Complement-activation fragment C4a mediates effector functions by binding as untethered agonist to protease-activated receptors 1 and 4

Abstract

C4a is a small protein released from complement component C4 upon activation of the complement system's classical and lectin pathways, which are important constituents of innate immune surveillance. Despite the structural similarity between C4a and well-described anaphylatoxins C3a and C5a, the binding partner and biological function of C4a have remained elusive. Using a cell-based reporter assay, we screened C4a against a panel of both known and orphan G protein-coupled receptors and now provide evidence that C4a is a ligand for protease-activated receptor (PAR)1 and PAR4. Whereas C4a showed no activity toward known anaphylatoxin receptors, it acted as an agonist for both PAR1 and PAR4 with nanomolar activity. In human endothelial cells, ERK activation by C4a was mediated through both PAR1 and PAR4 in a Gα_i-independent signaling pathway. Like other PAR1 activators, C4a induced calcium mobilization through the PAR1/Gα_q/PLCβ signaling axis. Moreover, C4a increased stress fiber formation and enhanced endothelial permeability, both of which were reduced by PAR1 antagonists. In sum, our study identifies C4a as an untethered agonist for PAR1 and PAR4 with effects on cellular activation and endothelial permeability, thereby revealing another instance of cross-talk between the complement system and other host defense pathways.

Keywords: C4a; PAR1; PAR4; complement; endothelial cells.

Fig. 1.

C4a is a putative agonist for protease-activated receptors (PAR)1 and PAR4. Screening of the gpcrMAX panel in agonist mode reveals that PAR1 and PAR4 meet the selective criteria as putative targets for C4a. Dash line represents 30% of activation. The data represent the mean ± SD of duplicate samples.

Fig. S1.

Neither antagonistic activity against any known GPCR nor agonistic activity for orphan GPCR was detected for C4a. (A) Schematic overview of the PathHunter β-arrestin assay: GPCR activity was monitored using a proprietary reporter cell assay based on enzyme fragment complementation upon interaction between β-arrestin and the activated GPCR. In this system, the GPCR of interest is fused in frame with a small, 42 amino acid fragment of β-galactosidase (β-gal) at the C terminus, and coexpressed in cells (e.g., CHO-K1) that are stably expressing a fusion protein of β-arrestin with a large, N-terminal deletion mutant of β-gal. Activation of the GPCR by an agonist stimulates binding of β-arrestin to the GPCR and enforces complementation of the two enzyme fragments, resulting in the formation of active β-gal, which processes a substrate to a product that can be measured using chemiluminescent readouts. In antagonist mode, the same assay is performed with a known agonist of each GPCR in presence of an unknown compound, and the loss of reporter activity is measured. (B) When C4a (600 nM) was added to known GPCRs in presence of their corresponding agonists (Table S2), no significant antagonistic activity was found for C4a in the gpcrMAX GPCR assay panel. A dashed line represents 35% of antagonistic activity, which is considered relevant in antagonist mode. The data represent mean ± SD of duplicate samples. (C) No putative orphan GPCRs for C4a were found using the orphanMAX GPCR assay panel, which can only be screened in agonist mode. The dashed line represents the recommended activity threshold of 50% agonism. The data represent mean ± SD of duplicate samples.

Fig. 2.

C4a acts as an agonist for PAR1 and PAR4. CHO-K1 cells expressing either PAR1 or PAR4 were seeded in a 96-well plate and stimulated with PAR1 agonist TFLLR-NH₂, PAR4 agonist AY-NH₂, human C4a, C3a, or C5a. After stimulation, the chemiluminescent signal was detected. Culture medium alone was used as a control. (A) C4a and PAR1 agonist TFLLR-NH₂ dose-dependently activate PAR1. (B) PAR1 antagonist RWJ56110 dose-dependently inhibits C4a-induced PAR1 activation. (C) C4a and PAR4 agonist AY-NH₂ dose-dependently activate PAR4. (D) PAR4 antagonist tcY-NH₂ has no effect on C4a-induced PAR4 activation. The data are expressed as fold change compared with control or relative light unit (RLU) and represent the mean ± SE of three independent experiments.

Fig. S2.

The N-terminal area of C4a is essentially involved in the activation of PAR1 and PAR4. In contrast to the anaphylatoxins (C3a and C5a) and their corresponding receptors, the desarginated form of C4a (i.e., C4a-desArg) remains to activate both PAR1 (A) and PAR4 (B). Moreover, the C-terminal 20 amino acids of C4a (i.e., C4a-CT20) do not exert agonistic activity on either PAR1 or PAR4. The data are expressed as fold change compared with the control group and represent the mean ± SE of three independent experiments.

Fig. 3.

On CHO-K1 cells expressing either PAR1 or PAR4, recombinant human C4a colocalizes with the corresponding receptor. His6-C4a (red, anti-6His Ab) colocalizes with PAR1 (A; green, anti-PAR1 Ab) or PAR4 (B; green, anti-PAR4 Ab) on CHO-K1 cells expressing PAR1 or PAR4, respectively, but does not bind to wild-type CHO-K1 cells (Fig. S4A). The experiment was performed three times with similar results, and one representative experiment is shown.

Fig. S3.

Expression and purification of human His6-C4a for colocalization studies on PAR1/4-expressing cells. (A) Western blots reveal the expression of PAR1 in CHO-K1 PAR1 cells. (B) Western blots reveal the expression of PAR4 in CHO-K1 PAR4 cells. The cell lysate of CHO-K1 cells was used as control. (C) Recombinant human C4a fused to an S-tag and a His6-tag was expressed in E. coli strain Rosetta-gami B (DE3) Lys-S and purified using a His-Trap column and S-protein agarose (EMD). His6-S-tag-C4a, S-tag-C4a, and C4a were analyzed by SDS/PAGE (15%) and stained with Coomassie blue (white spacers indicate noncontiguous lanes from different Coomassie blue staining gels). (D) Amino acid sequence of the recombinant human 6His-C4a fusion protein used for the colocalization experiments. The protein containing 6His and S-tags was expressed in E. coli using pET-32a-hC4a construct based on a previous publication (9). Functional parts of the protein are marked as follows: blue, 6His-tag; red, enterokinase cleavage site; yellow, S-tag; and green, human C4a.

Fig. S4.

Colocalization of human His6C4a with PAR1 and PAR4 in CHO-K1 and HMEC-1 cells. (A) No colocalization of 6His-C4a with PAR1 (Upper) or PAR4 (Lower) was observed in wild-type CHO-K1 cells (for experiments using PAR1- and PAR4-expressing CHO-K1 cells see Fig. 3). (B) His6-C4a colocalizes with PAR1 in HMEC-1 cells. (C) His6-C4a colocalizes with PAR4 in HMEC-1 cells. The experiment was performed three times with similar results, and one representative experiment is shown.

Fig. 4.

In human HMEC-1 endothelial cells, C4a increases ERK phosphorylation through a Gα_i-independent signaling pathway. (A) C4a dose-dependently enhances ERK phosphorylation. (B) PAR1 antagonist RWJ56110 (10 μM), but not pertussis toxin (PTX; 0.3 μg/mL, 10 h), inhibits ERK activation upon C4a exposure (7 min). The data are expressed as fold change in densitometry of Western blots from the control group and represent the mean ± SE of three independent experiments. (n = 3; *P < 0.05 vs. control; pairwise two-sided Student’s t test.)

Fig. S5.

ERK activation by C4a on human endothelial cells. (A) C4a time-dependently induces ERK activation in HMEC-1 cells. The data are expressed as −fold change compared with the control group and represent the mean ± SE of three independent experiments. (B) In EA.hy926 cells, C4a induces ERK activation, and the PAR1 antagonist RWJ56110 (10 μM), but not the PAR4 antagonist tcY-NH₂ or pertussis toxin (0.3 μg/mL for 10 h), inhibits ERK activation after C4a treatment for 7 min. The experiment was performed three times with similar results, and one representative experiment is shown. (C) PAR4 agonist AY-NH₂ and thrombin activate ERK on EA.hy926 cells. The experiment was performed three times with similar results, and one representative experiment is shown.

Fig. S6.

PAR1 and PAR4 are involved in C4a-induced ERK activation in human endothelial cells. (A) Selective PAR4 receptor agonist peptide AY-NH₂ (AYPGKF-NH₂) significantly increases ERK phosphorylation in HMEC-1 cells. The data are expressed as fold change in densitometry of Western blots from the control group and represent the mean ± SE of three independent experiments. (n = 3; **P < 0.01 vs. control; pairwise two-sided Student’s t test.) (B) Pretreatment with PAR4 polyclonal blocking antibody (H-120) significantly decreases ERK phosphorylation by C4a treatment (7 min) in HMEC-1 cells. The experiment was performed three times. The data are expressed as fold change in densitometry of Western blots from the control group and represent the mean ± SE of three independent experiments. (n = 3; **P < 0.01 vs. control; pairwise two-sided Student’s t test.) White spacers indicate noncontiguous lanes of the same Western blot. (C) HMEC-1 cells were pretreated with the PAR1 antagonist RWJ56110 (10 mM), PAR4 blocking antibody (4 μg/mL), or a combination of both inhibitors. The data are expressed as fold change compared with the control group and represent the mean ± SE of three independent experiments. [n = 3; **P < 0.01 vs. C4a treatment (7 min); pairwise two-sided Student’s t test.]

Fig. 5.

C4a increases [Ca²⁺]_i and endothelial permeability via PAR1 activation. (A) C4a dose-dependently increases [Ca²⁺]_i in HMEC-1 cells. The data are expressed as relative fluorescence [Δ(ΔRFU)] and represent the mean ± SE of four to seven independent experiments. (B) PAR1 antagonist RWJ56110 (10 μM) significantly inhibits C4a-mediated elevation of [Ca²⁺]_i. The data are expressed as relative fluorescence [Δ(ΔRFU)] and represent the mean ± SE of seven independent experiments. (n = 7; **P < 0.01 vs. control; pairwise two-sided Student’s t test.) (C) Phospholipase C inhibitor U73122 inhibits C4a-induced elevation of [Ca²⁺]_i. The data are expressed as relative fluorescence [Δ(ΔRFU)] and represent the mean ± SE of eight independent experiments (n = 8; **P < 0.01 vs. control; pairwise two-sided Student’s t test). (D) In EA.hy926 cells, C4a dose-dependently increases endothelial permeability, and the PAR1 antagonist RWJ56110 significantly inhibits C4a-induced endothelial permeability. The data are expressed as OD at 650 nm and represent the mean ± SE of four independent experiments. [n = 4; *P < 0.05 vs. control; **P < 0.01 vs. control; ^##P < 0.01 vs. C4a (300 nM); pairwise two-sided Student’s t test.] (E) C4a-induced stress fiber formation is significantly decreased by pretreatment with PAR1 antagonist RWJ56110. The experiment was performed three times with similar results, and one representative experiment is shown.

Fig. S7.

C4a disrupts endothelial barrier function in HMEC-1 cells. (A) In HMEC-1 endothelial cells, C4a dose-dependently increases endothelial permeability, and the PAR1 antagonist RWJ56110 significantly inhibits C4a-induced endothelial permeability. The data are expressed as OD at 650 nm and represent the mean ± SE of three independent experiments. [n = 3; **P < 0.01 vs. control; ^##P < 0.01 vs. C4a (300 nM); pairwise two-sided Student’s t test.] (B) C4a-induced stress fiber formation is significantly decreased by pretreatment with PAR1 antagonist RWJ56110. The experiment was performed three times with similar results, and one representative experiment is shown.

Fig. S8.

C4a-induced PAR1 and PAR4 activation signals are not caused by thrombin contamination. The thrombin activity assay reveals that the C4a preparation used in this study fails to activate the chromogenic thrombin substrate even at a concentration of 3 μM, whereas thrombin itself shows activity at less than 1 nM, which can be reversed by adding a thrombin inhibitor. The data are expressed as fold change compared with the control group and represent the mean ± SE of three independent experiments.