do not produce a cytolytic infection of lymphocytes, as does the pathogenic MDV (Witter, 1985.) The majority of infected cells do not bear 8 cell or macrophage markers, although both latently-infected B and T lymphocytes may be present. The cell types infected by different serotypes may not be the same (Calnek et al, 1981b; Witter, 1985.) A transient innunosuppression, detected as a depression in the PHA response of T cells, has been noted at about 7 days post vaccination, but the response is quickly recovered (Lee et al, 1978.) The vaccination inmune response involves both antibody and cell-nediated responses to viral antigens, but probably does not include anti-MATSA antibodies. Protection against MD is related more to cell-mediated than antibody response, and is directed more to viral than tumor antigens (Witter, 1985.) 96 GENETIC RESISTANCE TO MD If a chicken is infected with an MDV strain other than a very virulent one, susceptibility or resistance of the host strain of chicken often nediates the outcone of the MDV infection (Schat et al, 1981,) which proceeds through the early stages of infection described above regardless of the genetic background. There are at least two genetic loci responsible for resistance, probably both are involved in regulating inmune response (Lee et al, 1981,) although there is evidence that ability of MDV to infect T lymphocytes is also controlled by a locus which affects the incidence or severity of M) (Powell et a1, 1982.) . A major host gene influence on MD succeptibility is the major histoconpatibility canplex, the _B_-locus; this probably involves an active rejection of the tumor cells (Pazderka et al, 1975; Longenecker et a1, 1976; 8ri1es et al, 1977; Briles et al, 1983.) Pevzner et al (1981) have also observed a linkage between Marek's resistance and response to GAT, which would be an inmune response gene, probably within the B—complex. Another locus possibly involved in MD susceptibility is the Ly-4 locus, controlling a T cell antigen (Fredericksen et al, 1977.) In general, the effects of genetic resistance and vaccination are additive, and selection of resistant strains is important, in an effective vaccination program (Witter, 1985.) However, in at least one experiment involving B—congenic chickens, sate B-haplotypes conferred 97 increased vaccination response, and were ranked as more resistant than other haplotypes after vaccination, even though they were more susceptible before vaccination (Bacon, 1987.) OTHER HERPESVI RUSES Other animal herpesviruses have been described; sone of them also appear to be oncogenic. Luecke's frog herpesvirus was the first oncogenic herpesvirus to be described (Granoff, 1982); it causes a renal adenocarcinoma in 1-9% of frogs in high incidence areas. The biology has not been completely characterized, but a unique temperature dependence of cytolytic activity exists. At warm ten'peratures, the virus is latent and tumors develop, at cold temperatures during hibernation, a percentage of tmnor cells develop cytolytic infections (Sharma, 1976; Granoff, 1982.) Other viruses have been associated with oncogenicity in the natural hosts, including Herpes sylvilagus in cottontail rabbits (Hinze and Chipnan, 1971,) Herpes papio in baboons (Falk, 1980,) Herpesvirus saimiri in primates (Falk, 1980,) and Epstein-Barr and Herpes simplex type 2 in humans (zur Hausen, 1980; Nahmias and Norrild, 1980.) Most studies of these viruses in regard to mechanism of transformation has used in vitro transformation assays. For example, hamster cells can be transformed constitutively by a small (about 6%) portion of the Equine herpesvirus type I genane, which integrates into the cellular genone (O'Callaghan et al, 1983.) Similar experinents with herpes 98 simplex and cytanegalovirus have yielded inconsistent, puzzling results with apparently different regions being involved in unknown functions related to transformation (Bishop, 1985.) A 2 kb region of Herpesvirus saimiri is required for transformation of T cells in culture, and for oncogenicity in primates, but it is not known whether this region is directly responsible for transformation (Desrosiers, 1985.) Another applicable in vitro study of interest is the finding that an innediate early gene product of pseudorabies virus (and a similar protein from Herpes simplex virus) could substitute for the activity of the adenovirus Ela product (Ben-Porat and Kaplan, 1982; Kingston et al, 1985.) However, that the gene products are not totally hanologous is shown by the fact that Ela cannot substitute for the herpesvirus gene products. Nevertheless, the parallelism is intriguing. Herpes simplex is the herpesvirus for which the most molecular genetic information is available. HSV-l DNA consists of a long unique region flanked by inverted repeats, which is joined to a short unique region, also flanked by inverted repeats. The long and short regions can invert with respect to each other, and all four possible configurations exist in equimolar amounts (Roizman, 1979.) The entire genone has been cloned (Goldin et al, 1981; Post et al, 1980) and sequencing is in progress (mGeoch et al, 1985; Rixon and McGeoch, 1985.) A considerable amount is known about specific gene transcription in HSV during productive infection (for review, see Wagner, 1982;) a very brief synopsis will be given here. HSV-1 mRNA is produced in the nucleus, is 99 capped and polyadenylated, and most is colinear with its gene, i.e., there is very little splicing that occurs in HSV mRNA. Overlapping mRNA families are cannon. The mechanisms for generation of these families include multiple pranotors, inefficient termination at a polyadenylation site, an additional promotor within a gene initiating transcription of mRNA for a shorter peptide product, and in a few cases, differential splicing patterns. Whether these gene families encode a single peptide, closely related peptides, or non-related peptides appears to vary, as all three possibilities have been identified in HSV. Intnediate early (alpha) genes are transcribed prior to viral genate activity, and have been mapped in or near the repeat sequences flanking both unique regions. At least one of these, ICP-4, appears to be necessary throughout infection, and codes for a protein which is autoregulatory, in analagy to the T-antigens of papova viruses. The beta genes are expressed after viral protein synthesis begins, and the ganlna gene products appear following viral genane replication. Several of the cell surface glycoproteins are encoded within the short unique region. Specific regions of the HSV genome involved in transformation have not been clearly defined; in vitro transformation has yielded varying results (Bishop, 1985,) studies of cervical carcinomas induced by HSV-2 may implicate a 35-38 kD peptide, but its gene is not yet defined (Gilman et al, 1980.) 100 HSV gene expression in lytically-infected cells is interrupted in neuronal latency, and viral transcripts are undetectable in Northern blots (Puga et al, 1978.) However, more recent studies have indicated the continued expression of the ICP-0 gene in latently-infected mouse trigeminal ganglia, using various nethods (Puga and Notkins, 1987; Deatley et al, 1987.) The potential importance of sone inmediate early genes in not only lytic infection, but also latency, is echoed by their potential role in tumor- induction. Both adenoviruses and papovaviruses contain genes which have been involved in grins-activation of oncogenes--the Ela, Elb of adenoviruses, and the large T—antigen of polyoma and SV-40 (Kingston et al, 1985.) There are inmediate early gene products of both pseudorabies and HSV which appear to have similar activities in vitro (Ben-Porat and Kaplan, 1982; Bishop, 1985.) This leads naturally to the possibility of herpesvirus innediate early gene products also playing a role in tumor induction. MOLECULAR BIOLCBY STUDIES OF MDV Serotype I MDV has a genone which consists of linear, double-stranded DNA of approximately 120x106 daltons (Lee et al, 1971,) or about 180 kb. The genane structure appears similar to that of Herpes simplex virus (HSV) with both long unique and short unique regions, each of which is flanked by inverted repeats (Cebrian et al, 1982,) although there is no evidence for isoners representing different arrangements between the 101 long and short unique sequences. Electron microscope conparison measurements indicated that the serotype III (HVT) genane is about the sane total size as HSV DNA, with serotype I being about 15% larger. The use of molecular cloning technology has enabled researchers to look carefully at structural characteristics of the viral genone. Fukuchi et al (1984) have cloned nearly all of the viral genane in B_gnHI fragments. Gibbs et al (1984) cloned a sizable number of the Echl fragments. The canposite restriction enzyme map (adapted from Silva and Witter, 1985) which shows the location of these fragments is shown in Figure 81. Dr. Nonoyama and Dr. Kung have kindly made these clones available to other researchers for further studies. Restriction enzyme digest comparisons of DNAs from different strains has shown very similar patterns within strains of a given serotype, but totally different patterns between serotypes (Hirai et al, 1979; Ross et a1, 1983; J. Carter, personal cannunication.) This carparison is in agreement with monoclonal antibody data on proteins in which certain antibodies can differentiate between serotypes, but very little difference is seen among strains within a serotype (Lee et a1, 1983; Silva and Lee, 1984; Ikuta et al, 1983.) However, nest antibody preparations recognize conmon determinants between serotypes, and the vaccines are apparently effective due to conlnon epitopes between serotypes. Therefore, DNA hybridization studies have been carried out to further investigate the extent of homology between DNAs of the various serotypes. Initial hybridization studies indicated very little honelogy, less than 5% (Lee et al, 1979; Hirai et al, 1979.) However, 102 using very low stringency conditions, it has been determined that one can detect at least 70—80% similarity between HVT and serotype I MDV, over 90-95% of the genone (Gibbs et al, 1984; Fukuchi et al, 1985a; Hirai et al, 1984.) Fruitful comparisons have also been made between the DNA of pathogenic serotype I strains and the attenuated strains derived from then (Hirai et al, 1981b; Ross et al, 1983; Fukuchi et al, 1985b; Silva and Witter, 1985; Maotani et a1, 1986.) In particular, all groups found alterations within a region from the inverted repeats flanking the long unique region, and located within the cloned fragments BanHI-D, BamHI-H, and EcoRI-F. It was shown that the loss of this fragment in restriction analysis was correlated with the number of passages in tissue culture, and with loss of pathogenicity (Silva and Witter, 1985.) The altered region has been cloned from high passage DNA, and the specific area has been identified and sequenced (Maotani et al, 1986.) The region conmon to BamHI-H, BamHI-D, and EcoRI-F contains a 132bp repeated segment which apparently becones duplicated a large number of tines. This fragment does not contain an open reading frame, but does contain short inverted repeats which could indicate secondary structures of importance in regulation of gene expression. In addition to expansion of this region, H-J. Kung and collaborators (personal conmunication) have detected regions within the same fragnents which show homology by hybridization with the long terminal repeats of reticuloendotheliosis virus (REV.) Since REV is capable of inducing, in susceptible chickens, T cell lymphomas which appear quite similar to the 103 MDV lymphomas (Witter et a1, 1986,) the sequence similarities within this region may also indicate that it is involved in significant regulatory functions. Preliminary evidence that this region may play a significant regulatory role was discussed at the Herpesvirus Workshop in England (1986). Transcription of a specific 2 kb message beginning about 600 nt downstream from this expansion region of BamHI-H has been postulated to occur in CEF infected with pathogenic MDV strains, but not with high- passage, attenuated serotype I strains. This is distinct fran a 1.8 kb mRNA, which is supposed to begin several kilobases upstream of the expanded region and is transcribed from the opposite strand. Since most of the molecular biological studies have been done on i_n yi_t£c_>_ infected fibroblasts, the status of viral DNA and its transcription in tumor cells is still largely unknown. Nazerian and Lee (1974) looked at the number of genanes present in the MSB-l tumor cell line, which does support a limited amount of viral replication, and virus-specific antigens are detectable in 1-2% of the cells. They estimated an average of 60-90 genones per cell. Previous estimates for MDV genanes in in vivo tumors was 3-15 per cell (Nazerian et al, 1973), but this estimate is probably too low because of the presence of non- tumor cells in the tumor. Whether transformation requires the integration of MDV DNA into the genone is not clear; there are reports of at least sone of the MDV DNA being integrated in tumor cell lines (Kaschka-Dierich et al, 1979; Hughes et al, 1980,) although most is probably episomal (Tanaka et al, 1978.) 194 Specific efforts to detect transcription from the MDV genome in tumors or in cell lines derived from tumors has not been reported to date. Efforts have begun, however, to identify and study the genes for inmunogenic proteins produced by MDV infection. P. Coussins and L.Velicer (personal comnunication) have identified, mapped, and sequenced the gene for the A antigen from both MDV and HVT. It is located within the BamHI-B fragment of serotype 1, and contains an open reading frame which apparently codes for a secreted glycoprotein of the expected size. The amino acid sequence homology between the proteins of the two serotypes is quite strong, but the nucleotide sequence contains many changes in silent or conservative positions. I. Sithole and L. Velicer are also making considerable progress in identifying and sequencing the gene for the B antigen (personal coumunication.) 165 MATERIALS AND METHODS Many of the methods used have already been described. Only those which were unique to this project will be described here. Cbll lines: A synopsis of cell lines used and their characteristics is listed in Figure B3. The MSB—l cell line was developed by Akiyana and Kato (1974.) The RP-l cell line was developed from a transplantable tumor, JMV, which had been passaged in birds an unknown number of times before adaptation to tissue culture (Nazerian et al, 1977.) RP-4 was also developed from a transplantable tumor, induced by GA strain of MDV in a E: homozygous bird (Nazerian et al, 1978.) RP-l9 is a turkey MDV; GA-induced cell line, which apparently is of immunoglobulin-secreting B cell origin (Elmubarak et a1, 1981.) RP-l3 is a B lymphoblastoid cell line, induced by REV (Nazerian et a1, 1982.) RP-9 has been descibed previously; it is an ALV-induced B lymphoblastoid cell line. All cell lines were grown in Liebowitz-McCoy media (1:1; GIBOO Laboratories, Grand Island NY) supplemented with 20% chicken serum, 19% bovine fetal serum, 5% tryptose phosphate broth, 2w glutamine, 1mM sodium pyruvate, and 5.011!!! 2-mercaptoethanol. They were grown at 41C in 5% C02. Infected fibroblasts: Duck embryo fibroblasts (DEF) were prepared from 13 day embryos, and grown at 37C in Liebowitz-McCoy media supplemented with 4% calf serum. After reaching confluency, they were maintained in the same media with 1% calf serum. 1fl6 Virus stocks used were cells infected with the following strains: GA (Eidson and Schmittle, 1968,) Mdll (Witter, 1983,) and Mdll/7SC (Witter and Lee, 1984; Silva and Witter, 1985.) Multiplicity of infection is difficult to determine with cell—associated virus, but the cultures were infected by adding one infected cell to about 3-6 uninfected fibroblasts. Usually, cells were harvested when the cultures displayed extensive cytopathic effects, although for one experiment, the RNA was purified fran cultures only 8 hours post-infection in order to see transcripts from inmediate early viral genes. 197 RESULTS A series of cloned DNA fragments representing a large portion of the Marek's disease virus genome was selected fran the wl (Fukuchi et al, 1984) and the §_c_o_R1 (Gibbs et a1, 1984) clones available to us. A genomic map showing the locations of the cloned fragments used in this study is dagranmed in Figure Bl. The plasmid DNA was purified, the inserts were cut out with either EcoRI or BamHI, and the fragments separated in an agarose gel to verify the presence of an insert of the correct size (Figure B2.) The plasmid insert sizes and vectors are listed in Table B1. For the survey experiment, total RNA was isolated from three MD—induced lymphoblastoid cell lines using the method of centrifuging guanidinium isothiocyanate extracts through a 5.7M CsCl cushion. For each probe used, 15ug total RNA from each line was separated by electrophoresis through agarose gels containing formaldehyde, in adjacent lanes, and the RNA was blotted to nitrocellulose. In order to crudely differentiate between early and late viral gene expression from Md11/750 MDV, a confluent culture of duck embryo fibroblasts (DEF) was infected using a multiplicity of infection (moi) of one infected cell per four DEF. RNA was harvested as before at 8 hours (for "early" transcripts) and at 48 hours (for "late" transcripts), when the cells exhibited substantial cytopathic effect. Because of the cell-associated nature of the virus, it is difficult to obtain cultures in which the cells are synchronized in the time of 168 7mm?" 595:5 cam m>:m «vmma .Hm uw Evian Scum vmummvmv macaw mum mmcoHo Hmoom new Hus—mm mo cofiumooH new cogmnficmmuo msocmm Hmuwcwm one .msu? wmmmmwv mlmumz m0 m9: cojofiuumwm ”Hm musmwm .o a: m o a~......u zzsosx “_c [C [ [[[ttl Ec[[ [[ «meow < .3 :Fxsmz m c613. m 9an u. o ode N_ ._ t‘f NOne $6 «of. £58 OS 9. OO— 0: cap 09 ca 00 CV ON 0 109 Eco Bam BCEABCDEFGH Eco Born I JKL‘IQXI2I3K2K3L Figure B: Photograph of ethidium bromide stained gel showing BamHI or $9121 digested plasmid clones. llfl Cell Transforming mmunofluorescent reactions Reference line Species Virus anti anti Rabbit Monoclonal B T anti anti cell cell MATSA MATSA MSBl chicken MDVeBCI - + 95% 95% Akiyama and Kato, 1974 RPl chicken MDVeJM - + 75 95 Nazerian et a1, 1977 RP4 chicken MDVAGA - + 85 85 Nazerian et a1, 1978 RP9 chicken ALV-RAVZ + - 0 0 Okazaki et a1, 1989 RP13 chicken REV-CS + - 0 25 Nazerian et a1, 1982 RP19 turkey MDV—GA + - 3S <5 Elmubarak et a1, 1981 Figure B 3: MDV tumor cell lines used in transcription survey. 111 Clone insert vector Ampr Tetr Camr EcoRI-B 12.5 pBR328 + + — -C 11.8 " + + - -E 7.8 " + + _ -F 6.9 " + + _ -I 5.15 " + + _ -J 4.4 " + + _ -K 4.1 " + + - -Ll 3.35 " + + _ ~Q 2.2 " + + — -X+F .9+6.9 " + + — BamHI-A 23.5 pACYC184 - - + —B 18.5 " _ _ + -C 15.8 " _ - + -D 11.5 " - _ + -E 9.7 " - - + -F 8.9 " - _ + -G 7.1 pBR322 + - _ -H 5.5 " + - _ ~12 5.2 pACYC184 +/¥ - + -13 5.2 " - - + -K2 3.6 pBR322 + - _ —K3 3.6 pHC79 + - - -L 3.8 " + - _ Table Bl: MDV genomic clones used in transcription survey. 112 infection. Therefore, immediately after infection in this system, the infected cells which are added will be producing late gene products, while the newly-infected cells begin producing early gene transcripts. As a result, the "early" and "late" designations for the RNA obtained from these cultures only refer to the probable enrichment of the RNA for either early or late gene expression. Approximately 15ug of the early and late MDll/75C—infected DEF total RNA was electrophoresed and blotted in the same manner as the cell line RNA. The total plasmid DNAs of the 23 §a__mHI and E29121 clones listed in Table B1 were individually nick—translated and hybridized at high stringency with filters containing the three cell line RNAs and the two infected cell RNAs. After washing in 0.1xSSC at 65C, all filters were exposed to x-ray film with intensifying screens for 4, 18, and 48 hours. The most useful exposure of each was photographed, and is shown in Figures B4, B5, and B6, together with a map of the probe location on the genome. Major RNA bands seen in the lanes are marked by dots; the calculated sizes of these transcripts are tabulated in Table BZ. Because each gel was run separately, and some of the molecular weight markers were not clearly visible, the sizes are all approximate, and in some cases, it may be misleading to directly compare the sizes obtained with different probes . Further studies were undertaken to better understand the transcriptional pattern of genes located in the repeats flanking the long unique region. RNA blots were prepared as before with the following differences. Additional cell lines were used, to provide more information and better 113 Figure B4: Northern blots of RNA from cell lines and infected cells, probed with nick-translated DNA from the short region of the genane. 1=RP1, 2=RP4, 3=MSB1, 4=48hr culture of Md11/75C-infected DEF, 5=8hr culture of MDll/75C-infected DEF. Dots indicate major RNA transcript, and sizes are listed in Table B2. 114 _ cum 3m 9553 115 Figure BS: Northern blots probed with DNA from the long unique region. RNA is the same as in Figure B4. 116 117 Figure B6: Northern blots probed with DNA from the repeats flanking the long unique region. RNA is the same as in Figure B4. 118 _ "mm magma 119 Table B2: MDV transcripts identified in survey. CEIl lines infected cells Probe size RP-l RP-4 MSB-l size 48hr 8hr l. BamHI-A 4.25 + + + 4.25 + + 1.9 + + + 1.9 + + 8.7 + + + 2.EcoRl-I ? - + - 6.3 - + 3045 " + 2.8 + + 1.5 - + 3. EcoRl—J — — — - >7.5 — + 3.7 +/; + 1.5 +/- + 4. BamHI-G - - - - 3.1 - H- 2.6 + ++ 1.8 + + 5. BamHI-C - - - - ? +/- +/- 6. BamHI-F - - - - ? + + 2.8 +/- +/- 1.8 + + 8. BamHI-K3 - - - - ND ND 9. BamHI-I3 - - - - 5.2 - + 3.7 + + 2.8 + ++ 1.8 + + 18. BamHI-E - - - - - - - ll. EcoRI-E - - - - 3.8 + + 1.4 + + 12. EcoRl-Ll — - - — 3.7 + + 13. BamHI-B - - - - 3.6 + + 2.8 + + 2.5 + + 1.7 + + 14. EcoRl-C . - - 2.6 + + 1.8 + + h‘hDUJU1 O ¢.o\uua> + + + + I l 128 Table 82: continued 15. EcoRl-F 2.3 ++ - +/- 3.6 + + 1.1 + ++ 16. EcoRl-K - - - - 3.3 - ++ 17. EcoRl-B 8.1 + - - - — — 5.2 ++ '- '- 4.2 ++ - - 1.5 ++ ++ 19. EcoRl—Q 3.8 - - + 2.85 +/; + 204 "" - + 2.3 ++ +/— - 2.8 +/; + ++ 1.6 +/4 +/4 + 28. BamHI-L 8.4 +/— ? +/L ND 1.8 + + 23. BamHI-IZ 2.2 +/; - +/- 1.4 + ++ 1.9 - - + 1.8 ++ +/- + 1.6 - - +/- ND=not done -,+/;, +, ++ a arbitrary designations comparing presence of bands within a single probe only 121 Figure B7: Northern blots of RNA from various tumor cell lines probed with cloned DNA which maps to the inverted repeats flanking the long unique region of the MDV genome. - .u Fm 0H"..me - _ 2 . 2 O cum N. Eon "— cum 1 m—mmk ammo o—n—xm 5m<< V 0_ .—n_¢m 2:”. N 3:. — :50m @00m 123 omidczum £824 .8; .B 35m; 5 mm 3905 dawn on“. 53 @303 63? >9 ca? @3035 man Bonn 52 no 303 593qu “mm 352 . 7‘77‘1': J nu. an... o 8m N_ can u. 8“ 124 Table BS: Transcript sizes detected with probes specific to the long repeat regions of MDV. Tumor cell lines MDV infected DEF Probe RPl RP4 MSB RP19 GA Mdll Mdll/75C EcoRl-B — - - - 5.1 5.1 5.1 3.1 3.1 3.1 2.8 2.8 2.8 BamHI-H +/4 - - - 4.4 4.4 - 1.5-2.3 1.5-2.3 ? 1.2-1.3 1.1-1.3 1.1-1.3 EcoRI-F 2.1 - - - 4.4 4.4 - 1.5-2.3 1.5-2.3 ? 1.1-1.3 1.1-1.3 1.1-1.3 BamHI-IZ 3.3 3.8 4.8 4.8 - 2.7 2.8 2.6 3.7 3.7 - 2.2 2.1 2.3 2.1 2.2 2.2 2.2 1.6 1.7 2.8 1.7 1.5 1.5 1.5 1.5 1.3 EcoRI-Q 3.4 3.2 4.8 4.8 4.8 2.8 2.8 2.6 3.8 3.8 3.8 2.3 2.3 2.3 2.1 2.5 2.5 2.5 1.8 2.8 1.8 1.6 1.5 1.5 125 controls. RNA was not obtained from infected cells at "early" times, but both pathogenic and attenuated serotype I viruses were grown in infected DEF. Probes used from the repeat regions included EcoRl-B, BamHI-H, BamHI-IZ, EcoRl-F, and EcoRl-Q. The results are shown in Figures B7 and B8, and the sizes of the transcripts marked by dots are shown in Table BB. The Eco Rl-B transcription pattern was quite different from that previously obtained during the initial survey experiment. At that time, transcripts were detected only in RP-l cells, and not in other cell lines or infected cells. The data in Figures B7 and B8 indicate substantial transcription in all infected DEF, but not in any cell lines. The reason for this discrepancy in not known. In order to better understand the genomic clones used in probing transcription from this region, additional restriction mapping of some of the genomic clones was done; the composite map is shown in Figure B9a. The EcoRl-B fragment had not previously been located next to the EcoRl-X fragment, and the order of sane of the wI fragments within the EcoRl-B insert has been changed from that published (Fukuchi et a1 , 1984.) In order to confirm that the EcoRl-X fragment cloned adjacent to EcoRl—F and EcoRl—B is the same as that seen in the BaIrI-II-H clone, these three clones were digested with _Ec_oR1, and the fragments were separated on agarose gels, transferred to nitrocellulose, and hybridized with BamHI-H. As seen is Figure B18, the 988bp fragment digested frail all three of these clones hybridized to the BamHI-H probe, indicating their identity. The clones BamHI-H, EcoRl—Q, and EcoRl-B were digested with both E9931 and §a__mHI, and the 599121-9311 fragments were gel purified to produce the probes 1-5 marked on the map in Figure B9b. Probe 6 is a 126 1955 .m an woof/one unamoom no 83033 HmoomlHHHccw: m m“ o .mwcoHo 55.3 no Hmoom Eon—w 333% How mum: mLH “monoum now owns mucosamum $5 .m 63.835 mum mucosa—mum Hmoom new Hag—mm .mwocmsvmm umwmwu @332: no umoe @5039: 5:253 wag—.5 9.3 no new .3me um :oflmmn >9; no mm: .m "mm 9..ng m me. I. l 505: > N . I o _ l I .n 228 ’. :38 Av 127 IL “- + N N ca>, Sharma, J.M., and Nazerian, K. (1975) Demonstration of a tumor-associated surface antigen in Marek's disease. J. Immunol. 115:177. zurHausen, H. (1988) The role of Epstein-Barr virus in Burkitt's lymphama and nasopharyngeal carcinama. In: Rapp, F., editor, Oncogenic Herpesv i ruses , vol . 2 , QC Press , Boca Raton FL , l3 .