Poly (ADP-ribose) polymerase in yeasts: characterization and involvement in telomere maintenance

Abstract

Poly (ADP-ribose) polymerases (PARPs) are enzymes catalyzing the post-translational addition of chains of ADP-ribose moieties to proteins. In most eukaryotic cells, their primary protein targets are involved in DNA recombination, repair, and chromosome maintenance. Even though this group of enzymes is quite common in both eukaryotes and prokaryotes, no PARP homologs have been described so far in ascomycetous yeasts, leaving their potential roles in this group of organisms unexplored. Here, we characterize Pyl1 protein of Yarrowia lipolytica as the first candidate of PARP in yeasts. We show that the expression of PYL1 gene is increased in mutants lacking either subunit of telomerase and identified several of its candidate protein targets in vivo. We demonstrate that Pyl1p is a functional PARP that undergoes auto-PARylation and PARylates YlKu70/80 complex. We also show that overexpression of PYL1 in Y. lipolytica cells results in dissociation of YlKu80 from telomeres in vivo, supporting the role of Pyl1p in telomere protection and maintenance. Based on our observations, we propose Pyl1p and its homologs identified in other yeast species represent a distinct class of PARPs, thus substantiating a more detailed investigation of their roles in these organisms.

Graphical Abstract

Graphical Abstract.

Introduction

Poly (ADP-ribose) polymerases (PARPs) are enzymes responsible for the addition of linear and/or branched chains of ADP-ribose (PAR) to target proteins. These enzymes belong to a family of poly-ADP-ribosyltransferases, referred alternatively as poly-ARTs [1]. The reaction consumes nicotinamide adenine dinucleotide (NAD⁺) and releases nicotinamide (NAM) [2]. The members of the PARP family play a role in a wide range of cellular processes, such as DNA repair, genome maintenance, and cell death [3, 4]. The best-characterized enzyme of this family is human PARP1, which was first identified as a component of base excision repair machinery involved in the repair of single-strand DNA breaks [5, 6]. PARP1 has also been known as an attractive target for anti-cancer chemotherapy in tumors with a deficiency in homologous recombination (HR) [7, 8]. In humans, the PARP family consists of 17 proteins that differ in their enzymatic activity, substrate preferences, or mode of regulation [9]. According to the type of catalytic activity, they are classified as poly (ADP-ribose) polymerases/transferases (PARPs/poly-ARTs) or mono-ADP-ribosyltransferases (mono-ARTs) that can add only one ADP-ribosyl unit to the target protein [1]. Structurally, human PARP1 consists of three main domains: an N-terminal DNA-binding domain containing three Zn-fingers (Zn 1-3), a central BRCA1 C-terminus domain (BRCT), and a C-terminal catalytic domain. Other human PARPs lack Zn-finger motifs, but their DNA binding may be mediated by additional WGR domain [10]. Upon DNA damage, PARP1 undergoes massive auto-PARylation, and the branched poly (ADP-ribose) chain serves as a docking site for the recruitment of other proteins involved in the recognition and repair of DNA breaks [11]. Importantly, a variety of proteins involved in different types of DNA repair processes, such as XRCC1, DNA-PKCs, and Ku70/80 complex, are also directly PARylated in mammalian cells, resulting in their recruitment to the site of DNA damage or displacement and dissociation from DNA [12–14]. PARylation can be reversed by poly (ADP-ribosyl) glycohydrolase (PARG) which possesses both endo- and (preferentially) exo-glycohydrolase activity. Even though in human cells, the primary PAR hydrolase is PARG, de-PARylation can also be partially carried out by ADP-ribosylhydrolase 3 (ARH3) [15].

Several mammalian PARPs have been implicated in the maintenance and protection of telomeres, nucleoprotein complexes located at the ends of linear chromosomes [16–18]. The maintenance of telomeres in majority of eukaryotes is mediated by telomerase, whose catalytic subunit (TERT) elongates a 3′ single-stranded telomeric overhang, employing a template domain of its RNA subunit (TER) [19]. The protective function of chromosomal ends is mediated by both telomerase and the protein complex called shelterin [20, 21] or telosome [22]. The mammalian shelterin complex is composed of six proteins: the double-stranded DNA binding proteins TRF1 and TRF2, directly protecting the double-stranded region of telomeres composed of 5′-TTAGGG-3′ tandem repeats, their interaction partners TIN1 and RAP1, POT1 protein binding the 3′ single-stranded telomeric overhang of a G-rich strand, and TPP1 facilitating the connection between TIN2 and POT1 [20, 21].

In ascomycetous yeasts, the composition of telomeres exhibits a high degree of interspecies variability. For example, whereas telomeres in the fission yeast Schizosaccharomyces pombe are protected by a shelterin-like complex containing an ortholog of both TRF1/TRF2 (Taz1p) and POT1 (Pot1p) [23], the main telomere-binding proteins in Saccharomyces cerevisiae are Rap1 and Cdc13 [24]. One of the major driving forces behind the frequent replacement of telomere-associated proteins in this group of microorganisms seems to be a runaway evolution of telomeric repeats, resulting in a wide repertoire of sequence motifs constituting the terminal sequences of chromosomes in various species [25, 26]. One example that is particularly relevant for the present study is the oleaginous yeast Yarrowia lipolytica. Its telomeric arrays are composed of TTAGGG-like repeats (5′-TTagtcAGGG-3′), yet it lacks orthologs of TRF1/TRF2 (or Taz1p) and Pot1p. The only protein that was shown so far to be associated with the double-stranded region of telomeres in this yeast is Tay1p, which also exhibits a high affinity for TTAGGG-like sequences [27, 28]. The other two known players involved in telomere maintenance in Y. lipolytica are telomerase, composed of catalytic (Est2p) and RNA (TER) subunits [29, 30], and YlKu70/80 complex [29]. The absence of telomerase results in a rapid loss of telomeric repeats without an apparent senescence crisis [29] that was observed for telomerase mutants of S. cerevisiae or S. pombe [31, 32]. The nature of the mechanism(s) involved in an alternative, telomerase-independent maintenance of telomeres is currently under investigation. On the other hand, the absence of a functional YlKu70/80 complex yields over-elongated telomeric tracts of heterogeneous size. This is in contrast to the situation in S. cerevisiae, where the mutants lacking Ku70/80 exhibit shorter telomeres [33–37], and it is more similar to Ku-deficient mutants in other organisms such as Candida albicans, Arabidopsis thaliana, or mammalian cells [38–43]. These peculiarities make Y. lipolytica an attractive nonconventional model organism that can provide novel insights into the telomere maintenance mechanisms and their evolution. This claim is further substantiated by our observation that Y. lipolytica cells lacking telomerase exhibit an increased expression of the gene YALI0C17061g encoding a putative poly-ART [29]. Our interest in a detailed investigation of this protein was triggered by the fact that ascomycetous yeasts were thought to lack PARPs as enzymes involved in the recognition of DNA damage in general and of dysfunctional telomeres in particular (see also below).

In mammalian cells, PARylation is triggered by a plethora of signals such as DNA damage and oxidative or replication stress [44]. The first PARP enzyme that was shown to be directly associated with telomeres is Tankyrase 1 (TNK1 or PARP5a), which PARylates TRF1, thus mediating its dissociation from telomeres, which results in telomere elongation [45]. The sequences of TNK1 and a more recently identified tankyrase TNK2 (PARP5b) differ from other PARPs by the presence of ankyrin repeats [46]. In addition to tankyrases, PARP1 also participates in telomere maintenance by regulating the access of the components of HR and nonhomologous end-joining (NHEJ) pathways to telomeric DNA [47]. Retaining the balance between PARylation catalyzed by PARP1 and hydrolysis mediated by PARG is critical for the alternative telomerase-independent mechanism of telomere maintenance (ALT). Inhibition of PARP1 stimulates telomere recombination, however, subsequent inhibition of PARG and retention of PARylated proteins perturbs HR, which is essential for ALT [48]. Moreover, PARylation of TRF1 and TRF2 by PARP1 plays a role in proper telomere replication. Displacement of PARylated TRF1 from telomeric DNA duplex enables the formation of complexes recruiting WRN and BLM helicases that unwind secondary structures, such as G-quadruplexes, formed on the lagging strand of the proceeding replication fork [49]. It was also demonstrated that the inhibition of PARP1 reduces PARylation of TERT in human HeLa cells, leading to telomere shortening [50], indicating the role of PARP1 in the regulation of telomerase activity in vivo.

Various PARP homologs have been identified in metazoans, plants, protists, and filamentous fungi, suggesting the widespread evolutionary conservation of PARylation as a protein regulatory mechanism [51]. Two PARPs were previously characterized in filamentous fungi classified into the subphylum Pezizomycotina: PrpA in Aspergillus nidulans and Npo in Neurospora crassa. PrpA is an essential protein with a role in early DNA damage response, cell death, and asexual development, however, its involvement in telomere maintenance has not been studied [52]. Unlike PrpA, Npo is non-essential protein, and npo deletion mutants are not sensitive to DNA-damaging agents. On the other hand, the levels of Npo transcripts accumulate 30 min after treatment of the wild-type cells with methyl methanesulfonate (MMS), suggesting its involvement in DNA damage response. Interestingly, even though the strain lacking the functional Npo gene exhibits accelerated replicative aging, the length of its telomeres is comparable to wild type cells [53].

Until now, there has been no report on functional analysis of a PARP-like protein in ascomycetous yeasts. This is due to the fact that genomes of most of the species, including conventional models Saccharomyces cerevisiae (subphylum Saccharomycotina) and Schizosaccharomyces pombe (subphylum Taphrinomycotina), lack genes encoding a putative PARP. In this context, it was surprising that the transcriptomic analysis of telomerase-deficient strain of yeast Yarrowia lipolytica (subphylum Saccharomycotina, order Dipodascales) revealed overexpression of YALI0C17061g gene encoding a protein with predicted catalytic PARP domain [29].

In this study, we demonstrate that the YALI0C17061g protein product, which we named Pyl1 (PARP in Yarrowia lipolytica  1), is a functional PARP mediating PARylation of several protein targets in vivo, including both subunits of the YlKu70/80 complex, which are key components of NHEJ pathway. Using a recombinant version of the Pyl1p, we demonstrate that it is a functional PARP that, in addition to its auto-PARylation, PARylates purified Ku70/80 complex in vitro. Even though the PYL1 gene is not essential for cell survival, the double deletion mutants ΔterΔpyl1 exhibit a decreased growth rate, suggesting its role in the adaptation to the loss of telomerase. Importantly, the association of YlKu80 with telomeres in wild type cells is strongly reduced in cells overexpressing PYL1. This indicates that Pyl1-mediated PARylation of YlKu70/80 proteins may lead to the release of the YlKu70/80 complex from telomeric DNA, allowing the access of proteins involved in the maintenance of chromosomal ends. Finally, we identified PYL1 homologs in other ascomycetous yeasts, indicating that the occurrence of PARPs in this phylogenetic group is more frequent than previously thought.

Materials and methods

Y. lipolytica strains and cultivation

Y. lipolytica wild type (WT) strain H222-S4 (MATA, ura3-302 SUC2) and H222-SW6 strain (MATA, ura3-302 SUC2 Δku80) lacking YlKU80 gene (Δku80) [54] were kindly provided by Gerold Barth (Technische Universität Dresden, Dresden, Germany). The strains lacking TER locus (Δter, ΔterΔku80) were derived from H222-S4 and H222-SW6 strains, respectively [29]. The Δest2 H222-S4 -derived strain lacking a catalytic subunit of telomerase was constructed as described by Kinsky et al., (2010) [30].

The deletion cassette for PYL1 gene disruption was constructed using pUB4 plasmid, where 5′ (857 bp) and 3′ (604 bp) flanking regions of PYL1 gene were cloned upstream and downstream of HygB^R gene, respectively. The cassette was amplified by PCR (oligonucleotides are listed in Supplementary Table S1) and transformed into the WT, Δter, and Δku80 strains. Transformants were selected after 3-day cultivation at 29°C on a complex medium containing Hygromycin B (YPD + Hyg) (Supplementary Table S2). The correct integration of the deletion cassette into the genomic DNA was verified by PCR (Supplementary Table S1).

Construction of plasmids for the heterologous expression of PYL1 and YlKU70/80 genes in yeast and bacteria

For the construction of pYES_PYL1 plasmid allowing expression of PYL1 gene in S. cerevisiae, the PYL1 ORF (without STOP codon) was amplified from genomic DNA of Y. lipolytica strain H222-S4 (Supplementary Table S1) and inserted into pYES-2CT vector in frame with V5-tag using GeneArt™ Gibson Assembly HiFi Cloning Kit (Thermo Fisher Scientific). The deletion variant pYES_PYL1_Δcat was created by inverse PCR (Supplementary Table S1) designed to leave out the region encoding the catalytic domain of Pyl1. Both constructs were transformed into S. cerevisiae strain W303-1A (MATa, ade2-1, leu2-3, leu112, his3-11, his15, trp1, ura3-1) and transformants were selected after 3-day cultivation at 29°C on synthetic medium lacking uracil (SD-URA) (Supplementary Table S2).

For the overexpression of the PYL1 gene in Y. lipolytica (plasmid pUB4 + PYL1), the EYK1 gene promoter (P_EYK1) and PYL1 terminator were amplified from the genomic DNA of Y. lipolytica H222-S4 by PCR using primers listed in Supplementary Table S1. The PYL1 terminator was fused with the V5 tag at the 5′ end by adding the V5 tag sequence to forward primer (Supplementary Table S1). PYL1 ORF without a STOP codon was amplified from pYES-2CT_PYL1 plasmid. All PCR fragments were integrated into pUB4 plasmid by GeneArt™ Gibson Assembly HiFi Cloning Kit (Thermo Fisher Scientific). The version of the plasmid encoding catalytically inactive Pyl1p was constructed by inverse PCR using the primers del_cat_pYESYlPARP-F and del_cat_pYESYlPARP-R (Supplementary Table S1) and pUB4 + PYL1 as a template. Transformants were selected on a complex medium containing hygromycin B (YPD + Hyg) (Supplementary Table S2).

The plasmid carrying Myc-tagged version of YlKU80 was constructed by cloning of YlKU80 gene without STOP codon amplified by PCR from genomic DNA Y. lipolytica H222-S4 using primers KuTag_full2_fw and KuTag_full_rev (Supplementary Table S1) into the pUB4-based plasmid carrying 13xMyc tag coding sequence, and NAT^R gene enabling resistance of transformants to nourseothricine (CloNAT). Vector was linearized by inverse PCR (primers invPCR_vec_rev2 and KU80_Myc_fw2; Supplementary Table S1) and combined with YlKU80 PCR product using GeneArt™ Gibson Assembly HiFi Cloning Kit (Thermo Fisher Scientific).

Plasmids carrying P_EYK1-PYL1-V5 (named pUB4_PYL1) and YlKU80-Myc were transformed into Y. lipolytica cells, and transformants were selected on a complex medium supplemented with hygromycin B and CloNAT (YPD + Hyg + CloNAT) (Supplementary Table S2).

For the expression of PYL1 in Escherichia coli, PYL1 open reading frame lacking ATG codon was PCR amplified from Y. lipolytica genomic DNA with primers pGEX-PARP-Gibson-F and pGEX-PARP-Gibson-R (Supplementary Table S1) and cloned into SmaI-digested pGEX-6P-2 vector (Pharmacia) using Gibson Assembly Cloning kit (New England Biolabs). The resulting vector pGEX-PYL1 encodes Pyl1p fused at N-terminus with glutathione-S-transferase that can be cleaved off by PreScission protease.

For the production of YlKu70/80 complex in E. coli, the coding sequence of YlKU80 lacking STOP codon was amplified with YlKu80-Duet-F and YlKu80-Duet-R primers (Supplementary Table S1) from Y. lipolytica cDNA, and after its treatment with NdeI/BglII it was cloned into the pETDuet-1 plasmid (Novagen) linearized with NdeI and BglII. The resulting vector was linearized with BamHI and NotI and ligated with a BamHI/NotI-treated PCR fragment carrying the coding sequence of YlKU70 lacking ATG codon that was amplified from cDNA using the primers YlKu70-Duet-F and YlKu70-Duet-R (Supplementary Table S1). Subsequently, the plasmid carrying both YlKU70 and YlKU80 was used as a template for an inverse PCR with the primers YlKu-His-Pres-F and YlKu-His-Pres-R to introduce a recognition site for PreScission protease between the His₆-tag and YlKu70 coding sequence. The final plasmid pDuet-Ku70/80 encodes YlKu70 tagged at its N-terminus with His₆-tag (cleavable with PreScission protease) and C-terminally S-tagged YlKu80.

RNA isolation and quantitative PCR

Three independent cultures of Y. lipolytica H222-S4 WT, Δter, and Δest2 yeast strains were cultivated in 25 ml of liquid YPD medium until the exponential phase (OD₆₀₀= 0.6 – 0.8). Afterwards, total RNA was isolated using Direct-zol^TM RNA miniprep kit (Zymo Research). Around 2 μg of total RNA was transcribed into cDNA by Maxima First Strand cDNA Synthesis Kit (Thermo Fisher Scientific). RT-qPCR was performed in QuantStudio™ 3 Real-Time PCR System using PowerTrack™ SYBR Green Master Mix for qPCR (Thermo Fisher Scientific) according to the instruction manual. The thermocycling parameters were set as follows: 2 min at 95°C (initial denaturation) followed by 40 cycles of denaturation (2 min at 95°C) and annealing/polymerization (15 s at 60°C). ACT1 gene (Supplementary Table S1) was used for normalization, and the calculation of the relative expression of PYL1 gene was carried out using the relative standard curve method. The results were validated by two-tailed Student's t-test.

Macrodomain affinity pull-down

Y. lipolytica and S. cerevisiae cells were cultivated overnight at 29°C in 10 ml of YPD or SD medium lacking uracil (SD-URA), respectively. To test the effect of PARP inhibitor 6(5H)-phenanthridinone (PHE), yeasts were cultivated overnight in YPD containing 100 μM PHE. The MMS treatment was performed for 3 h in YPD medium with the addition of 0.05% (v/v) MMS (Supplementary Table S2). Cultures were centrifuged and cells were suspended in 300 μl of protein lysis buffer [50 mM Tris-HCl, pH 8.0, 200 mM NaCl, 1 mM EDTA, 1% (v/v) Triton-X100, 10% (v/v) glycerol, 1 mM DTT, 0.5% (w/v) deoxycholate, and 1 × cOmplete^TM Protease Inhibitor Cocktail (Roche)]. Afterward, 0.3 g of Lysing Matrix C beads (MP Biomedicals) were added, and cells were homogenized six times at 6.5 m/s for 20 s with 20 s breaks at ice using Fast Prep 24 (MP Biomedicals). Cell lysates were centrifuged at 2000 × g, 4°C for 8 min., and cell extracts were transferred into clean tubes and centrifuged again at 8000 × g, 4°C for 10 min. Protein concentration was measured by Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific). Magnetic macrodomain affinity resin (Af1521; Tulip Biolabs) was used to enrich the protein extracts for PARylated proteins. For each sample, 25 μl of magnetic beads were washed with 1 ml of protein lysis buffer, then 1 mg of yeast protein extract was added, and the samples were incubated overnight at 4°C by continuous stirring. The magnetic resin was washed according to the instruction manual, and proteins were eluted by adding 75 μl of 1 × SDS PAGE loading dye (62.5 mM Tris-HCl, pH 6.8, 2% (w/v) SDS, 10% (v/v) glycerol, bromphenol blue) and heated at 95°C for 1 min. As a negative control, 1 mg of protein extract of WT cells was added to the negative control magnetic affinity resin with macrodomain carrying mutation abolishing PAR binding.

Whole-cell protein extracts

For the verification of expression of PYL1 and YlKU80 in Y. lipolytica transformants carrying plasmid pUB4 + PYL1 or pUB4_Ku80_Myc, the whole-cell protein extracts were prepared from 5 ml of overnight cell culture incubated at 29°C in YNBD (5% (w/v) glucose) or YNBOL (5% (w/v) erythritol) medium with the addition of uracil and hygromycin B (for pUB4 + PYL1) or CloNAT (for pUB4NAT_KU80_Myc) (Supplementary Table S2). Cells were pelleted, suspended in 1 ml of 2 M lithium acetate, and incubated on ice for 5 min. The culture was sedimented and pellet was washed with 0.2 M NaOH. After 5-min incubation on ice, the culture was centrifuged under the same conditions and the pellet was mixed with 60 μl of 2 × SDS PAGE loading dye. Before loading into the gel, extracts were incubated at 95°C for 5 min.

Immunoblot analysis

Staining of polyacrylamide gels with Coomassie Brilliant Blue R-250 and immunodetection were performed as described in Nosek and Tomáška [56]. The protein samples (25 μl) were separated in 8% SDS-PAGE [55]. Proteins were transferred to nitrocellulose membrane (Amersham) in Trans-Blot Turbo Transfer System (Biorad). After the transfer, the membrane was first stained for 5 min with 5% (w/v) Ponceau S in 0.1% (v/v) acetic acid, its image was captured and then it was destained five times for 5 min with 1 × TTBS (25 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.05% (v/v) Tween-20). The membrane was then soaked in blocking solution (5% (w/v) BSA, 1 × TTBS) for 1.5 h and then incubated with anti-PAR/pADPr antibodies (RD systems, 4335-MC-100) diluted 1:1000 in blocking solution. After overnight incubation at 4°C, the membrane was washed 3 times with 1 × TTBS and once with 1 × TBS (1 × TTBS without Tween-20) for 5 min. followed by incubation with HRP-conjugated secondary anti-mouse IgG antibodies (Sigma-Aldrich, A9044) diluted 1:10000 in blocking solution. After the final washing (3 times with 1 × TTBS and once with 1 × TBS for 5 min), the ECL substrate was added and the signal was detected by Alliance Q9 Imager (UviTec). For in vitro PARylation experiments, secondary anti-mouse IgG antibodies (Sigma-Aldrich; A3562; 1:10000) conjugated with alkaline phosphatase were used. The P_EYK1-driven overexpression of V5-tagged PYL1 in Y. lipolytica transformants was verified by anti-V5 antibodies (Life Technologies, R96025) diluted 1:5000 and secondary anti-mouse IgG antibodies (Sigma-Aldrich, A3562) conjugated with alkaline phosphatase and diluted 1:10000. The expression of Myc-tagged YlKU80 was assessed analogously except that anti-Myc antibody (clone 9E10, Thermo Fisher Scientific) was diluted 1:500. The signal was detected by applying BCIP and NBP substrate [56]. As a control of protein loading, the membrane was incubated with anti-tubulin antibodies (Sigma-Aldrich, T9026) diluted 1:500 (the washing steps were performed as described for anti-PAR antibodies).

Mass spectrometry analysis of PARylated proteins

In-gel digestion. Selected 1D gel bands were excised manually and after destaining and washing procedures each band was subjected to protein reduction (10 mM DTT in 25 mM NH₄HCO₃, 45 min, 56°C, 750 rpm) and alkylation (55 mM IAA in 25 mM NH₄HCO₃; 30 min, laboratory temperature, 750 rpm). After further washing by 50% ACN/NH₄HCO₃ and pure ACN, the gel pieces were incubated with 125 ng of trypsin (sequencing grade; Promega) in 50 mM NH₄HCO₃. The digestion was performed for 2 h at 37°C on a Thermomixer (750 rpm; Eppendorf). Tryptic peptides were extracted into LC-MS vials by 2.5% formic acid (FA) in 50% ACN with the addition of polyethylene glycol (final concentration 0.001% (w/v)) [57] and concentrated in a SpeedVac concentrator (Thermo Fisher Scientific).

LC-MS/MS analysis. LC-MS/MS analyses were performed using Ultimate 3000 RSLCnano system connected to Orbitrap Exploris 480 spectrometer (Thermo Fisher Scientific) with EASY Spray ion source (Thermo Fisher Scientific). Prior to LC separation, tryptic digests were online concentrated and desalted using trapping column (300 μm × 5 mm, μPrecolumn, 5 μm particles, heated to 40°C, Acclaim PepMap100 C18, Thermo Fisher Scientific). After washing of trapping column with 0.1% FA, the peptides were eluted (experiment 1: flow 300 nl/min; experiment 2: flow 250 nl/min) from the trapping column onto Acclaim PepMap100 C18 column (experiment 1: 2 μm particles, 75 μm × 250 mm, heated to 40°C; experiment 2: 2 μm particles, 75 μm × 500 mm, heated to 50°C; Thermo Fisher Scientific) by 60 min long gradient (mobile phase A: 0.1% FA in water; mobile phase B: 0.1% FA in 80% acetonitrile).

MS data were acquired in a data-dependent strategy (cycle time 2 s). Survey scan range was set to m/z 350-2000 with the resolution of 120, 000 (at m/z 200), normalized target value of 250% and maximum injection time of 500 ms. HCD MS/MS spectra (isolation window 1.2 m/z, 30% relative fragmentation energy) were acquired from m/z 110 with a relative target value of 200% (intensity threshold 5 × 10³), resolution of 30, 000 (at m/z 200) and maximum injection time of 250 ms. Dynamic exclusion was enabled for 45 s.

The analysis of the mass spectrometric RAW data files was carried out using the MaxQuant software (version 2.0.3.0) using default settings unless otherwise noted. MS/MS ion searches were done against modified cRAP database (based on http://www.thegpm.org/crap, 112 protein sequences), and UniProtKB protein database [58] for Y. lipolytica (6 454 protein sequences). Modifications were set as follows: oxidation (M), deamidation (N, Q) and acetylation (protein N-term) as variable modifications, carbamidomethylation (C) as a fixed modification. Trypsin/P enzyme specificity with two allowed missed cleavages and minimal peptide length 6 amino acids were set. Peptides and proteins with FDR threshold < 0.01 and proteins having at least one unique or razor peptide were considered only.

Reported protein intensities were further processed using the software container environment (https://github.com/OmicsWorkflows). Processing workflow is available upon request. Briefly, it covered: (a) removal of decoy hits and contaminant protein groups, (b) protein group intensities log₂ transformation, (c) loessF normalization, (d) and calculation of control/sample ratio.

Analysis of PAR chains by treatment with PARG and SVP

Treatment of PARylated YlKu70/80 with PARG and SVP

The in vitro PARylation of recombinant YlKu70/80 was performed as described below (In vitro PARylation assays) for 30 min at room temperature in a final volume of 500 μl. The sample was dialyzed for 16 h at 4°C against 2000 volumes of a dialysis buffer (50 mM HEPES-NaOH, pH 7.2; 150 mM NaCl; 10% (v/v) glycerol). For treatment with poly (ADP-ribose) glycohydrolase (PARG), 120 μl of the sample was incubated for 30 min at 30°C with 10 nM of human PARG (Sigma-Aldrich, SRP8023) in the dialysis buffer containing 0.5 mM dithiothreitol. For treatment of snake venom phosphodiesterase (SVP), 120 μl of the sample was incubated for 30 min at 30°C with 1 U of SVP (Innovative Research, IDBRPDEILY100UN) in a dialysis buffer (pH adjusted to 8.8 with 3 M Tris-base) containing 15 mM MgCl₂. Aliquots of untreated, PARG-treated, and SVP-treated samples were subjected to SDS-PAGE electrophoresis followed by (i) immunoblot analysis using anti-PAR antibodies, and (ii) staining with Pro-Q™ Diamond Phosphoprotein gel Stain (ThermoFisher Scientific). The rest of the samples were subjected to mass spec analysis aimed at the detection of ADP-ribose (PARG-treated sample) and 2′-phosphoribosyl-AMP (PRAMP, for SVP-treated samples).

Purification of PARG and SVP-treated samples

Digested samples were fractionated by solid-phase extraction (SPE) using Strata-X cartridges (Phenomenex, USA). The cartridges were first conditioned with 100% ACN (1 × 500 μl), followed by 90% ACN with 0.05% TFA (4 × 500 μl), and then equilibrated with deionized water (5 × 500 μl). The sample (30 μl) was applied to a wet carrier and after penetrating onto the cartridge, the column was washed with deionized water (5 × 60 μl) and subsequently eluted with 10–40% ACN with 0.05% TFA (2 × 60 μl of each solvent). All eluted fractions were collected into individual tubes and concentrated to a final volume of approximately 20 μl.

MALDI-MS analysis

SPE fraction (2 μl) was spotted into the solution (0.6 μl) of 2.5-dihydroxybenzoic acid (DHB, 10 mg) in ACN/H₂O/TFA (500/500/0.1 μl) deposited on the ground steel MALDI target plate. After airdrying, samples were analyzed on the UltrafleXtreme^TM mass spectrometer (Bruker, Germany). The instrument was used in the reflectron positive mode and externally calibrated using a peptide mixture over the m/z 250–3000 range. Individual precursor ions were selected manually for MS/MS experiments. MS spectra were obtained through the FlexControl software. Detected peaks were annotated manually, and structural assignment was derived from their molecular masses and MS/MS fragmentation patterns. The assignment of fragment ions in the tandem mass spectra was adopted from the Domon & Costello nomenclature [59].

Growth rate measurements

Cell cultures were grown overnight in 5 ml of YPD media at 29°C, with or without the presence of 100 μM PHE. Then, 10⁶ cells from each culture were inoculated into 10 ml of fresh media and cultivated for 24 h at 29°C. The first cell counting was done after 8 h and then in 2-h periods using CellDrop (DeNovix). The results of three independent experiments were validated by the two-tailed Student’s t-test.

Spot tests

Y. lipolytica and S. cerevisiae cells were cultivated overnight at 29°C in 5 ml of YPD or SD medium lacking uracil (SD-URA), respectively (Supplementary Table S2). Cells were counted using CellDrop (DeNovix). The cultures were diluted to obtain 3 × 10⁵ cells in 3 μl-drop and subsequently serial 10-fold dilutions were prepared. Drops were spotted onto solid YPD medium for the control, and YPD medium containing DNA-damaging agents (Supplementary Table S2). To test the growth inhibition of S. cerevisiae, drops were spotted onto SD or SGal medium lacking uracil (SD-URA, SGal-URA) (Supplementary Table S2). The results were evaluated after 2-day cultivation at 29°C.

Terminal restriction fragment (TRF) analysis

For the isolation of genomic DNA, yeasts were cultivated overnight in 25 ml of YPD medium at 29°C. Cultures were transferred each day on fresh solid YPD medium to obtain passage 10, which represents ∼150 cell divisions. Total genomic DNA was isolated as described by Barth and Gaillardin [60]. Around 1.5 μg of gDNA was digested using a Shay–Wright mix of restriction enzymes [61] (5U of AluI, HaeIII, HhaI, HinfIII, MspI, RsaI; NEB) overnight at 37°C. The fragments were separated in 1% (w/v) agarose gel for 16 h at 1.6 V/cm and stained with 0.5 μg/ml ethidium bromide solution for 20 min (stained gel served as the loading control). The gel was then washed in (i) denaturation solution (1.5 M NaCl, 0.5 M NaOH) for 40 min, (ii) neutralization solution (1.5 M NaCl, 0.5 M Tris-HCl, pH 7.4) for 30 min, and (iii) 20 × SSC (3 M NaCl, 0.3 M Na-citrate, pH 7.0) for 30 min. The DNA was then transferred to Immobilon NY + membrane (EMD Millipore) with a VacuGene XL blotter (GE Healthcare) for 3 h and fixed by incubating the membrane at 80°C for 1 h. The membrane was incubated for 1 h at 65°C in hybridization solution [5 × SSC, 0.5% (w/v) SDS, 5 × Denhardt’s solution (0.1% (w/v) Ficoll 400, 0.1% (w/v) polyvinylpyrrolidone, 0.1% (w/v) BSA, fraction V)] and hybridized at 65°C overnight in the same buffer containing 50 ng of denaturated telomere-specific probe (YlTEL81X; Supplementary Table S1) labeled with [α-³²P]dCTP using the Prime-a-Gene®Labeling System (Promega). The membrane was then washed once with 5 × SSC, 0.1% (w/v) SDS for 15 min at room temperature, and twice for 15 min at 42°C. Finally, the membrane was exposed to detection screen and the signal was detected by Typhoon Biomolecular Imager (Amersham).

Purification of recombinant proteins from E. coli

The expression plasmid pGEX-PYL1 was transformed into One Shot BL21(DE3) cells (Thermo Fisher Scientific), and the transformants were grown on LB plates containing 100 μg/ml ampicillin. The cells were then inoculated into 30 ml of LBGBA medium [1% (w/v) Bacto-peptone, 1% (w/v) NaCl, 0.5% yeast extract (pH 7.1), 2% (w/v) glucose, 10 mM benzamide (Sigma-Aldrich), 100 μg/ml ampicillin] and cultivated overnight (15 h) at 37°C and 225 rpm. The cells were centrifuged for 5 min at 3000 rpm (rotor JA 25.5, Avanti JXN Series, Beckman Coulter) at 25°C, washed once with LBB medium [1% (w/v) Bacto-peptone, 1% (w/v) NaCl, 0.5% yeast extract (pH 7.1), 10 mM benzamide], inoculated into 1 liter of LBB medium containing 100 μg/ml ampicillin, and cultivated at 37°C and 275 rpm until the OD₆₀₀ reached a value of 0.5–0.6. The culture was cooled to 28°C followed by the addition of isopropyl β-D-1-thiogalactopyranoside (IPTG; final concentration 1 mM) and cultivated for 3 h at 28°C. The culture was then centrifuged for 15 min at 4000 rpm at 4°C (rotor JLA8.1, Avanti JXN Series, Beckman Coulter), the cells were washed once with 200 ml of ice-cold phosphate-buffered saline (pH 7.0), and the pellet was frozen at −20°C. The thawed cells were resuspended in buffer A [20 mM Tris-HCl (pH 7.5), 150 mM NaCl, 0.5 mM DTT) containing 1 mg/ml lysozyme (Sigma-Aldrich), 1x protease inhibitors cocktail cOmplete (Roche), 10 mM MgCl₂, 2 μg/ml RNase A (Invitrogen) and 1 U/ml DNase I (Applichem)]. The suspension was incubated on ice for 15 min followed by sonication [6 × 20 s, 30% amplitude, Model 120 Sonic Dismembrator (Thermo Fisher Scientific)]. Triton X-100 (Sigma-Aldrich) was added to the final concentration of 0.25% (v/v) and sonicated for three additional pulses. After 5-min incubation on ice, the suspension was centrifuged for 20 min at 10 000 rpm and 4°C in (rotor JA 25.5, Avanti JXN Series, Beckman Coulter). The resulting supernatant was mixed with 0.2 ml glutathione agarose (Sigma-Aldrich) equilibrated with buffer A containing 0.25% (v/v) Triton X-100 and incubated for 3 h end-over-end at 4°C. The beads were then washed 3-times with buffer A containing 0.1% (v/v) Triton X-100 and 4-times with buffer B [25 mM HEPES-NaOH (pH 7.4), 200 mM NaCl, 1 mM EDTA-NaOH (pH 7.5), 1 mM DTT, 10% (v/v) glycerol). The beads were resuspended in 1 ml of buffer B containing PreScission protease (GE27-0843-01, Cytiva) and incubated 3 h end-over-end at 4°C. The flow-through containing tag-free Pyl1p was collected, and protein concentration was determined by NanoDrop One. The sample was aliquoted and stored at −80°C.

The purification of YlKu70/80 complex was performed as in the case of Pyl1p, with the following modifications: (i) The plasmid pDuet-Ku70/80 was transformed into One Shot BL21(DE3) cells; (ii) the thawed cells were resuspended in buffer C [20 mM Tris-HCl pH 7.5, 150 mM NaCl] containing 1 mg/ml lysozyme (Sigma-Aldrich), 1x protease inhibitors cocktail cOmplete (Roche), 10 mM MgCl₂ (Sigma-Aldrich), 2 μg/ml RNase A (Invitrogen) and 1 U/ml DNase I (Applichem); (iii) the protein extract was mixed with 0.2 ml cobalt resin (Sigma-Aldrich) equilibrated with buffer C containing 0.25% (v/v) Triton X-100 and incubated for 3 h end-over-end at 4°C. The beads were then washed three times with buffer C containing 0.1% (v/v) Triton X-100) and four times with buffer D [50 mM HEPES-NaOH (pH 7.2), 150 mM NaCl, 1 mM DTT, 10% glycerol]; (iv) the beads were resuspended in 1 ml of buffer D containing PreScission protease (GE27-0843-01, Cytiva) and incubated 3 h end-over-end at 4°C.

Recombinant Tay1p was purified as described previously [27].

In vitro PARylation assays

In vitro PARylation assays were performed according to the protocol described in Langelier et al. [62] with the following modifications: (i) complementary oligonucleotides were annealed in 40 mM Tris-HCl (pH 7.5) containing 0.1 M NaCl; (ii) the proteins were diluted in the buffer used for their elution from the affinity beads; (iii) the automodification buffer contained DTT instead of TCEP; (iv) the reactions were performed for 30 min at room temperature; and (v) 10% SDS-PAGE followed by anti-PAR immunoblot was used to detect PARylated proteins.

Electrophoretic mobility shift assay (EMSA)

The EMSA reactions used for studying Pyl1p binding to DNA contained purified Pyl1p at indicated concentrations, 1 × PARP buffer [25 mM HEPES (pH 7.4), 200 mM NaCl, 1 mM EDTA (pH 7.5), 1 mM DTT, 10% (v/v) glycerol], 1 × automodification buffer [20 mM Tris-HCl (pH 7.5), 50 mM NaCl, 5 mM MgCl₂, 100 μM DTT], 5 mM β-NAD where indicated, and 16 nM radioactively labeled dsDNA telomeric probe (see below). The assays analyzing the binding of YlKu70/80 to DNA contained 0.25 × PARP buffer [6.25 mM HEPES (pH 7.4), 50 mM NaCl, 0.25 mM EDTA (pH 7.5), 0.25 mM DTT, 2.5% (v/v) glycerol], 0.5 × YlKu70/80 buffer [25 mM HEPES (pH 7.2), 75 mM NaCl, 0.5 mM DTT, 5% (v/v) glycerol], 1 × automodification buffer, 16 nM dsDNA telomeric probe and indicated combinations of 337 nM YlKu70/80, 150 nM Pyl1, and 5 mM β-NAD. All reactions were carried out for 10 min at room temperature in a final volume of 10 μl. Subsequently, glycerol was added to the final concentration of 4.5% (v/v) and the samples were separated in 6% polyacrylamide gel in 0.5 × TBE buffer [40 mM Tris-HCl (pH 8.3), 45 mM boric acid, 1 mM EDTA] in cold (4°C) 0.5 × TBE at 10 mA per gel for 20 min. Afterwards, the gels were fixed with 10% (v/v) methanol, 10% (v/v) acetic acid for 10 min, dried, inserted into a transparent plastic package and exposed to a phosphorimager screen. The signal was detected with Typhoon Biomolecular Imager (Amersham, UK). For the dsDNA telomeric probe preparation, the 20 nt G-rich single stranded oligonucleotide containing two Y. lipolytica telomeric repeats (YlipTEL_G) was labeled at the 5′ end with T4 polynucleotide kinase (Thermo Fisher Scientific) using [γ-³²P]ATP (Hartmann Analytic) and mixed with 3-fold molar excess of the unlabeled C-strand oligonucleotide (YlipTEL_C). Mixtures were heated for 5 min at 100°C and cooled down to room temperature overnight. Newly formed dsDNA probes were added to the EMSA reactions as described above.

Analysis of association of YlKu80 with telomeres in vivo

For the chromatin immunoprecipitation (ChIP) and dot blot we adapted a protocol of Misino et al. [63]. Briefly, 50 ml of yeast cells were grown in the corresponding media to reach the exponential growth phase (OD₆₀₀ ∼ 0.6). The cells were treated with 1% (v/v) formaldehyde (Roth) for 10 min at 24°C with shaking (200 rpm) and quenched with 125 mM glycine for 10 min (24°C, 200 rpm). The cells were centrifuged at 2634 g for 5 min at 4°C and the pellet was washed once with ice-cold HBS buffer [50 mM HEPES (pH 7.5), 140 mM NaCl] and once with ChIP lysis buffer [50 mM HEPES (pH 7.5), 140 mM NaCl, 1 mM EDTA (pH 8.0), 1% (v/v) IGEPAL, 0.1% (w/v) sodium deoxycholate]. The pellet was resuspended in 400 μl of ChIP lysis buffer supplemented with protease inhibitor cocktail (Sigma–Aldrich) and snap frozen in liquid nitrogen. For lysis, the cells were beaten for 1 min (2 × 6.0m/s) in the FastPrep-24 (MP Biomedicals). The samples were centrifuged at 32 212 g for 30 min at 4°C. The pellet was resuspended in 500 μl ChIP lysis buffer supplemented with a protease inhibitor cocktail. The chromatin was sheared in Bioruptor Pico (Diagenode) under high-power mode for eight cycles (30 s ON, 30 s OFF), the tubes were kept on ice. Samples were centrifuged at 16 435 g for 5 min at 4°C, and the supernatants were transferred to pre-chilled LoBind tubes (Eppendorf). Around 10 μl was set aside as Input. 5 μl of anti-Myc-tag antibody (9B11, mouse mAb Cell Signaling, #2276) were added and the samples were incubated for 1 h at 4°C. In the meanwhile, Dynabeads Protein G (#10009D, Thermo Fisher Scientific) were prepared (40 μl of the beads was used for one sample). The beads were pulled down on magnetic rack, washed 3 times with ChIP lysis buffer supplemented with 1% (w/v) BSA and resuspended in ChIP lysis buffer supplemented with 1% (w/v) BSA and protease inhibitor cocktail. Subsequently, the beads were added to each sample and incubated at 4°C on a rotator (Phoenix, RS-RD 5) for 2 h. They were pulled down on the magnetic racks, the supernatant was discarded, and the beads were washed three times with 500 μl of ice-cold wash buffer [100 mM KCl, 10 mM Tris-Cl (pH 7.4), 0.1% (v/v) Tween-20). The wash buffer was removed and 120 μl of TE buffer [10 mM Tris-Cl (pH 8.0), 1 mM EDTA] was added, the same buffer was added also to the Input samples. The immunoprecipitated (IP) samples were incubated for 30 min at 65°C, 1400 rpm (Eppendorf Thermomixer). The samples were pulled down on the magnetic racks and 1 μl of Proteinase K (20 mg/mL) and 1 μl of RNAase A (10 mg/ml) were added. The samples and inputs were incubated at 37°C for 1 h and 1400 rpm rotation, followed by another 8-h incubation at 65°C for reverse cross-linking. The IP and the input DNA samples were purified using MinElute Kit (Qiagen) following the manufacturer's protocol. DNA was eluted in 30 μl of double-distilled water.

For dot blot, each IP was carried out 3 × from each sample and pooled together before elution. Around 10 μl of each IP sample was used as input. Around 100 μl of the eluate was mixed with 100 μl of 2 × SSC buffer and serially diluted in four 1:1 steps. The samples were spotted onto an Amersham Hybond-N⁺ membrane using a Bio-Dot Apparatus (BioRad). After drying the membrane, the DNA was crosslinked using a UVP crosslinker with 120 mJ/cm². The membrane was incubated rotating in a hybridization oven at 45°C in 10 ml of DIG Easy Hybridization solution (Merck) for 1 h. One reaction volume of a DIG-labelled telomeric YlTEL30C probe (labelled using second-generation DIG Oligonucleotide 3′-End Labeling Kit, Roche) was added and incubated overnight at 45°C. The blot was washed twice for 5 min in 10 ml 2 × SSC + 0.1% (w/v) SDS, followed by two washes in 10 ml 0.5 × SSC + 0.1% (w/v) SDS for 20 min. The membrane was rinsed in DIG wash buffer [0.1 M maleic acid (pH 7.5), 150 mM NaCl, 0.3% (v/v) Tween-20] and blocked for 30 min in 10 ml of blocking solution [0.1 M maleic acid (pH 7.5), 1% (w/v) Blocking Reagent (Roche)] at room temperature. 1:5000 Anti-Digoxigenin-AP Fab fragments (Roche) were added and incubated for 30 min at room temperature. The membrane was washed four times in DIG wash buffer, followed by 5 min in DIG detection buffer [0.1 M Tris–HCl (pH 9.5), 0.1 M NaCl]. 2 ml of CDP-Star® (Roche) were evenly distributed on the blot, which was incubated in the dark for 5 min. The blot was transferred between plastic sheets and scanned using a ChemiDoc (Bio-Rad) in the Chemiluminescence mode.

Bioinformatic screen

Yeast homologs of PARP (Pyl1) and ADP-ribosyl glycohydrolase (ARH3) were identified using TBLASTN (2.13.0+) [64] searches of 1219 reference genomes of Saccharomycotina and 34 genomes of Taphrinomycotina species available in the NCBI’s GenBank + RefSeq database (https://www.ncbi.nlm.nih.gov/genome/ as of January 17, 2025) with PARP/Pyl1 from Y. lipolytica (YALI0C17061p/XP_501931), PARG from Aspergillus fumigatus (XP_751636), and ARH3 from Aspergillus oryzae (XP_003189211) as queries. The searches resulted in 44 PARP hits (in 42 species) and 5 ARH3 hits. No hit was identified with PARG query. The E-value score cut-off was set at 1e-6 in all searches. The phylogenetic relationship of Saccharomycotina lineages was inferred from single copy orthologs identified by BUSCO (v. 5.1.2; in genome mode with ascomycota_odb10 lineage) [65], the supermatrix was calculated by BUSCO_phylogenomics pipeline (https://github.com/jamiemcg/BUSCO_phylogenomics), the tree was built using FastTree (2.1.11) [66] and visualized using the ggtree package in R 4.3.1 [67].

Amino acid sequences of selected PARP proteins were downloaded from Uniprot database [58] (https://www.uniprot.org) and SMART tool [68] (http://smart.embl-heidelberg.de) was used for the identification of corresponding functional domains.

Results

PYL1 encodes a functional PARP and is overexpressed in Y. lipolytica cells lacking active telomerase

Investigation of cellular responses of Y. lipolytica to the absence of telomerase revealed that the mutant strains retain standard growth rates despite the very rapid loss of telomeric sequences indicating a robust means of adaptation to dysfunctional telomeres [29]. To get an insight into the nature of this adaptation, we previously used RNAseq to identify transcripts whose abundance increased early after the telomerase loss. One of the genes exhibiting overexpression was YALI0C17061g, encoding a putative homolog of PARPs [29], here named Pyl1. Employing RT-qPCR analysis, we confirmed that PYL1 is upregulated not only in the mutant lacking the TER gene for telomerase RNA (2.4-fold increase), but also in Δest2 mutant without the catalytic subunit of telomerase (3.1-fold increase) (Fig. 1A). This suggests that PYL1 overexpression is indeed a part of a response to the loss of telomerase activity. These results are particularly interesting, as so far, no PARP-like protein has been characterized in any ascomycetous yeast species. We therefore decided to perform a more detailed analysis of Pyl1p and its function(s) in Y. lipolytica cells.

Figure 1.

PYL1 gene is overexpressed in telomerase mutants resulting in increased PARylation of proteins. (A) RT-qPCR analysis of PYL1 gene expression normalized to ACT1 transcripts. Δter and Δest2 mutants show 2.4-fold and 3.1-fold increased ratio of PYL1/ACT1 gene expression compared to WT cells, respectively (*** P< 0.001, ** P< 0.01; two-tailed Student’s t-test) (B) Domain organization of PARP proteins. Amino acid sequences were downloaded from Uniprot database (https://www.uniprot.org) and SMART (http://smart.embl-heidelberg.de) searches were used for the identification of corresponding domains. CAEEL: Caenorhabditis elegans; ARATH: Arabidopsis thaliana; EMEND: Emericella (Aspergillus) nidulans; NEUCR: Neurospora crassa; YARLI: Yarrowia lipolytica. (C) Immunoblotting analysis of protein macrodomain pull-downs (Af1521) using anti-PAR antibodies. NC (negative control) – pull-downs of proteins from WT cells using mutated macrodomain with no binding activity.

Predicted Pyl1 protein structure exhibits similarities to other ARTs (Fig. 1B). The analysis using the SMART tool [68] identified N-terminal WGR domain (1-68 aa), presumably involved in DNA binding, two coiled-coil domains located within the central part of the protein (239-306 aa and 359-389 aa), and a catalytic ADP-ribosyltransferase domain (PARP) at its C-terminus (413-595 aa). The Pyl1p sequence apparently lacks the regulatory domain (PARP_reg) present in the PARP proteins from other species, however, the coiled-coil domains appear to be unique for Y. lipolytica. The catalytic domain contains the motif [H470-Y505-E576] and therefore, according to a recent nomenclature of ARTs [1], Pyl1 belongs to the diphtheria toxin-like (ARTD) family, which is further supported by overall structural similarity of the catalytic domains present in Pyl1p and human PARP1-3 proteins (Supplementary Fig. S1).

To test if Pyl1 is a functional enzyme able to catalyze PARylation in vivo, we analyzed Y. lipolytica protein extracts for the presence of PARylated proteins using immunoblotting with anti-PAR antibodies. In contrast to human cell extracts, where PARylation can be detected and is increased upon the MMS treatment of cells [69, 70], we did not identify PARylated proteins in whole cell extracts of Y. lipolytica, even in the presence of MMS, most likely due to a low overall level of PARylation. Therefore, we enriched the fraction of PARylated proteins by a pull-down using macrodomain magnetic affinity resin (Af1521). Using this approach, we detected PARylated proteins both in the WT cells and Δter mutant (Fig. 1C). In agreement with the increased expression of PYL1 gene, overall PARylation was enhanced in Δter mutant. We also observed elevated levels of PARylation in Δest2 strain lacking a catalytic subunit of telomerase (Supplementary Fig. S2). In contrast to human cells, a 3-h treatment of cells with 0.05% MMS did not yield any increase in the overall PARylation (Fig. 1C, lanes 3 and 6). Importantly, PARylation signal was completely lost in extracts from cells grown in the presence of 100 μM PARP inhibitor 6(5H)-phenanthridinone (PHE) (Fig. 1C, lanes 4 and 7). As expected, the PARylation activity was also abolished in the Δpyl1 mutant and the double mutant ΔterΔpyl1 (Fig. 1C, lanes 8 and 9). All these results support the conclusion that PYL1 codes for a functional PARP in Y. lipolytica.

Heterologous expression of PYL1 in Saccharomyces cerevisiae causes growth inhibition

Studies of human [71] and Arabidopsis thaliana [72] PARPs demonstrated that PARylation activity of heterologous PARP enzymes expressed in S. cerevisiae causes growth inhibition that is correlated with the appearance of PARylated proteins. To test if the heterologous expression of PYL1 causes a similar effect, we expressed the PYL1 gene in S. cerevisiae which naturally lacks a PARP homolog and, consequentially, protein PARylation as a post-translational modification. First, we cloned the PYL1 gene into the pYES-2CT plasmid placing the PYL1 open reading frame under the P_GAL1 promoter, which allows induction of heterologous gene expression when the cells are cultivated on media containing galactose. Indeed, the growth of S. cerevisiae cells upon heterologous PYL1 expression was partially inhibited (Supplementary Fig. S3A). When we inspected the level of PARylation by the pull-down approach described above, in contrast to the control strain we observed a strong signal in case of cells overexpressing PYL1 that disappeared when the cells were treated with PHE or when they expressed PYL1 lacking the catalytic domain (Supplementary Fig. S3C). However, the inhibitory effect of PYL1 overexpression on cell growth was not attenuated when PARylation was prevented (Supplementary Fig. S3A and B). This is in contrast with the studies, where it was shown that the toxicity of overexpression of human [71] and plant [72] PARPs is alleviated by the inhibition of catalytic activity. In spite of this discrepancy, the results of the immunoblot analysis (Supplementary Fig. S3C) clearly show that expression of PYL1 in S. cerevisiae cells yields elevated levels of PARylated proteins and thus represent additional evidence that PYL1 encodes a functional PARP.

YlKu70/80 complex and Pyl1 itself are enriched in the fraction containing PARylated proteins

To determine the PARylation targets in Y. lipolytica cells, the pull-downed PARylated proteins from the WT, Δter, and Δpyl1 cells were analyzed by mass spectrometry. In Δter mutant, a significant enrichment was observed for 16 proteins, from which 7 were also found to be PARylated in WT cells (Table 1, Supplementary Table S3). The most abundant PARylated protein in WT and Δter strain was Pyl1, indicating that, similarly to its human counterparts, it undergoes auto-PARylation [11, 73]. Among the enriched proteins we also identified both subunits of YlKu70/80 complex. Although it cannot be excluded that the enrichment of YlKu70/80 subunits may not be due to their PARylation, but rather interaction with other PARylated proteins, the results of in vitro experiments (see below, Fig. 3) are in line with a hypothesis that they are direct targets of Pyl1p. In any case, as the functional YlKu70/80 complex was previously shown to cause a defect in telomere homeostasis in Y. lipolytica, it is a strong candidate for a factor involved in Pyl1-dependent adaptation of cells to the loss of telomerase [29]. Among other PARylated proteins, we found a homolog of S. cerevisiae Rrm3 helicase, which has been implicated in the progression of subtelomeric and telomeric replication [74], and a group of proteins involved in ribosome biogenesis and RNA maturation process. This is consistent with the fact that human PARP1 has also been described to have a role in RNA biogenesis and splicing regulation [75, 76].

Table 1.

Proteins significantly enriched in macrodomain pull-downs of WT and Δter mutants analyzed by mass-spectrometry analysis. Sixteen proteins were found in Δter mutants, and seven of them were also present in WT protein extracts (X - indicates the presence of particular protein in WT and/or Δter strain)

Accession UniProt	Description	Wild-type extracts	Δter extracts
Q6CBN1	Poly [ADP-ribose] polymerase	X	X
Q6C7B9	Ku80	X	X
Q6C9N5	Dynamin 1-like protein		X
Q6CAI0	ATP-dependent DNA helicase (homologous to ScRrm3)		X
Q6CG44	Nucleolar GTP-binding protein 1	X	X
Q6CEY9	Nucleolar protein SRP40	X	X
Q6CCK2	Ku70		X
Q6C9H2	Nucleolar protein 9 (rRNA processing)		X
Q6CCT1	1, 4-alpha-glucan-branching enzyme	X	X
Q6CDI7	Protein involved in lipid metabolism		X
Q6CGD1	ATP-dependent RNA helicase DBP4		X
Q6C6T3	Ser/Thr kinase	X	X
Q6C3L5	Protein involved in RNA splicing		X
Q6C3H5	ATP citrate synthase	X	X
Q6CA00	Protein involved in cytoplasmic translation		X
Q6CEE6	DNA-directed RNA polymerase subunit beta		X

Figure 3.

Pyl1p exhibits PARylation activity in vitro. (A) Recombinant versions of Pyl1p, YlKu70/80 complex, and Tay1p were purified from E. coli as described in the “Materials and Methods” section, separated by electrophoresis in 10% polyacrylamide gel and stained with Coomassie Brilliant Blue. (B) Pyl1p was incubated with the indicated compounds for 30 min at room temperature, proteins were separated by electrophoresis in 10% polyacrylamide gel followed by an immunoblot analysis of PARylated proteins using anti-PAR antibodies. (C) PARylation of YlKu70/80 complex and Tay1p by Pyl1p in vitro was assessed as in (B) except that a lower concentration of Pyl1p was used to eliminate a signal from its auto-PARylation. (D) Treatment of the PARylated proteins with snake venom phosphodiesterase (SVP) and PARG results in a loss of signal recognized by anti-PAR antibodies. (E) SVP-treated (but not PARG-treated) proteins retain a phosphate group representing phosphoribose (Pro-Q Diamond phosphoprotein staining).

Deletion of PYL1 gene slows down the growth of Δter mutant but does not affect the telomere length

Based on the increased PARylation of proteins in the Δter mutant, it can be assumed that this post-translational modification plays a role in the adaptation of yeast cells to the loss of telomerase. Therefore, we compared the growth characteristics of mutants defective in telomere maintenance and those lacking functional PYL1 gene. The growth rates of the strains over the 24-h period were similar, except ΔterΔpyl1 mutant exhibiting a slower growth during the mid- and late-exponential phase (Fig. 2A).

Figure 2.

Effect of PYL1 deletion on growth and telomere length. (A) Cells were inoculated into the YPD medium at the same density and counted using Cell-drop counter (DeNovix) in 2-h time interval for 24 h. Statistical significance between Δter and ΔterΔpyl1 strains was assessed by two-tailed Student’s t-test (*** P< 0.001, ** P< 0.01, *P< 0.05) (B) TRF analysis of the corresponding strains of Y. lipolytica. Membrane was hybridized with [³²P]-labeled YlTEL81X probe representing 81 telomeric repeats of Y. lipolytica. Δpyl1_#1, Δpyl1_#2, two independent mutants lacking PYL1; ΔterΔpyl1_#1, ΔterΔpyl1_#2, two independent mutants lacking both TER and PYL1.

As an alternative approach, we compared the 24-h growth rate of the WT, Δter, and ΔterΔku80 mutant strains in the presence of PARP inhibitor PHE, which completely abolishes detectable PARylation of intracellular proteins as shown in Fig. 1C. Significant, yet transient growth delay was observed between the WT and ΔterΔku80 strain in the 10th and 14th hour of cultivation, and between Δter and ΔterΔku80 in the 22nd hour of cultivation in the presence of PHE inhibitor (Supplementary Fig. S4). As the absence of Pyl1 activity affects telomerase mutants only transiently (Fig. 2A and Supplementary Fig. S4), it seems that PARylation may play a role during the adaptation to the loss of telomerase activity. Nevertheless, it is not essential for viability of the cells adapted to telomerase deficiency.

To assess whether PYL1 directly participates in long-term telomere maintenance, we subjected the strains with various genetic backgrounds to terminal restriction fragment (TRF) analysis. We observed that in two independent Δpyl1 strains, the telomeres are of the same length as those from WT cells (Fig. 2B, lanes 1 and 2). Furthermore, the double mutants ΔterΔpyl1 (Fig. 2B – lanes 3 and 4) and Δku80Δpyl1 (Fig. 2B – lane 5) did not exhibit any alterations of the telomeric phenotype compared with their parental strains Δter (no signal) and Δku80 (long heterogeneous telomeres) (Fig. 2B – lanes 8 and 9). This indicates that Pyl1 is important for the adaptation to the absence of telomerase, but it does not participate in the long-term maintenance of telomeres or does so only under specific, yet unknown conditions.

Pyl1p exhibits PARylation activity in vitro

To assess that Pyl1p is a functional PARP, we expressed the PYL1 gene in E. coli and purified the recombinant protein by affinity chromatography (Material and Methods; Fig. 3A). Using the in vitro PARylation assay, we demonstrate that Pyl1p undergoes auto-PARylation that is dependent on NAD⁺ and independent from the addition of exogenous DNA (Fig. 3B – lanes 4–6). The stoichiometry of PARylation is not high as on gels stained with Coomassie Brilliant Blue R-250 we did not observe slower migrating species of Pyl1p that would represent its PARylated subpopulation (Supplementary Fig. S5A). We also noted that the purified Pyl1p contains PAR residues, indicating that it is auto-PARylated in E. coli cells, yet the level of its PARylation is clearly increased after the in vitro reaction. Moreover, in contrast to the basal PARylation, the in vitro auto-PARylation is inhibited by 100 μM PHE (Fig. 3B – lanes 7 and 9).

As YlKu70/80 complex is the major constituent of the protein fraction present in the anti-PAR pull-downs, we tested whether it is a direct substrate of Pyl1p in vitro. To this end, we purified the YlKu70/80 dimer from E. coli cells (Materials and Methods, Fig. 3A) and subjected it to PARylation reaction in vitro. Note that the concentration of Pyl1p in this experiment is 9-fold lower than that used for auto-PARylation assay (0.15 μM versus 1.4 μM), so the PARylation signal comes exclusively from the PARylated YlKu70/80. As in the case of Pyl1p auto-PARylation, the stoichiometry of PARylation is not high, as demonstrated by the lack of smeared bands on Coomassie-stained gels (Supplementary Fig. S5B). On the other hand, the results of the anti-PAR immunoblot analysis clearly demonstrate that the YlKu70/80 complex is PARylated by Pyl1p and that this modification is sensitive to 100 μM PHE (Fig. 3C – lanes 4–6).

To evaluate the substrate specificity of Pyl1p, we performed the in vitro reaction using another telomere-associated protein, Tay1p [27], that, in contrast to YlKu70/80 complex, was not enriched in anti-PAR pulldowns. We show that the purified Tay1p (Fig. 3A), when incubated with Pyl1p and NAD⁺ is not modified by PAR residues (Fig. 3C – lanes 8 and 9).

To demonstrate that the moieties recognized by anti-PAR antibodies are indeed represented by poly (ADP-ribose) chains, we performed analysis of de-PARylation products obtained by treatment of the PARylated YlKu70/80 with poly (ADP-ribose) glycohydrolase (PARG) and snake venom phosphodiesterase (SVP). Both of these enzymes should remove PAR chains from the PARylated proteins, albeit via different means. Whereas digestion of PAR with PARG should result in a release of ADP-ribose yielding mono-ADP-ribosylated protein substrate, treatment of PAR with SVP should yield 2′-phosphoribose-AMP (PRAMP) while the protein retains a phosphoribose at the modified amino acid. Indeed, treatment of the in vitro PARylated YlKu70/80 with either PARG or SVP resulted in a loss of a signal detected by anti-PAR antibodies (Fig. 3D). Using staining with Pro-Q™ Diamond Phosphoprotein gel Stain also showed that treatment of the PARylated YlKu70/80 with SVP yields 70 and 80 kDa polypeptides presumably containing phosphoribose [77] (Fig. 3E). The sensitivity of PARylated YlKu70/80 to PARG, that is highly specific for poly (ADP-ribose) [78], provides additional evidence for a presence of PARs on YlKu70/80.

To further confirm the presence of ADP-ribose and 2′-phosphoribosyl AMP in samples containing the PARylated proteins, digests treated with PARG and SVP were analyzed by MALDI-MS. Among numerous peaks, both samples showed a prominent signal at m/z 560.08 (MH⁺). However, as demonstrated in Fig. 4, tandem MS analysis of these precursor ions revealed distinct fragmentation patterns. The profile of the PARG-treated sample supported the presence of ADP-ribose (Fig. 4A), whereas the same m/z precursor ion in the SVP-digested sample produced fragment ions consistent with 2′-phosphoribosyl AMP (Fig. 4B).

Figure 4.

MALDI-TOF/TOF tandem mass spectra recorded for the precursor ions with m/z 560.08 (MH⁺) from protein sample treated for 30 min with (A) PARG, and (B) SVP. The fragmentation schemes represent the major isomeric structures detected in samples. The assignment of fragment ions in the spectra corresponds to the respective fragmentation scheme.

In summary, these results demonstrate that Pyl1p: (i) is a functional PARP; (ii) undergoes NAD⁺-dependent and PHE-sensitive auto-PARylation; (iii) PARylates YlKu70/80 complex; (iv) it seems to exhibit a substrate specificity; and (v) treatment of PARylated proteins with PARG and SVP yields expected products (Figs. 3 and 4).

Association of YlKu70/80 complex with Y. lipolytica telomeres is reduced in cells overexpressing PYL1

To assess the functional significance of PARylation of the YlKu70/80 in vivo, we investigated the association of one of its subunits (YlKu80) with telomeres in cells producing different levels of Pyl1p (Fig. 5A). Our attempts to tag YlKu80 at its genomic locus using different strategies repeatedly failed, most likely due to inaccessibility of this site to the HR machinery combined with a high activity of non-homologous end joining pathway [54]. To circumvent this problem, we have expressed a Myc-tagged version of YlKu80 in Y. lipolytica cells from a plasmid. To check that ectopic expression of YlKu80 does not affect telomere homeostasis, we measured TRFs in the transformants and found that their length is the same as in the wild-type cells bearing a control vector (Supplementary Fig. S6). Next, we performed a ChIP using anti-Myc antibodies followed by a dot blot analysis of telomeric DNA. The results demonstrate that telomeric DNA is enriched in the YlKu80-IP samples, indicating that YlKu80 is associated with telomeres in vivo (Fig. 5B and C – row 1/column 1). The specificity of the signal was verified by a ChIP of YlKu80 from Δter strain that lacks the telomeric repeats (Fig. 5B and C – row 2/column 2).

Figure 5.

Association of YlKu80 with Y. lipolytica telomeres in vivo is reduced in cells overproducing Pyl1p. (A) Different levels of Pyl1p were achieved by cultivating the cells harboring a plasmid vector (pUB4) or a derived construct carrying PYL1 under the P_EYK1 promoter (pUB4 + PYL1) in media containing glucose or erythritol. (B) YlKu80 binding to telomeres was assessed by ChIP followed by dot blot hybridization with the telomeric DIG-labelled telomeric probe (YlTEL30C) as described in the “Materials and Methods” section. (C) Quantification of the dot blot. The signal for precipitated telomeric DNA was normalized to total telomeric DNA (input) (n = 2, mean +/− SEM is displayed).

Next, we wanted to test, if the binding of YlKu70/80 to telomeres is affected by an increased level of PARylation. To this end we constructed a multicopy vector (pUB4 + PYL1) encoding PYL1 with a C-terminally fused V5 tag under the control of P_EYK1 promoter. EYK1 gene encodes L-erythrulose-1-kinase whose expression is induced when yeasts are grown on medium containing erythritol or erythrulose instead of glucose, so placing P_EYK1 promoter upstream of the PYL1 enables regulation of its expression by the specific carbon source (glucose versus erythritol; see below) [79]. We transformed pUB4 and pUB4 + PYL1 plasmids into Y. lipolytica cells and performed immunoblotting analysis of whole-cell protein extracts using anti-V5 antibodies. The results demonstrate a basal level of expression in cells grown on glucose that is increased in cells grown on erythritol (Supplementary Fig. S7). Thus, we can obtain cells expressing PYL1 at normal (cells transformed with an empty pUB4 vector), moderate (pUB4 + PYL1 transformants grown on glucose) and high (pUB4 + PYL1 transformants grown on erythritol) levels, respectively (Fig. 5A). When we compared the association of YlKu70/80 in the wild-type cells (Fig. 5B and C – row 1/column 1) with the cells expressing PYL1 at moderate or high levels (Fig. 5B and C – rows 3, 4/columns 3, 4), we observed a substantially reduced association of YlKu80 with telomeres in cells with an increased Pyl1 activity. This observation is in line with the hypothesis that an increased PARylation of YlKu70/80 results in its dissociation from telomeric regions. Interestingly, when we tested the DNA-binding properties of YlKu70/80 in vitro, PARylation did not decrease its binding to telomeric probe (Supplementary Fig. S8A). We also did not observe decreased levels of YlKu80 in cells overexpressing PYL1 (Supplementary Fig. S8B), ruling out the possibility that PARylation of YlKu70/80 decreases its overall stability. This indicates that the Pyl1-induced dissociation of YlKu70/80 from telomeres is mediated by yet unidentified factor.

To test the effect of PYL1 overexpression on Y. lipolytica telomeres, the cells transformed with a control plasmid (pUB4), the vector carrying standard PYL1 (pUB4 + PYL1) as well as PYL1 lacking the catalytic domain (pUB4 + PYL1Δcat) were passaged for 1 (passage 1) or 10 (passage 10) days on media with glucose or erythritol followed by TRF analysis. Under these conditions, we did not observe significant changes in telomere length in any of the tested strains (Supplementary Fig. S9).

Double mutant Δku80Δpyl1, but not single Δpyl1 mutant exhibits increased sensitivity to hydroxyurea and hydrogen peroxide

Several human PARPs are important factors in DNA repair and DNA damage recognition [9]. To elucidate whether the Pyl1 protein participates in any of the major DNA repair pathways in Y. lipolytica, we treated yeast cells with four different DNA-damaging agents. The WT cells, as well as two independent Δpyl1, Δku80, and Δku80Δpyl1 mutants, were grown overnight and spotted on media containing different concentrations of MMS, zeocin, hydrogen peroxide, or hydroxyurea (HU). As the RAD52 protein is involved in recombination-dependent repair pathways in Y. lipolytica [80], Δrad52 mutant served as a positive control. Surprisingly, our results did not reveal an increased sensitivity of Δpyl1 and/or Δku80 mutants to any of the DNA-damaging agents tested (Fig. 6). On the other hand, we observed a negative genetic interaction between YlKU80 and PYL1 genes in case of double deletion mutant Δku80Δpyl1 on medium containing H₂O₂ (>6 mM) or HU. Both agents can cause stalling of the replication forks at telomeres either through oxidative damage of these G-rich regions (H₂O₂) or via depletion of deoxyribonucleotides (HU), although it was shown recently that the latter also works by producing copious reactive oxygen species [81]. Earlier studies suggested that YlKu70/80 complex is not primarily involved in the repair of stalled or broken replication forks [82], so increased sensitivity of the double mutant to H₂O₂ and HU might be connected to problems with the replication of overelongated telomeres in Δku80 mutants, whose solution may be dependent on the functional Pyl1. However, it was shown in fission yeast that Ku70/80 can process reversed replication fork [83], thus the genetic interactions between YlKU70/80 and PYL1 may be more complex.

Figure 6.

Spot tests of sensitivity of mutant strains to DNA-damaging agents. 10-fold serial dilutions of WT, Δrad52, Δku80, two independent mutants Δpyl1_#1 and Δpyl1_#2, and double mutant Δku80Δpyl1 were cultivated for 48 h on YPD media containing methyl methanesulphonate (MMS) (0.02%, 0.03%), zeocin (50, 75, and 100 μg/ml), hydrogen peroxide (6, 7, and 8 mM), and hydroxyurea (40, 50, and 60 mM). Mutant Δrad52 serves as a positive control.

PARP homologs are present in several clades of Saccharomycotina

The presence of functional PARP in Y. lipolytica indicates that the occurrence of similar enzymes in ascomycetous yeasts may be more widespread than previously thought. To investigate the distribution of Pyl1 homologs, we performed a systematic bioinformatic search among species classified to the subphyla Saccharomycotina and Taphrinomycotina. These searches did not find any hit in Taphrinomycotina, yet they revealed candidate PARPs in three orders of Saccharomycotina (Alloascoideales, Dipodascales, Phaffomycetales) including two Alloascoidea species, eighteen species of the genus Yarrowia and closely related Candida hispaniensis, three Blastobotrys species, Trichomonascus vanleenenianus, fourteen species of the Barnettozyma/Candida clade, and two yeasts classified as Trichomonascaceae sp. Y4102 and Wickerhamomyces sp. AT-2023 (Fig. 7, Supplementary Fig. S10). All identified proteins have conserved PARP catalytic domain. The searches for potential de-PARylation enzymes PARG and ARH3 revealed only ARH3 homologs in a few species, including Candida hispaniensis (order Dipodascales) and four species belonging to the order Serinales (Candida glaebosa, C. fluviatilis, C. sphangicola, and Diutina catenulata). Interestingly, C. hispaniensis is the only species possessing homologs of both PARP and ARH3, yet its PARP is most likely non-functional because the corresponding open reading frame contains a nonsense mutation eliminating the entire PARP catalytic domain [84]. We tested the possibility that, in addition to PARP, Pyl1p also exhibits hydrolase activity in vitro. However, when we used PARylated YlKu70/80 complex as a substrate for Pyl1p in the absence of NAD⁺, we did not observe a decrease in a PARylation signal (Supplementary Fig. S12). Thus, the nature of an enzyme responsible for the removal of PARs in Y. lipolytica and other PARP-containing yeasts remains enigmatic.

Figure 7.

Distribution of PARP, ARH3, and PARG homologs in Saccharomycotina and Taphrinomycotina. The phylogeny was inferred from single-copy orthologs identified by BUSCO (v. 5.1.2) [64] in the reference genome assemblies of representative species of each clade, i.e.Alloascoidea hylecoeti, Ascoidea rubescens, Debaryomyces hansenii, Dipodascus albidus, Lipomyces starkeyi, Pachysolen tannophilus, Phaffomyces opuntiae, Pichia membranifaciens, Saccharomyces cerevisiae, Saccharomycodes ludwigii, and Trigonopsis variabilis. Schizosaccharomyces pombe and Aspergillus nidulans representing the subphyla Taphrinomycotina and Pezizomycotina, respectively, were used as an outgroup. The tree was built using FastTree (2.1.11) [65] from the supermatrix calculated by BUSCO_phylogenomics pipeline and visualized using ggtree in R 4.3.1 [66]. The number of species with PARP/Pyl1, ARH3, and PARG hits in TBLASTN searches (see “Materials and Methods”) is indicated for each clade. The number of reference genome sequences available in the GenBank + RefSeq database for each clade (as of January 17, 2025) is shown in square brackets.

Discussion

Although the presence of a putative PARP gene in Y. lipolytica was noticed before [51], PARylation as a process of post-translational modification in ascomycetous yeasts has not been reported so far. The absence of PARP in such a large group of organisms would be intriguing as this enzyme was found in species from all domains of life. Therefore, the identification of a functional PARP in Y. lipolytica and the detection of its homologs in other yeast species represents a starting point for their comparative analysis, as well as filling the gaps in understanding their evolutionary history.

The conclusion that Pyl1 is a functional PARP is based on the following results: first, the level of PARylated proteins correlates with the level of expression of the PYL1 gene (Fig. 1A,C). Second, PARylation in the protein extracts is completely lost in Δpyl1 strain and when WT cells are cultivated in the presence of PARP inhibitor PHE (Fig. 1C). Third, when S. cerevisiae cells that lack a native PARP are transformed with an expression vector carrying PYL1 gene under the P_GAL1 promoter, there is a galactose-induced presence of PAR signal in the protein extracts. The signal is lost when the PYL1-expressing cells are cultivated in the presence of PARP inhibitor and when they overproduce Pyl1 protein lacking catalytic domain (Supplementary Fig. S3C). We also show that purified Pyl1p undergoes NAD⁺ dependent auto-PARylation and also that it PARylates YlKu70/80 complex in vitro, and this reaction is inhibited by PHE (Fig. 3B and C). Finally, we provide the evidence that the side chains are indeed represented by poly (ADP-ribose) (Fig. 3D and E and 4). Interestingly, in vitro PARylation activity of Pyl1 is not dependent on exogenously added DNA (Fig. 3B). This may have two explanations. First, the purified Pyl1 protein can already be associated with bacterial DNA, even though the protein extracts were extensively treated with DNase I (see Material and Methods). When we tested its DNA-binding properties by EMSA, we found that the purified Pyl1p is able to bind to dsDNA in vitro (Supplementary Fig. S11). This indicates that at least a subpopulation of Pyl1 molecules has the capacity to bind the dsDNA. Interestingly, DNA-binding activity of Pyl1 is not affected by auto-PARylation (Supplementary Fig. S11). Alternatively, PARP activity of Pyl1p may be independent of DNA binding as is the case of several mammalian PARPs [4]. Addressing this and other, methodologically challenging biochemical characteristics of Pyl1p such as the preference for amino acid targets or the length/architecture (linear vs branched) of poly-ADP-ribose chains [85] will be subjects of the follow-up studies.

Since it was not possible to detect PARylation in the whole protein extracts, even upon MMS treatment, it was necessary to enrich the PARylated proteins by affinity pull-down. To this end, we employed an established experimental strategy based on a specific recognition of PARylated proteins by Af1521-macrodomain [86]. The advantage of using the macrodomain compared to anti-PAR antibodies is that its binding to the target proteins is not dependent on the size of the ADP-ribose chain. The results presented in Fig. 1C show that all PARylated proteins pulled-down by the Af1521-macrodomain beads are Pyl1 substrates as they are completely absent in extracts from cells treated with the PARP-inhibitor (PHE). Thus, we believe that our approach is efficient in terms of the identification of the most prominent Pyl1p substrates.

We assume PARylation in Y. lipolytica is relatively rare compared to human cells and it is not increased when cells are exposed to an alkylation agent [69]. Our observation that the strain lacking functional PYL1 gene does not exhibit an increased sensitivity to several genotoxic drugs also indicates that PYL1 is not involved in DNA damage response, at least under the tested conditions (Fig. 6). Increased sensitivity of double mutant Δku80Δpyl1 to H₂O₂ and HU opens a possibility, that Pyl1p is involved in solving problems with replication fork stalling, which can be a result of problems with replication of over-elongated telomeres in Δku80 mutant.

Several lines of evidence suggest that Pyl1 plays a role in the adaptation of Y. lipolytica cells to a crisis caused by dysfunctional telomeres. Although, similarly to Neurospora crassa [53] or Arabidopsis thaliana [72], the Δpyl1 mutant of Y. lipolytica does not exhibit telomere shortening (Fig. 2B), the gene is overexpressed in both Δter and Δest2 mutants lacking either the RNA or the catalytic subunit of telomerase (Fig. 1A). In addition, the growth rate of the WT, Δter, and ΔterΔku80 strains is not transiently affected by the inhibition of PARP (Fig. 2A, Supplementary Fig. S4), indicating that although Pyl1 may be important for the cell transition from telomerase-dependent to telomerase-independent maintenance of chromosomal ends, it is dispensable in cells adapted to the absence of telomerase.

In mammalian cells possessing multiple PARPs, PARylation has diverse effects on telomere maintenance. For example, PARP1, 2, and 3 have been shown to participate in the regulation of telomere length homeostasis [87], and long-term inhibition of tankyrases leads to telomere shortening [88]. Moreover, PARylation of TRF1 promotes its dissociation from telomeres, enabling telomere elongation [49]. In humans, overexpression of tankyrases causes telomere elongation and PARP1 overexpression stimulates hTERT transcription [88, 89]. Additionally, the decreased levels of PARP3 in human cancer cells correlate with the increase in telomerase activity [90], and human PARP2 interacts with TRF2, which enables t-loop opening. Nevertheless, similarly to Δpyl1 of Y. lipolytica, PARP2^−/− human cells show normal telomere length, yet they display increased frequency of chromosome and chromatid breaks [91]. These reports indicate that although PYL1 does not significantly contribute to telomere length regulation under standard conditions, it may be involved in the regulation of telomere dynamics under specific cellular contexts.

In mammalian cells, a balance between PARylation and de-PARylation appears to be important in the ALT mechanism [48]. Inhibition of PARP activity led to an increase of ALT-associated PML (promyelocytic leukemia protein) bodies (APBs) in ATRX-mutated cancer cells. On the other hand, PARylation was required to mediate HIRA (Histone Cell Cycle Regulator) association with ALT telomeres, which is crucial for the deposition of histone H3.3, an important step during homology-directed repair in ALT cells [48]. Due to different types of ALT mechanisms, ALT telomeres have frequent nicks and gaps [92], and double-strand DNA breaks (DSBs) at telomeres are substrates for PARP1 and Ligase-3 dependent reactions in human ALT cancer cells [47]. Trapping of proteins, including PARP1, on DNA by PARP inhibitors leads to ALT induction in cells lacking ATRX [93].

Despite a wealth of evidence supporting the role of mammalian PARPs in normal and ALT cells, the role of Pyl1 in the telomere homeostasis of Y. lipolytica is most likely more subtle. A clue about the potential role of Pyl1 in the transition of cells to a telomerase-independent state may be provided by YlKu70/80 complex. This is supported by the observation that the main targets of PARylation in Δter mutants are both subunits of this complex (Table 1, Supplementary Table S3). Ku70/80 interacts with mammalian telomeric DNA [94] and binds TRF1 [95] and TRF2 [96] proteins. Mouse cells lacking Ku proteins display telomere elongation [95] but not in a telomerase-deficient background [40], suggesting that Ku70/80 may be a negative regulator of telomerase. Specifically, the access of telomerase to telomeres may be mediated by the direct interaction of the Ku70/80 complex with the catalytic subunit of telomerase [97]. Similarly, Ku70/80 inactivation in A. thaliana leads to telomerase-dependent telomere elongation [41–43]. Deletion of YlKU80 gene in Y. lipolytica results in telomere elongation and heterogeneity, suggesting a similar function of the Ku protein complex as in mammals, A. thaliana, or yeast Candida albicans [38]. This finding is in contrast with S. cerevisiae and S. pombe, where an absence of Ku70 leads to telomere shortening, and it was also shown that Ku80 is a direct interaction partner of the RNA subunit of telomerase (TLC1) [33–37]. Therefore, the role of Ku proteins in S. cerevisiae and S. pombe could be to suppress the accessibility of telomeres to recombination machinery, as S. cerevisiae and S. pombe Δku mutants show an increase in subtelomeric and telomeric recombination [34, 98]. Overall, the deletion of Ku genes has different outcomes in different species or different genetic backgrounds. It was also reported, that PARP1 protein and Ku complex compete for binding to DSBs in mouse embryonic fibroblasts (MEFs) [99]. The demonstration that Ku complex is one of the main targets of PARylation upon the loss of telomerase combined with the observation that the overexpression of PYL1 results in its dissociation from telomeres (Fig. 5B and C), supports the idea that in Y. lipolytica an elevated PARylation of YlKu70/80 complex can be important for cell adaptation to the absence of telomerase. On the other hand, the observation that we did not detect changes in telomere length in cells with elevated levels of PYL1 (Supplementary Fig. S9) that would mimic the phenotype of Δku80 mutant indicates that PARylation of the YlKu70/80 complex results in a strongly reduced yet not completely abolished binding to telomeres, thus retaining its ability to control their length. Even though PARylation of Ku proteins has been previously described in the process of DSB repair [12], it was never before studied in the context of telomere maintenance. Elongated telomeres in Δku80 mutant may introduce problems in telomere replication. The increased sensitivity of double mutant Δku80Δpyl1 to hydroxyurea (Fig. 6) might be a consequence of replication in Δku80 cells where Pyl1 may participate in resolving the stalled replication forks. In addition to YlKu70/80 complex, some of the other Pyl1 targets (Table 1, Supplementary Table S3) may also mediate its functions on telomeres. For example, a DNA helicase Rrm3 was shown in S. cerevisiae to promote replication fork progression through telomeric and subtelomeric DNA [74], and the PARylated proteins involved in RNA metabolism may participate in telomerase biogenesis and/or homeostasis of telomeric transcripts (TERRA). PARylation also affects the activity of H/ACA protein complex, which is important for telomerase RNP assembly and activity in mammalian cells [100, 101]. The addition of PAR moieties to protein targets has been observed to contribute to phase separation [102], so the nuclear protein PARylation may be an important mechanism for the formation of molecular condensates and membrane-less organelles. These possible functions of Pyl1 will be further explored in the follow-up studies.

In addition to Pyl1 as the first characterized PARP in ascomycetous yeasts and identification of YlKu70/80 complex as a Pyl1 target involved in telomere maintenance, we also show that the occurrence of PARPs in this subphylum is more common than previously thought, yet it shows only a sparse distribution on the phylogenetic tree (Fig. 7, Supplementary Fig. S10). What remains enigmatic is the nature of the hydrolase that is responsible for the de-PARylation of Pyl1 targets. Paradoxically, the only candidate for an Arh3 homolog was identified in the genome of C. hispaniensis, which seems to possess only a PARP pseudogene. Although it is possible that PARylated proteins in Y. lipolytica and other yeasts may be degraded rather than de-PARylated, it seems more likely that the removal of PARs is catalyzed by a noncanonical, yet unknown hydrolase. Alternatively, Pyl1p could act as both transferase and hydrolase as was shown for human PARP14 [103]. However, when we tested the ability of Pyl1p to de-PARylate YlKu70/80, we did not observe any hydrolase activity (Supplementary Fig. S12), so the nature of the enzyme responsible for de-PARylation of the modified proteins in Y. lipolytica remains elusive. Its identification would underline the importance of exploring the yeast biodiversity for obtaining a more complete picture of the protein toolbox employed in the regulation of this post-translational modification.

Supplementary data

Supplementary data is available at NAR online.

Conflict of interest

None declared.

Funding

This work was funded by the Slovak Research and Development Agency [APVV-23-0056, APVV-19-0068, APVV 22-0144], the Scientific Grant Agency of the Ministry of Education, Science, Research and Sport of the Slovak republic [VEGA 1/0031/24, 1/0234/23], National Institutes of Health [1R01ES031635-01] and Operation Program of Integrated Infrastructure for the project, Advancing University Capacity and Competence in Research, Development and Innovation, ITMS2014+: 313021 × 329, co-financed by the European Regional Development Fund. CIISB, Instruct-CZ Centre of Instruct-ERIC EU consortium, funded by MEYS CR infrastructure project LM2023042, is gratefully acknowledged for the financial support of the measurements at the CEITEC Proteomics Core Facility. Computational resources were provided by the e-INFRA CZ project (ID:90254), supported by MEYS CR. Erika Lattova thanks for CEITEC institutional funding. Research in the Paeschke laboratory is funded by the CANTAR project received from the program “Netzwerke 2021,” an initiative of the Ministry of Culture and Science of the State of Northrhine Westphalia and second funding by the Fritz Thyssen Foundation (Az. 10.21.1.027MN).

Data availability

Mass spectrometric data underlying this article have been deposited to the ProteomeXchange Consortium via the PRIDE [104] partner repository with the dataset identifier PXD048801.