Subscriber access provided by UNIVERSITY OF THE SUNSHINE COAST
Article
Rational Design of an Anticalin-Type Sugar-Binding Protein Using a Genetically Encoded Boronate Side Chain Selvakumar Edwardraja, Andreas Eichinger, Ina Theobald, Carina Andrea Sommer, Andreas J. Reichert, and Arne Skerra ACS Synth. Biol., Just Accepted Manuscript • DOI: 10.1021/acssynbio.7b00199 • Publication Date (Web): 22 Sep 2017 Downloaded from http://pubs.acs.org on September 23, 2017
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 Synthetic 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 15
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 Synthetic Biology
Rational Design of an Anticalin-Type Sugar-Binding Protein Using a Genetically Encoded Boronate Side Chain Selvakumar Edwardraja,† Andreas Eichinger, Ina Theobald, Carina Andrea Sommer, Andreas J. Reichert and Arne Skerra* Munich Center for Integrated Protein Science (CIPS-M) and Lehrstuhl für Biologische Chemie, Technische Universität München, Emil-Erlenmeyer-Forum 5, 85354 Freising (Weihenstephan), Germany ABSTRACT: The molecular recognition of carbohydrates plays a fundamental role in many biological processes. However, the development of carbohydrate-binding reagents for biomedical research and use poses a challenge due to the generally poor affinity of proteins towards sugars in aqueous solution. Here, we describe the effective molecular recognition of pyranose monosaccharides (in particular, galactose and mannose) by a rationally designed protein receptor based on the human lipocalin scaffold (Anticalin). Complexation relies on reversible covalent cis-diol boronate diester formation with a genetically encoded L-boronophenylalanine (Bpa) residue which was incorporated as a non-natural amino acid at a sterically permissive position in the binding site of the Anticalin, as confirmed by X-ray crystallography. Compared with the metal-ion and/or avidity-dependent oligovalent lectins that prevail in nature, our approach offers a novel and promising route to generate tight sugar-binding reagents both as research reagents and for biomedical applications. KEYWORDS: amber suppression, borocalin, boronophenylalanine, carbohydrate recognition, non-natural amino acid, protein design
INTRODUCTION Sugar molecules constitute key components in a variety of biological as well as pathological events including devastating diseases like diabetes, HIV/AIDS and cancer. For example, HIV infection is mediated by virus glycoprotein binding to cell surface receptors1 while tumor progression involves glycan-mediated processes.2 The selective recognition, under physiological conditions, of oligosaccharides and even simple sugars bears significant potential for the development of innovative research tools, medical diagnostics, therapeutic agents as well as vectors for targeted drug or gene delivery.3, 4 Beyond their important roles in biology as part of the metabolism and for mechanical support, carbohydrates exert crucial functions in cell-cell interaction, cell adhesion, cell trafficking as well as signal transduction and immune response. Most often these physiological processes are mediated through protein-oligosaccharide interactions. Moreover, glycan structures that coat the surfaces of bacteria, fungi, parasites and viruses are crucial both for disease transmission via host receptors and for shielding these pathogens from the immune system. However, the nature of carbohydrate antigens, regarding their configurational and physicochemical properties, poses a challenge for high affinity binding by proteins.5 In fact, monosaccharides share chemical structures dominated by hydroxyl groups, with few distinct functional features, whereas the connections in oligosaccharides can
p. 1
be remarkably complex. On the other hand, their composition is not precisely encoded at the genetic level but controlled by the enzymatic activities of various glycoside transferases during biogenesis.6 Consequently, from the perspective of molecular recognition targeting carbohydrate antigens in aqueous solution is a daunting task.7, 8 Their high content of hydroxyl groups makes them hydro-mimetic and difficult to differentiate from the solvent. This chemical camouflage hampers the tight binding of glycans by protein receptors, including natural lectins.9 Despite increasing progress towards understanding protein-carbohydrate interactions,10 the precise nature of the forces that govern complexation of sugars by protein receptors are not fully understood. Sugar binding by lectins takes place at shallow grooves on the protein surface and is often mediated by calcium ions.11, 12 Selectivity is achieved by a network of hydrogen bonds to the sugar hydroxyl groups, involving direct metal ion contacts, van der Waals packing as well as C-H/π interactions with aromatic amino acid side chains.13, 14 Generally, lectin-carbohydrate interactions are relatively weak, with dissociation constants in the range of 10-3 to 10-4 M for an individual complex. To achieve satisfactory strength of interaction in a biological system, lectins are usually linked to oligomerization domains and utilize the avidity effect.15 Antibodies directed against carbohydrate antigens show a different mode of binding but often have
ACS Paragon Plus Environment
ACS Synthetic Biology
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
low affinity, depending on the type of sugar antigen, and have weak specificity, which explains the few documented applications.16, 17 Thus, there is clear need for an alternative platform to generate practically useful sugar-binding proteins based on advanced protein engineering. Anticalins are a novel class of binding proteins generated by combinatorial protein design from the small (~180 amino acids) lipocalin scaffold, which provides a robust βbarrel fold with four structurally hypervariable loops at the open end that shape the cup-like binding site. Anticalins have been successfully engineered to recognize all major classes of antigens with picomolar affinities, including proteins, peptides and small molecule haptens, hence providing a viable alternative to the much bigger and more complex immunoglobulins.18-20 The human lipocalin 2 (Lcn2) has proven particularly effective as a scaffold; corresponding Anticalins offer high tolerability for the human body and are already subject to clinical trials. To enable Anticalins for the recognition of sugar ligands, and to overcome the camouflage effect in aqueous solution, we endeavored the incorporation of boronic acid, which is known to form reversible covalent complexes with cis-diols as they usually occur in saccharides and their derivatives.21 This type of functional group provides an ideal reagent for selective molecular recognition and has led to the development of boronic acid-based small-molecule sensors.22 However, while the ability of boronic acids to react with sugars has been known for over hundred years, application to carbohydrate recognition remained limited to the development of synthetic receptor molecules.4 Recently, the possibility of site-specific incorporation of boronic acid as an artificial side chain into recombinant polypeptides was opened via expansion of the genetic code. In the initial example, an amber stop codon was introduced into the reading frames for the Z-domain of staphylococcal protein A and for T4 lysozyme. This nonsense codon was then suppressed in Escherichia coli by action of a foreign suppressor tRNA in combination with an engineered aminoacyl-tRNA-synthetase (aaRS) specific for the non-natural amino acid L-p-boronophenylalanine (Bpa), which was supplemented to the culture medium.23 This approach led to detectable binding of glucamine, an amino derivative of the acyclic sugar alcohol sorbitol, but did not enable complexation of biologically more relevant cis-diol pyranoses, even after attempting to evolve a Bpamodified single-chain antibody variable fragment (scFv) for such activity.24 We hypothesized that the lipocalin scaffold, with its deep ligand pocket formed by four structurally hypervariable loops,20 provides a more appropriate protein format to incorporate a functionally active boronate group for the binding of sugar compounds. Therefore, we devised a two-step strategy: first, rational design of a hybrid protein receptor equipped with a boronate side chain to provide basic chemical affinity towards cis-diol compounds and, second, combinatorial reshaping of the loop region, like in conventional Anticalin technology,25 to implement additional non-covalent interactions from neighboring
p. 2
Page 2 of 15
amino acid side chains within the ligand pocket, thus boosting affinity and conveying molecular specificity towards the sugar. To accomplish the first step, we describe here the development of a boronated Anticalin, dubbed "Borocalin", which carries Bpa in its ligand cavity and can effectively bind pyranose sugars (Scheme 1).
RESULTS AND DISCUSSION Design and Construction of Boronated Anticalins. The crystal structure of Lcn2 (PDB ID: 1L6M)26 was screened in silico for sites that sterically allow the incorporation of the structurally extended Bpa side chain, also taking into consideration the additional space needed for a resulting sugar adduct, resulting in the positions of Leu36 and Leu94 (Figure 1A,B). To site-specifically introduce Bpa at each of these positions into the recombinant wild-type Lcn2 (wtLcn2) produced in E. coli we applied the amber suppression strategy by utilizing an orthogonal pair of the amber suppressor tRNA from M. jannaschii and a tyrosyl-tRNA synthetase from the same host engineered for specificity towards Bpa23, all encoded on a oneplasmid expression system.27 Therefore, the codons at position 36 and 94, respectively, of the cloned wtLcn2 gene were mutated to the amber stop codon (UAG) and both mutant genes were cloned along with the OmpA signal sequence at the N-terminus to achieve periplasmic secretion while the Strep-tag II was appended at the Cterminus to facilitate detection as well as Strep-Tactin affinity purification.28 Scheme 1. Strategy for the development of Borocalins that can tightly and specifically bind sugar compounds in aqueous solution.
(i) Rational implementation of a covalent sugar-binding site by modelling L-p-boronophenylalanine (Bpa) in the cavity of human lipocalin 2 (Lcn2); (ii) site-directed incorporation of Bpa into the recombinant Lcn2 via bacterial in vivo amber suppression using an engineered aminoacyl-tRNA-synthetase (aaRS) and suppressor tRNA pair; (iii) rational/combinatorial engineering of the surrounding residues to optimize noncovalent (secondary) interactions with the sugar ligand within the binding site; (iv) complex formation between the resulting Borocalin and a target sugar is driven via reversible cyclic boronate diester formation with the cis-diol moiety, thus providing basic chemical affinity, whereas a suitably
ACS Paragon Plus Environment
Page 3 of 15
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 Synthetic Biology
shaped surrounding loop region of the lipocalin scaffold contributes additional non-covalent interactions to further enhance affinity and provide binding specificity toward a particular mono- or oligosaccharide of biomedical relevance.
The mutant proteins, Lcn2(36Bpa) and Lcn2(94Bpa), were produced in E. coli BL21 via recombinant gene expression in the presence of 1 mM 4-borono-D/Lphenylalanine (see Supplementary Materials and Methods). Western blot analysis using a Strep-Tactin/alkalinephosphatase (AP) conjugate revealed expression of the full length mutant proteins and also confirmed production exclusively in the presence of the non-natural amino acid (Figures S1A and S1B). The recombinant proteins were purified from the periplasmic cell extract of E. coli by Strep-Tactin affinity chromatography and appeared pure in SDS-PAGE (Figure S2A). Furthermore, size exclusion chromatography (SEC) revealed monodisperse size of 21 kDa, indicating a correctly folded and fully monomeric globular protein (Figure S2B). Yields after purification were 1.4 mg L-1 for Lcn2(36Bpa) and 0.5 mg L-1 for Lcn2(94Bpa), compared with 2.8 mg L-1 for wtLcn2 when expressed from the same vector. Successful incorporation of Bpa was verified by electrospray ionization mass spectrometry (ESI-MS), indicating masses of 21813.9 Da for Lcn2(36Bpa) and of 21814.9 Da for Lcn2(94Bpa) as expected (each calculated: 21814.6 Da) (Figure S3). Both recombinant proteins were tested for binding activity towards cis-diol sugars by means of Tyr/Trp fluorescence titration. Initially, the linear sugar alcohol sorbitol was employed, which has well documented affinity towards boronic acids, e.g. KD ≈ 2.7 mM for phenylboronic acid (PBA) in aqueous buffer at pH 7.4,29 and had served as model ligand for Bpa in previous studies.24
A
B Bpa36 Bpa94
Trp125
Bpa36
4.4 Å
Asn134
Figure 1. (A,B) Models of two positions suitable for the incorporation of Bpa (depicted as CPK model in the trihedral form of boronic acid) within the ligand pocket of the Lcn2 scaffold. Manual docking was performed with Bpa on the Lcn2 crystal structure (PDB ID: 1L6M) using Chimera 30 software. Among 75 sites analyzed in total, only the positions Leu36 (A) and Leu94 (B) appeared suitable for exchange with Bpa. (C) Model of the double mutant 36Bpa-
p. 3
Remarkably, Lcn2(94Bpa) showed a strong increase in fluorescence upon titration with sorbitol at pH 7.4 if compared with wtLcn2, revealing binding activity with a KD of 40.2 ± 2.5 mM (Figure 2A). This finding demonstrated the successful incorporation of the artificial boronate side chain into Lcn2(94Bpa) in an orientation that allows binding of a linear sugar alcohol compound. However, no corresponding fluorescence change was detectable for Lcn2(36Bpa) and, also, no binding signal could be detected for the cis-diol pyranose D-galactose with either engineered lipocalin (Figure 2E). Rational Design and Engineering of the Borocalin Sugar-Binding Site. To better understand the structural implications of the boronate side chain at the two different positions in the engineered Lcn2, the models of both mutant proteins (cf. Figure 1A,B) were extended to the complexes with cis-diol pyranose sugars such as galactose, or the disaccharide lactose, in order to visualize sterical accessibility of the lipocalin cavity and possible modes of chemical complex formation. As result, the ligand pocket of Lcn2(36Bpa) was found to provide sufficient space to encompass both types of carbohydrate and, hence, was chosen as more promising basis for further rational engineering compared with Lcn2(94Bpa) despite its missing sugar-binding activity in the first place. Notably, boronic acids can react with side chains of amino acids that are capable of donating an electron pair, in particular serine, histidine and lysine.31 Indeed, the structural model of Lcn2(36Bpa) indicated a potential intramolecular and, hence, entropically favored interaction between the side chains of Lys134 and Bpa36 (Figure S4A) which could have interfered with complexation of sorbitol in the experiment mentioned above. Consequently, Lys134 was substituted with the shorter Asn side chain, which lacks a free amino group (Figure 1C), resulting in the mutant Lcn2(36Bpa/K134N) (Figure S4A-C and S3C). This version, dubbed 36Bpa-N, showed clear fluorescence increase upon titration with sorbitol at pH 7.4, even though the curve lacked saturation behavior and could not be fitted according to the Law of Mass Action (Figure 2B). Introduction of Tryptophan as Spectroscopic Probe in Proximity to the Sugar-Binding Site. In the fluorescence titration assay, the major signal change arises from alterations in the polar environment of Trp side chains; hence, we inspected the vicinity of Bpa in the structural models of both Borocalins.
C 4.8 Å
NW highlights the substituted side chains Asn134 and Trp125 in the vicinity of Bpa36.
Indeed, residue Trp79 appeared in close distance of
