Factors which regulate transcription in immature myeloid cells are of great current interest for the light they may shed upon myeloid differentiation. In the course of screening for transcription factors which interact with the human myeloperoxidase (MPO) promoter we, for the first time, identified and cloned the cDNA and genomic DNA for human HBP1 (HMG-Box containing protein 1), a member of the high mobility group of non-histone chromosomal proteins. HBP1 cDNA was initially cloned from rat brain in 1994, but its presence in human cells or in myeloid tissue had not been described previously. The sequence of human HBP1 cDNA shows 84% overall homology with the rat HBP1 cDNA sequence. We have subsequently cloned the gene, which is present as a single copy, 25 kbp in length. Northern blotting reveals a single 2.6 kb mRNA transcript which is expressed at higher levels in human myeloid and B lymphoid cell lines than in T cell lines tested and is present in several non-myeloid human cell lines. Comparison of the mRNA and genomic sequences reveals the gene to contain 10 exons and 9 introns. The sequence of human HBP1 mRNA contains a single open reading frame, which codes for a protein 514 amino acids in length. The amino acid sequence specified by the coding region shows 95% homology with the rat HBP1 protein. The human protein sequence exhibits a putative DNA-binding domain similar to that seen in rat HBP1 and shows homology with the activation and repressor domains previously demonstrated in the rat protein. We have expressed human HBP1 protein both in vitro and in prokaryotic and eukaryotic cells. The expressed fusion protein binds to a sequence in a functionally important region within the basal human MPO promoter. In transient co-transfection experiments HBP1 enhances MPO promoter activity. Human HBP1 appears to be a novel transcription factor which is likely to play an important role in regulating transcription in developing myeloid cells.
The development of terminally differentiated white blood cells from multipotent precursors is a complex process which remains to be fully elucidated. In order to better understand the mechanisms which control myeloid differentiation, we have been studying the regulation of myeloperoxidase (MPO) transcription in immature hematopoietic cells and cell lines.123 Transcription of the MPO gene begins early during the myeloblast stage of white cell differentiation and ceases at the end of the promyelocyte stage, concomitantly with the cessation of cell proliferation and the onset of terminal differentiation.4 Identification of factors responsible for the maturation-specific activation and repression of MPO transcription should provide clues to the broader mechanisms which govern the maturation of myeloid cells.
As one aspect of our studies of MPO transcription, we recently carried out yeast one-hybrid screens to identify putative regulatory proteins which bind to functional elements within the human MPO promoter. We now report the isolation and cloning of the cDNA and the gene for the human homolog of HMG-Box containing protein 1 (HBP1),5 a recently described member of the high mobility group of non-histone chromosomal proteins.6 We further demonstrate that this protein binds to a specific sequence within the basal human MPO promoter and stimulates MPO promoter activity in co-transfection experiments. Our findings suggest that HBP1 may play an important role in regulating transcription in developing myeloid cells. Preliminary reports of these studies have appeared previously in abstract form789 and the cDNA sequence was submitted to GenBank (Accession number AF019214).
Propagation of cells and preparation of nuclear extracts
The human cell lines HL-60, KG1, and KG1a (myeloblastic), K-562 (erythroleukemic), U-937 (histiocytic), and HeLa (squamous carcinoma) were obtained from the American Type Culture Collection (Rockville, MD, USA). Additional human lymphoblast and myeloblast cell lines were obtained from Dr Harry Findley, Department of Pediatrics, Emory University School of Medicine. All of these cell lines were grown in humidified 5% CO2 at 37°C in RPMI 1640 (Gibco BRL, Gaithersburg, MD, USA) supplemented with 1% L-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 10% heat-inactivated fetal calf serum. HL-60 cells were induced to differentiate by growth for 48 h in 1.25% dimethyl sulfoxide (DMSO), or 10 ng/ml 12-O-tetradecanoylphorbol 13-acetate (TPA). Cells were harvested by centrifugation and washed in phosphate-buffered saline (pH 7.4) before processing. Nuclear extracts of cultured cells or peripheral blood leukocytes were prepared essentially according to the method of Dignam et al,10 as previously described. Protein concentrations of nuclear extracts were determined by measuring absorbance at 280 and 260 nm.
Yeast one-hybrid screening
We previously described and characterized a series of functionally important nuclear protein binding sites (DP1 to DP7) in the promoter region of the human MPO gene.12 To screen for proteins which bind to the DP1 segment of this promoter, three tandem repeats of the DP1 DNA sequence comprising this site were prepared by ligation and inserted upstream from a minimal promoter in a one-hybrid vector pHISi (Matchmaker One-Hybrid System; Clontech, Palo Alto, CA, USA). The resultant pHISi-(DP1)3 plasmid was then linearized with XhoI and integrated into the yeast strain YM4271 (Clontech). Individual colonies were picked and tested for background growth in the presence of 3-aminotriazole. A pGAD10-based bone marrow cell cDNA expression library (Clontech) was amplified in E. coli. This library was then screened on selective media for ability to produce large colonies in yeast cells bearing pHISi(DP1)3. Following purification of the large colonies by rescreening, plasmids were isolated from the yeast cells and purified by standard methods. Plasmid inserts were analyzed by restriction digestion with BglII, followed by agarose gel electrophoresis. The 5′-ends of inserts greater than 500 bp in length were sequenced by a modified Sanger procedure as previously described11 using Sequenase 2.0 (USB, Cleveland, OH, USA) and 35S as label. These sequences were then compared with sequences archived in GenBank. DNA motif searches were performed using the software program MacDNASIS Pro version 3.6 Sequence Analysis Software (Hitachi Software Engineering, San Bruno, CA, USA). The entire cDNA sequence of human HBP1 was then determined by dideoxy terminator sequencing using an automated procedure (Model 373A; Applied Biosystems, Foster City, CA, USA).
Expression and purification of HBP1 protein
Our human HBP1 cDNA clone was excised from a pGAD10 vector and ligated into pET32 vectors A, B, and C (Novagen, Madison, WI, USA) so that the resultant fusion protein could be expressed in all three possible frames. A reticulocyte lysate-based coupled transcription/translation system (TNT Quick-coupled Transcription/Translation System, Promega, Madison, WI, USA) with 35S methionine as label was used to express HBP1 protein. Expression of the HBP1 protein from the HBP1 cDNA/pET32C clone was confirmed by electrophoresis in 4–20% acrylamide gels under denaturing conditions. The expressed protein was labeled by 35S-methionine, allowing it to be detected directly by exposing the dried gel to X-ray film. The HBP1 insert from the pGAD10 vector was also ligated into a mammalian expression vector, pCI (Promega) and assayed in the same coupled transcription/translation system.
To achieve HBP1 expression in E. coli, HBP1/pET32C was transformed into bacterial hosts BL21 (DE3) and AD494 (DE3). Bacterial pellets were lysed and protein purified using the BugBuster protein extraction reagent (Novagen) and His-Bind kit (Novagen), which takes advantage of the His tag in the vector. Expression of HBP1 was demonstrated by acrylamide gel electrophoresis on 4–20% SDS gels followed by Western blotting using the S-Tag HRP Lumiblot kit to detect the S-Tag in the pET32 fusion protein and by partial sequencing of the expressed protein. Initially, only low levels of expression of the HBP1 fusion protein were obtained, suggesting that perhaps HBP1 might be inhibitory to bacterial growth or that bacterial cells might lack a factor (such as a specific tRNA) necessary for expression of this protein. However, expression of HBP1 protein in the BL21 (DE3) host was subsequently optimized. The following procedure yielded maximal HBP1 protein levels: cells were grown in LB medium with 200 μg/ml carbenecillin until an OD600 of 0.6 was reached. The cells were then induced with 100 mM IPTG and grown overnight in LB media containing 500 μg/ml carbenicillin. Bacteria were harvested and washed once with 0.25 culture volume of 20 mM Tris-HCl, pH 8.0 and the pellet stored at −70°C.
To prepare antiserum against HBP1 protein, 60 μg of purified protein (100 kDa band) was recovered from an acrylamide gel, mixed with complete Freund's adjuvant, and injected bilaterally into the base of the tail of Sprague–Dawley rats. A second booster injection of 60 μg HBP1 protein in incomplete Freund's adjuvant was performed 2 weeks following the first injection. Blood was obtained from the tail vein 2 weeks later and serum separated by centrifugation.
Cloning the human HBP1 gene
A 958 bp DNA segment obtained from our previously isolated HBP1 cDNA clone by restriction digestion was used as a probe for screening a pre-made human whole blood genomic library in phage Lambda Dash II (Stratagene, La Jolla, CA, USA) using E. coli XL1-Blue MRA as host strain. Plaques were transferred to Duralose-UV membranes (Stratagene) and after denaturation, neutralization, and washing, the DNA was affixed to the membranes by UV crosslinking. The membranes were hybridized with 32P-labeled probe and positive clones detected by autography. Strongly positive clones were subjected to secondary screening by the same procedure. DNA was purified from several positive clones and the presence of sequences homologous to HBP1 cDNA was verified by Southern hybridization and limited sequencing.
Construction of luciferase reporter plasmids containing normal or mutated segments of the 5′-flanking sequence of the human MPO gene
The published sequence of the 5′-flanking region of the human MPO gene was obtained from GeneBank (February 1999 version plus daily updates, Accession Number D14998) compiled by the US National Institutes of Health and distributed by Hitachi Software Engineering. To prepare luciferase reporter plasmids containing portions of MPO promoter DNA, selected segments of the MPO gene were amplified using human genomic DNA as template. Promoter segments were amplified using an upstream HindIII and a downstream BglII handle, cut with the appropriate restriction enzymes, isolated on agarose gels and inserted into HindIII and BglII sites within the multiple cloning site of pGL2-Basic or pGL3-Basic vectors (Promega). Dideoxy terminator sequencing was used to check the correct sequence of all mutations.
Plasmid pSV-β-galactosidase Control, which was used as an internal control for transfection studies, was obtained from Promega and was grown in E. coli HB101. Beta-galactosidase activity was measured by conversion of substrate ONPG (o-nitrophenyl-β-D-galactopyranoside) to a yellow product which was determined by measuring absorbance at 420 nm.12
Transient transfection by electroporation or using Lipofectamine reagent
Leukemic cell lines were transfected by electroporation as previously described,123 using a Gene Pulser System (BioRad, Richmond, CA, USA) or a PZ200 Progenator II (Hoeffer, San Francisco, CA, USA). Electroporetic transfections were carried out in a volume of 400 μl OPTI-MEM I (Gibco BRL Life Technologies, Gaithersburg, MD, USA) without serum, at 250 V, 960 μF, and Tau of 39 ms using 30–40 μg of plasmid DNA and 0.8 to 1.0 × 107 cells. HeLa cells were transfected using Lipofectamine Reagent (Gibco BRL) using 5 μg plasmid DNA and 106 cells. To normalize for efficiency of transfection, cells were co-transfected with plasmid pSV-β-galactosidase control vector (Promega) (0.5 μg for HeLa cells, 1.0 μg for the electroporetic procedure.12 Cells were harvested 24 to 48 h after transfection, washed in PBS, and lysed in 200 μl lysis buffer (Promega) for 15 min at room temperature. Luciferase was assayed using a luciferase assay kit (Promega), 20 μl cell extract and an ML3000 Microtiter Plate Luminometer (Dynatech Laboratories, Chantilly, VA, USA). Results were routinely expressed as relative light units (RLU) luciferase activity per OD420 β-galactosidase.
Preparation of oligonucleotides
DNA probes for hybridization experiments were labeled by random primer labeling (Stratagene) and purified using Microspin G-25 columns (Amersham Pharmacia Biotech, Arlington Heights, IL, USA). Northern and Southern blotting studies were carried out essentially as described by Ausubel et al,13 using 106 c.p.m./ml 32P-labeled HBP1 cDNA in a final volume of 10 ml.
Complementary oligonucleotides with 5′-overhangs for gel shift experiments were synthesized using a Model 394 DNA/RNA Synthesizer (Applied Biosystems). Annealing was performed by heating these oligonucleotides to 100°C for 2 min in water and allowing them to cool to room temperature. The annealed double-stranded DNA segments were then labeled with α-32P-dCTP (7500 Ci/mmol) by a fill-in reaction with DNA polymerase I, Klenow fragment (New England BioLabs, Beverly MA, USA). After incubation for 1 h at 37°C, the reaction tubes were placed on ice and the incorporated label was purified using Microspin G-25 columns.
Gel shift assays
Gel shift assays were carried out as described by Ausubel et al.13 DNA-protein binding interactions were carried out in 20 μl of 10 mM Tris-Cl, pH 8.0, containing 5 mM MgCl2, 100 mM KCl, 1 mM CaCl2, 2 μg poly dI:dC, 20 000 c.p.m. DNA, and 5 μg nuclear protein extract. The mixture was incubated at 25°C for 15 min, mixed with 4 μl 0.25% bromophenol blue and 30% glycerol, and applied to a 6% native polyacrylamide gel (19:1). Electrophoresis was carried out at 200 V for 20 min in Tris-glycine buffer, and the gel was dried and autoradiographed. Gel shift competition experiments were performed by adding a 100-fold molar excess of the competing oligonucleotide to the mixture containing the labeled oligonucleotide. Supershift studies were done by incubating the nuclear extract with monoclonal antibodies directed against specific transcription factors prior to incubation with the DNA probe.
Isolation of HBP1 cDNA clones
One round of screening using three tandem repeats of the DP1 segment of the human MPO promoter as bait yielded 200 larger colonies of which 120 were studied further. Plasmid DNA containing cDNA inserts was found in 91 colonies and restriction mapping with BglII revealed that 55 of the 91 colonies contained plasmids with cDNA inserts greater than 500 bp in length. The 5′-ends of those 55 clones were sequenced and compared with sequences archived in GenBank. The 5′-terminal sequences of four of the plasmid inserts were homologous to the cDNA of a novel member of the high mobility group proteins (a group of non-histone chromosomal proteins6), HMG-box-containing protein 1 (HBP1), which had been cloned recently from rat brain.5 The entire sequence of human HBP1 cDNA was determined and submitted to GenBank (Accession number AF019214, 2 October 1997) and is shown in Figure 1. This cDNA sequence shows 84% overall homology with the previously described sequence of rat HBP1 cDNA (GenBank Accession Number U09551). The corresponding human mRNA sequence contains a single open reading frame, which codes for a protein composed of 514 amino acids. The amino acid sequence of this protein (also shown in Figure 1) has 91% overall homology with the published sequence of rat HBP1. An HMG-box-like sequence which shows 100% homology with that of the rat HBP1 gene is located between amino acids 431 and 500. In addition, two pocket protein-binding sequences are noted in the human HBP1 protein: an LXCXE site at amino acids 35 to 39, and an IXCXE site at amino acids 323 to 327. Third, a 28 amino acid sequence between amino acids 93 and 120 shows 86% homology with an activation domain previously recognized in the rat HBP1 gene by Lavender et al.14 These observations suggest that the human HBP1 protein may function, in part, to activate other genes through interactions with pocket proteins.
Using the entire HBP1 cDNA as probe, DNA from normal human peripheral blood leukocytes was subjected to restriction digestion and Southern blotting to obtain a preliminary estimate of the size of the HBP1 gene. Genomic DNA was restricted by HindIII, EcoR1, and BamHI, respectively, and the fragments resolved by electrophoresis on 0.65% agarose gels. The results are consistent with the presence of a single copy of HBP1 DNA, and suggest a minimum molecular weight for the human HBP1 gene of at least 15 kb (data not shown).
None of the other 51 clones isolated in our one-hybrid screen was found more than once except for human beta globin cDNA which was found three times, as might be expected, considering the abundance of this cDNA in bone marrow precursors. Only three other known genes were isolated which appear to be candidate transcription factors. These include the human PHL-1 gene (a c-myc oncogene containing a cox III sequence), human transforming growth factor beta-1 receptor, and the gene for human ras-related protein, Rab5. None of these appears likely to be the factor which binds the MPO promoter sequence. (A detailed list of all of these clones may be obtained from the corresponding author of this paper.)
Isolation of HBP1 genomic clones
Using as a probe a 1000 bp double stranded DNA segment derived from our human HBP1 cDNA clone, a human whole blood genomic library was screened to obtain genomic clones of HBP1. Several clones corresponding to various segments of the human HBP1 gene were isolated and identified by limited sequencing. While sequencing of these clones was in progress in our laboratory, we became aware that portions of the DNA sequence of a human BAC clone (RG363E19) submitted to GeneBank from Human Genome Project studies (a collaboration between the NHGRI Chromosome 7 Mapping Project and the Washington University Genome Sequencing Center) showed segments of homology with the HBP1 cDNA sequence which we had previously submitted to GeneBank. This BAC clone (GeneBank accession number AC004492) is derived from human BAC library CITB-HS-A, and contains human DNA from chromosome 7 (7q31.1). By comparing the sequence of our HBP1 cDNA clone with that of the genomic BAC clone we were able to define the intron/exon structure of the human HBP1 gene, which is shown in Figure 2. Limited sequencing of the genomic clones of human HBP1 which we have isolated indicates their homology with the sequence of BAC clone RG363E19. The coding region of the HBP1 gene consists of 10 exons and nine introns. HBP1 protein is encoded by exons 1–9 plus a small segment at the 5′-end of exon 10. Most of exon 10, which is the largest exon, does not code for expressed protein sequences.
As illustrated in Figure 2, the coding sequence of the HBP1 gene is about 23 kb in length rather than as little as 15 kb as suggested by our Southern blotting experiments. The reason for this discrepancy is explained by analysis of the HindIII, EcoR1, and BamH1 restriction sites present within the gene sequence. Several of the larger bands seen on our Southern blot actually consist of two or more restriction fragments of nearly identical length, while a few very small restriction fragments are not visualized on the blot, because they consist of sequences not found within the cDNA and hence not recognized by the cDNA probe employed.
Expression of HBP1 RNA in hematopoietic cells and cell lines
Northern blotting studies using HBP1 cDNA as probe were carried out to determine the relative expression of HBP1 mRNA in normal peripheral blood leukocytes and hematopoietic cell lines of different lineages. Levels of HBP1 mRNA expression in several myeloid cell lines are compared in Figure 3a, while Figure 3b illustrates HBP1 mRNA expression in several cell lines of B lymphoid and T lymphoid lineage, as well as peripheral blood leukocytes and HeLa cells. Most hematopoietic cell lines showed a single HBP1 mRNA transcript, migrating at a position equivalent to a molecular weight of approximately 2.6 kb, in agreement with the size of our HBP1 cDNA clone. However, peripheral blood leukocytes and occasional hematologic cell lines showed a doublet band, suggesting possible processing of the transcript. The intensity of HBP1 mRNA expression varied from one cell line to another but in all cases was less than that of the glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mRNA used as control. There was no systematic difference in level of HBP1 mRNA expression between myeloid, B lymphoid and T lymphoid cell lines. However, levels of HBP1 expression differed substantially among different cell lines of the same lineage, suggesting that the extent of HBP1 expression might be related to the degree of maturation of the individual cell lines. In accordance with this hypothesis, induction of differentiation of HL-60 cells by treatment with 12-O-tetradecanoylphorbol 13-acetate (TPA) brought about an initial rapid decrease in HBP1 mRNA expression to almost undetectable levels by 3 h, with a return to uninduced levels by 16–24 h, followed by a dramatic (20-fold) increase in RNA levels from 48 to 96 h after induction (Figure 4). During this period the level of the mRNA for glyceraldehyde-3-phosphate dehydrogenase remained relatively constant, suggesting that the changes in HBP1 mRNA levels were not due to gross changes in total mRNA concentration within these cells. By comparison, myeloperoxidase (MPO) mRNA levels decreased markedly within 16 h following TPA treatment and remained low thereafter (Figure 4). Note that the decline in HBP1 mRNA level precedes the drop in MPO mRNA. However, the secondary rise in HBP1 RNA levels occurs after the decline in MPO RNA levels, suggesting that the late rise in HBP1 levels does not play a major role in the cessation of MPO transcription.
Expression and purification of HBP1 protein
In order to obtain purified HBP1 protein for studying its biological and biochemical properties, several approaches were employed. First, HBP1 was expressed in vitro in a reticulocyte lysate-based coupled transcription/translation system. The entire HBP1 cDNA was cloned into bacterial expression vectors pET32A, pET32B, and pET32C (see Experimental Procedures). The resulting construct with the insert in the correct frame should code for a fusion protein containing 514 amino acids of the human HBP1 protein. Using the transcription/translation system the parent vector Pet32C yielded an 18 kDa 35S-labeled protein (Figure 5a). The plasmids HBP1/Pet32A and HBP1/Pet32B, which contained the HBP1 sequences in incorrect frames for translation, produced prematurely terminated proteins slightly smaller than 18 kDa. However, the plasmid HBP1/pET32C yielded a predominant band representing a 100 kDa 35S-labeled fusion protein. The same double-stranded HBP1 DNA segment was also cloned into a mammalian expression vector, pCI (Promega). In the same transcription/translation system the mammalian expression vector containing the HBP1 insert produced a 90 kDa 35S-labeled protein, whereas the parent vector PCI produced no discernable protein product (Figure 5a). The reason for the larger size of the protein produced with the bacterial vector HBP1/pET32C is that this construct contains several marker (Tag) sequences located N-terminally to the HBP1 insert, whereas the pCI construct lacks these tags.
To obtain a larger amount of HBP1 protein for generation of antisera and other purposes, plasmid HBP1/pET32C was transformed into E. coli strains BL21 (DE3) and AD494(DE3). SDS gel electrophoresis and Western blotting of the bacterial extracts revealed HBP1 expression in cells transformed by the HBP1/pET32C construct, but not in cells transformed by the parent plasmid pET32C which lacked the HBP1 insert. Figure 5b illustrates a Coomassie blue-stained SDS gel showing proteins from crude and purified protein extracts of the transformed cells. The two major high molecular weight bands (arrows) seen in the gel of bacterial extracts purified with the His-Bind kit were identified as HBP1 fusion proteins by Western blotting using the S-tag (data not shown). The largest band, a 100 kDa protein, proved to be insoluble in aqueous buffer, while the more abundant 50 kDa protein, was soluble. The 50 kDa protein is presumably an early termination product of the 100 kDa band. Several lower molecular weight bands are also noted; these may be cleavage products of the intact fusion protein or may represent early termination of the transcription/translation reaction in E. coli.
HBP1 stimulates MPO promoter activity
Since we had isolated HBP1 cDNA clones on the basis of the interaction of HBP1 protein with a segment of MPO P1 promoter DNA in the yeast one-hybrid system, we wished to confirm the interaction of this protein with MPO promoter DNA in leukemic cells and to determine its physiologic significance. We first carried out co-transfection experiments to determine whether HBP1 expression affected MPO promoter activity and to verify that HBP1 interacted specifically with the DP1 segment of the MPO promoter. HL-60 cells were co-transfected with: (1) a pGL-3 series reporter plasmid (pGL3–178) containing a wild-type 189 bp MPO P1 promoter (bp −178 to +11) or one mutated in any of the four protein binding sites (DP1, DP2, DP3, or DP4) which we previously have identified within this basal promoter;2 and (2) a pCI vector containing or lacking an expressible HBP1 cDNA insert. The base pair positions of the four MPO promoter mutations are: DP1 (bp −30 to +7), DP2 (bp −47 to −33), DP3 (bp −100 to −83), and DP4 (bp −160 to −137). Mutation of any of the four major elements within the PI promoter (DP1, DP2, DP3, or DP4) produced a slight reduction in baseline promoter activity, observable in cells co-transfected with the pCI vector which lacks HBP1 expression (Figure 6). (These mutations produce more substantial reductions in MPO promoter activity when a 500 bp promoter segment is used,2 but show only a modest effect with the basal 189 bp MPO promoter.) On the other hand, co-transfection of HL-60 cells with plasmid pCI-HBP1 (which contains the expressible HBP1 cDNA insert) stimulated wild-type MPO promoter P1 activity by approximately three-fold. Mutation of the DP1 segment of the MPO promoter abolished the stimulation produced by co-transfection with pCI-HBP1. However, mutation of other segments of the MPO P1 promoter (DP2, DP3, or DP4) reduced but did not abolish the stimulation due to HBP1 expression. These results suggest that HBP1 may stimulate MPO promoter activity by interacting with the DP1 region of the P1 MPO promoter.
We have previously described a second MPO promoter (P3) located about 1 kb upstream from the P1 promoter, which shows significant activity in vitro3 but which appears to be physiologically less important than the P1 promoter (manuscript submitted for publication). As Figure 6 also illustrates, HBP1 expression appears to have no significant effect upon the activity of MPO promoter P3.
HBP1 protein binds specifically to the DP1 segment of the MPO promoter
As a first step toward verifying that HBP1 protein does, in fact, bind to the MPO promoter in myeloid cells, leukemic cell lines were transfected with an HBP1-expressing plasmid or the parent control vector lacking HBP1 sequences, and gel shift experiments conducted using protein extracts prepared from these transfected cells and a probe consisting of the DP1 segment of the MPO P1 promoter. Figure 7a shows a representative gel shift experiment comparing nuclear and cytoplasmic extracts prepared from K-562 cells transfected with the mammalian expression vector pCI (which lacks HBP1 sequences) and extracts prepared from K-562 cells transfected with the construct pCI-HBP1, which codes for the expression of HBP1 protein. As this figure illustrates, nuclear extracts prepared from K-562 cells transfected with the parent vector pCI contain a protein (or protein complex) which binds to and retards an oligonucleotide corresponding to the DP1 segment of the MPO promoter, whereas cytoplasmic extracts do not. The same results are seen using non-transfected K-562 cells (data not shown), indicating that this binding is due to endogenous proteins present in K-562 cell extracts. However, both nuclear and cytoplasmic extracts prepared from K-562 cells transfected with plasmid pCI-HBP1, which expresses HBP1 protein, show increased intensity of the principal retarded band, suggesting that HBP1 binds to this oligonucleotide. Furthermore, this binding is competable by a DP1 oligonucleotide competitor, but not by oligonucleotides corresponding to other segments of the MPO promoter (shown are oligonucleotides specific for an enhancer region which we term rIV and an AML1 binding site).
HBP1 also binds to nuclear and cytoplasmic extracts of HL-60 cells and the extent of this binding activity changes following induction of maturation of the cells by treatment with TPA. Figure 7b illustrates a gel shift experiment showing the transient reduction in HL-60 nuclear protein binding to a DP1 oligonucleotide following TPA treatment.
While such experiments provide preliminary evidence that HBP1 is one of the proteins which bind to the DP1 segment of the MPO promoter, their interpretation is complicated by the presence of endogenous HBP1 expression in all myeloid cell lines so far examined and by the probable binding of other regulatory proteins to this segment of the MPO promoter. Therefore, we next carried out gel shift experments replacing the crude cell extracts by purified HBP1 fusion protein prepared from E. coli cells transformed by plasmid HBP1/pET32C, or protein extracted from E. coli transformed by the control plasmid pET32C, which lacks HBP1 sequences. Figure 8a shows that a DP1 oligonucleotide probe is retarded by HBP1 protein but not by proteins prepared from E. coli transformed by the parent plasmid. Furthermore, preincubation of the purified extract of HBP1/pET32C with an excess of the homologous probe essentially eliminated the retarded band, whereas preincubation with probes corresponding to another segment of the MPO promoter (DP5) or a segment of the MPO upstream enhancer (F1) did not. These results provide strong evidence that HBP1 protein does bind specifically to the DP1 segment of the MPO promoter.
To begin to define the exact sequence within the DP1 segment which is recognized by HBP1 protein, preliminary mutational analysis of the 37 bp DP1 probe was carried out. Three mutants were prepared: mutA, in which bp 1–12 at the upstream end of the DP1 sequence were mutated; mutB, in which bp 13–26 in the middle of the DP1 sequence were mutated; and mutC, in which the downstream 13 bp of the DP1 sequence were mutated. The sequence of wild-type DP1 and the three mutants (mutA, mutB, and mutC) are shown in Figure 8b. A Mac DNASIS transcription motif search (Hitachi Software Engineering) revealed that no recognized binding sites were created by any of these mutations. Wild-type DP1 probe and each of the mutants was used as competitor in a gel shift competition experiment using purified HBP1 protein and labeled DP1 as probe (Figure 8b). The specific retarded band was competed substantially by an excess of unlabeled DP1, mutA, or mut C, but was not competed by mutB. This suggests that the sequence GCAGATTGAGCT (the central portion of DP1), which was altered in mutB, is critical for the binding of HBP1 to the DP1 segment of the MPO promoter.
These results provide the first comprehensive description of the structure of the human HBP1 gene, its cDNA, and the protein which it encodes. HBP1 protein appears to be the human homologue of a transcription factor recently described in rat5 and mouse1415 cells which, on the basis of its DNA binding domain, appears to be a member of the HMG-box-containing transcription factor family.616 These nuclear proteins, also known as high mobility group (HMG) proteins constitute a discrete and well-defined subclass of non-histone chromosomal proteins, some of which are transcription factors.161718 The HMG proteins are widely distributed among eukaryotic cells and have diverse functions including regulation of differentiation and mating type in yeast1920 and control of tissue-specific gene expression in mammals.2122
The human HBP1 gene is fairly large (about 26 kb), coding for 10 exons and nine introns. Exons 1 to 9 and a small portion of exon 10 code for the 514 amino acid HPB1 protein. The role, if any, of the remainder of exon 10 remains to be determined. The 5′-untranslated region of the HBP1 gene contains several putative cis-elements, and our preliminary results suggest that this part of the gene contains most, if not all, of the sequences which regulate the level of HBP1 mRNA expression (data not shown). Our data indicate the existence of only a single transcription start site and most cell lines show only a single mRNA band on Northern blots. However, the presence of a second, lower molecular weight band in blots from peripheral blood leukocytes and from occasional cell lines raises the possibility of post-transcriptional processing of the RNA in some cell types.
Comparison of the amino acid sequence of human HBP1 protein with that of the rat HBP1 protein reveals several interesting similarities. First, the HMG-Box sequence at the 3′-end of the human HBP1 molecule shows complete homology with the HMG-box sequence of the rat HBP1 protein, suggesting that the DNA binding sites of the human and rat proteins may be similar. Second, the human protein, like the rat protein, has both an LXCXE site and an IXCXE site. These sequences are well recognized as pocket protein binding sites, suggesting that the human protein, like its rat counterpart,1415 may interact functionally with pocket proteins such as Rb and p130. Third, there is a 28 amino acid sequence in the human HBP1 protein which shows 86% homology to an activation domain identified in the rat HBP1 gene by Lavender et al.14 Both rat and mouse proteins also contain inhibitor domains1415 and homologous sequences are present in the human protein. In the rat these inhibitor domains appear to flank the activation domain.14 However, the amino acids responsible for this inhibitory activity have not been accurately described in any species, precluding their exact comparison with sequences in the human gene. Nonetheless, the overall homology of the human and rodent proteins suggests that the human gene may contain similar functional elements to those present in the rodent genes.
The nucleotide sequence of the HMG-box of human HBP1 shows a high degree of sequence homology to that of several other HMG proteins such as IREABP, SRY, Mata1, TCF1, and Mc1.5 However, outside the HMG-box (and nuclear localization signals, see below) there is little sequence homology between HBP1 and these other HMG group proteins51415 suggesting that the functions, protein interactions and sites of action of HBP1 and the other proteins are likely to differ substantially.
Previous studies of the rodent HPB1 genes51415 suggested that HBP1 is widely expressed among different cell types. For example, the rat HBP1 gene has been shown to be active in brain, and expression of both rat and mouse HBP1 genes is upregulated during differentiation of fat and muscle cells.51415 Likewise, we found HBP1 RNA expression in normal peripheral blood leukocytes and in a wide variety of immature human hematopoietic cell lines. Our blotting studies showed relatively low levels of HBP1 RNA in all hematopoietic cells examined, compared with levels of the control glyceraldehyde-3-phosphate dehydrogenase (GAPDH) RNA. However, levels of HBP1 RNA were higher in immature myeloid and B lymphocyte lines than in T lymphoblasts, suggesting some variation in expression between different cell lineages. Lesage et al5 reported that the level of rat HBP1 RNA rises during maturation of immature muscle and fat cells. Interestingly, we found that induction of maturation of human myeloid cell lines such as HL-60 or K-562 by phorbol esters resulted in an initial rapid decline in the level of human HBP1 RNA, followed by a marked (20-fold) increase. The fact that expression of this putative transcription factor is modulated both up and down during myeloid differentiation suggests that HBP1 may be actively involved in the differentiation process in leukocytes, as has been suggested for developing muscle and fat cells.23
Our success in expressing HBP1 fusion proteins both in E. coli and in a coupled transcription/translation system has allowed us to generate polyclonal antiserum against HBP1 protein by injecting rats with these proteins. We plan to use this antiserum to further explore the tissue distribution of HBP1 protein in various human tissues and to study in detail the levels of HBP1 protein present at various stages of myeloid differentiation. This should provide additional clues that may guide studies of the functional roles of this factor in hematopoiesis.
Little is yet known about the specific functional roles of human HBP1 in any tissue, while the role of HPB1 in myeloid cells is totally unknown. However, evidence that HBP1 mRNA (and presumably HBP1 protein) is expressed, at least at some level, in most, if not all tissues, suggests that this putative transcription factor may be functionally important in a wide variety of cell types. HBP1 was originally isolated from rat brain on the basis of its ability to suppress the potassium transport defective phenotype of mutant yeast cells. Furthermore, the observation of Lesage et al5 that rodent HBP1 gene expression is upregulated during differentiation of adipocytes and muscle cells, and our finding that human HBP1 mRNA levels are modulated up and down during induced differentiation of myeloid cells, suggest a general role of HBP1 in the control of cell growth and maturation. However, much more remains to be learned about the specific interactions of HBP1 with other cellular growth regulators. Lavender et al14 showed that rodent HBP1 has enhancer activity for an exogenous promoter and that this activity is repressed by interaction of HBP1 with pocket proteins Rb, p130 or p107. On the other hand, Tevosian, Yee, and coworkers have provided evidence that HBP1 can act as a transcriptional repressor of N-myc, cyclin D1, and p21, and may contribute to cell cycle arrest at the onset of differentiation.15232425 Yee et al suggested that changes in the ratio of Rb to HBP1 in immature muscle cells may help determine whether or not cell cycle arrest is followed by differentiation.24
Analysis of the sequence of HBP1 protein reveals the presence of two nuclear localization signal (NLS) motifs (435-KRPMNAFMLFAKKYRVE-451 and 503-KRKR-506). These sequences are highly homologous with corresponding NLS motifs in other HMG domain proteins2627 On the other hand, our data suggest that HBP1 is found in both nucleus and cytoplasm. HBP1 appears similar in this regard to other HMG-group proteins such as HMG-1/-2, which have been shown to be present in both the nucleus and cytoplasm of Chinese hamster lung, rat liver, and bovine trachea cells.26 The possible significance of the shuttling of HMG protein between nucleus and cytoplasm is not yet clear. However, it has been speculated that the cytoplasm might act as a storage site for these proteins before they are transferred to the nucleus for a specific function, or that these proteins might play some role in the transport of cellular components across the nuclear membrane.6
Comparison of the sequence of the DP1 site of the MPO gene with binding sites for rat HBP1 in the rat N-myc gene reveal partial homology between these sites. The central portion and adjacent downstream segment of the DP1 site contain the sequence AGCTAAGA. By comparison, three binding sites for rat HBP1 in the rat N-myc gene23 contain the sequences GGTAAGA, GGGTAAAA, and GGTAACT. The binding site in the MPO promoter shares about 50–75% homology with each of these sites. Likewise, the binding site for human HBP1 in the human cyclin D1 promoter (GGGTAAGA)23 shares six of eight base pairs with the corresponding site in the human MPO gene. On the other hand Zhuma et al27 reported an entirely different sequence (TTCATTCATTCA) for a binding site for HBP1 which they identified in the locus control region of the human CD2 gene. The significance of these different binding sequences is unclear at present. The binding site for HBP1 in the p21WAF1/CIP1 promoter lies between bp −119 and +16 of that promoter but has not been accurately characterized, to our knowledge. However, comparison of that sequence with the consensus HBP1 biding site of Yee et al23 or with the HBP1 binding site in the human MPO gene reveals no significant homology.
Our data indicating that HBP1 binds to the myeloperoxidase promoter and enhances MPO promoter activity, suggest that HBP1 may play an additional role in enhancing tissue-specific gene expression in proliferating myeloid cells prior to the onset of terminal differentiation. The fact that HBP1 mRNA levels decline early after TPA treatment, prior to the drop in MPO RNA synthesis is consistent with this idea. Our observation that protein binding to the DP1 site of the MPO promoter showed only a transient reduction after TPA treatment suggests that HBP1 may help to trigger the drop in MPO RNA transcription after induction, but that other factors become more important at later stages in the process. Further studies designed to explore the effect of HBP1 on expression of other genes which are active in immature myeloid cells will be needed to assess the hypothesis that HBP1 helps to regulate early gene expression in developing myeloid cells.
It is of interest that human HBP1 appears to stimulate rather than inhibit MPO transcription. As mentioned above, other workers have found that rodent HBP1 can exhibit either enhancer14 or repressor15232425 activities. The presence of both activation and inhibitor domains within the HBP1 protein sequence (Refs 14,15 and our data), suggests that under different circumstances HBP1 should be capable of either enhancing or repressing the activity of various promoters.
How might HBP1 exert its enhancing or inhibiting effects upon expression of target genes? It has been suggested that several of the most extensively studied members of the HMG family of proteins, such as LEF1 and HMG-1, may function as ‘architectural’ elements involved in the assembly of multiprotein complexes at the site of enhancers or promoters.161718 Likewise, the HMG protein YY1 is thought to produce a bend in the promoter DNA of the c-fos gene, thereby modulating the binding of transcription factor TFIID to the nearby TATA box.28 Recently, Zhuma et al,27 on the basis of studies of the effect of HBP1 on position effect variegation in transgenic mice, have proposed that this protein may promote chromatin opening and remodeling activities by binding to and bending the DNA, thus allowing DNA–protein and/or protein–protein interactions which increase the probability of establishing an active locus.
A potentially important issue concerns the possibility that HBP1 may act as an oncogene or a tumor suppressor gene, and the question of whether mutation of this gene could be involved in the genesis of leukemias or other types of human cancer. In support of a possible oncogenic role, Lavender et al14 showed that overexpression of rodent HBP1 could bring about morphologic transformation of NIH3T3 cells. On the other hand, Tevosian et al15 showed that HBP1 could cause cell cycle arrest in C2 muscle cells, which might appear more consistent with a role as a tumor suppressor. Intriguingly, the chromosomal location of the HBP1 gene, the long arm of chromosome 7 at 7q31.1, is a site which is frequently deleted in many types of cancers, including malignant myeloid disorders.29303132333435 Ten percent of patients with either myelodysplastic syndrome or de novo acute myelogenous leukemia exhibit a loss of the entire chromosome 7 or a deletion of the long arm of that chromosome.29 The highest frequency of abnormalities of chromosome 7 is found in therapy-related myelodysplastic syndromes or therapy-related acute leukemias.2930 Leukemias occurring in patients with predisposing conditions such as Fanconi anemia, congenital neutropenia, and neurofibromatosis 1 also frequently exhibit deletions of 7q.2930 In addition, recent studies in several laboratories have demonstrated frequent loss of heterozygosity of 7q31.1 in prostate, breast, squamous, colon, and ovarian carcinomas.3132333435 Furthermore, Koike et al29 recently reported that the segments of chromosome 7 which were most commonly deleted in their series of acute myelogenous leukemias were 7q31.1 and 7q33–34, suggesting that tumor suppressor genes might exist at these sites. The genes already described at this site do not include any known tumor suppressor genes.29 On the basis of the evidence that the HBP1 gene has been localized to 7q31.1, that this gene appears to be involved in cell cycle arrest in muscle tissue, and that overexpression of HBP1 can morphologically transform NIH3T3 cells, it appears possible that abnormal expression of this gene could play a role in human neoplasia. Studies to look for mutations of the HBP1 gene in human acute leukemias appear justified.
In summary, the evidence presented in this paper adds to other available information suggesting that HBP1 is a novel transcription factor, and that it may play a significant role in regulating gene expression in developing myeloid cells.
Austin GE, Zhao W-G, Zhang W, Austin ED, Findley HW, Murtagh JJ Jr . Identification and characterization of the human myeloperoxidase promoter Leukemia 1995 7: 848–857
Zhao W-G, Regmi A, Austin ED, Braun JE, Racine M, Austin GE . Cis-elements in the promoter region of the human myeloperoxidase gene Leukemia 1996 10: 1089–1103
Zhao W-G, Regmi A, Lu J-P, Austin GE . Identification and functional analysis of multiple murine myeloperoxidase (MPO) promoters and comparison with the human MPO promoter region Leukemia 1997 11: 97–105
Austin GE, Chan WC, Zhao W-G, Racine M . Myeloperoxidase gene expression in normal granulopoiesis and acute leukemias Leuk Lymphoma 1994 15: 209–226
Lesage F, Hugnot J-P, Amri E-Z, Grimaldi P, Barhanin J, Lazdunski M . Expression cloning in K+ transport defective yeast and distribution of HBP1, a new putative HMG transcriptional regulator Nucleic Acids Res 1994 22: 3685–3688
Landsman D, Bustin M . A signature for the HMG-1 box DNA-binding proteins Bioessays 1993 15: 539–546
Austin GE, Zhao W-G, Lu J-P . Human HBP1, a new high mobility group protein, enhances the activity of the human MPO promoter FASEB J 1998 12: A131
Austin GA, Lu J-P, Zhao W-G . HBP1, a novel high mobility group protein enhances myeloperoxidase promoter activity in developing myeloid cells Blood 1998 92: 192a
Lin KM, Bhatnaggar J, Austin GE . Expression and purification of human HBP1 protein and cloning the HBP1 gene Blood 1999 94: 141b
Dignam JD, Lebovitz RM, Roeder RG . Accurate transcription initiation by RNA polymerase II in a soluble extract from isolated mammalian nuclei Nucleic Acids Res 1983 11: 1475–1489
Austin GE, Lam L, Zaki SR, Chan WC, Hodge T, Hou J, Swan D, Zhang W, Racine M, Whitsett C . Sequence comparison of putative regulatory DNA of the 5′-flanking region of the myeloperoxidase gene in normal and leukemic bone marrow cells Leukemia 1993 7: 1445–1450
Lim K, Chae CB . A simple assay for DNA transfection by incubation of the cells in culture dishes with substrates for β-galactosidase Biotechniques 1989 7: 576–579
Ausubel FM, Brent R, Kingston RE . Current Protocols in Molecular Biology Greene: New York 1988
Lavender P, Vandel L, Bannister J, Kouzarides T . The HMG-box transcription factor HBP1 is targeted by the pocket proteins and E1A Oncogene 1997 14: 2721–2728
Tevosian SG, Shih HH, Mendelson KG, Sheppard K-A, Paulson KE, Yee AS . HBP1: a HMG box transcriptional repressor that is targeted by the retinoblastoma family Genes Dev 1997 11: 383–396
Grosschedl R . Higher-order nucleoprotein complexes in transcription: analogies with site-specific recombination Curr Opinion Cell Biol 1995 7: 362–370
Tjian R, Maniatis T . Transcriptional activation: a complex puzzle with few easy pieces Cell 1994 77: 5–8
Wolffe AP . Architectural transcription factors Science 1994 264: 1100–1101
Arie T, Christiansen SK, Yoder OC, Turgeon BG . Efficient cloning of ascomycete mating type genes by PCR amplification of the conserved MAT HMG Box Fungal Genet Biol 1996 21: 118–130
Sugimoto A, Lino Y, Maeda T, Watanabe Y, Yamamota M . Schizosaccharomyces pombe ste11+ encodes a transcription factor with an HMG motif that is a critical regulator of sexual development Genes Dev 1991 11: 1990–1999
Van de Wetering M, Oostewegel M, Dooijes D, Clevers H . Identification and cloning of TCF-1, a T lymphocyte-specific transcription factor containing a sequence-specific HMG box EMBO J 1991 10: 123–132
Alexander-Bridges M, Dugast I, Ercolani L, Giere L, Nasrin N . Multiple insulin-responsive elements regulate transcription of the GAPDH gene Adv Enzyme Regul 1992 32: 149–159
Yee AS, Shih HH, Tevosian SG . New perspectives on retinoblastoma family functions in differentiation Frontiers Bioscience 1998 3: 532–547
Shih HH, Tevosian SG, Yee AS . Regulation of differentiation of HBP1, a target of the retinoblastoma protein Mol Cell Biol 1998 18: 4732–4743
Gartel AL, Goufman E, Tevosian SG, Shih H, Yee AY, Tyner AL . Activation and repression of p21WAF1/CIP1 transcription by RB binding proteins Oncogene 1998 17: 3463–3469
Sudbeck P, Scherer G . Nuclear localization signals are present in the A-binding high-mobility group domains of SRY and SOX9 J Biol Chem 1997 272: 27848–27852
Zhuma T, Tyrrell R, Sekkali B, Skavdis G, Saveliev A, Tolaini M, Roderick K, Norton T, Smerdon S, Sedgwick S, Festenstein R, Kioussis D . Human HMG box transcription factor HBP1: a role in hCD2 LCR function EMBO J 1999 18: 6396–6406
Natesan S, Gilman MZ . DNA bending and orientation-dependent function of YY1 in the c-fos promoter Genes Dev 1993 7: 2497–2509
Koike M, Tasaka T, Spira S, Tsuruoka N, Koeffler HP . Allelotyping of acute myelogenous leukemia: loss of heterozygosity at 7q31.1 (D7S486) and q33–34 (D7S498, D7S505) Leukemia Res 1999 23: 307–310
Fischer K, Fröhling S, Scherer SW, McAllister Brown J, Scholl C, Stilgenbauer S, Tsui LC, Lichter P Dohner H . Molecular cytogenetic delineation of deletions and translocations involving chromosome band 7q22 in myeloid leukemia Blood 1997 87: 2036–2041
Zenklusen JC, Thompson JC, Troncoso P, Kagan J, Conti CJ . Loss of heterozygosity in human primary prostate carcinomas: a possible tumor suppressor gene at 7q31.1 Cancer Res 1994 54: 6370–6373
Zenklusen JC, Bièche I, Lidereau R, Conti CJ . (C-A)n microsatellite repeat D7S522 is the most commonly deleted region in human primary breast cancer Proc Natl Acad Sci USA 1994 91: 12155–12158
Zenklusen JC, Thompson JC, Klein-Szanto JP, Conti CJ . Frequent loss of heterozygosity in human primary squamous cell and colon carcinomas at 7q 31.1. Evidence for a broad range of tumor suppressor genes Cancer Res 1995 55: 1347–1350
Zenklusen JC, Weitzel JN, Ball HG, Conti CJ . Allelic loss at 7q31.1 in human primary ovarian carcinomas suggests the existence of a tumor suppressor gene Oncogene 1995 11: 359–363
Koike M, Takeuchi S, Yokota J, Park S, Hatta Y, Miller CW . Frequent loss of heterozygosity in the region of the D7S523 locus in advanced ovarian cancer Genes Chromosome Cancer 1977 19: 1–5
This work was supported by a VA Merit Award to GEA.
About this article
Biochemical and Biophysical Research Communications (2019)
Matrix metalloproteinase-13 is a target gene of high-mobility group box-containing protein 1 in modulating oral cancer cell invasion
Journal of Cellular Physiology (2019)
The HMG box transcription factor HBP1: a cell cycle inhibitor at the crossroads of cancer signaling pathways
Cellular and Molecular Life Sciences (2019)
AllTrans-Retinoic Acid Mediates MED28/HMG Box-Containing Protein 1 (HBP1)/β-Catenin Signaling in Human Colorectal Cancer Cells
Journal of Cellular Physiology (2016)
The expression of the tumour suppressor HBP1 is down-regulated by growth factors via the PI3K/PKB/FOXO pathway
Biochemical Journal (2014)