Tetraspanin is required for generation of reactive oxygen species by the dual oxidase system in Caenorhabditis elegans

Abstract

Reactive oxygen species (ROS) are toxic but essential molecules responsible for host defense and cellular signaling. Conserved NADPH oxidase (NOX) family enzymes direct the regulated production of ROS. Hydrogen peroxide (H(2)O(2)) generated by dual oxidases (DUOXs), a member of the NOX family, is crucial for innate mucosal immunity. In addition, H(2)O(2) is required for cellular signaling mediated by protein modifications, such as the thyroid hormone biosynthetic pathway in mammals. In contrast to other NOX isozymes, the regulatory mechanisms of DUOX activity are less understood. Using Caenorhabditis elegans as a model, we demonstrate that the tetraspanin protein is required for induction of the DUOX signaling pathway in conjunction with the dual oxidase maturation factor (DUOXA). In the current study, we show that genetic mutation of DUOX (bli-3), DUOXA (doxa-1), and peroxidase (mlt-7) in C. elegans causes the same defects as a tetraspanin tsp-15 mutant, represented by exoskeletal deficiencies due to the failure of tyrosine cross-linking of collagen. The deficiency in the tsp-15 mutant was restored by co-expression of bli-3 and doxa-1, indicating the involvement of tsp-15 in the generation of ROS. H(2)O(2) generation by BLI-3 was completely dependent on TSP-15 when reconstituted in mammalian cells. We also demonstrated that TSP-15, BLI-3, and DOXA-1 form complexes in vitro and in vivo. Cell-fusion-based analysis suggested that association with TSP-15 at the cell surface is crucial for BLI-3 activation to release H(2)O(2). This study provides the first evidence for an essential role of tetraspanin in ROS generation.

Figure 1. bli-3 and doxa-1 mutants are similar to the tsp-15 mutant.

(A) The structure of the gene/proteins related to tsp-15 function. Schematic representation of the BLI-3 and MLT-7 protein with functional domain, and the genomic structure of the doxa-1 gene. The im10, im21, im32, im38 and im39 mutations are indicated. Previously identified missense mutations in the bli-3 gene including e767 (glycine to aspartic acid at 246) and n529 (aspartic acid to asparagine at 392) are shown. The bold line indicates the region of the gk141 deletion allele. The im10 mutation has a leucine instead of a proline at position 1311 within the NOX domain. TM and NOX refer to the transmembrane and NOX domains, respectively. The im21 mutation is characterized by a G to A transition in the splice donor site at the fifth intron. The im32 mutation is a G to T transversion in the splice acceptor site at the fourth intron. The im38 and im39 alleles are indicated in the MLT-7 protein. Both alleles contain the missense mutations in the peroxidase domain. (B) bli-3(im10) and doxa-1(im21), but not bli-2(e768) are similar to tsp-15(sv15). Hypodermal expression of GFP driven by dpy-7p::gfp in the mutants revealed an unusual accumulation of cellular materials in the blisters of bli-3, doxa-1 and tsp-15 mutants (indicated by black arrows), but not in bli-2 mutants (indicated by the white arrow). The scale bars represent 50 µm.

Figure 2. TSP-15 function is compensated with BLI-3 system.

Restoration of the tsp-15 defect by co-expression of bli-3 and doxa-1. Co-expression of bli-3 and doxa-1 in the tsp-15(sv15) hypomorph mutant effectively rescued the phenotype of sv15. The embryonically lethal tsp-15(ok881) null mutant was partially rescued and developed into larvae via bli-3 and doxa-1 co-expression. Green fluorescence at the tail tip represents expression of the lin-44::gfp injection marker. The ok881 deletion homozygosity of surviving transgenic larvae was confirmed by genomic PCR. The scale bars represent 50 µm.

Figure 3. Deterioration of dityrosine in the tsp-15 mutant.

(A) Representative immunofluorescent images showing the distribution of dityrosine in embryos of the tsp-15(ok854) null mutant. Embryos were obtained from the OB129 strain, which was the tsp-15(ok854) mutant rescued by a tsp-15p::(His)⁶Xp::tsp-15 extrachromosomal array. Nuclear GFP fluorescence by sur-5::gfp defined the rescued (tsp-15(+)) or spontaneously array-lost (null; tsp-15(0)) embryo. Micrographs on the left show merged Nomarski images exhibiting GFP and dityrosine immunolocalization. Right panels show the reconstruction of confocal images for dityrosine distribution in the same embryo that is displayed on the left. In tsp-15(+) normal embryos, dityrosine localization showed a regular pattern representing the cuticle surface structure. Fluorescence intensity was severely deteriorated in tsp-15(0) embryos. Scale bars indicate 10 µm. (B) DPY-7 localization was compared under the same conditions as in (A). In tsp-15(+) embryos, DPY-7 localized as regular bands in the cuticle. In the tsp-15(0) embryo, the expression of DPY-7 was comparable to the normal embryo despite its disorganized pattern. Scale bars indicate 10 µm.

Figure 4. Both TSP-15 and DOXA-1 are required for H₂O₂ production by BLI-3 in mammalian cells.

(A) Immunoblot analysis of the expression of Xpress-tagged TSP-15, FLAG-tagged DOXA-1, BLI-3 and BLI-3^P1311L in HT1080 stable transfectants. Xpress-tagged TSP-15 (30 kDa) is highly glycosylated (Figure S5). (B) Extracellular H₂O₂ production from stable transfectants. Fold-activation compared with non-transfected HT1080 cells was determined. Only cells expressing TSP-15, DOXA-1, and BLI-3 (HT1080^TDB) significantly generated H₂O₂. A 10 µM concentration of DPI inhibited H₂O₂ production in HT1080^TDB cells. BLI-3 carrying the G246D or P1311L mutation did not release H₂O₂. The graph shows the means ± SEM. The numbers in the bars indicate the number of independent experiments (n). *P<10⁻⁹. (C) BLI-3 was not activated by mammalian DUOX stimulators. Ionomycin (iono, 1 µM), forskolin (Fsk, 1 µM), or phorbol 12-myristate 13-acetate (PMA, 1 µM) was added to HT1080^DB and HT1080^TDB cells during culture. The graph shows the means ± SEM. The numbers in the bars indicate the number of independent experiments (n).

Figure 5. Direct association of BLI-3 with TSP-15 and DOXA-1.

(A) Direct association of BLI-3 with TSP-15, and BLI-3 with DOXA-1. BLI-3 was transiently expressed in COS-7 stable transfectants expressing Xpress::TSP-15, DOXA-1::FLAG, or both. Cell surface proteins were labeled with biotin. A 1% CHAPS cell lysate was used for immunoprecipitation or pull-down assay with anti-TSP-15 antibody, anti-FLAG antibody or streptavidin. BLI-3 was co-immunoprecipitated with both TSP-15 (indicated by asterisks) and DOXA-1, and TSP-15 and DOXA-1 were co-immunoprecipitated. Bands above asterisks are non-specific. Cell surface localization of BLI-3 was independent of TSP-15 and DOXA-1. ER-resident calnexin (CANX) was not detected on the cell surface. BLI-3 expression was not affected by TSP-15 and DOXA-1. (B) Direct association of BLI-3 with TSP-15, BLI-3 with DOXA-1, and TSP-15 with DOXA-1 was confirmed in C. elegans. Xpress::tsp-15 was expressed in OB129, and doxa-1::venus is expressed in the OB218 transgenic strain. Venus-tagged DOXA-1 was analyzed with anti-GFP antibody. BLI-3 was co-immunoprecipitated with endogenous and Xpress-tagged TSP-15 and Venus-tagged DOXA-1 in 1% Triton-X100 cell lysates. Endogenous TSP-15 also associated with DOXA-1::Venus. Normal rat and mouse IgG was used as a negative control.

Figure 6. Requirement of TSP-15 for reconstitution of BLI-3 function at cell surface.

(A) HVJ-mediated cell fusion. GFP-expressing HT1080 cells and HT1080^DB cells labeled with Cell Tracker Orange were fused with HVJ. Under HVJ(+) conditions, fused cells were large compared with HVJ(−) cells and exhibited yellow/orange fluorescence. Scale bar indicates 50 µm. (B) HT1080^DB cells fused with HT1080^T cells (T::DB) produced H₂O₂, which was inhibited by DPI. Mock::DB and T::DB^P1311L fusion cells did not produce H₂O₂. Treatment of T::DB fusion cells with 10 µg/ml cycloheximide (CHX) did not inhibit H₂O₂ production. The graph shows the means ± SEM. The number of independent experiments is indicated. *P<10⁻⁶. (C) Rapid H₂O₂ production from T::DB fusion cells. The recovery time after the fusion event was examined to determine when fusion cells acquired the ability to produce H₂O₂. Maximum H₂O₂ production was observed at 30–60 min, although production was observed at 15 min post-fusion. The graph shows the means ± SEM (n = 3).

Figure 7. Molecular regulation of BLI-3 by TSP-15 and DOXA-1.

TSP-15 and DOXA-1 are essential for H₂O₂ production by BLI-3. TSP-15 associates with BLI-3 at the cell surface or during trafficking. The role of DOXA-1 in BLI-3 targeting to the plasma membrane remains elusive. BLI-3/DOXA-1 complexes at the cell surface are inactive, but recruiting to the tetraspanin-microdomain facilitates the formation of a functional unit for generation of H₂O₂ that is utilized by innate host immunity, and cross-linking of extracellular matrix with peroxidase (MLT-7).