K15 protein of Kaposi's sarcoma-associated herpesvirus is latently expressed and binds to HAX-1, a protein with antiapoptotic function

Abstract

The Kaposi's sarcoma-associated herpesvirus (KSHV) (or human herpesvirus 8) open reading frame (ORF) K15 encodes a putative integral transmembrane protein in the same genomic location as latent membrane protein 2A of Epstein-Barr virus. Ectopic expression of K15 in cell lines revealed the presence of several different forms ranging in size from full length, approximately 50 kDa, to 17 kDa. Of these different species the 35- and 23-kDa forms were predominant. Mutational analysis of the initiator AUG indicated that translation initiation from this first AUG is required for K15 expression. Computational analysis indicates that the different forms detected may arise due to proteolytic cleavage at internal signal peptide sites. We show that K15 is latently expressed in KSHV-positive primary effusion lymphoma cell lines and in multicentric Castleman's disease. Using a yeast two-hybrid screen we identified HAX-1 (HS1 associated protein X-1) as a binding partner to the C terminus of K15 and show that K15 interacts with cellular HAX-1 in vitro and in vivo. Furthermore, HAX-1 colocalizes with K15 in the endoplasmic reticulum and mitochondria. The function of HAX-1 is unknown, although the similarity of its sequence to those of Nip3 and Bcl-2 infers a role in the regulation of apoptosis. We show here that HAX-1 can form homodimers in vivo and is a potent inhibitor of apoptosis and therefore represents a new apoptosis regulatory protein. The putative functions of K15 with respect to its interaction with HAX-1 are discussed.

FIG. 1.

K15 expression analysis. (A) Sequence comparison of K15 C-terminal domain for the P and M forms. The predicted C-terminal cytoplasm domains of both the P and M forms of K15 are shown. Grey shading indicates identical amino acids. Dashed lines indicate amino acids deleted for mutational analysis (see below). Predicted SH2-B and SH3-B domains are indicated. The boxed region indicates putative mitochondrial targeting sequence. (B) In vitro TNT in the presence of [³⁵S]methionine and pCR3.1 (vector only) or pCR3.1K15 as indicated. The full-length eight-exon form of K15 produces a protein of approximately 50 kDa. (C) HeLa and the endothelial cell line IE7 as indicated were transiently transfected with either vector-only control (pCR3.1) or full-length wild-type K15 (pCR3.1K15). The predominant protein sizes for K15 in all three cell types were 23 and 35 kDa. (D) Primary HMVEC were infected with lentivirus expressing K15 (Lenti-K15) and virus not expressing K15 (Lenti-empty [negative control]). Equal viral loads were determined by p24 ELISA (data not shown). Cell extracts were made 48 h postinfection and Western blotted with anti-K15 MAb (3B5/D7), developed using ECL, and exposed to autoradiographic film for the indicated times. The blot was subsequently stained with Coomassie blue to demonstrate equal protein input. (E) Infection of HMVEC with Lenti-empty, Lenti-K15, and Lenti-K15ΔAUG. At 48 h postinfection cell extracts were Western blotted with anti-K15 MAb. (F) Internal signal sequence cleavage site prediction for K15. For the prediction of cleavage sites within K15 we employed the SignalP V1.1 prediction program (see Materials and Methods). The arrows indicate amino acid positions of predicted cleavage sites and the molecular mass of the cleaved C-terminal protein product. Red arrows indicate those sites that most closely match the cleavage consensus sequence. (G) Endogenous expression of K15. Raji and BC3 (PEL) cell extracts equivalent to 10⁵ cells were also subjected to SDS-15% PAGE and blotted for K15 with the 3B5/D7 MAb as described above. Positions of molecular mass markers are indicated at left. The predominant protein detected was 23 kDa. (H) K15 protein kinetics of expression following lytic cycle induction. BC3 and DG75 cells (6.0 × 10⁵) were treated with TPA (20 ng/ml). At the time points indicated 10⁵ cells were harvested and lysed and subjected to SDS-PAGE and anti-K15 blotting as described above.

FIG. 1.

K15 expression analysis. (A) Sequence comparison of K15 C-terminal domain for the P and M forms. The predicted C-terminal cytoplasm domains of both the P and M forms of K15 are shown. Grey shading indicates identical amino acids. Dashed lines indicate amino acids deleted for mutational analysis (see below). Predicted SH2-B and SH3-B domains are indicated. The boxed region indicates putative mitochondrial targeting sequence. (B) In vitro TNT in the presence of [³⁵S]methionine and pCR3.1 (vector only) or pCR3.1K15 as indicated. The full-length eight-exon form of K15 produces a protein of approximately 50 kDa. (C) HeLa and the endothelial cell line IE7 as indicated were transiently transfected with either vector-only control (pCR3.1) or full-length wild-type K15 (pCR3.1K15). The predominant protein sizes for K15 in all three cell types were 23 and 35 kDa. (D) Primary HMVEC were infected with lentivirus expressing K15 (Lenti-K15) and virus not expressing K15 (Lenti-empty [negative control]). Equal viral loads were determined by p24 ELISA (data not shown). Cell extracts were made 48 h postinfection and Western blotted with anti-K15 MAb (3B5/D7), developed using ECL, and exposed to autoradiographic film for the indicated times. The blot was subsequently stained with Coomassie blue to demonstrate equal protein input. (E) Infection of HMVEC with Lenti-empty, Lenti-K15, and Lenti-K15ΔAUG. At 48 h postinfection cell extracts were Western blotted with anti-K15 MAb. (F) Internal signal sequence cleavage site prediction for K15. For the prediction of cleavage sites within K15 we employed the SignalP V1.1 prediction program (see Materials and Methods). The arrows indicate amino acid positions of predicted cleavage sites and the molecular mass of the cleaved C-terminal protein product. Red arrows indicate those sites that most closely match the cleavage consensus sequence. (G) Endogenous expression of K15. Raji and BC3 (PEL) cell extracts equivalent to 10⁵ cells were also subjected to SDS-15% PAGE and blotted for K15 with the 3B5/D7 MAb as described above. Positions of molecular mass markers are indicated at left. The predominant protein detected was 23 kDa. (H) K15 protein kinetics of expression following lytic cycle induction. BC3 and DG75 cells (6.0 × 10⁵) were treated with TPA (20 ng/ml). At the time points indicated 10⁵ cells were harvested and lysed and subjected to SDS-PAGE and anti-K15 blotting as described above.

FIG. 1.

K15 expression analysis. (A) Sequence comparison of K15 C-terminal domain for the P and M forms. The predicted C-terminal cytoplasm domains of both the P and M forms of K15 are shown. Grey shading indicates identical amino acids. Dashed lines indicate amino acids deleted for mutational analysis (see below). Predicted SH2-B and SH3-B domains are indicated. The boxed region indicates putative mitochondrial targeting sequence. (B) In vitro TNT in the presence of [³⁵S]methionine and pCR3.1 (vector only) or pCR3.1K15 as indicated. The full-length eight-exon form of K15 produces a protein of approximately 50 kDa. (C) HeLa and the endothelial cell line IE7 as indicated were transiently transfected with either vector-only control (pCR3.1) or full-length wild-type K15 (pCR3.1K15). The predominant protein sizes for K15 in all three cell types were 23 and 35 kDa. (D) Primary HMVEC were infected with lentivirus expressing K15 (Lenti-K15) and virus not expressing K15 (Lenti-empty [negative control]). Equal viral loads were determined by p24 ELISA (data not shown). Cell extracts were made 48 h postinfection and Western blotted with anti-K15 MAb (3B5/D7), developed using ECL, and exposed to autoradiographic film for the indicated times. The blot was subsequently stained with Coomassie blue to demonstrate equal protein input. (E) Infection of HMVEC with Lenti-empty, Lenti-K15, and Lenti-K15ΔAUG. At 48 h postinfection cell extracts were Western blotted with anti-K15 MAb. (F) Internal signal sequence cleavage site prediction for K15. For the prediction of cleavage sites within K15 we employed the SignalP V1.1 prediction program (see Materials and Methods). The arrows indicate amino acid positions of predicted cleavage sites and the molecular mass of the cleaved C-terminal protein product. Red arrows indicate those sites that most closely match the cleavage consensus sequence. (G) Endogenous expression of K15. Raji and BC3 (PEL) cell extracts equivalent to 10⁵ cells were also subjected to SDS-15% PAGE and blotted for K15 with the 3B5/D7 MAb as described above. Positions of molecular mass markers are indicated at left. The predominant protein detected was 23 kDa. (H) K15 protein kinetics of expression following lytic cycle induction. BC3 and DG75 cells (6.0 × 10⁵) were treated with TPA (20 ng/ml). At the time points indicated 10⁵ cells were harvested and lysed and subjected to SDS-PAGE and anti-K15 blotting as described above.

FIG. 2.

Immunohistochemical staining for K15. HEK 293 cells were transfected with either pCR3.1 vector-only control (A) or the full-length eight-exon P form of K15 (B). (C) High-power (×100) magnification of panel B. Cells were then fixed and stained with the anti-K15 MAb 3B5/D7 (1:200). (E to G) K15 staining of KSHV-positive PEL cell lines BC3 (F) and JSC-1 (G) and negative control cell line Raji (E). (D) LNA-1 staining of plasmablasts in KSHV-associated MCD with characteristic nuclear stippling. (H) K15 staining of an adjacent section shows that all the plasmablasts also express K15 with characteristic cytoplasmic staining.

FIG. 3.

Localization of K15 to the ER. (A) HeLa cells grown on coverslips were transfected with expression vectors for K15. Cells were stained 48 h posttransfection as described in Materials and Methods. K15 was detected with the anti-K15 MAb (3B5/D7) followed by a goat anti-mouse IgG AMCA conjugate (DAKO). ER was detected with the specific dye DiOC₆(3). (B) Staining of K15 and ER in transfected HeLa cells (as above) shows a cell that has undergone mitosis. This panel includes propidium iodide staining for the nucleus.

FIG. 4.

HAX-1 and K15 interaction. (A) Diagrammatic summary of HAX-1 cDNA clones obtained from yeast two-hybrid screens. Wild-type HAX-1 amino acid sequence and the regions of overlap with each of the HAX-1 clones (C2, 38, 57, and 83) obtained from the yeast two-hybrid screen are shown. Bcl-2 homology domains 1 and 2 (BH1 and -2) are indicated. W.T., wild type; TMD, transmembrane domain. (B) Alignment of the BH1 domain of Bcl-2 family members with human and murine HAX-1 (hHAX-1 and mHAX-1, respectively). Dark-grey-shaded areas indicate amino acid identity in all entries. Light-grey-shaded areas indicate amino acid identity in seven or more entries. (C) The specificity of the HAX-1-K15-CT interaction in the yeast two-hybrid assay was tested by combination of these two proteins with positive and negative controls and then assayed by prototrophy for histidine (His) and adenine (Ade). The indicated GAL4BD fusion (BD), GAL4AD (AD), or vector-only (VO) plasmids were cotransformed into yeast strain PJ69-4a. Growth on medium lacking His and Ade (−His/−Ade) is indicative of specific interactions. P53 and SV40-LT were included as a positive control. (D) HAX-1 binds GSH-K15-CT in vitro. BC3 cell extracts, which contain endogenous HAX-1, were incubated with GST or GST-K15-CT attached to glutathione-Sepharose beads as described in Materials and Methods. Specifically bound HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb. (E) The yeast two-hybrid system was used to determine the region in K15-CT necessary for the interaction with HAX-1. The interaction was assayed for prototrophy for His and Ade and by a liquid β-galactosidase assay (β-galactosidase activity obtained with the wild-type GAL4-K15-CT fusion protein was set at 100%). GAL4BD-HAX-1 (C38) was transformed with the indicated GAL4AD-K15-CT mutants and replica plated onto selection media with His and Ade (+His/+Ade) or lacking His and Ade (−His/−Ade). As an extra control the C-terminal cytoplasmic domain of a related membrane protein (EBV LMP-1) was included. Growth on medium lacking His and Ade is indicative of specific interactions. (F) The GST pull down assay was used to confirm the yeast two-hybrid data. BC3 cell extracts, which contain endogenous HAX-1, were incubated with GST, GSH-K15-CT, or the indicated GST-K15-CT mutants attached to glutathione-Sepharose beads as described in Materials and Methods. Specifically bound HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb. (G) Colocalization of K15 and HAX-1. HeLa cells grown on coverslips were cotransfected with expression vectors for K15 and GSH-mHAX-1. Cells were stained 48 h posttransfection as described in Materials and Methods. K15 was detected with the anti-K15 MAb (3B5/D7) followed by a goat anti-mouse IgG fluorescein isothiocyanate conjugate (DAKO). HAX-1 was detected with the rabbit anti-GST polyclonal followed by a donkey anti-rabbit IgG R-phycoerythrin conjugate. The cell nucleus was stained with Hoechst dye. Colocalization of the two proteins is seen upon overlay of the images. (H) Subcellular fractionation of HeLa cells transfected with K15 or expressing endogenous HAX-1 was analyzed. Both K15 and HAX-1 are detected in the ER and mitochondrial fractions. Bax is a cytoplasmic positive control and an ER negative control. Lane M, mitochondrial fraction; lane C, cytoplasmic fraction.

FIG. 4.

HAX-1 and K15 interaction. (A) Diagrammatic summary of HAX-1 cDNA clones obtained from yeast two-hybrid screens. Wild-type HAX-1 amino acid sequence and the regions of overlap with each of the HAX-1 clones (C2, 38, 57, and 83) obtained from the yeast two-hybrid screen are shown. Bcl-2 homology domains 1 and 2 (BH1 and -2) are indicated. W.T., wild type; TMD, transmembrane domain. (B) Alignment of the BH1 domain of Bcl-2 family members with human and murine HAX-1 (hHAX-1 and mHAX-1, respectively). Dark-grey-shaded areas indicate amino acid identity in all entries. Light-grey-shaded areas indicate amino acid identity in seven or more entries. (C) The specificity of the HAX-1-K15-CT interaction in the yeast two-hybrid assay was tested by combination of these two proteins with positive and negative controls and then assayed by prototrophy for histidine (His) and adenine (Ade). The indicated GAL4BD fusion (BD), GAL4AD (AD), or vector-only (VO) plasmids were cotransformed into yeast strain PJ69-4a. Growth on medium lacking His and Ade (−His/−Ade) is indicative of specific interactions. P53 and SV40-LT were included as a positive control. (D) HAX-1 binds GSH-K15-CT in vitro. BC3 cell extracts, which contain endogenous HAX-1, were incubated with GST or GST-K15-CT attached to glutathione-Sepharose beads as described in Materials and Methods. Specifically bound HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb. (E) The yeast two-hybrid system was used to determine the region in K15-CT necessary for the interaction with HAX-1. The interaction was assayed for prototrophy for His and Ade and by a liquid β-galactosidase assay (β-galactosidase activity obtained with the wild-type GAL4-K15-CT fusion protein was set at 100%). GAL4BD-HAX-1 (C38) was transformed with the indicated GAL4AD-K15-CT mutants and replica plated onto selection media with His and Ade (+His/+Ade) or lacking His and Ade (−His/−Ade). As an extra control the C-terminal cytoplasmic domain of a related membrane protein (EBV LMP-1) was included. Growth on medium lacking His and Ade is indicative of specific interactions. (F) The GST pull down assay was used to confirm the yeast two-hybrid data. BC3 cell extracts, which contain endogenous HAX-1, were incubated with GST, GSH-K15-CT, or the indicated GST-K15-CT mutants attached to glutathione-Sepharose beads as described in Materials and Methods. Specifically bound HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb. (G) Colocalization of K15 and HAX-1. HeLa cells grown on coverslips were cotransfected with expression vectors for K15 and GSH-mHAX-1. Cells were stained 48 h posttransfection as described in Materials and Methods. K15 was detected with the anti-K15 MAb (3B5/D7) followed by a goat anti-mouse IgG fluorescein isothiocyanate conjugate (DAKO). HAX-1 was detected with the rabbit anti-GST polyclonal followed by a donkey anti-rabbit IgG R-phycoerythrin conjugate. The cell nucleus was stained with Hoechst dye. Colocalization of the two proteins is seen upon overlay of the images. (H) Subcellular fractionation of HeLa cells transfected with K15 or expressing endogenous HAX-1 was analyzed. Both K15 and HAX-1 are detected in the ER and mitochondrial fractions. Bax is a cytoplasmic positive control and an ER negative control. Lane M, mitochondrial fraction; lane C, cytoplasmic fraction.

FIG. 4.

HAX-1 and K15 interaction. (A) Diagrammatic summary of HAX-1 cDNA clones obtained from yeast two-hybrid screens. Wild-type HAX-1 amino acid sequence and the regions of overlap with each of the HAX-1 clones (C2, 38, 57, and 83) obtained from the yeast two-hybrid screen are shown. Bcl-2 homology domains 1 and 2 (BH1 and -2) are indicated. W.T., wild type; TMD, transmembrane domain. (B) Alignment of the BH1 domain of Bcl-2 family members with human and murine HAX-1 (hHAX-1 and mHAX-1, respectively). Dark-grey-shaded areas indicate amino acid identity in all entries. Light-grey-shaded areas indicate amino acid identity in seven or more entries. (C) The specificity of the HAX-1-K15-CT interaction in the yeast two-hybrid assay was tested by combination of these two proteins with positive and negative controls and then assayed by prototrophy for histidine (His) and adenine (Ade). The indicated GAL4BD fusion (BD), GAL4AD (AD), or vector-only (VO) plasmids were cotransformed into yeast strain PJ69-4a. Growth on medium lacking His and Ade (−His/−Ade) is indicative of specific interactions. P53 and SV40-LT were included as a positive control. (D) HAX-1 binds GSH-K15-CT in vitro. BC3 cell extracts, which contain endogenous HAX-1, were incubated with GST or GST-K15-CT attached to glutathione-Sepharose beads as described in Materials and Methods. Specifically bound HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb. (E) The yeast two-hybrid system was used to determine the region in K15-CT necessary for the interaction with HAX-1. The interaction was assayed for prototrophy for His and Ade and by a liquid β-galactosidase assay (β-galactosidase activity obtained with the wild-type GAL4-K15-CT fusion protein was set at 100%). GAL4BD-HAX-1 (C38) was transformed with the indicated GAL4AD-K15-CT mutants and replica plated onto selection media with His and Ade (+His/+Ade) or lacking His and Ade (−His/−Ade). As an extra control the C-terminal cytoplasmic domain of a related membrane protein (EBV LMP-1) was included. Growth on medium lacking His and Ade is indicative of specific interactions. (F) The GST pull down assay was used to confirm the yeast two-hybrid data. BC3 cell extracts, which contain endogenous HAX-1, were incubated with GST, GSH-K15-CT, or the indicated GST-K15-CT mutants attached to glutathione-Sepharose beads as described in Materials and Methods. Specifically bound HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb. (G) Colocalization of K15 and HAX-1. HeLa cells grown on coverslips were cotransfected with expression vectors for K15 and GSH-mHAX-1. Cells were stained 48 h posttransfection as described in Materials and Methods. K15 was detected with the anti-K15 MAb (3B5/D7) followed by a goat anti-mouse IgG fluorescein isothiocyanate conjugate (DAKO). HAX-1 was detected with the rabbit anti-GST polyclonal followed by a donkey anti-rabbit IgG R-phycoerythrin conjugate. The cell nucleus was stained with Hoechst dye. Colocalization of the two proteins is seen upon overlay of the images. (H) Subcellular fractionation of HeLa cells transfected with K15 or expressing endogenous HAX-1 was analyzed. Both K15 and HAX-1 are detected in the ER and mitochondrial fractions. Bax is a cytoplasmic positive control and an ER negative control. Lane M, mitochondrial fraction; lane C, cytoplasmic fraction.

FIG. 4.

HAX-1 and K15 interaction. (A) Diagrammatic summary of HAX-1 cDNA clones obtained from yeast two-hybrid screens. Wild-type HAX-1 amino acid sequence and the regions of overlap with each of the HAX-1 clones (C2, 38, 57, and 83) obtained from the yeast two-hybrid screen are shown. Bcl-2 homology domains 1 and 2 (BH1 and -2) are indicated. W.T., wild type; TMD, transmembrane domain. (B) Alignment of the BH1 domain of Bcl-2 family members with human and murine HAX-1 (hHAX-1 and mHAX-1, respectively). Dark-grey-shaded areas indicate amino acid identity in all entries. Light-grey-shaded areas indicate amino acid identity in seven or more entries. (C) The specificity of the HAX-1-K15-CT interaction in the yeast two-hybrid assay was tested by combination of these two proteins with positive and negative controls and then assayed by prototrophy for histidine (His) and adenine (Ade). The indicated GAL4BD fusion (BD), GAL4AD (AD), or vector-only (VO) plasmids were cotransformed into yeast strain PJ69-4a. Growth on medium lacking His and Ade (−His/−Ade) is indicative of specific interactions. P53 and SV40-LT were included as a positive control. (D) HAX-1 binds GSH-K15-CT in vitro. BC3 cell extracts, which contain endogenous HAX-1, were incubated with GST or GST-K15-CT attached to glutathione-Sepharose beads as described in Materials and Methods. Specifically bound HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb. (E) The yeast two-hybrid system was used to determine the region in K15-CT necessary for the interaction with HAX-1. The interaction was assayed for prototrophy for His and Ade and by a liquid β-galactosidase assay (β-galactosidase activity obtained with the wild-type GAL4-K15-CT fusion protein was set at 100%). GAL4BD-HAX-1 (C38) was transformed with the indicated GAL4AD-K15-CT mutants and replica plated onto selection media with His and Ade (+His/+Ade) or lacking His and Ade (−His/−Ade). As an extra control the C-terminal cytoplasmic domain of a related membrane protein (EBV LMP-1) was included. Growth on medium lacking His and Ade is indicative of specific interactions. (F) The GST pull down assay was used to confirm the yeast two-hybrid data. BC3 cell extracts, which contain endogenous HAX-1, were incubated with GST, GSH-K15-CT, or the indicated GST-K15-CT mutants attached to glutathione-Sepharose beads as described in Materials and Methods. Specifically bound HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb. (G) Colocalization of K15 and HAX-1. HeLa cells grown on coverslips were cotransfected with expression vectors for K15 and GSH-mHAX-1. Cells were stained 48 h posttransfection as described in Materials and Methods. K15 was detected with the anti-K15 MAb (3B5/D7) followed by a goat anti-mouse IgG fluorescein isothiocyanate conjugate (DAKO). HAX-1 was detected with the rabbit anti-GST polyclonal followed by a donkey anti-rabbit IgG R-phycoerythrin conjugate. The cell nucleus was stained with Hoechst dye. Colocalization of the two proteins is seen upon overlay of the images. (H) Subcellular fractionation of HeLa cells transfected with K15 or expressing endogenous HAX-1 was analyzed. Both K15 and HAX-1 are detected in the ER and mitochondrial fractions. Bax is a cytoplasmic positive control and an ER negative control. Lane M, mitochondrial fraction; lane C, cytoplasmic fraction.

FIG. 4.

HAX-1 and K15 interaction. (A) Diagrammatic summary of HAX-1 cDNA clones obtained from yeast two-hybrid screens. Wild-type HAX-1 amino acid sequence and the regions of overlap with each of the HAX-1 clones (C2, 38, 57, and 83) obtained from the yeast two-hybrid screen are shown. Bcl-2 homology domains 1 and 2 (BH1 and -2) are indicated. W.T., wild type; TMD, transmembrane domain. (B) Alignment of the BH1 domain of Bcl-2 family members with human and murine HAX-1 (hHAX-1 and mHAX-1, respectively). Dark-grey-shaded areas indicate amino acid identity in all entries. Light-grey-shaded areas indicate amino acid identity in seven or more entries. (C) The specificity of the HAX-1-K15-CT interaction in the yeast two-hybrid assay was tested by combination of these two proteins with positive and negative controls and then assayed by prototrophy for histidine (His) and adenine (Ade). The indicated GAL4BD fusion (BD), GAL4AD (AD), or vector-only (VO) plasmids were cotransformed into yeast strain PJ69-4a. Growth on medium lacking His and Ade (−His/−Ade) is indicative of specific interactions. P53 and SV40-LT were included as a positive control. (D) HAX-1 binds GSH-K15-CT in vitro. BC3 cell extracts, which contain endogenous HAX-1, were incubated with GST or GST-K15-CT attached to glutathione-Sepharose beads as described in Materials and Methods. Specifically bound HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb. (E) The yeast two-hybrid system was used to determine the region in K15-CT necessary for the interaction with HAX-1. The interaction was assayed for prototrophy for His and Ade and by a liquid β-galactosidase assay (β-galactosidase activity obtained with the wild-type GAL4-K15-CT fusion protein was set at 100%). GAL4BD-HAX-1 (C38) was transformed with the indicated GAL4AD-K15-CT mutants and replica plated onto selection media with His and Ade (+His/+Ade) or lacking His and Ade (−His/−Ade). As an extra control the C-terminal cytoplasmic domain of a related membrane protein (EBV LMP-1) was included. Growth on medium lacking His and Ade is indicative of specific interactions. (F) The GST pull down assay was used to confirm the yeast two-hybrid data. BC3 cell extracts, which contain endogenous HAX-1, were incubated with GST, GSH-K15-CT, or the indicated GST-K15-CT mutants attached to glutathione-Sepharose beads as described in Materials and Methods. Specifically bound HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb. (G) Colocalization of K15 and HAX-1. HeLa cells grown on coverslips were cotransfected with expression vectors for K15 and GSH-mHAX-1. Cells were stained 48 h posttransfection as described in Materials and Methods. K15 was detected with the anti-K15 MAb (3B5/D7) followed by a goat anti-mouse IgG fluorescein isothiocyanate conjugate (DAKO). HAX-1 was detected with the rabbit anti-GST polyclonal followed by a donkey anti-rabbit IgG R-phycoerythrin conjugate. The cell nucleus was stained with Hoechst dye. Colocalization of the two proteins is seen upon overlay of the images. (H) Subcellular fractionation of HeLa cells transfected with K15 or expressing endogenous HAX-1 was analyzed. Both K15 and HAX-1 are detected in the ER and mitochondrial fractions. Bax is a cytoplasmic positive control and an ER negative control. Lane M, mitochondrial fraction; lane C, cytoplasmic fraction.

FIG. 4.

HAX-1 and K15 interaction. (A) Diagrammatic summary of HAX-1 cDNA clones obtained from yeast two-hybrid screens. Wild-type HAX-1 amino acid sequence and the regions of overlap with each of the HAX-1 clones (C2, 38, 57, and 83) obtained from the yeast two-hybrid screen are shown. Bcl-2 homology domains 1 and 2 (BH1 and -2) are indicated. W.T., wild type; TMD, transmembrane domain. (B) Alignment of the BH1 domain of Bcl-2 family members with human and murine HAX-1 (hHAX-1 and mHAX-1, respectively). Dark-grey-shaded areas indicate amino acid identity in all entries. Light-grey-shaded areas indicate amino acid identity in seven or more entries. (C) The specificity of the HAX-1-K15-CT interaction in the yeast two-hybrid assay was tested by combination of these two proteins with positive and negative controls and then assayed by prototrophy for histidine (His) and adenine (Ade). The indicated GAL4BD fusion (BD), GAL4AD (AD), or vector-only (VO) plasmids were cotransformed into yeast strain PJ69-4a. Growth on medium lacking His and Ade (−His/−Ade) is indicative of specific interactions. P53 and SV40-LT were included as a positive control. (D) HAX-1 binds GSH-K15-CT in vitro. BC3 cell extracts, which contain endogenous HAX-1, were incubated with GST or GST-K15-CT attached to glutathione-Sepharose beads as described in Materials and Methods. Specifically bound HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb. (E) The yeast two-hybrid system was used to determine the region in K15-CT necessary for the interaction with HAX-1. The interaction was assayed for prototrophy for His and Ade and by a liquid β-galactosidase assay (β-galactosidase activity obtained with the wild-type GAL4-K15-CT fusion protein was set at 100%). GAL4BD-HAX-1 (C38) was transformed with the indicated GAL4AD-K15-CT mutants and replica plated onto selection media with His and Ade (+His/+Ade) or lacking His and Ade (−His/−Ade). As an extra control the C-terminal cytoplasmic domain of a related membrane protein (EBV LMP-1) was included. Growth on medium lacking His and Ade is indicative of specific interactions. (F) The GST pull down assay was used to confirm the yeast two-hybrid data. BC3 cell extracts, which contain endogenous HAX-1, were incubated with GST, GSH-K15-CT, or the indicated GST-K15-CT mutants attached to glutathione-Sepharose beads as described in Materials and Methods. Specifically bound HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb. (G) Colocalization of K15 and HAX-1. HeLa cells grown on coverslips were cotransfected with expression vectors for K15 and GSH-mHAX-1. Cells were stained 48 h posttransfection as described in Materials and Methods. K15 was detected with the anti-K15 MAb (3B5/D7) followed by a goat anti-mouse IgG fluorescein isothiocyanate conjugate (DAKO). HAX-1 was detected with the rabbit anti-GST polyclonal followed by a donkey anti-rabbit IgG R-phycoerythrin conjugate. The cell nucleus was stained with Hoechst dye. Colocalization of the two proteins is seen upon overlay of the images. (H) Subcellular fractionation of HeLa cells transfected with K15 or expressing endogenous HAX-1 was analyzed. Both K15 and HAX-1 are detected in the ER and mitochondrial fractions. Bax is a cytoplasmic positive control and an ER negative control. Lane M, mitochondrial fraction; lane C, cytoplasmic fraction.

FIG. 4.

HAX-1 and K15 interaction. (A) Diagrammatic summary of HAX-1 cDNA clones obtained from yeast two-hybrid screens. Wild-type HAX-1 amino acid sequence and the regions of overlap with each of the HAX-1 clones (C2, 38, 57, and 83) obtained from the yeast two-hybrid screen are shown. Bcl-2 homology domains 1 and 2 (BH1 and -2) are indicated. W.T., wild type; TMD, transmembrane domain. (B) Alignment of the BH1 domain of Bcl-2 family members with human and murine HAX-1 (hHAX-1 and mHAX-1, respectively). Dark-grey-shaded areas indicate amino acid identity in all entries. Light-grey-shaded areas indicate amino acid identity in seven or more entries. (C) The specificity of the HAX-1-K15-CT interaction in the yeast two-hybrid assay was tested by combination of these two proteins with positive and negative controls and then assayed by prototrophy for histidine (His) and adenine (Ade). The indicated GAL4BD fusion (BD), GAL4AD (AD), or vector-only (VO) plasmids were cotransformed into yeast strain PJ69-4a. Growth on medium lacking His and Ade (−His/−Ade) is indicative of specific interactions. P53 and SV40-LT were included as a positive control. (D) HAX-1 binds GSH-K15-CT in vitro. BC3 cell extracts, which contain endogenous HAX-1, were incubated with GST or GST-K15-CT attached to glutathione-Sepharose beads as described in Materials and Methods. Specifically bound HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb. (E) The yeast two-hybrid system was used to determine the region in K15-CT necessary for the interaction with HAX-1. The interaction was assayed for prototrophy for His and Ade and by a liquid β-galactosidase assay (β-galactosidase activity obtained with the wild-type GAL4-K15-CT fusion protein was set at 100%). GAL4BD-HAX-1 (C38) was transformed with the indicated GAL4AD-K15-CT mutants and replica plated onto selection media with His and Ade (+His/+Ade) or lacking His and Ade (−His/−Ade). As an extra control the C-terminal cytoplasmic domain of a related membrane protein (EBV LMP-1) was included. Growth on medium lacking His and Ade is indicative of specific interactions. (F) The GST pull down assay was used to confirm the yeast two-hybrid data. BC3 cell extracts, which contain endogenous HAX-1, were incubated with GST, GSH-K15-CT, or the indicated GST-K15-CT mutants attached to glutathione-Sepharose beads as described in Materials and Methods. Specifically bound HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb. (G) Colocalization of K15 and HAX-1. HeLa cells grown on coverslips were cotransfected with expression vectors for K15 and GSH-mHAX-1. Cells were stained 48 h posttransfection as described in Materials and Methods. K15 was detected with the anti-K15 MAb (3B5/D7) followed by a goat anti-mouse IgG fluorescein isothiocyanate conjugate (DAKO). HAX-1 was detected with the rabbit anti-GST polyclonal followed by a donkey anti-rabbit IgG R-phycoerythrin conjugate. The cell nucleus was stained with Hoechst dye. Colocalization of the two proteins is seen upon overlay of the images. (H) Subcellular fractionation of HeLa cells transfected with K15 or expressing endogenous HAX-1 was analyzed. Both K15 and HAX-1 are detected in the ER and mitochondrial fractions. Bax is a cytoplasmic positive control and an ER negative control. Lane M, mitochondrial fraction; lane C, cytoplasmic fraction.

FIG. 5.

K15 and HAX-1 coprecipitate in vivo. HeLa cells were transiently transfected with the indicated combination of K15, GSH-HAX-1, or control expression vectors as indicated (A). Cell extracts were prepared as described in Materials and Methods, and then tested for expression of HAX-1 (A) and K15 (B) by Western blotting. (C) The same extracts were incubated with glutathione-Sepharose beads as described in Materials and Methods. Specifically bound GSH-HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-K15 MAb (3B5/D7). (D) Endogenous K15 and HAX-1 interact in vivo. Five micrograms of anti-K15 MAb and isotype control were incubated with 500 μl of BC3 cell extracts (prepared as described in Materials and Methods) at 4°C for 1 h with rotation. Fifty microliters of protein G (50:50 slurry) was then added and incubated for 2 h as before. Beads were washed extensively in RIPA buffer and then analyzed by Western blotting with anti-HAX-1 MAb. Input controls for K15 and HAX-1 expression in BC3 extracts have been shown (Fig. 1G and 4D, respectively). (E and F) HAX-1 forms homodimers in vivo. HeLa cells were transiently transfected with the indicated GST only or increasing concentrations of GSH-HAX-1 expression plasmid. Cell extracts were prepared as described in Materials and Methods and then checked for expression with the anti-HAX-1 MAb (H65220; Transduction Labs, Lexington, Ky.), which recognizes both human and mouse HAX-1 (E). Extracts were then incubated with glutathione-Sepharose beads as described in Materials and Methods. Specifically bound GSH-HAX-1 protein was eluted off beads with SDS-PAGE sample buffer and subjected to SDS-PAGE and then blotted with anti-HAX-1 MAb (F). Endogenous HAX-1 (35 kDa) coprecipitates in the presence of GSH-HAX-1 and not GST alone. An asterisk indicates the GSH-HAX-1 degradation product.

FIG. 6.

HAX-1 inhibits Bax-induced apoptosis. HeLa cells were transfected with the expression plasmids indicated in the top right portion of each panel. At 48 h posttransfection all cells, both adherent and in suspension, were pooled, washed, and stained with the mitochondrial membrane specific dye MitoTracker Red CMXRosamine (Molecular Probes, Leiden, The Netherlands) and analyzed by fluorescence-activated cell sorting. The percentage of apoptotic cells is indicated. (A to D) Fluorescence-activated cell sorter analysis of representative samples for each assay. (E) Summary of percent induced apoptosis for each of the indicated transfections. These data represent the average and standard deviations (error bars) of three independent experiments.