Isolation and characterization of a chicken homologue of the ... - Nature

12 downloads 0 Views 1MB Size Report
Spi-1/PU.1 is a member of the Ets family of transcription factors important in regulation of hemato- poiesis. We have isolated a chicken cDNA homologuous.
Oncogene (1998) 16, 1357 ± 1367  1998 Stockton Press All rights reserved 0950 ± 9232/98 $12.00

SHORT REPORT

Isolation and characterization of a chicken homologue of the Spi-1/PU.1 transcription factor Zoulika Kherrouche1, Anne Beuscart1, Christelle Huguet1, Anne Flourens1, FrancËoise Moreau-Gachelin2, Dominique Stehelin1 and Jean Coll1 1

Institut de Biologie de Lille, UMR 319 CNRS/Institut Pasteur de Lille, 1 rue Calmette, BP 447, 59021 Lille Cedex; 2Institut Curie, INSERM U248, 26 rue d'Ulm, 75231 Paris Cedex 05, France.

Spi-1/PU.1 is a member of the Ets family of transcription factors important in regulation of hematopoiesis. We have isolated a chicken cDNA homologuous to the mammalian Spi-1/PU.1 gene with an open reading frame of 250 amino acids (aa). The chicken Spi-1/PU.1 protein is 14 aa and 16 aa shorter than its human and mouse counterparts but is extremely well conserved with 78.8% and 75.2% identity respectively. The carboxy terminal DNA binding region, or ETS binding domain, is 100% identical to that of human and mouse. Some di€erences with the mammalian homologues are seen in the N-terminal part of the protein and in the PEST connecting domain. However, the di€erences are mainly conservative and all the features underlying functional aspects seem preserved. The major discrepancy lies in a 12 aa deletion in an already poorly conserved part of the PEST sequence. Spi-1/PU.1 transcripts were detected at high levels in spleen and Fabricius bursa of chick embryos by Northern blot and in situ hybridization. Our results show that the chicken Spi-1/PU.1 protein behaves like a bona®de Spi-1/PU.1 transcription factor in its DNA binding and transactivating properties. Keywords: Spi-1/PU.1; ets genes; transcription factor; chicken; in-situ-hybridization

The Spi-1 oncogene was ®rst isolated by its implication in the development of Spleen Focus Forming Virusinduced murine Friend erythroleukemia. Spleen Focus Forming Virus (SFFV) was found to integrate adjacent to the Spi-l locus in 95% of the tumors, resulting in an elevated expression of Spi-l mRNA (Moreau-Gachelin et al., 1988, 1989). The instrumental role of Spi-l overexpression for malignant erythropoiesis was recently con®rmed by transgenic experiments (Moreau-Gachelin et al., 1996). The Spi-l gene product is identical to the transcriptional factor PU.1 which has been identi®ed by its ability to bind a purine rich core sequence 5'-GAGGAA-3' in the SV40 transcriptional enhancer (Goebl, 1990; Klemsz et al., 1990). Spi-1/PU.1 is a member of the Ets protein family (Laudet et al., 1993; Lautenberger et al., 1992). All members of this family share a conserved domain, the ETS domain, which is responsible for sequence speci®c DNA-binding (Karim et al., 1990). However, it is the Correspondence: J Coll Received 12 May 1997; revised 15 October 1997; accepted 15 October 1997

most divergent member of the family de®ning with its most related gene Spi-B (Ray et al., 1992), the SPI subfamily (Laudet et al., 1993). Spi-1/PU.1 expression was originally described as Bcell and macrophage-speci®c (Klemsz et al., 1990). More recent analyses revealed that Spi-1/PU.1 expression is detectable in most hematopoietic lineages except T-lymphocytes (Ray et al., 1992). Studies with human CD34+ cells have shown that Spi-1/PU.1 is expressed in the earliest multipotential progenitors, upregulated during commitment to the myeloid lineage and downregulated during erythroid di€erentiation (Chen et al., 1995; Voso et al., 1994). Moreover, overexpression of Spi-1/PU.1 in developing erythroblasts or in transgenic mice results in a block of erythroid di€erentiation (Moreau-Gachelin et al., 1996; Schuetze et al., 1993). These results and others (Rao et al., 1997) strongly suggest that downregulation of Spi-1/PU.1 is a prerequisite to terminal di€erentiation of erythroid cells. Homologous recombination of Spi-1/PU.1 gene in mouse leads to the death at embryonic stage (Scott et al., 1994) or days after birth (McKercher et al., 1996). The homozygous mutant embryos show a multilineage defect in the maturation of B and T lymphocytes, macrophages and neutrophils. The recent data which imply Spi-1/PU.1 in osteoclast development reinforce the major role of this transcription factor as a transcriptional regulator in the hematopoietic development (Tondravi et al., 1997). Moreover, the capital role of this gene may be further stressed by the recent ®nding that Spi-1/PU.1 interacts with RNA-binding proteins and binds RNAs in vitro (Hallier et al., 1996). Target sequences for Spi-1/PU.1 have been identi®ed in viral enhancers or promoters (Klemsz et al., 1990; Carvalho et al., 1993; Laux et al., 1994). However, in accordance to its expression pattern, numerous target genes have been reported in di€erent hematopoietic lineages (reviewed in Moreau-Gachelin, 1994). This is for example the case for genes with expression restricted to B lymphocytes, like genes coding for immunoglobulin chains: kappa (Pongubala et al., 1992, 1993) and lambda (Eisenbeis et al., 1993) lightchains, mu heavy (Nelsen et al., 1993; Rivera et al., 1993) and J chains (Shin and Koshland, 1993); and genes regulated by Spi-1/PU.1 in myeloid lineages like interleukin lb (Kominato et al., 1995), CD18 (Rosmarin et al., 1995), c-Fes/ c-Fps tyrosine kinase (Ray-Gallet et al., 1995). In many cases, cell-type restricted expression of target genes was shown to result from combinatorial and cooperative action of various transcription factors, including Spi-1/PU.1. This is well documented for the B-cell restricted

Characterization of the chicken Spi-1/PU.1 gene Z Kherrouche et al

1358

action of the kappa 3' enhancer (Pongubala and Atchison, 1997), and for the murine neutrophil elastase promoter in immature myeloid cells (Oelgeschlager et al., 1996). Finally, it is interesting to note that Spi-1/PU.1 plays a role in its own restricted expression by positive autoregulation of its promoter (Chen et al., 1995; Kistler et al., 1995). In this communication, we report the isolation and sequence analysis of a cDNA encoding the complete sequence of a chicken Spi-1/PU.1 protein. The product of this gene was able to bind DNA and to transactivate a reporter plasmid containing a Spi-1/PU.1 binding site. Northern blot analysis of tissues from chicken embryos showed that high expression of Spi-1/PU.1 RNA was restricted to the spleen and the bursa. Moreover, in situ hybridizations localized the expression of the chicken Spi-1/PU.1 in the B-cells and macrophages areas of the hematopoietic organs. cDNA sequence, predicted protein structure and comparison of the chicken Spi-1/PU.1 with its mammalian homologues Amino-acid sequences of the human and mouse DNAbinding domains of the Spi-1/PU.1 proteins are strictly conserved. Moreover, this perfect identity re¯ected a good conservation of the nucleotide sequence corresponding to this domain in the two genes (94.7% identity). Assuming that this conservation could be preserved in other vertebrates, we used a probe encompassing this region to isolate the chicken Spi-1/ PU.1 gene. To this aim, a 404 bp SacI-NarI probe was derived from a mouse Spi-1/PU.1 cDNA and used to screen 106 phages from the lgtl0 chicken spleen cDNA library (provided by Klaus Strebhardt) as described (Sambrook et al., 1989). Among the cDNAs isolated and size-characterized, the longest clone was subcloned into the EcoRI site of the pBluescript polylinker and the nucleotide sequence was determined. This clone was 1457 bp long and ended with a 13 nucleotides-long poly(A) track (EMBL accession number Y12225) (Figure 1). The overall sequence homology between this sequence and the human and mouse Spi-1/PU.1 genes in the coding region is 73.8% and 72.5%, respectively. Analysis of the 3' end of the clone revealed the motif AATAAA placed 13 nucleotides (nt) upstream from the poly(A) track and characteristic of an appropriate polyadenylation signal. The longest open reading frame detected in the sequence is 768 nt long, ¯anked by a 462 nt-long 3' untranslated region (UTR) and 227 nt of a 5' UTR. As expected for untranslated sequences, these ¯anking sequences display signi®cant but weaker homology with the corresponding parts of the mammalian genes (39% and 31% for the human and mouse genes respectively). Two potential initiation codons spaced by only 15 nt are found close to the beginning of this open reading frame. The ®rst in-frame ATG codon is located just downstream from a terminator codon but is in a disadvantageous context for proper initiation. In contrast, the second ATG matches the classical Kozak rules with the presence of appropriate Adenine and Guanine bases at 73 and +4 positions, respectively. This codon could be considered as the actual translation initiation codon as was argued for the

murine Spi-1/PU.1 cDNA which displays these same features (Moreau-Gachelin et al., 1989). Translation initiated at this codon and terminating at a TGA codon should lead to the synthesis of a 250 amino acids protein with a predicted molecular mass of 29 kDa (Figure 1). The protein has an overall 78.8% and 75.2% identity with the human and mouse Spi-1/PU.1 proteins, respectively (Figure 2). The identity drops to 32.5-28.1% when the sequence was compared to the human and murine Spi-B proteins, the most closely related Ets family members. This clearly indicated that we did actually clone the chicken Spi-1/PU.1 counterpart. As expected, the highest identity (100-99.4%) is found in the 94 a.a. region (residues 146 to 240) containing the DNA binding domain (also named the ETS domain) and located in the C-terminal part of the protein. An aspartic to glutamic acid conservative substitution at position 176 is the unique di€erence found for this domain in the chicken protein. Amino acid substitutions are mainly clustered in the Nterminal part of the protein. Homology scores between chicken, human and mouse Spi-1/PU.1 proteins are 76.1% and 66.7% for the N-terminal 105 amino acids. Moreover, the majority of the substitutions found in this part are conservative. The

Figure 1 Nucleotide and predicted amino acid sequences of chicken Spi-1/PU.1 cDNA. The complete sequence was determined on both strands by the automatic sequencing PCR procedure with ¯uorescent primer coupled (Prism TM ready reaction, Dye Deoxy terminator kit from Applied Biosystems). The nucleotide and amino acid positions are indicated to the right of the sequence in normal and bold characters, respectively. The nucleotides corresponding to the open reading frame are in capital letters and the 5'- and 3'- untranslated regions are in small letters. The putative polyadenylation signal is in bold letters. The EcoRI site used to clone the cDNA into the pSG5 vector is underlined

Characterization of the chicken Spi-1/PU.1 gene Z Kherrouche et al

1359

Expression pattern of chicken Spi-1 mRNAs Tissue distribution in chick embryos of Spi-1/PU.1 mRNA was determined by Northern blotting (Figure 3a). It revealed a wide expression of two mRNA transcripts with a respective size of 2.1 kb and 2.9 kb in 19 day-old chick embryonic spleen and bursa. The other tissue tested (bone marrow, thymus, lung, gizzard, heart, muscle and liver) seem negative for Spi-1/PU.1 expression at this stage. However, overexposure reveals very faint signals for the thymus and the bone marrow (data not shown). In the same order, no expression was found in 3 day-old or in 9 day-old embryos. This result re¯ects probably, at least in the last case, the poor sensitivity of the method to detect expression in already developing hematopoietic organs. We also examined Spi-1/PU.1 mRNA expression in a wide series of chicken cell lines (Figure 3b). The macrophage-like HD11 cell line presented the highest expression observed in this study. Spi-1/PU.1 mRNAs were also detected in the myeloblasts BM2C3 cell line

Fabricius bursa E19

HD11 cell line

Body E9

Head E9

Liver E19

Total embryo E3

Muscle E19

Heart E19

Gizzard E19

Spleen E19

Lung E19

Thymus E19

a Bone marrow E19 Spi/PU.1

GAPDH

HD11

5YS

CEF

RPL12

OK10BM

BM2C3

MSB1

RP9

b

6C2

more striking di€erence is located beyond these ca. 100 ®rst a.a. in a region displaying a 12 a.a. deletion in the chicken protein. Our RT ± PCR experiments seemed to rule out the generation of the deleted protein as a result of an alternate splicing event (data not shown). Nevertheless, in terms of overall aminoacid composition attested by a richness in certain residues clustered in some particular regions, the three proteins are very similar. Therefore, three domains can be de®ned in the chicken Spi-1/PU.1 protein which roughly correspond to the three structural features ®rst delineated in the human and mouse proteins (Klemsz et al., 1990; Pongubala et al., 1992). The acidic sequence found in the amino-terminal part of the protein corresponding in chicken to residues 1-105, was shown to encode an activation domain in the mammalians homologues. The basic carboxyl-terminal sequence containing the DNA-binding or ETS-domain de®ned in the mouse protein is located in chicken between residues 141-250. In between these domains was found a PEST sequence to residues 106-140 implicated in the mouse Spi-1/PU.1 protein in protein-protein interaction (Pongubala et al., 1993). The PEST sequence, dubbed after a particular rich composition in proline, glutamic acid, serine and threonine residues, is also known as a destabilizing motif in numerous proteins (Rechsteiner and Rogers, 1996). The already cited deletion speci®c of the chicken protein is located in between the activation domain and the amino-terminal part of the PEST domain (Figure 2). Whether the deletion of 12 a.a. could a€ect the ability of Spi-1/PU.1 protein to interact via the PEST sequence to other protein remains to be addressed. However it is interesting to note that Ser127 (Ser148 for the mouse) required for a protein-protein interaction between Spi-1/PU.1 and NF-EM5/Pip-1 (Pongubala et al., 1993) was conserved in an identical surrounding context. Phosphorylation of this residue by type II Casein Kinase (CKII) was shown critical for the interaction. Similarly, other potential phosphorylation sites by CKII characterized by Pongubala et al. (1993), Ser41, Ser45, and Ser132 are all preserved in positions 34, 38, and 111 in the chicken protein.

Spi/PU.1

GAPDH

Figure 2 Comparison of the translated sequence of the chicken Spi-1/PU.1 cDNA with the human and mouse counterparts. Asterisks indicate sequence identity, whereas points indicate conservative modi®cation. AA residues are numbered relative to the ®rst putative amino terminus methionine. The boxed regions correspond to the a-helical and basic domains (or ETS domain) conserved in all Ets family members. Arrows delineate the di€erent structural or functional domains

Figure 3 Northern blot analysis of Spi-1/PU.1 total RNAs in chicken embryos tissues (a) and in the following chicken cell lines (b) 6C2 (AEV-transformed erythroblasts); BM2C3 and 5YS (myeloblasts transformed by AMV containing the myb oncogene); OK10 BM (macrophage-like cell line transtormed with the myc oncogene containing OK10 virus); HD11 (macrophages transformed by the myc oncogene containing MC29 virus); MSB1 (lymphoid cell line of the T-lineage infected by MDV virus); RPL12 and RP9 (cell lines of the B-lineage originated from ALV-induced lymphomas) (Coll et al., 1983). Total RNAs were isolated from frozen chicken tissues as described by Chirgwin et al. (1979) and using RNA PLUS system (Bioprobe systems) from cultured cells. RNAs were fractionated by electrophoresis on denaturing 2.2 M formaldehyde-1% agarose gels and transferred to Hybond N membranes (Amersham). Filters were hybridized with the 1138bp EcoRI fragment from the cDNA as a 32P-labeled probe. Spi-1/PU.1 is expressed as 2.9 kb and 2.1 kb transcripts. A glyceraldehyde 3-phosphate dehydrogenase (GAPDH) probe was used as a control for the amount of RNA applied

Characterization of the chicken Spi-1/PU.1 gene Z Kherrouche et al

1360

and to a lesser extend in B-lymphoma RP9, in myeloid 5YS and erythroid 6C2 cell lines. No expression was detected in the other cell lines tested, even after over exposure, including transformed T-cell (MSB1), B-cell

(RPL12) and myeloid (OK10BM) cell lines or primary culture of chicken embryos ®broblasts (CEF). Interestingly, positive (HD11) and negative (OK10BM) cells could be detected among v-myc transformed macro-

Figure 4 In vivo detection of Spi-1/PU.1 expression in El9 bursa of Fabricius and spleen. Dissected organs were ®xed in 4% paraformaldehyde, embedded in paran and processed for histological sections. In situ hybridization of the sections was adapted from the method of (Cox et al., 1984), as described by (Queva et al., 1992). Sense and antisense riboprobes were generated from Spi1/PU.1 ck inserts cloned in sense or antisense orientation in pSG5. The [35S]-labeled riboprobes were synthesized using the Promega in vitro transcription kit with the T7 primer according to the provided instructions. In situ hybridizations were counterstained with Hoechst 33258, a DNA ¯uorescent dye. Sections were observed either under U.V. illumination alone (b, f) or under both simultaneously U.V. and dark ®eld illumination (a, c, e, g). (a, b, c) Bursa of Fabricius at E19 show high levels of Spi-1/PU.1 mRNA in the B-cell follicles (f) while hybridization with the sense mRNA probe shows intense refringence around these follicles. (d) Adjacent section stained with Pappenheim method characterizes the B-cell follicles (f) in dark blue and granulocytes (g) in pink. (e, f) In the spleen, Spi-1/PU.1 transcripts are detected in cells of the nodal white pulp (w) which appeared in blue by the Pappenheim coloration (h). (g) The sense probe hybridization gives a refringent signal on the red pulp (r), colored in slightly violet on (h). (White bar: 160 mm/ black bar :80 mm)

Characterization of the chicken Spi-1/PU.1 gene Z Kherrouche et al

phage cell lines and very di€erent expression yields were observed in two AMV-transformed myeloid cells expressing the v-myb oncogene (BM2C3, 5YS). This stressed the heterogeneity in Spi-1/PU.1 expression between transformed cells of the same lineages independently from the expressed transforming oncogene. There is also no apparent correlation in our results with the di€erentiation state since v-myb transformed myeloid cells are supposed to be blocked at an earlier stage of myeloid di€erentiation than v-myc transformed macrophages (Ness et al., 1987). It is however worthy to note that this di€erentiation status was de®ned for freshly transformed cells and so probably hardly transposable to established cell lines. Another interesting result was brought by the faint expression in the AEV transformed 6C2 erythroblasts. These cells are blocked at a CFU-E stage of erythroid di€erentiation roughly corresponding to stages expressing Spi-1/PU.1 in mice (Galson et al., 1993). Moreover, on line with a role of Spi-1/PU.1 in erythroid progenitor cells in chicken as well, an enforced expression of human Spi-1/PU.1 protein in chicken erythroblasts was shown to inhibit normal differentiation (Quang et al., 1995). The low expression found in

6C2 could re¯ect a species speci®city or a blockade of these cells at a more advanced stage of di€erentiation since there is a marked drop in Spi-1/PU.1 expression in SFFV pro-erythroblasts induced to di€erentiate (Galson et al., 1993). However, further investigation is needed to clearly substantiate the role of Spi-1/PU.1 in chicken hematopoiesis. This could be achieved by analysis of valuable systems of hematopoietic lineage di€erentiation and commitment using the numerous conditional transforming avian retroviruses available in chicken (Graf et al., 1992). The 2.1 kb transcript appeared predominant in all positive cell lines and chicken tissues. The two mRNA detected with the Spi1/PU.1 probe could result from di€erential splicing of a primary Spi-1/PU.1 transcript. Alternatively, they could arise from the use of di€erent promoters or polyAdenylation sites. The possibility that one mRNA could derived from a known gene related to Spi-1/ PU.1, like Spi-B, is very unlikely since hybridizations were performed in stringent conditions. Indeed, these conditions do not allow cross reaction between the mouse Spi-1/PU.1 and Spi-B genes. To further characterize Spi-1/PU.1 in vivo expression, we performed in situ hybridization in isolated chicken

Figure 5 Spi-1/PU.1 mRNAs expression in thymus and heart from a chicken embryo at day 19. (a, b) In the thymus strong hybridization with the antisense Spi-1/PU.1 mRNA probe is detected in cells at the periphery of the medulla (m) and in the trabecula (t), and low signals are observed in cells of the cortex (c). (c) In contrast no signal above background is observed with the sense probe. (d) The Pappenheim coloration stained thymocytes in blue and macrophages in pink/violet (arrowhead). (e, f) Spi-1/ PU.1 transcripts are not detected in the heart. (White bar: 180 mm/ black bar: 80 mm)

1361

Characterization of the chicken Spi-1/PU.1 gene Z Kherrouche et al

hybridizations were performed using a Spi-1/PU.1 antisense mRNA probe, controlled by the Spi-1/PU.1 sense mRNA probe. Pappenheim coloration was used as mean to identify cell types in the hematopoietic organs (Kolmer and Boerner, 1945). Figure 4a showed a high expression of the Spi-1/PU.1 mRNAs in the bursa of Fabricius whereas sense mRNA probe revealed distinct highly refringent areas (Figure 4c). The expressing cells, were clearly identi®ed as B-cell follicles (f) in serial sections treated by the Pappenheim method of coloration, which stained in dark blue the immature lymphoid cells (Figure 4d). Bursa is, for the chicken before hatching, the site of early B-cell development. Thus as expected from B-cell expression of Spi-1/PU.1 (Galson et al., 1993), bursal follicles showed an intense signal. In the spleen at El9, Spi-1/PU.1 mRNAs were expressed in cells of the nodular white pulp (w) (Figure 4e), stained in blue by the Pappenheim coloration, whereas the red pulp (r), stained in violet, constituting the erythroid area, did not express Spi-1/PU.1 but also appeared highly refringent with control sense mRNA probe (Figure 4g). Indeed, the Pappenheim method identi®ed within the nodular white pulp some large dark blue cells, probably B-cells, and some large cells with a pink cytoplasm that could correspond to macrophages (Figure 4h). High expression of Spi-1/PU.1 mRNAs was detected in the thymus at E19, the lack of detection by Northern blot was probably due to the lesser amounts of loaded RNA as well as to the lesser sensitivity of this technique compared to in situ hybridization. At El9 the third and last period of colonization by thymocyte precursors is ongoing. However, the signal was restricted to large round cells localized at the cortical-medullary junction of the medulla (m) and in the trabecula (t) in between the

protein extract

HD11

control sense reticulocytes

PI 1.6F

1.6F

PI 1.5K

1.6F

1.5K

PI 1.6F

1.5K

1.6F

Serum

PI 1.5K

hematopoietic organs at 19 days of embryonic development: spleen, bursa of Fabricius and thymus. We also wanted to determine the Spi-1/PU.1 expression status in the chicken heart because of the controversial results previously obtained in the mouse (Klemsz et al., 1990; Paul et al., 1991; Galson et al., 1993). In situ

1.5K

1362

kDa — 62 — 47.5

— 32.5 — 25 Figure 6 Analysis of Spi-1/PU.1 proteins from in vitro transcription/translation and HD11 cell line. Chicken Spi-1/ PU.1 cDNA was transcribed and translated with the TNT T7 coupled rabbit reticulocyte lysate system according to the manufactor conditions (Promega). Proteins were double immunoprecipitated with two polyclonal antisera directed against the 110 N-terminal amino acids of the mouse Spi-1/PU.1 protein (15K and 1-6F) or with preimmune sera as control (PI). The products were separated on a SDS 10% polyacrylamide gel electrophoresis and visualized by autoradiography. The size (kDa) and positions of molecular weight standards are indicated on the right

a Competitor oligonucleotide Reticulocyte lysate

Spi/PU.1

Free probe

– –

– +

In Vitro 22 26 + +

EBS

+

21 +

– +

26 +

HD11 24 22 + +

21 +

Competitor oligonucleotide Nuclear extract

Spi/PU.1

Free probe

Characterization of the chicken Spi-1/PU.1 gene Z Kherrouche et al

lobules of the cortex (c) (Figure 5a), exhibiting the characteristic morphology and distribution of the thymic macrophages stained in pink/violet by Pappenheim coloration. The thymocytes, which represent the main thymic population, appeared as small round blue cells and neither the cortical immature thymocytes nor the more mature medullary thymocytes expressed the Spi-1/PU.1 gene con®rming the absence of Spi-1/PU.1 in the T cell lineage (Figure 5d). As a control, we obtained pratically no signal background with a sense probe (Figure 5c). In the heart at E l9, we could not detected any expression of the Spi-1/PU.1 mRNAs (Figure 5e). Thus, the localization and the identity of cells expressing the chicken Spi-1/PU.1 gene correlates

perfectly with the cell-type restricted pattern of the mammalian genes. The chicken Spi-1/PU.1 gene directs the synthesis of a 40 kDa protein To experimentally characterize the coding potential of the isolated chicken Spi-1/PU.1 cDNA, the pSG5-ckSpi-1/PU.1 construct was transcribed and translated in vitro using T7 RNA polymerase and rabbit reticulocytes lysate. Two proteins of 40 kDa and 35 kDa were detected by SDS ± PAGE and were both recognized by antisera directed against the 110 N-terminal amino-

b

c Serum Reticulocyte lysate

– –

In Vitro 1.6F – + +

PI +

– +

HD11 1.6F +

PI +

Serum Nuclear extract

Spi/PU.1

Spi/PU.1

Free probe

Figure 7 Electrophoretic mobility shift assays (EMSA) were performed with Spi-1/PU.1 proteins from programmed reticulocyte lysates (2 ml) or HD11 nuclear extract puri®ed as described (Sambrook et al., 1989) (5 mg). The oligonucleotide use as probe is the Spi-1/PU.1 site within the Ig-J chain promoter (Shin and Koshland, 1993). Complexes were visualized on 6% non-denaturing polyacrylamide gel. (a) Speci®city of binding was tested by using 200 fold molar excess of cold speci®c oligonucleotide (26) or the di€erent oligonucleotide competitors listed in (b) (Ansieau et al., 1993; Wasylyk et al., 1991). (c): Preimmune serum (PI) and speci®c antiserum (1.6F) against Spi-1/PU.1 were added before the probe

1363

Characterization of the chicken Spi-1/PU.1 gene Z Kherrouche et al

1364

acids of the murine Spi-1/PU.1 protein (Delgado et al., 1994) (Figure 6). Owing the conservation between the predicted chicken sequence and the murine protein in this region, the results attested that the detected proteins are actually derived from the presumed ORF. As was the case for the mammalian proteins, we found a discrepancy between the predicted (29 kDa) and the observed molecular mass for the slower migrating species. This is probably the consequence of the high amount of proline residues found in all these Spi-1/PU.1 proteins. The fast migrating species most likely derived by proteolytic cleavage of the most abundant and slower migrating protein. To assess that the Spi-1/PU.1 proteins observed in vitro have cellular counterparts, we performed immunoprecipitation analyses using the same antibodies on total extracts of HD11 cells expressing high amounts of RNA, labeled with 35S-methionine/ cysteine. Figure 6 shows that two proteins were speci®cally detected. Their size corresponded roughly to the size of the bands detected in programmed reticulocytes lysates. This holds especially for the 40 kDa species which thus likely corresponds to the protein encoded by the isolated cDNA. However, the smaller band displayed slight di€erences in size and in relative intensity when compared to reticulocytes lysate. This prompted us to question the synthesis in HD11 cells of two Spi-1/PU.1 isoforms which could be encoded by the two detected mRNAs. According to this hypothesis, the isolated cDNA would be representative of a mRNA encoding the 40 kDa species and the lowest species would be encoded by an alternative RNA still to be isolated. To test this hypothesis, we performed RT-PCR experiments with RNA from HD11 cells and using couples of oligonucleotides encompassing the coding sequence of the cDNA. We particularily emphasize on the 12 amino acids deletion by using couples of oligonucleotides ¯anking the concerned region. For each PCR, we failed to obtain any extra band in addition to the one expected from the known sequence. Without ruling out de®nitively this possibility, our results make very unlikely the existence of an alternate chicken Spi-1/PU.1 protein isoform. It is likely instead that the two products arose from posttranslational modi®cations like phosphorylation as well documented for the mammalian proteins (Carey et al., 1996). As a non mutually exclusive explanation, proteolytic cleavages events could participate to this diversity. Indeed, the PEST sequence is presumed to be targeted by endopeptidases involved in proteolysis of intracellular proteins (Rechsteiner and Rogers, 1996); this could explain our proteolytic product, already observed by Pongubala et al. (1992) and Carvalho et al. (1993). The slight size di€erence between in vitro and in vivo studies could be explained as well by in vivo speci®c post translational modi®cations. DNA-binding activity of the Spi-1/PU.1 protein We performed gel retardation assays, using as a probe the Spi-1/PU.1 oligonucleotide from Ig-J-chain promoter (n826), which contains a single Spi-1/PU.1 binding site. This binding site was reported to be

speci®c for Spi-1/PU.1 proteins since other members of the ETS family failed to recognize it (Shin and Koshland, 1993). For these experiments we used either HD11 cells nuclear extracts or reticulocytes lysates programmed with Spi-1/PU.1 RNAs transcribed in vitro. Figure 7 illustrates typical gel retardation assays obtained with the radiolabeled Ig-J Spi-1/PU.1 oligonucleotide and these two extracts. A single retarded complex was observed between the Ig-J Spi-1/PU.1 oligonucleotide and the Spi-1/PU.1 protein synthesized in vitro. Competitions experiments were realized with unlabeled oligonucleotides listed in Figure 7b. As expected, the protein-DNA complex corresponding to this band, was competed by a 200-fold molar excess of a nonlabeled oligonucleotide (n826) but not by the same excess of an unrelated Sp1 oligonucleotide (n821) (Figure 7a). The speci®city of interaction was further con®rmed by lack of competition with an excess of a canonical Ets binding site (EBS). On the contrary, we found a competition with a Spi-1/PU.1 binding site derived from the chicken c-mil/raf promoter (n822). A more complex pattern was found in HD11 nuclear extracts with the same probe. Thus, all the complexes detected were competed by an excess of unlabeled oligonucleotide, but not by the aspeci®c Sp1 competitor. These bands were also competed to some extend by oligonucleotides containing the c-mil/raf promoter Spi-1/PU.1 binding site (n822 and 24). The slight di€erence in competition obtained using these two competitors could be explained by a cooperative e€ect between the Spi-1/PU.1 and Sp1 sites in the cmil/raf promoter (unpublished data). The speci®city of the complex observed in vitro was further stressed by lack of detection of the corresponding bandshift after incubation with Spi-1/PU.1 antisera, whereas the speci®c complex was una€ected by incubation with the preimmune serum (Figure 7c). In nuclear extracts, the slower migrating complex was supershifted after incubation with the antiserum stressing the presence of Spi-1/PU.1 proteins in this complex. Another band speci®c of the antiserum treatment can be also observed, coming probably from the supershifting of part of the faster migrating bands. Using this probe, Shin and Koshland (1993) had shown a major single complex with nuclear extracts from S194 myeloma cells, but speci®ed that a longer exposure of the gel revealed additional slower migrating species which did not react with the anti-PU.1 antibody. And strikingly, two antibody-reacting complexes from HD11 cells were also shown to bind to a Spi-1/PU.1 binding site in the enhancer of the chicken lysozyme gene (Ahne and StraÈtling, 1994). The detection of these two complexes could be considered at the light of posttranslational events modifying the mammalian counterparts. This is particularly the case of phosphorylations by CKII and SAP kinases previously described by Pongubala et al. (1993) and Mao et al. (1996). Moreover some phosphorylations were shown at least in one case to trigger protein-protein interactions leading to multiproteins binding complexes. It is tempting to speculate that some of the speci®c Spi-1/ PU.1 complexes shown in HD11 were the result of such phosphorylation-dependent or -independent recruitment events. The other complexes, not recognized by the speci®c Spi-1/PU.1 antisera may occur

Characterization of the chicken Spi-1/PU.1 gene Z Kherrouche et al

from di€erential in vivo cleavage resulting in products with a preserved DNA binding domain but lacking the N-terminal part recognized by the antiserum. Alternatively, they could correspond to related Spi-1/ PU.1 proteins, since mammalian Spi-1/PU.1 and Spi-B exhibiting similar binding properties as was shown (Ray-Gallet et al., 1995). It is indeed worthy to note that the antisera used in this study do not react with mammalian Spi-B proteins. Thus, unless an hypothetical better conservation between chicken members of the SPI group, putative chicken Spi-B products were most likely present in shifted than in supershifted complexes. In the same way, two comigrating complexes, involving Spi-1/PU.1 and Spi-B, were obtained in EMSA performed with Raji cell nuclear extracts and Spi-1/PU.1 binding site of the EBNA2RE (Laux et al., 1994). Although we did not ®nd a competition of the HD11 complexes with an EBS oligonucleotide (data not shown) participation of other Ets proteins in these complexes could not be completely ruled out. The EBS site was indeed derived for high anity binding of ets-1 related members of the family (Wasylyk et al., 1991).

PU.1 antisense, showing that in HD11 cells the constitutive Spi-1/PU.1 transactivating capacity was mainly mediated by the responsive element mutated in this construct. Our transactivating experiments show that despite the less homology score between the amino-terminal part of the chicken and mammalians proteins, the transactivating function seems identical. This result is not really surprising because aminoacid residues mapped by Klemsz and Maki (1996) as required for the activation of transcription were conserved in chicken. Thus, 13 out of the 13 acidic residues located between positions 2 and 74 in the chicken and de®ning the acidic transactivation domain were preserved.

a

Transactivating properties of the chicken Spi-1/PU.1 protein In order to test the transactivation capacity of chicken Spi-1/PU.1 protein, we cotransfected HeLa cervical carcinoma cells which lack endogenous Spi-1/PU.1 proteins with the pSG5-ck-Spi-1/PU.1 expression vector and a luciferase reporter plasmid (Ray et al., 1992; Chen et al., 1995). In this reporter the luciferase expression was under the control of the thymidine kinase minimal promoter and three copies of the Spi1/PU.1 binding site present in the c-fes promoter (ptkluc fes) (Ray-Gallet et al., 1995; Righi, unpublished results). As shown in Figure 8a, we could obtain an eight-fold activation of the reporter by the chicken Spi-1/PU.1 in these cells. The e€ect was actually mediated by Spi-1/PU.1 binding site since it was not reproduced with a reporter plasmid carrying fes mutated elements (ptk-luc fes mut). Similar transactivation results (ninefold) were obtained in chicken HD11 cells with wild-type elements. However, we could detect in these cells a signi®cative activity of the reporter with mutated elements (fourfold activation). This cell-speci®c e€ect on the mutated reporter has to be considered at the light of recent data showing a putative second Spi-1/PU.1 responsive element in the c-fes promoter (Heydemann et al., 1996). In the mutated version used in our assay, this second binding site is present and could be functional in HD11 cells (Figure 8b). Transactivating properties of endogenous Spi-1/PU.1 proteins in HD11 cells were analysed by cotransfection of an antisense pSG5-Spi1/PU.1 expression vector and the reporter vectors described above. Figure 8c shows that antisense Spi-1/ PU.1 protein reduced about 50% the basal activation status of the wild-type reporter, pointing out the role of endogenous Spi-1/PU.1 protein in this e€ect. As expected, the mutated reporter displayed a residual activity similar to that obtained with the antisense construct on the wild-type reporter. This activity was not further decreased in the presence of the Spi-1/

b

c

Figure 8 Spi-1/PU.1 transactivate the expression of a reporter plasmid containing three copies of an oligonucleotide containing the Spi-1/PU.1 responsive element from the c-fes promoter (ptkluc fes) (Ray-Gallet et al., 1995). Transfections were carried out in HeLa cells (a) or HD11 cells (b) on 500 ng of ptk-luc fes or a mutated ptk-luc fes mut as reporters, with or without 100 ng of Spi-1/PU.1 expression vectors (pSG5- Spi-1/PU.1). Cells were transfected in 6 well-plates using 10 ml lipofectamine (GibcoBRL), harvested 18 h after transfection and lysed in the Promega reporter lysis bu€er. (c) E€ect of antisense Spi-1/PU.1 expression vector on ptk-luc fes or ptk-luc fes mut in HD11 cell line containing endogenous Spi-1/PU.1 proteins. For the three panels, the mean value of six expression experiments performed with pSG5 was used as standard to calculate the fold (a and b) or the percentage activations (c). All the experiments were performed at least twice using three di€erent plasmid preparations

1365

Characterization of the chicken Spi-1/PU.1 gene Z Kherrouche et al

1366

Between positions 69 to 78, the cluster of ®ve glutamine residues de®ning the glutamine-rich transactivating subdomain was strictly conserved as well. In conclusion, in this paper we showed that the chicken Spi-1/PU.1 gene displayed functional and restricted expression properties similar to its mammalian counterparts. These results stress the conservation of this transcription factor playing a key role in hematopoiesis of vertebrate species. They provide also interesting tools to further study this role using di€erentiation and developmental models of better or easier access in the chicken than in mammals.

Acknowledgements We wish to thank Dr Simon Saule and Dr Marco Righi for critical reading of the manuscript. We are grateful to Brigitte Quatannens for the dissections and to Jean Marc Merchez for preparation of the photographs. We feel particularly indebted to Helene Pelzar and Serge Plaza for helpful discussions. Claire Montpellier, Virginie Mattot and Frederic Gilles o€ered friendly advice. This study was supported by grants from the Centre National de la Recherche Scienti®que, the Institut National de la Sante et de la Recherche MeÂdicale, the Institut Pasteur de Lille, the Association pour la Recherche sur le Cancer and the Ligue Nationale contre le Cancer. EMBL accession number Y12225.

References Ahne B and StraÈtling WH (1994). J. Biol. Chem., 269, 17794 ± 17801. Ansieau S, Plaza S, Ferreira E, Dozier C and Stehelin D (1993). Genomics, 18, 537 ± 545. Carey JO, Posekany KJ, deVente JE, Pettit GR and Ways DK (1996). Blood, 87, 4316 ± 4324. Carvalho M, Kirkland M and Derse D (1993). J. Virol., 67, 6586 ± 6595. Chen H, Ray GD, Zhang P, Hetherington CJ, Gonzalez DA, Zhang DE, Moreau-Gachelin F and Tenen DG (1995). Oncogene, 11, 1549-1560. Chirgwin JM, Przybyla AE, MacDonald RJ and Rutter WJ (1979). Biochemistry, 18, 5294 ± 5299. Coll J, Saule S, Martin P, Raes MB, Lagrou C, Graf T, Beug H, Simon IE and Stehelin D (1983). Experimental Cell Research, 149, 151 ± 162. Cox KH, DeLeon DV, Angerer LM and Angerer RC (1984). Dev. Biol., 101, 485-502. Delgado MD, Hallier M, Meneceur P, Tavitian A and Moreau-Gachelin F (1994). Oncogene, 9, 1723 ± 1727. Eisenbeis CF, Singh H and Storb U (1993). Mol. Cell. Biol., 13, 6452 ± 6461. Galson DL, Hensold JO, Bishop TR, Schalling M, D'Andrea AD, Jones C, Auron PE and Housman DE (1993). Mol. Cell. Biol., 13, 2929 ± 2941. Goebl MK (1990). Cell, 61, 1165 ± 1166. Graf T, McNagny K, Brady G and Frampton J (1992). Cell, 70, 201 ± 213. Hallier M, Tavitian A and Moreau-Gachelin F (1996). J. Biol. Chem., 271, 11177-11181. Heydemann A, Juang G, Hennessy K, Parmacek MS and Simon MC (1996). Mol. Cell. Biol., 16, 1676 ± 1686. Karim FD, Urness LD, Thummel CS, Klemsz MJ, McKercher SR, Celada A, Van BC, Maki RA, Gunther CV and Nye JA (1990). Genes Dev., 4, 1451 ± 1453. Kistler B, P®sterer P and Wirth T (1995). Oncogene, 11, 1095 ± 1106. Klemsz MJ, McKercher SR, Celada A, Van BC and Maki RA (1990). Cell, 61, 113 ± 124. Klemsz MJ and Maki RA (1996). Mol. Cell. Biol., 16, 390 ± 397. Kolmer JA and Boerner F (1945). Approved Laboratory Technic. 4nd ed. D. Appleton-Century Company: New York, pp. 71 ± 72. Kominato Y, Galson D, Waterman WR, Webb AC and Auron PE (1995). Mol. Cell. Biol., 15, 59 ± 68. Laudet V, Niel C, Duterque CM, Leprince D and Stehelin D (1993). Biochem. Biophys. Res. Commun., 190, 8 ± 14. Lautenberger JA, Burdett LA, Gunnell MA, Qi S, Watson DK, O'Brien SJ and Papas TS (1992). Oncogene, 7, 1713 ± 1719. Laux G, Adam B, Strobl LJ and Moreau-Gachelin F (1994). EMBO J, 13, 5624 ± 5632.

Mao C, Ray-Gallet D, Tavitian A and Moreau-Gachelin F (1996). Oncogene, 12, 863 ± 873. McKercher SR, Torbett BE, Anderson KL, Henkel GW, Vestal DJ, Baribault H, Klemsz M, Feeney AJ, Wu GE, Paige CJ and Maki RA (1996). EMBO J, 15, 5647 ± 5658. Moreau-Gachelin F, Tavitian A and Tambourin P (1988). Nature, 331, 277 ± 280. Moreau-Gachelin F, Ray D, Mattei MG, Tambourin P and Tavitian A. (1989). Oncogene, 4, 1449 ± 1456. Moreau-Gachelin F. (1994). Biochem. Biophys. Acta, 1198, 149 ± 163. Moreau-Gachelin F, Wendling F, Molina T, Denis N, Titeux M, Grimber G, Briand P, Vainchenker W and Tavitian A. (1996). Mol. Cell. Biol., 16, 2453 ± 2463. Nelsen B, Tian G, Erman B, Gregoire J, Maki R, Graves B and Sen R. (1993). Science, 261, 82 ± 86. Ness SA, Beug H and Graf T. (1987). Cell, 51, 41 ± 50. Oelgeschlager M, Nuchprayoon I, Luscher B and Friedman AD. (1996). Mol. Cell. Biol., 16. 4717 ± 4725. Paul R, Schuetze S, Kozak SL, Kozak CA and Kabat D. (1991). J. Virol., 65, 464 ± 467. Pongubala JM, Nagulapalli S, Klemsz MJ, McKercher SR, Maki RA and Atchison ML. (1992). Mol. Cell. Biol., 12, 368 ± 378. Pongubala JM, Van BC, Nagulapalli S, Klemsz MJ, McKercher SR, Maki RA and Atchison ML. (1993). Science, 259, 1622 ± 1625. Pongubala JM and Atchison ML. (1997). Proc. Natl. Acad. Sci. U S A, 94, 127 ± 132. Quang CT, Pironin M, von Lindern M, Beug H, Ghysdael J. (1995). Oncogene, 11, 1229 ± 1239. Queva C, Ness SA, Grasser FA, Graf T, Vandenbunder B and Stehelin D. (1992). Developolent, 114, 125 ± 133. Rao G, Rekhtman N, Cheng G, Krasikov T and Skoultchi AI. (1997). Oncogene, 14, 123 ± 131. Ray D, Bosselut R, Ghysdael J, Mattei MG, Tavitian A and Moreau-Gachelin F. (1992). Mol. Cell. Biol., 12, 4297 ± 4304. Ray-Gallet D, Mao C, Tavitian A and Moreau-Gachelin F. (1995). Oncogene, 11, 303 ± 313. Rechsteiner M and Rogers SW. (1996). Trends Biochem. Sci., 21, 267 ± 271. Rivera RR, Stuiver MH, Steenbergen R and Murre C. (1993). Mol. Cell. Biol., 13, 7163 ± 7169. Rosmarin AG, Caprio DG, Kirsch DG, Handa H and Simkevich CP. (1995). J. Biol. Chem., 270, 23627 ± 23633. Sambrook J, Fritsch EF and Maniatis T. (1989). Molecular Cloning: A Laboratory Manual. 2nd ed. Cold Spring Harbor Laboratory Press: Cold Spring Harbor, NY. Schuetze S. Stenberg PE and Kabat D. (1993). Mol. Cell. Biol., 13, 5670 ± 5678. Scott EW, Simon MC, Anastasi J and Singh H. (1994). Science, 265, 1573 ± 1577.

Characterization of the chicken Spi-1/PU.1 gene Z Kherrouche et al

Shin MK and Koshland ME (1993). Genes Dev., 7, 2006 ± 2015. Tondravi MM, McKercher SR, Anderson K, Erdmann JM, Quiroz M, Maki R and Teitelbaum SL. (1997). Nature, 385, 81 ± 84.

Voso MT, Burn TC, Wulf G, Lim B, Leone G and Tenen DG. (1994). Proc. Natl. Acad. Sci. U S A, 91, 7932 ± 7936. Wasylyk C, Gutman A, Nicholson R and Wasylyk B. (1991). EMBO J, 10, 1127 ± 1134.

1367