BRUCE, a Giant E2/E3 Ubiquitin Ligase and Inhibitor of Apoptosis Protein of the trans-Golgi Network, Is Required for Normal Placenta Development and Mouse Survival

Construction of the targeting vector.

Genomic 5′ sequence from the BRUCE locus of 129/SvJ mice was obtained by screening a λ phage library with cDNA fragments according to standard procedures. Three partially overlapping clones spanning exon 2 to 6 (λ5, λ7, and λ9) were subcloned into pBluescript SK(+) and analyzed by restriction mapping, Southern blotting, sequencing, and PCR. A 5-kb XbaI-EcoRI fragment from clone λ7 spanning parts of intron 3 and exon 4 was subcloned into pBluescript SK(+) and elongated by a 1-kb PCR fragment generated with primer jp71 (GCT GGA GAA GTA TTG CTT TGC) and mutagenic primer ko2XbaI*.la.asense (GGT CTA GAG GTA AGA CAA TCT GTC C) to introduce an XbaI site and a stop codon into exon 4. This construct was further elongated by a 2.2-kb PCR fragment generated with primers ko2XbaI*.la.sense (ACC TCT AGA CCA GTG CAC GTC CAG) and ko2SalItd.la.asense (GAC GTC GAC GAC AGC CTT ATG AGA CTA CCT GG) to introduce the XbaI site and the stop codon into exon 4 and a SalI site into intron 4. These three regions together represent the 5′ arm of homology. A neo^R-tk cassette containing the neomycin resistance gene under control of the herpes simplex virus promoter, the 1.1-kb 3′ homology arm derived from exon 6 and intronic sequences by PCR with primers ko2XbaItd.sa.asense (CCG TCT AGA CTG TAC TGT AAA TAC TTC CAT GC) and ko2XbaItd.sa.sense (CCG TCT AGA TTA TGA TGT TAG TCC TAT GTC CTA C), and the negative selectable marker HSV-TK (herpes simplex virus thymidine kinase) was then placed into the SalI site of the pBluescript SK(+) construct. Thus, in the resulting targeting vector, exon 4 contains an introduced stop codon, whereas exon 5 is replaced with a neo^R cassette.

Gene targeting in ES cells.

Fifty micrograms of linearized (NotI) plasmid DNA was electroporated into 10⁷ R1 embryonic stem (ES) cells suspended in phosphate-buffered saline (PBS). The ES cells were then cultured on dishes treated with gelatin. Following double selection with G418 (300 μg/ml) (Sigma) and ganciclovir (2 μM) (Sigma), 64 ES cell clones were picked, expanded on mitomycin-treated embryonic fibroblasts, and analyzed by Southern blotting of genomic PvuII-digested DNA. An external probe comprising a 0.4-kb BamHI fragment of a PCR-amplified λ7 sequence (using primers jp101 [GAG CCT GTG ATT TAT AGT GTA C] and jp136 [CTT CAG CTG GAT AGT GAG TTC C]) was used to identify the 2.9-kb fragment of the wild-type (WT) allele and the 2.3-kb fragment of the mutated allele. In addition, a genomic PCR was performed to screen for homologous recombination events. One primer was chosen to be outside the targeting vector, and the other was chosen to be within the neo^R cassette (outer.33, TCT TCT AAA TTA AGG CTT CAT AGT AGA TCT TGG; jp82, GCC TCT CCA CCC AAG CGG CCG GAG AAC CTG CGT GC). For higher sensitivity, a second round of PCR was performed to amplify a 1.5-kb genomic fragment (inner.22, GTT GTA CAC TAT AAA CAG G; jp83, GCA ATC CAT CTT GTT CAA TGG C).

Generation of BRUCE-deficient mice.

Two different ES cell clones with the targeted BRUCE allele were injected into 3.5-day-old C57BL/6 blastocysts and then reimplanted into 2.5-day pseudopregnant NMRI females. The spermatides of an infertile male chimera (germ line transmission was observed in the coat color of the F₁ offspring) were used for in vitro fertilization of C57BL/6 oocytes, and fertilized oocytes were reimplanted into pseudopregnant B6D2F2 females. Male chimeras were then backcrossed to C57BL/6 females, and heterozygous male and female mice were interbred to produce homozygous mutant offspring. The genotypes of the mice were assessed by Southern blotting and by PCR analysis with genomic DNA from tail biopsies of 21-day-old pups with primers wt-0 (GTG TCT CCA CCT AAC CTA TGC) and wt-3.as (GGT GAT AAA ATC CAG CTT GAG C) to detect a 556-bp band for the WT allele and with additional primers neo4 (GGC TAT TCG GCT ATG ACT GGG C) and neo5 (GGG TAG CCA ACG CTA TGT CCT G) to amplify a 624-bp band for the mutated allele.

Weight study and histological examination.

After cervical dislocation of pregnant mice, embryos and placentas at different developmental stages (embryonic day 9.5 [E9.5] to E19.5) were dissected out of uterus and decidua and cleared of membranes and umbilical cords in PBS. Yolk sac was used for genotyping. After excess buffer was blotted with a paper towel, embryos and placentas were weighed. For histological examination, dissected tissue was fixed in Bouin's or 10% neutral buffered formalin overnight and then dehydrated and embedded in paraffin. Serial sections (4 to 8 μm thick) were stained with hematoxylin and eosin by using routine procedures.

Protein analysis and Western blotting.

Immunoblot analysis of BRUCE protein was performed either on E10 whole-embryo lysates obtained by solubilization of the embryo in sodium dodecyl sulfate (SDS) sample buffer or on extraembryonic tissue lysates obtained by homogenization of the tissue in PBS (500 μl per 100 mg of tissue, containing complete [Roche] protease inhibitor cocktail) followed by solubilization in SDS sample buffer. For fibroblast extracts, cells were harvested and counted, and equal numbers of cells were solubilized in SDS sample buffer. Proteins were resolved by SDS-polyacrylamide gel electrophoresis (in a 4 to 20% polyacrylamide gradient) and electroblotted onto polyvinylidene difluoride membrane (Immobilon P; Millipore). Blots were probed with a monoclonal antibody against the N terminus of BRUCE (Transduction Laboratories) and a rabbit polyclonal (affinity-purified) antibody recognizing the C terminus of BRUCE (12). Equal loading of the sample was assessed by probing blots with a goat antiserum against actin (I-19; Santa Cruz). Bands were visualized with peroxidase-conjugated secondary antibodies and the ECL detection kit (Amersham). Other antibodies used for immunoblotting were anti-XIAP (Transduction Laboratories), anti-cIAP1 (PharMingen), anti-cIAP2 (R&D Systems), antisurvivin (A-19; Santa Cruz), and anti-caspase 3 (K-19; Santa Cruz).

BrdU labeling, immunohistochemistry, and TUNEL assay.

Pregnant mice from heterozygous matings were injected intraperitoneally at E11.5, E13.5, E15.5, and E17.5 with 5-bromo-2-deoxyuridine (BrdU) (Sigma) at a dose of 50 μg per g of body weight. Pregnant mice were sacrificed 4 h after injection, and the uteri were removed. Embryos and corresponding placentas were shock frozen on dry ice and cut on a cryostat in 10- to 15-μm-thick sections. Detection of the incorporated BrdU was performed with a BrdU in situ detection kit (BD PharMingen) according to the manufacturer's instructions, followed by microscopic examination. For immunohistochemistry, the placenta was dissected on E9.5 to E17.5, fixed in 10% neutral buffered formalin, paraffin embedded, and serially sectioned (8 μm thick). Immunohistochemistry on dewaxed and rehydrated sections was performed with the Vectastain Universal Quick kit in combination with the M.O.M. Basic kit (Vector Laboratories) according to the manufacturer's protocol. Affinity-purified antilaminin antibody (1:25) (Sigma) was used as the primary antibody. Immune complexes were visualized with diaminobenzidine tetrahydrochloride precipitates, and the sections were subsequently counterstained with nuclear fast red. For analysis of programmed cell death, terminal deoxynucleotidyltransferase-mediated dUTP-biotin nick end labeling (TUNEL) assays were performed with the ApopTag in situ apoptosis detection kit (Oncor) according to the manufacturer's instructions. To determine apoptotic cell death, in addition to TUNEL staining, caspase 3 activity was detected by using an antibody that specifically binds to the cleaved and thereby activated form of caspase 3 (anti-active caspase 3 pAb; Promega). Immunohistochemistry was performed after permeabilization of the cells with 0.1% Triton X-100 by using the Vectastain Universal Quick kit (Vector Laboratories).

In situ hybridization.

Hybridization of paraffin-embedded sections was performed essentially as described previously (6) ³³P-labeled antisense and sense mRNA probes were transcribed in vitro by using T7, T3, or SP6 RNA polymerase (Roche) from plasmids containing cDNA sequences of mPl-1, Tpbp (4311), Flt-1, Mash 2, Gcm1, eHand, Flk-1, and BRUCE. For hybridization, sections were dewaxed and pretreated as described previously (6). Briefly, slides were pretreated with proteinase K (20 μg/ml) for 7 min, postfixed in 4% paraformaldehyde in PBS, and acetylated in acetic acid anhydride diluted 1:400 in 0.1 M triethanolamine (pH 8.0). After washing, slides were dehydrated in ethanol and air dried. Hybridizations were performed in defined hybridization buffer (InnoGenex) supplemented with ≥3 × 10⁷ cpm of sense or antisense riboprobe per ml at 57.5°C overnight. Slides were washed by standard protocols, treated with RNase A (20 μg/ml) for 30 min at 37°C, washed again in SSC buffer (2×, 1×, and 0.5×; 1× SSC is 0.15 M NaCl plus 0.015 M sodium citrate), and at least at 65°C in 0.1× SSC containing 0.1 mM dithiothreitol. Slides were finally rinsed, dehydrated, air dried, coated with autoradiographic emulsion (LM-1; Amersham), developed after 2 to 6 weeks, and counterstained with cresyl violet.

Cell culture of primary embryonic fibroblasts.

Primary mouse embryonic fibroblasts (MEFs) (at least four independent lines per genotype) were isolated from day 13.5 embryos. Part of the embryo was used for genotyping. The remaining embryonic tissue was minced by using a pair of scissors, syringe, and needle (19 gauge) and immersed in culture medium (Dulbecco's modified Eagle medium supplemented with 10% fetal calf serum, 2 mM glutamine, 50 μg of penicillin per ml, and 50 μg of streptomycin per ml). Cells were then incubated at 37°C with 95% humidity and 7.5% CO₂. A single-cell suspension of MEFs (excluding the embryonic cell clumps) was plated onto a new dish before reaching confluency, and these cells were regarded as passage 1 cells. In the growth curve experiments, passage 2 cells (four lines per genotype) were plated at 1 × 10⁵, 3 × 10⁵, and 9 × 10⁵ cells per 6-cm-diameter culture dish and counted on successive days. In parallel the same MEF lines were immortalized as described previously (26). The number of cells in S phase in asynchronous cultures of fibroblasts was determined by using the BrdU in situ detection kit (BD PharMingen) according to the manufacturer's instructions, followed by microscopic examination of a randomly chosen area. For analysis of programmed cell death in asynchronous cultures of fibroblasts, TUNEL assays were performed with the ApopTag in situ apoptosis detection kit (Oncor) according to the manufacturer's instructions. Apoptosis was induced in primary MEFs by cultivating the cells for 0 to 36 h in growth medium containing different concentrations of various inducers of apoptosis (actinomycin D [Sigma], 0.01 to 2.5 μg/ml; anisomycin [Sigma], 0.05 to 2.5 μM; brefeldin A [Sigma], 0.1 to 20 μg/ml; calcimycin [Calbiochem], 0.05 to 5.0 μM; camptothecin [Calbiochem], 0.1 to 20 μg/ml, UV radiation, 10 to 500 J/m²; etoposide [Sigma], 0.5 to 100 μM; tunicamycin [Sigma], 0.1 to 10 μg/ml; H₂O₂ [Merck], 10 to 500 μM; tumor necrosis factor alpha [TNF-α] [R&D Systems], 1.0 ng/ml to 1.0 μg/ml [alone or in combination with 1.0 μg of cycloheximide]; cycloheximide [Sigma], 1.0 to 50 μg/ml; and hamster anti-mouse Fas [PharMingen], 1.0 to 100 ng/ml). Measurements of cell death were performed at different time points in a colorimetric assay with crystal violet (0.5% [wt/vol] in 40% methanol in PBS). Cell death was expressed as the optical density at 595 nm.

Targeted disruption of the mouse BRUCE gene.

Murine BRUCE was originally identified by a homology-based strategy with DNA sequences for yeast ubiquitin-conjugating enzymes as a probe (12). The corresponding cDNA was cloned by cDNA walking with λ phage-based libraries. The genomic locus of BRUCE spans 175 kb on chromosome 17 and includes a 14,535-bp open reading frame (12). In order to design a knockout construct, we characterized the genomic organization of BRUCE's 5′ region. To this end, we cloned genomic sequences corresponding to the 5′ region from a λ library containing liver DNA of the mouse line 129/SvJ and characterized two nonconsecutive genomic regions in more detail. The first 20-kb region (R1) encompassed exon 1, whereas the second region of 30 kb (R2) contained exon 2 to exon 6 (Fig. 1A). We noticed that the first half of BRUCE's BIR domain is encoded by exon 5. Exon 6 corresponds to the metal-binding fold of the BIR domain; however it does not encode it completely. Region R1, and to a lesser extent also R2, contained multiple copies of B1 and B2 elements of the SINE family and were intensely rich in GC content. Based on this information, we chose a knockout strategy that took away a region encompassing exon 4 to exon 6, thereby eliminating essential portions of the BIR domain. In addition, a stop codon was introduced in the open reading frame of exon 4, which is expected to result in a complete null allele. The targeting construct (Fig. 1A) contained a neomycin resistance gene (neo^R) flanked by sequences corresponding to segments containing exon 4 and exon 6, followed by a thymidine kinase cassette (TK).

Following transfection of R1 ES cells, we obtained correctly targeted ES clones at a frequency of 40.5%. Two clones were used to generate chimeras, but due to pronounced mortality among the littermates, only a single male chimera was available for crossing. Among the progeny, only one male mouse carried the knockout allele in the germ line. Because this animal was apparently infertile, progeny were obtained after in vitro fertilization of C57BL/6 oocytes and implantation into pseudopregnant B6D2F2 females. After crossing with C57BL/6 WT animals, the BRUCE alleles were transmitted in Mendelian ratio. BRUCE heterozygotes, identified by PCR (Fig. 1B) and Southern analysis using WT- and mutant-specific probes, were viable and fertile and exhibited no obvious phenotypic defects.

BRUCE deficiency results in retarded growth and perinatal lethality.

Phenotypically normal heterozygotes were intercrossed to produce BRUCE-deficient mice. We observed Mendelian inheritance up to the time of birth (Table 1). However, during or shortly after birth BRUCE null mice died, and no surviving knockout offspring could be obtained, demonstrating that the absence of BRUCE causes perinatal death. Tissue samples of knockout (BRUCE^−/−) embryos verified that no detectable BRUCE protein was expressed (Fig. 1B), thus confirming that the targeting strategy led to a complete BRUCE knockout.

Newborn BRUCE^−/− mice were significantly smaller than WT mice but otherwise appeared morphologically normal (Fig. 2A). To characterize the phenotype of mutant embryos, different stages of gestation were analyzed and the genotypes were determined by PCR analysis of genomic DNA obtained from the yolk sacs of each embryo (Fig. 1B). BRUCE mutants up to E12 were hardly distinguishable from their heterozygous and WT littermates. However, from about E13.5 onwards, BRUCE-deficient embryos showed increasing growth retardation, reflected by a slowed body weight gain (Fig. 2B). In addition, they appeared pale, and major blood vessels in mutant embryos and in the corresponding yolk sacs were sometimes not clearly visible (Fig. 3). This suggests that vascularization and/or blood supply was poor at this stage. Between E12 and E14, about 34% of the knockout embryos exhibited temporary bleeding in the brain (ventricle system) and neural tube, yet at later stages (E15.5 to birth), no remarkable morphological changes besides growth retardation were observed in 96% of the BRUCE mutants. However, 4% exhibited subcutaneous hemorrhages and generalized edema. Despite the growth retardation, we observed no consistent overt defects in vital organs and tissues that might result in embryonic lethality. Histological examination of major organs and tissues, such as heart, lung, liver, kidney, and neuronal tissue, of living BRUCE-deficient E18.5 embryos revealed no obvious abnormal architecture (data not shown).

BRUCE is required for normal placenta development.

To determine the basis for the perinatal lethality of BRUCE mutants, we examined the morphology and histology of the extraembryonic tissue from E9.5 to E17.5 of progeny. The mouse placenta consists of the maternal decidual tissue and a component of embryonic origin. In the mature placenta the embryonic part is formed by three distinctive trophoblast cell structures that arise by E10 and remain until the end of gestation (5, 18). The inner labyrinthine layer, containing the network of fetal capillaries and maternal blood sinuses, develops into a highly folded and branched tissue, building up the maternal-fetal interface for nutrient and gas exchange. The mature labyrinth is established by E14.5, and after E17 it remains unchanged until birth (15, 18). A compact layer of trophoblast cells between the labyrinth and the outer giant cells, which derives largely from cells of the ectoplacental cone, forms the spongiotrophoblast zone (22). The trophoblast giant cells result from trophectoderm cells by endoreduplication and lie at the periphery of the placenta towards the maternal decidual tissue.

In situ hybridization analysis with a radiolabled antisense probe revealed a relatively high expression level of BRUCE in the different layers of the WT placenta, and also in the visceral yolk sac, throughout the different developmental ages (Fig. 4). Within the labyrinth and the chorionic plate, BRUCE mRNA is detectable in endothelial cells as well, but predominantly in labyrinthine and chorionic trophoblast cells. Increased levels of BRUCE mRNA were found in the spongiotrophoblast layer, whereas in trophoblast giant cells BRUCE expression was reduced. Notably, BRUCE expression is particularly abundant in those tissues that exhibit defects in the targeted mutants.

In comparison to placentas of WT and BRUCE heterozygous embryos, the homozygous mutant placenta appeared smaller and was frequently paler in color (Fig. 3 and 5A), suggesting a defect in vascularization and/or poor blood supply. The average weight of placenta of BRUCE-deficient embryos was 73% of that of WT embryos at E11.5, and this ratio further decreased to 65% until birth (Fig. 2B). The formation of the chorioallantoic connection appeared normal, and, except for a slight reduction of the spongiotrophoblast layer, no significant gross difference between WT and mutant placentas was detectable at early stages (Fig. 5B and C, E9.5 to E11.5). All structures, including the labyrinth as well as the trophoblast giant cells, appeared fairly normal, and the border of the spongiotrophoblast and the labyrinth layer was clearly noticeable. Also, infiltration of the labyrinth layer with embryonic blood vessels and maternal blood sinusoids occurred normally at this point.

However, at later developmental stages (E13.5 to E17.5), the reduction in thickness of the spongiotrophoblast became more evident. This observation could also be confirmed by spongiotrophoblast marker analysis using in situ hybridization (Fig. 6A and http://www.biochem.mpg.de/jentsch/JentschDataLotz.html). The fetal fms-like tyrosine kinase Flt-1 or the spongiotrophoblast-specific marker Tpbp/4311 (17) can be used to monitor differentiation of ectoplacental cone cells into spongiotrophoblast cells from E9.5 onwards (13). In mutant placenta, the layer formed by Flt-1- and 4311-expressing cells was clearly reduced at all embryonic stages examined. By analyzing the expression profiles of the specific trophoblast giant cell marker Pl-1 (4) and proliferin (22), we observed that alongside the normal continuous giant cell layer, additional Pl-1- and proliferin-expressing cells can be detected in BRUCE-deficient placenta from E11.5 onwards. Initially, those cells are mainly spread over the spongiotrophoblast, but they also specifically appear within the labyrinth layer after E15.5. Although these cells seem to be slightly enlarged, they do not reach the typical giant cell shape.

Additional markers were used to identify subsets of differentiated trophoblast and endothelial cells, such as Mash2 (10), Gcm1 (23), eHand (24), or Flk-1 (8). All cells were present in WT as well as in mutant placenta and were distributed in a similar manner throughout the placenta (Fig. 6A and http://www.biochem.mpg.de/jentsch/JentschDataLotz.html). In addition, no dysregulation in expression of the basic helix-loop-helix transcription factors (Mash2 and eHand), which are key regulators of spongiotrophoblast maintenance and giant cell differentiation (5), could be detected. However, correct organization of the labyrinth layer was apparently delayed in the mutant, resulting in a less dense structure, and the number of trophoblast cells within this layer seemed to be reduced. This phenotype became pronounced after E13.5, when a dilatation of fetal capillaries and maternal sinusoids could be detected, in contrast to WT tissue (Fig. 5C).

The development of the placental vascular network was further investigated immunohistochemically by using an antibody to laminin, which stains the basal lamina of fetal capillaries in the labyrinth layer. In mature WT placenta the fetal capillaries were elongated, of similar size, and evenly distributed throughout the labyrinth from the chorionic plate to the spongiotrophoblast layer. In contrast, the fetal blood vessels in mutant placenta showed less branching. They were highly variable in thickness and formed a less dense vasculature than in WT placenta (Fig. 6B). In addition, the elimination of the nuclei from red blood cells seemed to be delayed in BRUCE mutant placenta. In fact, in WT placenta, all red blood cells are nucleated before E11.5. However, between E12 and E15 these cells enucleate, and after E16 all red blood cells in the embryonic circulation are nonnucleated (15). In BRUCE mutants about 50% of cells in fetal blood vessels of E15.5 placenta were still nucleated, whereas in the WT almost all red blood cells had already lost their nuclei. Even at stage E17.5, single nucleated red blood cells were visible in the mutant labyrinth (Fig. 5C).

Apoptosis and proliferation.

Recently, it was shown that BRUCE inhibits apoptosis, that it possesses IAP activity, and that overexpression of the protein blocks apoptosis (2, 3, 20). We therefore considered the possibility that the observed abnormalities of BRUCE-deficient mice are caused by an increase of apoptosis events. To address this question, the rate of apoptosis in WT and BRUCE mutant placenta, specifically within the spongiotrophoblast and the labyrinth layer, was analyzed. Yet, by using a TUNEL assay and immunohistochemistry to monitor activated caspase 3 levels, no significant differences between BRUCE-deficient and control placentas were found (Fig. 7A and B). Apoptosis could be observed at equally low rates in WT and mutant maternal decidual tissue, yolk sac, chorionic plate, spongiotrophoblasts, and labyrinth layer. Other IAP family members, which might overlap with BRUCE function, are expressed at equal levels in WT and knockout placenta (Fig. 7C). In contrast, we detected significant differences in cell proliferation between WT and BRUCE-deficient cells as monitored by BrdU incorporation. Analysis of E11.5 to E17.5 placenta indicated a changed BrdU index of BRUCE-deficient trophoblasts in comparison to WT littermate controls (Fig. 8A and B). At early stages, the number of proliferating cells was significantly reduced within the mutant labyrinth and almost no proliferating cells were visible within the spongiotrophoblast layer, whereas in the corresponding WT layers high proliferation rates could be observed. The proliferating activity in the labyrinth of knockout placenta was reduced to 60% at E11.5 and to 55% at developmental stage E13.5. At later stages (E15.5 and E17.5), when cells in the mature WT placenta already ceased proliferating, BrdU-positive cells were still detectable in the mutant labyrinth (Fig. 8B). This effect seems to be predominantly restricted to trophoblast cells, since the numbers of proliferating endothelial cells within the chorionic plates of WT and knockout placenta were similar (Fig. 8C). However, intensive trophoblast proliferation and branching are a prerequisite for vascularization of the labyrinth (22), and thus the number of proliferating endothelial and trophoblast cells within the labyrinth layer seems to be reduced to the same extent. From these findings we conclude that the abnormalities observed in BRUCE mutant placenta are perhaps caused by an altered proliferation and/or differentiation pattern of trophoblast cells and are not due to enhanced apoptosis in this tissue.

Analysis of embryonic fibroblasts cultured in vitro.

We further investigated possible defects in proliferation and apoptosis in primary cell culture. We established primary MEFs prepared from E13.5 WT, heterozygous, and BRUCE knockout fetuses. The morphology and size of the mutant MEFs were indistinguishable from those of WT and heterozygous cells at the beginning of the culture period, but the number of cells obtained from knockout fibroblast cultures was significantly lower. Comparing growth behavior of fibroblasts at different suboptimal high and low and at optimal growth densities, we observed that knockout cells grow slower than WT and heterozygous cells and, in addition, that mutant MEFs reached only ∼70% of the normal culture density of WT cells (Fig. 9A).

To investigate the consequence of BRUCE deficiency for apoptosis and proliferation of fibroblasts, the number of cells in S phase (determined by BrdU labeling) and the number of apoptotic cells (determined by TUNEL staining) in continuously growing cultures of WT and knockout fibroblasts were determined (Fig. 9B). The numbers of S-phase cells in WT and mutant cultures were indistinguishable. Fluorescence-activated cell sorting analysis of asynchronous cells also revealed similar distributions of G₁, S, and G₂ cell cycle phases in knockout and WT cultures (not shown). The numbers of apoptotic cells in WT and mutant cultures were similar as well (Fig. 9B). In addition, primary BRUCE-deficient fibroblasts were not more susceptible than WT cells to apoptosis induced by different apoptotic stimuli, such as actinomycin D, oxidative stress, or TNF-α (Fig. 9C) or Fas, brefeldin A, or camptothecin (not shown). Next, we asked whether other IAP family members are able to compensate for the lack of BRUCE, as, for example, reported for the knockout of XIAP (11). To address this question, we investigated the levels of XIAP, c-IAP2, and survivin in WT or BRUCE-deficient MEFs that were either untreated or treated with TNF-α. No increase of IAP levels could be detected in mutant cells compared to the WT control (Fig. 9D). Moreover, caspase 3 levels also were similar in WT and knockout fibroblasts after induction of apoptosis.