Targeting and regulation of reactive oxygen species generation by Nox family NADPH oxidases

Abstract

Nox family NADPH oxidases serve a variety of functions requiring reactive oxygen species (ROS) generation, including antimicrobial defense, biosynthetic processes, oxygen sensing, and redox-based cellular signaling. We explored targeting, assembly, and activation of several Nox family oxidases, since ROS production appears to be regulated both spatially and temporally. Nox1 and Nox3 are similar to the phagocytic (Nox2-based) oxidase, functioning as multicomponent superoxide-generating enzymes. Factors regulating their activities include cytosolic activator and organizer proteins and GTP-Rac. Their regulation varies, with the following rank order: Nox2 > Nox1 > Nox3. Determinants of subcellular targeting include: (a) formation of Nox-p22(phox) heterodimeric complexes allowing plasma membrane translocation, (b) phospholipids-binding specificities of PX domain-containing organizer proteins (p47(phox) or Nox organizer 1 (Noxo1 and p40(phox)), and (c) variably splicing of Noxo1 PX domains directing them to nuclear or plasma membranes. Dual oxidases (Duox1 and Duox2) are targeted by different mechanisms. Plasma membrane targeting results in H(2)O(2) release, not superoxide, to support extracellular peroxidases. Human Duox1 and Duox2 have no demonstrable peroxidase activity, despite their extensive homology with heme peroxidases. The dual oxidases were reconstituted by Duox activator 2 (Duoxa2) or two Duoxa1 variants, which dictate maturation, subcellular localization, and the type of ROS generated by forming stable complexes with Duox.

FIG. 1.

Generation of reactive oxygen species (ROS) by Nox family NADPH oxidases. All are electrogenic enzymes that accept electrons from cytosolic NADPH, transport them through FAD and membrane-imbedded hemes, and donate single electrons to molecular oxygen, thereby producing superoxide anion. In the case of the dual oxidases (Duox1 and Duox2), a superoxide intermediate is not readily detected and the N-terminal peroxidase-like domain appears to affect its conversion into H₂O₂ by undefined mechanisms. The most structurally conservation features of these enzymes include regions of the C-terminal reductase domain that bind NADPH and FAD and the membrane-spanning helical segments thought to bind heme.

FIG. 2.

Molecular components of active Nox family NADPH oxidase complexes. Nox1, Nox2, and Nox3 function as regulated enzymes involving cytosolic adaptor proteins or “Nox organizers” (p47^phox or Noxo1 and p40^phox) and “Nox activators” (p67^phox or Noxa1) that bind GTP-Rac and affect the flow of electrons. The p22^phox component forms a stable heterodimeric complex with Nox core components (Nox1–4), required for post-translation processing or maturation into active oxidases. In Nox1–Nox3 systems, p22^phox also promotes plasma membrane targeting of the oxidases and provides a docking site for Nox organizers. Nox5 and Duox are calcium-responsive oxidases that contain calcium-binding EF-hands. The Duox activators (Duoxa) are maturation factors functionally similar to p22^phox recently shown to form stable complexes with Duox on the plasma membrane (72).

FIG. 3.

Nox 1 and Nox 3 expression enables plasma membrane targeting and stabilization of p22^phox. Immunofluorescence imaging of endogenous (A) and transfected (B) p22^phox in HEK293 cells, showing reticular (ER) and nuclear membrane (arrows) staining patterns. (C) Transfection of Nox1 results in a redistribution of endogenous p22^phox to the plasma membrane. (D) A similar redistribution of p22^phox occurs in Nox3-transfected HEK293 cells. The untransfected cells (*) show primarily cytosolic staining patterns and overall weaker staining. Western blot analysis shows increased endogenous p22^phox levels in Nox1- or Nox3- transfected cells (E), but not in mock or Noxo1β-transfected HEK293 cells (F). (Modified with permission from Fig. 6 of (100) and Fig. 6 of Ref. (102)).

FIG. 4.

Schematic representation of multiple modular domains in Nox organizer and activator proteins in phagocytic (A) and non-phagocytic systems (B). The Nox organizers (p47^phox and Noxo1) share common structures and functional properties of binding to phosphatidic acid (PA) and phosphoinositide lipids through PX domains, to p22^phox through two SH3 domains, and to SH3 domains of Nox activators through the proline-rich (PR) motifs at their carboxyl termini. The Nox activators (p67^phox and Noxa1) also share homologous tetratricopeptide repeat (TRP) scaffolds that present Rac binding sequences and activation domains (AD), and PB1 (Phox and Bem1) domain. The p40^phox component is a secondary adaptor that bridges contacts between membrane (PI(3)P) and p67^phox, using PX (phox) domain and PB1 domain heterodimerlization, respectively. Auto-inhibitory intramolecular interactions that maintain closed conformations in p47^phox and p40^phox are depicted with dashed lines. Striped box in p47^phox between AIR and PR motif represents residues 341–360, which enhance the autoinhibitory interaction between the PX domain and the structure constructed by the tandem SH3s+ the AIR (101).

FIG. 5.

Schematic representation of conformations changes in p47^phox and p40^phox allowing assembly of the ternary cytosolic complex into the active oxidase during the course of FcγR-mediated phagocytosis. The p47^phox component acts early-phase adaptor protein following PKC-dependent phosphorylation-induced conformational change by binding to p22^phox and membrane PA and PI(3,4)P₂. In contrast, p40^phox serves in the retention of the cytosolic complex on sealed phagosomes in later phases when PA and PI(3,4)P₂ have disappeared, due to its PX domain-PI(3)P and PB1 domain-p67^phox interactions (modified with permission from Fig. 11 of ref. 104). In the early phase, class I PI3-kinase (106) and PLD2 (15) serve in production of PI (3,4)P₂ and PA, while in the later phase class III PI3-kinase (106) and early endosomes (43) serve in accumulation of PI (3)P.

FIG. 6.

Predicted three-dimensional structures of alternatively spliced PX domain variants of Noxo1, based on alignments with sequences of p47^phox and other PX domain-containing proteins. (A) Alignment of PX domain amino acid sequences, showing locations of splice site 1 (deleted from Noxo1α and δ) in the middle of conserved alpha-helix 1 and splice site 2 (present in Noxo1γ and δ) inserted within a variable loop region frequently containing the proline-rich motif in other PX domains. Conserved residues involved in p47^phox binding to PA are marked with an asterisk (*) while those that bind phosphoinositide lipids are overlined (modified with permission from Fig. 2 of ref. 102). (B) Structural models of the Noxo1 isoforms based on the structure of p47^phox (PDB ID: 1KQ6). The critical residues of the PA binding site are unchanged between the β and γ isoforms, despite the nearby insertion five amino acids in Noxo1γ. The electrostatic character of the PA binding site is significantly altered in the α isoform by the replacement of lysine by glutamic acid (red arrow heads). The PI binding site is basically unchanged between p47^phox and the Noxo1 isoforms (blue arrows). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article at www.liebertonline.com/ars).

FIG. 7.

Subcellular localization of PX domain and full-length Noxo1 alternative splice variants in HEK293 (A) and COS7 (B) cells. Noxo1α(PX)-GFP and Noxo1α-GFP show a tendency to aggregate or are localized on intracellular vesicles. Noxo1β(PX)-GFP and Noxo1β-GFP are localized predominately along the plasma membrane. Noxo1γ(PX)-GFP and Noxo1γ-GFP localize in the nucleus and along the plasma membrane. Noxo1δ(PX)-GFP and Noxo1δ-GFP show localization patterns similar to that of Noxo1α. Bar: 10 μm. (modified with permission from Fig. 2 of ref. 102).

FIG. 8.

Laser scanning confocal immunofluorescence detection of Duox on the apical plasma membrane surface of ciliated primary human bronchial epithelial cells. Images show the apical X-Y plane of confluent cells grown on permeable transwells and re-differentiated for 25 days in an air–liquid interface culture system, as described (84). Ciliated cells, stained with anti β-tubulin antibody (red), are the same as those stained positive for Duox (green). 3-D reconstructed images depicting the rotation of the X-Y plane are available in the supplemental online video, showing Duox and β-tubulin accumulation within the apical aspect of this polarized cell layer. (See online supplemental video at www.liebertonline.com). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article at www.liebertonline.com/ars).

FIG. 9.

Role of Duox maturation factors in Duox isozyme biosynthesis and ROS-generating mechanism. DuoxA2 and DuoxA1α have similar sizes and predicted N-glycosylation sites. Duoxa1γ has an extended C-terminal sequence similar to that of Drosophila NIP (numb interacting protein). Duoxa1β and δ forms (bottom left) that lack exon 3-encoded sequence and two predicted gylcosylation sites remain in the ER and do not support biosynthesis of active Duox. Active Duox maturation factors (Duoxa1α, γ, Duoxa2) translocate to the plasma membrane, undergoing Golgi-based carbohydrate modifications and forming stable H₂O₂-generating Duox/Duoxa complexes (upper left). Less competent combinations of Duox2 overexpressed with Duoxa1α or Duoxa1γ do not form plasma membrane Duoxa complexes, show less carbohydrate processing, and produce superoxide (upper right).