J Biol Chem, Vol. 275, Issue 10, 6707-6711, March 10, 2000

Cell Division Regulation by BIR1, a Member of the Inhibitor of Apoptosis Family in Yeast^*

Fengzhi Li^§, Paul L. Flanary^§¶, Dario C. Altieri, and Henrik G. Dohlman^¶**

From the Departments of  Pathology and ^¶ Pharmacology, Boyer Center for Molecular Medicine, Yale University School of Medicine, New Haven, Connecticut 06536

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

The inhibitor of apoptosis (IAP) gene family comprises molecules that block the activity of pro-apoptotic caspase proteases. Paradoxically, yeasts contain IAP proteins but no caspases and no apoptotic program. To determine the function of these proteins in vivo, we disrupted the BIR1 gene, encoding the only known IAP in yeast Saccharomyces cerevisiae. Sporulation of heterozygous diploids yielded no viable mutant haploids, indicating that BIR1 is an essential gene. By flow cytometry, some heterozygous mutants were polyploid accumulating >4 N DNA content. These cells exhibited a 20-40% reduction in growth rate, which was rescued by plasmid-borne over-expression of BIR1 but not by its human counterpart, survivin. Deletion analysis revealed that the N-terminal domain of Bir1, containing the conserved baculovirus IAP repeat, was able to partially complement the cell growth defect caused by BIR1 deletion. Moreover, the full-length and truncated forms of Bir1 accelerated cell division in wild-type cells. Finally, BIR1 heterozygous mutants exhibited grossly altered cell morphology with mis-shapen or abnormally long buds connected to an unusually large mother cell. These findings identify a novel function of IAP proteins in the pleiotropic control of cell division, in addition to their role in the suppression of apoptosis.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

The genetic control of cell death and cell survival (apoptosis) ensures developmental morphogenesis and preserves homeostasis in differentiated organisms (1). Apoptosis is a highly evolutionarily conserved process, with homologous gene families of inhibitors and stimulators of cell death present in all metazoa. Genetic studies in model organisms Drosophila melanogaster (2) and Caenorhabditis elegans (3) provided fundamental paradigms for the assembly of the cell death machinery and helped identify critical molecular targets of apoptosis effector proteins (4). In addition to the Bcl-2 family of cell death regulators (5), the inhibitor of apoptosis (IAP)¹ gene family has recently attracted attention for its broad distribution and evolutionary conservation (6, 7).

Originally found in the baculovirus as factors that prevent the suicide of virally infected host cells (8), homologous IAP proteins have since been identified in yeast, worms, flies, and mammalian cells (6, 7). Each molecule contains one to three copies of an 80-amino acid baculovirus IAP repeat (BIR), which are potentially involved in Zn²⁺ ion coordination via a DX₃CX₂C/HX₆C consensus (9). In addition, BIR proteins often contain a C₃HC₄ RING finger (6, 7). Functionally, a Drosophila IAP protein, DIAP, has been shown to counteract apoptosis induced by the death genes, reaper, hid, and grim (10). In contrast, mammalian IAP proteins (XIAP, NAIP, c-IAP1, c-IAP2, and survivin) block apoptosis by inactivating terminal effector caspase-3 and -7 and by interfering with upstream activation of caspase-9 (6, 7). This suggested that, although IAP proteins are evolutionarily conserved apoptosis inhibitors, they may have distinct mechanisms in different organisms.

Recent evidence suggested that certain IAP proteins might participate in cellular functions other than apoptosis and, in particular, cell division. Specifically, survivin, the smallest mammalian IAP protein with a single BIR domain and no RING finger (11), was found to be expressed in mitosis in a strict cell cycle-regulated manner (12). Starting at early prophase, survivin was localized to mitotic spindle microtubules persisting until late telophase in midbodies (12). Moreover, RNA interference of a single BIR-containing IAP gene, BIR-1, in C. elegans did not result in dysregulated apoptosis but caused an embryonic lethal defect of cytokinesis with multinucleation (13). Intriguingly, a single BIR gene is also present in yeast S. cerevisiae (6, 7), which lacks caspases and is not believed to possess an apoptosis program (14).

To determine the function of prototypic, single BIR-containing IAP proteins in vivo, we disrupted the BIR1 gene in yeast S. cerevisiae. This action produced a lethal phenotype, indicating that BIR1 is an essential gene. Moreover, analysis of heterozygous diploid mutants revealed a pleiotropic cell division defect, which affected ploidy, cell growth, and cell morphology, revealing a role of IAP proteins in cell division control.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

Strains, Media, and Transformation-- Standard methods for the growth and maintenance of the yeast and bacteria were used throughout (15). S. cerevisiae strain YPH501 (MATa/MAT ura3-52/ura3-52 lys2-801;am/lys2-801;am ade2-101;oc/ade2-101;oc trp1-63/trp1-63 his3-200/his3-200 leu2-1/leu2-1) (16) was used for the characterization of BIR1. Yeast cells were maintained in rich (YP) medium or synthetic (SC) medium lacking appropriate nutrients to maintain plasmid selection and were supplemented with either 2% dextrose (YPD, SCD) or 2% galactose and 0.2% sucrose (YPG, SCG). Cells were grown at 30 °C unless otherwise indicated.

To construct the bir1::URA3 mutant, the "left" and "right" portions of the bir1 gene (lacking a 2938-base pair segment of the coding region) were PCR-amplified using the primer pairs of yBIR1-P1 (5'-TGC TCT AGA TTG GTT TGT TTC TTG CTA CG-3') and yBIR1-P2 (5'-GGA AGA TCT ATA AAC AAC CGA TGT GTC C-3') or yBIR1-P3 (5'-GGA AGA TCT GAA TGT TTC GCC GCA TCG CAG-3') and yBIR1-P4 (5'-GAC GCG TCG ACC AAG ATT ATG TCG TAG CAC C-3'). The amplification products were digested with XbaI-BglII or BglII-SalI and subcloned into the XbaI-SalI sites of pBluescript KS, yielding pBluescript KS-BIR1-LR. The URA3 gene was subcloned as a 3.8-kilobase pair BglII-BamHI fragment (from plasmid pNKY51) into the BglII site of pBluescript KS-BIR1-LR, yielding the bir1::URA3 plasmid, pBluescript KS-BIR1-LR-URA3.

YPH501 cells were transformed using the LiAc TRAFO method, as described previously (17). To confirm each gene disruption, genomic DNA was prepared (15) and PCR-amplified using primers flanking yBIR-P0 (5'-CTG CTT CAA ATG CAT GTC), as well as internal to the disrupting auxotrophic marker. Disruptions were further confirmed using a set of primers internal to the BIR1 gene, yBIR1-5' (5'-TCC CCC GGG ATA ATG GAT GGT CAA ATA GAT AAA ATG G-3') and yBIR1-3' (5'-ACG CGT CGA CCT GCG ATG CGG CGA AAC ATT C-3'). For tetrad dissection, diploids were grown overnight at 24 °C in pre-sporulation medium (1% potassium acetate, 0.1% yeast extract, 0.67% yeast nitrogen base without amino acids and ammonium sulfate, supplemented with amino acids and uracil) and then washed and transferred to SPM sporulation medium (1% potassium acetate) for an additional 3-5 days. Dissections were carried out as described (15).

Expression Plasmids-- For expression of wild-type BIR1, the gene was PCR-amplified from genomic DNA using the primer pair yBIR1-P1 (5'-TGC TCT AGA TTG GTT TGT TTC TTG CTA CG-3') and yBIR1-P4 (5'-GAC GCG TCG ACC AAG ATT ATG TCG TAG CAC C-3'). The 4930-kilobase pair PCR product was digested with XbaI-SalI and subcloned into the corresponding sites of pRS315. For overexpression of BIR1, the full-length open reading frame was amplified using yBIR1-5' and yBIR1-3' (described above). The amplified product was digested with SmaI-SalI and subcloned into the corresponding sites of pRS315-ADH, a yeast expression vector containing the promoter and terminator from ADH1 (18). Similar plasmids containing the Bir1 N-terminal domain (containing two BIR repeats, residues 1-251) or the C-terminal domain (residues 252-954) were constructed by PCR, as described above, using internal primers yBIR1N-3' (5'-ACG CGT CGA CCT ATG TCT GAA AGA AAT AAC AGT TCC C-3') or yBIR1C-5' (5'-TCC CCC GGG ATA ATG CAG ACA CGT AAT CGA TTT GAG AG-3').

For expression of wild-type survivin, the cDNA was PCR-amplified using the primer pair of HS-Y5' (5'-CGG GAT CCA GAA TGG GTG CCC CGA CGT TGC-') and HS-Y3' (5'-GG AAT TCT CAA TCC ATG GCA GCC AG-'). The 450-base pair products were digested with BamHI-coRI and subcloned into the corresponding sites of pRS316-ADH to yield pRS316-ADH-survivin. Similar plasmids containing either the N-terminal (residues 1-96, N-survivin) and C-terminal (65 residues 97-142, C-survivin) domains of survivin were constructed by PCR as described above, using internal primers HS-N3' (5'-GG AAT TCT CAA GAA AGG AAA GCG CAA CCG GA-') and HS-C5' (5'-GG GAT CCA CAA TGC TTT CTG TCA AGA AGC AG-'). For galactose-inducible expression, the ADH1 promoter was removed (KpnI, SmaI fragment) and replaced with the GAL1/10 promoter (KpnI, blunt-ended XbaI) from pRS315-GAL, to yield pRS316-GAL-survivin. Alternatively, the coding region and ADH1 terminator were moved as a BamHI-BglII fragment to the BamHI site of pRS315-GAL to yield pRS315-GAL-survivin (including N-survivin and C-survivin). To substitute the URA3 nutritional marker with the LEU2, URA3-containing PvuI-PvuI fragment of pRS316-based vectors was replaced with the corresponding LEU2-containing fragment of pRS315 to yield pRS315-ADH-survivin and pRS315-GAL-survivin. All PCR products were confirmed by DNA sequencing.

Fluorescence-activated Cell Sorting (FACS)-- Cultures for FACS analysis were grown to saturation and then diluted to A_600 nm ~ 0.2 in the same medium and grown to A_600 nm ~ 1.0. 1.5 ml of the culture was sonicated for 10 s, centrifuged for 10 s at maximum speed in a tabletop microfuge, resuspended in 2 ml of 70% EtOH, and incubated overnight at 23 °C with gentle shaking. Cells were then centrifuged, washed twice with 1 ml of 50 mM Tris-HCl, pH 7.8, resuspended in 0.8 ml of 50 mM Tris-HCl, pH 7.8, and 200 µg of heat-inactivated RNase A, and incubated at 37 °C overnight. Cells were centrifuged, resuspended in 0.5 ml of 50 mM Tris-HCl, pH 7.8, 2.5 mg pepsin, and incubated at 37 °C for 30 min. Cells were centrifuged, washed once with 1 ml of 200 mM Tris-HCl, pH 7.5, 211 mM NaCl, 78 mM MgCl₂, and resuspended in 0.55 ml of the same buffer containing 50 µg of propidium iodide. Flow cytometric analysis was performed with a FACS Vantage flow cytometer (Becton Dickinson, San Jose, CA). The cells were excited at 488 nm, and the emission was collected through a 630/22 nm band pass filter. A minimum of 10,000 cells was analyzed for each sample. Cell cycle analysis was performed using the Modfit 5.2 model (Verity Software House, Topsham, ME).

Microscopy-- For light microscopic examination, cells were grown to mid-log phase, fixed in growth medium with formaldehyde (3.7% final concentration), and photographed with Kodak TMax 100 film using a Zeiss Axioscop and MC80 camera at 400× magnification with DIC optics.

Cell Growth-- Cell growth was monitored by diluting a saturated overnight culture to A_600 nm ~ 0.2 and measuring the change in absorbance over time. Cell viability was determined by serially diluting cultures grown to saturation, spotting onto the appropriate medium, and incubating for 48 h at 24, 30, 34, or 37 °C in the presence or absence of 1 M sorbitol.

	RESULTS AND DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES

The sequencing of the S. cerevisiae genome (19) has revealed a single open reading frame with homology to IAP family proteins. This gene, designated BIR1 (SCYJR089w), encodes a protein of 954 amino acids and features two BIR domains at residues 12-116 and 145-240. Phylogenetic analysis (20) revealed the greatest similarity to Schizosaccharomyces pombe bir1, human survivin, and C. elegans BIR-2 and BIR-1. A comprehensive analysis of cell cycle-regulated genes in yeast indicates that BIR1 transcription does not change during the cell cycle (21), in contrast to human survivin (12).

To determine the in vivo function of BIR1, we constructed a gene disruption mutation in the diploid strain YPH501. Cells were transformed with a construct in which the entire coding region had been replaced with a nutritional marker (bir1::URA3, bir1). Gene disruption mutants were confirmed by PCR and sporulated, and the resulting tetrads were dissected and allowed to grow on rich (YPD) medium. For one of the BIR1/bir1 mutants (strain YPH501-B2), sporulation yielded two viable and two nonviable spore products. The viable cells failed to grow in medium lacking uracil, confirming that the bir1::URA3 mutation is lethal. However, another mutant (YPH501-B9) yielded a complex pattern of spore products, some of which were viable but also able to grow in uracil-free medium.

The segregation pattern for YPH501-B9 did not match that expected for a simple suppressor mutation. We thus considered whether this strain had undergone endoreduplication, resulting in a higher than normal number of chromosomes. To test for this possibility, we measured the DNA content of YPH501-B9 using FACS analysis. Both the BIR1 and the bir1 mutants and the corresponding wild-type strains were characterized in this manner, as described in Fig. 1. YPH501 and YPH499 are wild-type diploid and haploid strains, respectively. As expected, YPH501 and YPH501-B2 each exhibited a normal (2 N + 4 N) pattern of DNA content. In contrast, YPH501-B9 had major peaks corresponding to 4 and 6 N DNA. A third mutant, YPH501-B6, contained 3 or 5 N DNA. We also characterized a viable, URA⁺ spore product from YPH501-B9 (designated YPH499-B9). Compared with the normal haploid strain YPH499, the YPH499-B9 mutant strain had twice the normal DNA content (2 and 4 N) (Fig. 1). Again, bir1 haploid mutants were not viable and could not be analyzed by FACS.

View larger version (10K): [in this window] [in a new window]   Fig. 1.   Some BIR1/bir1 mutants are polyploid. FACS analysis was used to determine whether BIR1/bir1 mutants had a higher than normal DNA content. Data are represented as cell number (Counts) versus fluorescence intensity (DNA content). YPH501 and YPH499 are wild-type diploid and haploid strains, respectively. YPH501-B9 and YPH501-B2 are BIR1/bir1 mutants derived from YPH501. YPH499-B9 is a BIR1/bir1 spore product derived from YPH501-B9.

We then examined each of the mutants by DIC microscopy (Fig. 2). Compared with the wild-type parent strain, YPH501, both the YPH501-B2 and YPH501-B9 mutants exhibited aberrant cell morphology, with mis-shapen buds connected to an unusually large mother cell. The YPH499-B9 mutant also had an altered cell morphology, except that the bud neck was grossly elongated in most cases. This morphology is distinct from that seen during pseudohyphal differentiation, because disruption or overexpression of BIR1 does not influence the formation of pseudohyphae resulting from nitrogen starvation or from permanently differentiated elm1 or tec1 mutant cells (22).

View larger version (70K): [in this window] [in a new window]   Fig. 2.   BIR1/bir1 mutants exhibit an altered cell morphology. Wild-type and BIR1/bir1 mutants were grown to mid-log phase, fixed in formaldehyde, and examined by DIC microscopy. Strain designations are described in Fig. 1. The elongated bud neck in the YPH499-B9 mutant is indicated (arrows).

Enlarged cells are often the consequence of slower cell division. Thus, we examined the ability of each mutant to grow in liquid medium (Fig. 3A). Compared with the wild-type YPH501, which had a doubling time of 107 min, YPH501-B2, YPH501-B6, and YPH501-B9 grew somewhat more slowly, with doubling times of 122, 122, and 134 min, respectively. Compared with the parent haploid strain, YPH499, the doubling time of YPH499-B9 was similarly increased (from 124 to 166 min). To determine whether the bir1 mutant was responsible for this slow-growth phenotype, YPH499 and YPH499-B9 were transformed with a plasmid-borne copy of BIR1 or the parent vector alone. As described in Fig. 3B, the mutant carrying the empty vector grew slowly (180 min, doubling time), whereas mutants carrying BIR1 on a plasmid doubled at a rate (157 min) comparable with that of the wild type carrying the vector (152 min) or BIR1 (150 min) (Fig. 3B).

View larger version (10K): [in this window] [in a new window]   Fig. 3.   The bir1 mutation slows the rate of cell division. The abilities of wild-type and bir1 mutant cells to grow in liquid medium were compared. Cells were diluted to A600 nm ~ 0.2 and monitored for the indicated times. A, growth of wild-type (, YPH499 (124 min)) versus BIR1/bir1 heterozygous mutant (, YPH499-B9 (166 min)) in rich (YPD) medium. B, growth of wild-type and mutant cells in selective (SCD) medium containing a plasmid-borne copy of BIR1 (pBIR1) or the parent vector. Data are represented as cell density (log A600 nm) versus time (hours). Each experiment was performed at least three times, with similar results. Doubling times were determined by calculating the slope of the linear portion of the curve. The values given represent the average of three independent experiments, each of which varied by less than 1 min. , YPH499 + vector (152 min); , YPH499-B9 + pBIR1 (157 min); , YPH499-B9 + vector (180 min).

We then examined the consequences of BIR1 over-expression. Our rationale was that if a reduction in gene copy number leads to a slow-growth phenotype, an increase in expression might accelerate growth. Wild-type YPH501 cells were transformed with a plasmid containing full-length BIR1 expressed using a strong constitutive promoter from ADH1. As described in Fig. 4, BIR1 over-expressing cells doubled at a rate considerably faster (111 min) than non-over-expressing cells (146 min) (Fig. 4A). To determine whether the BIR homology was responsible for the enhanced growth, we tested a C-terminally truncated form of the protein containing only the two BIR domains (residues 1-251, designated N-Bir1). As a control, we also tested the reciprocal construct containing an N-terminally truncated protein lacking the BIR domains (residues 252-954, C-Bir1). As shown in Fig. 4, the rate enhancement by N-Bir1 nearly matched that of the full-length protein Bir1, whereas the C-Bir1 mutant behaved like the empty vector. N-Bir1 was also able to partially complement the slow growth phenotype of the YPH499-B9 and YPH501-B9 mutants (data not shown).

View larger version (9K): [in this window] [in a new window]   Fig. 4.   BIR domains of BIR1 and survivin confer function. Wild-type YPH501 cells were transformed with plasmids containing a truncated form of Bir1 protein (N-Bir1) containing only the two BIR domains (residues 1-251) or a truncated form of survivin (N-survivin) containing the single BIR domain (residues 1-96). Cells were diluted to A600 nm ~ 0.2 and monitored for the indicated times. A, growth of wild-type YPH501 cells bearing N-Bir1 (, 111 min) versus the parent vector (, 146 min) alone. B, growth of wild-type cells bearing N-survivin (, 243 min) versus the parent vector (, 163 min). Data are represented as cell density (log A600 nm) versus time (hours). Each experiment was performed at least three times, with similar results. Doubling times were determined by calculating the slope of the linear portion of the curve. The values given represent the average of three independent experiments, each of which varied by less than 1 min.

BIR proteins are remarkably well conserved across species. To determine whether this conservation of sequence translates into conservation of function, we tested the ability of survivin (12) to function in place of BIR1 in yeast. Survivin was selected because it is the closest human homologue and the smallest member of the family (6, 7) and is therefore likely to be efficiently expressed in yeast. Wild-type YPH501 cells were transformed with full-length survivin, a C-terminally truncated form of the protein containing only the BIR domain (N-survivin), an N-terminally truncated form lacking the BIR domain (C-survivin), or the empty vector alone. As described in Fig. 4B, the empty vector and wild-type human survivin had no effect on the doubling time of these cells. However, N-survivin (which contains the BIR domain) exhibited a dramatic reduction in the rate of cell doubling, from 163 to 243 min. C-survivin had only a modest growth inhibitory effect (181 min). Thus, the survivin BIR domain appears to act in a dominant-negative manner to inhibit cell division.

Finally, we noticed that yeast expressing survivin exhibit an increased susceptibility to lysis in low osmotic solutions. When cells were transferred from synthetic medium to water, approximately 10% of the cells underwent spontaneous lysis. A similar "cell integrity" defect has been described for a variety of mutants and is thought to indicate a defect in cell wall synthesis (21, 23). The effect was exacerbated by growth at high temperatures and ameliorated by growth at low temperatures and in an osmotically balanced solution of 1 M sorbitol. As shown in Fig. 5, YPH501 cells expressing full-length survivin showed a marked decrease in viability on plates maintained at 34 °C as compared with vector alone. A less severe defect was exhibited by the N-survivin and C-survivin constructs. This defect was reversed by growth in 1 M sorbitol, by growth in dextrose (which represses survivin expression), or both. A cell integrity defect was not seen in any of the bir1 mutant strains described above or in cells overexpressing wild-type Bir1. Thus, the phenotype is unique to yeast that express survivin or the survivin BIR domain (N-survivin). It is also distinct from the slow growth phenotype described in Fig. 4, where full-length survivin had no effect. Regardless, either of the growth phenotypes could be useful for yeast-based genetic screens for IAP-domain mutants in survivin.

View larger version (92K): [in this window] [in a new window]   Fig. 5.   Survivin confers a cell integrity defect. Wild-type YPH501 cells were transformed with plasmids containing the galactose-inducible GAL1/10 promoter with full-length survivin, a truncated form of survivin containing the single BIR domain (N-Survivin, residues 1-96), or a truncated form of survivin containing all but the BIR domain (C-Survivin, residues 97-142). Cells were grown to saturation in liquid medium (SCD) and then serially diluted (1:10, 1:100, 1:1000, 1:10,000) and spotted onto medium containing dextrose (to repress transcription) or galactose (to induce transcription) in the presence or absence of 1 M sorbitol. Plates were maintained at 34 °C and photographed after 48 h.

The main conclusion that can be drawn from these studies is that inactivation of the IAP gene, BIR1, causes a pleiotropic cell division defect in yeast, with dysregulation of cellular ploidy, cell growth rate, and overall cell morphology. The identification of IAP homologous genes in yeast, which are not believed to implement an apoptosis program (14), has suggested the potential participation of these molecules in functions other than apoptosis control, particularly in the cell division control pathway. In support of this view, we found that BIR1 is essential for cell viability and that an increase or decrease in the BIR1 copy number will accelerate or decelerate cell division, respectively. This view is further reinforced by the cell cycle-regulated expression and localization to the mitotic apparatus of the mammalian IAP homologue, survivin (12), and by the multinucleation and cytokinesis defect resulting from targeting a single BIR molecule, BIR-1, in C. elegans (13).

Very recently, Uren et al. (22) described their efforts to disrupt BIR1 in yeast. Consistent with our findings, they observed inefficient sporulation and germination of both heterozygous and homozygous diploid mutants, with only a small percentage of viable spore products and an unusual bud morphology. In striking contrast to our findings, however, they conclude that homozygous diploid and haploid mutants are viable. They acknowledge that this could have resulted from "maldistribution of DNA" during cell division, but they did not test that possibility directly (e.g. by FACS). Nor did they attempt to confirm the genotype of their "diploid" mutants. Finally, and not surprisingly, they were also unable to demonstrate complementation by plasmid-borne BIR1. Thus, the rare viable spore products that they obtained are likely to have been polyploid, as documented here.

Although the molecular ordering of the cell division control pathway coordinated by BIR1 in yeast has not been elucidated, the pleiotropic nature of the defect in diploid mutants reveals a likely role in cytokinesis and control of cytoskeletal and cell wall integrity. This function was recapitulated by the isolated BIR domain, a potential Cys/His-based zinc finger fold (9), known to participate in apoptosis inhibition by certain mammalian IAP proteins (6, 7). Interestingly, heterologous expression of survivin (11) also resulted in dysregulation of cell growth rate, thus emphasizing the high degree of evolutionary and functional conservation among IAP proteins. Altogether, these data are consistent with a potential unifying model in which "ancient" IAP proteins (20), containing a single BIR module, are primarily involved in cell division control rather than regulation of apoptosis. More recently diverged IAP proteins may have further evolved a specialized function of suppression of apoptosis through a direct BIR-dependent interaction with terminal effector caspases (6, 7). Intriguingly, human survivin may provide an adaptation between these two evolutionarily conserved processes for its cell cycle-dependent expression in mitosis, localization to the mitotic apparatus (12), and caspase-inhibitory properties (6, 7). The use of powerful yeast genetic approaches should facilitate the molecular dissection and identification of additional modulators of IAP-dependent cell division control.

	FOOTNOTES

^* This work was supported by National Institutes of Health Grants GM55316 and GM59167 (to H. G. D.), HL54131 and CA78810 (to D. C. A.), CA16359 (to the Yale Cancer Center Flow Cytometry Shared Resource), and T32-NS07136 (to P. L. F.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

^§ These authors contributed equally to this work.

To whom correspondence should be addressed: Dept. of Pathology, Boyer Center for Molecular Medicine, Yale University School of Medicine, 295 Congress Ave., New Haven, CT 06536-0812. Tel.: 203-737-2869; Fax: 203-737-2402; E-mail: dario.altieri@yale.edu.

^** An Established Investigator of the American Heart Association.

	ABBREVIATIONS

The abbreviations used are: IAP, inhibitor of apoptosis; BIR, baculovirus IAP repeat; YPD, PCR, polymerase chain reaction; FACS, fluorescence-activated cell sorting; DIC, differential interference contrast.

	REFERENCES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS AND DISCUSSION REFERENCES