| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
The Julius Friedrich Cohnheim-Minerva Center for Phagocyte Research and the Ela Kodesz Institute of Host Defense against Infectious Diseases, Sackler School of Medicine, Tel Aviv University, Israel
1Correspondence: Department of Human Microbiology, Sackler School of Medicine, Tel Aviv University, Tel Aviv 69978, Israel. E-mail: epick{at}post.tau.ac.il
ABSTRACT
Phagocytes generate superoxide (O2.–) by an enzyme complex known as reduced nicotinamide adenine dinucleotide phosphate (NADPH) oxidase. Its catalytic component, responsible for the NADPH-driven reduction of oxygen to O2.–, is flavocytochrome b559, located in the membrane and consisting of gp91phox and p22phox subunits. NADPH oxidase activation is initiated by the translocation to the membrane of the cytosolic components p47phox, p67phox, and the GTPase Rac. Cytochrome b559 is converted to an active form by the interaction of gp91phox with p67phox, leading to a conformational change in gp91phox and the induction of electron flow. We designed a new family of NADPH oxidase activators, represented by chimeras comprising various segments of p67phox and Rac1. The prototype chimera p67phox (1–212)-Rac1 (1–192) is a potent activator in a cell-free system, also containing membrane p47phox and an anionic amphiphile. Chimeras behave like bona fide GTPases and can be prenylated, and prenylated (p67phox-Rac1) chimeras activate the oxidase in the absence of p47phox and amphiphile. Experiments involving truncations, mutagenesis, and supplementation with Rac1 demonstrated that the presence of intrachimeric bonds between the p67phox and Rac1 moieties is an absolute requirement for the ability to activate the oxidase. The presence or absence of intrachimeric bonds has a major impact on the conformation of the chimeras, as demonstrated by fluorescence resonance energy transfer, small angle X-ray scattering, and gel filtration. Based on this, a "propagated wave" model of NADPH oxidase activation is proposed in which a conformational change initiated in Rac is propagated to p67phox and from p67phox to gp91phox.
Key Words: superoxide • oxygen radicals • respiratory burst • small GTPases • Rac • GTP/GDP • p67phox • p47phox
OVERTURE
A major mechanism by which phagocytic cells kill pathogenic microorganisms is the production of toxic reactive oxygen species. These are all derived from the primordial oxygen radical, superoxide (O2.–), which is produced, in response to appropriate stimuli, by a tightly regulated enzyme complex, known as the O2.–- generating reduced nicotinamide adenine dinucleotide phosphate (NADPH) oxidase (briefly, "oxidase"; reviewed in ref. [1 ]). The catalytic function of the oxidase consists of the NADPH-driven, one-electron reduction of molecular oxygen, leading to the formation of O2.–. The functionally competent oxidase complex is composed of a membrane-associated flavocytochrome b559 (a heterodimer consisting of two subunits, gp91phox and p22phox) and four cytosolic components: p47phox, p67phox, p40phox, and the small GTPase Rac (1 or 2; reviewed in ref. [2 ]). The component responsible for electron transport form NADPH to oxygen is gp91phox, which contains the NADPH-binding site and two redox stations, represented by flavin adenine dinucleotide and two nonidentical hemes. In resting phagocytes, the components of the complex to be formed exist as distinct entities. NADPH oxidase activation is the consequence of the interaction of cytochrome b559 with cytosolic components, requiring translocation of the latter to the membrane environment of the former. This process rests on a complex set of protein-protein interactions and is defined as oxidase assembly (reviewed in ref. [3 ]). Many of the component-to-component interactions have been dissected in great detail and their structural basis established by a variety of techniques culminating in X-ray crystallography (reviewed in ref. [4 ]).
The paramount interaction is that of gp91phox with one or more of the cytosolic components, resulting in a conformational change in gp91phox, which enables the electron flow from NADPH to oxygen and the generation of O2.–. The prevalent idea is that the cytosolic component solely or principally responsible for an activating interaction with gp91phox is p67phox and that this involves p67phox residues 199–210, known as the "activation domain" [5 ]. However, binding of Rac to gp91phox and its direct involvement in the regulation of electron flow were also proposed (ref. [6 ]; for review, see ref. [7 ]).
As far as a possible "activating" role of p47phox is concerned, there is a considerable body of evidence for protein-protein interactions between p47phox and gp91phox, but no direct effect of these interactions on electron flow in gp91phox has been demonstrated (ref. [8 ]; for review, see refs. [2 , 3 ]). The binding of p47phox to p22phox is one of the best-studied events in oxidase activation and involves the tandem src homology 3 (SH3) domains of p47phox and a consensus poly-proline motif at the C terminus of p22phox [9 ]. In the resting state, this interaction is prevented by the existence of an intramolecular bond between the SH3 domains and a polybasic region at the C terminus of p47phox. This autoinhibited state is relieved by phosphorylation of critical serines within the polybasic stretch, allowing the SH3 domains to bind to the proline-rich region of p22phox [10 ]. The only consequence of the binding of p47phox to p22phox appears to be the facilitation of the interaction of p67phox with gp91phox, with p47phox serving as a carrier for p67phox by virtue of tail-to-tail interactions. There is no experimental evidence for the induction of a secondary, conformational change in gp91phox following the interaction of p47phox with p22phox.
Lambeth and collaborators [11 ] and our group [12 ] have put forward the hypothesis that p67phox is the only cytosolic component responsible for an activating interaction with gp91phox, but as a result of the fact that p67phox does not possess a membrane-attachment signal of its own, it requires the assistance of Rac and p47phox. The roles of these two components in enabling the productive association of p67phox with gp91phox are not interchangeable. The differences in the "assistance" provided to p67phox by Rac and by p47phox will be discussed below.
Understanding the mechanism of activation of the phagocyte oxidase complex gained in general relevance by the discovery of a family of homologues of gp91phox, known as the NADPH oxidase (NOX)/dual oxidase enzymes (DUOX), found in a variety of nonphagocytic cells from humans (reviewed in refs. [2 , 13 ]) to Drosophila [14 ]. The NOX proteins function as components of enzyme complexes generating oxygen radicals with a variety of roles, in addition to host defense, such as oxidative modification of proteins, thyroid hormone synthesis, oxygen sensing, regulation of cell proliferation, and induction of apoptosis and angiogenesis. Recently, homologues of p47phox, known as Nox organizer 1, and of p67phox, known as Nox activator 1, were described in colon epithelial cells, and evidence became available that p22phox is required for the activity of most NOX enzymes (reviewed in refs. [2 , 13 ]).
A discovery that led to rapid advances in the identification of the components of the oxidase complex and in the understanding of their mechanism of assembly was the design of a cell-free system, which mimicked stimulus-elicited oxidase activation in intact cells. This consisted of phagocyte membranes or purified Cytochrome b559 and the cytosolic components p47phox, p67phox, and Rac (1 or 2), to which a critical concentration of arachidonic acid or other long-chain, unsaturated fatty acids was added [15 16 17 18 ]. It was next found that the activator does not have to be a fatty acid and that what endows a compound with activating capacity is its anionic amphiphile character [19 ]. Thus, sodium or lithium dodecyl sulfates are excellent inducers of activation [19 ]. There is good evidence for anionic amphiphiles acting by causing a conformational change in p47phox [20 ], imitating those induced by phosphorylation of C-terminal serines and possibly also by the induction of structural changes in cytochrome b559 [21 ]. Cell-free oxidase activation can also take place in the absence of an anionic amphiphilic activator under the following three conditions: truncation of p47phox, resulting in the removal of the polybasic autoinhibitory domain, and of p67phox, resulting in the removal of both SH3 regions [22 ]; prenylated Rac1 combined with p67phox in the presence or absence of p47phox [12 ]; and enrichment of the lipid membrane environment of cytochrome b559 with anionic phospholipids (ref. [23 ]; A. Mizrahi, Y. Berdichevsky, Y. Ugolev, S. Molshanski-Mor, E. Pick, unpublished results). The essential differences between the amphiphile-dependent and -independent pathways of oxidase activation and their relationship to the participation or lack of participation of p47phox in the two processes are reviewed in ref. [24 ].
THE POOR MAN’S NADPH OXIDASE—COMPLEX ASSEMBLY IN THE ABSENCE OF p47phox
Up to a decade ago, the accepted paradigm was that p47phox and p67phox are of similar importance in the activation of the oxidase. This view originated in part from the fact that patients with chronic granulomatous disease, lacking or expressing loss-of-function mutations in any of the two components, have impaired oxidase activity and consequent pathology [25 ]. An asymmetrical view of the respective roles of p47phox and p67phox was born in findings made in two laboratories, demonstrating that the phagocyte oxidase can be activated in vitro in the absence of p47phox [26 , 27 ]. Activation was amphiphile-dependent and required the presence of p67phox and Rac at concentrations much larger than those required for activation in the presence of p47phox. As opposed to the high level of oxidase activity found with the combination p67phox+ Rac, no activity was detected with the combination p67phox+ p47phox. This was surprising in view of the body of evidence showing that translocation of p67phox to the membrane was absolutely dependent on p47phox (reviewed in refs. [2 , 3 ]). These findings suggested that p67phox can reach the membrane in the absence of p47phox; association of p67phox with the membrane (a term that leaves open the question of the precise topography of such association) is required but not sufficient for productive interaction with gp91phox, and Rac can provide the services of membrane tether or carrier for p67phox and in addition, juxtapose p67phox with gp91phox and/or induce a conformational change in p67phox, enabling productive interaction with gp91phox.
Further support for a central role of the Rac-p67phox tandem in NADPH oxidase activation was provided by our finding that equimolar mixtures of prenylated Rac1 and p67phox or p67phox truncated at residue 212 (missing both SH3 domains and the proline-rich region) activate the oxidase in vitro in the absence of p47phox and amphiphile [12
]. Contrary to the initial impression, optimal activation requires the prenylated Rac1 to be in the guanosine 5'-triphosphate (GTP)-bound form. This can be achieved by exchange to a nonhydrolyzable form of GTP such as guanosine-5'-O-(3-thiotriphosphate; GTP
S) or guanylyl-imidodiphosphate (GMPPNP), the use of the Q61L mutant of Rac1, or enzymatic exchange to GTP by the guanine nucleotide exchange factors (GEF) Trio or Tiam1 [28
]. The ability of prenylated Rac1-GTPS to form a nucleus for oxidase assembly and activation is illustrated in Figure 1
. The most poignant aspect of this experiment is the ability to elicit NADPH-dependent O2.– production by the mere addition of p67phox (but not of p47phox) to membrane liposomes "coated" with prenylated Rac1-GTPS in the absence of any activator (Fig. 1B)
. Work in progress now reveals yet another way to activate the oxidase in the absence of p47phox and amphiphile. This consists of the artificial enrichment of the phagocyte membrane lipid with anionic phospholipids, such as phosphatidic acid or phosphatidyl glycerol, and their exposure to p67phox and nonprenylated Rac1. In this case, p47phox- and amphiphile-independent activation is based on the electrostatic attraction of Rac1, via its polybasic domain, to the anionic phospholipid in the membrane, which substitutes for and relieves the need for the prenylation of Rac (see the penultimate section of this overview).
|
In spite of positive reports in the literature [29
30
31
], many investigators found it difficult to express the association of p67phox with Rac in quantitative terms. A dissociation constant in the 2-µM range was found using isothermal titration calorimetry [31
]. To be able to concentrate on the study of the association of p67phox with cytochrome b559, we decided to design and produce recombinant fusion proteins consisting of selected segments of p67phox and Rac1. Such p67phox-Rac1 chimeras contain, by definition, an unbreakable bond between the two oxidase components, and we hoped that this will facilitate revealing the mechanisms of membrane attachment of and gp91phox "activation" by p67phox. The group of Tamura [32
], who generated a p67phox (1–210)-p47phox (1–286) chimera capable of oxidase activation, first reported NADPH oxidase chimeras. The group of Tamura [33
] also described the generation of a p67phox (1–210)-Rac1 (1–192) chimera [33
], and we reported the construction of 
20 variations of p67phox-Rac1 chimeras, designed to allow the dissection of the role of functionally significant residues or domains in both moieties of the chimera [34
, 35
]. Figure 2
lists the first of group of p67phox-Rac1 chimeras generated by us. These were examined for the ability to activate the oxidase in a cell-free system consisting of chimera, exchanged to a nonhydrolyzable GTP analog (when the length of the Rac1 moiety permitted such exchange), p47phox, and the anionic amphiphile lithium dodecyl sulfate.
|
We initially hypothesized that in the presence of full-length Rac1, chimeras containing only the activation domain of p67phox will be activating. It was found, however, that the presence of p67phox residues 199–212 was not sufficient for endowing the chimera with activating ability. Extending the p67phox sequence to residues 155–212 still resulted in an inactive chimera, proving that more than just the activation domain is required for oxidase activation by p67phox-Rac chimeras (see Fig. 2 ). This requirement will be studied in more detail in the next section.
An in-depth analysis of the effect of positive charge at the C terminus of p67phox-Rac1 chimeras on oxidase-activating capacity appears in ref. [35 ]. In this study, we mutated from one to six basic residues to neutral glutamines. As apparent in Figure 3 , a reduction in even a single basic residue (chimera 9) leads to a severe loss in oxidase-activating ability. Chimeras with three or fewer basic residues are totally inactive.
|
"WAHLVERWANDSCHAFTEN" (ELECTIVE AFFINITIES)—THE REQUIREMENT FOR INTRACHIMETRIC BONDS
"This affinity (the Captain replied) is sufficiently striking in the case of alkalis and acids, which although they are mutually antithetical, and perhaps because they are so, most decidedly seek and embrace one another, modify one another, and together form a new substance."—Johann Wolfgang von Goethe, "Die Wahlverwandschaften" (Elective Affinities), 1809
In the light of the findings that p67phox-Rac chimeras, lacking residues 1–154 in the p67phox moiety or residues 1–178 in the Rac1 moiety, are inactive, we proposed that for activity, domains participating in the canonical intermolecular interactions between p67phox and Rac1 have to be present and engaged in intramolecular (intrachimeric) interactions between the two moieties. These domains are the TPR motifs in the p67phox moiety [31 , 37 ] and the switch I and preswitch I region in the Rac1 moiety [31 , 38 ]. We, thus, proposed that p67phox-Rac chimeras exist in two conformations: a "closed" one, in which the TPR domains of the p67phox moiety interact with switch I residues in the GTP-bound form of the Rac1 moiety, and an "open" one, in which the two moieties do not interact with each other. The schematic view of these two conformations, shown in Figure 4 , should not be interpreted too literally; clearly the two situations are not of the "all or none" type but rather, the expression of a more compact and more extended shape, respectively.
|
|
S-bound form of chimera 3 [34
]. PAK competes with p67phox for binding to Rac, and inhibition by PAK would only be expected if it were assumed that intrachimeric bonds are important for activity and are disrupted by PAK. BACK TO NATURE—MAKING CHIMERAS STICKY
An essential condition for making conclusions drawn from in vitro work with p67phox-Rac1 chimeras applicable to the in vivo reality was to modify the C terminus by prenylation. In the situation that the C terminus ended in the sequence CLLL, the physiologic modification was geranylgeranylation. This was applied to chimeras 1–3 and to the chimera 3 mutants listed in Figure 5
. Prenylation was performed in vitro with the aid of recombinant geranylgeranyltransferase type I (for details, see ref. [36
]). The C terminus of the prenylated constructs resembled that of Rac1 prenylated in vivo, with the exception that cysteine 189 was not carboxyl-methylayed, and residues 190–192 were not removed by proteolysis. We found that prenylated chimera 3, exchanged to GTPS, is a potent oxidase activator in the absence of p47phox and an amphiphilic activator [36
]. Prenylated chimeras p67phox (1–212)-Rac1 (189–192; chimera 1) and p67phox (1–212)-Rac1 (178–192; chimera 2) were inactive. We found that oxidase activation by prenylated chimera 3 was accompanied by its spontaneous association with membrane liposomes in vitro (Fig. 6
). However, membrane association, although a prerequisite for oxidase activation, was not sufficient. Thus, as apparent in Figure 6
, prenylated chimera 2 was also bound to the membrane but lacked significant activating ability. Chimera 1 did not bind to the membrane nor activate the oxidase. These results are in good agreement with those discussed in the preceding section and illustrated in Figure 3
, namely, that membrane localization of the chimeras requires the presence at the C terminus of a prenyl group and a minimum positive charge. In the absence of C-terminal prenylation, the dependence on a positive charge at the C terminus is even more pronounced.
|
SUPERCHIMERAS ENTER THE SCENE
To find an explanation for the absolute need for intrachimeric interaction, in spite of the physical fusion among the moieties, we examined the ability of exogenous Rac1-GTPS to repair the lack of activity of open chimeras. Two experimental designs were chosen.
In the first, the inactive, prenylated chimeras 1 and 2 were supplemented with Rac1-GTPS and assayed for activity. As described in ref. [36
] and illustrated in Figure 7
, exogenous Rac1-GTPS reconstituted the oxidase-activating capacity of prenylated chimera 2 but not that of chimera 1. We explained this by the establishment of a bond between the switch I region of exogenous Rac1-GTPS and the TPR region of the inactive chimera, which substituted for the absent intrachimeric bond. The inactivity of the Rac1-supplemented chimera 1 is a result of the lack of a C terminal-positive charge in the chimera, an early indicator for the need for polybasic stretches on the target chimera and the exogenous Rac1 (see additional proof for this below). We reasoned that the Rac1 moiety within the chimera or the added exogenous Rac1 interacts with the TPR region of the p67phox moiety and causes a conformational change, which affects the activation domain of p67phox in a positive manner, promoting its interaction with gp91phox and the induction of an electron flow.
|
S chimera 3-GDP, exhibiting an open conformation, and mutants of chimera 3-GTPS, in which an open conformation was obtained by A27K or G30S mutations in the Rac1 moiety. In all cases, the oxidase-activating ability of the previously inactive chimeras was fully regenerated. Supplementation of inactive chimera 3 with Rac1-GDP, Rac1 A27K-GTPS, Rac1 G30S-GTPS, CDC42Hs-GTPS, or RhoA-GTPS did not repair activity, demonstrating that the exogenous partner must be able to bind to the p67phox moiety of the inactive chimera [35
]. We also found that only Rac1-GTPS with a C terminus containing the polybasic stretch (residues 183–188) was capable of repairing open, inactive chimeras; Rac1-GTPS, truncated at residue 178 or 185, was incapable of repair [35
]. Reciprocally, a p67phox-Rac1 chimera truncated at residue 178 of the Rac1 moiety (chimera 5; see Fig. 2
) could not be repaired by Rac1-GTPS. This is in good agreement with the inability to repair prenylated chimera 1 by Rac1-GTPS, as shown in Figure 7
. Thus, for effective repair, the initially inactive chimera and the exogenous Rac1 must both possess a full, polybasic stretch at their respective C termini. We call the hypothetical structure consisting of an open p67phox-Rac1 chimera and bound exogenous Rac1-GTPS a "superchimera." Attempts to isolate a stable superchimera by gel filtration have, so far, met with failure. Also, trials to reconstitute an inactive (open) chimera 3-GTPS, with a R102E mutation in the p67phox moiety, by supplementation with exogenous p67phox (1–212), failed. Possible reasons for this failure are discussed in ref. [35
]. All these results provide support for the proposal that Rac plays a dual role in the assembly of the NADPH oxidase complex of phagocytes: that of a membrane tether for p67phox and of an inducer of a conformational change in p67phox, essential for its productive interaction with gp91phox.
"THE PROOF OF THE PUDDING IS IN THE EATING"—PHYSICAL EVIDENCE FOR CLOSED/OPEN CHIMERAS
So far, proof for structural differences in p67phox-Rac1 chimeras in the GTP-bound versus GDP-bound and native versus mutated forms, in which the establishment of intrachimeric bonds was prevented, was based exclusively on functional criteria. We intended to provide physical evidence for the existence of closed and open conformations of the p67phox-Rac1 chimera.
The first clue for this came from the observation that when subjecting chimeras in the GMPPNP-bound and in the native (GDP-bound) forms to gel filtration on a Superdex 75 column, only the GMPPNP-bound form eluted at the expected molecular mass of 45 kDa. The GDP-bound form eluted at an apparent molecular mass of 49 kDa, suggesting a more extended conformation (see ref. [35 ] and Fig. 8 ). The nature of the bound nucleotides was confirmed for both preparations by anion exchange chromatography on a Partisil 10 SAX column (Fig. 8) .
|
S, or mant-GDP, attached to the Rac1 moiety. The performance of such experiments was only feasible if the occurrence of FRET from the two tryptophans present in the Rac1 moiety (W56 and W97) to the mant nucleotides, bound to the same moiety, could be prevented. To achieve this, we mutated W56 or W56 and W97 to phenylalanines in the Rac1 moiety. Most work was done with chimera p67phox (1–212)-Rac1 (1–192), W56F, W97F. The principles governing the particular FRET experiments and a schematic representation of the mutant chimeras appear in Figure 9
. Both mutant chimeras were found fully competent to support oxidase activation in a cell-free system and to bind mant-GMPPNP. The reality of the W56F mutation was ascertained by the inability of chimeras containing this mutation to respond to the Rac-specific GEF Trio by the uptake of mant-GMPPNP [39
]. Mutant chimeras were subjected to exchange to mant-GMPPNP, mant-GTPS, or mant-GDP, freed of unbound mant nucleotides by extensive centrifugal filtration or gel filtration and brought to a concentration of 1 or 2 µM. Fluorescence was measured in a spectrofluorometer under two excitation/emission conditions. In the first, FRET from tryptophans, present in the p67phox (1–212) moiety to mant nucleotide bound to the Rac moiety, was measured by exciting samples at 295 nm and recording the emission spectrum from 305 to 485 nm. In the second, direct, mant-derived fluorescence, as the result of irradiating the mant nuclotide bound to the Rac moiety, was measured by exciting samples at 361 nm and recording the emission spectrum from 400 to 500 nm. As apparent in the typical experiment shown in Figure 10
, chimera p67phox (1–212)-Rac1 (1–192), W56F, W97F in the mant-GMPPNP-bound form exhibits lower tryptophan fluorescence (with an emission maximum close to 340 nm) and higher mant-derived fluorescence (with an emission maximum close to 440 nm) than the mant-GDP-bound form of the same chimera at the same concentration. The two types of measurements are illustrated in the lower section of Figure 11
. The results are expressed in quantitative terms as F/W and F/M ratios and shown in the Table in the upper section of Figure 11
. The higher the ratio, the more intense is the FRET (see legend of Fig. 11
for a description of the method of calculating the ratios).
|
|
|
S-bound and mant-GDP-bound forms and between two mant-GTPS-bound forms of chimera p67phox (1–212)-Rac1 (1–192), W56F, W97F, one of which also had a A27K mutation in the Rac1 moiety preventing the establishment of intrachimeric bonds. Thus, FRET studies point clearly to a more compact conformation of the GTP-bound (active) chimera and a more extended shape of the GDP-bound (less-active or inactive) chimera. Finally, yet another approach to assess the structure of the various forms of the chimera is in progress. This is based on the application of small angle X-ray scattering to the closed and open chimeras [41 ]. Preliminary results are supporting the results obtained by FRET (M. Hirshberg, Günter Grossmann, A. Mizrahi, Y. Berdichevsky, E. Pick, unpublished results).
Figure 12 depicts a model of chimera p67phox (1–212)-Rac1 (1–192) in the resting (GDP-bound) and active (GTP-bound) conformations in relation to cytochrome b559 and its phospholipid membrane microenvironment. Also shown is the establishment of intrachimeric bonds in the active conformation, and the activation domain of the p67phox moiety makes contact with a yet-unidentified domain(s) in the cytosolic region(s) of gp91phox (marked on/off).
|
Among the cytosolic components of the oxidase complex, p47phox was the target of the most intense interest and of a large number of studies. As mentioned earlier in this article, the centrality of p47phox in the process of oxidase activation was questioned by reports by two groups describing the ability to activate the oxidase in vitro by p67phox in conjunction with Rac in the absence of p47phox [26 , 27 ]. There was never any question about the involvement of and requirement for p47phox in the respiratory burst in intact phagocytes. Thus, experiments dealing with oxidase activation in the absence of p47phox are not meant to serve as in vivo models but as a mean to pinpoint the precise role of p47phox in the assembly of the oxidase. p47phox is not required for assembly of the oxidase under the following conditions: high concentrations of p67phox and Rac-GTP in the presence of amphiphile [26 , 27 ]; p67phox and prenylated Rac-GTP in the absence of amphiphile [12 ], and prenylated p67phox-Rac1-GTP chimera in the absence of amphiphile [36 ].
Recently, we developed a novel, cell-free oxidase activation system, which was inspired by reports by the groups of Tamura [32 ] and Kleinberg [23 ] (A. Mizrahi, Y. Berdichevsky, Y. Ugolev, S. Molshanski-Mor, E. Pick, unpublished results). In this macrophage membrane, liposomes were enriched with anionic phospholipids, such as phosphatidic acid or phosphatidyl glycerol. The addition to such "acidified" membranes of a mixture of p67phox, p47phox, and nonprenylated Rac1, exchanged to a nonhydrolyzable GTP analog such as GMPPNP, resulted in dose-dependent oxidase activation in the absence of amphiphile (Fig. 13A ). Replacing Rac1-GMPPNP by native Rac1 (Rac1-GDP) also led to quite significant activation. Oxidase activation in vitro supported by Rac1-GDP and by p67phox-Rac1-GDP chimeras was observed on several occasions, albeit with higher effective concentration of 50% values than measured with the GTP-bound equivalents (see refs. [12 , 34 , 36 , 42 ]). When the experiment was repeated in the absence of p47phox, oxidase activation was found with Rac1-GMPPNP but was absent with Rac1-GDP (Fig. 13B) . It thus appears that in this particular experimental situation, oxidase activation in the absence of p47phox is possible only when Rac1 is in the GTP-bound conformation. It is also of interest that p47phox exerts its action in the absence of an anionic amphiphile and consequently, does not require the opening of intramolecular bonds between the SH3 tandem and the C terminus.
|
Such stabilization is of key importance when the p67phox-tethering role of Rac1 is lost as a result of conversion of Rac-GTP to Rac-GDP [by physiological or GTPase-activating protein-stimulated hydrolysis] and/or its removal from the membrane by Rho-GDP dissociation inhibitor (Rho-GDI).
A "PROPAGATED WAVE" MODEL OF NADPH OXIDASE ACTIVATION?
"... the phenomenon had the appearance of a...propagated wave of physico-chemical disturbance...."—Sir Henry Dale, Nobel Lecture, 1936
Figure 14 illustrates schematically a model of oxidase activation, with emphasis on the stabilizing role of p47phox, discussed above. Touching circles indicate occurring or potential protein-protein interactions. The connecting symbol between gp91phox and p22phox indicates a heterodimer. Figure 14A describes the physiological situation in which Rac is in the GTP-bound state, forms a complex with p67phox, and induces an activating, conformational change in p67phox (small green arrow). This leads to a productive interaction between p67phox and gp91phox, most likely expressed in a conformational change in gp91phox (large green arrow), the induction of electron flow from NADPH to oxygen, and ensuing O2.– generation. The model does not spell out the entire spectrum of protein-protein interactions; as an example, it is possible that there is direct contact between Rac and gp91phox, as suggested by Diebold and Bokoch [6 ]. When Rac is in the GTP-bound form, the absence of p47phox has no major effect on oxidase activation (Fig. 14B) . Hydrolysis of GTP to GDP on Rac causes dissociation of Rac from p67phox and the reversion of p67phox to its preactivation state. This means dissociation of p67phox from gp91phox or the reversion of gp91phox to its resting state, in the absence of dissociation of p67phox from gp91phox, with no electron flow occurring (Fig. 14C) . Data in ref. [5 ] suggest a dichotomy between activation by and binding of p67phox in that a V204A mutation in the activation domain of p67phox causes loss of activation by p67phox but not necessarily loss of binding to its most likely target, cytochrome b559. If the conversion of Rac-GTP to Rac-GDP takes place in the presence of p47phox, p67phox is maintained, at least in part, in its active conformation, further transmitted to gp91phox, resulting in persistent electron flow and O2.– generation (Fig. 14D) . It is equally possible that in this situation, the stabilizing effect of p47phox is exerted directly on gp91phox via p47phox-gp91phox contacts or indirectly through p22phox.
|
With the risk of oversimplification, we propose a model inspired by the propagated wave model, which originated in physics but found applications in biological systems. The essence of this model is that conformational changes occurring in one oxidase component are propagated to the next, ultimately reaching gp91phox. The model put forward here is Rac 
p67phox gp91phox. Clearly, alternative and/or additional (reinforcing or inhibitory) routes may be at work, such as p47phox gp91phox; p47phox p67phox gp91phox; or p47phox p22phox gp91phox. It will also have to be found out what is causing the conformational change in the first component of a propagated wave. As an example, in a situation in which the wave starts with Rac, it remains to be determined what causes the conformational change in Rac (dissociation from Rho-GDI, the action of a GEF, or the direct action of a phosphoinositide). Other important, unsolved questions are whether there are multiple propagated wave pathways leading to oxidase activation in intact phagocytes and nonphagocytic cells and whether different pathways reflect oxidase activation initiated by different stimuli acting via different signal transduction sequences.
ACKNOWLEDGEMENTS
The research described in this report was supported by the Julius Friedrich Cohnheim-Minerva Center for Phagocyte Research, the Ela Kodesz Institute of Host Defense against Infectious Diseases, Israel Science Foundation Grants 428/01 and 19/05, the Roberts-Guthman Chair in Immunopharmacology, and the Walter J. Levy Benevolent Trust.
Received October 3, 2005; accepted January 5, 2006.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  K. A. Gauss, L. K. Nelson-Overton, D. W. Siemsen, Y. Gao, F. R. DeLeo, and M. T. Quinn Role of NF-{kappa}B in transcriptional regulation of the phagocyte NADPH oxidase by tumor necrosis factor-{alpha} J. Leukoc. Biol., September 1, 2007; 82(3): 729 - 741. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Berdichevsky, A. Mizrahi, Y. Ugolev, S. Molshanski-Mor, and E. Pick Tripartite Chimeras Comprising Functional Domains Derived from the Cytosolic NADPH Oxidase Components p47phox, p67phox, and Rac1 Elicit Activator-independent Superoxide Production by Phagocyte Membranes: AN ESSENTIAL ROLE FOR ANIONIC MEMBRANE PHOSPHOLIPIDS J. Biol. Chem., July 27, 2007; 282(30): 22122 - 22139. [Abstract] [Full Text] [PDF]
|
