Catherine Muller, Caroline Dusseau, Patrick Calsou and Bernard Salles

Transcription

1 Oncogene (1998) 16, 1553 ± Stockton Press All rights reserved 0950 ± 9232/98 $12.00 Human normal peripheral blood B-lymphocytes are de cient in DNA-dependent protein kinase activity due to the expression of a variant form of the Ku86 protein Catherine Muller, Caroline Dusseau, Patrick Calsou and Bernard Salles Institut de Pharmacologie et de Biologie Structurale (CNRS, UPR 9062) 205 route de Narbonne, Toulouse cedex, France The heterodimeric Ku protein, which comprises a 86 kda (Ku86) amd a 70 kda (Ku70) subunits, is an abundant nuclear DNA-binding protein which binds in vitro to DNA termini without sequence speci city. Ku is the DNA-targeting component of the large catalytic sub-unit of the DNA-dependent protein kinase complex (DNA- PK CS ), that plays a critical role in mammalian doublestrand break repair and lymphoid V(D)J recombination. By using electrophoretic mobility shift assays, we demonstrated that in addition to the major Ku DNA complex usually detected in cell line extracts, a second complex with faster electrophoretic mobility was observed in normal peripheral blood lymphocytes (PBL) extracts. The presence of this faster migrating complex was restricted to B cells among the circulating lymphocyte population. Western blot analysis revealed that B cells express a variant form of the Ku86 protein with an apparent molecular weight of 69 kda, and not the 86 kda- full-length protein. Although the heterodimer Ku70/variant-Ku86 binds to DNA-ends, this altered form of the Ku heterodimer has a decreased ability to recruit the catalytic component of the complex, DNA-PK CS, which contributes to an absence of detectable DNA-PK activity in B cells. These data provide a molecular basis for the increased sensitivity of B cells to ionizing radiation and identify a new mechanism of regulation of DNA-PK activity that operates in vivo. Keywords: DNA-PK; Ku autoantigen; B-lymphocytes; V(D)J recombination Introduction Ku protein was originally identi ed as an autoantigen recognized by sera from patients with autoimmune diseases (Minori et al., 1981). Ku antigen was puri ed by immunoa nity and shown to be a heterodimer consisting of 86 kda (Ku86) and 70 kda (Ku70) subunits (Francoeur et al., 1986; Minori and Hardin, 1986). Biochemical analyses of Ku demonstrated that it bound in vitro to DNA termini or other discontinuities in the DNA structure without apparent sequence speci city (Minori and Hardin, 1986; Blier et al., 1993; Falzon et al., 1993). Ku has also been shown to possess DNA-dependent ATPase (Cao et al., 1994) and Correspondence: B Salles Received 27 August 1997; revised 24 October 1997; accepted 24 October 1997 helicase (Tuteja et al., 1994) activities. Lastly, Ku has also been reported to be capable of sequence-speci c binding within the promoter elements of genes (Gri n et al., 1996). Taken together, these biochemical properties have suggested a role for Ku in DNA metabolic processes such as recombination, repair, replication and/or transcription. It is now established that Ku is the DNA-targetting component of the DNA-dependent protein kinase (DNA-PK) complex. DNA-PK consists of the heterodimeric Ku protein and a large catalytic subunit (DNA PK CS ), a 460 kda serine/threonine kinase which belongs to the phosphatidyl (PI)-3 kinase (p110) family (Hartley et al., 1995). DNA-PK CS is recruited to DNA by Ku binding (Gottlieb and Jackson, 1993) and then acquires the capacity, at least in vitro, to phosphorylate many DNA-bound proteins in the vicinity (Anderson et al., 1994). Recently, a series of mutational and biochemical experiments have identi- ed DNA-PK as a crucial component of the mammalian double-strand breaks (DSBs) repair apparatus. Indeed, cell mutants with a defective Ku86 (CHO xrs cells) are sensitive to ionizing radiation (IR S ) due to a decreased ability to repair DSBs (for review see Jeggo, 1997). In xrs cells, a direct consequence of the Ku defect is a concomitant de ciency in the kinase activity of DNA-PK (Finnie et al., 1995) indicating that Ku DNA-binding represents the predominant, if only, mechanism for DNA-PK activation. Similarly, in the rodent scid and V-3 mutants, the loss of DNA- PK CS (Blunt et al., 1995; Peterson et al., 1995; Blunt et al., 1996; Danska et al., 1996) confers IR S and a de ciency in DNA DSBs repair (Biederman et al., 1991; Hendrickson et al., 1991). In addition to IR S, the observed defect in DNA DSB repair alters e cient joining of V(D)J recombination intermediates that result in a profound immunode ciency. The V(D)J recombination is a lymphoid and site-speci c DNA recombination process that assembles the non contiguous genomic segments that encode the Variable (V), Diversity (D), and Joining (J) elements of immunoglobulin and T cell receptor genes (reviewed in Weaver, 1995). Thus, mice homozygous for the murine severe combined immune de ciency (scid) mutation lack mature T- and B-lymphocytes, due to a defect in V(D)J recombination coding junction formation (Schuller et al., 1986). Lastly, knock-out mice for Ku86 have been generated and, as expected, these mice have a profound immunode ciency (Zhu et al., 1996). Although its precise role remains largely unknown, these results clearly identi ed DNA-PK as an essential component of the mammalian DNA repair apparatus that rejoin damage-induced, as well as V(D)J recombination intermediates, DSBs.

2 1554 In human cell lines, Ku represents the major doublestrand DNA end binding protein as detected by band shift experiments and its expression does not appear to be regulated within cell types (Finnie et al., 1996 and our unpublished results). In addition, in human lymphoblastoid cell lines, no changes in the levels of Ku polypeptides have been detected following treatment with ionizing radiation (Jongmans et al., 1996). All these results lead to the idea that Ku activity is present at high constitutive levels in human cells. Despite many studies performed with established cell lines, little is known about Ku activity in normal fresh human cells and tissues. Indeed, only a few studies of Ku gene expression have been reported to date (Yaneva and Jhiang, 1991; Cai et al., 1994). Regarding the major role of the Ku protein in DSBs repair, this appears to be an important question. In an attempt to explore this issue, we have chosen to study Ku function in normal peripheral blood lymphocytes (PBL) obtained from healthy donors. Our results show, for the rst time, that DNA-PK activity is regulated in vivo, through the expression of a variant form of the Ku86 protein in human normal B cells. Results CMulleret al Detection of an altered form of the Ku heterdomer in peripheral blood lymphocytes In human cell lines, Ku DNA-ends binding (DEB) activity can be easily detected by using dsdna fragments in an electrophoretic mobility shift assay (EMSA) (Zhang and Yanera, 1992). As shown in Figure 1, in the human cell lines AHH1 and HeLa, at least three complexes of di erent mobility were observed (A, A' and A''). Ku has the ability to translocate internally on DNA allowing another Ku heterodimer to bind to free DNA ends (devries et al., 1989). Thus, the increase in complex A' and A'' formation in association with increased protein concentration is likely to correspond to the binding of multiple Ku heterodimers to the same DNA-probe as previously described (Zhang and Yaneva, 1992). When PBL extracts were considered, the results consistently showed two features independent of the origin of the samples tested (shown for three di erent donors in Figure 1). Firstly, DEB activity appears lower in PBL extracts than the one observed for established cell lines. Secondly, in addition to the major complex usually detected in cell lines (complex A), a second complex with faster electrophoretic mobility (complex B) was observed in PBL extracts. In PBL, the signal due to these two di erent complexes appeared almost identical. This pro le of DEB activity was detected in all the PBL extracts tested in our laboratory (obtained to date from six di erent healthy donors). In our hands, this unusual pattern of DEB activity observed in PBL was not detected in a variety of transformed cell lines of human or rodent origin tested (data not shown). The pattern of DEB activity observed in PBL suggests either that an additional protein, expressed only in this cell population, is able to recognize double-stranded DNA-ends or that the Ku heterodimer is present in an altered form in PBL extracts. To test these two Extracts 2 µg 4 µg A" A' A F PBL PBL A H A H Figure 1 A second protein-dna ends complex with faster electrophoretic mobility is detected in PBL extracts. Increasing concentrations of protein extracts obtained from human cell lines (A, AHH1; H, HeLa) or PBL (three independent healthy donors) were incubated with 0.1 ng 32 P-labeled DNA probe in the presence of 1 mg of closed circular plasmid DNA, as a non speci c competitor. The electrophoretic mobility of the protein- DNA complexes were analysed in 5% polyacrylamide gel as described in Materials and methods. F, position of free DNA probe; A, A', A'' and B, positions of protein-dna complexes consisting of DNA-ends binding activity hypotheses, we rst evaluated the presence or absence of Ku70 and Ku86 proteins in the B complex, by using various antibodies directed against these proteins in EMSA supershift experiments. First, the autoantibodies (puri ed from the AF human serum), which were shown to recognize the Ku70 and Ku86 proteins (Francoeur et al., 1986), were used. In the presence of these antibodies, both A and B complexes were completely supershifted and retarded in the gel as shown in Figure 2 (panel I). Similar experiments were performed with antibodies directed either against Ku70 or Ku86, to further assess the presence in the B complex of each subunit of the Ku herodimer. For Ku86, two di erent antibodies were used: a polyclonal antibody, Ku86 pc, and mab S10B1 which recognizes an epitope de ned by aa 8 ± 221 on the N terminus of Ku86 (Wang et al., 1993). As shown in Figure 2 (panel II and III), all Ku antibodies supershifted the migration of the complex B present in PBL extracts and the migration of the complex A present in PBL and AHH1 extracts shown as a control. The mab S10B1 a ected the migration of the complex B less than the other Ku Abs (see Figure 2, panel III). However similar results were obtained for the complex A present in PBL and in AHH1 extracts suggesting that this lower reactivity is an inherent property of this antibody. Thus, this set of experiments showed that both Ku70 and Ku86 were present in the complex B, the faster migration pattern indicating that one or both Ku proteins are present in an altered/modi ed form in PBL extracts. The detection of the altered form of the Ku heterodimer is associated with B-lymphocyte population In PBL extracts, both forms of the Ku DNA complexes were detected by EMSA. Thus, the appearance of the faster migrating complex might be B

3 C Muller et al 1555 AHH1 PBL AHH1 PBL AHH1 PBL AF Abs (µg) C C C pc C pc C pc S10B1 C pc S10B1 Supershift A B F Ku 80/Ku 70 Ku 70 Ku 86 I II III Figure 2 EMSA supershift experiments demonstrate that the faster migrating complex contained both Ku70 and Ku86 proteins. 4 mg of whole cell extracts were incubated with 0.1 ng 32 P-labeled DNA probe in the presence of 1 mg of closed circular plasmid DNA, as a non speci c competitor. After 10 min, either no antibody (control) or antibodies speci c for (I) Ku70 and Ku86 proteins (mabs AF), (II) Ku70 protein (Ab Ku70 pc), (III) Ku86 protein, polyclonal (Ab Ku86 pc), or monoclonal that recognizes aa 8 ± 221 (mab S10B1) were added to the reaction mixture prior to gel electrophoresis. The electrophoretic mobility of the protein-dna complexes were analysed in 5% polyacrylamide gel as described in Materials and methods. F, position of free DNA probe; positions of protein-dna complexes consisting of DNA-ends binding activity and the region where the supershifted complexes migrate are indicated component of one of the sub-populations present in PBL. Indeed, when whole cell extracts from puri ed B cells were evaluated, only the faster migrating B complex was detected (Figure 3). Interestingly, in B cell extracts, the mobility change associated with increasing protein concentration was also observed (detection of two additional complexes B' and B'' when the protein concentration in the binding assay is 4 mg, Figure 3) suggesting that the altered form of the Ku heterodimer still has the ability to translocate on the DNA as previously described for the normal Ku protein (devries et al., 1989). In this representative experiment, the DEB activity is shown in proteins extracts obtained from the whole lymphocyte population and B cells, within the same blood sample. Furthermore, in one sample, we veri ed that in T cells the usual pattern of Ku DEB activity is detected (data not shown). Thus, these results showed that the presence of the faster migrating complex is largely con ned to B cells among the circulating lymphocyte population. The B-lymphocytes express a variant form of the Ku86 protein To further characterize the altered form of the Ku heterodimer, the expression of the Ku proteins was analysed by Western blot. The expression of the full length Ku70 protein, which migrates at approximately 72 kda in SDS polyacrylamide gels, was detected in PBL and B cells at a slightly lower level of expression (about twofold) than in the human cell lines AHH1 and HeLa (Figure 4, panel I). In contrast, the expression of Ku86 protein was altered in the extracts obtained from lymphocytes (Figure 4, panel II and III). In PBL extracts, in addition to the full-length 86 kda protein, a second band, which cross-reacts with anti- Ku86 pc Ab appeared at a lower position corresponding to about 69 kda. This band was absent in extracts Extracts 2 µg 4 µg A" A' A F A PBL LB A PBL LB B" B' Figure 3 The detection of the faster migrating complex is restricted to B cells within circulating lymphocyte population. Increasing concentrations of protein extracts obtained from the human cell line AHH1 (A), PBL or puri ed B cells were incubated with 0.1 ng 32 P-labeled DNA probe in the presence of 1 mg of closed circular plasmid DNA, as a non speci c competitor. The electrophoretic mobility of the protein-dna complexes were analysed in 5% polyacrylamide gel as described in Materials and methods. F, position of free DNA probe; the positions of protein-dna complexes consisting of DNA-ends binding activity are indicated from AHH1 and HeLa cells. In B cells extracts, the full-length protein was not detected and only the lower 69 kda band appeared. These results were con rmed B

4 1556 CMulleret al using the mab S10B1. Again, the full-length Ku86 protein was undetectable in B cells and a lower band, present as a doublet (at approximately 69 and 71 kda, with a higher expression of the 69 kda protein), was detected in lymphocytes that was not present in AHH1 or HeLa cells (Figure 4, panel III). The altered form of the Ku heterodimer has a decreased ability to recruit DNA-PK CS that results in a failure to detect DNA-PK activity in B cells The Ku heterodimer is the DNA-binding component of DNA-PK, the catalytic activity of which is carried by a 460 kda protein, DNA-PK CS. We have shown that the altered Ku heterodimer retains its capacity to bind to DNA termini. We then investigated the capacity of the altered heterodimer to contribute, in the presence of ds-dna, to the formation of an active DNA-PK complex. First, the protein kinase activity of the complex was evaluated in lymphocytes and in B-puri ed cell population using the `pull-down' assay as previously described (Finnie et al., 1995). In this assay, only proteins able to bind to double-stranded DNA ends on cellulose beads are considered. Furthermore, the speci city for DNA-PK activity is ensured through the use of both a consensus substrate peptide issued from the sequence of the murine p53 and a mutated peptide which is a poor substrate for DNA-PK phosphorylation in vitro. By using this assay, a detectable amount of DNA-PK activity was found in PBL but at a signi cantly lower level than the one observed in the AHH1 and HeLa human cell lines (Figure 5). In B cell extracts, no detectable DNA-PK activity was found in the three independent samples tested. Thus, it can be deduced from these experiments that the DNA-PK activity detected in PBL was solely due to the T cells present in the PBL samples. We then evaluated by Western blot experiments the expression of DNA-PK CS in these di erent samples. As shown in Figure 5, both PBL and B cell puri ed population extracts contained DNA-PK CS, although at a level which appeared lower (about twofold) than in the AHH1 cell line. Thus, these results suggested that the altered form of the Ku heterodimer present in B cells is not able to recruit the catalytic sub-unit of the complex, DNA-PK CS. To test this hypothesis, the proteins associated with the double-stranded DNA cellulose beads were eluted and subjected to Western analyses. Beads which have been incubated with AHH1 or PBL extracts contained DNA-PK CS as shown in Figure 6. In contrast, beads incubated with B cell extracts lacked DNA-PK CS. Furthermore, proteins associated with beads from all the extracts contained the Ku proteins. Beads which have been incubated with AHH1 extracts contained full length Ku70 and Ku86 whereas beads incubated with B cell extracts contained 250 kda A PBL LB Figure 5 B-lymphocytes are de cient in DNA-PK activity. Extracts (50 mg) were analysed by the DNA-PK pull-down assay in the presence of wild-type (black squares) or mutated (hatched squares) p53 peptide as indicated in Materials and methods. The established cell lines corresponded to AHH1 and HeLa cells. PBL extracts were obtained from three independent donors and B cell extracts from the puri ed B population of these three same samples. Mean of at least two experiments for all these independent extracts. The insert shows the expression of DNA- PK CS in these di erent cells as determined by Western blot experiments using mab 18 ± 2 (I) Ab Ku70 pc A H PBL LB AHH1 PBL LB DNA-PK cs (Ab 18 4) (II) Ab Ku86 pc Ku70 (Ab Ku70 pc) (III) Ab S10B1 K 86 (Ab S10B1) Figure 4 Western blot analysis in lymphocyte extracts revealed the expression of a variant form of the Ku86 protein. 50 mg of protein extracts obtained from human cell lines (A, AHH1; H, HeLa), PBL or B cells were subjected to electrophoresis on a 8% SDS polyacrylamide gel under reducing conditions. The proteins were then blotted onto a nitrocellulose lter and expression of the Ku proteins were detected as indicated in Materials and methods with indicated antibodies that recognize either the Ku70 protein (I) or the Ku86 protein (II and III) Figure 6 The variant Ku heterodimer is unable to interact with DNA-PK CS. After incubation of the indicated protein extracts with ds-dna cellulose beads, the proteins bound to the beads were eluted as indicated in Materials and methods and subjected to electrophoresis on a 8% SDS polyacrylamide gel under reducing conditions. The proteins were then blotted onto a nitrocellulose lter and the expression of DNA-PK CS, Ku70 and Ku86 was detected by Western blots using the indicated antibodies

5 full-length Ku70 and only the variant form of Ku86 (see Figure 6). These results imply that the altered form of the heterodimer fails to interact with DNA-PK CS. Mechanism(s) that contribute(s) to the expression of the Ku86 variant protein To test if the variant form of the Ku86 protein arises from a spliced form of Ku86, Northern blot experiments were performed using a Ku86 cdna probe. As shown in Figure 7, in the control cell line AHH1, Ku86 probe hybridized to two mrnas of 2.6 and 3.4 kb as previously described (Cai et al., 1994; Jongmans et al., 1996). The two Ku86 transcripts were also present in PBL and in puri ed B-lymphocytes and no smaller form of Ku86 mrna was detected in B-lymphocytes. Discussion In this study, we demonstrate that the Ku DNA-end binding activity is altered in peripheral blood B-lymphocytes with the expression of a faster mobility complex in electrophoretic mobility shift assays (EMSAs), as compared with human cell lines. Western blot analysis revealed that B cells express a variant form of the Ku86 protein with an apparent molecular weight of 69 kda, and not the 86 kda- full-length protein. Although the heterodimer Ku70/variant-Ku86 binds to DNA-ends, this altered form of the Ku heterodimer has a decreased ability to recruit the catalytic component of the complex, DNA-PK CS, which contributes to an absence of detectable DNA-PK activity in B cells. Characteristics of the variant Ku86 protein The results obtained in our study strongly suggest that the protein detected in B-lymphocytes indeed corresponds to a variant form of Ku86 rather than an unrelated protein which cross-reacts with Ku86 antibodies. It has been reported that the active DNA end binding member of the heterodimer is likely to be Ku70 (Zhang and Yaneva, 1992). One study raised the possibility that Ku70 might carry out this DEB function in the absence of Ku86 (Wang et al., 1994) but this remains controversial (Gri th et al., 1992). Indeed, recent work using in vitro-translated Ku products clearly showed that full-length individual subunits are inactive for DEB activity indicating that heterodimer assembly precedes, and is necessary for 3.4 kb 2.6 kb H PBL LB LB Figure 7 Northern blot experiments. 5 mg of total RNA obtained from the control cell line AHH1 (H), peripheral blood lymphocytes (PBL) or B-lymphocytes (LB, two independent healthy donors) were separated on a 1.2% agarose ± 7% formaldehyde gel, transferred to nitrocellulose and hybridized with a Ku86 cdna probe. The size of the two Ku86 mrnas transcripts detected is indicated C Muller et al DNA-binding (Wu et al., 1996). Thus, since Ku86 fulllength protein is undetectable in B-lymphocytes, it appears that the second band detected by the Western blot experiments is indeed a variant form of the Ku86 protein which results in DEB activity by the assembly with Ku70 protein. Similar expression of a variant form of the Ku86 protein (but in addition to the full-length protein) has been recently reported in the promyelocytic leukemia cell line, HL60 (Han et al., 1996). The variant form of the Ku86 protein detected in B-lymphocytes presents similarities with the protein detected in HL60 cell line. In both studies, the mab S10B1 (aa 8 ± 221) detected a doublet consisting of two lower forms of the Ku86 protein with apparent molecular weight of 69 and 71 kda. In the HL60 cell line, another mab which recognizes the 610 ± 705 aa of Ku86 (mab 111) failed to recognize the variant form of the protein. Interestingly, in a previous report, only the full-length Ku86 was detected in PBL using this antibody (Ajmani et al., 1995), suggesting that mab 111 also fails to recognize the Ku86 variant expressed in lymphocytes. Finally, we observed that mab 150 which recognizes an epitope within amino acid 490 to 707 of the Ku86 protein (Kamiya Biomedical Co, Seattle, Washington, USA) did not a ect the migration of the altered Ku heterodimer DNA complexes in EMSA supershift experiments (data not shown). At the present time, the domain(s) responsible for the interaction between Ku86 or the Ku heterodimer with the catalytic component of the DNA-PK CS complex remain(s) unknown. The inability of the mab 111 (aa 610 ± 705) to recognize the variant form of Ku86 in the HL60 cell line (Han et al., 1996) and in PBL (Ajmani et al., 1995) and of the mab 150 (aa 490 ± 707) to supershift Ku70/variant Ku86 DNA complexes (our results) may implicate the C-terminal part of the protein in this function. However, only reconstitution studies with various deletion mutants of the Ku86 protein will allow this question to be answered since conformational changes induced in the Ku86 variant (and not the region deleted) can be discounted for the decreased ability to recruit DNA-PK CS. Presently, the mechanism which contributes to the expression by B cells of this variant form of Ku86 remains unknown. This expression appeared however to result from a signi cant biological event occurring within the cells. The possibility that a nonspeci c proteolytic cleavage of Ku86 may have occurred during preparation of cell extracts can be ruled out since, (i) the expression of the variant form of Ku86 was observed in extracts from circulating blood lymphocytes and not from continuous human cell lines, even when the extract preparations were performed simultaneously; (ii) a large set of protease inhibitors was used during protein extraction (see Material and methods) that consistently would abrogate the cleavage events induced by the major classes of known proteases; (iii) it has been previously demonstrated that unstimulated B cells have low protease activity as compared to mitogen-activated cells (Kishi et al., 1983; Ku et al., 1983). Finally, our results suggest that the mechanism that contributes to the expression of the variant Ku86 protein occurs at the post-transcriptional level. Indeed, the fact that B-lymphocytes express the two Ku86 transcripts also detected in cell lines (Cai et al., 1994; 1557

6 1558 CMulleret al Jongmans et al., 1996 and our present results), and not a smaller form, strongly suggests that the variant Ku86 protein does not arise from a spliced form of Ku86 and favor the involvement of event such as speci c proteolysis. Implication of the absence of DNA-PK activity for DNA DSB repair activity in B cell lymphocytes The catalytic component of the DNA-PK complex is essential for DNA DSB repair activity. Indeed, mutant cell lines such as the hamster V-3 and the murine scid cell line that exhibited normal levels of Ku-DNA ends binding activity but are defective in the catalytic subunit, DNA-PK CS (Blunt et al., 1995; Peterson et al., 1995) show a defect in DSB repair after exposition to ionizing radiation. Interestingly, among normal fresh human cells, resting peripheral blood lymphocytes (PBL) are especially sensitive to IR and within subpopulations of lymphocytes, B cells are more sensitive than T cells (Prosser, 1976; Prusek and Astaldi, 1979; Schwartz and Gaulden, 1980). This relative radiosensitivity appears to be related to a defect in DNArepair capacity of these cells, especially in the rejoining of DSB (Mayer et al., 1986). In accordance with the major functional role of DNA-PK in DSBs repair, the observed defect in DNA-PK activity in our study provides a molecular basis for the increased sensitivity of B cells to ionizing radiation. Furthermore, since T cells, on the contrary, appear to contain detectable levels of DNA-PK activity (see Figure 5 and comments), our data may explain the known di erence in sensitivity towards IR between these two lymphocyte sub-populations. Besides lymphocytes, di erences in radiosensitivity are observed in cells derived from di erent tumor types, as well as in normal tissues within and between individuals. The molecular mechanism responsible for these di erences are not well understood (McMillan and Peacock, 1994). The level of DNA-PK activity is likely to play a role in this variation in radiosensitivity, as we have shown here for normal lymphocytes. The expression of the Ku86 variant protein may be important for the regulation of DNA-PK activity in vivo and studies of its expression and distribution amongst primary normal human cell and tissues may help to further understand the physiological mechanism(s) underlying the relative radiosensitivity in humans. Role of the Ku86 variant in B cell di erentiation Finally, our results suggest that DNA-PK activity is regulated during B cell di erentiation in a stage speci c manner, through the expression of a truncated form of the Ku86 protein. Indeed, V(D)J recombination occurs at de ned stages of lymphocyte development and this function is primarily regulated by the expression of the two recombination activating genes, RAG1 and RAG2 (for review see Weaver et al., 1995). Whether or not the expression and activity of DNA-PK holoenzyme is also actively regulated in human cells upon di erentiation and V(D)J rearrangement still remains unknown. However, the fact that DNA-PK is an essential component of the V(D)J recombination process suggests that this activity is present in actively recombining pro/pre-b cells in contrast to its absence observed in our study in unstimulated circulating B- lymphocytes. In the light of di erences in DNA-PK activity between circulating B and T cells observed in our study, it is interesting to note that recently published works underline unsuspected di erences in the molecular regulation of early B and T cell di erentiation, leading to the idea that partial bypass of DNA-PK-dependent pathway of DSBs rejoining may occur in the T cell, but not the B cell, lineage. Indeed, a single treatment of newborn scid mice with DNA-damaging agents restored functional, diverse T cell receptor b chain coding joints, as well as development and di erentiation of thymocytes, but did not promote B cell development (Danska et al., 1994). In addition, p53 de ciency promoted T cell development but did not appear to circumvent the block in B cell maturation conferred by the scid mutation (Guidos et al., 1996; Nacht et al., 1996). All these data provide evidence that the later steps involved in V(D)J joining are not regulated identically, in T and B cells. Thus, whether the relative expression of the variant form of Ku86 protein is used by human B cells to actively regulate DNA-PK activity upon di erentiation and V(D)J rearrangement appears an important issue. In conclusion, we have shown that DNA-PK activity is negatively regulated in normal circulating B cells through an original mechanism that involves the expression of a variant form of the Ku86 protein. Whether the expression of this variant protein is restricted to B cells amongst primary human cells and tissues, and the molecular mechanism that contributes to its expression, appears to give rise to two important issues that are currently investigated in our laboratory. Materials and methods Antibodies Monoclonal antibodies directed against Ku86 (mab S10B1, Interchim, France) and DNA-PK CS (mab 18 ± 2, provided by Dr T Carter, St-John's University, New York, USA) have been previously described (Wang et al., 1993; Carter et al., 1990). Polyclonal antibodies directed against Ku86 and Ku70 (designated Ku86 pc and Ku70 pc in our study) were provided by Dr Falaschi (UNIDO, Trieste, Italy) (Tuteja et al., 1994); Human serum (AF) that contained autoantibodies that cross-reacted with Ku70 and Ku86 antigens (Francoeur et al., 1986) were provided by Dr EM Tan (the Scripps Research Institute, La Jolla, USA) and puri ed antibodies were obtained in our laboratory by the protein A sepharose procedure. Peripheral blood lymphocytes and transformed cell lines Blood samples (50 to 100 ml) were collected in heparinized tubes from healthy consenting volunteers. Mononuclear cell suspensions were obtained after Hypaque- Ficoll (MSL, Eurobio, France) gradient centrifugation as previously described (Barret et al., 1995). Mononuclear cells at the interface were collected, diluted to cells/ ml in RPMI 1640 medium, and incubated in tissue culture asks for 2 h at 378C to remove monocytes by adherence. After monocyte depletion, the samples were divided into two parts. The rst part of the sample was used for immediate cell extract preparation of the whole lymphocyte

7 population (PBL). The second part of the sample was processed to obtain puri ed B cell suspension as follows: samples were incubated with antibody to CD3 (mab UCHT1, Dako, Trappes, France) and T cells then removed by incubation with goat anti-mouse IgG1 (Fc) antibody coupled to magnetic beads according to the manufacturer's recommendation (Dynabeads M-450, Dynal, Oslo, Norway). Under these conditions, less than 0.5% CD3 or CD14 positive cells were detected in the highly puri ed CD 19+ B cell population (data not shown). Cell viability was 490% by trypan blue exclusion. Following puri cation, B-lymphocytes were immediately used for cell extract preparation or RNA extraction. The Epstein-Barr virus immortalized human lymphoblastoid AHH1 and the human HeLa cell lines were grown in RPMI 1640 medium (Gibco BRL, Cergy Pontoise, France) supplemented with 10% fetal calf serum, 2 mm glutamine, 125 U/ml penicillin, and 125 mg/ml streptomycin. Cell extracts Whole cell extracts were prepared as previously described with slight modi cations (Finnie et al., 1995). Brie y, cells were resuspended in extraction bu er ( cells per 100 ml for transformed cell lines and cells per 100 ml for normal human lymphocytes) containing 50 mm NaF, 20 mm HEPES (ph 7.8), 450 mm NaCl, 25% glycerol, 0.2 mm EDTA, 0.5 mm dithiothreitol in the presence of proteinase inhibitors [0.5 mm phenylmethylsulfonyl fluoride, 0.5 mg/ml apoprotin, 0.5 mg/ml chymostatin, 0.5 mg/ml leupeptin and 1.5 mg/ml pepstatin], then frozen on dry ice and thawed at 308C three times. After centrifugation for 30 min at 48C, supernatants were stored at 7708C. Band shift assay The probe was prepared as follows: an aliquot of a 123-bp DNA ladder (Gibco BRL, France) was digested with AvaI (Gibco BRL, Cergy Pontoise, France) in order to generate 123 bp monomers. After puri cation from agarose gel, fragments were end-labeled with a- 32 P-dCTP using DNA polymerase I Klenow fragment (Gibco BRL, France). Endlabeled probes were separated from unincorporated nucleotides by chromatography through Sephadex G-50 (Sigma, Saint Quentin, France). The band shift assay was performed as previously described (Zhang and Yaneva, 1992). In the supershift experiments, antibodies (at the indicated concentration) were added to the reaction mixture following 10 min incubation at 308C and incubated for an additional 20 min prior to gel electrophoresis. C Muller et al Western blot Proteins were subjected to electrophoresis on a 8% SDS ± polyacrylamide gel under reducing conditions. The proteins were then electroblotted onto a nitrocellulose lter using a semi-dry apparatus (Bio-Rad Trans-Blot, Bio-Rad, Ivrysur-Seine, France). Western blot experiments were performed as previously described (Muller et al., 1994). Antibodies used in Western-blotting experiments were as follows: the polyclonal antibodies directed against Ku70 and Ku80 proteins were used at a 1 : 2500 dilution (secondary antibody mouse anti-rabbit at a 1 : 2500 dilution). The puri ed monoclonal antibody directed against Ku86 (mab S10B1) was used at a 1 : 1000 dilution and the anti DNA-PK CS antibody 18 ± 2 at a 1 : 5000 dilution (secondary antibody goat anti-mouse at a 1 : 2500 and 1 : 5000 dilution, respectively). DNA-PK assay DNA-PK `pulldown' kinase assay was performed as previously described (Finnie et al., 1995). In certain experiments, after incubation of the extracts with doublestranded DNA-cellulose beads (Sigma, Saint-Quentin, France), beads were extensively washed with Z' 0.05 bu er [25 mm HEPES (ph 7.9), 50 mm KCl, 10 mm MgCl2, 20% glycerol, 0.1% Nodinet P-40, 1 mm dithiothreitol] and incubated for 1 h in 50 mm Tris-HCl ph 9.0/ Triton 1% with constant agitation to elute proteins bound to the beads. Samples were then electrophoresed in an 8% SDS ± polyacrylamide gel under reducing conditions. Proteins were transferred onto nitrocellulose lters which were subjected to Western blot analysis for the presence of DNA-PK CS, and Ku proteins as indicated above. Northern blotting Total RNA was isolated from lymphocytes and control cell line (AHH1) using TRIZOL reagent (GIBCO BRL) according to the manufacturer's recommendation. Five to 10 mg of RNA were then separated on a 1.2% agarose ± 7% formaldehyde gel. RNA was transferred to nitrocellulose and hybridized as previously described (Muller et al., 1994) using a Ku86 cdna probe (from nucleotide 248 to nucleotide 1198). Before hybridization experiments the transfer membranes were photographed under U.V. light to assess the amount of rrna by ethidium bromide uorescence. Acknowledgements The excellent technical assistance by Helene Vinial is gratefully acknowledged. We thank Dr Defais, Dr Jonhson (IPBS, Toulouse, France) for critical reading of the manuscript. We are grateful to Dr EM Tan for the gift of the AF serum, to Dr T Carter and Dr A Falaschi for the gift of the antibodies directed against DNA-PK CS and Ku proteins, respectively. This study was supported by grants from the `Association pour la Recherche sur le Cancer', the `Ligue Nationale Contre le Cancer' and the `Fondation de France'. CM is a recipient of a post-doctoral fellowship from the `Comite leuceâ miedelafondationdefrance' References Ajmani AK, Satoh M, Reap E, Cohen PL and Reeves WH. (1995). J. Exp. Med., 181, 2049 ± Anderson CW, Connelly MA, Zhang H, Sipley JD, Leesmiller SP, Sakaguchi K, Ullrich SJ, Jackson SP and Appella E. (1994). J. Protein Chem., 13, 500 ± 501. Barret JM, Calsou P and Salles B. (1995). Carcinogenesis, 7, 1611 ± Biedermann KA, San JR, Giaccia AJ, Tosto LM and Brown JM. (1991). Proc. Natl. Acad. Sci. USA, 88, 1394 ± Blier PR, Gri th AJ, Craft J and Hardin JA. (1993). J. Biol. Chem., 268, 7594 ± Blunt T, Gell D, Fox M, Taccioli GE, Lehmann AR, Jackson SP and Jeggo PA. (1996). Proc. Natl. Acad. Sci. USA, 93, ± Blunt T, Finnie NJ, Taccioli GE, Smith GC, Demengeot J, Gottlieb TM, Mizuta R, Varghese AJ, Alt FW, Jeggo PA and Jackson SP. (1995). Cell, 80, 813 ± 823.

8 1560 CMulleret al Cai QQ, Plet AA, Imbert J, Lafagepochitalo M, Cerdan C and Blanchard JM. (1994). Cytogen. Cell. Genet., 65, 221 ± 227. Cao QP, Pitt S, Leszyk J and Baril EF. (1994). Biochemistry, 33, 8548 ± Carter T, Vancurova I, Sun I, Lou W and DeLeon S. (1990). Mol. Cell. Biol., 10, 6460 ± Danska JS, Holland DP, Mariathasan S, Williams KM and Guidos CJ. (1996). Mol. Cell. Biol., 16, 5507 ± Danska JS, P umio F, Williams CJ, Huner O, Dick JE, Guidos CJ. (1994). Science, 266, 450 ± 455. devries E, van Driel W, Bergsma WG, Arnberg AC and Van der Vliet PC. (1989). J. Mol. Biol., 208, 65 ± 78. Falzon M, Fewell JW and Ku EL. (1993). J. Biol. Chem., 268, ± Finnie NJ, Gottlieb TM, Blunt T, Jeggo PA and Jackson SP. (1995). Proc. Natl. Acad. Sci. USA, 92, 320 ± 324. Finnie NJ, Gottlieb TM, Blunt T, Jeggo PA and Jackson SP. (1996). Phil. Trans. Royal Soc., 351, 173 ± 179. Francoeur AM, Peebles CL, Gompper PT and Tan EM. (1986). J. Immunol., 136, 1648 ± Gi n W, Torrance H, Rodda DJ, Prefontaine GG, Pope L and Hache RJG. (1996). Nature, 380, 265 ± 268. Gottlieb TM and Jackson SP. (1993). Cell, 72, 131 ± 142. Gri th AJ, Blier PR, Minori T and Hardin JA. (1992). J. Biol. Chem., 267, 331 ± 338. Guidos CJ, Williams CJ, Grandal I, Knowles G, Huang MT and Danska JS. (1996). Genes & Dev., 10, 1054 ± Han Z, Johnston C, Reeves WH, Carter T, Wyche JH and Hendrickson E. (1996). J. Biol. Chem., 271, ± Hartley KO, Gell D, Smith GC, Zhang H, Divecha N, Connelly MA, Admon A, Leesmiller SP, Anderson CW and Jackson SP. (1995). Cell, 82, 849 ± 856. Henderickson EA, Qin XQ, Bump EA, Schatz DG, Oettinger M and Weaver DT. (1991). Proc. Natl. Acad. Sci. USA, 88, 4061 ± Jeggo PA. (1997). Mutation Res., 384, 1±14. Jongmans W, Artuso M, Vuillaume M, Bre sil H, Jackson SP and Hall J. (1996). Oncogene, 13, 1133 ± KishiH,MikiY,KikutaniH,YamamuraYandKishimoto T. (1983). J. Immunol., 131, 1961 ± Ku GS, Quigley JP and Sultzer BM. (1983). J. Immunol., 131, 2444 ± Mayer PJ, Bradley MO and Nichols WW. (1986). Mutation Res., 166, 275 ± 285. McMillan TJ and Peacock JH. (1994). Int. J. Radiat. Biol., 65, 49 ± 55. Minori T and Hardin JA. (1986). J. Biol. Chem., 261, ± Minori T, Akizuki M, Yamagata H, Inada S, Yoshida S and Homma M. (1981). J. Clin. Invest., 68, 611 ± 620. Muller C, Bailly JD, Goubin F, Laredo J, Ja rezou JP, Bordier C and Laurent G. (1994). Int. J. Cancer, 56, 749 ± 754. Nacht M, Strasser A, Chan YR, Harris AW, Schlissel M, Bronson RT and Jacks T. (1996). Gene & Dev., 10, 2055 ± Peterson SR, Kurimasa A, Oshimura M, Dynan WS, Bradbury EM and Chen DJ. (1995). Proc. Natl. Acad. Sci. USA, 92, 3171 ± Prosser JS. (1976). Int. J. Radiat. Biol., 30, 459 ± 465. Prusek W and Astaldi G. (1979). Arch. Immunol. Ther. Exp., 27, 59 ± 71. Schuller W, Weiler IJ, Schuller A, Phillips RA, Rosenberg N, Mak TW, Kearney JF, Perry RP and Bosma MJ. (1986). Cell, 46, 963 ± 972. Schwartz JL and Gaulden ME. (1980). Environ. Mutagen., 2, 473 ± 485. Tuteja N, Tuteja R, Ochem A, Taneja P, Huang NW, Simoncsits A, Susic S, Rahman K, Marusic L, Chen J, Zhang J, Wang S, Ponger S and Falaschi A. (1994). EMBO J., 13, 4991 ± Wang J, Chou CH, Blankson J, Satoh M, Knuth MW, Eisenberg RA, Pisetsky DS and Reeves WH. (1993). Mol. Biol. Rep., 18, 15 ± 28. Wang J, Satoh M, Chou CH and Reeves WH. (1994). FEBS Lett., 351, 219 ± 224. Weaver DT. (1995). Adv. Immunol., 58, 29 ± 84. Wu X and Lieber MR. (1996). Mol. Cell. Biol., 16, 5186 ± Yaneva M and Jhiang S. (1991). Biochim. Biophys. Acta, 1090, 181 ± 187. Zhang WW and Yaneva M. (1992). Biochem. Biophys. Res. Comm., 186, 574 ± 579. Zhu C, Bogue MA, Lim DS, Hasty P and Roth DB. (1996). Cell, 86, 379 ± 389.