Synthesis of Mannosylated Lipopeptides with Receptor Targeting

Article pubs.acs.org/bc

Synthesis of Mannosylated Lipopeptides with Receptor Targeting Properties Bita Sedaghat,† Rachel J. Stephenson,*,† Ashwini Kumar Giddam,† Sharareh Eskandari,† Simon H. Apte,‡ David J. Pattinson,‡ Denise L. Doolan,*,‡ and Istvan Toth*,†,§,∥ †

School of Chemistry and Molecular Biosciences and ∥Institute for Molecular Bioscience, The University of Queensland, St Lucia, Queensland 4072, Australia ‡ Infectious Diseases Program, QIMR Berghofer Medical Research Institute, Brisbane, Queensland 4029, Australia § School of Pharmacy, The University of Queensland, Woolloongabba, Queensland 4012, Australia S Supporting Information *

ABSTRACT: Present on the surface of antigen presenting cells (APCs), the mannose receptor (MR) has long been recognized as a front-line receptor in pathogen recognition. During the past decade many attempts have been made to target this receptor for applications including vaccine and drug development. In the present study, a library of vaccine constructs comprising fluorescently labeled mannosylated lipid-dendrimers that contained the ovalbumin CD4+ epitope, OVA323−339, as the model peptide antigen were synthesized using fluorenylmethyloxycarbonyl (Fmoc) solid phase peptide synthesis (SPPS). The vaccine constructs were designed with an alanine spacer between the O-linked mannose moieties to investigate the impact of distance between the mannose units on receptor-mediated uptake and/or binding in APCs. Uptake studies performed on F4/80+ and CD11c+ cells showed significant uptake and/or binding for lipopeptides containing mannose, and also the lipopeptide without mannose when compared to the control peptides (peptide with no lipid and peptide with no mannose and no lipid). Furthermore, mannan inhibition assays demonstrated that uptake of the mannosylated and lipidated peptides was receptor mediated. To address the specificity of receptor uptake, surface plasmon resonance studies were performed using biacore technology and confirmed high affinity of the mannosylated and lipidated vaccine constructs toward the MR. These studies confirm that both mannose and lipid moieties play significant roles in receptor-mediated uptake on APCs, potentially facilitating vaccine development.

■

INTRODUCTION

targeting peptide-based vaccines and drugs to the immune system via APCs to enhance cellular and humoral responses.4 The MR contains eight extracellular carbohydrate recognition domains (CRDs).5 The conformation of the CRDs has been shown to play an important role in the recognition of glycosylated ligands.3,6 The structural properties of the MR CRDs have been studied extensively, but the relationship between the antigen/receptor binding and recognition and the properties that preferentially target the antigen toward the receptor are less known.7 Natural O-mannosylated structures found on the surface of pathogens are known to be recognized by the MR. Many studies have investigated the targeting and uptake of mannosylated antigens through the MR.8−11

The mannose receptor (MR) is found on the surface of dendritic cells (DCs) and macrophages and has long been accepted as a front-line receptor in the recognition and internalization of pathogenic antigens.1 The MR and other Ctype lectin receptors (e.g., DC-SIGN) are part of a family of pattern recognition receptors found on the surface of APCs that recognize glycoconjugates from foreign pathogens, such as bacteria and fungi.2,3 These receptors have been a primary focus of research to develop targeted drugs and vaccines over many decades.2,3 Surface receptors (including MR) allow DC and macrophages to recognize and internalize glycosylated antigens. These antigens are processed and then presented by major histocompatibility complex (MHC) class I and class II molecules, activating the adaptive immune response. The inclusion of mannosylated moieties has been one strategy of © XXXX American Chemical Society

Received: October 12, 2015 Revised: December 17, 2015

A

DOI: 10.1021/acs.bioconjchem.5b00547 Bioconjugate Chem. XXXX, XXX, XXX−XXX

Article

Bioconjugate Chemistry

Figure 1. Schematic image of mannosylated lipopeptide vaccines used to target DCs and macrophages. The vaccine contained a linear arrangement of two mannose moieties separated by an alanine spacer that contained 0, 1, or 2 alanine units attached to the ovalbumin CD4+ epitope, OVA323−339. Lipids were included in the construct to confer self-adjuvanting properties, while a fluorescent tag [5(6)-carboxyfluorescein, FAM] facilitated in vitro cell tracking.

Label-free surface plasmon resonance (SPR) and competition studies using mannan, a known MR ligand, were also performed. Mannosylation of the OVA antigen has been shown to increase MR uptake, and thus enhance the immunogenicity of the construct, presumably due to better presentation of the antigen through MHC class I and II molecules.24 Furthermore, the conformation of the final constructs, which may include peptide antigens or other bound functionalities (e.g., conjugation with PEG, or branching mannose) used to create an effective vaccine, are all known to affect the arrangement and location of the mannose units required for effective receptor targeting and binding.22,25 In the present work, the constructs were designed to mimic the linear arrangement and O-linked properties of glycoconjugates found in nature.26 An alanine spacer was inserted between the mannose units to study the structural properties of the mannosylated vaccine in relation to APC uptake. Previous studies have reported that the distance between the mannose groups played an important role in the uptake of mannosylated compounds through the MR on monocytes.7,17,19 In a study by Kowalczyk et al., the optimal uptake of mannosylated compounds was achieved when three alanine amino acids were inserted between the linearly attached O-mannose groups.27 Two mannose units have been included in each construct as previous studies showed that synthetic constructs that contained two mannose units could target the MR with high affinity.7,17,18,28 Additionally, shorter spacers between the mannose units have been shown to be effective at targeting the MR; thus, we investigated this aspect relative to the presence of lipids and a peptide antigen which are known to enhance self-adjuvanting properties and induce particulate formation.11,17,28 Amphiphillic moieties have been shown to enhance nanoparticle formation, and the change in nanoparticle size was found to affect uptake.29 This study has shown that structures that contain mannose and lipids bind strongly to APCs in vitro, compared to unmannosylated compounds. To support MR binding by these constructs as observed in flow cytometry and confocal experiments, SPR (using biacore technology) on the binding specificity of this compound library to the recombinant human macrophage MR was used. Here, the presence of both mannosyl and lipidic moieties for enhanced MR binding affinity was observed.

However, the structural and spatial arrangement of glycosylation on the pathogen that triggers recognition by these receptors remains to be identified.1,12 Glycosylation, in particular, mannosylation, was an effective method for targeting antigenic peptide and protein vaccines to the MR, increasing their antigenic potential by enhancing uptake and processing by APCs.6,13,14 Although the exact binding requirements of the MR remain to be elucidated, studies aimed at targeting this receptor reported that polymannan derivatives, mannosylated dendrimers, and compounds that mimic a dendritic branched structure, known as the cluster effect, were often found to be antigenic.15−17 Furthermore, it was shown that branching, multimerization, and the number of mannose moieties affect receptor binding.15−17 In 2008, Kowalczyk et al. showed that bis-mannose compounds that contained a PEG spacer were taken up by human monocytes better than monomannosylated compounds. However, they also identified a monomannosylated compound that showed similar binding, suggesting that the incorporation of a second mannose group in the peptide chain was not essential.18,19 Studies by Frison et al. showed that the linear arrangement of the MRs CRD led to specific and high affinity binding between the MR and synthetic compounds that contained dimannose clusters and/or end-standing single mannose units.16,17 The MR has been shown to bind to mannosylated ligands with different branching and spacing arrangements in the presence of other functional groups. Nevertheless, the exact binding requirements of the MR have not been clearly defined.7,17 O-Linked glycosyls are the most common form of glycoproteins found in nature. The glycosylation of peptides or proteins is challenging and has been investigated extensively.11,20,21 For example, the successful O-mannosylation of proteins was achieved using enzymes such as O-Dmannosyltransferase or by complicated synthetic methods that required several complex steps and often generated low yields.22 However, both methods have been shown to generate the O-mannosylated building blocks serine and threonine.20 Routes developed to successfully synthesize glycosylated peptides (and proteins) are dependent on the location and step in the synthesis and are reviewed in detail elsewhere.23 This study aimed at investigating the effect of distance between the mannose units on receptor-mediated binding and uptake by APCs. A library of O-mannosylated lipopeptides that contained a linear arrangement of mannose and the model CD4+ antigen, ovalbumin (OVA323−339) (Figure 1) were synthesized. These vaccine constructs were characterized by measuring their uptake by CD11c+ and F4/80+ cells in vitro.

■

RESULTS AND DISCUSSION Synthesis and Characterization of Fluorescently Labeled Mannosylated (and Unmannosylated) Peptide Azides. O-Linked glycoproteins, where the glycosyl moiety is B


Article


Figure 2. Scheme for the synthesis of 1,2,3,4,6-penta-O-acetyl-α,β-D-mannopyranosyl (1) and N-Fmoc-O-(2,3,4,6-tetra-O-acetyl-α,β-Dmannopyranosyl)-L-serine (2) from D-mannose. (a) DMAP, Ac2O, 46 °C, 24 h, N2 gas; (b) Fmoc-Ser-OH, BF3·Et2O, RT, 4 h.

the orthogonal group of lysine because it is easily deprotected with 1% TFA and scavenger TIS in DCM without cleaving the peptide from the resin.34 Following this, the Fmoc group of lysine was removed and the Fmoc-serine-mannose amino acid (2, Figure 2) was coupled using standard in situ Fmoc coupling procedures.35 In this study, Fmoc-Ala-OH was used as a spacer between the mannose groups as shown in Figure 3. Fmoc amino acids were coupled by activation with HATU and DIPEA in DMF. HATU has been shown to be an efficient coupling reagent for generating products with high reactivity and low racemization.36 Following addition of the last amino acid in sequence, FAM was coupled and excess ester-bound FAM removed by treatment with 20% piperidine, enabling visualization in cellbased studies.37 The mannosylated peptides (3−5) were released from the solid support by treating the resin with a mixture of TFA, TIPS, and water. All the acid-labile amino acid side chain protecting groups were removed concomitantly. The resulting crude peptides were purified by preparative RP-HPLC on a C8 column and characterized using ESI-MS (Table 1). Unmannosylated fluorescent peptide 6 was synthesized using the same procedure as compounds 3−5; however, the Fmoc mannosylated serine amino acid (2) was replaced with FmocSer (tBu)-OH (Figure 3). Peptide 6 contained two alanine amino acids employed as a spacer between each serine unit (Figure 3) and was used a control for in vitro studies to investigate the role of mannose on receptor binding. Synthesis and Characterization of OVA323−339 Lipopeptide and Peptide Alkynes. An adjuvant is an essential component of vaccine design. Lipopeptide vaccines have been shown to be structurally stable and biologically active with effective adjuvant properties in a variety of vaccines, including those that target Group A streptococcus and cancer.38,39 Furthermore, glycosylated lipopeptides have been shown to be good peptide delivery systems with self-adjuvanting properties.40 Here, we synthesized alkyne-functionalized peptides that contained two lipid moieties (C12) and two branches of OVA323−339 as the model antigen using microwave-assisted Fmoc SPPS (7, Figure 5). Synthesis of the lipids and subsequent protection using the 1-(4,4-dimethyl-2,6dioxyacyclohexylidene)ethyl (Dde) protecting group was performed using published procedures.41 Following coupling of the Dde-protected C12 lipids using standard microwave assisted Fmoc synthesis, the Dde protecting group was manually removed by treating the resin with 2% hydrazine in DMF. Branching and concurrent addition of two OVA peptides (Figure 5) was achieved using Fmoc-lysine(Fmoc)-OH. Standard microwave-assisted Fmoc SPPS protocols were used for the synthesis of peptide 7, with amino acid activation by HATU and DIPEA. Peptide 8, which contained no lipid moieties, was synthesized as a control for the cell-based assays

attached through a serine or threonine, are the most common glycoproteins found in nature30 and are commonly isolated from the cell wall of yeasts and bacteria.31,32 In this study, mannose amino acid building blocks based on the amino acid serine were used to target the C-type lectins on APCs. Here, functionalization of the serine side chain with mannose enabled the stepwise addition of an O-mannosylated building block during solid phase peptide synthesis (SPPS). These vaccine constructs were designed with different length spacers (0, 1, 2) between the two mannose moieties to gain information about the relative spatial requirements of the sugars on their uptake by APCs. The effect of lipids and a peptide antigen on uptake were also investigated. The synthesis of multifunctional glycosylated peptides that contained two mannose moieties, an azido functionality (to enable conjugation to the OVA antigen using azido-alkyne click chemistry), and a fluorescent tag was challenging and required specific orthogonal protection during peptide synthesis, protection of reactive side chain functional groups, and the production of a glycosylated amino acid for use in Fmoc SPPS. An Fmoc-serine mannosylated amino acid 2 (N-Fmoc-O(2,3,4,6-tetra-O-acetyl-α,β-D-mannopyranosyl)-L-serine, Figure 2) was synthesized as a building block for the mannosylated constructs 3− 5 (Figure 3). This mannosylated amino acid has been shown to successfully generate high-yielding peptides by Fmoc SPPS.33 Here, the hydroxyl groups of D-mannose were acetylated to ensure stability of the sugar moiety during peptide cleavage from the resin in the presence of 95% TFA. Without protection, the glycosidic bond was susceptible to degradation.33 The synthesis of 1 was performed as previously described.33 The synthesis of 2 was optimized by monitoring the Lewis acid reaction using analytical RP-HPLC; a 4 h reaction was found to give a maximum yield of 42% while longer reactions resulted in degradation of the product (Figure 4). This degradation was associated with rearrangement of the glycosyl group and formation of new subsidiary structures.21 Additionally, two peaks were observed for the product which was attributed to the α- and β-analogues of the sugar (Figure 4). Furthermore, as the Fmoc-serine amino acid (Rt: 19.3 min) was in excess, it was expected that this peak would be detected at the conclusion of the reaction (Figure 4). Variations in the observed peak that corresponds with the starting material (Rt: 19.3 min, Figure 4) has been previously reported by Liu et al. and is related to the reversibility of the reaction and the presence of transient species at early reaction time points.21 Azido-functionalized mannosylated peptides 3−5 (Figure 3) were synthesized using manual Fmoc SPPS. The orthogonal methyltrityl (Mtt) side chain protecting group was removed from the lysine using mildly acidic conditions (Figure 3) and azdioacetic acid was coupled to allow for functionalization with the OVA lipo-peptide after cleavage. Mtt was chosen to protect C


Article


Figure 3. Reagents and conditions for the synthesis of peptides 3−6: (a) Piperidine in DMF, RT; (a*) 2 cycles of piperidine in DMF; (b) Fmoclys(Mtt)-OH, HATU, DIPEA; (c) 2% TFA: 1% TIPS in DCM, 20 cycles; (d) azidoacetic acid, HATU, DIPEA, DMF; (e) N-Fmoc-O-(2,3,4,6-tetraO-acetyl-α,β-D-mannopyranosyl)-l -serine, HATU, DIPEA, DMF; (f) Fmoc-Ala-OH, HATU, DIPEA, DMF; (g) 5(6)-carboxyflourscein (FAM), DIC, HOBt, DMF; (h) 95% TFA: 2.5% water: 2.5% TIS.

assays found that the triazole bond formed between the two products had high stability against proteolytic degradation and was able to mimic a native peptide bond.42,43 In this study, copper-mediated click chemistry was used to conjugate the OVA323−339 lipopeptide-alkyne (7) to fluorescently labeled mannosylated-azides (3−5), forming a library of mannosylated vaccine constructs (9−11, Figure 6) designed to test the impact of distance between the mannosyl groups on cell uptake. Additionally, vaccine constructs used as controls in in vitro studies were also synthesized. Here, conjugation of 6 to 8

using a similar method to peptide 7, without inclusion of the lipids. Conjugation of Mannosylated Lipopeptide Vaccine Constructs Using Copper-Mediated Azide−Alkyne Click Chemistry. Large dendritic peptides with complex branched structures can be difficult to synthesize and purify in one step. However, there are a number of techniques available to assist in their synthesis.42,43 In this study, copper-mediated cycloaddition (or “click chemistry”) was employed due to its safety, reliability, and high yielding reactions. Furthermore, cell-based D


Article


Figure 5. Structure of the OVA323−339 lipopeptide alkyne (7) and OVA323−339 peptide alkyne (8). Peptide 7 contained two C12 lipids at the C-terminus to enhance self-adjuvanting properties of the final vaccine constructs, an alkyne functionality for conjugation to the mannosylated peptide-azides, and two OVA323−339 CD4+ epitopes conjugated through a branching lysine residue. The amino acid sequence for OVA323−339 is ISQAVHAAHAEINEAGR. Peptide 8, a control peptide, contained the same functional groups as peptide 7 with the exception of the two C12 lipids that were omitted.

Figure 4. Monitoring of the boron trifluoride diethyl etherate (BF3· Et2O) Lewis acid reaction at 1, 2, 3, 4, 6, and 8 h by analytical RPHPLC on a C18 column (0−100% solvent B over 30 min). At 4 h, the optimum yield for the product (2, 42%) was observed with a retention time (Rt) of 23.6 min (peak indicated by an arrow) at 62% solvent B. Rt 19.3 min is Fmoc-serine-OH starting material and Rt: 21.5 min is the product missing one acetyl group. The graph was drawn using nudging in GraphPad Prism.

(24 h), there was no change in the product yield and byproducts were formed, potentially due to the breakdown of the sugar moiety from lengthy exposure to elevated temperatures (50 °C). In this study, copper wire was used as the source of copper catalyst because its advantages over traditional sources of copper because it obviates the need to use N-based ligands or reduce CuII to CuI in situ and facilitates the easy removal of excess copper from the reaction.29 At the completion of the reaction copper wires were removed by filtration, and the reaction was quenched with water and lyophilized to obtain the crude products. Vaccine constructs 9−14 (Figure 7) were purified by preparative RP-HPLC on a C8 column and isolated in yields of 34−51% (Table 1). The retention times were slightly longer than that of the reacting peptides due to their larger molecular size and higher hydrophobicity (Table 1). RP-HPLC traces of peptides 3−12 are provided in the Supporting Information. Figure 8 shows the RP-HPLC trace for construct 9 as an example. The peak of the product (9, Figure 8) and the lipopeptide starting material (7, Figure 8) was multivalent because the lipids were present as a diastereomeric mixture. Construct 9 also contained a mixture of both α- and βstereochemistries for the mannose units (Figure 8). Here, broad peaks are observed for all mannosylated peptides (3−6) and lapidated peptides (7−8) due to the anomeric mixture of the mannose moieties and the diastereoisomeric mixture of the lipids, respectively (Figures S1−6 in the Supporting Information). As a result, broad peaks are observed for the vaccine constructs 9−11 (Figures S7−12 in the Supporting Information). For all constructs, the number of peaks observed in RPHPLC is less than the expected number of peaks and this can be attributed to each diastereoisomer having a similar retention time, and hence the broader RP-HPLC peaks. Furthermore, the mannosylated peptide azide (3), used in excess, was present at the conclusion of the reaction (Figure 8). Vaccine constructs that contained the mannose groups had been acetylated for peptide synthesis (2, Figure 2) and this acetyl group was not removed before cell-based assays. Previous studies have shown that acetylation of glycosyl groups (including mannosyl) increased cell uptake and the acetyl groups were subsequently removed by esterases inside the cell.44,45 Previous studies have also confirmed no significant difference in biological activity

Table 1. ESI-MS Data, RP-HPLC Rt and Yields for Vaccine Constructs 3−14 ESI-MS m/z ionization

calculated

found

3

[M + H]+1 [M + H]+1 [M + H]+1 [M + H]+1 [M +3H]+3 [M +2H]+2 [M +4H]+4 [M +4H]+4 [M +4H]+4 [M +4H]+4 [M +4H]+4 [M +4H]+4

1563.45

1565.0

20.83

54, C8

1634.53

1634.5

20.57

62, C8

1705.61

1706.6

20.50

68, C8

1044.39

1045.8

15.04

48, C8

1439.98

1440.1

73, C4

1962.10

1962.1

17.47, 17.28, 16.96 17.79

1471.09

1472.0

1488.86

1489.3

1506.63

1506.3

1341.48

1341.5

1407.97 1242.82

4 5 6 7 8 9 10 11 12 13 14

Rt (min)

yield (%, column)

construct

83, C18

1408.1

21.68, 21.94, 22.42 21.26, 20.56, 21.97 19.88, 20.11, 20.47 20.13, 20.41, 21.36 13.50

33, C8

51, C8

1244.1

14.35

35, C8

35, C8 34, C8 45, C8

generated 12 (Figure 7) which contained no mannose moieties. Furthermore, a mannosylated construct that contained no lipids (13, Figure 7) and a construct that contained neither lipids nor mannose (14, Figure 7) were generated by conjugation of 8 with 5, and 8 with 6, respectively. Constructs 12−14 were designed to test the role of lipids and/or mannose in cell binding and uptake studies in vitro. The click reaction was performed in DMSO, which easily dissolved each reaction component. Reaction monitoring by analytical RP-HPLC and product identification via ESI-MS were used to quantify the optimal reaction duration. Here, the reactions were found to be completed in 5 h where formation of the product did not proceed any further with increased reaction times. In some cases, when the reaction was left longer E


Article


Figure 6. Mannosylated lipopeptide vaccine constructs (9−11) prepared from a copper-mediated click reaction (DMSO, 5 h, RT, N2 gas) of the lipopeptide-alkyne 7 with fluorescently labeled mannosylated-azides (3−5), respectively. The triazole ring that results is highlighted by a square.

between acetylated and nonacetylated mannosylated peptides.46 In addition, studies also show that both acetylated and nonacetylated hexoses inhibit the binding of mannan to mannose binding lectins.47 Size Characterization. Vaccine constructs 9−11 were investigated for their self-assembly and particle formation in PBS. Here, the hydrophilic OVA323−339 peptide conjugated to the hydrophobic lipid created amphiphillic constructs that selfassembled in aqueous media to form particles, providing access to peptide nanoparticles with discrete sizes.48 Self-assembled nanosized particulate vaccines have been shown to possess many advantages, including enhanced stimulation of immune responses owing to improved presentation of multivalent antigens on the surface of the particle and increased in vitro stability.49 Furthermore, particle formation has been shown to enhance receptor accessibility to the targeting mannose moieties.11 In this study, dynamic light scattering (DLS) measurements of peptides 9−11 showed a size range of 150 ± 50 nm with a polydispersity index (PDI) of 0.4−0.6. Transmission electron microscopy (TEM) imaging of these constructs (9−14) showed a smaller size range (80 ± 50 nm, Figure 9A−C) when compared to their size measured by DLS. This observed change in size has been attributed to the drying step during sample preparation. TEM of 12−14 showed a size range of 40 ± 10 nm (Figure 9D−F). In general, it was postulated that constructs with lower molecular weight (12− 14) had smaller particle size while the presence of lipid and sugar moiety caused a higher molecular weight and consequently bigger size. In Vitro Evaluation. Flow Cytometry Cell Uptake Study. In vitro uptake and targeting of the mannosylated vaccine constructs (9−14, Figure 7) was tested in APCs isolated from spleens of naive C578L/6 mice. Cells were isolated as a mixed population of APCs, including macrophages and DCs, using a standard protocol50 and were used immediately after isolation. Fluorescently labeled vaccine constructs (0.5 μM) were added to the isolated cells and allowed to incubate for 4 h before washing the cells to remove any excess compound. The washing step in this procedure is known to remove unbound

peptides from the surface of the cells.40,51The vaccine constructs were labeled with a fluorescent tag (FAM) to facilitate cell tracking. Cells were then stained with anti-CD11c and anti-F4/80 antibodies which bound to cells that contained CD11c and F4/80 markers. CD11c is present in high levels on DCs and F4/80 on macrophages.52,53 The concentration of vaccine constructs was optimized (data shown in the Supporting Information). Changing the distance between the mannose moieties from zero alanine units (9, Figure 7) to two alanine units (11, Figure 7) resulted in a negligible difference in uptake by both CD11c+ and F4/80+ cells (Figure 10A,B, respectively). Here, peptides 12−14 are control peptides designed with only lipids (12), only mannose moieties (13) and no mannose or lipids (14). However, a significantly lower binding was observed in both subsets of cells for the control peptides 13 and 14 that contained no lipid and/or mannose moieties (Figure 10A,B). Dextran has been shown to bind strongly to the MR and was used as a positive control for both subsets of cells.54 Results also confirmed a higher level of uptake by 9−14 in CD11c+ positive cells compared to F4/80+ cells. A mannan inhibition study was performed to confirm receptor-mediated uptake of the constructs on the surface of CD11c+ and F4/80+ cells. Here, mannan is a known ligand of the MR.15 Following preincubation of the cells with mannan (1 mg/mL), the cells were incubated for 4 h with 0.5 μM of the vaccine constructs (9−14) before washing and staining separately for CD11c+ or F4/80+ cells (Figure 10). Flow cytometry results showed a significant reduction in binding and/or uptake for all constructs when the cells were preincubated with mannan (Figure 10), implying that mannan specifically blocked the receptor-mediated uptake of these compounds. A significant difference in mannan inhibition studies was observed in the binding and/or uptake of constructs 9−11 in both CD11c+ and F4/80+ cells with increased binding observed when a longer spacer was employed between the mannose units. However, no significant difference was observed for constructs 9−11 in binding and uptake studies in the absence of mannan. From these lipidated and F


Article


Figure 7. Library of fluorescently labeled mannosylated lipopeptide vaccine test constructs (9−11) and control constructs (12−14). Here, test constructs contained a variable length linker of alanine used to investigate APC-targeted binding. The control constructs: 12 contained no mannose moieties, 13 contained no lipid moieties, and 14 contained neither lipid nor mannose moieties.

mannosylated constructs, binding inhibition in the presence of mannan was higher for construct 11 compared with constructs

9 and 10. It could only be concluded that 11 is the best mimic of mannan for CD11c+ and F4/80+ cells (11, Figure 10A,B). G


Article


of cell-binding observed for 12 (Figure 10).56 Here, nonspecific binding observed for peptide 12 could be associated with the affinity of the vaccine candidates for non-C-type lectin receptors (including Toll-like receptors) where previous studies on lipidated peptides have confirmed this finding.57 The mannan inhibition study indicated that receptor-mediated uptake through a mannan-inhibited receptor (e.g., MR) is likely. However, as uptake was not completely inhibited, other receptors and mechanisms were also likely to contribute to the uptake.58 Overall, results from this study suggest that both mannose and lipids play important roles in the uptake of vaccine constructs, with both structural features affecting receptor mediated uptake. Confocal Imaging. Confocal imaging was used to visualize the uptake and/or binding of the vaccine constructs to F4/80+ cells. Constructs (0.5 μM) were incubated with cells for 4 h followed by staining with an anti-F4/80 antibody. The cell nucleus was stained with HOECHST stain. Mannan-FITC was used to show the association of mannan to the cell-surface receptors (Figure 11A). The results showed that construct 9 was localized at the periphery of the cell and was observed inside the cell as punctate granules in the cytoplasm (Figure 11B). This was consistent for vaccine constructs 11 and 12 (Figure 11D). Construct 12 was primarily located at the periphery, consistent with nonspecific binding at the surface of the cells that was hypothesized in response to the results of the flow cytometry uptake experiments (Figure 11E). Receptormediated uptake was investigated; cells were preincubated with mannan (20 times excess) before incubation with 9. Preincubation significantly reduced uptake with the compound predominantly localized on the cells surface (Figure 11C). This correlated with the mannan competition assay results reported in Figure 10. Here, differences observed in the light emission by 11 when compared to 9 can be associated with resolution limitations in confocal microscopy.59 Real Time Surface Plasmon Resonance (SPR). Label-free SPR was performed to confirm the affinity of vaccine constructs 9−14 toward the recombinant human macrophage MR protein.

Figure 8. RP-HPLC trace for the copper-mediated click reaction between 3 and 7 for the formation of vaccine construct 9. (A) 3, Rt 20.83 min; (B) 9 (a diastereomeric mixture), Rt 21.2 min, 21.6 min, 21.8 min; (C) 7 Rt 23.5 min. Click reaction was performed in DMSO at 50 °C under an N2 atmosphere and stopped at 5 h.

Interestingly, even the constructs that contained no mannose units in their structure (12 and 14) had a lower uptake in the competition assay, indicating that the peptide and/or lipids have nonspecific binding and/or uptake into these cells (Figure 10), although in the case of 14, this difference was not significant. It has been shown that lipids can target and bind the CRDs on the C-type lectin receptors.55 Unlike 12 (lipids present in structure) whose high binding ability was significantly inhibited in the presence of mannan, the binding of 13 (no lipids) was not inhibited (Figure 10A,B). This indicated that lipids, and not the antigen, directed the binding and uptake by the tested cell lines. Our results showed that the uptake of vaccine constructs that contained both mannose and lipid moieties (9−11, Figure 10) was higher than that of constructs that contained no lipids (13) or no lipid and mannose moieties (14). However, 12 (which did not have any mannose units in its structure) had a higher uptake than 13 or 14. The addition of hydrophobic moieties (such as lipids) has been shown to increase nonspecific binding to the surface of cells, which could account for the higher level

Figure 9. TEM images of vaccine constructs 9−14. (A) 9, (B) 10, (C) 11, (D) 12, (E) 13, and (F) 14. The scale bar is 200 nm. H


Article


Figure 10. Mannan competition study. Vaccine constructs 9−11 (0.5 μM) and control constructs 12−14 (0.5 μM) were assessed in (A) CD11c positive and (B) F4/80 positive cells for uptake in the presence or absence of mannan (construct + mannan (1 mg/mL) vs construct alone). Dextran-FITC (1 mg/mL) was used as the positive control and PBS was used as the negative control. The study was performed in triplicate and the results were analyzed using one-way ANOVA with mean ± SD from 3 independent experiments, p < 0.0001.

shown to have a strong KD of 0.15 μM with an R2 value of 0.9943 (Figure 12A). Results indicate that vaccine construct 10 (KD = 0.83, R2 = 0.9950, Figure 12C) had the highest level of binding (RU value) and affinity (KD value) when compared with vaccine constructs 9−12 which had binding affinities between 2.9 and 22.8 (Figure 12B−E). Here, the lower the KD the better the affinity the construct has to the receptor. Furthermore, the consistency of the results at varying concentrations was indicated by the R2 values with a good value being close to 1.00. These results suggest that the optimal affinity and binding was dependent upon the compounds structural properties. Compound 10 was shown to have a similar particle size range as per compounds 9 and 11, despite 10 containing only one alanine spacer between the mannose moieties (Figure 9). Therefore, this increase in binding level can be associated with its conformation and structural properties. However, future work with vaccine constructs bearing different types of spacers (e.g., hydrophilic spacers) may provide more information about the relationship between linker geometry and binding affinity. Furthermore, construct 12 (control) containing only lipids in the structure (no mannose) also showed a very high KD (low affinity) of 128 (R2 = 0.9744, Figure 12G) when compared to constructs 9−11. It is well-known that the MR binds to the lipoarabinomannan structures, an exogenous glycolipid antigen moiety found on the cell walls of mycobacteria.67 The MR has also been shown to play an important role in the uptake of lipidic structures.68 Additional studies with the aim of investigating the role lipids play in MR binding are needed for future work. Interestingly, our results indicate that even with an absence of mannose clusters in the vaccine constructs, control compound 13, which contains two mannose moieties in close proximity and no lipids, had a strong affinity to the recombinant MR protein (Figure 12F). However, it can be observed from the Rmax that the level of binding is low (Figure 12F). Additionally, 14 containing no mannose and no lipids also showed a very low affinity (KD of 128 μM, R2 = 0.9744, Figure 12G). Overall, the poor level of binding obtained for constructs 13−14 suggests that, for a better interaction with the MR, OVA-based vaccine constructs need to contain both mannosyl and lipidic moieties.

SPR provides information pertaining to the specificity of molecular interactions, kinetics, and affinity data. The advantage of this technique compared to more traditional methods (e.g., ELISA) for binding-affinity measurements includes having fewer experimentation steps, and the generation of faster results and kinetics data.60 Measuring kinetic constants by optical biosensors based on the molecular interactions is a common valid method applied in many studies.59,61,62 Among these presented studies, Biacore appears to be the most widely used instrument to assess binding affinity of variable constructs.59 The MR protein contains 8 CRDs and has long been the center of interest for the design of targeted peptide, protein, or polymer-based vaccine constructs.16,63 Here, we investigated the differences in the binding affinities of the vaccine constructs toward the recombinant human macrophage MR protein using SPR on a CM-5 sensor chip coated with a dextran polymer. The recombinant MR has 8 CRDs that have been proven to have one (or two binding motifs) involved in the binding of carbohydrates.63 The relative distance between the two mannose moieties in each construct and the presence (or absence) of lipids on the ability to bind this recombinant human macrophage MR protein was investigated. Here, binding to some degree of all vaccine constructs toward the MR was observed (Figure 12). Here, binding interactions between ligand (receptor) and analyte (vaccine constructs) occur in two steps. In the first step, the analyte diffuses from the buffer to the surface of the chip and is known as “mass transfer”. The second stage is when the actual binding occurs. Mass transfer, one of the limiting factors in the application of SPR, is known to affect the accuracy of affinity (KD) calculations. To avoid this issue to achieve an accurate binding affinity calculation, an equilibrium binding measurement was performed to determine the steady state binding characteristics of peptides to the MR protein.64,65 Results are presented as the steady state binding level (Req) as a function of analyte concentration (Figure 12) and the KD (affinity constant). Mannan has a reported high binding affinity to the macrophage MR and has also been used as an activity control for binding assays on the recombinant human MR protein.63,66 In this study, mannan was used as the positive control and for the first time using Biacore technology was I


Article


Figure 11. Confocal images of F4/80+ cells (2 × 105) incubated with vaccine constructs 9, 11, and 12 (0.5 μM) in the presence or absence of mannan, and commercially available mannan-FITC as the positive control. Cells were plated on polylysine-treated culture slides. FAM-labeled vaccine constructs (green) were added to the cells and incubated in the presence and absence of mannan for 4 h at 37 °C. Nuclei were stained with HOECHST stain (blue). F4/80+ cells were identified with anti-F4/80 antibody (red): (A) mannan-FITC; (B) vaccine construct 9; (C) cells preincubated with mannan followed by addition of 9; (D) vaccine construct 11; (E) vaccine construct 12. Analysis was performed on a GE DeltaVision Deconvolution confocal microscope at 60× using oil immersion, graphical scales 15 and 5 μm.

■

Overall, interesting findings from this SPR study were obtained supporting observations made in the uptake and binding studies performed using flow cytometry and confocal microscopy experiments on DCs and macrophage isolated from the spleens of mice (Figures 10 and 11). Both mannose and lipidic moieties appear to be necessary for the enhanced uptake and binding of these vaccine constructs to the MR. It was also confirmed that structural properties, such as the distance between the mannose units, plays any important role in binding.

CONCLUSION

We have successfully developed a library of fluorescently labeled mannosylated lipopeptides using Fmoc SPPS. Peptides were designed to study the relationship between the distance separating the mannose units and the APC receptor-binding properties of the vaccine construct. A mannosylated Fmocserine amino acid enabled successful incorporation of the mannose unit into the peptide during synthesis on resin. Characterization of the library showed nanoparticle formation which was attributed to the amphiphilic properties of the vaccine constructs. A mannan inhibition study used flow J


Article


Figure 12. Surface plasmon resonance analysis for the concentration-dependent binding of mannosylated constructs 9−11 and controls 12−14 with the immobilized extracellular domain of the recombinant human macrophage mannose receptor (MR) protein. Solutions of 9−14 (0.78−50 μM concentration range) and mannan (18.75−600 μg/mL) was prepared in the running buffer (10 mM HEPES, 1 mM CaCl2, 1 mM MgCl2,150 mM NaCl, 0.005% P20; pH 7.4) and were injected over a period of 1.5 min (2 min for mannan), with a dissociation interval of 8 min. The graphs show the plotted level of binding in the steady state (Req) against different concentrations of (A) the positive control mannan, and the vaccine constructs (B) 9, (C) 10, (D) 11, (E) 12, (F) 13, and (G) 14. R2 indicates the fit of the results data to the curve with respect to concentration. The affinity constant (KD) was calculated using the response level at equilibrium using eq 1.

chem (Laufelfingen, Switzerland) and Mimotopes (Melbourne, Australia). N,N′-Dimethylformamide (DMF), trifluoroacetic acid (TFA), methanol (MeOH), N,N′-diisopropylethylamine (DIPEA), 1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b] pyridinium 3-oxid hexafluorophosphate (HATU), N,N,N′,N′-tetramethyl-O-(1H-benzotriazol-1-yl)uronium hexafluorophosphate O-(Benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate (HBTU), piperidine, and dichloromethane (DCM) were obtained from Merck (Hohenbrunn, Germany). N,N′-Diisopropylcarbodiimide (DIC), triisopropylsilane (TIS), 4-dimethylaminopyridine (DMAP), boron trifluoride diethyl etherate (BF3·Et2O), copper wires, and 5(6)-carboxyfluorescein (FAM) were purchased from Sigma-Aldrich (St. Louis, MO, USA). High-pressure liquid chromatography (HPLC)-grade acetonitrile (MeCN) was purchased from RCI Labscan Ltd. (Bangkok, Thailand). Fmoc-protected amino acids (L-configuration) and Rink amide MBHA resin (100−200 mesh, 0.8 mmol/g) were obtained from Novabiochem (Melbourne, VIC, Australia) or Mimotopes (Clayton, VIC, Australia). Phenol free IMDM

cytometry and confocal microscopy to confirm receptormediated uptake of the mannosylated constructs. This study confirmed that structures containing both mannose and lipids had enhanced uptake in F4/80 + and CD11c + cells. Furthermore, structures with lipid alone (no mannose) were also shown to have binding affinity toward the MR (in uptake studies) confirmed through mannan inhibition and SPR studies. SPR using Biacore technology was used to study the binding and affinity of binding between the vaccine constructs and the recombinant human macrophage MR. Both mannose and lipid moieties were found to be necessary components of an APC targeting vaccine, supporting previous findings in the uptake and inhibition studies. The findings also confirmed a low affinity between the OVA 323−339 peptide epitope and the MR. Future work aims to investigate the in vivo immunological properties of receptor targeting vaccine constructs.

■

MATERIALS AND METHOD General. All reagents used were of the highest available grade. Protected amino acids were purchased from NovabioK


Article


Fmoc-Deprotection. Each Fmoc-deprotection consisted of resin treatment with 20% piperidine in DMF (2 × 5 min, 50 °C) followed by extensive DMF washing. 1-(4,4-Dimethyl-2,6-dioxyacyclohexylidene)ethyl (Dde) and 1-(4,4-Dimethyl-2,6-dioxo-cyclohexylidene)-3-methylbutyl (ivDde) Deprotection. Dde and ivDde protecting groups were removed manually by treatment of the resin with 2% hydrazine in DMF solution (12 × 15 min) followed by extensive DMF washing. Amino Acid Coupling. For Manual Synthesis. N-Fmocprotected amino acids (4.2 equiv) were preactivated with HATU (4 equiv) and DIPEA (5 equiv) for 5 min and coupled to the resin twice for at least 30 min. Lipo-amino acids (4.2 equiv) were prepared by dissolving them in a HATU/DMF solution (4.0 equiv) then DIPEA (6.2 equiv) and coupling to the resin for 30 min. For Microwave Synthesis. N-Fmoc-protected amino acids (4.2 equiv) were preactivated with HATU (4 equiv) and DIPEA (5 equiv) for 5 min and coupled to the resin (2 × 5 min at 70 °C). The coupling of Fmoc-Arg(pbf)-OH was carried out at 25 °C (10 min) and 50 °C (10 min), respectively. For the peptide dendrimers, quantities of amino acids and coupling agents were used in ratios relative to the number of branches. Resin Cleavage. After the synthesis was completed and the terminal Fmoc group removed, the resin was washed with DMF, MeOH, and DCM, and dried in vacuo. The crude peptides were cleaved from the resin using a 3 h incubation with 95% TFA, 2.5% TIS, and 2.5% water. All cleaved peptides were precipitated, filtered, and washed thoroughly with ice-cold diethyl ether. The precipitated compounds were dissolved in MeCN:water (1:1) that contained 0.1% TFA and lyophilized overnight to give amorphous powders. General Methods for Peptide Azide Synthesis and Characterization (3−6). Peptides (3−5, Figure 3) were assembled on Rink amide MBHA resin (0.1 mmol scale) using the in situ neutralization protocol for Fmoc SPPS.69 After loading the resin with Fmoc-lysine(Mtt)-OH, the Mtt group was deprotected using TFA:TIS:DCM (1:2:97) in 20 cycles (5 min/cycle) and the resin was washed well with DCM after each the cycle. Azidoacetic acid (4.2 equiv) was synthesized according to a previously published method34 and then coupled to the resin using HATU (4.0 equiv) and DIPEA (6.2 equiv) in DMF overnight. N-Fmoc-O-(2,3,4,6-tetra-O-acetyl-α-D-mannopyranosyl)-L-serine ( 2, 1.5 equiv) was coupled to the resin using HATU (1.5 equiv) and DIPEA (4.5 equiv) in DMF for 12 h followed by washing with DMF (3 × 5 min). A subsequent coupling was performed when ninhydrin testing indicated it was required. FAM (1.5 equiv) was coupled to the resin using HOBt (2.5 equiv) and DIC (2.5 equiv) in DMF overnight. Removal of the unwanted FAM ester-bond was achieved by treating the resin with 20% piperidine (6 × 5 min or until no more color was observed in the washes) followed by extensive washing with DMF.37 Once the peptide sequence was complete, the resin was washed with DMF, MeOH, and DCM, and dried under vacuum overnight. Cleavage and purification of the crude peptides was achieved by dissolving the crude products in solvent B and purified by preparative RP-HPLC on a C8 column with a gradient of 30% to 80% solvent B over 45 min. Peptides 3−5 were characterized by ESI-MS (Table 1). Synthesis and Characterization of OVA323−339 Lipopeptide and Peptide Alkynes (7−8). Lipopeptide 7 was synthesized using microwave-assisted Fmoc SPPS on Rink Amide resin (0.2 mmol) using an in situ neutralization

Glutamax medium, 2-mercaptoethanol, and streptomycin was purchased from Gibco (Life Technologies). Anti-mouse CD11c Alexa Fluor 700 was purchased from eBioscience (Science Center Drive, San Diego, CA, USA) and APC/Cy7 anti-mouse F4/80 antibody, while erylysis buffer was purchased from Sigma-Aldrich. FACS buffer (PBS, 0.02% sodium azide, 0.5% BSA) was purchased from BD Biosciences (North Ryde, NSW, Australia). Analytical Reverse Phase (RP)-HPLC was carried out on a Shimadzu instrument with an LC-20AB pump, a SIL20AHT autosampler and an SPD-M10A detector set to a wavelength of 214 nm. Preparative RP-HPLC was carried out on a Shimadzu system equipped with a CBM-20A controller, LC-20AT pump, SIL-10A autosampler, SPD-20A UV/vis detector set to a wavelength of 214 nm and a FRC-10A fraction collector. Peptides 2−6 and 8 were purified using a Vydac C18 column (10 μm, 22 mm × 250 mm). Peptide 7 was purified using a Vydac C4 column (10 μm, 22 mm × 250 mm). Peptides 9−14 were purified using a Vydac C8 column (10 μm, 22 mm × 250 mm). Mobile phases were solvent A (0.1% TFA in water); solvent B (90:10:0.1% MeCN:water:TFA). Peptide purity was checked using analytical RP-HPLC on a Vydac C18 column (5 μm, 4.6 mm × 250 mm) or a Vydac C4 (5 μm, 4.6 mm × 250 mm). NMR studies were performed in CDCl3 on a Bruker Avance 300 MHz instrument (Bruker Biospin, Germany). Electrospray ionization mass spectrometry (ESIMS) was performed on a PE sciex AP13000 triple quadrupole mass spectrometer operated with a constant flow of 1:1 mixture of solvent A (0.1% acetic acid in water) and solvent B (90:10:0.1% acetonitrile:water:acetic acid) at a rate of 0.5 mL/ min. Synthesis of peptides 7 and 8 was carried out using a CEM Discovery microwave (20 W, CEM Corporation, Matthews, NC, USA) with temperature-resistant, open vessels. Biacore experimentation was performed using a biacore 3000 instrumentation (BIAcore, GE Healthcare, NSW, Australia) on a CM-5 chip provided from GE Healthcare-Australia (Silverwater, NSW, Australia). 1-Ethyl-3-(3-(dimethylamino)propyl)carbodiimide (EDC), N-hydroxysuccinimide (NHS), ethanolamine, 2-[4-(2-hydroxyethyl)piperazin-1-yl]ethanesulfonic acid (HEPES), and CaCl2 were obtained from Sigma-Aldrich NSW, Australia). Glycine, MgCl2, and NaCl were obtained from Chem-Supply (Gillman, South Australia). Surfactant P20 was obtained from GE Healthcare-Australia (Silverwater, NSW, Australia). The recombinant human macrophage mannose receptor (catalogue number 2534-MR/ CF) was purchased from In Vitro Technologies Pty. Ltd. (Brisbane, Australia). Synthesis of N-Fmoc-O-(2,3,4,6-tetra-O-acetyl-α,β-Dmannopyranosyl)-L-serine ( 2). The Fmoc-serine mannosylated amino acid, N-Fmoc-O-(2,3,4,6-tetra-O-acetyl-α,β-D-mannopyranosyl)-L-serine (2, Figure 2) was synthesized by first acetylating the mannose hydroxyl groups (1, Figure 2) to form α,β-mannose pentaacetate using a previously published method.28 This intermediate was then coupled with Fmocserine-OH in a BF3 Lewis acid catalyzed reaction to generate the title compound 2 (Figure 2 ).33 Compound 2 was purified in a 28% overall yield using preparative RP-HPLC/C18 column using a gradient of 20−80% over 45 min with a Rt of 23.6 min. NMR and mass spectrometry data were consistent with previously published results. Resin Loading. Fmoc SPPS was assisted by a CEM Discovery microwave peptide synthesizer and/or manual shaking (Shaker WS/180°, Glas-Col) on Rink amide MBHA resin. Resin was swollen in DMF overnight before use. L


Article


density of 2 × 105 cells per well. After 4 h incubation, cells were treated with vaccine constructs 9−14 at a concentration of 0.5 μM. After 24 h at 37 °C in a humidified atmosphere of 5% CO2 and 95% air, the adherent cells were scraped from the plate, centrifuged, and resuspended in buffer that contained antiCD11c and anti-F4/80 antibodies for 30 min at 4 °C.50 The cells were centrifuged and resuspended in 0.5 mL of FACS buffer and analyzed using a LSR II flow cytometer (BD Biosciences).50 To investigate receptor-mediated uptake, a mannan competition assay was performed.15 Cells were treated with mannan (1 mg/mL) 1 h prior to incubation with the vaccine constructs (0.5 μM). Cell washing, processing, and antibody staining were performed as stated previously. Data are presented as mean ± standard deviation for triplicate samples. Differences between groups were determined using one-way analysis of variance (ANOVA) followed by the Tukey test and were considered statistically significant if the P value was

Synthesis of Mannosylated Lipopeptides with Receptor Targeting

Recommend Documents