"Glyco-sulfo barcodes" regulate chemokine receptor function

Abstract

Chemokine ligands and receptors regulate the directional migration of leukocytes. Post-translational modifications of chemokine receptors including O-glycosylation and tyrosine sulfation have been reported to regulate ligand binding and resulting signaling. Through in silico analyses, we determined potential conserved O-glycosylation and sulfation sites on human and murine CC chemokine receptors. Glyco-engineered CHO cell lines were used to measure the impact of O-glycosylation on CC chemokine receptor CCR5, while mutation of tyrosine residues and treatment with sodium chlorate were performed to determine the effect of tyrosine sulfation. Changing the glycosylation or tyrosine sulfation on CCR5 reduced the receptor signaling by the more positively charged CCL5 and CCL8 more profoundly compared to the less charged CCL3. The loss of negatively charged sialic acids resulted only in a minor effect on CCL3-induced signal transduction. The enzymes GalNAc-T1 and GalNAc-T11 were shown to be involved in the process of chemokine receptor O-glycosylation. These results indicate that O-glycosylation and tyrosine sulfation are involved in the fine-tuning and recognition of chemokine interactions with CCR5 and the resulting signaling.

Keywords: Chemokine receptor; G protein-coupled receptor; O-glycosylation; Tyrosine sulfation.

Fig. 1

In silico analysis reveals a general and conserved “Glyco-Sulfo barcode” in C–C Chemokine Receptor N-termini. N-terminal sequences from human (h) and mouse (m) CCRs are aligned and predicted potential O-glycosylation and tyrosine sulfation sites are indicated on the human sequences with a yellow square or SO₄, respectively. Consensus sequences for N-glycosylation are underlined. Potential N-glycosylation motifs are underlined

Fig. 2

Neuraminidase treatment inhibits phosphorylation of ERK1/2. THP-1 cells were stimulated with A 32 mU/ml or B 16 mU/ml neuraminidase dissolved in PBS. A similar amount of PBS was added to the control group. The cells were incubated for 2 min in a medium with or without CCL5. The amount of phosphorylated ERK1/2 (pERK) was determined by ELISA. The median values were determined for 7 to 8 experiments. 100% corresponds to the amount of pERK in medium-treated cells. Phosphorylation of ERK1/2 induced by CCL5 was compared and statistical significance was evaluated by the sign test; comparison to the medium-treated cells (*p < 0,05); comparison between groups with or without neuraminidase (#p < 0,05); ns = not significant. C The mean ± SEM percentage CCR1 and CCR5 receptor positive THP-1 cells of 12 experiments determined by flow cytometry. D The mean values ± SEM of 10 experiments of live THP-1 cells were determined by flow cytometry with FVS620 staining and the percentage of live cells are shown for each group

Fig. 3

Western blot analysis of CCR transiently transfected CHO and HEK293 cells showing O-glycans. A Cells were transfected with a FLAG-tagged CCR1, CCR3 or CCR5 construct, harvested and analyzed by western blot 24 h later. The decrease in size between WT and 5XKO (CHO cell lacking 5 different GalNAc-Ts) shows that CCR1 and CCR5 are glycosylated in these CHO cells. B CCR5 was transiently expressed in HEK293 SCs and cellular lysates were prepared 24 h after transfection. Flag-tagged CCR5 was immunoprecipitated with FLAG M2 antibody and analyzed by WB and lectin blotting, using FLAG M2-HRP or biotinylated-VVA followed by streptavidin-HRP, respectively

Fig. 4

Analysis of CCR5-mediated signaling in CHO cell lines producing distinct O-glycoforms. CHO cell lines were transfected with CCR5 and stimulated with 100 nM A CCL3, B CCL5, C CCL8 or D 100 µM ZnBip. The figures show the radar plot presentation of the mean ΔBRET obtained in mutant CHO cell lines compared to the mean obtained in CHO WT cell line (normalized to 100 percent) in percentages of 3 (5xKO ZnBip = 2) independent experiments, performed in duplicate, 40 min after addition of the ligand. Statistically significant differences (p < 0.05) in comparison to WT for 10^–7 M chemokine are indicated with an *. Glycans are shown as GalNAc = yellow square, Galactose = yellow circle and Sialic acid purple diamond. E–H Cells were stimulated with E CCL3, F CCL5, G CCL8 or H ZnBip. The results show the mean ΔBRET + standard error of the mean (SEM) of 3 (5xKO ZnBip = 2) independent experiments performed in duplicate 40 min after the addition of the ligand Statistically significant differences by a two-way ANOVA test and a multiple comparison Tukey test between WT and another cell line are highlighted for the whole curve with stars: ****p-value < 0.0001, ***p-value ≤ 0.001 and *p-value ≤ 0.05. Legend of symbols: Yellow square: GalNAc, yellow circle: galactose, purple triangle: sialic acid

Fig. 5

Analysis of CCR5-mediated signaling in CHO cell lines with changed tyrosine sulfation status. A, B Analysis of CHO cell lines transiently transfected with CCR5 and grown for 24 h in CHO medium with various concentrations of NaClO₃. Cells were stimulated with 10^–7 M CCL5 or 10^–3.5 M ZnBip. The radar plots show the mean ΔBRET of, respectively, 3 (5xKO ZnBip = 1) independent experiment(s), performed in duplicate, 40 min after addition of the ligand. C–F Analysis of CHO cell lines transiently transfected with CCR5 (blue line), CCR5 4xF (red line) or CCR5 and grown for 24 h in CHO medium with 100 mM of NaClO₃ (green line). Cells were stimulated with 100 nM CCL3, CCL5, CCL8 or 100 µM ZnBip. The radar plots show the mean ΔBRET obtained in mutant CHO cell lines of 3 independent experiments, performed in duplicate 40 min after the addition of the ligand. Statistically significant differences (p < 0.05) in comparison to WT for 10^–7 M chemokine are indicated with an * for CCR5, # for 4xF CCR5 compared to CCR5 and ° for treatment with or without 100 mM NaClO₃. G–N Cells were stimulated with G, K CCL3, H, L CCL5, I, M CCL8 or J, N ZnBip. The results show the mean ΔBRET ± standard error of the mean (SEM) of 3 (5xKO ZnBip = 1) independent experiments performed in duplicate, 40 min after the addition of the ligand. Statistically significant differences by a two-way ANOVA test and a multiple comparison Tukey test between WT and another cell line are highlighted for the whole curve with stars: ****p-value < 0.0001, ***p-value ≤ 0.001, **p-value ≤ 0.01 and *p-value ≤ 0.05. Legend of symbols: Yellow square: GalNAc, yellow circle: galactose, purple triangle: sialic acid

Fig. 6

Differences in glycosyation are observed depending on the method of removal of tyrosine sulfation. CHO cells were transfected with FLAG-tagged CCR5, 4xF CCR5 or CCR5 and treated with 100 mM NaClO3, harvested and analyzed by western blot. It was revealed that the 4xF mutant in all cell lines apparently not only removes tyrosine sulfation but also O-glycosylation. The cells treated with 100 mM NaClO3 still have O-glycosylation. A downshift was detected in the cell lines transfected with the 4xF sequence. This is likely due to alterations in SDS binding which has been reported before in studies with tyrosine to phenylalanine mutations in GPCRs (13). (Full WB can be found in supplementary Figure. S4.)

Fig. 7

Analysis of CCR5-mediated signaling in Knock-out CHO cell lines affects cell signaling. Analysis of CHO cell lines transiently transfected with CCR5. Cells were stimulated with A CCL3, B CCL5 or C CCL8. The radar plots show the mean ΔBRET obtained in mutant CHO cell lines compared to mean obtained in the CHO WT cell line (normalized to 100%) in percentages of 3 independent experiments performed in duplicates. Statistically significant differences (p < 0.05) in comparison to WT for 10^–7 M chemokine are indicated with an * for CCR5. D–F Cells were stimulated with D CCL3, E CCL5 or F CCL8. The results show the mean ΔBRET ± standard error of the mean (SEM) of 3 independent experiments performed in duplicate, 40 min after the addition of the ligand. Statistically significant differences by a two-way ANOVA test and a multiple comparison Tukey test between WT and another cell line are highlighted for the whole curve with stars: ****p-value < 0.0001, ***p-value ≤ 0.001, **p-value ≤ 0.01 and *p-value ≤ 0.05. Legend of symbols: Yellow square: GalNAc, yellow circle: galactose, purple triangle: sialic acid

Fig. 8

Analysis of GalNAc-T specificity towards the N-terminus of CCR5. A Sequence of the designed and unmodified peptide with amino acids shown with Rasmol colours based on their properties. Potential sites of O-glycosylation are shown with a core-1 structure and potential tyrosine sulfation is shown by a sulfate group. B A 20 mer peptide corresponding to the human CCR5 N-terminus was incubated with different recombinant GalNAc-Ts at 37 °C for overnight reaction. Samples were analyzed using MALDI-TOF and show that GalNAc-T1 was able to glycosylate the peptide at up to three different positions, whereas GalNAc-T11 could glycosylate only at two positions, while GalNAc-T2 and GalNAc-T3 failed to glycosylate the CCR5 peptide

Fig. 9

Possible glycosylation sites and residues susceptible to tyrosine sulfation on crystal and NMR structures. Location of glycosylated and sulfated residues in CCR5:chemokine complexes. Potentially sulfated and glycosylated residues are highlighted with red and yellow spheres, respectively. CCL3 (PDB ID: 5COR), CCL5 (5COY) and CCL8 (1ESR) are colored by electrostatic potential and aligned with CCL3 in the CCR5:CCL3 complex (7F1T). Top and bottom panels show side and top views, respectively. The Cys residue at position 16 is highlighted as a glycosylation site since the native sequence has a Thr residue in this position. The first seven residues of CCR5 are not shown due to a lack of electron density from the X-ray data