Characterization of a novel proapoptotic caspase-2- and caspase-9-binding protein.

Caspases play important roles in regulating apoptotic signaling pathways. Here we report the cloning, by the yeast two hybrid system with dominant negative caspase-2 as "bait," of a proapoptotic molecule named proapoptotic caspase adaptor protein (PACAP), encoded by a 372-base pair open reading frame. Binding of this novel protein to caspase-2 (casp-2) was confirmed in yeast two hybrid, in vitro, and in vivo assays. The deduced amino acid sequence revealed homology to functional motifs, including ATP and cytochrome c binding sites. PACAP mRNA was widely expressed in most human tissues; in transfected cells, PACAP was diffusely expressed in the cytoplasm. Bindings studies with the PACAP recombinant protein demonstrated specific binding to casp-2 and casp-9 but not to casp-3, -4, -7, or -8 in cell extracts. Cotransfection experiments showed that PACAP binds to casp-2 and -9 in 293T cells. In addition, studies with truncated PACAP demonstrated a requirement for residues 39-72 of PACAP for specific binding to casp-2 and -9. Transient transfection of PACAP into 293T human kidney cells and rat-1 fibroblasts triggered apoptosis at 24 h, which was at least in part prevented by an inhibitor of casp-3-like enzymes. Transfection of PACAP into human B cell lines using a retroviral system also triggered apoptotic cell death. In addition, transcription of PACAP in primary human B cells was dramatically down-regulated early after cellular activation by CD40L and Staphylococcus aureus and markedly up-regulated as the cells apoptose. These findings identify a novel proapoptotic caspase adaptor protein.

Programmed cell death, or apoptosis, is essential for the development and homeostatic maintenance of many cell lineages, including the immune system, where it regulates the antigenic specificity of the system (1)(2)(3). Apoptotic signaling pathways are dependent on the activation, by proteolytic cleavage after key aspartic acid residues, of a family of cysteine proteases, termed caspases, which mediate cleavage and functional destruction of various essential intracellular proteins.
Cleavage-dependent activation of caspases is mediated by other caspases or autocatalytically with the assistance of adaptor proteins, such as Apaf-1, 1 Nod/caspase recruitment domain 4 (CARD4), RAIDD/CRADD, DEFCAP, and NAC, which facilitate the formation of multimolecular complexes, oligomerization of included caspases, and autocatalytic activation (4 -7). Fourteen structurally related mammalian caspases have been identified to date and subgrouped into three subfamilies based on phylogenetic considerations, sequence homology, predicted structure, and substrate specificity (4,8). Various death stimuli, such as TNF ␣ or Fas ligand, growth factor deprivation, DNA damage, certain drugs, etc., activate caspase-dependent intracellular signaling pathways (9). The "initiating" or "upstream" caspases in these pathways contain a large aminoterminal prodomain, termed a CARD, which contains motifs for protein-protein interactions, whereas effector or downstream caspases contain short prodomains (9). Activation of effector caspases leads to cleavage and functional destruction of various essential intracellular proteins and thus to apoptotic cell death (4,8).
Several apoptotic stimuli induce the release of cytochrome c (cyt c) from mitochondria into the cytosol, where it interacts with the WD40 motif of the adaptor protein Apaf-1. Apaf-1 and cyt c complexes apparently not only facilitate the hydrolysis of dATP/ATP, but they also oligomerize via CED4 domains into larger complexes, which bind and facilitate the activation of procaspase-9 (10 -12). Such apoptosome complexes also contain casp-3 and -7, indicating that casp-9mediated activation of these effector caspases occurs in the apoptosome (13,14). Other studies in cell-free lysates have shown that casp-9 is activated in cell free lysates in a Apaf-1-dependent manner after addition of cyt c and that Apaf-1independent sequential activation of casp-2, -3, -6, -7, -8, and 10, but not of casp-1, -4, and -5, follows (15). In the immune system, apoptosis triggered by ligation of T or B cell antigen receptors may trigger cellular proliferation and differentiation or, alternatively, negative signals leading to apoptosis, depending on many factors, including cellular maturation state, tissue location, and co-ligation of other cell surface receptors (2,16). In our previous studies of apoptosis triggered by ligation of the BCR on Epstein-Barr virus-negative mature and immature human B cell lines, we found that cell death proceeded via a previously unreported intracellular signaling pathway (17). Casp-2 was activated early after B cell receptor ligation and was required for apoptosis, whereas casp-3, an effector caspase involved in most apoptotic pathways, was activated subsequently. Casp-9 was activated much later and likely functioned to amplify the apoptotic signal. Casp-8 and -1, which are activated by ligation of the CD95 and TNF-R1 cell death receptors, were not involved (17). Although casp-2 has been found to become activated upstream of casp-3 in various primary cultures and cell lines in response to different apoptotic stimuli, including etoposide, ␥-irradiation, or growth factor withdrawal, and to induce apoptosis upon overexpression in various cell types (18 -25), its involvement early in apoptosis triggered by ligation of a membrane receptor is unprecedented.
Casp-2 is unique among the caspases because it exhibits sequence similarity and shares structural features of two of the caspase subfamilies with a CARD prodomain characteristic of initiator caspases and a substrate specificity more typical of downstream effector caspases; for these reasons, it is grouped in a separate subfamily. Two mechanisms for casp-2 activation have been reported. First, the CARD of casp-2 binds to the CARD of a 22-kDa death adaptor protein termed RAIDD or CRADD (26,27). Because RAIDD/CRADD also contains a death domain (DD) capable of binding the DD of a protein termed RIP, a component of the TNF-R1 and CD95 signaling complex, it is postulated that RAIDD/CRADD couples casp-2 to the death signaling complex, leading to casp-2 activation (26,27). Second, casp-2 has been shown to be activated by the cytotoxic lymphocyte granule proteins granzyme B and perforin (20). Furthermore, casp-2 is essential for apoptosis of B lymphoblasts triggered by granzyme B and oocytes triggered by doxorubicin (28) and for apoptosis of neurons stimulated by amyloid ␤-peptide (25).
The present studies were initiated in order to determine the mechanism of activation of casp-2 in apoptosis of B cells triggered by ligation of the antigen receptor. Because caspases interact with other proteins that regulate their activation and function, as noted above, we sought to identify a casp-2-interacting protein by screening a B cell library using the yeast two hybrid system with dominant negative casp-2 as "bait." The cloning and functional studies of the novel protein termed proapoptotic caspase adaptor protein (PACAP) are presented here.
Cloning of PACAP-A cDNA fragment encoding human dominant negative (DN) caspase-2 (17) was amplified by PCR and inserted into the SmaI site in pBD-GAL4 Cam (Stratagene, La Jolla, CA) to generate pBD-GAL4 Cam/DN casp-2. After verification of the sequence, Saccharomyces cerevisiae CG1945 cells (CLONTECH, Palo Alto, CA) were sequentially transformed with pBD-GAL4 Cam/DN casp-2 and a human B lymphocyte cDNA library (MATCHMAKER, CLONTECH; Ͼ10 8 cfu/ml) in pACT using a lithium acetate transformation protocol. Selection was done by growth on SD medium lacking histidine, uracil, and tryptophan (CLONTECH). Twenty clones exhibiting activation of the lacZ reporter gene were identified among 3 ϫ 10 6 transformants by the ␤-galactosidase assay, but only one shown a reproducible interaction with casp-2 DN. Plasmids were isolated from positive yeast colonies by a glass bead phenol-chloroform extraction protocol (CLONTECH). Escherichia coli DH5␣ cells were transformed with the plasmids, and cells containing the pACT vector were selected in ampicillin resistant plates. The pACT plasmids were isolated from E. coli and restrictionmapped (XhoI), and the sequence of the insert was determined by DNA sequencing.
The sequence upstream of the predicted ATG start site was determined through the use of the 5Ј rapid amplification of cDNA ends kit (Life Technologies, Inc., Gaithersburg, MD) with total RNA isolated with TRIzol (Life Technologies, Inc.) from human brain and peripheral blood leukocytes, both of which were found to express high levels of PACAP message. Three nested primers reflecting sequences ϳ300 bp downstream from the predicted start site (5Ј-AGACGTTTCACTTG-GTCCACTTCT-3Ј, 5ЈCTCCGGTCCAGGACATCCGTGTAG-3Ј, and 5Ј-AGTTTGGTCTCTGCC TTTGCCAGAT-3Ј) were assayed by PCR one at a time with abridged anchor primer provided in the kit. Identical sequences were contained in the 300 -500-bp fragments amplified by the three combinations of primers.
Assays for Binding of PACAP to DN Casp-2 and GST Recombinant Proteins-Yeast two hybrid binding assays were employed to confirm binding of PACAP to DN casp-2. In these experiments, S. cerevisiae CG1945 cells were sequentially transformed with purified pBD-GAL4 Cam/DN casp-2 and pACT containing the novel clone or other control proteins, including pCL1 (coding for the wild type GAL4 protein) and pVA3-1 (a fusion of murine p53 and GAL4DNA-BD). Interactions between the two proteins were confirmed by activation of both reporter genes (lacZ and LEU2).
Radiolabeled ( 35 S) DN casp-2 was generated using the TNT reticulocyte lysate system (Promega, Madison, WI). Unincorporated isotope was removed by passage through a G25 Sephadex column (Amersham Pharmacia Biotech, Piscataway, NJ). SDS-PAGE followed by autoradiography revealed the presence of a strongly labeled 47 kDa band, consistent with the presence of DN casp-2.
For the preparation of GST fusion proteins, full-length and truncated PACAPs were subcloned into the BamHI and EcoRI sites of the pGEX4-T bacterial expression vector (Amersham Pharmacia Biotech), in-frame with the amino-or carboxyl-terminal GST tag. After expression in E. coli DH5␣ or BL21 (Life Technologies, Inc.), bacterial cells from 400-ml overnight cultures grown at 37°C were resuspended in an equal volume of fresh LB medium. After 30 min of additional growth at 30°C, a final concentration of 1 mM isopropyl-1-thio-␤-D-galactopyranoside was added for 90 min. The induced bacteria were collected by centrifugation, resuspended in 30 ml of cold phosphate-buffered saline, and sonicated. The bacterial cells were further incubated with Triton X-100 (1%) and lysozyme at 0.2 mg/ml at room temperature for 1 h. The GST fusion products were purified and immobilized on glutathione-Sepharose beads (Amersham Pharmacia Biotech).
Distribution of PACAP in Human Tissues-The human Rapid-Scan-Gene expression panel (OriGene Technologies, Rockville, MD), a semiquantitative PCR method, was used to evaluate the expression of PACAP in different human tissues. The high quality first strand cDNAs from 23 tissues provided in the kit were normalized using ␤-actin as the internal control according to the manufacturer's protocol and evaluated undiluted and at a 100-fold dilution. The PCR settings were established according to the T m of the primers and the hot start Taq polymerase optimal temperature requirements. The upstream and downstream primers were 5Ј-CGAGGATCCATGAGGCTGTCACTGCCACTG-3Ј and 5Ј-CGGGAATTCGCCACGAAGGCCTTTTTTTTT-3Ј, respectively. The following PCR settings were used: 5 min at 72°C, 15 min at 95°C; and then 35 cycles of 1 min at 94°C, 1 min at 58°C, and 1 min at 72°C; ending with 1 min at 72°C. The products were analyzed on 12% agarose gels and stained with ethidium bromide.
Transient Transfection and Immunoprecipitation Studies-Human 293T cells and Rat-1 fibroblasts were transiently transfected with PACAP using the calcium phosphate and LipofectAMINE (Life Technologies, Inc.) methods, respectively, according to the manufacturer's instructions. For these studies, PACAP was subcloned into engineered EcoRI and XbaI sites in the pcDNA 3.1 Myc-His EpiTag vector (Invitrogen) in-frame with the carboxyl-terminal tags of this vector. 5-Bromo-4-chloro-3-indolyl-␤-D-galactopyranoside staining was employed to monitor transfection efficiency in controls. Apoptosis was assessed by nuclear morphology 24 h after transfection by optic and fluorescent microscope (29).
Human 293T cells were co-transfected with pcDNA3.1 Myc-His Epi-Tag-PACAP, pBactH37Z-casp-2, and pcDNA3.1-casp-9 using the calcium phosphate method. Eighteen hours posttransfection, cells were harvested and lysed in the buffer used in the pull-down experiments, and the lysates were precleared with protein G beads (Amersham Pharmacia Biotech) for 3 h at 4°C. After centrifugation, the supernatants were transferred into clean tubes and incubated with anti-Myc monoclonal antibodies for 3 h at 4°C. The mixtures were then incubated with protein G beads for 2 h at 4°C. After washing, the samples were subjected to SDS-PAGE analyses followed by immunoblotting with anti-casp-2 and -9 antibodies. Controls included cells transfected with PACAP alone and casp-2 and -9 alone, as well as a sample of the lysates of the triply transfected cells, prior to immunoprecipitation.
Transfection Studies in B Cell Lines by Retroviral Infection-Fulllength PACAP was cloned into a carboxyl-terminal GFP-expressing amphotropic Epstein-Barr virus/retroviral vector termed PINCO, obtained from Dr. P.G. Pelicci, Milan, Italy (30). Phoenix packaging cells, obtained from the ATCC (Manassas, VA) were grown in 100-mm dishes and transfected by the calcium phosphate method with control PINCO and with the PINCO/PACAP vector. Recombinant retroviral particles were collected from the supernatant 2 days later and used to infect ST486 and B104 Epstein-Barr virus-negative human B cell lines. Three days later, GFP-positive B cells were sorted on a FACS Vantage flow cytometer (Becton Dickinson), and the positive cells (Ͻ1% GFP-positive) were cultured in fresh medium.
Immunohistochemical Studies-Immunohistochemical studies were performed as earlier described (29). Briefly, coverslip cultures of 293T and Rat-1 cells were fixed with methanol:acetone (80:20) for 5 min and then incubated overnight with monoclonal IgG1 anti-Myc antibody (Invitrogen) at 4°C. After washing, the cells were incubated with streptavidin-conjugated anti-mouse IgG followed by FITC anti-mouse antibody or horseradish peroxidase-labeled biotin and substrate. TO-PRO3 (Molecular Probes, Eugene, OR) was used as a nuclear counterstain with FITC and hematoxylin with horseradish peroxidase. The percentage of FITC-positive cells was assessed by a laser scanning cytometer (Compucyte, Cambridge, MA).
Western Blotting Studies-Cells or proteins were lysed with 2ϫ SDS buffer and boiled, and cytosolic extracts were analyzed on SDS-PAGE gels under reducing and nonreducing conditions. After electrotransfer to Immobilon membranes and blocking (5% milk), the membranes were incubated with mouse anti-Myc (Invitrogen) or mouse anti-casp-2, -3, -9, or -8 (Pharmingen) antibodies overnight at 4°C. After washing, the membranes were incubated with horseradish peroxidase-conjugated antibody to mouse immunoglobulin followed by substrate.
Expression of PACAP mRNA in Primary Human B Cells-Positively selected, CD19 ϩ MACS-purified human B cells (AllCells, San Jose, CA) were resuspended at 1 ϫ 10 6 cells/ml in Yssel's medium without antibody (Gemini Bioproducts, Calabasas, CA) supplemented with 10% fetal bovine serum (Hyclone, Logan, UT). Aliquots of the B cell suspension were distributed into flat-bottomed 96-well microtiter plates (100 l/well). Activation was initiated by the addition of either 5 g/ml CD40 ligand-fusion protein (obtained from Dr. V. Brinkmann, Novartis Pharma, Basel, Switzerland) together with 10 ng/ml IL-4 (Sigma), or Staphylococcus aureus Cowan (CalBiochem, San Diego, CA) in Yssel's medium with 10% fetal bovine serum. Successful activation was monitored by assessing CD20, CD25, and CD69 expression on days 0, 1, 2, and 5 in a flow cytometer. B cell proliferation was assessed by the addition of 2 ϫ 10 4 Nite Red microsphere particles (Spherotech, Libertyville, IL) to each well just before staining. Time course studies were carried out by pooling the cells from several wells at 0 min, 30 min, 2 h, 6 h, 1 day, 2 days, and 5 days. Total RNA was extracted with TRIzol (Life Technologies, Inc.), and PACAP mRNA expression was evaluated by RT-PCR carried out on 1 g of RNA per sample using 5Ј-CGAG-GATCCATGAGGCTGTCACTGCCACTG-3Ј and 5Ј-CGGGAATCCGC-CACGAAGGCCTTTTTTTTT-3Ј as the upstream and downstream primers, respectively. These primers amplify an 800-bp sequence. Expression of glyceraldehyde-3-phosphate dehydrogenase was monitored for equal loading and amplification efficiency using 5Ј-CTGAGAACGG-GAAGCTTGTC-3Ј and 5Ј-CCTGCTTCACCACCTTCTTG-3Ј as the upstream and downstream primers, respectively; these primers amplify a 600-bp sequence.

RESULTS
Cloning of PACAP cDNA-Screening of a human B cell cDNA library (3 ϫ 10 6 transformants) by the yeast two hybrid system with human DN casp-2 (C303A) as bait yielded 20 yeast colonies, which activated both reporter genes (LEU2 and lacZ). We used DN casp-2 as bait, because the wild type casp-2 reduced the normal growth rate of the yeast cells. After transformation into E. coli, selection, isolation, and restriction mapping, the sequence of the insert was determined. Only one of these colonies, coding for a novel gene we named PACAP (Fig.  1), was confirmed to bind DN casp-2 by activation of both reporter genes when the yeast 2-hybrid binding assay was repeated with the plasmid from the initial screens and DN casp-2 ( Fig. 2A).
The cDNA sequence of the insert contains an initiator methionine, a stop codon, a polyadenylation signal, and a poly(A) tail (Fig. 1). The insert contains 905 bp from the initiator codon to the end of the poly(A) tail (Fig. 1). The initiator codon, which conforms to Kozak's rules for initiation (31), was confirmed by the 5Ј rapid amplification of cDNA ends studies, which also yielded 300 -400 additional bp of 5Ј sequence. The stop codon is located 371 bp downstream of the initiator methionine (Fig. 1). The deduced amino acid sequence consists of 123 amino acids with a calculated molecular mass of 13.6 kDa. A TGA stop codon is located 225 bp upstream of the initiator methionine. The GenBank TM accession number for the cDNA is AF338109.
Residues 1-22 of the deduced amino acid sequence conform to the consensus sequence of a cleavage site of a signal peptide for a secreted protein (32,33). The inability to detect aminoterminal Myc-tagged PACAP (1-123) coupled in the ease of detecting carboxyl-terminal Myc-tagged PACAP (1-123) in transfection studies in mammalian systems, as described below, suggests that residues 1-22 represent a signal peptide, which is absent from the mature protein. For these reasons, ⌬23-123 PACAP is considered to represent the mature protein.
Data base searches using the Blast server at the National Center for Biotechnology Information identified a partially homologous cDNA sequence to PACAP, which was cloned from a CD34 ϩ stem cell library (GenBank TM accession number AF151024). This cDNA sequence, which has not been published in the literature, is completely homologous to the 3Ј nucleotide sequence including the poly(A)-tail of PACAP (about 320 bp), but the sequence upstream sequence is completely different from the cDNA sequence reported here. In addition, the deduced amino acid sequence of AF151024 does not match that of the present clone due to a missing nucleotide, which produces a frameshift. In searching the GenBank TM expressed sequence tag data base, nine expressed sequence tag clones derived from eight different proliferative and nonproliferative human tissue cDNA libraries (AI698118, AI920943, AA363205-5, AI670058, AI004279, AW001552, AI203960, AI203981, and AI762046) showed 98 -100% homology to the untranslated 3Ј 300 -550 bp of PACAP downstream of the stop codon, as assessed by the National Center for Biotechnology Information Blast search.
In Vitro Demonstration of PACAP Binding to DN Casp-2-An independent assay system was used to demonstrate binding of full-length PACAP to DN casp-2. For these studies, PACAP was translated in-frame with a carboxyl-terminal GST tag, purified, and immobilized on glutathione-Sepharose beads. As shown in Fig. 2B, the beads bound 35 S-labeled recombinant DN casp-2 generated by the reticulocyte lysate system, confirming the ability of PACAP to bind casp-2.
Expression of PACAP mRNA in Human Tissues-A semiquantitative PCR method was used to evaluate PACAP mRNA expression in different human tissues. The high quality first strand cDNAs from 23 human tissues were subjected to RT-PCR analyses using PACAP specific primers. As indicated in Fig. 3, PACAP mRNA was detectable in all of the tissues except placenta and fetal brain. Expression was highest in the brain, whereas heart, spleen, skin, ovary, adrenal gland, pancreas, bone marrow, small intestine, muscle, stomach, testis, and peripheral blood lymphocytes exhibited intermediate levels of expression. Very low, but detectable, expression was found in kidney, liver, colon, lung, uterus, prostate, salivary gland, and thyroid.
In Vivo Binding of PACAP to Casp-2 and -9 -In order to determine whether PACAP binds to casp-2 and -9 in intact cells, 293 T cells were cotransfected with carboxyl-terminal Myc-tagged PACAP and casp-2 and -9. Eighteen hours later, the cells were lysed. Myc-tagged PACAP was immunoprecipitated with anti-Myc antibodies, and binding of casp-2 and -9 to PACAP was assessed by SDS-PAGE analysis followed by Western blotting studies with antibody to casp-2 and -9. As shown in Fig. 4C, PACAP bound to both caspases in the transfected cells. Identical results were obtained in three independent experi-ments. The experiments were performed in the reverse direction (immunoprecipitation with antibody to casp-2 and -9 and Western blotting with anti-Myc antibody) with identical results (not shown).
The Amino-terminal Portion of PACAP Is Required for Binding to Casp-2 and -9 -To evaluate binding sites in PACAP, we constructed several amino-terminal deletion constructs of the novel protein (Fig. 4D). Full-length PACAP cDNA (1-123), mature PACAP (⌬23-123), and deletion constructs ⌬25-123, ⌬39 -123, and ⌬72-123 were fused with an amino-terminal GST tag in the pGEX4T1 plasmid (Fig. 4D). The recombinant proteins were obtained using two different bacterial strains, DH5␣ and BL21, with the same results. Proteins were evaluated in stained SDS-PAGE gradient gels. The GST-mock construct exhibited an apparent molecular mass of 26 kDa, as expected (Fig. 4E). However, the full-length GST-PACAP (1-123) construct failed to show the predicted fusion protein band, but rather exhibited only the 26-kDa GST band plus a slightly larger band (Fig. 4E, lane 1). The GST-PACAP ⌬23-123, ⌬32-123, and ⌬39 -123 constructs showed major bands at ϳ34, 33, and 32 kDa (Fig. 4E, lanes 2-4). Two additional smaller bands with approximate molecular masses of 27 and 28 kDa were also evident. These data indicate that the amino-terminal portion of the truncated PACAP constructs fused to GST is further cleaved in the bacteria at two sites. Cleavage was not observed with the GST-⌬72-123 PACAP construct, which exhibited an apparent molecular mass of 36 kDa (Fig. 4E, lane 5). The data are consistent with a cleavage site of PACAP in the bacteria between residues 39 and 72. Despite their small size, the GST-PACAP (⌬23-123, ⌬32-123, and ⌬39 -123) mutants bound to casp-2 as well as casp-9 (Fig. 4, F and G, lanes 2-4). The GST-⌬72-123 PACAP, however, failed to bind either of these caspases (Fig. 4, F and G, lane 5). These data localize the binding site for casp-2 and -9 to the region of the PACAP protein encompassed by residues 39 -72. Two different concentrations (100ϫ and 1ϫ) of cDNA of the various tissues supplied in the kit were amplified, and the products were separated by electrophoresis on agarose gels followed by staining with ethidium bromide. ␤-Actin was detected by PCR with primers provided in the kit. efficiency in human 293T cells assessed a day after transfection was 60 Ϯ 8% (Fig. 5A). However, 69 Ϯ 7% of the transfected cells died during the incubation period (Fig. 5E). The transfection efficiency was lower (Ͻ10%) in Rat-1 fibroblasts. As observed with the 293T cells, a large proportion (55 Ϯ 5%) of the transfected cells died during the incubation time. Microscopic evaluation of transfected 293T cells showed that many of the cells were shrunken and had pyknotic nuclei, indicative of apoptotic cell death (Fig. 5B). Comparable findings were observed with transfected Rat-1 fibroblasts (not shown). These data indicate that overexpression of PACAP in human 293T cells and in rat fibroblasts leads to apoptotic cell death.

PACAP Induces Apoptosis upon Transient Transfection into Human Embryonic 293T Kidney Cells and Rat Fibroblasts-
Immunohistochemical studies demonstrated that Myctagged PACAP was expressed in the cytosol (Fig. 5B, C, arrows). Immunofluorescence studies showed that the protein exhibited a fine granular distribution in the cytosol surrounding the nucleus (Fig. 5C). Very similar findings were obtained with transfected Rat-1 fibroblasts (not shown). Apoptotic cell death was, at least in part, inhibited by pretreatment of the PACAP-transfected cells with 300 nM DEVD, because cell death was reduced from 69 Ϯ 7 to 41 Ϯ 9% (Fig. 5E). In addition, the status of PACAP in transfected 293T cells was evaluated in Western blotting studies with antibody to the Myc tag. Under reducing conditions, carboxyl-terminal Myc-tagged PACAP exhibited an apparent molecular mass of about 14 kDa (Fig. 5D), consistent with the predicted size of a PACAP (⌬23-123) Myc fusion protein.
Transfection of PACAP into Human B Cell Lines Using a Retroviral Vector-ST486 and B104 human B cells were stably infected with a PACAP retroviral construct for further studies of the function of PACAP. Previous efforts to obtain stable B cell transfectants using transient transfection methodologies have not been successful (17). In the present studies, therefore, we used a GFP carboxyl-terminal-expressing, amphotropic retroviral vector termed PINCO, which has been reported to yield high transfection efficiencies in human hematopoietic cells (30). Three days after infection of the Epstein-Barr virus-negative ST-486 human B cell line with the vector or with the vector containing PACAP, only ϳ1% of the cells expressed GFP (Fig. 6, A and B), in contrast to higher levels in the published report (30). The GFP and GFP-PACAP-positive cells were then sorted and evaluated for apoptosis. Although the control vector infected B cells grew rapidly, high levels of cell death were observed in the PACAP-PINCO infected cells (Fig. 6C). Comparable results were obtained with the B104 cell line. As a control, PACAP expression was detected by RT-PCR to evaluate the presence of the gene in the cell lines (not shown). These data indicate that PACAP overexpression in human B cell lines, as in 293T embryonic kidney cells and Rat-1 fibroblasts, triggers cell apoptosis.
Expression of PACAP mRNA during Activation of Primary Human B Cells-We evaluated PACAP expression by RT-PCR during the activation of human primary B cells with SAC, which binds to the antigen receptor, as well as during activation by CD40 ligand plus IL-4 (CD40L/IL-4), which mimics T helper cell-mediated B cell activation (37,38). Primary B cell activation by these agents is followed by proliferation and ultimately by cell death. PACAP mRNA was constitutively expressed in the primary B cells (Fig. 7, A and B). Expression abruptly decreased 3 h after activation by SAC and by CD40L/ IL-4 and then returned to resting levels 12 h (SAC) or 24 h (CD40L/IL-4) after addition of the activators (Fig. 7, A and B). By 24 h (SAC) or 48 h (CD40/IL-4) after activation, PACAP mRNA levels in the activated cells considerably exceeded constitutively expressed levels. Marked activation, documented by the expression of the CD69 and CD25 activation markers, was observed 24 h after addition of the ligands (Fig. 6, C and D). Marked cell death was evident in the cultures by 3 days (SAC) and 5 days (CD40L/IL-4) after stimulation (not shown). These findings indicate that primary human B cells constitutively express PACAP mRNA. They further show that PACAP mRNA expression decreases with B cell activation triggered by either ligand combination and increases as the cells begin to die. DISCUSSION Because multiple homophilic and heterophilic interactions with other proteins regulate caspase actions, we used the yeast two hybrid system with DN casp-2 as bait with the goal of identifying a possible casp-2-interacting protein. These experiments identified a novel protein, termed PACAP, with a calculated molecular mass of 13.6 kDa. This is consistent with an observed molecular mass of 14 kDa for the protein on reduced SDS-PAGE gels.
PACAP triggered apoptosis upon transient transfection of human kidney cells and Rat-1 fibroblasts and early in stable transfection of human B cell lines using a retroviral vector. These findings clearly identify PACAP as a member of an apoptotic pathway. PACAP does not contain a CARD, a motif that mediates interactions between some of the caspases and certain other apoptotic signaling proteins. PACAP does, however, contain an amino-terminal WD40 repeat, a P loop nucleotide binding motif, and a cyt c binding site, motifs that are found in several apoptotic signaling proteins, including Apaf-1. The WD40 motif is unlikely to be functionally significant, be-cause it is contained in the first 22 amino acids of the protein that appear to function as a signal peptide. PACAP also contains consensus sequences for phosphorylation by tyrosine and serine kinases, as well as potential myristylation sites. Finally, PACAP exhibits motif homologies to the active sites of a number of families of enzymes, including glycoproteases, cysteine proteases, serine proteases, and thiol proteases. The presence of multiple consensus sites characteristic of intracellular signaling pathways suggests the possibility not only that PACAP is a component of such a signaling pathway but also that it represents an adaptor protein that, in analogy to Apaf-1, interacts with cyt c and dATP and facilitates the activation of one or more caspases. However, because PACAP lacks a CARD, it clearly cannot bind via homophilic interactions to the CARDs of casp-2 or -9 or to another protein with this domain. The absence of a CARD plus the complete lack of sequence similarity differentiates PACAP from RAIDD/CRADD, a bipartite anchoring molecule that binds to casp-2 via its amino-terminal CARD, and to RIP, a component of the tumor necrosis factor-R1 signaling complex, via its carboxyl-terminal DD (26,27). The absence of a CARD, lack of sequence homology, and different molecular masses also differentiate PACAP from two other recently reported casp-2-binding proteins. These include two forms of a Ced-4/Apaf-1 family member termed DEFCAP-L (1473 residues), and DEFCAP-S (1429 residues), to which casp-2 binds via homophilic CARD interactions (5). The second casp-2-binding protein is caspase-2S-binding protein, a 191residue protein that binds to the carboxyl-terminal portion of the short form of casp-2 via an unknown type of interaction and blocks its apoptotic actions (39). The absence of sequence similarity and the different molecular masses distinguish PACAP from the casp-9 binding adaptor protein, Apaf-1.
Analyses of PACAP mRNA expression revealed its presence in many adult tissues with different levels of expression. For example, PACAP is most highly expressed in the adult brain as is casp-9, whereas casp-2 expression is barely detectable in this organ (23). On the other hand, casp-2 mRNA is most highly expressed in the developing brain and placenta, both of which lack detectable PACAP expression (23). The message for these proteins is also expressed in other tissues, but at differing relative levels (5,26,27,35,39).
Amino-terminal deletion mutants of PACAP were generated in two bacterial strains (DH5␣ and BL21) in-frame with an amino-terminal GST tag in order to begin to identify binding sites for casp-2 and -9 in the molecule. The full-length PACAP peptide was generated as cleaved fragments due to a cleavage site located between residues 22 and 23. The inability of the GST amino-terminal PACAP peptide construct to bind either casp-2 or -9 indicates that the WD 40 repeat contained in this portion of the protein is not important for caspase binding. The ⌬23-123, ⌬32-123, and ⌬39 -123 PACAP constructs were all cleaved in the bacteria, yielding GST constructs fused with short amino-terminal PACAP peptides. Those short peptides bound casp-2 and -9. The PACAP ⌬72-123 peptide, which remained largely intact in the bacteria, was not able to bind casp-2 or -9.Thus, the casp-2 and -9 binding site in PACAP is located between residues 39 and 72. Studies are under way to further localize the casp-2 and -9 binding site in PACAP within this 33-residue sequence. Cotransfection studies with casp-2, casp-9, and PACAP in 293T cells indicate that PACAP (23-123) binds both caspases in intact cells. These in vivo studies support a physiological role for these interactions.
Potential proapoptotic actions of the amino-terminal truncated forms of PACAP were not evaluated in the present study. However, the clear reduction in the amounts of detectable casp-3, -4, and 7 in Jurkat T cells lysates after exposure to the GST-PACAP ⌬23-123 construct, which maximally contains residues 23-72 of PACAP, strongly suggests that this portion of the protein indirectly activates these downstream effector caspases in the Jurkat T cell lysates. The ability of DEVD, an inhibitor of casp-3-like enzymes, to partially block apoptosis in PACAP-transfected 293 T cells is consistent with a role for downstream caspases in the PACAP apoptotic pathway.
The constitutive expression of an apoptotic gene in many tissues in the absence of an apoptotic stimulus is surprising. It is possible that PACAP is constitutively expressed in an inactive state and that activation (cleavage at position 22) is required in order for PACAP to bind to and regulate the acti- vation of casp-2 and -9. These could occur via binding of cyt c and dATP, in analogy to the activation of Apaf-1 (10), or via other processes, including membrane localization, proteolytic cleavage, and phosphorylation. We are evaluating these possibilities.
The rapid down-regulation of PACAP mRNA levels in primary B cells after stimulation with CD40L/IL-4 is also of interest. These findings suggest that PACAP expression must be down-regulated in order for the cells to divide and thus may also indicate a role for PACAP in B cell cycle regulation. Furthermore, PACAP mRNA expression is up-regulated during death. This and other aspects of the functions of PACAP in resting, activated, and cycling cells are under investigation.