Introduction

Prostate carcinoma (PCA) is the most frequently diagnosed malignancy of men in the United States. The molecular and genetic events underlying neoplastic transformation and progression in the prostate are not well understood. PCA at advanced stages, like other aggressive tumors, has acquired the ability not only to proliferate but also to resist apoptosis (Tenniswood, 1997; Tang and Porter, 1997). We have been interested in the mechanism of PCA resistance to apoptosis. Among the several major signal transduction pathways, the PI3-kinase pathway has been strongly implicated in cell survival and resistance to apoptosis (Franke and Kaplan, 1997; Alessi and Cohen, 1998; Ptasznik et al., 1997). How PI3-kinase affects cell survival has been a subject of intensive investigation. It was shown that the metabolic products of PI3-kinase, phosphatidyl inositol phosphates, bind a protein modular structure, called the pleckstrin-homology (PH) domain (Franke and Kaplan, 1997). Several PH-domain-containing proteins are found to be effectors for PI3-kinase (Franke et al., 1995; Cross et al., 1995; August et al., 1997; Li et al., 1997; Qiu et al., 1998). Of particular relevance is Akt/PKB, a serine/threonine kinase with an N-terminal PH domain, which was recently shown to phosphorylate and inactivate Bad, a protein involved in the apoptosis pathway. This offers one mechanism whereby PI3-kinase can reduce apoptosis (Datta et al., 1997; del Peso et al., 1997; Dudek et al., 1997). Here we describe the action of a PH-domain containing tyrosine kinase, Etk/Bmx, in the protection of PCA from apoptotic death.

Etk/Bmx is a newly discovered tyrosine kinase, commonly expressed in prostate epithelial and carcinoma cells. Our previous studies showed that this kinase is activated by PI3-kinase. Etk belongs to the Btk and Itk family of kinases. Btk (Bruton's tyrosine kinase) and Itk have been studied extensively in recent years and shown to play significant roles in B and T cell development by channeling B and T-cell receptor-mediated signals and protecting these cells from apoptosis. Btk was discovered as a gene whose germline mutation in the PH domain leads to XLA, X-linked agammaglobulinemia. Mice with a targeted deletion of this locus develop xid, an acquired immunodeficiency disease. Overexpression of a constitutively active Btk induces morphological transformation and protects the transfected cells from apoptosis (Li et al., 1995). Likewise, Itk is highly expressed in T-cells and involved in T-cell development. Interestingly, Etk is not expressed in either B or T cells, but rather in epithelial cells, endothelial cells, and monocytes/macrophages (Qiu et al., 1998; Tamagnone et al., 1994; Kaukonen et al., 1996; Weil et al., 1997). It is tempting to think that Etk plays a similar role in the growth, development and anti-apoptosis of epithelial cells. In this report we present data that suggest that Etk is involved in protection of prostate cancer cells from apoptosis.

To study the anti-apoptotic role of Etk, we used two approaches to induce apoptosis in LNCaP, an androgen-responsive PCA cell line. The first approach is PDT (photodynamic therapy), a novel cancer treatment that utilizes a photosensitizing drug, visible light, and oxygen to produce reactive oxygen species, including singlet oxygen, in cellular targets; the resultant oxidative stress leads to cell death and tumor ablation (Moan and Berg, 1992; Henderson and Dougherty, 1992; Hasan and Parrish, 1997). The initial targets of PDT damage are thought to reside in membranes. Although PDT causes a direct activation of the final stages of apoptosis (Kessel et al., 1997; Kessel and Luo, 1999; Oleinick and Evans, 1998), the sensitivity of cells to PDT can vary between closely related cell lines as a function of the expression of particular genes. Thus, HL60 cells expressing wild-type p53 were more sensitive to cell killing by PDT than were HL60 cells in which p53 genes were deleted or mutant, although all lines underwent a rapid apoptotic response (Fisher et al., 1997). Chinese hamster ovary cells expressing a human Bcl-2 gene were partially resistant to apoptosis and to overall cell death after PDT sensitized by the phthalocyanine Pc 4 (He et al., 1996). PDT is also an efficient activator of a variety of signal transduction pathways (Agarwal et al., 1993; Ryter and Gomer, 1993; Xue et al., 1997; Tao et al., 1996; Separovic et al., 1997), although the functions, if any, of these pathways in mediating or protecting from cell death have yet to be fully defined.

In addition to PDT, thapsigargin (TG) is frequently used to induce apoptosis in cancer cells, including prostate carcinoma cell lines (McConkey et al., 1996). The malignant transformation of prostate cancer cells has been found to correlate with their resistance to TG. TG inhibits an ATP-gated calcium pump controlling the influx of calcium from the cytosol to the endoplasmic reticulum (ER), thereby depleting the calcium pool of the ER and activating the apoptosis pathway (He et al., 1997). This process is accompanied by an induction of GRP78 expression and caspase activation, and cells can be rescued by overexpression of Bcl2 (Qi et al., 1997; Liu et al., 1997). Thus, although PDT and TG take different routes to initiate the apoptotic process, there is a converging of pathways in the downstream events of apoptosis.

In this report, we show that LNCaP cells are sensitive to apoptosis induced by either PDT or TG, as evidenced by DNA fragmentation, cleavage of poly(ADP-ribose) polymerase (PARP), and a cell viability test. We further demonstrate that PI3-kinase and its downstream substrate Etk are involved in the protection of LNCaP cells from apoptosis induced by these agents. Overexpression of wild-type Etk renders the cell more resistant to apoptosis, whereas overexpression of a kinase-dead, dominant-negative Etk sensitizes cells toward apoptosis. Our finding defines a new PI3-kinase pathway involved in survival and possibly tumor resistance found in malignant PCA.

Results

Development of LNCaP cells overexpressing wild-type and mutant Etk

We recently reported that Etk is an effector of PI3-kinase and plays an important role in IL6-induced signaling in human prostate cancer cells (Qiu et al., 1998). PI3-kinase has been shown to be important for the survival of a number of different cell types (Franke and Kaplan, 1997). The purpose of this study was to investigate whether Etk as well as PI3-kinase are involved in protection of prostate cancer cells from apoptosis. To this end, we developed two LNCaP cell lines overexpressing, respectively, a wild-type Etk (LNCaP-Etkwt) and a dominant-negative mutant of Etk (LNCaP-EtkDN). Both molecules were tagged with a T7-epitope at the N-termini. The dominant-negative (DN) Etk carries a substitution of K444 with R in the catalytic domain which abolishes the kinase activity and in addition, a second mutation in the PH domain which, by analogy with Btk, increases the affinity toward lipid. In transient transfections, EtkDN has proven highly effective in reducing the endogenous Etk kinase activity, presumably by sequestering the activating lipid moiety (Qiu et al., 1998). We show here that EtkDN is equally effective when stably transfected into the LNCaP cell line. The result is illustrated in Figure 1a, using an in vitro kinase assay of immunoprecipitated Etk from the various cell lines. The parental cells (lane LNCaP) display a basal level of endogenous Etk activity in the presence of serum. This activity is completely absent in EtkDN-expressing cells (lane EtkDN), whereas the wild-type Etk overexpressing cells (lane Etkwt) exhibit a high level of kinase activity. Western-blot analysis (Figure 1b) with anti-T7 antibody, which detects ectopically tagged Etk, reveals that similar amounts of wild-type and dominant-negative Etk were expressed in the two cell lines. These results set the stage for the analysis of the anti-apoptotic effect of Etk.

Figure 1
figure 1

Etk activity in the human prostate cancer cell line LNCaP and its derivatives Etkwt (Etk-overexpressing) and EtkDN (Etk-dominant-negative). (a) Kinase assay. The three cell lines were grown in medium with 10% serum. The cell lysates were immunoprecipitated with anti-Etk antibody, then the immunoprecipitates were assayed for kinase activity as described in Materials and methods. The arrows indicate the positions of Etk (autophosphorylated) and enolase (exogenous substrate). (b) Western blot analysis. Immunoprecipitates obtained with the anti-Etk antibody as described above were Western blotted with anti-T7 antibody

Full size image

PDT-induced apoptosis and the protection by Etk

Photodynamic treatment (PDT) is a potent inducer of apoptosis in many types of cells (Oleinick and Evans, 1998). Figure 2 shows the oligonucleosomal DNA fragmentation pattern in PDT-treated LNCaP, Etkwt and EtkDN cells. The cells were treated with 0.5 μM Pc 4 for ∼18 h, followed by red light irradiation with fluences of 15 kJ/m2 or 20 kJ/m2, then postincubated for 4 h. There was little or no DNA fragmentation in the non-treated control cells (Figure 2) or in cells treated with Pc 4 alone (not shown). After PDT, a marked increase in DNA ladders was found in samples from both LNCaP and EtkDN cells. However, the extent of DNA fragmentation 4 h after PDT was much less in the Etk-overexpressing cells. The results imply that Etk blocks apoptosis induced by PDT.

Figure 2
figure 2

PDT-induced apoptosis in LNCaP, Etkwt and EtkDN cells, as estimated by DNA fragmentation. The cells were treated with 0.5 μM Pc 4 for 18 h, irradiated with red light (15 or 20 kJ/m2), then returned to the 37°C incubator for an additional 4 h. DNA was isolated and analysed by agarose gel electrophoresis. Similar results were obtained in three additional experiments

Full size image

The execution phase of apoptosis consists of the activation of a series of cysteine-containing proteases, termed caspases, acting on substrates with specific four-amino acid sequences terminating in aspartic acid. One of the substrates for the caspases is PARP (Lazebnik et al., 1994; Kaufmann et al., 1993), which has been used fruitfully to monitor the progression of apoptosis. Cleavage of PARP at its DEVD motif results in the generation of an Mr 90 000 fragment from the Mr 116 000 native enzyme. To further characterize the block in the apoptotic pathway in Etk-overexpressing cells, the same treated cells used for Figure 2 were analysed for the extent of PARP cleavage by Western blotting with an anti-PARP antibody (Figure 3). For cells not treated with PDT, the majority of the PARP migrated at the position of intact PARP; a faint band corresponding to the cleavage product was observed, indicating a low level of constitutive cleavage of PARP. Four hours after PDT, PARP cleavage increased markedly in LNCaP and EtkDN cells. By contrast, only a slight increase in cleaved PARP was observed in PDT-treated Etkwt cells. The extent of PARP cleavage was quantified by densitometry and is shown at the bottom of Figure 3.

Figure 3
figure 3

PARP cleavage in PDT-treated LNCaP, Etkwt and EtkDN cells. Each cell line was exposed to 0.5 μM Pc 4 and 10 or 15 kJ/m2 of red light. Four hours later, the cells were collected. The total cell extracts were subjected to SDS – PAGE followed by Western blotting with anti-PARP antibody. The bottom of the figure shows the percentage of PARP cleaved after PDT treatment in three cell lines calculated as described in Materials and methods. Similar results were observed in at least three independent experiments, in which the standard deviations were⩽15% of the mean values

Full size image

Cell killing by PDT is dependent on both the light fluence and the photosensitizer concentration. For the present study, the Pc 4 dose in the medium was held constant, but it was important to ascertain whether or not the uptake of dye into the three cell lines differed as a result of the presence of Etk protein. Cultures of the three cell lines were incubated with various concentrations of Pc 4 overnight, then cell-associated dye was extracted and its level was estimated by absorption spectrophotometry. The uptake of Pc 4 increased linearly with the concentration in the range of 0.5 – 2 μM, and there was no significant difference among the three cell lines in the uptake of Pc 4 (data not shown). The result indicates that the lower level of apoptosis in Etkwt cells did not result from a reduced uptake of Pc 4, but from the function of Etk.

We next investigated the overall cell viability after PDT treatment, using the WST-1 assay, which measures the reduction of dye by intact mitochondria. Based on this assay, dose-dependent killing was found in all three cell lines. However, the Etkwt cells were considerably less sensitive to PDT than the parental LNCaP cells, and the EtkDN cells (Figure 4). This figure also shows that the cells expressing dominant-negative Etk were more sensitive to killing by PDT than were the parental cells. Taken together, these data suggest that Etk controls the sensitivity of LNCaP toward PDT-induced apoptosis.

Figure 4
figure 4

Dose-dependence of the killing of LNCaP, Etkwt and EtkDN cells by PDT. Cells were incubated in 0.5 μM Pc 4 for 18 h, then irradiated (2, 5, or 10 kJ/m2). After irradiation, cells were incubated for 24 h, and WST-1 reduction was measured. The data are the means (n=5)±s.d

Full size image

Thapsigargin-induced apoptosis and protection by Etk

In order to determine the generality of the role of Etk as an inhibitor of apoptosis, another potent inducer of apoptosis, thapsigargin (TG), which acts via the inhibition of the endoplasmic reticulum calcium ion pump, was studied in the three cell lines. Figure 5a shows the DNA fragmentation patterns at 48 h after addition of TG to the cultures. Both EtkDN and LNCaP cells are sensitive to TG-induced DNA degradation, whereas Etkwt cells are resistant. The PARP cleavage analysis showed a similar trend (Figure 5b). When the cells were incubated in 300 nM TG for 20 h, there was little or no PARP cleavage in any of the cell lines, compared to the untreated control. At 40 h, significant PARP cleavage was detected in all three cell lines with EtkDN cells displaying the highest level of cleavage (59%) and LNCaP cells the intermediate level (30%). PARP cleavage in Etkwt cells was the lowest at 10%. Similar graded differences were found in the cell viability analysis (Figure 5c). For instance, only 5% of EtkDN cells survived TG treatment for 48 h, whereas close to 50% of Etkwt cells remained viable. The parental cell line had an intermediate survival rate (10%). These data, like those described for PDT, provide strong indications that Etk activity is important in the resistance of LNCaP cells to apoptosis.

Figure 5
figure 5

Apoptosis in TG-treated LNCaP, Etkwt and EtkDN cells. (a) DNA fragmentation. Cells were incubated with or without 200 nM TG for 48 h, then DNA was isolated and analysed by gel electrophoresis. (b) PARP cleavage. Cells were incubated in the absence or presence of 300 nM thapsigargin for 20 or 40 h as indicated, then PARP cleavage was measured by immunoblotting. The percentage of PARP cleavage is shown at the bottom. The results are from a single experiment representative of three other experiments. (c) Cell viability. Cells were incubated with or without 200 nM TG for 48 h, then WST-1 reduction was measured as described in Materials and methods. Per cent viability was calculated based on the viability of the control cells as 100%. For each cell line, the results of three independent experiments are shown, and each datum represents the mean±s.d. of triplicate results within a single experiment

Full size image

PI3-kinase, Etk and anti-apoptosis

PI3-kinase has been shown to be very important for the anti-apoptotic response and survival of a number of different cell types (Franke et al., 1997). Previously, we showed that endogenous Etk is activated by PI3-kinase, which is in turn activated by growth factors present in the serum. If Etk is involved in anti-apoptosis, as suggested by the above data, we would predict that inactivation of PI3-kinase, which diminishes Etk activity, should sensitize cells toward apoptosis. One effective way of inactivating PI3-kinase is treatment with the PI3-kinase inhibitor, LY294002. Figure 6a shows Etk activity in parental LNCaP cells. In serum-starved LNCaP (lane 1), very little Etk activity was detected by the in vitro kinase assay. In the presence of serum (lane 2), Etk activity was stimulated, consistent with data shown in Figure 1. The serum-induced Etk activity was abolished upon treatment with LY294002 (lane 3), demonstrating that Etk activation requires PI3-kinase activity. PARP cleavage was then used to assess the apoptosis state of cells treated with LY294002 (Figure 6b). Cells were treated with 300 nM TG for 20 h, a time point at which little induction of PARP cleavage was detected as compared to the basal level (Figure 5b). It is evident from Figure 6b that addition of LY294002 to TG treated LNCaP cells resulted in markedly accelerated apoptosis (Figure 6b, lanes 3 and 4). Evidence that the enhanced cleavage was not due to LY294002 treatment per se is shown by the control in lane 2 where cells treated with LY294002 but without TG gave the basal cleavage pattern. The data thus implicate PI3-kinase and its effector in the protection of LNCaP cells from apoptosis.

Figure 6
figure 6

The effects of the PI3-kinase inhibitor LY294002 on Etk activity and TG-induced PARP cleavage in LNCaP cells. (a) Kinase assay. LNCaP cells were grown in the presence or absense of 10% serum. The cells grown in medium with serum were incubated with or without 20 μM LY294002 for 2 h. The kinase assay was performed as described in Materials and methods. (b) PARP cleavage. LNCaP cells were treated with or without 300 nM TG for 20 h. At the end of the incubation, 20 μM LY294002 was added to some cells for 2 h, then cells were collected and Western blotted with an anti-PARP antibody. Blotting with an antibody to actin was used as a loading control. Similar results were observed in two independent experiments

Full size image

Discussion

There is considerable evidence suggesting that the PI3-kinase pathway is involved in cell survival and tumor resistance to therapeutics. This connection, however, has not been established in prostate cancer cells. Recently, we reported the identification of Etk/Bmx, a PH-domain-containing tyrosine kinase, as a new PI3-kinase effector in prostate epithelial cells. In this study, we set out to determine the protective role of PI3-kinase and Etk in apoptosis of the prostate cancer cell line, LNCaP. We chose two very different apoptosis inducing agents, PDT and TG, both shown to be effective in inducing apoptosis in cancer cell lines. PDT works through oxidative damage to membranes, whereas TG acts as a calcium-gate inhibitor. PDT-induced apoptosis appears to involve the release of ceramide (Separovic et al., 1997) and the activation of JNK and p38/HOG (Oleinick and Evans, 1998). PDT-induced apoptosis is also characterized by a very rapid triggering of the terminal events (Agarwal et al., 1991), which may result from direct oxidative damage to the mitochondrial membrane (Kessel et al., 1997; Kessel and Luo, 1999). TG, on the other hand, depletes the calcium pool of the ER. Both pathways activate GRP78, a chaperone protein (Gomer et al., 1991; Li et al., 1993) and can be blocked by overexpression of Bcl2 (He et al., 1996; Qi et al., 1997; Liu et al., 1997). We show here that Etk is able to protect LNCaP cells from apoptosis induced by either agent. Exactly how Etk counteracts the effects of these two different agents is not clear. Recently, it was shown that Etk/Bmx is a potent activator of STAT3, a transcription factor which is known to be involved in anti-apoptosis and oncogenic transformation and with potential to augment the expression of Bcl-XL (R Jove, personal communication) and possibly Bcl2 (Fukada et al., 1996), thereby eliciting antiapoptotic effects.

Taken in the context of the work of McConkey et al. (1996), our data suggest that the PI3-kinase/Etk pathway may also be involved in the progression of prostate carcinoma. These authors isolated variants of LNCaP cells with different metastatic potentials and found that the more malignant phenotypes correlated with the higher resistance to TG-induced apoptosis. Thus, genetic or epigenetic alterations of the PI3-kinase pathway and Etk may be relevant to oncogenic transformation of prostate cancer cells. In this regard, it is interesting to note that a constitutively activated PI3-kinase catalytic subunit was recently identified as an oncoprotein of a retrovirus (Chang et al., 1997) and as the putative oncogene locus of ovarian cancers (Iwabuchi et al., 1995). The present data draw our attention to the possible involvement of the PI3-kinase pathway in prostate carcinogenesis.

Our studies also provide new insights regarding the PI3-kinase pathway and define a new component involved in anti-apoptosis. A number of proteins containing a PH domain are shown to bind phosphatidyl inositol polyphosphates and to be activated by PI3-kinase. These include rac-GEF, unconventional PKCs, Akt/PKB, PDK2 and the Btk family of tyrosine kinases (Franke et al., 1995; Cross et al., 1995; August et al., 1997; Li et al., 1997; Qiu et al., 1998; Ma et al., 1998; Falasca et al., 1998; Van Lint et al., 1998). Much of the attention has recently been focused on Akt as a protector against apoptosis. Accumulating evidence suggests that upon activation by PI3-kinase, Akt phosphorylates Bad, a dimerization partner of Bcl-XL. The phosphorylated Bad is sequestered by the chaperone protein 14-3-3, and releases Bcl-XL. The free Bcl-XL in turn dimerizes with and sequesters Bax, leaving few Bax dimers to trigger apoptosis. These studies were carried out in IL3-responsive hematopoietic cells and neuronal cells, and it is not clear whether the principle is directly applicable to prostate cancer cells. We found that Akt is expressed in LNCaP cells, but its activation by a PI3-kinase agonist is modest (data not shown). This prompted us to look for other candidate pathways by which PI3-kinase mediates its anti-apoptosis effects. We were helped by a tyrosine kinase expression profile that we have developed for prostate cancer cells. This profile identified Etk as a potential PI3-kinase effector by virtue of its PH domain. Although the present data provide strong evidence that Etk is involved in prevention of apoptosis in LNCaP cells, they do not exclude the participation of the Akt pathway. In fact, it is likely that Akt and Etk act synergistically or in parallel to promote cell survival. Experiments are in progress to study whether Etk and Akt pathways interact or converge with each other.

The induction of apoptosis by PDT occurs much more rapidly than that following TG treatment. The rapid induction of apoptosis has been observed in many types of cells with some lymphoid cell lines producing extensive DNA fragmentation in less than 1 h (Agarwal et al., 1991; Oleinick and Evans, 1998). It has been proposed that the exceptionally efficient entry of PDT-treated cells into apoptosis results from direct damage to mitochondria (Kessel et al., 1997; Kessel and Luo, 1998). The resultant immediate release of cytochrome C would then initiate the terminal irreversible phase of apoptosis (Kroemer et al., 1998; Liu et al., 1996). The present results, as well as the observed inhibition of PDT-induced apoptosis by inhibition of phospholipase C (Agarwal et al., 1993) or p38/HOG, a stress activated kinase (Oleinick et al., 1998), implicate these signal pathways in the control of apoptosis even when there is direct mitochondrial damage.

In summary, in this report, we have elucidated a new pathway involved in protection of a prostate carcinoma cell line from apoptosis. Our data show that PI3-kinase and its effector Etk function to render LNCaP cells more resistant to apoptosis induced by PDT or by TG. This finding is in resonance with the observations by others that Etk's hematopoietic analogs, Btk and Itk, are also strongly implicated in controlling the apoptosis process. Thus, PI3-kinase utilizes at least one PH-domain-containing tyrosine kinase (Etk) and one PH-domain-containing serine/threonine kinase (Akt) to mediate its anti-apoptosis function.

Materials and methods

Cell culture

LNCaP cells (ATCC) were grown in RPMI 1640 medium containing 10% fetal bovine serum (FBS). Etk-overexpressing cells (Etkwt) and Etk dominant-negative cells (EtkDN) were developed and selected as described (Qiu et al., 1998) and grown in RPMI 1640 medium with 10% FBS and 500 μg/ml of G418. The cultures were grown in a humidified atmosphere at 37°C with 5% CO2.

Photodynamic treatment

The phthalocyanine photosensitizer Pc 4, HOSiPcOSi(CH3)2(CH2)3N(CH3)2 (Oleinick et al., 1993), was provided by Drs Ying-syi Li and Malcolm E Kenney, Case Western Reserve University Department of Chemistry. It was dissolved in dimethyl formamide to 0.5 mM. Cells were loaded with Pc 4 by addition of an aliquot (1 μl per ml) of the stock solution to the culture medium ∼18 h before irradiation. In preparation for irradiation, the dye-containing medium was removed and replaced with 3 ml of Hank's balanced salt solution with Ca2+ and Mg2+. The light source was a 500-W tungsten-halogen lamp with a 600-nm long-pass filter. The fluence rate was 75 W/m2. All irradiations were performed at room temperature, and the temperature during irradiation did not exceed 34°C.

Cellular uptake of Pc 4

Cultures were treated with various concentrations of Pc 4 for 18 h at 37°C. Then cells were trypsinized, an aliquot was counted, and the remaining cells were collected onto a glass fiber filter. After air-drying the filters, Pc 4 was dissolved in 5 ml 100% ethanol and quantified by absorption spectrophotometry at λ=666 nm. The concentration of cell-associated Pc 4 in the extract was calculated assuming ε=2.4×105.

Determination of cell viability

Cells were seeded into 96-well microculture plates at 1×104 cells/well and allowed to attach overnight. For PDT, the medium was removed and replaced with fresh medium with or without 0.5 μM Pc 4. The cells were incubated for ∼18 h, then irradiated at various fluences. After irradiation, cells were incubated for an additional 24 h. For TG treatment, cells were incubated in 200 or 300 nM TG for 48 h. Cell viability was measured using the tetrazolium salt WST-1 (Boehringer Mannheim) according to the manufacturer's instructions.

DNA fragmentation analysis

Cells were collected at various times after PDT or TG treatment. DNA isolation and gel electrophoresis were performed as previously described (Agarwal et al., 1991). Briefly, cell pellets were resuspended in 500 μl of a solution containing 1×SSC and 10 mM EDTA, then 1% sodium lauryl sarkosinate and 0.1 mg/ml proteinase K were added, and the mixture was incubated at 50°C for at least 2 h. The DNA was precipitated by addition of 2 volumes of absolute ethanol, resuspended in 200 μl TE, and briefly treated with 1 mg/ml RNase before loading onto a 1.5% agarose gel. After electrophoresis, the gel was stained with 0.5 μg/ml ethidium bromide and photographed under UV light.

Western blot analysis for PARP cleavage

Following PDT and TG treatment, cells were lysed and sonicated in a solution containing 0.5% sodium deoxycholate, 0.2% SDS, 1% Triton X-100, 5 mM EDTA, 10 μg/ml leupeptin, 10 μg/ml aprotinin, and 1 mM phenylmethyl sulfonyl fluoride in phosphate buffered saline. An equal volume of 2×SDS sample buffer (125 mM Tris, pH 6.8, 4% SDS, 10% mercaptoethanol, and 20% glycerol) was added to the cell lysate. Equivalent amounts of protein were loaded onto an 8% polyacrylamide gel, subjected to electrophoresis, transferred to a PVDF membrane, and probed with a monoclonal anti-PARP antibody (a gift of C Whitacre and NA Berger, Department of Medicine, CWRU) or an anti-actin antibody (from Amersham, Arlington Heights, IL, USA) overnight at 4°C. The immune complexes were detected by ECL system (Amersham, Arlington Heights, IL, USA). The anti-PARP antibody recognizes an epitope in the NAD-binding domain and thus detects intact Mr 116 000 PARP and the Mr 90 000 cleavage product. The percentage of the total PARP cleaved in each lane was calculated as follows:

Immunoprecipitation

Cells were harvested and lysed for 30 min in 1 ml of ice-cold lysis buffer (0.5% NP-40, 1% Triton X-100, 50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1 mM sodium orthovanadate, 0.5 mM PMSF, 3 μg/ml aprotinin and 1 μg/ml leupeptin). The cleared cell extracts were incubated with an Etk antibody described previously (Qiu et al., 1998) overnight at 4°C followed by incubation with 10 μl of protein A+G Sepharose beads for 1 h. The immunoprecipitates were washed three times with ice-cold lysis buffer. The bound proteins were eluted from the beads in SDS sample buffer and separated on a 7.5% SDS – PAGE gel, then transferred and Western blotted.

In vitro kinase assay

Assay of the kinase activity of Etk was performed as described previously (Qiu et al., 1998). Briefly, the immunoprecipitates were washed twice with PAN buffer (20 mM PIPES, pH 7.0, 100 mM NaCl) and resuspended in 20 μl of kinase reaction buffer (30 mM PIPES, pH 7.0, 10 mM MnCl2, 10 μM ATP, 1 mM NaVO4 and 10 μCi γ-32P-ATP) containing 5 μg enolase (Sigma). The reaction mixture was incubated at room temperature for 10 min and stopped by adding an equal volume of 2×SDS sample buffer. The samples were separated and detected by autoradiography.