Deletion of Plasmodium falciparum ubc13 increases parasite sensitivity to the mutagen, methyl methanesulfonate and dihydroartemisinin

The inducible Di-Cre system was used to delete the putative ubiquitin-conjugating enzyme 13 gene (ubc13) of Plasmodium falciparum to study its role in ubiquitylation and the functional consequence during the parasite asexual blood stage. Deletion resulted in a significant reduction of parasite growth in vitro, reduced ubiquitylation of the Lys63 residue of ubiquitin attached to protein substrates, and an increased sensitivity of the parasite to both the mutagen, methyl methanesulfonate and the antimalarial drug dihydroartemisinin (DHA), but not chloroquine. The parasite was also sensitive to the UBC13 inhibitor NSC697923. The data suggest that this gene does code for an ubiquitin conjugating enzyme responsible for K63 ubiquitylation, which is important in DNA repair pathways as was previously demonstrated in other organisms. The increased parasite sensitivity to DHA in the absence of ubc13 function indicates that DHA may act primarily through this pathway and that inhibitors of UBC13 may both enhance the efficacy of this antimalarial drug and directly inhibit parasite growth.


Results
Comparison of the putative P. falciparum UBC13 with the human enzyme. An alignment of the HsUBC13 and putative PfUBC13 revealed that the 152-residue protein coded by PF3D7_0527100 is orthologous to HsUBC13 with 68% amino acid sequence identity (Fig. 1a). Most of the sequence differences are located away from the active site cysteine (Fig. 1b).
Generation of an inducible ubc13 knockout in P. falciparum. We used CRISPR-Cas9 to modify the 3' end of the putative ubc13 gene (PF3D7_0527100), inserting a loxP element in the fourth and final intron, inserting a recodonised final exon and placing a loxP site after the stop codon in the ubc13 sequence (Supplementary Fig. 1a). The changes were made in the II-3 parasite line, which expresses functional DiCre recombinase activity following rapamycin addition 32 . After transfection, the parasite population was screened by PCR for DNA integration using diagnostic primer pairs P1/P2, P1/P4, P3/P4, P1/P6 and P5/P4 ( Supplementary Fig. 1b, Supplementary Table S1) and products from both modified and unmodified locus were detected (Supplementary Fig. 1c). Parasites were cloned by limiting dilution and integration was examined in individual clones; for recombinant clones, primers P1/P6 and P5/P4, produced bands of 589 and 550 bp, respectively, diagnostic of integration and no product was obtained from primer pairs P1/P2, and P3/P4 indicating the absence of unmodified locus ( Supplementary Fig. 1c). The parasite clone 2E was used in subsequent experiments.
Induced excision of ubc13 shows that the gene is essential for the growth and survival of P. falciparum asexual blood stage. To  www.nature.com/scientificreports/ alone (the control) for 24 h. Excision of the floxed DNA was monitored by PCR amplification with primer pair P1 and P4 ( Supplementary Fig. 1d). For the control (DMSO-treated) sample a 1280 bp band was amplified, indicating no excision, and from the rapamycin-treated sample a 975 bp band was observed, indicating quantitative excision of the floxed DNA ( Supplementary Fig. 1e). This result indicates that by the late trophozoite stage at 28 h after invasion, only a truncated ubc13 gene remained following rapamycin-treatment. We next investigated whether functional ubc13 is essential for parasite growth and survival. Parasites were treated at 2 h post invasion with rapamycin or DMSO for 24 h and then further incubated for either one (48 h) or two (96 h) cell cycles and analysed by flow cytometry (FACS) and examination of Giemsa-stained thin blood smears by microscopy. At 48 h the parasitemia had increased in both samples ( Fig. 2a) but was significantly less (p < 0.008) following rapamycin treatment, with fewer new ring-stage forms resulting from merozoite release and reinvasion (Fig. 2b). There was a clear difference in the parasitemia of the control and rapamycin-treated parasites at the second cycle (p < 0.0002) (Fig. 2a). At 96 h most of the parasites were at the ring stage of development in the control cultures, whereas the rapamycin-treated parasites were largely at the schizont stage, and looking very unhealthy on Giemsa-stained smears. Little increase in parasitemia had occurred overall (Fig. 2b). and Plasmodium falciparum (Pf)UBC13 amino acid sequences are aligned, with identical residues in the two sequences highlighted in black. The active site cysteine residue that binds activated ubiquitin covalently and reacts with NSC697923 in human UBC13 is highlighted in yellow and marked with an asterisk. (b) Threedimensional structures with the active site cysteine highlighted in yellow and marked with an asterisk. In the PfUBC13 structure, amino acid residues where the two sequences differ are highlighted in blue, or in cyan if the substitution is conservative. www.nature.com/scientificreports/ Further development of the rapamycin-treated parasite was monitored for up to five cycles (Fig. 2c). Parasites were still detectable in the culture although the parasitemia gradually declined. Analysis by PCR of DNA collected at cycles 3 and 5 confirmed that parasites contained the truncated gene, rather than representing parasites in which the gene was still intact (Fig. 2d). After the second cycle the rapamycin-treated parasites lost synchrony of development and some visible abnormalities were present in late stages, such as pyknotic and deformed cells ( Supplementary Fig. 2). These results indicate that ubc13 is not absolutely essential for completion of the first growth cycle following rapamycin treatment, perhaps because of residual active UBC13 protein. However, for the second and subsequent cycles the absence of growth and decrease in parasitemia, coupled with the abnormal morphology of many parasites, indicate that UBC13 is essential for asexual blood stage growth of P. falciparum. ubc13 deletion affects K63 polyubiquitylation in the parasite. We examined whether or not deletion of the putative ubc13 affected polyubiquitylation in the parasite. We used two antibodies, one specific for ubiquitin ubiquitylation at lysine-48 (K48) and the other for ubiquitin ubiquitylation at lysine-63 (K63). Parasites were treated with rapamycin or DMSO as described above and then extracts were prepared from schizonts for western blotting with the two antibodies (Fig. 2e). In a comparison of treated and control parasites, there was no clear difference or a slight increase in the pattern of K48-linked ubiquitylation, but the level of K63linked ubiquitylation was clearly diminished in the rapamycin-treated cells. In the control-treated cell lysate, six discrete bands were clearly visible, with molecular masses in the range of 32 to ~ 150 kDa and the intensity of at least five of these was reduced. These results indicate that the putative UBC13 is largely responsible for synthesis of K63-linked ubiquitin ubiquitylation in blood stage parasites of P. falciparum, and suggest that the PF3D7_0527100 gene product is a functional homologue of UBC13, which is responsible for K63-linked ubiquitylation in human cells.
ubc13 deletion increases parasite sensitivity to DNA damage induced by treatment with methyl methanesulfonate. In human cells UBC13 plays an important role in DNA damage repair 19,22 and we were interested to examine whether this protein has a similar function in P. falciparum. Therefore, we induced DNA damage in the parasite following rapamycin treatment, by incubation with MMS, a mutagen that alkylates DNA bases and induces single and double-strand breaks, and which has been shown to cause DNA damage in P. falciparum 30,33 . Parasites were exposed for different periods to either 2.25 mM or 4.5 mM MMS, washed to remove the drug and then cultured further to 48 h after initial invasion, at which point parasite survival was measured. The rapamycin-induced ubc13-iKO parasite was more sensitive to MMS than the DMSOtreated parasite. This was particularly evident at 2.25 mM MMS, since at the higher MMS concentration 40 min or more exposure resulted in no survival of either parasite population (Fig. 3a).
The effect of adding rapamycin and a lower concentration of MMS together in the first intracellular cycle was also investigated. Parasites were treated with and without 20 nM rapamycin and with and without 500 µM MMS for 48 h and then parasitemia was monitored for up to four cycles of growth. The parasites not treated with rapamycin were severely inhibited by MMS but started to recover in cycle 3 at 144 h. In contrast the rapamycin and MMS treated parasites were killed and no recovery over the period of the experiment was observed (Fig. 3b).
Pfubc13 deletion increases parasite sensitivity to dihydroartemisin (DHA) but not chloroquine (CQ). DHA generates oxidative free radicals that can cause DNA damage 30 . In contrast, CQ is not known to act in this way. Therefore, we examined parasite sensitivity to DHA and CQ, in mock treated or rapamycintreated ubc13 inducible knockout parasites. The results (Fig. 3c) revealed that the ubc13-iKO parasite is more sensitive to DHA than the control parasite, with a significantly decreased DHA IC 50 (1.69 nM and 4.1 nM for the rapamycin and DMSO treated parasites; P < 0.005). In contrast, the CQ IC 50 was not affected by ubc13 deletion (18.51 and 20.19 for the two treatments). These results suggest that PfUBC13 functions, at least in part, in the DNA damage repair system. Treatment of parasites with NSC697923 mimics the induced ubc13 knockout. NSC697937 is a UBC13 inhibitor in human cells and forms a covalent adduct with the active site cysteine 28 . Since NSC697937 has been shown to inhibit the asexual blood stage development of P. falciparum 29 , we wished to compare the outcome of rapamycin and NSC697923 treatment on the growth of blood stage parasites. Assessment of the percent parasitemia by FACS at the start of the experiment (2 h post invasion) and at time points corresponding to the first cycle (50 hpi) and second cycle (98 hpi) showed that the treated parasites showed significantly reduced parasite numbers ( Fig. 4a; DMSO treatment compared to either rapamycin or NSC697937 treatments, p < 0.0001). Giemsa stained thin blood smears were prepared to examine parasite development and morphology after treatment. Both NSC697923 and rapamycin treatment resulted in a growth defect. Both treatments resulted in slowed growth; for example, at 49 h post-invasion, treated parasites with either treatment were largely at the schizont stage, whereas the control parasites were predominantly at the ring stage (Fig. 4b).

Discussion
Human UBC13 (HsUBC13) specifically catalyzes the formation of K63-linked polyubiquitin chains, which are important in, for example, cytoplasmic NF-κB signaling and the DNA repair system. In P. falciparum, the role of UBC13 is unclear although PF3D7_0527100 appears to be an essential gene 26 and the protein's function is controlled by the activity of PK9 24,27 . In this study, we generated a transgenic parasite to allow the inducible knockout of this gene using the DiCre recombinase strategy, and examined the resultant phenotype.
The structural alignment of the PfUBC13 and HsUBC13, and the fact that deletion of Pfubc13 affected K63-linked ubiquitylation in the parasite, indicate that the protein's catalytic function mirrors that of UBC13 www.nature.com/scientificreports/  www.nature.com/scientificreports/ in higher eukaryotes. These data, together with crystal structure information for PfUBC13 (https:// www. rcsb. org/ struc ture/ 2r0j) and PfUBC13 complexed with PfUEV1a (https:// www. rcsb. org/ struc ture/ 3e95) indicate that PF3D7_0527100 is a homolog of UBC13 in higher eukaryotes. This is the first demonstration of the potential role of PfUBC13 in K63-linked polyubiquitylation in the parasite because the previous studies focused on PK9 and PK9 inhibitors 24,27 . The six visible K63 ubiquitylated substrates ranged in size from 32 to ~ 150 kDa with no evidence of heterogeneity, suggesting that the ubiquitylation is either a chain of just two or a small discrete number of ubiquitin units. These substrates remain to be identified, although they are likely present in the experimentally determined trophozoite/schizont ubiquitomes 29 . Interestingly, K63-polyubiquitylation of histone H2A 34 and proliferating cell nuclear antigen (PCNA) 35 is important in DNA repair processes, and both these proteins are present in the P. falciparum ubiquitome. In future work it may be possible to identify the proteins directly following their affinity purification with K63 polyubiquitin-specific antibody. Their identification may illuminate further the biological processes in which this modification is involved in the parasite. The slight increase in K48 ubiquitylation in rapamycin treated cells may reflect a loss of proteostasis, leading to enhanced ubiquitylation of proteins destined for degradation in the proteasome. Rapamycin-induced Pfubc13 deletion affected parasite growth severely. After one cycle the number of new ring stage parasites was reduced significantly but not completely; and this effect was amplified at the second cycle and beyond, although morphologically intact parasites were still present in the culture at cycle 5. These parasites did not have an intact ubc13 gene and some visible abnormalities were present in late stages, such as pyknotic and deformed cells. Some parasites could develop from ring to schizont stage, and apparently survive to the www.nature.com/scientificreports/ next cycle although growth was severely retarded. Parasite largely failed to develop into viable late stages, which is when DNA replication, nuclear division and merozoite differentiation occur, suggesting a role of Pfubc13 in maintaining normal DNA replication. K63-linked polyubiquitylation has been identified to play a role in DNA repair, signal transduction, and kinase activation in eukaryotes 5,6,34 , while K48-linked polyubiquitylation targets protein substrates to the proteasome. Therefore, we chose to examine the importance of PfUBC13 in DNA repair processes for P. falciparum. In Plasmodium spp., DNA damage has been studied experimentally and can result from exposure to UV irradiation; ionizing radiation like X-rays and gamma rays; chemical mutagens such as alkylating agents (e.g., methyl methanesulfonate and cisplatin) and also exposure to oxidative free radicals 33,36 . The damage includes interstrand cross-linking, DNA strand breaks, and inhibition of DNA synthesis and replication, leading to cell cycle arrest, and parasite death 37 . Genome integrity is maintained in the parasite by repair mechanisms, for example, homologous recombinant repair (HR), mismatch repair (MMR), nucleotide excision repair (NER) and base excision repair (BER). Canonical non-homologous end joining (C-NHEJ) repair is absent in Plasmodium, but an alternative inefficient NHEJ pathway has been described 38,39 .
We chose to use the DNA alkylating agent, MMS, which predominantly modifies guanine to N7-methyl guanine (7meG) and O6-methyl guanine (O6meG), to study the effect of ubc13 deletion on DNA repair. These lesions can lead to the collapse of the replication fork and subsequent induction of DNA double-strand breaks 40 that are commonly repaired by BER, NER and MMR processes 37,41 . Up-regulation of genes that function directly in DNA repair mechanisms such as PfRAD51 and PfRAD54, and changes in histone modification were reported after MMS treatment of P. falciparum 30,42 . These changes in chromatin modulate the access of repair factors to the damage site resulting in DNA repair 43 . The up-regulation of Pfubc13 transcription in the MMS-treated parasite has not been reported 30 . However, our results show that the Pfubc13-inducible knockout parasite has greater sensitivity to MMS treatment, and fails to survive under conditions in which the intact parasite recovers. These results are consistent with a major role for UBC13 and K63-polyubiquitylation in DNA repair in P. falciparum.
DHA is the active metabolite of artemisinin and related compounds, and its mode of action against the malaria parasite is complex 44,45 . One property, following activation, is its ability to alkylate macromolecules, and recent studies have suggested that one cellular process where this is important is in DNA damage and repair 30,31 . The Pfubc13-inducible knockout parasite had increased sensitivity to DHA, with a significantly decreased IC 50 , and showing a similar response profile to that with MMS, while CQ did not show this effect. This suggests that DHA is acting at least in part through mechanisms where repair is dependent on UBC13 and K63 ubiquitylation.
NSC697923 is an inhibitor of HsUBC13 function, reacting with the sulfhydryl group of the active site Cys87 to form a 5-nitrofuran adduct 28 . It has activity against P. falciparum growth in vitro 29 , and therefore it was of interest to compare mock and rapamycin-treated Pfubc13-inducible knockout parasites to parasites treated with this drug. Both treatments produced a similar result with a similar delay in parasite development. Although NSC697923 is not a potent inhibitor, the results suggest that if the primary target of NSC697923 in the parasite is UBC13 then better compounds with higher affinity, specificity, and a good therapeutic index, could be potential antimalarial compounds. Furthermore, we would expect that inhibitors of PfUBC13 could be used in combination to enhance the efficacy of artemisinin drugs.
In conclusion, our results demonstrate that the PF3D7_0527100 gene product is PfUBC13, a functional homologue of HsUBC13, which is responsible for K63-linked ubiquitylation and is involved in DNA repair systems. Deletion of this gene results in a P. falciparum growth defect and increases the sensitivity of the parasite to agents that cause DNA damage, in particular MMS and DHA.
In vitro culture and synchronisation of P. falciparum. The inducible DiCre recombinase-expressing P. falciparum line II-3 32 was cultured in RPMI 1640 medium supplemented with 1% (w/v) Albumax at 37˚C in gassed (5% CO 2 , 5% O 2 , 90% N 2 ) flasks. The parasite population was synchronized by schizont centrifugation onto a 63% Percoll cushion. Purified schizonts were cultured for 2 to 3 h to allow merozoite invasion of fresh red blood cells and ring stage formation, and then residual schizonts were lysed by treatment with 5% D-sorbitol for 10 min. RBCs were purchased from the United Kingdom National Health Service Blood and Transplant (NHSBT).

Generation of an inducible Pfubc13 knockout parasite.
To generate an inducible knockout of ubc13, we used CRISPR-Cas9 to insert two loxP sites flanking the last exon of the gene and the rapamycin-inducible active DiCre-recombinase system to excise the intervening DNA so that a truncated inactive protein would be produced, using the methodology described previously 29,32 . The Guide RNA sequence was designed to direct Cas9 to introduce a double-stranded break in Exon IV (Supplementary Fig. 1  www.nature.com/scientificreports/ restriction sites at the 5' and 3' end, respectively. The repair plasmid (50 μg) was digested with EcoRI and XhoI, ethanol precipitated together with 20 μg plasmid to express the guide RNA, and redissolved in 10 μl sterile TE (10 mM Tris-HCl 1 mM EDTA pH 8.0). For transfection, purified P. falciparum II-3 schizonts (10-40 µl) were mixed with DNA dissolved in 100 µl of AMAXA primary cell solution P3 and electroporated using an AMAXA 4D nucleofector with program FP158. Following electroporation, schizonts were transferred to 2 ml culture medium containing 300 µl of erythrocytes, and incubated with shaking at 37 °C for 30 min. A further 8 ml of culture medium was added and then 24 to 48 h post-transfection at 37 °C, 2.5 to 10 nM WR99210 was added to provide selection. The parasite population was screened for integration by PCR, using pairs of primers (Supplementary Fig. 1  Parasite growth and survival study. One hundred microliters synchronised ring-stage parasites (2 h PI, 0.2% parasitemia; 3% hematocrit), treated with 20 nM rapamycin or 0.01% DMSO, were placed in wells of a round-bottom 96-well plate at time zero. At the start and after specific periods (48 and 96 h) under normal culture conditions, parasite samples were fixed with 100 µl 4% paraformaldehyde + 0.1% glutaraldehyde for 1 h at room temperature (RT). The plates were centrifuged and fixative was replaced with 50 µl PBS and the parasites were stained with 2 × SyBR Green I in the dark for 30 min. One microlitre of stained cells was diluted into 3 ml PBS in a FACS tube and analysed using a BD LSRFortessa flow cytometer, with a 488 nm UV laser and fluorescence at 530 nm. Duplicate samples on triplicate plates were analysed for each time point. Data were analysed with one-way ANOVA analysis (Dunnett's multiple comparison test) using Graphpad software. At various time points thin blood smears were prepared, stained with Giemsa's reagent and examined by microscopy to evaluate parasitemia and parasite morphology.
Western blotting of parasite lysates. Parasites (3-4 h PI) were treated with 20 nM rapamycin or 0.01% DMSO for 40 h, then schizonts were purified by centrifugation over a 63% Percoll cushion, suspended in PBS containing 0.15% (w/v) saponin to lyse erythrocyte membranes, and collected by centrifugation at 5,000 g for 5 min. The cell pellet was dissolved in 10 volumes of lysis buffer (150 mM NaCl, 1% Triton X-100, 0.1% SDS, 1 μl/ml benzonase [Roche], and 1X complete protease inhibitors [Roche]). After incubation on ice for 20 min, samples were centrifuged at 17,000 g for 30 min. The protein content of the supernatant was measured using a BCA Protein Assay (Pierce), and 10 μg total protein was resolved on a 4 to 12% Tris-acetate PAGE gel (Invitrogen), transferred to nitrocellulose, and probed with specific antibodies. Protein ubiquitylation was detected with antibodies to ubiquitin Lys48-and Lys63-specific linkages (Merck-Millipore rabbit mAbs 05-1307 and 05-1308, respectively). Antibodies to BiP 29 were used as a loading control.
Parasite survival following treatment with the mutagen, methyl methanesulfonate. Parasites pretreated with rapamycin or DMSO for 24 h were treated with 2.25 mM and 4.5 mM MMS for 0, 20, 40, 60 and 100 min, washed twice with incomplete RPMI 1640 medium and finally returned to culture in complete RPMI 1640 medium. At 48 h after the start of the experiment, thin blood smears were prepared, stained with Giemsa's reagent and examined by microscopy to calculate percent parasite survival.
To examine parasite recovery from MMS induced-DNA damaged, ring stage parasites 4 to 6 h PI were adjusted to 0.5% parasitemia, 3% hematocrit and then treated for 48 h with 20 nM rapamycin or DMSO and with or without 500 μM MMS. The medium was changed every 2 days, and the percentage parasitemia calculated every day from examination of Giemsa-stained thin blood smears.
Parasite drug sensitivity. Synchronised ring stage parasites (0.05-1% parasitemia; 3% hematocrit) were treated with either DMSO alone or 20 nM rapamycin, together with serial dilutions (from 200 nM) of DHA and CQ as 100 µl final volume of culture in a black 96-well plate. Following incubation under normal culture conditions for 48 or 96 h, parasite survival was determined by using the malaria SYBR Green I-based fluorescence assay (MSF), as described previously 48 . Briefly, 100 µl of SYBR Green I in lysis buffer (0.2 µl of SYBR Green I/ml of lysis buffer [20 mM Tris-HCl, 5 mM EDTA, 0.008% saponin, and 0.08% Triton X100]) was added to each well, then the plate was mixed and incubated in the dark at RT for 1 h. The fluorescence was measured with excitation and emission at 485 and 530 nm, respectively [48][49][50] . The data were analysed and the EC 50 for each compound was calculated using GraphPad Prism software.

Data availability
All plasmids and transgenic parasites described in this study are available upon request.