Selective Allosteric Modulation of N-Terminally Cleaved, but Not Full Length CCL3 in CCR1

Abstract

Chemokines undergo post-translational modification such as N-terminal truncations. Here, we describe how N-terminal truncation of full length CCL3^(1–70) affects its activity at CCR1. Truncated CCL3^(5–70) has 10-fold higher potency and enhanced efficacy in β-arrestin recruitment, but less than 2-fold increased potencies in G protein signaling determined by calcium release, cAMP and IP₃ formation. Small positive ago-allosteric ligands modulate the two CCL3 variants differently as the metal ion chelator bipyridine in complex with zinc (ZnBip) enhances the binding of truncated, but not full length CCL3, while a size-increase of the chelator to a chloro-substituted terpyridine (ZnClTerp), eliminates its allosteric, but not agonistic action. By employing a series of receptor mutants and in silico modeling we describe residues of importance for chemokine and small molecule binding. Notably, the chemokine receptor-conserved Glu287^7.39 interacts with the N-terminal amine of truncated CCL3^(5–70) and with Zn²⁺ of ZnBip, thereby bridging their binding sites and enabling the positive allosteric effect. Our study emphasizes that small allosteric molecules may act differently toward chemokine variants and thus selectively modulate interactions of specific chemokine subsets with their cognate receptors.

Introduction

Chemokine receptors belong to the class A G protein-coupled receptors (GPCRs) and their ligands, the chemokines, are peptides of 8–12 kDa. Most chemokines contain four conserved cysteines, which form disulfide bridges that are name-giving for their subgrouping (CC-, CXC-, CX₃C-, and XC-group).¹ The CC/CXC-motif moreover separates the rigid chemokine core structure from its more flexible N-terminus.² In broad terms, these two regions have different roles in receptor interaction as the core domain provides the major share of receptor specificity and -affinity to extracellular receptor domains, while the N-terminus contributes more to receptor activation by interacting with transmembrane receptor residues. The effect of N-terminal modifications on activity further underpins the role of the chemokine N-terminus in receptor activation.^3,4 Most prominently, AOP-CCL5 acts as a potent CCR5 antagonist in Ca²⁺-signaling and chemotaxis in monocytes, but is also more potent than wild type CCL5 at internalizing and desensitizing CCR5, thereby explaining its strong anti-HIV activity.^5,6 Many other N-terminal modifications, such as single amino acid additions,⁷ chemical modifications,⁸⁻¹⁰ or sequence alterations of the entire chemokine N-terminus¹¹⁻¹⁵ have been investigated, and emphasized the role of the N-terminus in the receptor-activating step.

Most endogenous nonpeptide ligands for class A GPCRs bind solely in the transmembrane receptor pocket.^16,17 Likewise, most synthetic small molecules for chemokine receptors possess a positive charge in their structure that interacts with a chemokine receptor-conserved glutamate (Glu) in transmembrane (TM) domain 7 (position 7.39 according to the Ballesteros/Weinstein nomenclature¹⁸). This Glu is located centrally in the transmembrane binding pocket where most chemokine N-termini are projected to interact.^19,20 Since the binding site of small molecules is most often located in the transmembrane pocket, the molecules’ effect on the chemokine is considered allosteric. However, recent crystal structures of chemokine receptors in complex with chemokine ligands, such as US28 with CX₃CL1,²¹ or CCR5 with an antagonistic CCL5-derivative,²² showed that the N-terminus of the chemokines reaches deeply into the transmembrane receptor domain. This raises the question of whether small molecules in the transmembrane binding pocket are solely allosteric or partially overlapping with the chemokines.

The chemokine system mediates leukocyte migration, and is central to homeostatic and inflammatory functions of the immune system. It also plays a role in hematopoiesis, angiogenesis, lymphoid organogenesis, and HIV-infection, and aberrant regulation of the chemokine system can result in allergic, inflammatory, or autoimmune diseases, or may promote cancer metastasis and tumor angiogenesis.^1,23 It has, therefore, been intensively studied in terms of its druggability. Numerous ligands with good in vitro antagonistic effects have been identified, but only three compounds have reached the market: the CXCR4 antagonist Plerixafor, the CCR5 antagonist Maraviroc, and Mogamulizumab, targeting CCR4 (only approved in Japan).¹ To fully exploit the pharmaceutical potential of chemokine receptors as drug targets, a better understanding of the complexity of the chemokine system is needed.

The chemokine receptor CCR1 binds several chemokines with CCL3 and CCL5 being most potent in addition to eight other CC-chemokines (CCL4, 7, 8, 13, 14, 15, 16, and 23) with lower affinity.¹ CCR1 has been investigated as a target for the treatment of multiple sclerosis, rheumatoid arthritis, and chronic obstructive pulmonary disease. Compounds such as BX471 (Schering AG),²⁴ MLN 3897 (Millenium),²⁵ CP-481,715 (Pfizer),²⁶ AZD4818 (Astra Zeneca),²⁷ C-4462 and C-6448 (Merck)²⁸ have been subject to clinical trials, but failed due to lack of in vivo efficacy.¹ The reasons could be insufficient receptor coverage,²⁹ or the need for chemokine neutralization rather than receptor inactivation, as suggested by the observation that CCL3 neutralization is completely protective in a mouse model for multiple sclerosis.³⁰

CCL3 is available in various forms: the commercially available long variant CCL3^(1–70), in addition to CCL3^(2–70), and N-terminally truncated CCL3^(5–70), numbered 23–92, 24–92, and 27–92, respectively, if the signal peptide is counted (Figure 1A,B). A specific role for the truncated variant in physiology is supported by the finding of CCL3^(5–70) in ascites of patients with ovarian cancer.³¹ We have previously described the pharmacological properties of a class of small molecule positive allosteric modulators consisting of complexes between a metal ion (Zn²⁺ or Cu²⁺) and the chelators 2,2′-bipyridine, 1,10-phenanthroline, or 2,2′:6′,2′′-terpyridine (and their derivatives) in chemokine receptors. ZnBip and ZnPhe are low-potent agonists at a subset of chemokine receptors, and act as allosteric enhancers of CCL3-binding to CCR1 and CCR5.³²⁻³⁶ Here, we investigate the impact of N-terminal CCL3 truncation for binding and activation of CCR1 and for the positive allosteric modulation of CCL3 in CCR1.

Figure 1.

N-terminal truncation CCL3^(5–70) of CCL3^(1–70) binds with unaltered affinity, but higher potency and efficacy in arrestin recruitment. (A) Amino acid sequence of CCL3^(1–70) and the truncated CCL3^(5–70). (B) Structure of CCL3^(1–70) (PDB: 2X69, manually added residues 1–3) and indication of the first four amino acids that will be cleaved off in CCL3^(5–70). (C) Homologous binding curves with ¹²⁵I-CCL3^(1–70) (squares) or ¹²⁵I-CCL3^(5–70) (circles) against respective cold ligand, normalized to the specific binding of each tracer as 100% and binding in mock-transfected cells as 0% (nonspecific binding). (D–G) Dose–response curves of CCL3^(1–70) (circles) and CCL3^(5–70) (squares) in different in vitro assays: IP₃-accumulation (D), cAMP-generation (E), Ca²⁺-release (F), and β-arrestin recruitment (G). All dose–response curves are normalized to the response obtained for CCL3^(1–70) (100%) and background signaling (0%). The homologous binding curves are normalized to the specific binding of tracer (100%) and nonspecific binding on mock-transfected cells (0%). The values of all IC₅₀ and EC₅₀-values are given. * p ≤ 0.0001 (unpaired t test), n ≥ 3.

Results

Truncation of CCL3^(1–70) to CCL3^(5–70) Results in More Potent and Efficacious Activation of CCR1

First, we determined the activity of CCL3^(1–70) and CCL3^(5–70) (Figure 1A,B) in a range of intracellular signal transduction assays: IP₃-accumulation (Figure 1D), cAMP-generation (Figure 1E), Ca²⁺-signaling (Figure 1F), and β-arrestin recruitment (Figure 1G). In all G protein-mediated pathways (Figure 1D–F), CCL3^(5–70) appeared to have slightly improved activity compared to CCL3^(1–70) although this difference did not reach significance. However, with regard to β-arrestin recruitment, CCL3^(5–70) was significantly improved with 10-fold higher potency and 130% of maximal efficacy compared to CCL3^(1–70). We moreover investigated the binding properties of both CCL3-variants in a homologous competition binding setting with iodination of both variants. Using this approach, we observed no significant difference in the binding affinities with log K_d −8.0 ± 0.22 (n = 8) and −7.7 ± 0.31 (n = 6), for CCL3^(1–70) and CCL3^(5–70), respectively (Figure 1C). Similar high affinities were obtained in the heterologous settings with log IC50 values of −8.2 ± 0.45 (n = 3) for CCL3^(5–70) in competition with ¹²⁵I-CCL3^(1–70) and −8.0 ± 0.04 (n = 37) for CCL3^(5–70) in competition with ¹²⁵I-CCL3^(1–70), confirming an overall similar binding site of the two CCL3 forms.

Allosteric Modulation of CCR1 Depends on the Length of the Chemokine N-Terminus

As most small molecule ligands for chemokine receptors bind to the transmembrane pocket,^19,20 we explored whether the length of the chemokine N-terminus, which interacts with this receptor area,^3,4 influenced the action of allosteric modulators on CCR1. We used the previously described allosteric modulators,^32,33 Zn²⁺ complexed with the metal ion chelators 2,2′-bipyridine (ZnBip) or 1,10-phenanthroline (ZnPhe) (Figure 2A). ZnBip and ZnPhe are ago-allosteric modulators as agonists and positive allosteric modulators of CCL3^(5–70) binding to CCR1.³² In contrast to their ability to enhance the binding of CCL3^(5–70), ZnBip and ZnPhe did not enhance the specific binding of the longer variant ¹²⁵I-CCL3^(1–70) (Figure 2B,C). This suggests that the truncation of CCL3 allows positive allosteric effects of the small molecules ZnBip and ZnPhe, whereas the extended variant, CCL3^(1–70), prevents this action.

Figure 2.

Truncation of the CCL3 N-terminus enables positive allosteric modulation by ZnBip and ZnPhe in CCR1. (A) The structure of ZnBip and ZnPhe with indication of their potency in IP₃-assays as determined previously.³⁴ (B,C) Heterologous competition binding assays with ¹²⁵I-CCL3^(5–70) (white symbols) or ¹²⁵I-CCL3^(1–70) (black symbols) in competition with ZnBip (circles, B) or ZnPhe (squares, C). Data are normalized to the specific binding of each tracer as 100% and nonspecific binding in mock-transfected cells as 0%. All experiments were repeated at least 3 times (n ≥ 3).

The Allosteric Enhancement of CCL3^(5–70) is Lost upon Increased Size of the Metal Ion Chelators

Considering that the long N-terminus in CCL3^1–70 excludes the positive allosteric modulation by ZnBip and ZnPhe (Figure 2), we speculated that an increase in the size of small molecule ligands could prevent the allosteric effect. We therefore included two additional metal ion chelator complexes, ZnTerp and ZnClTerp with a third pyridine ring (Figure 3A), compared to the smaller ZnBip with two pyridine rings (Figure 2A). These compounds were previously described as 6.7-fold and 56-fold more potent than ZnBip on CCR1.³⁴ Neither ZnTerp nor ZnClTerp were able to enhance the binding of ¹²⁵I-CCL3^(5–70). In fact, ZnClTerp competed with ¹²⁵I-CCL3^(5–70) with a 50%-displacement at 10 mM (Figure 3B). This suggests that the positive allosteric modulation between the small molecule ligands and CCL3 in CCR1 requires a short chemokine N-terminus and a small chelator size (Figure 2).

Figure 3.

Larger chelator structures abrogates the positive allosteric modulation of CCL3^(5–70). (A) The structure of ZnTerp with the change to the ZnBip structure circled in red and ZnClTerp with the added chlorine indicated, and indication of their agonistic potency in IP₃-assays, as determined previously.³⁴ (B) Heterologous competition binding assays with ¹²⁵I-CCL3^(5–70) in competition with ZnTerp (gray triangles), ZnClTerp (gray diamonds), ZnBip (white circles), or ZnPhe (white squares). Notably, the positive allosteric enhancement that was found for ZnBip and ZnPhe is not observed for ZnTerp or ZnClTerp. Graphs are normalized to the specific binding of the tracer as 100% and its nonspecific binding in mock-transfected cells as 0%. n ≥ 3.

Computational Model of the Binding Site of ZnBip and ZnClTerp in CCR1

To understand the differences in agonism and allosteric action of ZnBip and ZnClTerp, we docked both into a model of CCR1 (Figure 4A,B), which was based on the X-ray crystal structure of CCR5 in complex with [5P7]CCL5 (PDB: 5UIW). As previously described, the chemokine receptor-conserved Glu287^7.39 was involved in the coordination of Zn²⁺ in both metal ion chelator complexes.³² Zn²⁺ is preferentially coordinated in a tetrahedral manner³⁷ and, for ZnBip, the other three coordinating partners are the two nitrogens of the bipyridine (Bip), and the hydroxyl-group of Tyr113^3.32. In the chloro-substituted terpyridine ring (ZnClTerp), all three nitrogens contribute to the coordination of Zn²⁺. Although this does not display a tetrahedral geometry, we observed that all three nitrogens are indeed required for receptor activation of this ligand. Removal or repositioning of the nitrogen in one outer ring of terpyridine abolishes the activity of ZnClTerp (Figure 5), which is in support of a central role of all three nitrogens for ligand–receptor interaction. Furthermore, Trp90^2.60 makes edge-to-face aromatic interactions with the bipyridine of ZnBip. No other close interactions with the bipyridine were observed (Figure 4A). Like bipyridine, the chloro-terpyridine chelator interacts with Trp90^2.60 and additionally with Tyr114^3.33 and Tyr255^6.51(Figure 4B).

Figure 4.

ZnBip and ZnPhe bind to Glu287^7.39 and aromatic residues in the transmembrane domain of CCR1. (A) Computational model of ZnBip (green) and (B) of ZnClTerp (cyan) bound to CCR1. Receptor residues near the ligand are represented as sticks, helices are indicated, and the ion Zn²⁺ is shown as a gray sphere. Projected interactions are indicated by yellow dotted lines. (C,D) IP₃ activity on key CCR1 mutants by ZnBip relative to wild type (dashed line, C), and positive allosteric binding modulation of CCL3^(5–70) to the same residues (D). (E) Dissimilar IP₃ dose–response curves of the same mutations activated with ZnClTerp. IP₃ graphs are normalized to full CCL3^(1–70) activation as 100% and basal CCR1 activity as 0%. Binding curves are normalized to the specific binding of the radioligand, n ≥ 3.

Figure 5.

Role of the nitrogen in the third pyridyl-ring of ZnTerp. (A) Structure of Terp and Terp-analogues, which either lack the third nitrogen (PhBip), or in which this nitrogen is placed in a position that is unfavorable for zinc-complex formation (mTerp, pTerp). (B) Column diagram presenting the efficacy of these analogues alone or in complex with Zinc in comparison to the efficacy of ZnTerp, ZnClTerp, or no ligand. All compounds were given at 100 μM and the IP₃-accumulation of CCR1-expressing cells was tested. Data represent mean ± SEM of at least three repetitions and are normalized to no ligand as 0% and the efficacy of CCL3^(1–70) as 100%.

The Effect of CCR1 Receptor Mutagenesis on the Activity of CCL3, ZnBip, and ZnClTerp

To verify the suggested binding poses of ZnBip and ZnClTerp in CCR1 (Figure 4), we prepared 27 mutants within the transmembrane pocket of CCR1, and tested whether they were activated by CCL3^(1–70), ZnBip, and ZnClTerp in the IP₃-assay (Table 1). These residues were selected based on our modeling and previous work, where they were predicted to be available for chemokine binding in CCR1 (Leu87^2.57, Ser110^3.29, Gly207^5.46, and Glu287^7.39) or aligned well to similar small molecule binding in CCR1, CCR5, and CCR8 (Trp90^2.60, Lys94^2.64, Phe112^3.31, Tyr113^3.32, Tyr114^3.33, Trp195^5.34, Lys196^5.35, Phe206^5.45, Tyr255^6.51, and Tyr291^7.43).^{32,33,38−42} In support of the surface expression and general functionality of the mutant library, we observed that CCL3^(1–70) activated all mutants (Table 1). This also highlights that single point mutations within the transmembrane area of CCR1 do not significantly affect the activity of CCL3^(1–70). The predicted interaction with Tyr113^3.32 was supported by the functional test of Y113A^3.32, at which ZnBip changed from agonist to inverse agonist (Figure 4C). Furthermore, Y113F^3.32 led to a 31-fold reduced (agonistic) potency, and E287A^7.39 abolished activity (Table 1). This supports the model in which Glu287^7.39 and the OH-group of Tyr113^3.32 coordinate zinc (Figure 4A). The W90A^2.60 mutation had a severe effect on the activity of ZnBip, despite the fact that virtual mutation of Trp90^2.60 to alanine only resulted in an increase of the ICM_VLS docking score for ZnBip from 8.8 to 10.3 (for comparison, mutating Glu287^7.39 to alanine increases the score from 8.8 to 21.2). Like ZnBip, ZnClTerp turned into an inverse agonist at Y113A^3.32 (Figure 4E), but had only 6.8-fold decreased potency at Y113F^3.32. In accordance with the model, this suggests that the OH-group of Tyr113^3.32 is not as central for ligand-interaction with ZnClTerp. Similarly to ZnBip, the activity of ZnClTerp was abolished by E287A^7.39, and in addition its potency decreased 27-fold by W90A^2.60, 12-fold by Y255A^6.51, and 5.3-fold by Y114A^3.33. These residues interacted with Zn²⁺ or the chelator in the model (Figure 4B). None of the remaining mutants impaired the activity of ZnClTerp. Altogether, this is in support of its suggested binding pose.

Table 1. Activity of CCL3^(1-70), ZnBip and ZnClTerp at CCR1 and Mutant Receptors Measured in the IP₃-Assay.

			CCL3(1–70)					ZnBip					ZnClTerp
			EC50a					EC50					EC50
		location	log ± SEM	(nM)	Fmutb	(n)c		log ± SEM	(μM)	Fmut	(n)		log ± SEM	(μM)	Fmut	(n)
TM-2	WT		–8.42 ± 0.03	3.8	1.0	(118)		–4.49 ± 0.04	32	1.0	(52)		–6.11 ± 0.09	0.78	1.0	(12)
	F83A	2.53/II:13d	–8.37 ± 0.17	4.2	1.1	(6)		–4.70 ± 0.11	20	0.62	(3)		–6.43 ± 0.11	0.37	0.48	(3)
	L87F	2.57/II:17	–8.66 ± 0.14	2.2	0.57	(4)		–4.02 ± 0.09	95	2.9	(4)		n.d.
	W90A	2.60/II:20	–8.67 ± 0.08	2.1	0.56	(9)		–3.93 ± 0.06	119	3.7	(5)		–4.68 ± 0.15	21	27	(3)	*f
	K94A	2.64/II:24	–8.75 ± 0.12	1.8	0.47	(8)		–5.01 ± 0.07	9.7	0.30	(6)		n.d.
	K94L	2.64/II:24	–9.03 ± 0.03	0.9	0.25	(4)	*	–5.04 ± 0.03	9.1	0.28	(4)		n.d.
TM-3	K107A	3.26/III:02	–7.86 ± 0.05	14	3.6	(5)	*	–4.18 ± 0.10	66	2.0	(5)		n.d.
	L109A	3.28/III:04	–7.98 ± 0.07	10	2.7	(5)		–4.02 ± 0.08	97	3.0	(3)		–5.64 ± 0.09	2.3	2.9	(3)
	S110A	3.29/III:05	–8.73 ± 0.10	1.8	0.49	(5)		–4.53 ± 0.08	29	0.91	(5)		n.d.
	F112A	3.31/III:07	–8.47 ± 0.15	3.4	0.89	(7)		–4.96 ± 0.16	11	0.34	(4)		n.d.
	Y113Ae	3.32/III:08	–8.69 ± 0.09	2.1	0.54	(28)	*	-5.33 ± 0.19	4.7	0.15	(9)	*	-5.59 ± 0.26	2.6	3.3	(3)	*
	Y113F	3.32/III:08	–7.90 ± 0.06	13	3.3	(5)	*	>−3		31	(3)	*	–5.28 ± 0.18	5.3	6.8	(3)	*
	Y114A	3.33/III:09	–8.84 ± 0.16	1.4	0.38	(11)	*	–4.30 ± 0.10	50	1.5	(3)		–5.38 ± 0.15	4.1	5.3	(3)	*
	Y114F	3.33/III:09	–8.14 ± 0.09	7.2	1.9	(7)		–3.94 ± 0.08	114	3.5	(6)	*	–5.85 ± 0.07	1.4	1.8	(5)
	L117A	3.36/III:12	–8.22 ± 0.07	6.0	1.6	(5)		–3.96 ± 0.10	109	3.4	(4)		–5.82 ± 0.10	1.5	2.0	(4)
TM-5	W195A	5.31/V:-04	–7.56 ± 0.20	27	7.2	(4)	*	–3.72 ± 0.13	191	5.9	(3)	*	n.d.
	K196A	5.32/V:-03	–7.98 ± 0.02	11	2.8	(6)		–4.85 ± 0.17	14	0.43	(4)		n.d.
	F206A	5.42/V:08	–8.43 ± 0.16	3.7	0.97	(7)		–4.76 ± 0.22	17	0.54	(4)		n.d.
	G207A	5.43/V:09	–8.29 ± 0.13	5.1	1.3	(8)		–4.62 ± 0.14	24	0.73	(5)		n.d.
	G207F	5.43/V:09	–8.57 ± 0.18	2.7	0.71	(6)		–4.57 ± 0.24	27	0.84	(4)		n.d.
TM-6	Y255A	6.51/VI:16	–8.46 ± 0.10	3.4	0.90	(20)		–3.91 ± 0.08	124	3.8	(10)	*	–5.05 ± 0.17	9.0	12	(5)	*
	Y255F	6.51/VI:16	–8.45 ± 0.20	3.5	0.92	(5)		–4.99 ± 0.12	10	0.31	(4)		–6.82 ± 0.04	0.15	0.20	(3)	*
TM-7	D280A	7.32/VII:-02	–8.30 ± 0.16	5.0	1.3	(3)		–4.15 ± 0.08	71	2.2	(4)		–5.93 ± 0.14	1.2	1.5	(3)
	L281A	7.33/VII:-01	–8.05 ± 0.11	9.0	2.4	(3)		–4.76 ± 0.09	17	0.54	(3)		n.d.
	E287A	7.39/VII:06	–8.29 ± 0.12	5.1	1.4	(12)		>−3		31	(9)	*	–3.85 ± 0.07	142	183	(5)	*
	A290S	7.42/VII:09	–8.45 ± 0.07	3.5	0.93	(5)		–4.34 ± 0.10	45	1.4	(4)		n.d.
	Y291A	7.43/VII:10	–8.85 ± 0.14	1.4	0.37	(9)	*	–4.76 ± 0.08	17	0.53	(5)		n.d.
	Y291F	7.43/VII:10	–8.54 ± 0.10	2.9	0.76	(3)		–4.61 ± 0.03	24	0.75	(3)		n.d.

^aEC₅₀ values are given in log and nM/μM.

^bF_mut is the factor presenting the fold-decrease of EC₅₀ for the mutant compared to WT CCR1. F_mut changes greater than a 3-fold decrease are highlighted in bold.

^c(n) is the number of experiments.

^dPosition of mutants according to the Ballesteros/Weinstein (left) and Baldwin/Schwartz (right) numbering system.

^eThe ZnBip and ZnClTerp activity on this mutation refers to inverse agonism.

^fStatistics: one-way Anova of log ± SEM (n) of mutant compared to wt,

^*p ≤ 0.05.

Computational Model of Concomitant Binding of ZnBip and CCL3^(5–70) in CCR1

In order to understand the positive allosteric modulation of ZnBip for the binding of CCL3^(5–70), we made a model of both ligands simultaneously bound to CCR1. For this purpose, we modeled CCL3^(5–70) onto the chemokine coordinates from the crystal structure of 5P7-CCL5 with CCR5,²² and proceeded with the receptor model and metal ion chelator complex docking as described above (Figure 6A). To our surprise, we found that in the presence of CCL3^(5–70), there is no space for ZnBip to be coordinated by Glu287^7.39. Instead, Zn²⁺ is coordinated by Tyr255^6.51, Tyr113^3.32, and both nitrogens of bipyridine. However, the role of Glu287^7.39 is still central, as it coordinates the N-terminus of the chemokine on one side, and Tyr255^6.51 on the other. Thus, Glu287^7.39 bridges the chemokine to the metal ion chelator complex. In support of the lack of allosteric modulation of CCL3^(1–70), it was not possible to create an energetically favorable model of CCL3^(1–70) with the ago-allosteric modulator ZnBip (CCL3^(1–70)-ZnBip complex) or of the larger allosteric modulator with the shorter CCL3 (CCL3^(5–70)-ZnClTerp complex) in CCR1.

Figure 6.

In the ternary complex model of ZnBip with CCL3^(5–70) in CCR1, Glu287^7.39 interacts directly with the chemokine N-terminus and indirectly via Tyr255^6.51 with Zn²⁺ of ZnBip. (A) ZnBip bound to the CCR1-CCL3^(5–70) complex; Zn²⁺ is coordinated by Tyr255^6.51, Tyr113^3.32, and Bip. Glu287^7.39 mediates interaction between ZnBip and CCL3^(5–70). (B) ZnBip bound to the Y255A^6.51-CCR1-CCL3^(5–70) complex. In the absence of Tyr255^6.51 Zn²⁺ is coordinated by Glu287^7.39, Tyr113^3.32, and Bip. Receptor residues near the ligand are represented as sticks, helices are indicated, and the ion Zn²⁺ is shown as a sphere. Bip is green; CCL3^(5–70) is purple.

Receptor Residues Involved in the ZnBip-Mediated Positive Allosteric (PAM) Effect on CCL3^(5–70)

To validate the suggested binding mode of ZnBip and CCL3^(5–70) in CCR1, we screened 11 selected mutants for their impact on the allosteric enhancement of ¹²⁵I- CCL3^(5–70) binding by ZnBip (Table 2). As expected, Ala-substitution of Glu287^7.39 completely abrogated the ability of ZnBip to increase the binding of CCL3^(5–70). The same was observed for W90A^2.60, possibly by creating space for the chemokine to move away from Glu287^7.39, thereby breaking the link to ZnBip. None of the other tested residues had any effect, including Y113A^3.32 and Y255A^6.51. To investigate why we did not observe an effect of Y255A^6.51 although it was shown to coordinate Zn²⁺, we repeated the modeling in CCR1 Y255A^6.51 (Figure 6B), and found that in the absence of Tyr255^6.51, Glu287^7.39 directly coordinates Zn²⁺ and the N-terminus of CCL3^(5–70) explaining why its mutation is tolerated. A similar redundancy mechanism is conceivable for the role of Tyr113^3.32, whose mutation to alanine likewise is tolerated in the allosteric binding assay. In Y113A^3.32, it is possible that the neighboring Tyr114^3.33 overtakes the coordination of Zn²⁺.

Table 2. Competition Binding Assays of ¹²⁵I-CCL3^(5-70) to CCR1 WT and Mutants Shows That Only Small Molecule PAM Activity Relies on Transmembrane Receptor Residues.

			125I-CCL3(5–70) against CCL3(1–70)						125I-CCL3(5–70) against ZnBip
			IC50a				Bmax (fmol/100 000 cells)b		(n)c	IC50				% enhancementd		(n)
		location	log ± SEM	(nM)	Fmute		Mean ± SEM			log ± SEM	(μM)	Fmut		Mean ± SEM
	WT		–8.0 ± 0.03	10	1.0		1695 ± 214		(33)	–4.2 ± 0.10	60	1.0		235 ± 22		(17)
TM-2	W90A	2.60/II:20	–7.8 ± 0.24	15	1.5		6131 ± 2349	*g	(3)	no enhanced binding	*					(3)
TM-3	S110W	3.29/III:05	–7.8 ± 0.05	17	1.8		4869 ± 1048	*	(3)	–3 ± ####		17	*			(3)
	Y113A	3.32/III:08	–8.3 ± 0.06	5.0	0.52		758 ± 158		(11)	–4.5 ± 0.21	32	0.54		264 ± 29		(5)
	Y114A	3.33/III:09	–8.1 ± 0.11	7.8	0.82		1640 ± 701		(8)	–4.7 ± 0.28	19	0.31		169 ± 42		(4)
	L117F	3.36/III:12	–7.7 ± 0.26	19	2.0		3332 ± 666		(2)	–3.9 ± 0.41	131	2.2		174 ± 27		(2)
TM-6	F248A	6.44/VI:09	–7.7 ± 0.10	18	1.9		3313 ± 347		(3)	–3.8 ± 0.12	168	2.8		223 ± 42		(3)
	W252A	6.48/VI:13	–8.6 ± 0.07	2.4	0.25	*	135 ± 27		(6)	–4.6 ± 0.01	26	0.44		363 ± 28	*	(4)
	Y255A	6.51/VI:16	–8.9 ± 0.22	1.4	0.15	*	149 ± 79		(5)	–4.3 ± 0.04	48	0.80		187 ± 17		(4)
TM-7	E287A	7.39/VII:06	–9.0 ± 0.15	0.9	0.09	*	68 ± 19		(6)	no enhanced binding	*					(5)
	Y291A	7.43/VII:10	–8.3 ± 0.10	4.9	0.52		1134 ± 503		(5)	–4.2 ± 0.17	63	1.1		344 ± 51		(5)

^aIC₅₀ is concentration at which 50% of ¹²⁵I-CCL3^(5–70) is displaced, IC₅₀ values are given in log and nM or μM.

^bB_max is total number of binding sites calculated as fmol/100.000 cells.

^c(n) is the number of experiments.

^d%-Enhancement is the amount of ZnBip-induced enhanced binding relative to the amount of specific binding of CCL3^(5–70) to each mutant (set to 100%).

^eF_mut is the fold-decrease of IC₅₀ of mutant compared to WT. F_mut changes greater than a 2-fold decrease are highlighted in bold.

^fPosition of mutants according to the Ballesteros/Weinstein (left) and Baldwin/Schwartz (right) numbering system.

^gStatistics: one-way Anova of log ± SEM (n) of mutant compared to wt.

^*p ≤ 0.05.

Discussion and Conclusions

Effect of Chemokine Truncation on CCL3 Activity

The truncation of chemokines is often crucial for attenuation, neutralization, and termination of inflammatory responses and thus influences how immunological challenges are resolved. Chemokine truncation may occur at the N-terminal or C-terminal end and may have a multitude of consequences such as increasing or decreasing biological function, creation of antagonistic properties, complete inactivation, or change in receptor preference. Truncating proteases include CD13 and CD26, which exist in membrane-associated or soluble forms, for example, in plasma. CD13 cleaves N-terminal residues from, for example, CXCL5 and CXCL8 leading to an increased activity with moderate truncation, but loss of function after substantial truncation (reviewed in Vanheule et al.⁴³). CD26 is a serine-protease that cleaves dipeptides from substrates with a (hydroxy)Pro or Ala in the penultimate position. Chemokine substrates for CD26 are, for example, CCL3L1, CCL4, and CCL5. Furthermore, matrix-metalloproteases (MMPs), stored in intracellular vesicles and released upon appropriate stimulus, cleave CCL2, 7, and 8, which results in retained receptor binding, but reduced Ca²⁺-signaling and potentially antagonistic activities.⁴⁴⁻⁴⁶ Here, we find that N-terminal truncation of CCL3 results in similar affinities and a tendency toward higher activity through G proteins. A similar activating effect was described for CD26-mediated truncation of CCL4, where increased Ca²⁺-signaling properties via CCR1 and CCR2 were obtained.^47,48 In contrast CD26-mediated truncation of CCL5^(1–68) to CCL5^(3–68) impaired Ca²⁺-signaling through CCR1 and CCR3, while improving it through CCR5.⁴⁹⁻⁵¹ Of note, we find that the truncation of CCL3 induces selectivity toward β-arrestin signaling with a 10-fold higher potency for truncated compared to full-length CCL3, which could be important for retention of receptors at the cell surface.

CCL3 exists in several N-terminal variants, of which CCL3^(5–70) has been confirmed.⁵² However, it does not carry the typical cleavage site for CD13, CD26, and MMPs,^53,54 and CCL3^(5–70) found in ascites might therefore arise from CCL3L1 and resemble CCL3^(5–70) in its exact mass and amino acid composition, with only two amino acid differences (S61 to G62 and G69 to S70 in CCL3 and CCL3L1, respectively).³¹ These chemokines cannot be distinguished using mass spectrometry and the small alterations are unlikely to affect the allosteric modulation by small molecule drugs presented in this study. Consistent with our findings, increased affinity and potency of short compared to full length CCL3 at CCR1 have been reported previously for CCL3 and CCL3L1 truncations.^54,55 In addition, in a previous study, CCL3L1 displayed potent interactions with CCR1 and CCR5, with a 10-fold increase after N-terminal truncation, which is mirrored in our study of a CCL3 truncation.³⁶

Chemokine-Dependent Allosteric Action and the Role of Glu7.39^7.39

The chemokine system is well suited for the development of allosteric modulators, as the main chemokine binding domain is formed by extracellular receptor regions.^2,56 Therefore, small molecule ligands that bind to the transmembrane receptor domain have often been classified as allosteric, that is, not overlapping with the chemokine. However, it is well established that the flexible chemokine N-termini interact with specific receptor residues in the transmembrane area to induce receptor activation,^2,56 as confirmed by crystal structures of chemokines in complex with their receptors.^21,22,57 One residue that seems to be important for both chemokines and small molecule ligands is the chemokine receptor-conserved glutamate in the top of TM7 (position 7.39) that is found in 74% of endogenous chemokine receptors, but in less than 1% of nonchemokine class A GPCRs.¹⁹ This glutamate is crucial for activation by chemokines in a number of chemokine-receptor pairs such as CCL2 and −7 with CCR2, CCL11, with CCR3, CCL17, and −22 with CCR4 and CCL5 with CCR1 and CCR5.³ In contrast, as shown here, and previously by others, CCL3 does not depend on Glu287^7.39 for CCR1 or CCR5 activation,^32,33 whereas it coordinates Zn²⁺ in the ZnBip-CCR1 and ZnClTerp-CCR1 models, and contributes to the binding of CCL3^(5–70) in the ZnBip-CCR1 complex. Furthermore, we have previously shown that ZnBip and ZnPhe do not enhance ¹²⁵I-CCL5 binding, but rather compete with it, consistent with the fact that CCL5, like CCL3^(1–70), carries a nine amino acid long N-terminus, which competes with the allosteric modulators for the binding to Glu287^7.39, as corroborated by mutagenesis studies proving this residue be necessary for CCL5 activity.³²

To summarize, the transmembrane binding pocket of chemokine receptors, with Glu287^7.39 at its center, is occupied by the chemokine N-termini to varying degrees depending on the chemokine and truncation variants. Small allosteric ligands for chemokine receptors often use this Glu as anchor,¹⁹ and it is therefore around this residue that a competitive situation may occur between chemokines and small molecule ligands.

From Agonist to Inverse Agonist and the Role of the Tyrosine Network for Efficacy

Mutation Y113A^3.32 results in a shift from agonist to inverse agonist for ZnBip and ZnClTerp (Table 1), while having no effect on the allosteric action of ZnBip on CCL3^(5–70). The efficacy switch might result from an altered, conceivably deeper binding, of the metal ion chelator complexes in the transmembrane pocket that lock the receptor in an inactive conformation. Furthermore, one could argue that ZnBip is no longer able to establish a connection between helix 3 and helix 6 in Y113A^3.32, resulting in an inactivation of CCR1. This is in accordance with the earlier proposed activation mechanisms for GPCRs, suggesting a movement of the extracellular-facing part of TM-6 inward toward TM-3 for activation.⁵⁸ Interestingly, Tyr113^3.32 is one of four tyrosine residues at the bottom of the transmembrane binding pocket. The other three are the adjacent Tyr114^3.33, and Tyr255^6.51, and Tyr291^7.43. These positions have previously been shown to form an aromatic zipper of importance for a small molecule binding site in CXCR3, where a metal ion site was introduced between Asp^4.60 and His^3.29.⁵⁹ Upon binding of ZnBip via Zn²⁺ to this engineered metal ion site, an aromatic zipper consisting of the chelator 2,2′-bipyridine and Tyr^6.51, Phe^3.32, and Tyr^7.43 (in this order), pulls TM-6 toward TM-3, leading to the stabilization of the active conformation.⁵⁹ It is remarkable that these residues are also present in CCR1 and CCR5, which both are activated by metal ion chelator complexes.^32,34,35 However, in contrast to the site in CXCR3, the chelator in our CCR1 model binds centrally and above Glu287^7.39 and, instead of forming a zipper, the tyrosines are located below Zn²⁺. This may allow the tyrosines to substitute for each other in their interaction with Zn²⁺. Such redundancy would explain why mutation of either tyrosine alone has a relatively small effect. Furthermore, they might be of general importance for aromatic interactions of small molecules with the transmembrane receptor domain.

Clinical Relevance

In this study, we show that the action of allosteric modulators depends on the length of the chemokine N-terminus. To our knowledge, this is the first report of its kind, and it emphasizes that drug development programs for chemokine receptors and other peptide receptors, such as C3A and C5A receptors, or glucagon-like receptors (class B1 GPCRs) need to consider that differential post-translational ligand modification may influence the action of a drug candidate. To give an example; if chemokine receptor drug candidates were developed as antagonists for a long-variant chemokine, but, at the relevant sites in vivo, would meet a truncated chemokine variant, this could change their activity either in a biased manner, or even result in complete inactivity. This is further complicated by the difficulties involved in determining the exact amount and location of truncated chemokines in vivo. In the chemokine field in general, N-terminal truncation is often a ubiquitous and progressive set of events, and has been identified in more than half of all chemokines.⁴³ So far, it seems that the effect of chemokine truncations can only be determined on a case-by-case basis, and as a consequence, the same will hold true for the allosteric modulators in this GPCR group. Therefore, our study highlights that future drug development targeting chemokine (and other peptide) receptors should take this in vivo phenomenon into account on a pharmacological level to optimize the likelihood of finding compounds with promising clinical effects.

Experimental Section

Materials

Human CCL3^(5–70) was purchased from R&D Biotechne (MN, U.S.), and human CCL3^(1–70) was from Peprotech (NJ, U.S.). ¹²⁵I-CCL3^(5–70) was produced in-house, and ¹²⁵I-CCL3^(1–70) purchased from PerkinElmer (MA, U.S.). ZnCl₂, 2,2′-bipyridine, 4′-chloro-2,2′:6′,2″-terpyridine, and DMSO was purchased from Sigma-Aldrich and used without further purification. 6-Phenyl-2,2′-bipyridine (PhBip) was synthesized from 2,2′-bipyridine by reaction with phenyl lithium followed by oxidation with manganese oxide.⁶⁰ 2,2′:6′,3′′-Terpyridine (mTerp) and 2,2′:6′,4′′-terpyridine (pTerp) were synthesized by condensation of 3-(dimethylamino)-1-(pyridin-2-yl)prop-2-en-1-one⁶¹ with 3-acetylpyridine or 4-acetylpyridine, respectively.⁶² The highest concentrations of metal-ion chelator complexes were 20 mM for ZnBip, and 2 mM for ZnClTerp, and were made from 0.2 or 0.02 M ZnCl₂ in water, and 100 mM Bip or 10 mM ClTerp in DMSO, respectively, and were supplemented with water and 70% ethanol. The ratio of Zn²⁺/chelator was 1:2. Dilutions were made in water. The human CCR1 cDNA was kindly provided by Tim Wells (Serono Pharmaceutical Research Institute, Geneva, Switzerland). The promiscuous G protein Gα_Δ6qi4myr (abbreviated as G_qi4myr) was kindly provided by Evi Kostenis (University of Bonn, Germany) Myo[³H]inositol (PT6–271) and 125-iodine were purchased from PerkinElmer (MA, U.S.). Polyethyleimine (PEI) for transfection was purchased from Polysciences (PA, U.S.). CHO-K1 cells stably expressing CCR1 were kindly provided by Hans Rudolf von Lüttichau (University of Copenhagen, Denmark).

Molecular Biology

Receptor mutations were introduced by the polymerase chain reaction overlap extension technique or the QuikChange technique (Agilent Technologies, CA, U.S.) using wild-type (WT) CCR5. All reactions were carried out using Pfu polymerase (Stratagene, CA, U.S.) under conditions recommended by the manufacturer. The mutations were cloned into pcDNA3.1+ for use in inositol-trisphosphate (IP₃) and binding assays (Invitrogen, CA, U.S.). All constructs were verified by restriction endonuclease digestion and DNA sequencing (GATC Biotech, Germany).

Transfection and Tissue Culture

COS-7 cells were grown in DMEM with Glutamax (Invitrogen) supplemented with 10% fetal bovine serum (FBS), 180 U/mL penicillin and 45 μg/mL streptomycin at 37 °C in a 10% CO₂/90% air-humidified atmosphere. Transfection of 6 × 10⁶ COS-7 cells in a T175 culture flask was carried out by the calcium phosphate precipitation method.^63,64 Briefly, plasmid DNA (20 μg of receptor cDNA and 30 μg G_qi4myr for IP₃-assays, or 40 μg receptor cDNA for ¹²⁵I-CCL3-binding assays) was mixed with TE buffer (10 mM Tris-HCl, 2 mM EDTA-Na₂, pH 7.5) and 30 μL calcium chloride (2 M) to a total volume of 480 μL, and was then added to the same amount of HEPES buffered saline (280 mM NaCl, 50 mM HEPES, 1.5 mM Na₂HPO₄, pH 7.2). The mixture was allowed to precipitate for 45 min at room temperature, after which the precipitate and 300 μL chloroquine (2 mg/mL) in 10 mL culture media were added to the 6 × 10⁶ COS-7 cells seeded the day before. Transfection was stopped after 5 h by replacing to fresh medium, and cells were incubated overnight. HEK293 cells were grown in Dulbecco’s Modified Eagle Medium (DMEM), containing 1% GlutaMAX, and supplemented with 10% FBS and 1% penicillin (180 U/mL)/streptomycin (45 μg/mL), and incubated at 37 °C, 10% CO₂ and 95% air humidity. CHO K1 cells stably transfected with CCR1 were grown in RPMI1640 supplemented with 10% FBS and 1% penicillin (180 U/mL)/streptomycin (45 μg/mL) and incubated in 37 °C, 5% CO₂ and 95% air humidity. For transfection of HEK293 or CHO K1 cells (1 000 000 cells/well for HEK293 and 500 000 cells/well for CHO-K1) were seeded in pretreated 6-well tissue culture plates before a polyethylenimine (PEI) DNA transfection was performed.⁶⁵ For the cAMP assays, CHO-K1 cells stably expressing CCR1 were transfected with 1 μg of CAMYEL (YFP-Epac-RLuc). For the ß-arrestin-2-assays, HEK293 cells were transfected with 0.33 μg of human CCR1 receptor, 0.042 μg of Rluc8-Arrestin3-SP1, 0.8 μg of mem-citrine-SH3 and 0.8 μg of GRK2. DNA constructs were mixed with PEI (ratio 1:2 or 1:1.5 (DNA/PEI) for cAMP- and ß-arrestin assays, respectively) and nonsupplemented DMEM, incubated at room temperature (RT) for 15 min, before the sample was added dropwise to the cells. The following day, transfection was terminated by replacing the cell transfection-medium with fresh supplemented DMEM medium.

Inositol-Trisphosphate (IP₃) Assay

One day after transfection, COS-7 cells (35 000 cells/well) were incubated with [³H]myo-inositol (5 μL/mL, 2 μCi/mL) in 0.1 mL of media overnight in a 96-well plate. The following day, cells were washed twice in PBS and were incubated in 0.1 mL of Hank’s balanced salt solution (Invitrogen) supplemented with 10 mM LiCl at 37 °C in the presence of various concentrations of ligands for 90 min. Assay medium was then removed, and cells were extracted by the addition of 50 μL of 10 mM formic acid to each well, followed by incubation on ice for 30–60 min. The [³H]IPs in the formic acid cell lysates were thereafter quantified by adding YSi-poly-d-Lys-coated SPA beads. Briefly, 35 μL of cell extract was mixed with 80 μL of SPA bead suspension in H₂O (12.5 μg/μL) in a white 96-well plate. Plates were sealed, agitated for at least 30 min, and centrifuged for 5 min at 402 rcf. SPA beads were allowed to settle and react with the extract for at least 8 h before radioactivity was determined using a Packard Top Count NXT scintillation counter (PerkinElmer). All determinations were made in duplicate. These overall readouts have earlier been used effectively for CCR5, CXCR4, and other chemokine receptors.^33,34,66,67

cAMP-Assay

Two days after transfection, cells were resuspended in PBS with 1% glucose (5 mM), before they were aliquoted at 80 μL into a white 96-well culture plate (PerkinElmer). Subsequently, the cells were stimulated to produce cAMP by adding 10 μM forskolin. Five minutes after incubation with forskolin, Coelenterazine h (Nanolight Technologies, Pinetop, AZ) was added in a final concentration of 5 μM. The reaction was started 10 min after forskolin stimulation by the addition of ligands (concentrations were ranging from 1 pM to 0.1 μM). After a 30 min incubation at room temperature (RT), the inhibition of forskolin-stimulated cAMP by the activated Gα_i-coupled chemokine receptor was measured as luminescence (Rluc 485/40 nm and YFP 530/25 nm) on a LB 940 Mithras Multimode Microplate Reader (Berthold Technologies GmbH & CO. KG, Bad Wildbad, Germany).

β-Arrestin Assay

The ß-arrestin recruitment assay was performed as previously described.⁶⁸ In short, 2 days after transfection, cells were resuspended in PBS with 1% glucose (5 mM), before they were aliquoted at 85 μL into a 96-well black-white iso-plate (PerkinElmer). Subsequently, Coelenterazine h (Nanolight Technologies) was added in a final concentration of 5 μM and the reaction was started immediately after, by the addition of ligands (concentrations were ranging from 1 pM to 0.1 μM). After 30 min incubation at RT, the luminescence (Rluc 485/40 nm and YFP 530/25 nm) was measured on a LB 940 Mithras Multimode Microplate Reader (Berthold Technologies GmbH & CO. KG).

Ca²⁺-Assay

Two days prior to the assay, CHO-K1 cells stably expressing CCR1 were seeded in clear-bottom black 96-well plates (Thermo Fisher, Invitrogen) at 35 000 cells/well. One vial (77 mg) probenecid (Thermo Fisher, Invitrogen) was dissolved in 550 μL of 0.6 M NaOH; 1 vial of Fluo-4 AM (50 μg) (Thermo Fisher, Invitrogen) was dissolved in 23 μL of DMSO. Wash buffer was prepared consisting of 50 mL of HBSS + 20 mM Hepes, 1 mM CaCl₂, and 1 mM MgCl₂, supplemented with 250 μL of probenecid. Loading buffer contains wash buffer +0.2% Fluo-4 (e.g., 5.5 mL of wash buffer +11 μL of Fluo-4 solution). Upon start of the assay, the cell culture medium from the cell-plate was removed and 50 μL/well loading buffer was added and incubated in the dark at 37 °C for 1 h. During that time, a ligand plate (96-well v-bottom, Greiner bio-one, Austria) was prepared containing 5-fold the desired concentration of ligand diluted in wash buffer. Thereafter, cells were washed twice with 100 μL/well wash buffer. Finally, 100 μL/well wash buffer was added, and cell- and ligand-plates were placed into a Flexstation 3 (Molecular Devices, CA, U.S.), which adds ligands and measures fluorescence over 90 s.

Iodination of CCL3

Seventeen micrograms (ca. 2 nmol) of carrier-free CCL3 (RnD systems, Bio-Techne, MN, U.S.) was dissolved in 10 μL of iodination buffer (300 mM phosphate buffer pH 7.4). A 4 μL aliquot of 125-iodine (PerkinElmer, NEX033A, MA, U.S.) was added. Five microliters of a 3 μg/mL chloramine T solution in 300 mM phosphate buffer pH 7.4 was added six times at 1 min intervals while occasionally stirring. After 6 min, the reaction was stopped by the addition of 400 μL of water containing 0.1% trifluoroacetic acid. The reaction mixture was then purified on C18 column with an acetonitrile gradient from 20 to 80% over ca. 45 min.

Competition Binding

On the day after transfection, cells were transferred to 24- or 12-well culture plates. The number of cells seeded per well was determined by the apparent expression efficiency of the receptors, aiming for 5–10% specific binding of ¹²⁵I-CCL3. The number of cells thus ranged from 10 000 to 300 000. Two days after transfection, cells were assayed by competition binding for 3 h at 4 °C using 10–15 pM ¹²⁵I-CCL3 plus unlabeled ligand in 0.2 mL (24-well-plates) or 0.3 mL (12-well-plates) of 50 mM Hepes buffer, pH 7.4, supplemented with 1 mM CaCl₂, 5 mM MgCl₂, and 0.5% (w/v) bovine serum albumin. After incubation, cells were washed quickly two times with 4 °C binding buffer supplemented with 500 mM NaCl. Nonspecific binding was determined in the presence of 0.1 μM unlabeled CCL3. Determinations were made in duplicate.

Computational Modeling

A homology model of CCR1 (UniProt ID P32246) in complex with CCL3 was generated using the X-ray crystal structure of CCR5 in complex with 5P7-CCL5 (PDB 5UIW).²² The N- and C-termini of CCR1 not covered by the template were not considered during model generation, and the structural waters of CCR5 were omitted. The models were built using the Full Model Builder of ICM 3.8-7b (Molsoft L.L.C.) and subsequently refined through 200 steps of all-atom Monte Carlo-minimization. The best model was selected by having the lowest overall energy strain based on the “Protein Health” tool in ICM, which assesses the relative energy of each amino acid residue. The chelator metal ion receptor complexes were optimized through 5000 Monte Carlo global optimization steps, each involving a random move of one of four types (pseudo-Brownian move, change of internal variable, group of angles or loop conformation), followed by a local energy minimization (100 calls, gradient RMSD < 0.05, tolFunc = 6 consecutive steps without energy improvement), a calculation of the complete energy including surface and advanced electrostatic terms and, finally, acceptance or rejection of the move based on the energy and the temperature. The chelator, metal ion and side chains were left free to move, which resulted in chelator and metal ion receptor-mediated interactions in agreement with metal ion interaction geometries observed in experimental structures.

Data Analysis

Nonlinear regression concentration–response curves, EC₅₀ values, and IC₅₀ values were determined using GraphPad Prism 8 for binding and all signaling assays. If included concentrations of small molecule or chemokine ligands did not reach a plateau, the EC₅₀/IC₅₀ of the interaction was qualified as nondeterminable (no sign of sigmoidal plateau) or nearing saturation, in which case the Hill coefficient was set to 1 (or −1 for competitive binding) to determine the plateau. The values of efficacy, B₀, IC₅₀, and EC₅₀ of the CCL3 variants were compared within each assay format using the unpaired two-tailed t test.

Glossary

Abbreviations

CCR1

C–C-chemokine receptor 1

EAE

experimental autoimmune encephalomyelitis

FBS

fetal bovine serum

GAG

glycosaminoglycan

GPCR

G protein-coupled receptor

HBSS

Hanḱs balanced salt solution

RT

room temperature

TE

Tris-EDTA

ZnBip

Zn²⁺ in complex with 2,2′-bipyridine

ZnPhe

Zn²⁺ in complex with 1,10-phenantrholine

ZnTerp

Zn²⁺ in complex 2,2′:6′,2′′-terpyridine

ZnClTerp

Zn²⁺ in complex with 4′-chloro-2,2′:6′,2′′-terpyridine.

Author Present Address

^# Current author addresses for Matjaz Brvar: Helmholtz Zentrum München, Institute of Medicinal Chemistry, Schneiderberg 1B, 30167 Hannover.

The authors declare no competing financial interest.