- Email: [email protected]
Dual-targeting peptide probe for sequence- and structure-sensitive sensing of serum albumin
Biosensors and Bioelectronics 94 (2017) 657–662
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Dual-targeting peptide probe for sequence- and structure-sensitive sensing of serum albumin ⁎
Yang Yua,b, Yanyan Huanga,b, , Yulong Jina,b, Rui Zhaoa,b,
MARK
⁎
a Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Analytical Chemistry for Living Biosystems, CAS Research/Education Center for Excellence in Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, China b University of Chinese Academy of Sciences, Beijing 100049, China
A R T I C L E I N F O
A BS T RAC T
Keywords: Peptide probe Surface plasmon resonance imaging Sequence-sensitive Serum albumin Dual-targeted interaction
Peptide-protein interactions mediate numerous biologic processes and provide great opportunity for developing peptide probes and analytical approaches for detecting and interfering with recognition events. Molecular interactions usually take place on the heterogeneous surface of proteins, and the spatial distribution and arrangement of probes are therefore crucial for achieving high specificity and sensitivity in the bioassays. In this study, small linear peptides, homogenous peptide dimers and hetero bivalent peptides were designed for sitespecific recognition of human serum albumin (HSA). Three hydrophilic regions located at different subdomains of HSA were chosen as targets for the molecular design. The binding affinity, selectivity and kinetics of the candidates were screened with surface plasmon resonance imaging (SPRi) and fluoroimmuno assays. Benefiting from the synergistic effect from the surface-targeted peptide binders and the flexible spacer, a heterogenetic dimer peptide (heter-7) with fast binding and slow dissociation behavior was identified as the optimized probe. Heter-7 specifically recognizes the target protein HSA, and effectively blocks the binding of antibody to HSA. Its inhibitory activity was estimated as 83 nM. It is noteworthy that heter-7 can distinguish serum albumins from different species despite high similarities in sequence and structure of these proteins. This hetero bivalent peptide shows promise for use in serum proteomics, disease detection and drug transport, and provides an effective approach for promoting the affinity and selectivity of ligands to achieve desirable chemical and biological outcomes.
1. Introduction Peptide-protein interactions mediate a wide range of biological processes including signaling pathways, DNA replication machinery, protein trafficking and immune response, and also control many protein-protein interactions (Pawson and Nash, 2003; Wu et al., 2010; Xiong et al., 2002). Based on such multifarious function, peptide-protein interactions have generated interest in the fields of chemistry, biology and medicine (Ding et al., 2016; Hossain et al., 2016; Lewin et al., 2000). Understanding the principles of these interactions as they occur naturally, and discovery of new interaction partners is critical for drug screening, disease treatment, as well as for development of effective bioanalytical techniques (An et al., 2015; Gray and Brown, 2014; Healy et al., 2015). Considerable effort has been made to advance understanding of basic and applied aspects of peptide-protein binding events. Peptide libraries (Larman et al., 2011; Wu et al., 2014; Gray and Brown, 2014; Peng et al., 2006) and array technologies (Doran and Kodadek, 2014)
⁎
provide plentiful resources to identify peptide binders to target proteins. Progress in analytical methods provides effective strategies for discovery and characterization of these interactions (Debnath et al., 2011; Diana et al., 2015; Frei et al., 2013; Goh et al., 2014). With increasing knowledge, peptide-mediated interactions have been shown to have small binding interfaces, fast response to stimuli, and operation through short linear sequences on the protein surface (Petsalaki and Russell et al., 2008). These are beneficial for manipulating cellular processes, blocking disease pathways and detecting target molecules (Hu et al., 2014; Zhou et al., 2016; Lipp et al., 2015; Hwang et al., 2017). However, the above features also bring moderate affinity and specificity as compared with protein-protein interactions, and may be difficult to handle biochemically. This makes application of these peptides in bioassays of complex biosystems challenging. To address the problems of affinity and specificity in peptide probes, mimicking natural protein interactions is an attractive way (Pelay-Gimeno et al., 2015; Shiba, 2010; Chen et al., 2014; Grigoryan et al., 2009; Lawson et al., 2013). Multimetric ligands in particular are
Correspondence to: CAS Key Laboratory of Analytical Chemistry for Living Biosystems, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, China. E-mail addresses: [email protected] (Y. Huang), [email protected] (R. Zhao).
http://dx.doi.org/10.1016/j.bios.2017.03.067 Received 9 December 2016; Received in revised form 24 March 2017; Accepted 31 March 2017 Available online 02 April 2017 0956-5663/ © 2017 Published by Elsevier B.V.
Biosensors and Bioelectronics 94 (2017) 657–662
Y. Yu et al.
pargyl-Gly-2 and azido-PEG-3 were obtained and characterized (the cleavage conditions were provided in Supplementary Material). Click reaction was carried out with 1 equiv of azido-PEG-3 (3.0 mg, 2.5 mM) and 1.2 equiv of propargyl-Gly-2 (1.5 mg, 3 mM) in DMF containing 1% (v/v) trimethylamine. CuSO4 (0.3 mg, 3 mM) and sodium ascorbate (6 mM) were added to catalyze the reaction. The mixture was degassed and purged with nitrogen. The reaction vessel was sealed and the mixture was stirred for 2 h at room temperature. After purification, heter-7 was obtained as a white solid in 10.7% yield, and characterized with HPLC and high resolution mass spectrometry (Table S1 and Fig. S14).
effective due to the benefit of synergistic effects (Janssen et al., 2013; Joshi et al, 2008; Vance et al., 2009). Covalently attaching multiple copies of recognition elements to a molecular scaffold yields peptide dendrimers which possess enhanced binding ability and functionality. Successful application of multimeric peptides has been demonstrated in small antibody mimetics, bioimaging and drug delivery (Cai et al., 2014; Chu et al., 2015; Rao et al., 1998; Wan and Alewood, 2016). Currently various branched structures such as polylysine, polyglycerol and poly(amidoamine) are available, and afford diverse functional surface groups for chemical ligation and fine control of conjugation chemistry (Cheng et al., 2011; Choi et al., 2000). The spatial distribution and surface arrangement of these ligands are critical for achieving the desired binding behavior towards different target proteins (Kojima et al., 2014; Kwok et al., 2013; Wan et al., 2015; Xu et al., 2016). Moreover, compared to well-characterized homomeric dendrimers with multiple identical peptide sequences, branched peptides with heterogeneous targeting ability still need further exploration. Human serum albumin (HSA) is the most abundant protein in plasma, and plays crucial roles in maintaining blood pressure, buffering plasma pH and transporting various endogenous and exogenous species (Fanali et al., 2012; He and Carter, 1992; Yu et al., 2016; Zhang et al., 2008). HSA is also frequently used as a therapeutic reagent in patients during the restoration of lost fluid and blood (Fan et al., 2014). Recent studies have revealed HSA has a long half-life and an intrinsic capability to extravasate and accumulate in solid tumors. These findings render HSA useful for cancer treatment (Byeon et al., 2014; Li et al., 2016). BSA with high structural similarity to HSA but much cheaper is usually used as an alternative to HSA in many biochemical applications. However, misuse of BSA instead of HSA to patients can lead to life-threatening injury (Fan et al., 2014; Sohl and Splittgerber, 1991; Bujacz, 2012). Therefore, it is of great importance to develop targeting probes and biosensors which recognize HSA but not BSA (Reja et al., 2016; Fan et al., 2014; Prasad et al., 2013). In this study, small linear peptides, homo peptide dimers and hetero targeting peptides were designed to recognize three hydrophilic regions located on the surface of different HSA subdomains. The affinity and selectivity of these molecules were screened with surface plasmon resonance imaging (SPRi) and fluoroimmuno assays. To give more insight into these peptide-protein interactions, binding kinetics were also monitored in real-time. A hetero dimer peptide was identified as the optimum probe. Its binding specificity, structure sensitivity and inhibitory activity against antibody binding to HSA was also examined. The interactions between the hetero dimer and serum albumins were further analyzed with molecular modeling.
2.3. SPRi sensing SPRi analysis was measured on a Plexera PlexArray HT system (Plexera LLC, Bothell, WA). The system was operated at a flow rate of 2 μL/s. PBS buffer was employed as the running buffer throughout given its compatibility with the analysis of serum albumins (Anees et al., 2014; Canoa et al., 2015; Shim and Reaney, 2015; Jiang et al., 2014). The chip surface was first balanced with PBS to obtain a baseline. HSA sample solutions of different concentrations were then injected for interaction with immobilized peptides. After 300 s, dissociation was brought about by injection of PBS. The sensor chip surface was regenerated with 0.5% (v/v) H3PO4. Real-time binding signals were recorded and analyzed with PlexArray HT software. Kinetic analysis was performed using BIA evaluation. 2.4. Fluorescence-Linked Immunosorbent Assay (FLISA) In a 96-well plate, aliquots (100 μL) of 5% HSA (w/v) in sodium carbonate buffer (0.1 M, pH 9.6) were added to each well. After incubation at 4 °C overnight, the plates were washed with PBST (PBS with 0.05% Tween-20) once, and then washed twice with PBS. Different peptides (0.15 mM, in PBS) were then added into the wells and incubated for 2 h at 37 °C. After washing with PBS three times, 100 μL FITC-labeled antibody against HSA (10 μg/mL) was added and incubated for 1 h at 37 °C to allow binding with HSA. After the final washing step, fluorescence was measured using a microplate reader. The excitation wavelength was set at 490 nm, and the emission intensity at 525 nm was recorded. A control experiment was carried out following the same procedure except for the addition of peptide. These experiments were repeated, and all data were obtained in triplicate. 3. Results and discussion
2. Experimental 3.1. Design of peptide monomers 2.1. Peptide synthesis HSA contains three homologous domains (I-III), and each of these is further divided into two separate helical subdomains (A and B) (Carter et al., 1989). This modular structural organization of HSA is the basis of its unique ligand binding ability and biological function. As peptide-protein interactions are mediated by short linear sequences (London et al., 2010; Stein and Aloy, 2008), three 6-mer hydrophilic fragments distributed in different domains of HSA were chosen as the targets (Scheme 1). Residues 82–87 (ETYGEM) located in domain IA were defined as target 1 (T1); Residues 268–273 (QDSISS) and 499– 504 (PKEFNA) are distributed in domain IIA and III and were selected as target 2 (T2) and target 3 (T3) respectively. Evidence suggests that antisense peptides can interact specifically with their corresponding sense peptides (Heal et al., 2002; Miller, 2015). Considering the extended conformation T1-T3, use of antisense peptides provides a suitable means for the effective design of affinity peptide binders. Based on the genetic sequences of T1-T3, three hexapeptides mono-1 (HFTIGF), mono-2 (TGDRIL) and mono-3 (SIKLFG), were readily designed by direct reading of corresponding
Small linear peptides (mono-1–mono-3), homo peptide dimers (homo-4, homo-5) and a hetero targeting peptide (heter-6) were synthesized manually using FMOC solid phase peptide synthesis. FMOC-amino acid-Wang resins were used as the starting material. 4Methylmorpholine was used as activating reagents. HBTU was used as the coupling agent. After elongation, peptides were cleaved from the resins and analyzed with HPLC and MS. The cleavage conditions, yields and purities of the peptides were provided in Supplementary Material. 2.2. Synthesis of the hetero bivalent dimer (heter-7) The copper-catalyzed azide-alkyne cycloaddition reaction was employed for synthesis of hetero bivalent dimer (heter-7). FMOCpropargyl-Gly-OH was attached to the N-terminal of mono-2 during the solid phase peptide synthesis. Azido-PEG(24) carboxylic acid was coupled to the N-terminal amino group of mono-3 using the on-resin condensation reaction. After FMOC deprotection and cleavage, pro658
Biosensors and Bioelectronics 94 (2017) 657–662
Y. Yu et al.
Scheme 1. Schematic illustration of the location of the target sequences in HSA and the design of different peptide binders.
complementary codens (Table S2). All the synthesized peptides were characterized by HPLC and MS (Fig. S1–S3).
Table 1 Kinetic and affinity constants for peptide probes and HSA based on SPRi analysis.
3.2. SPRi sensing of interactions between HSA and peptide monomers The real-time binding behavior of these peptides with HSA was monitored with SPRi. To fabricate the sensor chip, an additional cysteine (Cys) residue was added to the N-terminal of each peptide. This facilitated peptide assembly on the gold surface of the chip with Au-S bonds. SPRi analysis was carried out by injecting different concentrations of HSA onto the chip with peptide arrays. Signal changes were detected for all three peptides (Fig. 1), suggesting their interactions with HSA. The SPR response intensified with increasing concentrations of HSA. Under all the concentrations, mono-2 and mono-3 induced greater signal changes than mono-1. Interaction kinetics were analyzed and fitted with BIA evaluation 4.1 software (Table 1). Mono-2 and mono-3 showed faster association rates and slower dissociation rates, demonstrating they were bound to HSA more tightly than mono-1. The calculated equilibrium dissociation constant (KD) of 19.75 ± 0.45 μM demonstrates that mono-3 had the highest binding affinity for HSA. With a KD of 72.15 ± 2.05 μM, mono-2 showed slightly weaker affinity. However, the KD value of mono-1 is one order of magnitude higher (421.5 ± 88.38 μM), indicating its HSA binding is the weakest of these three peptides. To examine the specificity of the peptide sequence, scrambled peptides were synthesized and analyzed with SPRi (Supplementary material, Fig. S4). Although with the same amino acid composition, the change in the arrangement led to greatly weakened binding of the scrambled peptides with HSA. These results manifest the interactions between the designed peptide and HSA are sequence dependent. It has been proposed that peptides with longer sequences usually have higher affinity than shorter ones (Huang et al., 2011; Shai et al., 1987). In an attempt to enhance the moderate binding affinity of these peptides, mono-1~ mono-3 were elongated by adding four residues at the N-terminal. However, there was no enhancement of the response intensity (Fig. S5-S8, Table S3 and S4), indicating that simple elongation does not augment affinity for HSA.
Compounds
ka (× 102 M−1 s−1)
kd (× 10−3 s−1)
KD (μM)
mono-1 mono-2 mono-3 homo-4 homo-5 heter-6 heter-7
0.56 ± 0.03 1.60 ± 0.01 2.35 ± 0.03 13.7 ± 1.65 39.4 ± 0.18 7.67 ± 0.02 126.3 ± 13.7
23.75 ± 3.74 11.6 ± 0.2 4.66 ± 0.35 29.7 ± 0 22.7 ± 1.7 2.55 ± 0.005 0.91 ± 0.02
421.5 ± 88.38 72.15 ± 2.05 19.75 ± 0.45 21.9 ± 2.7 5.38 ± 0.68 3.32 ± 0.02 0.073 ± 0.01
3.3. Design and affinity evaluation of homo-bivalent peptides The high affinity and specificity of antibody depends largely on its “Y”-shape dimeric structure (Calarese et al., 2003). Mono-2 and mono-3 demonstrating stronger affinity for HSA were chosen as recognition units for the design of homo dimer peptides (homo-4 and homo-5) respectively (Fig. S9 and S10). The sequence GGGCGGG was used as the dimer dendron. These multiple glycine residues act as a spacer and provide the recognition moieties with enough flexibility during interaction. Cysteine was incorporated with the intension of immobilizing these dimeric peptides onto the gold surface of the SPR chip through an Au-S bond. After assembly, the antibody mimetic “Y”shaped configuration was established for these bivalent peptides on the sensor surface (Fig. 2). For comparison, a peptide-arrayed sensor chip was prepared by printing hexapeptide monomers (mono-2 and mono-3) and bivalent peptides (homo-4 and homo-5) on the chip, and lower concentrations of HSA were examined. As shown in Fig. 2b, homo-4 and homo-5 induced distinctly higher SPR responses than mono-2 and mono-3. Kinetic analysis demonstrated that these homo bivalent peptides bind HSA much faster (Table 1). Their association rates were more than 10-fold higher than those of monomers, which contributed to affinity enhancement. However, their dissociation rates also became faster, which tends to lower the affinity (Table 1). With 3–4-fold lower KD values, the bivalent peptides homo-4 and homo-5 bound HSA
Fig. 1. SPRi sensing of the interactions of HSA and peptides mono-1 (a), mono-2 (b) and mono-3 (c). [HSA]=9.4, 18.8, 37.6 and 75.3 μM, respectively.
659
Biosensors and Bioelectronics 94 (2017) 657–662
Y. Yu et al.
Fig. 2. (a) Illustration of the SPRi chip with immobilization of peptide monomers and homo dimers for interaction with HSA. (b) SPRi sensing the binding of peptide monomers (mono-2 and mono-3) and homo bivalent peptides (homo-4, homo-5) with HSA (2.4 μM).
fluorescein isothiocyanate (FITC)-labeled antibody to HSA induced significant fluorescence emission (Fig. 3a). Various peptides were introduced to compete with this antibody for HSA binding, and peptides with higher affinity for HSA reduced the amount of bound FITC-labeled antibody leading to decreased fluorescence intensity. Only heter-7 markedly reduced the fluorescence from FITC-labeled HSA antibody (Fig. 3a) with resultant fluorescence intensity of only 28% of that induced by direct binding of FITC-labeled antibody to HSA. This reflects the strong interaction of heter-7 with HSA. No significant change in fluorescence was found with the GGGCGGG-linked heter-6, short linear peptides (mono-2 and mono-3) or homo bivalent peptides (homo-4 and homo-5), demonstrating their lack of capacity to interfere with the binding of antibody to HSA and lower antibody binding affinity. The competition between heter-7 and HSA antibody also suggests that heter-7 and the antibody may recognize the same epitope in HSA. To confirm this, a reversed competitive FLISA assay was carried out (Supplementary material, Fig. S15). The results that heter-7 can compete with the pre-bounded antibody for the binding sites in HSA molecule verify the high specificity of heter-7. It is notable that none of the structural components of heter-7 (PEG, azido-PEG-3 and propargylglycine-modified mono-2) can inhibit fluorescence intensity from FITC-labeled antibody. This implies the strong interaction of heter-7 with HSA results from the cooperative effect of the hetero bivalent structure. To investigate the potency of the hetero dimer peptide heter-7, its inhibitory activity was measured with competitive FLISA. As shown in Fig. 3b, fluorescence decreases as the concentration of heter-7 increases, suggesting the enhancement in blocking the interactions between antibody and HSA. The half maximal inhibitory concentration (IC50) was estimated as 83 nM, reflecting the effectiveness of the hetero bivalent peptide.
more tightly. The use of two copies of recognition units increased the chance for accessing target sites on HSA, which contributes to stronger interactions. Although homo-4 and homo-5 affinity was enhanced, such homo dimeric structures may be more suitable for homomeric targets which have multiple identical sites for interaction with multimeric ligands (Choi et al., 2000). However, the polypeptide chain of an HSA molecule only contains one T2 or T3 sequence. The homo bivalent homo-4 and homo-5 therefore cannot yield an increase in affinity comparable to what has been previously reported (Choi et al., 2000; Wan and Alewood, 2016).
3.4. Hetero-bivalent peptides with high affinity Linking multiple heterologous ligands that bind different receptors is attractive to achieve high affinity and enhance biologic function (Berkov-Zrihen et al., 2013; Hart et al., 2014). Design of heterobivalent peptides targeting different domains within HSA molecule was thus carried out. T2 and T3 located at domains II and III of HSA were chosen as ideal targeting sites for the heterodimer. Accordingly, mono-2 and mono-3 were again used as site-specific recognition moieties. For peptide conjugation, two kinds of linkers are considered, including the peptidic GGGCGGG sequence and polyethylene glycol (PEG). The GGGCGGG-linked hetero dimer heter-6 was prepared with the solid phase synthesis and characterized with HPLC and MS (Fig. S11). The synthesis, structure and characterized of PEG-linked heter-7 are provided in Supplementary Material (Scheme S1 and Fig. S12–S14). To give an overall view of the binding of these peptides with HSA, the fluorescence-linked immunosorbent assay (FLISA) featuring high sensitivity and high throughput was employed. The direct binding of
Fig. 3. (a) Competitive FLISA of various compounds against antibody for HSA binding. (1) FITC-labeled antibody only, with addition of competitors: mono-1 (2), mono-2 (3) and mono-3 (4); homo-4 (5) and homo-5 (6); (7) heter-6; (8) azido-PEG acid; (9) propargylglycine-modified mono-2; (10) azido-PEG-3; (11) heter-7. (b) Inhibitory activity of heter-7 against antibody based on FLISA. I and I0: fluorescence from binding of HSA and its FITC-labeled antibody with and without the addition of heter-7. (c) SPRi sensing curves for the heter-7-immobilized chip after injection of HSA and BSA.
660
Biosensors and Bioelectronics 94 (2017) 657–662
Y. Yu et al.
Fig. 4. Structural analysis of HSA and BSA (PDB code: 2BXM for HSA, 4LUF for BSA). Overviews of the binding sites T2 and T3 in HSA (a), and corresponding sequences in BSA (b). (cf) Conformational comparison of the target sequences in HSA and BSA. The target sequences are shown in red. The structures of albumin molecules were analyzed with autodock 4.2 (Sanner, 1999). Ribbon drawings of HSA and BSA were generated and analyzed with Pymol.
significantly lower than that from HSA (Fig. 3c). The affinity constant (KD) of heter-7 for BSA was estimated as 1335 μM, which is orders of magnitude larger than that for HSA (KD 73 nM). After saturation, the surface bound amount of BSA on the chip was also distinctly smaller than that of HSA (Supplementary Material). These results demonstrate the ability of heter-7 for HSA differentiation. The different responses of heter-7 to HSA and BSA can arise from the tiny differences in the sequence and structure of the target sites in HSA and BSA. For sequence comparison, BSA shares 76% identity with HSA in amino acid composition. The T2 sequences are QDSISS and QDTISS (residues 267–272) in HSA and BSA, respectively (Fig. 4). The one-only residue difference between T2 sites in these two albumins is further narrowed because Ser269 in HSA and Thr269 in BSA both contain a hydroxyl side chain. However, distinct differences are observed in the secondary structures of these sequences in albumins. As shown in Fig. 4c, QDSISS in HSA forms an extended conformation, which is favored by peptide-peptide interactions. The higher content of α-helix conformation in QDTISS of BSA hampers molecular recognition between peptides (Fig. 4d). As for T3, the sequences are PKEFNA and PKAFDE in HSA and BSA respectively (Fig. 4). The target sequences in both HSA and BSA adopt extended conformations (Figs. 4e, f). However, the amino-acid composition and arrangement are different. By substituting asparagine (N) in HSA with aspartic acid (D), the corresponding sequence in BSA contains more negative charge and is more acidic. Moreover, the different location of glutaminic acid (E) and alanine (A) in these two proteins also affects the affinity recognition of the peptide ligand. Based on the above analysis, the selectivity of heter-7 is sensitive to both the amino-acid arrangement and the conformation of the target sites. By targeting two sites on two domains of HSA, the designed bi-branched hetero dimer peptide shows desirable enhancement in affinity and selectivity owing to the cooperation recognition effect.
3.5. Binding kinetics and high selectivity of the hetero dimer To acquire binding kinetics on the interaction of heter-7 with HSA, SPRi was employed. A mercaptoundecanoic acid (MUA)-decorated sensor chip was used for immobilization of heter-7. The preassembled MUA played a role as the spacer and provided heter-7 with flexibility, thus avoiding potential steric hindrance during interaction with HSA. Kinetic data and equilibrium dissociation constants were calculated by fitting the sensor curves. As listed in Table 1, with a rate constant of 126.3× 102 M−1 s−1, heter-7 binds HSA significantly faster than the other peptides. Its dissociation rate (0.91× 10−3 s−1) is one order of magnitude lower, showing heter-7 dissociates from HSA with difficulty. The rapid binding and slow dissociation of heter-7 yield a lowered KD of 73 nM, which is in accordance with its high affinity in competitive FLISA binding assays. Although having the same two recognition moieties as heter-7, heter-6 binds HSA slowly and has a strong tendency to dissociate, thus the affinity is significantly lower (KD 3.32 μM). This is most likely attributable to the difference in the linker. Based on molecular modeling, a minimum length of 99 Å is needed for the linker to achieve effective binding at two sites (T2 and T3) on one HSA molecule (Fig. 4a). PEG has a molecular length of 105.6 Å and can thus accommodate this two-site binding. In contrast, the linker GGGCGGG is only 20.1 Å in length, which is too short to reach both T2 and T3 sites simultaneously on one HSA molecule. Such shortness of the linker impairs the bivalent binding of heter-6 to HSA. Although heter-6 is disabled from two-site binding on one HSA molecule, the hetero bivalent structure increases the chance for contact with target sites during interaction. This may contribute to higher affinity of heter-6 for HSA than other short linear peptides and homo dimers (Table 1). Therefore, both hetero bivalency and a spacer of proper length serve to strengthen the interaction of heter-7 and HSA. To examine the specificity of heter-7, various biomolecules including proteins, peptides, carbohydrate and nucleotides were used as the substitutes for HSA and injected into the SPRi sensor system. Although with different charging characters and hydrophobicity, none of these compounds could induced comparable signal change as HSA did (Fig. S16). It is interesting to find that despite the high sequence and structural similarity of BSA with HSA, the SPRi signal from BSA is
4. Conclusion In this study site-specific and sequence-sensitive peptide binders towards human serum albumin (HSA) were designed, synthesized and analyzed. In comparison with the monomers and homo dimers which were evaluated, the hetero dimer showed fast binding and slow 661
Biosensors and Bioelectronics 94 (2017) 657–662
Y. Yu et al.
Grigoryan, G., Reinke, A.W., Keating, A.E., 2009. Nature 458, 859–864. Hart, N.J., Chung, W.J., Weber, C., Ananthakrishnan, K., Anderson, M., Patek, R., Zhang, Z., Limesand, S.W., Vagner, J., Lynch, R.M., 2014. ChemBioChem 15, 135–145. He, X.M., Carter, D.C., 1992. Nature 358, 209–215. Heal, J.R., Roberts, G.W., Raynes, J.G., Bhakoo, A., Miller, A.D., 2002. ChemBioChem 3, 136–151. Healy, A.R., Houston, D.R., Remnant, L., Huart, A., Brychtova, V., Maslon, M.M., Meers, O., Muller, P., Krejci, A., Blackburn, E.A., Vojtesek, B., Hernychova, L., Walkinshaw, M.D., Westwood, N.J., Hupp, T.R., 2015. Chem. Sci. 6, 3109–3116. Hossain, M.A., Kocan, M., Yao, S.T., Royce, S.G., Nair, V.B., Siwek, C., Patil, N.A., Harrison, I.P., Rosengren, K.J., Selemidis, S., Summers, R.J., Wade, J.D., Bathgate, R.A.D., Samuel, C.S., 2016. Chem. Sci. 7, 3805–3819. Hu, F., Huang, Y., Zhang, G., Zhao, R., Yang, H., Zhang, D., 2014. Anal. Chem. 86, 7987–7995. Huang, Y., Zhao, R., Fu, Y., Zhang, Q., Xiong, S., Li, L., Zhou, R., Liu, G., Chen, Y., 2011. ChemBioChem 12, 1209–1215. Hwang, H.J., Ryu, M.Y., Park, C.Y., Ahn, J., Park, H.G., Choi, C., Ha, S., Park, T.J., Park, J.P., 2017. Biosens. Bioelectron. 87, 164–170. Janssen, B.M.G., Lempens, E.H.M., Olijve, L.L.C., Voets, I.K., J. van Dongen, J.L., A. de Greefc, T.F., Merkx, M., 2013. Chem. Sci., vol. 4, pp. 1442–1450. Jiang, Z., Qin, Y., Peng, Z., Chen, S., Chen, S., Deng, C., Xiang, J., 2014. Biosens. Bioelectron. 62, 268–273. Joshi, A., Vance, D., Rai, P., Thiyagarajan, A., Kane, R.S., 2008. Chem. Eur. J. 14, 7738–7747. Kojima, C., Irie, K., Tada, T., Tanaka, N., 2014. Biopolymers 101, 603–612. Kwok, A., Eggimann, G.A., Reymond, J.L., Darbre, T., Hollfelder, F., 2013. ACS Nano 7, 4668–4682. Larman, H.B., Zhao, Z., Laserson, U., Li, M.Z., Ciccia, A., Gakidis, M.A.M., Church, G.M., Kesari, S., LeProust, E.M., Solimini, N.L., Elledge, S.J., 2011. Nat. Biotechnol. 29, 535–541. Lawson, K.V., Rose, T.E., Harran, P.G., 2013. Proceedings Natl. Acad. Sci. USA, 110, E3753-E3760. Lewin, M., Carlesso, N., Tung, C.H., Tang, X.W., Cory, D., Scadden, D.T., Weissleder, R., 2000. Nat. Biotechnol. 18, 410–414. Li, R., Yang, H., Jia, D., Nie, Q., Cai, H., Fan, Q., Wan, L., Li, L., Lu, X., 2016. J. Control. Release 228, 96–106. Lipp, A., Ji, B., Hager, R., Haas, S., Schweiggl, S., Sonnleitner, A., Haselgrübler, T., 2015. Biosens. Bioelectron. 74, 757–763. London, N., Movshovitz-Attias, D., Schueler-Furman, O., 2010. Structure 18, 188–199. Miller, A.D., 2015. Expert Opin. Biol. Ther. 15, 245–267. Pawson, T., Nash, P., 2003. Science 300, 445–452. Pelay-Gimeno, M., Glas, A., Koch, O., Grossmann, T., 2015. Angew. Chem. Int. Ed. 54, 8896–8927. Peng, L., Liu, R., Marik, J., Wang, X., Takada, Y., Lam, K.S., 2006. Nat. Chem. Biol. 2, 381–389. Petsalaki, E., Russell, R.B., 2008. Curr. Opin. Biotechnol. 19, 344–350. Prasad, B.B., Prasad, A., Tiwari, M.P., 2013. Biosens. Bioelectron. 39, 236–243. Rao, J., Lahiri, J., Isaacs, L., Weis, R.M., Whitesides, G.M., 1998. Science 280, 708–711. Reja, I.S., Khan, I.A., Bhalla, V., Kumar, M., 2016. Chem. Commun. 52, 1182–1185. Sanner, M.F., 1999. J. Mol. Graph Model. 17, 57–61. Shai, Y., Flashner, M., Chaiken, I.M., 1987. Biochemistry 26, 669–675. Shiba, K., 2010. Chem. Soc. Rev. 39, 117–126. Shim, Y.Y., Reaney, M.J.T., 2015. J. Agric. Food Chem. 63, 1099–1106. Sohl, J.L., Splittgerber, A.G., 1991. J. Chem. Educ. 68, 262–264. Stein, A., Aloy, P., 2008. PLoS One 3, E2524. Vance, D., Martin, J., Patke, S., Kane, R.S., 2009. Adv. Drug Deliv. Rev. 61, 931–939. Wan, J., Alewood, P.F., 2016. Angew. Chem. Int. Ed. 55, 5124–5134. Wan, J., Huang, J.X., Vetter, I., Mobli, M., Lawson, J., Tae, H.S., Abraham, N., Paul, B., Cooper, M.A., Adams, D.J., Lewis, R.J., Alewood, P.F., 2015. J. Am. Chem. Soc. 137, 3209–3212. Wu, B., Chien, E.Y.T., Mol, C.D., Fenalti, G., Liu, W., Katritch, V., Abagyan, R., Brooun, A., Wells, P., Bi, F.C., Hamel, D.J., Kuhn, P., Handel, T.M., Cherezov, V., Stevens, R.C., 2010. Science 330, 1066–1071. Wu, S., Núnez, D., Hu, S., Domingo, M.P., Chen, Y., Wei, P., Pardo, J., Gálvez, E.M., Chiou, A., 2014. Biosens. Bioelectron. 56, 159–166. Xiong, J.P., Stehle, T., Zhang, R., Joachimiak, A., Frech, M., Goodman, S.L., Arnaout, M.A., 2002. Science 296, 151–155. Xu, X., Wang, L., Huang, Y., Gao, W., Li, K., Jiang, W., 2016. Anal. Chem. 88, 9885–9889. Yu, Y., Huang, Y., Hu, F., Jin, Y., Zhang, G., Zhang, D., Zhao, R., 2016. Anal. Chem. 88, 6374–6381. Zhang, Q., Huang, Y., Zhao, R., Liu, G., Chen, Y., 2008. Biosens. Bioelectron. 24, 48–54. Zhou, J., Du, X., Yamagata, N., Xu, B., 2016. J. Am. Chem. Soc. 138, 3813–3823.
dissociation behavior, which results in strong binding with HSA. According to SPRi analysis, the KD value of heter-7 was 73 nM, which is orders of magnitude lower than those of other peptides. In competitive fluoroimmuno assays with antibody, heter-7 demonstrated the site-specific and competing binding behavior for HSA. Benefiting from the cooperative effect of the site-specific binders and the PEG linker, heter-7 can differentiate HSA from BSA with sequence and conformation sensitivity. These results demonstrate the advantage hetero bivalent peptide in affinity and selectivity for bioanalysis. With the combination with nanomaterial, sensing techniques or functional agents, such molecular bioprobes shows promise for use in serum proteomics, disease detection and drug transport. Acknowledgements Financial support from National Natural Science Foundation of China (21375134, 21475140, 21675161 and 21621062), the Ministry of Science and Technology of China (2015CB856303), and the Chinese Academy of Sciences is gratefully acknowledged. We also express our gratitude to Dr. Michael A. McNutt from U.S. for his help. Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at doi:10.1016/j.bios.2017.03.067. References An, D., Su, J., Weber, J.K., Gao, X., Zhou, R., Li, J., 2015. J. Am. Chem. Soc. 137, 8412–8418. Anees, P., Sreejith, S., Ajayaghosh, A., 2014. J. Am. Chem. Soc. 136, 13233–13239. Berkov-Zrihen, Y., Green, K.D., Labby, K.J., Feldman, M., Garneau-Tsodikova, S., Fridman, M., 2013. J. Med. Chem. 56, 5613–5625. Bujacz, A., 2012. Acta Crystallogr. Sect. D: Biol. Crystallogr. 68, 1278–12289. Byeon, H.J., Min, S.Y., Kim, I., Lee, E.S., Oh, K.T., Shin, B.S., Lee, K.C., Youn, Y.S., 2014. Bioconjug. Chem. 25, 2212–2221. Cai, H., Sun, Z.Y., Chen, M.S., Zhao, Y.F., Kunz, H., Li, Y.M., 2014. Angew. Chem. Int. Ed. 53, 1699–1703. Calarese, D.A., Scanlan, C.N., Zwick, M.B., Deechongkit, S., Mimura, Y., Kunert, R., Zhu, P., Wormald, M.R., Stanfield, R.L., Roux, K.H., Kelly, J.W., Rudd, P.M., Dwek, R.A., Katinger, H., Burton, D.R., Wilson, I.A., 2003. Science 300, 2065–2071. Canoa, P., Simón-Vázquez, R., Popplewell, J., González-Fernández, Á., 2015. Biosens. Bioelectron. 74, 376–383. Carter, D.C., He, X.M., Twigg, P.D., Casale, E., 1989. Science 244, 1195–1198. Chen, S., Bertoldo, D., Angelini, A., Pojer, F., Heinis, C., 2014. Angew. Chem. Int. Ed. 53, 1602–1606. Cheng, Y., Zhao, L., Li, Y., Xu, T., 2011. Chem. Soc. Rev. 40, 2673–2703. Choi, J.S., Joo, D.K., Kim, C.H., Kim, K., Park, J.S., 2000. J. Am. Chem. Soc. 122, 474–480. Chu, D.S.H., Bocek, M.J., Shi, J., Ta, A., Ngambenjawong, C., Rostomily, R.C., Pun, S.H., 2015. J. Control. Release 205, 155–161. Debnath, S., Chatterjee, S., Arif, M., Kundu, T.K., Roy, S., 2011. J. Biol. Chem. 286, 25076–25087. Diana, D., Russomanno, A., Rosa, De. L., Stasi, R.D., Capasso, D., Gaetano, S.D., Romanelli, A., Russo, L., D’Andrea, L.D., Fattorusso, R., 2015. Chem. Eur. J. 21, 91–95. Ding, B., Jasensky, J., Li, Y., Chen, Z., 2016. Acc. Chem. Res. 49, 1149–1157. Doran, T.M., Kodadek, T., 2014. ACS Chem. Biol. 9, 339–346. Fan, J., Sun, W., Wang, Z., Peng, X., Li, Y., Cao, J., 2014. Chem. Commun. 50, 9573–9576. Fanali, G., di Masi, A., Trezza, V., Marino, M., Fasano, M., Ascenzi, P., 2012. Mol. Asp. Med. 33, 209–290. Frei, A.P., Moest, H., Novy, K., Wollscheid, B., 2013. Nat. Protoc. 8, 1321–1336. Goh, W.L., Lee, M.Y., Joseph, T.L., Quah, S.T., Brown, C.J., Verma, C., Brenner, S., Ghadessy, F.J., Teo, Y.N., 2014. J. Am. Chem. Soc. 136, 6159–6162. Gray, B.P., Brown, K.C., 2014. Chem. Rev. 114, 1020–1081.
662
Contents lists available at ScienceDirect
Biosensors and Bioelectronics journal homepage: www.elsevier.com/locate/bios
Dual-targeting peptide probe for sequence- and structure-sensitive sensing of serum albumin ⁎
Yang Yua,b, Yanyan Huanga,b, , Yulong Jina,b, Rui Zhaoa,b,
MARK
⁎
a Beijing National Laboratory for Molecular Sciences, CAS Key Laboratory of Analytical Chemistry for Living Biosystems, CAS Research/Education Center for Excellence in Molecular Sciences, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, China b University of Chinese Academy of Sciences, Beijing 100049, China
A R T I C L E I N F O
A BS T RAC T
Keywords: Peptide probe Surface plasmon resonance imaging Sequence-sensitive Serum albumin Dual-targeted interaction
Peptide-protein interactions mediate numerous biologic processes and provide great opportunity for developing peptide probes and analytical approaches for detecting and interfering with recognition events. Molecular interactions usually take place on the heterogeneous surface of proteins, and the spatial distribution and arrangement of probes are therefore crucial for achieving high specificity and sensitivity in the bioassays. In this study, small linear peptides, homogenous peptide dimers and hetero bivalent peptides were designed for sitespecific recognition of human serum albumin (HSA). Three hydrophilic regions located at different subdomains of HSA were chosen as targets for the molecular design. The binding affinity, selectivity and kinetics of the candidates were screened with surface plasmon resonance imaging (SPRi) and fluoroimmuno assays. Benefiting from the synergistic effect from the surface-targeted peptide binders and the flexible spacer, a heterogenetic dimer peptide (heter-7) with fast binding and slow dissociation behavior was identified as the optimized probe. Heter-7 specifically recognizes the target protein HSA, and effectively blocks the binding of antibody to HSA. Its inhibitory activity was estimated as 83 nM. It is noteworthy that heter-7 can distinguish serum albumins from different species despite high similarities in sequence and structure of these proteins. This hetero bivalent peptide shows promise for use in serum proteomics, disease detection and drug transport, and provides an effective approach for promoting the affinity and selectivity of ligands to achieve desirable chemical and biological outcomes.
1. Introduction Peptide-protein interactions mediate a wide range of biological processes including signaling pathways, DNA replication machinery, protein trafficking and immune response, and also control many protein-protein interactions (Pawson and Nash, 2003; Wu et al., 2010; Xiong et al., 2002). Based on such multifarious function, peptide-protein interactions have generated interest in the fields of chemistry, biology and medicine (Ding et al., 2016; Hossain et al., 2016; Lewin et al., 2000). Understanding the principles of these interactions as they occur naturally, and discovery of new interaction partners is critical for drug screening, disease treatment, as well as for development of effective bioanalytical techniques (An et al., 2015; Gray and Brown, 2014; Healy et al., 2015). Considerable effort has been made to advance understanding of basic and applied aspects of peptide-protein binding events. Peptide libraries (Larman et al., 2011; Wu et al., 2014; Gray and Brown, 2014; Peng et al., 2006) and array technologies (Doran and Kodadek, 2014)
⁎
provide plentiful resources to identify peptide binders to target proteins. Progress in analytical methods provides effective strategies for discovery and characterization of these interactions (Debnath et al., 2011; Diana et al., 2015; Frei et al., 2013; Goh et al., 2014). With increasing knowledge, peptide-mediated interactions have been shown to have small binding interfaces, fast response to stimuli, and operation through short linear sequences on the protein surface (Petsalaki and Russell et al., 2008). These are beneficial for manipulating cellular processes, blocking disease pathways and detecting target molecules (Hu et al., 2014; Zhou et al., 2016; Lipp et al., 2015; Hwang et al., 2017). However, the above features also bring moderate affinity and specificity as compared with protein-protein interactions, and may be difficult to handle biochemically. This makes application of these peptides in bioassays of complex biosystems challenging. To address the problems of affinity and specificity in peptide probes, mimicking natural protein interactions is an attractive way (Pelay-Gimeno et al., 2015; Shiba, 2010; Chen et al., 2014; Grigoryan et al., 2009; Lawson et al., 2013). Multimetric ligands in particular are
Correspondence to: CAS Key Laboratory of Analytical Chemistry for Living Biosystems, Institute of Chemistry, Chinese Academy of Sciences, Beijing 100190, China. E-mail addresses: [email protected] (Y. Huang), [email protected] (R. Zhao).
http://dx.doi.org/10.1016/j.bios.2017.03.067 Received 9 December 2016; Received in revised form 24 March 2017; Accepted 31 March 2017 Available online 02 April 2017 0956-5663/ © 2017 Published by Elsevier B.V.
Biosensors and Bioelectronics 94 (2017) 657–662
Y. Yu et al.
pargyl-Gly-2 and azido-PEG-3 were obtained and characterized (the cleavage conditions were provided in Supplementary Material). Click reaction was carried out with 1 equiv of azido-PEG-3 (3.0 mg, 2.5 mM) and 1.2 equiv of propargyl-Gly-2 (1.5 mg, 3 mM) in DMF containing 1% (v/v) trimethylamine. CuSO4 (0.3 mg, 3 mM) and sodium ascorbate (6 mM) were added to catalyze the reaction. The mixture was degassed and purged with nitrogen. The reaction vessel was sealed and the mixture was stirred for 2 h at room temperature. After purification, heter-7 was obtained as a white solid in 10.7% yield, and characterized with HPLC and high resolution mass spectrometry (Table S1 and Fig. S14).
effective due to the benefit of synergistic effects (Janssen et al., 2013; Joshi et al, 2008; Vance et al., 2009). Covalently attaching multiple copies of recognition elements to a molecular scaffold yields peptide dendrimers which possess enhanced binding ability and functionality. Successful application of multimeric peptides has been demonstrated in small antibody mimetics, bioimaging and drug delivery (Cai et al., 2014; Chu et al., 2015; Rao et al., 1998; Wan and Alewood, 2016). Currently various branched structures such as polylysine, polyglycerol and poly(amidoamine) are available, and afford diverse functional surface groups for chemical ligation and fine control of conjugation chemistry (Cheng et al., 2011; Choi et al., 2000). The spatial distribution and surface arrangement of these ligands are critical for achieving the desired binding behavior towards different target proteins (Kojima et al., 2014; Kwok et al., 2013; Wan et al., 2015; Xu et al., 2016). Moreover, compared to well-characterized homomeric dendrimers with multiple identical peptide sequences, branched peptides with heterogeneous targeting ability still need further exploration. Human serum albumin (HSA) is the most abundant protein in plasma, and plays crucial roles in maintaining blood pressure, buffering plasma pH and transporting various endogenous and exogenous species (Fanali et al., 2012; He and Carter, 1992; Yu et al., 2016; Zhang et al., 2008). HSA is also frequently used as a therapeutic reagent in patients during the restoration of lost fluid and blood (Fan et al., 2014). Recent studies have revealed HSA has a long half-life and an intrinsic capability to extravasate and accumulate in solid tumors. These findings render HSA useful for cancer treatment (Byeon et al., 2014; Li et al., 2016). BSA with high structural similarity to HSA but much cheaper is usually used as an alternative to HSA in many biochemical applications. However, misuse of BSA instead of HSA to patients can lead to life-threatening injury (Fan et al., 2014; Sohl and Splittgerber, 1991; Bujacz, 2012). Therefore, it is of great importance to develop targeting probes and biosensors which recognize HSA but not BSA (Reja et al., 2016; Fan et al., 2014; Prasad et al., 2013). In this study, small linear peptides, homo peptide dimers and hetero targeting peptides were designed to recognize three hydrophilic regions located on the surface of different HSA subdomains. The affinity and selectivity of these molecules were screened with surface plasmon resonance imaging (SPRi) and fluoroimmuno assays. To give more insight into these peptide-protein interactions, binding kinetics were also monitored in real-time. A hetero dimer peptide was identified as the optimum probe. Its binding specificity, structure sensitivity and inhibitory activity against antibody binding to HSA was also examined. The interactions between the hetero dimer and serum albumins were further analyzed with molecular modeling.
2.3. SPRi sensing SPRi analysis was measured on a Plexera PlexArray HT system (Plexera LLC, Bothell, WA). The system was operated at a flow rate of 2 μL/s. PBS buffer was employed as the running buffer throughout given its compatibility with the analysis of serum albumins (Anees et al., 2014; Canoa et al., 2015; Shim and Reaney, 2015; Jiang et al., 2014). The chip surface was first balanced with PBS to obtain a baseline. HSA sample solutions of different concentrations were then injected for interaction with immobilized peptides. After 300 s, dissociation was brought about by injection of PBS. The sensor chip surface was regenerated with 0.5% (v/v) H3PO4. Real-time binding signals were recorded and analyzed with PlexArray HT software. Kinetic analysis was performed using BIA evaluation. 2.4. Fluorescence-Linked Immunosorbent Assay (FLISA) In a 96-well plate, aliquots (100 μL) of 5% HSA (w/v) in sodium carbonate buffer (0.1 M, pH 9.6) were added to each well. After incubation at 4 °C overnight, the plates were washed with PBST (PBS with 0.05% Tween-20) once, and then washed twice with PBS. Different peptides (0.15 mM, in PBS) were then added into the wells and incubated for 2 h at 37 °C. After washing with PBS three times, 100 μL FITC-labeled antibody against HSA (10 μg/mL) was added and incubated for 1 h at 37 °C to allow binding with HSA. After the final washing step, fluorescence was measured using a microplate reader. The excitation wavelength was set at 490 nm, and the emission intensity at 525 nm was recorded. A control experiment was carried out following the same procedure except for the addition of peptide. These experiments were repeated, and all data were obtained in triplicate. 3. Results and discussion
2. Experimental 3.1. Design of peptide monomers 2.1. Peptide synthesis HSA contains three homologous domains (I-III), and each of these is further divided into two separate helical subdomains (A and B) (Carter et al., 1989). This modular structural organization of HSA is the basis of its unique ligand binding ability and biological function. As peptide-protein interactions are mediated by short linear sequences (London et al., 2010; Stein and Aloy, 2008), three 6-mer hydrophilic fragments distributed in different domains of HSA were chosen as the targets (Scheme 1). Residues 82–87 (ETYGEM) located in domain IA were defined as target 1 (T1); Residues 268–273 (QDSISS) and 499– 504 (PKEFNA) are distributed in domain IIA and III and were selected as target 2 (T2) and target 3 (T3) respectively. Evidence suggests that antisense peptides can interact specifically with their corresponding sense peptides (Heal et al., 2002; Miller, 2015). Considering the extended conformation T1-T3, use of antisense peptides provides a suitable means for the effective design of affinity peptide binders. Based on the genetic sequences of T1-T3, three hexapeptides mono-1 (HFTIGF), mono-2 (TGDRIL) and mono-3 (SIKLFG), were readily designed by direct reading of corresponding
Small linear peptides (mono-1–mono-3), homo peptide dimers (homo-4, homo-5) and a hetero targeting peptide (heter-6) were synthesized manually using FMOC solid phase peptide synthesis. FMOC-amino acid-Wang resins were used as the starting material. 4Methylmorpholine was used as activating reagents. HBTU was used as the coupling agent. After elongation, peptides were cleaved from the resins and analyzed with HPLC and MS. The cleavage conditions, yields and purities of the peptides were provided in Supplementary Material. 2.2. Synthesis of the hetero bivalent dimer (heter-7) The copper-catalyzed azide-alkyne cycloaddition reaction was employed for synthesis of hetero bivalent dimer (heter-7). FMOCpropargyl-Gly-OH was attached to the N-terminal of mono-2 during the solid phase peptide synthesis. Azido-PEG(24) carboxylic acid was coupled to the N-terminal amino group of mono-3 using the on-resin condensation reaction. After FMOC deprotection and cleavage, pro658
Biosensors and Bioelectronics 94 (2017) 657–662
Y. Yu et al.
Scheme 1. Schematic illustration of the location of the target sequences in HSA and the design of different peptide binders.
complementary codens (Table S2). All the synthesized peptides were characterized by HPLC and MS (Fig. S1–S3).
Table 1 Kinetic and affinity constants for peptide probes and HSA based on SPRi analysis.
3.2. SPRi sensing of interactions between HSA and peptide monomers The real-time binding behavior of these peptides with HSA was monitored with SPRi. To fabricate the sensor chip, an additional cysteine (Cys) residue was added to the N-terminal of each peptide. This facilitated peptide assembly on the gold surface of the chip with Au-S bonds. SPRi analysis was carried out by injecting different concentrations of HSA onto the chip with peptide arrays. Signal changes were detected for all three peptides (Fig. 1), suggesting their interactions with HSA. The SPR response intensified with increasing concentrations of HSA. Under all the concentrations, mono-2 and mono-3 induced greater signal changes than mono-1. Interaction kinetics were analyzed and fitted with BIA evaluation 4.1 software (Table 1). Mono-2 and mono-3 showed faster association rates and slower dissociation rates, demonstrating they were bound to HSA more tightly than mono-1. The calculated equilibrium dissociation constant (KD) of 19.75 ± 0.45 μM demonstrates that mono-3 had the highest binding affinity for HSA. With a KD of 72.15 ± 2.05 μM, mono-2 showed slightly weaker affinity. However, the KD value of mono-1 is one order of magnitude higher (421.5 ± 88.38 μM), indicating its HSA binding is the weakest of these three peptides. To examine the specificity of the peptide sequence, scrambled peptides were synthesized and analyzed with SPRi (Supplementary material, Fig. S4). Although with the same amino acid composition, the change in the arrangement led to greatly weakened binding of the scrambled peptides with HSA. These results manifest the interactions between the designed peptide and HSA are sequence dependent. It has been proposed that peptides with longer sequences usually have higher affinity than shorter ones (Huang et al., 2011; Shai et al., 1987). In an attempt to enhance the moderate binding affinity of these peptides, mono-1~ mono-3 were elongated by adding four residues at the N-terminal. However, there was no enhancement of the response intensity (Fig. S5-S8, Table S3 and S4), indicating that simple elongation does not augment affinity for HSA.
Compounds
ka (× 102 M−1 s−1)
kd (× 10−3 s−1)
KD (μM)
mono-1 mono-2 mono-3 homo-4 homo-5 heter-6 heter-7
0.56 ± 0.03 1.60 ± 0.01 2.35 ± 0.03 13.7 ± 1.65 39.4 ± 0.18 7.67 ± 0.02 126.3 ± 13.7
23.75 ± 3.74 11.6 ± 0.2 4.66 ± 0.35 29.7 ± 0 22.7 ± 1.7 2.55 ± 0.005 0.91 ± 0.02
421.5 ± 88.38 72.15 ± 2.05 19.75 ± 0.45 21.9 ± 2.7 5.38 ± 0.68 3.32 ± 0.02 0.073 ± 0.01
3.3. Design and affinity evaluation of homo-bivalent peptides The high affinity and specificity of antibody depends largely on its “Y”-shape dimeric structure (Calarese et al., 2003). Mono-2 and mono-3 demonstrating stronger affinity for HSA were chosen as recognition units for the design of homo dimer peptides (homo-4 and homo-5) respectively (Fig. S9 and S10). The sequence GGGCGGG was used as the dimer dendron. These multiple glycine residues act as a spacer and provide the recognition moieties with enough flexibility during interaction. Cysteine was incorporated with the intension of immobilizing these dimeric peptides onto the gold surface of the SPR chip through an Au-S bond. After assembly, the antibody mimetic “Y”shaped configuration was established for these bivalent peptides on the sensor surface (Fig. 2). For comparison, a peptide-arrayed sensor chip was prepared by printing hexapeptide monomers (mono-2 and mono-3) and bivalent peptides (homo-4 and homo-5) on the chip, and lower concentrations of HSA were examined. As shown in Fig. 2b, homo-4 and homo-5 induced distinctly higher SPR responses than mono-2 and mono-3. Kinetic analysis demonstrated that these homo bivalent peptides bind HSA much faster (Table 1). Their association rates were more than 10-fold higher than those of monomers, which contributed to affinity enhancement. However, their dissociation rates also became faster, which tends to lower the affinity (Table 1). With 3–4-fold lower KD values, the bivalent peptides homo-4 and homo-5 bound HSA
Fig. 1. SPRi sensing of the interactions of HSA and peptides mono-1 (a), mono-2 (b) and mono-3 (c). [HSA]=9.4, 18.8, 37.6 and 75.3 μM, respectively.
659
Biosensors and Bioelectronics 94 (2017) 657–662
Y. Yu et al.
Fig. 2. (a) Illustration of the SPRi chip with immobilization of peptide monomers and homo dimers for interaction with HSA. (b) SPRi sensing the binding of peptide monomers (mono-2 and mono-3) and homo bivalent peptides (homo-4, homo-5) with HSA (2.4 μM).
fluorescein isothiocyanate (FITC)-labeled antibody to HSA induced significant fluorescence emission (Fig. 3a). Various peptides were introduced to compete with this antibody for HSA binding, and peptides with higher affinity for HSA reduced the amount of bound FITC-labeled antibody leading to decreased fluorescence intensity. Only heter-7 markedly reduced the fluorescence from FITC-labeled HSA antibody (Fig. 3a) with resultant fluorescence intensity of only 28% of that induced by direct binding of FITC-labeled antibody to HSA. This reflects the strong interaction of heter-7 with HSA. No significant change in fluorescence was found with the GGGCGGG-linked heter-6, short linear peptides (mono-2 and mono-3) or homo bivalent peptides (homo-4 and homo-5), demonstrating their lack of capacity to interfere with the binding of antibody to HSA and lower antibody binding affinity. The competition between heter-7 and HSA antibody also suggests that heter-7 and the antibody may recognize the same epitope in HSA. To confirm this, a reversed competitive FLISA assay was carried out (Supplementary material, Fig. S15). The results that heter-7 can compete with the pre-bounded antibody for the binding sites in HSA molecule verify the high specificity of heter-7. It is notable that none of the structural components of heter-7 (PEG, azido-PEG-3 and propargylglycine-modified mono-2) can inhibit fluorescence intensity from FITC-labeled antibody. This implies the strong interaction of heter-7 with HSA results from the cooperative effect of the hetero bivalent structure. To investigate the potency of the hetero dimer peptide heter-7, its inhibitory activity was measured with competitive FLISA. As shown in Fig. 3b, fluorescence decreases as the concentration of heter-7 increases, suggesting the enhancement in blocking the interactions between antibody and HSA. The half maximal inhibitory concentration (IC50) was estimated as 83 nM, reflecting the effectiveness of the hetero bivalent peptide.
more tightly. The use of two copies of recognition units increased the chance for accessing target sites on HSA, which contributes to stronger interactions. Although homo-4 and homo-5 affinity was enhanced, such homo dimeric structures may be more suitable for homomeric targets which have multiple identical sites for interaction with multimeric ligands (Choi et al., 2000). However, the polypeptide chain of an HSA molecule only contains one T2 or T3 sequence. The homo bivalent homo-4 and homo-5 therefore cannot yield an increase in affinity comparable to what has been previously reported (Choi et al., 2000; Wan and Alewood, 2016).
3.4. Hetero-bivalent peptides with high affinity Linking multiple heterologous ligands that bind different receptors is attractive to achieve high affinity and enhance biologic function (Berkov-Zrihen et al., 2013; Hart et al., 2014). Design of heterobivalent peptides targeting different domains within HSA molecule was thus carried out. T2 and T3 located at domains II and III of HSA were chosen as ideal targeting sites for the heterodimer. Accordingly, mono-2 and mono-3 were again used as site-specific recognition moieties. For peptide conjugation, two kinds of linkers are considered, including the peptidic GGGCGGG sequence and polyethylene glycol (PEG). The GGGCGGG-linked hetero dimer heter-6 was prepared with the solid phase synthesis and characterized with HPLC and MS (Fig. S11). The synthesis, structure and characterized of PEG-linked heter-7 are provided in Supplementary Material (Scheme S1 and Fig. S12–S14). To give an overall view of the binding of these peptides with HSA, the fluorescence-linked immunosorbent assay (FLISA) featuring high sensitivity and high throughput was employed. The direct binding of
Fig. 3. (a) Competitive FLISA of various compounds against antibody for HSA binding. (1) FITC-labeled antibody only, with addition of competitors: mono-1 (2), mono-2 (3) and mono-3 (4); homo-4 (5) and homo-5 (6); (7) heter-6; (8) azido-PEG acid; (9) propargylglycine-modified mono-2; (10) azido-PEG-3; (11) heter-7. (b) Inhibitory activity of heter-7 against antibody based on FLISA. I and I0: fluorescence from binding of HSA and its FITC-labeled antibody with and without the addition of heter-7. (c) SPRi sensing curves for the heter-7-immobilized chip after injection of HSA and BSA.
660
Biosensors and Bioelectronics 94 (2017) 657–662
Y. Yu et al.
Fig. 4. Structural analysis of HSA and BSA (PDB code: 2BXM for HSA, 4LUF for BSA). Overviews of the binding sites T2 and T3 in HSA (a), and corresponding sequences in BSA (b). (cf) Conformational comparison of the target sequences in HSA and BSA. The target sequences are shown in red. The structures of albumin molecules were analyzed with autodock 4.2 (Sanner, 1999). Ribbon drawings of HSA and BSA were generated and analyzed with Pymol.
significantly lower than that from HSA (Fig. 3c). The affinity constant (KD) of heter-7 for BSA was estimated as 1335 μM, which is orders of magnitude larger than that for HSA (KD 73 nM). After saturation, the surface bound amount of BSA on the chip was also distinctly smaller than that of HSA (Supplementary Material). These results demonstrate the ability of heter-7 for HSA differentiation. The different responses of heter-7 to HSA and BSA can arise from the tiny differences in the sequence and structure of the target sites in HSA and BSA. For sequence comparison, BSA shares 76% identity with HSA in amino acid composition. The T2 sequences are QDSISS and QDTISS (residues 267–272) in HSA and BSA, respectively (Fig. 4). The one-only residue difference between T2 sites in these two albumins is further narrowed because Ser269 in HSA and Thr269 in BSA both contain a hydroxyl side chain. However, distinct differences are observed in the secondary structures of these sequences in albumins. As shown in Fig. 4c, QDSISS in HSA forms an extended conformation, which is favored by peptide-peptide interactions. The higher content of α-helix conformation in QDTISS of BSA hampers molecular recognition between peptides (Fig. 4d). As for T3, the sequences are PKEFNA and PKAFDE in HSA and BSA respectively (Fig. 4). The target sequences in both HSA and BSA adopt extended conformations (Figs. 4e, f). However, the amino-acid composition and arrangement are different. By substituting asparagine (N) in HSA with aspartic acid (D), the corresponding sequence in BSA contains more negative charge and is more acidic. Moreover, the different location of glutaminic acid (E) and alanine (A) in these two proteins also affects the affinity recognition of the peptide ligand. Based on the above analysis, the selectivity of heter-7 is sensitive to both the amino-acid arrangement and the conformation of the target sites. By targeting two sites on two domains of HSA, the designed bi-branched hetero dimer peptide shows desirable enhancement in affinity and selectivity owing to the cooperation recognition effect.
3.5. Binding kinetics and high selectivity of the hetero dimer To acquire binding kinetics on the interaction of heter-7 with HSA, SPRi was employed. A mercaptoundecanoic acid (MUA)-decorated sensor chip was used for immobilization of heter-7. The preassembled MUA played a role as the spacer and provided heter-7 with flexibility, thus avoiding potential steric hindrance during interaction with HSA. Kinetic data and equilibrium dissociation constants were calculated by fitting the sensor curves. As listed in Table 1, with a rate constant of 126.3× 102 M−1 s−1, heter-7 binds HSA significantly faster than the other peptides. Its dissociation rate (0.91× 10−3 s−1) is one order of magnitude lower, showing heter-7 dissociates from HSA with difficulty. The rapid binding and slow dissociation of heter-7 yield a lowered KD of 73 nM, which is in accordance with its high affinity in competitive FLISA binding assays. Although having the same two recognition moieties as heter-7, heter-6 binds HSA slowly and has a strong tendency to dissociate, thus the affinity is significantly lower (KD 3.32 μM). This is most likely attributable to the difference in the linker. Based on molecular modeling, a minimum length of 99 Å is needed for the linker to achieve effective binding at two sites (T2 and T3) on one HSA molecule (Fig. 4a). PEG has a molecular length of 105.6 Å and can thus accommodate this two-site binding. In contrast, the linker GGGCGGG is only 20.1 Å in length, which is too short to reach both T2 and T3 sites simultaneously on one HSA molecule. Such shortness of the linker impairs the bivalent binding of heter-6 to HSA. Although heter-6 is disabled from two-site binding on one HSA molecule, the hetero bivalent structure increases the chance for contact with target sites during interaction. This may contribute to higher affinity of heter-6 for HSA than other short linear peptides and homo dimers (Table 1). Therefore, both hetero bivalency and a spacer of proper length serve to strengthen the interaction of heter-7 and HSA. To examine the specificity of heter-7, various biomolecules including proteins, peptides, carbohydrate and nucleotides were used as the substitutes for HSA and injected into the SPRi sensor system. Although with different charging characters and hydrophobicity, none of these compounds could induced comparable signal change as HSA did (Fig. S16). It is interesting to find that despite the high sequence and structural similarity of BSA with HSA, the SPRi signal from BSA is
4. Conclusion In this study site-specific and sequence-sensitive peptide binders towards human serum albumin (HSA) were designed, synthesized and analyzed. In comparison with the monomers and homo dimers which were evaluated, the hetero dimer showed fast binding and slow 661
Biosensors and Bioelectronics 94 (2017) 657–662
Y. Yu et al.
Grigoryan, G., Reinke, A.W., Keating, A.E., 2009. Nature 458, 859–864. Hart, N.J., Chung, W.J., Weber, C., Ananthakrishnan, K., Anderson, M., Patek, R., Zhang, Z., Limesand, S.W., Vagner, J., Lynch, R.M., 2014. ChemBioChem 15, 135–145. He, X.M., Carter, D.C., 1992. Nature 358, 209–215. Heal, J.R., Roberts, G.W., Raynes, J.G., Bhakoo, A., Miller, A.D., 2002. ChemBioChem 3, 136–151. Healy, A.R., Houston, D.R., Remnant, L., Huart, A., Brychtova, V., Maslon, M.M., Meers, O., Muller, P., Krejci, A., Blackburn, E.A., Vojtesek, B., Hernychova, L., Walkinshaw, M.D., Westwood, N.J., Hupp, T.R., 2015. Chem. Sci. 6, 3109–3116. Hossain, M.A., Kocan, M., Yao, S.T., Royce, S.G., Nair, V.B., Siwek, C., Patil, N.A., Harrison, I.P., Rosengren, K.J., Selemidis, S., Summers, R.J., Wade, J.D., Bathgate, R.A.D., Samuel, C.S., 2016. Chem. Sci. 7, 3805–3819. Hu, F., Huang, Y., Zhang, G., Zhao, R., Yang, H., Zhang, D., 2014. Anal. Chem. 86, 7987–7995. Huang, Y., Zhao, R., Fu, Y., Zhang, Q., Xiong, S., Li, L., Zhou, R., Liu, G., Chen, Y., 2011. ChemBioChem 12, 1209–1215. Hwang, H.J., Ryu, M.Y., Park, C.Y., Ahn, J., Park, H.G., Choi, C., Ha, S., Park, T.J., Park, J.P., 2017. Biosens. Bioelectron. 87, 164–170. Janssen, B.M.G., Lempens, E.H.M., Olijve, L.L.C., Voets, I.K., J. van Dongen, J.L., A. de Greefc, T.F., Merkx, M., 2013. Chem. Sci., vol. 4, pp. 1442–1450. Jiang, Z., Qin, Y., Peng, Z., Chen, S., Chen, S., Deng, C., Xiang, J., 2014. Biosens. Bioelectron. 62, 268–273. Joshi, A., Vance, D., Rai, P., Thiyagarajan, A., Kane, R.S., 2008. Chem. Eur. J. 14, 7738–7747. Kojima, C., Irie, K., Tada, T., Tanaka, N., 2014. Biopolymers 101, 603–612. Kwok, A., Eggimann, G.A., Reymond, J.L., Darbre, T., Hollfelder, F., 2013. ACS Nano 7, 4668–4682. Larman, H.B., Zhao, Z., Laserson, U., Li, M.Z., Ciccia, A., Gakidis, M.A.M., Church, G.M., Kesari, S., LeProust, E.M., Solimini, N.L., Elledge, S.J., 2011. Nat. Biotechnol. 29, 535–541. Lawson, K.V., Rose, T.E., Harran, P.G., 2013. Proceedings Natl. Acad. Sci. USA, 110, E3753-E3760. Lewin, M., Carlesso, N., Tung, C.H., Tang, X.W., Cory, D., Scadden, D.T., Weissleder, R., 2000. Nat. Biotechnol. 18, 410–414. Li, R., Yang, H., Jia, D., Nie, Q., Cai, H., Fan, Q., Wan, L., Li, L., Lu, X., 2016. J. Control. Release 228, 96–106. Lipp, A., Ji, B., Hager, R., Haas, S., Schweiggl, S., Sonnleitner, A., Haselgrübler, T., 2015. Biosens. Bioelectron. 74, 757–763. London, N., Movshovitz-Attias, D., Schueler-Furman, O., 2010. Structure 18, 188–199. Miller, A.D., 2015. Expert Opin. Biol. Ther. 15, 245–267. Pawson, T., Nash, P., 2003. Science 300, 445–452. Pelay-Gimeno, M., Glas, A., Koch, O., Grossmann, T., 2015. Angew. Chem. Int. Ed. 54, 8896–8927. Peng, L., Liu, R., Marik, J., Wang, X., Takada, Y., Lam, K.S., 2006. Nat. Chem. Biol. 2, 381–389. Petsalaki, E., Russell, R.B., 2008. Curr. Opin. Biotechnol. 19, 344–350. Prasad, B.B., Prasad, A., Tiwari, M.P., 2013. Biosens. Bioelectron. 39, 236–243. Rao, J., Lahiri, J., Isaacs, L., Weis, R.M., Whitesides, G.M., 1998. Science 280, 708–711. Reja, I.S., Khan, I.A., Bhalla, V., Kumar, M., 2016. Chem. Commun. 52, 1182–1185. Sanner, M.F., 1999. J. Mol. Graph Model. 17, 57–61. Shai, Y., Flashner, M., Chaiken, I.M., 1987. Biochemistry 26, 669–675. Shiba, K., 2010. Chem. Soc. Rev. 39, 117–126. Shim, Y.Y., Reaney, M.J.T., 2015. J. Agric. Food Chem. 63, 1099–1106. Sohl, J.L., Splittgerber, A.G., 1991. J. Chem. Educ. 68, 262–264. Stein, A., Aloy, P., 2008. PLoS One 3, E2524. Vance, D., Martin, J., Patke, S., Kane, R.S., 2009. Adv. Drug Deliv. Rev. 61, 931–939. Wan, J., Alewood, P.F., 2016. Angew. Chem. Int. Ed. 55, 5124–5134. Wan, J., Huang, J.X., Vetter, I., Mobli, M., Lawson, J., Tae, H.S., Abraham, N., Paul, B., Cooper, M.A., Adams, D.J., Lewis, R.J., Alewood, P.F., 2015. J. Am. Chem. Soc. 137, 3209–3212. Wu, B., Chien, E.Y.T., Mol, C.D., Fenalti, G., Liu, W., Katritch, V., Abagyan, R., Brooun, A., Wells, P., Bi, F.C., Hamel, D.J., Kuhn, P., Handel, T.M., Cherezov, V., Stevens, R.C., 2010. Science 330, 1066–1071. Wu, S., Núnez, D., Hu, S., Domingo, M.P., Chen, Y., Wei, P., Pardo, J., Gálvez, E.M., Chiou, A., 2014. Biosens. Bioelectron. 56, 159–166. Xiong, J.P., Stehle, T., Zhang, R., Joachimiak, A., Frech, M., Goodman, S.L., Arnaout, M.A., 2002. Science 296, 151–155. Xu, X., Wang, L., Huang, Y., Gao, W., Li, K., Jiang, W., 2016. Anal. Chem. 88, 9885–9889. Yu, Y., Huang, Y., Hu, F., Jin, Y., Zhang, G., Zhang, D., Zhao, R., 2016. Anal. Chem. 88, 6374–6381. Zhang, Q., Huang, Y., Zhao, R., Liu, G., Chen, Y., 2008. Biosens. Bioelectron. 24, 48–54. Zhou, J., Du, X., Yamagata, N., Xu, B., 2016. J. Am. Chem. Soc. 138, 3813–3823.
dissociation behavior, which results in strong binding with HSA. According to SPRi analysis, the KD value of heter-7 was 73 nM, which is orders of magnitude lower than those of other peptides. In competitive fluoroimmuno assays with antibody, heter-7 demonstrated the site-specific and competing binding behavior for HSA. Benefiting from the cooperative effect of the site-specific binders and the PEG linker, heter-7 can differentiate HSA from BSA with sequence and conformation sensitivity. These results demonstrate the advantage hetero bivalent peptide in affinity and selectivity for bioanalysis. With the combination with nanomaterial, sensing techniques or functional agents, such molecular bioprobes shows promise for use in serum proteomics, disease detection and drug transport. Acknowledgements Financial support from National Natural Science Foundation of China (21375134, 21475140, 21675161 and 21621062), the Ministry of Science and Technology of China (2015CB856303), and the Chinese Academy of Sciences is gratefully acknowledged. We also express our gratitude to Dr. Michael A. McNutt from U.S. for his help. Appendix A. Supporting information Supplementary data associated with this article can be found in the online version at doi:10.1016/j.bios.2017.03.067. References An, D., Su, J., Weber, J.K., Gao, X., Zhou, R., Li, J., 2015. J. Am. Chem. Soc. 137, 8412–8418. Anees, P., Sreejith, S., Ajayaghosh, A., 2014. J. Am. Chem. Soc. 136, 13233–13239. Berkov-Zrihen, Y., Green, K.D., Labby, K.J., Feldman, M., Garneau-Tsodikova, S., Fridman, M., 2013. J. Med. Chem. 56, 5613–5625. Bujacz, A., 2012. Acta Crystallogr. Sect. D: Biol. Crystallogr. 68, 1278–12289. Byeon, H.J., Min, S.Y., Kim, I., Lee, E.S., Oh, K.T., Shin, B.S., Lee, K.C., Youn, Y.S., 2014. Bioconjug. Chem. 25, 2212–2221. Cai, H., Sun, Z.Y., Chen, M.S., Zhao, Y.F., Kunz, H., Li, Y.M., 2014. Angew. Chem. Int. Ed. 53, 1699–1703. Calarese, D.A., Scanlan, C.N., Zwick, M.B., Deechongkit, S., Mimura, Y., Kunert, R., Zhu, P., Wormald, M.R., Stanfield, R.L., Roux, K.H., Kelly, J.W., Rudd, P.M., Dwek, R.A., Katinger, H., Burton, D.R., Wilson, I.A., 2003. Science 300, 2065–2071. Canoa, P., Simón-Vázquez, R., Popplewell, J., González-Fernández, Á., 2015. Biosens. Bioelectron. 74, 376–383. Carter, D.C., He, X.M., Twigg, P.D., Casale, E., 1989. Science 244, 1195–1198. Chen, S., Bertoldo, D., Angelini, A., Pojer, F., Heinis, C., 2014. Angew. Chem. Int. Ed. 53, 1602–1606. Cheng, Y., Zhao, L., Li, Y., Xu, T., 2011. Chem. Soc. Rev. 40, 2673–2703. Choi, J.S., Joo, D.K., Kim, C.H., Kim, K., Park, J.S., 2000. J. Am. Chem. Soc. 122, 474–480. Chu, D.S.H., Bocek, M.J., Shi, J., Ta, A., Ngambenjawong, C., Rostomily, R.C., Pun, S.H., 2015. J. Control. Release 205, 155–161. Debnath, S., Chatterjee, S., Arif, M., Kundu, T.K., Roy, S., 2011. J. Biol. Chem. 286, 25076–25087. Diana, D., Russomanno, A., Rosa, De. L., Stasi, R.D., Capasso, D., Gaetano, S.D., Romanelli, A., Russo, L., D’Andrea, L.D., Fattorusso, R., 2015. Chem. Eur. J. 21, 91–95. Ding, B., Jasensky, J., Li, Y., Chen, Z., 2016. Acc. Chem. Res. 49, 1149–1157. Doran, T.M., Kodadek, T., 2014. ACS Chem. Biol. 9, 339–346. Fan, J., Sun, W., Wang, Z., Peng, X., Li, Y., Cao, J., 2014. Chem. Commun. 50, 9573–9576. Fanali, G., di Masi, A., Trezza, V., Marino, M., Fasano, M., Ascenzi, P., 2012. Mol. Asp. Med. 33, 209–290. Frei, A.P., Moest, H., Novy, K., Wollscheid, B., 2013. Nat. Protoc. 8, 1321–1336. Goh, W.L., Lee, M.Y., Joseph, T.L., Quah, S.T., Brown, C.J., Verma, C., Brenner, S., Ghadessy, F.J., Teo, Y.N., 2014. J. Am. Chem. Soc. 136, 6159–6162. Gray, B.P., Brown, K.C., 2014. Chem. Rev. 114, 1020–1081.
662