The use of membrane translocating peptides to identify sites of interaction between the C5a receptor and downstream effector proteins

Abstract

The complement fragment C5a is a potent leucocyte chemoattractant and activator, mediating its effects through a G-protein-coupled receptor. Whilst the C-terminal domain of this receptor has been shown to be essential for receptor desensitization and internalization, it is not known which domains couple to the receptor's heterotrimeric G proteins. In this report we have used a membrane translocating sequence (MTS) to examine the effects of the four intracellular domains of the human C5a receptor (C5aR) on the receptor's signalling via G_αi family heterotrimeric G proteins in intact RBL-2H3 cells. The results indicate that all of the intracellular domains couple to downstream signalling, with the proximal region of the C terminus being a major binding site and intracellular loop 3 playing a role in G protein activation or receptor desensitization.

Introduction

Recruitment of phagocytic leucocytes to sites of infection and injury is vital to host defence. The ligands involved in this process include bacterial formyl peptides, the complement fragment C5a, leukotrienes and the large family of chemokines.^1,2

These ligands activate a family of G-protein-coupled receptors (GPCRs), which act as molecular switches initially relaying the presence of extracellular ligands to their associated heterotrimeric (αβγ) G proteins in the cytoplasm. Upon activation, the GPCR catalyses the exchange of guanosine triphosphate (GTP) for guanosine diphosphate (GDP) on the α subunit, causing dissociation of the α-GTP subunit and the βγ dimer from the receptor. The dissociated α-GTP and βγ then subsequently activate a cascade of effector molecules leading to a variety of functional responses.^3,4

Chemoattractant receptors are able to couple to a variety of different G-α subunits in vitro. However, in phagocytes the coupling is generally pertussis-toxin-sensitive,^5,6 indicating that this process is largely mediated via the G_αi family, comprising G_αi2 and G_αi3 in these cells. Despite this there is little sequence conservation in the cytoplasmic domains of the receptors and no known consensus sequences for G protein binding.

Like all GPCR, the receptor for the chemotactic complement fragment C5a (C5aR) has seven transmembrane helices with four intracellular and four extracellular domains (Fig. 1), assumed to associate into the barrel-like structure of rhodopsin.⁷ The C5aR has been characterized in some detail. The extracellular domains create a definitive ligand-binding pocket^8,9 transducing information via critical residues in the transmembrane helices to the intracellular domains^10,11 which ultimately interact with the G-proteins. However, apart from the C-terminus, which appears to play a role in receptor internalization and recycling,^12,13 little is known about the role of the different intracellular domains in receptor coupling.

Figure 1.

Schematic representation of the C5a receptor showing the intracellular domains (ICL1, ICL2, ICL3 and C-terminus) and the serpentine nature of the receptor weaving through the plasma membrane.

One approach frequently taken to identify critical sites of interaction with effector molecules is to make mutations in the target receptor. While in many cases this has proved informative, such mutations may indirectly affect other regions of the receptor leading to loss of expression or inaccurate attribution of sites of molecular interaction. Indeed when we constructed a series of alanine triplet mutations for the intracellular domains of C5aR the majority failed to express (unpublished observations). In this report we have taken an alternative approach of introducing peptides corresponding to the four intracellular domains of C5aR into intact cells expressing the receptor and examining their effect on C5a-induced signalling via G_αi. This was achieved by exploiting the ability of a short membrane translocating sequence (MTS) of the hydrophobic region in the signal sequence of Kaposi fibroblast growth factor^14–16 to translocate peptides across the plasma membrane. C5aR-transfected RBL-2H3 cells were chosen as we have shown that coupling of C5aR to degranulation in these cells is fully pertussis-toxin-sensitive and is readily measured by loading the cells with ³H-labelled 5-hydroxytryptamine ([³H]5-HT).¹⁷

Materials and methods

Cells and media

LD10, a stable clone of the RBL-2H3 cell line transfected with wild-type C5a receptor, was produced as described previously¹⁷ and cultured in Dulbecco's modified Eagle's minimum essential medium, supplemented with 10% heat inactivated fetal calf serum.

Construction of glutathione S-transferase (GST)-MTS expression plasmids

The DNA sequence for the MTS protein bounded by relevant restriction sites was chemically synthesized (Fwd: 5′-gat ccc cat ggt acc cgg ggc agc cgt tct tct ccc tgt tct tct tgc cgc acc caa gct tct aga gat cg-3′ Rev: 5′-aat tcg atc tct aga agc ttg ggt gcg gca aga aga aca ggg aga aga acg gct gcc ccg ggt acc atg gg-3′). Using standard molecular biological techniques the two MTS primers where then annealed and inserted into pGEX-3X (Amersham Pharmacia Biotech UK Ltd., Little Chalfont, UK) cut with BamHI and EcoRI (Fig. 2).

Figure 2.

Diagram detailing the insertion of the membrane translocating sequence (MTS) and associated restriction sites into pGEX-3X. The DNA fragments corresponding to the ICL1, ICL2, ICL3 and CT were inserted at the point represented by an asterisk.

The intracellular domains of the C5aR were amplified by polymerase chain reaction with relevant primers; ICL1 (Fwd: 5′-agc ttg ccg cca cca tgg gaa cgg cat tcg aag cca agc gga cca tca atc ccg ggt g-3′ Rev: 5′-aat tca ccc ggg att gat ggt ccg ctt ggc ctc gaa tgc cgt tcc cat ggt ggc ggc a-3′), ICL2 (Fwd: 5′-gca gca gca aag ctt gcc gcc acc atg gga gac cgc ttt ctg ctg gtg t-3′ Rev 5′-gca gaa ttc acc cgg ggg ccc ctc gga agt tct g-3′), ICL3 (Fwd 5′-gca gca gca aagctt gcc gcc acc atg gga cgg acg tgg agc cgc ag-3′ Rev 5′-gca gaa ttc acc cgg gct tga gtg tct tgg tgg acc-3′) and C-terminus (CT) (Fwd: 5′-gca gca gca aag ctt gcc gcc acc atg gga ggc ttc cag ggc gca ctg-3′ Rev: 5′-gca gaa ttc acc cgg gca ctg cct ggg tct tct gg-3′). The products of these reactions resulted in the relevant intracellular domain bounded by HindIII and EcoRI sites. These fragments were then ligated into the pGEX-3X-MTS plasmid (Fig. 2) and verified by automated DNA sequencing.

Bacterial growth and induction

Overnight cultures of DH5α transformed with the relevant pGEX-3X plasmid construct were used to inoculate fresh media (1 litre) to an OD₆₀₀ of 0·1. These were grown for 90 min at 37° with shaking, gene expression was then induced by the addition of isopropyl β-D-thiogalacto-pyranoside (IPTG) (0·5 mm). The bacteria were then grown for a further 4 hr prior to harvesting. Pellets were washed twice with 50 ml STE buffer (10 mm Tris–HCl, pH 7·5, 10 mm NaCl, 1 mm ethylene diaminetetraacetic acid pH 8·0).

GST purification

Thawed cells were resuspended in 25 ml STE buffer containing peptidase inhibitors (Complete™, Boehringer Mannheim UK, Bracknell, UK) and lysozyme (50 μg/ml, Sigma UK, Sigma-Aldrich Company Ltd., Gillingham, Dorset, UK) and incubated on ice for 30 min with mixing. Sarkosyl (0·2%) and dithiothreitol (100 mm) were then added and the mixture was vortexed for 2 min Disruption of the bacterial cells was achieved using a sonicating probe at 4°. Cell lysates were centrifuged (10 000 g for 15 min) in polypropylene tubes, the supernatant was transferred to a fresh tube and Triton X-100 (4%) was added. Pre-soaked glutathione–agarose resin (Sigma-Aldrich UK) was then added to the supernatant and gently mixed at room temperature for 15 min The supernatant agarose mix was passed though a column and the retained resin was washed with 3 × 5 ml of STE buffer. GST protein was eluted by the addition of 3 × 5 ml of 50 mm Tris–HCl, 5 mm glutathione, pH 7·5. Protein was then concentrated by ultra filtration (Centriprep, Amicon YM-10) and the final protein concentration was estimated by OD₂₈₀.

Factor Xa protein restriction

Confirmation that the purified GST-MTS-ID proteins had retained their receptor domains was achieved by the mobility shift observed on sodium dodecyl sulphate–polyacrylamide gel electrophoresos (SDS–PAGE) after treating with Factor Xa which cleaves between the GST and MTS sequences. Protein constructs (50 μg) were cleaved with Factor Xa (1 μg, New England Biolabs, Hitchin, UK) by incubation at room temperature for 24 hr in 20 mm Tris–HCl, 100 mm NaCl and 2 mm CaCl₂.

Western blots

Cells, treated according to the figure legends, were washed five times with cold phosphate-buffered saline (PBS) to remove extracellularly associated protein and lysed with 0·5% Triton X-100 lysis buffer containing protease and phosphatase inhibitors. Samples were separated by SDS–PAGE, transferred to nitrocellulose membranes, and probed with a rabbit anti-GST antibody (Sigma-Aldrich UK) followed by horseradish peroxidase-linked anti-rabbit secondary (DAKO Ltd., Ely, UK) and visualized by chemiluminescence using Supersignal West Dura kit (Perbio Science UK Ltd., Tattenhall, UK).

Fluorescence microscopy

LD10 cells, seeded the previous day at 20 000 per well, were incubated on chambered coverglass slides (Nunc 155411; Nunc, Paisley, UK) with 20 μm of the GST-MTS-ID proteins in growth media for 3 hr at 37° then washed with PBS containing 0·1% bovine serum albumin (BSA). Prior to staining, the cells were either incubated in PBS alone (non-permeabilized) or with 30 μg/ml of l-α-lysophosphatidylcholine (Sigma-Aldrich) (permeabilized) in PBS, at 4° for 20 min, followed by washing with PBS/1% BSA. The cells were then stained with monoclonal anti-glutathione-S-transferase (Sigma-Aldrich, G-1160) in PBS/20% rabbit serum for 2 hr at 4°. Cells were washed twice in PBS/0·1% BSA and the secondary antibody, fluorescein isothiocyanate (FITC)-labelled anti-mouse immunoglobulins (DAKO, F0313) was added for 1 hr at 4° in PBS/20% rabbit serum. Visualization was performed using a Molecular Dynamics 2010 confocal microscope.

Cell viability assay

LD10 cells, seeded the previous day onto 96-well tissue culture plates, were incubated with and without the different peptides as above (20 μm for 3 hr at 37°) and then viability was determined by colorimetric assay using WST-1 (Roche Diagnostics Ltd., Lewes, UK), according to the manufacturer's instructions.

Release assay

LD10 cells were suspended in growth medium at 4 × 10⁵ cells/ml containing 1·33 μl/ml of [³H]5-HT, 80 μl of cell suspension was added to each well of a 96-well tissue culture plate. Cells were incubated at 37° in a CO₂ incubator over night. After 16 hr the media was aspirated and replaced with 40 μl of pre-warmed fresh medium containing the experimental proteins as required and left for 3 hr at 37°. The wells were then washed with two 150 μl changes of prewarmed (37°) release buffer (balanced salt solution+1 mm CaCl₂) and a final 150 μl left in the wells which were then returned to the incubator for 10 min Wash solution was removed from the plate and 50 nm of C5a (Sigma-Aldrich) in release buffer (pre-warmed to 37°) was added to the wells and incubated for 30 min at 37° then placed on ice for 5 min Finally, 30 μl of release buffer was transferred from each well into a 96-well Lumaplate™ (Becton Dickinson UK Ltd., Oxford, UK) and dried in an oven overnight before being read in a scintillation counter. Background release was determined without ligand and maximum release by cell lysis with 30 μl of 2% Triton-X100 in release buffer. All samples were assayed in quintuplicate.

Receptor internalization assay

LD10 cells were incubated in growth media with or without the various GST-MTS-ID proteins as required (20 μm, 3 hr at 37°) and subsequently with or without the addition of C5a (50 nm, 30 min at 37°). Internalization of the C5a receptor was assessed by specific staining with a rabbit anti-C5aR serum (Serotec AHP353; Serotec Ltd., Oxford, UK) for 1 hr at 4°, and secondary FITC-labelled anti-rabbit IgG (Serotec, STAR34B) for 1 hr at 4°. Stained cells were then examined by flow cytometry (FACSort; Becton Dickinson UK Ltd., Oxford, UK), and assessed using cell quest software (Becton Dickinson).

Results

Protein expression and purification

Constructs containing the intracellular domains of the C5a receptor (ID) fused to the C terminus of the glutathione S-transferase-membrane translocating sequence (GST-MTS) were expressed and purified and the protein concentrations were estimated by OD₂₈₀. SDS–PAGE of the GST-MTS-ID constructs showed the eluted protein to be approximately 95% pure (Fig. 3 lane U). Factor Xa digestion of the purified proteins and subsequent SDS–PAGE revealed the release of fragments of the correct size, for example for the GST-MTS-ICL1 protein the initial 30 000 MW construct was reduced to 26 000 as the appropriate 4000 MW fragment was released (Fig. 3).

Figure 3.

Purified GST-MTS-ICL1 protein stained with Coomassie blue. Lane U, uncut and lane Xa, Factor Xa digested.

Cellular uptake of purified GST-MTS fusion proteins

The purified GST-MTS fusion protein (0–100 μm) was added to LD10 cells in growth media. After a period of incubation (0·5–120 min) cells were washed and harvested. Extracts from cells presented with protein were run on SDS–PAGE and specifically probed for GST by Western blotting. Entry of the GST-MTS into cells was proportional to the protein concentration (Fig. 4a) and the time of incubation with the protein (Fig. 4b). In the absence of the MTS sequence virtually no GST was detectable in the cell extract (Fig. 5). LD10 cells incubated with the GST-MTS-ID proteins (20 μm protein for 2 hr) also confirmed uptake of these constructs, albeit at a reduced level compared to GST-MTS alone (Fig. 5). Protein translocation inside the cell was confirmed by fluorescence microscopy. Neither non-permeabilized nor permeabilized cells incubated with GST alone showed detectable staining. In contrast all the constructs containing the MTS sequence showed a strong, diffuse intracellular staining pattern with the permeabilized cells, irrespective of the additional ID attached to the sequence, whereas staining was undetectable in non-permeabilized cells (data not shown).

Figure 4.

Western blots showing (a) the uptake of GST-MTS into LD10 cells with increasing concentration (0–100 μm) of added protein and (b) the uptake of 20 μm GST-MTS construct with time.

Figure 5.

Western blot showing uptake of the different GST protein constructs by LD10 cells (20 μm for 2 hr).

Effect of intracellular domain peptides of C5aR on release

Having shown that the GST-MTS protein constructs were translocated across the plasma membrane, their effect on C5a receptor signalling was examined. The unmodified GST-MTS control and the constructs associated with each intracellular domain were incubated with LD10 cells (2 hr), and then a 5-HT release assay was performed. The addition of the GST-MTS had no significant effect when compared to samples without protein (Fig. 6). Compared to the cells with the addition of GST-MTS, ICL1 and ICL3 both activated release in proportion to the concentration added, with ICL3 having almost twice the effect of ICL1 (Fig. 6). Using 20 μm of protein the release induced by the ICL1 and ICL3 constructs over the control was approximately 50% and 100%, respectively. ICL2 seemed to have little effect on activity throughout the concentration range examined compared to the GST-MTS control. The CT construct had little effect at concentrations under 10 μm, but over this level caused a dramatic drop in 5-HT release to one-third of that seen with the GST-MTS control, which remained constant up to the highest concentrations of protein used (Fig. 6). Addition of the GST-MTS-IDs induced no detectable release in the absence of C5a (data not shown).

Figure 6.

Release assay of LD10 cells presented with a concentration range of the GST-MTS protein constructs prior to stimulation with C5a. Loop 1 (▪), loop 2 (▴), loop 3 (▾), C-terminus (•) and GST-MTS control (O). n = 4 ± SEM. The P-values of one-way anova tests against the GST-MTS control, for 20 μm of protein constructs added, are represented by asterisks (*P < 0·05, **P < 0·001).

To determine whether the effects seen on release were the result of internalization of the C5a receptor, LD10 cells were incubated with 20 μm of each of the GST-MTS-ID peptides and internalization with and without C5a was assessed. No significant difference was seen in receptor expression between cells incubated with any of the peptides on their own, or in the reduction in receptor expression seen when stimulated with C5a (data not shown). The GST-MTS-IDs were also investigated for their cytotoxicity as this may affect the cells ability to release 5-HT. Non of the peptides showed any effect on cell viability as determined by WST-1 staining (data not shown).

Effect of combination of intracellular loops

In the cell the intracellular domains of C5aR probably act in a co-ordinated manner when binding and activating their coupled G-proteins. Thus, the effect of adding the protein constructs to cells in combination was examined. GST-MTS-ID constructs were added (all at 20 μm) to LD10 cells singularly and in all combinations possible for the four intracellular domains. Results for the double, triple and quadruple mixes were compared to the GST-MTS control with 20 μm protein (Fig. 7); however, addition of GST-MTS alone at concentrations of up to 100 μm was also without effect (data not shown). Each of the ID constructs added singularly produced results similar to previous experiments (Fig. 7); ICL1 and ICL3 activating release of ICL2 having no effect and the CT exhibiting a repressive effect (Fig. 7, open bars), although in these particular experiments the differences between the control and test samples were not as great as before (Fig. 6).

Figure 7.

Release assay of LD10 cells presented with mixes of GST-MTS-ID protein constructs: L1, loop 1; L2, loop 2; L3, loop 3; and CT, C-terminus (20 μm each, 2 hr). Combinations are single (open bars), double (hatched bars), triple (cross-hatched bars), and all of the intracellular domains (solid bar). n = 5 ± SEM. The P-values of one-way anova tests against the GST-MTS control are represented by asterisks (*P < 0·01, **P < 0·001).

Surprisingly, the ICL1 and ICL2 combination caused a repressive effect even though ICL1 alone caused activation and ICL2 had no effect (Fig. 7, hatched bars). The ICL3 double mixes with ICL1 or ICL2 each showed an activation as was previously seen for ICL3 on its own, although the level of activation was not as high as previously (Fig. 7). All of the CT combinations showed a marked reduction in activity, which exceeded the levels seen by the CT alone and counteracted the activation seen for ICL1 and ICL3 when alone (Fig. 7).

The triplet combination of ICL1, ICL2 and ICL3 showed no effect, presumably because of the activating effect of ICL3 alone counteracting the repression observed with the ICL1 and ICL2 mix (Fig. 7, cross-hatched bars). However, like the double combinations, all the CT-containing groups had a repressive effect. The reduction in release for the triplet group was most pronounced for the ICL2, ICL3, CT combination (Fig. 7, cross-hatched bars). Finally the largest reduction in the release of 5-HT was seen when all the intracellular domains were added together (Fig. 7, black bar).

Discussion

The MTS sequence derived from the hydrophobic region in the signal sequence of Kaposi fibroblast growth factor^14,15 has been used to import a variety of small peptides into cells to study intracellular signalling pathways.^15,18,19 The same sequence was also shown to be capable of efficiently transporting the relatively large glutathione S-transferase (29 000 MW) protein into cells, both on its own or when coupled to an additional 12 000 protein.¹⁶ The addition of GST permits high level expression of target sequences in bacteria and rapid, single-step purification by glutathione–agarose affinity chromatography. Whilst the vector constructed here was primarily designed to permit cleavage of the MTS-IDs from the GST fusion protein after purification, this was not found to be necessary and GST-MTS-ID sequences were still imported into the cell, albeit with a reduced efficacy compared to the GST-MTS alone (Fig. 5). Leaving the GST attached to the target sequences also facilitated detection of the proteins using an anti-GST antibody.

When added individually ICL3, and to a lesser extent ICL1, enhanced C5a-mediated 5-HT release, the CT inhibited release and ICL2 had no detectable effect (Fig. 7). This would tentatively implicate ICL1, ICL3 and the receptor's CT in interacting with G_αi heterotrimeric G proteins. This contrasts with studies on the formyl peptide receptor (FPR) (the chemoattractant receptor with most sequence homology to C5aR) where ICL2 but not ICL3 or ICL1 appears to couple to G_αi G proteins.²⁰ However the latter study did implicate the membrane proximal region of the C terminus. Our data indicate this may also be true for the C5aR. We have shown previously that the distal 24 amino acids of the C5aR CT are required for receptor internalization but do not appear to be involved in coupling to secretion (and by implication G_αi).^12,17 This region contains serine residues critical for phosphorylation dependent desensitization.²¹ If the CT peptide was interfering with desensitization it would be expected to increase rather than decrease the C5a-mediated response and so it is likely that the proximal region does indeed interact with the G protein. Interestingly, combining the CT with any of the other domains always further inhibited C5a mediated secretion indicating a dominant role for this domain. This was not merely a non-specific additive effect of the GST-MTS sequence because concentrations of GST-MTS alone at up to 100 μm had no discernible effect (data not shown). Although higher concentrations of CT than ICL1 or ICL3 were required to produce a discern-ble effect (Fig. 6) this almost certainly reflects the reduced uptake of this peptide compared to the others (Fig. 5).

The fact that ICL1 and particularly ICL3 enhance, rather than suppress C5a-mediated release but do not trigger release on their own could indicate that these peptides stabilize the activated conformation of the G protein. However, an alternative possibility is that they interfere with receptor desensitization in some way. Both domains contain threonine and/or serine residues, particularly abundant in ICL3 (Fig. 1), which could act as targets for phosphorylation-dependent desensitization. The peptide may therefore compete for phosphorylation and/or arrestin binding.

ICL2 contains the sequence DRF. In most G-protein-coupled receptors this is highly conserved as DRY and the D and R residues are thought to play a role in receptor activation and G protein binding, respectively. In the FPR (where the sequence is DRC), this sequence has also been implicated in phosphorylation-dependent arrestin binding.²² The lack of effect of ICL2 alone initially indicated that this region may not be so important for C5aR coupling. However, the results of combining the different peptides reveal a much more complex picture. Whilst ICL1 apparently enhanced C5a-mediated coupling and ICL2 was without effect, in combination they inhibited the response, while combining ICL2 with ICL3 decreased the activating effect of ICL3. Furthermore, combining ICL2 with ICL1 and ICL3 or with ICL3 and CT or with all three other domains together increased the inhibition seen with these combinations alone. Again this is indicative of a direct interaction of this domain with the G protein as any interference with arrestin binding²² should it occur at this site in C5aR, would be expected to enhance signalling.

Taken together these data present a complex picture of C5aR/G protein interaction. The additive effects of the different domains indicate that all have unique sites of interaction with the G protein. The strongly inhibitory effect of the CT may indicate that this is a major G-protein-binding site, whilst the ability of ICL1 and particularly ICL3 to enhance secretion suggests a role for these domains in G protein activation.

Abbreviations

C5aR

complement fragment C5a receptor

CT

C-terminus

GPCRG

G protein-coupled receptor

5-HT

5-hydroxytryptamine

ICL

intracellular loop