Duox maturation factors form cell surface complexes with Duox affecting the specificity of reactive oxygen species generation

Abstract

Dual oxidases (Duox1 and Duox2) are plasma membrane-targeted hydrogen peroxide generators that support extracellular hemoperoxidases. Duox activator 2 (Duoxa2), initially described as an endoplasmic reticulum resident protein, functions as a maturation factor needed to deliver active Duox2 to the cell surface. However, less is known about the Duox1/Duoxa1 homologues. We identified four alternatively spliced Duoxa1 variants and explored their roles in Duox subcellular targeting and reconstitution. Duox1 and Duox2 are functionally rescued by Duoxa2 or the Duoxa1 variants that contain the third coding exon. All active maturation factors are cotransported to the cell surface when coexpressed with either Duox1 or Duox2, consistent with detection of endogenous Duoxa1 on apical plasma membranes of the airway epithelium. In contrast, the Duoxa proteins are retained in the endoplasmic reticulum when expressed without Duox. Duox1/Duoxa1alpha and Duox2/Duoxa2 pairs produce the highest levels of hydrogen peroxide, as they undergo Golgi-based carbohydrate modifications and form stable cell surface complexes. Cross-functioning pairs that do not form stable complexes produce less hydrogen peroxide and leak superoxide. These findings suggest Duox activators not only promote Duox maturation, but they function as part of the hydrogen peroxide-generating enzyme.

Figure 1.

Alternate mRNA splice variants of DUOXA1. A) Schematic of alternatively spliced DUOXA1 ORFs, showing exon sizes. Arrowheads represent alternative splicing sites; black boxes show distinct C-terminal coding sequences; arrows denote positions of PCR primers. B) PCR amplification (41 cycles) of short (α,β) and long (γ, δ) transcripts from human tissue cDNA sources. β and δ transcripts lack the third coding exon. C) Predicted proteins encode 5 transmembrane sequences. Black circles delineate varying C-terminal sequences; gray denote exon 3-encoded sequence (45 aa) including two N-glycosylation sites (Y) absent from β and δ forms. Dotted circles represent peptide (Duoxa1α aa 122–139) used to raise the anti-Duoxa1 serum.

Figure 2.

Detection of Duoxa1 protein variants. A) Western blot analysis of protein lysates (10 μg) extracted from stable Flp-In 293 clones expressing C-terminal V5-tagged proteins shows long forms (Duoxa1γ and Duoxa1δ) are produced less efficiently than short forms (Duoxa1α and Duoxa1β). Susceptibilities to endoglycosidase H (H) or N-glycosidase F (F) confirm N-glycosylation patterns proposed in Fig. 1C and indicate β and δ forms are incompletely glycosylated. B) Western blot analysis shows anti-Duoxa1 serum specificity for Duoxa1 but not Duoxa2. Protein extracts (10 μg) derived from Flp-In 293 cells transfected with empty, Duoxa1α^V5, or Duoxa2^V5 coding vector. Primary sera were diluted 1:500. C) Duoxa1 Western blot analysis (1:500) shows Duoxa1α is detected in IL-13-induced NHBE cells. Controls included protein lysates from Flp-In 293 stable clones expressing Duoxa1α^V5 or Duoxa1γ^V5. Flp-In 293 (10 μg) and NHBE (75 μg) cell extracts were subjected to N-glycosidase F treatment prior to SDS-PAGE.

Figure 3.

Epitope tagging of Duox1 and Duoxa1 does not affect reconstitution of oxidase activity. A) Ionomycin-stimulated H₂O₂ production [10-min integrated relative fluorescence units (RFU)] from stable Flp-In 293 clones expressing ^HADuox1 protein, transiently transfected (48 h) with tagged and untagged Duoxa1α. Error bars reflect mean ± sd; data show a representative triplicate assay of 4 independent experiments. Right panel: corresponding Western blot analysis of expressed proteins (20 μg). B) Ionomycin-stimulated H₂O₂ production (10-min integrated RFU) from stable Flp-In 293 clones expressing Duoxa1α^V5 protein, transiently transfected (48 h) with tagged and untagged Duox1 proteins. Error bars reflect mean ± sd; data show a representative triplicate assay of 3 independent experiments. Right panel: corresponding Western blot analysis of expressed proteins (20 μg).

Figure 4.

Transfected Duoxa1 is retained in the ER, whereas endogenous Duoxa1 localizes to the cell surface. A) IF staining of permeabilized stable Flp-In 293 clones expressing Duoxa1^V5 variants resembles that of ER marker calnexin. B) IF staining of human tracheal sections shows Duoxa1 colocalizes with the apical plasma membrane marker ezrin along ciliated epithelial cells. Primary sera diluted 1:100.

Figure 5.

Duox1 is retained in the ER when coexpressed with Duoxa1β or δ. A) Ionomycin-stimulated H₂O₂ production (10-min integrated RFU) from stable Flp-In 293 clones expressing ^HADuox1 protein, transiently transfected (48 h) with tagged or untagged Duoxa1 proteins. Error bars reflect mean ± sd; data show a representative triplicate assay of 4 independent experiments. Right panel: corresponding Western blot analysis of expressed proteins (20 μg). B, C) IF staining of permeabilized (B) and unpermeabilized (C) stable Duoxa^V5-^HADuox Flp-In 293 clones.

Figure 6.

Active Duox maturation factors are transported to the plasma membrane when coexpressed with Duox. A) IF staining of permeabilized stable Duoxa^V5-^HADuox Flp-In 293 clones reveals distinct localization patterns for each Duoxa variant. B) Cell surface IF staining of unpermeabilized ^HADuox1 or ^HADuox2 Flp-In 293 clones transiently transfected (48 h) with ^V5Duoxa coding vectors.

Figure 7.

Active Duoxa1α/Duox1 and Duoxa1γ/Duox1 combinations translocate to the cell surface. A, B) IF staining of permeabilized (A) and unpermeabilized (B) COS-7 cells transiently cotransfected (48 h) with ^FLAGDuoxa1 and ^HADuox1 coding vectors. Pictures were acquired on a Zeiss LSM 510 confocal laser-scanning microscope with an ×40 oil-immersion objective, NA 1.4. Anti-HA (MBL, Woods Hole, MA, USA) and anti-FLAG (Sigma) antibodies were diluted 1:1000. C) IF staining of untagged Duoxa1α, transiently expressed (48 h) in ^HADuox1 Flp-In 293 clone, detects Duox activator at the cell surface of unpermeabilized cells (left). When transiently expressed in Flp-In 293 cells, untagged Duoxa1α is not detected at the plasma membrane (right). Anti-Duoxa1 serum diluted 1:100.

Figure 8.

Duox maturation and reconstitution correlate with Golgi apparatus-based carbohydrate modifications on both Duox and Duoxa. A) Flow cytometric analysis of cell-surface-exposed ^HADuox on stable Duoxa^V5-^HADuox Flp-In 293 clones. Numbers indicate fluorescence geometric mean. Flp-In 293 clone stably transfected with ^HADuox1 was used as negative control (shaded peak). B) Ionomycin-stimulated (▪, ×) or unstimulated (−) H₂O₂ release by corresponding cloned lines in A. Oxidase activity was inhibited by preincubation (10 min) with 10 μM DPI (×). Arrow, 1 μM ionomycin stimulation. Error bars reflect mean ± sd of a representative triplicate assay. Activities in parenthesis derived from 3 experiments performed in triplicate. C) Western blot analysis of N-deglycosylation-treated protein lysates (10 μg) extracted from stable Duoxa^V5-^HADuox Flp-In 293 clones.

Figure 9.

Duox maturation factors affect specificity of reactive oxygen species generation and form stable complexes with Duox on cell surface. A) Kinetics of extracellular O₂^·− release by Diogenes-based luminescence assay (RLU, relative luminescence unit) following 1 μM ionomycin stimulation (arrow) from stable Duoxa^V5-^HADuox Flp-In 293 clones. Error bars reflect mean ± sd of a representative triplicate assay of 3 independent experiments. B) Bar graph comparing extracellular O₂^·− release activities from stable Duoxa^V5-^HADuox Flp-In 293 clones detected by Diogenes (10 min integrated RLU; ▪) from A and cytochrome c reduction (10 min assay; □) shows O₂^·− production by cross-functioning Duoxa1α or γ /Duox2 pairs. Activities represent means of 3 experiments performed in triplicate. C) Western blot analysis does not detect Duoxa1α and Duoxa1γ immunoprecitated in a complex with surface-exposed ^HADuox2. Stable Duoxa^V5-^HADuox Flp-In 293 clones were incubated 30 min at 4°C with anti-HA before lysis. Surface [anti-HA-^HADuox] complexes were then immunoprecipitated (HA-IP) with protein G Sepharose beads from total lysates (Tot). Negative controls included Flp-In 293 cells stably expressing either Duoxa1α^V5 or ^HADuox1 alone. Data show representative results from one of three experiments.