Recapitulation of the Cellular Xeroderma Pigmentosum-Variant Phenotypes Using Short Interfering RNA for DNA Polymerase Η¹

Patients with the hereditary sun-sensitivity and cancer-prone syndrome XP³ of the V group are deficient in a damage-specific DNA polymerase, POLH. DNA photoproducts are blocks to the replicative DNA polymerases because these enzymes cannot replicate past large distortions such as DNA photoproducts or adducts (1, 2). Replicative bypass of photoproducts is instead achieved mainly by POLH (2, 3) that has a high capacity for replicating many kinds of DNA lesions, and preferentially inserts adenines opposite thymine-containing lesions (4). XP-V cells contain mutations in POLH (5, 6) that cause severe truncations of the protein or prevent the enzyme from relocating to replication foci after UV irradiation (7, 8, 9). The phenotype of XP-V cells includes an enhanced arrest of DNA replication at pyrimidine dimer sites and increased mutagenesis, recombination, and UV sensitivity in the presence of caffeine (10, 11, 12, 13, 14, 15, 16). To study the signaling events involved in the response of XP-V cells to UV damage, we developed an inhibitor of POLH expression, using an RNA interference approach that used both RNA and a RNA polymerase III-based expression vector system (17).

SV40-transformed human fibroblasts that exhibited normal (GM637) or replication-deficient UV-repair response (XP30RO) were used in all of the studies and have been described elsewhere (14, 15).

Pairs of RNA oligonucleotides were designed using the Oligoengine design tool (Oligoengine, Seattle, WA) and were synthesized commercially (Xeragon, Huntsville, AL). The sequences of the oligonucleotides were: A-F, r(GCU GCA GAA AGG CAG AAA G)dTT; A-R, r(CUU UCU GCC UUU CUG CAG C)dTT; B-F, r(ACG UGC UCG AUU CUC CAG C)dT; B-R, r(GCU GGA GAA UCG AGC ACG U)dTT; C-F, r(UCC d-F, r(GCA ACC CCA AAU CCA AGU C)dTT; and d-R, r(GAC UUG GAU UUG GGG UUG C)dTT. Oligonucleotides were annealed according to the manufacturer’s instructions.

An RNA polymerase III-based expression vector (pSUPER; Oligoengine; Ref. 17) was linearized with HindIII and BglII. Pairs of DNA oligonucleotides encoding hairpin RNAs were designed using the Oligoengine design software, and were annealed and ligated into the vector using standard techniques. The sequences were directed against exons 7 and 11 of the Pol H sequence. Numbers indicate the starting nucleotide position of the targeted sequence, using the GenBank identifier for POLH, NM_006502. The sequences for the pairs were: 1087F, GAT CCC CTC CCA GCT CCA GAG TCA TTT TCA AGA GAA ATG ACT CTG GAG CTG GGA TTT TGG AAA; 1087R, AGC TTT TCC AAA AAT CCC AGC TCC AGA GTC ATT TCT CTT GAA AAT GAC TCT GGA GCT GGG AGG G; 1696F, GAT CCC CGC TGC AGA AAG GCA GAA AGT TCA AGA GAC TTT CTG CTT TCT GCA GCT TTT TGG AA; 1696R, AGC TTT TCC AAA AAG CTG CAG AAA GGC AGA AGT CTC TTG AAC TTT CTG CCT TTC TGC AGC GGG; 1770F, GAT CCC CTT CAC CAT CCA AGC CCT CAT TCA AGA GAT GAG GGC TTG GAT GGT GAA TTT TTG GAA A; 1770R, AGC TTT TCC AAA AAT TCA CCA TCC AAG CCC TCA TCT CTT GAA TGA GGG CTT GGA TGG TGA AGG C; 2095F, GAT CCC CGC AAC CCC AAA TCC AAG TCT TCA AGA GAG ACT TGG ATT TGG GGT TGC TTT TTG GAA A; and 2095R, AGC TTT TCC AAA AAG CAA CCC CAA ATC CAA GTC TCT CTT GAA GAC TTG GAT TTG GGG TTG CGG G. Recombinants were confirmed by sequencing.

For SiRNA, 8 μg of RNA were added to a 100-mm² culture dish of normal human GM637 fibroblasts, using TransMessenger transfection reagent (Qiagen, Valencia, CA). For plasmids, 10 μg of DNA were added to a 100-mm² culture dish of normal human GM637 fibroblasts, using Fugene transfection reagent (Invitrogen, Carlsbad, CA). For stable clones, 2 μg of siRNA-encoding plasmid DNA along with 200 ng of the PvuII-linearized puromycin resistance plasmid pPUR (BD Biosciences, Palo Alto, CA) was added to a single well of a six-well culture dish of normal human GM637 fibroblasts cells, and clones were selected with puromycin (500 ng/ml).

At 48 h after transfection (RNA transfections), 72 h after transfection (transient plasmid transfections), or after 4 weeks of drug selection (stable clones), POLH protein was immunoprecipitated from nuclear lysates using the rabbit polyclonal antibody H-300 (Santa Cruz Biotechnology, Santa Cruz, CA), resolved on a 10% Tris-HCl acrylamide gel, blotted with a POLH mouse monoclonal (B-7) antibody (Santa Cruz Biotechnology), and detected with ECL-Plus (Amersham Pharmacia, Piscataway, NJ) using standard techniques. Densitometric quantitation of band intensities was performed using QuantityOne software (Bio-Rad Laboratories, Hercules, CA).

Cell survival after UV irradiation was determined from the number of cells that formed visible colonies 2 weeks after UV irradiation (254 nm, 1.3 J.m⁻²; Ref. 14). In some experiments, cells were grown in caffeine (1 mm) after irradiation, which is known to sensitize XP-V cells to UV light (10, 13, 14).

GM637, XP30RO, and stable siRNA clones were irradiated with 5.2 J.m⁻² UV and incubated with 1 mm of caffeine in culture medium or medium alone for 1 h. The medium was replaced with 1 μm bromodeoxyuridine (Roche Molecular Biologicals) dissolved in culture medium, with or without caffeine (1 mm) supplementation, and cells were incubated for 18 h. Cells were fixed and stained according to the instructions of BD PharMingen, using a 1:20 dilution of a FITC-conjugated mouse antibromodeoxyuridine antibody (Phoenix Flow Systems, San Diego, CA) and counterstained with propidium iodide (2 ng/ml). Cells were analyzed by flow cytometry (FACSCalibur 2; BD PharMingen), and 10,000 events were recorded for each sample using CellQuest Pro (BD PharMingen). For HU experiments, cells were cultured in medium supplemented with combinations of 2 mm HU and/or 1 mm caffeine for 18 h before fixation.

Cells were seeded onto dual-chambered slides (Nalgene) 16 h before irradiation with 13 J.m⁻² UV. At 4 h after irradiation, cells were fixed and stored at −80°C. Cells were stained for hMre11 using rabbit polyclonal anti-hMre11 antibody (Novus Biologicals, Littleton, CO) at a dilution of 1:100, using a protocol described previously (15).

We initially used siRNAs in the form of four independent 21-bp RNA duplexes that target POLH mRNA. To ascertain which siRNA sequences might be most effective at inhibiting POLH expression, SV40-transformed normal human fibroblasts were transfected with these duplexes, and Western blot analysis of POLH (Fig. 1) indicated a substantial reduction of POLH protein for two of the four RNA oligonucleotides (Fig. 1,A, Lanes 2 and 3). Because the use of siRNA-expressing plasmids can eliminate the need to synthesize RNA oligonucleotides for SiRNA experiments, an RNA polymerase III-based expression vector system encoding siRNAs (17) was also used. Inhibition was significant for one of the four plasmids (Fig. 1,B, Lane 7). The cotransfection of a selectable marker can lead to stable expression of siRNA in plasmid-transfected cells, and stable siRNA expression rather than transient expression is likely necessary for the evaluation of long-term end points, especially when the time course of inhibition is unknown. Accordingly, we engineered stable clones containing the plasmid that proved most effective in transient assay (Fig. 1,C). Two clones showed especially strong inhibition (Fig. 1 C, Lanes 5 and 6) and were used in subsequent experiments.

The most effective siRNA sequence corresponded to exon 11, the final exon of the POLH gene. This is in the region that corresponds to the proliferating cell nuclear antigen binding site and nuclear foci-formation region of the protein (8), but beyond the sites of all of the known mutations in POLH (7). Cleavage of the mRNA by the siRNA at a site just before the polyadenylic acid signal must, in this case, initiate a particularly efficient degradation of the message to reduce significantly the amounts of protein synthesized.

Fibroblasts of XP-V patients do not exhibit a significant increase of UV sensitivity in colony formation assays, but a dramatic increase in UV sensitivity becomes apparent if XP-V cells are incubated in the presence of caffeine (7, 10, 13, 14). Caffeine inhibits various kinases that are required for checkpoint activation (18), and after UV irradiation the sensitization is dependent on the presence of normal or elevated levels of p53, such as seen in these SV40-transformed cells (14). With the addition of caffeine to the culture medium after UV, normal (GM637; Fig. 2,A) and XP-V (XP30RO; Fig. 2,C) can be distinguished clearly. In stable clones containing vector-encoded siRNA against POLH, caffeine sensitivity was clearly evident at doses above ∼1.3 J.m⁻² (Fig. 2,B), indicating that the siRNA inhibitor was effective. The sensitivity was dependent on UV dose, and was intermediate between normal and XP-V cells. Colony survival in the presence of caffeine is perhaps the most sensitive method for distinguishing an XP-V phenotype and can be used as a diagnostic marker for the testing of XP patients. The stable clones that show a substantial but not complete inhibition of POLH via Western blot analysis (Fig. 1) demonstrate an intermediate phenotype in this particular assay.

We examined S phase arrest using a stable clone containing vector-encoded siRNA against POLH to determine whether the S phase arrest after UV that has been observed previously in XP-V cells (19) could also be observed with our siRNA inhibitor. UV damage initiated an arrest in the S phase in normal SV40-transformed cells as has been observed previously (Ref. 19; Fig. 3, a–c). XP-V cells undergo a much more protracted arrest at the S phase checkpoint(s) after UV damage (19), especially after caffeine (1 mm) sensitization (Fig. 3, e–h). A novel observation in the current experiment is the accumulation of cells with high (polyploid) DNA content in XP-V cells treated with both UV and caffeine (Fig. 3,h), indicating the inability of these cells to exit the cell cycle appropriately. A stable clone containing vector-encoded siRNA against POLH exhibited a caffeine-sensitive accumulation of cells in S phase after UV (Fig. 3, i–l), demonstrating the efficacy of this siRNA-based inhibitor. The caffeine-sensitized S phase arrest phenotype of this stable clone (Fig. 3,l) was intermediate between wild-type (Fig. 3,d) and XP-V (Fig. 3 h) cells.

To confirm that the S phase arrest observed in this siRNA clone was a specific response to UV damage, a similar experiment was conducted using the metabolic inhibitor HU (2 mm) that activates S phase checkpoints by a different mechanism than UV. Wild-type cells, the siRNA clone, and XP-V cells all showed an identical response to HU (Fig. 3, m–x) demonstrating that the prolonged S phase arrest observed after caffeine and UV irradiation in XP-V cells and the siRNA clone was specific for UV-induced DNA damage.

Mre11 is a component of the Mre11/Rad50/Nbs1 recombination complex, which mediates DNA recombination at DNA double-strand breaks and at arrested replication forks. In γ-irradiated or UV-irradiated cells, this complex undergoes relocalization into foci (15, 16, 20). In XP-V cells these foci associate with proliferating cell nuclear antigen (15, 16) indicating that they are located at replication forks. In response to UV irradiation, the frequency of foci-positive cells was similar in stable clones and XP-V cells, and significantly different from the wild-type parental cells (Fig. 4). For this end point, the phenotype of siRNA clones was indistinguishable from that of XP-V cells that are null for POLH, indicating that even a partial deficiency in POLH is sufficient for the recruitment of Mre11 complexes after UV damage.

These results demonstrate that the inhibition of DNA POLH using siRNA recapitulates the cellular XP-variant phenotypes. An important finding of these studies is that the threshold of PolH protein reduction required for an XP-V phenotype differs depending on the biological end point under study; greater reduction is required before cells exhibit full UV sensitivity than is required to initiate Mre11 recombination foci. Second, the development of these inhibitors lays the foundation for improving our understanding of the mechanism of skin cancer susceptibility in XP-V patients. These siRNA-based inhibitors can be applied to numerous signal transduction-mutant cell lines to study UV-induced signaling pathways under conditions of POLH-deficiency.

We thank the XP Society, Poughkeepsie, NY, for their continued support and encouragement to one of us (J. E. C.). We also thank Dr. William Hyun and the staff of the Laboratory for Cytometric Analysis, University of California San Francisco Cancer Center, for assistance in immunofluorescence and fluorescence-activated cell sorting.