Skip to main content
NCBI home page
NAR Molecular Medicine logoLink to NAR Molecular Medicine
. 2026 Feb 4;3(1):ugag010. doi: 10.1093/narmme/ugag010

Altered transposon element-derived genes distort oxygen-free radical scavenger systems in FXD

Tamaki Suganuma 1,✉, Huzaifa Hassan 2, Selene K Swanson 3, Sunil Laxman 4, Jerry L Workman 5
PMCID: PMC12936792  PMID: 41768973

Abstract

The altered expression of transposon element (TE)-derived genes that regulate immune responses implied the involvement of previous viral infections in the onset of Fragile X disorders (FXD), often bearing sustained inflammation. Here, we discovered that hypoxia greatly modifies the expression of TE-derived genes that act as oxygen-free radical scavengers, resulting in an alkylating environment under normoxia in FXD. Molybdenum cofactor synthesis 3 (MOCS3) and selenocysteine lyase (SCLY) stabilize the association of Kelch-like ECH-associated protein 1 (KEAP1) with PGAM family member 5 (PGAM5). This rewires connections between the oxidative stress response, cysteine-sulfur and selenium metabolism, and oxidoreductases. Desensitized tRNA thiolation against hypoxia in FXD suggests that the rewired radical scavenging system induces resistance to ROS in FXD.

Graphical Abstract

Graphical Abstract.

Graphical Abstract

Open in a new tab

Introduction

Fragile X syndrome (FXS) is the most common cause of intellectual disability and autism spectrum disorder (ASD) [1, 2]. FXS arises from > 200 CGG repeats expansion (‘full mutation’) in FMR1 (Xq27.3), resulting in the absence or deficiency of FMR1 protein [1, 2]. There is a spectrum of disorders associated with the FMR1 premutation (55–200 CGG repeats) [1], leading to Fragile X-associated tremor/ataxia syndrome (FXTAS) and fragile X-associated primary ovarian insufficiency (FXPOI). These three disorders are referred to as Fragile X-associated disorders (FXD) [3]. The prevalence of FXS in males is higher than in females; therefore, the penetrance of FXS does not accommodate the X-linked dominant pattern [4]. Immune-mediated disorders are over-represented in FXS patients and the patients carrying the FMR1 premutation [5, 6]; however, the causes of these phenotypes remain to be elucidated.

Previously, we found that DNA-alkylation-damage signaling is impaired in Lymphoblastoid Cell Lines (LCLs) of FXD by disruption of the molybdopterin synthase associating complex (MPTAC) or dissociation of MPTAC from the Ada2a-containing histone acetyltransferase complex (ATAC) [7]. MPTAC, in which MOCS2 and MOCS3 are representative subunits, is required for catabolism of sulfur containing amino acids [8]. Importantly, the activity of xanthine oxidase (XO), which is converted from xanthine dehydrogenase (XDH) upon inflammation or low O2 tension [9], is increased in FXD compared to the healthy donors [7]. Therefore, inflammation and aerobic restriction are sustained in FXD patients [7]. Recently, we found that the expression of specific transposon element (TE)-derived genes regulating molecular oxygen signals, purine metabolism, and immune responses against viral infections, is altered in FXD [10]. In addition, an increased C G composition in those TE-derived genes has suggested an involvement of genes from past viral infections that are latently expressed as activated transposons in the pathogenesis of FXD [10].

Sulfur metabolism plays a pivotal role in redox homeostasis [11]. The sulfur atom in sulfur-containing amino acids, cysteine and methionine, is reduced to thiols/sulfhydryl [12]. Cysteine is catabolized by two pathways, oxidative cysteine catabolism and the H2S pathway [11] (Fig. 1A). The terminal products of the oxidative pathway are taurine and sulfite. The sulfite, which impairs mitochondrial bioenergetics with induced reactive oxygen species (ROS) and has neurotoxicity, is further oxidized to sulfate by sulfite oxidase (SUOX) in a molybdenum cofactor (Moco) dependent manner [11].

Figure 1.

Figure 1.

Open in a new tab

TE-derived genes with changed expression upon hypoxia in healthy cells greatly overlap with those genes changed in FXD compared to healthy donors under normoxia. (A) Diagram of cysteine and selenium metabolism. GSR, glutathione reductase; GPXs, Glutathione peroxidases; SUOX, sulfite oxidase; TXNRD, thioredoxin reductase; SEPHS2, Selenophosphate synthase 2; SCLY, selenocysteine lyase. (B) Diagram of the sample comparisons in this study. (C) Venn diagrams show the numbers of TE-derived genes, genes with splicing changes, and genes with changed expression levels (DE genes) via comparison between healthy donor cells cultured under normoxia and FXD cells cultured under normoxia in males (Venn11) and females (Venn12). TE-derived genes with DE and changed splicing or non-TE-derived genes (normal genes) with DE and changed splicing were compared between normoxia treatment and 1% hypoxia treatment for 22 h in healthy donor cells (Venn1 [males] and Venn3 [females]) or in FXD cells (Venn2 [males] and Venn4 [females]) in each sex. (D) Weighted Venn diagrams show the number of TE-derived genes with changed expression (DE and splicing) upon hypoxia in healthy donors and TE-derived genes with changed expression under normoxia in FXD (versus healthy donors) in each sex (Supplementary Table S2).

A thiol or a thiolate group of cysteine residues in proteins forms a disulfide (-S-S-) [12] and is a binding site of metals [13]. The thiol/ thiolate group is also reactive for alkylation by donating electrons to electrophiles of cysteine and for oxidation by ROS and nitrogen species [12]. S-Alkylation of cysteine residues regulates redox signaling [13]. For example, an endogenous electrophile binds to cysteine residues of the Kelch like ECH associated protein 1 (KEAP1), playing crucial roles in redox signaling [14, 15]. In addition, thiol-containing small molecules, including glutathione (GSH) and coenzyme A, are important for redox-mediated cellular processes [12].

The chemistry of cysteine and selenocysteine is similar [12] (Fig. 1A). However, selenium is a stronger nucleophile and can react with ROS faster than sulfur, but the oxidized selenium compound is much more readily reduced compared to oxidized sulfur compounds [16]. However, it is unknown whether sulfur and selenium metabolism are coordinated.

In humans, Moco biosynthesis proceeds in four steps. It begins with the conversion of GTP to cyclic PMP (cPMP) by MOCS1, and cPMP is then converted to molybdopterin (MPT) by insertion of two sulfur atoms catalyzed by MPT synthase, which consists of a heterodimer of MOCS2A and MOCS2B. The MPT moiety is sequentially adenylated, generating MPT-AMP that is subsequently used for molybdate (Mo) insertion catalyzed by Gephyrin, yielding mature Moco [17, 18]. After MPT synthase transfers the two sulfurs to cPMP, the E1-like domain (adenyltransferase domain) of MOCS3 (UBA4) adenylates MOCS2A, and the sulfur bound to a Rhodanese homology domain (RHOD) of MOCS3 is transferred to MOCS2A, reactivating MPT synthase [19, 20]. Notably, MOCS3 plays a key role in tRNA thiolation [21, 22].

Modification of the wobble position uridine 34 (U34) of tRNA has been shown to respond to a variety of stresses, including oxidative stress and DNA damage agents [23]. The levels of U34 thiolation, mcm5s2u34, represent sulfur consumption and modulate amino acid codons [24]. Low levels of mcm5s2u34 resulted from a limited supply of methionine/cysteine decrease carbohydrate metabolism and increasing methionine/cysteine/lysine synthesis [24]. For tRNA U34 thiolation in eukaryotes, sulfur from Nsf1 is transferred to MOCS3, which adenylates the C-terminus of URM1 and sequentially thiocarboxylates URM1. Then, URM1 further donates sulfur to uridine at U34 of tRNA, generating mcm5s2u34 [21, 25]. It has been studied that the enzymes for tRNA U34 thiolation, ELP3 or CTU1 and/or CTU2, directly regulate the translation of HIF1A mRNA and maintain high levels of HIF1α protein [21, 25]. Moreover, reduced mcm5s2u34 formation has been found to be involved in familial dysautonomia and bronchial asthma [26]. However, the involvement of tRNA thiolation in neurodegenerative diseases like FXD is largely unknown.

Here, we investigated the causal relationships between hypoxia and expression of TE-derived genes to understand the causes of sustained inflammation in FXD.

Materials and methods

Ethics approval

Human lymphoblastoid cell lines (LCLs) from B-lymphocytes from FXS and apparently healthy donors used in this study, provided by the Coriell Institute for Medical Research, were commercially available. Human HEK293 cells are commercially available at the American Type Culture Collection (ATCC). Examinations of the LCLs and HEK293 cells in this study were approved by the Institute Biosafety Committee (IBC) in the National Institute of Health (NIH) office through Stowers Institute for Medical Research.

Examination of knockdown HEK293 cells

HEK293 cells transfected with target gene-specific CRISPR-Cas9 expression clones were cultured in DMEM-High Glucose (4500 mg/L) (D6429, SigmaAldrich, Missouri, USA) with 10% fetal bovine serum and 500 μg/mL Geneticin (Life Technologies) at 37°C for 21 days. A part of the knockdown cells was treated with 1% hypoxia for 20 h. The cells were lysed in a 5x volume of ice-cold Extraction Buffer I (10 mM 4-(2-Hydroxyethyl)-1-piperazine ethanesulfonic acid (pH 7.5), 1.5 mM Magnesium chloride, 10 mM Potassium Chloride, 1 mM Dithiothreitol, and proteinase inhibitors) for 10 min, followed by incubation with 1% Nonoxynol-40 for 5 min. Cytoplasmic extracts were obtained from supernatants after centrifugation at 2655 g for 5 min [27]. The pellets were then incubated with a 3x volume of Extraction Buffer II (20 mM 4-(2-Hydroxyethyl)-1-piperazine ethanesulfonic acid, pH 7.5, 25% (v/v) glycerol, 0.42 M Sodium chloride, 1.5 mM Magnesium chloride, 10 mM Potassium Chloride, 1 mM Dithiothreitol, and proteinase inhibitors) for 1 h. Nuclear extracts were obtained from supernatants after centrifugation at 20 817 g for 20 min. The extracts were examined by Western blotting.

Target gene-specific CRISPR-Cas9 expression clones (GeneCopoeia):

-3 × sgRNA/Cas9 all-in-one expression clones targeting:

MOCS3 (NM_014484.3) (HCP207660-CG01-3)

LOC84661 (NM_032574.2) (HCP220858-CG01-3)

Examinations of human lymphoblastoid cell lines (LCLs)

LCLs from B-lymphocytes from FXS and apparently healthy donors listed in Supplementary Table S1 were provided by the Coriell Institute for Medical Research and were cultured in RPMI1640 containing 2 mM L-glutamine and 15% fetal bovine serum at 37 °C under 5% CO2. Details are described on the website below. A part of the knockdown cells was treated with 1% hypoxia for 22 h. Cytoplasm and nuclear extracts were obtained as described above. The extracts were examined by Western blotting with antibodies against MOCS3 (NBP2-17320, Novus Biologicals), HSD17B10 (MABN2416, MilliporSigma), KEAP1 (60027–1-Ig, Proteintech), PGAM5 (sc-515880, Santa Cruz), eEF2 (2332, Cell Signaling Technology), SCLY (10667–1-AP, Proteintech), TXNRD1 (67728–1-Ig, Proteintech), and eIF3β (sc-374156, Santa Cruz). Total RNA in human lymphoblastoid cell lines (LCLs) was extracted by TRIzol (15 596 026, Invitrogen/ThermoFisherScientific). The gel-purified tRNA from total RNA was analyzed for tRNA thiolation, as described below. Information on the cell lines is provided on the Coriell Institute website: https://www.coriell.org/0/Sections/BrowseCatalog/DiseaseDetail.aspx?PgId=403&omim=MEN30955&coll=Ribonucleoside (mcm5s2U) Analysis using a HPLC/MS2 approach

Experimental methods and materials

HPLC solvents/reagents and nucleoside standards (RNs) were purchased commercially with the highest quality possible.

RNA hydrolysis and dephosphorylation

Isolated tRNA was hydrolyzed and dephosphorylated using benzonase, phosphodiesterase, and alkaline phosphatase as described [28].

HPLC/MS2 analysis (QE-plus)

The analysis was performed using a Q Exactive™ Plus Hybrid Quadrupole-Orbitrap™ Mass Spectrometer coupled to an UltiMate 3000 RSLCnano HPLC system. Ribonucleosides were separated on an ACQUITY UPLC CSH C18 column (1.7 µm particle size, 1 mm ID × 100 mm, Waters) maintained at 42°C with a flow rate of 35 μL /min. A 40-min gradient elution was applied using Mobile Phase A (5 mM ammonium acetate in water) and Mobile Phase B (5 mM ammonium acetate in methanol). The gradient profile was as follows: 0–5 min at 0% B, 12–14 min at 32% B, 15–21 min at 99% B, followed by a 19-min re-equilibration at 0% B.

The mass spectrometer was operated in positive ion mode using a Full MS/dd-MS² method with an inclusion list targeting 18 specific ribonucleosides. The scan range was set to 70–1050 m/z with a normalized collision energy (NCE) of 30%. Parallel ddMS2 was performed only on selected ions specified in the inclusion list.

Raw data were imported into Skyline (version 24.1.0.199) along with a transition list for interpretation and quantification. Calibration curves were generated using standard RNs with serial dilutions in triplicate. Calculated R-squared for individual calibration curve was >0.95 with analyte concentration ranging from 0.2 to 25 ng/ μL. Peak area intensity was used to determine the relative abundance of mcm5U and mcm5s2U, normalized against the peak intensity of inosine (I) in each dataset.

RNA sequencing

For short-read RNA sequencing, total RNA from LCLs was ribo-depleted, and stranded 100-base single-end reads were generated on a NextSeq 2000 (n = 3). RNA-seq reads were demultiplexed into fastq format, allowing up to one mismatch using Illumina’s bcl2fastq2 (v 2.20). Subsequently, the reads were aligned to the GRCh38 reference genome from Ensembl using STAR (v 2.7.3a) [29]. The gene model retrieved from Ensembl, release 110, was used to generate gene read counts.

For long-read RNA sequencing, total RNA was used for RNA library preparation for each synthetic poly-A tail (n = 2) and normal RNA library (n = 2). The RNA was directly sequenced as a single read on the PacBio RS Sequel- II-3000 provided by Cold Spring Harbor Laboratory. The reads from each sample were aligned to the human genome (hg38).

Statistical information

For differential expression analysis using short read RNA sequencing (Illumina) of ribo-depleted RNA from LCLs, the gene model retrieved from Ensembl, release 110, was used to generate gene read counts. The transcript abundance TPM (Transcript per Million) was quantified using RSEM (v 1.3.0) [30]. To correct for batch effects arising from samples prepared and sequenced at two different time points, the sva package (v 3.48.0) [31] in R was used. Differentially expressed genes were determined using the R package edgeR (v 3.36.0) [32] after filtering low-expressed genes with a “CPM” (counts per million) of 0.5 in at least one library. The resulting P-values were adjusted with the Benjamini–Hochberg method using R function p.adjust. Genes with an adjusted P-value less than 0.05 and a fold change of at least 2 were considered as differentially expressed. Gene functional enrichment analysis for LCLs was performed using a custom script built over the R package “clusterProfiler” (v 4.4.4). Venn diagrams were generated using the Python package “pyvenn,” and Weighted venn diagrams were generated using python package “matplotlib”.

Previously identified transposable elements using PacBio long reads were aligned with all the samples [10].

Results and discussion

Expression of TE-derived genes responding to hypoxia has been modified in FXD under normoxia

We recently found that altered expression of specific TE-derived genes in FXD leads to an alkylating environment resembling sustained viral infection [10, 33]. To explore the mechanism, we examined the expression of TE-derived genes in LCLs from FXD patients and healthy donors in normoxic or hypoxic conditions for 22 h by RNA-sequencing. To find TE-derived genes, we examined the total RNA and mRNA (polyA selection) in all LCLs cultured under normoxia by long-read RNA sequencing (PacBio) as done previously [10]. We searched annotated TE-derived genes using the Dfam database [34] in individual cells cultured under normoxia and then looked for them within healthy donor cells (TAS10/42 in males and TAS11/43 in females) and FXD cells (TAS18/20 in males and TAS2/19 in females) cultured under normoxia or hypoxia (Fig. 1B and C). Human transposons are known to contribute to diseases, including cancer and genetic disease, through perturbation of genome functions such as transcription and RNA splicing [35]. Hence, we compared the differential expression levels (DE) and altered splicing (using rMATS (v 4.1.1) [36]) of all the genes by short-read RNA-seq (Illumina) between FXD cells and healthy donor cells in each sex (Fig. 1B).

We first compared the genes with changed expression (DE and splicing) to TE-derived genes in FXD compared to healthy donors under normoxia in each sex [10] (Fig. 1C Venn11 and 12). We next compared all genes with changed expression upon hypoxia versus normoxia in healthy donors (Fig. 1C Venn1 and 3) and FXD (Fig. 1C Venn2 and 4) separately. We further compared TE-derived genes with changed expression upon hypoxia in healthy donor cells (Fig. 1C Venn1 “51, 22, and 508”/Venn3 “106, 40, and 370”) to TE-derived genes with changed expression under normoxia in FXD cells (Fig. 1C Venn11 “18 and 61/ Venn12 “11(A) and 81”) in each sex based on the 4venns in Fig. 1C (Fig. 1D). We found that more than half of TE-derived genes with changed expression under normoxia in FXD were also the TE-derived genes with changed expression upon hypoxia in healthy donor cells (Fig. 1D “44” and “57”, Supplementary Table S2), indicating that expression of those TE-derived genes is altered by a latent hypoxic environment in FXD.

To better understand the changes in the expression levels of TE-derived genes by hypoxia, we compared the TE-derived genes with DE upon hypoxia in healthy donors (Fig. 1C Venn 1/3) to TE-derived genes with DE or without any changes (DE and splicing) upon hypoxia in FXD (Fig. 1C Venn 2/4). We then compared the TE-derived gene with DE upon hypoxia in each FXD cell versus healthy donor cells (Fig. 2A and B). We found that the majority of TE-derived genes with DE upon hypoxia in healthy donors overlapped with the TE-derived genes with/without changes upon hypoxia in FXD in each sex (Supplementary Fig. S1A and S1B). The expression levels of TE-derived genes coding virus suppressor (RSAD2), immune responses/ inflammatory responses (IL10/19, MSR1, ROBO2), and reductase (CYBB) were greatly reduced; however, methionine sulfoxide (MSRB3) and SARS-CoV-2 associating protein (TLCD4) were robustly increased upon hypoxia in FXD (TAS18/20/19) compared to healthy donors (TAS42/11) (Fig. 2A and B, Supplementary Table S3). Remarkably, the expression levels of TE-derived genes regulating production of immunoglobulins and cytokines (LAIR1, BTNL9, and HLA-L) increased upon hypoxia in healthy cells but not in FXD cells (Fig. 2A and B, Supplementary Table S3).

Figure 2.

Figure 2.

Open in a new tab

Expression of TE-derived genes regulating virus responses and sulfur and selenium metabolism is changed in FXD. (A and B) Scatter plots display a log fold change between two groups of TE-derived genes with changed expression levels (DE), highlighted on the plots. The highlighted genes are from different sections of the 2venn diagrams (Supplementary Fig. S1A and B), which correspond to males (Supplementary Fig. S1A): Venn1: 51, 22, 76, 66, 12, and 9; Venn2: 1093, 1681, 415, 62, 31, 87, 43, 30, and 9; and females (Supplementary Fig. S1B): Venn3: 106, 4, 190, 88, 47, and 16; Venn4: 825, 1616, 493, 109, 53, 225, 119, 72, and 22 (Fig. 1C, Supplementary Table S3). (C and D) Venn diagrams show numbers of splicing changed TE-derived genes upon hypoxia in healthy donors (Venn1 and Venn3 in Fig. 1C), splicing changed TE-derived genes upon normoxia and 1% hypoxia in FXD (Venn2 and Venn4 in Fig. 1C), TE-derived genes without changed expression upon hypoxia in FXD, and TE-derived genes with changed expression under normoxia in FXD (versus healthy donors) (Venn2, Venn4, Venn 11, and Venn 12 in Fig. 1C, Supplementary Table S4). (E) Dot plots display GO analysis clustered by biological process (BP) for genes with changed transcript (DE and splicing) under normoxia in male FXD versus upon 1% hypoxia in male healthy donors (males venn “35” in Fig. 1D). The top 20 processes based on the lowest P-value are listed (Supplementary Table S5). (F and G) Weighted Venn diagrams show the numbers of normal genes (F) or TE-derived genes (G) with changed expression (DE and splicing) upon hypoxia in healthy donors (Venn1 versus Venn3) and in FXD (Venn2 and Venn4), compared between males and females.

We next compared TE-derived genes with splicing changes between FXD and healthy donors. Interestingly, 415 (in males, Fig. 2C) and 352 (in females, Fig. 2D) TE-derived genes with changed splicing upon hypoxia in healthy donors were unchanged upon hypoxia in FXD. This indicates that a latent hypoxic environment has already changed the splicing of these TE-derived genes in FXD, so there was no further change upon hypoxia in these cells. Notably, splicing of selenocysteine lyase (SCLY) was changed in male FXD in normoxia (Fig. 1A and C Venn11 “61,” and 2C “15,” Supplementary Table S4). Splicing of selenoprotein S/T (SELENOS/T) was changed upon hypoxia in both FXS and healthy donors in males (Figs 1A and 2C, Supplementary Table S4). In females, splicing of genes encoding selenoproteins, thioredoxin reductase 1 (TXNRD1) and SELENOT, was changed in FXD and healthy donors upon hypoxia (Figs 1A and 2D, Supplementary Table S4). Moreover, splicing of pyruvate dehydrogenase kinase 1 (PDK1), which regulates the sulfur metabolic process, was changed under normoxia in FXD and upon hypoxia in healthy donors in both sexes (Fig. 2C and D, Supplementary Table S4). PDK1 has been shown to phosphorylate pyruvate dehydrogenase (PDH) to suppress its activity, resulting in decreased pyruvate levels in mitochondria. Hence, PDK1 inhibition induces oxidative phosphorylation and also increases ROS [37]. Notably, splicing of HIF1A was changed upon hypoxia in FXD in both sexes and male healthy donors (Fig. 2C and D, Supplementary Table S4). However, splicing of hypoxia inducible factor 1 subunit alpha inhibitor (HIF1AN) was not changed upon hypoxia in female FXD (Fig. 2D, Supplementary Table S4). Additionally, splicing of VHL was changed upon hypoxia in FXD and healthy individuals, and splicing of BNIP3 was not further changed upon hypoxia in FXD. VHL and BNIP3 have been shown to be transcriptionally activated by HIF1, and their products induce mitophagy [38] (Fig. 2C and D), suggesting that mitophagy does not respond to hypoxia in FXD cells. Notably, the expression of TE-derived genes, which regulate tRNA modification and tRNA metabolic processes were changed under normoxia in FXD and upon hypoxia in healthy donors. These include TP53RK binding protein (TPRKB) and tRNA isopentenyltransferase 1 (TRIT1) (Fig. 2C and D, Supplementary Table S4). Splicing of tRNA mitochondrial 2-thiouridylase (TRMU) was also changed upon hypoxia in healthy donors but not in FXD (Fig. 2D, Supplementary Table S4).

Through GO term-analysis corresponding to the Venn diagrams in Fig. 1D, we found that expression of TE-derived genes, which function in the catabolism of amyloid precursor protein (ADAM17 and CLN3) and interferon production (DHX33 and IRF5), was commonly changed in normoxia-treated FXD cells and hypoxia-treated healthy donor cells in males (Fig. 1D “44” and Supplementary Fig. S1C, Supplementary Tables S2 and S5). Thus, expression of TE-derived genes regulating redox signaling through both sulfur and selenium metabolism, mitophagy, anti-viral systems, inflammatory responses, and tRNA modifications, including tRNA thiolation, is modified in FXD, enabling maintenance of a hypoxic environment in FXD. These alterations occurred in healthy donors upon hypoxia, suggesting consequences of latent virus expression in FXD.

TE-derived genes distinguish sex specific responses

We found that expression of TE-derived genes regulating spermatid development, including CCNBIP1 and YTHDC2, and selenium metabolic enzymes, including SCLY, was changed in male FXD cells (Fig. 1D males “35”, Fig. 2C and E, Supplementary Tables S2 and S5). In addition, expression of MSRB1, which encodes methionine sulfoxide reductase B1 selenoprotein that reduces methionine-R-sulfoxides to methionines, was changed upon hypoxia in male FXD versus male healthy donors (Supplementary Fig. S1D). The reduction of MSRB1 levels was TE-dependent in a part of male healthy (TAS42) and male FXD (TAS20) (Fig. 2A). Previous studies have shown that deletion of the selenoprotein, glutathione peroxidase 4 (GPX4), in the epididymis results in sperm that are unable to fertilize oocytes, causing male infertility in mice [39]. The high expression of selenogenes: GPX4, SELENOV, and TXNRD3 in the testis has suggested their importance in male reproduction [40]. Therefore, males may be sensitive to redox status, particularly selenium metabolism.

In females, expression of TE-derived genes encoding proteins that catalyze dealkylation/demethylation (JMJDC1 and KDM6A) was changed under normoxia in FXD cells and upon hypoxia in healthy donor cells (Fig. 1D female “57,” Fig. 2D, and Supplementary Fig. S1E, Supplementary Tables S2 and S5). We compared all genes with changed expression upon hypoxia between males and females. Interestingly, the ratio of common populations of normal genes with changed expression upon hypoxia in male and female (Fig. 2F “1368” and “1776”) overlapped much more than the ratio of common populations of TE-derived genes with changed expression in male and female (Fig. 2G “107” and “115”) in both healthy donors and FXD, suggesting that TE-derived genes are biased towards sex specific responses to hypoxia. Remarkably, expression of TE-derived genes associated with disulfide oxidoreductase activity, such as CCS, PDIA5, and QSOX1, was significantly changed in male FXD upon hypoxia (Fig. 2G “546” and Supplementary Fig. S2A, Supplementary Table S5). Conversely, expression of TE-derived genes related to dehydrogenases acting on carbon-carbon bonds, such as ACAD9 and BDH2, and NAD-dependent enzymes, including SIRT3, was significantly changed in female FXD upon hypoxia (Fig. 2G “451” and Supplementary Fig. S2B, Supplementary Tables S5). Thus, redox pathways using sulfur and selenium are sensitive in males, and these pathways, regulated by TE-derived genes, are disrupted in male FXD. By contrast, redox-active nucleotides/coenzymes, NAD/NADH and FAD/FADH2, appear to be preferentially used for antioxidation in response to hypoxia in females; however, pathways using these coenzymes are changed by alterations of TE-derived genes in female FXD. Perturbing these redox/antioxidation systems may lead to sustained viral expression and hypoxia resistance in FXD. It remains unclear whether these differences in preferentially used redox systems relate to the higher occurrence with greater severity of FXS in males [4].

Altered sulfur metabolism causes hypoxia resistance of tRNA thiolation in FXD

Sulfur transfer by MOCS3 is crucial for tRNA thiolation27. MPTAC is disrupted by loss of some subunits, including MOCS3, in some FXD cells [7]. In this study, we found that expression of TE-derived genes regulating tRNA modification and sulfur metabolism was changed in FXD (Fig. 2C and D, Supplementary Figs S1C and S2A). Therefore, we monitored tRNA thiolation in FXD. We first examined MOCS3 protein expression in LCLs from FXD and healthy donors treated with hypoxia for 22 h or normoxia. We did not observe obvious differences in MOCS3 protein expression in any cells under normoxia (Fig. 3A and B). However, upon hypoxia, the expression levels of full-length (FL) MOCS3 (50.6 KDa) were significantly reduced, and smaller sizes of MOCS3 (Fig. 3B * and **) were observed in all cells. FL and two shortened MOCS3 products were recognized by the MOCS3 antibody, suggesting that MOCS3 was cleaved upon hypoxia (Fig. 3A and B). Importantly, the overall expression levels of MOCS3 (FL and cleavages) in TAS18 and 19 FXD cells were higher than those in others (Fig. 3B). We also found that the expression of HSD17B10, which is a subunit of MPTAC and functions in the TCA cycle, was totally suppressed upon hypoxia in all cells (Fig. 3B).

Figure 3.

Figure 3.

Open in a new tab

tRNA thiolation levels are not affected by hypoxia in some FXD. A. Diagram of MOCS3. B. Cytoplasmic extracts of indicated LCLs treated with (+) or without (-) 1% hypoxia for 22 h were analyzed by western blots. eIF3β was the loading control. C. Graph shows relative abundance of mcm5s2u34 normalized with mcm5u34 and inosine levels in indicated LCLs (n = 4–11). Two-tailed P-values compared between TAS42/18/20/39 and TAS10 or TAS43/2/19/21 and TAS11 treated with normoxia, or hypoxia are indicated (Supplementary Table S6). D. Graphs show the proportion of relative abundance of averaged mcm5s2u34/ mcm5u34/ inosine (C) by analyzing the percentage of one to the numbers in normoxic condition in each sample. Two-tailed P-values compared between normoxic condition versus hypoxic conditions in each sample are indicated in parentheses.

We next used mass spectrometry to measure tRNA mcm5s2u34 (U34 thiolation) levels normalized with unthiolated mcm5u34 and inosine levels in each LCL from FXD (TAS18/20/39 in males and TAS2/19/21 in females, Supplementary Table S1) and healthy donors (TAS10/42 in males and TAS11/43 in females) cultured under normoxia or hypoxia. We found that mcm5s2u34 levels in TAS20 were lower than TAS10/42, while mcm5s2u34 levels in TAS2/21 were higher than TAS11/43 under normoxia (Fig. 3C). The levels of mcm5s2u34 decreased upon hypoxia treatment compared to those treated with normoxia in healthy donor cells (TAS10/42/11/43) and some FXD cells (TAS20/2/21) (Fig. 3C). Unexpectedly, mcm5s2u34 levels upon hypoxia were similar to those upon normoxia in TAS 18/39/19 (Fig. 3C and D). Thus, mcm5s2u34 did not appear to respond to hypoxia in TAS18/39/19. These data suggested that sulfur consumption was reduced upon hypoxia, decreasing tRNA thiolation in healthy donors and TS20/2/21 FXD. However, sulfur consumption did not appear to be reduced upon hypoxia in TAS18/39/19 FXD versus healthy donors, implying that the sulfur-relay, including tRNA U34 thiolation, is resistant to hypoxia in TAS18/39/19 FXD. It is also possible that cleaved MOCS3 under hypoxia accommodates sulfur-relay in the process of mcm5s2u34 formation, and increased expression of cleaved MOCS3 might facilitate formation of mcm5s2u34 in TAS18/19. It has been shown that methylation of tRNA U34 by FtsJ RNA 2′-O-methyltransferase 1 (FTSJ1) promotes selenocysteine insertion at the UGA stop codon, leading to UGA recording during translation [41]. This methylation induces oxidative stress resistance in melanoma [41]. What is interesting about tRNA U34 is how its thiolation and methylation are coordinated under the alkylating environment.

Preventions of KEAP1-PGAM5 interaction in the nucleus by MOCS3 are diminished in FXD

KEAP1 serves as an adaptor of ubiquitin E3 ligase14,15. Previous studies found that KEAP1 is associated with PGAM5 under the low levels of ROS [42, 43]. PGAM5 is a mitochondrial serine/threonine protein phosphatase. Under high levels of ROS, dissociation of KEAP1 from PGAM5 leads to PGAM5 binding to apoptosis-inducible factor (AIFM1) in mitochondria, inducing oxeiptosis/mitochondria-associated cell death [43]. However, oxidized KEAP1 remains associated with PGAM5 and prevents PGAM5 proteolysis upon oxidative stress, eliciting mitophagy in healthy mitochondria [44]. Furthermore, it was suggested that PGAM5 is cleaved and released from mitochondria to translocate into the nucleus during mitophagy [45]. Importantly, proinflammatory cytokines and morbidity have been shown to be increased in PGAM5-knockout mice in response to influenza A virus [42]. This suggests that the role of PGAM5 in mitophagy is important for the prevention of an exaggerated immune response upon virus-mediated ROS.

The MOCS3 E1-like domain and RHOD domain play roles in sulfur transfer and ubiquitin-like conjugation of MOCS3 [46] (Fig. 3A). Therefore, we examined whether KEAP1 is associated with sulfur-dependent redox signaling by analyzing proteins that co-purify with endogenously expressed MOCS3 in the cytoplasm of the LCLs. We find that under normoxia, KEAP1 strongly co-purified with MOCS3 in healthy donor (TAS42) and more weakly from FXD (TAS18) (Fig. 4A lanes 5 and 6). KEAP1 was slightly detected in the purified MOCS3-associated proteins in TAS19 under normoxia (Fig. 4B lane 6). We did not observe KEAP1 in purified MOCS3 treated with hypoxia (Fig. 4A and B, lanes 7 and 8). Thus, MOCS3-bound KEAP1 might be more stable than unbound KEAP1, and truncated MOCS3 might not stabilize KEAP1. We next examined KEAP1-IP in the cytoplasm of the LCLs. We found that MOCS3 (FL) interacted with KEAP1 in TAS42/43 healthy donors and TAS20/2/19 FXD under normoxia (Fig. 4C lanes 5 and 7, and 4D lanes 5–7). These interactions were greater in FXD cells (TAS20/2/19) than in healthy donor cells (TAS42/43) in each sex (Fig. 4C lanes 7 versus 5, and 4D lanes 6 and 7 versus 5). By contrast, PGAM5 (both FL and cleaved) interacted with KEAP1 in only TAS20/2/19 FXD under normoxia (Fig. 4C lane 7 and 4D lanes 6 and 7). Deficiency of Eukaryotic translation elongation factor 2 (EEF2) through its diphthamide modification has been shown to reduce selenocysteine insertion into selenoproteins [47]. We found that EEF2 interacted with KEAP1 in TAS42/43/20/2/19 in normoxia and TAS42 in hypoxia (Fig. 4C and D). However, these interactions were not observed in normoxia-treated TAS18 and hypoxia-treated TAS43, as immunoprecipitated KEAP1 was greatly reduced in these cells (Fig. 4C lanes 6 and 8, and 4D lane 8). Hence, the interactions of KEAP1 with MOCS3, PGAM5, and EEF2 were enhanced in TAS20/2/19 FXD and were lost in TAS18 FXD compared to healthy donors.

Figure 4.

Figure 4.

Open in a new tab

KEAP1-MOCS3 interactions are modified in FXD. (A and B) Elution from MOCS3 immunoprecipitants (IP) in cytoplasmic extracts from indicated LCLSs treated with (+) 1% hypoxia for 22 h or normoxia (−) were examined by Western blot. IgG immunoprecipitants were used as a control. eIF3β was used as the loading control. LCLs from FXD patients are indicated with blue labels (A–D). (C–E) KEAP1-IP in cytoplasmic extracts of LCLs from male (C) and female (D) FXD patients/ healthy donors and in nuclear extracts from indicated knockdown HEK293 cells (E) were examined by Western blot.

We next examined whether MOCS3 influences the interactions of KEAP1 with PGAM5 in the nucleus. We did not observe the association of PGAM5 with KEAP1 under normoxia or hypoxia in control HEK293 cells (Fig. 4E lanes 5 and 7). Remarkably, PGAM5 was clearly associated with KEAP1 in MOCS3-KD cells under both normoxia and hypoxia; however, this association was reduced upon hypoxia, leading to decreased immunoprecipitated KEAP1 (Fig. 4E lanes 6 versus 8). Thus, loss of MOCS3 induced KEAP1-PGAM5 interactions in the nucleus. These data suggested that KEAP1-PGAM5 induces mitophagy by disruption of cysteine-sulfur catabolism and increased ROS [8]. Therefore, we examined KEAP1-PGAM5 interactions in the nucleus of LCLs. PGAM5 interacted with KEAP1 in all cells under normoxia, and these interactions were increased in FXD (TAS18/2) versus healthy donors (TAS42/43) in each sex (Fig. 5A lanes 6 versus 5 and Fig. 5B lanes 6 versus 5). Part of these increased interactions were mediated by increased PGAM5-FL (Fig. 5A lane 6, and Fig. 5B lane 6). The KEAP1-PGAM5 complex was likely translocated into the nucleus in TAS18 (Fig. 4C lane 6 versus Fig. 5A lane 6). These data suggested that KEAP1-PGAM5 interactions in the nucleus are facilitated by latent ROS in FXD.

Figure 5.

Figure 5.

Open in a new tab

KEAP1 and PGAM5 interacted with SCLY in FXD. (A–E) KEAP1-IP (A and B) or SCLY-IP (C–E) in nuclear extracts of LCLs from male (A and C) and female (B and D) FXD patients, healthy donors, or from knockdown HEK293 cells (E) were examined by Western blot. eIF3β was used as the loading control. LCLs from FXD patients are indicated with blue labels (A–D). (F) Model illustrates how altered expression of TE-derived genes in FXD modifies the radical scavenging systems where cysteine thiol/ thiolate groups and selenoproteins neutralize ROS [49]. In healthy donor cells, the radical scavenging system of cysteine-sulfur and selenium metabolism is suppressed (blue arrows) upon hypoxia. TE-derived genes regulating immune-responsive genes and selenogenes are upregulated (red arrows). tRNA thiolation is also suppressed by hypoxia. Hypoxia induces ROS (ex. H2O2) and damages mitochondria, leading to PGAM5-mediated mitophagy or oxeiptosis. However, in FXD (TAS20/2/19) cells cultured under normoxia, TE-derived genes have been altered, including genes regulating sulfur metabolism, oxidoreductases, immune response, and selenium metabolism. In FXD cells, PGAM5 is associated with KEAP1 and SCLY in the nucleus under normoxia. These changes maintain a mild hypoxic environment and cell survival in FXD. Selenium (Se), hydrogen selenide (H2Se), thioredoxin reductase (TXNRD), glutathione (GSH), selenodiglutathione (GSSeSG), glutathione reductase (GSR) [49–51].

The PGAM5-KEAP1 pathway and redox signaling via cysteine-sulfur and selenium metabolism are rewired in FXD

We wondered whether redox signaling through cysteine-sulfur and selenium metabolism was connected to the oxidative stress response. We tested whether KEAP1 links to selenium redox signaling. Since we found that TE-derived genes with modified expression include SCLY (Figs 1A and 2C and E), we examined the interactions of SCLY with KEAP1. We found that SCLY interacted with KEAP1 under normoxia in all cells, and the interaction was reduced upon hypoxia in healthy donor cells, as indicated by reduced immunoprecipitation with KEAP1 (Fig. 5A and B). Therefore, we examined these interactions using SCLY-IP. The associations of KEAP1 and PGAM5 with SCLY were increased in TAS18/19 under normoxia and in healthy donor cells upon hypoxia compared to healthy donor cells under normoxia (Fig. 5C lanes 6 and 8 versus lane 5, and Fig. 5D lanes 7 and 8 versus lane 5). TXNRD1 is encoded by the TE-derived TXNRD1 gene that was found to have altered splicing upon hypoxia in FXD and healthy donors in females (Fig. 2D “1242”). TXNRD1 association with SCLY was decreased in TAS18 versus TAS42 in males and was slightly increased in TAS2/19 versus TAS43 in females under normoxia (Fig. 5C lanes 6 versus 5, and Fig. 5D lanes 6 and 7 versus lane 5). Hence, SCLY-bound KEAP1 appeared to be stabilized in the nuclear extracts from healthy donor cells upon hypoxia (Fig. 5C and D lanes 8).

We next examined whether the interaction between KEAP1, PGAM5, and SCLY was linked to sulfur metabolism. Importantly, we found that the association of KEAP1 and PGAM5 with SCLY1 in the nucleus was increased in MOCS3-KD cells versus controls under normoxia and hypoxia, although the increase was less upon hypoxia (Fig. 5E lanes 5 versus 6 and 8, and 6 versus 8, respectively). Thus, the KEAP1-PGAM5 pathway connects to redox signaling of cysteine-sulfur and selenium metabolism. KEAP1 may bind PGAM5 while it senses selenium-dependent redox signaling under loss of sulfur-dependent redox signaling. Importantly, this crosstalk is facilitated in FXD. It might be due to excess ROS via aberrant sulfur consumption and hampered selenium metabolism. Notably, expression of HIF1A-dependent mitophagy activators, such as PDK1, was modified by altered expression of TE-derived genes in FXD (Fig. 2C and D). Hence, the altered expression of TE-derived genes changes to the cysteine-sulfur and selenium redox signaling, which rewires the KEAP-PGAM5 pathway. This may allow cell survival while enduring inflammation in FXD (Fig. 5F). Excessive mitophagy causes deleterious depletion of mitochondria [38]; however, moderately induced mitophagy, restricting respiration, may be beneficial for viral infection.

Altogether, alteration of these radical scavenging systems by modified expression of TE-derived genes sustains a hypoxic environment. This may lead to resistance to ROS and prolonged TE-derived gene expression; in short, viral expression is sustained in FXD. Technically, increases in the replicates of long-read RNA-sequencing may lead to finding additional TEs, of which expression might be modified in FXD. Additionally, the increased incidence of male FXD does not adhere to the rules in the onset of X-linked genetic disease [48]. Supportively, our data show sex-specific differences in the responses of TE-derived genes. However, the variances within the same population with limited parameters in our study restrict some potential conclusions, for instance, the high demand for free radical scavengers in males compared to females. Regarding the involvement of infection in the onset of FXD, somatic infection or gamete infection should be studied at the organism level. Prospective and retrospective cohort studies for virus infections in FXD patients and familial FXD donors would support elucidating the mechanism of FXD. Furthermore, alteration of sulfur and selenium metabolism should be considered in the mechanisms of virus infection and FXD. In addition, alteration of nucleotide metabolism, inducing sustained inflammation, is caused by modified TEs expression and is behind the onset of FXD [10]. Considering these alterations may guide the discovery of the definitive therapy rather than symptomatic treatment, such as anti-inflammation, for FXD.

Supplementary Material

ugag010_Supplemental_Files
ugag010_supplemental_files.zip (43.5MB, zip)

Acknowledgements

We thank Ms. Madelaine Gogol and Workman Lab members for the discussion. This research was supported by NIH grant R35GM118068 and the Stowers Institute for Medical Research.

Author contributions: T.S. (Conceptualization [lead], Data curation [lead], Formal Analysis [equal], Funding acquisition [equal], Investigation [lead], Methodology [lead], Project administration [lead], Resources [lead], Software [equal], Supervision [lead], Validation [lead], Visualization [lead], Writing – original draft [lead], Writing – review & editing [equal]), H.H. (Data curation [equal], Formal Analysis [equal], Investigation [supporting], Methodology [supporting], Software [equal], Validation [supporting], Visualization [supporting]), S.K.S. (Formal Analysis [equal], Investigation [supporting], Methodology [equal], Project administration [supporting], Resources [supporting], Software [supporting], Validation [supporting]), S.L. (Data curation [supporting], Methodology [supporting], Validation [supporting]), J.W. (Conceptualization [supporting], Data curation [supporting], Funding acquisition [equal], Investigation [supporting], Project administration [supporting], Resources [equal], Supervision [supporting], Validation [supporting], Visualization [supporting], Writing – review & editing [equal]).

Contributor Information

Tamaki Suganuma, Stowers Institute for Medical Research, Workman Lab, 1000 E. 50th Street, Kansas City, MO 64110, United States.

Huzaifa Hassan, Stowers Institute for Medical Research, Workman Lab, 1000 E. 50th Street, Kansas City, MO 64110, United States.

Selene K Swanson, Stowers Institute for Medical Research, Workman Lab, 1000 E. 50th Street, Kansas City, MO 64110, United States.

Sunil Laxman, Institute for Stem Cell Science and Regenerative Medicine, Sunil Laxman Lab, Bangalore 560065, India.

Jerry L Workman, Stowers Institute for Medical Research, Workman Lab, 1000 E. 50th Street, Kansas City, MO 64110, United States.

Supplementary data

Supplementary data are available at NAR Molecular Medicine online.

Conflict of interest

None declared.

Funding

This research was supported by National Institute of General Medical Sciences [R35GM118068] and the Stowers Institute for Medical Research.

Data availability

Original data associated with this manuscript is available from NCBI GEO with GSE264091 and GSE307217 and will be accessible from the Stowers Original Data Repository at https://www.stowers.org/research/publications/LIBPB-2579. Original blots of Western blots are available in Supplementary Figs S3–S7.

References

  • 1. Berry-Kravis  E, Levin  R, Shah  H  et al.  Cholesterol levels in fragile X syndrome. Am J Med Genet A. 2015;167A:379–84. 10.1002/ajmg.a.36850. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Santoro  MR, Bray  SM, Warren  ST. Molecular mechanisms of fragile X syndrome: a twenty-year perspective. Annu Rev Pathol. 2012;7:219–45. 10.1146/annurev-pathol-011811-132457. [DOI] [PubMed] [Google Scholar]
  • 3. Hagerman  PJ, Hagerman  RJ. Fragile X-associated tremor/ataxia syndrome. Ann N Y Acad Sci. 2015;1338:58–70. 10.1111/nyas.12693. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Wheeler  AC, Bailey  DB  Jr, Berry-Kravis  E  et al.  Associated features in females with an FMR1 premutation. J Neurodev Disord. 2014;6:30. 10.1186/1866-1955-6-30. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Yu  KH, Palmer  N, Fox  K  et al.  The phenotypical implications of immune dysregulation in fragile X syndrome. Eur J Neurol. 2020;27:590–3. 10.1111/ene.14146. [DOI] [PubMed] [Google Scholar]
  • 6. Winarni  TI, Chonchaiya  W, Sumekar  TA  et al.  Immune-mediated disorders among women carriers of fragile X premutation alleles. Am J Med Genet. Part A 2012;158A:2473–81. 10.1002/ajmg.a.35569. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 7. Suganuma  T, Workman  JL. MPTAC links alkylation damage signaling to sterol biosynthesis. Redox Biol. 2022;51:102270. 10.1016/j.redox.2022.102270. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8. Suganuma  T, Swanson  SK, Gogol  M  et al.  MPTAC determines APP fragmentation via sensing sulfur amino acid catabolism. Cell Rep. 2018;24:1585–96. 10.1016/j.celrep.2018.07.013. [DOI] [PubMed] [Google Scholar]
  • 9. Cantu-Medellin  N, Kelley  EE. Xanthine oxidoreductase-catalyzed reactive species generation: a process in critical need of reevaluation. Redox Biol. 2013;1:353–8. 10.1016/j.redox.2013.05.002. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Suganuma  T, Hassan  H, Gogol  M  et al.  C G composition in transposon-derived genes is increased in FXD with perturbed immune system. NAR Mol Med. 2024;1:ugae015. 10.1093/narmme/ugae015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11. Kohl  JB, Mellis  AT, Schwarz  G. Homeostatic impact of sulfite and hydrogen sulfide on cysteine catabolism. Br J Pharmacol. 2019;176:554–70. 10.1111/bph.14464. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12. Poole  LB. The basics of thiols and cysteines in redox biology and chemistry. Free Radical Biol Med. 2015;80:148–57. 10.1016/j.freeradbiomed.2014.11.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Satoh  T, Lipton  SA. Redox regulation of neuronal survival mediated by electrophilic compounds. Trends Neurosci. 2007;30:37–45. 10.1016/j.tins.2006.11.004. [DOI] [PubMed] [Google Scholar]
  • 14. Saito  R, Suzuki  T, Hiramoto  K  et al.  Characterizations of three major cysteine sensors of keap1 in stress response. Mol Cell Biol. 2016;36:271–84. 10.1128/MCB.00868-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Eggler  AL, Liu  G, Pezzuto  JM  et al.  Modifying specific cysteines of the electrophile-sensing human Keap1 protein is insufficient to disrupt binding to the Nrf2 domain Neh2. Proc Natl Acad Sci USA. 2005;102:10070–5. 10.1073/pnas.0502402102. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Palla  T, Mirzahosseini  A, Noszal  B. Properties of selenolate-diselenide redox equilibria in view of their thiolate-disulfide counterparts. Antioxidants. 2023;12:822. 10.3390/antiox12040822. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Hercher  TW, Krausze  J, Hoffmeister  S  et al.  Insights into the Cnx1E catalyzed MPT-AMP hydrolysis. Biosci Rep. 2020;40: 10.1042/BSR20191806. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Mendel  RR. The molybdenum cofactor. J Biol Chem. 2013;288:13165–72. 10.1074/jbc.R113.455311. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Neukranz  Y, Kotter  A, Beilschmidt  L  et al.  Analysis of the cellular roles of MOCS3 identifies a MOCS3-independent localization of NFS1 at the tips of the centrosome. Biochemistry. 2019;58:1786–98. 10.1021/acs.biochem.8b01160. [DOI] [PubMed] [Google Scholar]
  • 20. Noma  A, Sakaguchi  Y, Suzuki  T. Mechanistic characterization of the sulfur-relay system for eukaryotic 2-thiouridine biogenesis at tRNA wobble positions. Nucleic Acids Res. 2009;37:1335–52. 10.1093/nar/gkn1023. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Leidel  S, Pedrioli  PG, Bucher  T  et al.  Ubiquitin-related modifier Urm1 acts as a sulphur carrier in thiolation of eukaryotic transfer RNA. Nature. 2009;458:228–32. 10.1038/nature07643. [DOI] [PubMed] [Google Scholar]
  • 22. Schlieker  CD, Van der Veen  AG, Damon  JR  et al.  A functional proteomics approach links the ubiquitin-related modifier Urm1 to a tRNA modification pathway. Proc Natl Acad Sci USA. 2008;105:18255–60. 10.1073/pnas.0808756105. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23. Ranjan  N, Rodnina  MV. tRNA wobble modifications and protein homeostasis. Translation. 2016;4:e1143076. 10.1080/21690731.2016.1143076. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 24. Laxman  S, Sutter  BM, Wu  X  et al.  Sulfur amino acids regulate translational capacity and metabolic homeostasis through modulation of tRNA thiolation. Cell. 2013;154:416–29. 10.1016/j.cell.2013.06.043. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 25. Chowdhury  MM, Dosche  C, Lohmannsroben  HG  et al.  Dual role of the molybdenum cofactor biosynthesis protein MOCS3 in tRNA thiolation and molybdenum cofactor biosynthesis in humans. J Biol Chem. 2012;287:17297–307. 10.1074/jbc.M112.351429. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 26. Torres  AG, Batlle  E, Ribas de Pouplana  L. Role of tRNA modifications in human diseases. Trends Mol Med. 2014;20:306–14. 10.1016/j.molmed.2014.01.008. [DOI] [PubMed] [Google Scholar]
  • 27. Suganuma  T, Mushegian  A, Swanson  SK  et al.  The ATAC acetyltransferase complex coordinates MAP kinases to regulate JNK target genes. Cell. 2010;142:726–36. 10.1016/j.cell.2010.07.045. [DOI] [PubMed] [Google Scholar]
  • 28. Su  D, Chan  CT, Gu  C  et al.  Quantitative analysis of ribonucleoside modifications in tRNA by HPLC-coupled mass spectrometry. Nat Protoc. 2014;9:828–41. 10.1038/nprot.2014.047. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29. Dobin  A, Davis  CA, Schlesinger  F  et al.  STAR: ultrafast universal RNA-seq aligner. Bioinformatics. 2013;29:15–21. 10.1093/bioinformatics/bts635. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 30. Li  B, Dewey  CN. RSEM: accurate transcript quantification from RNA-Seq data with or without a reference genome. BMC Bioinformatics. 2011;12:323. 10.1186/1471-2105-12-323. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Zhang  Y, Parmigiani  G, Johnson  WE. ComBat-seq: batch effect adjustment for RNA-seq count data. NAR Genom Bioinform. 2020;2: 10.1093/nargab/lqaa078lqaa078.[CrossRef] [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Robinson  MD, McCarthy  DJ, Smyth  GK. edgeR: a bioconductor package for differential expression analysis of digital gene expression data. Bioinformatics. 2010;26:139–40. 10.1093/bioinformatics/btp616. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 33. Akaike  T, Noguchi  Y, Ijiri  S  et al.  Pathogenesis of influenza virus-induced pneumonia: involvement of both nitric oxide and oxygen radicals. Proc Natl Acad Sci USA. 1996;93:2448–53. 10.1073/pnas.93.6.2448. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 34. Storer  J, Hubley  R, Rosen  J  et al.  The Dfam community resource of transposable element families, sequence models, and genome annotations. Mobile DNA. 2021;12:2. 10.1186/s13100-020-00230-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Payer  LM, Burns  KH. Transposable elements in human genetic disease. Nat Rev Genet. 2019;20:760–72. 10.1038/s41576-019-0165-8. [DOI] [PubMed] [Google Scholar]
  • 36. Shen  S, Park  JW, Lu  ZX  et al.  rMATS: robust and flexible detection of differential alternative splicing from replicate RNA-Seq data. Proc Natl Acad Sci USA. 2014;111:E5593–5601. 10.1073/pnas.1419161111. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Deng  X, Wang  Q, Cheng  M  et al.  Pyruvate dehydrogenase kinase 1 interferes with glucose metabolism reprogramming and mitochondrial quality control to aggravate stress damage in cancer. J Cancer. 2020;11:962–73. 10.7150/jca.34330. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 38. Wilhelm  LP, Ganley  IG. Eating your mitochondria-when too much of a good thing turns bad. EMBO J. 2023;42:e114542. 10.15252/embj.2023114542. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 39. Imai  H, Hakkaku  N, Iwamoto  R  et al.  Depletion of selenoprotein GPx4 in spermatocytes causes male infertility in mice. J Biol Chem. 2009;284:32522–32. 10.1074/jbc.M109.016139. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 40. Liu  A, Li  F, Xu  P  et al.  Correction to: gpx4, Selenov, and Txnrd3 are three most testis-abundant selenogenes resistant to dietary selenium concentrations and actively expressed during reproductive ages in rats. Biol Trace Elem Res. 2023;201:260–1. 10.1007/s12011-022-03167-w. [DOI] [PubMed] [Google Scholar]
  • 41. Nease  LA, Church  KP, Delclaux  I  et al.  Selenocysteine tRNA methylation promotes oxidative stress resistance in melanoma metastasis. Nature Cancer. 2024;5:1868–84. 10.1038/s43018-024-00844-8. [DOI] [PubMed] [Google Scholar]
  • 42. Holze  C, Michaudel  C, Mackowiak  C  et al.  Oxeiptosis, a ROS-induced caspase-independent apoptosis-like cell-death pathway. Nat Immunol. 2018;19:130–40. 10.1038/s41590-017-0013-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 43. Pallichankandy  S, Thayyullathil  F, Cheratta  AR  et al.  Targeting oxeiptosis-mediated tumor suppression: a novel approach to treat colorectal cancers by sanguinarine. Cell Death Discov. 2023;9:94. 10.1038/s41420-023-01376-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44. Qi  Y, Rajbanshi  B, Hao  R  et al.  The dual role of PGAM5 in inflammation. Exp Mol Med. 2025;57:298–311. 10.1038/s12276-025-01391-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 45. Baba  T, Tanimura  S, Yamaguchi  A  et al.  Cleaved PGAM5 dephosphorylates nuclear serine/arginine-rich proteins during mitophagy. Biochimica et biophysica acta. Mol Cell Res. 2021;1868:119045. [DOI] [PubMed] [Google Scholar]
  • 46. Pabis  M, Termathe  M, Ravichandran  KE  et al.  Molecular basis for the bifunctional Uba4-Urm1 sulfur-relay system in tRNA thiolation and ubiquitin-like conjugation. EMBO J. 2020;39:e105087. 10.15252/embj.2020105087. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47. Mayer  K, Mundigl  O, Kettenberger  H  et al.  Diphthamide affects selenoprotein expression: diphthamide deficiency reduces selenocysteine incorporation, decreases selenite sensitivity and pre-disposes to oxidative stress. Redox Biol. 2019;20:146–56. 10.1016/j.redox.2018.09.015. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 48. Dobyns  WB, Filauro  A, Tomson  BN  et al.  Inheritance of most X-linked traits is not dominant or recessive, just X-linked. Am J Med Genet A. 2004;129A:136–43. 10.1002/ajmg.a.30123. [DOI] [PubMed] [Google Scholar]
  • 49. Kang  D, Lee  J, Wu  C  et al.  The role of selenium metabolism and selenoproteins in cartilage homeostasis and arthropathies. Exp Mol Med. 2020;52:1198–208. 10.1038/s12276-020-0408-y. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50. Schoenmakers  E, Carlson  B, Agostini  M  et al.  Mutation in human selenocysteine transfer RNA selectively disrupts selenoprotein synthesis. J Clin Invest. 2016;126:992–6. 10.1172/JCI84747. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 51. Shimada  BK, Swanson  S, Toh  P  et al.  Metabolism of selenium, selenocysteine, and selenoproteins in ferroptosis in solid tumor cancers. Biomolecules. 2022;12:1581. 10.3390/biom12111581. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ugag010_Supplemental_Files
ugag010_supplemental_files.zip (43.5MB, zip)

Data Availability Statement

Original data associated with this manuscript is available from NCBI GEO with GSE264091 and GSE307217 and will be accessible from the Stowers Original Data Repository at https://www.stowers.org/research/publications/LIBPB-2579. Original blots of Western blots are available in Supplementary Figs S3–S7.

Articles from NAR Molecular Medicine are provided here courtesy of Oxford University Press

RESOURCES