S-Phase Checkpoint Pathways Stimulate the Mobility of the Retrovirus-Like Transposon Ty1

The mobility of the Ty1 retrotransposon in the yeast Saccharomycescerevisiae is restricted by a large collection of proteins thatpreserve the integrity of the genome during replication. Severalof these repressors of Ty1 transposition (Rtt)/genome caretakersare orthologs of mammalian retroviral restriction factors. Inrtt/genome caretaker mutants, levels of Ty1 cDNA and mobilityare increased; however, the mechanisms underlying Ty1 hypermobilityin most rtt mutants are poorly characterized. Here, we showthat either or both of two S-phase checkpoint pathways, thereplication stress pathway and the DNA damage pathway, partiallyor strongly stimulate Ty1 mobility in 19 rtt/genome caretakermutants. In contrast, neither checkpoint pathway is requiredfor Ty1 hypermobility in two rtt mutants that are competentfor genome maintenance. In rtt101 ${Delta}$ mutants, hypermobility isstimulated through the DNA damage pathway components Rad9, Rad24,Mec1, Rad53, and Dun1 but not Chk1. We provide evidence thatTy1 cDNA is not the direct target of the DNA damage pathwayin rtt101 ${Delta}$ mutants; instead, levels of Ty1 integrase and reversetranscriptase proteins, as well as reverse transcriptase activity,are significantly elevated. We propose that DNA lesions createdin the absence of Rtt/genome caretakers trigger S-phase checkpointpathways to stimulate Ty1 reverse transcriptase activity.

INTRODUCTION

Top

INTRODUCTION

Long terminal repeat (LTR) retrotransposons and retroviral provirusesare related families of mobile DNA elements found dispersedin eukaryotic genomes. Their structures consist of LTRs flankinga central coding domain, including gag and pol genes and, inthe case of infectious retroviruses, an env gene (33). The gaggene encodes structural proteins that form the retrotransposonvirus-like particle (VLP) or the retroviral core particle, whilethe pol gene encodes the enzymes necessary for mobility, includingprotease (PR), reverse transcriptase (RT), and integrase (IN).LTR retrotransposons and integrated retroviral proviruses aretranscribed by RNA polymerase II to form a terminally redundantRNA, which is translated and also packaged into the virion coreparticle or VLP, where it functions as a template for RT. INcatalyzes integration of the resulting cDNA into the host cellgenome by a homology-independent mechanism. The similaritiesbetween LTR retrotransposon transposition and retrovirus replicationsuggest that these elements are susceptible to common mechanismsof regulation by their host organisms (58).

The genome of the budding yeast Saccharomyces cerevisiae harborsfive families of LTR retrotransposons, named Ty1 to Ty5. Ty1elements are the most abundant, the most highly expressed, andthe most mutagenic (88). Over 30 genes that encode repressorsof Ty1 transposition (RTT genes) have been identified, despitethe fact that an exhaustive screen has yet to be performed (58).Many of the known RTT gene products function during S phaseto maintain the integrity of the yeast genome. For example,the Rtt factor Rad27 is the ortholog of human FEN-1, a 5'-to-3'exonuclease, 5' flap endonuclease required for Okazaki fragmentprocessing and maturation during replication (29, 39, 72). TheRrm3 helicase promotes replication through nonnucleosomal protein-DNAcomplexes, while the cullin Rtt101 is involved in replicationthrough natural and MMS-induced pause sites (38, 55). Mms1/Rtt108is a constituent of a functional module containing Rtt101 andis important for replication-dependent DNA repair (36, 70).Rtt109 is a histone acetyltransferase that modifies Lys56 onhistone H3 and is required for cells to survive DNA damage duringS phase (12, 20, 31, 76, 86). Rtt107/Esc4 is a BRCT domain proteinimplicated in replication restart after DNA damage (11, 73).Elg1/Rtt110 is a replication factor C homolog that is importantfor recovery from DNA damage during replication (4, 5, 42).Members of the Rad52 and Rad50 epistasis groups are requiredfor recombination-mediated restart of stalled or broken replicationforks (34, 69, 83). The roles of these conserved replication-and repair-related factors in the regulation of Ty1 retrotransposonspoints to a connection between preserving the integrity of thegenome during S phase and maintaining the dormancy of Ty1 elements.

The Rtt/genome integrity factors described to date inhibit transpositionat posttranscriptional steps, resulting in low levels of Ty1cDNA (10, 52, 53, 71, 77, 81). Two models, which are not mutuallyexclusive, have been proposed to explain how Rtt/genome caretakerslimit the levels of Ty1 cDNA and retrotransposition. The firstmodel is one in which Rtt/genome caretakers interact directlywith Ty1 cDNA in the nucleus to promote its degradation. Thismodel is exemplified by the TFIIH helicases Rad3 and Ssl2 andRad27, which block Ty1 cDNA accumulation at a step subsequentto reverse transcription (52, 81). However, an important rolefor cDNA degradation in restricting transposition is difficultto reconcile with the remarkably long half-life of Ty1 cDNA,determined to be between 93 and 252 min (52, 78, 81). A secondmodel posits that Rtt/genome caretakers prevent the formationof DNA lesions that activate a DNA integrity checkpoint pathwaythat stimulates Ty1 cDNA synthesis. This model was proposedto explain the DNA damage checkpoint pathway-dependent increasein Ty1 cDNA levels in telomerase-negative mutants (78). Becausethe half-life of Ty1 cDNA in a telomerase-negative strain wasequivalent to that in a wild-type strain, it was proposed thatthe DNA damage pathway stimulates the activity of a Ty1 or Ty1-associatedprotein, resulting in increased cDNA synthesis.

In this study, we ask whether the high levels of Ty1 mobilityin rtt mutants that have defects in maintaining genome integrityduring replication are triggered by either of two S-phase checkpointpathways (Fig. 1). The highly integrated S-phase checkpointpathways are activated by delays in DNA replication (replicationstress pathway) or by DNA lesions that obstruct replication(DNA damage pathway) (66). In both pathways, replication blocksare bound independently by a sensor complex containing the essentialkinase Mec1 and a second sensor complex, known as the slidingclamp (46, 59). The sliding clamp is loaded onto DNA by a five-subunitclamp loader (56). The clamp loader in the DNA damage pathwayconsists of RFC2-5 subunits and the replication factor C-likeprotein Rad24, while Ctf18 and Elg1 play partially redundantroles with Rad24 (4, 5, 42). In the replication stress pathway,Rad24 functions redundantly with Ctf18 (64, 70). Binding ofthese sensor complexes to DNA at sites of damage or replicationpausing promotes the hyperphosphorylation and activation ofthe Chk kinases Rad53 and Chk1 (74, 75, 79, 80). The activationof Chk kinases also requires the mediator protein Rad9 in theDNA damage pathway or Mrc1 in the replication stress pathway(1, 6, 28, 68, 82, 85). In the absence of Mrc1, Rad9 becomesessential for S-phase checkpoint activity, suggesting that theDNA damage pathway functions as a backup when the replicationstress pathway is compromised (1, 43, 66). Both checkpoint pathwaysblock cell cycle progression and initiate a transcriptionaland posttranslational response to the DNA lesion (66).

View larger version (15K):
[in this window]
[in a new window]

FIG. 1. The S-phase checkpoint pathways are activated when replication is impeded. Components of each pathway that are relevant to this study are shown. The replication stress and the DNA damage pathways are activated by replication pausing and replication blocks, respectively. The DNA is bound independently by two sensor complexes, represented by the individual components Mec1 and Rad24. Binding of sensor complexes activates the Chk kinases Rad53 and Chk1. The mediator proteins Mrc1 and Rad9 stimulate the activities of the Chk kinases in the replication stress and DNA damage pathways, respectively. Rad53 activates the Dun1 kinase, which has multiple downstream targets, including the ribonucleotide reductase inhibitor Sml1. The pathways are drawn separately but function in an integrated fashion.

We show below that Ty1 hypermobility in 19 rtt/genome integritymutants is partially or strongly dependent on the presence ofthe S-phase checkpoint proteins Rad9 and Rad53 and, in somecases, Rad24. To understand how S-phase checkpoint pathwaysregulate the mobility of Ty1, we have focused on rtt101 ${Delta}$ mutants,in which the DNA damage checkpoint is activated by defects inreplication fork progression (55, 61). Our data indicate thatneither Rtt101 nor an effector protein in the DNA damage pathwayinteracts with Ty1 cDNA to regulate mobility. Instead, the activatedDNA damage pathway triggers increases in processed Ty1 IN andRT proteins and in RT activity in Ty1 VLPs. These findings pointto a major role for S-phase checkpoint pathways in triggeringTy1 mobility when the integrity of DNA replication is compromised.

MATERIALS AND METHODS

Top

MATERIALS AND METHODS

Plasmids. Plasmid pGTy1^PR-^d¹his3AI[ ${Delta}$ 1], a derivative of pGTy1his3AI[ ${Delta}$ 1](77) in which amino acids 405 to 434 in the Ty1 PR domain arereplaced by the peptide sequence GAP (60), was constructed bycloning a 1.5-kb XhoI-BstEII fragment of plasmid pJB1105-d1,kindly provided by J. Boeke, in place of the 1.5-kb XhoI-BstEIIfragment in pGTy1his3AI[ ${Delta}$ 1].

Plasmid pYES2-3xHA:RTT101 was described previously (47). Derivativesof the URA3-marked 2µ plasmid BG1805, carrying the MMS1,ELG1, RAD27, and RRM3 open reading frames (ORFs) (27), wereobtained from Open Biosystems.

S. cerevisiae strains. The strains used in this study and details of their constructionare listed in Table S1 in the supplemental material. With theexception of the strains used in one experiment, all strainsare derivatives of the congenic strains BY4741 and BY4742 (9).Strain JC3212 is a derivative of strain BY4741 containing achromosomal Ty1his3AI[ ${Delta}$ 1]-3114 element. Strain JC3787 also harborsTy1his3AI[ ${Delta}$ 1]-3114 but has the opposite mating type (63). Derivativesof these strains harboring Ty1his3AI[ ${Delta}$ 1]-3114 and various genedeletions were made by genetic crossing with yeast ORF deletionstrains or by one-step gene replacement. Transformants weretested for loss of the wild-type allele and gain of the mutantallele by PCR. A minimum of three segregants or transformantsof each genotype were assayed in a semiquantitative patch testto determine the frequency of Ty1his3AI mobility. One representativesegregant or transformant was chosen for quantitative analysis.The oligomer sequences used for one-step gene replacement areavailable upon request.

To construct strains JC4331, JC4333, JC4335, and JC4337, anrtt101 ${Delta}$ ::LEU2 allele was introduced by PCR-mediated gene disruptioninto the spt3 ${Delta}$ ::kanMX derivative of BY4741, obtained from OpenBiosystems. Plasmids pGTy1his3AI[ ${Delta}$ 1] and pGTy1^PR-^d¹ his3AI[ ${Delta}$ 1]were then transformed into the spt3 ${Delta}$ ::kanMX and spt3 ${Delta}$ ::kanMX rtt101 ${Delta}$ ::LEU2derivatives of BY4741.

In contrast to the strains described above, strain JC4438 isa derivative of strain GRF167 (8). This strain and two congenicderivatives were constructed as follows. A single chromosomalTy1kanMXAI element was introduced into a trp1::hisG derivativeof GRF167 by inducing the expression of a Ty1kanMXAI elementfused to the GAL1 promoter on a yeast 2µ vector, as describedpreviously (15). The kanMXAI marker consists of a 104-bp artificialintron (AI) (94) inserted in antisense orientation into theSmaI site in the ORF of kanMX4 (89). The kanMXAI marker wasintroduced into plasmid pGTy1-H3Cla (26) between the TYB1 ORFand the 3' LTR in the transcriptional orientation opposite thatof Ty1 by gap repair in yeast. Plasmid pGTy2his3AI (16) wasintroduced into the Ty1kanMXAI-2910 trp1::hisG derivative ofstrain GRF167. An isolate with one chromosomal Ty2his3AI elementwas derived following induction of transposition to obtain strainJC4438. Strain JC4389 is an ade2 ${Delta}$ rtt101::ADE2 derivative ofJC4438 made by PCR-mediated disruption of ADE2 followed by PCR-mediateddisruption of RTT101. JC4639 is an rrm3 ${Delta}$ ::URA3 derivative ofJC4438 made by PCR-mediated gene disruption.

Ty1his3AI mobility assay. Strain JC3787 was used as the wild-type strain in each experiment.Each strain was grown in yeast extract-peptone-dextrose (YPD)broth at 30°C overnight, diluted 1:1,000 into 3 to 12 culturesof YPD broth, and grown at 20°C for 3 days. For samplesthat were treated with hydroxyurea (HU), various amounts ofa freshly prepared 1 M HU solution were added to the YPD cultureat the time of dilution. Aliquots of each culture were platedonto YPD and synthetic complete agar without histidine (SC-Hisagar) and grown for 3 to 4 days at 30°C. The frequency ofTy1his3AI transposition is the number of His⁺ prototrophs dividedby the number of viable cells in the same culture volume.

The mobilities of Ty1his3AI in strains carrying URA3-marked2µ-based plasmids were determined as follows. StrainsJC3212, JC3890, JC3901, and JC3913 were transformed with thepYES2 vector and pYES2-3xHA:RTT101, BG1805-ELG1, or BG1805-MMS1plasmid DNA. Four transformants from each transformation weregrown in SC-Ura glucose broth overnight at 30°C. Cultureswere diluted 1:100 into SC-Ura galactose broth and grown for2 days at 20°C. An equal volume of YPD was added to eachculture, which was grown for 18 h at 20°C. Aliquots of eachculture were plated on SC-Ura glucose agar and SC-Ura-His glucoseagar to determine the frequencies of Ty1HIS3 mobility eventsoccurring in cells that retained the URA3-based plasmid.

Quantitative analysis of Ty1 RNA and cDNA levels. Northern blot analysis of Ty1, Ty1his3AI, and PYK1 RNAs wasperformed as described previously (77). For Ty1 cDNA analysis,independent colonies of each strain were inoculated into 10ml YPD broth or YPD broth containing various concentrationsof HU, and cultures were grown at 20°C for 2 days. GenomicDNA prepared from each culture was digested with SphI, and Ty1cDNA was quantified as described previously (77).

Ty1kanMXAI and Ty2his3AI mobility assay. To determine the frequencies of Ty1kanMXAI and Ty2his3AI transposition,YPD cultures of strains JC4438, JC4389, and JC4639 grown overnightat 30°C were each diluted 1:1,000 into four independentYPD cultures and grown for 3 days at 20°C. Aliquots of eachculture were plated on YPD agar to determine the total numberof cells, YPD agar containing 200 µg/ml G418 to determinethe number of G418^r cells, and SC-His agar to determine thenumber of His⁺ cells.

Gag-green fluorescent protein (GFP) activity assay. Strain JC3808, in which GFP is fused to gag at nucleotide position+1203 within a single chromosomal Ty1 element (78), was crossedwith the rtt101 ${Delta}$ ::kanMX derivative of BY4741, and congenic wild-typeand rtt101 ${Delta}$ segregants containing Ty1-GFP-3566 were obtainedby tetrad dissection. The mean fluorescence intensity per cellwas determined for three wild-type and six rtt101 ${Delta}$ segregantsas described previously (78).

Purification of Ty1 VLPs. Ty1 VLPs were isolated from strains JC4331, JC4333, JC4335,and JC4337 by established methods (22, 25), but with the followingchanges in the galactose induction step. A 400-ml culture ofeach strain was grown for 24 h at 30°C in SC-Ura with 5%raffinose. Cells were pelleted and used as an inoculum for two1-liter cultures in SC-Ura with 1.6% galactose and 0.5% raffinose.The cultures were grown at 20°C for 12 h, at which time15 ml of a galactose-raffinose mixture (16%:5%) was added toboost induction, and the cultures were grown for an additional18 h at 20°C.

RT assays. Exogenous and endogenous assays for RT activity associated withTy1 VLPs were carried out as previously described (24, 95),using [ ${alpha}$ -³²P]dGTP incorporation to monitor cDNA synthesis.

Western blot analysis. Ty1 proteins were separated on 10% Tris-glycine gels (Invitrogen)and transferred to polyvinylidene difluoride membranes (Immobilon-P;Millipore). The membranes were hybridized with antibody B2 todetect IN, B8 to detect RT (25), or anti-VLP to detect Ty1 Gag(95). Horseradish peroxidase-linked anti-rabbit immunoglobulinG from donkey (Amersham) was used as a secondary antibody. Visualizationof antibody complexes was performed using chemiluminescence.

RESULTS

Top

RESULTS

The hypermobility of Ty1 in rtt/genome integrity mutants is triggered by Rad9. To investigate the hypothesis that S-phase checkpoints activateTy1 mobility when DNA replication is compromised, we analyzed19 rtt mutants with roles in maintaining DNA integrity duringS phase, including rtt101 ${Delta}$ , rtt107/esc4 ${Delta}$ , mms1/rtt108 ${Delta}$ , rtt109 ${Delta}$ ,elg1 ${Delta}$ /rtt110 ${Delta}$ , ctf4 ${Delta}$ , rrm3 ${Delta}$ , wss1 ${Delta}$ , tel1 ${Delta}$ , sae2 ${Delta}$ , rad18 ${Delta}$ , rad27 ${Delta}$ , rad50 ${Delta}$ ,xrs2 ${Delta}$ , mre11 ${Delta}$ , rad51 ${Delta}$ , rad52 ${Delta}$ , rad55 ${Delta}$ , and rad57 ${Delta}$ . This set includesfour new rtt mutants identified during this study: ctf4 ${Delta}$ , wss1 ${Delta}$ ,rad18 ${Delta}$ , and rad55 ${Delta}$ . Ctf4 is a component of the replisome progressioncomplex that links sister chromatid cohesion to DNA synthesis(23, 32). Wss1 interacts genetically with Ctf4 and is thoughtto play a role in stabilizing stalled or broken replicationforks (67). Rad18 is an E3 ubiquitin ligase that modifies thereplication processivity factor Pol30 to promote translesionsynthesis of DNA (35). Rad55 is a recombinational repair proteinthat stimulates strand exchange in association with Rad57. Weoriginally identified these four rtt strains by screening mutantswith S-phase defects for elevated Ty1 cDNA levels and subsequentlydemonstrated that they have elevated levels of Ty1 mobility(Table 1). As shown previously for other rtt/genome integritymutants, the absence of Ctf4, Wss1, Rad18, or Rad55 does notcause an increase in Ty1his3AI RNA or total Ty1 RNA, and infact, both RNAs are decreased $~$ 3-fold in the wss1 ${Delta}$ mutant (Fig.2). In addition to these 19 rtt/genome integrity mutants, twortt mutants, fus3 ${Delta}$ and med1 ${Delta}$ , which have no reported defects ingenome maintenance, were analyzed (Table 1). Fus3 is a mitogen-activatedprotein kinase, while Med1 is a component of the RNA polymeraseII transcription mediator complex.

View this table:
[in this window]
[in a new window]

TABLE 1. Suppression of rtt hypermobility phenotype by rad9 ${Delta}$

View larger version (24K):
[in this window]
[in a new window]

FIG. 2. Ty1 RNA levels are not increased in rtt mutants. Northern blot analysis is shown for total RNA from congenic strains whose relevant genotypes are indicated above each lane. The spt3 ${Delta}$ strain is a negative control. The Northern blot was probed with ³²P-riboprobes that hybridize to Ty1his3AI RNA (top), total Ty1 RNA (middle), and PYK1 RNA (bottom), the last as a loading control. Autoradiograms of the blot hybridized to each probe are shown. The ratio of ³²P activity in the Ty1his3AI band to ³²P activity in the PYK1 band and the ratio of ³²P activity in the Ty1 band to ³²P activity in the PYK1 band were determined by phosphorimaging. Ty1his3AI/PYK1 and Ty1/PYK1 RNA ratios for each strain, normalized to that for wild-type strain JC3787, are provided below each lane. WT, wild type.

To measure Ty1 mobility in rtt mutants, strains harboring thechromosomal Ty1his3AI-3114 element as well as an rtt ORF deletionwere generated by genetic crossing or one-step gene disruption.The his3AI retrotranscript indicator gene consists of the HIS3gene interrupted by an AI in the antisense orientation. Thehis3AI gene is inserted in the 3' untranslated region of a genomicTy1 element in the opposing transcriptional orientation; consequently,splicing of the AI from the Ty1his3AI RNA and reverse transcriptionof the spliced transcript generate Ty1 cDNA carrying a functionalHIS3 gene (15). When Ty1HIS3 cDNA is inserted into the genomeby transposition or recombination with genomic Ty1 sequences,the cell becomes His⁺. Therefore, the frequency of His⁺ prototrophformation is a measure of the retromobility of the Ty1his3AIelement.

To determine whether the DNA damage checkpoint is required forhypertransposition in rtt mutants, we deleted RAD9 and measuredthe frequency of His⁺ prototroph formation in congenic RAD9and rad9 ${Delta}$ strains (Table 1). Individual rtt mutations conferred3.6 (tel1 ${Delta}$ )- to 105 (elg1 ${Delta}$ )-fold increases in Ty1his3AI mobility.Deletion of RAD9 reduced Ty1 mobility about twofold in the wild-typeRTT strain or the rad24 ${Delta}$ RTT strain, and this deletion had asimilar or smaller effect in rtt mutant fus3 ${Delta}$ or med1 ${Delta}$ . Deletionof RAD9 had stronger effects in all 19 rtt/genome integritymutants tested, which showed 2.6 (wss1 ${Delta}$ )- to 73 (mms1 ${Delta}$ )-fold suppressionsof Ty1his3AI mobility. Complete suppression of the hypermobilityphenotype was seen only in mms1 ${Delta}$ rad9 ${Delta}$ mutants. Thus, Rad9 ispartially or strongly required for the hypertransposition phenotypein rtt/genome caretaker mutants but not in rtt mutants med1 ${Delta}$ and fus3 ${Delta}$ , in which genome integrity is not perturbed. In eightrtt/genome caretaker mutants, Ty1 hypermobility was mainly Rad9independent (rtt ${Delta}$ RAD9/rtt ${Delta}$ rad9 ${Delta}$ < rtt ${Delta}$ rad9 ${Delta}$ /RTT rad9 ${Delta}$ ) (Table1), while in the other 11 rtt/genome integrity mutants, hypermobilitywas predominantly dependent on the presence of RAD9 (rtt ${Delta}$ RAD9/rtt ${Delta}$ rad9 ${Delta}$ > rtt ${Delta}$ rad9 ${Delta}$ /RTT rad9 ${Delta}$ ) (Table 1).

Next, we asked whether rad9 ${Delta}$ suppresses Ty1 hypermobility inrtt/genome integrity mutants by reducing Ty1 cDNA levels (Fig.3). Deletion of RAD9 did not reduce the level of Ty1 cDNA ina wild-type or fus3 ${Delta}$ strain, and this deletion suppressed Ty1cDNA only 1.1-fold in a med1 ${Delta}$ mutant. In contrast, rad9 ${Delta}$ suppressedTy1 cDNA levels 2.5-fold or more in 16 of the 19 of the rtt/genomecaretaker mutants. These data suggest that Rad9, and thus theDNA damage checkpoint pathway, is involved in stimulating bothTy1 cDNA levels and hypermobility in the majority of rtt/genomeintegrity mutants. In elg1 ${Delta}$ , rtt109 ${Delta}$ , and rad27 ${Delta}$ mutants, however,Ty1 cDNA levels were similar or elevated in the absence of Rad9,even though Ty1his3AI mobility was reduced (Table 1). Hence,Rad9 may restrict Ty1 mobility in ways other than altering cDNAlevels in a minority of the rtt/genome integrity mutants.

View larger version (53K):
[in this window]
[in a new window]

FIG. 3. High Ty1 cDNA levels are suppressed in the absence of S-phase checkpoint proteins in many rtt/genome integrity mutants. (A) Schematic of a genomic Ty1 element with flanking genomic DNA, denoted by solid black bars. LTRs are indicated by tripartite boxes. A gray rectangle indicates the location of the Ty1 riboprobe used in Southern analysis. The approximate locations of the furthest 3' SphI site in Ty1 and the next SphI site in genomic DNA 3' of the Ty1 element are indicated, as is the end of the unintegrated Ty1 cDNA. (B) A representative Southern blot containing SphI-digested genomic DNA hybridized to a ³²P-labeled Ty1 riboprobe. Each lane contains genomic DNA from an independent colony grown to a 10-ml saturated culture at 20°C. The ³²P activity in the 1.7-kb Ty1 cDNA band (denoted with a C), which extends from the SphI site in Ty1 to the end of the unintegrated cDNA molecule, was quantified by phosphorimaging. Bands from two genomic Ty1 elements (denoted by G1 and G2) were quantified to determine the level of Ty1 cDNA relative to genomic DNA in each sample. The rtt mutation present in each set of strains is indicated below the blot. In addition, the genotypes at the RAD9 and RAD24 loci in each strain are indicated (+, RAD9 or RAD24; ${Delta}$ , rad9 ${Delta}$ or rad24 ${Delta}$ ). (C) Table showing the average Ty1 cDNA level in one to five independent DNA samples of rtt ${Delta}$ , rtt ${Delta}$ rad9 ${Delta}$ , rtt ${Delta}$ rad24 ${Delta}$ , and rtt ${Delta}$ rad53 ${Delta}$ sml1 ${Delta}$ strains relative to the average Ty1 cDNA level in two independent DNA samples of wild-type strain JC3787 run on each Southern blot. The average level of Ty1 cDNA in wild-type strain JC3787 on each Southern blot was assigned a value of 1. Data from the Southern blot shown, and eight additional Southern blots are compiled in the table. The control strains med1 ${Delta}$ and fus3 ${Delta}$ are rtt mutants with no known effect on genome integrity. Asterisks indicate values determined from one DNA sample.

Rad24 is partially required for Ty1 hypermobility in most rtt/genome integrity mutants. In contrast to Rad9, which functions primarily in the DNA damagepathway, Rad24 is involved in sensing DNA damage in both S-phasecheckpoint pathways but is not essential for either. Therefore,we asked whether Ty1 hypermobility is suppressed in a subsetof rtt mutants in the absence of Rad24 and, if so, whether thepattern of suppression differs from that for the correspondingrtt rad9 ${Delta}$ mutants. Deletion of RAD24 caused a 1.7-fold decreasein Ty1 mobility in a wild-type strain, and this deletion hada similar effect in a med1 ${Delta}$ or fus3 ${Delta}$ mutant. In contrast, therad24 ${Delta}$ mutation suppressed hypermobility more than fourfold in12 of 14 rtt mutants tested (Fig. 4). Complete suppression ofhypermobility was observed when RAD24 was deleted in an rtt107 ${Delta}$ mutant, which showed only partial suppression by rad9 ${Delta}$ , and inrtt101 ${Delta}$ and rad51 ${Delta}$ mutants, which displayed strong suppressionby rad9 ${Delta}$ . It is likely that Rad24 is required for most of thecheckpoint pathway activation of Ty1 hypermobility in thesestrains, since suppression in an rtt107 ${Delta}$ rad24 ${Delta}$ rad9 ${Delta}$ mutant wasequivalent to that in an rtt107 ${Delta}$ mutant (data not shown). Onthe other hand, the rad24 ${Delta}$ mutation failed to suppress Ty1his3AIhypermobility in a sae2 ${Delta}$ mutant, which was fully suppressed byrad9 ${Delta}$ (Fig. 4, Table 1). Hence, the effect of rad24 ${Delta}$ on Ty1 mobilitycan differ from that of rad9 ${Delta}$ in some rtt mutants.

View larger version (20K):
[in this window]
[in a new window]

FIG. 4. A rad24 ${Delta}$ mutation suppresses Ty1 hypermobility in most rtt/genome integrity mutants. Ty1his3AI mobility is shown for RAD24 (light gray, dark gray, or black bars) and rad24 ${Delta}$ (white bars) derivatives of a wild-type strain and isogenic rtt derivatives. Light gray bars, RTT strains; dark gray bars, rtt mutants without defects in genome maintenance; black bars, rtt/genome integrity mutants. Error bars; standard error.

Next, we asked whether the rad24 ${Delta}$ mutation reduces the levelof Ty1 cDNA in each of 12 rtt/genome integrity mutants (Fig.3C). Deletion of RAD24 reduced Ty1 cDNA 2.5-fold in the wild-typestrain and only slightly in the med1 ${Delta}$ mutant. In the majorityof rtt/genome integrity mutants tested, rad24 ${Delta}$ did not reduceTy1 cDNA levels any further than it did in a wild-type strain.In contrast, Ty1 cDNA was substantially reduced in rtt101 ${Delta}$ , rad57 ${Delta}$ ,and rad51 ${Delta}$ mutants when RAD24 was deleted. Thus, in most rtt/genomeintegrity mutants, Rad24 appears to suppress Ty1 mobility subsequentto the accumulation of cDNA.

The replication stress checkpoint pathway triggers Ty1 hypermobility. We intended to analyze the epistatic relationship between rttmutations and the mrc1 ${Delta}$ mutation to determine whether the replicationstress pathway, like the DNA damage pathway, triggers Ty1 mobilityin rtt/genome integrity mutants. However, unlike all the otherS-phase checkpoint mutations analyzed here and in a previousstudy (78), the mrc1 ${Delta}$ mutation conferred a hypertranspositionphenotype on the wild-type strain. The Ty1his3AI mobility frequencywas 8.00 x 10^–8 ± 1.25 x 10^–8 in the wild-typestrain and 50.8 x 10^–8 ± 9.1 x 10^–8 in theisogenic mrc1 ${Delta}$ strain. Since Mrc1 has a role in restarting ofstalled replication forks in addition to its role in replicationcheckpoint signaling (43, 68), stalled replication forks couldtrigger the DNA damage checkpoint in mrc1 ${Delta}$ mutants, resultingin Ty1 hypermobility. In fact, Ty1 hypermobility was reducedwhen a plasmid carrying the mrc1^AQ allele that is defectivefor replication checkpoint signaling but competent for replicationrestart (68) was introduced into the mrc1 ${Delta}$ strain (data not shown).Activation of the DNA damage checkpoint in mrc1 ${Delta}$ mutants complicatesthe use of this mutation to examine the role of the replicationstress checkpoint pathway in triggering Ty1 hypertransposition.Therefore, we determined the effect of deleting RAD53, whichis required for the activities of both S-phase checkpoints,in a subset of rtt/genome caretaker mutants. The rad53 ${Delta}$ mutationwas introduced subsequent to deletion of SML1, which suppressesthe inviability caused by a rad53 ${Delta}$ mutation. Deletion of RAD53and SML1 had only a minor effect on Ty1his3AI mobility in awild-type strain or med1 ${Delta}$ mutant (Fig. 5A). In an elg1 ${Delta}$ mutant,Ty1his3AI hypermobility was completely suppressed in the absenceof Rad53 and Sml1, even though neither rad9 ${Delta}$ nor rad24 ${Delta}$ causedstrong suppression. These results suggest that the replicationstress checkpoint pathway activates Ty1 in elg1 ${Delta}$ mutants. Ty1his3AIhypermobility in an rtt101 ${Delta}$ or rad55 ${Delta}$ strain was fully suppressedin the absence of Rad53, as it was in the absence of Rad9, consistentwith our conclusion that the DNA damage checkpoint pathway activatesTy1 in these mutants. In rrm3 ${Delta}$ and rtt109 ${Delta}$ mutants, Ty1his3AImobility was moderately decreased in the absence of Sml1 andRad53, indicating that Ty1 hypermobility in rrm3 ${Delta}$ and rtt109 ${Delta}$ mutants is partially dependent on the S-phase checkpoints. Comparisonof the Ty1 cDNA levels in four of these five rtt ${Delta}$ mutants tothose in the corresponding rtt ${Delta}$ rad53 ${Delta}$ sml1 ${Delta}$ mutant fully supportsthe conclusions reached by Ty1his3AI mobility assays (Fig. 3C).

View larger version (30K):
[in this window]
[in a new window]

FIG. 5. Ty1 is activated through S-phase checkpoint pathways in rtt/genome integrity mutants. (A) Ty1his3AI mobility in RAD53 SML1 (light gray bars, RTT strains; dark gray bars, rtt mutants without defects in genome integrity; black bars, rtt/genome integrity mutants) and rad53 ${Delta}$ sml1 ${Delta}$ (white bars) derivatives of a wild-type strain and congenic rtt mutants. WT, wild type. (B) Ty1his3AI mobility in wild-type and isogenic rad9 ${Delta}$ strains grown in the presence of increasing concentrations of HU. The level of Ty1 cDNA relative to that of genomic Ty1 DNA in two independent DNA samples was determined as described in the legend to Fig. 2. (C) Ty1his3AI mobility in a wild-type strain and three mutant rtt strains carrying the pGAL1 vector or one of three pGAL1:RTT gene expression constructs. Ty1his3AI mobility, which is the average of the number of His⁺ Ura⁺ prototrophs divided by the number of total Ura⁺ cells per culture, was determined for each strain. Error bars, standard errors.

To further test the hypothesis that replication stress can induceTy1 mobility, we measured Ty1his3AI mobility and Ty1 cDNA levelswhen cells were treated with low levels of HU, which causesnucleotide depletion and replication fork pausing (Fig. 5B).The frequency of Ty1his3AI mobility increased progressivelyas the concentration of HU was increased from 25 to 100 mM.The level of Ty1 cDNA was also elevated at 25 mM HU and peakedat 50 mM. The lower level of Ty1 cDNA at 100 mM HU may indicatethat not only DNA replication but also synthesis of cDNA inVLPs is inhibited at higher HU concentrations. The HU-mediatedincreases in Ty1 cDNA and mobility do not require Rad9, consistentwith activation occurring through the replication stress pathwayrather than the DNA damage pathway.

Mec1 and Dun1 are essential for Ty1his3AI hypermobility in rtt101 ${Delta}$ mutants. To examine how the DNA damage checkpoint pathway induces Ty1mobility in response to DNA lesions, we focused on the rtt101 ${Delta}$ mutant, in which the DNA damage checkpoint is known to be activatedby defects in replication fork progression (55, 61). Elevatedlevels of Ty1 mobility and cDNA in rtt101 ${Delta}$ mutants are stronglysuppressed by rad9 ${Delta}$ , rad24 ${Delta}$ , and rad53 ${Delta}$ sml1 ${Delta}$ mutations (Table 1and Fig. 3C, 4, and 5A), suggesting that the DNA damage checkpointpathway specifically induces Ty1 mobility in rtt101 ${Delta}$ mutants.In support of this model, deletion of MEC1 and SML1 also fullysuppressed Ty1his3AI mobility in an rtt101 ${Delta}$ mutant, while ansml1 ${Delta}$ mutation alone modestly increased Ty1his3AI mobility inthe rtt101 ${Delta}$ mutant (Table 2). Chk1 is phosphorylated by Mec1and regulates the transition from metaphase to anaphase in theDNA damage checkpoint pathway (74). A chk1 ${Delta}$ mutation had littleeffect on Ty1his3AI mobility in a wild-type strain and suppressedTy1his3AI hypermobility in the rtt101 ${Delta}$ mutant only 1.9-fold.Thus, Chk1 does not play a pivotal role in the activation ofTy1 in rtt101 ${Delta}$ mutants. The kinase Dun1 acts downstream of Mec1and Rad53 to induce expression of damage-inducible genes anddegradation of Sml1, an inhibitor of ribonucleotide reductase(Fig. 1). Deletion of DUN1 or DUN1 and SML1 slightly increasedTy1 mobility in a wild-type strain, but Ty1his3AI mobility inan rtt101 ${Delta}$ mutant was partially suppressed by the dun1 ${Delta}$ mutationand completely suppressed by the dun1 ${Delta}$ sml1 ${Delta}$ mutations (Table2). Ty1 cDNA levels in an rtt101 ${Delta}$ mutant were also strongly suppressedby either the mec1 ${Delta}$ sml1 ${Delta}$ mutations or the dun1 ${Delta}$ sml1 ${Delta}$ mutations(data not shown). Thus, checkpoint proteins Mec1 and Dun1 arerequired to mobilize Ty1 elements in rtt101 ${Delta}$ mutants.

View this table:
[in this window]
[in a new window]

TABLE 2. Effect of checkpoint gene deletions on Ty1his3AI mobility in rtt101 ${Delta}$ mutants

Overexpression of Rtt101 does not repress Ty1his3AI mobility. One prediction of the hypothesis that Rtt101 inhibits Ty1 mobilityindirectly by suppressing DNA lesions that trigger the DNA damagecheckpoint is that overexpression of Rtt101 will not enhancethe inhibition of Ty1 transposition. In contrast, overexpressionof Rtt101 should enhance the repression of Ty1his3AI mobilityif Rtt101 interacts directly with Ty1 cDNA. When RTT101, orELG1 or MMS1 for comparison, was expressed from the GAL1 promoteron a high-copy-number vector, Ty1his3AI mobility in a wild-typestrain increased (Fig. 5C). Thus, overexpression of RTT101,ELG1, or MMS1 failed to enhance the repression of Ty1 mobilityconferred by endogenous levels of the same protein. This wasnot due to lack of pGAL1:RTT expression, since expression ofELG1, RTT101, or MMS1 reduced the hypermobility phenotype ofan elg1 ${Delta}$ , rtt101 ${Delta}$ , or mms1 ${Delta}$ mutant, respectively (Fig. 5C). Wealso attempted to overexpress RAD27 and RRM3, because both havebeen proposed to interact with Ty1 cDNA (77, 81). However, expressionof pGAL1:RAD27 failed to complement the rad27 ${Delta}$ mutation, andstrains harboring the pGAL1:RRM3 plasmid did not grow in mediacontaining galactose. Nonetheless, our results are consistentwith the conclusion that Rtt101, Mms1, and Elg1 act indirectlyto reduce Ty1 cDNA levels.

Mobility of Ty1 but not Ty2 is elevated in rtt101 ${Delta}$ mutants. Next, we tested the hypothesis that a downstream protein inthe DNA damage pathway directly interferes with degradationof Ty1 cDNA by nuclear DNA repair enzymes in the rtt101 ${Delta}$ mutant.If cDNA were the direct target of the DNA damage checkpointpathway, we would expect the mobilities of elements from differentTy families to be increased in the absence of Rtt101, sincecDNA is an intermediate in all retrotransposition events. Therefore,we asked whether the mobilities of both Ty1 and Ty2 elements,which are closely related, are increased in an rtt101 ${Delta}$ mutant.An rtt101 ${Delta}$ mutation was introduced into a strain harboring agenomic Ty1kanMXAI element and a genomic Ty2his3AI element.Ty1kanMXAI mobility was increased 34-fold upon deletion of RTT101,while Ty2his3AI increased only 2-fold (Table 3). In comparison,the difference between the stimulation of Ty1kanMXAI mobilityand that of Ty2his3AI mobility was much less significant whenRRM3 was deleted (21-fold and 8-fold, respectively). The findingthat rtt101 ${Delta}$ preferentially activates the mobility of a Ty1 elementargues against a direct role for the DNA damage pathway in preventingdegradation of Ty cDNA by DNA repair enzymes.

View this table:
[in this window]
[in a new window]

TABLE 3. Effect of rtt101 ${Delta}$ and rrm3 ${Delta}$ mutations on Ty1 and Ty2 element mobility^a

Levels of RT and IN proteins and RT activity are increased in rtt101 ${Delta}$ mutants. In rtt101 ${Delta}$ mutants, Ty1 cDNA levels and levels of cDNA integrationupstream of tRNA genes, which are preferred regions for integration,are elevated, yet Ty1 RNA is not increased (77). Furthermore,we found that deletion of RTT101 did not increase the levelof Ty1 Gag-GFP fusion protein produced from the chromosomalTy1-GFP-3566 element (78). (The mean fluorescence intensityper particle was 5.70 ± 0.02 in RTT101 strains versus4.44 ± 0.23 in congenic rtt101 ${Delta}$ mutants.) Thus, the elevatedlevel of Ty1 cDNA in an rtt101 ${Delta}$ mutant is not explained by anincrease in translation of Ty1 RNA. Therefore, we asked whetherthe rtt101 ${Delta}$ mutation confers an increase in VLP-associated RTprotein or activity levels that could result in increased cDNAlevels. The RT protein produced from endogenous Ty1 elementsis difficult to detect by Western blotting of whole-cell proteinsor sucrose gradient fractions enriched for Ty1 VLPs, particularlyin the S288C strain background used in this study (14, 71) (datanot shown). To overcome this problem, a Ty1 element was expressedfrom the GAL1 promoter on a high-copy-number plasmid (pGTy1).In addition, a pGTy1 element containing a 30-amino-acid deletionin the C terminus of Gag and the overlapping N terminus of PR(PR-d1) was expressed (51). Since the N-terminal region of PRis divergent between Ty1 and Ty2 elements, we reasoned thatthis domain of Ty1 was a possible target of the DNA damage checkpointin rtt101 ${Delta}$ mutants.

The pGTy1 and pGTy1^PR-^d¹ elements were introduced into RTT101spt3 ${Delta}$ and rtt101 ${Delta}$ spt3 ${Delta}$ strains. (The spt3 ${Delta}$ mutation blocks transcriptionof endogenous Ty1 elements but not pGTy1 elements.) Ty1 VLP-enrichedsucrose gradient fractions were analyzed by Western blottingwith anti-VLP, anti-RT, and anti-IN polyclonal antibodies (Fig.6A). Anti-VLP antibodies detected both the p49-Gag precursorprotein and proteolytic processed p45-Gag in VLP fractions fromcells expressing the wild-type Ty1 element. Only one band wasdetected in VLPs derived from Ty1^PR-^d¹, because the Gag proteinencoded by the Ty1^PR-^d¹ gene lacks most of the C terminus ofGag, which is removed by proteolytic processing (51). Totallevels of Gag produced from wild-type Ty1 or Ty1^PR-^d¹ were notincreased in rtt101 ${Delta}$ mutants, consistent with the result thatRtt101 affects a posttranslational step in transposition. However,there was less p49-Gag precursor and more processed p45-Gagin the rtt101 ${Delta}$ strain expressing wild-type Ty1 than in the RTT101strain (Fig. 6A, top). Thus, Gag processing may be more efficientin rtt101 ${Delta}$ mutants.

View larger version (18K):
[in this window]
[in a new window]

FIG. 6. Deletion of RTT101 results in increased levels of Ty1 IN and RT proteins and RT activity in VLPs derived from a pGTy1 element but not a pGTy1^PR-^d¹ element. (A) Western blot analysis of VLP-enriched sucrose gradient fractions. (Top) 0.7 µg of each VLP-enriched fraction was probed with anti-VLP ( ${alpha}$ -VLP) antibody, which recognizes p49-Gag and p45-Gag. (Middle) 2.1 µg of each VLP-enriched fraction was probed with anti-RT antibody, which recognizes the Gag-Pol fusion protein, processing intermediates (not labeled), and RT. (Bottom) 2.1 µg of each VLP-enriched fraction was probed with anti-IN antibody, which recognizes Gag-Pol, processing intermediates (not labeled), and IN. (B) RT activity of VLP fractions from each strain on an exogenously added template. Units are 10⁵ ³²P counts µg^–1 h^–1.

In contrast to what was found for Gag proteins, marked increasesin the levels of processed Pol proteins RT and IN relative tothe levels in the RTT101 strain were observed when wild-typeTy1 was expressed in the rtt101 ${Delta}$ mutant (Fig. 6A, middle andbottom, lanes 1 and 3). The activity of RT on an exogenouslyadded RNA-DNA primer template was also increased 21-fold inthe absence of RTT101 (Fig. 6B). In addition, the ratio of totalGag to p199-Gag:Pol is elevated in the rtt101 ${Delta}$ mutant (Fig. 6A,middle, lanes 1 and 3), and this relative increase in Gag-Polmay contribute to the elevated levels of mature RT and IN. Intriguingly,these rtt101 ${Delta}$ -mediated increases in Gag-Pol, RT, and IN werecompletely blocked when Ty1^PR-^d¹ was expressed. There were moderateincreases in both RT protein and exogenous RT activity in VLPsfrom Ty1^PR-^d¹ relative to the levels in VLPs from wild-typeTy1 in an RTT101 strain (Fig. 6A, middle, lanes 1 and 2, andB); however, the levels of RT, IN, and exogenous RT activityin VLPs from the Ty1^PR-^d¹ element were equivalent in the rtt101 ${Delta}$ and RTT101 strains. These data raise the possibility that theDNA damage checkpoint pathway activated in an rtt101 ${Delta}$ mutanttargets the N-terminal domain of PR that is deleted in Ty1^PR-^d¹.

DISCUSSION

Top

DISCUSSION

A large and diverse group of genome caretakers, particularlyproteins that function during replication, have been implicatedin the regulation of Ty1 (10, 52, 53, 71, 77, 81). Several Rtt/genomecaretakers, including Elg1, Rtt101, Rtt109, Rrm3, and Sgs1,also restrict the mobility of the Ty3 retrotransposon, whichis more closely related to mammalian retroviruses than is Ty1(37). Moreover, human orthologs of several other Rtt/genomecaretakers, including Rad3, Ssl2, Rad52, and Rad18, have beenfound to impede retroviral replication (48, 54, 93). In thisstudy, we tested the hypothesis that transposition is derepressedin the absence of Rtt/genome integrity factors by a common mechanism,i.e., stimulation of Ty1 mobility by S-phase checkpoint pathways.In mammalian systems, some studies have suggested that DNA damageresponse proteins are required for retroviral integration orfor survival of cells after retroviral infection (17, 18, 49,65), but other studies found no evidence for a contributionof checkpoint proteins to retroviral transduction (2, 19). Inyeast, S-phase checkpoint pathways are dispensable for the mobilityof Ty1 elements (78), and accordingly, we observed no more thana 2.5-fold decrease in Ty1 retromobility or cDNA levels in rad9,rad24, mec1, rad53, or dun1 mutants (Fig. 3, 4, and 5A and Tables1 and 2). We have not ruled out the possibility that defectsin the S-phase checkpoint pathway alter the ratio of Ty1 cDNAthat is integrated into the genome to cDNA that recombines withpreexisting Ty1 sequences. Clearly, though, Ty1 cDNA integrationcan occur in rad9 ${Delta}$ and rad24 ${Delta}$ mutants, since the frequency ofTy1his3AI mobility in a rad9 ${Delta}$ or rad24 ${Delta}$ mutant is not reducedwhen RAD52, a gene required for Ty1 cDNA recombination, is deleted(Table 1 and Fig. 4).

Although S-phase checkpoint proteins are not essential for Ty1mobility in wild-type cells, they are required to differentdegrees for the elevated levels of Ty1 cDNA and retrotranspositionin all of the 19 rtt/genome integrity mutants tested. In contrast,S-phase checkpoint proteins play no role in the Ty1 hypermobilityphenotypes of two rtt mutants without known defects in DNA replication(Table 1 and Fig. 3 and 5A). The various degrees of influenceof Rad9, Rad24, and Rad53 on transposition in rtt/genome integritymutants suggest that the DNA damage pathway, the replicationstress pathway, or a combination of the two can activate Ty1cDNA and mobility. Moreover, the variation is consistent withthe diverse roles of Rtt/genome caretakers in the maintenanceof genome integrity during S phase, since their absence is likelyto result in a variety of DNA lesions.

Various contributions of the replication stress pathway and DNA damage pathway to hypertransposition in rtt/genome caretaker mutants. Strong suppression of the Ty1 hypermobility phenotype by rad9 ${Delta}$ indicated that hypermobility is triggered primarily throughthe DNA damage checkpoint pathway in 11 different rtt/genomeintegrity mutants, including the recombinational repair mutantsrad50 ${Delta}$ , mre11 ${Delta}$ , xrs2 ${Delta}$ , rad51 ${Delta}$ , rad52 ${Delta}$ , and rad55 ${Delta}$ (Table 1). Rattrayet al. (71) found that recombinational repair proteins blockTy1 transposition at a posttranslational step and suggestedthat these repair proteins promote degradation of Ty1 cDNA.However, the half-life of Ty1 cDNA in a rad50 ${Delta}$ mutant is equivalentto that in a wild-type strain (78), arguing against the hypothesisthat Ty1 cDNA is stabilized in recombinational repair mutants.Studies of human cells using a recombinant human immunodeficiencyvirus type 1 (HIV-1) vector indicate that the human RAD52 proteininhibits HIV-1 integration by binding HIV-1 cDNA and preventingthe binding of factors necessary to promote integration (48).The model in which IN and Rad52 compete for binding to retroelementcDNA is not inconsistent with our model in which Rad52 inhibitsTy1 by repressing lesions that activate the DNA damage pathway.Perhaps Ty1 cDNA synthesis is stimulated through the activatedDNA damage pathway in rad52 ${Delta}$ mutants, as we have proposed forrtt101 ${Delta}$ mutants, and in addition, the fraction of Ty1 cDNA thatis utilized for integration is increased because Rad52 cannotcompete with IN for binding Ty1 cDNA. An argument against thishypothesis is that Rattray et al. (71) found no evidence forincreased levels of RT protein or activity in rad52 ${Delta}$ and otherrecombinational repair mutants. However, the DNA damage responseis strongly required in both rtt101 ${Delta}$ and rad52 ${Delta}$ mutants, and ourdata on rtt101 ${Delta}$ mutants suggest that activation of the DNA damagepathway leads to higher levels of Ty1 RT and IN proteins (Fig.5A). Thus, we suggest that the difference in our results isdue to the fact that Rattray et al. (71) quantified the levelsof RT protein and activity in endogenous VLPs, whereas we haveanalyzed VLPs from pGTy1 elements. The yield of VLPs from endogenousTy1 elements is low, and the purity varies from one preparationto the next (13, 71). Moreover, accurate measurement of RT activityin endogenous VLP preparations is complicated by the low concentrationsof RT protein. Therefore, an effect of the rad52 ${Delta}$ mutation onTy1 proteins may have gone undetected.

In elg1 ${Delta}$ mutants, it is likely that Ty1 hypermobility is triggeredprimarily through the replication checkpoint pathway, as theelevated Ty1 cDNA and mobility levels were reduced only slightlyby a rad9 ${Delta}$ mutation, but they were completely suppressed by arad53 ${Delta}$ mutation (Table 1 and Fig. 3C and 5A). Consistent withElg1 acting indirectly to modulate the mobility of Ty1, overexpressionof ELG1 failed to enhance repression of Ty1 in a wild-type strain.Elg1 interacts with Rad27 and Pol30 and likely functions atthe replication fork, but Elg1 is not required for the replicationstress checkpoint (4, 42). Therefore, DNA lesions arising inthe absence of Elg1 may activate the replication stress checkpoint,which in turn promotes Ty1 mobility. Further evidence that thereplication stress pathway can trigger Ty1 cDNA accumulationand retromobility is provided by the finding that treatmentof cells with low levels of HU caused a Rad9-independent increasein Ty1 cDNA and retromobility (Fig. 5B).

Both the DNA damage checkpoint and the replication stress checkpointcould contribute to the hypertransposition phenotype of an rtt107 ${Delta}$ mutant, which was partially suppressed by rad9 ${Delta}$ but stronglysuppressed by rad24 ${Delta}$ . In rtt/genome integrity mutants such asrtt109 ${Delta}$ , rrm3 ${Delta}$ , and rad27 ${Delta}$ , S-phase checkpoint pathways appearto play a modest role in stimulating Ty1 mobility. Ty1 mobilityand cDNA levels remained partially elevated even when RAD53,which is required for the activities of both pathways, was deletedin rtt109 ${Delta}$ and rrm3 ${Delta}$ mutants (Fig. 5A). Hypertransposition inthese mutants could be influenced indirectly by an S-phase checkpointpathway or be regulated by more than one mechanism. In the rad27 ${Delta}$ mutant, neither the rad9 ${Delta}$ nor the rad24 ${Delta}$ mutation strongly suppressedthe elevated Ty1 mobility or cDNA levels. (We have not analyzeda rad27 ${Delta}$ rad53 ${Delta}$ mutant.) A previous study found that Rad27 repressesTy1 retrotransposition at a step subsequent to reverse transcription,consistent with Rad27 interacting directly with Ty1 cDNA (81).

The DNA damage pathway causes increased Ty1 cDNA synthesis by altering the level and activity of Ty1 RT. Although Ty1 and Ty2 elements are closely related LTR retroelementsin yeast (40, 41, 45), both the yeast protein kinase Fus3 andAPOBEC3G, a human retroviral restriction factor that interactswith Ty1 Gag when expressed in yeast, have been shown to preferentiallyrepress the activity of Ty1 (13, 21). Remarkably, Rtt101 alsorepresses Ty1 more strongly than Ty2. In contrast, the retromobilitiesof both Ty1 and Ty2 elements were elevated in rrm3 ${Delta}$ mutants,although a modest preference for stimulation of Ty1 mobilitywas observed. Our interpretation is that Rrm3 interacts directlywith Ty1 and Ty2 cDNA to inhibit mobility or that Ty1 and Ty2cDNAs are incorporated into the genome at DNA lesions createdin the absence of Rrm3. On the other hand, it is unlikely thatRtt101 represses Ty cDNA directly or that an effector proteinin the DNA damage signaling pathway prevents degradation ofTy1 cDNA in rtt101 ${Delta}$ mutants. Since Ty1 and Ty2 cDNAs are structurallyequivalent molecules, the DNA repair enzymes that recognizethe unprotected ends of linear cDNA and cause their degradationare not expected to differentiate between them. The observationthat Ty1 mobility is not repressed by overexpression of RTT101is also inconsistent with a direct interaction between Rtt101and cDNA. Therefore, we propose that the preferential activationof Ty1 elements in rtt101 ${Delta}$ mutants results from a differencein amino acid sequence between analogous Ty1 and Ty2 proteins,resulting in a difference in their abilities to interact withan effector protein in the DNA damage pathway.

If an effector protein in the DNA damage pathway does not interactdirectly with Ty1 cDNA, then how does the checkpoint pathwayactivate Ty1? Hypermobility in an rtt101 ${Delta}$ mutant was fully suppressedby rad9 ${Delta}$ , rad24 ${Delta}$ , rad53 ${Delta}$ sml1 ${Delta}$ , mec1 ${Delta}$ sml1 ${Delta}$ , and dun1 ${Delta}$ sml1 ${Delta}$ mutations.Thus, one possibility is that a Ty1 protein is the target ofthe Dun1 kinase and that phosphorylation of this protein resultsin hypermobility. We were unable to identify an amino acid motifin Ty1 Gag or Pol that matched the characterized Dun1 phosphorylationmotif in Sml1 (87), and Dun1 is a nuclear protein, while Ty1VLPs are cytoplasmic. Since Sml1 is not necessary for the rtt101 ${Delta}$ hypertransposition phenotype, Dun1 may work through anotherdownstream effector protein or protein cascade, which in turninteracts directly with Ty1 proteins in the cytoplasm to regulatetheir activity. Indeed, activation of the S-phase checkpointscan regulate the nuclear export of downstream target proteinssuch as ribonucleotide reductase subunits Rnr2 and Rnr4 (92).

Analysis of Ty1 VLP proteins in an rtt101 ${Delta}$ mutant indicates thata Ty1 protein(s) could be the target of the activated DNA damagepathway. Processed RT and IN proteins and RT activity in VLP-enrichedfractions from cells expressing a pGTy1 element were substantiallyincreased in an rtt101 ${Delta}$ mutant (Fig. 6). The Gag-Pol precursorprotein was also modestly elevated. The higher levels of processedPol proteins are probably due to enhanced protein stabilityrather than increased translational frameshifting, because conditionsthat increase frameshifting efficiency are inhibitory to transposition(3, 44, 91). Perhaps modification of Pol proteins or the Gag-Polprotein precursor by an effector protein in the DNA damage pathwayresults in increased Pol protein stability and processing. Incontrast to the results obtained with the pGTy1 element, levelsof RT and IN proteins and RT activity in Ty1 VLPs derived froma mutant pGTy1^PR-^d¹ element were not increased by the rtt101 ${Delta}$ mutation. It is possible, therefore, that the 30-amino-aciddomain in the N terminus of PR and the C terminus of p49-Gag,which is absent from proteins derived from pGTy1^PR-^d¹ (50),is a target of the DNA damage pathway. Because the 30-amino-aciddomain is divergent between Ty1 and Ty2 elements, posttranslationalmodification of the domain could underlie the preferential mobilizationof Ty1 in rtt101 ${Delta}$ mutants. Furthermore, the N-terminal regionof PR plays a nucleocapsid-related function in the initiationof reverse transcription (51), so modification of this regioncould conceivably affect the efficiency of cDNA synthesis. Ifthis domain is modified within Gag-Pol, enhanced processingfor forming mature RT and IN might result. Modification of the30-amino-acid domain in the C terminus of p49-Gag could alsoenhance the processing of p49-Gag to p45-Gag in the rtt101 ${Delta}$ mutant(Fig. 6A).

It is important to note that although levels of RT and IN proteinsderived from a pGTy1 element were substantially elevated inan rtt101 ${Delta}$ mutant, the frequency of pGAL1Ty1his3AI mobility wasnot increased (data not shown). The latter result was anticipated,since most, if not all, repressors of chromosomal Ty1 elementsdo not inhibit pGTy1 elements (30). The fact that pGTy1 RT andIN levels, but not mobility, increase in rtt101 ${Delta}$ mutants suggeststhat transposition of pGTy1 elements is limited at a late step,such as integration.

A role for the S-phase checkpoint pathways in activating transpositionprovides a common explanation for involvement of diverse genomecaretakers in restricting Ty1 mobility. The observation thatTy1 reverse transcription and mobility are activated throughS-phase checkpoint pathways supports the hypothesis that Ty1activity is part of the cellular response to DNA damage (78).Fragments of Ty1 cDNA are captured at sites of DNA breaks inthe absence of homologous recombination (62, 84, 96), suggestinga potential role for cDNA in the repair of double-strand breaks.Also, cDNA of the Y' subtelomeric repeat, synthesized in Ty1VLPs, is introduced into the genome at high levels in telomerase-negativesurvivors, where it may play a role in telomere maintenance(57). Ty1 transposition itself, while often inconsequentialor deleterious, is also capable of creating adaptively favorablemutations (7, 90). Hence, Ty1 transposition or components ofTy1 VLPs could play functional roles in genome reorganizationor cell survival when the genome is compromised.

ACKNOWLEDGMENTS

We thank the Wadsworth Center Immunology core for flow cytometryand Patrick Maxwell for comments on the manuscript.

Our work is supported by National Institutes of Health grantGM52072 to M.J.C. The work of D.J.G. and S.M. was sponsoredby the Center for Cancer Research of the National Cancer Institute,Department of Health and Human Services.

The contents of this publication do not necessarily reflectviews or policies of the Department of Health and Human Services,nor does mention of trade names, commercial products, or organizationsimply endorsement by the U.S. Government.

FOOTNOTES

* Corresponding author. Mailing address: Center for Medical Sciences, Wadsworth Center, P.O. Box 2002, Albany, NY 12201-2002. Phone: (518) 473-6078. Fax: (518) 474-3181. E-mail: curcio{at}wadsworth.org

Published ahead of print on 8 October 2007.

Supplemental material for this article may be found at http://mcb.asm.org/.

REFERENCES

Top