Leukemia (2002) 16, 1507–1518  2002 Nature Publishing Group All rights reserved 0887-6924/02 $25.00 www.nature.com/leu Functional cloning of SPIN-2, a nuclear anti-apoptotic with roles in cell cycle progression BS Fletcher1, C Dragstedt1, L Notterpek2 and GP Nolan3

1Department of Pharmacology and Therapeutics, University of Florida, College of Medicine, Gainesville, FL, USA; 2Department of Neuroscience, University of Florida, College of Medicine, Gainesville, FL, USA;and 3Departments of Molecular Pharmacology and Microbiology and Immunology, Stanford University School of Medicine, Stanford, CA, USA.

The balance between hematopoietic cell viability and apoptosis plasm.8,9 Apoptosis then proceeds by the association of cyto- is regulated by exogenous growth factors, however, the mol- chrome c with the mammalian apaf-1 protein, which results ecular mechanisms by which these trophic factors exert their effects remain obscure. A functional retroviral cDNA - in the activation of caspase 9 and other members in the 10 based screen was employed to identify that prevent caspase cascade. growth factor withdrawal-mediated apoptosis in the myeloid In hematopoietic tissues, in addition to controlling cell progenitor cell 32Dcl3. This approach identified three classes number and differentiation, apoptosis likely plays a role in of genes: those with known roles in apoptosis (bcl-X and orni- L leukemogenesis, as transformed cells must be able to over- thine decarboxylase); genes previously identified but not linked directly to apoptotic signaling (O-linked N-acetylglucosamine come pro-apoptotic cellular signals. The pro-survival pheno- transferase); and a previously uncharacterized we termed type may be due to the up-regulation of intact anti-apoptotic SPIN-2. In 32Dcl3 cells, expression of exogenous SPIN-2 pro- genes or to the expression of mutant with similar vides 25% protection from apoptosis following growth factor functions. An example of this is seen in follicular B cell lym- ෂ withdrawal compared to controls which show 1–2% survival. phomas in which chromosomal translocations (ie t(8:14) and SPIN-2 overexpression slows cell growth rates and increases the percentage of cells in G /M (32% vs control cells at 12%). t(8:22)) result in the transcriptional upregulation of the anti- 2 11,12 Immunolocalization studies indicate that myc-epitope tagged apoptotic gene product bcl-2. Bcl-2 overexpression has SPIN-2 proteins, which retain their anti-apoptotic function, also been demonstrated in patients with acute myelogenous reside in the nucleus, whereas a C-terminal deletion mutant leukemia and was shown to correlate with poor prognosis.13,14 that loses its anti-apoptotic activity is located in the cytoplasm. Several members of the bcl-2 family have been identified that These studies suggest that SPIN-2 is a novel nuclear protein 15 16 that functions to regulate cell cycle progression and this have anti-apoptotic properties, including bcl-XL, A1 and 17 activity is related to the inhibition of apoptosis following the mcl-1. A second family of proteins that can inhibit apoptosis removal of essential growth factors. by blocking activation of the caspase cascade are the IAP pro- Leukemia (2002) 16, 1507–1517. doi:10.1038/sj.leu.2402557 teins (inhibitors of apoptosis proteins), such as survivin, XIAP, Keywords: apoptosis; growth factor withdrawal; retroviral cDNA cIAP1 and cIAP2.18,19 Finally, several recently described pro- libraries; spindilin teins have been demonstrated to affect apoptosis in hemato- poietic cells including cFLIP and Toso.20,21 While expression of anti-apoptotic bcl-2 family members can prolong survival Introduction following growth factor withdrawal, few other proteins besides transforming oncogenes (ie bcr/abl) are able to Programmed cell death, or apoptosis, has emerged as an promote survival or growth in these paradigms. essential mechanism by which an organism controls cell num- To identify novel proteins with potential anti-apoptotic ber and fate. In hematopoietic systems, apoptosis is highly regulated as it can function to restrict cell number, eliminate activity, a functional was devised utilizing auto-reactive immune cells, and control differentiation. With- retroviral cDNA libraries. We report on the characterization drawal of essential cytokines that function as growth factors of SPIN-2, a novel nuclear gene that can protect the factor- has been shown to initiate the apoptotic cascade in both dependent hematopoietic progenitor cell lines 32Dcl3 and lymphoid and myeloid lineages.1,2 Cytokines, such as NSF/N1-H7 from apoptosis following interleukin (IL)-3 with- interleukins and growth factors, play roles in regulating cell drawal. The anti-apoptotic property of SPIN-2 appears specific number and promoting differentiation of specific hemato- for growth factor withdrawal-induced apoptosis as over- poietic lineages. expression in 32Dcl3 cells, or the lymphoid cell line Jurkat, Following growth factor withdrawal one of the earliest failed to protect against cytotoxic drug (doxorubicin)-induced physiological changes is a time-dependent increase in phos- apoptosis, or Fas-mediated apoptosis, respectively. Over- phatidylserine externalization.3 Subsequently, there is a expression of SPIN-2 in 32Dcl3 cells slowed cell cycle pro- 4,5 decrease in mitochondrial membrane potential and a cal- gression and increased the percentage of cells in G2/M phase cium influx-dependent redistribution of mitochondria.6 of the cell cycle. This effect of SPIN-2 on cell cycle pro- Additional changes to the mitochondria include interactions gression appears to be linked to its anti-apoptotic activity, as with the protein bax, a pro-apoptotic bcl-2 family member, mutants that are unable to promote G2/M slowing are no that translocates from the cytoplasm and associates with mito- longer able to prevent apoptosis following growth factor with- chondrial bound bcl-2.7 Ultimately, these cellular changes drawal. Thus, it is possible to define genes that have specific result in mitochondrial release of cytochrome c into the cyto- anti-apoptotic activities under defined circumstance that dis- tinguish the actions of these genes from those of more broadly acting anti-apoptotic genes. It is likely that an understanding Correspondence: BS Fletcher, Department of Pharmacology and of such genes will help refine our knowledge of the many Therapeutics, Box 100267, University of Florida, College of Medicine, mechanisms by which leukemias become resistant to Gainesville, FL 32610-0267, USA; Fax: 352 392–9696 pro-apoptotic signaling. Received 20 November 2001; accepted 20 March 2002 SPIN-2 effects on cell cycle and apoptosis BS Fletcher et al 1508 Materials and methods derivative of pBabeMN-lacZ in which the lacZ gene has been removed and the polylinker has been modified to create a SalI Cell culture site upstream of the NotI site.21,22 Following ligation, the cDNA was transformed into chemically competent UltraMAX 32Dcl3 cells (a gift of Dr Alan Friedman, Johns Hopkins Uni- DH5␣ cells (Gibco BRL). Transformants were amplified on versity, Baltimore, MD, USA) were grown in Iscove’s modified solid LB ampicillin plates and DNA was purified. A Dulbecco’s media (IMDM) with 1.5 g sodium bicarbonate per sampling of 20 random clones revealed that Ͼ95% of the liter. Additional supplements included 10% fetal bovine clones contained inserts with an average insert size of 2.0 kb. serum (FBS), 1 ng/ml murine IL-3 (R & D Systems, Minnea- Overall complexity of the library was estimated by counting polis, MN, USA) and /streptomycin/glutamine the number of independent colonies from a dilution of each (Pen/Strept/Glut) (Gibco BRL, Rockville, MD, USA). The bacterial transformation that was plated and harvested. The amphotropic and ecotropic retroviral packaging cell lines total number of colonies that were pooled together to generate (Phoenix A and E, developed by Garry Nolan, Stanford the library was ෂ350 000 independent clones. University) and NIH3T3 cells were grown in high glucose Dulbecco’s modified Eagle’s media (DMEM) supplemented with 10% FBS and Pen/Strept/Glut. The NSF/N1-H7 cell line Retroviral cDNA library screening (a gift of Dr Deng, University of Florida) was grown in RPMI- 1640 supplemented with 10% FBS, Pen/Strept/Glut and The plasmid cDNA library was transiently transfected using 1 ng/ml murine IL-3. calcium phosphate into the amphotropic packaging cell line Phoenix A that produces infectious retrovirus with titers approaching ෂ5 × 106 infectious units/ml.21 Briefly, the pack- Apoptosis assays aging cell line was plated at 5 × 106 cells per 10 cm plate (or 2 × 106 cells/6 cm plate), 18 h prior to transfection. Plasmid To demonstrate apoptotic cell death, cells were either grown DNA (20 ␮g/10 cm plate or 8 ␮g/6 cm plate) was added by in the presence (+IL-3) or absence of IL-3 (−IL-3) for 24 and standard Hepes calcium phosphate transfection in the pres- 48 h. One million cells from each condition were collected ence of chloroquine 50 ␮M. Eight hours after the addition of by centrifugation and washed once in PBS. The cell pellets the precipitate, the media was replaced with 10 ml (3 ml/6 cm were resuspended in 100 ␮l of labeling solution (containing plate) of fresh DMEM with 10% FBS. Sixteen hours later, the Hepes buffer with calcium, annexin-V-FLUOS and propidium medium was replaced again and the cells were placed at 32°C iodide (PI)) per manufacturer’s recommendations (Roche, Indi- for ෂ30 h. Retroviral supernatant was then collected and fil- anapolis, IN, USA). After a 15 min room temperature incu- tered through a 0.45 ␮m syringe filter to prevent transfer of bation, 0.3 ml of additional Hepes buffer containing PI was the packaging cell to the target culture. added to each sample followed by fluorescence activated cell Target 32Dcl3 cells were spin-infected with the recombi- sorting (FACS) analysis (FACScan; Becton Dickinson, Immun- nant retroviral library supernatant or specific cDNA super- ocytometry Systems, San Jose, CA, USA). For the DNA lad- natant using the following procedure. Target 32Dcl3 cells dering assay, ෂ8 million cells from the two conditions were (1 × 106) were collected by centrifugation and all medium was pelleted and resuspended in 50 ␮l TE buffer (10 mM Tris, pH removed by aspiration. The cell pellet was resuspended in 8 and 10 mM EDTA). Lysis buffer (900 ␮l TE buffer plus 0.5% 2 ml of retroviral supernatant and placed in two wells of a 24- Triton X-100) was added and the samples were put on ice for well plate (1 ml/well). Additional additives included poly- 15 min. Following centrifugation (12 000 r.p.m.) for 15 min at brene (50 ␮M) to enhance the transduction efficiency and IL- 4°C, the supernatants were transferred to a new tube. The 3 (1 ng/ml). For the library screen, the procedure was scaled samples were incubated with RNase (50 ␮g/ml) for 1 h at 37°C up by 10-fold using 20 wells total (10 × 106 cells). The 24-well and then with proteinase K (100 ␮g/ml) and SDS (0.5%) for plates were then placed in a temperature-controlled centrifuge 2 h at 50°C. Following phenol/chloroform extraction the able to accommodate microtiter plates and spun for 90 min at samples were ethanol precipitated overnight at −20°C. DNA 2500 r.p.m. and 32°C. Following the spin-infection, the plates was collected by centrifugation (13 000 r.p.m.) at 4°C for 20 were placed at 32°C for an additional 18 h. After this incu- min and the pellets were washed once with 70% cold ethanol. bation, the library-transduced cells were pooled together and Following air-drying, the pellets were resuspended in 30 ␮lTE collected by centrifugation. The cell pellet was resuspended buffer and 10 ␮l of each sample was loaded on a 2% agarose and plated (75 cm2 flask in 25 ml) in fresh IL-3 containing gel. Following electrophoresis, the gel was stained with medium and incubated at 37°C for 48 h to allowfor inte- ethidium bromide and photographed. gration and expression of the retrovirus. Utilizing these con- ditions, the infection efficiency of the target cells was consist- ently between 30 and 40% as determined by infection with Retroviral cDNA library construction the marker retrovirus pBSF-GFP (green fluorescent protein) transduced under similar conditions. A retroviral cDNA library was constructed from RNA derived Library-infected 32Dcl3 cells were then cycled through sev- from HL-60 cells. Total RNA was isolated from ෂ2 × 108 HL- eral rounds of selection for survival in the absence of IL-3 60 cells using TRIzol reagent (Gibco BRL) and poly A+ mess- by the following procedure. The total population of cells was enger RNA was enriched using the MessageMaker kit (Gibco collected by centrifugation and placed into two 75 cm2 flasks BRL). Five ␮g of mRNA was used to generate cDNA utilizing in media depleted of IL-3 for 2 days (density ෂ500 000 the Superscript Plasmid cDNA Library Kit (Gibco BRL). The cells/ml). This resulted in the majority (ෂ98%) of the cells cDNA was size selected using cDNA Size Fractionation Col- undergoing cell death. Surviving cells were collected by cen- umns (Gibco BRL) to enrich for fragments Ͼ500 bases. cDNA trifugation and resuspended in fresh medium supplemented fractions from the columns were ligated directionally into the with IL-3 and again placed in two 75 cm2 flasks. After several SalI and NotI sites of the retroviral plasmid pBSF. pBSF is a days of growth (ෂ4–5 days), the revived cells were again taken

Leukemia SPIN-2 effects on cell cycle and apoptosis BS Fletcher et al 1509 through another round of IL-3 withdrawal, but for extended ized to construct the bcl-2-IRES-GFP construct was a gift of periods of time (3–4 days) to reach ෂ95% cell death. This pro- Dr Stratford May (University of Florida). The pBMN-IRES-GFP cess was repeated for a third time with even longer periods vector is a derivative of pBabeMN-lacZ in which the lacZ of IL-3 withdrawal (4–7 days). With each round of selection gene has been removed and replaced with IRES-GFP (Garry the number of flasks increased by a factor of two, so by the Nolan, unpublished data). Cell cycle analysis of the control end of three rounds there were eight independent flasks, or and retrovirally transduced and GFP sorted 32Dcl3 cells was pools. Cells, from every pool, that survived each round of performed using the CycleTEST PLUS kit (Becton Dickinson) selection had their genomic DNA extracted (Micro DNA per the manufacturer’s recommendations. The data were col- Extraction kit; Invitrogen, Carlsbad, CA, USA) to provide tem- lected by FACS (FACScan; Becton Dickinson) and analyzed plate for chain reaction (PCR) rescue of the inte- using the software Modfit Version 2.0. grated retroviral provirus. PCR from genomic DNA of pools of cells that had undergone one, two or three rounds of selection demonstrated that enrichment of specific bands on PCR Tissue array blotting required at least three rounds of selection. Furthermore, differ- ent pools produced both overlapping and divergent PCR A 30 base pair oligonucleotide that is antisense to the 5Ј products thereby increasing the odds of identifying novel untranslated (UTR) region of the SPIN-2 gene was obtained (5Ј- candidates. GGTCAGCAGGCGACCGGCCGAGTGAATGCT-3Ј) (Qiagen Operon, Alimeda, CA, USA). The oligonucleotide was radiola- beled using gamma 32P-ATP (NEN, Boston, MA, USA) and T4 Rescue of integrated provirus kinase (NewEngland Biolabs, Beverly, MA, USA). Following labeling, the oligonucleotide was purified using Nensorb prep Genomic DNA (ෂ500 ng) from cells that had undergone the columns (NEN) and hybridized to the multiple tissue selection was used as template to PCR amplify the integrated expression array according to the manufacturer’s recommen- cDNAs. The Taq/Pwo thermopolymerase blend (Expand High dations (Clontech, Palo Alto, CA, USA). Fidelity from Roche) and primers that flank the cDNA inserts designated 5ЈO (CAGCTTGGATACACGCCG) and 3ЈO (CCCCCCTTTTTCTGGAGAC), were utilized in the PCR reac- Construction of deletion and myc-epitope tagged tions. The PCR reactions were separated on agarose gels and SPIN-2 constructs amplified bands were isolated and purified (Qiagen Qiaex II, Valencia, CA, USA). The PCR fragments were then blunt Oligonucleotide primers were designed to PCR amplify spe- cloned into the vector pCR XL-Topo (Invitrogen) and sub- cific regions of the SPIN-2 cDNA, including the full open sequently sequenced. Clones that contained the correct retro- reading frame (ORF), an N-terminal truncation of 26 aa viral flanking sequences were subcloned back into the retrovi- (amino acids) which corresponds to the second in frame meth- ral backbone pBSF (using EcoRI and NotI sites) for functional ionine, a 9 aa C-terminal truncation and a double N- and C- analysis in the IL-3 withdrawal assay. The cDNA sequence for terminal deleted protein. Additional oligonucleotides were the clone SPIN-2 has been submitted to GenBank under the designed to generate several myc-epitope tagged proteins. The accession number AF356353. myc epitope (EQKLISEEDL) was added in frame following the first methionine in the SPIN-2 ORF and in frame following the second methionine in the 26 aa N-terminal deletion construct. IL-3 withdrawal assay and cell cycle analysis The myc epitope was also added to the N-terminus of the C- terminal truncation mutant. Following PCR amplification, the 32Dcl3 cells were transduced with amphotropic retroviral products were cut with restriction endonucleases and direc- supernatants expressing the unique cDNA clones as described tionally cloned into the retroviral vector pBSF (EcoRI/NotI above in the section ‘Retroviral cDNA library screening’. Typi- sites). One additional construct, designated SPIN-2 wild type cally, transfection of the packaging cell line occurred in 6 cm (corrected), was made by PCR to correct a point mutation plates and during the spin-infection two wells were used for found in all the SPIN-2 clones at position 628 (replacing the each particular cDNA clone. Following the 32°C incubation mutated G with the consensus A that results in a correction step for ෂ18 h, the cells were collected by centrifugation and of the Gly to Glu aa change at position 167). All SPIN-2 con- the pellet was resuspended in 3 ml of IL-3 containing medium structs were completely sequenced prior to testing in the IL- and plated in a 6-well plate. Two days later, the cells were 3 withdrawal assay. collected by centrifugation and all medium was removed by aspiration. The cell pellet was resuspended in 8 ml of fresh medium lacking IL-3 (density ෂ500 000 cells/ml), placed in a Characterization of the myc-tagged SPIN-2 protein 25 cm2 flask and incubated for 48 h. The percentage of cells that died was quantitated by propidium iodide staining fol- The three myc-tagged SPIN-2 constructs described above lowed by FACS analysis (FACScan; Becton Dickinson). To were transduced into the murine fibroblast cell line NIH3T3. determine the percentage of live cells, a gate was drawn For transduction of NIH3T3 cells, 200 000 cells were plated around the PI-negative cells and the percentage this fraction in a 6 cm dish ෂ18 h prior to infection. On the day of infec- represented over the total number of events was calculated. tion, the medium was removed and replaced with 2.5 ml fresh For experiments utilizing the IRES-GFP retroviral vectors retroviral supernatant containing polybrene (50 ␮M). The cells (pBMN-IRES-GFP, constructed in the laboratory of Garry were incubated at 32°C for 20 h followed by aspiration of the Nolan, unpublished results), 32Dcl3 transduced cells were remaining retroviral supernatant and addition of 3 ml of fresh first sorted for GFP expression (FACSort; Becton Dickinson) media. The cells were incubated at 37°C for an additional 2 and then analyzed for cell death by PI staining following 2 days to allowfor integration and expression of the retrovirally days of IL-3 withdrawal. The control murine bcl-2 gene util- introduced genes. These cells were then expanded on to

Leukemia SPIN-2 effects on cell cycle and apoptosis BS Fletcher et al 1510 10 cm plates and directly lysed for analysis or shown in Figure 1a, demonstrate that cells first become plated on glass cover slips for subcellular immunolocalization annexin V positive (an early marker of apoptosis) and then of the protein. subsequently stain for PI (a marker of dead cells). Agarose gel electrophoresis of genomic DNA derived from control cells (Figure 1b, lane 1) and cells deprived of IL-3 for 24 h (lanes Western blotting: NIH3T3 cells on tissue culture plastic 2) illustrate a marked increase in DNA laddering following IL- (10 cm plates) expressing either full length, N- or C-terminal 3 withdrawal (lane 2). In the control sample (lane 1), there is truncated myc-tagged SPIN-2 proteins (transduction faint DNA laddering observed suggesting that a small pro- efficiencies ~90%), as well as GFP, were rinsed twice with portion of cells are undergoing spontaneous apoptosis. This is ice-cold PBS and lysed in SDS gel sample buffer (3% SDS, in agreement with the small percentage of cells that are stain- 62.5 mM Tris, pH 6.8). Protein concentrations were determ- ing positive for annexin and PI even in the presence of IL-3 ined using the BCA protein assay kit (Pierce, Rockford, IL, (Figure 1a). These results confirm that apoptosis is occurring USA) and samples (75 ␮g/lane) were analyzed on 12.5% SDS following IL-3 withdrawal and indicate that PI staining can polyacrylamide gels under reducing conditions. Gels were serve as a surrogate marker of apoptotic cell death in this transferred to nitrocellulose membranes and processed for model system. immunoblotting with monoclonal anti-myc (Roche To identify genes that promote cell survival after IL-3 with- Biochemicals) at 1:400 diluted in 5% normal goat serum drawal we established the following functional genetic screen. (NGS) in PBS (5% NGS/PBS). After several washes in PBS, and A retroviral cDNA library was generated utilizing mRNA from incubation with goat anti- HRP-conjugated secondary the human promyelocytic leukemia cell line HL-60 for trans- 27 (Sigma, St Louis, MO, USA) diluted in 5% duction into the factor dependent cell line 32Dcl3. Follow- NGS/PBS, bound primary antibodies were detected with the ing transduction with the retroviral cDNA library into 32Dcl3 ECL chemiluminescent reagents (Amersham, Piscataway, NJ, cells, the population was subjected to several rounds of IL-3 USA). withdrawal for extended periods of time (2–7 days). This repeated IL-3 withdrawal strategy enriches for cells expressing specific cDNAs that provided a survival advantage over the parental cell line following growth factor withdrawal. To res- Immunostaining procedures: In parallel with the bio- cue these integrated cDNAs, genomic DNA from genetically chemical studies, full length, or truncated SPIN-2 and GFP screened cell populations was used as template to PCR ampl- infected NIH3T3 cells were plated on glass coverslips and ify the proviral sequences using primers that flank the cDNA allowed to adhere for 1–2 days. Cells were rinsed and fixed inserts. Following the third round of genetic selection, clear in 4% paraformaldehyde in L-15 (Leibovitz-15) balanced salt bands arose from these PCR reactions (data not shown). The (Gibco BRL) for 10 min at room temperature. After three rinses DNAs representing these bands were isolated, recloned into in PBS, samples were permeabilized by incubation with either the retroviral vector backbone, and tested for their ability to 100% methanol at −20°C, or 0.1% saponin in PBS at room transfer protection from apoptosis following IL-3 withdrawal. temperature, each for 5 min. After blocking with 10% NGS As summarized in Table 1, this approach yielded three known containing PBS, samples were incubated with monoclonal genes and one unknown gene. An example of a known gene anti-myc antibody at 1:400 (Roche Biochemicals) in the with strong anti-apoptotic activity included the bcl-2 family blocking solution overnight at 4°C. After four washes in PBS, member bcl-X , which provided 50% survival.15 Another Alexa Texas red-conjugated anti-mouse IgG (Molecular L known gene with a role in apoptosis was a truncated form of Probes, Eugene, OR, USA) at 1:300 and Hoechst nuclear dye the enzyme ornithine decarboxylase (ODC). ODC has been at 1 ␮g/ml were added (both diluted in 10% NGS/PBS). In shown to be upregulated by c-myc and the enhanced order to determine the subcellular localization of the overex- apoptosis seen in c-myc overexpressing cells is thought, in pressed SPIN-2 protein, fluorescein isothiocyanate (FITC)- part, to be due to elevated levels of ODC.2,28 A gene that had conjugated lectins were used. Concavalin A was used to label not previously been demonstrated to influence apoptosis was the endoplasmic reticulum (ER),23 while vicia villosa lectin the O-linked N-acetylglucosamine transferase gene (ONT). (VVL) was used to label the Golgi24 (both from Vector Labora- This protein functions to post-translationally attach the sugar tories, Burlingame, CA, USA). Glass cover slips were mounted N-acetylglucosamine to serine and threonine residues and using ProLong Antifade kit (Molecular Probes) and images may compete with protein phosphorylation thereby affecting were acquired with a Spot digital camera attached to a Nikon signal transduction.29 Both the ODC and ONT clones ident- Eclipse 1000 microscope. Images were printed from Adobe ified by this screen were N-terminally truncated suggesting Photoshop 5.0. they may function as dominant negatives, thus lowering their effective enzymatic activities. One uncharacterized cDNA clone consistently demon- Results strated modest protection from cell death following IL-3 with- drawal with transduction efficiencies of ෂ40% (Table 1). This Retroviral cDNA library screen for anti-apoptotic clone was designated SPIN-2 because it is the second member genes in a gene family that bears similarity to the spindlin gene.30 Although initially, the degree of protection conferred by SPIN- In agreement with previous studies, removal of IL-3 from the 2 appeared low( ෂ5% vs 1% for control), studies belowindi- media of 32Dcl3 cells results in the induction of cate that cells expressing SPIN-2 were selected against due to apoptosis.25,26 In our laboratory, Ͼ95% of the cells are non- an effect of SPIN-2 overexpression on cell growth. Therefore, viable by 48 h after IL-3 withdrawal as measured by a number the actual percentage of cells that are protected by SPIN-2 of independent assays; including trypan blue exclusion, overexpression is greater when cells are enriched for SPIN-2 annexin V or propidium iodide (PI) staining followed by flu- expressing cells (see Figure 2 below). The cDNA and protein orescence activated cell sorting (FACS) analysis. The results sequence data for the human SPIN-2 gene have been submit-

Leukemia SPIN-2 effects on cell cycle and apoptosis BS Fletcher et al 1511

Figure 1 Apoptosis of 32Dcl3 cells following IL-3 withdrawal. FACS analysis of propidium iodide (PI) and annexin V stained cells was performed on exponentially growing (+IL-3) cells as well as cells depleted of IL-3 (−IL-3) for 24 and 48 h (panel a). To demonstrate that these cells are undergoing cell death by apoptosis, panel b shows DNA laddering in control (+IL-3, lane 1) and IL-3-depleted (24 hours) cells (−IL- 3, lane 2).

Table 1 Genes identified in the anti-apoptotic screen

HL-60 library Relative Controls Relative screen survival survival

bcl-XL 50% Positive control: SPIN-2 5.0% ± 0.9 bcl-2 25% (25%) (74%) ODC 2.0% (N-term. truncated) Negative control: ONT 1.8% Mock transduced 1.0% (N-term. truncated) cells

Following transduction (ෂ30-40% efficiency) IL-3 was withdrawn for 48 h and the percentage of viable cells was quantitated using pro- pidium iodide staining and flow cytometry. The data are expressed as a relative percentage by setting the control cells to 1%. The data are the mean of at least two independent experiments and for the SPIN-2 gene, the mean and standard deviation rep- resent data from four independent experiments. The percentages in the bold numbers are the cell viability when Ͼ90% of the cells express the transgene (Figure 1). Figure 2 FACS analysis of retrovirally transduced 32Dcl3 cells fol- ODC, ornithine decarboxylase; ONT, O-linked N-acetylglucosam- lowing IL-3 withdrawal. 32Dcl3 cells were transduced with either the ine transferase. parental plasmid pBMN-IRES-GFP or the experimental constructs SPIN-2-IRES-GFP and bcl-2-IRES-GFP. Three days after transduction, the cells were sorted for GFP expression. These sorted pools were ted to GenBank database under the accession number then subjected to 48 h of IL-3 withdrawal and evaluated for cell death AF356353. by FACS analysis of propidium iodide (PI) and GFP fluorescence. The gated region identifies live GFP expressing cells and the number above each gate is the percentage of the total events counted that SPIN-2 and the conferral of an anti-apoptotic this fraction represents. Panel a illustrates that in the presence of IL- 3 following the sort, Ͼ90% of the cells are GFP positive. The results phenotype of three independent experiments done in triplicate are shown in panel b. Panel c shows data from a representative experiment follow- The SPIN-2 transgene was inserted into a retroviral vector that ing IL-3 withdrawal for each of the three constructs. contained an internal ribosomal entry site (IRES) upstream of the reporter gene GFP. Transduction of cells with the IRES- GFP vectors allowed for the enrichment (Ͼ90%) of GFP- statistically significant with a P value of 0.01 (Student’s one- positive cells by cell sorting (Figure 2a). In addition to the tailed paired t-test). Cells expressing the positive control bcl-2 SPIN-2-IRES-GFP construct, a murine bcl-2-IRES-GFP and an provided more robust protection up to 74% (P value of 0.001). empty vector (pBMN-IRES-GFP) were used as positive and The pro-survival effect of SPIN-2 overexpression was inves- negative controls, respectively. Following retroviral transduc- tigated in additional models of apoptosis. To determine tion and FACS sorting for GFP-positive cells, each construct whether the protective effects of SPIN-2 following growth fac- was tested for anti-apoptotic activity in the IL-3 withdrawal tor withdrawal could be observed in cell types other than assay. Figure 2c demonstrates one representative experiment 32Dcl3, a second factor-dependent cell line was tested. with the gated region indicating the percentage of live cells NFS/N1-H7, an IL-3-dependent murine myeloid progenitor quantitated by flowcytometry of PI staining. The mean and cell line, was utilized in a similar IL-3 withdrawal assay fol- standard error from three independent experiments are shown lowing transduction.31 Overexpression of SPIN-2 in NFS/N1- in Figure 2b, illustrating that SPIN-2 expressing cells consist- H7 cells (transduction efficiency ෂ50%) was protective follow- ently provided ෂ25% protection from apoptosis compared to ing 72 h of IL-3 withdrawal (30.7% alive), as was bcl-2 (44%), IRES-GFP transduced control cells (ෂ1–2%). This finding was vs the control cells (17.1%), indicating that SPIN-2 protection

Leukemia SPIN-2 effects on cell cycle and apoptosis BS Fletcher et al 1512 is not a cell line-dependent event (data not shown). In con- lated region of the SPIN-2 gene was identified. This region trast, SPIN-2 was unable to block the pro-apoptotic effects of of the X chromosome has been implicated in several genetic Fas ligand stimulation in the human T cell line Jurkat (data disorders including incontinentia pigmenti type 1 (MIM not shown). Furthermore, SPIN-2 had no protective effect on 30830), however, the genes involved have yet to be ident- 32Dcl3 cells treated with the chemotherapy agent doxorub- ified.34 Analysis of the genomic sequences immediately icin (115 ␮M) (not shown). SPIN-2 expressing cells were actu- upstream of the presumed SPIN-2 transcriptional initiation site ally more sensitive to doxorubicin-mediated cell death (6% revealed several elements consistent with a . This alive) compared to the negative control (16%), whereas bcl- included numerous SP1 sites as well as a binding site for the 2 was protective (47%) (data not shown). CAATT enhancer binding protein.35 At the message level, the SPIN-2 and DXF34 mRNAs are highly similar but differ substantially at their extreme 5Ј SPIN-2 protein homologies, genetic locus and untranslated regions (5ЈUTR). The 5ЈUTR of SPIN-2 is only expression patterns 128 nucleotides in length, whereas the 5ЈUTR of DXF34 is ෂ400 bases long and they share no sequence homologies. The Basic local alignment search technique (BLAST) analysis of the sequences of the two genes become identical 4 bp upstream SPIN-2 protein sequence demonstrated homology with two of the start methionine and remain identical throughout the proteins involved in germ cell development (spindlin and Ssty) open-reading frame (ORF), except for two nucleotide changes. and a third protein expressed in brain (DXF34). Figure 3 shows One of these nucleotide changes results in an aa substitution; an amino acid (aa) alignment of SPIN-2 and the related pro- Thr-234 in DXF34 is changed to an Ala in SPIN-2. Following teins DXF34, spindlin and Ssty. Of these three proteins, only the coding region, the 3Ј untranslated region of SPIN-2 is spindlin has been characterized biochemically, and was 170 bp long and is highly similar to DXF34, differing by only found to be abundant in developing oocytes.30 The DXF34 4 bp. The SPIN-2 gene initiates polyadenylation early com- gene was identified by screening a brain cDNA library with pared to the much longer 3Ј untranslated region of the DXF34 a probe for a repeated sequence on the human X chromosome gene (1.3 kb). The non-coding 5Ј and 3Ј UTRs were likely (GenBank Accession No. NP 061876).32 The Ssty protein under less selective pressure and were thus free to diverge. (spermatogenesis specific Y-linked transcript) is reported to be The high degree of identity within the coding regions of these expressed only during spermatogenesis and is essential for two genes and their close proximity on the X chromosome normal sperm head development.33 The SPIN-2 protein (258 implies that they formed as a result of gene duplication. aa) is almost identical to the DXF34 protein (258 aa) except We determined the steady-state SPIN-2 mRNA levels in for one aa difference described below. Shown in the align- 32Dcl3 cells by Northern blot analysis. In control 32Dcl3 ment plot in Figure 3 is a corrected version of the DXF34 pro- cells, both with and without IL-3, no expression of SPIN-2 tein that was reannotated based on sequence data from could be detected from total RNA using a probe for the full genomic DNA (GenBank Accession No. XP 010226). length SPIN-2. In contrast, retrovirally transduced cells The location of the SPIN-2 gene was identified by sequence expressing full length SPIN-2 had abundant message (data not homology to a region 13 kb telomeric of the DXF34 gene on shown). To characterize the SPIN-2 transcriptional expression the short arm of the X chromosome at Xp11.21. Utilizing the patterns, a human multiple tissue expression array (Clontech) human X chromosome sequence (GenBank Accession was probed with a radiolabeled oligonucleotide that is spe- No. AL022157) and comparing that with the SPIN-2 cDNA cific for the 5Ј UTR of SPIN-2. The tissue blot revealed low- sequence, the presence of a 661 bp intron in the 5Ј untrans- level baseline of SPIN-2 in all tissues tested

Figure 3 Alignment plots of SPIN-2 and related proteins. Blast homology searching of the SPIN-2 coding region against the GenBank database revealed extensive homologies with the proteins DXF34, spindlin and Ssty. Amino acid identities amongst all listed proteins are shaded in black and similar aa in gray. A potential ribonucleotide binding site (RNP-1) and a tyrosine phosphorylation site within the C-terminus are shown. The GenBank accession numbers for spindlin and the Ssty proteins are NP 006708 and NP 033246, respectively. Accession numbers for the other proteins are described within the text.

Leukemia SPIN-2 effects on cell cycle and apoptosis BS Fletcher et al 1513 (Figure 4), suggesting SPIN-2 is a constitutively expressed the oligonucleotide used contains a contiguous 14 bp region gene. This would be expected for a gene that is required to that is identical to E. coli DNA. continuously pass positive growth factor signals. This finding agrees with database results indicating that ESTs identical to SPIN-2 have been detected in numerous tissues, whereas no The C-terminus is essential for the anti-apoptotic ESTs identical to the 3Ј UTR region of DXF34 have been ident- activity of SPIN-2 ified. The level of expression of SPIN-2 appears to be the high- est in adult liver (A9) and is also strongly expressed in fetal To delineate the functional domains of SPIN-2 that elicit the liver (D11). Other tissues with elevated expression of SPIN-2 anti-apoptotic effect, a series of truncation mutants were gen- include the heart (A4), stomach (B5), kidney (A7), skeletal erated and assayed in the IL-3 withdrawal assay (Figure 5). muscle (B7), placenta (B8) and pancreas (B9). All transformed Alignment plots of SPIN-2 and related proteins revealed cell lines analyzed, A10 to H10, express relatively lowlevels slightly divergent N-termini and conservation within the other of SPIN-2. Several of the controls within the array gave a posi- domains of the proteins (see above). Of note is a conserved tive signal when probed with the SPIN-2 oligonucleotide 8 aa region within the extreme C-terminus of the protein that (A12–H12). The human DNA spots G12 and H12 are may contain a possible ribonuclear-binding domain (RNP-1, expected to be positive, as this probe is identical to genomic Figure 3). In addition, the C-terminus contains a conserved

DNA. The spot F12 that contains human C0t–1 DNA is also tyrosine residue that is a potential phosphorylation site in positive, likely because this region of the X chromosome has spindlin (residue 250).30 To test the hypothesis that the C-ter- surrounding repetitive elements. The strong positive signal minus is essential for the anti-apoptotic function of SPIN-2, a within the E. coli DNA spot (D12), may be due to the fact that retroviral expression construct was made in which the last nine aa of SPIN-2 were deleted (SPIN-2 ⌬C-term). Additional deletion constructs included a N-terminal truncation of 26 aa (SPIN-2 ⌬N-term) and a double C and N-terminal truncation (SPIN-2 ⌬N/C-term). The functional activities of the two c-myc epitope-tagged proteins were also investigated (mycSPIN-2 full ORF and mycSPIN-2 ⌬N-term). As shown in Figure 5, deletion of the C-terminal residues markedly diminished the anti-apoptotic activity of SPIN-2, whereas deletion of the N- terminal residues only slightly impaired its function. The myc- tagged constructs retained their full functional activity. During the sequencing of these expression clones we noted that there was a one-nucleotide change in all the SPIN-2 expression

Figure 4 Tissue expression patterns of SPIN-2. An oligonucleotide specific for the 5ЈUTR region of SPIN-2 was hybridized to the Multiple Tissue Expression Array (Clontech). mRNA from a variety of human tissues, as indicated below, was dot blotted into the array and nor- malized based on the expression of eight different housekeeping genes. A1, whole brain; B1, cerebral cortex; C1, frontal lobe; D1, parietal lobe; E1, occipital lobe; F1, temporal lobe; G1, paracentral gyrus; H1, pons; A2, left cerebellum; B2, right cerebellum; C2, corpus callosum; D2, amygdala; E2, caudate nucleus; F2, hippocampus; G2, medulla oblongata; H2, putamen; A3, substantia nigra; B3, accum- bens nucleus; C3, thalamus; D3, pituitary gland; E3, spinal cord; A4, heart; B4, aorta; C4, left atrium; D4, right atrium; E4, left ventricle; F4, right ventricle; G4, ventricular septum; H4, apex of heart; A5, esophagus; B5, stomach; C5, duodenum; D5, jejunum; E5, ileum; F5, iliocecum; G5, appendix; H5, ascending colon; A6, transverse colon; B6, descending colon; C6, rectum; A7, kidney; B7, skeletal muscle; C7, spleen; D7, thymus; E7, blood leukocyte; F7, lymph node; G7, bone marrow; H7, trachea; A8, lung; B8, placenta; C8, bladder; D8, uterus; E8, prostate; F8, testis; G8, ovary; A9, liver; B9, pancreas; C9, Figure 5 Protein mapping of the anti-apoptotic domains in SPIN- adrenal gland; D9, thyroid gland; E9, salivary gland; F9, mammary 2. A series of deletion constructs of the SPIN-2 gene were generated gland; A10, leukemia, HL-60; B10, HeLa S3; C10, leukemia, K-562; and tested in the IL-3 withdrawal assay. The data graphed are the D10, leukemia, MOLT-4; E10, Burkitt’s lymphoma, Raji; F10, Burkitt’s percentage of live cells quantitated by PI staining and FACS analysis lymphoma, Daudi; G10, colorectal adenocarcinoma, SW480; H10, following 2 days of IL-3 withdrawal. The values represent the mean lung carcinoma, A549; A11, fetal brain; B11, fetal heart; C11, fetal and standard deviation of three independent experiments done in kidney; D11, fetal liver; E11, fetal spleen; F11, fetal thymus; G11, duplicate. The no insert control consisted of 32Dcl3 cells transduced fetal lung; A12, total RNA; B12, yeast tRNA; C12, E. coli rRNA; with retrovirus expressing no protein. The sizes of the various con-

D12, E. coli DNA; E12, poly r(A); F12, human C0t-1 DNA; G12, structs are depicted graphically and are described in detail in the human DNA 100 ng; H12, human DNA 500 ng. The following pos- method section. The symbol ⌬ refers to deletion of a specific region itions contained no blotted RNA or DNA: F3, G3, H3, D6, E6, F6, of the SPIN-2 protein. A paired Student’s t-test was performed on G6, H6, H8, G9, H9 and H11. selected constructs and is described in the Results.

Leukemia SPIN-2 effects on cell cycle and apoptosis BS Fletcher et al 1514 constructs compared to the consensus genomic sequence for SPIN-2. This mutation results in a single amino acid substi- tution of glycine for a glutamine, at position 167. Expression of a SPIN-2 protein with this aa corrected (SPIN-2 ORF wild type) showed similar activity to the mutated form (SPIN-2 full ORF), indicating that this aa substitution does not modify SPIN-2’s functional activity. Of note is that these experiments were performed without positive selection for cells expressing the various transgenes (comparable to Table 1 and in contrast to Figure 2) and the infection efficiency, as determined by infection with GFP expressing retrovirus under similar con- ditions, was only 30–40%. Statistical analysis between the control and the full length SPIN-2 cDNA or full ORF was sig- nificant with a P value of 0.05 and 0.03, respectively (one- tailed paired Student’s t-test). A comparison between the SPIN-2 full ORF and the ⌬C-terminal deletion mutant (or the double ⌬N/C-mutant) was not significant (P value of 0.07), however, a clear trend towards diminished activity was seen. To further support the hypothesis that the C-terminus of SPIN- 2 is essential for its survival promoting activity, a ⌬C-term- IRES-GFP construct was tested in the IL-3 withdrawal assay following FACS sorting for GFP-positive cells. This construct provided no survival advantage compared to the control (data not shown).

Spin-2 overexpression perturbs cell cycle progression

One potential mechanism by which SPIN-2 may elicit its anti- apoptotic activity is through altering cell cycle progression. This hypothesis stems from the fact that some anti-apoptotic molecules function to alter cell cycle progression such as galectin-3.36 Bcl-2 has also been shown to restrain cell cycle

progression at the G0/G1 and G2/M transitions, however, these functions have been found to be independent of its anti-apop- totic function.37,38 To test the effects of SPIN-2 overexpression on cell cycle regulation, the growth kinetics of control 32Dcl3 cells were compared to FACS sorted GFP-positive (Ͼ90%) 32Dcl3 cells expressing either the IRES-GFP, bcl-2-IRES-GFP Figure 6 Cell growth and cell cycle analysis of SPIN-2 expressing or SPIN-2-IRES-GFP constructs. Equal numbers (1 × 105 cells. The growth characteristics of control 32Dcl3 cells, or those cells/ml) of these various cell lines were plated in separate transduced with the various pBMN-IRES-GFP expressing constructs is flasks and the cell densities were quantitated every 24 h for 3 demonstrated in panel a. Panels b and c illustrate cell cycle analysis days. These studies revealed that SPIN-2 expressing cells have of 32Dcl3 cells expressing either the control IRES-GFP or the SPIN- 2-IRES-GFP construct in the presence or absence of IL-3 for 24 h, longer doubling times and slower growth rates (Figure 6a). respectively. In panel c, intact cells from identical samples were ana- Analysis of the cell cycle distribution of exponentially growing lyzed for viability by staining with PI and the percentage of PI positive cells by propidium iodide staining revealed that SPIN-2-IRES- (dead) cells is shown. Panel d shows the growth kinetics and cell cycle GFP expressing cells had a decreased percentage of cells in S analysis of the C-terminal deletion construct designated SPIN⌬CT. In phase (35%) compared to control GFP expressing cells (44%, the cell cycle analyses, the Y-axis is number of events and the X-axis Figure 6b). This finding is in agreement with results of the is PI staining. The percentage of cells in each phase of the cell cycle was calculated using Modfit software. A color key for this software is growth studies. Notably, SPIN-2-IRES-GFP expressing cells as follows: red, diploid G0/G1 or G2/M; hatched blue, diploid S-phase; had a marked increase in the percentage of cells in G2/M yellow, polyploid G0/G1 or G2/M; teal, debris; purple, apoptotic (31.8%) compared to control GFP cells (12.3%). This increase DNA fragmentation.

in the percentage of cells in G2/M was even more pronounced following IL-3 withdrawal (25% vs Ͻ1%) (Figure 6c). Further- more, while control IRES-GFP cells deprived of IL-3 had already begun fragmentation of their DNA (purple color in Figure 6c), a large fraction of the SPIN-2-IRES-GFP cells were bation and cell survival, cell cycle analysis of 32Dcl3 cells still in G1/G0 (62%), with minimal DNA fragmentation. PI staining of intact cells from identical samples showed less PI- overexpressing the C-terminal deletion construct was perfor- ⌬ positive cells in the SPIN-2 expressing cells (51% vs 81%). med. SPIN CT-IRES-GFP cells grewsimilar to controls, These results support the hypothesis that the SPIN-2-mediated whereas SPIN-2-IRES-GFP expressing cells grew more slowly ⌬ perturbation of cell cycle progression influences cell survival, as before (Figure 6d). Furthermore, SPIN CT expressing cells did not exhibit an increase in the percentage of cell in G /M as G2/M arrested cells appear more viable and only after pro- 2 gression to G /G do they undergo DNA fragmentation. To (15%) (Figure 6d), compared to SPIN-2 expressing cells with 1 0 ෂ further support this relationship between cell cycle pertur- aG2/M fraction of 31% (Figure 6b).

Leukemia SPIN-2 effects on cell cycle and apoptosis BS Fletcher et al 1515 SPIN-2 is a nuclear protein methods). It has been shown previously that treatment with saponin does not permeabilize the nuclear membrane; there- To determine the subcellular localization of the SPIN-2 pro- fore antibodies against intranuclear antigens will not label the tein, the myc-epitope tag (EQKILSEEDL) was inserted after the cells.39 As predicted from panel a, saponin permeabilization start methionine of the complete SPIN-2 coding sequence. of SPIN-2 infected cells did not result in positive nuclear anti- Retrovirally transduced NIH3T3 cells were processed for dou- myc immunoreactivity (insert in panel a). After treatment with ble immunolabeling with monoclonal anti-myc antibodies saponin, all cells, including GFP-infected samples, gave low, and various organelle markers, 48 h post infection. NIH3T3 non-specific immunoreactivity with the anti-myc antibodies. cells were employed for these experiments due to their high Overexpression of the N-terminal truncated myc-tagged SPIN- transduction efficiencies, adherence to glass coverslips and 2 protein in NIH3T3 cells gave an identical nuclear localiz- classic morphology. As a positive control for transduction ation (Figure 7d). In comparison, the subcellular localization (ෂ90% transduction efficiency by FACS), GFP infected cells of the C-terminal truncated myc-tagged SPIN-2 was dramati- were examined in parallel for GFP fluorescence, which dem- cally different, showing a uniform mostly cytoplasmic onstrated cytoplasmic staining (data not shown). Figure 7, staining (Figure 7e). panels a (anti-myc), b (ConA) and c (Hoechst) showthe same The predicted molecular mass of the SPIN-2 protein is viewof a triple labeled sample. As Figure 7a indicates, the ෂ32 kDa. To verify equal expression of the full length, N- and majority of the myc-tagged SPIN-2 protein was localized in C-terminal truncated SPIN-2 proteins, in parallel with the the nucleus of transduced cells. An uninfected cell (arrowin immunolocalization studies, which showed equal transduc- panel a, b and c) showed low levels of non-specific immuno- tion efficiencies, Western blot analyses of cell lysates were reactivity with the anti-myc antibody over the cytoplasm performed. As shown in Figure 7f, the monoclonal anti-myc (panel a). Double labeling the samples with FITC-conjugated antibody detected a single protein product for each construct. concavalin A (an ER marker) (Figure 7b) revealed that over- The migration patterns of the resulting immunoreactive pro- expression of the SPIN-2 protein did not affect the normal ER ducts are as predicted from the aa sequences, at ෂ34 kDa for morphology of the cells (compare infected and uninfected the full length form (lane 2), ෂ30 kDa for the N-terminal trunc- cells in panel b). Panel c shows the location of the cell nucleus ated form (lane 3) and ෂ33 kDa for the C-terminal truncation by staining with the Hoechst nuclear dye. In order to further mutant (lane 4). As revealed in the Western blot, all three verify the nuclear localization of SPIN-2, parallel cultures forms of SPIN-2 are about equally expressed given the equal were immunostained with anti-myc antibodies after permeabi- transduction efficiencies and equal protein loading (75 ␮g lization with saponin, rather than methanol (see Materials and protein loaded in each lane). Samples infected with the GFP control vector did not react with the anti-myc antibody (Figure 7f, lane 1).

Discussion

Apoptosis can be activated by at least two independent path- ways. The extrinsic pathway involves ligands that bind to plasma membrane receptors (ie TNF␣ or FasL receptor), whereas the intrinsic pathway is activated by numerous stim- uli including the removal of essential survival factors. Many of the genes that are involved in the extrinsic pathway of apoptosis have been elucidated, while those regulating the intrinsic pathway are less well understood. Utilizing func- tional expression cloning, here we report on the identification of a novel gene, SPIN-2, that can retard apoptosis following growth factor withdrawal. In addition to its anti-apoptotic function, SPIN-2 overexpression alters cell cycle progression, causing decreased growth rates and perturbing cell cycle distributions. Retroviral cDNA library approaches have been successfully utilized to clone genes that regulate apoptosis; specifically Fas-mediated apoptosis, Toso,21 and c-myc-induced Figure 7 Immunolocalization and Western analysis of myc-tagged apoptosis, twist.40 An advantage of functional cloning over SPIN-2 protein in NIH3T3 cells. NIH3T3 cells transduced with myc- other methods of detecting altered gene expression is that tagged SPIN-2 constructs were grown on glass cover slips and stained functional cloning allows for identification of point mutations, with a monoclonal anti-myc antibody (panel a), concavalin A, an ER deletions or inversions that may alter protein activities. An marker (panel b), and the Hoechst nuclear dye (panel c). The insert within panel a demonstrates the lack of signal when the cells were example includes the TrkA deletion mutation that was ident- permeabilized with saponin, which does not penetrate the nuclear ified in a retroviral cDNA library derived from a patient with membrane. Panels d and e illustrate the localization of the myc-tagged leukemia.41 These types of modifications would not be N-terminal and C-terminal truncation proteins, respectively. Panel f detected by other wide approaches, such as cDNA is a Western blot probed with anti-myc monoclonal antibody. Lane chip technologies, the latter of which also does not determine 1 contains total lysate from GFP transduced cells, lane 2 is lysate from causal events vs caused events. The selection process utilized cells expressing full-length myc-tagged SPIN-2 protein, lane 3 is lysate from the N-terminal truncation mutant of SPIN-2 and lane 4 is lysate in this study involved repetitive rounds of IL-3 withdrawal fol- from the C-terminal truncation mutant. The locations of molecular lowed by recovery of the remaining viable cells by the re- mass standards are indicated in kDas. addition of IL-3. The screen yielded genes with known anti-

Leukemia SPIN-2 effects on cell cycle and apoptosis BS Fletcher et al 1516 apoptotic functions (bcl-XL, ODC) and two genes that pre- and degraded thereby altering its activity and allowing or faci- viously have not been demonstrated to be involved in apop- litating progression of meiosis. We are currently investigating totic processes (SPIN-2, ONT) (see Table 1). Although some whether the activities of SPIN-2 might be controlled by phos- of these clones were less efficacious at protecting 32Dcl3 cells phorylation. The exact mechanisms by which these proteins from apoptosis as compared to bcl-2, the fact that they sur- carry out their cell cycle regulatory functions have yet to be vived in an IL-3 withdrawal screen shows that the phenotype determined. A recent article by Staub and colleagues,45 sug- is sufficient to confer survival. gests that members of this family of proteins are composed of In the functional screen for anti-apoptotic genes, the SPIN- repeated domains (three repeats) of ෂ50 amino acids, which 2 gene exhibited the next strongest survival activity after that form a four-stranded ␤-structure. Howthis structure relates to of the bcl-2 family members. This ability of SPIN-2 to protect function awaits determination. cells undergoing apoptosis was tested in several paradigms Several potential mechanisms exist by which SPIN-2 over- including extrinsic pathway stimulation (Fas ligand expression may enhance cell survival following IL-3 with- stimulation), intrinsic pathway activation (growth factor drawal. One possible mechanism is that the anti-apoptotic withdrawal) and cytotoxic cell death (chemotherapy). Interest- effect of SPIN-2 overexpression is mediated through its ability

ingly, SPIN-2 protected cells only in factor withdrawal models to cause alterations at the G2/M transition. Just as oocytes in of apoptosis, whereas it moderately enhanced apoptosis fol- metaphase II of meiosis must stay dormant for many months lowing exposure to the chemotherapeutic agent doxorubicin. until they fully mature and are ovulated, it is possible that

In comparison, bcl-2 was protective in all of the tested para- cells slowed as they pass through G2/M by spindlin-like genes digms. The enhanced killing by doxorubicin may be due to become more resistant to apoptotic signals. This idea is diffi-

the fact that this agent is maximally cytotoxic in S phase and cult to test as pharmacologic agents that cause G2/M arrest 42 promotes cell cycle arrest and death in G2 phase. SPIN-2 (such as nocodozole) can induce apoptosis, likely due to toxic overexpression likely enhanced this susceptibility due to its effects from microtubule inhibition.6 SPIN-2 overexpression

effect on cell cycle progression and partial arrest in G2/M allows for a biological suppression of cell cycle progression (Figure 6b). These data imply that SPIN-2 is protective only if that is non-toxic and thereby allows the protective effects of

apoptosis is triggered under certain criteria. Although SPIN-2 G2/M arrest to be fully realized. Recently, it has been shown is a nuclear gene, there are no data to support the notion that that the expression of the IAP gene, survivin, is cell cycle regu- SPIN-2 is a transcription factor or that it can regulate the lated.46 The highest levels of survivin mRNA and protein

expression of survival cytokines such as IL-3 or occur in G2/M phase where the protein is found to associate granulocyte/macrophage colony-stimulating factor (GM-CSF). with the mitotic spindle. Like spindlin, and possibly SPIN-2, At present, the mechanism of SPIN-2 action remains survivin is regulated by phosphorylation.47 It is postulated that uncharacterized. survivin may function not only to inhibit caspases, but also to Of interest is the homology of SPIN-2 with the protein stabilize the mitotic spindle apparatus. Thus, it is possible that spindlin. The spindlin gene was cloned and characterized as the anti-apoptotic effects of SPIN-2 may be due only to its an abundant protein found in developing oocytes and early effect on cell cycle progression and subsequent upregulation 30 embryos. Spindlin has been demonstrated to undergo differ- of G2/M specific genes, like survivin. Our current characteriz- ential phosphorylation during the meiotic and first mitotic cell ation of SPIN-2 allows for us to proceed with this hypothesis. cycles. The localization of the spindlin protein was mainly In summary, we have identified by functional expression cytoplasmic, however, with resumption of the meiotic cell cloning a novel protein, SPIN-2, that partially protects cells cycle, a portion of the protein was found to associate with the from growth factor withdrawal-mediated apoptosis. The anti- meiotic spindle apparatus.30 This association of spindlin with apoptotic activity of SPIN-2 is dependent upon the C-terminus the meiotic spindle was found to be phosphorylation depen- of the protein and is correlated with a change in progression

dent as MOS kinase-deficient oocytes showed weak spindlin– of cells through the G2/M phase of the cell cycle. Given that spindle interactions.43 The spindlin protein and message dis- SPIN-2 is highly similar to spindlin, a phosphoprotein associa- appear upon fertilization of the oocytes and formation of the ted with cell cycle events during gametogenesis and meiosis, early embryo. Investigations into the localization of SPIN-2 in it is possible that SPIN-2 plays a fundamental role regulating NIH3T3 cells indicate that SPIN-2 is a nuclear protein and we cell division in a variety of cell types. Since overcoming factor found no evidence of microtubule association even in cells dependence and other pro-apoptotic stimuli is thought to be undergoing mitosis (data not shown). Recent characterization a crucial requirement of leukemia formation, the identification of the chicken ortholog of SPIN-2, termed chSpin-Z, was also of anti-apoptotic genes, like SPIN-2, may reveal general found to be a nuclear protein.44 Further investigations are mechanisms that could be targeted for anti-leukemic needed to characterize the expression levels and intracellular strategies. locations of the endogenous SPIN-2 protein in hematopoietic cells and in other cells throughout the cell cycle. Although spindlin and SPIN-2 appear to have different sub- cellular localizations, they may have similar roles in reg- Acknowledgements ulating cell cycle progression. Whereas overexpression of

SPIN-2 in mitotic cells results in slowing of the cells in G2/M, the effects of overexpression of spindlin are not known. Since We thank Alan Friedman for the 32Dcl3 cells, Vijay Antharam meiosis and mitosis are similar cellular processes, the func- for technical assistance and Drs Stratford May and Steven tions of spindlin in meiosis might lead to a better understand- Baker for critical reading of the manuscript. The nucleotide ing of howSPIN-2 functions, if at all, in mitosis. By analogy sequence for the SPIN-2 gene has been submitted to the Gen- to the effects of SPIN-2 on mitosis, it is possible that spindlin Bank database under the Accession Number AF356353. A functions in developing oocytes to maintain cells arrested in portion of this work was supported by a grant to BSF from metaphase II of meiosis. Upon fertilization and subsequent the Howard Hughes Medical Institute, Postdoctoral Research resumption of meiosis, spindlin becomes dephosphorylated Fellowship for Physicians.

Leukemia SPIN-2 effects on cell cycle and apoptosis BS Fletcher et al 1517 References GP. The T cell activation factor NF-ATc positively regulates HIV- 1 replication and gene expression in T cells. Immunity 1997; 6: 1 Li J, Eastman A. Apoptosis in an interleukin-2-dependent cytotoxic 235–244. T lymphocyte cell line is associated with intracellular acidifi- 23 Griffiths G, Brands R, Burke B, Louvard D, Warren G. Viral mem- + + cation. Role of the Na( )/H( )-antiport. J Biol Chem 1995; 270: brane proteins acquire galactose in trans Golgi cisternae during 3203–3211. intracellular transport. J Cell Biol 1982; 95: 781–792. 2 AskewDS, Ashmun RA, Simmons BC, Cleveland JL. Constitutive 24 Roth J, Wang Y, Eckhardt AE, Hill RL. Subcellular localization of c-myc expression in an IL-3-dependent myeloid cell line sup- the UDP-N-acetyl-D-galactosamine: polypeptide N-acetylgalacto- presses cell cycle arrest and accelerates apoptosis. Oncogene saminyltransferase-mediated O-glycosylation reaction in the sub- 1991; 6: 1915–1922. maxillary gland. Proc Natl Acad Sci USA 1994; 91: 8935–8939. 3 van Engeland M, Nieland LJ, Ramaekers FC, Schutte B, Reutelings- 25 Blandino G, Scardigli R, Rizzo MG, Crescenzi M, Soddu S, Sacchi perger CP. Annexin V-affinity assay: a reviewon an apoptosis A. Wild-type p53 modulates apoptosis of normal, IL-3 deprived, detection system based on phosphatidylserine exposure. Cytome- hematopoietic cells. Oncogene 1995; 10: 731–737. try 1998; 31: 1–9. 26 Gaozza E, Baker SJ, Vora RK, Reddy EP. AATYK: a novel tyrosine 4 Bojes HK, Feng X, Kehrer JP, Cohen GM. Apoptosis in hematopo- kinase induced during growth arrest and apoptosis of myeloid ietic cells (FL5.12) caused by interleukin-3 withdrawal: relation- cells. Oncogene 1997; 15: 3127–3135. ship to caspase activity and the loss of glutathione. Cell Death 27 Greenberger JS, Sakakeeny MA, Humphries RK, Eaves CJ, Eckner Differ 1999; 6: 61–70. RJ. Demonstration of permanent factor-dependent multipotential 5 Lilly M, Sandholm J, Cooper JJ, Koskinen PJ, Kraft A. The PIM-1 serine kinase prolongs survival and inhibits apoptosis-related mito- (erythroid/neutrophil/basophil) hematopoietic progenitor cell chondrial dysfunction in part through a bcl-2-dependent pathway. lines. Proc Natl Acad Sci USA 1983; 80: 2931–2935. Oncogene 1999; 18: 4022–4031. 28 Packham G, Cleveland JL. Ornithine decarboxylase is a mediator 6 Garland J, Brown G, Beasley J, Brown R. Apoptosis in factor- of c-Myc-induced apoptosis. Mol Cell Biol 1994; 14: 5741–5747. dependent haematopoietic cells is linked to calcium-sensitive 29 Wells L, Vosseller K, Hart GW. Glycosylation of nucleocyto- mitochondrial rearrangements and cytoskeletal modulation. Br J plasmic proteins: signal transduction and O- GlcNAc. Science Haematol 2000; 109: 221–234. 2001; 291: 2376–2378. 7 Wolter KG, Hsu YT, Smith CL, Nechushtan A, Xi XG, Youle RJ. 30 Oh B, Hwang SY, Solter D, Knowles BB. Spindlin, a major Movement of Bax from the cytosol to mitochondria during maternal transcript expressed in the mouse during the transition apoptosis. J Cell Biol 1997; 139: 1281–1292. from oocyte to embryo. Development 1997; 124: 493–503. 8 Kluck RM, Bossy-Wetzel E, Green DR, Newmeyer DD. The 31 Deng X, Ito T, Carr B, Mumby M, May WS, Jr. Reversible phos- release of cytochrome c from mitochondria: a primary site for Bcl- phorylation of Bcl2 following interleukin 3 or bryostatin 1 is 2 regulation of apoptosis. Science 1997; 275: 1132–1136. mediated by direct interaction with protein phosphatase 2A. J Biol 9 Liu X, Kim CN, Yang J, Jemmerson R, Wang X. Induction of apop- Chem 1998; 273: 34157–34163. totic program in cell-free extracts: requirement for dATP and cyto- 32 Laval SH, Reed V, Blair HJ, Boyd Y. The structure of DXF34, a chrome c. Cell 1996; 86: 147–157. human X-linked sequence family with homology to a transcribed 10 Li P, Nijhawan D, Budihardjo I, Srinivasula SM, Ahmad M, mouse Y-linked repeat. Mamm Genome 1997; 8: 689–691. Alnemri ES, Wang X. Cytochrome c and dATP-dependent forma- 33 Conway SJ, Mahadevaiah SK, Darling SM, Capel B, Rattigan AM, tion of Apaf-1/caspase-9 complex initiates an apoptotic protease Burgoyne PS. Y353/B: a candidate multiple-copy spermiogenesis cascade. Cell 1997; 91: 479–489. gene on the mouse Y chromosome. Mamm Genome 1994; 5: 11 Hockenbery DM, Oltvai ZN, Yin XM, Milliman CL, Korsmeyer SJ. 203–210. Bcl-2 functions in an antioxidant pathway to prevent apoptosis. 34 Reed V, Rider S, Maslen GL, Hatchwell E, Blair HJ, Uwechue IC, Cell 1993; 75: 241–251. Craig IW, Laval SH, Monaco AP, Boyd Y. A 2-Mb YAC contig 12 Hawkins CJ, Vaux DL. Analysis of the role of bcl-2 in apoptosis. encompassing three loci (DXF34, DXS14, and DXS390) that lie Immunol Rev 1994; 142: 127–139. between Xp11.2 translocation breakpoints associated with incon- 13 Kohler T, Leiblein S, Borchert S, Eller J, Rost AK, Lassner D, Krahl tinentia pigmenti type 1. Genomics 1994; 20: 341–346. R, Helbig W, Wagner O, Remke H. Absolute levels of MDR-1, 35 Quandt K, Frech K, Karas H, Wingender E, Werner T. MatInd and MRP, and BCL-2 MRNA and tumor remission in acute leukemia. MatInspector: newfast and versatile tools for detection of consen- Adv Exp Med Biol 1999; 457: 177–185. sus matches in nucleotide sequence data. Nucleic Acids Res 1995; 14 Lauria F, Raspadori D, Rondelli D, Ventura MA, Fiacchini M, Vis- 23: 4878–4884. ani G, Forconi F, Tura S. High bcl-2 expression in acute myeloid 36 Yoshii T, Fukumori T, Honjo Y, Inohara H, Kim HR, Raz A. Galec- leukemia cells correlates with CD34 positivity and complete tin-3 phosphorylation is required for its anti-apoptotic function remission rate. Leukemia 1997; 11: 2075–2078. and cell cycle arrest. J Biol Chem 2002; 277: 6852–6857. 15 Yang E, Zha J, Jockel J, Boise LH, Thompson CB, Korsmeyer SJ. 37 Huang DC, O’Reilly LA, Strasser A, Cory S. The anti-apoptosis Bad, a heterodimeric partner for Bcl-XL and Bcl-2, displaces Bax function of Bcl-2 can be genetically separated from its inhibitory and promotes cell death. Cell 1995; 80: 285–291. effect on cell cycle entry. EMBO J 1997; 16: 4628–4638. 16 Lin EY, Orlofsky A, Wang HG, Reed JC, Prystowsky MB. A1, a 38 Furukawa Y, Iwase S, Kikuchi J, Terui Y, Nakamura M, Yamada Bcl-2 family member, prolongs cell survival and permits myeloid H, Kano Y, Matsuda M. Phosphorylation of Bcl-2 protein by CDC2 differentiation. Blood 1996; 87: 983–992. kinase during G2/M phases and its role in cell cycle regulation. J 17 Reynolds JE, Yang T, Qian L, Jenkinson JD, Zhou P, Eastman A, Biol Chem 2000; 275: 21661–21667. Craig RW. Mcl-1, a member of the Bcl-2 family, delays apoptosis 39 Pichon B, Mercan D, Pouillon V, Christophe-Hobertus C, Christo- induced by c-Myc overexpression in Chinese hamster ovary cells. phe D. A method for the large-scale cloning of nuclear proteins Cancer Res 1994; 54: 6348–6352. 18 Deveraux QL, Reed JC. IAP family proteins – suppressors of and nuclear targeting sequences on a functional basis. Anal apoptosis. Genes Dev 1999; 13: 239–252. Biochem 2000; 284: 231–239. 19 LaCasse EC, Baird S, Korneluk RG, MacKenzie AE. The inhibitors 40 Maestro R, Dei Tos AP, Hamamori Y, Krasnokutsky S, Sartorelli V, of apoptosis (IAPs) and their emerging role in cancer. Oncogene Kedes L, Doglioni C, Beach DH, Hannon GJ. Twist is a potential 1998; 17: 3247–3259. oncogene that inhibits apoptosis. Genes Dev 1999; 13: 2207– 20 Irmler M, Thome M, Hahne M, Schneider P, Hofmann K, Steiner 2217. V, Bodmer JL, Schroter M, Burns K, Mattmann C, Rimoldi D, 41 Reuther GW, Lambert QT, Caligiuri MA, Der CJ. Identification and French LE, Tschopp J. Inhibition of death receptor signals by cellu- characterization of an activating TrkA deletion mutation in acute lar FLIP. Nature 1997; 388: 190–195. myeloid leukemia. Mol Cell Biol 2000; 20: 8655–8666. 21 Hitoshi Y, Lorens J, Kitada SI, Fisher J, LaBarge M, Ring HZ, 42 Ritch PS, Occhipinti SJ, Cunningham RE, Shackney SE. Schedule- Francke U, Reed JC, Kinoshita S, Nolan GP. Toso, a cell surface, dependent synergism of combinations of hydroxyurea with adria- specific regulator of Fas-induced apoptosis in T cells. Immunity mycin and 1-beta-D-arabinofuranosylcytosine with adriamycin. 1998; 8: 461–471. Cancer Res 1981; 41: 3881–3884. 22 Kinoshita S, Su L, Amano M, Timmerman LA, Kaneshima H, Nolan 43 Oh B, Hampl A, Eppig JJ, Solter D, Knowles BB. SPIN, a substrate

Leukemia SPIN-2 effects on cell cycle and apoptosis BS Fletcher et al 1518 in the MAP kinase pathway in mouse oocytes. Mol Reprod Dev from gamete to embryo. Genome Biol 2001; 3: 0003.1–0003.6. 1998; 50: 240–249. 46 O’Connor DS, Grossman D, Plescia J, Li F, Zhang H, Villa A, Tog- 44 Itoh Y, Hori T, Saitoh H, Mizuno S. Chicken spindlin genes on W nin S, Marchisio PC, Altieri DC. Regulation of apoptosis at cell and Z chromosomes: transcriptional expression of both genes and division by p34cdc2 phosphorylation of survivin. Proc Natl Acad dynamic behavior of spindlin in interphase and mitotic cells. Sci USA 2000; 97: 13103–13107. Chromosome Res 2001; 9: 283–299. 47 Li F, Ambrosini G, Chu EY, Plescia J, Tognin S, Marchisio PC, 45 Staub E, Mennerich D, Rosenthal A. The Spin/Ssty repeat: a new Altieri DC. Control of apoptosis and mitotic spindle checkpoint motif identified in proteins involved in vertebrate development by survivin. Nature 1998; 396: 580–584.

Leukemia