Two distinct DNA ligase activities in mitotic extracts of the yeast Saccharomyces cerevisiae
Two distinct DNA ligase activities in mitotic extracts of the yeast Saccharomyces cerevisiae William Ramos, Nancy Tappe1, Jose Talamantez, Errol C. Friedberg1and Alan E. Tomkinson*
Department of Molecular Medicine, Institute of Biotechnology, The University of Texas Health Science Center at San Antonio, 15355 Lambda Drive, San Antonio, TX 78245, USA and 1Laboratory of Molecular Pathology, The University of Texas Southwestern Medical Center, Dallas, TX 75235, USA
Received January 13, 1997;Revised and Accepted March 5, 1997
ABSTRACT
Four biochemically distinct DNA ligases have been identified in mammalian cells. One of these enzymes, DNA ligase I, is functionally homologous to the DNA ligase encoded by the Saccharomyces cerevisiaeCDC9 gene. Cdc9 DNA ligase has been assumed to be the only species of DNA ligase in this organism. In the present study we have identified a second DNA ligase activity in mitotic extracts of S.cerevisiae with chromatographic properties different from Cdc9 DNA ligase, which is the major DNA joining activity. This minor DNA joining activity, which contributes 5-10% of the total cellular DNA joining activity, forms a 90 kDa enzyme-adenylate intermediate which, unlike the Cdc9 enzyme-adenylate intermediate, reacts with an oligo (pdT)/poly (rA) substrate. The levels of the minor DNA joining activity are not altered by mutation or by overexpression of the CDC9 gene. Furthermore, the 90 kDa polypeptide is not recognized by a Cdc9 antiserum. Since this minor species does not appear to be a modified form of Cdc9 DNA ligase, it has been designated as S.cerevisiae DNA ligase II. Based on the similarities in polynucleotide substrate specificity, this enzyme may be the functional homolog of mammalian DNA ligase III or IV.
INTRODUCTION
DNA ligation is an essential step in DNA replication, DNA excision repair and genetic recombination. Studies in prokaryotes such as Escherichia coli suggest that the cellular requirement for DNA ligation is fulfilled by a single DNA ligase (1 ). In contrast, two and four distinct species of DNA ligase have been characterized in extracts from Drosophila (2 ) and mammalian cells (3 ,4 ), respectively. DNA ligase II appears to be derived from DNA ligase III by a proteolytic mechanism (5 -7 ). Thus, the three human genes encoding DNA ligases identified so far can account for the four characterized enzyme activities (3 -5 ,8 ).
Mammalian DNA ligase I is the most extensively studied eukaryotic DNA ligase. This enzyme appears to be required for the joining of Okazaki fragments at the replication fork (9 -12 ) and is also involved in some pathways of DNA repair (12 ,13 ). Human DNA ligase I cDNA complements the conditional lethal phenotype of a S.cerevisiae cdc9 mutant defective in DNA ligase (14 ). Not surprisingly, mammalian DNA ligase I and yeast Cdc9 DNA ligase possess identical polynucleotide substrate specificities and so Cdc9 can be considered to be DNA ligase I of S.cerevisiae (4 ,15 ,16 ). Enzymes with biochemical properties similar to mammalian and yeast DNA ligase I have also been purified from extracts of Drosophila (2 ,17 ,18 ), Xenopus laevis (19 ) and trypanosomes (20 ).
Mammalian DNA ligase II was initially described as a minor activity in extracts of calf thymus that bound more firmly to hydroxylapatite than the major activity, DNA ligase I (21 ). The 70 kDa DNA ligase II has been purified to homogeneity from both bovine thymus and liver (7 ,22 ). It is immunologically distinct from DNA ligase I (22 ,23 ) and also possesses different catalytic properties, including the ability to join oligo (pdT) molecules hybridized to a poly (rA) template (4 ,24 ). The human LIG3 and LIG4 genes encode DNA ligases with calculated molecular weights of ~100 kDa. These enzymes are also active with the oligo (pdT)/poly (rA) substrate (3 -5 ,8 ). Amino acid sequencing of peptides from bovine DNA ligases II and III revealed the presence of identical amino acid sequences within these polypeptides (6 ,7 ). Subsequent cloning experiments resulted in the identification of a cDNA whose translated open reading frame contains regions homologous to all the DNA ligase II and III peptides (5 -8 ). The absence of different sized transcripts in tissues such as the liver, in which DNA ligase II is the predominant activity, suggests that DNA ligases II and III are not generated by an alternative splicing mechanism (5 ,7 ). While the formal possibility that DNA ligase II is encoded by a distinct gene cannot be ruled out, it appears more likely that DNA ligase II is derived from DNA ligase III by proteolytic processing.
The interaction between 103 kDa DNA ligase III and the human DNA strand-break repair protein XRCC1 implicates DNA ligase III in the repair of DNA single-strand breaks which arise as a consequence of DNA damage by alkylating agents and ionizing radiation (25 -27 ). A 96 kDa form of DNA ligase III with a C-terminus distinct from that in the 103 kDa form of the enzyme is produced by an alternative splicing event which is restricted to male germ cells undergoing meiotic recombination (28 ). The 96 kDa protein does not interact with XRCC1 and thus may play a specific role in the completion of meiotic recombination with protein partners other than XRCC1 (28 ). The open reading frame encoding DNA ligase IV was identified by screening a human EST database with a peptide sequence that is conserved in eukaryotic DNA ligases (8 ). At the present time, the biological function(s) of DNA ligase IV is unknown. In summary, three mammalian LIG genes encode distinct polypeptides designated DNA ligases I, III and IV. Two species of DNA ligase III exist as a consequence of alternative processing of the LIG3 transcript whilst DNA ligase II appears to be derived from DNA ligase III by proteolysis.
A second DNA ligase activity that is clearly distinct from the replicative DNA ligase has been partially purified from Drosophila pupae and embryos (2 ,29 ), indicating that multiple species of DNA ligase occur in other multicellular eukaryotes. At present there is no compelling biochemical or genetic evidence for DNA ligases in other lower eukaryotes besides the previously described Cdc9 DNA ligase of S.cerevisiae and the Cdc17 DNA ligase of S.pombe (30 ,31 ). To address this question we have begun to investigate the repertoire of DNA ligases in S.cerevisiae. In this report we describe the partial purification of a DNA ligase activity from extracts of haploid S.cerevisiae cells. Since this activity is not recognized by a Cdc9 antibody and is not affected by a cdc9 mutation, we conclude that it is distinct from Cdc9 DNA ligase (DNA ligase I). Consequently, we have designated this activity, which may be functionally homologous to mammalian DNA ligase III or IV, as S.cerevisiae DNA ligase II.
MATERIALS AND METHODS
Plasmids and strains
The S.cerevisiae strain, BJ2168 (MATa prc1-407 prb1-1122 pep4-3 leu2 trp1 ura3-52) was from the Yeast Genetic Stock Center, Berkeley, CA. The S.cerevisiae strains, M1-2B (CDC9 ura3 trp1) and L94-4D (cdc9-7 ura3 trp1) and the plasmid p207Sclig1BH (32 ), a derivative of pJDB207, were supplied by Dr L. H. Johnston, Laboratory of Cell Propagation, National Institute of Medical Research, UK. Strain BJ2168 and BJ2168 harboring the plasmid p207Sclig1BH were grown in a 60 l New Brunswick Fermentor in the appropriate yeast minimal media as described previously (16 ). Approximately 350 g cells (wet weight) were obtained from 59 l of culture. Strains M1-2B and L94-4D were grown in YPD at 25oC.
The E.coli strain SW1064 harboring the plasmid pKK223-3-RLGX was supplied by Dr E. M. Phizicky, Department of Biochemistry, University of Rochester. Expression of the plasmid-encoded yeast tRNA ligase gene was induced with isopropyl-[beta]-D-thiogalactopyransoside (IPTG) as described (33 ).
Preparation of polynucleotide substrates
Oligonucleotides (pdT)16, (dT)16 and (rA)12-18 and polynucleotides (dA), (rA) and (dT) were purchased from Pharmacia. Dephosphorylated oligonucleotides were labeled and annealed with the appropriate polynucleotide (16 ). A DNA substrate that contains a single defined nick was prepared by annealing two oligonucleotides, a labeled 16mer and an 18mer with a complementary 34mer as described previously (16 ).
DNA ligase assays
Reaction mixtures (60 [mu]l) contained 50 mM Tris-HCl (pH 8.0), 10 mM MgCl2, 5 mM dithiothreitol (DTT), 1 mM ATP, 50 [mu]g/ml bovine serum albumin, polynucleotide substrate (20 000 c.p.m.) and a limiting amount of DNA ligase (4 ,16 ). Incubations were at 20oC for 15 min. The conversion of 32P-labeled phosphomonoesters to alkaline phosphatase-resistant diesters was measured using bacterial alkaline phosphatase (Sigma). One unit of DNA ligase activity catalyses the conversion of 1 pmol terminal phosphate residues to a phosphatase-resistant form in 15 min at 20oC.
Formation of DNA ligase-adenylate complex
Reaction mixtures (20 [mu]l) contained 60 mM Tris-HCl (pH 8.0), 10 mM MgCl2, 5 mM DTT, 50 [mu]g/ml bovine serum albumin, 0.5 [mu]Ci [[alpha]-32P]ATP (3000 Ci/mmol, Amersham) and DNA ligase. Incubations were at room temperature for 15 min. After the addition of 5 [mu]l of sodium dodecyl sulfate (SDS) sample buffer, reaction mixtures were heated at 90oC for 10 min. Proteins were separated by SDS-polyacrylamide gel electrophoresis (PAGE)(34 ). Gels were fixed for 10 min in 10% acetic acid and dried. Adenylylated proteins were detected by autoradiography.
To examine the reactivity of the enzyme-adenylate intermediate, the adenylylation reaction was performed as described above. Aliquots (10 [mu]l) were incubated with 0.8 [mu]g unlabelled polynucleotide or with 20 nmol sodium pyrophosphate for 2 h at 20oC. Reactions were stopped by the addition of SDS sample buffer and adenylylated polypeptides were detected as above.
Purification of yeast DNA ligase II
Protein purification was carried out on ice or at 4oC. All buffers contained a cocktail of protease inhibitors (35 ). Protein concentrations were measured by the method of Bradford using bovine serum albumin as the standard (36 ). A cleared lysate (510 ml, 6.12 g) was prepared from 200 g wet weight of BJ2168 cells and fractionated by phosphocellulose chromatography as described previously (16 ). After the addition of 0.5 M KPO4 to a final concentration of 1 mM, the high salt eluate from the phosphocellulose resin (550 ml, 1.1 g) was applied to a 2.6 * 40 cm hydroxylapatite column that had been pre-equilibrated with 50 mM Tris-HCl (pH 7.5), 0.75 M NaCl, 10 mM 2-mercaptoethanol, 1 mM KPO4. After washing with this buffer, proteins were eluted stepwise with 50 mM KPO4 (pH 7.5), 200 mM KPO4 (pH 7.5) and 400 mM KPO4 (pH 7.5) buffers, each containing 0.5 mM DTT. Fractions from each of the eluates that contained significant amounts of protein were pooled. Aliquots from these pools were dialyzed against 50 mM Tris-HCl (pH 7.5), 30 mM NaCl, 1 mM EDTA, 0.5 mM DTT, 50% glycerol and assayed for DNA ligase activity.
The 400 mM KPO4 eluate (300 ml, 150 mg) was dialyzed against 50 mM Tris-HCl (pH 7.5), 30 mM NaCl, 1 mM EDTA, 0.5 mM DTT, 10% glycerol (buffer A) and then loaded onto a 1.6 * 25 cm phosphocellulose that had been pre-equlibrated with buffer A. Bound proteins were eluted with a 500 ml linear gradient from buffer A to buffer A containing 1 M NaCl and fractions were assayed for DNA ligase activity. The DNA joining activity (48 ml, 15 mg, 11 U) was dialyzed against 50 mM Tris-HCl (pH 7.5), 1 mM EDTA, 0.5 mM DTT (buffer B) containing 0.9 M ammonium sulfate prior to loading onto a 15 ml phenyl-Sepharose column. Bound proteins were eluted with a 150 ml linear gradient from 0.9 to 0 M ammonium sulfate in buffer B and then with a 150 ml linear gradient from buffer B to buffer containing 70% ethylene glycol. No DNA ligase activity was detected in the flow through, wash and gradient eluted fractions. DNA ligase activity (7 ml, 560 [mu]g, 1.5 U) was eluted by buffer B containing 80% ethylene glycol.
Partial purification of yeast tRNA ligase
Yeast tRNA ligase was partially purified from 15 g (wet weight) E.coli SW1064 pKK223-3-RLGX cells by polymin P precipitation and heparin sepharose chromatography as described by Xu et al. (33 ). tRNA ligase activity was monitored by formation of the enzyme-AMP complex. The reactivity of the enzyme-adenylate intermediate formed by the peak fractions from the heparin-Sepharose column with different polynucleotide substrates and pyrophosphate was examined as described above.
Immunization procedures
Cdc9 protein, which was purified as described previously (16 ), was used as the antigen to produce a rabbit polyclonal antiserum by a standard immunization protocol (15 ).
Immunoprecipitation
Proteins were incubated with (5 [mu]l) pre-immune serum or (5 [mu]l) immune serum raised against Cdc9 protein for 90 min at 4oC in a final volume of 500 [mu]l. Protein A-Sepharose beads, mixed with Sepharose CL-4B beads in a ratio of 1:3 (100 [mu]l of a 50% suspension) were added and the incubation was continued for 60 min. The beads, which were collected by centrifugation and washed three times with lysis buffer, were resuspended in an equal volume of 60 mM Tris-HCl (pH 8.0), 10 mM MgCl2, 5 mM DTT, 50 [mu]g/ml bovine serum albumin and 2 mM sodium pyrophosphate. After incubation for 15 min at room temperature to remove AMP residues that had bound in vivo or during purification, the beads were washed with lysis buffer (16 ) and resuspended in an equal volume of 60 mM Tris-HCl (pH 8.0), 10 mM MgCl2, 5 mM DTT, 50 [mu]g/ml bovine serum albumin containing 5 [mu]Ci [[alpha]-32P]ATP (3000 Ci/mmol, Amersham). After further incubation for 30 min, the beads were washed and resuspended in an equal volume of SDS sample buffer. Labeled polypeptides were detected as described above.
Immunoblotting
Polypeptides were separated by denaturing gel electrophoresis and transferred to nitrocellulose membranes. The membrane was incubated with the primary antibody and the resultant antigen-antibody complexes were detected by enhanced chemiluminescence (35 ).
RESULTS
Saccharomyces cerevisiae extracts contain a DNA ligase that is active with the oligo (pdT)/poly (rA) substrate
Four biochemically distinct DNA ligase activities have been identified in mammalian cell extracts (3 ,4 ). DNA ligases II, III and IV can be distinguished from the enzyme that functions in DNA replication (DNA ligase I in mammals and Cdc9 in S.cerevisiae) by their ability to catalyze the joining of oligo (pdT) molecules hybridized to a poly (rA) template (3 ,4 ,16 ). Therefore, we examined extracts of S.cerevisiae for a DNA joining activity that was active with the oligo (pdT)/poly (rA) substrate.
Partial purification and characterization of S.cerevisiae DNA ligase II
Because of the extreme lability and relatively low levels of the oligo (pdT)/poly (rA) joining activity a purification scheme was developed that was more suited to the rapid processing of large quantities of yeast whole cell extracts. After lysis and batch adsorption to P11 phosphocellulose (16 ), proteins in the high salt eluate from the phosphocellulose resin were loaded directly onto a hydroxylapatite column. Consistent with previous observations, Cdc9 DNA ligase (yeast DNA ligase I) was eluted by 200 mM KPO4 and contributed the majority of the total cellular DNA joining activity (16 ). We consistently observed an activity that could catalyze joining of the oligo (pdT)/poly (rA) substrate in the 400 mM KPO4 eluate. However, the low level of this activity and the presence of contaminating nucleases, in particular RNases, precluded accurate quantitative measurements of the DNA joining activity in this fraction (data not shown). No joining of an oligo (prA)/poly (dT) substrate was detected in assays with the eluates from the hydroxylapatite column. This also appears to reflect nuclease contamination since purified Cdc9 DNA ligase is active with this substrate (16 ).
Saccharomyces cerevisiae DNA ligase II activity is not affected by alterations in the level of Cdc9 DNA ligase
DNA ligase II is not recognized by antisera against Cdc9 DNA ligase
To provide further evidence that Cdc9 DNA ligase (DNA ligase I) and DNA ligase II are the products of distinct genes we examined the immunoreactivity of the polypeptides in the 200 and 400 mM KPO4 hydroxylapatite fractions that formed enzyme-AMP complexes (Fig. 2 ) with a polyclonal Cdc9 antiserum. This antiserum immunoprecipitated a single polypeptide that formed an 87 kDa enzyme-AMP complex from yeast whole cell extracts (Fig. 5 A). In similar experiments the Cdc9 polypeptide in the 200 mM KPO4 eluate was also specifically immunoprecipitated and then adenylylated (Fig. 5 B, lane 3). However, no polypeptides were adenylylated in the immunoprecipitate from the 400 mM KPO4 eluate (Fig. 5 B, lane 5). To exclude the possibility that the 400 mM KPO4 eluate contains an activity which inhibits specific immunoprecipitation and/or adenylylation, we preincubated mixtures of the 200 and 400 mM KPO4 eluates which by themselves yielded similar amounts of labeled polypeptide prior to immunoprecipitation. A single labeled 87 kDa polypeptide corresponding in size and quantity to that expected from the 200 mM KPO4 eluate was detected (data not shown).
DISCUSSION
We have presented evidence for the presence of two distinct species of DNA ligase in mitotic extracts of S.cerevisiae. One of these enzymes, encoded by the CDC9 gene, has previously been shown to be functionally homologous to mammalian DNA ligase I and therefore can be considered to be the DNA ligase I of S.cerevisiae (14 ,16 ). This type of DNA ligase is required for DNA replication (9 ,11 ,12 ,14 ,37 ,38 ) and also functions in DNA repair (12 ,13 ,37 ). A second minor DNA joining activity was resolved from Cdc9 DNA ligase by fractionation on several different chromatography resins. We have designated this enzyme S.cerevisiae DNA ligase II, since it appears to be distinct from Cdc9 DNA ligase based on the following observations: (i) unlike Cdc9 DNA ligase, this enzyme was active on an oligo (pdT)/oligo (rA) substrate; (ii) the level of S.cerevisiae DNA ligase II activity was unaffected by overexpression or mutational inactivation of the CDC9 gene; (iii) S.cerevisiae DNA ligase II was not recognized by Cdc9 antiserum.
The ability to join oligo (dT) molecules hybridized to a poly (rA) template is a characteristic feature of S.cerevisiae DNA ligase II and mammalian DNA ligases II, III and IV (3 ,4 ) and specifically distinguishes these enzyme activities from the enzymes that function in DNA replication such as mammalian DNA ligase I and Cdc9 DNA ligase (4 ,16 ). Although mammalian DNA ligases III and IV can be distinguished by the ability of DNA ligase III to join oligo (rA) molecules hybridized to a poly (dT) template (3 ), activity with this substrate cannot be used to define a DNA ligase as a type III or type IV enzyme because mammalian DNA ligase II, which appears to be derived from DNA ligase III by a proteolytic mechanism (5 -7 ), is also unable to join oligo (rA) molecules hybridized to a poly (dT) template (4 ). Hence, we do not interpret the absence of joining activity with the oligo (rA)/poly (dT) substrate as evidence that yeast DNA ligase II is more similar to mammalian DNA ligase IV than to mammalian DNA ligase III.
The very low levels and lability of yeast DNA ligase II have hindered our attempts to purify and characterize this enzyme more thoroughly. We have not been able to isolate sufficient quantities of this protein from mitotic extracts for peptide sequencing, which in turn would facilitate gene cloning. However, the recent availability of the complete sequence of the S.cerevisiae genome offers an alternative approach to identifying genes that may encode DNA ligases. A six amino acid motif required for formation of the enzyme-adenylate intermediate (7 ,39 ) and a 16 amino acid sequence of unknown function (5 ,8 ,14 ,15 ) are highly conserved in all known eukaryotic DNA ligases. The latter sequence was used to identify human DNA ligase cDNAs encoding DNA ligases I, III and IV in a search of a human EST database (8 ), but does not identify open reading frames other than the one encoded by the CDC9 gene in the S.cerevisiae genome. In contrast, searches of the S.cerevisiae genome with the active site motif KYDGER identified the CDC9 gene as well as a second open reading frame designated YOR005c which encodes a 944 amino acid polypeptide. There are additional regions within this open reading frame that share less extensive homology with conserved regions (including the 16 amino acid sequence of unknown function) in the catalytic domains of other eukaryotic DNA ligases. The relationship between this putative DNA ligase, which most closely resembles human DNA ligase IV, and S.cerevisiae DNA ligase II is currently being investigated.
The relative ease of genetic manipulation in S.cerevisiae makes this organism an attractive model system for higher eukaryotes. For example, molecular and genetic analysis of the yeast genome has identified functional homologues for mammalian DNA polymerases [alpha], [beta], [delta] and [epsilon] (40 -44 ). The identification of a yeast open reading frame that may encode a DNA ligase coupled with our biochemical demonstration of a yeast DNA ligase activity with similar properties to mammalian DNA ligases III and IV, suggests that genetic analysis of the multiple species of eukaryotic DNA ligase should be tractable in S.cerevisiae, and that the results of such studies will be informative with respect to DNA metabolism in mammalian cells.
ACKNOWLEDGEMENTS
We thank Doug Johnston for assistance with yeast fermentation. These studies were supported by United States Public Health Service grants GM47251 (A.E.T.) and CA-12428 (E.C.F.) from the Department of Health and Human Services.
REFERENCES
1 Kornberg,A. and Baker,T. (1992) DNA Replication. W.H. Freeman and Co., New York, NY.
2 Takahashi,M. and Senshu,M. (1987) FEBS Lett.213, 345-352.MEDLINE Abstract
3 Robins,P. and Lindahl,T. (1996) J. Biol. Chem.271, 24257-24261.MEDLINE Abstract
4 Tomkinson,A.E., Roberts,E., Daly,G., Totty,N.F. and Lindahl,T. (1991) J. Biol. Chem.286, 21728-21735.
