13962
Langmuir 2008, 24, 13962-13968
Surfaces for Immobilization of N-Terminal Cysteine Derivatives via Native Chemical Ligation Sally Anderson† Sharp Laboratories of Europe, Ltd., Edmund Halley Road, Oxford Science Park, Oxford OX4 4GB, U.K. ReceiVed July 10, 2008. ReVised Manuscript ReceiVed September 21, 2008 This paper presents an improved method for fast, reliable, and quantitative native chemical ligation (NCL - the reaction between an N-terminal cysteine and a thioester to yield an amide bond) at thiophenyl ester-functionalized glass surfaces. For the first time, the degree of surface functionalization has been measured and can be readily controlled by varying the concentration of thiophenyl ester groups on the surface. This methodology facilitates the preparation of tailor-made functionalized glass surfaces for diverse applications. S-Phenyl 11-(chlorodimethylsilyl)undecanethiolate, and the benzyl analogue, are readily prepared from 10-undecanoic acid via thioester formation and platinum-catalyzed hydrosilylation of the terminal alkene functionality. Thioesters covalently bound to glass surfaces were then formed by submerging clean glass in toluene-solutions of the relevant thioester silylchloride (1%) containing the non-nucleophilic base, ethyldiisopropylamine (1%). NCL was explored with a cysteine-lissamine conjugate, and the degree of surface functionalization was quantified by UV/vis absorption spectroscopy, using the lissamine chromophore. NCL at thiophenyl ester surfaces proved fast (half-life less than 10 min) and yielded levels of surface functionalization consistent with a dense monolayer of dimethyloctylsilane (ca. 2.0 molecules/nm2). Water contact angles on thiophenyl ester surfaces were found to decrease after NCL reaction with the cysteine-lissamine conjugate, whereas surfaces treated with the same buffer solution containing an unreactive alanine-lissamine conjugate showed no significant changes in contact angle, indicating that thioester hydrolysis is not significant during the course of the reaction. NCL with the cysteine-lissamine conjugate at mixed surfaces containing both thiophenyl esters and inert octyl chains showed lower levels of surface functionalization as expected. Plotting the proportion of thiophenyl esters used in the preparation of the substrates against integrated absorption from the surface yielded a linear relationship, demonstrating that NCL on these surfaces occurs in a controllable manner.
Introduction The preparation of tailor-made functionalized surfaces is important in many research fields, e.g., (a) stationary phases for liquid and gas chromatography,1 (b) as a pathway to the preparation of self-assembled multilayer structures,2 (c) to facilitate site-selective adsorption of molecules and mesoscopic objects onto surfaces for the fabrication of electronic, mechanical, and photonic devices,3-5 (d) in the preparation of biocompatible surfaces with controlled surface charge and precise control over surface composition,6 (e) in the preparation of dye functionalized surfaces for sensing applications,7 and (f) in the preparation of microarrays for the high-throughput evaluation of biomolecular interactions.8 Progress in all these areas is underpinned by new and efficient methods for immobilizing molecules on surfaces through covalent attachment.9 The chemical strategies that have been reported for the preparation of functionalized sur† E-mail:
[email protected]. (1) Schiel, J. E.; Mallik, R.; Soman, S.; Joseph, K. S.; Hage, D. S. J. Sep. Sci. 2006, 29, 719–737. (2) Bertrand, P.; Jonas, A.; Laschewsky, A.; Legras, R. Macromol. Rapid Commun. 2000, 21, 319–348. (3) del Campo, A.; Boos, D.; Spiess, H. W.; Jonas, U. Angew. Chem., Int. Ed. 2005, 44, 4707–4712. (4) Auletta, T.; Dordi, B.; Mulder, A.; Sartori, A.; Onclin, S.; Bruinink, C. M.; Pe´ter, M.; Nijhuis, C. A.; Beijleveld, H.; Scho¨nherr, H.; Vancso, G. J.; Casnati, A.; Ungaro, R.; Ravoo, B. J.; Huskens, J.; Reinhoudt, D. N. Angew. Chem., Int. Ed. 2004, 43, 369–373. (5) Vilan, A.; Cahen, D. Trends Biotechnol. 2002, 20, 22–29. (6) Ganesan, R.; Yoo, S. Y.; Choi, J.-H.; Lee, S. Y.; Kim, J.-B. J. Mater. Chem. 2008, 18, 703–709. (7) (a) Onclin, S.; Ravoo, B. J.; Reinhoudt, D. N. Angew. Chem., Int. Ed. 2005, 44, 6282–6304. (b) Zhang, C.; Suslick, K. S. J. Am. Chem. Soc. 2005, 127, 11548–11549. (8) MacBeath, G.; Schreiber, S. L. Science 2000, 289, 1760–1763. (9) Woo, Y.-H.; Camarero, J. A. Curr. Nanosci. 2006, 93–103.
faces include Staudinger ligation,10 native chemical ligation (NCL),11-13 the Diels-Alder reaction,14 reaction between surface-bound silyl chlorides and alcohols,15 reaction between diazobenzylidene-functionalized glass and heteroatoms with acidic protons,16 Michael reaction between a maleimidefunctionalized surface and thiols,17 “click chemistry” (Huisgen 1,3-dipolar cycloadditions)18 between acetylenes and azides, and noncovalent strategies.19 NCL,20 the reaction between an N-terminal cysteine and a thioester to yield an amide bond, has several features that make (10) (a) Ko¨hn, M.; Wacker, R.; Peters, C.; Schro¨der, H.; Soule`re, L.; Breinbauer, R.; Niemeyer, C. M.; Waldmann, H. Angew. Chem., Int. Ed. 2003, 42, 5830– 5834. (b) Soellner, M. B.; Dickson, K. A.; Nilsson, B. L.; Raines, R. T. J. Am. Chem. Soc. 2003, 125, 11790–11791. (11) (a) Lesaicherre, M.-L.; Uttamchandani, M.; Chen, G. Y. J.; Yao, S. Q. Bioorg. Med. Chem. Lett. 2002, 12, 2079–2083. (b) Uttamchandani, M.; Chan, E. W. S.; Chen, G. Y. J.; Yao, S. Q. Bioorg. Med. Chem. Lett. 2003, 13, 2997– 3000. (c) Girish, A.; Sun, H.; Yeo, D. S. Y.; Chen, G. Y. J.; Chua, T.-K.; Yao, S. Q. Bioorg. Med. Chem. Lett. 2005, 15, 2447–2451. (12) Camarero, J. A.; Kwon, Y.; Coleman, M. A. J. Am. Chem. Soc. 2004, 126, 14730–14731. (13) Wojtyk, J. T. C.; Morin, K. A.; Boukherroub, R.; Wayner, D. D. M. Langmuir 2002, 18, 6081–6087. (14) Houseman, B. T.; Huh, J. H.; Kron, S. J.; Mrksich, M. Nat. Biotechnol. 2002, 20, 270–274. (15) Hergenrother, P. J.; Depew, K. M.; Schreiber, S. L. J. Am. Chem. Soc. 2000, 122, 7849–7850. (16) Barnes-Seeman, D.; Park, S.-B.; Koehler, A. N.; Schreiber, S. L. Angew. Chem., Int. Ed. 2003, 42, 2376–2379. (17) MacBeath, G.; Koehler, A. N.; Schreiber, S. L. J. Am. Chem. Soc. 1999, 121, 7967–7968. (18) Devaraj, N. K.; Dinolfo, P. H.; Chidsey, C. E. D.; Collman, J. P. J. Am. Chem. Soc. 2006, 128, 1794–1795. (19) Schmid, E. L.; Keller, T. A.; Dienes, Z.; Vogel, H. Anal. Chem. 1997, 69, 1979–1985. (20) (a) Yeo, D. S. Y.; Srinivasan, R.; Chen, G. Y. J.; Yao, S. Q. Chem.;Eur. J. 2004, 10, 4664–4672. (b) Dawson, P. E.; Muir, T. W.; Clark-Lewis, I.; Kent, S. B. H. Science 1994, 266, 776–779. (c) Johnson, E. C. B.; Kent, S. B. H. J. Am. Chem. Soc. 2006, 128, 6640–6646.
10.1021/la8022024 CCC: $40.75 2008 American Chemical Society Published on Web 11/12/2008
Immobilization of N-Terminal Cysteines via NCL Scheme 1. Mechanism for NCL
it uniquely attractive as a potential methodology for grafting molecules to surfaces: • N-terminal cysteines can easily be incorporated into peptides, proteins and other nonpeptide based systems during conventional or automated solid-phase synthesis. • It generates a stable native amide bond from an N-terminal amine and a C-terminal thioester (Scheme 1). The structure and function of complex native proteins should therefore be unaffected by surface immobilization under these conditions.21 • NCL can be carried out in water under physiological conditions. • It has been used widely in solution to join oligopeptides to generate larger proteins.22 Kent and co-workers20c have explored the kinetics of the NCL reaction in solution and optimized the reaction conditions. Examples of NCL reactions on surfaces are rare. Yao and coworkers11a published the first report of NCL at surfaces. In this study, a cysteine-fluorescein conjugate was reported to be immobilized on a thioester-functionalized surface; subsequently, the authors described the preparation of peptide libraries for the rapid determination of kinase specificity.11b Wayner and coworkers13 also described the preparation of thioethyl esterfunctionalized surfaces on porous silicon and subsequently followed the reaction with cysteine by diffuse reflectance Fourier transform infrared spectroscopy (FT-IR). Camarero and coworkers12 reported the inverse procedure, in which a fluorescent thioester protein conjugate was immobilized on a cysteine functionalized surface. More recently, Helms and co-workers21 have explored functionalization of surfaces for surface-plasmon resonance via NCL and have monitored binding between a streptavidin-binding peptide and streptavidin. Our interest in NCL at surfaces grew from a need to prepare surfaces with precise control over surface properties and composition, with particular interest in the UV/vis absorption characteristics of the surface. Any molecules containing an N-terminal cysteine, e.g., peptides, proteins, and peptide derivatives, can be attached to surfaces using this reaction. Initially, the methods of Yao and co-workers11a were explored, choosing the combination of a thioesterfunctionalized surface and an N-terminal cysteine in solution, because thioester surfaces are more stable than thiol-functionalized surfaces, since they are not susceptible to oxidation. Thioester surfaces were prepared from amino-functionalized glass slides in a multistep procedure following Yao and co-workers’ method, but when NCL was attempted with a simple cysteine-fluorescein conjugate (CK-fluorescein, see Supporting Information for structure) the reaction was found to be very poor. Similar levels of fluorescence were observed for thioester (thiobenzyl ester), (21) Helms, B.; van Baal, I.; Merkx, M.; Meijer, E. W. ChemBioChem 2007, 8(15), 1790–1794. (22) Torbeev, V. Y.; Kent, S. B. H. Angew. Chem., Int. Ed. 2007, 46, 1667– 1670. (23) Fadeev, A. Y.; McCarthy, T. J. Langmuir 2000, 16, 7268–7274.
Langmuir, Vol. 24, No. 24, 2008 13963
or control amide (ethyl hexyl amide), or amino-functionalized glass surfaces. Dye surface coverages calculated from the very small and noisy signals observed in the UV/vis absorption spectra were never greater than 0.15 molecules/nm2. This surface-density corresponds to significantly less than monolayer coverage (ca. 1/10) for fluorescein, even with the molecules lying flat on the surface. Another problem with this system is that the fluorescence from fluorescein is very dependent on pH and solvent conditions, leading to problems quantifying the amount of functionalization occurring during NCL. This paper presents the preparation of improved thioester surfaces for NCL. The surfaces were prepared in a single step from thioester silyl chlorides. The preparation of the more reactive thiophenyl ester surfaces proved particularly important. To the best of our knowledge, this is the first report of the direct preparation of thiophenyl ester surfaces. A new cysteine-lissamine conjugate was used to interrogate these surfaces and provided good evidence for fast and controllable NCL at the surface.
Experimental Section General. Unless otherwise stated, the starting materials were purchased from Sigma Aldrich and used without further purification. 4-Mercaptophenylacetic acid (MPAA) was purchased from Toronto Research Chemicals, Canada. Preparative flash column chromatography was performed with 60 µm Merck silica gel (230-400 mesh), and reversed phase chromatography was performed with C18 modified silica gel purchased from Fluka. Dry triethylamine (Et3N) and diisopopylethylamine (DIEA) were obtained by distillation from calcium hydride under nitrogen, and dry thiophenol and benzylmercaptan were obtained by drying over anhydrous sodium sulfate, decanting from the drying agent and then distilling trapto-trap under reduced pressure. The thiols were collected by cooling the receiver flask to liquid nitrogen temperatures. Mass spectrometry was carried out at the University of Southampton. 1H NMR spectra were recorded on a Bruker DPX 300. Overlapping signals in 13C NMR spectra (determined by integration comparison of similar environments) are denoted with an asterix. Absorption spectra were measured on a Shimadzu UV - 2410PC UV/vis spectrophotometer. Fluorescence scanning was carried out on a Typhoon Trio Plus (GE Healthcare) exciting samples with the green laser (532 nm) and detecting at 580 nm with a 30 nm bandpass filter. A photomultiplier tube (PMT) voltage of 400 V was used, and scanning was carried out at normal sensitivity with 200 µm resolution. 4: 10-Undecylenic acid (5.25 g, 28 mmol) was placed in a threenecked round-bottomed flask (100 mL). Diethylether (50 mL, anhydrous, and nitrogen-saturated) was then added, followed by oxalyl chloride (18.1 g, 12.5 mL, 143 mmol), and the mixture was stirred until no more gas was evolved. After 2 h of stirring at room temperature, the solvent was removed in Vacuo, and the acid chloride was redissolved in diethylether (50 mL). The ether acid chloride solution was then transferred to a nitrogen-filled dropping funnel over a mixture of thiophenol (3.8 g, 3.6 mL, 35 mmol) and DIEA (4.5 g, 6.1 mL, 35 mmol) at 0 °C. On adding the acid chloride, the reaction mixture became warm and turned cloudy. 1H NMR showed that all the acid chloride had been transformed into the thioester. The reaction mixture was then poured into water, and extracted with diethylether (4 × 20 mL). The combined organic fractions were then washed with additional water (50 mL) and brine (50 mL), and then dried over sodium sulfate. The solvent was removed via distillation. The yellow oil was then transferred to a small roundbottomed flask and distilled using a Ku¨gelrohr at 5 × 10-2 mbar and 140 °C yielding a colorless oil (5.9 g, 76%). 1H NMR (300 MHz, CDCl3) δ 7.4 (m, 5 H, phenyl), 5.7-5.9 (m, 1 H, alkene), 4.9-5.0 (m, 2 H, alkene), 2.65 (t, 2 H, J ) 7.33), 2.04 (dt, 2 H, J ) 7, J ) 7, next to alkene), 1.7 (m, 2 H, PhSCOCH2CH2), 1.2-1.4 (m, 10 H, alkyl-H). 13C NMR (100 MHz, CDCl3) δ 197.98, 139.57, 134.88, 129.68, 129.55, 128.36, 114.57, 44.12, 34.18, 29.64, 29.59, 29.43, 29.33, 29.28, 25.98.
13964 Langmuir, Vol. 24, No. 24, 2008 1: Thioester 4 (5.9 g, 0.021 mol), chlorodimethylsilane (34.04 g, 40 mL, 0.28 mol), and chloroplatinic acid (500 µL of a 4% solution in dry isopropanol) were stirred at 25 °C under nitrogen for 48 h. 1H NMR showed that all the alkene signals had disappeared. The reaction mixture was then transferred to a small round-bottomed flask for Ku¨gelrohr distillation. Distillation at 4 × 10-2 mbar and 200 °C yielded 1 as a colorless oil (5.97 g, 77%). 1H NMR (300 MHz, CDCl3) δ 7.41 (m, 5 H, Ph), 2.65 (t, J ) 7.3 Hz, 2 H, CH2CH2OCSPh), 1.6-1.8 (m, 2 H, CH2CH2OCSPh), 1.2-1.5 (m, 14 H, alkyl H), 0.86-0.75 (m, 2 H, ClMe2SiCH2), 0.4 (s, 6 H, ClMe2SiCH2). 13C NMR (100 MHz, CDCl3) δ 197.90, 134.86, 129.65, 129.53, 128.42, 44.12, 33.32, 29.80, 29.70, 29.61, 29.58, 29.34, 26.0, 23.36, 19.39, 2.06. CI MS Calcd for C19H31ClOSSi 371.05; found 261.2 [M - Ph - Cl]+, 371.2 [MH]+, 388.2 [M + NH4]+. 2: 1H NMR (300 MHz, CDCl3) δ 7.3-7.2 (m, 5 H), 4.12 (s, 2 H), 2.56 (t, J ) 7.5 Hz, 2 H), 1.7-1.2 (m, 16 H), 0.81 (m, 2 H), 0.40 (s, 6 H). 13C NMR (75 MHz, CDCl3) δ 199.37, 138.16, 129.20, 129.01, 127.60, 44.25, 33.52, 33.34, 29.80, 29.78, 29.59*, 29.32, 26.01, 23.36, 19.38, 2.07. Substrate Preparation: Glass Cleaning. Glass microscope slides were cleaned in “piranha solution” overnight.24 The substrates were then washed in copious amounts of deionized (DI) water and dried under vacuum at room temperature. Preparation of Thioester-Coated Glass Microscope Slides: In the Vapor Phase. Clean glass microscope slides were placed in a wide-necked glass tube with the relevant thioester silylchloride (100 µL) and then heated to 200 °C under vacuum (3 × 10-2 mbar). The thioester silylchloride vapor was distilled over the slides and condensed further up the glass tube. After the substrates had cooled to room temperature, they were washed in dichloromethane and dried in a stream of nitrogen and then stored in a freezer (-18 °C). Preparation of Thioester-Coated Glass Microscope Slides: In Solution (This Procedure Was Carried out in a Glovebox). Piranha-cleaned glass microscope slides were placed in a staining jar containing a 1% solution of the relevant thioester silylchloride and/or alkylsilylchloride and DIEA in toluene. The solution was then left for a week. Substrates were then washed with clean toluene (×2) and dried in a stream of nitrogen before being stored in a freezer (-18 °C). NCL Reaction Conditions. Sodium phosphate buffer (pH 9, 200 mM) was added to tris(2-carboxyethyl)phosphine hydrochloride (TCEP) and MPAA until the pH was about 7.5. Sodium phosphate buffer (pH 7, 200 mM) was then added until the required concentrations of TCEP and MPAA were achieved (10 mM each). Cys-Liss 6 and Ala-Liss 7 solutions were prepared in methanol (1-3 mM), and the relevant amount was transferred to the buffer solution (see Supporting Information for details). Functionalized glass microscope slides were cut up into substrates (0.9 cm × 2.56 cm), and these substrates were bundled together with a short plane glass spacer at each end. A maximum of three substrates and four glass spacers were used. One end of the bundle was then bound with Teflon tape, and the spacers from the unbound end were removed to generate a comb. These substrates were then submerged in NCL solution, and the comb-like structure enabled free movement of the NCL solution around the substrates. After 2 h under NCL conditions, the substrates were washed in dimethylformamide (DMF; ×2) and then sonicated in nitrogen saturated sodium phosphate buffer (pH 7.5, 200 mM) containing TCEP (10 mM). On removing the substrates from the aqueous buffer, they were rinsed in methanol and dried in a stream of nitrogen. This procedure led to substrates being functionalized on both sides. For the substrates functionalized with mixtures of octyl and thiophenyl ester groups, NCL reaction was carried out on one side only by spotting NCL solution onto the substrate and leaving the substrate in a humid box for 2 h. During workup, the NCL reaction mixture was removed carefully by pipet before the substrates were submerged in DMF, following the workup procedure outlined above. (24) “Piranha solution” was prepared by carefully adding hydrogen peroxide solution (27%, 20 mL) to concentrated sulfuric acid (45 mL) and used while still warm. Explosive mixtures may form with organic material, so great caution should be taken.
Anderson Scheme 2. Details of the Preparation of Thioester Silylchlorides 1 and 2
UV/Vis Absorption Spectra. UV/vis absorption spectra of dry substrates were measured from 350 to 630 nm using a Shimadzu UV-2401PC UV-vis spectrophotometer. UV/vis absorption measurements were taken from various random locations on the substrates (size 0.9 cm × 2.56 cm), and in all cases results from at least two independently prepared substrates were compared. The integrated absorption measurements were within 10% of each other. Care was taken that the functionalized region of the substrate was in the beam for substrates prepared by spotting rather than substrate submersion. Contact Angles. Contact angle goniometry measurements were made on a Kru¨ss DSA10 - Mk2 using a 3 µL drop of DI water delivered by syringe pump. Drop profiles were photographed and angles analyzed by DSA1 v1.8. Three measurements were made at random points on the substrates, and the values were averaged.
Results and Discussion Thioester silylchlorides 1 and 2 were prepared in excellent yield from commercially available 10-undecanoic acid via thioester formation and platinum(IV) catalyzed hydrosilylation of the double bond (Scheme 2). These reactions proved extremely clean, and the monochlorosilane thioesters were readily purified by distillation and stored under an inert atmosphere. Monochlorosilane thioesters are preferred for glass functionalization, over di- and trichlorosilanes, because they cannot undergo polymerization in the presence of moisture.23 Only two possible products can be formed from monochlorosilanes: the silane can either react productively with the glass surface to yield the required surface bound thioester, or in the presence of a trace amount of water, it can form Si-O-Si linked dimers that are readily washed away. Surface functionalization with monochlorosilane thioesters may lead to lower surface coverage than di- and trichlorosilane thioesters, but gives less ambiguity in the surface structure. The absence of surface-bound polymers is expected to lead to less nonspecific binding when attempting NCL on these surfaces. Thiobenzyl and thiophenyl ester functionalized surfaces were prepared in a single step. Solutions of thioester silyl chlorides 1 or 2 (1% v/v in toluene containing 1% iPr2NEt) were prepared and glass microscope slides (that had been cleaned in H2O2/H2SO4 piranha solution,24 washed in water, and then dried under nitrogen) were submerged in these solutions. The whole operation was carried out in a glovebox under nitrogen. After a week in solution, the substrates were then washed in toluene and dried under nitrogen. Gas-phase thioester surface preparation was also investigated by depositing the monochlorosilane thioester onto clean glass microscope slides from the vapor at 200 °C and 2 × 10-2 mbar pressure. These surfaces were then used without further modification for NCL experiments.
Immobilization of N-Terminal Cysteines via NCL
Langmuir, Vol. 24, No. 24, 2008 13965 Scheme 4. Illustration of NCL on a Surface, between a Surface-Bound Thioester (R ) Bn, Ph) and Cys-Liss 6 in NCL Buffer at pH 7.5a
Figure 1. Structural formulas for the Cys-Liss and Ala-Liss conjugates. Scheme 3. NCL in Solution
Cysteine-lissamine (Cys-Liss) conjugate 6 was prepared (details in Supporting Information; for structure, see Figure 1) and used to explore the reactivity of thioester surfaces under NCL reaction conditions. Alanine-lissamine conjugate (Ala-Liss) 7 was also prepared, as a control compound, because it is unable to undergo the NCL reaction. Lissamine was chosen as the dye label because it has no overall charge and shows absorption and emission characteristics that are almost independent of solvent and pH.25 As a first step, the NCL reaction was explored in solution (Scheme 3) using a slight modification of the procedure described by Kent and co-workers.20c Thioester exchange occurs between MPAA in solution and 8 to generate a more reactive thiophenyl ester. A thioester-linked intermediate is then formed by transthioesterification with the cysteine thiol of 6; this intermediate rapidly undergoes an intramolecular rearrangement to yield a native amide bond. Within 2 h, NCL reaction between 6 (0.1 mM) and 8 (0.1 mM) in sodium phosphate buffer (200 mM, pH 7.5) containing TCEP (10 mM) and MPAA (10 mM) was complete. The reaction was readily monitored by high-performance liquid chromatography (HPLC; details are given in the Supporting Information page S17). MPAA has three roles: it acts as a catalyst for NCL, reverses nonproductive thioester formation (with the product) and, with TCEP, maintains the N-terminal cysteine in its reduced form. These NCL conditions were then used to test the reaction on thiophenyl and thiobenzyl ester functionalized surfaces (Scheme 4). (25) Hermanson, G. T. Bioconjugate Techniques; Academic Press, Inc.: San Diego, CA, 1996; Chapter 8.
a NCL proceeds via trans-thioesterification with the side-chain thiol of Cys-Liss 6 to yield a thioester-linked intermediate that undergoes a spontaneous rearrangement to yield an amide bond.
Thioester functionalized surfaces were submerged in NCL buffer (sodium phosphate buffer (200 mM, pH 7.5), TCEP (10 mM) and MPAA (10 mM)) containing Cys-Liss 6 (100 µM) for two hours before quenching with large volumes of DMF and washing the substrates with NCL buffer and methanol. Reactions were tested under air and under nitrogen. The integrity of the Cys-Liss 6 ligation solution was checked by analytical HPLC after each reaction. Interestingly, only when the Cys-Liss 6 ligation solutions were left in the open laboratory for a period of a week did significant oxidation of Cys-Liss 6 to the corresponding disulfide dimer occur; Cys-Liss 6 solutions showed almost no change when left in the glovebox under nitrogen for the same period. UV/vis absorption spectroscopy was chosen to monitor the surface reactions because it yields a direct measurement of dye concentration on the surface, which, along with careful controls, can generate reliable data on the efficiency of the NCL reaction at a surface. The UV/vis absorption spectra of dry thioester substrates after treatment with NCL buffer containing Cys-Liss 6 are summarized in Figure 2. Estimates of lissamine surface coverage were calculated by comparing the (26) (a) Estimates of lissamine surface coverage were made using the following equation: Surface density (mol cm-1) ) 0.001 × (Integrated absorption of lissamine (cm-1)/Integrated extinction coefficient of lissamine (M-1 cm-2). The integrated extinction coefficient for lissamine (sulforhodamine B sodium salt) in aqueous solution was measured to be 9.3 × 107 M-1 cm-2. This method assumes that the integrated extinction coefficient (oscillator strength) of lissamine in solution is the same as that of surface-bound lissamine. (b) McCallien, D. W. J.; Burn, P. L.; Anderson, H. L. J. Chem. Soc., Perkin Trans. 1 1997, 2581–2586.
13966 Langmuir, Vol. 24, No. 24, 2008
Anderson Table 2. Summary of Contact Angles Measured with Water Using the Static Sessile Drop Methoda substrate surface octyl (solution) thiophenyl ester (solution)
chemical history
contact angle
as prepared NCL with Cys-Liss 6 NCL with Ala-Liss 7 as prepared NCL with Cys-Liss 6 NCL with Ala-Liss 7
101.5 ( 0.9° 101.5 ( 0.7° 100.0 ( 1.7° 82.0 ( 0.6° 56.7 ( 1.5 ° 80.3 ( 0.9°
a Octyl and thiophenyl ester surfaces were prepared in solution; contact angle measurements were then taken before and after submerging substrates into NCL buffer containing Cys-Liss 6 or Ala-Liss 7 for 2 h.
Figure 2. UV/vis absorption spectra of various thioester surfaces after functionalization with Cys-Liss 6 (a,b) and Ala-Liss 7 (c). Thioester substrates shown in (a) and (c) were prepared in solution, and those in (b) were prepared in the gas-phase. Both sides of the substrates were exposed to NCL reaction conditions. Table 1. Estimates of Dye (Lissamine) Surface Coverage, Calculated from Integrated Absorption Spectra26,a substrate
lissamine molecules/nm2 Solution
thiophenyl ester thiobenzyl ester octyl
2.03 0.17 0.00 Gas Phase
thiophenyl ester thiobenzyl ester octyl
0.70 0.09 0.00
a In each case, the absorption spectra of at least two different substrates were measured, and several measurements were made on the same substrate. Surface coverages were reproducible to within about 10%.
integrals of the absorption spectra with the integrated extinction coefficient of lissamine measured in aqueous solution.26 Lissamine surface densities from these experiments are summarized in Table 1. Substrates functionalized with thiophenyl ester groups (prepared in solution), treated with Cys-Liss 6 in NCL buffer showed the highest surface lissamine densities (largest absorption signals), followed by thiophenyl ester surfaces prepared in the gas phase; thiobenzyl ester surfaces performed poorly with almost no lissamine absorption observed on surfaces prepared in the gas
phase. These results indicate, as expected, that thiophenyl ester surfaces are more reactive than thiobenzyl ester surfaces, even in the presence of MPAA, which suggests that the rate of transthioesterification with MPAA is slower at a surface than in solution where MPAA acts as an efficient NCL catalyst with thiobenzyl ester derivative 8. They also show that substrates for NCL, prepared in solution, had a higher surface density of reactive groups than the corresponding substrates prepared in the gas phase. Two types of control experiments confirmed that nonspecific binding to surfaces was not taking place: unreactive Ala-Liss 7 was used instead of Cys-Liss 6 in the NCL reaction solution, and octyl surfaces (from glass functionalization with chloro(dimethyl)octylsilane) were used instead of thioester functionalized surfaces with Cys-Liss 6 in solution. In both cases, the surface-coverage of lissamine was below the detection limit for UV/vis absorption spectroscopy (ca. 5% of the value obtained for a fully functionalized surface). Only for the combination of a thiophenyl ester-functionalized surface and Cys-Liss 6 conjugate was a significant absorption band from the lissamine dye observed. These results are consistent with the surfaces being functionalized via NCL. McCarthy and co-workers23 estimated the footprint of the dimethylsilyl group to be 32-38 Å2; this corresponds to 2.0-2.4 molecules/nm2. The functionalization densities calculated in Table 1 for thiophenyl ester surfaces prepared in solution fall within this range and indicate that essentially all possible reaction sites on these surfaces react with Cys-Liss 6 during NCL. All NCL reactions at thiophenyl ester surfaces seemed to progress rapidly, and when the reaction was monitored, it was found to exhibit pseudo-first-order kinetics with a half-life of about 8 min (see Supporting Information).29 Water contact angles on thiophenyl ester- and octyl-coated substrates were measured using the static sessile drop method both, before and after treatment with NCL buffer containing Cys-Liss 6 or the control Ala-Liss 7 (see Table 2). Contact angles fell significantly for thiophenyl ester surfaces exposed to Cys-Liss 6 under NCL conditions. However, when the control Ala-Liss 7 was used instead of Cys-Liss 6 in the NCL buffer, substrates showed essentially no change in contact angle before and after exposure to the reaction mixture. This strongly suggested that significant hydrolysis or trans-thioesterification with MPAA of surface-bound thiophenyl esters was not occurring during the course of the NCL reaction, and the change in contact angle observed is due to the lissamine dye on the surface. The substrates were also analyzed by fluorescence scanning at a resolution of 200 µm. Unfortunately fluorescence quenching of lissamine is observed at all but the lowest surface loadings, so that fluorescence intensity provides a poor measure of surface coverage. It does, however, show that lissamine coverage is reasonably uniform over the whole substrate area exposed to the reaction mixture (images of scanned substrates are shown in the
Immobilization of N-Terminal Cysteines via NCL
Langmuir, Vol. 24, No. 24, 2008 13967
Figure 3. UV/vis absorption spectra for octyl/thiophenyl ester surfaces after functionalization with Cys-Liss 6 (a-e) or Ala-Liss 7 (f) (100 µM) during NCL (measurements were carried out in air). (a) 100% thiophenyl ester substrate treated with Cys-Liss 6, (b) 50% thiophenyl ester substrate treated with Cys-Liss 6, (c) 25% thiophenyl ester substrate treated with Cys-Liss 6, (d) 10% thiophenyl ester substrate treated with Cys-Liss 6, (e) 100% octyl substrate treated with Cys-Liss 6, and (f) 100% thiophenyl ester substrate treated with Ala-Liss 7. Table 3. Estimates of Dye (Lissamine) Surface Coverage for Substrates Functionalized with Varying Amounts of Thiophenyl Estersa lissamine molecules/nm2 % thiophenyl [Cys-Liss] [Cys-Liss] [Cys-Liss] ester (100 µM) (10 µM) (1 µM) 100 50 25 10 0
2.10 1.15 0.62 0.30 0.06
1.50 1.04 0.62 0.30 0.06
0.45 0.53 0.42 0.27 0.05
[Cys-Liss] (100 µM) Tween 20 (1%) 0.82 0.29 0.20 0.08 0.00
a These estimates were calculated from integrated absorption spectra.26 In each case, the absorption spectra of at least two different substrates were measured, and several measurements were made on the same substrate. Surface coverages were reproducible to within about 10%.
Supporting Information, Table S1), particularly for substrates prepared in solution. To further confirm that NCL took place on the surfaces and demonstrate that the surface-coverage can be controlled, a second series of substrates displaying mixtures of thiophenyl ester and octyl chains was prepared. The reactivities of these surfaces under NCL conditions were explored at various concentrations of Cys-Liss 6 and in the presence of the surfactant Tween 20. Tween 20 was included in the reaction mixture to reduce nonspecific binding between Cys-Liss 6 and thiophenyl ester functionalized surfaces. To prepare these new surfaces, clean glass was placed in toluene solutions containing mixtures of octyl dimethylsilylchloride and thiophenyl ester silyl chloride 1 in the presence of iPr2EtN for a week in a glovebox. Surfaces with the following proportions of thiophenyl esters were prepared: 100%, 50%, 25%, and 10%. After NCL with Cys-Liss 6, the surfaces were analyzed by UV/vis absorption spectroscopy (Figure 3). These data confirm that NCL occurs in a controllable manner at the thiophenyl ester sites on the surface, and that nonspecific binding accounts for less than 5% of the integrated absorption signal.27 Table 3 summarizes the dye surface densities measured for these surfaces at Cys-Liss 6 concentrations of 100 µM, 10 µM, and (27) The lowest level of lissamine functionalization for reliable integrated absorption measurements was found to be about 5% of the value obtained for a fully functionalized substrate, i.e., 2 molecules/nm2. At lower functionalization densities, the spectra are too noisy.
Figure 4. Graph showing lissamine surface density against percentage mole-fraction of thiophenyl ester on the surface after carrying out NCL with Cys-Liss 6 at various concentrations for the same time period. (a) 100 µM Cys-Liss, (b) 10 µM Cys-Liss, (c) 1.0 µM Cys-Liss, and (d) 100 µM Cys-Liss with 1% Tween 20.
1.0 µM, and Figure 4 shows these surface densities plotted against the percentage mole-ratio of thiophenyl ester on the surface. At a Cys-Liss 6 concentration of 10 µM (Figure 4, line (b)) full functionalization of thiophenyl ester sites was observed for 25% thiophenyl ester slides. When the concentration of Cys-Liss 6 was reduced to 1 µM, full functionalization of the phenyl thioester sites was only observed for the 10% thiophenyl ester slides. Lissamine absorption remained constant for 10% thiophenyl ester surfaces after treatment with Cys-Liss 6 solutions at 1, 10, and 100 µM, and nonspecific binding would be expected to be concentration dependent, providing further evidence for surface NCL. Introducing a nonionic detergent, (Tween 20, 1% v/v) resulted in much lower levels of NCL at the surface. Presumably, Tween 20 coats the thiophenyl ester surface and interacts with Cys-Liss 6 to form micelles in solution.28 In both instances, Cys-Liss 6 will be hindered from reaching the thiophenyl ester surface and hence from undergoing NCL.
Conclusions This work demonstrates the first single-step preparation of thiophenyl ester functionalized glass surfaces, and their characterization through UV/vis absorption spectroscopy after functionalization with Cys-Liss conjugate 6 under NCL conditions. The preparation of thioester silylchlorides 1 and 2 for the direct functionalization of glass greatly simplifies this approach. Thiophenyl ester functionalized surfaces prepared in solution reacted rapidly (half-life less than 10 min) with Cys-Liss 6 under NCL conditions, giving ca. 2 molecules of lissamine/nm2. This surface density corresponds to functionalization of essentially all the active thioester sites on the surface. Reaction between thiobenzyl ester substituted surfaces and Cys-Liss 6 under NCL conditions progressed slowly, and substrates prepared in the gas phase showed low levels of functionalization. When a series of surfaces displaying both thiophenyl esters and octyl groups was prepared, lissamine absorption was found to be linearly proportional to the percentage of thioester on the surface. The focus so far has been on establishing a robust method for the preparation of thioester functionalized surfaces and their characterization through functionalization with a simple Cys-Liss conjugate. Further work will explore thioester surface functionalization with proteins, peptides, and other molecules that contain an N-terminal (28) (a) A 1% solution of Tween 20 has a concentration of 8 mM, significantly higher than its critical micelle concentration (0.08 mM); see. (b) Malik, W. U.; Jhamb, O P. Kolloid. Z. Z. Polym. 1970, 242, 1209–1211. (29) For an interesting example of monitoring NCL at a surface see Wieczerzak, E., Hamel, R., Jr.; Chabot, V.; Aimez, V.; Grandbios, M.; Charette, P. G.; Escher, E. Biopolymers, 2008, 90, 415-420.
13968 Langmuir, Vol. 24, No. 24, 2008
cysteine suitable for NCL, with a view to preparing surfaces with specific surface characteristics, e.g., charge density, hydrophobicity/hydrophilicity, and optical properties for a number of applications. Supporting Information Available: Experimental procedures and characterization data for all new compounds, structure of CK-
Anderson fluorescein, surface characterization data including surface energy measurements, contact angles, fluorescence scanning, and surface reaction kinetics are all included in the Supporting Information. This material is available free of charge via the Internet at http://pubs.acs. org. LA8022024
