Structure/Function Analysis of the Saccharomyces cerevisiae Trf4/Pol DNA Polymerase

Corresponding author: Michael F. Christman, E603, Boston University Medical Center, 715 Albany St., Boston, MA 02118., mfc{at}bu.edu (E-mail)

Communicating editor: M. LICHTEN

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The Trf4p/Pol DNA polymerase (formerly Trf4p/Pol ) couples DNA replication to the establishment of sister chromatid cohesion. The polymerase is encoded by two redundant homologs in Saccharomyces cerevisiae, TRF4 and TRF5, that together define a fourth essential nuclear DNA polymerase in yeast and probably in all eukaryotes. Here we present a thorough genetic analysis of the founding member of this novel family of DNA polymerases, TRF4. Analyses of mutants carrying 1 of 34 "surface-targeted" alanine scanning mutations in TRF4 have identified those regions required for Pol 's essential function, for its role in DNA double-strand break repair, and for its association with chromosomes. The data strongly support the importance of the regions of predicted structural similarity with the Pol ß superfamily as critical for Trf4p/Pol 's essential and repair functions. Surprisingly, five lethal mutations lie outside all polymerase homology in a C-terminal region. The protein possesses Mg²⁺-dependent 3' to 5' exonuclease activity. Cell cycle analysis reveals that Trf4p/Pol associates with chromosomes in G1, S, and G2 phases, but that association is abolished coincident with dissolution of cohesion at the metaphase-to-anaphase transition.

IN eukaryotes, sister chromatids are held together during or shortly after their formation until they separate into mother and daughter cells during anaphase. Sister chromatid cohesion is established coincident with DNA replication (UHLMANN and NASMYTH 1998 ) in a process that requires Eco1p (Ctf7p) and Scc1p (Mcd1p; SKIBBENS et al. 1999 ; TOTH et al. 1999 ). Loading of cohesins onto chromosomes occurs in late G1 (CIOSK et al. 2000 ), meaning that as replication forks move along chromosomes they will encounter cohesin complexes. We have shown that a novel DNA polymerase, Pol , is necessary to establish cohesion during S phase (Z. WANG et al. 2000 ; CARSON and CHRISTMAN 2001 ). Thus, Pol may directly couple replication to the action of cohesion factors at replication forks (TAKAHASHI and YANAGIDA 2000 ). We have proposed a polymerase switch model for cohesion establishment in which Pol replaces replicative polymerases upon encounter of the fork with a precohesion site (Z. WANG et al. 2000 ; CARSON and CHRISTMAN 2001 ). Polymerase switching from Pol to Pol on Okazaki fragments is thought to be accomplished by the replication factor C (RFC) complex (WAGA and STILLMAN 1994 ). Recent data indicate that a modified RFC complex, in which CTF18 replaces RFC1, is required for cohesion (HANNA et al. 2001 ), lending strong support to the model.

TRF4/Pol is conserved in all eukaryotes and we have shown that the human enzyme possesses DNA polymerase activity in vitro (Z. WANG and M. F. CHRISTMAN, unpublished observation). Pol represents the fourth essential nuclear DNA polymerase, in addition to Pol , Pol , and Pol , in yeast and probably in all eukaryotes. While mutation of both TRF4 and TRF5 is lethal in Saccharomyces cerevisiae (CASTANO et al. 1996 ), a trf4 single mutant displays a defect in DNA repair (WALOWSKY et al. 1999 ). The relationship of the repair function of TRF4 to its essential function is unknown at present. A Schizosaccharomyces pombe homolog, cid1, appears to play a role in the S-phase checkpoint (S. W. WANG et al. 2000 ).

Analysis of the highly conserved TRF4 gene family has led to the prediction that the TRF4/Pol gene family is distantly related to the ß-polymerase superfamily (ARAVIND and KOONIN 1999 ), which consists of proteins that catalyze a variety of nucleotidyltransferase reactions, including DNA synthesis. The homology was discovered using iterative database searches combined with computer modeling based on the crystal structure of known ß-polymerase family members (ARAVIND and KOONIN 1999 ). Homology is not apparent in standard "BLAST" database searches. It is important to determine the significance of this predicted structural similarity since some small molecule inhibitors of Pol ß have antitumor activity (HSU et al. 1997 ; DENG et al. 1999 ). In addition, a genetic analysis will allow us to better define those regions of TRF4/Pol required for its various roles in cohesion and DNA repair.

Here we present a thorough genetic analysis of TRF4/Pol . The data strongly support the prediction that regions of structural similarity to the ß-polymerase superfamily are important for TRF4/Pol function. Furthermore, the analysis defines a C-terminal domain that is also essential for TRF4/Pol function.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Site-directed mutagenesis:
The TRF4 gene was cloned into a pALTER-derived phagmid vector (Promega, Madison, WI) containing the HIS3 gene (CB905) and site-directed mutagenesis was carried out using the Altered Sites II in vitro mutagenesis system (Promega). The trf4 alleles produced by this method were sequenced (Biomolecular Research Facility, University of Virginia) to confirm the presence of the desired mutation. Sequences of mutagenic oligos are available upon request.

Integration of trf4 mutant alleles by one-step gene replacement:
Fragments containing mutant trf4 alleles and the HIS3 marker were isolated by digesting the mutated plasmid with MluI and NotI restriction enzymes. The linear fragments were transformed into the recipient strain CY1035 (MATa trf4:: TRP1 trf5::LEU2 pTRF5.URA3). Integrants that replace the parent trf4::TRP1 allele with the trf4 surface-targeted allele linked to HIS3 were selected on SC-his plates. Colonies were then purified on SC-his plates and tested for their growth on SC-trp medium to confirm that the parental trf4::TRP1 allele had been replaced. Mutant alleles that are integrated correctly at the normal TRF4 locus should be His⁺ and Trp^- since the trf4 allele linked to HIS3 should have replaced the trf4::TRP1 allele on chromosome XV (Fig 1A). The genomic structure of the candidate mutant strains was further confirmed by PCR analysis (not shown).

View larger version (49K): In this window In a new window Download PPT slide   Figure 1. Construction of Trf4/Pol "surface-targeted" mutants: (A) The scheme used to integrate each of the trf4 surface-targeted mutant alleles into the natural TRF4 chromosomal locus. This was performed by transformation of CY1187 with a linear DNA fragment (MluI-NotI) carrying the trf4 mutant allele and the HIS3 gene cloned 3' to the TRF4 open reading frame (Table 1). His+ transformants were isolated and screened to identify those that had simultaneously become Trp-. Correct genomic structure was confirmed by PCR analysis of genomic DNA from candidate transformants using primers that flank the TRF4 locus. (B) CY1187 (TRF4::HIS3), CY1035 (trf4:: TRP1), CY1291 (trf4-282::HIS3), and CY1297 (trf4-332::HIS) were streaked onto SC-his plates at 30° and 5-FOA plates at both 30° and 37° and incubated for 3 days. Only those cells that lost the URA3 plasmid were able to grow on 5-FOA-containing medium (BOEKE et al. 1984 ). (C) Strains carrying the following mutant alleles of TRF4: trf4-182 (CY1193), trf4-194 (CY1194), trf4-224 (CY1625), trf4-282 (CY1291), trf4-309 (CY1296), trf4-378 (CY1264), trf4-425 (CY1248), trf4-444 (CY1295), trf4-486 (CY1196), trf4-491 (CY1270), trf4-502 (CY1293), trf4-508 (CY1294), and wild-type (CY184) strains were grown to log phase and cell extracts were made. Western blots were performed to detect Trf4p using a polyclonal antibody to the C-terminal 15 amino acids.

Western blotting:
Yeast were grown to log phase in YPD medium. Cell extracts were made by glass-bead disruption (DUNN and WOBBE 1997 ), run on an 8% SDS-PAGE gel, and transferred to a nitrocellulose membrane. The nitrocellulose membrane was blocked with 5% milk- Tris-borate SDS (TBS) and incubated for 1 hr with a rabbit anti-Trf4p polyclonal antibody, followed by a 1-hr incubation with secondary antibody. The membrane was washed three times with TBS before and after adding the second antibody. Immunoreactive bands were visualized by enhanced chemiluminescence (Amersham, Buckinghamshire, UK).

Exonuclease activity assays:
Cofractionation with polymerase: DNA polymerase assays were performed with 1-µl aliquots of each fraction. The oligo(dT) primer was end labeled using T4 polynucleotide kinase. Aliquots were incubated with 100 nM oligo(dT)/poly(dA) [20:1 molar excess of poly(dA): oligo(dT)] for 5 min at 37° in 50 mM Tris-HCl pH 8.0 (pH at 22°), 10 mM MgCl₂, 2 mM dithiothreitol, 20 mM NaCl, 20 mM KCl, 2.5% glycerol, 0.2 mg/ml BSA, and 1 mM deoxy thymidine triphosphate (dTTP).

Metal dependency: A 35-mer primer was 5' labeled with ³²P and hybridized with a 75-mer template to form the exonuclease substrate as described in Z. WANG et al. 2000 . Exonuclease reactions were performed for 30 min at 37° in 50 mM Tris-HCl pH 8.0 (pH at 22°), 10 mM MgCl₂, 2 mM dithiothreitol, 20 mM NaCl, 20 mM KCl, 2.5% glycerol, 0.2 mg/ml BSA, and 1 mM dTTP. Reaction products were resolved on an 8 M urea/20% acrylamide gel.

Chromosome spreads:
Chromosome spreads were performed as described (MICHAELIS et al. 1997 ) with the following minor modifications. Cells were grown in YPD to early log phase and 10 ml of the culture was harvested, washed in spheroplasting solution (1.2 M sorbitol, 0.1 M potassium phosphate pH 7.4, 0.5 M MgCl₂), and then digested in the same solution containing 0.5% 2-mercaptoethanol and 50 µg/µl zymolyase 100-T (ICN Biomedicals) at 23° for 10 min. The digestion was stopped and the cells were spread as described (MICHAELIS et al. 1997 ).

Tubulin immunofluorescence:
Mitotic spindles were visualized by immunofluorescence as previously described (CASTANO et al. 1996 ), using an antibody to yeast -tubulin, YOL1/34 (Serotec, Oxford), and a Texas red-conjugated secondary antibody. Cells were viewed using a Nikon Eclipse 800 epifluorescence microscope. Images were captured digitally using a Princeton Instruments camera and IP Lab Spectrum software.

GFP chromosome tagging assay for sister chromatid cohesion:
Sister chromatid cohesion was monitored essentially as described in STRAIGHT et al. 1996 . Yeast strains containing the lacO repeat integrated at the LEU2 locus, and the GFP-lacI fusion under the HIS3 promoter integrated at the HIS3 locus, were grown overnight in SC-histidine at 30°. Cells were diluted to an A₆₀₀ of 0.2 and allowed to double once more. Cells were spun and resuspended in the same volume of YPD plus 20 µg/ml of nocodazole or 25 µg/ml of -factor. After 4 hr, cells were fixed by adding 0.1 volume of 37% formaldehyde and incubated for 5 min. Cells were then collected by centrifugation in a microfuge and washed three times with water. The cell suspension was diluted and sonicated briefly, cells were then adhered to polylysine-coated slides, and 5 µl of antifade solution with 4',6-diamidino-2-phenylindole (DAPI) was added to each well. Slides were viewed and images were captured as described above. A GFP LP filter from Chroma Technology (Brattleboro, VT) was used to visualize the GFP signal.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Systematic construction of "surface-targeted" mutations in TRF4:
On the basis of iterative database searching and structural modeling (ARAVIND and KOONIN 1999 ) it has been proposed that the Trf4p/Pol DNA polymerase family of proteins is distantly related to the ß-polymerase superfamily of nucleotidyltransferases. However, virtually no primary amino acid sequence similarity is evident upon alignment of the two sequences despite the fact that both encode DNA polymerases. To determine the importance of these regions to TRF4 function, we constructed 34 "surface-targeted" mutations in TRF4 and examined their phenotypes.

Surface-targeted mutagenesis is based on the assumption that clusters of charged residues are likely to reside on a protein's surface and not in the hydrophobic interior (WERTMAN et al. 1992 ). "Surface regions" of a protein are predicted by screening the primary sequence of amino acids for the presence of two or more charged residues in a window of any five amino acids. Each of the charged residues in a predicted surface region is then changed to alanine by site-directed mutagenesis. Mutagenic oligos were synthesized and site-directed mutagenesis was performed to make the 34 mutant alleles of TRF4 that alter predicted surface regions (Table 1 and Fig 2). Each allele was then integrated into the natural TRF4 locus by single-step gene replacement made possible by the presence of a selectable HIS3 gene inserted 84 nucleotides 3' to the TRF4 stop codon (Fig 1A). Western blots confirmed that mutant strains produced normal levels of Trf4p (Fig 1C). Mutants carrying each of the 34 trf4 mutant alleles were then examined in both TRF5 and trf5 mutant backgrounds for a variety of phenotypes including cell viability, double-strand break (DSB) repair capacity, sister chromatid cohesion, and chromosome association. The most amino- terminal amino acid mutated in a given charge cluster on Trf4p was used to designate the allele number.

View larger version (29K): In this window In a new window Download PPT slide   Figure 2. Phenotypes and locations of surface-targeted mutant alleles in TRF4: The position of each of the 34 surface-targeted mutations in TRF4 is marked with the allele number. Each allele number corresponds to the most amino-terminal residue that is mutated to alanine. The specific residues changed to alanine are shown in red in the primary sequence. The allele numbers are highlighted in green for those that do not affect viability in a trf5 background, yellow for those that cause a growth defect or temperature sensitivity, and red for those that are inviable in the absence of TRF5. Mutations that cause CPT hypersensitivity in a TRF5 background are designated as DSBs. The eight regions of predicted structural similarity between Trf4p/Pol and the Pol ß superfamily are underlined in blue and marked with Arabic numerals. The five lethal mutations clustered outside the polymerase domains are underlined and labeled "C-terminal domain."

The ß-polymerase-like domains in TRF4 are essential for its function:
We have previously shown that a null mutation in trf4 is lethal in combination with a trf5 null mutation (CASTANO et al. 1996 ). The predicted Trf4p and Trf5p proteins are 55% identical and 72% similar and clearly act as redundant homologs. Thus, we examined the viability of the 34 surface-targeted mutant alleles of TRF4 in a trf5 background to define those regions of TRF4 needed for its essential function.

To facilitate construction of yeast strains carrying mutant alleles of trf4 at the normal TRF4 locus in a trf5 background, we generated a trf4::TRP1 trf5::LEU2 double-mutant strain covered with a TRF5.URA3 plasmid to make the strain viable (CY1035, shown schematically in Fig 1A). Transformants that replace the parent trf4-TRP1 allele with the trf4 surface-targeted allele linked to HIS3 were selected as described in MATERIALS AND METHODS.

The viability of mutants carrying a trf4 allele was then examined by streaking three independent colonies of each mutant on SC plates containing (5-fluoroorotic acid 5-FOA) to select cells that have spontaneously lost the pTRF5.URA3 cover plasmid (BOEKE et al. 1984 ). Thus, the inability of an integrant to give rise to 5-FOA- resistant segregants indicates that the mutant allele confers a lethal phenotype.

Examples of the analysis of 2 of the 34 alleles are shown in Fig 1B. A positive control with a wild-type TRF4 locus readily yields 5-FOA-resistant segregants and a negative control carrying a trf4 null mutation fails to yield 5-FOA-resistant segregants as expected (Fig 1B). The three independent trf4-282 transformants are His⁺, but are unable to give rise to 5-FOA-resistant segregants at either 30° or 37°. This indicates that the trf4-282 mutation abolishes the essential function of TRF4 at both 30° and 37°. In contrast, the trf4-332 mutant does yield slow-growing 5-FOA-resistant segregants at 30°, indicating that it causes a growth defect but does not eliminate the essential function. However, at 37° the trf4-332 mutant does not yield 5-FOA-resistant segregants. Thus, trf4-332 confers a growth defect at 30° and a temperature-sensitive inviability at 37°. A similar analysis was performed for all 34 mutant alleles in TRF4.

Of the 34 alleles that were integrated at the TRF4 locus in a trf5 mutant background, 13 are inviable, 8 display temperature-sensitive growth or a growth defect as observed by significantly slower colony formation compared to the wild-type parent, and 13 have no obvious growth defect. Western blot analysis of the 13 inviable mutants demonstrates that each produces wild-type levels of Trf4p (Fig 1C). Thus, the mutants are defective for TRF4 function and not simply for protein stability.

The eight regions of predicted structural similarity between the ß-polymerase superfamily and TRF4 (ARAVIND and KOONIN 1999 ) are underlined and numbered in Fig 2 to show the positions of these regions with respect to the surface-targeted mutations in TRF4. Ten mutations lie in the eight ß-polymerase-like regions. Of these, 7 are inviable and 3 show growth defects. In contrast, of the 5 mutations that lie between, but not within, the eight ß-polymerase-like regions, 4 have no obvious phenotype and 1 is inviable. This striking contrast clearly shows the importance of the predicted Pol ß-like domains for the essential function of TRF4/Pol .

The C terminus and central polymerase domain are highly evolutionarily conserved, whereas the N terminus is not. Of the eight mutations constructed in the N-terminal region, all remain viable in the absence of TRF5: Six have no obvious phenotype and two show a growth defect. In contrast, the conserved C-terminal region (amino acids 436–584), which lies outside the polymerase domain, is critical for the common essential function between TRF4 and TRF5, because five of the inviable mutant alleles occur in this region.

The DNA polymerase and C-terminal domains are required for TRF4's role in DNA damage repair:
A null mutant strain of trf4 is hypersensitive to the DNA-damaging agents camptothecin (CPT) and methyl methanesulfonate (MMS) even in the presence of TRF5 (WALOWSKY et al. 1999 ; Fig 3). Both CPT and MMS induce S-phase-specific DNA lesions that cause subsequent collapse of approaching replication forks (HSAING et al. 1989 ; LIU 1989 ; SCHWARTZ 1989 ; NITISS and WANG 1996 ; TERCERO and DIFFLEY 2001 ). Since DSB repair pathways are required for resistance to MMS and the repair of replication-fork-associated DSB caused by CPT (XIAO et al. 1996 ; ARNAUDEAU et al. 2001 ), it is likely that TRF4 is necessary for DSB repair. To determine whether the surface-targeted mutations are defective for this repair, all of the trf4 mutant alleles were integrated into the natural TRF4 locus in a TRF5 background, and mutant strains were assayed for their sensitivity to CPT (Fig 3). Mutant strains were grown to log phase and serial 10-fold dilutions were made on plates containing CPT.

View larger version (66K): In this window In a new window Download PPT slide   Figure 3. Regions of Trf4p/Pol are required for repair. trf4 (CY1000), trf4-236 (CY1262), trf4-182 (CY1193), trf4-194 (CY1194), trf4-378 (CY1264), trf4-502 (CY1293), trf4-508 (CY1294), and wild-type (CY184) strains were grown to mid-log phase in YPD and serial 10-fold dilutions were spotted onto plates containing 10 µg/ml camptothecin or with no drug. Plates were incubated at 30° for 3 days.

The mutants show a spectrum of repair capacity. For example, the trf4-182 mutant is defective for DNA repair, whereas the trf4-502 allele is not. Analysis of all 34 mutants indicates that the regions of Trf4p/Pol required for CPT resistance are a subset of the essential regions of the gene. Of the 13 alleles that are inviable in combination with trf5, 9 of these are repair defective in a TRF5 background (at least 10-fold more sensitive to CPT than the wild-type parent strain). In contrast, none of the 21 viable mutants have a repair defect. Alleles that are repair defective are marked "DSB^s" in Fig 2. These mutations reside in both the DNA polymerase and C-terminal domains, but not in the N-terminal domain.

Crystal structures of ß-polymerase superfamily proteins indicate that the conserved "GS...DXD" motif forms a metal-binding site that is crucial for their polymerase activity (PELLETIER et al. 1994 ; SAWAYA et al. 1994 ). We mutated the DXD motif (trf4-236) and found that it greatly diminishes the DNA polymerase activity in vitro (Z. WANG et al. 2000 ). The data in Fig 3 show that the trf4-236 mutant is camptothecin sensitive, demonstrating that polymerase activity is required for Trf4p/Pol 's repair function. Consistent with this, another polymerase motif mutant, trf4-224, which mutates the GS residues to alanine, is also repair defective (data not shown). In addition, another four mutants within the polymerase domain are repair defective, suggesting that the DNA polymerase domain plays a crucial role in TRF4's DNA repair function.

However, the DNA polymerase domain alone is not sufficient for DNA repair, because the trf4-444, trf4-491, and trf4-508 mutants, which alter amino acids outside the polymerase domain, are also repair defective. Each of these alleles resides in the C-terminal domain, further demonstrating the importance of this region in TRF4 function.

Trf4p/Pol displays 3' to 5' exonuclease activity in vitro:
During purification of recombinant Trf4p from Escherichia coli we observed consistent cofractionation of an exonuclease activity through the final purification step. In Fig 4A, lanes 6 and 7, cofractionation of Trf4p with DNA polymerase activity and a truncated 5'-end-labeled primer (arrow at bottom of Fig 4A) are observed. To investigate this further, assays of purified recombinant protein were performed using a substrate consisting of a 75-mer annealed to a complementary, 5'-end-labeled 35-mer primer. These assays show that incubation of purified Trf4p with the oligo/labeled primer in the absence of dNTPs results in removal of nucleotides from the 3' end of the primer (Fig 4B). Thus, Trf4p possesses 3' to 5' exonuclease activity. The activity does not require ATP (lane 4), but is completely dependent on exogenous Mg²⁺ (lanes 4 and 7). Neither Mn²⁺ (lane 5) nor Ca²⁺ (lane 6) are effective substitutes for Mg²⁺ in the assay.

View larger version (46K): In this window In a new window Download PPT slide   Figure 4. Trf4/Pol displays Mg2+-dependent exonuclease activity in vitro. (A, top) A Coomassie-stained SDS-PAGE gel of the Trf4p fractions eluted from a Mono Q column following ammonium sulfate and Ni-agarose affinity chromatography (Z. WANG et al. 2000 ). The oligo(dT) primer is the dark band near the bottom of the gel while a truncated primer (arrow), the product of the exonuclease activity, is observed just below full-length oligo(dT) primer. (B) A 35-mer primer was 5' labeled with 32P and hybridized with a 75-mer template as the substrate as in Z. WANG et al. 2000 . Trf4p was added to lanes 3–7; Mg2+ was added to lanes 2, 4, and 7; Mn2+ was added to lane 5; and Ca2+ was added to lane 6; and 1 mM ATP was added to lane 7. Reaction products were resolved on an 8 M urea/20% acrylamide gel and autoradiography was performed.

Cell-cycle-regulated association of Trf4p/Pol with chromosomes:
To begin to understand whether Trf4p/Pol associates with other chromosome-bound proteins, we sought to identify the regions of Trf4p/Pol required for its chromosome localization. Association of nuclear proteins with chromosomes can be monitored using the chromosome spread method (MICHAELIS et al. 1997 ). Cells are partially fixed, lysed, and then extensively washed such that all cytoplasm and nucleoplasm are removed. Under these conditions only chromatin-associated proteins remain and their presence can be detected by immunostaining.

To monitor Trf4p/Pol chromosome association, we used a strain that expresses Trf4p epitope tagged with two repeats of the IgG-binding domain of protein A, referred to as "TRF4-ZZ" (CY1344), and visualized the protein using FITC-coupled IgG. The association of the wild-type protein with chromosomes is readily detected as expected (Fig 5A). Negative controls included a strain with an untagged TRF4 gene that carried the ZZ vector (CY1346) and a strain carrying cytoplasmically located Pho4-ZZ mutant protein (CY1349). Chromosome spreads were then performed on the nuclei from the 13 trf4 charge-to-alanine mutants that we knew to be lethal in combination with a TRF5 deletion. To do this, each allele was fused in frame to the ZZ epitope tag either at the natural TRF4 locus or at the TRP1 locus in a trf4 deletion background, so that chromosome association could be monitored (MATERIALS AND METHODS and Table 2).

View larger version (38K): In this window In a new window Download PPT slide   Figure 5. Trf4p/Pol binds to chromosomes. (A) TRF4-ZZ (CY1344), ZZ-vector (CY1346), and PHO4-ZZ strains (CY1349) were grown to log phase and then fixed to perform chromosome spreads. The ZZ-tagged protein was visualized with FITC-conjugated rabbit IgG. DNA was visualized with DAPI. (B) A small region of Trf4p is required for chromosomal association. The chromosome spread assay was performed as in A. The trf4-182-ZZ (CY1435) and trf4-194-ZZ (CY1436) mutants fail to associate with chromosome spreads, whereas the trf4-491-ZZ mutant (CY1432) remains associated.

Of the 13 mutant Trf4 proteins examined, 11 retain their ability to bind to chromosomes. However, the trf4-182-ZZ and trf4-194-ZZ alleles produced proteins that completely failed to associate with chromosomes in chromosome spreads (Fig 5B) despite the fact that the proteins are made at normal levels (Fig 1C). Thus, a relatively discrete region of TRF4 is critical for its binding to chromosomes. Intriguingly, this region flanks the only cyclin-dependent kinase consensus phosphorylation site in the protein at serine 191, suggesting that phosphorylation of Trf4p may be important for its chromosome association. Indeed, an activating mutation in the yeast cyclin-dependent kinase CDC28-Y19F suppresses the temperature-sensitive growth defect of a trf4-ts trf5 strain (I. B. CASTAÑO, C. ADAMS and M. F. CHRISTMAN, unpublished observation), but the mechanism of suppression remains to be determined.

To determine whether Trf4p is chromosome associated at all points in the cell cycle, we monitored the association of Trf4p with chromosome spreads throughout the course of a single synchronous cell cycle. G1 daughter cells obtained by centrifugal elutriation were released into fresh medium and monitored every 30 min thereafter for the presence of Trf4p on chromosome spreads, for nuclear morphology, and for spindle morphology (Fig 6A). Association of Trf4p with chromosomes begins 30 min after release, becomes stronger at 60 min (onset of DNA synthesis; not shown), and continues through 90 min. At 120 min after release the metaphase-to-anaphase transition takes place as evidenced by the presence of highly elongated anaphase spindles. Coincident with the metaphase-to-anaphase transition, Trf4p's association with chromosome spreads is abolished.

View larger version (96K): In this window In a new window Download PPT slide   Figure 6. Association of Trf4p/Pol with chromosomes is cell cycle regulated. (A) G1 cells of strain CY1344 expressing TRF4-ZZ were collected by elutriation and released into YPD at 30°. At the indicated times, samples were collected and fixed for chromosome spreads and tubulin immunofluorescence. (B, top) A strain containing TRF4-ZZ and a temperature-sensitive cdc15-2 mutation (CY1447) was grown to log phase at 30° in YPD and either treated with 20 µg/ml nocodozole for 3 hr or shifted to 37° for 3 hr. Cells were fixed for chromosome spreads and Trf4-ZZ protein was detected with FITC-conjugated rabbit IgG. DNA was visualized with DAPI. (Bottom) Cells were harvested after 3 hr in nocodazole or at 37°. Protein extracts were prepared and separated by SDS-PAGE and Trf4-ZZ proteins were detected by Western blot using anti-IgG antibodies.

To further examine the apparent dissociation of Trf4p from chromosomes at the metaphase-to-anaphase transition we compared metaphase- and telophase- arrested cells for Trf4-ZZp association with chromosomes. A strain containing Trf4-ZZp and a temperature-sensitive cdc15-2 mutation (CY1447) was grown at permissive temperature and treated with nocodazole to synchronize cells in metaphase. To achieve telophase arrest, cells were shifted to the nonpermissive temperature (37°) for 3 hr (KABACK et al. 1984 ). As shown in Fig 6B, Trf4-ZZp is associated with metaphase chromosome preparations but is nearly absent from telophase chromosome preparations. Quantitation of 167 preparations from nocodazole-treated cells and 341 preparations from telophase-arrested cells shows that 95% of metaphase chromosome spreads stain intensely with FITC-conjugated IgG whereas only 10% of telophase spreads show any staining and the staining that is observed is generally weak. Wild-type CDC15 controls demonstrate that Trf4-ZZp association is not lost simply due to the temperature shift to 37° (data not shown).

Western blot analysis of metaphase and telophase cell extracts shows that the level of the Trf4-ZZp protein is the same at both points in mitosis (Fig 6B). Thus, Trf4p is being removed from chromosomes in some manner and not simply degraded. The slight reduction of Trf4p in G1 cells must occur after its removal from chromosomes. This behavior is similar to the cohesin Scc1p/Mcd1p, which dissociates from chromosomes at the metaphase-to-anaphase transition prior to being degraded rapidly (MICHAELIS et al. 1997 ).

Defective sister chromatid cohesion in TRF4 point mutants:
If the Trf4-182 and Trf4-194 proteins have completely lost their ability to bind to chromosomes, then we anticipate that they would behave as null mutants. To address this we examined sister chromatid cohesion in the trf4-182 and trf4-194 mutants using the "GFP chromosome tagging assay" developed by Murray and colleagues (STRAIGHT et al. 1996 ). In the GFP chromosome assay, 256 tandem repeats of the lac operator are integrated on one of the yeast chromosomes and chromosomes are visualized using immunofluorescence microscopy to detect the GFP fluorescence signal generated from binding of a GFP-lac repressor fusion. Using this assay the status of sister cohesion can be monitored by determining whether one or two fluorescence signals are observed within a single nucleus following DNA replication (STRAIGHT et al. 1996 ). We made the mutant strains with the lac operators integrated on chromosome III and GFP-LacI fusion integrated at the HIS3 locus by a cross with a wild-type strain carrying those genes (Table 2). The wild-type strain showed a single GFP signal in nearly all of the cells examined. Of 110 nuclei examined, two signals were present in <2.5% of wild-type nuclei. In contrast, the trf4 mutant cells showed a single signal in 82% of nuclei and two signals in 18% of nuclei (Fig 7A and Fig B). G1 controls were performed in all trf4 mutants and confirm that the presence of two signals is not due to aneuploidy (data not shown). trf4-182 and trf4-194 mutants show two signals in 18 and 20% of nuclei, respectively (Fig 7B). The cohesion defect in trf4-182 and trf4-194 is nearly identical to that of a trf4 deletion, consistent with a failure of these alleles to bind chromosomes. Both alleles are CPT sensitive and cause inviability in a trf5 background. Thus, mutation near the cyclin-dependent kinase (CDK) consensus site results in null phenotypes, indicating that chromosome association is likely to be completely abolished.

View larger version (28K): In this window In a new window Download PPT slide   Figure 7. Cohesion-defective alleles of Trf4/Pol reside in the polymerase and C-terminal domains. (A) Wild-type (CY1244) and trf4::TRP1 (CY1247) strains expressing a GFP-lacI fusion and containing 256 tandem repeats of lacO at the LEU2 locus on chromosome III were grown in SC complete to early log phase. After induction of the GFP-lacI fusion, cells were pelleted, washed, and resuspended in YPD containing 20 µg/ml nocodazole for an additional 4 hr. After the arrest, cells were fixed and adhered to glass slides. Cells were analyzed for the presence of one or two fluorescence signals. (B) Wild type (CY1244); trf4::TRP1 (CY1247); and strains containing alleles trf4-182 (CY1351), trf4-194 (CY1451), and trf4-491 (CY1355) were analyzed as in A. Quantitation of the cohesion defects in each strain is shown.

While association of Trf4p with chromosomes is necessary for cohesion, it is clearly not sufficient. This conclusion stems from the observation that the trf4-491-ZZ mutant product associates normally with chromosome spreads (Fig 5B), but is nonetheless defective in cohesion (24% of cells at the nocodazole block, Fig 7B). Thus, Trf4p must associate with chromosomes but association alone is not sufficient for proper cohesion.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

While Pol and Pol ß are likely to possess some level of structural similarity, the Pol family is clearly a distinct family of enzymes. TRF4/Pol genes are highly conserved in all eukaryotes and show much greater similarity to each other than any of the genes do to Pol ß. For example, while the yeast and human Trf4p/Pol proteins are 39% identical and 51% similar in primary amino acid sequence (WALOWSKY et al. 1999 ), neither protein has significant primary amino acid similarity to its species' Pol ß.

Genetic analysis of 34 surface-targeted mutations in TRF4/Pol has led to the following conclusions:

1. The regions of predicted structural similarity between Pol and Pol ß are indeed critical to the function of TRF4/Pol . This finding validates the predicted structural similarity (ARAVIND and KOONIN 1999 ) in spite of the paucity of primary sequence homology between these two enzymes.

2. The analysis has identified a novel C-terminal domain in TRF4/Pol that lies primarily outside the polymerase homology. Five mutant alleles in this region demonstrate that it, too, is essential to the function of TRF4/Pol .

3. Both the polymerase and C-terminal domains are necessary for DSB repair.

4. A discrete area in the N-terminal region of Trf4p is critical for chromosome association.

Functional significance of Trf4/Pol exonuclease activity:
Our genetic analysis of Trf4/Pol has uncovered a cluster of mutations outside of the polymerase domain that cause inviability in the absence of TRF5 and additional defects in DNA repair and sister chromatid cohesion. Recent analysis of the recBCD enzyme has identified residues in the primary sequence that are crucial for exonuclease activity (J. WANG et al. 2000 ). A closer examination of the C-terminal region of Trf4/Pol revealed limited similarity with the recBCD exonuclease domain (not shown), suggesting that the exonuclease activity we observed with Trf4/Pol in vitro has functional significance.

Trf4/Pol activities required for DNA damage repair:
DSB repair is known to require the activity of both leading- and lagging-strand DNA polymerases (HOLMES and HABER 1999 ). We have previously shown that TRF4 is also required for the repair of CPT- and MMS-induced lesions (WALOWSKY et al. 1999 ). The genetic data presented here demonstrate that the polymerase active site motifs, trf4-236 (DXD) and trf4-224 (GS), and several residues in the C terminus that may affect exonuclease activity, trf4-444, trf4-491, and trf4-508, are required for this repair.

It is possible that the repair defect is an indirect result of defective sister chromatid cohesion. However, the two mutant alleles located in the polymerase active site motif both cause hypersensitivity to CPT, consistent with a role for nucleotide polymerization per se in the repair process. All of the CPT-sensitive alleles are found within either a region of similarity with Pol ß, including the polymerase active site motif, or the putative C-terminal exonuclease domain, suggesting that polymerase and exonuclease activities are both necessary for Trf4/Pol 's role in DNA repair. None of the mutant alleles located outside of these regions are CPT sensitive. Analysis of sister chromatid cohesion in the remaining CPT-sensitive mutants will resolve whether or not the repair defect is secondary to the cohesion defect.

Cell-cycle-regulated association of Trf4/Pol with chromosomes:
Trf4/Pol is found associated with chromosomes from late G1 until late G2. The disappearance of Trf4/Pol from chromosomes coincides with spindle elongation at the metaphase-to-anaphase transition of mitosis. This is identical to what is observed with Scc1p/Mcd1p (MICHAELIS et al. 1997 ). The dissolution of cohesion allows sister chromatids to be separated by the pulling forces of the mitotic spindle. If the activities that maintain cohesion are not regulated, breakage of the chromosomes would result from pulling apart sister chromatids that are still "glued" together. The timed dissociation of Trf4/Pol from chromosomes suggests that its function also must be regulated for proper execution of these events in mitosis.

Only 2 of the 13 mutant alleles in TRF4 that are inviable in a trf5 background abolish its association with chromosomes. These two alleles, trf4-182 and trf4-194, flank a consensus CDK phosphorylation site in Trf4p at serine 191. This also defines "region 1" of the putative structural homology with Pol (Fig 2). The coincident positions of the CDK consensus site with the residues important for chromosome association indicate that Trf4p/Pol ß function may be regulated in some way by CDK phosphorylation.

Establishment of cohesion between sister chromatids is coupled with replication fork passage. Emerging evidence suggests that this coupling represents more than a coincident timing of independent events, but rather that the establishment of cohesion involves the active participation of replication-related activities (reviewed in CARSON and CHRISTMAN 2001 ). These include the Trf4p/Pol DNA polymerase; PCNA, a processivity clamp for some DNA polymerases; and a modified RFC clamp-loader complex. The genetic analysis of TRF4/Pol presented here provides us with tools for experiments aimed at determining how this is accomplished.

	FOOTNOTES

¹ Present address: The Johns Hopkins Oncology Center, Molecular Genetics Laboratory, Baltimore, MD 21231.
² Present address: Department of Molecular Biology and Genetics, Johns Hopkins University, Baltimore, MD 21205-2185.

	ACKNOWLEDGMENTS

The authors thank Nikki Levin, David Pellman, Mitch Smith, and members of the Christman lab for helpful comments. We also thank Erin O'Shea, Bruce Futcher, and Gavin Sherlock for strains and plasmids. This work was supported by grants from the National Institutes of Health and the Human Frontiers Science Program to M.F.C.

Manuscript received April 20, 2001; Accepted for publication October 9, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ARAVIND, L. and E. V. KOONIN, 1999  DNA polymerase beta-like nucleotidyltransferase superfamily: identification of three new families, classification and evolutionary history. Nucleic Acids Res. 27:1609-1618[Abstract/Free Full Text].

ARNAUDEAU, C., C. LUNDIN, and T. HELLEDAY, 2001  DNA double-strand breaks associated with replication forks are predominantly repaired by homologous recombination involving an exchange mechanism in mammalian cells. J. Mol. Biol. 307:1235-1245[Medline].

BOEKE, J. D., F. LACROUTE, and G. R. FINK, 1984  A positive selection for mutants lacking orotidine-5'-phosphate decarboxylase activity in yeast: 5-fluoro-orotic acid resistance. Mol. Gen. Genet. 197:345-346[Medline].

CARSON, D. R. and M. F. CHRISTMAN, 2001  Evidence that replication fork components catalyze establishment of cohesion between sister chromatids. Proc. Natl. Acad. Sci. USA 98:8270-8275[Abstract/Free Full Text].

CASTAÑO, I. B., S. HEATH-PAGLIUSO, B. U. SADOFF, D. J. FITZHUGH, and M. F. CHRISTMAN, 1996  A novel family of TRF (DNA topoisomerase I-related function) genes required for proper nuclear segregation. Nucleic Acids Res. 24:2404-2410[Abstract/Free Full Text].

CIOSK, R., M. SHIRAYAMA, A. SHEVENKO, T. TANAKA, and A. TOTH et al., 2000  Cohesin's binding to chromosomes depends on a separate complex consisting of Scc2 and Scc4 proteins. Mol. Cell 5:243-254[Medline].

DENG, J. Z., S. R. STARCK, and S. M. HECHT, 1999  DNA polymerase beta inhibitors from Baeckea gunniana. J. Nat. Prod. 62:1624-1626[Medline].

DUNN, B., and C. R. WOBBE, 1997 Preparation of protein extracts from yeast, pp. 13.13.1–13.13.9 in Current Protocols in Molecular Biology, edited by F. M. AUSUBEL, R. BRENT, R. E. KINGSTON, D. D. MOORE, J. G. SEIDMAN et al. John Wiley & Sons, New York.

HANNA, J. S., E. S. KROLL, V. LUNDBLAD, and F. A. SPENCER, 2001  Saccharomyces cerevisiae CTF18 and CTF4 are required for sister chromatid cohesion. Mol. Cell. Biol. 21:3144-3158[Abstract/Free Full Text].

HOLMES, A. M. and J. E. HABER, 1999  Double-strand break repair in yeast requires both leading and lagging strand DNA polymerases. Cell 96:415-424[Medline].

HSAING, Y. H., M. G. LIHOU, and L. F. LIU, 1989  Arrest of replication forks by drug-stabalized topoisomerase I-DNA cleavable complexes as a mechanism of cell killing by camptothecin. Cancer Res. 49:5077-5082[Abstract/Free Full Text].

HSU, H. Y., J. J. YANG, and C. C. LIN, 1997  Effects of oleanolic acid and ursolic acid on inhibiting tumor growth and enhancing the recovery of hematopoietic system postirradiation in mice. Cancer Lett. 111:7-13[Medline].

KABACK, D. B., P. W. OELLER, H. Y. DE STEENSMA, J. HIRSCHMAN, and D. RUEZINSKY et al., 1984  Temperature-sensitive lethal mutations on yeast chromosome I appear to define only a small number of genes. Genetics 108:67-90[Abstract/Free Full Text].

LIU, L. F., 1989  DNA topoisomerase poisons as antitumor drugs. Annu. Rev. Biochem. 58:351-375[Medline].

MICHAELIS, C., R. CIOSK, and K. NASMYTH, 1997  Cohesins: chromosomal proteins that prevent premature separation of sister chromatids. Cell 91:35-45[Medline].

NITISS, J. L. and J. C. WANG, 1996  Mechanisms of cell killing by drugs that trap covalent complexes between DNA topoisomerases and DNA. Mol. Pharmacol. 50:1095-1102[Abstract].

PELLETIER, H., M. R. SAWAYS, A. KUMAR, S. H. WILSON, and J. KRAUT, 1994  Structures of ternary complexes of rat DNA polymerase beta, a DNA template-primer, and ddCTP. Science 264:1891-1903[Abstract/Free Full Text].

SAWAYA, M. R., H. PELLETIER, A. KUMAR, S. H. WILSON, and J. KRAUT, 1994  Crystal structure of rat DNA polymerase beta: evidence for a common polymerase mechanism. Science 264:1930-1935[Abstract/Free Full Text].

SCHWARTZ, J. L., 1989  Monofunctional alkylating agent-induced S-phase dependent DNA damage. Mutat. Res. 16:111-118.

SKIBBENS, R. V., L. B. CORSON, D. KOSHLAND, and P. HIETER, 1999  Ctf7p is essential for sister chromatid cohesion and links mitotic chromosome structure to the DNA replication machinery. Genes Dev. 13:307-319[Abstract/Free Full Text].

STRAIGHT, A. F., A. S. BELMONT, C. C. ROBINETT, and A. W. MURRAY, 1996  GFP tagging of budding yeast chromosomes reveals that protein-protein interactions can mediate sister chromatid cohesion. Curr. Biol. 6:1599-1608[Medline].

TAKAHASHI, K. and M. YANAGIDA, 2000  Cell cycle: replication meets cohesion (comment). Science 289:735-736[Free Full Text].

TERCERO, J. A. and J. F. X. DIFFLEY, 2001  Regulation of DNA replication fork progression through damaged DNA by the Mec1/Rad53 checkpoint. Nature 412:553-557[Medline].

TOTH, A., R. CIOSK, F. UHLMANN, M. GALOVA, and A. SCHLEIFFER et al., 1999  Yeast cohesin complex requires a conserved protein, Eco1p(Ctf7), to establish cohesion between sister chromatids during DNA replication. Genes Dev. 13:320-333[Abstract/Free Full Text].

UHLMANN, F. and K. NASMYTH, 1998  Cohesion between sister chromatids must be established during DNA replication. Curr. Biol. 8:1095-1101[Medline].

WAGA, S. and B. STILLMAN, 1994  Anatomy of a DNA replication fork revealed by reconstitution of SV40 DNA replication in vitro. Nature 369:207-212[Medline].

WALOWSKY, C., D. J. FITZHUGH, I. B. CASTAÑO, J. Y. JU, and N. A. LEVIN et al., 1999  The topoisomerase-related function gene TRF4 affects cellular sensitivity to the antitumor agent camptothecin. J. Biol. Chem. 274:7302-7308[Abstract/Free Full Text].

WANG, J., R. CHEN, and D. A. JULIN, 2000  A single nuclease active site of the Escherichia coli RecBCD enzyme catalyzes single-stranded DNA degradation in both directions. J. Biol. Chem. 275:507-513[Abstract/Free Full Text].

WANG, S. W., T. TODA, R. MACCALLUM, A. L. HARRIS, and C. NORBURY, 2000  Cid1, a fission yeast protein required for S-M checkpoint control when DNA polymerase delta or epsilon is inactivated. Mol. Cell. Biol. 20:3234-3244[Abstract/Free Full Text].

WANG, Z., I. B. CASTAÑO, A. DE LAS PEÑAS, C. ADAMS, and M. F. CHRISTMAN, 2000  Pol kappa: a DNA polymerase required for sister chromatid cohesion. Science 289:774-779[Abstract/Free Full Text].

WERTMAN, K. R., D. G. DRUBIN, and D. BOTSTEIN, 1992  Systematic mutational analysis of the yeast ACT1 gene. Genetics 132:33.

XIAO, W., B. L. CHOW, and L. RATHGEBER, 1996  The repair of DNA methylation damage in Saccharomyces cerevisiae. Curr. Genet. 30:461-468[Medline].

This article has been cited by other articles:

				C. C. Reis and J. L. Campbell Contribution of Trf4/5 and the Nuclear Exosome to Genome Stability Through Regulation of Histone mRNA Levels in Saccharomyces cerevisiae Genetics, March 1, 2007; 175(3): 993 - 1010. [Abstract] [Full Text] [PDF]

				S. KADABA, X. WANG, and J. T. ANDERSON Nuclear RNA surveillance in Saccharomyces cerevisiae: Trf4p-dependent polyadenylation of nascent hypomethylated tRNA and an aberrant form of 5S rRNA RNA, March 1, 2006; 12(3): 508 - 521. [Abstract] [Full Text] [PDF]

				S. Kadaba, A. Krueger, T. Trice, A. M. Krecic, A. G. Hinnebusch, and J. Anderson Nuclear surveillance and degradation of hypomodified initiator tRNAMet in S. cerevisiae Genes & Dev., June 1, 2004; 18(11): 1227 - 1240. [Abstract] [Full Text] [PDF]

				S. Edwards, C. M. Li, D. L. Levy, J. Brown, P. M. Snow, and J. L. Campbell Saccharomyces cerevisiae DNA Polymerase {varepsilon} and Polymerase {sigma} Interact Physically and Functionally, Suggesting a Role for Polymerase {varepsilon} in Sister Chromatid Cohesion Mol. Cell. Biol., April 15, 2003; 23(8): 2733 - 2748. [Abstract] [Full Text] [PDF]

				M. A. Kenna and R. V. Skibbens Mechanical Link between Cohesion Establishment and DNA Replication: Ctf7p/Eco1p, a Cohesion Establishment Factor, Associates with Three Different Replication Factor C Complexes Mol. Cell. Biol., April 15, 2003; 23(8): 2999 - 3007. [Abstract] [Full Text] [PDF]