Facile Synthesis of Dimeric Thioether–Macrocyclic Peptides with

May 1, 2018 - https://chemport.cas.org/services/resolver?origin=ACS&resolution=options&coi=1%3ACAS%3A528%3ADC%252BC2sXksFSiu7k% ...
0 downloads 0 Views 4MB Size
Subscriber access provided by UCL Library Services

Communication

Facile synthesis of dimeric thioether-macrocyclic peptides with antibody-like affinity against Plexin-B1 Nasir Bashiruddin, Yukiko Matsunaga, Masanobu Nagano, Junichi Takagi, and Hiroaki Suga Bioconjugate Chem., Just Accepted Manuscript • DOI: 10.1021/acs.bioconjchem.8b00219 • Publication Date (Web): 01 May 2018 Downloaded from http://pubs.acs.org on May 2, 2018

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 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 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.

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 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

Bioconjugate Chemistry

Facile synthesis of dimeric thioether-macrocyclic peptides with antibody-like affinity against Plexin-B1 Nasir K. Bashiruddin†, Yukiko Matsunaga‡, Masanobu Nagano†, Junichi Takagi‡*, Hiroaki Suga†* †

Department of Chemistry, Graduate School of Science, The University of Tokyo, Tokyo 113-0033, Japan



Laboratory of Protein Synthesis and Expression, Institute for Protein Research, Osaka University, Osaka 565-0871, Japan

ABSTRACT: Macrocyclic peptides have gained increasing attention due to their ease of discovery through various in vitro display platforms as well as their potential in possessing favorable properties of both small molecule and antibody drug classes. It is well known that the avidity achieved through the bivalent binding mode of antibodies gives rise to their slow dissociation rates and thus high potency as drug molecules. Here, we report the synthesis of dimeric thioether-macrocyclic peptides through a branched synthesis approach allowing for synthesis of dimeric peptides in a comparable number of steps as monomers and tunability of linker lengths from 30 Å to 200 Å. Applying this method to synthesize dimers of a model PlexinB1-binding macrocyclic peptide showed close to 300-fold increases in their apparent binding affinity, bringing the KD down from 8 nM to 30 pM as well as affording improved biological activities when compared to their monomeric counterparts. These enhancements demonstrate that this is a simple synthetic strategy to harness the benefits of bivalence that antibodies naturally possess.

In vitro display platforms have made possible the selection of macrocyclic peptides from libraries with diversities in the trillions generated from random DNA templates1. With the advent of these technologies, numerous macrocyclic peptides have been reported with high affinities and bioactivity against various pharmaceutically relevant proteins1. These macrocyclic peptides also share many beneficial properties of both small molecule and antibody drug classes. Although macrocyclic peptides, such as those selected via the RaPID (Random Non-standard Peptide Incorporated Discovery) system, can possess favorable drug-like properties including high peptidase stability and target selectivity through the incorporation of non-proteinogenic structures1, they differ from antibodies in that they are monovalent. The avidity gained from bivalence of IgGs begets their slow dissociation rates when binding their targets resulting in superior biological activity2,3. Indeed, several studies have shown decreases up to 2 to 3 orders of magnitude in binding affinity, predominantly originating from slow dissociation rates, when IgGs are separated into individual variable fragments4–6. This utilization of bivalency is a clever evolutionary workaround whereby spending double the energy requirement in the biosynthesis of an extra variable domain results in a gain of over 100-fold efficiency in the end product. It would therefore be convenient to be able to readily harness the affinity gained from bivalence with chemically synthesized macrocyclic peptides as well. Previous reports of synthesizing dimeric macrocyclic peptides has involved synthesizing and purifying a monomer cyclic peptide containing a free C-terminal cysteine which is subsequently used to fuse two monomers together via commercially available dimaleimide PEG-linkers7. This dimerization step requires very high peptide concentrations as well as working with low microliter scale volumes. Monitoring of the reaction progression results in losing

non-negligible quantities of peptide as does the requirement of an additional purification step resulting in reduced yields. Further, remaining unreacted monomers, disulfide-dimers and PEG-linked monomers are formed which can be difficult to separate via HPLC. In this study, we utilize a branched peptide synthesis approach to generate dimeric thioether macrocyclic peptides in one step as opposed to dimerizing pre-synthesized and purified monomer macrocyclic peptides. We have demonstrated this synthesis approach by generating dimers of a PlexinB1-binding macrocyclic peptide to show their enhancements in biological activity compared to the monomer8. Synthesis of thioether macrocyclic peptides is initiated by solidphase peptide synthesis (SPPS) using Fmoc-protected amino acids followed by N-terminal chloroacetylation via choloroacetyl-NHS8. Subsequently, the peptides are deprotected, ether precipitated and dissolved in a DMSO/water mixture for the macrocyclization step. We have observed when synthesizing macrocyclic peptides that it is possible to carry out the macrocyclization reaction in exceedingly low volumes, which correspond to theoretical concentrations on the order of 10-2 molar, and still yield mainly intramolecular macrocyclization products with little or no observable intermolecular thioether formation (Figure S1). These concentrations equate to an average distance on the order of only 10-100 Å between each peptide in solution. Thus, we hypothesized that two linear peptides connected by a flexible linker within the same distance range would still properly self-macrocyclize, which would make possible the branched synthesis of dimeric peptides in one step. For flexible linkers we chose commercially available Fmoc-NH-PEG5-COOH (Figure 1A, S2). Previous reports of synthesizing dimeric peptides through a branched approach have used Fmoc-lysine(Fmoc) as the C-terminal node 9–13. In this study, we chose Fmoc-2,3 diaminopropionic acid(Fmoc)-OH for use as

ACS Paragon Plus Environment

Bioconjugate Chemistry 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

Figure 1. Synthesis scheme and structural information of peptides synthesized in this study. (A) General synthesis scheme of homodimeric thioether macrocyclic peptides via branched synthesis (B) 3D stick model of monomeric PB1m6 (taken from PDB ID: 5B4W) (C) 3D structure of human IgG for size comparison. Stars indicate complementarity determining regions (CDRs). (PDB ID: 1IGT) (D-H) Computer generated 3D stick models of PEG-linked dimers synthesized in this study based on the previously reported crystal structure of PB1m6. PEG linkers were rendered fully-extended for length comparison (E=PB1d6P44, F=PB1d6P32, G=PB1d6P22, H=PB1d6P10, I=PB1d6P2). Panels B-H are illustrated to scale. (I) Structural formula of PB1m6 showing the site of dimer linkage attachment. (J) Composition of dimers and the approximate lengths of the total linker regions.

the initial C-terminal node simply due to its greater symmetry between the primary amino groups and its commercial availability (Figure1, S3). In addition, since this method essentially doubles the quantity of peptide being synthesized per coupling resin weight, we used half of the usual quantity of coupling resin (Figure S2, methods). We have previously reported a selection campaign against the transmembrane receptor Plexin-B1(PlxnB1)8. This selection

produced a thioether-macrocyclic peptide, PB1m6, which binds to PlxnB1 and allosterically inhibits its interaction with the axonal guidance factor, Semaphorin4D (Sema4D). In the study, PB1m6 was able to block the PlxnB1-Sema4D interaction in living cells resulting in the complete absence of Sema4D-dependent growth cone collapse in PlxnB1 expressing cells. We therefore used PB1m6 for dimerization studies due to it being readily characterizable in our hands (Figure 1, S2). The aforementioned branched synthesis

ACS Paragon Plus Environment

Page 2 of 5

Page 3 of 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

Bioconjugate Chemistry

Figure 2. Biological activity of PB1m6 -based dimeric cyclic peptides (A) Summary of binding kinetics and dissocation constants obtained from SPR analysis and PPI inhibitory activity data obtained from PlxnB1-Sema4D pulldown assays. (B) Scatter plot showing the effects of increasing concentrations of PB1m6 (closed circles) and PB1d6P10 (open circles) on Sema4D-dependent growth cone collapse in HEK293 cells expressing exogenous PlxnB1. Data were obtained from 3 independent experiments (mean ± SD). The star indicates low concentration inhibitory maximum. (C) Representative microscopy images of PlxnB1 expressing HEK293 cells treated with mock medium (top-left), 1 nM Sema4D-Fc (top-right), 1 μM PB1m6 and 1 nM Sema4D (bottom-left) and 1 μM PB1d6P10 and 1 nM Sema4D (bottom-right).

approach was used in an attempt to generate a homodimeric PB1m6 linked through two NH-PEG5-COOH linkers connected by the amide form of 2,3-diaminoproinic acid (DAP) (PB1d6P10, Figure 1). MALDI-TOF MS analysis of the major product in the preparative HPLC spectrum showed a peak corresponding to the calculated MW of PB1d6P10 (Figure S3-4). Further, ultra high performance liquid chromatography (UHPLC) analysis of the purified product was performed to confirm that a single species was purified (Figure S5). We then chose to use surface plasmon resonance (SPR) analysis to see if there was improvement in binding affinity through bivalency. Indeed, when compared to the monomeric peptide, PB1m6, PB1d6P10 showed a 188-fold enhancement in binding affinity (Figure 2A, S6) with most of the gain in affinity coming from slower dissociation rates which is consistent with previous reports regarding the role of bivalence in antibodies3. We then set out to test what range of linker lengths would be possible for branched synthesis of dimeric peptides. The shortest linker attempted was PB1d6P2 containing two NH-PEG1-COOH segments as linkers. Longer linkers were made using combinations of Fmoc-NH-PEG5-COOH and Fmoc-NH-PEG11-COOH to generate PB1d6P22 which contains two NH-PEG11-COOH as linkers, PB1d6P32 which contains two NH-PEG11-COOH and two NH-PEG5-COOH as linkers, and finally, PB1d6p44 which uses a total of four NH-PEG11-COOH as linkers (Figure 1, Figure S3). These linker lengths span between approximately 30 Å for PB1d6P2 and approximately 200 Å for PB1d6P44 which is longer than the average distance between the IgG Fab arms (120-170 Å)14 (Figure 1). MALDI-TOF MS analysis showed successful synthesis of the aforementioned dimeric peptides and UHPLC analysis of

the purified products showed single peaks without serious contaminants (Figure S4-5). SPR analyses of these dimeric peptides with varying linker-lengths all showed similar improvements in affinity as PB1d6P10 did, displaying 70–150-fold decreases in their dissociation constant, KD (Figure 2A, Figure S6). Similar to the results with PB1d6P10, the large majority of the increases in binding affinity came from slower dissociation rates (Figure 2A, S6). The aforementioned Fmoc-NH-PEGX-COOH linkers used afforded long linker lengths while minimalizing coupling steps, however, their cost and availability can be somewhat limiting. Therefore, we tested whether branched synthesis of dimeric peptides can still be performed by replacing the NH-PEGX-COOH linkers with multiple repeats of glycine and β-alanine (GβA) which are significantly less expensive, readily available and have relatively higher coupling efficiencies compared to Fmoc-protected amino acids with protected side chains15. A GβA-linker containing PB1m6-based dimeric peptide (PB1d6-28A) was synthesized with a linker length comparable to that of PB1d6P10 (Figure 1J, S3). MALDI-TOF MS analysis was able to confirm the successful synthesis of PB1d6-28A and UHPLC analysis showed a single peak (Figure S4-5). Subsequent SPR analysis of PB1d6-28A also showed a substantial increase in binding affinity (285-fold) and a much slower dissociation rate (124-fold) as with the other PB1m6-based dimeric peptides (Figure 2A, S6). In vivo, PlxnB1 is expressed on neuronal cells and acts as an axon guidance molecule and when coming in contact with its binding partner, Sema4D, growth cone collapse occurs16. As mentioned above, PB1m6 is able to bind PlxnB1 and allosterically inhibit its

ACS Paragon Plus Environment

Bioconjugate Chemistry 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

protein-protein interaction (PPI) with its biological binding partner Sema4D and therefore prevents Sema4D-dependent growth cone collapse8. To see whether this increase in binding affinity through dimerization translates to increased PPI inhibition, we used a previously developed binding assay which detects the relative quantity of alkaline phosphatase-fused human Sema4D ectodomain that is pulled down by bead-immobilized hPlxnB18. Applying the PEG- linked PB1m6-based dimeric peptides, as well as monomeric PB1m6, to this assay showed no substantial differences in IC50 values between PB1m6 and its dimeric counterparts (Figure 2A, S7). However, while the maximal PPI inhibitory activity of PB1m6 tops out at approximately 65% inhibition, the dimeric versions of PB1m6 all showed improved maximal PPI inhibitory activity between 86 to 92% (Figure 2A, S7). We also compared the GβA-linked PB1d6-28A and its IC50 and maximal PPI inhibition was comparable to that of the PEG-based linkers (Figure 2A, S7).

linker is desired. We were also able to show that dimers with various linker compositions and lengths could be prepared. FmocNH-PEGX-COOH based linkers made possible the synthesis of dimers spaced as close as 30 Å apart, to as far as 200 Å apart. Importantly, dimerization of thioether-macrocyclic peptides shared the same benefits of bivalence as antibodies by exhibiting increases in affinity up to nearly 300-fold. In addition, dimeric peptides with shorter linker lengths showed higher PPI inhibitory activity which is likely due to dimerization of the target PlxnB1 to an nonfunctional outward-facing geometry indicated by the inhibition of Sema4D-dependent growth cone collapse in PlxnB1-expressing cells. The method reported in this work would find wide use in quickly enhancing the potency of in vitro selected cyclic peptides as well as to generate molecular tools to probe the effects of protein homo-dimerization.

To further examine the enhanced PPI inhibitory activity of these dimeric cyclic peptides, we used a full-length PlxnB1 expressing HEK293 cell line which exhibits growth cone collapse when exposed to Sema4D8. The monomeric peptide PB1m6 exerted no inhibition of Sema4D dependent growth cone collapse at 1 µM concentration but showed complete inhibition at 30 µM with an IC50 of 3 µM which is consistent with our previous reports (Figure 2B-C)8. As a representative dimer of PB1m6, we chose PB1d6P10. Interestingly, PB1d6P10 showed much improved inhibitory activity against PlxnB1 and complete inhibition of Sema4ddependent growth cone collapse was observed at a nearly 30-fold lower concentration than PB1m6 (1 µM, IC50 240 nM). Moreover, it showed a bimodal inhibition curve with another local inhibitory maxima at 3 nM PB1d6P10 concentration reaching approximately 35% inhibition (denoted by a star in Figure 2B). We speculate that this first phase of inhibition is caused by the ability of PB1d6P10 to induce inactive PlxnB1 dimers rather than by blocking the binding of the Sema4D. We and others, using the structural information gained by crystallographic and electron microscopic analyses, have postulated that plexin receptors transduce signals by forming particular dimeric configurations on the cell surface upon engagement with dimeric semaphorins17–20. Interestingly, PB1m6 binds the PlxnB1 sema domain opposite from the binding site for the physiological ligand, Sema4D8, indicating that the configuration of the PlxnB1 dimers induced by dimeric peptides would be drastically different from that induced by the physiological "activating" ligand (Figure S8). Thus we hypothesize that the presence of ~3 nM PB1d6P10 leads to a 1:2 clamping of two receptor molecules tethered into orientations that hinder the correct formation of Sema4D-induced dimers, resulting in the signal inhibition. This effect may be further compounded by the innate allosteric PPI activity of PB1m68. However, with increasing concentrations of PB1d6P10, the dimeric peptides outcompete each other and are no longer able to tether PlxnB1 into inactive dimers and the inhibition mode reverts to that of monomeric PB1m6, albeit much stronger due to the enhanced affinity gained from bivalency.

Supporting Information

In conclusion, we have developed a fast and efficient method of generating dimeric thioether-macrocyclic peptides through branched peptide synthesis. This method requires the same number of steps as synthesizing a monomeric macrocyclic peptide with the only addition being the initial coupling of a “node” amino acid (Fmoc-2,3 diaminopropionic acid in this study) and whatever

ASSOCIATED CONTENT The Supporting Information is available free of charge on the ACS Publications website. Detailed experimental procedures and Figures S1-S6.

AUTHOR INFORMATION Corresponding Author *E-mail: [email protected], [email protected]

Notes The authors declare no competing financial interest.

ACKNOWLEDGMENTS This research was supported by the Platform Project for Supporting in Drug Discovery and Life Science Research (Platform for Drug Discovery, Informatics, and Structural Life Science) from the Ministry of Education, Culture, Sports, Science (MEXT) and Japan Agency For Medical Research And Development (AMED) (to H.S. and J.T.).

REFERENCES (1) (2) (3) (4) (5)

(6) (7) (8) (9)

(10)

(11)

Obexer, R.; Walport, L. J.; Suga, H. Curr. Opin. Chem. Biol. 2017, 38, 52. Bobrovnik, S. A. J. Mol. Recognit. 2007, 20 (4), 253. Vauquelin, G.; Charlton, S. J. Br. J. Pharmacol. 2013, 168 (8), 1771. Hornick, C. L.; Karush, F. Immunochemistry 1972, 9 (3), 325. Klein, J. S.; Gnanapragasam, P. N. P.; Galimidi, R. P.; Foglesong, C. P.; West, A. P.; Bjorkman, P. J. Proc. Natl. Acad. Sci. 2009, 106 (18), 7385. Greenbury, C. L.; Moore, D. H.; Nunn, L. A. C. Immunology 1965, 8, 420. Ito, K.; Sakai, K.; Suzuki, Y.; Ozawa, N.; Hatta, T.; Natsume, T.; Matsumoto, K.; Suga, H. Nat Commun 2015, 6. Matsunaga, Y.; Bashiruddin, N. K.; Kitago, Y.; Takagi, J.; Suga, H. Cell Chem. Biol. 2016, 23 (11), 1341. Lorenzón, E. N.; Cespedes, G. F.; Vicente, E. F.; Nogueira, L. G.; Bauab, T. M.; Castro, M. S.; Cilli, E. M. Antimicrob. Agents Chemother. 2012, 56 (6), 3004. Monreal, I. A.; Liu, Q.; Tyson, K.; Bland, T.; Dalisay, D. S.; Adams, E. V.; Wayman, G. A.; Aguilar, H. C.; Saludes, J. P. Chem. Commun. 2015, 51 (25), 5463. Duggineni, S.; Mitra, S.; Lamberto, I.; Han, X.; Xu, Y.; An, J.; Pasquale, E. B.; Huang, Z. ACS Med. Chem. Lett. 2013, 4 (3), 344.

ACS Paragon Plus Environment

Page 4 of 5

Page 5 of 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

Bioconjugate Chemistry (12) (13) (14) (15) (16) (17)

(18)

(19)

(20)

Thumshirn, G.; Hersel, U.; Goodman, S. L.; Kessler, H. Chem. - A Eur. J. 2003, 9 (12), 2717. Groß, A.; Hashimoto, C.; Sticht, H.; Eichler, J. Front. Bioeng. Biotechnol. 2016, 3 (January). Doores, K. J.; Fulton, Z.; Huber, M.; Wilson, I. A.; Burton, D. R. J. Virol. 2010, 84 (20), 10690. Collins, J. M.; Porter, K. a; Singh, S. K.; Vanier, G. S. Org. Lett. 2014, 16 (3), 940. Worzfeld, T.; Offermanns, S. Nat. Rev. Drug Discov. 2014, 13 (8), 603. Janssen, B. J. C.; Robinson, R. a; Pérez-Brangulí, F.; Bell, C. H.; Mitchell, K. J.; Siebold, C.; Jones, E. Y. Nature 2010, 467 (7319), 1118. Nogi, T.; Yasui, N.; Mihara, E.; Matsunaga, Y.; Noda, M.; Yamashita, N.; Toyofuku, T.; Uchiyama, S.; Goshima, Y.; Kumanogoh, A.; Takagi, J. Nature 2010, 467 (7319), 1123. Kong, Y.; Janssen, B. J. C.; Malinauskas, T.; Vangoor, V. R.; Coles, C. H.; Kaufmann, R.; Ni, T.; Gilbert, R. J. C.; PadillaParra, S.; Pasterkamp, R. J.; Jones, E. Y. Neuron 2016, 91 (3), 548. Suzuki, K.; Tsunoda, H.; Omiya, R.; Matoba, K.; Baba, T.; Suzuki, S.; Segawa, H.; Kumanogoh, A.; Iwasaki, K.; Hattori, K.; Takagi, J. PLoS One 2016, 11 (6), 1.

ACS Paragon Plus Environment