A minimal serpin promoter with high activity in ... - Semantic Scholar

ã

The Hematology Journal (2001) 2, 150 ± 160 2001 The European Haematology Association All rights reserved 1466 ± 4680/01 $15.00 www.nature.com/thj

A minimal serpin promoter with high activity in haematopoietic progenitors and activated T cells Lynne Hampson*,1,5, Ian N Hampson2,5, Charolyn K Babichuk3, Laura Cotter1, R Chris Bleackley3, T Michael Dexter1 and Michael A Cross4 1

CRC Department of Hematopoietic Cell and Gene Therapeutics, Paterson Institute for Cancer Research, Wilmslow Road, Manchester M20 4BX, UK; 2Department of Obstetrics and Gynaecology, St. Marys Hospital, Hathersage Road, Manchester M13 0JH, UK; 3Department of Biochemistry, University of Alberta, Edmonton, Alberta T6G 2H7, Canada; 4IZKF, University of Leipzig, 04103 Leipzig, Germany

Introduction: The serine protease inhibitor Serpin 2A is highly expressed in ex vivo bipotent granulocyte/macrophage progenitor cells and in cultured myeloid stem cells. The gene undergoes rapid down-regulation as these cells are induced to dierentiate, and constitutive expression in cultured myeloid stem cells retards maturation. Serpin 2A is also expressed in T cells as a consequence of activation. We now report analysis of the upstream regulatory elements that control Serpin 2A transcription. Materials and methods: Using primer extension and rapid ampli®cation of cDNA ends the transcription start site of the Serpin 2A gene was mapped, and a 1.2 Kb genomic upstream fragment cloned and sequenced. Promoter activity and protein binding of deletion and sitedirected mutant constructs were analysed by transient transfection and by electrophoretic mobility shift assays. Results: A minimal promoter fragment was identi®ed with high activity dependent on NF-k and Moloney murine leukaemia enhancer factor LVa binding sites in both myeloid stem cells and activated T cells. NF-k was shown to be the main DNA binding protein in T cells, whereas that in haematopoietic stem cells appears to be novel. Conclusion: Serpin 2A promoter activity in T cells is due predominantly to NF-k binding to its consensus site. Activity in haematopoietic stem cells appears to be mediated by a novel protein, which recognises the NF-k consensus only in the context of ¯anking sequences. This concise regulatory element may be of potential value in gene therapeutic applications. The Hematology Journal (2001) 2, 150 ± 160 Keywords:

Serpin; T cells; promoter; NF-k; LVa; stem cells

Introduction The continuous generation of mature haematopoietic cells is maintained by a small population of primitive stem cells.1 Features which distinguish these stem cells from their mature progeny are their multipotentiality and capacity for self-renewal.2 The mechanisms which balance self-renewal versus dierentiation in haematopoietic stem cells are unclear, but are likely to involve regulated gene expression, with transcription factors playing a pivotal role. The use of multipotent haematopoietic cell lines (FDCP-Mix) to investigate the molecular changes *Correspondence: Dr L Hampson, Department of Obstetrics and Gynaecology, St Mary's Hospital, Hathersage Road, Manchester M13 0JH, UK Tel: +44 161 276 6478; Fax: +44 161 276 6134; E-mail: [email protected] 5 L Hampson and I Hampson contributed equally to this work Received 22 May 2000; accepted 20 November 2000

which accompany haematopoietic development has been previously described.3 ± 5 These cell lines are nonleukaemic, karyotypically normal and can either undergo extensive self-renewal when cultured in high levels of IL-3 or dierentiate along several haematopoietic lineages in response to varying combinations of growth factors. Using the technique of chemical cross-linking subtraction (CCLS),6 we have isolated a cohort of genes which down-regulate during FDCP-Mix differentiation. One such gene is the previously reported Serpin 2A (S2A)7,8 which shows extensive transcriptional down-regulation with corresponding decreases in the levels of both RNA and protein during differentiation of FDCP-Mix cells.8 S2A is also highly expressed in primary ex vivo bi-potent haematopoietic progenitors8 and belongs to a family of serine protease inhibitors which regulate proteolysis in a wide range

S2A transcriptional regualtion L Hampson et al

151

of circumstances including dierentiation of neuronal cells, muscle cells and enteric villus.9 ± 11 Indeed, forced expression of S2A in multipotent stem cells is associated with an increase in clonogenic potential and a delay in maturation, implicating a role for this Serpin in the control of haemopoiesis.8 It is interesting in this context that caspases, which can be activated by serine protease activity, have recently been reported to play a role in the feedback regulation of erythropoiesis via the degradation of speci®c transcription factors.12 Cytotoxic T cells have also been shown to express S2A at high levels and this expression is concomitant with the acquisition of cytotoxic function.8 While the precise role of S2A in the regulation of haematopoietic dierentiation and T-cell killing remains unclear, studies of the regulation of progenitor speci®c genes should further our understanding of the transcriptional environment which determines stem cell behaviour, and help to identify regulatory elements of potential value in gene therapeutic applications.

Materials and methods Cell culture FDCP-Mix cells The origin and detailed characterisation of the FDCP-Mix cell line have been described previously.3,13 Cells were maintained with a selfrenewing phenotype by culturing in 2% (vol:vol) X63 AG81653 cell-conditioned medium as a source of IL-3, 20% (vol:vol) horse serum (various suppliers: Gibco, Paisley, UK and Flow Labs. Rickmansworth, UK) and Iscoves medium (Gibco). Myeloid dierentiation was achieved by washing the cells twice in PBS, then transferring to Iscoves medium containing 20% foetal calf serum (Sigma, Poole, UK), 1 u/ml IL-3, 50 u/ml GM-CSF (Biogen, Geneva, Switzerland) and 100 u/ml G-CSF (Amgen, Thousand Oaks, CA, USA). All cultures were grown at 378C in 5% CO2 (vol:vol) in air.

129/St with a rat Serpin cDNA was kindly provided by Dr J Inglis (MRC Human Genetics Unit, Edinburgh, UK). This had previously been shown to contain the S2A gene7 in addition to a second related Serpin. A 5.8 Kb EcoRI-BamHI restriction fragment showing homology to a 5' S2A cDNA probe was sub-cloned into the low copy number vector PCS100. This fragment was initially restriction mapped and the location of S2A exon 1 identi®ed by hybridisation of restriction fragments with a radioactively labelled RACE product.

Thermostable reverse transcriptase primer extension Fifty ng of primer 1 (5' GACATCTGGGCAGCCAAAGAC 3') was end labelled with T4 kinase and g-32P-ATP (Amersham Pharmacia Biotech, Little Chalfont, UK). Twenty-®ve ng of this was mixed with 1 mg of FDCP-Mix poly (A)+ mRNA prepared from a self-renewing population, or with control RNA, in 20 mM Tris/HCl pH 8.3, and 180 mM KCl in a volume of 25 ml, and heated to 658C for 5 min in a thermocycler (Coy Corporation, Grass Lake, USA). A further 25 ml volume containing 0.2 mM of each dNTP, 2 mM MnCl2, 10 units of thermostable rTth polymerase (Perkin Elmer Applied Biosystems, Warrington, UK) was added and incubated at 708C for 10 min. Following this the reaction was subjected to 5 cycles of: 1 min 938C, 1 min 20 s 638C and 10 min 708C. Products were separated on a 3.5% sequencing gel and visualised by phosphorimaging (Molecular Dynamics, Amersham Pharmacia Biotech).

Thermostable reverse transcriptase RACE

MTL-2.8.2. The IL-2 dependent cytotoxic T-cell line was generated from CBA/J mice as previously described.14 Cells were cultured as for primary splenocytes.

FDCP-Mix mRNA was subjected to a modi®ed version of the RACE technique, using a thermostable transcriptase to minimise artefacts associated with secondary structure, and the resulting products cloned. Brie¯y 0.5 mg of FDCP-Mix poly (A)+ mRNA was mixed with 0.5 mg of an oligonucleotide complementary to the 5' coding sequence of S2A, (primer 2: 5' AGTCGAGGACAGCTCCT 3') in a volume of 50 ml containing 10 mM Tris/HCl (pH 8.3), 90 mM KCl, 200 mM of each dNTP, 1 mM MnCl2 and 10 units of thermostable rTth polymerase (Perkin Elmer Applied Biosystems). This was incubated for 5 cycles of 10 min at 708C and 1 min at 948C after which the product was eluted through a Sephadex G50 column, phenol/ chloroform extracted and precipitated. This material was then A-tailed and ampli®ed using the method of Frohman15 with the following primers: A (5' TCTGCACGCGTAATACGACTCGAGT163'), B (5' GATCTGCACGCGTAATACGACTCGAG 3'). Products were separated on a 1.5% agarose gel and cloned into the PCR cloning vector pGEM-T (Promega, Southampton, UK).

Isolation and mapping of the S2A upstream region

Plasmids

A Cosmid clone (Cos2A) obtained by screening a cosmid genomic library from the inbred mouse strain

The promoterless luciferase reporter plasmid pXP2 used in FDCP-Mix transfections was obtained from Steve

Primary splenocytes Primary splenocytes were obtained from 6 ± 12 week-old Balb/c mice. Spleen tissue was ground through a ®ne wire screen in RHFM medium (RPMI supplemented with 20 mM HEPES (pH 7.5), 100 mM b-mercaptoethanol and 10% foetal bovine serum) and 60 u/ml human recombinant IL-2 and the cells pelleted. Red blood cells were lysed with ammonium chloride lysis buer. Primary splenocytes were stimulated with 5 mg/ml conconavalin A (Sigma) and 1 : 500 dilution of hamster anti-mouse aCD3 monoclonal antibody supernatant.

The Hematology Journal


152

Sequence analysis Routine sequence analysis and searches for consensus transcription factor binding sites were performed using software from the University of Wisconsin GCG package.

tion conditions; cells from six electroporations for each construct were pooled, maintained for a recovery period of 4 h at 378C in medium supporting self renewal3 and then puri®ed over Nycoprep gradients. The washed cells were then aliquoted in triplicate into self renewal and dierentiation medium3 ± 5 and incubated for a further 44 h at 378C before assay. Each reporter construct was assayed in triplicate. Cells were harvested, washed in PBS and re-suspended in 100 ml of 16 Lysis buer (100 mM Potassium Phosphate Buer pH 7.8, 8 mM MgCl2, 1% Triton6100, 1% BSA, 15% Glycerol, 0.1 mM PMSF and 1 mM DTT). After a brief spin 80 ml of cleared lysate over 4 s following injection of 100 ml of 1 mM DLuciferin, 0.8 mM ATP in 16 Lysis buer as above. Since internal control plasmids of the type suitable for T-cell transfections can interfere with test promoter activity in FDCP-Mix cells (unpublished), FDCP-Mix transfections were performed in triplicate and repeated with independent DNA preparations to con®rm results.

Transient transfection assays

Generation of deletion fragments

T cells and primary splenocytes Transient transfections were performed using a previously described DEAE dextran transfection procedure optimised for cytotoxic T cells.19 Three aliquots of cell lysates were measured for 20 s following injection of Luciferin reagent in a LUMAT LB9501 luminometer (Berthold Systems Incorporated, Pforzheim, Germany). b-galactosidase assays were performed as described.20 Final activities are given as luciferase/b-galactosidase values.

A PCR strategy was employed using the 1.6 Kb StuStu fragment (Figure 1) as template for all reactions. The single 3' primer used (ctcggtaccGATATCTGGCTCCTG, S2A position 5' to 3' is +46 to +32 nucleotides) incorporated a 5' CTC clamp and a KpnI site (lowercase) and was paired with one of the following 2A 5 primers each having a 5' CTC clamp and a BamHI site (lowercase) for each deletion fragment: (ctcggatccTTCCCAGAAATCACC 2A position 7117 to 7112); (ctcggatccTCCCACAGCTCCTT 2A position 7258 to 7244); (ctcggatccGCTAAAGCCTGTCAA 2A position 7451 to 7437); (ctcggatccTGCTCCAGTCTCTCT 2A position 7645 to 7631); (ctcggatccAGGCATGGGTGGATA 2A position 7885 to 7871); (ctcggatccCAGGCATCAAGAATC 2A position 71170 to 71157). These primer mixes were used to amplify the various deletion fragments using Taq polymerase (Boehringer Mannheim, Lewis, UK) according to the manufac-

Nordeen16 as was the positive thymidine kinase control plasmid pTK109. For the T-cell assays the positive control vector was an 828 bp promoter fragment of CCP1 (a cytotoxic T-cell serine protease)17 cloned into p19luc. The p19luc plasmid was supplied by JR De Wet.18 PCS 100 is a general purpose cloning vector and was obtained from Prof D Sherratt (Dept. Biochemistry, Glasgow University, UK).

Sequencing All sequencing reactions were prepared using an ABI PRISM dye terminator cycle sequencing kit according to manufacturer's instructions and analysed on an ABI 373A DNA sequencer.

FDCP-mix cells Cells grown to 46105 cells/ml were re-suspended in complete medium at a density of 16107 per ml. Twenty mg of each DNA was added to 400 ml of cell suspension in an electroporation cuvette, mixed gently then pulsed (Easiject plus, Eurogentec, Seraing, Belgium) at 1050 mF/280 V.21 Cells were normally re-suspended in 2 ml of fresh medium and incubated at 338C overnight. For comparisons of promoter activity under self-renewal and differentia-

Figure 1 Restriction map of the 5' end of the S2A gene. The position of exon 1 is indicated, as are the StuI-StuI and StuI-EcoRV fragments utilized for promoter isolation. The Hematology Journal


153

turer's instructions. All products were cloned into the promoter-less expression vector pXP2 and sequenced. All constructs were transiently transfected into (i) FDCP-Mix myeloid stem cells; (ii) activated primary T cells; (iii) activated cultured T cells MTL 2.8.2. and reporter activity assessed. To determine more stringently whether these fragments retain not just the tissue speci®city but also the dierentiation-stage speci®c activity of the endogenous gene, transiently transfected FDCPmix cells were transferred either to self renewal or dierentiation medium for two days prior to reporter gene assay.

Site directed mutagenesis The technique of overlapping internal primer PCR was used to perform site directed mutagenesis22 on fragment 7259 and the following primer combinations; (Amplify 1+2 and 3+4, combine products, denature, anneal and amplify with 1+3): NF-kB, (1) ctcggtaccGATATCTGGCTCCTG (2A positions +32 to +46); (2) CTGTTAGTCTGTTTCCGAACAGG (2A positions 7186 to 7164); (3) ctcggatccTCCCACAGTCTCCT (2A positions 7258 to 7244); (4) CGGAAACAGACTAACAGTTCTAAA (7170 to 7193). LVA, (1) ctcggtaccGATATCTGGCTCCTG (2A positions +32 to +46); (2) GTTTCCGTCCAGGAAATGGTTTT (2A positions 7176 to 7154); (3) ctcggatccTCCCACAGTCTCCTT (2A positions 7258 to 7244); (4) TCCTGGACGGAAACCCCCTAA (2A positions 7163 to 7183). All mutated fragments were gel puri®ed, digested with BamHI and KpnI and ligated into pXP2luc.

Electrophoretic mobility shift assay In an attempt to identify proteins bound to those sites implicated in transcriptional control, electrophoretic mobility shift assays were performed using nuclear extracts prepared from expressing cell types. Oligonucleotide probes corresponding to the wild type sequence and sequence in which either the NF-kB, LVa or both sites had been mutated were assessed. The double stranded oligos listed below were 5' end labelled using T4 Polynucleotide Kinase (Boehringer Mannheim) and g-32P-ATP (Amersham Pharmacia Biotech). Base substitutions are in bold face and underlined: S2A wild type (5' TGTTAGGGGGTTTCCGAACAGGAAATGGTT3); NF-kB mutant (5' TGTTAGGTCTTTTCCGAACAGGAAATGGTT3'); LVa mutant (5'TGTTAGGGGGTTTCCGTCCAGGAAATGGTT3); S2A LVa/NF-kB mutant (5' TGTTAGGT-CTTTTCCGTCCAGGAAATGGTT3'). Nuclear extracts were prepared from activated primary T cells and the FDCP-Mix cell line as described previously.23 For EMSA 56105 c.p.m. of the labelled probe was incubated for 5 min at 48C and 5 min at 228C with 10 mg of either of the above extracts in the presence of 3 mg of poly (dI-dC) (Sigma) and 50 mM KCl in Buer D (20 mM N-(2 hydroxyethyl) piperazine-N'-(2-ethane-sulphonic acid), pH

7.9, 20% glycerol, 0.2 mM EDTA, 0.5 mM phenylmethylsulphonyl ¯uoride. The reaction mixtures were then loaded on a 10% native polyacrylamide gel and subjected to electrophoresis against Tris-Glycine buer. Gels were visualised with a Packard Instant Imager (Meriden, USA) using linear scaling.

Competition experiments and supershifts Similar conditions were used to those for standard EMSA, except that the labelled probe was incubated in the presence of unlabelled competitor oligonucleotides. These were either to the NF-kB consensus oligonucleotide: (5' AGTTGAGGGGACTTTCCCAGGC3'), or the CK-1 element known to bind NF-GMa: (5' AATTCTGATAAGGGCCAGGAGATTCCACA GTTCAGGTAGTTG3'). Supershift reactions were pre-incubated with the appropriate antibody (Promega) for 10 min prior to addition of probe.

Results Isolation, mapping and sequencing of the 5' end of the S2A mRNA A restriction map of the 5.8 Kb S2A cosmid fragment is shown in Figure 1. Using primer extension with a 21 bp oligonucleotide corresponding to nucleotide positions 81 ± 101 of the published S2A sequence8 a total of nine termination products were detected, 41 to 123 bases upstream from the start site of translation (Figure 2A). The major product corresponds to position-91. Sequence analysis of the longest RACE clone, which was represented several times, is shown in Figure 2B. It is clear that the initial 15 bases correspond to exon 2 and the remaining upstream sequence incorporates exon 1 (splice junction indicated in Figure 2B). Blot hybridisation analysis of the longest RACE clone to restriction fragments generated from the 5.8 Kb genomic EcoRI-BamHI fragment localised homology to within a StuI-StuI fragment of approximately 1.6 Kb. This entire region was sequenced and compared to the RACE clone allowing localisation of exon 1 (Figure 3). Although 77 bp of the RACE product showed 91% homology to the genomic sequence, 21 bases at the most 5' end of the RACE clone showed no sequence similarity. This sequence does, however, share very high similarity to part of exon 2 (indicated in Figure 2B). Furthermore, there is a short but signi®cant region of complementarity between the termini of exon 1. It therefore seems likely that the 21 5' bases have been generated by a spurious `snap-back' priming event to the junction between exons 1 and 2. Similar events may account for some of the larger, less abundant, primer extension products (Figure 2A). The 9% disparity between genomic and cDNA exon 1 nucleotide sequence is most likely attributable to the high mismatch base incorporation rate of the high temperature reverse transcriptase used. The Hematology Journal


154

Figure 2 5' Primer extension analysis and comparison of S2A RACE clone with genomic sequence. (A) Primer extension products were electrophoresed through a 3.5% PAGE sequencing gel. The gel was dried and signals visualised by phosphorimaging. The numbers on the left refer to the transcript sizes determined as position upstream of the ATG translation start. Lane 1, FDCP-Mix cell poly (A)+ RNA, lane 2, control, RNA derived from 3T3 ®broblasts which do not express S2A. (B) A sequence comparison of the longest S2A RACE clone and the corresponding genomic region. Additional sequence thought to be attributable to mis-priming within exon 2 is boxed.

The proposed transcription start site indicated in Figure 2B is based on the sequence of the major RACE product and is in agreement with the primer extension result in which the major band of 91 bps would comprise 15 nucleotides from exon 2 and 76 nucleotides from exon 1. Transcription would then initiate at an A residue, as is common for eukaryotic genes.24 There are no consensus TATA or CAAT box sequences in the promoter proximal region.

The S2A promoter and upstream regions direct cell type-speci®c expression A convenient EcoRV site within the 5' UTR of the 2A gene was utilized to generate an upstream StuI-EcoRV The Hematology Journal

genomic fragment of approximately 1.2 Kb (indicated in Figure 1) which was cloned into the promoterless luciferase vector pXP2.16 A series of deletion constructs derived from this fragment yielded similar patterns of activity in each of the cell types listed (Figure 4). The relatively low activity of the 7128 construct increased 8 ± 10-fold with the inclusion of sequences to 7259. Further extension to 7453 depressed activity 2 ± 3-fold in all cases, whilst inclusion of sequence to 7648 largely overcame this repression. These results suggest the presence of at least two positive control elements (one between 7128 and 7259 and the second between 7453 and 7648) and one negative regulatory element which lies between 7259 and 7453. Although the


155

relative strength of these elements is variable between the dierent cell types assayed, all are active in both FDCP-Mix progenitors and activated T cells. Transfection of the 71174 and 7259 constructs into 3T3 ®broblasts revealed barely detectable activity (53% of thymidine kinase promoter control) con®rming that expression is tissue speci®c (data not shown). As shown in Figure 5, the activity of both the 71174 and 7259 promoter fragments reduced approximately threefold upon transfer of transfected FDCPmix cells to dierentiation medium, compared to that in cells transferred to self-renewal medium, while the activity of a CMVLuc control plasmid remained constant. This marked down-regulation over the ®rst two to three days of dierentiation (during which the selfrenewal potential of the FDCPmix cells is lost without any gross changes in morphology) re¯ect the behaviour of the endogenous gene. This shows that the 7259 bp promoter fragment (and more precisely the region between 7128 and 7259) contains regulatory elements which mediate the high level tissue and dierentiation-stage speci®c expression typical of the S2A gene.

The S2A promoter proximal region contains consensus binding sites for Nf-kB and LVa A computer search for known transcription factor binding site consensus sequences within the 7259 fragment revealed sites for NF-kB25 and the Moloney murine leukaemia virus, MoMuLV, enhancer binding factor LVa26 at approximately 7172 and 7165 nucleotides relative to the transcription start site respectively (Figure 3). To assess the functional signi®cance of these transcription factor binding motifs, site directed mutagenesis was used to alter each sequence individually within the context of the 7259 fragment. The NF-kB site was changed from GGGGGTTTC to GTCTGTTTC, since this had previously been reported to abolish protein binding.27 In the absence of published data on the binding requirements of the LVa motif we altered two nucleotides of the six base sequence (GAACAG to GTCCAG). The mutant constructs were transiently transfected into (i) FDCP-Mix cells; (ii) activated primary splenocytes and (iii) MTL 2.8.2 cells after which reporter gene activity was compared with that of the wild-type 7259 fragment. In FDCP-Mix cells disruption of the NF-kB or LVa sites reduced luciferase activity to 25 and 50% of the wild type construct respectively (Figure 6A). Similar results were obtained using primary splenocytes (Figure 6B) indicating that both sites are required for maximal promoter activity. In the MTL cell line (Figure 6C) LVa site disruption also reduced activity approximately two-fold, but the mutation of the NF-kB site had less eect, reducing activity to 77% that of the intact promoter fragment. Mutation of both the LVa and NF-kB motifs completely abrogated promoter activity in all cell types tested.

Figure 3 Nucleotide sequence of S2A exon 1 and upstream promoter region. Sequence of the 5' upstream region used in deletion assays and exon-1 sequence (shadowed). The numbers to the left refer to the positions of the nucleotides relative to the transcription start site. Double underline indicates the EcoRV restriction site used in promoter cloning. Potential transcription factor binding sites are in boldface (NF-kB site=GGGGTTTC; LVa site=GAACAG).

Dierent proteins bind to the NFkB site in T cells and haematopoietic progenitors Two major DNA-protein binding complexes were formed, with crude nuclear extracts from both activated T cells and FDCP-Mix cells binding predominantly to those fragments containing an intact NF-kB sequence (Figure 7A, lanes 1 and 3). Complex A was much stronger in T cells than Complex B and the converse was true of FDCP-Mix cells (compare lanes 1 and 3 in Figure 7C). Supershift assays with anti p50 NF-kB con®rmed that binding at Complex A was NF-kB speci®c (Figure 7D). This was further supported with competition experiments which demonstrated that an NF-kB consensus oligo could compete eciently with the wild type probe for protein-DNA complex A formation at a 50-fold molar excess (data not shown). Although the NF-kB complex present with the T cell extract was completely absent when both the NF-kB and LVa sites were simultaneously modi®ed, a weak signal was still evident when the NF-kB site alone was mutated. Since the binding of recombinant NF-kB could be completely abrogated by mutation of the NF-kB site alone (Figure 7A, lanes 5 and 6), it appears that the binding of NF-kB to the oligo may be more ecient in the presence of other nuclear proteins which bind to the region containing the LVa site. In the absence of an independent, LVa-speci®c complex, this was the only indication from the EMSA assays The Hematology Journal


156

Figure 4 Transient transfection analysis of S2A promoter deltion fragments in Murine myeloid stem cells and activated T cells. The S2A deletion series in the luciferase reporter vector pXP2luc were transiently transfected into (A) the murine myeloid stem cell line FDCP-Mix; (B) activated primary splenocytes; (C) the cytotoxic T-cell line M.T.L.2.8.2. Reporter activities in (A) are expressed as percentage of the positive thymidine kinase promoter control (pTK109) whilst those in (B) and (C) are shown as a luciferase/bgalactosidase activity ratio. All standard deviation values for replicate T-cell assays were less than 5% of the mean. The T-cell positive control (C11-828) carries 828 bp of the CCP1 promoter driving luciferase expression. Figure 6 Site-directed mutagenesis of transcription factor binding motifs. Constructs in which the NF-kB and/or LVa motifs had been altered by site-directed mutagenesis, within the context of the 7259 S2A proximal promoter fragment were transiently transfected into (A) the murine myeloid stem cell line FDCP-Mix; (B) activated primary splenocytes and (C) the cytotoxic T-cell line M.T.L.2.8.2. Activities are expressed as percentage of the 7259 bp promoter control. In (A) the results represent the mean of at least three independent experiments. In (B) and (C) the results represent the ratio of luciferase/bgalactosidase activity. All standard deviation values for replicate T-cell assays were less than 5% of the mean.

Figure 5 Dierential activity of the S2A promoter in FDCPmix cells. FDCPmix cells transfected with the full-length (71174 bp) and deleted (7259 bp) Serpin promoter-luciferase constructs and with a CMV promoter-luciferase control were cultured for two days under either self renewal conditions (solid bars) or granulocyte/macrophage dierentiation conditions (open bars) before assay. The Hematology Journal

that protein binding to the regulatory region can be in¯uenced by the LVa site. Complex B, which was also dependent on an intact NF-kB site, and which was the predominant complex formed by FDCP-Mix extracts, could not be supershifted by antibody to NF-kB p50. Nor did addition of unlabelled competitor consensus NF-kB oligo have any


157

Figure 7 EMSA to determine protein binding to the S2A promoter element. (A) Nuclear extracts from FDCP-Mix cells incubated with labelled probes to: lane 1, wild type promoter element, lane 2 N NF-kB site mutant, lane 3 LVa site mutant, lane 4 NF-kB and LVa site mutant, lane 5 NF-kB mutant probe with recombinant NF-kB and lane 6 wild type probe with recombinant NF-kB. (B) Lanes 1 ± 4 as above with activated T-cell extracts. (C) FCDP-Mix extracts incubated with: lane 1 wild type probe and lane 2 NF-kB consensus, lane 3 T-cell extracts incubated with wild type probe and lane 4 NF-kB consensus. (D) Supershift assay. T-cell extracts with: one wild type probe and anti NF-kB p50, lane 2 NF-kB/LVa mutant probe and anti NF-kB p50 and lane 3 wild type probe only.

signi®cant eect on complex formation (data not shown). The formation of a complex between an oligonucleotide containing a canonical NF-GMa28 binding site and FDCP-Mix/activated T-cell extracts was not demonstrated. Furthermore this oligonucleotide could not eciently compete for the formation of complex B. Thus, the protein binding in complex B appears not to be NF-GMa. To establish whether or not the binding sequence for NF-kB is adequate to form complex B EMSA was performed using nuclear extracts from both T cells and FDCP-Mix cells together with the wild type Serpin promoter oligo containing both NF-kB and LVa sites and an oligo carrying the NF-kB consensus motif only (Figure 7C). Complex B was formed on the Serpin promoter sequence only, indicating that the NF-kB site alone was insucient for binding and that additional adjacent sequence in the wild type probe is required.

Discussion and conclusion We have identi®ed a 77 bp untranslated exon 1 sequence from the serine protease inhibitor S2A gene

and de®ned a putative start site for transcription. Although the primer extension result raises the possibility of multiple transcription start sites, which has been reported for other Serpin genes such as ovalbumin,29 the existence of repeat sequences in the 5' untranslated sequence of the S2A gene suggests that the multiple extension products observed may be artefacts caused by secondary structure in the RNA template. The high degree of secondary structure at the 5' end of the S2A gene, indicated from RACE and primer extension products, suggests that, in addition to transcriptional control, S2A may also be regulated at the level of translation. Indeed, this supports our previously reported results with T cells activated with concanavalinA/IL-2 or CD3/IL-2 in which the mRNA transcripts for S2A appear 24 h before the protein.8 Additional studies will be required to establish the contribution of post-transcriptional regulation to S2A expression. Since high levels of S2A mRNA have been found in both FDCP-Mix cells and activated T cells,8 the full length promoter/reporter construct was transiently transfected into cells of this type in order to assess promoter activity. The results of these initial experiThe Hematology Journal


158

ments indicated that the 1.2 Kb fragment is able to direct high levels of reporter gene expression in FDCPMix cells, primary splenocytes and the constitutively activated T-cell line MTL 2.8.2 (data not shown). Deletion analysis of sequences upstream of the transcription start as far as nucleotide 71174 has indicated that the region to 7648 contains at least two positive and one negative control elements that are active in both myeloid progenitor and activated T cells. Given the ability of the 7259 bp promoter fragment to drive appropriate expression in transiently transfected cells, it is likely that the additional upstream elements are involved either in establishing active gene expression in the chromatin context during development, or in ®ne-tuning the transcription. The interplay between the negative and positive elements may be particularly interesting in this respect, since the ultimate down regulation of inducible genes such as S2A is expected to be an important part of a co-ordinated activation response. The pattern of activity of the deletion mutants suggests that the negative element between 7259 and 7453 is able to suppress promoter activity, but is itself overruled by the positive element further upstream. Nonetheless, a 259 bp fragment is sucient to direct high level expression which retains both the cell type speci®city and dierentiation stage speci®city of the endogenous gene. Mutation of two transcription factor binding motifs within the 7259 promoter fragment has clearly shown that the S2A gene is regulated by similar elements in both myeloid stem cells and activated T cells. Potential roles for the transcription factor NF-kB and the MoMuLV factor LVa were indicated in transient transfection assays. The weaker contribution of the NF-kB site to reporter gene expression in MTL cells compared to primary splenocytes, may be indicative of dierences in transcription factor levels between the freshly activated primary lymphocytes (which are expected to re¯ect the in vivo situation more closely), and the transformed cell line. Coincident mutation of both NF-kB and LVa sites completely abolished reporter gene activity in FDCP-Mix cells and activated T cells, indicating that factors binding at these motifs are absolutely required for all activity seen with the 7259 fragment. Relatively concise and highly active elements of this nature should be of potential use in gene therapy applications. In this respect, our initial results indicate that the promoter is also highly active in the human CML progenitor cell line K562, and that this activity is also dependent on the NF-kB and LVa motifs. The NF-kB/Rel family of transcription factors are known to control the expression of many cellular and viral genes including those involved in immune response and acute in¯ammatory processes.30 Targeted gene knockouts and transgenic animal experiments have indicated many functions of this transcription factor family including immune cell activation, control of apoptosis and a role in the development of various haematopoietic cells.25 Many viruses are able to activate NF-kB and it has been suggested that The Hematology Journal

induction of NF-kB transcription factor activity may prove relevant to the mechanism of MoMuLV induced leukaemia.31 In this regard it is interesting to note that the FDCP-Mix cell lines have an active MoMuLV infection and we have shown that MoMuLV RNA transcripts, like those of S2A, also down-regulate extensively during FDCP-Mix dierentiation (unpublished). Using EMSA with probes which span the wild type or mutated NF-kB and LVa sites, we have shown that NF-kB from T cells does bind to the consensus sequence in the S2A promoter. However, we were unable to demonstrate in vitro binding of the LVa transcription factor. This factor was originally identi®ed as a protein which could bind to the 75 bp repeat of the MoMuLV enhancer26 and was shown to be present in a variety of murine tissues. Absence of a shift under various EMSA conditions (not shown) may be indicative of low levels of LVa within activated T cells and multipotent haematopoietic progenitor cells. Alternatively, it may be that LVa binding to the site within the context of the S2A promoter fragment is destabilised by ¯anking sequences and/or protein DNA interactions. Previous studies have reported that LVa only binds to a DNA fragment which spans a PvuII site (where the point of enzyme cleavage lies immediately adjacent to the LVa minimal binding site GAACAG).26 There is no such PvuII site in the S2A promoter. If this is absolutely necessary for factor binding, then the lower reporter activity seen in transient transfection assays, when the LVa site has been disrupted, may be due to altered binding of a dierent protein with a similar binding recognition sequence. An interesting result of the gel shift assays was the identi®cation of an apparently novel complex (B) formed between FDCPmix nuclear proteins and the NF-kB consensus within the context of the S2A promoter sequence. No complex B was formed with FDCP-mix extracts on labelled oligo carrying the NFkB consensus site alone. This, together with our inability to eectively compete complex B with NFkB competitor oligos, or to supershift complex B with NF-kB antibodies, makes it unlikely that the novel complex is due to binding of a degradation product. Transcription factors with binding sites similar to that of NF-kB have been reported. One such factor, shown to migrate faster than NF-kB on EMSA gels, is NFGMa28 which is found in the promoter region of several haematopoietic growth factor genes. The potential involvement of S2A in haematopoietic stem cell self renewal and T-cell activation prompted us to explore the possibility that this lower shift might be the result of NF-GMa binding. However binding of this factor could not be demonstrated. It therefore seems more likely that complex B involves a novel protein binding to a sequence that extends beyond the NF-kB consensus. This is currently under further investigation. In conclusion we have identi®ed and characterised a 259 bp promoter fragment which is highly active in


159

murine myeloid stem cells and activated T cells. Activity in both cell types is dependent on the integrity of closely associated consensus binding sites for NF-kB and LVa. Activity in T cells appears to re¯ect NF-kB binding to its consensus site, as may be expected, but in a manner which is in¯uenced by the adjacent LVa site. In contrast, the binding activity in haematopoietic progenitor cells appears to be a novel protein which recognises the NF-kB consensus only in the context of ¯anking sequences.

Acknowledgements We would like to thank Brenda Duggan, (University of Alberta) for provision of T-cell extracts. RC Bleackley is an Alberta Heritage Foundation for Medical Research (AHFMR) Scientist, a Distinguished Scientist of the MRC and an International Scholar of the Howard Hughes Medical Institute. This work was supported by the Cancer Research Campaign, Manchester, UK, the Medical Research Council of Canada, and BMBF-IZKF grant number 01KS9504N2.

References 1 Till JE, McCulloch EA. Hemopoietic stem cell differentiation. Biochimica Biophysica Acta 605: 431, 1980. 2 Dexter TM, Heyworth CM, Spooncer E, Ponting IL. The role of growth factors in self-renewal and dierentiation of hematopoietic stem cells. Philosophical Transactions of the Royal Society of London: Biological Sciences 327: 85, 1990. 3 Spooncer E, Boettiger D, Dexter TM. Continuous in vitro generation of multipotential stem cell clones from src infected cultures. Nature 310: 228, 1985. 4 Spooncer E, Heyworth CM, Dunn A, Dexter TM. Selfrenewal and dierentiation of interleukin-3-dependent multipotent stem cells are modulated by stromal cells and serum factors. Dierentiation 31: 111, 1986. 5 Heyworth CM, Dexter TM, Kan O, Whetton AD. The role of hemopoietic growth factors in self-renewal and dierentiation of IL-3-dependent multipotential stem cells. Growth Factors 2: 197, 1990. 6 Hampson IN, Pope L, Cowling GJ, Dexter TM. Chemical cross linking subtraction (CCLS): a new method for the generation of subtractive hybridisation probes. Nucleic Acids Research 20: 2899, 1992. 7 Inglis JD, Lee M, Davidson DR, Hill RE. Isolation of two cDNAs encoding novel alpha 1-antichymotrypsinlike proteins in a murine chondrocytic cell line. Gene 106: 213, 1991. 8 Hampson IN, Hampson L, Pinkoski M, Cross M, Heyworth CM, Bleackley RC, Atkinson E, Dexter TM. Identi®cation of a Serpin speci®cally expressed in multipotent and bipotent hematopoietic progenitor cells and in activated T cells. Blood 89: 108, 1997. 9 Steele FR, Chader GJ, Johnson LV, Tombran Tink J. Pigment epithelium-derived factor: neurotrophic activity and identi®cation as a member of the serine protease inhibitor gene family. Proceedings of the National Academy of Sciences USA 90: 1526, 1993. 10 Festo BW, Rao JS, Rayford A, Hantai D. Plasminogen activators and their inhibitors in the neuromuscular system: II. Serpins and Serpin protease complex receptors increase during in vitro myogenesis. Journal of Cell Physiology 144: 272, 1990. 11 Perlmutter DH, Daniels JD, Auerbach HS, De Schryver Kecskemeti K, Winter HS, Alpers DH. The alpha 1antitrypsin gene is expressed in a human intestinal epithelial cell. Journal of Biological Chemistry 264: 9485, 1989. 12 De Maria R, Zeuner A, Eramo A, Domenichelli C, Bonci D, Grignani F, Srinivasula SM, Alnemri ES, Testa U, Peschle C. Negative regulation of erythropoiesis by caspase-mediated cleavage of GATA-1 [see comments]. Nature 401: 489, 1999.

13 Boettiger D, Anderson S, Dexter TM. Eect of src infection on long-term marrow cultures: increased selfrenewal of hemopoietic progenitor cells without leukemia. Cell 36: 763, 1984. 14 Bleackley RC, Havele C, Paetkau V. Cellular and molecular properties of an antigen-speci®c cytotoxic T lymphocyte line. Journal of Immunology 128: 758, 1982. 15 Frohman M. PCR Protocols. In: Innis MA, Gelfand DH, Sninsky JJ, White TJ, (eds) San Diego: Academic Press, 1990, p 28. 16 Nordeen SK. Luciferase reporter gene vectors for analysis of promoters and enhancers. Biotechniques 6: 454, 1988. 17 Babichuk CK, Duggan BL, Bleackley RC. In vivo regulation of murine granzyme B gene transcription in activated primary T cells. Journal of Biological Chemistry 271: 16485, 1996. 18 de Wet JR, Wood KV, DeLuca M, Helinski DR, Subramani S. Fire¯y luciferase gene: structure and expression in mammalian cells. Molecular and Cellular Biology 7: 725, 1987. 19 Fregeau CJ, Bleackley RC. Transcription of two cytotoxic cell protease genes is under the control of dierent regulatory elements. Nucleic Acids Research 19: 5583, 1991. 20 Sambrook J, Fritsch EF, Maniatis T. Molecular Cloning: A Laboratory Manual, vol. 1 (2nd edn). NY, Cold Spring Harbor Labs., 1989. 21 Baum C, Forster P, Hegewisch Becker S, Harbers K. An optimized electroporation protocol applicable to a wide range of cell lines. Biotechniques 17: 1058, 1994. 22 Ho SN, Hunt HD, Horton RM, Pullen JK, Pease LR. Site-directed mutagenesis by overlap extension using the polymerase chain reaction. Gene 77: 51, 1989. 23 Dyer RB, Herzog NK. Isolation of intact nuclei for nuclear extract preparation from a fragile B-lymphocyte cell line. Biotechniques 19: 192, 1995. 24 Breathnach R, Chambon P. Organization and expression of eucaryotic split genes coding for proteins. Annual Revues of Biochemistry 50: 349, 1981. 25 Baeuerle PA, Baltimore D. NF-kB: ten years after. Cell 87: 13, 1996. 26 Speck NA, Baltimore D. Six distinct nuclear factors interact with the 75-base-pair repeat of the Moloney murine leukemia virus enhancer. Molecular and Cellular Biology 7: 1101, 1987. 27 Tschachler E, Bohnlein E, Felzmann S, Reitz Jr MS. Human T-lymphotropic virus type I tax regulates the expression of the human lymphotoxin gene. Blood 81: 95, 1993. The Hematology Journal


160

28 Shannon MF, Pell LM, Lenardo MJ, Kuczek ES, Occhiodoro FS, Dunn SM, Vadas MA. A novel tumor necrosis factor-responsive transcription factor which recognizes a regulatory element in hemopoietic growth factor genes. Molecular and Cellular Biology 10: 2950, 1990. 29 Malek LT, Eschenfeldt WH, Munns TW, Rhoads RE. Heterogeneity of the 5' terminus of hen ovalbumin messenger ribonucleic acid. Nucleic Acids Research 9: 1657, 1981.

The Hematology Journal

30 Gilmore TD. Role of rel family genes in normal and malignant lymphoid cell growth. Cancer Surveys 15: 69, 1992. 31 Pak J, Faller DV. Moloney murine leukemia virus activates NF-kappa B. Journal of Virology 70: 4167, 1996.