Id2 Promotes Apoptosis by a Novel Mechanism Independent of Dimerization to Basic Helix-Loop-Helix Factors

Abstract

Members of the helix-loop-helix (HLH) family of Id proteins have demonstrated roles in the regulation of differentiation and cell proliferation. Id proteins inhibit differentiation by HLH-mediated heterodimerization with basic HLH transcription factors. This blocks their sequence-specific binding to DNA and activation of target genes that are often expressed in a tissue-specific manner. Id proteins can also act as positive regulators of cell proliferation. The different mechanisms proposed for Id-mediated promotion of entry into S phase also involve HLH-mediated interactions affecting regulators of the G₁/S transition. We have found that Id2 augments apoptosis in both interleukin-3 (IL-3)-dependent 32D.3 myeloid progenitors and U2OS osteosarcoma cells. We could not detect a similar activity for Id3. In contrast to the effects of Id2 on differentiation and cell proliferation, Id2-mediated apoptosis is independent of HLH-mediated dimerization. The ability of Id2 to promote cell death resides in its N-terminal region and is associated with the enhanced expression of a known component of the programmed cell death pathway, the proapoptotic gene BAX.

The signals regulating apoptosis during the development of individual tissues are poorly understood, but a number of experimental systems have provided insights into the molecular mechanisms of several apoptotic pathways and the molecules which mediate them. Observations demonstrating that inappropriate entry into S phase is frequently associated with programmed cell death have been interpreted as indicating that these two cellular activities are coordinately regulated (3, 22, 43, 51, 73). Several genes have products which regulate cell cycle progression and also modulate apoptosis. Among these are the tumor suppressor proteins E2F-1, pRb (retinoblastoma protein), and p53 and oncoproteins such as E6, c-Jun, and c-Myc (3, 7, 22, 27, 31, 51, 56, 61, 67, 75). Evidence that the G₁ checkpoint regulator p53 can cooperate with E2F to augment apoptosis and that c-Myc- and E2F-1-induced apoptosis occurs when cells are deprived of mitogenic factors (3, 22, 30) supports a model in which activation of the cell death pathway is due to a conflict between growth-promoting and growth-inhibitory signals (31, 43, 51, 75). According to this view, the tendency of cells to undergo apoptosis when receiving divergent growth signals may result from a dysregulated expression of cell cycle modulators and the unscheduled activation of their target genes.

Recent observations, however, have challenged this model by demonstrating that dual regulators of proliferation and apoptosis promote these processes by separate mechanisms (73). Active G₁ cyclin-Cdk complexes are necessary for proliferation but not apoptosis mediated by c-Myc. Also, c-Myc-induced cell cycle progression and apoptosis can be separated pharmacologically (55, 65). More recently, it has been demonstrated that the ability of E2F-1 to transactivate genes required for progression into S phase can be uncoupled from its ability to activate apoptosis (9, 33, 59). Similar observations have been reported for the p53 tumor suppressor, whose abilities to induce growth arrest and apoptosis are genetically separable (9). These findings support the notion that while several regulators of the cell death and growth pathways are shared, their functional activities in these pathways are distinct. How these divergent activities are coordinated remains unknown.

Id proteins constitute a family of helix-loop-helix (HLH) transcription factors that are important regulators of cellular differentiation and proliferation (5, 6, 10, 21, 62, 70, 77). Id proteins lack a basic DNA binding region and are capable of inhibiting gene expression by forming inactive heterodimers with basic HLH (bHLH) transcription factors, thereby blocking the binding of bHLH factors to specific DNA sequences, E boxes, found in the regulatory regions of their target genes (10, 40, 48, 52, 57, 71). During cellular proliferation and prior to the onset of differentiation, Id genes are typically expressed at high levels (5, 19, 35, 36, 42, 71), and enforced overexpression of Id genes inhibits differentiation of a variety of cell lineages (14, 18, 38, 42, 47, 50, 70). In addition, constitutive expression of Id1 in B cells of transgenic mice inhibits B-cell maturation during the establishment of the immune system (70).

Id gene expression is enhanced in response to mitogenic stimuli (4, 5, 10, 16) and has been implicated in the induction of DNA synthesis (58). Ectopic expression of Id1 or Id2 enhances proliferation of several cell types (34, 60). Furthermore, treatment of cells with antisense Id1 oligonucleotides or microinjection of Id1 neutralizing antibodies prevents reentry of serum-deprived cells into the cell cycle (4, 28). Some aspects of the mechanisms by which Id proteins enhance proliferation have been described: Id2 binds to the unphosphorylated pRb family members and abolishes their growth-suppressing function (34, 45), and Id1a inhibits expression of p21/WAF1, an inhibitor of cyclin-dependent kinases (60). Moreover, Id2 can antagonize the growth inhibition of several tumor suppressor proteins whose action is mediated by pRb, including p21 and p16 (45). Other components of the cell cycle machinery interact with Id2 as well. Recently, Id2 was shown to be an in vivo substrate of the G₁ cyclin-dependent kinases, cyclin A/Cdk2 and cyclin E/Cdk2 (29). The only other substrates identified for these kinases are p27 (68), the Rb family of tumor suppressors, pRb, p107, and p130, and certain E2F family members, all of which are important for the G₀-to-S-phase transition (for a review, see reference 74).

Here we report that overexpression of Id2 augments apoptosis of 32D.3 cells, an interleukin-3 (IL-3)-dependent myeloid progenitor cell line, and U2OS cells, an osteogenic sarcoma-derived cell line. To date, all Id protein functions have been ascribed to the HLH sequence motif, which mediates heterodimerization with bHLH transcription factors (5) and the association of Id2 with pRb family members (34, 45). Strikingly, Id2 transmits an apoptotic death signal that is independent of its ability to dimerize with bHLH members. Since the Id2 HLH domain is required for interactions with the pRb family (34, 45), it is likely that the cell death- and growth-promoting activities of Id2 are separable. We found that the N-terminal region of Id2 is required for its apoptotic functions and that this activity is associated with the expression of a known modulator of the programmed cell death cascade, Bax, suggesting a role for Id2 in integrating the divergent cellular activities of growth and apoptosis.

MATERIALS AND METHODS

cDNA constructs and establishment of cell lines.

The mouse 32D.3 myeloid cell line was maintained in RPMI 1640 medium (GIBCO) supplemented with 10% fetal bovine serum (FBS) and 20 U of IL-3 per ml as previously described (3, 12). The human Id1, Id2, and Id3 cDNAs (5, 6, 11) were introduced into the pMAM-Neo expression plasmid (Clontech, Palo Alto, Calif.) at the XhoI site. Mutations in Id2 and Id3 were generated by PCR. To prepare an N-terminus deletion mutant, a 0.3-kb SalI-XhoI fragment lacking the region encoding amino acids 1 to 34 was inserted into the pMAM vector. A C-terminally truncated Id2 was prepared by removing a SmaI fragment containing cDNA corresponding to amino acids 93 to 135. Id2ΔHLH was prepared as previously described (34) and subcloned into the pMAM vector at the XhoI site. The primary structures of all mutants were confirmed by DNA sequencing. 32D.3 cells were electroporated as previously described with 10 μg of linearized plasmid DNA (12). Transfected cells were selected in medium containing G418 (0.4 mg/ml), and single-cell clones, as well as pools of transfectants, were isolated by limiting dilution. To prepare additional cultures, drug-resistant clones and pools of clones were prepared by transfecting cells with pMAMNeo DNA alone and pMAMNeo containing Id2 cDNAs in the antisense orientation. 32D.3 cells were grown in suspension and maintained at equivalent exponential growth phases by seeding at 10⁵ cells per ml of culture medium. All subsequent analyses were performed with cells harvested at 0.5 × 10⁶ to 1 × 10⁶ cells per ml.

To prepare a recombinant expression construct encoding Id2 which could be regulated by the addition of tetracycline to the culture media, the KpnI/HindIII filled-in fragment of Id2 cDNA was inserted into the pUHD10-3 vector (26) at EcoRI blunt-ended sites to prepare a tetracycline-regulatable transactivator (tTA)-responsive Id2 expression vector. DNA from pUHD-10-3Id2 was used in addition to DNA from pTet-TAk (GIBCO), encoding the tTA protein. Stable transfectants were prepared in the U2OS osteosarcoma cell line by performing a three-vector transfection using LipofectAmine (GIBCO) at a DNA ratio of 5:5:1 between pTet-Id2, ptet-TAk, and pcDNA-3, respectively. Transfectants were selected in the presence of Dulbecco modified Eagle medium (DMEM)–10% FBS–G418 (0.6 mg/ml)–tetracycline (5 μg/ml) after 2 days. A parallel transfection in which pUHD10-3Id2 DNA was replaced by a comparable pUHD10-3-derived recombinant plasmid containing a cDNA encoding luciferase (Tet-luciferase) was used to prepare control cell lines. To show expression of Id2, cells were trypsinized and plated with various concentrations of tetracycline (0 to 1,000 ng/ml).

RNA preparation and Northern blot analysis.

Total RNA was extracted from exponentially growing 32D.3 myeloid progenitor cells or from these cells deprived of IL-3 and analyzed by Northern blot hybridization. Aliquots of 20 μg of total RNA were fractionated by electrophoresis in 1% agarose-formaldehyde gels, blotted, and hybridized to full-length cDNAs of mouse Id1a, Id2, Id3, and Id4 and a 192-bp fragment of Id1b cDNA that uniquely recognizes Id1b (a region corresponding to amino acids 135 to 168 and 3′ untranslated sequences). The cDNAs were radiolabeled with [³²P]dCTP by random priming (Amersham Corporation). Membranes were washed as previously described (12) and exposed for autoradiography. Northern blot analysis to look for changes in BAX expression was carried out in the same manner, by using a full-length cDNA of mouse BAX.

Expression of Id genes and immunoblot analysis.

Id gene expression was induced by addition of 1 μM dexamethasone (Dex; Sigma) for 16 h. Cells were then collected at the indicated times by centrifugation and washed in phosphate-buffered saline (PBS), and whole-cell lysates were prepared by resuspending the pellets in radioimmunoprecipitation assay buffer (150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 1% sodium deoxycholate, 20 mM Tris-HCl [pH 7.2]) containing protease inhibitors (1 μM phenylmethylsulfonyl fluoride, 0.01 μM benzamidine HCl, 1 μg of o-phenanthroline per ml, 0.5 μg each of antipain, leupeptin, pepstatin A, and aprotinin per ml). The lysates were then clarified by centrifugation and examined by sodium dodecyl sulfate–12.5% polyacrylamide gel electrophoresis (SDS-PAGE). Western blot analysis was carried out to detect expression of Id proteins, by using the rabbit polyclonal antibody sc-489 against Id2 (Santa Cruz Biotechnology, Santa Cruz, Calif.) and rabbit polyclonal antibodies generated against bacterially expressed Id1 or Id3 (30a). The expression of Bax proteins was monitored by using a rabbit anti-Bax polyclonal antibody (sc-930; Santa Cruz Biotechnology). Briefly, blots were washed with blocking buffer (4% milk powder and 0.2% Tween 20 in PBS), incubated with anti-Id2 or -Bax antibody (used at a 1:1,000 dilution in blocking buffer), washed, and incubated with peroxidase-conjugated rabbit immunoglobulin G. The blots were washed again and developed by use of the DuPont enhanced chemiluminescence system.

Viability and apoptosis assays.

To investigate the effects of Id gene expression on viability, cells growing exponentially in the presence of IL-3 (20 U/ml) were treated with Dex for 16 h. The cells were collected by centrifugation at 200 × g for 10 min, washed twice in PBS, and resuspended in RPMI 1640 medium with 10% FBS without IL-3. The number of viable cells was determined at various times after IL-3 withdrawal by trypan blue dye exclusion. After cytospin, DNA fragmentation was visualized following incorporation of digoxigenin-labeled oligonucleotides (mediated by terminal deoxynucleotidyltransferase [66]), by using the Apotag TUNEL (terminal deoxynucleotidyltransferase-mediated dUTP nick end labeling) assay kit (Oncor).

To determine the extent of apoptosis in U2OS cells transfected with pTet-Id2 DNA, cells were plated at a density of 5 × 10⁵/100-mm plate and allowed to adhere for 24 h in the presence of tetracycline (1,000 ng/ml). Then the medium was changed to contain tetracycline at either 0 or 1,000 ng/ml after two washes with PBS. The cells were maintained in the culture for 4 days before harvesting by trypsinization (floating cells were also harvested). These cells were washed in PBS, fixed with 70% ethanol at 4°C for at least 1 h, washed with PBS, and stained with PBS-RNase (50 μg/ml)-propidium iodide (25 μg/ml). DNA content was determined by flow cytometry on a FACSort (53).

RESULTS

The levels of Id1a, Id1b, and Id2 in 32D.3 myeloid cells are differentially regulated by IL-3.

To assess the potential role of Id family proteins in myeloid cell apoptosis, we examined Id gene expression in 32D.3 myeloid progenitor cells before and after the removal of IL-3, a cytokine required for the survival of these cells in culture (3, 63). Northern blot analysis revealed a large decrease in the level of Id1a mRNA after 8 h of IL-3 withdrawal, whereas there was a detectable increase in the level of Id2 transcripts (Fig. 1A). Id3 and Id4 expression was not detected in these cells, although Id1b, which is expressed in 32D.3 cells, did not change over the course of these experiments (Fig. 1A). To ensure equal loading, we examined the levels of a control Nedd2 mRNA, a caspase which is not regulated at the mRNA level during IL-3 withdrawal (11). Id1a and Id2 protein levels were analyzed by Western blot analysis at these same time points after IL-3 withdrawal, and we observed changes that paralleled those occurring at the RNA level (Fig. 1B). The observation that Id1a and Id2 mRNA and protein levels are regulated during survival factor withdrawal raised the possibility that Id1a and Id2 may be regulators of the apoptotic response of 32D.3 cells to IL-3 deprivation.

FIG. 1.

Expression of Id2 and Id1a is regulated by IL-3 in 32D.3 myeloid cells. (A) The endogenous levels of Id1a, Id1b, Id2, Id3, and Id4 mRNAs in 32D.3 grown in the presence of 10% FBS and IL-3 were analyzed by Northern blot analysis at 0, 8, and 16 h after IL-3 withdrawal as indicated. As a control for loading, the levels of Nedd2 RNA were also determined. The results shown are representative of three independent experiments. (B) Immunoblot analysis of Id1a and Id2 protein levels in 32D.3 cells growing in IL-3 (0 time point) and after deprivation of IL-3 for 10 or 21 h.

Id2 enhances apoptosis in 32D.3 myeloid cells following IL-3 withdrawal.

Since Id2 levels in 32D.3 cells increased in association with IL-3 withdrawal, we examined the effect of increasing Id2 expression on the apoptotic response of these cells. 32D.3 myeloid progenitor cells were transfected with the Dex-inducible pMAMNeo expression vector containing a full-length Id2 cDNA. The expression of Id2 in clones and pools of transfectants was determined in immunoblots, and we observed a three- to fourfold increase in Id2 protein levels upon treatment of the cells with Dex for 16 h compared to cells transfected with vector DNA or with a pMAMNeo recombinant plasmid with Id2 in the antisense orientation (Fig. 2A). When examined for the extent of apoptosis by the TUNEL assay following IL-3 withdrawal (25), both Id2-transfected clones and pools of clones exhibited significantly elevated apoptosis relative to what was observed in vector-only or Id2 antisense transfectants (Fig. 2B). To confirm that these changes were due to increased apoptosis, we examined these same cell lines by flow cytometry following IL-3 removal and observed that the subdiploid population, indicative of apoptotic DNA fragmentation, was significantly greater in each of the cell lines expressing high levels of Id2 than in control cells (data not shown).

FIG. 2.

Id2 augments apoptosis in 32D.3 cells following growth factor withdrawal. (A) Immunoblot analysis of 32D.3 myeloid progenitors stably transfected with the pMAMNeo expression plasmid vector, an antisense Id2 pMAMNeo plasmid, or an Id2 pMAMNeo plasmid. Individual clones and pools of transfectants were maintained in RPMI 1640 medium containing G418 (0.4 mg/ml), 10% FBS, and IL-3 (20 U/ml). The expression of Id2 in these transfectants was determined after induction with Dex for 16 h. (B) Transfectants used for panel A were treated with Dex for 16 h and then deprived of IL-3. The number of apoptotic cells in individual transfectants and pools of transfectants was determined by the TUNEL assay (Oncor), 16 h after the withdrawal of IL-3. In each experiment, 300 cells were counted in duplicate specimens. The data shown are averages of three experiments. The brackets depict the standard deviation from the mean. (C) The viability of these cell cultures at several time points after IL-3 deprivation was determined. Cell cultures of antisense, vector-alone, and full-length Id2 transfectants were examined, and the percentage of live cells was determined by the trypan blue exclusion method using a hemocytometer. At each time point, 300 cells were counted. These values are from a single experiment done in triplicate and are representative of five independent experiments of this sort. (D) Immunoblot analysis of Id2 protein levels in 32D.3 cells grown in the presence of IL-3, deprived of IL-3 for 25 or 43 h, and in cells transfected with the Id2 expression plasmid (Id2 clone 2).

To extend our assessment of the effects of Id2 gene expression on the rates of myeloid cell death, we examined these transfectants by trypan blue dye exclusion for evidence of cell death when they were grown under optimal conditions and at various times following IL-3 withdrawal. Constitutive expression of Id2 had only minor effects on viability in the presence of IL-3, during the 16 h of treatment with Dex. Following the withdrawal of IL-3, however, we observed dramatically accelerated rates of death relative to control cells (Fig. 2C). These rates were comparable to those documented for IL-3-deprived 32D.3 cells ectopically expressing E2F-1 and c-myc, both of which are potent inducers of apoptosis (3, 31). The morphology of cells taking up dye in both the control and transfected cultures was characterized by a shrunken cytoplasm, condensed chromatin, and the presence of apoptotic bodies, indicating that all of these cells had died an apoptotic cell death (data not shown). Levels of exogenous Id2 in these transfectants were comparable to those seen in cells deprived of IL-3 (Fig. 2D). Therefore, enforced expression of Id2 exacerbates the rate of apoptotic death of 32D.3 myeloid cells following IL-3 withdrawal. We then examined more closely whether enforced Id2 expression could compromise 32D.3 cell survival in the presence of IL-3. The various cellular transfectants used for Fig. 2 were examined after 3 days of induction of Id expression with Dex in the presence of IL-3. Approximately 10 to 15% of the cells in each of the Id2 cultures exhibited nuclear fragmentation characteristic of apoptotic cells (data not shown). These data indicate that IL-3 is able to effectively antagonize the proapoptotic effects of Id2.

Id2-induced apoptosis of 32D.3 cells and U2OS osteosarcoma cells is concentration dependent.

To examine more closely the association between enhanced Id2 expression and the increased rate of cell death upon IL-3 withdrawal from 32D.3 cells, we obtained transfected clones of 32D.3 cells expressing various levels of Id2 protein after treatment with Dex (clones 2 to 4 and a pool of Id2 transfectants) (Fig. 3A, inset). These cells were subjected to IL-3 deprivation and analyzed for cell death by the TUNEL assay 16 h after IL-3 withdrawal (Fig. 3A). The Id2 pool of transfectants and clone 3 exhibited the highest rate of cell death and also expressed the highest levels of Id2 protein. Lower levels of Id protein were correlated with less apoptosis. Interestingly, even a very small increase in the intracellular concentration of Id2 increased the extent of apoptosis compared with vector-transfected cells. Using Pearson’s correlation coefficient, we related the percentage of apoptotic cells to Id2 levels of expression in the different transfectants determined by densitometry and obtained an r value of 0.98, indicating a close correlation between Id2 expression and the degree of apoptosis. These results suggest that Id2-induced apoptosis is concentration dependent.

FIG. 3.

Id2 promotes apoptosis in 32D.3 cells and U2OS cells in a dose-dependent manner. (A) Id2 expression in different clones and a pool of 32D.3 cells transfected with Id2 cDNA was determined by immunoblot analysis and is shown in the inset. The relative levels of Id2 protein in lysates prepared from the different transfectants were determined by densitometry. The extent of apoptosis was measured by the TUNEL assay and was determined 16 h after IL-3 withdrawal. Pearson’s correlation coefficient of these two variables is shown (r = 0.98). (B) U2OS osteosarcoma cells transfected with pTet-Id2 or Tet-luciferase were maintained in DMEM containing 10% FBS, G418 (0.4 mg/ml), and tetracycline (0.005 mg/ml). To induce expression of Id2, the cells were trypsinized and plated at various concentrations of tetracycline (0 to 1,000 ng/ml of culture medium). After 12 h of Id2 induction, the expression of Id2 in lysates of clone 9 was determined by SDS-PAGE and immunoblot analysis using an anti-Id2 antibody (shown in inset). To determine the extent of apoptosis in cultures in which Id2 expression (clone 9) or luciferase (upper inset) was regulated by tetracycline, we incubated such cells in the presence (5 μg/ml) or absence of tetracycline (after 12 h of removal). The cells were harvested by trypsinization and stained with propidium iodide (1 μg/ml of PBS). DNA content was determined by flow cytometry on a FACSort. The shaded regions represent the DNA profiles of cells grown in the presence of tetracycline, and the blank regions represent cells grown in the absence of tetracycline. The bracketed region labeled Sd indicates cells with a subdiploid DNA content. The percentage of cells with a subdiploid DNA content is indicated in parentheses. The experiment shown is representative of two independent experiments.

To further examine the effects of Id2 on apoptosis in a different cell type, we used the osteosarcoma cell line U2OS. We prepared stable transfectants in which the level of Id2 expressed from a recombinant expression vector could be regulated by tetracycline. Transfected cultures of Tet-luciferase were also examined as controls. We examined different clones of transfected cells for Id2 expression following treatment with various concentrations of tetracycline (0 to 1,000 ng/ml) in the culture medium (Fig. 3B, lower inset, and data not shown). To determine whether cells underwent apoptosis when Id2 levels were raised, we measured the DNA content of the cells 4 days after tetracycline (1,000 ng/ml) was removed from the medium. The results in Fig. 3B demonstrate a large proportion of apoptotic cells with subdiploid DNA content in cultures of U2OS clone 9 cells following removal of tetracycline and the consequent induction of Id2 expression. This effect was not observed in control cells transfected with Tet-luciferase and grown in the presence or absence of tetracycline (Fig. 3B, upper inset). Therefore, the proapoptotic activity of Id2 is observed in other cell types and appears to be dose dependent (Fig. 3A). Interestingly, a low level of exogenous Id2 seems to be well tolerated by U2OS cells (34), but expression at higher levels, such as those observed in this experiment, leads to apoptosis. Furthermore, the observation that Id2 overexpression can promote apoptosis in U2OS cells in the presence of serum suggests that high levels of Id2 can override the protective effects of survival factors utilized by these tumor cells.

Different members of the Id gene family have distinct apoptotic activities in 32D.3 myeloid cells.

Having observed that Id2 promotes cell death, we examined whether other members of the Id gene family had similar activities. We prepared stable transfectants of 32D.3 cells expressing Id1a and Id3 in pMAMNeo. In the presence of Dex, both Id1a and Id3 proteins were elevated to levels three to four fold higher than those observed in comparable cells transfected with vector DNA and selected in neomycin (Fig. 4A). Apoptosis and the rate of cell death in these transfectants were examined as before (Fig. 4B and data not shown). While Id1a had a detectable activity in augmenting the rate of cell death, expression of Id3 was without effect (Fig. 4B). Therefore, in 32D.3 myeloid cells, Id proteins are not functionally equivalent. Furthermore, although Id1a overexpression can augment apoptosis, its decreased expression following IL-3 withdrawal (Fig. 1) suggests that it may not play a physiologic role in the programmed cell death of this particular cell type. In contrast, Id2 expression is increased following IL-3 withdrawal, an observation which is consistent with its role in promoting apoptosis.

FIG. 4.

Id1a and Id2, but not Id3, augment apoptosis of 32D.3 cells following IL-3 withdrawal. (A) 32D.3 myeloid progenitors were stably transfected with DNA from pMAMNeo plasmid vector, an antisense Id2 pMAMNeo plasmid, an Id2 pMAMNeo plasmid, an Id1a pMAMNeo plasmid, or an Id3 pMAMNeo plasmid. Individual clones and pools of transfectants were maintained in RPMI 1640 medium containing G418 (0.4 mg/ml), 10% FBS, and IL-3 (20 U/ml). The expression of Id1a, Id2, and Id3 in these transfectants was determined by immunoblot analysis after induction with Dex for 16 h. (B) A comparison of the roles of different Id genes during programmed cell death of 32D.3 myeloid progenitors was carried out by measuring the rate of cell death in pools of cells stably transfected with Id1a, Id2, and Id3. The individual cultures were treated with Dex for 16 h to induce Id expression. An assay of cell viability was carried out as described above, and the percentage of viable cells was determined over time after IL-3 removal. The values plotted are from one experiment carried out in triplicate and are representative of three different experiments.

HLH-mediated dimerization of Id2 is not required for Id2-enhanced apoptosis.

Id genes heterodimerize with bHLH transcription factors blocking the formation of bHLH protein dimers required for DNA binding and transcriptional activation (40, 48). Since the HLH domains through which dimerization is mediated are highly conserved in all known members of the Id gene family, our finding that Id1a and Id2, but not Id3, enhanced apoptosis of 32D.3 cells was unexpected. We therefore sought to determine which regions of Id2 were required for this biologic activity. Recombinant expression plasmids were prepared in pMAMNeo containing cDNAs encoding Id2 proteins harboring deletions of either the N-terminal region (pId2ΔN-ter), the C-terminal region (pId2ΔC-ter), or the HLH region (pId2ΔHLH) (Fig. 5). These plasmids were transfected into 32D.3 cells, and pools of clones and individual clones were examined for expression of these altered Id2 proteins following 16 h of treatment with Dex (Fig. 6A and unpublished data). In the pId2ΔN-ter transfectant pool, we observed levels of expression slightly higher than those observed in cells expressing transfected full-length Id2. The Id2ΔHLH mutant was also detected by immunoblotting but was present at lower levels. We could not detect the Id2ΔC-ter protein in immunoblots despite repeated attempts with several different antibodies which can recognize Id2, although the precise epitopes recognized are unknown. Our interpretation of this result is that the Id2ΔC-ter may be present in smaller quantities than the other Id2 proteins that we have studied. This inference is consistent with previous observations by others indicating that deletion of the C-terminal portion of Id3 renders this protein unstable (8).

FIG. 5.

Functional domains of Id2 and Id2 mutants. A series of DNA constructs encoding mutant Id2 proteins was generated for this study including Id2ΔHLH with a deletion of the HLH region (residues 35 to 77), Id2ΔN-ter bearing a deletion of the N-terminal region (residues 2 to 34), and Id2ΔC-ter bearing a deletion of the C-terminal region (residues 93 to 134).

FIG. 6.

The apoptotic function of Id2 is associated with expression of the N-terminal domain. An analysis of mutants lacking the HLH domain (Id2ΔHLH), the C terminus (Id2ΔCter), or the N terminus (Id2ΔN-ter) was carried out. (A) Expression of the proteins in pools of transfected cells selected with G418 and treated for 16 h with Dex was determined by Western blot analysis. Vec-pool, vector pool. (B) The percentage of cells undergoing apoptosis was determined by TUNEL staining (Oncor) 16 h after IL-3 withdrawal. (C) Viability assays were carried out as described above. The experiment shown was performed in triplicate and is representative of three independent experiments.

When these pools of transfectants were examined for apoptosis following IL-3 withdrawal, we observed the expected enhancement in apoptosis in cells expressing exogenous Id2, and a similar level of apoptosis was observed in cells expressing the pId2ΔC-ter-encoded protein (Fig. 6B). Surprisingly, cell cultures transfected with the Id2 mutant lacking the HLH domain exhibited much higher rates of cell death than those observed in the wild-type Id2 transfectants, despite the fact that these cells expressed lower levels of exogenous Id2 protein. This increased activity was observed in several independent transfectants (Fig. 6C and data not shown). Deletion of the HLH domain renders the Id2 protein less stable (23a), yet even low levels of this molecule promote high levels of apoptosis in myeloid progenitors (Fig. 6A and B). In contrast, cells expressing the pId2ΔN-ter-encoded protein did not exhibit enhanced apoptosis, despite the easily detectable levels of both Id2ΔN-ter and endogenous Id2 protein seen in pools and several clones of Id2ΔN-ter transfectants (Fig. 6B and unpublished data). To extend this assessment, we examined the rates of cell death of these transfectants by trypan blue dye exclusion following IL-3 withdrawal. Rates of cell death following IL-3 withdrawal of these clones paralleled closely the pattern of TUNEL-positive cells (Fig. 6B and C). Therefore, Id2 enhancement of apoptosis requires its N-terminal region but does not require HLH-mediated functions.

Id2-induced apoptosis is associated with the expression of Bax.

We surveyed the expression of several known regulators of apoptosis in an effort to identify candidate mediators of Id2-induced apoptosis of 32D.3 myeloid cells. We compared Id2-transfected cells cultured in IL-3 with cells transfected with vector alone and observed an elevated steady-state level of the proapoptotic Bcl-2 family member Bax but no changes of other known regulators, including Bad, Bcl-2, Bcl-X_L, and Bcl-X_S (Fig. 7 and data not shown). As shown in Fig. 7A (upper panel), Bax levels were higher in cells expressing either Id2 or Id2ΔHLH than in 32D.3 cells transfected with the vector alone, with the same vector containing Id2 in the antisense orientation, or with the Id2ΔN-ter expression plasmid. In each case, we examined the expression of Id2 proteins in lysates of these same cells (Fig. 7A, lower panel). As shown, these pools of transfectants generally expressed comparable levels of Id2 wild-type and mutant proteins, although the Id2ΔHLH mutant was again present at lower levels. These changes in Bax protein levels were paralleled by changes in BAX mRNA levels (Fig. 7B), indicating that Id2 influences BAX gene expression in these cells.

FIG. 7.

Id2-induced apoptosis is associated with upregulation of Bax levels. (A) Immunoblot analysis was carried out with lysates from pools of transfectants and individual clones of 32D.3 cells transfected with pMAMNeo vector, pMAMNeoId2, pMAMNeo Id2ΔHLH, and pMAMNeo Id2ΔN-terminus to examine changes in the levels of Bax protein expression. Expression of Id genes was induced with 1 μM Dex for 16 h. The expression of Bax protein (upper panel) and of Id2 and Id2 mutant proteins (lower panel) is shown. (B) Northern blot analysis of the levels of BAX mRNA in 32D.3 cells transfected with pMAMNeo vector, pMAMNeo Id2 (clone 3 [cl3] and a pool), pMAMNeo Id2ΔHLH (pool), and pMAMNeo Id2ΔN-ter (clone 10 and pool) and treated with 1 μM Dex for 16 h. (C) Immunoblot analysis was carried out to compare Bax expression levels in 32D.3 cells stably transfected with pMAMNeo containing an antisense (AS) Id2 cDNA (a pool and an individual clone were examined), grown in the presence or absence of 1 μM Dex for 24 h (upper panel). The same immunoblot was probed with anti-Id2 antibodies to indicate levels of endogenous Id2 expression (lower panel).

To further assess the role of Id2 as a regulator of Bax, we examined a pool and an independent clone (clone 5) of 32D.3 cells stably transfected with antisense Id2 cDNA constructs (Fig. 7B). Cultures of these transfectants were examined by Western blot analysis following 24 h of Dex treatment in the presence of IL-3. Induced antisense expression by treatment with Dex decreased levels of Id2 and Bax (Fig. 7B), whereas no changes in Bax levels were observed in vector transfected cells treated with Dex (Fig. 7A and data not shown). We therefore examined the viability of these cells following IL-3 withdrawal. As shown in Table 1, expression of antisense Id2 was associated with a dramatically increased survival following IL-3 withdrawal from these cells. Taken together, these results suggest that modulating Id2 levels leads to parallel changes in Bax levels, and these levels correlate with Id2-induced apoptosis.

TABLE 1.

Change in survival following induction of antisense Id2

Time (h) following IL-3 withdrawal	Change in % survival following Dex treatmenta
	32D.3 vector	Antisense Id2c.5	Antisense Id2 pool
1	1.0	−4.0	1.0
16	4.1	26.6	40.6
19	4.4	40.0	37.2
22	6.8	29.0	33.3

^aDetermined by calculating the difference in viable cell percentage between cells expressing antisense Id2 and cells not expressing antisense Id2. Cells transfected with vector alone were similarly treated and examined as controls. These values are from a single experiment and are representative of three independent experiments. 

DISCUSSION

A major biological challenge throughout development is the coordination of cellular proliferation and death. Apoptosis is a form of regulated cell death required throughout development for the maintenance of tissue homeostasis. Genes that regulate the initiation or execution of apoptotic cell death have been identified (for a review, see reference 69), and among these are some which also participate in other important cellular processes. We have identified a role for Id2, a known modulator of both differentiation and cell proliferation, in the regulation of apoptosis of myeloid progenitors and osteosarcoma cells (Fig. 2 and 3). Cell death caused by Id2 is recognizable as apoptosis by several criteria, including morphology, analysis of DNA fragmentation, and analysis of subdiploid DNA content. Antisense inhibition of Id2 expression protects 32D.3 cells from cell death following IL-3 withdrawal, supporting the likelihood of a physiologic role for Id2 in mediating apoptosis (Fig. 7). This interpretation is consistent with experiments demonstrating that the proapoptotic activity of Id2 is sensitive to Id2 expression levels (Fig. 3). Moreover, high levels of Id2 override the protective effects survival factors present in serum in U2OS tumor cells of (Fig. 3). Our difficulty in isolating transfectants expressing high levels of exogenous Id2 from some cell lines (data not shown) therefore may have been due to the induction of apoptosis in these cells. Nonetheless, some tumor cell lines can tolerate high levels of Id2 protein (2, 16, 78), and it is possible that such cell lines have undergone selection for changes that suppress Id-mediated apoptosis.

The observation that Id2 mediates both cellular proliferation and apoptosis is consistent with work identifying other multifunctional proteins as gatekeepers of the G₁/S transition. Two important molecular features of the mechanisms utilized by Id genes to promote entry into S phase have been described: (i) Id2 can bind to members of the pRb tumor suppressor family and overcome their growth inhibition and (ii) Id1 can antagonize the bHLH-mediated activation of a known inhibitor of cell cycle progression, p21/WAF1 (34, 45, 60). These observations are most compatible with a model of Id function which proposes that during the G₁-to-S transition Id proteins are released from binding partners such as pRb and then become available to inhibit bHLH transcription factor-mediated gene expression (34). Regulation of Id activity at this juncture in the cell cycle might be further mediated by posttranslational modifications such as phosphorylation of the Id protein by cyclin-dependent kinases (29). Previous studies of c-Myc, E2F-1, and p53 have suggested that the proliferative and apoptotic activities of these molecules are likely associated with their ability to effect transcription, yet their effects on cell death and proliferation are dissociable events (9, 30, 33, 55, 59, 64, 65). Activation of either the proliferative or apoptotic pathway may thus result from differential transcriptional regulation of distinct sets of target genes (23, 33, 59, 64, 72). It is likely that the effects of Id2 on apoptosis and proliferation are also separable, since the HLH region, which is required for all known activities of Id2, is dispensable for enhancing apoptosis. Furthermore, our observation of increased endogenous Id2 in cells expressing the Id2ΔN-terminal mutant (Fig. 6B) raises the interesting possibility of Id2 autoregulation and a dominant negative action of this mutant. Our finding that cells expressing the ΔN-ter mutant die at rates comparable to parental cell rates despite elevated levels of endogenous Id2 (our unpublished observations) is consistent with this possibility.

Several models whereby Id2 functions to integrate signals mediating different biologic activities can be envisioned. Our observations that the N-terminal region of Id2 is required for its apoptotic activity and that the level of expression of a proapoptotic molecule, Bax, is closely correlated with the level of Id2 expression raise the possibility that Id2 can influence gene expression through a heretofore unknown transcriptional activity of its N-terminal domain. Consistent with this is our observation that Id2 expression leads to increases in BAX mRNA levels. In such a model, Id2 might act in trans to either activate transcription of bona fide cell death genes or repress genes important for survival (Fig. 8A). In a related scenario, Id2 protein may sequester transcriptional regulators of survival genes or of apoptotic genes such as BAX (Fig. 8A). It remains unclear whether Id2 can directly regulate BAX. However, 32D.3 cells are highly sensitive to changes in the levels of this proapoptotic molecule and will undergo enhanced apoptosis when deprived of IL-3 if Bax protein is ectopically expressed (11). Our observation that Id2 raises the levels of this protein makes it likely that these effects contribute to the apoptosis caused by Id2 in 32D.3 cells. We cannot exclude the possibility, however, that there are other targets responsible for the apoptotic signal induced by Id molecules.

FIG. 8.

Models for an Id2 apoptotic function. (A) The Id2 N-terminal region may function within the nucleus to promote apoptosis by directly interacting with the transcription complex (TF) mediating the expression of cell death genes. Alternatively, it might act indirectly by binding to nuclear factors which effect gene expression. (B) The cytoplasmic pools of Id2 may regulate cell death by binding molecules mediating or regulating apoptosis. In both models, these interactions would occur independently of dimerization via the HLH domain. Id2 enhancement of programmed cell death may still be antagonized by proteins that normally bind its HLH domain and may be physiologically regulated by the availability of HLH binding partners.

Our finding that deletion of the Id2 HLH domain increases the rate of apoptotic death by Id2 suggests that inhibition of Id function may be mediated by heterodimerization with Rb family members and bHLH proteins, which require the Id2 HLH domain for binding (Fig. 8A). Such a possibility is supported by the well-documented role of Rb in the inhibition of apoptosis (27, 49, 75) and the finding that bHLH transcription factors can also inhibit apoptosis (13). Others have recently reported that the bHLH factor E47 is capable of reversing the activity of Id3 to promote apoptosis in rat embryo fibroblasts (54). We report that Id2 but not Id3 promotes apoptosis in myeloid progenitors and these observations highlight the cell-type-specific function of the different Id molecules. Moreover, it is very likely that E47 or another bHLH protein can also negatively regulate the apoptotic activity of Id2 in 32D.3 cells. The observation indicating that Id gene expression is high in replicating cells and low in most well-differentiated tissues (4, 10, 16, 42) is consistent with this model and predicts that the dynamic balance between Id2 and its binding partners such as E47 or Rb influences whether a cell dies, proliferates, or differentiates.

An important alternative model of Id2 function should also be considered. It has been previously demonstrated that Id genes are detectable in both the cytoplasm and nucleus (5, 34), and E47, a bHLH protein, has been shown to function as a nuclear chaperone for Id1 and Id3 (16). Id genes lack a nuclear localization signal and translocation of Id genes to the nucleus is thought to be mediated by HLH-mediated binding of Id genes to bHLH proteins. Id2ΔHLH enhances apoptosis (Fig. 6) and is likely to be localized exclusively in the cytoplasm. Our observation that the apoptotic function of Id2 does not require the HLH region implies that the apoptotic targets of the N-terminal domain of Id molecules may also be cytoplasmic (Fig. 8B). Such targets could be apoptotic effectors such as the Bax protein itself or, more likely, molecules that regulate the expression of genes important for apoptosis, such as BAX. In such a model, bHLH molecules or pRb present in the cytoplasm may still have an inhibitory effect on Id function (72), as might the transport of Id2 into the nucleus.

These models are not entirely exclusive of one another, and both invoke an apoptotic function for the N-terminal domain of Id2 (Fig. 8). There are 35 amino acids in Id2 which are N-terminal to the HLH domain, and among these are the 32 amino acids deleted in pId2Δ-Nter (Fig. 5). While we are not aware of direct evidence indicating the presence in cells of a polypeptide corresponding to the N-terminal domain of Id2, it is noteworthy that two Id2 loci, Id2A and Id2B, have been found in the human genome (44). While Id2A encodes a polypeptide of 134 amino acids, Id2, Id2B encodes a 36-amino-acid peptide which would correspond to the first 36 amino acids of Id2. While the identification of a cDNA encoded by Id2B containing an in-frame stop codon has been interpreted as suggesting that Id2B may be a transcribed pseudogene (44), it is possible that in humans Id2B encodes a novel Id2-related peptide. Other Id gene family members do not include comparable loci encoding shortened Id polypeptides, and there are only limited sequence similarities in the N-terminal domains of other Id family members. This suggests that if there is a biologically active peptide corresponding to the N-terminal domain, it most commonly arises by other mechanisms. It is possible that under some circumstances Id proteins are processed to yield such a peptide and a residual protein of approximately 100 amino acids. In this regard, our occasional observation in murine cells of a second Id2 immunoreactive species which is slightly smaller than the full-length 14-kDa Id2 (e.g., Fig. 3A, inset, and our unpublished data) may be of importance.

We observed that three different Id transcripts, Id1a, Id1b, and Id2, are expressed at detectable levels in 32D.3 myeloid cells, yet only Id1a and Id2 are regulated during apoptosis following IL-3 withdrawal (Fig. 1). In contrast to Id2 and Id1a, Id3 does not augment the apoptotic program in 32D.3 myeloid cells deprived of IL-3 (Fig. 4). Hence, it is likely that all Id homologs will not be equally active in inducing apoptosis in some cell types and that each Id gene family member will mediate different activities in different cell types. This interpretation is in concert not only with previous work demonstrating that the ability of different Id genes to enhance proliferation is regulated by different cellular mechanisms (34) but also with the observation that there is little sequence similarity among Id homologs outside of their HLH domains (see, e.g., reference 77). Id proteins are widely expressed throughout development, and multiple Id genes are expressed in various tissues (35, 37, 39, 46, 62, 77). Our findings suggest that the Id gene family may function not only to promote cellular proliferation and inhibit differentiation but also to coordinate tissue maturation by promoting apoptosis.

Our identification of a distinct domain of Id2 being required for apoptosis suggests how the different functions of Id genes might be regulated and provides a strong foundation for future studies. By binding to different classes of partners, discrete functional domains within the Id family of molecules may be responsible for mediating distinct cellular activities. Given the variety of cellular processes that Id genes participate in and the number of signals to which they can respond (4, 6, 11, 16), we predict that Id genes might regulate the integration of these signals by activating and inactivating genes important for different cell fates, including proliferation, differentiation, apoptosis, and possibly malignant transformation. An important role during development has already been defined for the Drosophila emc gene (15), a functional homolog of Id which antagonizes the bHLH factors encoded by Daughterless and achaete-scute (20, 24). Since Id genes are most extensively expressed during early embryogenesis, it is tempting to speculate that they too play important roles in modulating multiple cellular activities including apoptosis during early mammalian development.