| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Published online before print October 4, 2005
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,
,
,,,1
* Department of Surgery, Denver Health Medical Center, Colorado; Departments of
Surgery and
Pediatrics, University of Colorado School of Medicine, Denver; and
Bonfils Blood Center, Denver, Colorado
1Correspondence: Bonfils Blood Center, 717 Yosemite St., Denver, CO 80230. E-mail: christopher.silliman{at}uchsc.edu
 |
ABSTRACT |
|---|
Key Words: review • innate immunity • respiratory burst • oxidase assembly
|
HISTORICAL BACKGROUND |
|---|
|
OXIDASE ACTIVITY: GENERATION OF THE SUPEROXIDE ANION (O2–) AND DOWNSTREAM BYPRODUCTS |
|---|
|
THE PMN NADPH OXIDASE COMPONENTS |
|---|
|
|
|
|
THE MEMBRANE COMPONENTS OF NADPH OXIDASE |
|---|
-subunit) and gp91phox (ß-subunit) but also a flavin adenine dinucleotide (FAD), which serves as a NADPH-binding site, and two heme prosthetic groups, one of which selectively binds gp91phox, and the other binds gp91phox and p22phox [42
43
44
45
46
47
]. The FAD and heme groups serve as the redox pathway, which enables the transfer of electrons across the membrane. The two heme groups are functionally distinct from each other [48 ], and there is controversy concerning the role of the heme prosthetic groups as obligatory intermediates for oxidase activity [42 , 49 50 51 52 ]. On one side, if heme oxidation is an obligatory participant in electron transfer, then the maximal rate of O2– production would be significantly diminished; furthermore, O2– generation is not inhibited by the heme antagonists CN–, N3–, CO, and butyl isonitrile [45 , 53 54 55 56 ]. Conversely, the heme prosthetic groups serve as an intermediate between the FAD and O2– generation, as the two heme prosthetic groups have binding sites to gp91phox and p22phox [42 , 50 , 51 , 57 58 59 60 61 ], affect or are affected by the cytosolic oxidase components upon translocation [42 , 50 51 52 , 56 , 60 , 61 ], and therefore, may be important in the assembly of the NADPH complex. In a cell-free system, the heme group was found to be reduced by NADPH in the plasma membrane, but it may not be involved in the catalytic activity of the oxidase [54 , 62 , 63 ]. However, despite the controversy regarding the heme subunits, there is a consensus among investigators that the FAD subunit is an electron carrying intermediate in O2– generation, as oxidase activity is lost when the FAD is removed but restored when added back. Moreover, the oxidase activity can be inhibited by flavin antagonists (e.g.. deaza-FAD and diphenylene iodonium) [10 , 64 , 65 ]. It is important to understand the function and role of these subunits to better understand and regulate the NADPH oxidase and O2– generation.
In resting PMNs, 15% of the cytochrome b558 subunits gp91phox and p22phox are located in the plasma membrane and the remaining 85%, within the membrane of the specific granules and secretory vesicles [66 67 68 69 ] and then translocated to the plasma membrane for oxidase activation [70 , 71 ]. Moreover, a defect in either of the subunits results in the complete absence of the heterodimer from the plasma membrane [52 , 72 ] and also a deficiency of that subunit in the specific granules, resulting in decreased oxidase activity [73 74 75 ]. The gp91phox and p22phox proteins are reviewed briefly below.
The mature p22phox (-subunit) protein binds to gp91phox (ß-subunit) at the membrane, creating a stable cytochrome b558 complex [27
, 71
]. The p22phox protein contains a proline-rich region on its cytoplasmic tail, which interacts with src homology region 3 (SH3) domains, including the SH3 domain of phosphorylated p47phox [76
, 77
]. Along with gp91phox, p22phox binding of the cytosolic oxidase components, specifically p47phox and p67phox, is essential for proper function of the respiratory burst [78
79
80
81
]. Defects and/or deficiencies in p22phox have been reported in CGD patients [72
, 82
83
84
85
86
]. However, these defects tend to be autosomal instead of X-linked like defects in gp91phox [85
, 86
]. It is interesting that there is a significant decrease (33%) in O2– production with the TT genotype of the C242T polymorphism and no link to a decrease in immune function; however, an increased risk of coronary artery disease is observed [87
].
The gp91phox (ß-subunit) is post-translationally glycosylated on three of the potential five N-linked glycosylation sites [88 89 90 ], contains five hydrophobic domains on the N-terminus, which likely represent membrane-spanning domains [91 , 92 ], and has a cytosolic, hydrophilic region at the C-terminus, which contains the heme moieties and interacts with p47phox [91 , 93 ]. Blocking of the C-terminus inhibits oxidase activity but not translocation or binding of the cytosolic subunits to the plasma membrane; thus, gp91phox appears essential for oxidase function [93 ]. The role of gp91phox is speculated to be a H+ channel, permitting charge compensation across the membrane, coinciding with electron transfer, regulated by the translocated Rac2 and p67phox [94 ], and gp91phox may also regulate the steady-state FAD reduction [95 ].
Mutations in gp91phox are X-linked and the most common defect in CGD. These mutations include A57E, E309K, C537R, P339H, 
F215, and F216 [96
97
98
]. The F215 and F216 mutations result in aberrant phenylalanine residues, and affected cells contain no trace of cytochrome b558, suggesting that one or both of these phenylalanine residues are essential for the binding of cofactors to the membrane-bound cytochrome [97
]; however, there is a normal amount of gp91phox present in the PMN. The carrier state for a gp91phox defect is also detrimental, leading to a depressed immune response, colonization with fungi, and resultant granulomatosis [99
]. In gp91phox–/– mice, there is a lack of O2– production, no change in p22phox, greater PMN sequestration in the lungs, and a down-regulation in cyclooxygenase 2 (COX-2) [100
101
102
103
104
].
Rap1A (Krev 1)
Rap1A (also referred to as Krev1) is a low molecular weight (22 kDa), guanosine 5'-triphosphate (GTP)-binding protein, which is a member of the Ras superfamily and found in large quantities in PMNs [37
, 105
106
107
]. Its role in the regulation of the oxidase was initially suggested by its copurification and coimmunoprecipitation with cytochrome b558 (1:1 stoichiometry) [37
, 38
, 91
, 108
]. Rap1A translocates with cytochrome b558 from the specific granules to the plasma membrane upon cell stimulation, but this interaction is inhibited by cyclic adenosine monophosphate (CAMP)-dependent protein kinase A (PKA) phosphorylation [38
, 106
, 109
]. When Rap1A binds to cytochrome b558, the guanosine-5'-O-3-thiotriphosphate (GTP-gamma-s)-bound form of Rap1A binds more tightly to cytochrome b558 than the guanosine diphosphate-bound form (GDP) [108
]. Phospholipase C (PLC) activation, leading to elevated levels of intracellular-free Ca2+ and diacylglycerol (DAG), has been implicated as a mediator of Rap1A activation via PKC phosphorylation [105
, 110
, 111
]. The role of Rap1A in oxidase regulation is controversial; some reports show that depleting the cell of Rap1A decreases oxidase activity [36
, 37
, 105
106
107
], and other reports indicate that there is no effect on the oxidase [108
, 112
]. In HL-60 cells, Rap1A was suggested to act as the final activation switch of the NADPH oxidase by its interaction with the cytochrome b558 [36
]. The truncated form of Rap1A in a cell-free system was functional in NADPH activation but may not have a direct role in the regulation of the oxidase, as there is no change in expression in patients with CGD [36
, 113
]. Further work is required to fully define the role of Rap1A in oxidase activation and the role of protein kinase isoforms in overall oxidase function.
Nox proteins
Nox proteins, or NADPH oxidase enzymes, are homologs to gp91phox (Nox2), which are found mainly in nonphagocytic cells [45
, 114
, 115
]. In humans, there are seven members of this family, but Nox1, Nox3, and Nox4 have the most homologous and similar structure to Nox2 [114
, 116
, 117
]. Nox1, along with Nox4, are structurally the most similar to Nox2, are found in epithelial cells, and colocalize with p22phox for O2–generation [117
118
119
120
], and Nox3 is found primarily in fetal tissues and is possibly related to development [114
, 116
]. Recently, however, Nox proteins have been shown to be responsible for reactive oxygen species (ROS) generation in nonphagocytic cells [114
, 121
122
123
124
125
126
] along with binding sites for FAD and NADPH [114
]. Furthermore, Rho-GTPase regulation of Nox proteins is comparable with its regulation in phagocytes, i.e., gp91phox (Nox2) regulation [114
]. Moreover, nonphagocytic cells contain homologs to p47phox and p67phox, known as p41 and p51, respectively, which can regulate Nox proteins, specifically Nox1 [127
]. Specific regulation of O2– generation has been located in the charged amino acids in the D-loop region and in the -helical loop of the C terminus [128
]. The regulation of the NADPH oxidase and the generation of ROS are most noted in PMNs and leukocytes; therefore, more research needs to done to fully delineate the regulation of NADPH oxidase activity and the correlation between phagocytic and nonphagocytic cells.
|
CYTOSOLIC COMPONENTS OF THE NADPH OXIDASE |
|---|
, ß, 
, and 
have been shown to cause the phosphorylation and translocation of p47phox with nonphysiologic stimuli, such as phorbol myristate acetate (PMA) in primary and transformed cell lines [145
146
147
]. One form of CGD is a deficiency in p47phox from mutations on the p47phox protein, which include a GT or TG deletion at exon 2 [148 149 150 ] or an Arg-to-Gln mutation in the PX domains of [135 ]. In p47phox-deficient CGD, p67phox, p40phox, and Rac2 are unable to translocate from the cytosol to the membrane with formyl-methionyl-leucyl-phenylalanine (fMLP) or PMA activation [53 , 151 ]. The characteristics of p47phox-deficient CGD are mimicked in p47phox–/– mice [152 , 153 ]. However, unlike in CGD, the activation of the NADPH oxidase in some cell-free systems does not always require the translocation of p47phox [154 155 156 ], but it is needed for maximal oxidase activity.
p67phox
The 526 amino acid peptide [129
] p67phox contains five known domains: 1) four N terminus tetratricopeptide repeat domains (TPRs), which interact with Rac [133
]; 2) an activation domain, which binds to gp91phox following phosphorylation and translocation [127
]; 3) an internal praline-rich domain [133
]; 4) two SH3 domains on the N-terminus side, which can bind to the proline-rich domain of p47phox in resting and activated PMNs [53
, 129
, 157
, 158
]; and 5) the PB1 domain, a 150 amino acid stretch between the two SH3 domains and the site of the C-terminus PC motif for p40phox interaction [159
160
161
]. The p67phox-p40phox interaction is maintained even in the presence of anionic amphiphiles, raising the possibility that this interaction mediates a constitutive association in resting and activated cells [160
]. Phosphorylation of p67phox is not only a prerequisite for translocation from the cytosol to the membrane but is also required for the translocation of p40phox and Rac2 to the membrane [151
, 162
]. Also, p67phox contains a catalytic NADPH binding site for electron transfer to the FAD in the cytochrome b558 complex [50
, 155
, 163
, 164
]. More specifically, amino acids 1–210 of p67phox are required for a steady-state reduction of FAD, indicating a dominant effect on hydride/electron transfer, and amino acids 199–210 are important for regulating the electron flow [50
]. The dehydrogenation (reduction) activity of NADPH by p67phox is proportional to the enzyme concentration, independent of FAD, insensitive to O2–dismutase, and inhibited by high concentrations of ferricyanide [165
]; moreover, these findings suggest that p67phox is involved in the transfer of electrons between NADPH and the oxidase flavin [165
].
CGD with a deficiency in p67phox is rare and results in a decrease in oxidase production and immunosuppression [166 , 167 ]. Various mutations have been identified that result in the absence of p67phox [168 ], such as a mutation that results in a deletion of the AAG sequence along with a deletion in another allele [169 ] or a T-to-C point mutation in intron 3 of the p67phox prevents the accumulation of p67phox mRNA [170 ]. A deficiency of p67phox in CGD fails to translocate p40phox in response to PMA, also confirmed by cell-free assays [154 , 171 ], whereas p47phox translocation and phosphorylation are unaffected [172 173 174 ].
p40phox
The p40phox protein is 339 amino acids in length [175
], its expression is restricted to hematopoietic cells, with the exception of erythroid cells [176
], and was originally identified as a protein tightly associated with p67phox [177
, 178
]. There are three known domains on p40phox: 1) an N terminus PX domain, specific for PI(3)P, 2) an internal SH3 domain, and 3) a C-terminus PC motif, which can interact with the PB1 domain of p67phox [133
, 137
, 161
, 175
, 178
, 179
]. In resting cells, p40phox is basally phosphorylated, but the onset and extent of additional phosphorylation strongly correlate with the level of O2–generation [180
]. Phosphorylation of p40phox occurs on the Ser315 and Thr154 residues and is inhibited markedly by the PKC inhibitor H-7, indicating PKC as a possible direct phosphorylator of p40phox in the p40phox-p47phox-p67phox trimer [173
, 181
]. Additionally, p40phox appears to down-regulate oxidase function by competing with the SH3 domain interactions between other essential oxidase components [172
].
There have not been any p40phox defects or deficiencies demonstrated to date; however, in CGD patients lacking p67phox, the amount of cytosolic p40phox is decreased significantly [151 , 174 , 175 ]. Therefore, the exact functional role of p40phox in oxidase activation and regulation has not been fully elucidated and may be dependent on multiple variables.
Rac 1/2
Rac is a small G-protein (21 kDa) member of the Ras superfamily, which is required for activation of the NADPH oxidase [37
, 109
, 182
, 183
]. In resting neutrophils, Rac2, the primary isoform in human PMNs [98%; although Rac1 is also present (2%)], is located in the cytosol [95
, 151
, 183
]. Upon activation, Rac2 rapidly converts from a GDP- to GTP-bound state, dissociates from Rho guanine-nucleotide-dissociation inhibitor, and migrates to the membrane [184
, 185
]. Rac2 (amino acid residues 170–199) binds directly to p67phox at the membrane in a 1:1 stoichiometry and a dissociation constant value of 60 nM, but it does not bind to p40phox or p47phox [109
, 186
, 187
]. The N-terminus (amino acids 1–192) of p67phox can be used as a specific inhibitor of Rac2 signaling, reducing the ability of Rac2-GTP to disrupt the p67phox-p40phox binding [188
, 189
]. Moreover, mutational studies have identified two regions in Rac2, which are important for activity: 1) the "effector region" (residues 26–45) and 2) the "insert region" (residues 124–135) [164
, 190
]. Proteins mutated in the effector region (N26H, I33N, and D38N) inhibit Rac2 binding to p67phox, and the insert region mutations (K132E and L134R) bind with normal affinity to p67phox [182
]. The motif RQQKRP in the C-terminus region and the D150 amino acid have recently been identified as being essential for O2–production, chemotaxis, and F-actin assembly [191
]. Based on this information, a model is postulated, whereby the Rac2 effector region binds p67phox, the Rac2 C-terminus binds to the membrane, and the insert region likely interacts with cytochrome b558.
Rac2 deficiency has been implicated in CGD. Rac2 deficiency has been mainly studied in mice and has been shown to be important for degranulation, chemotaxis, actin formation, and O2– production [192 193 194 195 ]. PMNs from Rac2–/– mice have decreased responsiveness to activating stimuli [i.e., fMLP, leukotriene B4 (LTB4), complement 5a (C5a)], although p67phox still binds to the membrane, but still do not produce O2– or release MPO or elastase [196 197 198 199 ]. In humans, however, a D57N point mutation was identified in a patient, which resulted in the inability of GTP to bind to Rac2, although GDP binding was normal, and decreased O2– production was observed [200 ]. In cell-free systems, the importance of Rac2 in the activation of the NADPH oxidase system has been shown [107 , 109 , 201 ].
Cdc42
Cdc42 is a small G-protein with 70% homology to Rac1/2 [202
]. The main difference is in the effector domain, also known as Switch I, where there is a four amino acid difference [201
202
203
]. The two amino acids that have the most effect, however, are amino acids 27 and 30, which when mutated in the Cdc42 protein, become fully functional in activating the oxidase, similar to Rac2 [202
203
204
]. Thus, Cdc42, which can be stimulated by ß2-integrins [205
], acts as an inhibitor of oxidase activation. Recently, it was suggested that Rac and Cdc42 act as antagonists, competing through the insert domain for binding to cytochrome b558 or, more specifically, the translocated p67phox [201
].
p29 peroxiredoxin
This is a newly discovered oxidase-associated protein, which coimmunoprecipitated with p67phox [39
]. It has phospholipase and peroxidase activity and when preincubated with recombinant cytosolic oxidase proteins Rac1, p47phox, and p67phox, increased oxidase activity [39
]. It is thought that p29 peroxiredoxin has an effect on the oxidase by its peroxiredoxin activity as a result of its ability to reduce thioredoxin [39
]. Further studies must be conducted to elucidate the role of the p29 peroxiredoxin as an oxidase-associated protein; there are currently no known deficiencies of this protein.
|
THE PMN STAGES FOR NADPH OXIDASE ACTIVITY: RESTING, PRIMED, AND ACTIVATED |
|---|
The complexities of NADPH oxidase are many-fold, and further elucidation of the various signaling circuitry, kinases, phosphatases, and lipases involved will be required before firm statements regarding the precise mechanism(s) underlying in vivo oxidase priming/activation can be made. However, each of these phenotypes (resting, primed, and activated) will be discussed below as they relate to the cellular structure (cytosol, membrane, and granule) important for the NADPH oxidase.
|
RESTING PMNs (Fig. 1) |
|---|
Membrane
The membrane in resting PMNs contains a small portion of gp91phox and p22phox, along with the FAD and heme moieties of the cytochrome b558 [53
, 213
, 216
, 217
]. Also, in some studies using Trition X-100 for subcellular fractions, p67phox, p40phox, gp91phox, and p22phox were found in the insoluble fraction, or the cytoskeleton [218
, 219
]. However, in nondetergent, subcellular fractions, the p67phox and p40phox were located in the cytosol, as described above.
Granules
The specific granules and the secretory vesicles contain the majority of the gp91phox and p22phox [53
, 213
, 216
, 220
].
|
THE PRIMED PMN (Fig. 2) |
|---|
(TNF-), GM-CSF, IL-18], which takes 15–60 min to manifest their effect [141
, 211
, 223
224
225
226
227
]. Different priming agents have a range of effects on the structural organization of the oxidase and will be reviewed briefly below.
Cytosol and membrane
Binding of a "priming" agent (e.g., PAF, LPC, C5a, IL-8, or LPS) to a G-protein-coupled receptor (GPCR) on PMNs may, pertaining to the oxidase, take from seconds to several minutes to up to 2 h (e.g., interferon-
) for maximal augmentation to a subsequent stimulus and thus, priming of the PMN [209
, 211
, 225
, 226
, 228
229
230
231
]. Priming of the oxidase with long-acting priming agents, such as LPS, GM-CSF, and IL-18, translocates p47phox to the plasma membrane [141
, 211
, 231
]. It is thought that the degree of p47phox phosphorylation, a prerequisite for translocation, correlates with the rigor of the priming agent [232
, 233
]; however, with TNF- priming (15–60 min), there is only partial phosphorylation of p47phox, not translocation, but TNF- is still considered an effective priming agent [224
, 231
, 233
234
235
]. Conversely, PAF, a rapid (3–5 min) primer, phosphorylates p67phox, p40phox, and Rac2, but not p47phox [234
]. Preliminary data suggest that p67phox then translocates to the plasma membrane with PAF priming [236
]. When PMNs are incubated with LPS (30–60 min), there is an increase in the association of cytochrome b558 with the plasma membrane along with phosphorylation and translocation of p47phox, but not p67phox, p40phox, or Rac2 [141
]. However, other priming agents, such as homocysteine, angiotensin II, opsonized zymosan (OpZ), and stimulation through the ß2-integrins, will cause phosphorylation of p47phox and p67phox [219
, 237
238
239
]. Phosphorylation and translocation of the cytosolic oxidase components to the plasma membrane during priming augment oxidase activity when a second stimulus is used [141
, 211
, 222
, 225
, 227
, 231
, 240
241
242
].
Along with the influence of the cytosolic oxidase components on priming, this phenomenon is regulated by additional proteins, which are vital to the various signaling pathways. One of the key regulators is from the family of MAPKs, specifically, p38 MAPK and extracellular signal-regulated kinase (Erk)1/2 (p42 and p44 MAPK, respectively). Many studies have shown that when either protein is inhibited, O2– production is inhibited [144
, 234
, 237
, 238
, 243
244
245
246
247
248
249
]. The exact mechanism for p38 MAPK and Erk1/2 phosphorylation of the cytosolic oxidase components, especially p40phox, p67phox, and p47phox, has yet to be understood fully, but there may be a dual role in PMN oxidase release. TNF- priming of human and porcine PMNs activates p38 MAPK and directly phosphorylates p47phox and p67phox [234
, 243
, 250
, 251
], whereas with PAF priming, there is only p67phox phosphorylation through p38 MAPK activation [234
]. Other priming agents, such as ionomycin and angiotensin II, activate p38 MAPK and Erk1/2 [238
, 252
], However, p38 MAPK and Erk1/2 do not always phosphorylate the oxidase components directly, but instead, these MAPKs can activate a secondary protein or lipid mediator in a signaling pathway [144
, 253
254
255
256
257
].
Lipid mediators have been implicated in priming of the NADPH oxidase. LPC, arachidonic acid (AA), and the leukotrienes, specifically LTB4, have been shown to cause an increase in O2– release in response to a subsequent stimulus [209
, 223
, 228
, 241
, 253
, 254
, 256
, 258
259
260
261
262
263
]. The exact mechanism of action is still relatively unknown, but it appears that there is involvement of a G-protein [probably Gi, based on pertussis toxin (PTX) inhibition] and the MAPK pathway [221
, 241
, 257
, 264
265
266
267
]. LTB4 has been shown to activate Erk1/2 [210
, 268
, 269
] and cause the translocation of Rac2 [270
]. Also, these lipid mediators may activate PKC and phosphoinositide-3 kinase (PI-3K) and are Ca2+-dependent [110
, 142
143
144
, 215
, 225
, 254
, 271
, 272
]. Further studies are required to determine the exact mechanism by which these lipids signal.
Another key component in human PMN priming is cytosolic calcium. The activation of many of the proteins required for priming is calcium-dependent, such as PKC, MAPK, and tyrosine kinases [211 , 238 , 241 , 245 , 247 , 273 274 275 276 277 278 279 ]. The influx of calcium to the cytosol from the internal and external stores creates a voltage gradient and a change in the membrane potential [280 ]. The rise in cytosolic calcium triggers the start of many signaling cascades essential for priming and assembly of the oxidase components [222 , 241 , 247 , 252 , 281 , 282 ]. Inhibiting calcium release or influx not only inhibits the priming of PMNs but also activation of the oxidase [241 , 252 , 282 ].
Priming of PMNs changes the physical structure of the cell by the rearrangement of actin, especially in F-actin. The actin cytoskeleton may provide a means for coordinating the process of NADPH oxidase assembly [80 , 283 284 285 286 ]. Currently, exact mechanisms of oxidase activation, with respect to cytoskeleton reorganization, are not fully understood. Initial studies indicate an association with the cytoskeleton based on the phox protein detergent insolubility in whole cells or at the membrane [69 , 142 ]. In human PMNs, the phox proteins particularly seem to interact with coronin, a cytoplasmic, actin-associated protein involved in the dynamics of the actin system, at the cytoskeleton [287 288 289 ]. Coronin is selectively solubilized when the PMN oxidase is activated by fMLP or PMA; however, it is not solublized in the absence of cytochrome b558 [290 ]. In addition, p67phox copurifies with coronin via the C-terminus half of p40phox and accumulates around the phagocytic cup in PMNs [218, 290 291 292 ]. Moesin, an F-actin-binding protein, binds to p47phox and p40phox in a phosphoinositide-dependent manner via the N-terminal of the PX domains [218, 293 ]. In addition, cofilin has been found to associate with p67phox [294 ]. Cofilin dephosphorylation occurs following PMA and fMLP stimulation of PMNs, and the kinetics are similar to those observed for O2– generation. Coflin dephosphorylation is also associated with movement toward F-actin-rich areas at the cell periphery [218, 295, 296] and colocalizes with areas high in ROS generation, implicating a role for cofilin and F-actin in oxidase activity [218, 296 ]. Oxidase subunit association with actin or actin-binding proteins has also been confirmed by microscopy and Western blot analysis [218 , 288 , 295 , 297 298 299 300 ]. However, more studies, especially with digital microscopy of whole, fixed PMNs, need to be completed to better understand the interaction amongst actin, the oxidase components, and membrane rearrangement.
There is increasing evidence that signal transduction in PMNs may occur in multiprotein signalosomes, or preassembled signaling complexes [301 , 302 ] and may regulate actions on a spatio-temporal basis. An obvious requisite is the assembly of some kind of backbone, a classic being the assembly of the cytoskeleton. However, unlike the cytoskeleton, which may support assembly on a more global level, clatherin, caveolin, and the emergence of lipid domains serve as specific microdomains, assembling at and as in a response to external stimuli [303 304 305 ]. There are two major concerns for PMN signaling, oxidase assembly, and the subsequent release of O2–. First, there is evidence for assembly in the region of receptor ligation, which is internalized via clathrin-mediated endocytosis [306 307 308 ] or lipid rafts {Fc receptors (FcRs) [309 310 311 312 ]}. Second, human PMNs do not contain caveolin [313 ], a significant subunit of lipid rafts, whereas many "model" cell lines do. The use of HL-60 cells, when properly differentiated, exhibits 90% homology to human PMNs, conferring their practical purpose [314 ]. However, when using phagocytic cells of a different delineation (i.e., monocyte precursors, U798, RAW39393, adipocytes, and endothelial derivatives) or even across different species of PMNs, this homology is not the case. For example, when comparing human PMNs with bovine PMNs, one of the big differences is the presence of caveolin in bovine PMNs [315 ]. Therefore, it is important to consider the cell linage when comparing in vitro and in vivo studies.
Granules
When a priming agent, such as LPS, binds to the PMN, the internal granules (specific and azuraphilic granules and secretory vesicles) fuse with the plasma membrane to form a phagosome, thus allowing gp91phox and p22phox to interact with the membrane [73
, 220
, 316
]. The increase of cytochrome b558 at the plasma membrane as a result of degranulation is well documented [241
, 317
318
319
320
]; however, the role of the specific granules in oxidase generation is still debatable. Upon incubation with PMA, the cytosolic oxidase components, p47phox, p67phox, and Rac2, translocate, not only to the plasma membrane but also the specific granules, where they are able to produce O2– for a short period of time [39
, 74
, 321
]. Oxidase activity in the granules is dependent on PKC and PI-3K activity to allow for proper assembly of the oxidase components [215
, 322
]. Also, the release of MPO and elastase from the azuraphilc granules during degranulation is important for the phagocytic activity of PMNs [323
324
325
326
327
].
|
THE ACTIVATED PMN (Fig. 3) |
|---|
Cytosol
Activation of oxidase requires additional phosphorylation, mainly on the serine and threonine residues of p47phox, p67phox, and p40phox, via kinases such as PKC, p21-activated kinase (PAC), p38 MAPK, PI-3K, and PA-activated protein kinase [45
, 78
, 144
, 329
, 330
], followed by translocation to the membrane. Activation of the NADPH oxidase also requires the dissociation of p40phox from p67phox via the association of an activated Rac2 with p67phox [174, 189
]. However, blocking the bond between p40phox and p67phox inhibits cell-free oxidase activation but not the translocation of the cytosolic proteins, including p47phox to the membrane, nor does it inhibit activated oxidase activity in vivo [178
]. In some cases, tyrosine phosphorylation is important for oxidase activation, such as the FcR for immunoglobulin G IIA (FcRIIA)-stimulated phosphorylation of p72syk as an early signal that precedes activation [330
, 331
]. Activation of PMNs results in complete translocation of all cytosolic oxidase components to the plasma or vesicle membrane.
Membrane
Activation of the oxidase through binding of soluble ligands (i.e., fMLP) to surface receptors usually results in O2– production lasting less than 5 min, whereas receptor-independent activation (i.e., PMA) causes prolonged generation of the O2– until depletion of necessary substrates and cofactors [332
]. Furthermore, in intact cells, the activated oxidase is experiencing a continuous process of activation and deactivation [333
].
Complete assembly of the oxidase components at the membrane is the final stage of activation. As previously established with priming, gp91phox and p22phox are already at the membrane, and there is translocation and partial phosphorylation of the cytosolic components. It is well defined that the SH3 domain of the p47phox interacts with the proline-rich domain of p22phox [61 , 77 , 81 , 134 , 138 , 334 , 335 ]. The polyproline region of p47phox binds to the C-terminus SH3 domain of p67phox [61 , 81 , 131 , 138 , 156 ], and the TPR region of p67phox binds to Rac2 at its effector region; Rac2 binds to the membrane at its C-terminus, and the insert region probably interacts with gp91phox [140 , 188 , 336 ]. The SH3 domain of p40phox binds to the polyproline region of p47phox, although p40phox binds more tightly to p67phox, and p40phox acts as bridge between p47phox and p67phox at the membrane [156 , 175 , 177 , 179 , 189 , 212 , 335 ]. Although Rap1A has been implicated in oxidase activity, its role in the oxidase complex is controversial but may bind to the cytochrome b2 complex at its carboxy terminus [38 , 105 , 108 , 213 , 337 ]. The PX domain of p40phox and p47phox also interacts with F-actin-binding proteins at the membrane via phosphoinositol-mediated signaling [177 , 212 , 290 , 293 ]. These interactions are proposed to be constant, regardless of the activating factor.
PMA has become the most commonly used positive control for oxidase activation [332
, 338
]; however, it is a nonphysiological stimulus. Originally considered to be involved in tumor promotion, although by themselves were not carcinogenic, phorbol esters resulted in many genetic and phenotypic changes, including an increase in oxidase production and the copurification of PKC with the putative phorboid receptor [339
340
341
342
343
]. Thus, PMA is a PKC agonist. Moreover, it is possible that the efficacy of ROS production in response to PMA has overridden its own irrelevancy in in vivo PMN function. The response to multiple stimuli (i.e., fMLP) over a reasonable amount of time will exhibit multiple oxidative responses; however, PMA uses all available resources in the production of ROS. Essentially, PMA, once the substrates are diminished, lacks any further oxidative capacity in the PMN, whereas other activators may allow for multiple oxidative responses. This is important to consider, as the lifespan of PMNs is 
24 h in vivo and 12 h in vitro.
The bacterial peptide fMLP is another oxidase activator that signals in a receptor-dependent mechanism. The fMLP receptor, a GPCR, regulates various activities in PMNs via a PTX-sensitive G-protein, which includes chemotaxis/chemokinesis, exocytosis (degranulation) [330 ], and multiple signaling pathways including MAPK, lipid kinases, the production of second messengers by various phospholipases (e.g., PLC, PLA2, and PLD) [4 , 5 ], and the activation of PKC [11 , 344 345 346 ]. For example, the binding of fMLP to its receptor activates PLC, which generates DAG and inositol trisphosphate and the release of intracellular Ca2+ [271 , 347 348 349 ]. DAG can also be converted to PA by the action of a stimulus-responsive, translocatable diacylglycerol kinase [350 ] or through the MAPK activation of PLD [351 352 353 354 ], which is involved in oxidase activation [355 , 356 ]. PMNs that are treated sequentially with LPS (priming) and fMLP (activation) have a three- to sixfold increase (compared with either agent alone) in the plasma membrane content of p47phox, p67phox, and Rac2 and augmented O2– generation by intact PMNs.
OpZ has also been shown to activate the NADPH oxidase and cause phagocytosis in PMNs; however, unlike PMA, which activates PKCs, OpZ activates cytosolic PLA2 (cPLA2) and the release of AA for oxidase activation [338 , 357 , 358 ] and an efflux of protons [359 , 360 ], with further regulation by Erk1/2. The absence of cPLA2 inhibits O2– generation [261 ], but the addition of AA restores O2– generation and efflux of protons [261 , 361 ]. Moreover, Erk1/2 and p38 MAPK are required for the onset of cPLA2 activation through the engagement of the tyrosine kinase Pyk2 [362 ]; however, activation of the proton channel and the NADPH oxidase is unaltered by the presence of COX and lipoxygenase inhibitors, indicating that the opening of the proton channel and the activation of O2– generation are not mediated via AA metabolites of these enzymes [362 ]. OpZ has also been shown to activate the C3 receptor and the FcR [363 , 364 ].
Granules
NADPH oxidase activation also occurs in the granules. The cytosolic components translocate to the granular membranes for the production of O2–, which is probably for the intracellular destruction of bacteria. Oxidase generation could also contribute to the apoptotic signaling pathway [39
, 73
, 317
]. The activation of PMNs with fMLP and OpZ has a greater response than with PMA and the formation of the NADPH complex, similar to the plasma membrane [73
].
|
SUMMARY |
|---|
TOP
ABSTRACT
HISTORICAL BACKGROUND
