Boronic Acids as Probes for Investigation of Allosteric Modulation of

Subscriber access provided by University of Guelph Library

Article

Boronic acids as probes for investigation of allosteric regulation of the chemokine receptor CXCR3 Viachaslav Bernat, Tizita Haimanot Admas, Regine Brox, Frank W Heinemann, and Nuska Tschammer ACS Chem. Biol., Just Accepted Manuscript • DOI: 10.1021/cb500678c • Publication Date (Web): 18 Sep 2014 Downloaded from http://pubs.acs.org on September 29, 2014

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

ACS Chemical Biology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

ACS Chemical Biology

Boronic acids as probes for investigation of allosteric modulation of the chemokine receptor CXCR3 Viachaslau Bernat1, Tizita Haimanot Admas1, Regine Brox1, Frank W. Heinemann2, Nuska Tschammer*1 1

Department of Chemistry and Pharmacy, Medicinal Chemistry, Emil Fischer Center, Friedrich Alexander University, Schuhstraße 19, 91052 Erlangen, Germany.

2

Department of Chemistry and Pharmacy, Inorganic Chemistry, Friedrich Alexander University, Egerlandstraße 1, 91058 Erlangen, Germany

[email protected]

KEYWORDS: Mutagenesis, chemokine receptor, CXCR3, allosteric modulation, biased signaling, boronic acid.

ACS Paragon Plus Environment

1


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 47

ABSTRACT The chemokine receptor CXCR3 is a G protein-coupled receptor, which conveys extracellular signals into cells by changing its conformation upon agonist binding. To facilitate the mechanistic understanding of allosteric modulation of CXCR3, we combined computational modeling with the synthesis of novel chemical tools containing boronic acid moiety, site-directed mutagenesis, and detailed functional characterization. The design of boronic acid derivatives was based on the predictions from homology modeling and docking. The choice of the boronic acid moiety was dictated by its unique ability to interact with proteins in a reversible covalent way, thereby influencing conformational dynamics of target biomolecules. During the synthesis of the library we have developed a novel approach for the purification of drug-like boronic acids. To validate the predicted binding mode and to identify amino acid residues responsible for the transduction of signal through CXCR3, we conducted a site-directed mutagenesis study. With the use of allosteric radioligand RAMX3 we were able to establish the existence of a second allosteric binding pocket in CXCR3, which enables different binding modes of structurally closely related allosteric modulators of CXCR3. We have also identified residues Trp1092.60 and Lys3007.35 inside the transmembrane bundle of the receptor as crucial for the regulation of the G protein activation. Furthermore we report the boronic acid 14 as the first biased negative allosteric modulator of the receptor. Overall, our data demonstrate that boronic acid derivatives represent an outstanding tool for determination of key receptor-ligand interactions and induction of ligand-biased signaling.


2

Page 3 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Introduction The G protein-coupled receptors (GPCRs) are natural allosteric proteins1–3. Allostery represents a mechanism by which even small molecules can exert profound effects on large proteins4. Allosteric modulators can stabilize different global conformations of a receptor, and thereby elicit countless functional responses4–6. In chemokine receptors, the class A GPCRs, allosteric modulators can disturb protein-protein interactions between a receptor and its endogenous chemokine ligands that have a size of 8-10 kDa. An outstanding example for the supremacy of allosteric modulators is aplaviroc. It alters the conformation of the chemokine receptor CCR5, rendering it unable to bind and support HIV-1 viral fusion and the binding of 125I-CCL3, but at the same time does not interfere with the binding and function of CCL57,8. The chemokine receptor CXCR3, which is mainly expressed on activated T cells, plays an important role in the regulation of the human immune system in response to different endogenous and exogenous stimuli9. CXCR3 binds chemokines CXCL9, CXCL10, and CXCL1110. Of these CXCL11 has the highest affinity and yields the highest maximal response10. The ligands elicit different types of functional response upon binding to CXCR311,12 and provide an example of ligand-biased signaling as a way to orchestrate the diverse functions of the receptor. Anomalies in the regulation and response of CXCR3 are associated with numerous pathologies including autoimmune diseases, cancer, vascular disease and transplant rejection; CXCR3 is thus an attractive pharmacological target13. Nevertheless, only one molecule (AMG487, Figure 1) from the number of high-throughput screening campaigns and extensive hit-to-lead optimization efforts has reached phase IIa clinical trials for rheumatoid arthritis and psoriasis, but did not surpass the treatment with placebo13,14. Although its metabolic profile was held responsible for the compound’s lack of


3


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 47

efficiency15,16, the complex pharmacology of chemokine signaling may also have contributed to the failure. With the aim to improve the understanding of the complex molecular mechanisms that govern the functions of CXCR3, we decided to create novel chemical tools for the investigation of these subtle mechanisms of allosteric modulation. We began with homology modeling of the CXCR3 structure and docking of a prototype high affinity small-molecule allosteric ligand, CHEMBL254083 (Figures 1 and 2). As such, the homology model of CXCR3 was constructed and predicated upon the crystal structure of the human chemokine receptor CXCR4 co-crystallized with 1T1t17 (PDB ID: 3odu). Identified binding pocket was addressed in the design of novel allosteric ligands, where we placed the boronic acid moiety into the well-characterized 8-azaquinazolinone scaffold18,19. We have chosen boronic acid group because of its unusual ability to form reversible covalent interactions with nucleophilic amino acid residues20,21. Recently reversible covalent ligands have received significant attention in drug development as they represent an intermediate class of biologically active molecules between non-covalent and covalent binders22. In particular, boronic acids are often used as reversible covalent inhibitors, e.g. for proteolytic enzymes23–25. The unique binding properties of boron arise from its ability to easily switch between hybridization states sp2 (trigonal geometry, trivalent) and sp3 (tetrahedral geometry, tetravalent26). Boronic acids can interact with O- and N-nucleophilic amino acid residues of proteins (Ser27, Thr, Tyr27,28, Lys21), N-protonated side chains (Arg, His29) and neutral hydrogen bond donors (Asn30), forming up to three covalent bonds to the protein target21. These unusual features suggest that the boronic acid moiety is an attractive tool for the exploration and stabilization of protein conformations and thus its function. Nevertheless, the literature on the GPCR ligands that contain this moiety is scarce and mainly consists of nonsystematic application of the group31–33.


4

Page 5 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


To identify the anchor amino acid residue(s) used by boronic acid derivatives, we mutated the residues within the putative allosteric binding pocket. The receptor mutants and boronic acid derivatives were extensively characterized in the allosteric radioligand (RAMX3)34 displacement assay, [35S]GTPγS incorporation and β-arrestin 2 recruitment assays. The use of allosteric radioligand is superior to radioactive chemokine displacement, because it enables one to perform the competitive binding assay between the allosteric radioligand and other novel allosteric modulators, which target the same binding pocket. Furthermore, it is important to note that structure-activity relationships (SAR) that govern classic orthosteric effects do not apply to allosteric sites. The SAR studies on allosteric modulators need to differentiate chemical modifications, which influence compound affinity from those, which influence the cooperativity exhibited towards the orthosteric ligand, as the two properties are not correlated1,35. The use of boronic acids as chemical probes led us to the discovery of a second allosteric binding site in CXCR3 and to the identification of the residue Lys300 as an important anchor for the boronic acid moiety that dictates functional properties of these derivatives. Importantly, our series of boronic acid derivatives resulted in the first biased negative allosteric modulator of CXCR3, which preferentially inhibited the CXCL11-mediated recruitment of β-arrestin 2 to CXCR3, but not the activation of G proteins. Our data demonstrate that boronic acid derivatives represent an outstanding tool for determination of key receptor-ligand interactions and induction of ligand-biased signaling.

RESULTS AND DISCUSSION The main aim of our efforts was to explore the delicate molecular mechanisms of allosteric modulation of CXCR3 with the combination of computational modeling, de novo synthesis of chemical tools containing boronic acid moieties, site-directed mutagenesis and detailed


5


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 47

functional characterization followed by the analysis of biased signaling. We created novel potent NAMs based on the boronic acids, which are scarcely used as moieties in GPCR ligands. With the aid of these boronic acid derivatives we were able to establish the existence of a second allosteric binding pocket in CXCR3 and to identify residues Trp1092.60 and Lys3007.35 inside the transmembrane bundle of the receptor as crucial for the regulation of the G protein activation. We also report the boronic acid 14 as the first biased NAM of CXCR3. Synthesis and purification of boronic acid derivatives. Allosteric ligands were synthesized according to the strategy that was developed for 8-azaquinazolinone derivatives19 and applied previously in our laboratory34. For boron-free compounds 20 and 21 the primary amine 1 was reductively alkylated using corresponding substituted benzaldehyde and triacetroxyborohydride. For the phosphonic acid derivative 19 the primary amine 1 was added to diethyl vinylphosphonate. Resulting secondary amines 9-11 were subsequently acylated by activated 2-[4-fluoro-(3-trifluoromethyl)phenyl]acetic acid (Scheme 1). For the synthesis of boronic acid derivatives 12-18 we used the trifluoroborate moiety as a boronic acid protecting group. The moiety was introduced into the 8-azaquinazolinone scaffold according to the protocol for reductive amination developed by Molander et al.36. The required trifluoroborate-substituted aromatic aldehydes were purchased or prepared from commercially available boronic acids by treatment with potassium hydrogen difluoride in methanol. Secondary amines 2-6 were acylated according to the protocol that was used for boron-free derivatives. Trifluoroborate moieties were converted to boronic acids by the treatment with trimethylsilyl chloride in aqueous acetonitrile. The application of the standard protocol, which employs borane-pyridine complex as the reducing agent, to the reductive amination of acrolein derivative (potassium (E)-trifluoro(3-oxoprop-1-en-1-yl)borate), resulted in the reduction of the conjugated C=C double bond. The obtained alkyltrifluoroborate 8 was then acylated and converted to alkylboronic acid 18. For the


6

Page 7 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


synthesis of alkenylboronic acid 17 we used an alternative reductive amination approach with sodium borohydride that preserved the C=C double bond. Further transformations of the secondary amine 8 were analogous to 18. The syntheses of building blocks and all the intermediates are described in the Supporting Information. For the purification of boronic acid derivatives 12-18 we tested different approaches (see Supporting Information) and found that a combination of solid extraction with dry-column vacuum chromatography (DCVC37) was the most suitable. The latter technique combines the robustness of TLC grade silica gel (particle size 2–25 µm vs. 35–75 µm for common flash silica), scalability, and operational simplicity. According to our optimized protocol, in the first step washing of the crude sample, which is placed on the top of the silica gel column, with polar aprotic solvent (acetonitrile) extracts all boron-free organic impurities. Because of the high retentivity of boronic acids, they remain on the stationary phase until gradient elution with the mixture that contains the protic solvent begins (typically from 2.5 to 10% ethanol in chloroform). In the second step the gradient elution chromatographically separates boron-containing derivatives according to their polarity. The detailed protocol of purification is described in the Supporting Information. CXCR3 homology model and proposed binding modes of allosteric modulators. Despite extensive structure-activity data on the small-molecule modulators of CXCR3 (642 bioactivity entries in ChEMBL database*), only few reports on the mode of interaction between the receptor and its small molecule ligands emerged38,39. To fill in this gap and to provide the basis for the rational design of allosteric modulators we created a homology model of the CXCR3 structure. The homology model of CXCR3 was constructed with aid of the Modeller 9.10 software40, based on the X-ray structure of CXCR4 co-crystallized with


7


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 47

small-molecule antagonist IT1t17. The QMEAN41 scoring function was chosen for assessment of modeling quality. Local optimization of the polypeptide backbone, protonation states and geometry of amino acid side chains was performed by Yasara NOVA42 and Protein Local Optimization Program (PLOP)43 software. Initial docking of the small-molecule ligand of CXCR3 (CHEMBL254083, Figure 1) was performed by Autodock Vina44; subsequent optimization of the binding pocket and docking poses was done with PLOP. A detailed procedure of the model construction and refinement is described in the Supporting Information. We identified two putative docking poses that allosteric modulators of CXCR3 can adopt upon binding (Figure 2A, B). According to the binding mode A (Figure 2A), the heterocyclic moiety of the small molecule ligand forms face-to-face π-π interaction with Trp1092.60, the amide oxygen interacts with an NH in the backbone of the extracellular loop 2 (ECL2), and the fluorinated aromatic ring points to the hydrophilic subpocket formed by Gln2195.42, Asp1864.60, and Asn1323.33, with Phe1313.32 lining the bottom of the subpocket and dividing it from another cluster of hydrophilic amino acids. The sulfone moiety of CHEMBL254083 interacts with the subpocket formed by the amino acid residues on TM6 and TM7: Tyr2716.51, Lys3007.35, Ser3047.39, and Tyr3087.43 (nucleophilic ‘hot spot’). The cyanophenyl substituent is directed between TM1 and TM7, forming mainly Van der Waals contacts. According to the binding mode B (Figure 2B), the orientation of the allosteric ligand is inverted relative to the central amide moiety: the heterocyclic core is accommodated in the subpocket formed by Gln2195.42, Asp1864.60, Leu2155.38, Gln204ECL2, and Tyr205ECL2. The fluorinated aromatic moiety forms face-to-edge π-π interaction with Trp1092.60 and His202ECL2, as well as Van der Waals contacts with Tyr601.39, Asp1122.63, and Phe47N-terminus. The cyanophenyl substituent forms face-to-edge π-π interaction with Phe1313.32 and nonpolar contacts with backbone of TM3, Trp1092.60, and Trp117ECL1. Notably, the orientation of


8

Page 9 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the polar sulfone moiety in the binding mode B is similar to the binding mode A, which suggests an important role of corresponding subpocket in anchoring allosteric modulators of CXCR3. Compared to the known structures of chemokine receptors co-crystallized with allosteric modulators, the identified binding modes of CXCR3 modulators are closer to the binding mode of maraviroc in CCR5

45

than IT1t in CXCR4. This may be explained by the size and

polarity of our templating ligand, which are closer to maraviroc than to IT1t. Another factor is the difference in amino acid sequence that determines the overall shape of the binding pocket. For example, both CCR5 and our homology model of CXCR3 have an aromatic binding subpocket formed by Phe1313.32, Phe1353.36, Trp2686.48, Tyr2716.51 (numbering according to the CXCR3 sequence), whereas CXCR4 has shallower binding pocket (Supplementary Figure S4). Generally, CHEMBL254083 (NAM of CXCR3) was predicted to bind slightly higher in its CXCR3 binding pocket than maraviroc in CCR5, but deeper than IT1t in CXCR4. In the ligand-free homology model of CXCR3 Lys3007.35 forms a salt bridge with Asp2977.32 (Figure 3), which is located one helical turn closer to the extracellular end (previously suggested by Rosenkilde et al.46). In the ligand-bound state Lys300 interacts with the negatively polarized moiety of the small molecule and Tyr2716.51. According to our binding hypotheses, the boronic acid moiety that replaced the sulfone group in CHEMBL254083 should interact with the identified nucleophilic hot-spot, where Lys300 plays a crucial role. To validate the predicted binding poses of allosteric ligands, we mutated three amino acid residues in the putative binding pocket: Lys3007.35, Gln2195.42, and Trp1092.60 and subjected them to detailed biological characterization. Allosteric radioligand RAMX3 displacement and proposed binding mode of boronic acid derivatives. The ability of novel ligands to displace the allosteric radioligand RAMX334


9


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 47

was evaluated in the binding assay with the membrane preparations of HEK293 cells, which expressed the wild type or the mutants of CXCR3. RAMX3 was developed in our group with the intention to measure the affinity of the ligands for their allosteric binding pocket in a competitive manner34. This is an advantage, because the affinity of the allosteric modulators is not influenced by the cooperativity between the allosteric ligand and orthosteric chemokines (e.g., [125I]–CXCL11 or [125I]–CXCL104,5). Note: the probe dependent antagonism has been well-described for CCR5, where the small molecule aplaviroc blocks the binding of 125I–CCL3 but not 125I–CCL57,8. In a good agreement with both proposed binding modes, most of the synthesized boronic acid derivatives had a high affinity to the wild type CXCR3 (e.g. pKb of 8.03 for 12) and completely displaced the allosteric radioligand (Table 1). Three compounds (13, 14, and 16) were unable to suppress RAMX3 binding completely, despite close structural similarity to the other members of the series (Figure 4, Table 1). This implied an allosteric interaction with a radioligand and, hence, a second allosteric binding site in CXCR3. As a consequence we applied the ternary complex model of allosterism to analyze the radioligand binding data, where the pKb is the equilibrium dissociation constant of modulator binding and α denotes cooperativity factor4 (Table 1). Values of α > 1 denote positive cooperativity, whereas α < 1 denotes negative cooperativity. Values of α approaching zero are indistinguishable from competitive antagonism. When α approaches zero, the Kb value approaches the Ki value4. An α value equal to 1 denotes an allosteric interaction that results in unaltered ligand affinity. The allosteric ligands 12, 15, 17, and 18 demonstrated the behavior undistinguishable from competitive antagonism with α equaling or approaching zero at the wild type CXCR3. Derivatives 13, 14, and 16 showed the α values of 0.33, 0.29 and 0.26, that indicated weaker negative cooperativity toward the radioligand RAMX3. The existence of the second allosteric binding pocket was strengthened by the mutation of Q219E that led to incomplete


10

Page 11 of 47

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


displacement of the allosteric radioligand RAMX3 by the majority of synthesized ligands (Table 1). Only the alkylboronic acid 18 completely displaced RAMX3. The possibility that the observed profiles in binding assays are caused by allosteric cooperativity across CXCR3 homodimers is currently under investigations. The data from the site-directed mutagenesis indicated that the affinity of cRAMX3 (‘cold’ RAMX3) did not change significantly upon mutations K300M and K300R, but decreased for 8-fold upon mutation Q219E (pKb value 7.96 (wild type) vs. 7.06, one-way ANOVA, p

Boronic Acids as Probes for Investigation of Allosteric Modulation of

Recommend Documents