Chrolnosoma (1992) 101:618-624

CHROMOSOMA 9 Springer-Verlag1992

A eDNA clone for a novel nuclear protein with D N A binding activity Tiziana Mattioni*, Clifford R. Hume**, Susanna Konigorski, Paula Hayes, Zvi Osterweil, and Janet S. Lee Immunology Program, Memorial Sloan Kettering Cancer Center, 1275 York Avenue, New York, NY 10021, USA Received May 6, 1992 Accepted May 12, 1992 by A. Bird

Abstract. In an effort to identify trans-acting factors regulating specific genes, we cloned a novel human gene, DBP-5. The cDNA clone contains a predicted open reading frame coding for a potential 1,179 amino acid protein. The m R N A corresponding to DBP-5 is ubiquitously distributed, and the gene is phylogenetically conserved. Immunofluorescence analyses with several cell lines indicate that the protein is localized to the nucleus. Sequence analysis revealed unusual features of the predicted protein structure, including four completely conserved repeats. The phylogenetic conservation of DBP-5, the ubiquity of its expression, its nuclear localization, and its ability to bind D N A sequences, raise the possibility that DBP-5 may play a role in the organization of interphase chromatin and/or in transcriptional regulation.

In our attempts to identify new factors specific for HLA class II promoters, we isolated a cDNA clone encoding a D N A binding protein by virtue of the affinity of its product for the X consensus sequence and flanking region within the H L A - D R A promoter. Here we describe the clone and its properties: four series of tandem repeats in the predicted protein structure, its expression in all tested cell lines, and the localization of its product to interphase nuclei with a peculiar punctate distribution. Although a fusion protein produced by the cDNA clone shows some specificity for M H C class II promoters in DNA/protein blotting assays, we suspect that this clone may encode a more general factor, possibly involved in transcription, replication, or in a structural role within the nucleus.

Materials and methods Introduction The recent development of powerful techniques for cloning cDNAs that code for D N A binding proteins has provided a substantial body of information on the mechanisms of transcriptional regulation in eukaryotes. Examples of clones identified by this method include widely expressed factors, such as C/EBP (Vinson et al. 1988) and H2TFI (Singh et al. 1988), and also a B cell specific factor, PU-I (Klemsz et al. 1990). cDNAs for the transcription factors RF-X and HXBP-1, which are likely candidates for regulators of major histocompatibility complex (MHC) class II gene expression, have also been isolated because they bind to specific sequences within the M H C class II promoter (Liou et al. 1988; Reith et al. 1989). * Present address: Department of Molecular Biology University of Geneva, CH- 1211 Geneva 4, Switzerland ** Present address." Departments Physiology and Biophysics BBll03, Columbia University, 630 West 168 Street, New York, NY 10032, USA Correspondence to: J.S. Lee

Construction and screening o f the eDNA libraries. A .~ gtll eDNA library generated from a normal B cell line was obtained from Clontech (San Diego, Calif.). Plating, induction, transfer to nitrocellulose filters, and screening with ~,.32p labeled probes were done as described (Vinson et al. 1988). Poly(A) + RNA was purified from the deletion mutant cell line 721.180 (Ehrlich et al. 1986) using a kit from Invitrogen. Subsequently, a library in pCDM8 (Seed and Aruffo 1987) was constructed from this RNA by Invitrogen (La Jolla, Calif.). Probe preparation and sequences. The - 1 3 2 / - 8 0 , XI, X2, and Y1 oligonucleotides used to screen the 2 gtll library were catenated by ligation, dephosphorylated, and labeled with [~:_32p] ATP by T4 polynucleotide kinase. Oligonucleotides were synthesized on a Dupont Coder 300 DNA Synthesizer and column purified (Nensorb Prep. Dupont NEN, Boston, Mass.) as recommended by the manufacturer. The following sequences were chosen :

- 132/-- 80 5'CTGGACCCTTTGCAAGAACCCTTCC CCTAGCAACAGATGCGTCATCTAAAATA3' XI 5'TGCAAGAACCCTTCCCCTAGCAACAGAT3' X2 5'TAGCAACAGATGCGTCATCTCA3' Y1 5'CTCAAAATATTTTTCTGATTGGCCAAAGAGTA3'

619 Both the sense and the antisense strands were synthesized with G A T C at each 5' end. The probe used to screen the pCDM8 library was derived from the 2 gtl 1 clone 5 by digestion with EcoRI and subsequent labeling of the purified insert. The fragment was labeled with [:~-32p]dCTP by random hexanucleotide priming. Dideoxy-sequencing was performed on double stranded plasmid templates and single stranded M13 templates; a Cyclone Kit (IBI, New Haeven, Conn.) was used to produce a sequential series of overlapping M13 clones (Dale et al. 1985). Overlapping sequences for both strands were obtained.

RNA blotting analysis and primer extension. Total cytoplasmic RNAs (10 gg) were fractionated on 1.3% agarose/formaldehyde gels containing 0.05 gg/ml ethidium bromide as described (Maniatis et al. 1982) and transferred to Zetabind membranes ( A M F Cuno Meriden, Conn.) in 20 x SSC (3 M NaC1, 0.3 M sodium citrate). The membrane was hybridized with the 32p-labeled insert from clone 2 gtl 1-5 as recommended by the manufacturer. Primer extension analysis was used to map the 5' terminus of the R N A transcript. Thirty micrograms of total R N A from the Raft and 180 celll lines and t R N A were hybridized in 10 mM TrisHC1, pH 8, 100 m M NaC1, 1 mM E D T A with the 5' end labeled specific oligonucleotide 5 ' C A G G A G G C A A A G G T G G C A C C A TATATGCCT3'. The hybridization mix was heated to 85 ~ C for 10 rain and allowed to anneal by cooling to 56 ~ C over 12 h. After annealing, the samples were cooled to room temperature and brought to a final concentration of 50 mM Tris-HCl, pH 8.3, 10 mM MgCI2, 1 mM DTT, 1 mM dNTPs, 0.5 gl RNAsin (Promega), 10 gg/ml of actinomycin D and 40 U of M-MuLV reverse transcriptase (Perkins Elmer Cetus, Norwalk, Conn.). The reactions were incubated at 37 ~ C for 90 rain, and after ethanol precipitation, the samples were analyzed on a 7 M urea, 8% polyacrylamide denaturing gel.

was added to 5 gg/ml. Samples were digested for 1 h at 37 ~ C. After digestion 2% SDS and 4 M fl-mercaptoethanol were added and the sample was boiled for 2 min. Then 0.09 OD26o units of treated nuclei were separated by electrophoresis on a 7.5% SDSpolyacrylamide gel using a mini Protean apparatus from Biorad and electrophoretically transferred to nitrocellulose filters overnight at 100 mA constant current. On completion of the transfer, nitrocellulose sheets were incubated with 3 mg/ml PBS-T (phosphate buffered saline, 0.1% Tween-20) or with 10 mg/ml BSA in PBS-T for 1 h at room temperature to block free protein-binding sites. Filters were incubated for 2 h at room temperature with rabbit anti DBP-5 serum diluted 1:200 or with normal rabbit serum in PBS. Primary antibodies were visualized by incubation with biotinylated anti-rabbit antibodies followed by streptavidin-horseradish peroxidase (Vectastin ABC Kit- Vector, Calif.) before subsequent chemiluminescent development using the ECL protein blotting detection reagents (Amersham, Arlington Heights, Ill.).

Polyclonal antibodies. The c D N A insert from the clone 2 g t l l - 5 was subcloned in frame into the EcoRI site in the prokaryotic expression vector pATH-11 within the trpE gene of the tryptophan operon (Dieckmann and Tzagoloff 1985) and a clone named p A T H 5 15 was isolated. After induction, a 500 ml culture containing p A T H 5-15 was harvested, lysed and fractionated on a 7.5% preparative SDS-polyacrylamide gel. Fusion protein was visualized on the gel using a high concentration of sodium acetate (Ratchrard et al. 1979), and excised. Bands containing approximately 100 gg of the fusion protein were homogenized in 1 ml of complete Freund's adjuvant and used to inject female New Zealand rabbits subcutaneously. At 14 days after the first injection, animals were boosted once with !00 gg of fusion protein homogenized in 1 ml of incomplete Freund's adjuvant and bled after 10 days from the second injection.

Immunofluoreseence. Exponentially growing HeLa cells were fixed for 10 rain at - 2 0 ~ in a 1:1 (vol:vol) solution of methanol/ acetone. The cellular distribution of DBP-5 was assessed by incubating the fixed cells with rabbit anti DBP-5 antibodies diluted 1:200 or normal serum diluted 1:200 for 60 min, followed by a 45 rain incubation with fluorescein conjugated affinity purified F(ab')2 fragment goat anti-rabbit IgG (Cappel) diluted 1:40. Cells were visualized with a Nikon Optiphot microscope using a x 40 Fluor Nikon lens. Fluorescent images were recorded on Kodak Trimax 400 ASA and developed conventionally.

Isolation of cell nuclei and immunoblotting. Suspension cultures of exponentially growing HeLa cells at a concentration of about 4 x 106 cells/ml, were collected and nuclei were isolated as described (Mirkovitch et al. 1984). Then 15 OD26 o units of nuclei were washed four times in isolation buffer (Mirkovitch et al. 1984) containing 0.1% digitonin without E D T A followed by two washes in FSB buffer [100 m M Tris-HC1, pH 6.8, 3 m M MgCI2, 12% glycerol, 1 mM phenylmethylsulfonyl fluoride (PMSF)]. The pellet was resuspended in 1 ml of FSB buffer and DNAse I (Boehringer)

Fig. 1A, B. Binding characteristics of DBP-5. A Fusion proteins from the 2 gt11-5 plaques were screened on nitrocellulose filters with 132/80, X1, X2, and Y probes as described in Materials and methods. B Fusion proteins were prepared from uninduced ( indole acetic acid, - I A A ) and induced ( + I A A ) tryptophan starved bacterial cultures with control or pATH-11 constructs. CH5-S designates the 2 gt11-5 EcoRI insert cloned in the sense orientation; CH5-N is the same insert cloned in the nonsense orientation. The left panel shows the total Coomassie blue stained proteins; the middle panel shows a filter with identical total proteins probed with the 132/80 catenated oligonucleotide (X box); the right panel shows a filter probed with the Y oligonucleotide (Y box)

620

Fig. 2A, B. Analysis of DBP-5 mRNA. A RNA blot: 10 gg of total RNA from the indicated cell lines was loaded in each lane and the filter was probed with the 2 kb EcoRI insert of the pATH 5-15 clone. The left panel shows the stained gel. The molecular weight of DBP-5 was determined by comparison with RNA size markers (0.24-9.5 kb RNA ladder, Bethesda Research Laborato-

ry). B Primer extension analysis of the 5' end of DBP-5 mRNA: an end labeled 30-mer oligonucleotide was hybridized for 12 h at 56~ C to 30 gg of RNA derived from the 721.180 and Raji cell lines and extended with reverse transcriptase, pBR322 digested with MspI was used as molecular weight marker

Results

Expression and tissue distribution of DBP-5

Isolation of 2 gtll clones encoding DNA binding proteins

Because positive trans-acting factors that regulate H L A class I! expression are mutated in certain B cell lines lacking H L A class II antigens (Hume and Lee 1989), we examined RNAs from several of these lines for expression of sequences related to 2 gt11-5. Total RNAs derived from BLS-1 (Hume and Lee 1989), BLS-2 (Hume et al. 1989), 6.1.6 (Gladstone and Pious 1980) and SJO (Bull et al. 1990) ( H L A class II negative mutant B cell lines), and from a control cell line Raji (Burkitt Lymphoma, H L A class II positive), were subjected to R N A blot analysis. An m R N A of approximately 7.5 kb was detected in all of the cell lines (Fig. 2A). Although a weaker band was seen with BLS-2 in the blot, this was not a reproducible result. To investigate the expression of DBP-5 in lines and tissues that do not express H L A class II genes constitutively, both an R N A blotting experiment, and reverse transcription coupled with the polymerase chain reaction (PCR) were performed. RNAs from several normal B cell lines, Jurkat (T cell), H e L a (epithelial) and teratocarcinoma lines were analyzed and all were found to express DBP-5 m R N A at similar levels (data not shown). Thus, DBP-5 appears to be ubiquitously expressed in humans.

In our effort to identify new factors specific for H L A class II promoters, a c D N A library constructed in 2 g t l l by Clontech (Palo Alto, Calif.) from a normal human B cell line was screened with a multimerized, double-stranded D N A probe containing the class II D R A X consensus element and flanking sequences. As a negative control, a similar probe (Y1) containing the Y element was used. Two clones, 2 gtl 1-5 and 2 gtl 1-26, were isolated. Both clones generated fusion proteins that specifically bound the X, but not the Y oligonucleotide probe. Figure 1 A shows that the 2gt11-5 fusion protein binds the - 1 3 2 / 8 0 and X1 probes, which overlap containing sequences from the class II X box and 5' pyrimidine tract, but not the X2 or Y probes, both containing sequences 3' of the X box. In this report, we describe the characterization of 2 gt11-5, which contains a 2 kb c D N A fragment. The other clone, 2 gt11-26, will be described elsewhere (T. Mattioni, manuscript in preparation). DNA/protein blotting analysis of a fusion protein generated by subcloning the insert of 2 gtl 1-5 into another prokaryotic expression vector, pATH-11 (Dieckmann and Tzagoloff 1985), confirmed the finding that a protein produced from this insert recognizes the D R A X box probe (Fig. 1 B). However, since we have been unable to solubilize the protein, it has not been possible to define its precise binding sites more rigorously. The experiments described in the following sections were directed toward general characterization of this gene, DBP-5.

Isolation of cDNA clones corresponding to the complete coding region R N A blotting experiments indicated that DBP-5 m R N A (7.5 kb) is significantly larger than the 2 kb 2 gt11-5 c D N A insert. To obtain longer cDNAs we therefore used

621

>m

m ~

.

.

.

.

xu

~

NN

Fig. 3. Nucleotide sequence and predicted amino acid sequence of DBP-5. The first methionine is boxed. The four types of repeats (A, B, C, D) are underlined. The nucleotide sequence data reported

in this paper have been submitted to the EMBL Data Library and assigned the accession number X63071

622 the 2 gt11-5 cDNA to screen a cDNA library constructed in pCDM8 by Invitrogen (San Diego, Calif.). The library was size-selected for inserts greater than 2 kb and was estimated to contain 2 x 10 6 clones. We isolated two overlapping clones, 5.1 and 5.9, with 2 and 4 kb inserts respectively. The 3' end of clone 5-i overlaps with approximately 1 kb of the 5' end of clone 5-9, such that the two cDNAs represent 5 kb of the DBP-5 mRNA. This region contains a 3,536 bp open reading frame starting with a methionine codon at nucleotide 73 and a TAG stop codon at nucleotide 3609. Although three methionine codons are found within the first 10 codons, we assume that the methionine codon at nucleotide 73 is the translation initiation codon for the following reason. The 5' end of the mRNA coincides approximately with nucleotide 25 of the sequence for DBP-5. This was determined by means of primer extension experiments (Fig. 2B) in which an antisense primer (situated 65-94 bp downstream of the 5' end of clone 5-1) was annealed to total Raji RNA and extended with reverse transcriptase. A 70 bp fragment was produced, indicating that clone 5.1 is essentially complete at its 5' end. We suspect that the excess 25 nucleotides at the 5' end of our cDNA represent an artifact of cloning, and thus the first three possible methionine codons are unlikely to be sites for translation initiation. Since DBP-5 mRNA has a size of 7.5 kb, an additional 2.5 kb of the 3' untranslated sequences must be missing from our cDNA clone. This is consistent with the observation that clone 5-9 contains neither a poly(A) tail nor a polyadenylation signal.

The sequence of DBP-5 reveals repeated motifs Nucleotide and predicted amino acid sequences of DBP5 are shown in Fig. 3. The cDNA potentially encodes a 1,179 amino acid protein that contains several interesting regions: 1. The amino-terminus contains a proline rich region that comprises a motif of 11 amino acids repeated three times (repeat A). 2. Between residues 233 and 264, a stretch of 8 amino acids is repeated four times (repeat B). 3. A highly basic motif of 7 amino acids is repeated six times between residues 833 and 874 (repeat C). 4. The cluster of repeats C is flanked at each end by a new motif of 19 amino acids (repeat D). A computerized scan of the amino acid sequence for basic and acidic regions in the predicted protein sequence revealed the presence of an acidic domain located in the amino-terminal region of the protein (24-350) and a large basic region extending from amino acid 700 to 920. The sequence of DBP-5 contains no previously identified DNA binding motifs such as zinc finger, helixturn-helix or helix-loop-helix domains (Struhl 1989). When we carried out a computer assisted homology search, the large basic region appeared to be homologous to a poorly characterized component of chicken spermatozoa, known as gallin (Nakano etal. 1976).

Fig. 4. Localization of DBP-5 in the cell. Immunofluorescence: exponentiallygrowingHeLAcellswere stainedusing the anti DBP5 antiserum (upper panel) and with a normal rabbit serum (lower panel) However, the majority of residues were arginine residues, leading us to question the significance of this finding. Comparison of the 5 kb sequence with that of the 2 gtl 1-5 clone showed that the 2 kb insert extended from amino acid 161 to amino acid 782. Consequently, the fusion protein generated by pATH 5-15 is truncated at its carboxyl-terminus within the basic region. The fusion protein is, however, capable of binding, suggesting that the complete basic region is not absolutely required. Nevertheless, since the specificity of many of the DNA binding proteins is at least partly dependent on basic regions (Johnson and McKnight 1989), the pATH 5-15 fusion protein may have a truncated basic DNA binding domain resulting in binding specificity that differs from that of the native protein. We have been unable to generate full length fusion proteins for further study owing to persistent rearrangement events in the basic region in several cloning attempts.

Specific antibodies recognizing DBP-5 Using the fusion protein derived from the pATH 5-15 subclone, we generated a high titer rabbit antiserum and

623

Fig. 5. Immunological analysis of DBP-5. For protein blotting assays, two different blocking solutions were utilized : PBS-T containing 3% BSA or PBS-T containing 10% BSA. Proteins were from the following cell lines : (1) HeLa nuclei, stained with normal rabbit serum (blocking solution: PBS-T, 3% BSA); (2) HeLa nuclei, stained with anti DBP-5 antiserum (blocking solution: PBS-T, 3% BSA); (3) HeLa nuclei, stained with anti DBP-5 antiserum (blocking solution: PBS-T, 10% BSA). The higher amount of BSA used in (3) reduced background staining, yet specific staining of the Mr 180,000 and ~ Mr 42,000 bands was still observed [compare with lane (1)]

tested its specificity by immunofluorescence. We observed a punctate pattern localized to the nuclei when we stained interphase HeLa cells. The same pattern was absent when we used a normal serum as a first antibody in a parallel experiment (Fig. 4). Acetone powder from induced cultures of p A T H 5 15 was prepared and used to preadsorb the anti DBP-5 antiserum (Harlow and Lane 1988). As predicted, after preadsorption the antiserum lost its ability to stain interphase nuclei, confirming that the punctate pattern observed was due to the recognition of the native protein corresponding to the cloned cDNA (data not shown). Preadsorption of the antiserum with acetone powder from pATH11 cultures (control) had no effect on staining. We also performed protein blot analysis using anti DBP-5 and normal serum as a negative control (Fig. 5). A specific band at Mr 180,000 and a smaller band at approximately Mr 42,000 were observed. We suspect that the large band represents the native protein and the smaller one (Mr 42,000) may be a specific cleavage product or a degradation product. Alternatively, the smaller band could be the result of cross-reactivity of anti DBP-5 with an unrelated protein.

Discussion

In our attempt to elucidate the mechanisms that control expression of M H C class II genes, we set out to clone

D N A binding proteins potentially involved in the regulation of these genes. We screened a c D N A library constructed in 2 gtl 1 using an oligonucleotide for the M H C class II X consensus sequence and flanking region within the H L A D R A promoter. We isolated and characterized a clone, 2 gtl 1-5. We have been unable to examine the binding specificity of the native protein because we could produce neither the recombinant protein nor the native protein products from human cells in a soluble form. Nevertheless, when expressed as a fusion protein, the product of this cDNA exhibited D N A binding with a preference for sequences from H L A class II promoters over unrelated sequences (T.M., unpublished observations). However, we cannot claim with certainty that DBP-5 is specific to class II promoters because we could not produce the fusion protein in a soluble form. This prevented us from performing analyses that would have allowed us to define essential contacts between the protein and specific nucleotides. Also, D N A sequences derived from the entire cDNA and the 2 gtl 1-5 insert indicated that the prokaryotic fusion protein generated with the 2 gt11-5 insert lacked part of the basic region, as well as considerable amino-terminal sequence, either or both of which could conceivably affect the ability of DBP-5 to recognize and bind stably its target sequence. R N A blotting and PCR experiments showed that the m R N A for DBP-5 is ubiquitous. The size of the unique DBP-5 m R N A is remarkable as is its extensive untranslated region, which makes up just over half of the sequence. There is no evidence for alternative splicing or polyadenylation of this transcript and the cDNA sequence only accounts for 5 kb of the 7.5 kb mRNA. Untranslated regions of R N A have been implicated in the control of m R N A stability and efficiency of translation (Jackson and Standart 1990). Whether either of these functions can be attributed to the 4.5 kb of the non-coding sequence of DBP-5 remains to be determined. Polycistronic mRNAs, although prevalent in some mammalian viruses, have not been described in eukaryotes. Recently, translation initiation in the absence of a cap structure has been reported in an artificial bicistronic m R N A in eukaryotic cells (Macejak and Sarnow 1991). This suggests that the eukaryotic translational machinery is capable of processing such messages and raises the possibility that mRNAs with large apparently untranslated regions, such as DBP-5 mRNA, may, in fact, contain a second open reading frame. A specific antiserum raised against the fusion protein specifically binds an Mr 180,000 protein in protein blots. This product is larger than the molecular weight predicted from the amino acid sequence but, because of the basic region, the protein may migrate anomalously in SDS-polyacrylamide gel electrophoresis. Moreover, posttranslational modifications, such as phosphorylation or glycosylation, could increase the apparent molecular weight, hnmunofluorescent staining provides a more dramatic demonstration of the presence of DBP-5 in human cells and localizes the protein to the nucleus. It is possible that this pattern is the result of the association of DBP-5 with nuclear envelope proteins. Alternatively, the localization of the protein to particular re-

624 gions o f c h r o m a t i n could give rise to p n n c t a t e staining. T h e change f r o m a p u n c t a t e to a m o r e diffuse pattern in cells u n d e r g o i n g mitosis (data n o t shown) could be consistent with either hypothesis. Sequence analysis revealed several interesting features o f the predicted protein structure, including some c o m pletely conserved repeats. N o n e o f the canonical binding motifs represented in other D N A binding transcription factors are present (e.g. zinc finger or helix-loop-helix) (Struhl 1989), but acidic and basic d o m a i n s are evident. The acidic d o m a i n , including the proline-rich region, is located in the a m i n o - t e r m i n u s suggesting a possible activ a t i o n function (Mitchell a n d Tjian 1989). T h e basic region is located in the middle o f the protein and p r o b a b ly includes the D N A binding d o m a i n . A n h o m o l o g y with c-mos, p r o p o s e d some years ago (Berdichevskii et al. 1988) when a smaller segment o f this clone was described, is in o u r o p i n i o n t o o weak to be considered informative. The only h o m o l o g y o f possible relevance is with a p o o r l y characterized structural c o m p o n e n t o f the nuclei o f chicken s p e r m a t o z o a k n o w n as gallin (Nakano et al. 1976). W h e t h e r its role is in the regulation o f transcription or the o r g a n i z a t i o n o f higher order chrom a t i n structures, D B P - 5 is strongly conserved as assessed b y S o u t h e r n blotting with various D N A s . Crosshybridization was detected with species as evolutionarily distant as Drosophila. The final determination o f the function o f this ubiquitous, highly conserved D N A binding protein will require further analysis.

Acknowledgements. We thank Drs. J. DiMartino and S. Vijayasazadhi for advice and critical reading of the manuscript. This work was supported by NIH R29 GM 39698.

References Berdichevskii FB, Chumakov IM, Kiselev LL (1988) Determination of the nucleotide sequence of the son3 fragment of the human genome: identification of a new protein with an unusual structure and homology with DNA-binding proteins. Mol Biol 22: 639-646 Bull M, van Hoef A, Gorski J (1990) Transcription analysis of class II human leukocyte antigen genes from normal and immunodeficient B lymphocytes, using polymerase chain reaction. Mol Cell Biol 10:3792-3796 Dale RM, McClure BA, Houchins JP (1985) A rapid singlestranded cloning strategy for producing a sequential series of overlapping clones for use in DNA sequencing: Application to sequencing the arm mitochondrial 18 S rDNA. Plasmid 13:31-40 Dieckmann CL, Tzagoloff A (1985) Assembly of the mitochondrial membrane system. CBP6, a yeast nuclear gene necessary for synthesis of cytochrome b. J Biol Chem 260:1513 1520 Ehrlich H, Lee JS, Petersen JW, Bugawan T, DeMars R (1986) Molecular analysis of HLA class I and class II antigen loss

mutants reveals a homozygous deletion of the DR, DQ and part of the DP region: Implication for class II gene order. Hum Immunol 16:205 219 Gladstone P, Pious D (1980) Identification of a transacting function regulating HLA-DR expression in a DR-negative B cell variant. Somatic Cell Genet 6:285-298 (Abstract) Harlow E, Lane D (1988) Antibodies. A laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY Hume CR, Lee JS (1989) Congenital immunodeficiencies associated with absence of HLA class II antigens on lymphocytes result from distinct mutations in trans-acting factors. Hum Immunol 26: 288-309 Hume CR, Shookster LA, Collins N, O'Reilly R, Lee JS (1989) Bare Lymphocyte Syndrome: Altered class II expression in two B cell lines. Hum ImmunoI 25:1-11 Jackson RJ, Standart N (1990) Do the poly(A) tail and 3' untranslated region control mRNA translation? Cell 62:15-24 Johnson PF, McKnight SL (1989) Eukaryotic transcriptional regulatory proteins. Annu Rev Biochem 58 : 79%839 Klemsz M J, McKercher SR, Celada A, Van Beveren C, Maki RA (1990) The macrophage and B cell-specific transcription factor PU.1 is related to the ets oncogene. Cell 61 : 113-124 Liou H-C, Boothby MR, Glimcher LH (1988) Distinct cloned class II MHC DNA binding proteins recognize the X box transcription element. Science 242:69-71 Macejak DG, Sarnow P (1991) Internal initiation of translation mediated by the 5' leader of a cellular mRNA. Nature 353:9094 Maniatis T, Fritsch EF, Sambrook J (1982) Molecular cloning. A laboratory manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY Mirkovitch J, Mirault M-E, Laemmli UK (1984) Organization of the higher-order chromatin loop: Specific DNA attachment sites on nuclear scaffold. Cell 39:223-232 Mitchell PJ, Tjian R (1989) Transcriptional regulation in mammalian cells by sequence-specific DNA binding protein. Science 245:371 378 Nakano M, Tobita T, Ando T (1976) Int J Pept Protein Res 8:565578 Ratchrard C, Dahmus H, Dahmus ME (1979) Rapid visualization of protein bands in preparative SDS-polyacrylamide gels. Anal Biochem 93 : 257-260 Reith W, Barras E, Satola S, Kobr M, Reinhart D, Herrero Sanchez C, Mach B (1989) Cloning of the major histocompatibitity complex class II promoter binding protein affected in a hereditary defect in class II gene regulation. Proc Natl Acad Sci USA 86:4200-4204 Seed B, Aruffo A (1987) Molecular cloning of the CD2 antigen, the T-cell erythrocyte receptor, by a rapid immunoselection procedure. Proc Natl Acad Sci USA 84:3365-3369 Singh H, LeBowitz JH, Baldwin AS, Sharp PA (1988) Molecular cloning of an enhancer binding protein: Isolation by screening of an expression library with a recognition site DNA. Cell 52: 415-423 Struhl K (1989) Helix-turn-helix, zinc-finger, and leucine-zipper motifs for eukaryotic transcriptional regulatory proteins. TIBS 14:137-140 Vinson CR, LaMarco KL, Johnson PF, Landschulz WH, McKnight SL (1988) In situ detection of sequence-specific DNA binding activity specified by a recombinant bacteriophage. Genes Dev 2: 801-806