The TEX44-CPT1B axis regulates mitochondrial sheath assembly and fatty acid oxidation in sperm

Abstract

Mitochondrial fatty acid β-oxidation (FAO) is essential for energy production and cellular homeostasis, yet its role in sperm function has remained unclear. Through whole-exome sequencing (WES) of 800 patients with asthenozoospermia, we identified biallelic Testis-Expressed Protein 44 (TEX44) variants in six individuals, all of whom exhibited defective mitochondrial sheath assembly and impaired sperm motility. Using Tex44 knockout mice, we show that TEX44 interacts with carnitine palmitoyltransferase 1B (CPT1B) to form a mitochondrial glue, anchoring adjacent mitochondria and facilitating the assembly of the sperm-specific mitochondrial sheath. In vitro, we show that purified TEX44 protein can modulate CPT1B enzymatic activity, limiting the conversion of long-chain fatty acids such as palmitic acid and myristic acid into acyl-carnitines, thereby reducing reactive oxygen species (ROS) production. Loss of TEX44 disrupts this regulatory mechanism, leading to unregulated FAO, excessive ROS generation, and severe oxidative damage to sperm DNA and flagellar structure. Additionally, germ cell-specific Cpt1b knockout mice exhibit phenotypes similar to TEX44 deficiency, including mitochondrial sheath defects and reduced sperm motility. These findings reveal a sperm-specific mechanism by which TEX44 regulates CPT1B activity to balance FAO and ROS generation, providing critical insights into energy metabolism, mitochondrial integrity, and male infertility.

Fig. 1. Asthenozoospermia and mitochondrial sheath defects in spermatozoa associated with human TEX44 variants.

A Flow chart of the identification of human TEX44 variants related to asthenozoospermia. Created in BioRender. f Zl, v}. (2025) https://BioRender.com/kbr0lu4. B Distribution of patient categories based on semen analysis. C Pedigree analysis of six families of individuals harboring biallelic TEX44 variants identified through whole-exome sequencing. Infertile males are denoted by filled black squares. Sanger sequencing results are shown beneath the pedigrees, with mutated residues highlighted by red arrows and boxes. D HE staining was used to assess the morphology of spermatozoa from a fertile control man and patients with biallelic TEX44 variants via light microscope. Scale bar: 10 μm. E Quantification of abnormal mitochondrial sheaths in fertile controls and patients with biallelic TEX44 variants (P = 1.45 × 10⁻⁹, n = 5). F Sperm tail length comparison between fertile controls and patients harboring biallelic TEX44 variants (ns not significant, n = 5). G Immunofluorescent staining for TEX44 in spermatozoa from fertile controls and males harboring biallelic TEX44 variants. Anti-TEX44 (green) and anti-TOMM20 (magenta) antibodies were used, with Hoechst (blue) for nuclei staining. Scale bar: 5 μm. H Western Blot analysis of TEX44 in spermatozoa from a fertile control and patients harboring biallelic TEX44 variant. TEX44 is missing in spermatozoa from patients, with β-Tubulin as a loading control. Uncropped blots are provided in Source data. I Transmission electron microscopy of longitudinal sperm flagellar midpiece sections from fertile controls and individuals with biallelic TEX44 variants. Red arrows indicate sloughed-off mitochondria; yellow arrowheads indicate the annulus. Scale bar: 500 nm. J Schematic diagram of mitochondrial sheath defect in spermatozoa from fertile controls and patient harboring biallelic TEX44 variant. K Immunofluorescence analysis showing the localization of TOMM20 (green), SEPT4 (magenta), and Hoechst (blue) in control and Family IV-1 spermatozoa. Scale bar: 10 μm. For (E, F), data are mean ± s.e.m. P values were determined using two-tailed Student’s t-tests. n values represent the number of biologically independent experiments. **** indicates statistical significance at P < 0.0001. H, I Experiments were performed on samples from three unrelated individuals with biallelic TEX44 variants. Similar results were observed across all biological replicates. K Images are representative of multiple spermatozoa from the same individual, with consistent staining patterns observed.

Fig. 2. IVF and ICSI outcomes with Tex44^−/− and Tex44^+/+ mice.

A Schematic diagram of in vitro fertilization (IVF). B–E Representative images of zygotes, 2-cell embryos, and blastocysts obtained from IVF in mice (B). The percentage of zygotes (C), 2-cell embryos (D), and blastocysts (E) in Tex44^+/+ and Tex44^−/− mice is illustrated (P = 0.0136 for (C), P = 0.0108 for (D), P = 0.0395 for (E); n = 3). Scale bars: 200 μm. F Schematic diagram of intracytoplasmic sperm injection (ICSI). G–J Representative images of zygotes, 2-cell stage, and blastocysts obtained from ICSI in mice (G). Comparable percentages of zygotes (H), 2-cell embryos (I), and blastocysts (J) were observed in Tex44^+/+ and Tex44^−/− mice (ns not significant, n = 3). Scale bar: 200 μm. For (C–E, H–J), data are presented as mean ± s.e.m. P values were determined using two-tailed Student’s t-tests. n values represent the number of biologically independent animals. * indicates statistical significance at P < 0.05.

Fig. 3. TEX44 interacts with CPT1B to function as an inter-mitochondrial linker.

A, B Co-immunoprecipitation assays assessing interactions between FLAG-TEX44 and HA-CPT1B. TEX44-FLAG and CPT1B-HA were immunoprecipitated from HEK293T cell lysate transfected with the indicated vectors with anti-FLAG (A) and anti-HA (B) antibodies. Interactions between proteins were detected with antibodies to FLAG or HA. Uncropped blots are provided in Source data. C Immunofluorescent staining for TEX44-EGFP (green) or CPT1B-HA (magenta) with antibodies to EGFP or HA in HeLa cell lines. TOMM20 served as a mitochondrial marker (blue). The magnified region in the right panel shows TEX44 assemblies between mitochondria (marked by white arrows). Scale bars: 10 μm (left), 2 μm (right). D Immunofluorescent staining was performed to examine the expression of TEX44 and CPT1B during various stages of spermiogenesis (steps 1–16) in the testes of adult wild-type mice. TEX44 (magenta) and CPT1B (green) antibodies were used, along with PNA (white) to highlight acrosomal structures. Nuclei were stained with Hoechst (blue). Scale bars: 20 μm. E Immunofluorescent staining for CPT1B (magenta) and TEX44 (green) in human and mice sperm samples. Nuclei were stained with Hoechst (blue). Scale bars: 5 μm. F Graph showing disordered regions (IDRs) of TEX44 identified by IUPred3. A score of ≥0.5 indicates disordered regions. A schematic representation of TEX44 protein and truncated mutants is shown, with orange boxes indicating IDRs. The numbers represent amino acid residues. G Co-immunoprecipitation assay for full-length or truncated TEX44-FLAG and CPT1B-HA with antibodies to FALG or HA. Uncropped blots are provided in Source data. H Immunofluorescent staining for full-length or truncated TEX44-FLAG (green) and CPT1B-HA (magenta) in HeLa cells. Nuclei were stained with Hoechst (blue). Scale bars: 10 μm. A–E, G, H All experiments were independently repeated at least three times with consistent results. Representative data are shown.

Fig. 4. Loss of CPT1B in male germ cells leads to TEX44 assembly failure and abnormal mitochondrial sheath formation.

A Generation of Cpt1b gene knockout mice through Cre-LoxP mediated exon deletion. B Scanning electron microscopy of spermatozoa from the cauda epididymis of control and Cpt1b gKO mice. White arrows indicate regions lacking mitochondria. Scale bars: 10 μm (left), 2 μm(right). C Transmission electron microscopy of spermatozoa from the cauda epididymis of control and Cpt1b gKO mice. The black boxed areas in the upper panels of each group are shown at higher magnification in the corresponding lower panels. Scale bars: 2 μm (upper), 1 μm (lower). D Immunofluorescence analysis of TEX44 (green) in spermatozoa from control and Cpt1b gKO mice. The mitochondrial sheath was labeled with TOMM20 (magenta), and nuclei were stained with Hoechst (blue). Scale bars: 50 μm. E Immunofluorescence analysis of TEX44 (green) from control and Cpt1b gKO testis. The mitochondrial sheath, acrosome, and nuclei were stained with GPX4 (magenta), WGA (yellow), and Hoechst (white), respectively. The regions demarcated by white dashed boxes in the middle panels of each group are presented at higher magnification in the corresponding right panels. The lower panels display the fluorescence intensity profiles corresponding to the areas indicated by the light blue lines. Scale bars: 20 μm (left), 10 μm (right). F Immunofluorescence analysis of CPT1B (green) from Tex44^+/+ and Tex44^−/− testis. TOMM20 (magenta) were used to stain the mitochondrial sheath, Hoechst (white) were used to stain nucleus, wheat germ agglutinin (WGA; yellow) were used to determine the step of spermiogenesis. The regions demarcated by white dashed boxes in the middle panels of each group are presented at higher magnification in the corresponding right panels. The lower panels display the fluorescence intensity profiles corresponding to the areas indicated by the light blue lines. Scale bars: 20 μm (left), 10 μm (right). B–F All experiments were independently repeated at least three times with consistent results. Representative data are shown.

Fig. 5. TEX44–CPT1B limits the generation of L-palmitoylcarnitine and myristoyl-l-carnitine of CPT1B in sperm.

A Schematic diagram of fatty acid transport. CPT1B, located on the mitochondrial outer membrane, converts long-chain acyl-CoA into long-chain acyl-carnitine. Short-chain fatty acids freely diffuse across mitochondrial membrane. B Quantification of carnitine, carnitine-conjugated long-chain fatty acids (LCFAs), and carnitine-conjugated short-chain fatty acids (SCFAs) in spermatozoa from control (n = 16), Cpt1b gKO (n = 9), and Tex44^−/− (n = 9) mice. For L-palmitoylcarnitine, significant differences were observed between Cpt1b gKO and control (P = 3.64 × 10⁻⁵) and between Tex44^−/− and control (P = 0.0019). For myristoyl-L-carnitine, significant differences were also observed between Cpt1b gKO and control (P = 0.0002) and between Tex44^−/− and control (P = 0.0002). Asterisks indicate statistical significance: P < 0.01 (**), P < 0.001 (***), P < 0.0001 (****); ns not significant, P ≥ 0.05. C Immunofluorescence analysis of CPT1B (green) in spermatozoa from Tex44^+/+ and Tex44^−/− mice. TOMM20 (magenta) marks mitochondria, Hoechst (blue) stains nuclei. Scale bars: 50 μm. D Western blot analysis showing comparable CPT1B protein levels in sperm from Tex44^+/+ and Tex44^−/− mice, with β-tubulin as a loading control. Uncropped blots are provided in Source data. E Quantitative analysis of CPT1B protein level was performed and normalized with β-tubulin protein level (n = 3). No statistically significant difference was observed (ns, P ≥ 0.05). F Schematic workflow for protein expression, purification, and subsequent CPT1B enzyme activity assay. CPT1B enzyme activity was tested alone and in combination with TEX44-FL or TEX44-C. G Coomassie brilliant blue staining to assess protein purity (CPT1B, TEX44-FL, TEX44-C), with BSA used as a reference for protein quantification. H CPT1B enzyme activity at incremental concentrations showing a linear relationship (R² = 0.9884). Relative enzyme activity of CPT1B in the presence of TEX44-FL (I) or TEX44-C (J) at various TEX44-to-CPT1B ratios. TEX44-FL inhibits CPT1B activity in a dose-dependent manner, whereas TEX44-C enhances activity. For (B, E), data are presented as mean ± s.e.m. P values were determined using two-tailed unpaired Student’s t-tests. n values represent the number of biologically independent animals. C–E Experiments were performed using samples from three biologically independent mice, with consistent results observed. Representative images and blots are shown.

Fig. 6. Predicted complex structures between CPT1B and TEX44 by AlphaFold3.

A Predicted structure of full-length CPT1B and TEX44 colored by chain (left) or pLDDT scores (right). The top hit was presented. B, C Predicted structure of full-length CPT1B and TEX44 (385–530). Two representative results were presented. The N-terminal of TEX44 insert through CPT1B in (B) but not in (C). D All five predicted results of the complex structure between CPT1B and TEX44 (385–530) show an invariant segment of TEX44 (424–450) highlighted in red. E Electrostatic surface of CPT1B (left) and TEX44 (424–450) (right). Positively and negatively charges are indicated by blue and red colors, respectively.

Fig. 7. Tex44^−/− sperm exhibit elevated ROS levels, abnormal morphology, and DNA damage when cultured in a medium containing palmitoyl-CoA and carnitine.

A Schematic workflow for detecting ROS levels, morphological defects, and apoptosis in control, Tex44^−/−, and Cpt1b gKO spermatozoa after treatment with palmitoyl-CoA and carnitine. B Mean fluorescence intensity of ROS levels in sperm treated with increasing concentrations of palmitoyl coenzyme A (0, 10 nM, 100 nM, and 1 µM) in control, Tex44^−/−, and Cpt1b gKO mice (n = 9 per group). Compared to controls, Tex44^−/− sperm showed significantly higher ROS levels at 0 nM (P = 9.15 × 10⁻¹⁵), 10 nM (P = 2.38 × 10⁻¹⁷), 100 nM (P = 2.88 × 10⁻²⁰), and 1 µM (P = 2.81 × 10⁻¹⁹); Cpt1b gKO sperm showed significantly lower ROS levels at 0 nM (P = 4.88 × 10⁻¹⁸), 10 nM (P = 4.32 × 10⁻¹⁴), 100 nM (P = 6.94 × 10⁻²⁰), and 1 µM (P = 1.86 × 10⁻¹⁹). The concentration of L-carnitine in each group was 1 mM. C, D HE staining of spermatozoa from control, Tex44^−/− and Cpt1b gKO mice (C) before and after treatment with 1 µM palmitoyl-CoA and 1 mM carnitine (n = 3, each). Quantification of bent tail rates in spermatozoa after treatment with 1 µM palmitoyl-CoA and 1 mM carnitine. Tex44^−/− spermatozoa show a significant increase after treatment (P = 2.73 × 10⁻⁵), while a moderate increase is seen in controls (P = 0.0341). No significant change is observed in Cpt1b gKO spermatozoa (P = 0.1304). Scale bars: 50 μm. E, F TUNEL staining (red) of control, Tex44^−/− and Cpt1b gKO spermatozoa, before and after treatment with 1 µM palmitoyl-CoA and 1 mM carnitine (n = 3, each). The quantification of TUNEL-positive spermatozoa is comparable between the control (P = 0.2236) and Cpt1b gKO (P = 0.5059) mice after treatment with 1 µM palmitoyl-CoA and 1 mM carnitine, but a significant increase in the Tex44^−/− mice (P = 1.24 × 10⁻⁷). The nuclei were stained with Hoechst (blue). Scale bars: 100 μm. For (B, D, F), data are presented as mean ± s.e.m. P values were determined using two-tailed unpaired Student’s t-tests. n values represent the number of biologically independent animals. * indicates statistical significance at P < 0.05. **** indicates statistical significance at P < 0.0001. ns not significant.