Yeast cultures

Yeast cells were grown at 30 °C. Media used in this study: YPAD (1% yeast extract, 2% bacto-peptone, 2% glucose, and 20 mg/ml adenine), synthetic defined (SD) (0.17% yeast nitrogen base without amino acids, 0.5% ammonium sulphate, supplemented with amino acids), synthetic complete (SC) (SD with 2% glucose, 2% galactose or 2% raffinose). Solid media were prepared by adding 2% agar. Genotoxic agents were added at the appropriate concentration to molten selective SC medium and poured into petri dish plates to get solidified.

Western blot

Western blots were performed on proteins extracted with 10% TCA following standard procedures (Kushnirov 2000). Samples were loaded on acrylamide gels, migrated in an SDS-containing buffer, transferred to nitrocellulose membranes and hybridized with anti-FLAG (1:20,000, F3165, Sigma-Aldrich) or anti-Actin (1:2000, ab8224, Abcam) primary antibodies and anti-mouse HRP-conjugated secondary antibody (Mouse: 1:6000, A4416, Sigma; Rabbit: 1:10,000, A6154, Sigma).

Detection of Rad52 and Rad51 foci

Spontaneous Rad52-YFP foci from mid-log cells carrying plasmid pWJ1344 were visualized and counted by fluorescence microscopy as previously described ((Lisby et al 2001) using a Leica DC 350F microscope. More than 200 S/G2 cells were inspected for each experimental replica. Experiments were performed with yeast cultures of transformants carrying plasmids with tet::RECQL5 or GAL::RECQL5 grown in selective SC medium with 2% glucose or SC with 2% galactose, respectively. To study the effect of RNase H1 (RNH1) overexpression on Rad52 foci formation cells were also transformed with a plasmid carrying RNH1 under the control of tet promoter (pCM184RNH1) (Santos-Pereira et al 2013). For Rad51-YFP foci the strain ML149-8A was used (Torres-Rosell et al 2007). The average and SEM of at least three independent transformants was plotted.

Analysis of recombination frequencies

Recombination frequencies were determined as described (Gómez-González et al 2011b). Chromosomal leu2-k::ADE2-URA3::leu2-k system and plasmids pRS314L, pRS314LY, and pSch204, pRS314GL-lacZ were used to determine recombination frequencies. Briefly, transformants with indicated plasmids were grown for 3–4 days (until similar colony size) in the appropriate selective media SC with 2% glucose (for plasmids with tet::RECQL5 or tet::RECQL5-ID) or 2% galactose supplemented with 0.05% glucose (for plasmids with GAL::RECQL5 or GAL::RECQL5-ID) when applicable. In the case of pSch204, pRS314GL-lacZ plasmid-systems yeast cells expressing tet::RECQL5 or tet:RECQL5-ID were grown in selective SC with 2% glucose or in selective SC with 2% galactose to determine recombination frequencies in low or high transcription conditions, respectively. Recombinants were obtained by plating appropriate dilutions and selected as Leu+colonies for the plasmids containing LEU2 truncated repeat systems. Recombination analyses for the chromosomal leu2-k::ADE2-URA3::leu2-k system (Lk-AU) were performed using 6 to 12 independent colonies grown in synthetic complete medium SC with appropriate amino acid selection, and recombinants were selected in SC+FOA.

For each strain, the recombination frequencies are given as the average and standard deviation of the median frequency value obtained from fluctuation tests performed in 3–4 different transformants using 6 independent colonies per transformant. These transformants come from one or two independent transformations; in any case, they are different and the analysis of 6 colonies of each one provides enough sample to perform the test according to the standard methodology (Gómez-González et al. 2011b).

Co-immunoprecipitation assays

Strains carrying plasmids with GAL::RECQL5, GAL::RECQL5-ID, or the empty vector (pYES) were grown in selective SC medium with 2% raffinose to mid-log phase and then 2% galactose was added for 4 h to induce protein expression. For protein complex precipitation, 50 ml of exponentially growing cells were harvested and washed. 100 µL of glass beads and 166 µL of Lysis Buffer (50 mM Tris HCl pH 7.5, 100 mM NaCl, 1.5 mM MgCl₂, 0.0075% NP40 10%, 100 mM DTT) were added and samples were vortexed in the FastPrep (intensity 5 m/s) 4×20″. After separating the beads, samples were centrifuged 10′ at 13,000 rpm. Protein amount in supernatant (crude extract) was quantified using Bradford. Extracts were homogenised to 5 mg/ml, 10 µl were separated for INPUT and the remainder was incubated with 10 µl of a FLAG antibody attached to Dynabeads Protein A (Invitrogen) for 3 h at 4 °C. Magnetic beads were centrifuged, washed 3×with ice-cold PBS 1×and resuspended in 40 µl Laemmly Buffer 1x. Samples were boiled 5′ before loading the gel.

Chromatin immunoprecipitation experiments

Recruitment of RECQL5-FLAG, RECQL5-ID-FLAG and RNAPII to chromatin (GAL1 locus) was analyzed in exponentially growing cells carrying plasmids for ectopic expression of helicase Flag-tagged proteins under the control of tet promoter (tet::RECQL5 or tet::RECQL5-ID) or the empty vector in the appropriate synthetic medium. Yeasts cultures grown were grown in selective SC medium with 2% raffinose to mid-log phase and then 2% glucose or 2% of galactose was added for 4 h to analyze protein recruitment at GAL1 gene under repressed or active conditions, respectively. ChIP analyses were performed as previously described (Gómez-González et al 2011a) using the monoclonal antibodies anti-FLAG-M2 and anti-Rpb1-CTD antibody 8WG16 (Berkeley Antibody Company, Richmond, CA, USA) and Dynabeads protein A (Invitrogen). For qPCR, the intergenic region at positions 9716–9863 of chromosome V was used as a negative control. Average and SEM of two independent experiments are shown.

RNA extraction and northern blot

Strains carrying plasmids with GAL::RECQL5, GAL::RECQL5-ID, or the empty vector (pYES) were grown selective in SC medium with 2% raffinose to mid-log phase and then 2% galactose was added to induce transcription from the GAL1 promoter. Samples were taken at different times and RNA extraction was performed by acid phenol following standard procedures. Northern blot of kinetics was carried out as previously described (Chávez and Aguilera 1997). RNA was separated by agarose electrophoresis in a gel containing MOPS 1×and 2% formaldehyde and transferred to Hybond-N nitrocellulose membranes (GE Healthcare). GAL1 and GAL10 probes were amplified by PCR using primers detailed in Supplementary Table 1. Probes were ³²P-labelled using Klenow polymerase (Roche) and ³²P-dCTP. Signal was acquired using a FLA-5100 Imager Fluorescence Analyzer (Fujifilm) and quantified with MultiGauge 2.0 analysis software (Science Lab). Signal was measured as arbitrarily units (a.u.) and normalized to control condition.

Microarray gene expression analysis

Strains carrying plasmids with tet::RECQL5, or the empty vector were grown in selective SC medium with 2% glucose and without doxycycline (tet ON). Total RNA was isolated from 50 ml of mid-log growing cultures using the RNeasy Midi kit (Qiagen). Global expression analyses were performed with Affymetrix platform GeneChip Yeast Genome 2.0 Array. The relative RNA levels for all yeast genes were determined using an Affymetrix microarray scanner. For the RECQL5-expressing cells (tet::RECQL5), microarray analysis was conducted in triplicate and the values presented represent the average of these three determinations, control not expressing RECQL5 was performed only once and standardized with published data of experiments performed in the same conditions GSE22644 (Tu et al 2011) (Pearson’s R correlation coefficient of 0.914). A total of 7.5 µg of RNA were used for cDNA synthesis, labelling and microarray hybridization, performed by the CABIMER’s Genomics Unit according to the manufacturer’s instructions. Microarray data were normalized by RMA (robust microarray average) and statistically analyzed by LIMMA (Linear Models for Microarray Analysis) comparing the RECQL5-expressing cells profile with its isogenic wild-type strain. Genes with expression levels below 600 (the median expression value of non-expressed meiotic genes in these experiments) were removed from the analysis to reduce false positive. Genes showing at least a 1.5-fold expression change with a FDR<0.1 were considered as altered. GO analyses were performed using the GO Term Finder tool from the Saccharomyces Genome Database (http://www.yeastgenome.org). A p-value<0.01 was established to consider GO terms as significantly enriched. The data generated are available under Gene Expression Omnibus accession number (GSE180112).

ChIP-chip experiments

ChIP-chip analysis were carried out in cells expressing RECQL5-FLAG tagged protein under the tet promoter (tet::RECQL5). Yeast cultures were grown to mid-log phase in selective SC medium containing 2% glucose and without doxycycline. S. cerevisiae oligonucleotide tiling microarrays were provided by Affymetrix (GeneChip S. cerevisiae Tiling 1.0R array). The high-density oligonucleotide arrays used allows the analysis of yeast chromosomes at a 300-bp resolution, each of the 300-bp region being covered by at least 60 probes. ChIP-chip of asynchronously growing cells was carried out as described (Bermejo et al 2009; Gómez-González et al. 2011a). For immunoprecipitation with RECQL5-FLAG, cells growing in SC-T were harvested at mid-log phase. 1.5×10⁷ cells were disrupted by multi-beads shocker (MB400U, Yasui Kikai, Japan), which maintained cells precisely at lower than 4 °C during disruption. Anti-FLAG antibody M2 (Sigma-Aldrich) was used for ChIP. ChIP DNA was purified and amplified by random priming using a WGA2 kit (Sigma- Aldrich) and following the manufacturer’s procedure. A total of 4 μg of amplified DNA was digested with DNaseI to a mean size of 100 bp and the purified DNA fragments were end-labelled with biotin-N6-ddATP23. The ChIP-chip data can be accessed at Gene Expression Omnibus (GSE180107). Annotation of RECQL5-enriched regions were done using in-house scripts with the latest version of the SGD_features.tab (based on Genome Version R62-2-1) hosted by SGD (SGD Project. http://www.yeastgenome.org/download-data/ [April 2021]).

RECQL5 expression leads to transcription-associated genome instability in yeast

The yeast sensitivity to genotoxic agents caused by RECQL5 expression suggests an interference with the DNA repair or replication processes that could potentially affect genome stability. This prompted us to assay whether RECQL5 expression leads to genome instability associated with transcription. We first analyzed whether RECQL5 expression generated a hyper-recombination phenotype in wild-type yeast. For this, we used different leu2 direct-repeat constructs to measure single strand annealing (SSA), a highly efficient mechanism of DSB repair independent of Rad51, which serves to infer whether DNA breaks are stimulated. We assayed SSA recombination in the chromosomal Lk-AU system which is based on two heteroallelic 2.16-kb leu2 repeats flanking the URA3 and ADE2 genes (Aguilera and Klein 1989). Recombination between the two leu2 repeats leading to URA3 deletions was scored as the frequency of Ura⁻ cells, detected as FOA-resistant colonies. A clear increase of the recombination frequencies was observed upon expression of both wild-type RECQL5 and RECQL5-ID respect to the control cells, the values being slightly lower upon expression of the mutant allele. The levels of hyper-recombination were dependent on the RECQL5 expression levels (Fig. 2A), proportional to the strength of the promoter used, whether GAL1 or tet promoters.

Open in a separate window

Fig. 2

Genome instability in RECQL5-expressing cells. A Recombination analysis in the WT strain (AYW3-1B) with the chromosomal Lk-AU system and transformed with the indicated plasmids. Cells were grown for 3–4 days in selective media SC with 2% glucose (for tet::RECQL5 and tet::RECQL5-ID) or 2% galactose supplemented with 0.05% glucose (for GAL::RECQL5 and GAL::RECQL5-ID). Control refers to the respective empty plasmid. B Recombination analysis of WT cells (W303-1AR5) carrying L and LY plasmid system and expressing GAL::RECQL5 or GAL:RECQL5-ID constructs. Other details as in panel A. C Recombination analysis of WT cells (W303-1AR5) expressing tet::RECQL5 or tet::RECQL5-ID constructs in the GL-lacZ system under the control of the inducible GAL1 promoter in glucose (low transcription-GAL1p OFF), in galactose (High transcription-GAL1p-ON) or the L-lacZ system expressed under the control of the constitutive LEU2 promoter in glucose (Medium transcription-LEU2p-ON). Average and SEM of at least three different fluctuation tests from six independent colonies each one are shown. *, p≤0.05; **, p≤0.01; ***, p≤0.005; (Student’s t-test)

For a better association of hyper-recombination with transcription levels we next analyzed SSA recombination in the plasmid-born systems L and LY. These systems are based on the same direct-repeats (600-bp internal fragments of the LEU2 gene sharing 300 bp of homology) that are transcribed from the LEU2 promoter, but they differ in the length of the transcribed intervening sequence (31 bp for L, and 5.57 kb for LY) (Prado and Aguilera 1995) (Fig. 2B). Expression of both RECQL5 and RECQL5-ID increased SSA recombination in the LY system (5.6-4 fold), whereas in the L system the increase was weak (twofold for RECQL5) or non-existing (RECQL5-ID). This result indicates that increase of SSA events is significant in the systems, like LY, generating a long transcript. Then, we determined the effect of RECQL5 and RECQL5-ID expression in the L-lacZ and GL-lacZ plasmid recombination systems that contain the GC-rich sequence lacZ between 0.6-kb leu2 direct-repeats transcribed from different promoters, LEU2 promoter (L-lacZ system) or GAL1 promoter (GL-lacZ system) (Fig. 2C). We have previously used these systems to report the transcription-associated recombination (TAR) phenotype of different mRNP biogenesis mutants (Chávez and Aguilera 1997; González-Aguilera et al 2008; Piruat and Aguilera 1998). Similar increases were obtained in the L-lacZ and GL-lacZ systems that also generated a long transcript (>4 kb). Importantly, hyper-recombination was significantly higher in the GL-lacZ system, compared to the L-lacZ, consistent with the higher transcription (Fig. 2C). The results therefore, clearly show that RECQL5 expression increases recombination in a transcription-dependent manner, a phenotype also observed for the mutant RECQL5-ID, although to a lesser extent. The high SSA levels suggest that RECQL5 expression leads to the accumulation of recombinogenic DSBs. To assess this more directly we tested Rad52-YFP foci accumulation in yeast expressing RECQL5 and RECQL5-ID. A fourfold increase in Rad52-YFP foci was detected in both cases (Fig. 3A), consistent with an increase in recombinogenic DSBs. However, Rad52 foci could also accumulate as a consequence of a lower DSB repair efficiency. To test this possibility, we analyzed accumulation Rad51 foci, provided that Rad51 acts in the HR reaction after Rad52 by forming nucleoprotein filaments. Interestingly Rad51 foci were clearly reduced compared with the wild-type control (Supplementary Fig. 2), suggesting that instability generated by RECQL5 expression in yeast may be produced by both the occurrence of DNA breaks and interference with the DNA repair process.

Open in a separate window

Fig. 3

Expression of human helicase RECQL5 in S. cerevisiae leads to R-loop dependent genome instability. A Spontaneous Rad52-YFP foci formation in WT (W303-1AR5) expressing GAL::RECQL5 or GAL::RECQL5-ID. B Effect of RNase H1 (RNH1) overexpression on Rad52 foci formation in cells carrying plasmids with GAL::RECQL5 or tet::RECQL5 constructs and transformed with either the plasmid pCM184RNH1 (tet::RNH1) or the empty vector pCM184. Control refers to the respective empty plasmid without RECL5. Average of at least three independent experiments are shown (*, p<0.05; **, p<0.0005). Yeast cultures were grown in selective SC medium (tet::RECQL5) or switched from S-Raffinose to SC with 2% galactose until exponential (4–6 h) for GAL::RECQL5

Taking into account the roles of RECQL5 in transcription and replication processes and the impact of R-loops as a source of transcription-replication conflicts and genome instability, we asked whether the increase in DNA breaks associated to RECQL5 expression was dependent on R-loops. We assayed whether Rad52 foci accumulation was suppressed by RNase H1 (RNH1), an enzyme that degrades the RNA strand of a DNA-RNA hybrid (Cerritelli and Crouch 2009). Since there were no consistent differences in the phenotype when expressing the wild-type of the ID mutant version of RECQL5, we only assayed the effect of the wild-type RECQL5 expression from the GAL1 or tet promoters (Fig. 3B). A reduction of Rad52 foci formation by RNH1 expression was observed in GAL::RECQL5 and tet::RECQL5 expressing cells suggesting that RECQL5 expression leads to R-loop-dependent genome instability.

Human RECQL5 interacts in vivo with yeast RNAPII impairing transcription

Since RECQL5 is the unique RecQ helicase that interacts with RNAPII and functions in transcription elongation in human cells (Saponaro et al. 2014), we next asked whether this capability to interact with the transcription machinery and chromatin was conserved in yeast. For this, we did co-immunoprecipitation (Co-IP) assays in cells transformed with plasmids expressing RECQL5 or RECQL5-ID fused to the FLAG epitope, using an anti-FLAG antibody and an antibody against the Rpb1 subunit of RNAPII. As can be seen (Fig. 4A), human RECQL5 immunoprecipitated with yeast Rpb1, implying a physical interaction with the RNAPII transcription machinery. The interaction could be seen regardless of whether we pulled down proteins with RECQL5 or Rpb1 (Fig. 4A). To confirm that this occurred in active chromatin, we performed ChIP-qPCR in the inducible GAL1 gene. The results clearly showed that when RECQL5 was expressed from the tet::RECQL5 construct it was recruited to the GAL1 gene when transcription was active (galactose) but not when it was repressed (glucose) (Fig. 4B upper panel). Unexpectedly, however, both the wt RECQL5 and mutant RECQL5-ID proteins interacted similarly with yeast chromatin in a transcription-dependent manner. Finally, ChIP analysis indicate that the expression of RECQL5 did not significantly affect the recruitment of RNAPII to the transcribing gene (Fig. 4B lower panel).

Open in a separate window

Fig. 4

Interaction of the human helicase RECQL5 with active chromatin and its impact on transcription in S. cerevisiae. A RNAPII protein interaction with either RECQL5 or RECQL5-ID FLAG-tagged versions detected by co-immunoprecipitation with anti-Rpb1 (8WG16) or anti anti-FLAG antibodies from whole cell extracts of the yeast carrying the empty vector (control) or the indicated GAL::RECQL5 constructs. Yeasts carrying plasmids with GAL::RECQL5, GAL:RECQL5-ID, or the control (empty pYES) were grown in selective SC medium with 2% raffinose to mid-log phase and then 2% galactose was added for 4 h to induce protein expression. B ChIP analysis of RECQL5 FLAG-tagged proteins (upper panel) or RNAPII (lower panel) within the GAL1 gene in yeasts transformed with the empty plasmid (control) or the indicated tet::RECQL5 constructs. Assays were performed from yeast cells grown under conditions of active (GAL) or repressed (GLU) transcription for GAL1 locus. Values represent the relative ratio to empty vector of precipitated DNA (IP) to input DNA (INPUT) for the FLAG antibody, or the ratio of IP to INPUT for the RNAPII antibody. Mean±SEM of two independent experiments are shown. Schematic diagram of gene is depicted. The black line indicates the region where ChIP-qPCR analyses were performed. C Northern blot analysis of GAL10 mRNAs in WT strain transformed with GAL::RECQL5 or the empty vector (Control). Graphics represent the corresponding mRNA signal relative to rDNA. Average and SEM of n=3. Other details as in (A)

To further explore how human RECQL5 could functionally affect yeast RNAPII in S. cerevisiae we analyzed the possible genetic interactions between RECQL5 expression and mutations of different RNAPII subunits, rpb1-1, rpb1S751F, rpb2-10 and rpb9, all of which are defective in transcription elongation (Felipe-Abrio et al 2015; Powell and Reines 1996; Strathern et al 2013; Van Mullem et al 2002). We assayed growth of these mutants when they expressed RECQL5. No clear differences in growth were observed in any of the mutants tested when expressing RECQL5 at low (2% Gal+0.05% Glu) or high levels (2% Gal), although a slight effect in the growth of rpb1S751F mutant upon RECQL5 repression conditions (2%Glu) was apparent (Supplementary Fig. 3). Interestingly, this mutation was reported to cause RNAPII slippage as also described for mutants of the elongation factor TFIIS (Strathern et al. 2013). Considering the described competition of RECQL5 with the binding of TFIIS to RNAPII (Kassube et al. 2013) data suggest that expression of human RECQL5 could interfere with the interaction of RNAPII and TFIIS in yeast.

For a more accurate analysis of transcription, we determined the kinetics of activation of the GAL1, 10 locus in wild-type cells expressing GAL::RECQL5 (Fig. 4C). Northern blots of the GAL10 and GAL1 genes showed no significant defects of transcription activation upon galactose addition, but interestingly, a larger mRNA resulting from readthrough at termination could be detected (Fig. 4C) (Supplementary Fig. 4). A similar effect was also observed in cells expressing RECQL5-ID (Supplementary Fig. 4). Altogether, these data suggest a transcriptional termination defect upon RECQL5 expression.

We would like to thank CABIMER’s Genomic Unit for technical assistance with microarray analysis.