SEARCH

SEARCH BY CITATION

Keywords:

  • plant G(alpha) protein;
  • phospholipase A2;
  • plant signal transfer;
  • membrane protein complexes;
  • plasma membrane proteins

Summary

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Conclusions
  6. Experimental procedures
  7. Acknowledgements
  8. References

Plant heterotrimeric G-proteins are involved in a variety of signaling pathways, though only one α and a few βγ isoforms of their subunits exist. In isolated plasma membranes of California poppy (Eschscholzia californica), the plant-specific Gα subunit was isolated and identified immunologically and by homology of the cloned gene with that of several plants. In the same membrane, phospholipase A2 (PLA2) was activated by yeast elicitor only if GTPγS (an activator of Gα) was present. From the cholate-solubilized membrane proteins, PLA2 was co-precipitated together with Gα by a polyclonal antiserum raised against the recombinant Gα. In this immunoprecipitate and in the plasma membrane (but not in the Gα-free supernatant) PLA2 was stimulated by GTPγS. Plasma membranes and immunoprecipitates obtained from antisense transformants with a low Gα content allowed no such stimulation. An antiserum raised against the C-terminus (which in animal Gαs is located near the target coupling site) precipitated Gα without any PLA2 activity. Using non-denaturing PAGE, complexes of solubilized plasma membrane proteins were visualized that contained Gα plus PLA2 activity and dissociated at pH 9.5. At this pH, PLA2 was no longer stimulated by GTPγS. It is concluded that a distinct fraction of the plasma membrane-bound PLA2 exists in a detergent-resistant complex with Gα that can be dissociated at pH 9.5. This complex allows the Gα-mediated activation of PLA2.


Introduction

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Conclusions
  6. Experimental procedures
  7. Acknowledgements
  8. References

Despite intensive research, it is still not understood how a single or very few isoforms of the subunits of plant heterotrimeric G-proteins (one Gα, one Gβ and two Gγ exist in Arabidopsis; Jones, 2002) can be involved in a variety of regulatory pathways as suggested by pharmacological, mutant or knock-down experiments (Adjobo-Hermans et al., 2006;Assmann, 2004, 2005; Jones and Assmann, 2004;Perfus-Barbeoch et al., 2004; Warpeha et al., 2006). While their animal counterparts generate a vast structural diversity by combining different subunits (23 Gα, 6 Gβ, and 12 Gγ proteins) to make task-specific trimers, plant cells obviously use different mechanisms to control various effector proteins with the same or structurally closely related G-proteins. Two alternative strategies emerge from the findings of the last years: on the one hand, Gα or the Gβγ complex function as modulators of signal conditions or components, without being directly involved in the signal transduction sequence, as shown for several hormone-triggered signal cascades (Assmann, 2004, 2005; Ullah et al., 2001, 2003). An alternative mode, the physical interaction of the same G-protein subunit with different effector proteins, has also been exemplified in a few cases. This is true for phospholipase D (Lein and Saalbach, 2001; Ritchie and Gilroy, 2000; Zhao and Wang, 2004), ion influx channels for potassium (Wang et al., 2001) and calcium (Aharon et al., 1998) and for prephenate dehydratase, an enzyme of aromatic amino acid biosynthesis (Warpeha et al., 2006). Seed germination was found to require the interaction of Gα with a protein of the pirin group, but no change of activity of this interactor has been reported (Lapik and Kaufman, 2003).

Phospholipase A2 (PLA2) is involved in a number of signaling processes in plants, including wounding and pathogen defense (Narvaez-Vasquez et al., 1999; Ryu, 2004; Wang, 2001). A G-protein-dependent control of this enzyme is not yet known. Both proteins are present in the plasma membrane of Eschscholzia californica (California poppy), the object of the present study, but no regulatory interaction has hitherto been established (Roos et al., 1999). Phospholipase A2 functions in a signal path initiated by a yeast glycoprotein elicitor. Lysophosphatidylcholine (LPC) produced by the elicitor-activated enzyme is an essential second messenger that causes an efflux of vacuolar protons which in turn triggers the pH-dependent overexpression of benzophenanthridine alkaloid biosynthesis. The requirement of a G-protein for this signal pathway was suggested from regulatory deficiencies of cell lines that were transformed to express either antisense-Gα mRNA or anti-Gα antibodies. These cells were lacking both the elicitor activation of PLA2 and the elicitor induction of alkaloid biosynthesis (Viehweger et al., 2006). Whether PLA2 is directly controlled by physical contact with Gα and the mode of putative interaction of these proteins are unknown. Our present work provides new evidence for a functional coupling of both proteins in pre-formed complexes of the plasma membrane.

Results and discussion

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Conclusions
  6. Experimental procedures
  7. Acknowledgements
  8. References

The Gα protein of the plasma membrane of E. californica represents the canonical plant Gα

Among the solubilized plasma membrane proteins of E. californica, a band of 44.5 kDa was detected by two different polyclonal anti-Gα antisera (Figure 1a). These sera had been raised against the Gα from Arabidopsis, i.e. against the C-terminus (W6) or the total recombinant Gα protein (P33), respectively (White et al., 1993).

image

Figure 1.  Immunological detection of Gα in the plasma membrane of Eschscholzia californica. (a) Detection of the Gα protein by different antisera. Western blotting of solubilized plasma membrane proteins after SDS-PAGE. Left: detection with antiserum P33; right: detection with antiserum W6. Intensities are normalized to the Gα bands (44.5 kDa) which are set to a similar and typical gray value. Intensities detected outside these bands are usually below 5%. Size markers are in kDa. (b) Effect of ADP-ribosylation at the immunological detection of Gα. Lanes 1–4 show western blotting with antiserum W6 (lane1 and 2) and P33 (lane 3 and 4) following SDS-PAGE of untreated or cholera toxin (CTX)-treated plasma membrane solubilizate. Lanes 2 and 4: plasma membrane vesicles pre-treated by CTX-catalyzed ADP-ribosylation. Lane 5: 32P autoradiography of SDS-PAGE-separated plasma membrane vesicles following CTX-catalyzed ADP-ribosylation. Insert: non-labeled and labeled NAD+ compete for the ribosylation site. Experiments like that of lane 5: a, labeled NAD+; c, as in a, +200 μm NAD+; d, as in a, +500 μm NAD+; b, absence of CTX.

Download figure to PowerPoint

The putative Gα protein was enriched by immunoprecipitation on matrix-coupled W6 antibodies and yielded a protein fraction that banded at 43–45 kDa on an SDS-PAGE gel.

For identification, this band was digested with trypsin and two of the resulting peptides were separated and sequenced by liquid chromatography/quantitative time-of-flight mass spectrometry. Their amino acid sequences, together with a non-overlapping peptide derived from a 300-bp cDNA fragment of the Eschscholzia Gα gene (see Experimental procedures) fitted with high coincidence into the Gα sequences known from different plants. The complete Gα gene of Eschscholzia was amplified by PCR with specific primers from a cDNA library. The derived amino acid sequence, aligned to several plant Gα homologs, is shown in Table 1. The high homology explains the cross-reactivity and justifies the use of the above-mentioned antibodies raised against the Arabidopsis protein.

Table 1.   Sequence alignment of Gα proteins from Eschscholzia, Lupinus and Arabidopsis
  1. Search results obtained with the NCBI BLASTp algorithm: first row, protein derived from the E. californica gene sequence; second row, GPA1 of Lupinus luteus (accession no. Q40224), identity 84%, similarity 93%; third row, GPA1 of Arabidopsis thaliana (accession no. NP_180198.1), identity 82%, similarity 91 %.

  2. Two peptides obtained from the trypsin digestion of the isolated Eschscholzia Gα protein are underlined, as well as the sequence 188–287 which is derived from the DNA probe used for the cDNA library screening (see text). An arrow marks the site of cholera toxin (CTX)-catalyzed ADP ribosylation (Arg 191).

  3. Gα proteins of many other species likewise share high sequence identities with the Eschscholzia protein, among them Phaseolus lunatus (BAF30964) 84%; Lotus japonicus (P49082) 84%; Pisum sativum, PGA1 (O04279) 83%; Pisum sativum, PGAII (AAM97353) 82%; Solanum tuberosum (P93564) 82%; Nicotiana tabacum (AAG12329) 82%.

  1 MGSLCS-RHNHRNEGDSEENVQAAEIEKRI-AQETKAEKHIQKLLLLGAGESGKSTIFKQ 58
  1 ..L...-.NRRY.DA.A...A......R..-EL........................... 58
  1 ..L...-.SR.HT.-.TD..T......R..-E..A......R.................. 57
59 IKLLFQTGFDDAELKSYTSVIYANVYQTIKVLHDGAKEFSETEEDSSKYLLSPENRDIGE 118
59 ..........E......LP..H...F....L....S..LAQNDV.....VI.D..K.... 118
58 ..........EG.....VP..H........L....T...AQN.T..A..M..S.SIA... 117
119 KLSGIGSRLDHPRLTAELAHEIETLWKDSAVQETYTRGNELQLPDCAHYFMENLQRLSDV 178
119 ...E...K..Y.Y..T...K......E.A.I....A......V.G..............A 178
118 ...E..G...Y....KDI.EG.......P.I...CA......V...TK.L....K....I 177
                    [DOWNWARDS ARROW]
179 NYVPTKEDVLHARVRTSGVVEIQFSPVGENKKSGEVYRLFDVGGQRNERRKWIHLFEGVN 238
179 ..........Y.....T..............R...........................S 238
178 ..I.......Y.....T..........................................T 237
239 AVIFCAAISEYDQMLFEDENKNRMMETKELFEWVLKQPCFEKTSFMLFLNKFDVFEKKVL 298
239 .............T..........T........I...................I....I. 298
238 .............T.....Q...........D.....................I...... 297
299 QVPLNVCDWFK-DYQPVSTGKQEIENAYEFVKKKFEEMYFQSTGPDRVDRVFKVYRTTAL 357
299 K......E...-.............H...........L.....A.E.............. 357
298 D......E..R-......S......H...........L.Y.N.A.........I...... 356
358 DQKLVKKTFKLVDETLRRRNLLEAGLL 384
358 ....I.........S......F..... 384
357 ........................... 383

The aligned proteins share the conserved site of cholera toxin (CTX)-catalyzed ADP-ribosylation around Arg 191 (Ma et al., 1990). The transfer of the adenyl moiety of NAD+ to this Arg is a specific activity of CTX that is often used to characterize animal and plant Gαs (Assmann, 2005; Jones, 2002; Ma et al., 1990). In our experiments, incubation of the plasma membrane with CTX and 32P-NAD+ yielded two radiolabeled protein bands during SDS-PAGE. The strongest 32P incorporation was found in a 44-kDa band that migrated in a similar way to the one detected by the anti-Gα antibodies. The incorporation of labeled NAD+ was diminished by excess unlabeled NAD+ and lacking in the absence of CTX (Figure 1, lane 5 and insert). These results not only identified the immunologically detected 44.5-kDa protein as a plant Gα but also substantiated that only one protein of this type was present in the Eschscholzia cells.

The elicitor stimulation of PLA2 in the plasma membrane requires GTP

In intact cells, activity of PLA2 at the cellular surface is rapidly stimulated by contact with a yeast glycoprotein elicitor (Roos et al., 2006; Viehweger et al., 2002, 2006). In the isolated plasma membrane, the enzyme is active but no such stimulation was observed.

In order to test whether a heterotrimeric G-protein was required for stimulation, we made use of a common feature of the Gα subunit: a conformational change (‘activation’) triggered by GTP or (permanently) by its non-hydrolyzable analog GTPγS, that can be conveyed into an activation of target proteins. In contrast, GDP or its non-hydrolyzable analog GDPβS stabilize the inactive state of Gα (Assmann, 2002; Warpeha et al., 2006; Zhao and Wang, 2004). In the above experiments, the stimulating effect of the elicitor required the presence of low micromolar concentrations of GTP or GTPγS, that alone caused less but significant stimulation. Thus, yeast elicitor plus GTPγS evoked a higher PLA2 activity than either component alone. No stimulation occurred with GDPβS or ATP (Figure 2). This result was considered to hint at the GTP-dependent interaction of Gα and PLA2 in the plasma membrane, although at this stage a direct influence of the nucleotide on the phospholipase could not be excluded.

image

Figure 2.  Phospholipase A2 (PLA2) activity of isolated plasma membranes, influence of elicitor and effectors of Gα. Plasma membrane vesicle suspensions were assayed for PLA2 activity by monitoring hydrolysis of the substrate bis-BODIPY® FL C11 PC (BPC) via increase in fluorescence. Substrate cleavage was validated in selected samples by high-performance thin layer chromatography (see Experimental procedures). One microgram per milliliter of yeast elicitor and/or GTPγS, as indicated, was added 10 min before the PLA2 substrate. Prolongation to 30 min did not significantly change the results. Concentrations of 0.3–10 μM GTPγS proved to be stimulating, with a maximum at 3 μm. Yeast elicitor is a glycoprotein fraction from bakers’ yeast of 30–42 kDa and a mannose content of 40%. Each column displays a ratio of PLA2 activities in the presence of the indicated effectors versus the non-stimulated activity of PLA2 (dashed line) which is about 5 nkat mg−1 membrane protein. Data are means of four experiments with SD.

Download figure to PowerPoint

In order to prove a functional coupling of both proteins in the plasma membrane, two sets of experiments were undertaken: (i) co-immunoprecipitation with the above-characterized anti-Gα antibodies, and (ii) non-denaturing PAGE to test for membrane-bound complexes of both proteins. The activation of PLA2 by GTPγS was used to report a conformation-dependent switching via Gα.

The Gα protein and PLA2 co-precipitate after solubilization of the plasma membrane

About 85% of the PLA2 activity measurable in the isolated plasma membrane could be solubilized with 0.5% sodium cholate in a carefully optimized procedure. Using the solubilized protein fraction, immunoprecipitates like those of Figures 1(b) and 3 were obtained with either of the aforementioned anti-Gα antisera. They contained similar amounts of the Gα protein as seen by cross-detection with western blots of each other serum. As an anti-PLA2 antibody of sufficient specificity for the plasma membrane-bound enzyme is not available (an anti-patatin antibody, provided by Dr T. Heitz, Université Louis Pasteur, Strasbourg, France, detected several proteins with and without PLA2 activity and precipitated neither significant PLA2 activity nor Gα), the presence of this enzyme in the immunoprecipitates was assayed by its activity (see Experimental procedures).

image

Figure 3.  Gα content and phospholipase A2 (PLA2) activity of the solubilized plasma membrane and different immunoprecipitates. Top: Western blot with W6 antiserum following SDS-PAGE, size markers in kDa. Bottom: high-performance thin layer chromatography separation of hydrolysis products following incubation of the same samples with the fluorescent PLA2 substrate BODIPY-1-O-alkyl-2-acyl-sn-glycerophosphocholine (BEPC, see Experimental procedures). The selection shown here includes the substrate (upper band) and the produced lysophospholipid BELPC (lower band). PM, solubilized plasma membrane proteins. Lanes 1–3, immunoprecipitates obtained with the following immobilized antisera: lane 1, anti-V-ATPase (negative control, cf. text); lane 2, anti-Gα, W6; lane 3, anti-Gα, P33; lane 4, supernatant remaining after immunoprecipitation as in lane 3.

Download figure to PowerPoint

The immunoprecipitate obtained with antiserum P33 clearly contained PLA2 activity, whereas the proteins precipitated with the W6 antiserum did not (Figure 3, lanes 2 and 3). No co-precipitation of PLA2 activity occurred with several antisera that detected no Gα but did detect other proteins of the plasma membrane, e.g. V-ATPase subunits (Figure 3, lane 1; the antibody detects V-ATPase (EC 3, 6, 1, 34) proteins that were shown to occur in both tonoplast and plasma membrane; Robinson et al., 1996), or of the cytoplasm, e.g. glucose-6-phosphate dehydrogenase (data not shown). Thus, unspecific adsorption of PLA2 to the immunomatrix was not likely, and the selective presence of PLA2 in the P33 precipitate suggested different binding properties of this antibody compared with W6.

The different epitope selectivities of the anti-Gα antibodies used were corroborated if the immunoprecipitation was performed after CTX-catalyzed ADP-ribosylation of the plasma membrane preparation. As shown in Figure 1(b) (lanes 2 and 4), this treatment significantly reduced the amount of Gα bound by antiserum P33 but not by antiserum W6. Thus, W6 and P33 antibodies bind Gα at distinct, non-overlapping epitopes and only the latter include the ADP-ribosylation site. This is again consistent with the different antigens used for raising these antisera: the C-terminus in case of W6 (White et al., 1993) and the whole (recombinant) Gα protein in case of P33.

A genuine three-dimensional structure of a plant Gα is not yet available, but homology models based on the existing sequence data indicate a high coincidence with the consensus structure of animal Gα proteins (Ullah et al., 2003). In the latter, the effector coupling site is close to the C-terminus (Millner, 2001; Ullah et al., 2003) and we therefore assume that Eschscholzia Gα interacts with PLA2 via this domain. This is expected to prevent the binding of W6 antibodies but still allow the P33 antibodies – which display broader selectivities – to find sufficient binding domains to a Gα–PLA2 complex.

Co-precipitation with Gα captured 50–60% of the total PLA2 activity of the solubilized plasma membrane (Figure 3, lanes 3 and 4) as confirmed with saturating concentrations of the P33 antiserum. In the remaining supernatant (containing 40–50% of PLA2) Gα was no longer detectable with either antibody. Thus, among the solubilized plasma membrane proteins, PLA2 molecules tightly bound to Gα coexist with others that are not coupled to this protein.

A functional Gα–PLA2 complex can be enriched in an immunoprecipitate

The ability to separate a Gα-bound from a non-Gα-bound form of PLA2 opened a way to test whether the contact with Gα provided the GTPγS-dependent control of enzyme activity. The effect of this nucleotide was compared between the isolated, non-solubilized plasma membrane, PLA2-containing immunoprecipitates and the remaining, solubilized plasma membrane preparation, which was ‘purged’ from Gα by saturating immunoprecipitation. As seen in Figure 4 (top), the PLA2 activity of immunoprecipitates was stimulated several-fold upon addition of this nucleotide, whereas almost no stimulation was seen with the PLA2 of the Gα-free supernatant (Figure 4, right). The different behavior of PLA2 in the presence and absence of Gα suggested that the stimulation by GTPγS was mediated via Gα and did not reflect a direct influence of the nucleotide on the PLA2 enzyme.

image

Figure 4.  Stimulation of phospholipase A2 (PLA2) activity by GTPγS. Each column gives a ratio of PLA2 activities in the presence versus absence of 10 μm GTPγS. Experiments were performed with the following: IP, immunoprecipitates obtained with antiserum P33; PM, isolated, non solubilized plasma membrane; no Gα, Gα-free supernatant of solubilized PM proteins obtained by saturating immunoprecipitate with P33; TG11, plasma membrane and precipitates from the antisense Gα transformant TG11. The Gα content was 11–14% of wild type, confirming estimates by Viehweger et al. (2006). Upper panel: solubilization and incubation at pH 7.5. Lower panel: solubilization and incubation at pH 9.5. PLA2 activity was determined as indicated in Figure 2 and Experimental procedures. Data are means of five experiments and the bar indicates the minimum and maximum values.

Download figure to PowerPoint

The higher degree of stimulation by GTPγS in the immunoprecipitate (almost fourfold) compared with the whole plasma membrane (about twofold) is consistent with the finding that only a distinct proportion of the enzyme activity of the solubilized plasma membrane is coupled to Gα; this part, as expected, is enriched in the immunoprecipitate. The stimulation of PLA2 was lost if GTPγS was replaced by GDPβS, both in the immunoprecipitate and in the plasma membrane.

At this point, our data indicated a stable, detergent-resistant complex in the plasma membrane that contains Gα and PLA2 and allows a stimulation of the enzyme via the activation of Gα.

The availability of recombinant cell strains that displayed a reduced content of Gα due to antisense suppression (11–24%; Viehweger et al., 2006) allowed a further proof of the role of Gα in the suggested complex. In plasma membranes obtained from the transgenic cells, PLA2 activity was not diminished, but was not stimulated by GTP or GTPγS. Precipitates with the P33 anti-Gα antiserum were still obtained (although in lower amounts) but likewise did not show significant activation of PLA2 (Figure 4). Thus, plasma membranes and co-precipitates with a reduced ratio of Gα to PLA2 lack the stimulatory effect of GTPγS. Even if some unspecific precipitation of PLA2 might have contributed to the immunoprecipitates, this finding clearly indicates that Gα is responsible for the nucleotide stimulation of PLA2.

Gα- and PLA2-containing complexes and the functional coupling of both proteins are released at higher pH

Several types of experiments indicate that the complexing of PLA2 and Gα can be released by an alkaline pH shift.

First, the stimulating effect of GTPγS and PLA2 in immunoprecipitates and non-solubilized plasma membranes disappeared if the pH was adjusted to a more alkaline level (e.g. pH 9.5 instead of the usual pH 7.5; Figure 4, bottom) This experiment was possible as the alkaline pH shift changed neither the specific activity of solubilized PLA2 nor the co-precipitation of Gα and PLA2 by the antiserum P33. Readjusting the pH to 7.5 recovered the stimulation (data not shown). The Gα-mediated control of PLA2 thus appears to be relieved at an alkaline pH.

Second, none of the PLA2 in the solubilized plasma membrane proteins was bound to immobilized Gα protein at pH 7.5. However, if solubilization and binding were performed at pH 9.5, a substantial amount of PLA2 was captured at a Ni–Sepharose matrix pre-loaded with recombinant, His-tagged Gα from E. californica (Figure 5a). This binding of PLA2 to an excess of immobilized Gα at pH 9.5 appeared specific (at least in part) as it was enhanced by pre-treating the Gα-loaded matrix with the activator GTPγS but not with the inhibitor GDPβS or with ATP (Figure 5b). Once bound, PLA2 was not eluted from the Gα-loaded matrix during extensive washing at pH 7.5, indicating a high binding strength between both proteins at the pH of the cytoplasm. Hence, the lack of capture at pH 7.5 did not reflect a pH dependence of binding but rather the lack at this pH of free PLA2 molecules with binding affinities for Gα among the solubilized proteins. At alkaline pH, PLA2 is most probably released from the complex with Gα and is thus able to rebind to an excess of immobilized Gα. The results further suggest that the binding strength of the active (GTP bound) form of Gα is higher than that of its inactive (GDP bound) form.

image

Figure 5.  Binding of phospholipase A2 (PLA2) solubilized at different pHs to immobilized Gα. One-hundred-microliter Ni-sepharose columns were loaded with recombinant Eschscholzia Gα protein, C-His tagged (about 2.6 μg, determined by BCA protein assay) and subsequently blocked with BSA. Each column contained the same amount of Gα and received the same amount of plasma membrane proteins that had been solubilized at pH 7.5 or 9.5. After incubation overnight at 8°C and the respective pH, followed by extensive washing with phosphate buffer of pH 7.5, the loaded matrix was incubated for 1 h with the PLA2 substrate bis-BODIPY® FL C11 PC (BPC), extracted and assayed for products of PLA2 activity (see Experimental procedures for details). (a) High-performance thin layer chromatography separation of the substrate (upper bands) and the hydrolysis product BLPC (lower band). Left, solubilization and incubation at pH 7.5. Right, solubilization and incubation at pH 9.5. Lanes 1 and 3: columns pre-loaded with Gα; lanes 2 and 4: no Gα loaded. (b) Phospholipase A2 activity captured by the Gα-loaded matrix. Prior to the incubation with solubilized plasma membrane at pH 9.5, the immobilized Gα was treated with the indicated nucleotides (10 μm, 2 h). Data are normalized to the PLA2 activity measured after pre-treatment with nucleotide-free buffer, which is set to 1.0. Hatched columns show the PLA2 activity of the supernatant, i.e. of the non-Gα-bound fraction, assayed under similar conditions. A typical experiment is shown, that was repeated twice with similar outcomes.

Download figure to PowerPoint

Finally, protein complexes of the plasma membrane that contain both Gα and PLA2 could be visualized after non-denaturing gel electrophoresis and showed pH-dependent differences in size. At pH 7.5, the immunologically detected Gα protein almost exclusively exists in bands of high molecular mass, i.e. >120 kDa (Figure 6, top). Phospholipase A2 mainly co-migrated with the Gα-containing bands, indicating the aggregation of both proteins in complexes of unknown structure. The presence of GTPγS did not influence the size of these complexes. In contrast to these findings, gels run at pH 9.5 displayed a significant part of PLA2 that migrated to low-molecular-weight (MW) regions of around 40 and 15 kDa, where no Gα was detectable – and surely is not present, as the MW of Gα is 44.5 kDa (Figure 6, bottom). Thus, at pH 9.5 a much larger part of PLA2 exists as low-MW species that are not associated with Gα. The protein amount of these enzymes in the native gel was below the detection limit, but their activity was clearly measurable. Seen together, stable complexes of plasma membrane proteins detectable at pH 7.5 dissociate at higher pH, but not upon activation of Gα via GTPγS.

image

Figure 6.  Non-denaturing PAGE of Gα- and phospholipase A2 (PLA2)-containing plasma membrane protein complexes. Cholate-solubilized plasma membrane vesicles (each containing 10 μg of total protein) were separated on polyacrylamide gels in the presence of 0.5% sodium cholate at pH 7.5 (upper panel) or pH 9.5 (lower panel) as described in Experimental procedures. Gels were run in duplicate and subsequently either electroblotted followed by immunodetection of Gα, or cut in pieces and screened for PLA2 activity. Lane a, MW standards; lanes b and d, solubilized proteins; lanes c and e, solubilized proteins, plus 10 μm GTPγS. Detection: lanes a–c, PageBlue™ protein stain; lanes d, e, western blotting with anti Gα antibody P33. Columns (right): PLA2 activity, given as fluorescence increase within a 2-min incubation of gel slices with the PLA2 substrate bis-BODIPY® FL C11 PC (BPC).

Download figure to PowerPoint

Conclusions

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Conclusions
  6. Experimental procedures
  7. Acknowledgements
  8. References

The present findings afford novel biochemical evidence in plant plasma membranes of stable protein complexes that contain Gα together with PLA2 and provide the GTP activation of this enzyme. Due to the high specific activity and stability of the plasma membrane-bound PLA2 (unimpaired by cholate solubilization, immunoprecipitation and PAGE) its interaction with Gα could be followed at the activity level, even before data for the isolated protein were available (see below).

After solubilization, the stimulation by GTPγS is maintained in a PLA2-containing protein complex that co-precipitates with Gα. The complexing of both proteins reflects a high binding strength as it is not released by a detergent (0.5% sodium cholate) that solubilized most plasma membrane proteins. The selectivity of interaction is indicated by the fact that: (i) antibodies raised against the effector coupling site precipitate Gα protein that is not associated with PLA2, and (ii) the non-precipitated, Gα-free fraction of PLA2 is not stimulated by GTPγS. For these reasons, we conclude that the complex of Gα and PLA2 enriched in the P33 immunoprecipitate upholds the site of interaction of both proteins in vivo and transmits the GTP-triggered conformational change of Gα into the stimulation of PLA2.

Further work is required, and it is planned in our laboratory to prove whether more proteins (with stabilizing and/or facilitating functions), and especially the Gβ/γ subunit, are included and required for the functional complex. Our preliminary data show that the Gβ protein is contained in the P33 immunocomplex (MH and WR, unpublished data).

The functional coupling between Gα and PLA2 is released at alkaline pH (Figure 4). Under the same conditions, the electrophoretic co-migration of both components is lost (Figure 6) and PLA2 with affinity for Gα appears among the solubilized membrane proteins (detected by its capture with an excess of immobilized Gα, Figure 5). These findings all point to a pH-dependent component in the PLA2/Gα-containing protein complex. Although its molecular structure is not known, the pH-dependent stability will be helpful in its investigation. (It is not intended, and would be premature, to speculate about any physiological relevance of the pH-dependent stability of these protein associations.)

There is little doubt that the plasma membrane of Eschscholzia contains PLA2 in both Gα-bound and Gα-free forms. If solubilized at the pH of the cytoplasm (7.5, confirmed by pH topography; Roos et al., 1998) about half of the plasma membrane’s PLA2 activity neither co-precipitates with Gα (Figure 3, lane 4) nor binds to immobilized Gα (Figure 5a, pH 7.5), which is suggestive of one or more PLA2 isoforms that lack affinity to Gα. This property explains the lack of stimulation by GTP of this PLA2 in the Gα-free solubilizate.

The coexistence of Gα-controlled and Gα-independent PLA2 species may now explain earlier data from signal transduction studies (summarized in Roos et al., 2006). The antisense Gα transformants used in the present experiments – as well as cell lines expressing an anti-Gα scFv antibody – are only deficient in the activation of PLA2 and downstream responses that are triggered by low elicitor concentrations (Viehweger et al., 2006). High elicitor concentrations activate PLA2 and induce alkaloid biosynthesis and the hypersensitive response in a Gα-independent manner. This includes overproduction of jasmonate (Färber et al., 2003), a well-known inducer of secondary biosynthetic enzymes (Haider et al., 2000). Therefore, the coexistence of Gα-controlled and Gα-free PLA2 species in the plasma membrane probably marks the molecular origin of branches in downstream signaling that use different transduction mechanisms and converge at the induction of phytoalexin biosynthesis.

Further progress in understanding the detailed mechanisms that control the activity of the plasma membrane-bound PLA2 requires structural data of the enzyme protein, which, to our knowledge has not yet been cloned and sequenced or shown to be identical with a known phospholipase. As the very low content in the membrane complicates the biochemical detection of this highly active enzyme (see Figure 6) recombinant and knock-down strategies are to be considered. From other plants (mainly Arabidopsis), we have sequence data for the sPLA2 family (secretory, low MW, Ca2+-dependent enzymes) and the patatin-like PLAs (Ca2+-independent enzymes, cytosolic or membrane bound, combined A1 and A2 specificity) (Ryu, 2004). The plasma membrane-bound enzyme of Eschscholzia investigated in this study is Ca2+ independent and shows no A1 activity (Viehweger et al., 2002).

Functional complexes of Gα with target proteins have been characterized only in a few cases. Phospholipase D1 (PLD) activity is inhibited by Gα added to the enzyme assay (Lein and Saalbach, 2001). Zhao and Wang (2004) demonstrated the site-specific interaction between this enzyme (isoform α1) and a recombinant Gα–GST fusion protein via co-precipitation experiments. Binding required the DRY motif in the PLD protein and was released by GTP. While these studies made use of recombinant proteins, our approach shows the interaction of native Gα and PLA2 under the genuine conditions of the Eschscholzia plasma membrane.

Another recent example is the interaction of the Arabidopsis Gα-protein (GPA1) with a specific prephenate dehydratase (PD1) as discovered in a yeast two-hybrid screen by Warpeha et al. (2006). This enzyme functions in a blue light-triggered signal chain that induces the biosynthesis of phenylalanine. Similar to our findings, GTPγS increased the activity of the target enzyme PD1.

Looking over the present examples, it appears that pre-formed complexes of the same Gα with different target proteins are used as signal modules to serve different downstream paths. Due to their shared Gα component, these modules make use of the conserved GTP-triggered activation mechanism that would gain selectivity due to different binding affinities towards GTP and the local ratio of GTP/GDP.

It is still a matter of debate whether the Gβγ subunits are involved in the interaction of Gα with distinct target proteins. Recently, FRET studies with Arabidopsis plants expressing constitutively active Gα suggested that the heterotrimeric complex in plants might not dissociate upon activation by GTP (Adjobo-Hermans et al., 2006). In contrast, Kato et al. (2004) found the Gα of rice plasma membranes incorporated into very large protein complexes (about 400 kDa), that contained β and γ subunits together with unknown proteins. This study supports the tendency of inactive Gα to associate with other plasma membrane proteins from which Gα is liberated upon activation by GTPγS. Our electrophoretic studies do not indicate a GTPγS-triggered dissociation of the Gα/PLA2 complexes, but rather support their pH-dependent dissociation. Whatever the solution to this question will be, pre-formed complexes of the heterotrimeric protein(s) with target and/or modulating proteins now appear to be plant-specific signal modules that deserve, and will surely attract, increasing research interests in signal biology.

Experimental procedures

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Conclusions
  6. Experimental procedures
  7. Acknowledgements
  8. References

Plant cell culture and plasma membrane preparation

Suspension cultures of E. californica were grown as previously described (Roos et al., 1999; Viehweger et al., 2002). After 7 days of growth, 200 g (fresh weight) of cells were harvested, stored and processed as outlined in Roos et al. (1999). From the resulting microsomal fraction, plasma membrane vesicles were prepared by sucrose density gradient centrifugation according to Sandelius and Morre (1990), subsequently pelleted at 100 000g for 40 min at 4°C, finally resuspended in 1 ml 25 mm 2-(N-morpholine)-ethanesulfonic acid (MES)/bis-tris-propane, pH 7.0, 0.1 mm EDTA, 10% glycerol, and stored at −80°C until further use. Marker enzyme activities, assayed as described in Pönitz and Roos (1994), indicated that the obtained fractions contained more than 80% plasma membrane vesicles.

Solubilization of plasma membrane-associated proteins

Plasma membrane fractions were treated by 3 × freeze/thaw (−80°C/4°C) prior to solubilization and then collected at 100 000g for 40 min at 4°C and resuspended in an equal volume of TE buffer [20 mm 2-amino-2-(hydroxymethyl)-1,3-propanediol (TRIS)–HCl, 1 mm EDTA, pH 7.5, or pH 9.5 if indicated], supplemented with 50 mm NaCl and 0.5% sodium cholate. Proteins were solubilized for 2 h at 8°C with gentle stirring and centrifuged at 100 000 g for 40 min at 4°C. The supernatant, which usually contained 500 μg protein ml−1 and about 80% of the total PLA2 activity of the plasma membrane, is further referred to as solubilized plasma membrane.

Immunoprecipitation procedure

The following antibodies were used: (i) antisera against Arabidopsis Gα (kindly provided by P. Millner, University of Leeds, UK): W6, raised against a peptide corresponding to the C-terminus of Gα (White et al., 1993); P33, raised against whole recombinant Gα; (ii) antiserum against cytosolic subunits of V-ATPase of Kalanchoe daigremontiana (Robinson et al., 1996; provided by R. Ratajczak, Darmstadt, Germany) as a control.

Preparation of the immunomatrix.  Protein A–Sepharose 4 Fast Flow beads (GE Healthcare, http://www.gehealthcare.com) were prepared according to the manufacturer’s protocol, washed three times and finally resuspended in an equal volume of PBS buffer pH 7.5. To 100 μl Sepharose slurry, 200 μl of solubilized plasma membrane vesicles and 200 μl of PBS were added. After adding the antiserum (final dilution 1:50 to 1:500), the sample was incubated at 700 rpm and 8°C for 60 min. The beads were washed with excess PBS until substantial amounts of neither total serum protein nor IgG (assayed by the Bradford method or western blot, respectively, see below) were detectable in the supernatant.

Immunoprecipitation.  The immunomatrix was incubated with the solubilized plasma membrane proteins (2 μg μl−1 bead suspension) overnight at 8°C with gentle shaking (5000 g). The beads were separated by brief centrifugation and washed five times in each five volumes of PBS buffer. An aliquot of the immunoprecipitates was used for assaying PLA2 activity as described below. For SDS-PAGE and subsequent western blotting for immunological detection of the precipitated material, the washed beads were treated with two volumes of SDS-PAGE sample buffer including 1.5% SDS (without dye) for 5 min at 90°C. The eluted proteins were concentrated by ultrafiltration using Vivaspin concentrators (cut-off 5 kDa; Vivascience, http://www.vivascience.com). For preparation of sufficient quantities of the putative Gα (e.g. for sequencing, see below), the eluted material obtained from several precipitations was pooled.

After elution of the bound plasma membrane proteins, the immunomatrix was washed and stored as described above.

Identification of the Gα protein

Mass spectrometry of tryptic fragments.  After SDS-PAGE of the immunoprecipitate, the gel region corresponding to the anti-Gα (W6) immunoblot signal (45 kDa) was excised, washed and digested with trypsin (Promega, http://www.promega.com/) in 30 μl of 50 mm ammonium bicarbonate, pH 8.0, overnight at 37°C (Shevchenko et al., 1996). The tryptic peptides were extracted from the gel pieces and analyzed by liquid chromatography-tandem mass spectrometry (LC/MS/MS) using a Cap-LC column, 75 μm internal diameter ×15 cm (Micromass, http://www.micromass.com) and 3 μm C18 100 μm□ PepMap®(Dionex, http://www.dionexsoftron.de) coupled to a quadrupole time-of-flight mass spectrometer (Q-TOF2; Micromass).

Uninterpreted peptide fragment ion spectra generated by LC/MS/MS were searched against the non-redundant protein data base for exact matches using the MASCOT search engine (http://www.matrixscience.com). Multiple sequence alignment with plant Gα sequences obtained from SWISS-PROT was performed using ClustalW (http://www.ebi.ac.uk/clustalw).

Identification of the E. californica Gα gene.  An internal cDNA fragment (847 bp) of EcGα was generated by RT-PCR using total RNA isolated from cell cultures of E. californica with the degenerated oligonucleotides GGIAAA(G)A(T)G(C)IACIATA(C/T)TTC(T)AAA(G)CAA(G)AT and CA(G)AAC(T)TTC(T)TTC(T)TTIACA(G)AAC(T)TCA(G)TAIGC (I = inosin). This fragment was ligated into the pCR 2.1 vector, followed by insertion into a plant transformation vector (BinAr, with the CaMV 35S promoter) using the BamHI and SalI sites. This vector, which was also successfully used for the antisense-Gα transformation (Viehweger et al., 2006) was digested with EcoRI and SalI. The resulting 300-bp fragment (594–894) showed high homology to a central region of several plant Gα genes. It was random-labeled by d-32P-CTP and used to probe a cDNA library that had been prepared from mRNA of the Eschscholzia cell culture by Gateway technology (CloneMiner® cDNA Library Construction Kit, Invitrogen, http://www.invitrogen.com/). This library contained 0.75 colony-forming units (cfu) ml−1 and an average insert size of 1350 bp. Hybridization screening identified a clone with the complete gene sequence inserted. It was amplified by PCR with stringent primers and completely sequenced. In parallel, the PCR product was ligated into the expression vector pET23a that served the heterologous expression of the C-His-tagged protein in Escherichia coli, BL21. The recombinant protein was purified at Ni-nitrilotriacetic (NTA) agarose and characterized by its specific binding to GTP-γ35S, fixed at nitrocellulose membranes.

ADP-ribosylation.  The CTX-catalyzed ADP-ribosylation of Gα was assayed essentially as described by Gill and Woolkalis (1991). A standard reaction mixture (20 μl) contained up to 30 μg of plasma membrane or 1.5 μg of immunoprecipitated total protein, 0.4 μg activated CTX (Calbiochem, http://www.emdbiosciences.com/), 0.5 μm [32P]NAD (30 nCi, New England Nuclear; http://www.las.perkinelmer.com), 0.1 mm GTPγS, and 10 mm thymidin. After incubation at 25°C for 30 min with gentle shaking the mixtures were separated by SDS-PAGE (see below). The gel was stained with Coomassie Blue, vacuum dried, and radioactivity was monitored by phosphor imaging (Storage Phosphor System, Canberra-Packard, http://www.canberra.com). Detected spots were semi-quantified by densitometry.

Electrophoresis

Non-denaturing electrophoresis was performed according to Schaegger and von Jagow (1991) without the Coomassie dye (which interfered with the immunostaining).

To 40 μl of solubilized plasma membrane, 10 μl of gel loading buffer was added to yield final concentrations of 4% glycerol, 4 mm dithioerythritol, 0.5% sodium cholate and 100 mmε-aminocaproic acid in 20 mm bis-tris-propane. The pH was adjusted to pH 7.5 or 9.5 as indicated. Samples of 25 μl were applied to a stacking gel made of 4% acrylamide, 0.1 m TRIS and 0.5% cholate and separated on a gradient gel (5–15% acrylamide) with 0.5% sodium cholate and 0.4 m TRIS/HCl. The cathode buffer contained 50 mm Tricine and 50 mm bis-tris-propane and the anode buffer 50 mm bis-tris-propane. The pH of all gel solutions was adjusted to either 7.5 or 9.5. Electrophoresis was performed with a MiniVE vertical electrophoresis system (gel size 90 mm × 100 mm × 1.0 mm; Amersham Biosciences, http://www5.amershambiosciences.com/) at 8°C, 110 V const. for 1 h, 130 V const. for 2 h and finally at 150 V const.

The SDS-PAGE procedure was according to Laemmli (1970).

Western blotting was done as described in Roos et al. (1999) with the Gα antisera W6 or P33 (usually diluted 1:300).

Phospholipase A2activity assays

Phospholipase A2 activity in plasma membrane vesicle suspensions, cholate-solubilized supernatants or in loaded immunomatrices was determined according to Viehweger et al. (2002) using fluorescent artificial phospholipids (from Molecular Probes, http://www.probes.invitrogen.com), either bis-BODIPY® FL C11 PC (BPC) or BODIPY-1-O-alkyl-2-acyl-sn-glycerophosphocholine (BEPC, ether bond at position 1) as indicated. A 230-μl reaction mixture contained 50 μl of enzyme test sample (loaded immunomatrix, plasma membrane preparation or supernatant after immunoprecipitation) in 400 nm substrate, 0.02% 3[(3-cholamidopropyl)dimethylammonio]-propanesulfonic acid (CHAPS) and 20 mm TRIS/HCl, pH 7.5.

Slices of cholate-PAGE gels (each 20 mm2, three parallels) were extracted by freezing at −20°C (2 h) and thawing (gentle shaking at 8°C overnight) in 200 μl of the same buffer, centrifuged (1000g, 2 min at 4°C) and 50 μl of supernatant used as above.

Samples with low PLA2 activity were incubated for 30 min to 2 h by gentle shaking at 30°C, extracted with methanol/chloroform and the phospholipids and hydrolysis products were separated by high-performance thin layer chromatography (HPTLC) and quantified densitometrically as detailed in Viehweger et al. (2002). In order to exclude non-enzymatic cleavage as a reason for fluorescent product formation, heat-denatured controls (15 min, 90°C) were included in any of the measurements.

For quantification of high PLA2 activity, the increase in fluorescence resulting from cleavage of BPC was monitored in a FLX 800 fluorescence microplate reader (Bio-Tek Instruments, http://www.biotek.com) over 10 min and the initial rates (estimated either at 1 min of reaction or by non-linear curve fitting) were taken as a measure of enzyme activity. The increase in fluorescence reflects the extent of substrate hydrolysis as shown earlier (Roos et al., 1999).

Effects of GTPγS on PLA2 activity

Two hundred microliters of solubilized plasma membrane vesicles or immunoprecipitate was diluted with 250 μl solubilization buffer and divided in two aliquots, one of them received 25 μl of 100 μm GTPγS (Sigma-Aldrich, http://www.sigmaaldrich.com/) and PLA2 activity was determined as described.

Binding of PLA2 to immobilized Gα

Recombinant, His-tagged Gα (3.5 μg) in 50 μl washing buffer (20 mm sodium phosphate, 500 mm NaCl and 20 mm imidazole, pH 7.5) were loaded onto a Illustra® MicroSpin column containing 100 μl of Ni-Sepharose (High Performance, GE Healthcare; http://www.gelifesciences.com). After 5 min incubation with shaking at room temperature, columns were washed twice with 600 μl washing buffer. The second wash contained 3% BSA to block unspecific binding sites. Each step was terminated by brief centrifugation at 430 g.

The loaded columns were incubated overnight at 8°C with each 400 μl plasma membrane preparation (100 μg protein per column) in solubilization buffer at a pH of either 7.5 or 9.5 (see above) . After washing five times with washing buffer (pH 7.5) and brief centrifugation at 430 g, 80 nmol BPC (PLA2 substrate) in 80 μl of the same buffer was added. After incubation with shaking at room temperature for 2 h, the phospholipids were extracted and the products of PLA2 activity separated by HPTLC as described.

The protein content of the detergent-free samples was estimated using Coomassie Blue G-250 according to Bradford (1976) and of the solubilized samples by the bicinchoninic acid (BCA) method as described by Smith et al. (1985).

Acknowledgements

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Conclusions
  6. Experimental procedures
  7. Acknowledgements
  8. References

We wish to thank Dr P. Millner, University of Leeds, for providing us with two anti-Gα antisera, Dr Heitz, Strasbourg, France, for providing us with anti-patatin-antibodies and Dr A. Schierhorn, Max Planck Arbeitsgruppe ‘Enzymatik der Proteinfaltung’, Halle, for the MS-based peptide sequencing.

References

  1. Top of page
  2. Summary
  3. Introduction
  4. Results and discussion
  5. Conclusions
  6. Experimental procedures
  7. Acknowledgements
  8. References