SLC22A14 is a mitochondrial riboflavin transporter required for sperm oxidative phosphorylation and male fertility

Abstract

Ablation of Slc22a14 causes male infertility in mice, but the underlying mechanisms remain unknown. Here, we show that SLC22A14 is a riboflavin transporter localized at the inner mitochondrial membrane of the spermatozoa mid-piece and show by genetic, biochemical, multi-omic, and nutritional evidence that riboflavin transport deficiency suppresses the oxidative phosphorylation and reprograms spermatozoa energy metabolism by disrupting flavoenzyme functions. Specifically, we find that fatty acid β-oxidation (FAO) is defective with significantly reduced levels of acyl-carnitines and metabolites from the TCA cycle (the citric acid cycle) but accumulated triglycerides and free fatty acids in Slc22a14 knockout spermatozoa. We demonstrate that Slc22a14-mediated FAO is essential for spermatozoa energy generation and motility. Furthermore, sperm from wild-type mice treated with a riboflavin-deficient diet mimics those in Slc22a14 knockout mice, confirming that an altered riboflavin level causes spermatozoa morphological and bioenergetic defects. Beyond substantially advancing our understanding of spermatozoa energy metabolism, our study provides an attractive target for the development of male contraceptives.

Keywords: SLC22A14 transporter; energy metabolism; fatty acid β-oxidation; male infertility; riboflavin.

Graphical abstract

Figure 1

Slc22a14 KO results in dysregulated spermatozoa motility, deformities, and infertility in male mice (A) Slc22a14 mRNA levels in mouse and human tissues. (B) Map of Slc22a14 KO mouse. (C) Fertility was analyzed by mating males with WT females using standard methods (n = 8). (D) SEM (i and iii) and TEM (ii and iv) imaging of spermatozoa showing the morphology of normal spermatozoa flagellum (top), in contrast to the angulated flagellum (bottom) in Slc22a14 KO spermatozoa. (E) TEM imaging of caudal epididymis showing a mature flagellum (blue arrows) of a WT spermatozoa in comparison with a hairpin bending (red arrows) of Slc22a14 KO flagellum. (F and G) IVF with cumulus-intact oocytes (F). Slc22a14 KO spermatozoa cannot fertilize cumulus-intact oocytes, even when exposed to zona pellucida with artificial assistance; they cannot penetrate the zona pellucida (G). (H) The IVF success rates of male mice were analyzed with cumulus cells (n = 3). (I) Birth rates assessed by ICSI test (n = 4). (J) Immunofluorescence staining of sp56, an acrosome marker of epididymal spermatozoa. Blue, DAPI; red, sp56. (K) A CASA system was used to measure spermatozoa motility parameters. VAP, average path velocity; VSL, straight-line velocity; VCL, curvilinear velocity; ALH, amplitude of lateral head displacement; BCF, beat cross frequency; STR, straightness of cell track; LIN, linearity of cell track (n = 5). Student’s t test; ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001; error bars, SEM. See also Figures S1–S3.

Figure 2

Knockout of Slc22a14 disrupts spermatozoa energy production, resulting in reduced TCA cycle but elevated glycolytic pathway activity (A) The rates of undamaged (calcein [+]/ Pi [−]), damaged (calcein [+]/ Pi [+]), and dead spermatozoa (calcein [−]/ Pi [+]) were calculated and analyzed. Quantification of calcein/Pi double staining in spermatozoa (n = 3). (B) ATP generation from caudal epididymal spermatozoa (n = 4). (C) Intracellular ROS levels in spermatozoa were assayed using the dye DCFH-DA and analyzed (n = 5). (D) The MMP of spermatozoa was determined using JC-1 probes and analyzed (n = 4). (E and F) The significant differential (p < 0.05) metabolite profile was analyzed by the pathway analysis of MetaboAnalyst, suggesting a strong perturbation of energy-related metabolic pathways in spermatozoa (Slc22a14 KO versus WT); increased (E), decreased (F). (G and H) Changes of metabolites in Slc22a14 KO spermatozoa compared with WT. (G) glycolysis. (H) TCA cycle (n = 3). Student’s t test; ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001; error bars, SEM. See also Figures S3 and S4.

Figure 3

Characterization of the metabolic regulation of glucose in Slc22a14 KO and WT spermatozoa (A) Scheme outlining the path of [U-¹³C₆]-glucose in glycolysis and the TCA cycle. (B and C) Significantly changed spermatozoa metabolites involved in ¹³C-labeled metabolic flux (in glycolysis, B; the TCA cycle, C) (n = 3). m+1 means ¹³C-labeled compound, m+2 means two ¹³C-labeled compounds, and m+3 means three ¹³C-labeled compounds. (D) Spermatozoa glucose uptake was measured using 2-deoxy-2-((7-nitro-2,1,3-benzoxadiazol-4-yl) amino)-D-glucose (2-NBDG), a D-glucose analog, diluted in glucose-free BWW medium. Results are presented as geometric mean fluorescence intensity of live cells measured by flow cytometry (n = 4). (E) Spermatozoa lactate production was examined using an L-lactate assay kit (n = 6). (F) ECAR was measured using a Seahorse XF96 analyzer. Each data point represents the mean (±SEM) of three independent spermatozoa samples. A total of 10 mM glucose, 1 μM oligomycin, and 50 mM 2-DG were injected sequentially to the sample plate at the time points indicated (n = 4). (G) ECAR values are presented as a bar graph, normalized to spermatozoa counts. (H) OCR was measured using a Seahorse XF96 analyzer. Each data point represents the mean (±SEM) of three independent spermatozoa samples. A total of 1 μM oligomycin, 1 μM FCCP, and 4 μM antimycin with rotenone were injected sequentially to the sample plate at the time points indicated (n = 4). (I) OCR values are presented as a bar graph, normalized to spermatozoa counts. Student’s t test; ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001; error bars, SEM.

Figure 4

Knockout of Slc22a14 causes lipid accumulation in spermatozoa resulting from suppressed FAO (A) LC-MS-based untargeted metabolomics analysis showing that short-chain acylcarnitine levels were significantly increased in Slc22a14 KO spermatozoa compared with WT (n = 3). (B) LC-MS-based untargeted metabolomics analysis showing that long-chain acylcarnitine levels were significantly reduced in Slc22a14 KO spermatozoa compared with WT (n = 3). (C) LC-MS-based untargeted metabolomics analysis showing that acylcarnitine levels were significantly elevated in HEK-SLC22A14 cells compared with control group. All p < 0.05 (n = 4). (D) Bar graph illustrating the lipidome of Slc22a14 KO and WT spermatozoa. (E) Saturated long-chain FFA contents significantly altered in Slc22a14 KO spermatozoa compared with WT (n = 3). (F) ATP levels of spermatozoa treated with 640 μM etomoxir, 25 mM 2-DG, and combined usage of two agents. Eto, etomoxir (n = 3). (G) Radioactive ¹⁴CO₂ measurements revealed significantly reduced FAO in Slc22a14 KO spermatozoa (n = 5). (H) Radioactive ¹⁴CO₂ measurements revealed significantly increased FAO in HEK-overexpressing SLC22A14 cells compared with control group (n = 3). (I and J) OCR in Slc22a14 KO and WT spermatozoa. Each data point represents the mean (±SEM) of three independent samples. Where presented, 40 μM etomoxir and 1 μM oligomycin were added. The summary data are shown at the right (J) (n = 4). Student’s t test; ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001; error bars, SEM. See also Figure S4.

Figure 5

SLC22A14 is localized at the inner membrane of spermatozoa mitochondria (A) SLC22A14 subcellular co-localization with mito-tracker red in HEK293T cells overexpressing the EGFP-SLC22A14 fusion protein. (B and C) Identification of hSLC22A14 (B) and mSlc22a14 (C) in mitochondria from HEK-hSLC22A14 and HEK-mSlc22a14 cells. The MS spectra of representative peptides are presented. (D) Identification of mSlc22a14 in mitochondria isolated from spermatozoa. The MS characteristics of representative peptides are presented. (E) Immunofluorescent staining of Slc22a14 in epididymal spermatozoa showed a mid-piece localization of Slc22a14. Blue, DAPI; red, Slc22a14 antibody. (F) Super resolution confocal images of mitochondria isolated from HEK293T cells overexpressing a FLAG-SLC22A14 fusion protein. The FLAG-SLC22A14 fusion protein was labeled with an anti-FLAG-tag antibody conjugated with Alexa Fluor 488 (green); inner membranes were labeled with an anti-COX-IV antibody and visualized with a secondary antibody conjugated to Alexa Fluor 594 (red); outer membranes were labeled with an anti-TOMM20 antibody and visualized with a secondary antibody conjugated to Alexa Fluor 647 (blue). (G) Plot profile of the COX-IV, FLAG-tag, and TOMM20 fluorescence intensity along the axis (line with arrow in F). (H) Super resolution confocal images of single mitochondria in HEK293T cells. FLAG-SLC22A14 fusion protein was labeled with an anti-FLAG-tag antibody conjugated with Alexa Fluor 488 (green); IMM was labeled with an anti-COX-IV antibody and visualized with a secondary antibody conjugated to Alexa Fluor 594 (red); outer mitochondrial membranes were labeled with anti-TOMM20 antibody and visualized with a secondary antibody conjugated to Alexa Fluor 647 (blue). (I) Plot profile of the COX-IV, FLAG-tag, and TOMM20 fluorescence intensity along the axis (line with arrow in H). See also Figures S5–S7.

Figure 6

SLC22A14 is identified as a riboflavin (Rf) transporter (A) Levels of Rf and Rf-derivative compounds in the mitochondrial metabolome of HEK-SLC22A14 cell lines (n = 3). (B) Rf content of spermatozoal mitochondria was measured (n = 4). (C) Uptake of radioactive Rf by HEK-EV, HEK-hSLC22A14, and HEK-mSlc22a14 cells (n = 3). (D and E) Homology-based 3D structural model of human SLC22A14 (D) and mouse Slc22a14 (E) based on PDB: 6H7D as the template, shown in ribbon representation. The transmembrane regions of the N domains of human (i) and mouse (iv) models are colored darker, and their respective C domains are colored lighter. Cut-through section of human SLC22A14 (ii) and mouse Slc22a14 (v) models depicted in surface representation showing the protein in an outward-occluded conformation, docked with Rf and the residues comprising the binding pocket. Close-up view of the central binding pocket showing a docked pose of Rf in the human SLC22A14 (iii) and mouse Slc22a14 (vi) model, also showing the SLC22A14 binding site for Rf. (F) Kinetics of Rf uptake by mitochondria isolated from HEK-EV, HEK-hSLC22A14, and HEK-mSlc22a14 cells (n = 3). (G) Uptake of radioactive Rf by HEK-EV, HEK-hSLC22A14, and HEK-mSlc22a14 and triple-mutated HEK-hSLC22A14 and HEK-mSlc22a14 cells (n = 3). Student’s t test; ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001; error bars, SEM.

Figure 7

Knockout of Slc22a14 reduces the activity of spermatozoa flavoenzymes, and a Rf-deficient diet induces Slc22a14-KO-like phenotypes in WT spermatozoa (A) Activity of ACADL in Slc22a14 KO and WT spermatozoa (n = 4). (B) Enzymatic activity of complex I in Slc22a14 KO and WT spermatozoa (n = 4). (C) Enzymatic activity of complex II in Slc22a14 KO and WT spermatozoa (n = 3). (D) Rf levels in testis were measured using ELISA (n = 3). (E) Spermatozoa abnormalities were categorized based on flagellum angulation (n = 4). (F) SEM imaging of spermatozoa showing the morphology of normal spermatozoa flagellum in contrast to the angulated flagellum in Rf-deficient spermatozoa. The asterisks indicate the bending point of the spermatozoa flagellum. (G) TEM imaging of caudal epididymis tissue showing a mature flagellum (blue arrows) of a normal spermatozoon in comparison with an Rf-deficient flagellum (red arrows). (H) Spermatozoa motility was analyzed using the CASA system (n = 3). (I) ATP generation from caudal epididymal spermatozoa (n = 3). (J) Gene expression levels (mRNA) of testicular Rf transporters and Slc22a14 (n = 8). (K) A proposed working model for SLC22A14. Rf is the precursor of FMN and FAD, which are coenzymes of many enzymes in the TCA cycle (FAD: OGDH), complex I (FMN: NADH dehydrogenase), and complex II (FAD: succinate dehydrogenase) of the ETC and in FAO (FAD: ACADL). Ablation of Slc22a14 disrupts Rf transport into the mitochondrial matrix, leading to FMN and FAD depletion and inactivity of flavoenzymes. Consequently, ATP generation from both FAO and OXPHOS are toned down, resulting in compensatory glycolysis upregulation. However, the increase of glycolysis pathway activity is insufficient to compensate for spermatozoa energy needs, resulting in spermatozoa disfunction and male infertility. Red dot, FAD; red triangle, FMN. Student’s t test; ^∗p < 0.05, ^∗∗p < 0.01, ^∗∗∗p < 0.001; error bars, SEM. See also Figure S7.