- Email: [email protected]
Distinct functions of two isoforms of a homeobox gene, BP1 and DLX7, in the regulation of the β-globin gene
Gene 278 (2001) 131–139 www.elsevier.com/locate/gene
Distinct functions of two isoforms of a homeobox gene, BP1 and DLX7, in the regulation of the b-globin gene Sidong Fu a, Holly Stevenson a, Jeff W. Strovel b, Susanne B. Haga c, Judy Stamberg b, Khanh Do a, Patricia E. Berg a,* a
Department of Biochemistry and Molecular Biology, The George Washington University Medical Center, Washington, DC, USA b Department of Pediatrics, Division of Genetics, University of Maryland School of Medicine, Baltimore, MD, USA c National Institute of Health, Office of Biotechnology Activities, Bethesda, MD, USA Received 26 March 2001; received in revised form 4 September 2001; accepted 10 September 2001 Received by E. Boncinelli
Abstract Homeotic proteins are transcription factors that regulate the expression of multiple genes involved in development and differentiation. We previously isolated a cDNA encoding such a protein from the human leukemia cell line K562, termed Beta Protein 1 (BP1), which is involved in negative regulation of the human b-globin gene. Sequence comparison revealed that BP1 is a member of the distal-less (DLX) family of homeobox genes and that it shares its homeodomain and 3 0 sequences with another DLX cDNA, DLX7. BP1 and DLX7 exhibit unique 5 0 regions, diverging at nucleotide 565 of BP1. We mapped this new distal-less family member BP1 to chromosome 17q21-22 by FISH and PCR, which is the same locus to which DLX7 has been mapped. These results strongly suggest that BP1 and DLX7 are isoforms (derived from the same gene). Since our previous data demonstrated that BP1 and DLX7 are frequently co-expressed, we determined whether DLX7 is also involved in the negative regulation of the b-globin gene. Mobility shift assays demonstrated that both BP1 and DLX7 proteins, synthesized in vitro, bind to the same BP1 binding site. However, using transient assays, we showed that although BP1 represses activity of a reporter gene through either of two silencer DNA sequences upstream of the b-globin gene, DLX7 did not show repressor activity against the b-globin promoter. Further characterization of these apparent isoforms is of significance since they are jointly expressed in acute myeloid leukemia and in many leukemia cell lines. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Hemoglobin; Homeodomain; Distal-less, Mapping
1. Introduction Homeobox genes encode a group of transcription factors with a highly conserved 60 amino acid DNA binding motif, the homeodomain, which confers on the resulting proteins the ability to modulate the expression of a variety of target Abbreviations: aa, amino acid(s); BAC, bacterial artificial chromosome; bp, base pairs; CAT, Cm acetyltransferase; cDNA, DNA complementary to RNA; DAPI, 4,6-diamidino-2-phenylindole; DNase, deoxyribonuclease; dNTP, deoxynucleotide triphosphate; EMSA, electrophoretic mobility shift assay; FISH, fluorescence in situ hybridization; HB, homeobox; HD, homeodomain; kDa, kilodalton(s); oligo, oligodeoxyribonucleotide; ORF, open reading frame; PAGE, polyacrylamide gel electrophoresis; RT-PCR, reverse transcriptase-polymerase chain reaction; SDS, sodium dodecyl sulfate * Corresponding author. Department of Biochemistry and Molecular Biology, The George Washington University Medical Center, Ross Building, Room 533, 2300 Eye Street N.W., Washington, DC 20037, USA. Tel.: 11-202-994-2810; fax: 11-202-994-8974. E-mail address: [email protected] (P.E. Berg).
genes (van Oostveen et al., 1999; Akam, 1989). Homeobox genes were initially discovered in Drosophila, where they control segment identity (Lewis, 1978), and have been cloned from many species, including humans, where they play a crucial role in development (Gehring et al., 1994; Levine et al., 1984). For example, disruption of the murine homeobox gene Hoxd-3 results in severe developmental malfunctions (Condie and Capecchi, 1993). Homeobox genes fall into distinct families, based on the sequence of their homeobox. The largest family in humans is called HOX, with 39 members (van Oostveen et al., 1999). We have cloned a cDNA encoding a protein, called Beta Protein 1 (BP1), which represses the human b-globin gene (Accession number: AF254115; Chase et al., 2001). BP1 acts through two DNA binding sites located upstream of the b-globin gene, Silencers I and II (Berg et al., 1989; Chase et al., 2001). Sequencing of the BP1 cDNA revealed that it contains a homeobox homologous to the Drosophila Distal-less gene and therefore belongs to the Distal-less
0378-1119/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0378-111 9(01)00716-8
132
S. Fu et al. / Gene 278 (2001) 131–139
(DLX) family (Cohen et al., 1989). In humans, seven DLX genes have been identified; they are expressed in branchial arches, forebrain, limbs, eyes, teeth, bones and facial mesenchyme (Dolle et al., 1992; Robinson and Mahon, 1994; Simeone et al., 1994). Interestingly, there is extensive sequence identity when BP1 is compared with DLX7, a previously cloned cDNA (Nakamura et al., 1996; Price et al., 1998): the DNA sequences of part of their open reading frames (ORF) are identical and their 3 0 flanking sequences differ by only a few bases. Both are co-expressed in variety of malignant cell lines and tissues (Haga et al., 2000). Based on the sequence identity and similarity in expression patterns, we hypothesize that BP1 and DLX7 are isoforms (RNA splice variants transcribed from a single gene). Therefore, we mapped BP1 by FISH and PCR. The locus of DLX7 is known to be on chromosome 17q21-22 (Nakamura et al., 1996). We also performed functional studies on DLX7 to discover whether it, too, has the ability to act as a repressor of the b-globin promoter. 2. Materials and methods 2.1. Bacterial artificial chromosome library screening and FISH A human bacterial artificial chromosome (BAC) library was screened using a 360 bp probe containing sequences unique to BP1 (Genome Systems). Three positive clones were obtained. Genomic DNA fragments from these clones were amplified with BP1 primers (Chase et al., 2001) using the eLONGase amplification system (Gibco BRL). A 4 kb fragment was obtained, purified from low melting agarose gel and sequenced. Partial sequence comparison with the BP1 cDNA sequence confirmed that one of the BAC clones (BAC19035) contained the BP1 gene. The total BAC genomic DNA was direct labeled with Spectrum Orange by nick translation according to the manufacturer’s protocol (Vysis). A chromosome 17 paint DNA FISH probe was purchased from Vysis. Hybridization was performed with 1 ml of each labeled probe in the same reaction. DAPI I counterstain (125 ng/ml) was used and slides were examined by fluorescence microscopy using a triple band pass filter. 2.2. PCR amplification of DLX7 genomic fragment by eLONGase Specific DLX7 primers were designed as follows: forward, 5 0 -CCTACACCGTGTTGTGCTGC-3 0 in exon 1 (Fig. 1, nt 230–249 of DLX7); reverse, 5 0 -CTGTTGCC ATAGCCACTG-3 0 in exon 2 (Fig. 1, nt 618–635 of DLX7) (Price et al., 1998). This primer set was used to amplify a DLX7 genomic fragment by the eLONGase Amplification System (Gibco BRL) from both BAC19035 and K562 cell line DNA. The final Mg 21 concentration for the PCR reaction used was 1.8 mM, achieved by mixing Buffer A and B with a ratio of 2:8. The PCR conditions
are as follows: pre-amplification denaturation at 948C for 30 s, one cycle; denaturation at 948C for 30 s, annealing at 608C for 30 s and extension at 688C for 4 min, 35 cycles. 2.3. Cloning of DLX7 and BP1 ORFs Primers were designed to amplify the DLX7 complete ORF using the PRIMER program. Restriction enzyme cloning sites HindIII and XbaI were added to facilitate the subsequent cloning. Primer sequences are (italics stand for the added restriction site sequences): forward, 5 0 -CCCAAGCTTGGG CGACGGCATGCAGACGAGAT-3 0 ; reverse, 5 0 -TGCTCTAGAGCA CCTCTAACTGCTTGTCCTCC-3 0 . A 693 bp fragment was amplified from K562 cDNA by PCR and purified from a 0.8% Low-Melting-Point agarose gel using a Pre-A-Gene DNA Purification System (BioRad). The DLX7 cDNA product was directionally cloned into pRc/CMV2 (Invitrogen) after cleavage of both DNAs with HindIII and XbaI. The cloned DLX7 cDNA was verified by sequencing. The BP1 complete ORF was previously cloned into pRc/RSV vector (Chase et al., 2001). We subcloned it into pRc/CMV2 vector using HindIII and XbaI restriction sites. 2.4. In vitro transcription and translation The BP1 and DLX7 ORFs were subcloned into pGEM7Zf(2) (Promega) for in vitro transcription. The TNT Coupled Wheat Germ Extract Transcription/Translation System (Promega) was used for in vitro protein synthesis. One microgram of plasmid DNA containing BP1 or DLX7 ORF and 20 mCi of L-[ 35S]methionine and cysteine mixture (Amersham) were added to an aliquot of the TNT wheat germ extract and incubated in a 50 ml reaction volume for 90 min at 308C. The synthesized proteins then were analyzed by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and autoradiography. 2.5. Electrophoretic mobility shift assays The binding reaction included 2 ml of in vitro translated BP1 or DLX7 protein product mixed with the binding buffer (Berg et al., 1989). Oligonucleotide sequences of probes and competitors for electrophoretic mobility shift assay (EMSA) analyses were as described (Chase et al., 2001). 2.6. Transient transfection and CAT assays Transient transfection assays were performed in K562 cells using DMRIE-C reagent and OPTI-MEM serum-free medium (Invitrogen). CAT assays were performed as described (Berg et al., 1989). Each sample included 1 mg of the internal control plasmid encoding b-galactosidase, 1 mg of the target vector, and varying amounts of the plasmid containing the BP1 or DLX7 ORF, as described in the text.
S. Fu et al. / Gene 278 (2001) 131–139
133
Fig. 1. cDNA sequence homology comparison between BP1 and DLX7. This comparison is based on the DLX7 sequence of Price et al. (1998) and the BP1 sequence of Chase et al. (2001). Dots indicate that the same nucleotide is present in DLX7 and in BP1 cDNA. A dash indicates a missing nucleotide.
134
S. Fu et al. / Gene 278 (2001) 131–139
2.7. RT-PCR Total RNA was isolated using TRIzol Reagent (Invitrogen) according to the manufacturer’s specifications. One microgram of RNA was reverse-transcribed using SuperScript II RT (Invitrogen) in a total reaction volume of 20 ml. PCR was performed with 1 ml of reverse-transcription product in 25 ml. The PCR conditions and primers for DLX7 amplification were as described (Haga et al., 2000).
1988). It was suggested that these mRNAs were transcribed from a common promoter and generated by alternative splicing of a common primary transcript (Simeone et al., 1988). Because BP1 and DLX7 transcripts display distinct 5 0 ends, we speculate that they are either generated by an alternative splicing event or transcribed from two distinct promoters. To further demonstrate that BP1 and DLX7 are isoforms, BP1 was mapped. 3.2. Chromosome localization of BP1
3. Results and discussion 3.1. Sequence comparison between BP1 and DLX7 BP1 and DLX7 exhibit 685 common bases, including their homeodomains, and diverge upstream of nucleotide 565 of the BP1 sequence, shown in Fig. 1. The few mismatches between BP1 nt 565 and 1250 are 3 0 to the BP1 ORF. This extensive sequence identity and common point of divergence indicates that BP1 and DLX7 are likely isoforms, consistent with the fact that homeobox genes are known to display isoforms (van Oostveen et al., 1999). For example, multiple HOX transcripts belonging to genes in the HOXC cluster have been identified (Simeone et al.,
We obtained a series of BACs, which were screened for the presence of BP1 sequences by hybridization, followed by PCR. One of them, BAC19035, contained BP1 sequences (data not shown). FISH analysis with BAC19035 as a probe was used to map BP1. This probe hybridized to chromosome 17q (Fig. 2). The two white arrows point to painted chromosomes 17, and the two orange dots on each chromosome indicate the BP1 hybridization signal. To verify whether this genomic BAC clone also contains DLX7 sequences, we used the DLX7-specific primer pair to amplify the genomic sequence from both K562 cells and BAC DNA (Fig. 3). The same size PCR products indicate that this BAC clone also contains DLX7
Fig. 2. Mapping of BP1 to chromosome 17q. Chromosomes 17, indicated by two arrows, were painted in red as described in Section 2. Two orange dots on each chromosome show the BP1 hybridization signals.
S. Fu et al. / Gene 278 (2001) 131–139
135
negative cells (Haga et al., 2000), which are early hematopoietic progenitors. BP1 and DLX7 are co-expressed in many malignant hematopoietic cell lines, as well as in 84% of the patients with AML (Haga et al., 2000). The significance of their frequent co-expression is as yet unknown. In Drosophila, isoforms of the homeobox gene Ultrabithorax are functionally equivalent (Busturia et al., 1990). In humans, it has been found that HOXC6 encodes two isoforms that both act as transcriptional repressors (Chariot et al., 1996). In previous studies, we demonstrated BP1 repressor function in K562 cells (Chase et al., 2001). Since DLX7 is also expressed in this cell line, we explored whether DLX7 has similar repressor activity. 3.3. In vitro binding of BP1 and DLX7 proteins
Fig. 3. PCR amplification of DLX7 using genomic DNA from the K562 cell line and the BAC19035 clone. Lanes 1 and 2 contain the PCR product from K562 and BAC19035 genomic DNA, respectively.
genomic sequences. These data demonstrate that BP1 maps to the same region as DLX7, 17q21-22, consistent with the hypothesis that they are isoforms. Although unlikely, the formal possibility exists that BP1 and DLX7 are duplicated genes, since an intact genomic sequence is not yet available in this region (S.F. and P.E.B., unpublished data). The mapping of BP1 has clinical significance. BP1 is overexpressed in 47% of adult patients with acute myeloid leukemia (AML), and in 81% of children with AML (Haga et al., 2000). Using these mapping data, we found that BP1 positive bone marrow samples from AML patients did not reveal abnormalities at 17q21-22, implying that translocation may not be an important mechanism for activating BP1 in AML. Overexpression of BP1 in solid tumors has been observed as well (P.E.B. and S.F., unpublished data); knowledge of the locus of BP1 will also be important in cytogenetic analysis of those tumors which are BP1 positive. Human HOX genes are arranged in four clusters, HOXA, HOXB, HOXC and HOXD, on chromosomes 7, 17, 12 and 2, respectively (van Oostveen et al., 1999). In humans, DLX genes have been identified and localized near the HOX clusters: DLX1 and DLX2 on chromosome 2q near the HOXD cluster; DLX5 and DLX6 on chromosome 7 near the HOXA cluster; and DLX3 and now BP1 and DLX7 on chromosome 17q21-22, 3 0 to the HOXB cluster (Scherer et al., 1995; Nakamura et al., 1996). It has been shown that HOX genes at the 3 0 end of each cluster are expressed early in development (van Oostveen et al., 1999). It will be interesting to determine whether BP1 and DLX7, located 3 0 to the HOXB cluster, are also expressed early. Our data support this hypothesis, since BP1 is expressed in CD34
First, BP1 and DLX7 ORFs were separately cloned into pGEM-7Zf(2), designed for coupled in vitro transcription and translation in a wheat germ system. Labeled proteins were synthesized, then analyzed by PAGE (Fig. 4). Two BP1 protein bands were observed, one of higher intensity at 32 kDa, compared with the predicted size of 26 kDa. There is an internal, out-of-frame ATG within BP1, so we speculate that the minor band may be due to translation at that site or degradation or cleavage of the larger band. Two
Fig. 4. BP1 and DLX7 in vitro translated protein products. The 35S-labeled in vitro translated BP1 and DLX7 proteins were analyzed on a 12% SDSpolyacrylamide gel and exposed to X-ray film. A rainbow marker was used to determine protein molecular weights. The sizes of BP1 and DLX7 are 32 kDa (lane 1) and 24 kDa (lane 2), respectively.
136
S. Fu et al. / Gene 278 (2001) 131–139
protein bands were also seen for DLX7. Upstream of the DLX7 translational start, there is an out-of-frame ATG, with no stop codons in that reading frame within our clone, which could potentially be translated into a larger protein than the correctly translated DLX7 (the reading frames of both BP1 and DLX7 are determined by translation of the homeodomain). The smaller DLX7 protein band was 24 kDa, also larger than the predicted size of 19 kDa. The differences observed between the predicted and observed molecular weight of BP1 and DLX7 proteins may be due to posttranslational modifications (Raman et al., 2000). Next, we performed an EMSA to determine whether these in vitro produced proteins bind to a BP1 binding site in Silencer I DNA (Fig. 5). Wheat germ extract alone, when tested for its ability to bind to the probe and cause a shifted band, was negative and used as a control (lane 1). In vitro translated BP1 did produce a shifted band (lane 2), which was competed by a 250 fold excess of a high affinity BP1 binding site (lane 3) (Elion et al., 1992; Berg et al., 1991). There was no competition by a DNA to which BP1 does not
Fig. 5. Binding of BP1 and DLX7 in vitro translated proteins to a known BP1 binding site. EMSA was used to determine the binding of proteins to the BP1 binding site in Silencer I. Lane 1 shows the wheat germ extract alone. Lanes 2–4 contain BP1 protein and lanes 5–7 contain DLX7 protein. Competitors were added at a 250 fold molar excess and included a specific competitor with a high affinity BP1 binding site and a non-specific competitor lacking a BP1 binding site. The arrow indicates the position of the shifted band.
bind (lane 4) (Berg et al., 1989). DLX7 protein also bound, producing a shifted band (lane 5) which was competed by the high affinity BP1 binding site but not by the negative control (lanes 6 and 7, respectively). Thus, both BP1 and DLX7 bind in vitro to this BP1 binding site. Interestingly, although the in vitro proteins are of different sizes, they produce shifted bands at the same position. One possible explanation for this could be that one or both of these proteins bends the DNA, which is known to change the mobility of a shifted band, depending upon where in the DNA fragment the protein binds (Wu and Crothers, 1984; Chase et al., 1999). This possibility was not tested in these experiments. 3.4. Functional comparison of BP1 and DLX7 Transfection of BP1 or DLX7 expressing plasmids into K562 cells was performed to determine whether DLX7 is also a repressor of the b-globin gene. Each cDNA was subcloned into a vector in which the CMV2 promoter drives gene expression, giving BP1 and DLX7 proteins. We previously demonstrated the repressor function of BP1 in K562 cells by co-transfection of a plasmid expressing BP1 protein (driven by the Rous sarcoma virus long terminal repeat) with target DNAs (Chase et al., 2001). Two of the target plasmids used here, described by Berg et al. (1989), contain the CAT reporter gene driven by the 1-globin promoter and include either Silencer II or Silencer I (located 300 or 500 bp upstream of the b-globin gene, respectively). As an internal control, a b-galactosidase plasmid was included in every transfection reaction. The target plasmids and internal control plasmid were transfected at constant amounts of 1 mg, while pDLX7/ORF and pBP1/ORF were transfected in increasing amounts. If repression occurs through either Silencer, we would expect to observe decreased CAT expression as the amount of plasmid encoding DLX7 or BP1 increases. Indeed, when increasing amounts of pBP1/ORF DNA were added to reactions which included Silencer II (Fig. 6A) or Silencer I (Fig. 6B) in the target plasmid, CAT activity was repressed in a dose-dependent manner. Co-transfection of pBP1/ORF and 1/CAT lacking a silencer did not repress CAT activity (data not shown). The addition of pDLX7/ORF, on the other hand, had almost no effect on CAT gene expression, whether the target included Silencer II or Silencer I (Fig. 6A,B, respectively). An additional target was tested which included b-globin promoter sequences to 2338 bp fused to CAT in the presence of the SV40 enhancer. Although the plasmid containing the BP1 ORF repressed CAT activity, the plasmid containing the DLX7 ORF had no effect (data not shown). To verify that the DLX7 ORF was active, analysis of DLX7 transcripts in transfectants was performed using RT-PCR (Fig. 6C). Increased DLX7 mRNA was seen upon transfection with 1, 2 or 5 mg of the plasmid containing the DLX7 ORF, indicating that transcription of the transfected plasmid did occur (Fig. 6C).
S. Fu et al. / Gene 278 (2001) 131–139
137
Fig. 6. Functional analysis of DLX7 and BP1. (A) K562 cells were co-transfected with an 1/CAT reporter plasmid containing Silencer II and with 1, 2, 3 or 4 mg of pDLX7/ORF (left) or pBP1/ORF (right). Bars represent CAT activity relative to the empty vector. Each point was repeated in duplicate in at least two separate experiments, with standard errors as indicated. (B) K562 cells were co-transfected with 1 mg of 1/CAT reporter plasmid containing Silencer I along with 1, 2 or 5 mg of pDLX7/ORF (left) or pBP1/ORF (right). (C) DLX7 RNA was measured following a transient assay in which pDLX7/ORF was cotransfected with the reporter plasmid, RNA was isolated, and RT-PCR was performed. Lane 1 is the control, in which only the reporter was transfected, and lanes 2–4 contain RT-PCR products after transfection of 1, 2 or 5 mg of pDLX7/ORF. The loading control was b-actin.
Thus, although DLX7 protein binds in vitro to a BP1 binding site, it lacks repressor function towards the b-globin gene. There are several possible reasons for this difference. Certain characteristic amino acids have been associated with activator or repressor function. In the case of BP1 and DLX7, functional differences may be explained by analysis of the predicted ORFs. In BP1, there are two proline-rich sequences characteristic of both activation and repression domains (Latchman, 1990; Hanna-Rose and Hansen, 1996), located between aa 81–113 and 188–
214 (Fig. 7). A sub-region of 21 amino acids (aa 87–107) which includes seven acidic amino acids, may have functional significance since acidic regions are also associated with both activation and repression (Latchman, 1990; Hanna-Rose et al., 1997). In addition, the presence of alanines is associated with repression domains (Licht et al., 1990; Dufort and Nepveu, 1994), and the predicted BP1 amino acid sequence contains an alanine-rich region (aa 13–30), which contains six alanine residues. The DLX7 ORF includes only 22 additional amino acids
138
S. Fu et al. / Gene 278 (2001) 131–139
4. Conclusions BP1 and DLX 7 appear to be isoforms since they not only map to the same locus, 17q21-22, but also share extensive sequence identity, including their homeoboxes. Although both BP1 and DLX7 proteins bind to a BP1 repressor site in vitro, only BP1 exhibits repressor activity towards the b-globin gene. BP1 and DLX7 isoforms exhibit similar binding specificity in vitro because they share the same homeodomains, but they have distinct functions possibly due to spliced transcripts encoding different activation or repression domains, or because of binding to different partner proteins. Fig. 7. Schematic diagram of the predicted BP1 and DLX7 proteins showing potential functional regions. The HD is indicated by a hatched box, and an alanine-rich region (Ala), two proline-rich regions (Pro) and an acidic region are boxed. The dashed line indicates the region of DNA sequence identity, beginning at nt 565 of BP1. In DLX7, the dark hatched area is a short proline-rich region which is fused to part of the same proline-rich region found in BP1.
5 0 to the area of sequence identity with BP1 (denoted by the dashed line in Fig. 7). The ORF of DLX7 shares 19 of the 32 amino acids in the proline-rich region of BP1 located upstream of the HB; this sub-region includes five prolines. DLX7 contributes four additional prolines in the 16 amino acids immediately adjacent to this area, giving a proline-rich region which is overlapping but different from that of BP1. Only 13 of 21 amino acids from the acidic region of BP1 (aa 87–107) remain in DLX7. It is questionable whether such a short region has biological significance. The alanine-rich region of BP1 (aa 13–30) is absent from DLX7 protein. The downstream proline-rich sequence is in the region of sequence identity and is present in both BP1 and DLX7. Functional differences in BP1 and DLX7 may be related to differences in their potential repression/activation domains. Alternatively, the functional differences we observed may be due to partner proteins. Many of the homeodomain proteins studied thus far interact with partner proteins, most commonly Meis or Pbx (Shen et al., 1997). The partner protein may determine or influence the activity of the homeodomain protein. It is not yet known whether BP1 or DLX7 has partners, but with BP1, we predict this may be the case since several specific shifted bands are observed in EMSAs using nuclear extract from K562 cells and a BP1 probe (Berg et al., 1989; Chase et al., 2001). As discussed earlier, BP1 and DLX7 are co-expressed in a high percentage of bone marrow samples from AML patients. In addition, overexpression of BP1 in the leukemia cell line K562 leads to an increased ability to grow in soft agar, an indicator of leukemogenic potential (Haga et al., 2000). Thus, it will be important to further delineate the functions of these apparent isoforms in order to discover whether their inappropriate expression is oncogenic.
Acknowledgements This work was supported in part by National Institutes of Health Grant R01 DK53533 (P.E.B.).
References Akam, M., 1989. Hox and HOM: homologous gene clusters in insects and vertebrates. Cell 57, 347–349. Berg, P.E., Williams, D.M., Qian, R.L., Cohen, R.B., Cao, S.X., Mittelman, M., Schechter, A.N., 1989. A common protein binds to two silencers 5 0 to the human beta-globin gene. Nucleic Acids Res. 17, 8833–8852. Berg, P.E., Mittelman, M., Elion, J., Labie, D., Schechter, A.N., 1991. Increased protein binding to a 2530 mutation of the human beta-globin gene associated with decreased beta-globin synthesis. Am. J. Hematol. 36, 42–47. Busturia, A., Vernos, I., Macias, A., Casanova, J., Morata, G., 1990. Different forms of Ultrabithorax proteins generated by alternative splicing are functionally equivalent. EMBO J. 9, 3551–3555. Chariot, A., Castronovo, V., Le, P., Gillet, C., Sobel, M.E., Gielen, J., 1996. Cloning and expression of a new HOXC6 transcript encoding a repressing protein. Biochem. J. 319, 91–97. Chase, M.B., Haga, S.B., Hankins, W.D., Williams, D.M., Bi, Z., Strovel, J.W., Obriecht, C., Berg, P.E., 1999. Binding of HMG-I(Y) elicits structural changes in a silencer of the human beta-globin gene. Am. J. Hematol. 60, 27–35. Chase, M.B., Fu, S., Haga, S.B., Davenport, G.J., Morgan, D., Mah, A.L., Berg, P.E., 2001. BP1, a novel homeobox gene, represses the b-globin gene, submitted for publication. Cohen, S.M., Bronner, G., Kuttner, F., Jurgens, G., Jackle, H., 1989. Distalless encodes a homeodomain protein required for limb development in Drosophila. Nature 338, 432–434. Condie, B.G., Capecchi, M.R., 1993. Mice homozygous for a targeted disruption of Hoxd-3 (Hox-4.1) exhibit anterior transformations of the first and second cervical vertebrae, the atlas and the axis. Development 119, 579–595. Dolle, P., Price, M., Duboule, D., 1992. Expression of the murine Dlx-1 homeobox gene during facial, ocular and limb development. Differentiation 49, 93–99. Dufort, D., Nepveu, A., 1994. The human cut homeodomain protein represses transcription from the c-Myc promoter. Mol. Cell. Biol. 14, 4251–4257. Elion, J., Berg, P.E., Lapoumeroulie, C., Trabuchet, G., Mittelman, M., Krishnamoorthy, R., Schechter, A.N., Labie, D., 1992. DNA sequence variation in a negative control region 5 0 to the b-globin gene correlates with the phenotypic expression of the b s mutation. Blood 79, 787–792.
S. Fu et al. / Gene 278 (2001) 131–139 Gehring, W.J., Affolter, M., Burglin, T., 1994. Homeodomain proteins. Annu. Rev. Biochem. 63, 487–526. Haga, S.B., Fu, S., Karp, J.E., Ross, D.D., Williams, D.M., Hankins, W.D., Behm, F., Ruscetti, F.W., Chang, M., Smith, B.D., Becton, D., Raimondi, S.C., Berg, P.E., 2000. BP1, a new homeobox gene, is frequently expressed in acute leukemias. Leukemia 14, 1867–1875. Hanna-Rose, W., Hansen, U., 1996. Active repression mechanisms of eukaryotic transcription repressors. Trends Genet. 12, 229–234. Hanna-Rose, W., Licht, J.D., Hansen, U., 1997. Two evolutionarily conserved repression domains in the Drosophila Kruppel protein differ in activator specificity. Mol. Cell. Biol. 17, 4820–4829. Latchman, D.S., 1990. Eukaryotic transcription factors. Biochem. J. 270, 281–289. Levine, M., Rubin, G.M., Tjian, R., 1984. Human DNA sequences homologous to a protein coding region conserved between homeotic genes of Drosophila. Cell 38, 667–673. Lewis, E.B., 1978. A gene complex controlling segmentation in Drosophila. Nature 276, 565–570. Licht, J.D., Grossel, M.J., Figge, J., Hansen, U.M., 1990. Drosophila Kruppel protein is a transcriptional repressor. Nature 346, 76–79. Nakamura, S., Stock, D.W., Wydner, K.L., Bollekens, J.A., Takeshita, K., Nagai, B.M., Chiba, S., Kitamura, T., Freeland, T.M., Zhao, Z., Minowada, J., Lawrence, J.B., Weiss, K.M., Ruddle, F.H., 1996. Genomic analysis of a new mammalian distal-less gene: Dlx7. Genomics 38, 314–324. Price, J.A., Bowden, D.W., Wright, J.T., Pettenati, M.J., Hart, T.C., 1998. Identification of a mutation in DLX3 associated with tricho-dentoosseous (TDO) syndrome. Hum. Mol. Genet. 7, 563–569.
139
Raman, V., Martensen, S.A., Reisman, D., Evron, E., Odenwald, W.F., Jaffee, E., Marks, J., Sukumar, S., 2000. Compromised HOXA5 function can limit p53 expression in human breast tumours. Nature 405, 974–978. Robinson, G.W., Mahon, K., 1994. Differential and overlapping expression domains of Dlx-2 and Dlx-3 suggest distinct roles for Distal-less homeobox genes in craniofacial development. Mech. Dev. 48, 199–215. Scherer, S.W., Heng, H.H., Robinson, G.W., Mahon, K.A., Evans, J.P., Tsui, L.C., 1995. Assignment of the human homolog of mouse Dlx3 to chromosome 17q21.3-q22 by analysis of somatic cell hybrids and fluorescence in situ hybridization. Mamm. Genome 6, 310–311. Shen, W.F., Montgomery, J.C., Rozenfeld, S., Moskow, J.J., Lawrence, H.J., Buchberg, A.M., Largman, C., 1997. AbdB-like Hox proteins stabilize DNA binding by the Meis1 homeodomain proteins. Mol. Cell. Biol. 17, 6448–6458. Simeone, A., Pannese, M., Acampora, D., D’Esposito, M., Boncinelli, E., 1988. At least three human homeoboxes on chromosome 12 belong to the same transcription unit. Nucleic Acids Res. 16, 5379–5390. Simeone, A., Acampora, D., Pannese, M., D’Esposito, M., Stornaiuolo, A., Gulisano, M., Mallamaci, A., Kastury, K., Druck, T., Huebner, K., 1994. Cloning and characterization of two members of the vertebrate Dlx gene family. Proc. Natl. Acad. Sci. USA 91, 2250–2254. van Oostveen, J., Biji, J., Raaphorst, F., Walboomers, J., Meijer, C., 1999. The role of homeobox genes in normal hematopoiesis and hematological malignancies. Leukemia 13, 1675–1690. Wu, H.M., Crothers, D.M., 1984. The locus of sequence-directed and protein-induced DNA bending. Nature 308, 509–513.
Distinct functions of two isoforms of a homeobox gene, BP1 and DLX7, in the regulation of the b-globin gene Sidong Fu a, Holly Stevenson a, Jeff W. Strovel b, Susanne B. Haga c, Judy Stamberg b, Khanh Do a, Patricia E. Berg a,* a
Department of Biochemistry and Molecular Biology, The George Washington University Medical Center, Washington, DC, USA b Department of Pediatrics, Division of Genetics, University of Maryland School of Medicine, Baltimore, MD, USA c National Institute of Health, Office of Biotechnology Activities, Bethesda, MD, USA Received 26 March 2001; received in revised form 4 September 2001; accepted 10 September 2001 Received by E. Boncinelli
Abstract Homeotic proteins are transcription factors that regulate the expression of multiple genes involved in development and differentiation. We previously isolated a cDNA encoding such a protein from the human leukemia cell line K562, termed Beta Protein 1 (BP1), which is involved in negative regulation of the human b-globin gene. Sequence comparison revealed that BP1 is a member of the distal-less (DLX) family of homeobox genes and that it shares its homeodomain and 3 0 sequences with another DLX cDNA, DLX7. BP1 and DLX7 exhibit unique 5 0 regions, diverging at nucleotide 565 of BP1. We mapped this new distal-less family member BP1 to chromosome 17q21-22 by FISH and PCR, which is the same locus to which DLX7 has been mapped. These results strongly suggest that BP1 and DLX7 are isoforms (derived from the same gene). Since our previous data demonstrated that BP1 and DLX7 are frequently co-expressed, we determined whether DLX7 is also involved in the negative regulation of the b-globin gene. Mobility shift assays demonstrated that both BP1 and DLX7 proteins, synthesized in vitro, bind to the same BP1 binding site. However, using transient assays, we showed that although BP1 represses activity of a reporter gene through either of two silencer DNA sequences upstream of the b-globin gene, DLX7 did not show repressor activity against the b-globin promoter. Further characterization of these apparent isoforms is of significance since they are jointly expressed in acute myeloid leukemia and in many leukemia cell lines. q 2001 Elsevier Science B.V. All rights reserved. Keywords: Hemoglobin; Homeodomain; Distal-less, Mapping
1. Introduction Homeobox genes encode a group of transcription factors with a highly conserved 60 amino acid DNA binding motif, the homeodomain, which confers on the resulting proteins the ability to modulate the expression of a variety of target Abbreviations: aa, amino acid(s); BAC, bacterial artificial chromosome; bp, base pairs; CAT, Cm acetyltransferase; cDNA, DNA complementary to RNA; DAPI, 4,6-diamidino-2-phenylindole; DNase, deoxyribonuclease; dNTP, deoxynucleotide triphosphate; EMSA, electrophoretic mobility shift assay; FISH, fluorescence in situ hybridization; HB, homeobox; HD, homeodomain; kDa, kilodalton(s); oligo, oligodeoxyribonucleotide; ORF, open reading frame; PAGE, polyacrylamide gel electrophoresis; RT-PCR, reverse transcriptase-polymerase chain reaction; SDS, sodium dodecyl sulfate * Corresponding author. Department of Biochemistry and Molecular Biology, The George Washington University Medical Center, Ross Building, Room 533, 2300 Eye Street N.W., Washington, DC 20037, USA. Tel.: 11-202-994-2810; fax: 11-202-994-8974. E-mail address: [email protected] (P.E. Berg).
genes (van Oostveen et al., 1999; Akam, 1989). Homeobox genes were initially discovered in Drosophila, where they control segment identity (Lewis, 1978), and have been cloned from many species, including humans, where they play a crucial role in development (Gehring et al., 1994; Levine et al., 1984). For example, disruption of the murine homeobox gene Hoxd-3 results in severe developmental malfunctions (Condie and Capecchi, 1993). Homeobox genes fall into distinct families, based on the sequence of their homeobox. The largest family in humans is called HOX, with 39 members (van Oostveen et al., 1999). We have cloned a cDNA encoding a protein, called Beta Protein 1 (BP1), which represses the human b-globin gene (Accession number: AF254115; Chase et al., 2001). BP1 acts through two DNA binding sites located upstream of the b-globin gene, Silencers I and II (Berg et al., 1989; Chase et al., 2001). Sequencing of the BP1 cDNA revealed that it contains a homeobox homologous to the Drosophila Distal-less gene and therefore belongs to the Distal-less
0378-1119/01/$ - see front matter q 2001 Elsevier Science B.V. All rights reserved. PII: S 0378-111 9(01)00716-8
132
S. Fu et al. / Gene 278 (2001) 131–139
(DLX) family (Cohen et al., 1989). In humans, seven DLX genes have been identified; they are expressed in branchial arches, forebrain, limbs, eyes, teeth, bones and facial mesenchyme (Dolle et al., 1992; Robinson and Mahon, 1994; Simeone et al., 1994). Interestingly, there is extensive sequence identity when BP1 is compared with DLX7, a previously cloned cDNA (Nakamura et al., 1996; Price et al., 1998): the DNA sequences of part of their open reading frames (ORF) are identical and their 3 0 flanking sequences differ by only a few bases. Both are co-expressed in variety of malignant cell lines and tissues (Haga et al., 2000). Based on the sequence identity and similarity in expression patterns, we hypothesize that BP1 and DLX7 are isoforms (RNA splice variants transcribed from a single gene). Therefore, we mapped BP1 by FISH and PCR. The locus of DLX7 is known to be on chromosome 17q21-22 (Nakamura et al., 1996). We also performed functional studies on DLX7 to discover whether it, too, has the ability to act as a repressor of the b-globin promoter. 2. Materials and methods 2.1. Bacterial artificial chromosome library screening and FISH A human bacterial artificial chromosome (BAC) library was screened using a 360 bp probe containing sequences unique to BP1 (Genome Systems). Three positive clones were obtained. Genomic DNA fragments from these clones were amplified with BP1 primers (Chase et al., 2001) using the eLONGase amplification system (Gibco BRL). A 4 kb fragment was obtained, purified from low melting agarose gel and sequenced. Partial sequence comparison with the BP1 cDNA sequence confirmed that one of the BAC clones (BAC19035) contained the BP1 gene. The total BAC genomic DNA was direct labeled with Spectrum Orange by nick translation according to the manufacturer’s protocol (Vysis). A chromosome 17 paint DNA FISH probe was purchased from Vysis. Hybridization was performed with 1 ml of each labeled probe in the same reaction. DAPI I counterstain (125 ng/ml) was used and slides were examined by fluorescence microscopy using a triple band pass filter. 2.2. PCR amplification of DLX7 genomic fragment by eLONGase Specific DLX7 primers were designed as follows: forward, 5 0 -CCTACACCGTGTTGTGCTGC-3 0 in exon 1 (Fig. 1, nt 230–249 of DLX7); reverse, 5 0 -CTGTTGCC ATAGCCACTG-3 0 in exon 2 (Fig. 1, nt 618–635 of DLX7) (Price et al., 1998). This primer set was used to amplify a DLX7 genomic fragment by the eLONGase Amplification System (Gibco BRL) from both BAC19035 and K562 cell line DNA. The final Mg 21 concentration for the PCR reaction used was 1.8 mM, achieved by mixing Buffer A and B with a ratio of 2:8. The PCR conditions
are as follows: pre-amplification denaturation at 948C for 30 s, one cycle; denaturation at 948C for 30 s, annealing at 608C for 30 s and extension at 688C for 4 min, 35 cycles. 2.3. Cloning of DLX7 and BP1 ORFs Primers were designed to amplify the DLX7 complete ORF using the PRIMER program. Restriction enzyme cloning sites HindIII and XbaI were added to facilitate the subsequent cloning. Primer sequences are (italics stand for the added restriction site sequences): forward, 5 0 -CCCAAGCTTGGG CGACGGCATGCAGACGAGAT-3 0 ; reverse, 5 0 -TGCTCTAGAGCA CCTCTAACTGCTTGTCCTCC-3 0 . A 693 bp fragment was amplified from K562 cDNA by PCR and purified from a 0.8% Low-Melting-Point agarose gel using a Pre-A-Gene DNA Purification System (BioRad). The DLX7 cDNA product was directionally cloned into pRc/CMV2 (Invitrogen) after cleavage of both DNAs with HindIII and XbaI. The cloned DLX7 cDNA was verified by sequencing. The BP1 complete ORF was previously cloned into pRc/RSV vector (Chase et al., 2001). We subcloned it into pRc/CMV2 vector using HindIII and XbaI restriction sites. 2.4. In vitro transcription and translation The BP1 and DLX7 ORFs were subcloned into pGEM7Zf(2) (Promega) for in vitro transcription. The TNT Coupled Wheat Germ Extract Transcription/Translation System (Promega) was used for in vitro protein synthesis. One microgram of plasmid DNA containing BP1 or DLX7 ORF and 20 mCi of L-[ 35S]methionine and cysteine mixture (Amersham) were added to an aliquot of the TNT wheat germ extract and incubated in a 50 ml reaction volume for 90 min at 308C. The synthesized proteins then were analyzed by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and autoradiography. 2.5. Electrophoretic mobility shift assays The binding reaction included 2 ml of in vitro translated BP1 or DLX7 protein product mixed with the binding buffer (Berg et al., 1989). Oligonucleotide sequences of probes and competitors for electrophoretic mobility shift assay (EMSA) analyses were as described (Chase et al., 2001). 2.6. Transient transfection and CAT assays Transient transfection assays were performed in K562 cells using DMRIE-C reagent and OPTI-MEM serum-free medium (Invitrogen). CAT assays were performed as described (Berg et al., 1989). Each sample included 1 mg of the internal control plasmid encoding b-galactosidase, 1 mg of the target vector, and varying amounts of the plasmid containing the BP1 or DLX7 ORF, as described in the text.
S. Fu et al. / Gene 278 (2001) 131–139
133
Fig. 1. cDNA sequence homology comparison between BP1 and DLX7. This comparison is based on the DLX7 sequence of Price et al. (1998) and the BP1 sequence of Chase et al. (2001). Dots indicate that the same nucleotide is present in DLX7 and in BP1 cDNA. A dash indicates a missing nucleotide.
134
S. Fu et al. / Gene 278 (2001) 131–139
2.7. RT-PCR Total RNA was isolated using TRIzol Reagent (Invitrogen) according to the manufacturer’s specifications. One microgram of RNA was reverse-transcribed using SuperScript II RT (Invitrogen) in a total reaction volume of 20 ml. PCR was performed with 1 ml of reverse-transcription product in 25 ml. The PCR conditions and primers for DLX7 amplification were as described (Haga et al., 2000).
1988). It was suggested that these mRNAs were transcribed from a common promoter and generated by alternative splicing of a common primary transcript (Simeone et al., 1988). Because BP1 and DLX7 transcripts display distinct 5 0 ends, we speculate that they are either generated by an alternative splicing event or transcribed from two distinct promoters. To further demonstrate that BP1 and DLX7 are isoforms, BP1 was mapped. 3.2. Chromosome localization of BP1
3. Results and discussion 3.1. Sequence comparison between BP1 and DLX7 BP1 and DLX7 exhibit 685 common bases, including their homeodomains, and diverge upstream of nucleotide 565 of the BP1 sequence, shown in Fig. 1. The few mismatches between BP1 nt 565 and 1250 are 3 0 to the BP1 ORF. This extensive sequence identity and common point of divergence indicates that BP1 and DLX7 are likely isoforms, consistent with the fact that homeobox genes are known to display isoforms (van Oostveen et al., 1999). For example, multiple HOX transcripts belonging to genes in the HOXC cluster have been identified (Simeone et al.,
We obtained a series of BACs, which were screened for the presence of BP1 sequences by hybridization, followed by PCR. One of them, BAC19035, contained BP1 sequences (data not shown). FISH analysis with BAC19035 as a probe was used to map BP1. This probe hybridized to chromosome 17q (Fig. 2). The two white arrows point to painted chromosomes 17, and the two orange dots on each chromosome indicate the BP1 hybridization signal. To verify whether this genomic BAC clone also contains DLX7 sequences, we used the DLX7-specific primer pair to amplify the genomic sequence from both K562 cells and BAC DNA (Fig. 3). The same size PCR products indicate that this BAC clone also contains DLX7
Fig. 2. Mapping of BP1 to chromosome 17q. Chromosomes 17, indicated by two arrows, were painted in red as described in Section 2. Two orange dots on each chromosome show the BP1 hybridization signals.
S. Fu et al. / Gene 278 (2001) 131–139
135
negative cells (Haga et al., 2000), which are early hematopoietic progenitors. BP1 and DLX7 are co-expressed in many malignant hematopoietic cell lines, as well as in 84% of the patients with AML (Haga et al., 2000). The significance of their frequent co-expression is as yet unknown. In Drosophila, isoforms of the homeobox gene Ultrabithorax are functionally equivalent (Busturia et al., 1990). In humans, it has been found that HOXC6 encodes two isoforms that both act as transcriptional repressors (Chariot et al., 1996). In previous studies, we demonstrated BP1 repressor function in K562 cells (Chase et al., 2001). Since DLX7 is also expressed in this cell line, we explored whether DLX7 has similar repressor activity. 3.3. In vitro binding of BP1 and DLX7 proteins
Fig. 3. PCR amplification of DLX7 using genomic DNA from the K562 cell line and the BAC19035 clone. Lanes 1 and 2 contain the PCR product from K562 and BAC19035 genomic DNA, respectively.
genomic sequences. These data demonstrate that BP1 maps to the same region as DLX7, 17q21-22, consistent with the hypothesis that they are isoforms. Although unlikely, the formal possibility exists that BP1 and DLX7 are duplicated genes, since an intact genomic sequence is not yet available in this region (S.F. and P.E.B., unpublished data). The mapping of BP1 has clinical significance. BP1 is overexpressed in 47% of adult patients with acute myeloid leukemia (AML), and in 81% of children with AML (Haga et al., 2000). Using these mapping data, we found that BP1 positive bone marrow samples from AML patients did not reveal abnormalities at 17q21-22, implying that translocation may not be an important mechanism for activating BP1 in AML. Overexpression of BP1 in solid tumors has been observed as well (P.E.B. and S.F., unpublished data); knowledge of the locus of BP1 will also be important in cytogenetic analysis of those tumors which are BP1 positive. Human HOX genes are arranged in four clusters, HOXA, HOXB, HOXC and HOXD, on chromosomes 7, 17, 12 and 2, respectively (van Oostveen et al., 1999). In humans, DLX genes have been identified and localized near the HOX clusters: DLX1 and DLX2 on chromosome 2q near the HOXD cluster; DLX5 and DLX6 on chromosome 7 near the HOXA cluster; and DLX3 and now BP1 and DLX7 on chromosome 17q21-22, 3 0 to the HOXB cluster (Scherer et al., 1995; Nakamura et al., 1996). It has been shown that HOX genes at the 3 0 end of each cluster are expressed early in development (van Oostveen et al., 1999). It will be interesting to determine whether BP1 and DLX7, located 3 0 to the HOXB cluster, are also expressed early. Our data support this hypothesis, since BP1 is expressed in CD34
First, BP1 and DLX7 ORFs were separately cloned into pGEM-7Zf(2), designed for coupled in vitro transcription and translation in a wheat germ system. Labeled proteins were synthesized, then analyzed by PAGE (Fig. 4). Two BP1 protein bands were observed, one of higher intensity at 32 kDa, compared with the predicted size of 26 kDa. There is an internal, out-of-frame ATG within BP1, so we speculate that the minor band may be due to translation at that site or degradation or cleavage of the larger band. Two
Fig. 4. BP1 and DLX7 in vitro translated protein products. The 35S-labeled in vitro translated BP1 and DLX7 proteins were analyzed on a 12% SDSpolyacrylamide gel and exposed to X-ray film. A rainbow marker was used to determine protein molecular weights. The sizes of BP1 and DLX7 are 32 kDa (lane 1) and 24 kDa (lane 2), respectively.
136
S. Fu et al. / Gene 278 (2001) 131–139
protein bands were also seen for DLX7. Upstream of the DLX7 translational start, there is an out-of-frame ATG, with no stop codons in that reading frame within our clone, which could potentially be translated into a larger protein than the correctly translated DLX7 (the reading frames of both BP1 and DLX7 are determined by translation of the homeodomain). The smaller DLX7 protein band was 24 kDa, also larger than the predicted size of 19 kDa. The differences observed between the predicted and observed molecular weight of BP1 and DLX7 proteins may be due to posttranslational modifications (Raman et al., 2000). Next, we performed an EMSA to determine whether these in vitro produced proteins bind to a BP1 binding site in Silencer I DNA (Fig. 5). Wheat germ extract alone, when tested for its ability to bind to the probe and cause a shifted band, was negative and used as a control (lane 1). In vitro translated BP1 did produce a shifted band (lane 2), which was competed by a 250 fold excess of a high affinity BP1 binding site (lane 3) (Elion et al., 1992; Berg et al., 1991). There was no competition by a DNA to which BP1 does not
Fig. 5. Binding of BP1 and DLX7 in vitro translated proteins to a known BP1 binding site. EMSA was used to determine the binding of proteins to the BP1 binding site in Silencer I. Lane 1 shows the wheat germ extract alone. Lanes 2–4 contain BP1 protein and lanes 5–7 contain DLX7 protein. Competitors were added at a 250 fold molar excess and included a specific competitor with a high affinity BP1 binding site and a non-specific competitor lacking a BP1 binding site. The arrow indicates the position of the shifted band.
bind (lane 4) (Berg et al., 1989). DLX7 protein also bound, producing a shifted band (lane 5) which was competed by the high affinity BP1 binding site but not by the negative control (lanes 6 and 7, respectively). Thus, both BP1 and DLX7 bind in vitro to this BP1 binding site. Interestingly, although the in vitro proteins are of different sizes, they produce shifted bands at the same position. One possible explanation for this could be that one or both of these proteins bends the DNA, which is known to change the mobility of a shifted band, depending upon where in the DNA fragment the protein binds (Wu and Crothers, 1984; Chase et al., 1999). This possibility was not tested in these experiments. 3.4. Functional comparison of BP1 and DLX7 Transfection of BP1 or DLX7 expressing plasmids into K562 cells was performed to determine whether DLX7 is also a repressor of the b-globin gene. Each cDNA was subcloned into a vector in which the CMV2 promoter drives gene expression, giving BP1 and DLX7 proteins. We previously demonstrated the repressor function of BP1 in K562 cells by co-transfection of a plasmid expressing BP1 protein (driven by the Rous sarcoma virus long terminal repeat) with target DNAs (Chase et al., 2001). Two of the target plasmids used here, described by Berg et al. (1989), contain the CAT reporter gene driven by the 1-globin promoter and include either Silencer II or Silencer I (located 300 or 500 bp upstream of the b-globin gene, respectively). As an internal control, a b-galactosidase plasmid was included in every transfection reaction. The target plasmids and internal control plasmid were transfected at constant amounts of 1 mg, while pDLX7/ORF and pBP1/ORF were transfected in increasing amounts. If repression occurs through either Silencer, we would expect to observe decreased CAT expression as the amount of plasmid encoding DLX7 or BP1 increases. Indeed, when increasing amounts of pBP1/ORF DNA were added to reactions which included Silencer II (Fig. 6A) or Silencer I (Fig. 6B) in the target plasmid, CAT activity was repressed in a dose-dependent manner. Co-transfection of pBP1/ORF and 1/CAT lacking a silencer did not repress CAT activity (data not shown). The addition of pDLX7/ORF, on the other hand, had almost no effect on CAT gene expression, whether the target included Silencer II or Silencer I (Fig. 6A,B, respectively). An additional target was tested which included b-globin promoter sequences to 2338 bp fused to CAT in the presence of the SV40 enhancer. Although the plasmid containing the BP1 ORF repressed CAT activity, the plasmid containing the DLX7 ORF had no effect (data not shown). To verify that the DLX7 ORF was active, analysis of DLX7 transcripts in transfectants was performed using RT-PCR (Fig. 6C). Increased DLX7 mRNA was seen upon transfection with 1, 2 or 5 mg of the plasmid containing the DLX7 ORF, indicating that transcription of the transfected plasmid did occur (Fig. 6C).
S. Fu et al. / Gene 278 (2001) 131–139
137
Fig. 6. Functional analysis of DLX7 and BP1. (A) K562 cells were co-transfected with an 1/CAT reporter plasmid containing Silencer II and with 1, 2, 3 or 4 mg of pDLX7/ORF (left) or pBP1/ORF (right). Bars represent CAT activity relative to the empty vector. Each point was repeated in duplicate in at least two separate experiments, with standard errors as indicated. (B) K562 cells were co-transfected with 1 mg of 1/CAT reporter plasmid containing Silencer I along with 1, 2 or 5 mg of pDLX7/ORF (left) or pBP1/ORF (right). (C) DLX7 RNA was measured following a transient assay in which pDLX7/ORF was cotransfected with the reporter plasmid, RNA was isolated, and RT-PCR was performed. Lane 1 is the control, in which only the reporter was transfected, and lanes 2–4 contain RT-PCR products after transfection of 1, 2 or 5 mg of pDLX7/ORF. The loading control was b-actin.
Thus, although DLX7 protein binds in vitro to a BP1 binding site, it lacks repressor function towards the b-globin gene. There are several possible reasons for this difference. Certain characteristic amino acids have been associated with activator or repressor function. In the case of BP1 and DLX7, functional differences may be explained by analysis of the predicted ORFs. In BP1, there are two proline-rich sequences characteristic of both activation and repression domains (Latchman, 1990; Hanna-Rose and Hansen, 1996), located between aa 81–113 and 188–
214 (Fig. 7). A sub-region of 21 amino acids (aa 87–107) which includes seven acidic amino acids, may have functional significance since acidic regions are also associated with both activation and repression (Latchman, 1990; Hanna-Rose et al., 1997). In addition, the presence of alanines is associated with repression domains (Licht et al., 1990; Dufort and Nepveu, 1994), and the predicted BP1 amino acid sequence contains an alanine-rich region (aa 13–30), which contains six alanine residues. The DLX7 ORF includes only 22 additional amino acids
138
S. Fu et al. / Gene 278 (2001) 131–139
4. Conclusions BP1 and DLX 7 appear to be isoforms since they not only map to the same locus, 17q21-22, but also share extensive sequence identity, including their homeoboxes. Although both BP1 and DLX7 proteins bind to a BP1 repressor site in vitro, only BP1 exhibits repressor activity towards the b-globin gene. BP1 and DLX7 isoforms exhibit similar binding specificity in vitro because they share the same homeodomains, but they have distinct functions possibly due to spliced transcripts encoding different activation or repression domains, or because of binding to different partner proteins. Fig. 7. Schematic diagram of the predicted BP1 and DLX7 proteins showing potential functional regions. The HD is indicated by a hatched box, and an alanine-rich region (Ala), two proline-rich regions (Pro) and an acidic region are boxed. The dashed line indicates the region of DNA sequence identity, beginning at nt 565 of BP1. In DLX7, the dark hatched area is a short proline-rich region which is fused to part of the same proline-rich region found in BP1.
5 0 to the area of sequence identity with BP1 (denoted by the dashed line in Fig. 7). The ORF of DLX7 shares 19 of the 32 amino acids in the proline-rich region of BP1 located upstream of the HB; this sub-region includes five prolines. DLX7 contributes four additional prolines in the 16 amino acids immediately adjacent to this area, giving a proline-rich region which is overlapping but different from that of BP1. Only 13 of 21 amino acids from the acidic region of BP1 (aa 87–107) remain in DLX7. It is questionable whether such a short region has biological significance. The alanine-rich region of BP1 (aa 13–30) is absent from DLX7 protein. The downstream proline-rich sequence is in the region of sequence identity and is present in both BP1 and DLX7. Functional differences in BP1 and DLX7 may be related to differences in their potential repression/activation domains. Alternatively, the functional differences we observed may be due to partner proteins. Many of the homeodomain proteins studied thus far interact with partner proteins, most commonly Meis or Pbx (Shen et al., 1997). The partner protein may determine or influence the activity of the homeodomain protein. It is not yet known whether BP1 or DLX7 has partners, but with BP1, we predict this may be the case since several specific shifted bands are observed in EMSAs using nuclear extract from K562 cells and a BP1 probe (Berg et al., 1989; Chase et al., 2001). As discussed earlier, BP1 and DLX7 are co-expressed in a high percentage of bone marrow samples from AML patients. In addition, overexpression of BP1 in the leukemia cell line K562 leads to an increased ability to grow in soft agar, an indicator of leukemogenic potential (Haga et al., 2000). Thus, it will be important to further delineate the functions of these apparent isoforms in order to discover whether their inappropriate expression is oncogenic.
Acknowledgements This work was supported in part by National Institutes of Health Grant R01 DK53533 (P.E.B.).
References Akam, M., 1989. Hox and HOM: homologous gene clusters in insects and vertebrates. Cell 57, 347–349. Berg, P.E., Williams, D.M., Qian, R.L., Cohen, R.B., Cao, S.X., Mittelman, M., Schechter, A.N., 1989. A common protein binds to two silencers 5 0 to the human beta-globin gene. Nucleic Acids Res. 17, 8833–8852. Berg, P.E., Mittelman, M., Elion, J., Labie, D., Schechter, A.N., 1991. Increased protein binding to a 2530 mutation of the human beta-globin gene associated with decreased beta-globin synthesis. Am. J. Hematol. 36, 42–47. Busturia, A., Vernos, I., Macias, A., Casanova, J., Morata, G., 1990. Different forms of Ultrabithorax proteins generated by alternative splicing are functionally equivalent. EMBO J. 9, 3551–3555. Chariot, A., Castronovo, V., Le, P., Gillet, C., Sobel, M.E., Gielen, J., 1996. Cloning and expression of a new HOXC6 transcript encoding a repressing protein. Biochem. J. 319, 91–97. Chase, M.B., Haga, S.B., Hankins, W.D., Williams, D.M., Bi, Z., Strovel, J.W., Obriecht, C., Berg, P.E., 1999. Binding of HMG-I(Y) elicits structural changes in a silencer of the human beta-globin gene. Am. J. Hematol. 60, 27–35. Chase, M.B., Fu, S., Haga, S.B., Davenport, G.J., Morgan, D., Mah, A.L., Berg, P.E., 2001. BP1, a novel homeobox gene, represses the b-globin gene, submitted for publication. Cohen, S.M., Bronner, G., Kuttner, F., Jurgens, G., Jackle, H., 1989. Distalless encodes a homeodomain protein required for limb development in Drosophila. Nature 338, 432–434. Condie, B.G., Capecchi, M.R., 1993. Mice homozygous for a targeted disruption of Hoxd-3 (Hox-4.1) exhibit anterior transformations of the first and second cervical vertebrae, the atlas and the axis. Development 119, 579–595. Dolle, P., Price, M., Duboule, D., 1992. Expression of the murine Dlx-1 homeobox gene during facial, ocular and limb development. Differentiation 49, 93–99. Dufort, D., Nepveu, A., 1994. The human cut homeodomain protein represses transcription from the c-Myc promoter. Mol. Cell. Biol. 14, 4251–4257. Elion, J., Berg, P.E., Lapoumeroulie, C., Trabuchet, G., Mittelman, M., Krishnamoorthy, R., Schechter, A.N., Labie, D., 1992. DNA sequence variation in a negative control region 5 0 to the b-globin gene correlates with the phenotypic expression of the b s mutation. Blood 79, 787–792.
S. Fu et al. / Gene 278 (2001) 131–139 Gehring, W.J., Affolter, M., Burglin, T., 1994. Homeodomain proteins. Annu. Rev. Biochem. 63, 487–526. Haga, S.B., Fu, S., Karp, J.E., Ross, D.D., Williams, D.M., Hankins, W.D., Behm, F., Ruscetti, F.W., Chang, M., Smith, B.D., Becton, D., Raimondi, S.C., Berg, P.E., 2000. BP1, a new homeobox gene, is frequently expressed in acute leukemias. Leukemia 14, 1867–1875. Hanna-Rose, W., Hansen, U., 1996. Active repression mechanisms of eukaryotic transcription repressors. Trends Genet. 12, 229–234. Hanna-Rose, W., Licht, J.D., Hansen, U., 1997. Two evolutionarily conserved repression domains in the Drosophila Kruppel protein differ in activator specificity. Mol. Cell. Biol. 17, 4820–4829. Latchman, D.S., 1990. Eukaryotic transcription factors. Biochem. J. 270, 281–289. Levine, M., Rubin, G.M., Tjian, R., 1984. Human DNA sequences homologous to a protein coding region conserved between homeotic genes of Drosophila. Cell 38, 667–673. Lewis, E.B., 1978. A gene complex controlling segmentation in Drosophila. Nature 276, 565–570. Licht, J.D., Grossel, M.J., Figge, J., Hansen, U.M., 1990. Drosophila Kruppel protein is a transcriptional repressor. Nature 346, 76–79. Nakamura, S., Stock, D.W., Wydner, K.L., Bollekens, J.A., Takeshita, K., Nagai, B.M., Chiba, S., Kitamura, T., Freeland, T.M., Zhao, Z., Minowada, J., Lawrence, J.B., Weiss, K.M., Ruddle, F.H., 1996. Genomic analysis of a new mammalian distal-less gene: Dlx7. Genomics 38, 314–324. Price, J.A., Bowden, D.W., Wright, J.T., Pettenati, M.J., Hart, T.C., 1998. Identification of a mutation in DLX3 associated with tricho-dentoosseous (TDO) syndrome. Hum. Mol. Genet. 7, 563–569.
139
Raman, V., Martensen, S.A., Reisman, D., Evron, E., Odenwald, W.F., Jaffee, E., Marks, J., Sukumar, S., 2000. Compromised HOXA5 function can limit p53 expression in human breast tumours. Nature 405, 974–978. Robinson, G.W., Mahon, K., 1994. Differential and overlapping expression domains of Dlx-2 and Dlx-3 suggest distinct roles for Distal-less homeobox genes in craniofacial development. Mech. Dev. 48, 199–215. Scherer, S.W., Heng, H.H., Robinson, G.W., Mahon, K.A., Evans, J.P., Tsui, L.C., 1995. Assignment of the human homolog of mouse Dlx3 to chromosome 17q21.3-q22 by analysis of somatic cell hybrids and fluorescence in situ hybridization. Mamm. Genome 6, 310–311. Shen, W.F., Montgomery, J.C., Rozenfeld, S., Moskow, J.J., Lawrence, H.J., Buchberg, A.M., Largman, C., 1997. AbdB-like Hox proteins stabilize DNA binding by the Meis1 homeodomain proteins. Mol. Cell. Biol. 17, 6448–6458. Simeone, A., Pannese, M., Acampora, D., D’Esposito, M., Boncinelli, E., 1988. At least three human homeoboxes on chromosome 12 belong to the same transcription unit. Nucleic Acids Res. 16, 5379–5390. Simeone, A., Acampora, D., Pannese, M., D’Esposito, M., Stornaiuolo, A., Gulisano, M., Mallamaci, A., Kastury, K., Druck, T., Huebner, K., 1994. Cloning and characterization of two members of the vertebrate Dlx gene family. Proc. Natl. Acad. Sci. USA 91, 2250–2254. van Oostveen, J., Biji, J., Raaphorst, F., Walboomers, J., Meijer, C., 1999. The role of homeobox genes in normal hematopoiesis and hematological malignancies. Leukemia 13, 1675–1690. Wu, H.M., Crothers, D.M., 1984. The locus of sequence-directed and protein-induced DNA bending. Nature 308, 509–513.