Phox activity of differentiated PLB-985 cells is enhanced, in an agonist specific manner, by the PLA2 activity of Prdx6-PLA2

Authors


Full correspondence: Dr. Daniel R. Ambruso ,Bonfils Blood Center, 717 Yosemite, St, Denver Colorado, United States, 80230.

Fax: +1-303-340-2616

e-mail: daniel_ambruso@bonfils.org

Abstract

Peroxiredoxin 6-phospholipase A2 (Prdx6-PLA2) is a bi-functional enzyme with peroxi-redoxin (Prdx) and phospholipase A2 (PLA2) activities. To investigate its impact on phagocyte NADPH oxidase (phox) activity in a neutrophil model, the protein was knocked down in PLB-985 cells using stable expression of a small hairpin RNA (shRNA) and phox activity was monitored after cell differentiation. The knockdown cells had reduced oxidase activity in response to stimulation with the formylated peptide fMLF, but the response to the phorbol ester PMA was unchanged. Reintroduction of shRNA-resistant Prdx6-PLA2 into the knockdown cells by stable transfection with a Prdx6-PLA2 expression plasmid restored the fMLF response, as did reintroduction of Prdx6-PLA2 mutated in the Prdx active site; reintroduction of PLA2 active site mutants, however, failed to restore the response. Thus, the PLA2 activity of Prdx6-PLA2 in intact cells mediates its ability to enhance phox activity in response to fMLF. In combination with previous publications by other groups, our work indicates that various PLA2 isoforms can enhance oxidase activity but they are differentially important in different cell types and in the response to different agonists.

Introduction

The phagocyte NADPH oxidase (phox) is the major source of reactive oxygen species (ROS) in neutrophils, macrophages, and other phagocytic cells [[1]] and has also been described in endothelial cells [[2, 3]]. It is the founding member of a family of superoxide anion (O2) generating enzymes that has a broad tissue distribution [[4-7]]. Phox is composed of multiple protein subunits; in the inactive state, the membrane-associated subunit, cytochrome b558, a dimer of gp91phox/Nox2 and p22phox, is separate from various cytosolic subunits which include Rac1 or Rac2 plus, p47phox, p67phox, and p40phox [[1, 4, 8-10]]. Phosphorylation of phox subunits leads to translocation of the cytoplasmic proteins and their tight interaction with cytochrome b558. The resulting active complex catalyzes transmembrane electron transfer from cytosolic NADPH to molecular oxygen (the oxidative burst) forming O2 [[1, 4, 8-10]]; subsequent enzymatic and nonenzymatic steps produce a series of other ROS from the O2. A range of agonists activate signaling pathways that are responsible for the phosphorylation events leading to assembly and activation; these include the formylated peptide, fMLF, and the phorbol ester, PMA.

Phox has been previously studied in the PLB-985 myeloid leukemia cell line [[11]]. These cells can be differentiated to a mature neutrophil-like phenotype which possesses an agonist-stimulated oxidative burst that is abolished by suppression of gp91phox/Nox2 [[3]]; differentiated PLB-985 cells are thus a useful model for the phox activity of neutrophils which are short-lived and thus not easily amenable to genetic manipulation.

Accumulating evidence suggests that phospholipase A2 (PLA2) enzymes play a crucial role in activation of phox in myeloid and endothelial cells [[12-17]]. PLA2 enzymes constitute a large family of proteins [[18]] that catalyze the hydrolysis of the sn-2 acyl bond of phospholipids, thus generating the corresponding free fatty acid and lyso-phospholipid; both the fatty acid arachidonate and various lysophospholipids have been implicated in direct or indirect modulation of oxidase activity [[14, 16, 19-22]]. The specific PLA2 isoform that is involved in phox activation remains unclear; as outlined below, some studies indicate that cytosolic PLA2 (cPLA2) is the key enzyme while others implicate other PLA2 isoforms including peroxiredoxin 6-PLA2 (Prdx6-PLA2), a bifunctional enzyme with a peroxiredoxin (Prdx) active site in addition to its PLA2 active site [[23]].

In neutrophils from patients with rheumatoid arthritis, a cPLA2 inhibitor was found to block the respiratory burst triggered by a Ca2+ ionophore whereas a Ca2+-independent PLA2 activity was implicated in the oxidative burst triggered by PMA [[16]]. In differentiated PLB-985 cells, stable expression of an antisense RNA that suppressed cPLA2 expression was found to inhibit the oxidative burst triggered by fMLP, PMA, and serum-opsonized zymosan (SOZ). The extent of inhibition of O2 production was proportional to the extent of cPLA2 knockdown and the addition of the fatty acid arachidonic acid restored the PMA response, suggesting that it is this product of cPLA2 that promotes phox activity [[15, 24]]. In human neutrophils and PLB-985 cells, gp91phox-dependent binding of cPLA2 to membrane associated phox subunits was demonstrated after stimulation with fMLF, PMA, and SOZ suggesting that cPLA2 is physically associated with the phox complex; additionally, the time course of cPLA2 association with the active phox complex was shown to correlate with the onset of O2 production [[15, 24]]. Supporting the role of cPLA2 in the phox activity of phagocytic cells other than neutrophils, suppression of cPLA2 in human monocytes by antisense oligodeoxyribo-nucleotides inhibited O2 production and p47phox and p67phox translocation following SOZ stimulation; again the defects caused by cPLA2 suppression were reversed by the addition of exogenous arachidonic acid [[12, 13]].

In contrast to the above studies which suggest that cPLA2 is crucial for phox activity, our laboratory identified the PLA2 enzyme Prdx6-PLA2 as a binding partner of p67phox as well as an enhancer of phox activity in an in vitro reconstituted, system [[25]]. Subsequently, it was shown that macrophages and pulmonary microvascular endothelial cells (PMVECs) from Prdx6-PLA2 knockout mice have dramatically reduced oxidase activity in response to Angiotensin II (Ang II) and PMA stimulation [[17]]. In addition, in Ang II-stimulated PMVECs from the knockout mice, Rac1 and p47phox translocation was abolished suggesting that Prdx6-PLA2 plays a crucial role in phox assembly. Further, in response to Ang II treatment, activation of Prdx6-PLA2 by phosphorylation was shown to be necessary for phox activity and PLA2-isoform-specific inhibitors indicated that Prdx6-PLA2, rather than other PLA2 enzymes, was necessary for the oxidative burst. Further evidence against a role for cPLA2 in phox activity was obtained when neutrophils from cPLA2 knockout mice were found to have normal oxidative bursts triggered by fMLF, PMA and SOZ [[26]].

The studies outlined above indicate that both cPLA2 and Prdx6-PLA2 have been implicated in phox activity and thus reveal ambiguity regarding the role of PLA2 enzymes in the oxidative burst.

Because PLB-985 cells [[11]] have been shown to depend on cPLA2 for PMA and fMLF [[15, 24]] stimulated phox activity, we investigated the possibility of overlapping roles for cPLA2 and Prdx6-PLA2 in these cells. Prdx6-PLA2 was knocked down in PLB-985 cells by stable expression of a small hairpin RNA (shRNA) and the effects on agonist-stimulated phox activity were measured.

Results

Prdx6-PLA2 knockdown

A PLB-985 cell line with knockdown of Prdx6-PLA2 (KD) and a nonsilenced, wild type, control cell line (WT) were established, as detailed in the Materials and methods, by stable transfection with a plasmid expressing a Prdx6-PLA2-targeting shRNA (for KD) or a nonsilencing shRNA (for WT). The KD and WT cells were differentiated with DMSO (1.3%, 4 days) and western blot analysis revealed that Prdx6-PLA2 expression was suppressed in the KD cells but the phox components gp91phox, p67phox, p47phox, p40phox, p22phox, and rac2 were unchanged (Fig. 1). Densitometry of western blots from seven independently differentiated WT and KD cultures was used to quantitate Prdx6-PLA2 knockdown; the protein was suppressed by 68 ± 6% (mean ± SEM).

Figure 1.

Prdx6-PLA2 suppression using shRNA. A PLB-985 cell line with stable knockdown of Prdx6-PLA2 (KD) and a nonsilenced, wild type, control cell line (WT) were established. Following DMSO-mediated differentiation, western blot analysis was used to confirm suppression of Prdx6-PLA2 in the KD cell line (top left) and equal expression of other oxidase proteins in the two cell lines (bottom 6 panels). Actin immuno-reactivity was used as a loading control. The Prdx6-PLA2 knockdown is representative of western blots done following seven independent differentiations of the two cell lines and the equivalence of the other oxidase proteins is representative of two such experiments.

Knockdown of Prdx6-PLA2 did not affect terminal differentiation induced by DMSO as assessed by two measures. During 4 days of DMSO exposure, p67phox, p47phox, and p40phox rose from undetectable levels (p67phox and p40phox) or low levels that were equal in WT and KD cells (p47phox, data not shown) to the higher levels that were the same in WT and KD cells (Fig. 1). Thus, the rate at which these proteins were up-regulated by differentiation was not altered by Prdx6-PLA2 knockdown. Also, while DMSO-mediated terminal differentiation caused a progressive increase in the fraction of dead cells as measured by permeability to trypan blue, the increase in dead cells over time was no different between the WT and KD lines, even when they were exposed to DMSO for 7 days rather than the 4 days used to prepare cells for measurement of phox activity. Less than 1% of WT and KD cells were permeable before differentiation whereas 40 ± 6% of WT and 36 ± 8% of KD (mean ± SEM, n = 5) were permeable after 7 days of differentiation.

Prdx6-PLA2 enhances Phox activity stimulated by fMLF and but not PMA

PMA and fMLF are agonists that promote assembly and activation of phox; fMLF binds cell surface receptors triggering downstream signaling pathways that promote phosphorylation of phox components and activation the complex, PMA on the other hand promotes phox assembly by direct activation of intracellular protein kinase C enzymes that carry out the necessary phosphorylation. Stimulation of the WT and KD cells with 1 μM fMLF resulted in transient bursts of O2 production that were detected by superoxide dismutase (SOD) inhibitable luminescence of Diogenes reagent. As can be seen (Fig. 2A and B), these bursts were smaller in the KD cells. The difference in O2 production by the WT and KD cells was significant (p < 0.001). Reduced O2 production by the KD cells was also seen when the maximal rate of SOD inhibitable cytochrome C reduction was used to measure O2 form-ation (Fig. 2C) and the difference was again significant (p < 0.05). In contrast, PMA caused prolonged O2 production and no difference was observed between WT and KD cells stimulated with 10 ng/mL, 200 ng/mL, or 1 μg/mL PMA (Fig. 2D). Since the different concentrations of PMA produced responses with very different initial rates and overall kinetics, the data show that Prdx6-PLA2 does not enhance phox activity specifically under conditions of weak or strong PMA-induced activation. The data in Fig. 2 represent the comparison of one WT and KD cell line. To confirm the result, a second wild type (WT2) and knockdown (KD2) cell line were characterized, Prdx6-PLA2 suppression in KD2 relative to WT2 was confirmed by western blot (not shown) and reduced O2 production in response to fMLP by the KD2 cell line was confirmed using the luminescence-based assay; levels of O2 prod-uction, normalized as described in the Materials and methods, were 1.00 ± 0.04 by WT2 and 0.58 ± 0.03 by KD2 (mean ± SEM, n = 6 from two independently differentiated cultures) and are significantly different (p < 0.01). PMA induced superoxide production was no different in WT2 and KD2 (not shown).

Figure 2.

Effect of Prdx6-PLA2 knockdown on O2 production in response to fMLF and PMA. (A) DMSO-differentiated WT or KD cells were stimulated with 1 μM fMLF at time zero and O2 was monitored by luminescence (WT, black line; KD, dotted line). The data are representative of multiple experiments as outlined in (B). (B) The average peak areas for the WT and KD cells, obtained in the multiple experiments like that in (A), are shown with normalization such that the average peak area for the WT cells is 1 (see Materials and methods). Data are plotted as mean ± SEM of n = 34 measurements pooled from seven independently differentiated cell cultures. (C) The average maximum rate of O2 production in response to stimulation by 1 μM fMLF, as measured by SOD inhibitable cytochrome C reduction, is shown as mean ± SEM of n = 16 measurements pooled from three independently differentiated cell cultures. (D) A representative experiment in which DMSO-differentiated WT (black line) or KD (dotted line) cells were stimulated with PMA (1 μg/mL, 200 ng/mL, or 10 ng/mL; the upper, middle, and lower pair of curves, respectively) at time zero and O2 was monitored by luminescence. The data are representative of six experiments from two independently differentiated cell cultures.

The PLA2 activity of Prdx6-PLA2 is necessary for maximal O2 production in response to fMLF

To determine whether the Prdx or PLA2 activity of Prdx6-PLA2 [[23]] is required for phox activation, the KD cell line was stably transfected with plasmids (thus generating knock in [KI] cell lines) expressing wild-type Prdx6-PLA2 (referred to as “WT(KI)” in Fig. 3), various mutants (H26A, S32A, and D140A) which abolish PLA2 activity [[27]] and C47S-Prdx6-PLA2 which abolishes Prdx activity [[27]]. An empty vector (EV) cell line with no reintroduction of Prdx6-PLA2 was also established. All protein-expressing constructs were designed to be resistant to silencing by the shRNA in the KD cell line. Expression of the exogenous proteins was confirmed by western blot analysis (Fig. 3A). Reintroduction of wild type and C47S-Prdx6-PLA2 reversed (rescued) the reduction in the response to fMLF seen in the KD cells but reintroduction of the various PLA2 active site mutants and the EV control did not (Fig. 3B, C, and D); O2 production by WT(KI) and C47S was significantly different from each of the PLA2 active site mutants and the EV control (p values < 0.05). These data thus suggest that the PLA2 activity but not the Prdx activity of Prdx6-PLA2 stimulates fMLF-mediated phox activity in DMSO-differentiated PLB-985 cells. Each bar in Fig. 3D represents data from a single clonal KI cell line although we were able to isolate a second cell line for EV, H26A, and S32A and their responses to fMLP were also reduced relative to the WT(KI) and C47S cell lines; levels of O2 production, normalized as described in the Materials and methods, were 0.56 ± 0.02, 0.54 ± 0.09, and 0.57 ± 0.14, respectively (mean ± SEM, n of at least 6). These amounts of O2 were significantly different from those produced by the WT and C47S cell lines (p values < 0.05). Gross defects in folding of the PLA2 active site mutants do not account for their failure to “rescue” the fMLF response because we (unpublished observations) and others [[27]] have shown that recombinant H26A and D140A-Prdx6-PLA2 have circular dichroism spectra very similar to wild type.

Figure 3.

O2 production by the oxidase is dependent on the PLA2 active site of Prdx6-PLA2. The KD cell line was stably transfected with plasmids expressing wild-type Prdx6-PLA2 (“WT(KI)” in the figure to distinguish it from the cell line with no knockdown of Prdx6-PLA2), various mutants of Prdx6-PLA2, or an empty vector (EV) control to create a set of knock in (KI) cell lines. (A) Reintroduction of wild-type Prdx6-PLA2 and the various mutants was confirmed by western blot analysis of DMSO-differentiated cells; for each KI cell line, immunoreactivity for Prdx6-PLA2 (bottom) and actin-loading control (top) are compared with same-blot immunoreactivity from the parental cell line, that is KD. The blots are representative of at least two repetitions from independently differentiated cell cultures. (B) DMSO-differentiated EV cells (dotted line) or WT(KI) cells (black line) were stimulated with 1 μM fMLF at time zero and O2 was monitored by luminescence. The data are representative of multiple experiments as outlined in (D). (C) DMSO-differentiated WT(KI) cells (black line), cells with reintroduction of the Prdx active site mutant (C47S, long dashes), and cells with reintroduction of the PLA2 active site mutants D140A (dots), S32A (short dashes), and H26A (open dashes) were stimulated with 1 μM fMLF at time zero and O2 was monitored by luminescence. The data are representative of multiple experiments as outlined in (D). (D) The average peak areas obtained in multiple experiments like those in (B) and (C) are shown with normalization such that the average peak area for the WT(KI) cells is 1 (see Materials and methods). Data are plotted as mean ± SEM of at least seven measurements pooled from at least two independently differentiated cell cultures.

Retinoic acid (RA) differentiation does not change the relative responses to fMLF and PMA

In a previous study [[15]], stable expression of an antisense RNA was used to knockdown cPLA2 in PLB-985 cells and deficiency of this protein was found to result in reduced O2 production in response to PMA and fMLF. In this earlier work, the PLB-985 cells were differentiated toward a granulocyte-like phenotype using RA. To understand our results in the context of this study, the experiments presented in Fig. 2A and B were repeated with differentiation of our WT and KD cells using RA instead of DMSO. Additionally, fMLF was applied at 0.1 μM and PMA was applied at 100 ng/mL, the concentrations used in the earlier study. As shown in Fig. 4A and B, the peaks of O2 stimulated by fMLF were sharper after RA differentiation than DMSO differentiation; nevertheless, knockdown of Prdx6-PLA2 caused a relative reduction in O2 production which was similar to that seen with DMSO-differentiated cells stimulated by fMLF (Fig. 1A and B). The responses of the WT and KD cells were significantly different (p < 0.001). As with DMSO-differentiated cells, RA-differentiated cells generated O2 in a sustained and Prdx6-PLA2-independent fashion in response to PMA. Therefore, the effect of Prdx6-PLA2 knockdown on agonist-induced phox activity in PLB-985 cells does not appear to depend on the nature of the differentiation agent; both DMSO and RA treatment results in cells with an fMLF response that is reduced in the KD line and a PMA response that is unchanged.

Figure 4.

The effect of Prdx6-PLA2 knockdown on responses to stimuli are the same in RA-differentiated cells and DMSO-differentiated cells. (A) RA-differentiated WT or KD cells were stimulated with 0.1 μM fMLF at time zero and O2 was monitored by luminescence. The data are representative of multiple experiments as outlined in (B). (B) The average peak areas for the WT and KD cells, obtained in the multiple experiments like that in (A), are shown with normalization such that the average peak area for the WT cells is 1 (see the Materials and methods). Data are plotted as mean ± SEM of n = 17 measurements pooled from three independently differentiated cell cultures. (C) RA-differentiated WT (black line) or KD (dotted line) cells were stimulated with 100 ng/mL PMA added at time zero and O2 was monitored by luminescence. The data are representative of three experiments from two independently differentiated cell cultures.

Discussion

This study indicates that, in PLB-985 cells, the PLA2 activity of Prdx6-PLA2 is necessary for maximal oxidase activity in response to fMLF but does not impact the response to PMA. This is in contrast to the previous study by Dana et al. [[15]] in which antisense RNA was used to knockdown cPLA2 in PLB-985 cells and various levels of depletion of cPLA2 resulted in matching reductions in O2 production stimulated by fMLF and PMA. Together, our study and the earlier work indicate that the fMLF response of differentiated PLB-985 cells is dependent on both cPLA2 and Prdx6-PLA2 but that the PMA response is either not dependent on Prdx6-PLA2 or is less effected by depletion of this protein (undetectably so in this study). Therefore, the role of PLA2 enzymes in oxidase activation in differentiated PLB-985 cells appears to be complex with Prdx6-PLA2 and cPLA2 necessary for the fMLP response but only cPLA2 needed for the PMA response.

A possible reason for the selective role of Prdx6-PLA2 in the fMLP response is differential regulation of Prdx6-PLA2 and cPLA2. The phospholipase activity of Prdx6-PLA2 is known to be modulated by phosphorylation [[17, 28]] and that of cPLA2 by phosphorylation, Ca2+, various lipids and various protein-binding partners [[29]]. In differentiated PLB-985 cells, the intracellular signaling pathways triggered by fMLP and PMA likely result in different patterns of activation/mobilization of intracellular signaling factors including kinases, signaling lipids, and Ca2+. Those factors agonized by fMLP may activate Prdx6-PLA2 and cPLA2 whereas those activated by PMA might activate only cPLA2 thus accounting for the involvement of Prdx6-PLA2 and cPLA2 in the fMLP response but only cPLA2 in the PMA response.

Recently, Chatterjee et al. [[17]] demonstrated that PMVECs from Prdx6-PLA2 knockout mice show dramatically reduced ROS production in response to stimulation with PMA due to loss of the PLA2 activity associated with the enzyme; in contrast, inhibitors of other PLA2 isoforms, including cPLA2, did not reduce the response to PMA. In contrast, our current data reveal no reduction in the PMA response following Prdx6-PLA2 knockdown indicating cell type (PMVEC vs. PLB-985) and/or species (human vs. mouse) differences in the role Prdx6-PLA2 plays in mediating the PMA response. In contrast, similar to our finding that fMLF-stimulated O2 production was reduced in the KD PLB-985 cells, Chatterjee et al. [[17]] observed that fMLF-stimulated ROS production by macrophages from Prdx6-PLA2 knockout mice was reduced. Thus, the relative roles of different PLA2 enzymes in oxidase activity are even more complex when considered across different cell types and species, some oxidase agonists are consistently dependent on a given PLA2 isoform in various cell types and species but others are not.

The influence of cell type, species, and antagonist on the relative importance of different PLA2 isoforms in oxidase activity could be the result of a number of factors. One possibly significant factor is different relative levels of PLA2 enzymes in different cell types; a given PLA2 isoform might be crucial for the response to a given phox agonist in cells where it is abundant but other PLA2 isoforms might take over its role in cells where it is not abundant. Another possibility is that distinct intracellular signaling pathways activated by different phox agonists and specifically activating different PLA2 isoforms could be present at different levels in different cell types. Thus, a particular PLA2 isoform might be important in the oxidative burst triggered by a particular agonist only in cells with an intact signaling pathway that is stimulated by that agonist and which activates that PLA2. Additionally, across species, variations in the catalytic efficiency and activation mechanism of a given PLA2 isoform might govern the relative importance of that isoform, a lower specific activity might reduce the relative importance of the isoform in phox stimulation as might the absence of sequence features necessary for its activation by relevant signaling molecules.

In summary, the PLA2 activity of Prdx6-PLA2 is necessary for fMLF-stimulated phox activity in a neutrophil cell model; in contrast, the PMA response was not detectably altered by suppression of Prdx6-PLA2. Our current data, in combination with previous work [[12, 13, 15, 17, 24, 26]] suggests that Prdx6-PLA2 and cPLA2 play a role in phox activation but that they are differentially important depending on the cell type, species, and agonist and warns against definitive assignment of a particular PLA2 as a global cofactor for phox activity.

Materials and methods

Western blot

Cells (0.5 × 106) were pelleted (250G, 5 min) from media in microcentrifuge tubes, supernatant was aspirated and the cell pellets were washed twice by resuspension in 1 mL of ice cold kRPD (12.5 mM Na2HPO4, 3 mM NaH2PO4, 4.8 mM KCl, 120 mM NaCl, 1.3 mM CaCl2, 1.2 mM MgSO4, 0.2% dextrose, pH 7.3–7.4) followed by recentrifugation. The pellets were resupended in 20 μL kRPD and 10 μL of previously described [[30]] SDS-polyacrylamide gel electrophoresis (SDS-PAGE) sample buffer. The cells were disrupted by vortexing and incubation at 90°C for 5 min. The disrupted cells were subjected to SDS-PAGE as described previously [[30]] and proteins were blotted to nitrocellulose and probed with antibodies (see below) as described earlier [[31]]. Immune complexes on the nitrocellulose were detected using enhanced chemiluminescence detection (GE Healthcare). The primary antibodies used were rabbit antihuman p22phox (sc-20781), rabbit antihuman Rac2 (sc-96), and goat antihuman gp91phox (sc-27635) from Santa Cruz Biotechnology; rabbit antihuman p40phox (P1001-20C) and mouse anti-Xenopus laevis beta actin (A0760-40) from US Biological (the loading control); goat antihuman 47phox and goat antihuman p67phox, described previously, [[32]], were gifts from Dr. TL Leto; and, rabbit antihuman Prdx6-PLA2 which was prepared commercially (Global Peptide Services) in rabbits using previously described recombinant Prdx6-PLA2 [[25]] as antigen. The primary antibodies were used at a 1:1000 dilution except the antibody against beta actin which was used at a 1:2000 dilution. The secondary antibodies used were peroxidase linked antirabbit IgG and antimouse IgG (GE Healthcare) and peroxidase linked antigoat IgG (Rockland Immunochemicals Inc.). The secondary antibodies were used at a 1:1000 dilution. All blots were reprobed for actin after the initial development to ensure equal loading.

Cell culture

PLB-985 cells (from the German Collection of Microorganisms and Cell Cultures) were grown in stationary suspension culture in RPMI 1640 medium containing 10% bovine serum, 2 mM L-glutamine, 100 units/mL penicillin, and 100 mg/mL streptomycin at 37°C in a humidified 5% CO2 atmosphere. Light microscopy was used to determine cell number and, by trypan blue exclusion, cell viability.

Stable transfection of PLB-985 cells

For each cell line established, 2 × 106 log phase PLB-985 cells were transfected (solution V, program C-023) with 2 μg of plasmid DNA using an Amaxa Nucleofector (Lonza) according to the manufacturer's instructions. Transfected cells were maintained in the absence of penicillin and streptomycin for 7 days and were then moved to the standard culture media supplemented with additional selective antibiotics (see below). The cells were transferred to 96-well tissue culture plates at approximately 3 × 104 cells per well and antibiotic resistant colonies appeared after about 5 weeks. Cells from wells in which a single colony had developed were expanded in the same media used for initial selection and were screened by Prdx6-PLA2 western blot for suppression of Prdx6-PLA2 or subsequent reintroduction of Prdx6-PLA2 or Prdx6-PLA2 mutants (see Suppression of Prdx6-PLA2 in PLB-985 cells and Reintroduction of various Prdx6-PLA2 mutants in the KD cell line in Materials and methods).

Suppression of Prdx6-PLA2 in PLB-985 cells

Selection using 1 μg/mL puromycin established PLB-985 cell lines stably transfected with four different pRS-based plasmids encoding shRNA molecules predicted to silence Prdx6-PLA2 (Origene). Cell lines stably transfected with a pRS-based plasmid (TR30003) encoding a noneffective, scrambled shRNA (Origene) were established in parallel. Western blot analysis (see Results) indicated that cell lines arising from transfection with plasmid TI340760 demonstrated the greatest knockdown of Prdx6-PLA2. One of these, the KD cell line, was used in further studies as was a cell line arising from transfection with plasmid TR30003, the nonsilenced WT control.

Reintroduction of various Prdx6-PLA2 mutants in the KD cell line

As discussed in Molecular Biology, the open reading frame (ORF) for wild-type Prdx6-PLA2 and various mutants was cloned into pCDNA3.1- (Invitrogen). Each ORF included a series of silent mutations to block annealing of the Prdx6-PLA2 mRNA and the shRNA that triggers Prdx6-PLA2 silencing in the KD cell line, thus making the various exogenous Prdx6-PLA2 molecules resistant to knockdown. The silent mutations are indicated in the following sequence where the first A is nucleotide 690 of the Prdx6-PLA2 cDNA, the nucleotide positions are separated by commas and nucleotide changes are indicated by “X to Y” in the mutated nucleotide positions (X is the nucleotide found in the wild-type cDNA and Y is the nucleotide incorporated into the shRNA resistant cDNAs); A, T to C, A to T, G to C, T to C, T, G to C, A, T, G, G, T, C to G, C to T, T, T to G, C, C, A to T. The KD cell line was transfected with the various pCDNA3.1-based plasmids and stable cell lines were selected in the presence of 800 μg/mL G418 plus 1 μg/mL puromycin as outlined above in “Stable Transfection of PLB-985 Cells.”

Molecular biology

An EcoRI/AflII fragment of a Prdx6-PLA2 ORF with an N-terminal tag [[25]] was subcloned into pCDNA3.1- such that Prdx6-PLA2 expression (without the tag) could be driven from the plasmids CMV promoter and the native 13 nucleotides upstream of the initiating methionione were present. QuickChange mutagenesis (Agilent Technologies) was carried out according to the manufacturer's instructions to create the silent mutations necessary for shRNA resistance (see Reintroduction of various Prdx6-PLA2 mutants in the KD cell line in Materials and methods) and to create the following Prdx6-PLA2 mutants; C47S, H26A, S32A, and D140A. All constructs were confirmed by DNA sequencing at a University of Colorado at Denver core facility.

Cell differentiation

The cell lines WT, KD, and KD with various versions of Prdx6-PLA2 reintroduced were differentiated with 1.3% (vol:vol) DMSO (4 days) or 1 μM RA (5 days). Cell viability of the differentiated cells used to measure phox activity was always greater than 75% and never varied by more than 10% between different cell lines differentiated at the same time and used for direct comparison in a given experiment.

Measurement of O2 production

Cells were pelleted from media (90G, 10 min) in 50 mL conical tubes, washed once by resuspension in 20 mL of ice cold kRPD, repelleted, and resuspended to approximately 15 million cells per milliliter in the kRPD. Cells were kept on ice and were used within 3 h.

O2 production was measured by luminescence using a LUMIstar Optima luminometer (BMG Labtech); 5 × 105 cells were included in 150 μL assays in the wells of white 96-well plates (Nunc). Each assay was composed of 30 μL of Diogenes (National Diagnostics), an O2 specific enhanced luminol that produces light in direct proportion to the O2 concentration, and 120 μL of cells and agonist (fMLF or PMA, both from Sigma) in kRPD. Agonist concentrations were as indicated in the text. The PMA was dissolved in DMSO such that the final concentration of DMSO in each assay was 1%. The cells were preincubated at 37°C for at least 10 min (pilot studies indicated that 10- and 30-min preincubations made no difference to the response) and fMLF and Diogenes (prewarmed to 37°C) were then injected, in 1 s, by the fluid-handling system of the instrument or PMA and Diogenes were added manually. After shaking, the amount of light detected over the course of 1 s was measured every second for 4 min (fMLF) or every 20 s (after additional shaking) for at least 30 min (PMA). In response to fMLF, peaks like that in Fig. 1A were observed and the peak areas were determined by summing all light measurements made during assay.

The magnitude of luminescence produced in response to agonist by a given cell line varied quite widely between experiments which were preformed on different days that were weeks to months apart. For example, compare the KD cell line stably transfected to express wild-type Prdx6-PLA2 (black lines) in Fig. 3B and C which are representative traces from experiments done on different days using independently differentiated cells. On the other hand, the relative magnitudes of luminescence produced by different cell lines were consistent day to day. Thus, the following normalization procedure was used. On each day that assays were done comparing WT and KD cells, the average peak area for WT cells (n of at least 3) was set at 1 and all peak areas recorded on that day were normalized to this value. On each day that KD cells with reintroduced Prdx6-PLA2 were assayed, cells with reintroduced WT Prdx6-PLA2 were included (n of at least 3) and the average peak area recorded in response to fMLF for these cells was set to 1; all peak areas recorded on this day were then normalized to this value.

SOD from Sigma, at 15 μg/mL, was included in control experiments and >98% (fMLF response) or >90% (PMA response) of the luminescence at every time point was SOD inhibitable.

For some studies, O2 production in response to fMLF was measured spectrophotometrically by following SOD inhibitable reduction of cytochrome C in a THERMOmax microplate reader (Molecular Devices) as described previously for PMN [[31]].

Statistical analysis

The two-tailed, unpaired Student's t-test was used in all cases.

Acknowledgements

Funding for this work was provided by Bonfils Blood Center, the Stacy Marie True Memorial Trust, and NIH grant NHLBI, NIH, K07 HL088968. The DNA samples were sequenced by the University of Colorado Cancer Center DNA Sequencing and Analysis Core (http://DNASequencingCore.ucdenver.edu), which is supported by an NIH/NCI Cancer Center Core Support Grant (P30 CA046934). We thank Christian Snyder for help with manuscript preparation and submission and Thomas Leto (PhD) for helpful comments.

Conflict of interest

The authors declare no financial or commercial conflict of interest.

Abbreviations
Ang II

angiotensin II

cPLA2

cytosolic phospholipase A2

KD

knockdown of Prdx6-PLA2 in PLB-985 cell line . KD2 second knockdown

Prdx6-PLA2

peroxiredoxin 6-phospholipase A2

Prdx

peroxiredoxin

phox

phagocyte NADPH oxidase

PLA2

phospholipase A2

PMVEC

pulmonary microvascular endothelial cell

RA

retinoic acid

SOD

superoxide dismutase

WT2

second WT.

Ancillary