The FHA domain protein ArnA functions as a global DNA damage response repressor in the hyperthermophilic archaeon Saccharolobus islandicus

ABSTRACT Forkhead-associated (FHA) domain proteins specifically recognize phosphorylated threonine via the FHA domain and are involved in signal transduction in various processes especially DNA damage response (DDR) and cell cycle regulation in eukaryotes. Although FHA domain proteins are found in prokaryotes, archaea, and bacteria, their functions are far less clear as compared to the eukaryotic counterparts, and it has not been studied whether archaeal FHA proteins play a role in DDR. Here, we have characterized an FHA protein from the hyperthermophilic Crenarchaeon Saccharolobus islandicus (SisArnA) by genetic, biochemical, and transcriptomic approaches. We find that ΔSisarnA exhibits higher resistance to DNA damage agent 4-nitroquinoline 1-oxide (NQO). The transcription of ups genes, encoding the proteins for pili-mediated cell aggregation and cell survival after DDR, is elevated in ΔSisarnA. The interactions of SisArnA with two predicted partners, SisvWA1 (SisArnB) and SisvWA2 (designated as SisArnE), were enhanced by phosphorylation in vitro. ΔSisarnB displays higher resistance to NQO than the wild type. In addition, the interaction between SisArnA and SisArnB, which is reduced in the NQO-treated cells, is indispensable for DNA binding in vitro. These indicate that SisArnA and SisArnB work together to inhibit the expression of ups genes in vivo. Interestingly, ΔSisarnE is more sensitive to NQO than the wild type, and the interaction between SisArnA and SisArnE is strengthened after NQO treatment, suggesting a positive role of SisArnE in DDR. Finally, transcriptomic analysis reveals that SisArnA represses a number of genes, implying that archaea apply the FHA/phospho-peptide recognition module for extensive transcriptional regulation. IMPORTANCE Cellular adaption to diverse environmental stresses requires a signal sensor and transducer for cell survival. Protein phosphorylation and its recognition by forkhead-associated (FHA) domain proteins are widely used for signal transduction in eukaryotes. Although FHA proteins exist in archaea and bacteria, investigation of their functions, especially those in DNA damage response (DDR), is limited. Therefore, the evolution and functional conservation of FHA proteins in the three domains of life is still a mystery. Here, we find that an FHA protein from the hyperthermophilic Crenarchaeon Saccharolobus islandicus (SisArnA) represses the transcription of pili genes together with its phosphorylated partner SisArnB. SisArnA derepression facilitates DNA exchange and repair in the presence of DNA damage. The fact that more genes including a dozen of those involved in DDR are found to be regulated by SisArnA implies that the FHA/phosphorylation module may serve as an important signal transduction pathway for transcriptional regulation in archaeal DDR.

damage. SisArnA inhibits the transcription of UV-induced pili genes (ups) via direct binding to the promoters of several ups genes together with SisArnB. The interac tions of SisArnA with SisArnB and SisArnE were stimulated by phosphorylation of the vWA proteins in vitro. In addition, during DDR, the interaction between SisArnA and SisArnB was reduced resulting in the removal of the transcriptional inhibition, while the interaction of SisArnA and SisArnE was enhanced. Finally, our transcriptomic analysis indicates that SisArnA inhibits dozens of genes, implying that it may function as a global repressor in the cell.

RESULTS
The SisarnA deletion strain exhibits higher resistance to NQO compared to the wild-type strain FHA domain-containing proteins in eukaryotes play critical roles in signal transduction during DDR and DSB repair (4). Gene annotation indicates that SiRe_1010 of S. islandicus REY15A encodes an FHA-containing protein which we name as SisArnA (24). A 70-amino acids flexible loop links the FHA domain and a Zinc-ribbon domain, which are located at the C-terminal and N-terminal end, respectively (Fig. 1A). The flexible loop contains multiple QQ motif and is conserved in Sulfolobales. Structural modeling using Alpha fold2 showed that the FHA domain of SisArnA was highly similar to those of bacterial and eukaryotic FHAs, with the former being a little closer both in the arrangement of loop/β-sheet and pThr binding residues (Fig. 1B).
To understand whether SisArnA is also involved in DDR and/or DNA repair in Sulfolobales, the SisarnA gene was deleted by genome editing using the endogenous CRISPR-Cas-based system of S. islandicus (Fig. S1), and the phenotype of the deletion mutant was analyzed. Firstly, we determined the growth of ΔSisarnA in the presence and absence of DNA-damaging agent 4-nitroquinoline 1-oxide (NQO), which can cause UV-mimic DNA lesions (1). As shown in Fig. 2A, there was no growth difference between the wild-type E233S and ΔSisarnA under normal conditions, suggesting that the generation time of E233S did not change after SisarnA deletion. Interestingly, ΔSisarnA grew faster than the wild-type strain in the medium supplemented with NQO. Moreover, cell viability analysis revealed that ΔSisarnA had a higher survival ratio compared with was performed by the AlphaFold2 program (25), which produces a per-residue confident score (pLDDT, 0-100) on the local Distance Difference Test (lDDT-Cα) (25). The model confidence is shown in color, and the regions below 50 pLDDT may be unstructured in solution. Four cysteine residues (Cys- 5, 8, 19, and 22)  Research Article mBio E233S in the presence of NQO (40.91% ± 3.02% vs 31.72% ± 1.39%, Student t-test P < 0.05), indicating that ΔSisarnA is more resistant to the DNA damage agent than the wild type. To further confirm the function of SisArnA, we constructed a SisArnA overexpression strain (E233S/pSeSD-N-His-SisArnA) and tested its growth in the presence of NQO when the protein is highly induced in the medium containing arabinose. As expected, the SisArnA overexpression strain exhibited higher sensitivity to NQO, while it grew similar to the wild type when the strain was grown in the sucrose-containing medium ( Fig. S2A and C). In addition, we also performed SisArnA complementation assay and showed that the complementation strain ΔSisarnA/pSeSD-His-SisArnA restored the NQO sensitivity compared with the deletion strain containing an empty vector (ΔSisarnA/ pSeSD) ( Fig. S2B and C). To know why the growth was affected, we performed flow cytometry analysis (Fig. 2B). It was shown that ΔSisarnA had less cells arrested in G1/S phase (1 copy of chromosome peak, 1C peak) than E233S at 6 and 9 h after NQO treatment. As time passed, the proportion of the DNA less cells (<1C, representing dead L, DNA-less cells; 1, cells containing one genome copy; 2, cells containing two genome copies. The ratios between the numbers of dead cells (boxed) and the total cell numbers at 6,9,21, and 27 h are indicated on the right in NQO-treated samples. (C) Cell aggregation analysis. Representative pictures are shown for the samples untreated or treated with NQO for 6 and 12 h. (D) Quantification of the results in (C). Aggregates containing more than three cells were counted. The values were calculated based on three independent experiments. One-tailed Student's t-test was used for the statistical analysis. Significant: *P < 0.05, **P < 0.01, ***P < 0.001, N.S., not significant.
Research Article mBio cells) in ΔSisarnA sample was less than that in E233S (31.5% vs 37.3% at 20 h, 28.8% vs 40.4% at 27 h). We find that ΔSisarnA also had slightly more cells with >2C peak than that for E233S in the presence and absence of NQO, which may contain some cell aggregates. Since no growth difference was observed between the wild type and ΔSisarnA in the absence of NQO, the aggregated cells in the sample of ΔSisarnA should be all alive. The result suggested that SisarnA deletion is beneficial for cell survival in the presence of the DNA damage agent. One of the best-studied DDR response in Sulfolobales cells is that they are prone to aggregation in the presence of DNA lesions caused by UV irradiation or NQO, by which DNA exchange occurs between cells (26)(27)(28). Cell aggregation is mediated by UV-induced pili, the components of which are encoded by a cluster of ups genes (26,27). To analyze the DDR response in ΔSisarnA, cell aggregation with or without NQO treatment was examined. We found that ΔSisarnA cells already aggregated at a higher ratio (11.7%) than E233S cells (4.4%) in the absence of NQO ( Fig. 2C and D). After the cells are treated with NQO for 6 and 12 h, the cell aggregation in ΔSisarnA increased further than that in the wild type (44.5% vs 21.6% at 6 h, 47.5% vs 26.7% at 12 h) ( Fig. 2C and D). However, at 24 h, the aggregation of ΔSisarnA was decreased to 13%, while that of the wild type was still around 30%. This indicated that the response of ΔSisarnA to DNA damage was faster and more efficient than that of the wild type.

Sisfha deletion derepresses the transcription of the DDR genes
It was demonstrated that there exists an Orc1-2-centered DDR network in Sulfolobus and Saccharalobus through which dozens of genes are regulated, facilitating DNA repair (15,26,29). These consist of up-regulated genes such as tfb3, ups, and ced which are involved in DNA damage regulation and DNA exchange as well as down-regulated genes which are involved in DNA replication and cell division (15,16). To test the effect of SisArnA on the transcription of DDR genes, RT-qPCR was employed to detect the transcription levels of orc1-2, tfb3, and upsX genes at the early log phase (OD 600 = 0.2-0.3). Orc1-2 activates tfb3 and ups which are also activated by Tfb3. Consistent with the growth and cell morphology of ΔSisarnA, the mRNA levels of all the three genes in ΔSisarnA were higher than those in E233S. After 6-h NQO treatment at OD 600 = 0.2-0.3, the mRNA levels of tfb3 and upsX further increased (Fig. 3A). However, the transcription of orc1-2 was not higher than E233S in the presence of NQO. In addition, we also analyzed the transcriptions of two down-regulated genes in Orc1-2-centered DDR network, orc1-1 and parA, which encode the proteins for DNA replication initiation and DNA segregation, respectively. But no difference was observed between E233S and ΔSisarnA (Fig. 3B). The results suggested that the phenotype of higher cell aggregation and more efficient response to DNA damage agent in ΔSisarnA was due to elevated transcription levels of ups. It may be also due to the higher transcription level of tfb3.
To confirm that SisArnA has a repressive role in the transcription of orc1-2, tfb3, and ups genes, we conducted a promoter activity assay based on the β-glucosidase activity system. The promoter regions of orc1-2, tfb3, and ups (upsA, upsE, and upsX) (200 bp upstream of their start codons) were used for the analysis. E233S and ΔSisarnA harboring pSe-Porc1-2/tfb3/upsA/upsE/upsX-SsolacS at the early log phase were treated or mock treated with NQO. The cells were collected at 6 h after treatment, and the β-galactosidase activity of cell extracts was measured. As shown in Fig. 3C, the promoter activities of orc1-2, upsA, upsE, and upsX in ΔSisarnA increased compared with those in E233S in both the presence and absence of NQO, while the promoter activity of tfb3 did not change significantly in ΔSisarnA. These results reinforce that SisArnA is a transcriptional repressor of the DDR genes, with the ups genes being the most strongly repressed. However, the results from the promoter activity assay for orc1-2 and tfb3 were not consistent with their qPCR data. SisArnA inhibited orc1-2 promoter activity but did not affect its mRNA in the presence of NQO implying that elevated downstream regulation of Orc1-2 in ΔSisarnA might be high enough for the DNA exchange and repair, and there may be a complicated regulation network through which the RNA level of Orc1-2 is reduced for turning off the DDR. The result that SisArnA inhibited tfb3 transcription but not its promoter region may suggest that the 200 bp promoter of tfb3 was not a main region targeted by SisArnA.

Genetic interaction analysis of SisarnA, SisarnB, and SisarnE
It was reported that SacArnA and SacArnB (homologs of SisArnA and SisArnB, respec tively, in S. acidocaldarius) interacted with each other and worked together in the regulation of archaellum (arl) genes which encodes the proteins for cell mobility (19)(20)(21). As interactions of SisArnA with both SisArnB and SisArnE have been detected in our previous study (18) and SisArnA is found to be involved in DDR, it is reasonable to ask whether SisArnB and SisArnE participate in DDR together with SisArnA. To answer this question, we constructed additional strains ΔSisarnB, ΔSisarnE, ΔSisarnAΔSisarnB, and ΔSisarnAΔSisarnE (Fig. S1). The sensitivities of these strains to NQO were analyzed with growth curves (Fig. 4). We found that ΔSisarnB exhibited slightly higher resistance to NQO than E233S within 30 h after treatment, while ΔSisarnE was more sensitive to NQO than E233S. In addition, the growth of ΔSisarnAΔSisarnB was better than that of the single mutants ΔSisarnA and ΔSisarnB in the presence of NQO, suggesting that SisArnA and SisArnB may work together for the regulation of DDR genes. SisArnA probably plays a major role in this process although other proteins or pathways may also be involved. On the other hand, the sensitivity of ΔSisarnAΔSisarnE was similar to that of the wild-type strain. Considering that ΔSisarnA and ΔSisarnE exhibited opposite responses to NQO and the proteins interact with each other, we speculate that SisArnE functioned as an activator during DDR in association with SisArnA.
To further confirm the phenotype of ΔSisarnB, the DNA content, cell aggregation, and transcription levels of the DDR genes were also analyzed. As expected, the ratios of DNA less cells (<1C) in ΔSisarnB were lower than those of wild type in the presence of NQO (Fig. 5A). We found that the proportion of cells in aggregation in ΔSisarnB increased at 3 and 6 h after NQO treatment compared with those of E233S although it did not change apparently in the absence of NQO (Fig. 5B). RT-qPCR analysis on the transcription of orc1-2, tfb3, and upsX genes revealed that the mRNA levels of tfb3 and upsX genes also increased compared with those in E233S, especially after NQO treatment for 6 h (Fig. 5C). These results indicated that SisArnB interacted with SisArnA and worked as an inhibitor complex during S. islandicus DDR. were obtained based on data from three independent experiments. Error bars indicate standard deviations. One-tailed Student's t-test was used for statistical analysis. Significant: *P < 0.05, **P < 0.01, ***P < 0.001, N.S., not significant.

Research Article mBio
We also analyzed the phenotype of ΔSisarnE. Consistent with the growth curve, ΔSisarnE had more DNA less cells than E233S (Fig. S4A). However, we found that the ratios of cells in aggregation in ΔSisarnE were higher than those in the wild type after NQO treatment at 6 and 12 h (Fig. S4B). This is consistent with the RT-qPCR result that the mRNA levels of tfb3 and upsX genes were further increased in ΔSisarnE compared with those in E233S in the presence of NQO (Fig. S4C). Notably, under normal growth conditions, the transcriptions of tfb3 and upsX were not significantly different between E233S and ΔSisarnE (Fig. S4C), suggesting that SisArnE did not repress the expression of these genes in vivo. The higher cell aggregation ratio and transcriptional levels of tfb3 and upsX genes in ΔSisarnE may be due to more lesions induced by NQO in this strain than E233S, leading to higher sensitivity to NQO.

The interaction between SisArnA and SisArnB was reduced but that with SisArnE was enhanced after DDR
Our previous proteomic study revealed that the phosphorylation levels of SisArnB and SisArnE were reduced after UV treatment for 30 min (18). Therefore, we speculated that the interactions of SisArnA with SisArnB and SisArnE may be inhibited in the presence of To determine this, we cultivated the complementation strains ΔSisarnB/pSeSD-Flag-ArnB and ΔSisarnE/pSeSD-Flag-ArnE and purified Flag-tagged SisArnB and SisArnE from their corresponding strains treated or mock-treated with NQO for 6 h. The interacted SisArnA pulled down by SisArnB or SisArnE was detected by Western blot with anti-SisArnA antibody. We found that the interaction of SisArnA-ArnB was stronger (about nine times) than that of SisArnA-ArnE under normal growth NQO for 6 h, and their total RNA was extracted for RT-qPCR with corresponding qPCR primers of orc1-2, tfb3, and upsX. The data were obtained from three independent experiments. Error bars indicate standard deviations. The comparative Ct value of tbp cDNA was used as a reference. One-tailed Student's t-test was used for statistical analysis. Significant: *P < 0.05, **P < 0.01, ***P < 0.001, N.S., not significant.
Research Article mBio conditions, indicating that SisArnB was a main interactor of SisArnA in vivo ( Fig. 6A and B). Strikingly, after NQO treatment, the amount of SisArnA pulled by SisArnB was reduced to 40% of the untreated sample, while that of SisArnA interacted with SisArnE increased to threefold of the untreated sample ( Fig. 6A and B). To confirm that phosphorylation would affect the interactions of SisArnA with SisArnB and SisArnE, we purified these proteins from Escherichia coli and performed in vitro pull-down assay in the absence or presence of the kinase ePK1 (SiRe_2056) (18) (Fig. S5A). The result showed that phos phorylation of SisArnB and SisArnE stimulated their interactions with SisArnA, respec tively (Fig. S5B). Since the phosphorylation levels of SisArnB and SisArnE were reduced at 30 min after UV treatment in our previous study (18), we could not exclude that their phosphorylation would be changed at the later stage of DDR, resulting in different interaction capabilities with SisArnA. The different responses of the interactions of SisArnA-ArnB and SisArnA-ArnE were in agreement with the contrasting phenotypes of ΔSisarnB and ΔSisarnE, further suggesting that SisArnB was a repressor, whereas SisArnE was a positive regulator after DDR.

SisArnA in combination with SisArnB exhibited higher DNA-binding activity toward the promoter DNA
To reveal how SisArnA and SisArnB inhibit the transcription of orc1-2, tfb3, and ups genes, electrophoretic mobility shift assay (EMSA) was performed to detect the DNA-binding activity of SisArnA, SisArnB, and SisArnE on the promoters of these DDR genes. His-SisArnB, His-SisArnE, and untagged SisArnA were heterologously expressed and purified from E. coli (Fig. S5A). FAM-labeled promoter DNA of orc1-2, tfb3, and upsX was used as substrates. All SisArnA, SisArnB, and SisArnE were unable to bind any promoter individually (Fig. S6A). Interestingly, DNA binding was observed in the presence of both SisArnA and SisArnB, or SisArnA and SisArnE, although the activity was low (Fig. S6A). Next, we detected the activities of the proteins purified from S. islandicus. His-SisArnA, Flag-SisArnB, and Flag-SisArnE were purified from their corresponding overexpression strains, individually (Fig. S5A). Surprisingly, the EMSA assay revealed that the SisArnB sample exhibited DNA-binding activity on all the three promoters (Fig. 7A), different from the protein from E. coli. Because SisArnB interacted with SisArnA more strongly than that of SisArnE in vivo, we have detected that a few amount of SisArnA was co-purified together with Flag-SisArnB from S. islandicus (Fig. 6A). Therefore, we speculated that To confirm this, Flag-SisArnB was purified from an overexpression strain with the arnA deletion background (ΔarnA/pSeSD-Flag-ArnB). As expected, this sample without ArnA was unable to bind the three promoters at the same concentrations ( Fig. S6B; Fig.  7A and B). Then, all three promoters were examined, and the result showed that the SisArnB sample had higher affinity to the upsX promoter (P upsX ) compared to the other two promoters (Fig. 7A and B). Although SisArnA did not exhibit DNA-binding activity toward different promoters, the addition of SisArnB in the reaction with the SisArnA sample further enhanced the DNA-binding capability ( Fig. 7A and B), consistent with the EMSA result using proteins purified from E. coli. To detect the specificity of DNA-binding ability by SisArnA-SisArnB, the FAM-labeled promoter of SiRe_1719 was used for the DNA-binding assay. SiRe_1719 encodes the histidyl-tRNA synthetase, and its transcription was not affected by either NQO treatment or SisarnA deletion [(15) and Table S3]. We found that the promoter P SiRe_1719 could also be bound by SisArnA and SisArnB comparable with the promoters of orc1-2, tfb3, and upsX (Fig. 7C). In addition, a DNA competition assay was performed using unlabeled promoters of orc1-2, tfb3, upsX, and a gene fragment (sire_0197). We found that the bound FAM-labeled P orc1-2 , P tfb3 , and Research Article mBio P upsX can be replaced with their corresponding unlabeled promoters (P orc1-2 , P tfb3 , or P upsX ), while the gene fragment of SiRe_0197 was unable to compete with P orc1-2 and P upsX and only partially compete with P tfb3 in the EMSA assay (Fig. 7D), suggesting that SisArnA-SisArnB exhibited higher DNA-binding activity toward promoter sequences. The lower binding activity of SisArnA-SisArnB on P tfb3 was consistent with the promoter activity assay (Fig. 3C) showing that the selected 200 bp promoter region of tfb3 may not be the main target of SisArnA-SisArnB. Unexpectedly, we found that the SisArnE sample from S. islandicus degraded these DNA substrates, and the degradation was inhibited by the addition of EDTA (Fig. S6C). Our results indicated that SisArnA-SisArnB inhibited the transcription of orc1-2 and ups genes by directly binding to their promoters. We found that the DNA-binding activity of SisArnA-ArnB purified from S. islandicus was stronger than those from E. coli. Given the fact that SisArnB was phosphorylated at multiple residues in its host cell, the higher DNA-binding activity should be due to the stronger interaction of SisArnA-ArnB than the proteins purified from E. coli, which was not phosphorylated or in low phosphorylation (18). As expected, Western blot analysis demonstrated that the phosphorylation level of the SisArnB sample purified from S. islandicus was much higher than SisArnB purified from E. coli, in which the phosphoryla tion can be removed by the phosphatase PP2A (Fig. S6D). To further confirm this, SisArnB, purified from E. coli, was phosphorylated by ePK1 in vitro for EMSA. We found that the phosphorylated SisArnB exhibited slightly higher DNA-binding activity in the presence of SisArnA, compared to its un-phosphorylated form (Fig. S6E).

Transcriptomic analysis of SisarnA deletion in the presence or absence of NQO
To further understand whether there are other genes or pathways regulated by SisArnA, we conducted transcriptomic analysis of SisarnA deletion mutant treated with NQO or without treatment compared with those of the wild-type E233S. The culture samples were taken at 6 h after treatment, and total RNA was isolated for reverse transcription and sequencing. It was shown that 72 genes were up-regulated and eight genes were down-regulated after SisarnA deletion. Consistent with our RT-qPCR results, the transcripts of tfb3 and the ups genes significantly increased (tfb3 2.23-fold, upsA 3.46-fold, upsB 2.66-fold, and upsE 3.39-fold) ( Table 1). However, no transcriptional change was observed for orc1-2. We found that the transcriptional levels of several arl genes, which encode for Sulfolobus archaella components, were also slightly increased (SiRe_0121 2.95-fold, SiRe_0122 1.87-fold, and SiRe_0124 1.88-fold) (Table S3) in agreement with the previous study that S. acidocaldarius FHA protein ArnA inhibited the biosynthesis of Sulfolobus archaella (19). Strikingly, a dozen of putative DDR genes highly induced in the presence of NQO was also up-regulated in the untreated ΔSisarnA (Table 1). In addition, multiple up-regulated genes were significantly enriched in several metabolic pathways, including oxidative phosphorylation, TCA cycle, butanoate and pyruvate metabolism, and glycolysis/gluconeogenesis, in ΔSisarnA (Fig. S7). Moreover, there were three gene clusters transcriptionally increased, including those involved in terpenoid backbone biosynthesis (SiRe_1459-1462) and sulfur metabolism (SiRe_2307-2309 and SiRe_2310-2313) (Table S3). SiRe_1459 in the cluster SiRe_1459-1462 was annotated as a 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase which was supposed to be involved in terpenoid backbone biosynthesis. Terpenoid compounds, such as sterols, carotenoids, or the prenyl groups of various proteins, are synthesized via the mevalonate pathway (30). A rate-limiting step of this pathway is the conversion of HMG-CoA to mevalonic acid catalyzed by the HMG-CoA reductase (31). It was reported that the activity of HMG-CoA may affect various biological processes including cellular adaption to environmental changes (32,33). Another two clusters, SiRe_2307-2309 and SiRe_2310-2313, encode proteins participating in sulfur metabolism. The sulfur oxidation and reduction are critical metabolic pathways for this model strain living in the environments containing sulfur (34,35). Therefore, all these genes up-regulated may stimulate cell growth under certain stress.
To analyze the putative effect of SisArnA on the Orc1-2-dependent DDR network, we compared the transcriptomic changes of E233S_NQO vs E233S with those of ΔSisarnA_NQO vs ΔSisarnA. As previously reported, NQO treatment triggered the Orc1-2-centered DDR network which induced orc1-2, tfb3, ups, ced, dpo2, herA-mre11-rad50-nurA operon for DNA exchange and repair and inhibited those genes involved in DNA replication, chromosome segregation, and cell division (Table 1). Nearly, all of these genes and those with unknown functions were changed in the same way in ΔSisarnA_NQO vs ΔSisarnA (Table 1), indicating that the major DDR was not affected by SisarnA deletion. In addition, the transcript levels of tfb3 and ups genes were further increased in ΔSisarnA_NQO vs ΔSisarnA compared to those in E233S_NQO vs E233S (Table 1), consistent with our RT-qPCR results (Fig. 3A). However, a major part (about  . 8; Table S3). Since the main DDR in the fundamental processes was unaffected and a number of the DDR genes were transcriptionally increased in ΔSisarnA in the absence of DNA damage, we speculated that since the response and recovery of ΔSisarnA to DNA damage was faster than the wild type, it was not necessary to further change the transcript levels of other genes such as those involved in cellular metabolism for the DDR.

DISCUSSION
FHA domain is a specific module for phospho-Thr recognition which is found in many eukaryotes. The signal transduction roles of FHA domain proteins were extensively studied in eukaryotes, especially those involved in DDR for signal transduction and enzymatic activity (4). In addition, there is a family of transcription factors, containing both FHA and DBD domains, participating in multiple pathways in addition to transcrip tion regulation. For example, the well-studied FHA proteins Fkh1 (forkhead homolog Research Article mBio 1) and Fkh2 of Saccharomyces cerevisiae regulate the cell cycle by targeting various genes in different cell cycle phases (36,37). Moreover, Fkh1 and Fkh2 also play a transcription-independent role in regulating DNA replication origin activation and yeast mating-type switch, via interaction with multiple proteins mediated by its FHA domain (38)(39)(40). Therefore, Fkh proteins participate in multiple pathways through interactions with target DNA and various proteins via their DNA binding and FHA domains. The bacterial FHA domain is fused with diverse domains and only widely distributed in three main groups: the mycobacteria and their relatives, the cyanobacteria, and the Gram-neg ative proteobacteria together with the Chlamydias (41). The FHA proteins in bacteria also exhibited diverse functions, including glutamate and lipid production, regulation of cell shape, type III and Type VI secretion, sporulation, and host-bacterium interactions. Nearly, all studies showed that FHA proteins regulate the activities of target proteins via protein-protein interaction (42)(43)(44)(45). We found that FHA proteins are only conserved in Crenarchaeota and occasionally identified in other phyla in archaea. As previously reported, the FHA domain might originate in eukaryotes after the separation of archaeal and eukaryotic lineages and move to certain prokaryotes by horizontal gene transfer (2). And, the functions of FHA domains evolved diversely together with their fused domains in a late stage of evolution.
Here, we report that the FHA protein from S. islandicus (SisArnA) regulated gene transcription and interact with several proteins. Interestingly, our study revealed that SisArnA worked as a repressor for a number of DDR genes together with its partner SisArnB. It was revealed that SisArnB was phosphorylated at multiple residues in vivo (18), which would mediate the interaction with SisArnA. Consistently, we found that complementation of wild-type SisArnA, but not pThr-binding-deficient mutant R134A/ S148A, inhibited cell viability in the presence of NQO ( Fig. S2B and C). EMSA analysis revealed that both proteins in combination display DNA binding activity, indicating that the interaction between SisArnA and SisArnB mediated by phosphorylation is essen tial for the transcriptional inhibition of the DDR genes. In addition, SisArnA harbors a Zn-ribbon domain containing four cysteines (Cys) at its N-terminus (Fig. 1). This domain often functions as a nucleic acid-binding module (46). However, we found that DNAbinding activity of SisArnA was undetectable (Fig. S7A), indicating that the Zn-ribbon itself could not bind DNA. This is different from the FHA domain-containing eukaryotic transcription factors which can directly bind DNA via its DBD for their roles in replication initiation and transcription regulation (39). The function of Zn-ribbon domain in SisArnA needs further investigation. It is also interesting to investigate how SisArnA and SisArnB in combination exhibited DNA-binding activity, especially toward promoter sequences.
In addition to SisArnB, another vWA domain protein SisvWA2, which we designate as SisArnE in this study, was also one of the interactors of SisArnA in vivo. We found that SisArnA-SisArnB interaction was stronger than that of SisArnA-SisArnE interaction under normal growth conditions (Fig. 6). However, the interaction between the SisArnA and SisArnE homologs in S. acidocaldarius was not detectable (19), implying a species-specific model, or the interaction was too weak to be detected. Our genetic work revealed that, in contrast to SisArnB, SisArnE acts as a positive regulator for genes involved in DDR since its deletion reduced cell viability in the presence of NQO. ΔSisarnA grew better, while ΔSisarnA/SisarnE exhibited similar growth compared with that of the wild-type strain after NQO treatment. In vivo pull-down showed that SisArnA-SisArnB interaction was reduced, while SisArnA-SisArnE interaction was enhanced in the presence of DDR (Fig.  6), indicating that SisArnA and SisArnE might be involved in a same process after DDR occurred. One potential mechanism was that SisArnE interacted strongly with SisArnA and released the repression of SisArnA-SisArnB on some DDR genes in the presence of DNA damage. However, we cannot exclude that SisArnA-SisArnB and SisArnA-SisArnE may function on different groups of genes. Thus far, we did not know whether SisArnA-SisArnE could bind certain DDR gene promoters because the DNA substrates were degraded in the EMSA assay with the SisArnE sample purified from S. islandicus. Since SisArnE purified from E. coli did not degrade DNA and the vWA domain is a universal scaffold for PPI (47), the DNA degradation activity of the SisArnE sample from S. islandicus might be due to certain SisArnE partner in vivo. The specific function of SisArnE in DDR needs further investigation.
The transcriptomic analysis of SisarnA deletion mutants revealed that a number of genes, including many DDR genes, were up-regulated, indicating that SisArnA was a global inhibitor in S. islandicus. Although most of these genes were up-regulated less than sixfold in ΔSisarnA, the increased levels of these DDR genes might deal with DNA damage and regulate cell cycle more quickly, resulting in increased cell viabil ity, as reported for the phenotype of Orc1-2 overexpression strain (15). According to our data, the Orc1-2-centered DDR network was not affected in ΔSisarnA after NQO treatment. Moreover, in the previous report, the transcription level of SisarnA did not change apparently with either NQO treatment or orc1-2 deletion (15). These imply that phosphorylation-dependent regulation mediated by SisArnA might be in parallel with the Orc1-2-centered transcriptional regulation. Furthermore, we also analyzed the phenotypes (cell growth, cell aggregation, and transcription of the DDR genes) of ΔSisarnA with methyl methanesulfonate, which resulted in similar phenotypes with those in the presence of NQO (Fig. S8). On the other hand, although we did not have cell mobility data on ΔSisarnA, our transcriptomic result is in agreement with the previous study showing that S. acildocaldarius FHA protein inhibited the transcription of arl genes and regulated cell mobility during starvation, which is totally different environmental conditions from the ups pili (19,20). It was reported that deletion of the genes encoding the adhesive pilus resulted in induction of the archaellum operon in S. acidocaldarius (48), implying that there is a complex regulatory network for different archaeal surface structures in which protein phosphorylation may be involved. Together, these indicate that archaeal FHA proteins might function in the regulation of various pili responding to various cellular stresses.
In conclusion, we establish that S. islandicus FHA domain protein SisArnA is a global repressor of various genes, many of which were involved in DDR. Based on our results, we propose a model for DDR in S. islandicus (Fig. 9). In the presence of DDR, in addition to the Orc1-2-centered regulatory network, certain signals may induce a Arn-depend ent regulation mediated by protein phosphorylation. The inhibition of a number of DDR genes is mediated by SisArnA-ArnB interaction under normal growth conditions. When DNA damage occurs, the interaction of SisArnA-ArnB is reduced by PP2A-medi ated dephosphorylation of SisArnB, whereas SisArnA-ArnE interaction is enhanced via phosphorylation of SisArnE possibly by SisePK1. Then, the repression of the DDR genes by the SisArnA-ArnB complex is removed by ArnB dephosphorylation, or the transcrip tion of the DDR genes is stimulated through SisArnA-ArnE interaction, facilitating DNA exchange/DNA repair in vivo and ensuring cell survival.

Construction of plasmids for gene knockout by the CRISPR-Cas system
The plasmids for the gene knockout were constructed based on the vector pGE (from Prof. Qunxin She's lab) (49). Two complementary ssDNA of the protospacers (40 bp) within the target genes were synthesized by BGI (Beijing Genomics Institute, Beijing, China) and annealed to each other. The resulting protospacer DNA was inserted into pGE between two repeat sequences, yielding pGE-Sp. The L-arm and R-arm for recombination to delete the target gene were amplified and joined via splicing by overlap extension PCR and inserted into the SphI and XhoI sites of pGE-Sp after the restriction enzyme digestion and purification. The sequences of PCR primers are listed in Table S2.

Construction of plasmids for protein expression
The plasmids for expression of S. islandicus ArnA, ArnB, and ArnE using pSeSD (50) were as described previously (18). The vectors pET15b and pET22b were applied for the construction of the plasmids expressing SisArnA and C-terminal His-tagged ArnB and ArnE, respectively, in E. coli. The genes were amplified using their corresponding primers (Table S2). The PCR products were digested with their corresponding restriction enzymes and ligated into the same restriction sites in pET15b or pET22b.

DNA damage agents sensitivity assay
For drug treatment in a liquid medium, the strains were inoculated and cultured to the logarithmic phase for three times before treatment. DNA damage agent 4-nitroquinoline 1-oxide (NQO, 3 µM) was added into a 30-mL aliquot of each strain at an OD 600 = 0.2. These flasks were then incubated at 75°C. The OD 600 values were measured every 6 or 12 h thereafter. The growth curves were derived based on data from three biological repeats. For the cell viability assay, cultures treated with NQO for 6 h were collected and diluted properly. Each aliquot of 500-1,000 cells was spread on plates and cultivated at 75°C for 5-7 days. The survival ratio was calculated by the number of colonies from NQO-treated sample divided by that from the untreated sample. The data were obtained from three independent experiments.  (15). Another pathway revealed from this study is regulated via phosphorylation. SisArnA-ArnB interaction mediated by SisArnB phosphorylation represses a number of DDR genes. In the presence of DDR, the phosphorylation of SisArnB, a repressor, is eliminated by the phosphatase PP2A, resulting in decreased SisArnA-ArnB interaction. At the same time, the SisArnA-ArnE interaction, which might be mediated by SisePK1 phosphorylation, is enhanced resulting in further derepression of SisArnA-ArnB. By Orc1-2 and phosphorylation-mediated processes, the DDR genes encoding for proteins in DNA exchange/ repair are up-regulated, ensuring genome integrity maintenance and cell survival. In addition, Orc1-2 may also be regulated by protein phosphorylation.

Flow cytometry
The Saccharolobus cells were collected and treated with 70% ethanol. Flow cytometry analysis was performed as described previously (51).

Microscopy analysis and cell aggregation formation assay
The wild-type strain and ΔSisarnA were cultivated at an initial OD 600 = 0.04. After overnight cultivation, the samples at the early log phase (OD 600 = 0.2) were taken and observed under a Nikon Eclipse80i Microscope (Nikon Corporation, Tokyo, Japan) or Nikon Eclipse Ti-E Inverted Microscope. The cell diameters were measured by the software NIS-Elements AR 3.1. Cell aggregation ratio was calculated from images taken with the inverted microscope. Cell clusters containing more than three cells were considered cell aggregation. More than 500 cells were counted for each sample. One-tailed Student's t-test was used for statistical analysis on the software GraphPad Prism 5.

Promoter activity assay
The promoter activity of DDR genes was assayed according to the previous study (52). Briefly, the gene of S. solfataricus β-glycosidase (SsolacS) was amplified from its genome and inserted into the multiple cloning sites (MCS) of pSeSD. The promoter regions (200 bp upstream of start codons) of target DDR genes were amplified and inserted into the SphI /NdeI site of pSeSD-SsolacS to replace the arabinose promoter, yielding the reporter plasmids. The plasmids were transformed into S. islandicus E233S or ΔSisarnA individually, and three single colonies were picked for each strain. They were grown in MTSV (Mineral salt medium + Typtone + Sucrose + Vitamin) for 6 h in the presence or absence of NQO. Cells were collected and lysed by sonication. The β-glycosidase activity in the cell extracts of different S. islandicus strains was determined using substrate β-nitrophenyl-b-D-galactopyranoside (ONPG) as described previously (52). One-tailed Student's t-test was used for statistical analysis on the software GraphPad Prism 5.

RT-qPCR
Total RNA was isolated by the Trizol agent. Briefly, cells were resuspended with 1 mL Trizol and vortex. After incubation for 5 min at room temperature, 200 µL chloroform was added to remove proteins. Then, RNA was isolated from the mixture by centrifugation with 13,000 × g for 15 min at 4°C and precipitated with an equal volume of isopropa nol. RNA was washed with 70% ethanol and dissolved into RNase-free water. Reverse transcription was performed using EvoM-MLV RT Mix Kit (Accurate Biotechnology Cp., Ltd, Hunan, China). Firstly, genomic DNA was removed by incubating the RNA with five gDNA Clean Reaction Mix at 42°C for 2 min. First-strand cDNAs were then synthesized at 37°C for 15 min in a mixture containing EvoM-MLV RTase, RNase inhibitor, dNTPs, oligo dT(18T) Primer, and Random 6 mers Primer, followed by 85°C for 5 s. The mRNA levels of DDR genes were estimated by the cDNA sample using SYBR Green Premix Pro Taq HS qPCR Kit (Accurate Biotechnology Cp., Ltd, Hunan, China). The primers for qPCR were listed in Table S2. PCR was performed in CFX Connect Real-Time PCR Detection System (Bio-Rad, Hercules, CA, USA) with the following steps: denatured at 95°C for 30 s, 40 cycles of 95°C for 5 s, and 60°C for 30 s. The comparative Ct value of tbp cDNA was used as reference. One-tailed Student's t-test was used for statistical analysis on the software GraphPad Prism 5.

Protein purification
The plasmids for protein expression were transformed into E. coli BL21(DE3)-Codonplus-RIL. Escherichia coli harboring pET15b plasmids was cultivated in 1 L of LB medium containing ampicillin (100 µg/mL) and chloramphenicol (34 µg/mL). Proteins were induced by the addition of IPTG and incubation at 37°C for 4 h. The cells were collected by centrifugation at 10,000 × g for 3 min and disrupted by sonication in lysis buffer (50 mM Tris-HCl pH 8.0, 200 mM NaCl, 5% glycerol). The His-tagged proteins were purified with Ni-NTA agarose (Invitrogen) column and pooled for further purification by gel filtration with Superdex 200 increase 10/300 Gl column (GE Healthcare, Boston, MA, USA). Flag-tagged proteins were purified with Flag magnetic beads (GE Healthcare, USA) pre-equilibrated with buffer A (50 mM Tris-HCl pH 7.5, 150 mM NaCl, 5% glycerol). Proteins were eluted with Gly-HCl buffer (0.1 M Glycine pH 3.0 adjusted with HCl). The protein concentration was determined by the Bradford method with bovine serum albumin as the standard, and proteins were frozen by liquid nitrogen for storage at −80°C.
For protein purification in S. islandicus, the strains harboring corresponding expres sion plasmids were transferred in 1 L of Sulfolobus medium without sugar (MTV). D-arabinose was added into the culture when it reached OD 600 = 0.2. The cells were cultivated for further 12-16 h before collected by centrifugation at 5,000 × g for 5 min. His-tagged and Flag-tagged proteins were purified as described above.

Pull-down assay
For in vitro pull-down assay, indicated amounts of SisArnA (Flag-tag) and its putative interacting proteins (N-His) were mixed in buffer A and incubated at 65°C for 30 min. The mixture was then mixed with 100 µL of Ni-NTA beads (Life Technologies, Carlsbad, CA, USA) pre-equilibrated with buffer A and incubated at room temperature for 10 min by gently shaking. Unbound protein was removed by centrifugation at 3,000 × g for 5 min. After being washed with 400 µL of wash buffer (buffer A supplemented with 40 mM imidazole) for four times, the His-tagged protein and its putative interacted protein were eluted with 200 µL of elute buffer (buffer A supplemented with 250 mM imidazole). The fractions were analyzed by SDS-PAGE and Western blot. For the pull-down assay with phosphorylated proteins, SisArnB and SisArnE were incubated with the eukaryotes-like protein kinase SiRe_2056 (SisePK1) in a reaction containing 25 mM Tris-HCl pH 8.0, 25 mM NaCl, 5 mM MgCl 2 , 1 mM DTT, 10 mM ATP at 65°C for 30 min before applied for pull-down assay.
For the in vivo pull-down assay to detect the interactions of SisArnA with SisArnB and SisArnE, 300 mL of the complementation strains ΔSisarnB/pSeSD-Flag-ArnB and ΔSisarnE/pSeSD-Flag-ArnE was cultivated to OD 600 = 0.2-0.3. The cells were treated with 3 µM NQO for 6 h before collection. The cells were resuspended in 20 mL buffer A for sonication. The soluble fractions (20 mL, input) were subjected to Flag magnetic beads for purification as described above. The eluted fractions (10 mL) were concentrated by 10 times. Twenty microliters of input and eluted samples was analyzed by Western blot with anti-Flag or anti-SisArnA antibodies. The gels were imaged with Amersham ImageQuant 800. The protein polyclonal antibody against SisArnA was raised in rabbits by Dai-An Biotechnology Co., Ltd (Wuhan, China) according to their standard procedure. The His-tagged SisArnA used for immunization was expressed and purified from E. coli BL21 as described above. The experiments were performed for three times independ ently. The ratios of SisArnA/SisArnB and SisArnA/SisArnE were quantified according to three independent experiments.

Electrophoretic mobility shift assay
The fragments of the DDR gene promoters (200 bp upstream from the start codon of the DDR genes) were amplified using the corresponding primers and inserted into EcoRI and HindIII of pUC19, yielding pUC19 plasmids harboring DDR promoters. FAM-labeled substrates were obtained via PCR with the primers FAM-pUC19-MCS-F/pUC19-MCS-R and various pUC19 plasmids as templates. EMSA was performed in 20-µL reaction mixture containing 25 mM Tris-HCl pH 8.0, 25 mM NaCl, 5 mM MgCl 2 , 1 mM DTT, 5% glycerol, FAM-labeled dsDNA (200 bp), and various concentrations of the target proteins. The mixture was incubated at 37°C for 30 min and analyzed in 6% or 8% native PAGE as indicated. The gels were imaged by Amersham ImageQuant 800 (Cytiva, North Logan, UT, USA).

In vitro kinase and phosphatase assay
The assays for protein phosphorylation and dephosphorylation in vitro were performed as described previously (17) with slight modification. Briefly, ATP was used for reaction, and phosphorylation signals were detected by Western blot with anti-pThr antibody (# 9381S; Cell Signaling Technology Inc., Danvers, MA, USA).

Transcriptomic analysis of SisarnA deletion strain
The wild-type and SisarnA deletion strains were grown in an MTSV medium. When the cultures reached OD 600 = 0.2-0.3, 3 µM NQO was added into the cultures. After incubation for further 6 h, the cells were collected by centrifugation (5,000 × g, 10 min). Total RNA extraction and RNA sequencing of the samples were carried out as described previously (53). The adjusted P-adjust (0.05) and log2 (fold change) of 1 were set as a threshold for significantly differential expression based on three independent cultures of each strain.