Protein Science
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Baumann, H.
Right arrow Articles by Carredano, E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Baumann, H.
Right arrow Articles by Carredano, E.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?
Protein Science (2003), 12:784-793.
Copyright © 2003 The Protein Society

Rational design, synthesis, and verification of affinity ligands to a protein surface cleft

Herbert Baumann1, Sara Öhrman1, Yasuro Shinohara2, Oguz Ersoy1, Devapriya Choudhury3,4, Andreas Axén1, Ulf Tedebark1 and Enrique Carredano1

1 Amersham Biosciences, Björkgatan 30, S-751 84 Uppsala, Sweden
2 Amersham Biosciences, 3-25-1, Hyakunincho, Shinjuku-ku, Tokyo 169-0073, Japan
3 Department of Molecular Biology, Swedish University of Agricultural Sciences, Uppsala, Sweden

Reprint requests to: Andreas Axén, Ulf Tedebark, or Enrique Carredano, Amersham Biosciences R&D, Björkgatan 30, S-751 84 Uppsala, Sweden; e-mail: andreas.axen{at}eu.amershambiosciences.com; ulf.tedebark{at}eu.amershambiosciences.com; or enrique.carredano{at}eu.amershambiosciences.com; fax: +46 18 612 18 44.

(RECEIVED October 21, 2002; FINAL REVISION December 13, 2002; ACCEPTED December 16, 2002)

Article and publication are at http://www.proteinscience.org/cgi/doi/10.1110/ps.0236603.

4 Present address: Centre of Biotechnology, Jawaharlal Nehru, New Delhi 110067, India. Back


   Abstract
TOP
 Abstract
Introduction
 Results and discussion
 Materials and methods
 References
 
The structure-based design, synthesis, and screening of a glucuronic acid scaffold library of affinity ligands directed toward the catalytic cleft on porcine pancreas {alpha}-amylase are presented. The design was based on the simulated docking to the enzyme active site of 53 aryl glycosides from the Available Chemicals Directory (ACD) selected by in silico screening. Twenty-three compounds were selected for synthesis and screened in solution for binding toward {alpha}-amylase using nuclear magnetic resonance techniques. The designed molecules include a handle outside of the binding site to allow their attachment to various surfaces with minimal loss of binding activity. After initial screening in solution, one affinity ligand was selected, immobilized to Sepharose (Amersham Biosciences), and evaluated as a chromatographic probe. A column packed with ligand-coupled Sepharose specifically retained the enzyme, which could be eluted by a known inhibitor.

Keywords: Affinity; {alpha}-amylase; chromatography; glucuronic acid; ligand; NMR; separation; Sepharose; structure-based design; synthesis

Abbreviations: ACD, Available Chemicals Directory • DMSOd6, perdeuterated dimethyl sulfoxide • DTT, dithiothreitol • MALDI-TOF, matrix-assisted laser desorption/ionization time-of-flight • NMR, nuclear magnetic resonance • PPA, porcine pancreas {alpha}-amylase • RP-HPLC, reversed phase high pressure liquid chromatography • Sepharose HP, Sepharose high performance • STD, saturation transfer difference • TFA, trifluoroacetic acid


   Introduction
TOP
 Abstract
 Introduction
Results and discussion
 Materials and methods
 References
 
Affinity chromatography has been used widely in biomedical research and biotechnology (Wilchek and Chaiken 2000). The inherent high specificity often enables purification factors over 1000 in a single step. Another powerful method, structure-based design, has become a standard tool in the development and improvement of drug candidates. However, despite earlier reports (Li et al. 1997; Sproule et al. 2000), this approach has not yet found broad application in the development of affinity media. One reason may be that such a challenge often involves additional complexity in regard to the need to develop ligands capable of functioning while surface localized, and target proteins difficult to design ligands for. This paper describes a successful attempt to use structure-based design as a fast and convenient route to develop new affinity ligands to be used in the face of such challenges.

PPA was selected as a suitable target protein. To the authors’ knowledge, there has been no earlier attempt to develop inhibitors or affinity ligands toward PPA by means of structure-based design. {alpha}-Amylase (EC 3.2.1.1) belongs to Family 13 of glycosyl hydrolases and catalyzes the hydrolysis of internal {alpha}-1,4-glucan links in preferably large linear polysaccharides containing three or more {alpha}-1,4-linked D-glucose units (Davies and Henrissat 1995). Protein- and carbohydrate-based inhibitors are known and have been extensively reviewed (e.g., Machius et al. 1996). Inhibitors belonging to the first category are generally very potent with typical KI in the nanomolar range. Acarbose as a representative for carbohydrate inhibitors has a reported KI in the order of 10 µM (Wilcox and Whitaker 1984).

For effective affinity chromatography, ligand-protein affinity must be high enough to provide selectivity over nonspecific interactions (e.g., binding to other sites on the same protein, as well as to other proteins) without hindering elution under acceptable chromatographic conditions. It has been indicated earlier that the ligands for affinity chromatography typically have an affinity constant in the range of 0.01–100 µM (Scopes 1987).

Here, we present the rationale, design, and testing of a glucuronic-acid, scaffold-based library of affinity ligands directed toward the PPA active site. The work includes synthesis of the directed library, and its testing for binding in solution using NMR techniques, immobilization of one compound with desired properties to Sepharose (Amersham Biosciences), and chromatographic evaluation of the rationally designed affinity media.


   Results and discussion
TOP
 Abstract
 Introduction
 Results and discussion
Materials and methods
 References
 
Rationale behind the design of the directed library
PPA ligand design was based on in silico screening of the MDL (MDL Information Systems, Inc.) ACD for ligands to porcine pancreas {alpha}-amylase (Amersham Biosciences, unpubl.). An initial docking of 53 possible aryl glycoside inhibitors from that study indicated that no available glycosides fulfilled the demands for both suitable affinity and the presence of a handle, that is, an available media attachment site. However, visual inspection inspired construction of a novel library based on a glucuronic-acid scaffold where the C-6 of an aryl glycoside is elongated, preferably with an aryl-type compound. The rationale of this new directed library was that an optimal ligand would have aryl groups on both sides of the carbohydrate moiety (Fig. 1Go).



View larger version (111K):
[in this window]
[in a new window]
  		Figure 1. Compounds (in stick model) docked into the active site cleft of PPA. The protein is shown in hydrophobic potential surface where blue denotes charged, green polar, and brown hydrophobic surface areas. Dotted-yellow lines suggest hydrogen bond possibilities. Top: Docked hits from the initial virtual screening of ACD. The binding site can accommodate aryl groups on both sides of the central carbohydrate moiety. Bottom: Docked compound G11-A44 from the virtual directed library. The brownish surface on one side of the cavity (right in the figures) corresponds to the side chain of Trp-59 amenable to interaction with aromatic rings through hydrophobic stacking.

 
Glycuronic acids, which have a 6-COOH as compared to glycosides that have a 6-CH2OH, can be coupled to an arylamine and thereby create an amide. To ensure further coupling capability to an insoluble matrix, all the arylamines must be bifunctional, that is, they must also include a handle (e.g., COOH, OH, SH, or NH2 in meta or para position). This position of the handle was chosen because from the modeling, it was suggested that it could then become available for coupling to gel without interfering with the binding. The glycuronic acid moiety was thought to fill the central and deepest sugar-binding subsite in the active-site cleft with the hydroxyl groups donating protons to all three carboxylate residues at the active site. The aryl part from each of the reactants were, on the other hand, expected to fill the lateral sugar-residue binding subsites in the active site for instance by interacting with the side chain of Trp-59 through {pi}-stacking interactions (Fig. 1Go).

Docking of the virtual combinatorial products
Nineteen commercial glycuronic acid compounds based on aryl ß-D glucuronic-acid type were found. However, after accounting for redundancy and absence of forbidden reactive groups, a total of nine were selected. Corresponding numbers for the arylamines were 47 and 26. A virtual combinatorial library consisting of 234 compounds was built using the 26 arylamines and the nine glucuronic acids. Of these, 229 were successfully docked into the active site. After visual inspection, 23 combinatorial virtual products, built from seven glucuronic-acid derivatives and eight arylamines (Fig. 2Go) were selected for synthesis (for details regarding the criteria used for docking and the selection of the ultimate library vide infra). Thirteen compounds fulfilling the rationale of the design, with accessible handle and complementing the binding site in charge, hydrogen-bond pattern, hydrophobicity, and conformation, were chosen as potential binders (one example shown in the bottom of Fig. 1Go) and 10 as negative controls.



View larger version (29K):
[in this window]
[in a new window]
  		Figure 2. Selected building blocks. Seven glucuronic-acid derivatives (left) and eight arylamine compounds used in the synthesis of a small library directed toward the active site of porcine pancreas {alpha}-amylase. The actual compounds screened, as shown in Table 1Go, are a product of condensing the carboxylic group of the glucuronic derivative with the amino group of the arylamine compound to form an amide bond.

 
Selection criteria of the ultimate library
Despite recent advances in the development of both first principles and knowledge-based docking programs and related scoring functions, these are still not able to discriminate false from true hits in an effective way in part because of steric effects and to neglecting dessolvation effects (Pérez and Ortiz 2001). However, docking functions are still useful because they do produce reasonable conformational sampling (Dixon 1997). Because of these limitations, in the selection of the ultimate library, the docking score of the compounds was given secondary importance. Instead, molecules were selected which, by visual inspection, provided complementarity to the active site in all of the following aspects: hydrophobicity, hydrogen-bonding pattern, charges, and conformation. The accessibility of the handle to the outside of the protein was considered a necessary prerequisite for selection. Also, synthetic feasibility was given consideration in the selection. Accordingly, molecules created with aromatic amino alcohols or aromatic di-amines were discriminated against harder because of additional synthetic steps in the solution-phase synthesis of the library, and instead, similar molecules with more feasible handles were chosen if they were considered equivalent based on the named criteria. Negative controls were chosen so as to enhance the expected structure-activity relationship information from the experimental screening and were alternative combinations of the same components forming the potential binders. Many of these combinations had favorable docking scores despite of an obvious hydrophobic mismatch: an aromatic carbon in v.d.w. contact with the {varepsilon}-nitrogen of His-101.

Synthesis of ligands
Twenty-three glucuronic-acid derivatives were synthesized in solution and purified by RP-HPLC prior to NMR screening. All non-natural amino acids (arylamines) were derivatized as methyl esters prior to coupling with glucuronic acids. Because of low reactivity, standard amide-forming reagents such as carbodiimide derivatives were not useful, yielding side products (Beuvery et al. 1986). Hence, the glucuronic acid was activated by reaction with isobutyl chloroformate, while the amine component was treated with sodium hydride prior to mixing with the activated glucuronic acid (Fig. 3Go).



View larger version (11K):
[in this window]
[in a new window]
  		Figure 3. Synthesis of ligands.

 
Screening in solution using NMR
The results of the STD-NMR experiments are summarized in Table 1Go. The 1D STD method was used as screening assay (Mayer and Meyer 1999). The resulting STD-NMR spectrum shows the difference between spectra recorded with on- and off-resonance irradiation of the protein, respectively. The two spectra are recorded in the same experiment in an interleaved fashion. If the resulting STD-NMR spectrum shows the same signals as the reference 1H-NMR spectrum of the ligand, the result is regarded as positive, that is, the ligand must have contacted the protein. Ligands that do not have any contact with the protein or are very tightly bound to the protein will not give any signal in the resulting STD-NMR spectrum. It has been shown that the method is capable of detecting ligands with dissociation constants between 10-3 and 10-8 M (Mayer and Meyer 1999). The strength of the STD-NMR signal depends upon several factors including protein size, offset, and duration of the on-resonance irradiation, the dissociation rate constant, and the excess of ligand. The STD-NMR method is advantageous in that the detection limits can be tuned for binding by varying the protein concentration while keeping the ligand concentration constant. Under such conditions, at higher protein concentrations, the weak to medium binders are detected, whereas at lower protein concentrations, only medium binders are detected. For instance, it has been shown before for another enzyme system that both micromolar and millimolar binders were detected at a protein concentration of 35 µM, whereas only micromolar binders were detected at protein concentrations of 1 µM and 100 nM (Peng et al. 2001). It should be noted that the signal intensity at one specific protein concentration should not be taken as a direct measure of the binding strength. For instance, in the same study, a millimolar binder showed a stronger signal as compared to a micromolar binder at 35 µM protein concentration, whereas when the protein concentration was reduced to 1 µM, the signal from the weaker binder vanished. On the other hand, the signal of the µM binder became even stronger than before (Peng et al. 2001). Here, three different protein concentrations were used, namely, 2, 8, and 25 µM. The compounds were screened in solution as methyl esters and not as free acids because a handle consisting of a methyl ester was thought to better mimic the immobilized ligand. However, only 14 of the 23 compounds in the original library were soluble in the experimental conditions and could thus be tested. Three of these compounds gave a positive binding result at the lowest protein concentration and were designated as the most promising affinity candidates. Five additional compounds gave a positive binding result at the higher protein concentrations and were thus designated as less promising affinity candidates. Figure 4Go shows the STD NMR spectra of the compound that was finally chosen as a potential affinity ligand, then coupled to Sepharose and evaluated chromatographically.


View this table:
[in this window]
[in a new window]
  		Table 1. Outcome of NMR-analysis of affinity properties for synthesised compounds
 


View larger version (27K):
[in this window]
[in a new window]
  		Figure 4. Reference and STD NMR spectra of compound G5-A33. The spectra have been expanded between 5 and 8.5 ppm showing the signals for the aromatic protons and the signal for the anomeric proton of the glucuronic acid moiety. Samples contained 1 mM ligand, Tris buffer pH 7.5 and protein concentrations as described for each spectrum (B–D). (A) Reference 1D 1H-NMR spectrum. (B) Corresponding STD NMR spectrum collected using 25 µM protein. (C) Corresponding STD NMR spectrum collected using 8 µM protein. (D) Corresponding STD NMR spectrum collected using 2 µM protein.

 
Structure activity relationships
By activity here is meant the binding behavior as measured by the NMR method in solution, still aware of the limitations in affinity window that the chosen conditions give. This activity criterion was chosen because the information obtained would fulfill our ends (vide supra). Given this definition, affinities outside this window (i.e., too high or too low affinity) are not detected and considered as not active.

In general, the chosen glucuronic acids displayed binding. The only exception, G3, deploys activity from all compounds of which it is part. Interestingly, G18a, which differs from G3 only in possessing a naphthalene aglycon structure instead of a quinoline structure, contributes to binding activity. On the other hand, the difference in structure between G18a and G18b {alpha}- and ß-naphthyl, respectively, does not seem to be of importance for binding affinity. Among the amines, it seems that A42 gives the best contribution to binding. Especially, the N-acetyl group in the structure of A42 contributes more to binding than the nitro group located in the corresponding position in the structure of A41.

Potential effects of DMSO
The presence of DMSOd6 in the NMR screening could potentially have an effect on the protein and the binding of small molecules to the protein (contra, an aqueous environment indirectly implied in the knowledge-based scoring function of the virtual docking process). To investigate this, binding experiments in pure buffer and buffer with 10% DMSOd6 were performed with the known inhibitor acarbose showing that the interaction was retained even in the presence of 10% DMSOd6 (data not shown). Therefore, it was concluded that the presence of 10% DMSOd6 was not significant for the interaction between the ligands and the protein.

Comparison of data with prediction from docking
In 10 out of 14 possible cases, it was possible to predict the presence or absence of some type of interaction in the lowest affinity window (Table 1Go). Also, the ligands regarded as most promising candidates, namely G5-A33, G18a-A42, and G18b-A42, were all predicted to be positive. There were three predicted positive compounds without any measurable interaction (all involving G3) and one negative control that showed binding in the experiment (G8-A37). This is a more than acceptable outcome taking into account the limitations of the modeling, and that the experimental screening has only a limited affinity window. It also verifies the wisdom of including visual inspection as an aid to docking score interpretation. Also, this library yielded a sufficient high number of compounds with properties that fulfilled the a priori criteria to allow a selection of one master from several possible candidates. In that sense, the modeling met the challenge to be a useful tool for finding suitable affinity ligands.

Strategy for selection of ligand to be coupled to matrix
The selection of a candidate for coupling to a solid support was based on fulfillment of several criteria. Affinity was considered the most important factor. Because of this, only the compounds designated as most promising from the NMR analysis were considered. Among these three compounds, G5-A33 (Fig. 5Go) was chosen based on synthetic considerations such as chemical yield and availability of starting materials.



View larger version (8K):
[in this window]
[in a new window]
  		Figure 5. Ligand G5-A33 selected for chromatographic evaluation.

 
Coupling of ligand to Sepharose and chromatographic evaluation
The selected affinity ligand was prepared on a larger synthetic scale and attached to a Sepharose HP gel by one of Amersham Biosciences’ standard coupling techniques.

In the chromatographic evaluation, it was considered important to distinguish nonspecific interactions between the target protein and the matrix from specific interactions between the affinity ligand and the target protein. Hence, a buffer system containing both high salt concentration (to prevent nonspecific ionic interactions) and an organic solvent (to prevent nonspecific hydrophobic interactions) was chosen. It turned out that {alpha}-amylase was captured in a chromatographic system until acarbose, as a specific competitor to the interaction, was introduced (Fig. 6Go). This was compared to the behavior of collagenase used as a reference protein, which was not retarded at all under the conditions chosen (data not shown).



View larger version (14K):
[in this window]
[in a new window]
  		Figure 6. Chromatographic evaluation. By-pass injection of {alpha}-amylase at 2 min. Injection of {alpha}-amylase on to column at 4 min. Elution of {alpha}-amylase with 10 mM acarbose at 12.5 min. Blank as inset. X-axis in minutes.

 
The results from the chromatographic evaluation were confirmed by collecting fractions of material eluted with acarbose. This material was subjected to further analysis by RP-HPLC. The RP-HPLC analysis showed that the material eluted with acarbose from the affinity column contained material with retention times similar for those of acarbose and {alpha}-amylase, respectively (Fig. 7Go). The peak with a retention time corresponding to {alpha}-amylase in the RP-HPLC chromatogram was collected and analyzed using MALDI-TOF mass spectrometry. The result indicated that the eluted material contains {alpha}-amylase, when comparing the mass corresponding to the single charged peak with the calculated mass from the amino-acid sequence in the PDB file (Fig. 8Go). To further confirm the identity of this material, a trypsin finger printing was performed, unambiguously showing that the material consisted of {alpha}-amylase from porcine pancreas (data not shown).



View larger version (116K):
[in this window]
[in a new window]
  		Figure 7. RP-HPLC analysis. Top chromatogram showing RP-HPLC analysis of the material eluted with acarbose during affinity chromatography. Middle chromatogram showing the corresponding analysis of {alpha}-amylase (SIGMA). Bottom chromatogram showing the corresponding analysis for buffer B, containing acarbose.

 


View larger version (19K):
[in this window]
[in a new window]
  		Figure 8. MALDI-TOF mass spectrum showing analysis of material eluted at 6.2 min in the RP-HPLC analysis. Calc. Mw. ({alpha}-amylase) 55384.

 
These results indicate the interaction between the prepared affinity chromatography medium and {alpha}-amylase has an appropriate binding constant, and is both specific and selective.

Conclusions
Structure-based design was proven to be an efficient route to develop a new affinity ligand for capture of {alpha}-amylase, a protein whose surface cleft active site presents a possible but challenging affinity target site. Such a rational approach allowed rapid development of a ligand with a molecular handling for coupling to a matrix. A selected inhibitor could specifically elute the enzyme after retention in a column packed with affinity gel. NMR proved particularly valuable in relating the structures of suitable ligands to their performance.

This study demonstrated that virtual ligands with high probable affinity also bind to target protein in solution and, if so verified, might also be suitable as ligands in affinity chromatography when the position of their matrix attachment is evaluated at the earlier stages of the design and selection process.


   Materials and methods
TOP
 Abstract
 Introduction
 Results and discussion
 Materials and methods
References
 
Molecular modeling
The program package SYBYL version 6.7 (Tripos, Inc.) was used for all modeling. All computations have been carried out in an OCTANE (Silicon Graphics, Inc.) workstation provided with two 195 MHz R10000 processors.

Preparation of molecules for docking
The molecules to be docked were prepared by first transforming the two-dimensional (2D) coordinates into three dimensions (3D) with the program CONCORD (Tripos, Inc.), ionized to consider their protonation state at neutral pH and finally minimized (500 cycles) using the MMFF94 force field (Halgren 1996).

Docking of prepared molecules
All docking simulations have been carried out with the program FlexX (FhI SCAI / BioSolveIT GmbH; Rarey et al. 1996), which is part of the SYBYL package. The default FlexX scoring function was used throughout the simulations. FlexX uses a fast docking method that allows flexibility in the ligands, keeping the receptor fixed. FlexX uses formal charges, which were turned on during the docking simulations. The protein structure used was the 2.03 Å resolution structure of porcine pancreatic {alpha}-amylase in complex with a substrate analog methyl 4,4'-dithio-{alpha}-maltotrioside (Qian et al. 1997) with accession code 1jfh to the Protein Data Bank (PDB). In the protein structure, the {varepsilon} carbon of His-101 is located 2.8Å away from the carbonyl oxygen of Tyr-62. Because this is probably an error as a result of misinterpretation of the electron density in the imidazole ring of His-101, this ring was flipped around 180°. In this modified structure, the {varepsilon} nitrogen of His-101 is at favorable hydrogen bonding distance to the carbonyl oxygen of Tyr-62. Otherwise, defaults have been used when creating the receptor definition file and no special customizations have been done. The residues belonging to the active-site file are Trp-58, Trp-59, Tyr-62, Gln-63, Asp-96, Val-98, His-101, Tyr-151, Val-157, Leu-162, Val-163, Leu-165, Arg-195, Asp-197, Ala-198, Lys-200, His-201, Glu-233, Ile-235, Phe-256, Asn-298, His-299, Asp-300, Asn-301, His-305, Gly-306, and Ala-307. In the case of the virtual combinatorial library, the SYBYL LINE NOTATION core option of FlexX in SYBYL was applied with input C[1]H(CH(CHOCHCH@ 1OH)OH)OH to indicate to the program to start fragment build-up using the central sugar substructure of the compounds.

Docking of aryl glycosides from virtual screening
A total of 53 aryl glycosides obtained from virtual screening (Amersham Biosciences, unpubl.) were docked into the active site. After visual inspection, nine inspired the generation of a small, directed library.

Enumeration and docking of virtual directed library
The virtual combinatorial library consisting of 234 compounds was created with the program LEGION (Tripos, Inc.) of the TRIPOS package by combining 26 amines with nine glucuronic-acid derivatives, all starting material being commercially available. Prior to docking, the compounds in the combinatorial library with carboxylic acid functionality were further transformed into methyl esters because later the compounds were tested in solution as such. In this way, the formal charge of the carboxylic acid was removed in the modeling work, thereby avoiding an artificial interaction.

Synthesis
General methods
1H and 13C NMR spectra were recorded on a Bruker Avance 300 or 500 MHz spectrometer. The glucuronic acid derivatives were purified on a Gilson HPLC system (ACE C18 column, water-acetonitrile system with the addition of 0.05% TFA). Product mass was verified by mass spectrometry on a Hewlett Packard 1100 MSD LC-MS. Analytical RP-HPLC were run on a Shimadzu 10A system with ultraviolet detection at 214 and 250 nm (water-acetonitrile +0.05% TFA, ACE C18).

Esterification of carboxylic acids
Sulfuric acid was added to a solution of the amino-acid component in methanol until the pH reached 2. The stirred reaction was left at reflux overnight. The solvent was removed in vacuo, and the residue dissolved in chloroform, washed with saturated aqueous sodium carbonate solution, dried (MgSO4), and evaporated to give the methyl ester in high yield.

Synthesis of glucuronic acid derivatives
Pyridine (4.0 µL, 50 µM) and isobutylchloroformate (8.0 µL, 60 µM) was added to a solution of the glucuronic acid derivative (50 µM) in DCM (1 mL) at 0°C. The reaction was stirred for 1 min before the addition of the deprotonated amine component. Deprotonation of the amine component was carried out by dissolving amine (75 µM) in DCM (1 mL), adding NaH (100 µM) at 0°C followed by stirring for 1 min.

The stirred mixture was left for 30 min at 0°C and an additional hour at room temperature. Solvent was evaporated and the residue dissolved in methanol. Purification by RP-HPLC gave the desired product in yields varying between 10% and 30%.

Three of the glucuronic acids (G11, G8, and G17) were supplied as cyclohexylammonium salts and were desalted on RP-HPLC prior to use.

Screening using NMR
All NMR experiments were performed at 298 K on a Bruker Avance 500 MHz spectrometer. The 1D saturation transfer difference method (STD NMR) was used as screening assay (Mayer and Meyer 1999) using several protein concentrations in order to differentiate between compounds with different binding strength (Peng et al. 2001). Because of the small size of the library, compounds could be individually screened. On-resonance irradiation was set at 0 ppm and off-resonance irradiation was set at -40 ppm. Irradiation time in each scan was 2 sec and 16 K data points were collected with 1024 scans in total. Compounds for testing were dissolved in DMSOd6 to a concentration of 50 mM. Ten microliters of the concentrated ligand solution was added to 490 µL buffer solution, thus yielding 1 mM ligand and 10% DMSOd6. The buffer consisted of 20 mM Tris, 3 mM DTT, and 8% DMSOd6 in D2O at pD 7.5, uncorrected reading on pH meter.

Compounds were initially tested for binding with 25 µM protein. Interesting ligands were further tested with protein concentrations of 8 or 2 µM. A one-dimensional 1H-spectrum was acquired first, and subsequently a STD spectrum was acquired. Each analysis took 90 min on the spectrometer. A positive result was obtained if signals from the ligand were observed in the difference spectrum.

Coupling of ligand to Sepharose
Ligand G5-A33 was synthesized according to the general description above and purified on RP-HPLC before coupling to an EAH-Sepharose HP, 34 µM/mL, gel (Allyl glycidyl ether activation of Sepharose HP, followed by coupling of 1,6-diamino hexane according to one of Amersham Biosciences standard protocols).

Coupling was performed by dissolving ligand (75 µM) in 50% of dioxane (4 mL) followed by the addition of 34 µM gel washed with 50% dioxane and the addition of EDC (500 µM). The resulting suspension was put on a shaker over night, occasionally checked, and when necessary, adjusted the pH to ~7. Finally, the gel was washed with water, 50% dioxane, water, and 20% ethanol.

Chromatographic evaluation
Medium with coupled ligand G5-A33 was packed in a column (0.5 x 2 cm) that was mounted on an |f#KTA explorer 10S system (Amersham Biosciences) and equilibrated with 10 mM HEPES buffer pH 7 containing 3 mM CaCl2, 1 mM DTT, 1 M NaCl, and 5% ethanol.

As test sample, porcine pancreas {alpha}-amylase (SIGMA) diluted 10 times from the storing buffer was used. Approximately 30 µg protein was applied in each injection.

As reference, protein collagenase (SIGMA) dissolved in the starting buffer to a final concentration of ~4 mg/mL. Approximately 40 µg protein was applied in each injection.

As elution buffer, the starting buffer with addition of 10 mM of acarbose was used. The column was regenerated with 8 M of guanidine between every analytical run.

Fraction collected was analyzed on RP-HPLC, and after desalting with a ZipTip C4 (Millipore) with a Bruker Reflex 4 MALDI-TOF mass spectrometer calibrated with bovine serum albumin.

Trypsin finger printing was performed by digestion of the sample with trypsin followed by analysis with of the cleavage products with MALDI-TOF mass spectroscopy. The pattern found in the mass spectrometry analysis was compared to a similar analysis from a database.


   Acknowledgments
 
We acknowledge Dr. James Van Alstine for a valuable linguistic and scientific revision of the manuscript.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 USC section 1734 solely to indicate this fact.


   References
TOP
 Abstract
 Introduction
 Results and discussion
 Materials and methods
 References
 
Beuvery, E.C., Speijers, G.J., Lutz, B.I., Freudenthal, D., Kanhai, V., Haagmans, B., and Derks, H.J. 1986. Analytical, toxicological and immunological consequences of the use of N-ethyl-N'-(3-dimethylaminopropyl) carbodiimide as coupling reagent for the preparation of meningococcal group C polysaccharide-tetanus toxoid conjugate as vaccine for human use. Dev. Biol. Stand. 63: 117–128.[Medline]

Davies, G. and Henrissat, B. 1995. Structures and mechanisms of glycosyl hydrolases. Structure 3: 853–859.[Medline]

Dixon, J.S. 1997. Evaluation of the CASP2 docking section. Proteins 1: 198–204.

Halgren, T. 1996. Merck molecular force field. I. Basis, form, scope, parameterization, and performance of MMFF94. J. Comp. Chem. 17: 490–519.[CrossRef]

Li, R., Dowd, V., Stewart, D.J., Burton, S.J., and Lowe, C.R. 1997. Design, synthesis, and application of a protein A mimetic. Nat. Biotechnol. 16: 190–195.

Machius, M., Vertesy, L., Huber, R., and Wiegand, G. 1996. Carbohydrate and protein-based inhibitors of porcine pancreatic {alpha}-amylase: Structure analysis and comparison of their binding characteristics. J. Mol. Biol. 260: 409–421.[CrossRef][Medline]

Mayer, M. and Meyer, B. 1999. Characterization of ligand binding by saturation transfer difference NMR spectroscopy. Angew. Chem. Int. Ed. 38: 1784–1788.[CrossRef]

Peng, J.W., Lepre, C.A., Fejzo, J., Abdul-Manan, N., and Moore, J.M. 2001. Nuclear magnetic resonance-based approaches for lead generation in drug discovery. Methods Enzymol. 338: 202–230.[Medline]

Pérez, C. and Ortiz, A.R. 2001. Evaluation of docking functions for protein-ligand docking. J. Med. Chem. 44: 3768–3785.[CrossRef][Medline]

Qian, M., Spinelli, S., Driguez, H., and Payan, F. 1997. Structure of a pancreatic {alpha}-amylase bound to a substrate analogue at 2.03 Å resolution. Protein Sci. 6: 2285–2296.[Abstract]

Rarey, M., Kramer, B., Lengauer, T., and Klebe, G. 1996. A fast flexible docking method using an incremental construction algorithm. J. Mol. Biol. 261: 470–489.[CrossRef][Medline]

Scopes, R.K. 1987. Protein purification: Principles and practice, Springer Verlag, New York.

Sproule, K., Morrill, P., Pearson, J.C., Burton, S.J., Hejnæs, K.R., Valore, H., Ludvigsen, S., and Lowe, C.R. 2000. New strategy for the design of ligands for the purification of pharmaceutical proteins by affinity chromatography. J. Chromatogr. B. 740: 17–33.[CrossRef]

Wilchek, M. and Chaiken, I. 2000. An overview of affinity chromatography. Methods Mol. Biol. 147: 1–6.[Medline]

Wilcox, E.R. and Whitaker, R.J. 1984. Some aspects of the mechanism of complexation of red kidney bean {alpha}-amylase inhibitor and {alpha}-amylase. Biochemistry 23: 1783–1791.[CrossRef][Medline]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us   Add to Digg Digg   Add to Reddit Reddit   Add to Technorati Technorati    What's this?


This article has been cited by other articles:


Home page
 Protein Sci.Home page
E. Carredano, H. Baumann, A. Gronberg, N. Norrman, G. Glad, J. Zou, O. Ersoy, E. Steensma, and A. Axen
A novel and conserved pocket of human {kappa}-Fab fragments: Design, synthesis, and verification of directed affinity ligands
Protein Sci., June 1, 2004; 13(6): 1476 - 1488.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Baumann, H.
Right arrow Articles by Carredano, E.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Baumann, H.
Right arrow Articles by Carredano, E.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS