Sci. Aging Knowl. Environ., 3 April 2002
The Werner Syndrome Helicase-Nuclease--One Protein, Many Mysteries
The author is in the Department of Biochemistry, Rappaport Faculty of Medicine, Technion-Israel Institute of Technology, Post Office Box 9649, Bat Galim Haifa 31096, Israel. E-mail: email@example.com://sageke.sciencemag.org/cgi/content/full/sageke;2002/13/re2
Key Words: Werner helicase exonuclease progeroid progeria DNA replication DNA repair
Abstract: Werner syndrome (WS) is an autosomal recessive condition characterized by an early onset of age-related symptoms that include ocular cataracts, premature graying and loss of hair, arteriosclerosis and atherosclerosis, diabetes mellitus, osteoporosis, and a high incidence of some types of cancers. A major motivation for the study of WS is the expectation that elucidation of its underlying mechanisms will illuminate the basis for "normal" aging. In 1996, the gene responsible for the syndrome was positionally cloned. This advance launched an explosion of experiments aimed at unraveling the molecular mechanisms that lead to the WS phenotype. Soon thereafter, its protein product, WRN, was expressed, purified, and identified as a DNA helicase-exonuclease, a bifunctional enzyme that both unwinds DNA helices and cleaves nucleotides one at a time from the end of the DNA. WRN was shown to interact physically and functionally with several DNA-processing proteins, and WRN transgenic and null mutant mouse strains were generated and described. The substantial number of excellent reviews on WRN and WS that were published in the past 2 years (1-7) reflects the rapid pace of advances made in the field. Unlike those comprehensive articles, this review focuses on the biochemistry of the WRN protein and some aspects of its cell biology. Also considered are the putative functions of WRN in normal cells and the consequences of the loss of these functions in WS.
A Short History of Werner Syndrome Back to TopSpace allows only the outlining of prominent highlights in the evolving story of the study of Werner Syndrome (WS). For a riveting, detailed rendition of the history of WS research, the reader should turn to George Martin's recent review (8).
Otto Werner's 1904 doctoral dissertation was the first description of the syndrome that came to bear his name. He identified the disorder in four out of five siblings in one generation who were distinguished by premature ocular cataracts and scleroderma-like changes in their skin. A 1945 systematic analysis of the syndrome by Thannhauser and a detailed study of two affected sisters formed the basis for a 1966 landmark review by Epstein et al. (9) that brought WS to the fore as a rare autosomal recessive condition that displays some features of early aging.
Subsequent experimental work indicated that the shortened life-span and genomic instability that mark WS patients are also expressed at the cellular level. Cultured fibroblasts from such individuals have an extended S phase of the cell cycle, lose their proliferative capacity at an accelerated rate, and display a decreased life-span in culture relative to cells from age-matched controls (10, 11). In addition, dermal fibroblasts and lymphoblastoid cell lines established from circulating lymphocytes of WS patients carry an increased number of chromosomal translocations and deletions as compared to the corresponding lines developed from normal individuals (10, 12, 13). Interestingly, WS cells exhibit heightened sensitivity to the DNA-damaging agent 4-nitroquinoline 1-oxide (4-NQO); to camptothecin, an inhibitor of topoisomerase I; and to VP-16 and amascarine, both inhibitors of topoisomerase II. In contrast, WS cells are not especially sensitive to DNA-damaging agents, such as ultraviolet light (UV), alkylating agents, and bleomycin. Furthermore, they display no sensitivity or limited sensitivity to x-ray irradiation and hydroxyurea [reviewed in (1, 5)].
After the assignment of the WRN gene to chromosome 8p in 1992 (14), it was positionally cloned in 1996 and shown to possess seven consensus motifs characteristic of the RecQ family of DNA helicases. This family of proteins is present in every examined species from Escherichia coli to humans (15). In short order, the gene was expressed in insect cells. Its product protein was demonstrated to have 3' 5' DNA helicase activity (16, 17) and, rather surprisingly, to also possess 3' 5' exonuclease activity (18, 19). The WRN protein was also shown to interact physically and functionally with several proteins that play key roles in DNA replication, repair, and recombination, including DNA polymerase , proliferating cell nuclear antigen (PCNA), replication protein A (RPA), 5'-flap endonuclease (FEN-1), topoisomerase I, p53, and the Ku protein complex. The functions of each of these proteins are discussed in detail below.
Parallel investigations were dedicated to the study of the cellular functions of RecQ helicases from E. coli, Saccharomyces cerevisiae, Xenopus laevis, mouse, and human cells [for review, see (20, 21)]. Lastly, data on the phenotypes of Wrn mice with a deletion in part of the Wrn helicase domain (22) (WRN-hel), of Wrn null mice (23), and of transgenic mice harboring a mutant version of human WRN (24) (K577M-WRN ) were published. These impressive advances brought about an understanding that WRN is probably involved in more than one pathway of DNA metabolism. However, we are still in search of a unified model that describes the physiological functions of WRN and explains the molecular, cellular, and organ-level consequences of its absence in WS.
WRN DNA Helicase-Exonuclease Back to TopThe WRN gene and its homologs. The exceptionally large human WRN gene, which consists of 35 exons, encodes a protein of 1432 amino acids. Sequence homology marks WRN as a member of the RecQ subfamily of DNA helicases. Because homology between WRN and E. coli RecQ helicase (25), S. cerevisiae Sgs1 (26), and Schizosaccharomyces pombe Rqh1 (27) is confined to the helicase domain, these proteins are considered WRN orthologs rather than homologs. That is, while originating from a common ancestral gene, they became differentiated in the course of evolution. By contrast, with homology spanning their full sequence, the X. laevis FFA-1 (28) and mouse Wrn (29) gene products are authentic WRN homologs. The main sequence motifs in WRN, schematically presented in Fig. 1, are an NH2-terminal exonuclease domain spanning amino acids 78 through 219, a centrally located RecQ helicase domain consisting of seven conserved motifs (amino acids 569 through 859), a COOH-terminal nuclear localization signal, and a direct repeat of 27 amino acids between the nuclease and helicase regions. As in other RecQ helicases, motif I of the helicase domain accommodates the adenosine triphosphatase (ATPase) activity, whereas motif II contains a conserved DHxA (aspartate-histidine-any amino acid-alanine) sequence. The WRN promoter, which has two transcription initiation sites, exhibits features characteristic of constitutive promoters, such as the absence of the common promoter elements TATA and CAAT boxes (30). All WS patients carry mutations in both copies of the WRN gene that, regardless of their type (nonsense, frame-shift, or insertion-deletion), generate truncated protein products (3, 31). These mutant proteins lack a nuclear localization signal and fail to translocate to the nucleus (32). In addition to WRN, human cells contain four other known RecQ helicases: BLM, RecQ4, RecQL, and RecQ5 [reviewed in (33)]. Mutations in two of these helicase-encoding genes are associated with genetic diseases that are characterized by genomic instability and predilection for cancer: BLM for the Bloom syndrome (34) and RecQ4 for the Rothmund-Thompson syndrome (35).
WRN DNA helicase. After recombinant WRN was expressed and purified from insect cells, the WRN protein was shown to act in vitro as a DNA helicase (16, 17). WRN hydrolyzes ATP to drive DNA unwinding in the 3' 5' direction in a Mg2+-dependent reaction (16, 17, 36). A distinguishing feature of WRN is its ability to resolve diverse DNA substrates (that is, separate two or more hydrogen-bonded DNA strands), some of which are schematically depicted in Fig. 2. WRN unwinds forked DNA molecules (17), partial DNA duplexes with a single-stranded 3'-overhang (36), and partial DNA/DNA and DNA/RNA duplexes (17). In addition, WRN resolves, with similar or higher efficiency, tetrahelical DNA (37, 38), triplex DNA (39), and DNA with a four-way Holliday junction (38, 40). Notably, other RecQ helicases such as yeast Sgs1 (41) and human BLM (38, 42) unwind tetraplex DNA structures, and BLM also resolves triplex DNA (39) and Holliday junctions (38, 43). It was suggested that the penchant of WRN and its homologous enzymes for efficiently unwinding alternate DNA structures reflects their role in resolving such structures during DNA synthesis in vivo (1).
WRN is also a 3' 5' exonuclease. The first hints that WRN might possess nuclease activity in addition to its helicase activity came from a hidden Markov model analysis of the protein's amino acid sequence. This study revealed sequence homology between the NH2-terminal region of WRN and several 3' 5' exonucleases (44). Biochemical, genetic, and immunochemical experiments demonstrated that a 3' 5' exonuclease activity resides in the NH2-terminal region of the same polypeptide that accomodates the helicase activity (45, 46). The reported 3' 5' directionality of the nuclease, which was subsequently confirmed by several research groups, contrasts with an independent claim that the WRN exonuclease acts in a 5' 3' direction (47). WRN is distinguished from all other members of the RecQ helicase family by being the only protein in this group that possesses nuclease activity in addition to helicase activity (1). Whereas ATP does not affect the activity of most known exonucleases, WRN's exonuclease is stimulated significantly by ATP but not by the nonhydrolyzable compound ATPS (48, 49). The WRN exonuclease degrades 3'-recessed strands of DNA/DNA or DNA/RNA partial duplexes, but fails to digest single-stranded DNA or double-stranded DNA with blunt ends or 3'-protruding strands (48, 50).
Remarkably, as is the case with the helicase activity, the WRN exonuclease is most active on unusual DNA structures; bubble-containing double-stranded DNA, DNA with a single-stranded loop, and stem-loop DNA molecules are digested at higher efficiencies than are 3'-recessed strands of duplex DNA (49). Whereas the WRN exonuclease does not normally degrade blunt-ended double-stranded DNA, the introduction of a melted region at the center of a DNA duplex allows WRN to bind to the substrate and to efficiently digest both its blunt ends (50). Some features of the WRN exonuclease hint at a possible role in DNA repair. The enzyme removes a single mismatched terminal nucleotide from a recessed 3' end (48, 51) and is active at nicks and gaps (51). Also in line with a possible function in the repair of oxidative DNA damage, the nuclease efficiently removes a 3' PO4-containing terminal nucleotide (48). Contrasting with this evidence that WRN functions in DNA repair, other results show that the WRN protein does not bind preferentially to DNA damaged by the reagent 4-NQO (this agent damages DNA in a variety of ways, most of which remain chemically undefined) (52). In addition, the WRN exonuclease is blocked by some 3'-terminal oxidative modifications and bulky lesions in DNA (53). All told, WRN is distinguished from most known exonucleases by its covalent linkage to a helicase, its preference for alternate DNA structures, and the stimulation of its activity by ATP hydrolysis.
Self-association of WRN. After their association with DNA, nearly all known helicases oligomerize. Recent results from size-exclusion chromatography experiments suggest that WRN self-associates to form trimers (53). Electron microscopy of WRN-DNA complexes should provide substantiation of the proposed oligomeric nature of WRN.
Coordination of the WRN helicase and nuclease activities. Bohr and colleagues reported last year that the helicase and exonuclease activities of the WRN protein act in coordination in the unwinding of model forked DNA substrates at the forked end, while attacking exonucleolytically the 3' blunt end of the DNA (54). Interestingly, the length of the forked DNA substrate determines which activity, helicase or exonuclease, predominates (54). A quandary that remains, however, is how these two activities are able to act in concert while progressing in opposite directions along the DNA molecule. Whereas the WRN helicase proceeds in a 3' 5' direction, as defined by the "template" strand, the exonuclease migrates with a 3' 5' polarity, as defined by the "primer" strand. Hence, starting at a DNA gap, the two activities seem to move away from each other (Fig. 3A), while appearing to migrate toward each other along a forked DNA molecule (Fig. 3B). One possible mechanism that accommodates these opposing directionalities might be the warping of a forked DNA substrate, so that the WRN helicase and exonuclease domains face opposite ends of the DNA and can thus proceed in the same direction. Another possibility is that the large size of WRN, particularly if it acts as an oligomer, allows the protein to cover a DNA stretch long enough for the helicase and exonuclease to simultaneously process, as reported (54), short tracts at opposite ends of the substrate molecule. Lastly, the threading of mobile DNA through static WRN protein could permit coordinate action of the helicase and exonuclease activities despite their opposing directionalities.
Guilt by Association (with Other Proteins) Back to TopIn searching for cellular functions of the WRN protein, data are rapidly accumulating that demonstrate the physical and functional association of WRN with multiple other proteins. Early work demonstrated that wild type, but not mutant, WRN protein copurifies with a 17S multiprotein DNA replication complex (55). Immunochemical analysis detected two WRN-interacting components in the complex: PCNA and topoisomerase I (55). Subsequent reports from various laboratories identified a number of additional WRN-associating proteins (vide infra). As evident from data summarized in Fig. 4, most of these proteins are recognized participants in various aspects of DNA metabolism. Unfortunately, because all the interacting proteins take part in more than a single DNA metabolic pathway, assigning one definitive role for WRN based on its association with any particular protein remains difficult.
DNA polymerase . WRN interacts physically and functionally with DNA polymerase (pol ). Because pol is a major player in both DNA replication and repair, its association with WRN provides a strong direct argument for the involvement of the helicase-exonuclease in DNA synthesis. Direct physical interaction between the COOH-terminal portion of WRN and the p50 subunit of human pol was detected by two-hybrid screening, and association of WRN with the p50 and p125 subunits of this enzyme was demonstrated by coimmunoprecipitation (56). In addition, ectopically introduced tagged WRN colocalizes in the nucleolus of HeLa cells, together with subunits p50 and p125 of pol (56). Parallel experiments showed that WRN increases the velocity of nucleotide incorporation by yeast pol in the absence of PCNA. By contrast, WRN does not affect the rate of DNA synthesis catalyzed by the eukaryotic replicative polymerases or or by a variety of other DNA polymerases (57). The progression of every tested DNA polymerase is arrested by tetraplex or hairpin structures in template DNA, and WRN exclusively alleviates the blocking of yeast pol , allowing only this polymerase to proceed through and past these DNA secondary structure impediments (58). The physical association of WRN and pol , their nucleolar colocalization, and the specific stimulation of the polymerase by WRN argue for joint action of the two proteins. However, because pol participates in both DNA repair and replication, its association with WRN does not reveal whether the helicase-exonuclease takes part in only one or both DNA transactions.
PCNA. The trimeric PCNA acts as a scaffold protein in both DNA replication and repair by wrapping itself around double-stranded DNA and recruiting many proteins, including pol , to sites of DNA synthesis [reviewed in (59)]. WRN, which contains an NH2-terminal PCNA binding motif, copurifies with PCNA in a 17S DNA replication complex (55). Physical association of PCNA with WRN is inferred from their coimmunoprecipitation (51, 55). It is impossible, however, to conclude that WRN acts in either DNA replication or repair, because PCNA operates in both processes.
Replication protein A (RPA). Human RPA is a heterotrimeric, single-stranded, DNA binding protein that plays essential roles in DNA replication, repair, and recombination [reviewed in (60)]. Human RPA stimulates DNA strand displacement by the WRN helicase more efficiently than do single-stranded DNA binding proteins from E. coli or bacteriophage T4 (36, 61). RPA also stimulates WRN to unwind a forked telomeric DNA structure (54) and to disassemble a large, noncovalent, cohesive complex formed by the telomeric DNA sequence (62). Physical interaction between WRN and RPA, initially suggested by their stoichiometric relations (36), was subsequently demonstrated directly by their coimmunoprecipitation (61). When HeLa cells are treated with the DNA replication inhibitor hydroxyurea, microscopy experiments show that WRN and RPA are superimposed in the nucleoplasm, suggesting that upon inhibition of DNA synthesis, WRN migrates from the nucleolus to colocalize with RPA in the nucleoplasm (40). Again, because RPA participates in all the major DNA transactions, no specific role can yet be assigned to the RPA-WRN complex.
5'-Flap endonuclease 1(FEN-1). FEN-1 is a 5' endonuclease/5'-3' exonuclease involved in the processing of Okazaki fragments (the short stretches of DNA produced during discontinuous replication) (63), in long patch base excision repair (64), and in nonhomologous DNA end joining (65). FEN-1 cleaves DNA in a structure-specific fashion, at single-stranded 5' flaps that form at the junction of single-stranded and double-stranded DNA [for review, see (66)]. WRN and FEN-1 interact physically through a COOH-terminal 144-amino acid domain of WRN that shares homology with other RecQ helicases (67). Upon association with WRN, the nucleolytic activity of FEN-1 is enhanced more than 80-fold, and this stimulation is independent of the helicase or ATPase activities of WRN (67) (see "Protective Pas de Deux?") The involvement of FEN-1 in DNA replication, repair, and nonhomologous end joining obscures any specific role that the FEN-1/WRN complex might play.
Topoisomerases I and II. Topoisomerases are enzymes that catalyze changes in the linking number of DNA (that is, the number of times that two strands of a closed DNA duplex cross over each other). Similar to PCNA, topoisomerase I also constitutes a component of the WRN-containing 17S replication complex and coimmunoprecipitates with WRN (55). The WRN-topoisomerase I association might explain the reported sensitivity of WS cells, which lack functional WRN protein, to the topoisomerase I inhibitor camptothecin during S phase (68). It is likely, therefore, that the WRN-topoisomerase I complex plays some topological role during DNA replication.
On the basis of the reported physical interaction between the yeast RecQ homolog Sgs1 and topoisomerases II and III (26), researchers sought evidence of an interaction between WRN and topoisomerase II. WS cells were shown to be hypersensitive to chromosome damage induced by the topoisomerase II inhibitors VP-16 and amascarine (69). This hypersensitivity was restricted to the G2 phase of the cell cycle, when most DNA synthesis is due to repair and not replication. This finding suggests that WRN cooperates with topoisomerase II in topological functions during DNA repair.
The observation that WS cells are sensitive to inhibitors of topoisomerases I and II, which act at distinct phases of the cell cycle, suggests that WRN contributes to DNA topological transformations during both DNA replication and repair.
The Ku-DNA-dependent protein kinase complex. Ku antigen, a heterodimeric protein composed of 86- and 70-kD subunits, helps to stabilize broken DNA ends and bring them together so they can be ligated. The Ku antigen binds to DNA ends and to other discontinuities in duplex DNA. Association of a ~470-kDa catalytic subunit (DNA-PKcs) of a DNA-dependent protein kinase activates the kinase function (70). The Ku-kinase complex is designated DNA-PK. Together with the kinase subunit, Ku participates in repairing DNA double-strand breaks caused by physiological oxidation reactions, recombination, ionizing radiation, and chemotherapeutic insults to DNA. Ku-dependent, nonhomologous, DNA end joining constitutes the major DNA double-strand break repair mechanism in mammalian cells (70).
Two groups of investigators independently and simultaneously used affinity binding and coimmunoprecipitation assays to demonstrate a physical interaction between WRN and Ku (71, 72). Binding of Ku by WRN deletion mutant proteins revealed that a 50-amino acid tract at the NH2-terminal region of WRN is necessary and sufficient to bind the Ku heterodimer at amino acids 215 through 276 of its 86-kD subunit (73). Whereas Ku binding does not affect the helicase activity of WRN (71), it greatly enhances the processivity of its exonuclease (71, 72). Ku also broadens the specificity of the WRN exonuclease so that in addition to recessed 3' ends, it also attacks blunt ends and protruding 3' single strands, neither of which can be degraded by WRN alone (72). Use of a WRN deletion mutant demonstrated that Ku-mediated stimulation of the WRN exonuclease activity does not require an intact helicase domain (73).
Immunochemical and electrophoretic mobility shift assays were recently used to show that WRN interacts directly with DNA-PKcs independently of Ku (74). DNA-PKcs phosphorylates WRN in vivo and in vitro and inhibits both its helicase and exonuclease activities. Interestingly, Ku reverses this DNA-PKcs-mediated inhibition (74).
The binding of WRN to Ku and DNA-PKcs and the modulation of its activities by the two proteins suggest a role for WRN in the repair of DNA double-strand breaks. One should note, however, that WS cells exhibit only limited (74) or no (1, 5) hypersensitivity to ionizing radiation. Hence, the absence of WRN does not appear to sensitize cells to DNA double-strand breaks. Interestingly, Ku was shown last year to enable WRN to degrade DNA through 8-oxoguanine- and 8-oxoadenine-containing regions that block WRN alone (75). Also, WRN and Ku colocalize in the nucleus, especially after 4-NQO insult, suggesting that they collaborate to process damaged DNA (75). It is thus possible that Ku activates WRN to participate in the removal of some lesions from damaged DNA.
p53. One view of cellular senescence is that it evolved to suppress the generation of tumors, and thus tumor suppressors are proposed to regulate both tumorigenesis and cell aging. Indications that the key tumor suppressor p53 controls cellular senescence [see (76) for review] initiated studies of its possible link to WRN in determining cell aging. WS cells maintain normal amounts of p53 (77, 78), but overexpression of WRN in normal fibroblasts increases the concentration of p53 and initiates p53-mediated apoptosis (79). Also, transcription of the WRN gene, which is regulated by the Sp1 transcriptional control system, is repressed by p53 (80). Direct physical association between the COOH-terminal portion of WRN and p53 was demonstrated by coimmunoprecipitation (78, 81). The binding of p53 to intact recombinant WRN protein inhibits its exonuclease activity, but has no effect on a WRN deletion mutant protein that lacks the p53-interacting COOH-terminal segment (82). Similarly, two naturally occurring mutant versions of p53 have a reduced ability to inhibit the exonuclease activity of full-length WRN protein (82). Migration of WRN from the nucleolus to the nucleoplasm of hydroxyurea-treated S phase-arrested cells, and its colocalization with p53 supports the notion that WRN interacts with p53 during the blocking of DNA replication (82).
Knockout mice provide a clue about how WRN and p53 might cooperate to control genome stability and determine life-span. Recent findings demonstrate that p53 by itself affects the aging process (see Campisi Perspective). Mutant mice that express a truncated form of p53 that augments wild type p53 activity exhibit enhanced resistance to spontaneous tumors but begin to age prematurely and have an abnormally short life-span (83). In examining the effect of WRN on aging in genetically altered mice, it was reported that Wrn null mice that bear a mutation that abolishes the expression of the WRN helicase COOH-terminal domain do not exhibit overt pathologies, are not hypersensitive to camptothecin or to 4-NQO, and have life-spans that exceed the usual 2 years. However, introduction of an additional p53-null mutation increases the mortality rate of these mice (23). Another Wrn knockout mouse strain with a different deletion in the helicase domain appears to have a higher rate of prenatal lethality than do wild type mice, and embryonic stem (ES) cells from this knockout mouse are hypermutable, hypersensitive to camptothecin, and have a shortened life-span in vitro (see WRN-hel ) (22). When a p53 null mutation is introduced into these mice, they develop tumors at an accelerated rate, and the neoplasms are more variable in type than in p53 null mutant mice (84). It will be interesting to test the effects of a p53 null mutation in recently described genetically altered mice that carry a version of human WRN that encodes a putative dominant-negative mutant protein and confers a WS-like phenotype on the mouse cells (24).
SUMO1 and WHIP. Use of mouse cDNA as bait in yeast two-hybrid screening identified three WRN-interacting proteins: SUMO1 (small ubiquitin related modifier-1), Ubc9, and Werner helicase interacting protein (WHIP) (85, 86). SUMO-1 is covalently conjugated to WRN through the action of the conjugating enzyme Ubc9 (86). WHIP is an evolutionarily conserved protein that shares similarity with replication factor C, which is required to load PCNA onto DNA during replication. Physical association of WHIP with the NH2-terminal portion of WRN was indicated by their coimmunoprecipitation. Also, ectopically expressed WHIP colocalizes with WRN in granular nuclear structures (86). The function of these granular nuclear structures is unknown. Disruption of the yeast WHIP gene enhances the early aging phenotype of SGS1 null mutants, suggesting functional cooperation between WHIP and Sgs1 in determining cell life-span (86).
What Functions Does WRN Serve in Cells? Back to TopThe intricate biochemistry and cell biology of the WRN protein, its association with a large variety of DNA-processing proteins, and the complex phenotype of WS give credence to the idea that WRN plays more than a single role in DNA metabolism. The baffling multiple faces exhibited by WRN might reflect the fact that it is a component of diverse multiprotein complexes that perform distinct DNA transactions.
Arguments for the participation of WRN in DNA replication. Several lines of evidence suggest that WRN functions as a component of the DNA replication machinery. Some data also suggest that it might act to level the path for the replication machinery through its distinctive capacity to unwind replication-blocking alternate structures of DNA.
The reasons to think of WRN as a player in DNA replication are as follows: (i) at the cellular level, WS cells display an extended S phase (87); (ii) WRN is isolated together with other replication proteins from a 17S multiprotein DNA replication complex (55); (iii) the Xenopus laevis WRN homolog FFA-1 is necessary for the assembly of replication foci and colocalizes at these sites together with RPA (28, 88); and (iv) as detailed above, WRN interacts physically and functionally with RPA, pol , PCNA, FEN-1, and topoisomerase I, all of which participate (although not exclusively) in DNA replication. Other evidence suggests a potential role for WRN in enabling the progression of the replication machinery through alternate structures in the DNA template: (i) both the helicase (36-39) and 3' 5' exonuclease (49) activities of WRN preferentially process noncanonical DNA structures; (ii) WRN selectively collaborates with pol to allow progression of DNA synthesis through a blocking tetraplex DNA region in the template strand (58); and (iii) the co-purification of WRN and topoisomerase I in a 17S replication complex (55) and the sensitivity of WS cells to the topoisomerase I inhibitor camptothecin (68) pronouncedly at S phase (89) might hint at WRN-topoisomerase I cooperation in resolving DNA structures that topologically perturb replication.
Arguments for the participation of WRN in DNA repair and recombination. Accumulating data suggest role(s) for the WRN helicase-exonuclease in some modes of DNA repair or recombination. These include the following: (i) WS cells accumulate chromosome rearrangements (12, 13, 31, 90) and somatic mutations (31, 91, 92) in an age-dependent fashion and at an elevated rate relative to normal cells. In addition, microsatellite instability--enhanced mutagenesis at highly repetitive, short nucleotide tracts in genomic DNA--that might result from defective mismatch repair occurs at a higher frequency in WS cells than in normal cells (93). (ii) WS cells are hypersensitive to some but not all types of DNA-damaging agents [reviewed in (1, 5, 31, 91)]; these include 4-NQO (94-96) and drugs that induce DNA intrastrand cross-links (97). Hence, the repair deficit in WS and the potential role of WRN in repairing DNA are damage-specific rather than general. (iii) Whereas WS fibroblasts were reported to proficiently repair UV-induced pyrimidine dimers, WS lymphoblastoid cell lines were shown to be deficient in gene-specific and strand-specific repair of this type of damage (98). Other reports indicate that some, but not all, of the WS lymphoblastoid cell lines tested exhibit deficient repair of base-base and insertion/deletion mismatches (99). These findings are consistent with the documented capacity of the WRN exonuclease to remove a single mismatched terminal nucleotide from a recessed 3' end (48, 51) and with the observed nuclease activity of WRN at nicks and gaps (51). An alternative explanation for the deficient repair activity of these cells, however, is that they are predisposed to accumulate secondary mutations in mismatch repair genes other than WRN. (iv) WRN protein associates with proteins that (nonexclusively) participate in DNA repair functions, including pol , PCNA, FEN-1, RPA, and topoisomerase II. Reported colocalization of WRN and RPA in nuclear foci in response to DNA damage suggests that these two proteins cooperate in repairing damaged DNA (100). Also, evidence that the hypersensitivity of WS cells to topoisomerase II inhibitors is restricted to the G2 phase is in line with a possible cooperation of the two proteins in resolving topologically tangled DNA structures during DNA repair (69). (v) WRN protein interacts physically and functionally with the DNA double-strand break repair proteins Ku and its associated DNA-PKcs (70). The possibility that WRN collaborates with Ku to end-join broken DNA double strands is contradicted, however, by the lack of a prominent sensitivity of WS cells to ionizing radiation (1, 5, 74). At the same time, Ku enables the WRN exonuclease to digest DNA containing 8-oxoguanine and 8-oxoadenine lesions, which normally obstruct WRN alone, and Ku colocalizes with WRN in the nuclei of 4-NQO-treated cells (75). Being a multifunctional repair protein (70), Ku might thus cooperate with WRN in pathways for the removal of DNA lesions generated by 4-NQO or oxidative stress. (vi) Drawing on analogy with the initiation and disruption of homologous recombination by E. coli RecQ (101), Shen and Loeb proposed that WRN participates in recombination-mediated gap repair after replication forks stall (7). In line with the role of RecQ helicases from diverse species in ensuring genome stability, WS cell lines were reported recently to have a reduced capacity to resolve mitotic recombination products (102). Although these results are consistent with the unwinding, by WRN, of three- or four-stranded DNA junctions generated during DNA repair, the observations could also reflect the resolution of structures generated during recombination or replication.
Is WRN involved in transcription? Evidence suggests that WRN might also have a role in transcriptional activation. (i) Human WRN protein was shown to activate transcription in a yeast system, and the WRN transcriptional activation domain was identified and mapped to amino acids 315 through 403 (103); (ii) WS cells, permeabilized to allow entry of radioactive nucleotides that would allow subsequent measurement of transcription, exhibit reduced transcription rates; (iii) transcription is restored to normal levels by addition of normal cell extracts to chromatin isolated from WS cells (104); and (iv) a mutant version of the WRN protein that lacks the 27-amino acid direct repeat is unable to stimulate transcription, suggesting that WRN acts directly to permit RNA synthesis (104). The localization of WRN protein to transcriptionally active nucleoli (105-107) and the dependence of its nuclear-nucleolar trafficking on the transcriptional state (107) might hint at a role in the transcription of ribosomal DNA (rDNA). Evidence to the contrary indicates that steady-state quantities of 28S rRNA remain constant during the life-span of WS fibroblasts, although methylation of the rDNA genes is enhanced in senescent WS cells (108). All told, the roles for WRN in transcription remain to be substantiated.
The WRN-telomere connection. Several lines of evidence indicate that the shortening of telomeres with each division of normal cells determines the onset of replicative senescence (109-112). A possible connection between the WS phenotype and telomere metabolism is therefore an appealing subject for exploration. An early report showed that although telomeres of WS skin fibroblasts shorten at an accelerated rate, the cells cease to divide when the mean length of their telomeres is greater than in normal senescent cells (113). Thus, although the stepped-up loss of telomeric sequences in WS cells is consistent with the accelerated decline in their proliferative capacity, telomere shortening cannot explain their exit from the cell cycle.
In comparing telomere shortening in Epstein-Barr virus-transformed WS cells to that in normal cells, the former exhibited accelerated telomere trimming (114). The transformed WS cells had more heterogeneous lengths of telomeres as compared with the normal cells when both ceased to divide. Three recent reports showed that forced expression of human telomerase in WS cell lines extends their life-span and culminates in their immortalization (115-117). The capacity of telomerase to overcome the accelerated senescence of WS cells suggests that part of the WS defect is the speeding up of telomere-triggered replicative senescence. Telomeres don't solve the whole WS conundrum, however. Patterns of mRNA expression in telomerase-immortalized WS and normal cells differ substantially, suggesting that telomerase expression does not faithfully reverse the WS genotype (117).
Additional evidence for a link between WRN and telomere metabolism is provided by the colocalization of WRN and the telomere repeat binding factors TRF1 and TRF2 in nuclear foci of immortalized human cells that lack telomerase (118). And finally, one can draw analogies between the proposed WRN-telomerase connection and the requirement for Sgs1 activity in the maintenance of telomeres in telomerase-deficient yeast cells (118).
Epilogue Back to TopThe promise of understanding "normal" aging by uncovering the functions of a single gene product has drawn investigators to the study of WS and the WRN protein. Indeed, our knowledge regarding the function of the WRN protein has expanded rapidly in recent years. However, to paraphrase Winston Churchill, the WRN helicase remains a tantalizing riddle wrapped in a mystery inside a puzzling enigma. Even as more becomes known about this beguiling enzyme, we continue to struggle with fundamental questions related to its physiological function(s) and the basis for the multifaceted pathophysiology that its absence creates.
First, one is intrigued by the experimental indications that WRN participates in more than a single DNA metabolic pathway. It is possible that WRN is recruited into different multiprotein complexes to conduct distinct DNA transactions. Yet an important question remains unanswered: Why, despite its multitasking, does the absence of WRN become phenotypically detectable only with the onset of puberty (9, 119)? Do homologous RecQ helicases substitute for WRN at earlier ages? Do the activities of these other helicases decline at puberty, making WRN an essential DNA-processing enzyme? Are there functions specific to puberty or to later stages of life that only WRN can fulfill?
Second, what are the relative contributions of the WRN helicase and exonuclease activities to the pathology of WS? Because mutations that cause WS generate truncated protein products (3, 31) that fail to translocate to the nucleus (32), it is not clear whether it is the absence of the helicase or the exonuclease or both functions that produce the WS phenotype.
A most tantalizing question concerns the link between defective function(s) of DNA metabolism and the plethora of aging-like features of WS. The susceptibility of WS patients to neoplasia is consistent with the aberrant DNA transactions and the genomic instability that the absence of WRN induces. However, as pointed out elsewhere (119), it is not readily apparent how the lack of WRN helicase-exonuclease activities is coupled to susceptibility at an early age to ocular cataracts, atherosclerosis, osteoporosis, or diabetes mellitus. These and other equally intriguing questions will certainly hold the attention of many researchers for years to come.
April 3, 2002
Suggested ReadingBack to Top
THIS ARTICLE HAS BEEN CITED BY OTHER ARTICLES:
Science of Aging Knowledge Environment. ISSN 1539-6150