The Saccharomyces cerevisiae RAD6 Group Is Composed of an Error-Prone and Two Error-Free Postreplication Repair Pathways

Corresponding author: Wei Xiao, Department of Microbiology and Immunology, University of Saskatchewan, 107 Wiggins Rd., Saskatoon, SK, S7N 5E5 Canada., xiaow{at}sask.usask.ca (E-mail)

Communicating editor: M. HAMPSEY

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The RAD6 postreplication repair and mutagenesis pathway is the only major radiation repair pathway yet to be extensively characterized. It has been previously speculated that the RAD6 pathway consists of two parallel subpathways, one error free and another error prone (mutagenic). Here we show that the RAD6 group genes can be exclusively divided into three rather than two independent subpathways represented by the RAD5, POL30, and REV3 genes; the REV3 pathway is largely mutagenic, whereas the RAD5 and the POL30 pathways are deemed error free. Mutants carrying characteristic mutations in each of the three subpathways are phenotypically indistinguishable from a single mutant such as rad18, which is defective in the entire RAD6 postreplication repair/tolerance pathway. Furthermore, the rad18 mutation is epistatic to all single or combined mutations in any of the above three subpathways. Our data also suggest that MMS2 and UBC13 play a key role in coordinating the response of the error-free subpathways; Mms2 and Ubc13 form a complex required for a novel polyubiquitin chain assembly, which probably serves as a signal transducer to promote both RAD5 and POL30 error-free postreplication repair pathways. The model established by this study will facilitate further research into the molecular mechanisms of postreplication repair and translesion DNA synthesis. In view of the high degree of sequence conservation of the RAD6 pathway genes among all eukaryotes, the model presented in this study may also apply to mammalian cells and predicts links to human diseases.

THE Saccharomyces cerevisiae RAD6 DNA postreplication repair (PRR) and mutagenesis pathway consists of RAD5(REV2), RAD6(UBC2), RAD18, REV1, REV3, and REV7 (LAWRENCE 1994 ; FRIEDBERG et al. 1995 ). It is now generally agreed that the Rad18 single-stranded DNA-binding protein (BAILLY et al. 1994 ) and the Rad6 ubiquitin-conjugating enzyme (JENTSCH et al. 1987 ) form a stable complex (BAILLY et al. 1994 , BAILLY et al. 1997A , BAILLY et al. 1997B ), which is required for both PRR and mutagenesis. The mutagenesis pathway (rev) mutants were initially isolated by their reduced mutations after UV treatment (LEMONTT 1971 , LEMONTT 1972 ). REV1 encodes a deoxycytidyl transferase (NELSON et al. 1996A ) with a stretch of amino acid sequence homologous to Escherichia coli UmuC (LARIMER et al. 1989 ). rev2 did not reduce mutation frequency in most mutagenesis assays and is allelic to RAD5, encoding a protein with DNA helicase and zinc-binding domains (JOHNSON et al. 1992 ) and DNA-dependent ATPase activity (JOHNSON et al. 1994 ). REV3 encodes the catalytic subunit of a nonessential DNA polymerase (MORRISON et al. 1989 ; NELSON et al. 1996B ). Purified Pol (consisting of Rev3 and Rev7) is capable of bypassing thymine dimers more efficiently than Pol (NELSON et al. 1996B ). Thus, the yeast mutagenesis pathway appears to rely on a specific DNA polymerase (Pol) to bypass DNA replication blocks at the cost of increased mutations.

A large body of evidence argues for the existence of an error-free PRR pathway distinct from mutagenesis. The repair pathway mediated by the RAD5 gene is referred to as error free, since deletion of RAD5 does not strongly interfere with UV-induced mutagenesis; however, the rad5 mutation limits instability of simple repetitive sequences (JOHNSON et al. 1992 ) and enhances nonhomologous end-joining of double-strand breaks (AHNE et al. 1997 ). In addition, several yeast genes have been recently reported to belong to the RAD6 pathway and participate in error-free PRR. First, an allele-specific POL30 mutation, pol30-46, is epistatic to rad6 and rad18, but is synergistic with rev3. The pol30-46 mutant is normal in UV-induced mutagenesis and DNA synthesis but displays significantly reduced PRR activity (TORRES-RAMOS et al. 1996 ). POL30 is essential and encodes proliferating cell nuclear antigen (PCNA) required for both Pol and Pol DNA synthesis (PRELICH et al. 1987 ; LEE et al. 1991 ; AYYAGARI et al. 1995 ). Inactivation of Pol, but not Pol, results in impaired PRR activity (TORRES-RAMOS et al. 1997 ). Hence, PCNA and Pol may be required in the later stages of error-free PRR. Second, the RAD30 gene is placed into the error-free PRR pathway on the basis of genetic analysis of the rad30 mutant (MCDONALD et al. 1997 ). RAD30 encodes a novel DNA polymerase (Pol), which is homologous to the E. coli DinB, UmuC, and S. cerevisiae Rev1, and can efficiently bypass a thymine-thymine dimer in vitro with high fidelity (JOHNSON et al. 1999B ). Mutations in its human homolog hRAD30 were found in all XP-V patients (JOHNSON et al. 1999A ; MASUTANI et al. 1999B ) whose cells display defective Pol activity (MASUTANI et al. 1999A ). Third, strains with a mutation in a newly identified MMS2 gene encoding a Ubc-like protein were found to share many phenotypes with pol30-46 (BROOMFIELD et al. 1998 ). In addition, the mms2 mutant exhibited significantly increased spontaneous mutation rates in a REV3-dependent manner (BROOMFIELD et al. 1998 ; XIAO et al. 1999 ), which would be expected if MMS2 plays a role in error-free PRR parallel to the REV3 mutagenesis pathway. More recently, Mms2 and Ubc13 have been shown to form a complex, which is responsible for in vitro Lys-63 ubiquitin chain assembly (HOFMANN and PICKART 1999 ). It has been proposed that this unique Lys-63 polyubiquitination on target protein(s) may be part of a novel signal transduction mechanism to recruit PRR proteins to the site of DNA damage (HOFMANN and PICKART 1999 ). To understand how the error-free PRR pathway is constituted, whether or not the above error-free PRR genes belong to the same pathway, and how enzymatic activities associated with these gene products contribute to error-free PRR, we conducted extensive genetic analysis in the hope of defining subpathways within the RAD6 group. Our results support a model in which the RAD6/RAD18 PRR/mutagenesis pathway consists of three rather independent subpathways represented by REV3, RAD5, and POL30. In addition, MMS2 and UBC13 may be required for both RAD5 and POL30 error-free PRR pathways. In contrast, the RAD30 gene plays a rather minor and specific role in the protection of yeast cells from UV damage and does not appear to belong to any of the above subpathways.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Yeast strains and cell culture:
Haploid S. cerevisiae strains used in this study are listed in Table 1. Three parental strains used in this study are DBY747, originally obtained from Dr. D. Botstein (Stanford University); BY448, from Dr. B. Andrews (University of Toronto, Canada); and PY39-0, from Dr. Burgers (Washington University, St. Louis). Other strains are all isogenic derivatives of the above strains created by targeted gene disruption. Yeast cells were cultured at 30° in either a rich YPD medium or a synthetic SD medium supplemented with various nutrients (SHERMAN et al. 1983 ). Intact yeast cells were transformed by a modified lithium acetate method. For one-step targeted gene disruption (ROTHSTEIN 1983 ), plasmid DNA containing the desired disruption cassette was cleaved with restriction enzymes prior to yeast transformation. All targeted gene disruption mutants were confirmed by Southern hybridization prior to phenotypic analysis.

Plasmids and plasmid construction:
A plasmid containing the rev3::LEU2 cassette was obtained from Dr. A. Morrison (National Institute of Environmental Health Sciences). The rev3::LEU2 cassette contains the REV3 coding region (MORRISON et al. 1989 ) with an internal 1.7-kb SnaBI fragment replaced by a 2.1-kb fragment containing the LEU2 gene (A. MORRISON, personal communication). Plasmid pBJ22 containing the rad5::hisG-URA3-hisG (JOHNSON et al. 1992 ) was received from Dr. L. Prakash (University of Texas Medical Branch, Galveston). Plasmid prad181 containing the rad18::LEU2 cassette (FABRE et al. 1989 ) was obtained from Dr. B. Kunz (Deakin University, Geelong, Victoria, Australia). Strategies for creating mms2::LEU2 (BROOMFIELD et al. 1998 ) and mms2::HIS3 (XIAO et al. 1999 ) mutations are as previously described.

The ubc13 disruption cassettes were made as follows. A 1.7-kb yeast genomic DNA at the UBC13 coding region was PCR amplified with oligonucleotides UBC13-1(5'-CTTGGGCATG CTGACAATG-3') and UBC13-2 (5'-CGGAATTAAACGTG GACCC-3'). After SphI-XhoI digestion, the DNA fragment was cloned into SphI-SalI sites of pTZ18R (Pharmacia, Piscataway, NJ). A 0.8-kb BssHII-NruI fragment containing essentially the entire UBC13 coding region from the resulting pTZ-UBC13 was deleted and converted into a BglII site with a BglII linker to form pubc13Bg. BglII-linearized pubc13Bg was used as a vector to clone either a 1.16-kb BamHI fragment from YDp-H or a 1.6-kb BamHI fragment from YDp-L (BERBEN et al. 1991 ) to form pubc13::HIS3 and pubc13::LEU2, respectively. The ubc13::HIS3 and ubc13::LEU2 cassettes were released by XbaI-MluI digestion.

The rad30 disruption cassettes were made as follows. Plasmid pJM80 (MCDONALD et al. 1997 ) was a gift from Dr. R. Woodgate (National Institute of Child Health and Human Development, National Institutes of Health). The 2.46-kb RAD30 PCR product was isolated from pJM80 as an SpeI fragment and cloned into the SpeI site of pBlueScript (Strategene, La Jolla, CA). A 1.0-kb AflII fragment within RAD30 was deleted and replaced by a BglII linker to form prad30Bg. BglII-linearized prad30Bg was used as a vector to clone either the 1.6-kb BamHI fragment from YDp-L (BERBEN et al. 1991 ) or the 3.8-kb BamHI-BglII fragment from pNKY51 (ALANI et al. 1987 ) to form prad30::LEU2 and prad30::hisG-URA3-hisG, respectively. The rad30::LEU2 disruption cassette was released by StuI-NarI digestion and the rad30::hisG-URA3-hisG disruption cassette was released by SspI digestion.

Cell killing by DNA-damaging agents:
Methyl methanesulfonate (MMS) and UV-induced quantitative killing experiments were performed at 30° in YPD. Overnight yeast cultures were used to inoculate fresh YPD at a 10-fold dilution and cells were allowed to grow for another 4–6 hr. For MMS treatment, MMS was added to the culture at a final concentration as specified and aliquots were taken at given intervals. Cells from each sample were collected via centrifugation, washed, diluted, and plated in duplicate on YPD. For UV treatment, cells were plated in duplicate at different dilutions and then exposed to 254 nm UV light in a UV crosslinker (Fisher Science model FB-UVXL-1000 at ~2400 µW/cm²) at given doses in the dark. The colonies were counted after a 3-day incubation. Untreated cells were also plated and scored as 100% survival.

MMS-induced killing was also measured by a gradient plate assay. Thirty milliliters of molten YPD agar were mixed with the appropriate concentration of MMS to form the bottom layer; the gradient was created by pouring the media into tilted square petri dishes. After brief solidification, the petri dish was returned flat and 30 ml of the same molten agar without MMS was poured to form the top layer. A 0.1-ml sample was taken from an overnight culture, mixed with 0.9 ml of molten 1% agar, and immediately imprinted onto freshly made gradient plates via a microscope slide. Gradient plates were incubated at 30° for the time indicated before taking photographs.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

MMS2 and UBC13 belong to the same error-free PRR pathway:
It was recently reported (HOFMANN and PICKART 1999 ) that Ubc13 and Mms2 form a complex in vitro, which is involved in the ubiquitin chain assembly through lysine 63. Epistatic analyses of yeast ubc13 and mms2 mutations also suggest that these two genes belong to the same pathway (HOFMANN and PICKART 1999 ). We deleted UBC13 from various mutant strains to further characterize the UBC13 gene function using the same criteria that defined MMS2. The ubc13 deletion mutant was indeed moderately sensitive to UV (data not shown) and to MMS (Fig 1A). Like mms2, the ubc13 mutation is synergistic with rev3 (Fig 1A) and belongs to the RAD6 pathway (BRUSKY et al. 2000 ). On a 0.005% MMS gradient plate, both ubc13 and rev3 single mutants grow to full length, whereas the ubc13 rev3 double mutant does not grow at all. These results are consistent with a recent report (BRUSKY et al. 2000 ) placing UBC13 within the error-free PRR pathway. The ubc13 mutant appears to be slightly more sensitive to MMS than either the mms2 single mutant or the mms2 ubc13 double mutant by a gradient plate assay (Fig 1A) and this result is reproducible, suggesting that mms2 is epistatic to ubc13. A similar result was also observed by UV killing (HOFMANN and PICKART 1999 ). These results would indicate that MMS2 acts upstream of UBC13, which is inconsistent with the model (HOFMANN and PICKART 1999 ) in which the Ubc13-Mms2 complex formation is required for its function(s). Furthermore, ubc13, mms2, and ubc13 mms2 mutants are indistinguishable in an MMS-induced liquid killing experiment (BRUSKY et al. 2000 ). Hence, the significance of the observed difference and genetic interactions between UBC13 and MMS2 remains to be elucidated. It should be borne in mind that although the ubc13 rev3 double mutant is strikingly more sensitive to DNA-damaging agents than its respective single mutants, it is still less sensitive than the rad6 or rad18 single mutant to killing by UV (BRUSKY et al. 2000 ) and MMS (Fig 1B). Under conditions of extremely low concentration of MMS (0.001%) and extended time of incubation, the ubc13 rev3 double mutant grows to full length, while rad6 and rad18 mutants only grow partially. The same phenomenon was also observed for the mms2 rev3 double mutant (BROOMFIELD et al. 1998 ; XIAO et al. 1999 ).

View larger version (68K): In this window In a new window Download PPT slide   Figure 1. Phenotypes of the ubc13 mutants by a gradient plate assay. Yeast cells were printed onto YPD or YPD gradient plates containing different concentrations of MMS as indicated, and the plates were photographed after (A) 42 hr or (B) 63 hr incubation at 30°. Strain genotypes are indicated. All the strains are isogenic derivatives of DBY747. The arrow points toward higher MMS concentration. DBY747 (WT); WXY850 (ubc13); WXY642 (mms2); WXY382 (rev3); WXY861 (mms2 ubc13); WXY862 (rev3 ubc13); WXY326 (rad18) and WXY376 (rad6). Individual colonies along the length of the MMS plate in (B) are revertants in the rad6 and rad18 mutants and have been repeatedly observed; these revertants are probably derived from srs2/radH mutations that suppress rad6 and rad18 sensitivity to DNA-damaging agents (LAWRENCE and CHRISTENSEN 1979 ; ABOUSSEKHRA et al. 1989 ).

RAD5 and POL30 represent two distinct error-free PRR pathways:
Both rad5 (JOHNSON et al. 1992 ) and pol30-46 (TORRES-RAMOS et al. 1996 ) mutations are additive to rad3 and rad52 group mutations and synergistic with rev3 and both genes belong to the RAD6/RAD18 pathway. In addition, rad5 and pol30-46 mutants are normal in UV-induced mutagenesis. These observations place RAD5 and POL30 within the error-free PRR pathway. To see if these two genes act in the same PRR pathway, we created isogenic single and double mutant strains and found that when the rad5 and pol30-46 mutations are combined, the double mutant is extremely sensitive to killing by either UV (Fig 2A) or MMS (Fig 2B), and the effect is considered to be highly additive (UV) or synergistic (MMS). This result would agree with the notion that RAD5 and POL30 constitute two parallel error-free PRR pathways within the RAD6/RAD18 pathway.

View larger version (22K): In this window In a new window Download PPT slide   Figure 2. RAD5 and POL30 belong to different DNA repair pathways. (A) UV-induced killing; (B) MMS-induced killing. () PY39-0 (wt); () PY39-46 (pol30-46); () WXY857 (rad5); (•) WXY858 (pol30-46 rad5). All the results are the average of at least three independent experiments with standard deviations.

MMS2 is common to the RAD5 and POL30 pathways:
To see if either RAD5 or POL30 acts in the same pathway as MMS2/UBC13, we performed epistatic analyses with respect to killing by either UV or MMS. The rad5 mutant is significantly more sensitive to UV (Fig 3A) and to MMS (Fig 3B) than its isogenic mms2 mutant; nevertheless, the rad5 mms2 double mutant is more sensitive than either of the corresponding single mutants, and the killing effect appears to be simply additive. This result indicates that MMS2 and RAD5 act in related but distinct pathways, although it does not rule out the possibility of overlapping functions. Similarly, inactivation of the mms2 gene enhances pol30-46 mutant sensitivity to either UV (Fig 3C) or MMS (Fig 3D) to a comparable extent as it does to the rad5 mutant, suggesting that MMS2 and POL30 act in different or overlapping error-free PRR pathways. It should be noted that the POL30 gene is essential for cell survival and that pol30-46 may be a partial loss-of-function mutation with respect to error-free PRR. On the other hand, Pol30/PCNA also physically interacts with factors involved in nucleotide excision repair (GARY et al. 1997 ), mismatch repair (JOHNSON et al. 1996 ; UMAR et al. 1996 ), and base excision repair (LI et al. 1995 ; WU et al. 1996 ), which may further complicate the above epistatic analyses.

View larger version (30K): In this window In a new window Download PPT slide   Figure 3. Genetic interactions of MMS2 with RAD5 and POL30 pathways. (A and B) rad5 vs. mms2. () DBY747 (wt); () WXY642 (mms2); () WXY731 (rad5); and (•) WXY732 (rad5 mms2). (C and D) pol30-46 vs. mms2. () PY39-0 (wt); () WXY859 (mms2); () PY39-46 (pol30-46); () WXY860 (pol30-46 mms2). All the results are the average of at least three independent experiments with standard deviations except D, which was from two sets of experiments.

Although MMS2 is not assigned to either the RAD5 or the POL30 pathway, it may belong to both error-free PRR pathways. This hypothesis is consistent with the observed additive effects between mms2 and rad5 or pol30-46 single mutations (Fig 3). Indeed, the rad5 pol30-46 mms2 triple mutant is no more sensitive than the rad5 pol30-46 double mutant to either UV (Fig 4A) or MMS (Fig 4C). We therefore propose that the Ubc13/Mms2 complex promotes both error-free PRR pathways represented by Rad5 and PCNA. In this model, Ubc13/Mms2 may act as a signal transducer to sense DNA damage or stalled replication, but is not absolutely required for the PRR activity via either Rad5 or PCNA.

View larger version (32K): In this window In a new window Download PPT slide   Figure 4. Epistatic analyses with the rad18 mutation. (A and C) mms2 is epistatic to the rad5 pol30-46 double mutations. (B and C), rad5, pol30-46, and rev3 are epistatic to rad18. (A and B) UV-induced killing. () PY39-0 (wt); () WXY859 (mms2); (•) WXY858 (pol30-46 rad5); () WXY880 (pol30-46 rad5 mms2); () WXY876 (rad18); (*) WXY879 (pol30-46 rad5 rad18); and () WXY887 (pol30-46 rad5 rev3). All the strains are isogenic to PY39-0, and results presented in A and B are from the same sets of experiments with standard deviations. (C) MMS-induced killing by a gradient plate assay. The gradient plates were photographed after 70 hr incubation at 30°. Lane 1, WXY857; lane 2, WXY858; lane 3, WXY880; lane 4, WXY887; lane 5, WXY876; and lane 6, WXY879. Strain genotypes are indicated at the bottom.

RAD30 is specific for UV damage and is distinct from all PRR pathways:
RAD30 encodes a novel DNA polymerase, Pol, which is able to synthesize DNA in vitro past thymine-thymine dimers in an error-free manner (JOHNSON et al. 1999C ). Previous epistatic analyses placed RAD30 within the error-free branch of RAD6 pathway (MCDONALD et al. 1997 ). However, unlike mms2, ubc13, rad5, and pol30-46, which are synergistic with the rev3 mutation, the rad30 mutation is only slightly additive to rev mutations (MCDONALD et al. 1997 and our unpublished data). Furthermore, the rad30 mutants are only sensitive to killing by UV, but not to a variety of other DNA-damaging agents including MMS, ionizing radiations, and a UV-mimetic agent 4-nitroquinoline-N-oxide, and do not display an increased spontaneous mutation rate (ROUSH et al. 1998 and our unpublished data). These results suggest that RAD30 differs from all other RAD6 pathway genes. Indeed, the rad30 mutation appears to be simply additive to mms2 (Fig 5A), rad5 (Fig 5B), or pol30-46 (Fig 5C) with respect to killing by UV. We note that MCDONALD et al. 1997 reported a strong synergistic interaction between rad5 and rad30 at low UV doses. Our results with 10 J/m² UV treatment (Fig 5B) also suggest a synergistic interaction between rad30 and rad5. However, at higher doses, the interaction is apparently additive.

View larger version (20K): In this window In a new window Download PPT slide   Figure 5. The rad30 mutation is additive to mms2 (A), rad5 (B), and pol30-46 (C) mutations with respect to killing by UV. (A) BY448 derivatives. () BY448 (wt); () T43 (mms2); () WXY724 (rad30); and (•) WXY725 (mms2 rad30). (B and C) PY39-0 derivatives. () PY39-0 (wt); () PY39-46 (pol30-46); (+) WXY858 (rad5); () WXY1004 (rad30); () WXY1005 (pol30-46 rad30); and (X) WXY1006 (rad5 rad30). All the results are the average of three independent experiments with standard deviations.

Reconstitution of the RAD6/RAD18 pathway by three distinct PRR/mutagenesis subpathways:
Having established a working hypothesis of two separate error-free PRR pathways, we attempted to construct a comprehensive model for RAD6/RAD18 PRR and mutagenesis. Both rad6 and rad18 mutants are extremely sensitive to killing by a variety of DNA-damaging agents and share other phenotypes such as increased spontaneous mutation rates but decreased UV-induced mutagenesis (LAWRENCE 1994 ; FRIEDBERG et al. 1995 ). However, the RAD6 gene is also involved in functions other than DNA postreplication repair, such as sporulation (MONTELONE et al. 1981 ), N-end rule protein degradation (DOHMEN et al. 1991 ; SUNG et al. 1991 ), polyubiquitination of histone H2B (WATKINS et al. 1993 ; ROBZYK et al. 2000 ), and telomere silencing (HUANG et al. 1997 ). The rad6 mutation also confers a slow-growth phenotype not shared by rad18 (LAWRENCE 1994 ; FRIEDBERG et al. 1995 ). Thus, the rad18 mutation instead of rad6 was used to represent complete defects in the PRR and mutagenesis. We have previously shown that yeast cells carrying both mms2 and rev3 mutations, although extremely sensitive to either UV or MMS, are still not as sensitive as the rad18 single mutant (XIAO et al. 1999 ). In the present study, we also found that the rad18 mutant is more sensitive than the ubc13 rev3 (Fig 1B) and rad5 pol30-46 (Fig 4) double mutants. If the RAD6/RAD18 PRR pathway consists of three subpathways represented by RAD5, POL30, and REV3, the rad5 pol30-46 rev3 triple mutant would be phenotypically equivalent to the rad18 single mutant, and the combination of any subpathway mutations with rad18 will be no more sensitive than the rad18 single mutant. Indeed, both UV (Fig 4B) and MMS (Fig 4C) killing experiments show that the pol30-46 rad5 rev3 triple, pol30-46 rad5 rad18 triple, and rad18 single mutants are indistinguishable, providing key support to our three-subpathway hypothesis.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The yeast S. cerevisiae has proved to be a paradigm for the study of DNA repair and mutagenesis in eukaryotes. Of three major DNA radiation damage repair pathways, namely, the RAD3 nucleotide excision repair, the RAD6 PRR and mutagenesis, and the RAD52 recombinational repair pathways, the RAD6 pathway is the most complicated and least characterized (FRIEDBERG et al. 1995 ). Historically, the RAD6 pathway has included all RAD genes that do not belong to either of the well-defined RAD3 and RAD52 groups. However, unlike the RAD3 pathway mutants, which are extremely sensitive to UV but less sensitive to MMS and ionizing radiation, and RAD52 pathway mutants, which are extremely sensitive to MMS and ionizing radiation but are less sensitive to UV, the RAD6 pathway mutants are sensitive to a broad range of DNA-damaging agents that probably share a common feature that inhibits DNA synthesis. It has been proposed that the RAD6 group consists of more than one subpathway (MCKEE and LAWRENCE 1979 ; FRIEDBERG 1988 ); however, these subpathways have not been exclusively defined, especially in the branch of error-free PRR. On the basis of previous reports and results obtained from this study, we present a comprehensive model of yeast RAD6/RAD18 DNA PRR and mutagenesis pathway, which is illustrated in Fig 6. In this model, we propose that the RAD6 group genes are responsible for the cellular tolerance to a variety of DNA replication-blocking lesions. The REV1,3,7 genes constitute a well-defined translesion synthesis pathway that replicates bypass lesions with low fidelity (LAWRENCE and HINKLE 1996 ). Mutations in the error-free PRR pathway genes are synergistic with rev mutations with respect to killing by DNA-damaging agents and are proficient in UV-induced mutagenesis. RAD5 and POL30 are assigned to two distinct PRR pathways based on the synergistic interaction between rad5 and pol30-46. Pol is included along with PCNA on the basis of reports that certain pol3 (e.g., pol3-13) mutations are epistatic to rad6 (GIOT et al. 1997 ) and that Pol is required for PRR, while Pol is not (TORRES-RAMOS et al. 1997 ).

View larger version (19K): In this window In a new window Download PPT slide   Figure 6. A model of the error-free postreplication repair and mutagenesis pathways in yeast.

Probably the most significant finding of this study is that the RAD6/RAD18 PRR and mutagenesis pathway can be exclusively defined by three subpathways represented by REV3, RAD5, and POL30. This conclusion is primarily based on the fact that REV3, RAD5, and POL30 all belong to the RAD6 epistasis group (JOHNSON et al. 1992 ; LAWRENCE 1994 ; FRIEDBERG et al. 1995 ; TORRES-RAMOS et al. 1996 ) and on our observations that rev3 rad5 pol30-46 and rad5 pol30-46 rad18 triple mutants and the rad18 single mutant are indistinguishable in response to killing by either UV or MMS.

Although RAD30 has been placed in the RAD6 group (MCDONALD et al. 1997 ), we are unable to assign it into any of the three existing subpathways. We argue that RAD30 is probably not a typical RAD6 pathway gene, since it only protects cells from a specific type of DNA damage. Although RAD30 functions in an error-free manner, the rad30 mutation is not synergistic with rev mutations (MCDONALD et al. 1997 ). However, since an allele-specific pol30-46 mutation instead of the pol30 null mutation was used in our epistatic analysis, we are unable to rule out the possibility that RAD30 belongs to the POL30 subpathway. It is of interest to note that Rad30, like other recently discovered UmuC DNA polymerase superfamily proteins (NELSON et al. 1996A ; JOHNSON et al. 1999A ; MASUTANI et al. 1999B ; TANG et al. 1999 ; WAGNER et al. 1999 ), synthesizes DNA in a distributive manner (JOHNSON et al. 1999B ) and consequently its in vivo function may be related to a cognate non-UmuC family DNA polymerase. For instance, UmuC mutagenesis requires PolIII (FRIEDBERG et al. 1995 ) and the Rev1 function is dependent on Rev3 (WAGNER et al. 1999 ). Hence, Rad30 may indeed require Pol for its in vivo function.

The PRR and mutagenesis pathway appears to be highly conserved within eukaryotes; thus a model established in budding yeast likely applies to other eukaryotic organisms. Numerous homologs of the RAD6 pathway genes have been identified in various organisms. In particular, RAD6, POL30, MMS2, UBC13, and REV3 homologs have been reported (KOKEN et al. 1991 ; YAMAGUCHI et al. 1996 ; GIBBS et al. 1998 ; XIAO et al. 1998A , XIAO et al. 1998B ; JOHNSON et al. 1999A ; MASUTANI et al. 1999B ), some of which are able to functionally complement the corresponding yeast defects (KOKEN et al. 1991 ; XIAO et al. 1998B ). Furthermore, human cells or animals compromised for the yeast RAD6 group genes display phenotypes reminiscent of the corresponding yeast mutants (ROEST et al. 1996 ; GIBBS et al. 1998 ; JOHNSON et al. 1999A ; MASUTANI et al. 1999B ). It is of great interest to note that while human diseases have been linked to mutations in both nucleotide excision repair genes (FRIEDBERG et al. 1995 ) and recombination repair genes (CARNEY et al. 1998 ; VARON et al. 1998 ), as well as a UV-specific PRR gene, hRAD30 (JOHNSON et al. 1999A ; MASUTANI et al. 1999B ), there is yet no disease linked to mutations within other PRR pathway genes. This is not to say that the PRR pathway fails to contribute significantly to the protection of cells against DNA damage; on the contrary, the rad6 and rad18 mutants are as sensitive to different DNA-damaging agents as any of the other severe DNA repair mutants (LAWRENCE 1994 ; FRIEDBERG et al. 1995 ). It is possible that the PRR pathway is of such vital importance that mammalian cells may have developed additional mechanisms to prevent loss of such gene functions. For example, each of the RAD6 (KOKEN et al. 1991 ) and MMS2 (XIAO et al. 1998B ) genes has two mammalian homologs with >90% amino acid sequence identity and functional redundancy. Furthermore, the important DNA repair/tolerance genes may have been rendered essential by playing additional roles in mammalian cells. Such examples have been found with members involved in nucleotide excision repair (e.g., XPB and XPD, DE LAAT et al. 1999 ), recombination repair (e.g., hMRE11, hRAD50, and hRAD51, PAQUES and HABER 1999 ), and base excision repair (e.g., REF1, XANTHOUDAKIS et al. 1996 ) pathways and will probably be demonstrated with some PRR pathway genes as well. The elucidation of the yeast PRR and mutagenesis pathways will greatly facilitate the full understanding of this most challenging DNA damage tolerance pathway in eukaryotic cells and shed light on cancer and genetic diseases related to the genetic defects of this pathway.

	FOOTNOTES

¹ Present address: Gene-Cell, Inc., 1010 Hercules Ave., Houston, TX 77058.

	ACKNOWLEDGMENTS

We thank J. Brusky and T. Fontanie for technical assistance and T. Hryciw for helpful discussion. We also thank many researchers for the yeast strains and plasmids. This work was supported by the Medical Research Council of Canada operating grant MT-15076 to W.X. W.X. is a Research Scientist of the National Cancer Institute of Canada and S.B. is supported by University of Saskatchewan and College of Medicine Graduate Scholarships.

Manuscript received February 17, 2000; Accepted for publication April 21, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ABOUSSEKHRA, A., R. CHANET, Z. ZGAGA, C. CASSIER-CHAUVAT, and H. HEUDE et al., 1989  RADH, a gene of Saccharomyces cerevisiae encoding a putative DNA helicase involved in DNA repair. Characteristics of radH mutants and sequence of the gene. Nucleic Acids Res. 17:7211-7219[Abstract/Free Full Text].

AHNE, F., B. JHA, and F. ECKARDT-SCHUPP, 1997  The RAD5 gene product is involved in the avoidance of non-homologous end-joining of DNA double strand breaks in the yeast Saccharomyces cerevisiae.. Nucleic Acids Res. 25:743-749[Abstract/Free Full Text].

ALANI, E., L. CAO, and N. KLECKNER, 1987  A method for gene disruption that allows repeated use of URA3 selection in the construction of multiply disrupted yeast strains. Genetics 116:541-545[Abstract/Free Full Text].

AYYAGARI, R., K. J. IMPELLIZZERI, B. L. YODER, S. L. GARY, and P. M. BURGERS, 1995  A mutational analysis of the yeast proliferating cell nuclear antigen indicates distinct roles in DNA replication and DNA repair. Mol. Cell. Biol. 15:4420-4429[Abstract].

BAILLY, V., J. LAMB, P. SUNG, S. PRAKASH, and L. PRAKASH, 1994  Specific complex formation between yeast RAD6 and RAD18 proteins: a potential mechanism for targeting RAD6 ubiquitin-conjugating activity to DNA damage sites. Genes Dev. 8:811-820[Abstract/Free Full Text].

BAILLY, V., S. LAUDER, S. PRAKASH, and L. PRAKASH, 1997a  Yeast DNA repair proteins Rad6 and Rad18 form a heterodimer that has ubiquitin conjugating, DNA binding, and ATP hydrolytic activities. J. Biol. Chem. 272:23360-23365[Abstract/Free Full Text].

BAILLY, V., S. PRAKASH, and L. PRAKASH, 1997b  Domains required for dimerization of yeast Rad6 ubiquitin-conjugating enzyme and Rad18 DNA binding protein. Mol. Cell. Biol. 17:4536-4543[Abstract].

BERBEN, G., J. DUMONT, V. GILLIQUET, P. A. BOLLE, and F. HILGER, 1991  The YDp plasmids: a uniform set of vectors bearing versatile gene disruption cassettes for Saccharomyces cerevisiae.. Yeast 7:475-477[Medline].

BROOMFIELD, S., B. L. CHOW, and W. XIAO, 1998  MMS2, encoding a ubiquitin-conjugating-enzyme-like protein, is a member of the yeast error-free postreplication repair pathway. Proc. Natl. Acad. Sci. USA 95:5678-5683[Abstract/Free Full Text].

BRUSKY, J., Y. ZHU, and W. XIAO, 2000  UBC13, a DNA damage-inducible gene, is a member of the error-free postreplication repair pathway in Saccharomyces cerevisiae.. Curr. Genet. 37:168-174[Medline].

CARNEY, J. P., R. S. MASER, H. OLIVARES, E. M. DAVIS, and M. LE BEAU et al., 1998  The hMre11/hRad50 protein complex and Nijmegen breakage syndrome: linkage of double-strand break repair to the cellular DNA damage response. Cell 93:477-486[Medline].

DE LAAT, W. L., N. G. JASPERS, and J. H. HOEIJMAKERS, 1999  Molecular mechanism of nucleotide excision repair. Genes Dev. 13:768-785[Free Full Text].

DOHMEN, R. J., K. MADURA, B. BARTEL, and A. VARSHAVSKY, 1991  The N-end rule is mediated by the UBC2(RAD6) ubiquitin-conjugating enzyme. Proc. Natl. Acad. Sci. USA 88:7351-7355[Abstract/Free Full Text].

FABRE, F., N. MAGANA-SCHWENCKE, and R. CHANET, 1989  Isolation of the RAD18 gene of Saccharomyces cerevisiae and construction of rad18 deletion mutants. Mol. Gen. Genet. 215:425-430[Medline].

FRIEDBERG, E. C., 1988  Deoxyribonucleic acid repair in the yeast Saccharomyces cerevisiae.. Microbiol. Rev. 52:70-102[Free Full Text].

FRIEDBERG, E. C., G.C. WALKER and W. SIEDE, 1995 DNA Repair and Mutagenesis. ASM Press, Washington, DC.

GARY, R. D., D. L. LUDWIG, H. L. CORNELIUS, M. A. MACINNES, and M. S. PARK, 1997  The DNA repair endonuclease XPG binds to proliferating cell nuclear antigen (PCNA) and shares sequence elements with the PCNA-binding regions of FEN-1 and cyclin-dependent kinase inhibitor p21. J. Biol. Chem. 272:24522-24529[Abstract/Free Full Text].

GIBBS, P. E., W. G. MCGREGOR, V. M. MAHER, P. NISSON, and C. W. LAWRENCE, 1998  A human homolog of the Saccharomyces cerevisiae REV3 gene, which encodes the catalytic subunit of DNA polymerase zeta. Proc. Natl. Acad. Sci. USA 95:6876-6880[Abstract/Free Full Text].

GIOT, L., R. CHANET, M. SIMON, C. FACCA, and G. FAYE, 1997  Involvement of the yeast DNA polymerase delta in DNA repair in vivo. Genetics 146:1239-1251[Abstract].

HOFMANN, R. M. and C. M. PICKART, 1999  Noncanonical MMS2-encoded ubiquitin-conjugating enzyme functions in assembly of novel polyubiquitin chains for DNA repair. Cell 96:645-653[Medline].

HUANG, H., A. KAHANA, D. E. GOTTSCHLING, L. PRAKASH, and S. W. LIEBMAN, 1997  The ubiquitin-conjugating enzyme Rad6 (Ubc2) is required for silencing in Saccharomyces cerevisiae.. Mol. Cell. Biol. 17:6693-6699[Abstract].

JENTSCH, S., J. P. MCGRATH, and A. VARSHAVSKY, 1987  The yeast DNA repair gene RAD6 encodes a ubiquitin-conjugating enzyme. Nature 329:131-134[Medline].

JOHNSON, R. E., S. T. HENDERSON, T. D. PETES, S. PRAKASH, and M. BANKMANN et al., 1992  Saccharomyces cerevisiae RAD5-encoded DNA repair protein contains DNA helicase and zinc-binding sequence motifs and affects the stability of simple repetitive sequences in the genome. Mol. Cell. Biol. 12:3807-3818[Abstract/Free Full Text].

JOHNSON, R. E., S. PRAKASH, and L. PRAKASH, 1994  Yeast DNA repair protein RAD5 that promotes instability of simple repetitive sequences is a DNA-dependent ATPase. J. Biol. Chem. 269:28259-28262[Abstract/Free Full Text].

JOHNSON, R. E., G. K. KOVVALI, S. N. GUZDER, N. S. AMIN, and C. HOLM et al., 1996  Evidence for involvement of yeast proliferating cell nuclear antigen in DNA mismatch repair. J. Biol. Chem. 271:27987-27990[Abstract/Free Full Text].

JOHNSON, R. E., C. M. KONDRATICK, S. PRAKASH, and L. PRAKASH, 1999a  hRAD30 mutations in the variant form of xeroderma pigmentosum. Science 285:263-265[Abstract/Free Full Text].

JOHNSON, R. E., S. PRAKASH, and L. PRAKASH, 1999b  Efficient bypass of a thymine-thymine dimer by yeast DNA polymerase, Pol. Science 283:1001-1004[Abstract/Free Full Text].

JOHNSON, R. E., S. PRAKASH, and L. PRAKASH, 1999c  Requirement of DNA polymerase activity of yeast Rad30 protein for its biological function. J. Biol. Chem. 274:15975-15977[Abstract/Free Full Text].

KOKEN, M. H., P. REYNOLDS, I. JASPERS-DEKKER, L. PRAKASH, and S. PRAKASH et al., 1991  Structural and functional conservation of two human homologs of the yeast DNA repair gene RAD6.. Proc. Natl. Acad. Sci. USA 88:8865-8869[Abstract/Free Full Text].

LARIMER, F. W., J. R. PERRY, and A. A. HARDIGREE, 1989  The REV1 gene of Saccharomyces cerevisiae: isolation, sequence, and functional analysis. J. Bacteriol. 171:230-237[Abstract/Free Full Text].

LAWRENCE, C., 1994  The RAD6 DNA repair pathway in Saccharomyces cerevisiae: what does it do, and how does it do it? Bioessays 16:253-258[Medline].

LAWRENCE, C. W. and R. B. CHRISTENSEN, 1979  Metabolic suppressors of trimethoprim and ultraviolet light sensitivities of Saccharomyces cerevisiae rad6 mutants. J. Bacteriol. 139:866-876[Abstract/Free Full Text].

LAWRENCE, C. W. and D. C. HINKLE, 1996  DNA polymerase and the control of DNA damage induced mutagenesis in eukaryotes. Cancer Surv. 28:21-31[Medline].

LEE, S. H., Z. Q. PAN, A. D. KWONG, P. M. BURGERS, and J. HURWITZ, 1991  Synthesis of DNA by DNA polymerase epsilon in vitro. J. Biol. Chem. 266:22707-22717[Abstract/Free Full Text].

LEMONTT, J. F., 1971  Mutants of yeast defective in mutation by ultraviolet light. Genetics 68:21-33[Free Full Text].

LEMONTT, J. F., 1972  Induction of forward mutations in mutationally defective yeast. Mol. Gen. Genet. 119:27-42[Medline].

LI, X., J. LI, J. HARRINGTON, M. R. LIEBER, and P. M. BURGERS, 1995  Lagging strand DNA synthesis at the eukaryotic replication fork involves binding and stimulation of FEN-1 by proliferating cell nuclear antigen. J. Biol. Chem. 270:22109-22112[Abstract/Free Full Text].

MASUTANI, C., M. ARAKI, A. YAMADA, R. KUSUMOTO, and T. NOGIMORI et al., 1999a  Xeroderma pigmentosum variant (XP-V) correcting protein from HeLa cells has a thymine dimer bypass DNA polymerase activity. EMBO J. 18:3491-3501[Medline].

MASUTANI, C., R. KUSUMOTO, A. YAMADA, N. DOHMAE, and M. YOKOI et al., 1999b  The XPV (xeroderma pigmentosum variant) gene encodes human DNA polymerase . Nature 399:700-704[Medline].

MCDONALD, J. P., A. S. LEVINE, and R. WOODGATE, 1997  The Saccharomyces cerevisiae RAD30 gene, a homologue of Escherichia coli dinB and umuC, is DNA damage inducible and functions in a novel error-free postreplication repair mechanism. Genetics 147:1557-1568[Abstract].

MCKEE, R. H. and C. W. LAWRENCE, 1979  Genetic analysis of gamma-ray mutagenesis in yeast. II. Allele-specific control of mutagenesis. Genetics 93:375-381[Abstract/Free Full Text].

MONTELONE, B. A., S. PRAKASH, and L. PRAKASH, 1981  Recombination and mutagenesis in rad6 mutants of Saccharomyces cerevisiae: evidence for multiple functions of the RAD6 gene. Mol. Gen. Genet. 184:410-415[Medline].

MORRISON, A., R. B. CHRISTENSEN, J. ALLEY, A. K. BECK, and E. G. BERNSTINE et al., 1989  REV3, a Saccharomyces cerevisiae gene whose function is required for induced mutagenesis, is predicted to encode a nonessential DNA polymerase. J. Bacteriol. 171:5659-5667[Abstract/Free Full Text].

NELSON, J. R., C. W. LAWRENCE, and D. C. HINKLE, 1996a  Deoxycytidyl transferase activity of yeast REV1 protein. Nature 382:729-731[Medline].

NELSON, J. R., C. W. LAWRENCE, and D. C. HINKLE, 1996b  Thymine-thymine dimer bypass by yeast DNA polymerase . Science 272:1646-1649[Abstract].

PAQUES, F. and J. E. HABER, 1999  Multiple pathways of recombination induced by double-strand breaks in Saccharomyces cerevisiae. Microbiol. Mol. Biol. Rev. 63:349-404[Abstract/Free Full Text].

PRELICH, G., C. K. TAN, M. KOSTURA, M. B. MATHEWS, and A. G. SO et al., 1987  Functional identity of proliferating cell nuclear antigen and a DNA polymerase-delta auxiliary protein. Nature 326:517-520[Medline].

ROBZYK, K., J. RECHT, and M. A. OSLEY, 2000  Rad6-dependent ubiquitination of histone H2B in yeast. Science 287:501-504[Abstract/Free Full Text].

ROEST, H. P., J. VAN KLAVEREN, J. DE WIT, C. G. VAN GURP, and M. H. KOKEN et al., 1996  Inactivation of the HR6B ubiquitin-conjugating DNA repair enzyme in mice causes male sterility associated with chromatin modification. Cell 86:799-810[Medline].

ROTHSTEIN, R. J., 1983  One-step gene disruption in yeast. Methods Enzymol. 101:202-211[Medline].

ROUSH, A. A., M. SUAREZ, E. C. FRIEDBERG, M. RADMAN, and W. SIEDE, 1998  Deletion of the Saccharomyces cerevisiae gene RAD30 encoding an Escherichia coli DinB homolog confers UV radiation sensitivity and altered mutability. Mol. Gen. Genet. 257:686-692[Medline].

SHERMAN, F., G. R. FINK and J. HICKS, 1983 Methods in Yeast Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY.

SUNG, P., E. BERLETH, C. PICKART, S. PRAKASH, and L. PRAKASH, 1991  Yeast RAD6 encoded ubiquitin conjugating enzyme mediates protein degradation dependent on the N-end-recognizing E3 enzyme. EMBO J. 10:2187-2193[Medline].

TANG, M., X. SHEN, E. G. FRANK, M. O'DONNELL, and R. WOODGATE et al., 1999  UmuD'(2)C is an error-prone DNA polymerase, Escherichia coli pol V. Proc. Natl. Acad. Sci. USA 96:8919-8924[Abstract/Free Full Text].

TORRES-RAMOS, C. A., B. L. YODER, P. M. BURGERS, S. PRAKASH, and L. PRAKASH, 1996  Requirement of proliferating cell nuclear antigen in RAD6-dependent postreplicational DNA repair. Proc. Natl. Acad. Sci. USA 93:9676-9681[Abstract/Free Full Text].

TORRES-RAMOS, C. A., S. PRAKASH, and L. PRAKASH, 1997  Requirement of yeast DNA polymerase in post-replicational repair of UV-damaged DNA. J. Biol. Chem. 272:25445-25448[Abstract/Free Full Text].

UMAR, A., A. B. BUERMEYER, J. A. SIMON, D. C. THOMAS, and A. B. CLARK et al., 1996  Requirement for PCNA in DNA mismatch repair at a step preceding DNA resynthesis. Cell 87:65-73[Medline].

VARON, R., C. VISSINGA, M. PLATZER, K. M. CEROSALETTI, and K. H. CHRZANOWSKA et al., 1998  Nibrin, a novel DNA double-strand break repair protein, is mutated in Nijmegen breakage syndrome. Cell 93:467-476[Medline].

WAGNER, J., P. GRUZ, S. R. KIM, M. YAMADA, and K. MATSUI et al., 1999  The dinB gene encodes a novel E. coli DNA polymerase, DNA pol IV, involved in mutagenesis. Mol. Cell 4:281-286[Medline].

WATKINS, J. F., P. SUNG, S. PRAKASH, and L. PRAKASH, 1993  The extremely conserved amino terminus of RAD6 ubiquitin-conjugating enzyme is essential for amino-end rule-dependent protein degradation. Genes Dev. 7:250-261[Abstract/Free Full Text].

WU, X., J. LI, X. LI, C. L. HSIEH, P. M. BURGERS, and M. R. LIEBER, 1996  Processing of branched DNA intermediates by a complex of human FEN-1 and PCNA. Nucleic Acids Res. 24:2036-2043[Abstract/Free Full Text].

XANTHOUDAKIS, S., R. J. SMEYNE, J. D. WALLACE, and T. CURRAN, 1996  The redox/DNA repair protein, Ref-1, is essential for early embryonic development in mice. Proc. Natl. Acad. Sci. USA 93:8919-8923[Abstract/Free Full Text].

XIAO, W., T. LECHLER, B. L. CHOW, T. FONTANIE, and M. AGUSTUS et al., 1998a  Identification, chromosomal mapping and tissue-specific expression of hREV3 encoding a putative human DNA polymerase . Carcinogenesis 19:945-949[Abstract/Free Full Text].

XIAO, W., S. L. LIN, S. BROOMFIELD, B. L. CHOW, and Y. F. WEI, 1998b  The products of the yeast MMS2 and two human homologs (hMMS2 and CROC-1) define a structurally and functionally conserved Ubc-like protein family. Nucleic Acids Res. 26:3908-3914[Abstract/Free Full Text].

XIAO, W., B. L. CHOW, T. FONTANIE, L. MA, and S. BACCHETTI et al., 1999  Genetic interactions between error-prone and error-free postreplication repair pathways in Saccharomyces cerevisiae.. Mutat. Res. 435:1-11[Medline].

YAMAGUCHI, T., N. S. KIM, S. SEKINE, H. SEINO, and F. OSAKA et al., 1996  Cloning and expression of cDNA encoding a human ubiquitin-conjugating enzyme similar to the Drosophila bendless gene product. J. Biochem. 120:494-497[Abstract/Free Full Text].

This article has been cited by other articles:

				P. Langerak, A. O.H. Nygren, P. H.L. Krijger, P. C.M. van den Berk, and H. Jacobs A/T mutagenesis in hypermutated immunoglobulin genes strongly depends on PCNAK164 modification J. Exp. Med., August 6, 2007; 204(8): 1989 - 1998. [Abstract] [Full Text] [PDF]

T. Fukushima, S.-i. Matsuzawa, C. L. Kress, J. M. Bruey, M. Krajewska, S. Lefebvre, J. M. Zapata, Z. Ronai, and J. C. Reed Ubiquitin-conjugating enzyme Ubc13 is a critical component of TNF receptor-associated factor (TRAF)-mediated inflammatory responses PNAS, April 10, 2007; 104(15): 6371 - 6376. [Abstract] [Full Text] [PDF]