Affi Lig Proteomics

DTD 5

ARTICLE IN PRESS

Biotechnology Advances xx (2005) xxx – xxx www.elsevier.com/locate/biotechadv

Advances and applications of de novo designed affinity ligands in proteomics A. Cecı´lia A. Roque, Christopher R. Lowe* Institute of Biotechnology, University of Cambridge, Tennis Court Road, CB2 1QT, Cambridge, UK Received 29 April 2005; accepted 9 May 2005

Abstract Affinity chromatography represents a promising technique for decoding the proteomics universe. While conventional affinity purification is being used in conjunction with two-dimensional electrophoresis (2D-PAGE) and mass spectrometry (MS) for the study of proteomes and subproteomes, scientists are still confronted with the need for specific and tailor-made affinity ligands to target desired groups and families of proteins. Evidence has shown that, in many situations, synthetic affinity ligands can circumvent inconveniences associated with the utilisation of biological ligands for the chromatography-based purification of biomolecules. This review will highlight the potential applications of affinity chromatography and synthetic de novo designed ligands as separation tools for proteomics. D 2005 Elsevier Inc. All rights reserved. Keywords: Affinity chromatography; Synthetic ligands; Biomimetics; Proteomics

1. Introduction The availability of complete genome sequences of many eukaryotic and prokaryotic organisms builds enthusiasm and challenges to biological researchers who wish to

* Corresponding author. Tel.: +44 1223 334157; fax: +44 1223 334162. E-mail address: [email protected] (C.R. Lowe). 0734-9750/$ - see front matter D 2005 Elsevier Inc. All rights reserved. doi:10.1016/j.biotechadv.2005.05.001

JBA-05983; No of Pages 14

ARTICLE IN PRESS 2

A.C.A. Roque, C.R. Lowe / Biotechnology Advances xx (2005) xxx–xxx

unravel the secrets of the gene products, the proteins. The equivalence between genes and the proteins they code for is generally a non-straightforward process, not only because many post-transcriptional and post-translational modifications take place, but many cellular regulation mechanisms also control the expression of genes. Proteomics can be defined as the analysis of part or all of the protein complements of a complex biological system at any given moment in time and it includes the characterization and quantitation of protein expression, function and structure (Shi et al., 2004). Proteome analysis requires techniques that could provide high throughput, high sensitivity and resolution, as well as the possibility of quantifying and identifying PTM (post-translational modifications) in proteins. Conventional experimental techniques used in proteomics include two-dimensional electrophoresis (2D-PAGE) and mass spectrometry (MS). Two-dimensional electrophoresis is widely used for the separation and quantitation of thousands of intact proteins in a single run. It also presents the possibility of applying labelling techniques for the detection of PTM and comparative protein expression in selected cell populations. It is therefore possible to use general protein detection reagents for a holistic approach (organic dyes, silver stain, radiolabelling and fluorescent stains, amongst others) and specific detection methods to reveal the status of proteins, such as the level of phosphorylation (e.g. autoradiography 32P and 33P), glycosylation (e.g. dansyl hydrazine), proteolytic modifications (e.g. zymographic assays for serine proteases), S-nitrosylation (e.g. Biotin Switch method), arginine methylation (e.g. immunoblotting) and ADP-ribosylation (e.g. 1,6-etheno NAD+) (Patton, 2002). Another option employs recombinant-DNA strategies to couple affinity tags or reporter enzymes to proteins and this approach has proven useful both at the detection and purification levels. Examples of these affinity tags include the FLAG and oligohistidine peptides, the green fluorescent protein and the enzyme h-glucuronidase. Differential display proteomics is based on the comparison of different protein profiles — examples of this approach using gel electrophoresis are difference gel-electrophoresis (DIGE) and multiplexed proteomics (MP). Isotope-encoded affinity tagging is another powerful strategy that combines affinity chromatography, gel electrophoresis and peptide mass profiling by MS (Patton, 2002). Despite the developments achieved in 2D-PAGE, there are still major technical disadvantages such as: poor reproducibility, difficult detection of extremely acidic or basic proteins (pI b 3.5 and N9.5), hydrophobic or membranelocated proteins, as well as proteins with extreme range of molecular weights, and limitations on the amount of proteins that can be loaded and deficient detection of low abundance proteins. Mass spectrometry-based matrix-assisted laser desorption/ionization (MALDI) and electrospray ionization (ESI) have been the most popular MS methodologies used for protein and peptide identification in proteomics (Newton et al., 2004). Other strategies have included, for example, quadrupole ion-trap (LCQ) and Fourier transform ion cyclotron resonance (FTICR). MS-based techniques for the separation and sequencing of peptides are usually a bbottom-upQ approach, complementary to upstream electrophoretic or chromatographic methods (for the separation of intact proteins) and digestion of proteins to smaller peptide fragments (Jensen, 2004). This bbottom-upQ tactic to proteomics has proved highly flexible and versatile. Intact-protein btop-downQ

ARTICLE IN PRESS A.C.A. Roque, C.R. Lowe / Biotechnology Advances xx (2005) xxx–xxx

3

approaches, where individual proteins are selected for MS analysis without any previous chemical or enzymatic proteolysis, have also been used in some situations as a means to complement peptide-level measurements with more information on the protein modification state. However, not all intact proteins are amenable to this btop-downQ methodology mainly due to intrinsic factors related with low solubility and molecular weight (Jacobs et al., 2005, Loo et al., 2005, Fountoulakis and Vlahou, 2005).

2. Affinity chromatography in proteomics Liquid chromatography is already considered an indispensable tool in proteomics by allowing the separation of macromolecules by different properties (Shi et al., 2004). The type of interactions and techniques explored may include reverse-phase, ion exchange, size-exclusion, and hydrophobic or affinity chromatography. The affinity concept relies on specific, reversible and non-covalent interactions between two biological entities forming a complex. Affinity technology exploits these natural specific recognition phenomena and the predictive and rational character of the binding between the proteins to purify and the complementary ligand. Association constants between the ligand and the protein in the range 103–108 (M 1) are normally the most suitable for purification purposes (Janson, 1984). As this binding is highly specific, the use of affinity purification reduces non-specific interactions, increases operational yields and facilitates the elimination of undesirable contaminants. The adsorption is efficient even from diluted crude extracts, which makes possible a one-step purification of the target proteins. Therefore, affinity chromatography, in particular, represents an appealing option to complement proteomics analysis, not only by separating target proteins based on their specific interaction with immobilized ligands but also by allowing the identification of protein–protein interactions that can allow the function or activity of the proteins to be inferred. Whilst the first protein to be purified by affinity chromatography was a-amylase (1910) by adsorption onto insoluble starch, the term baffinity chromatographyQ was coined in 1968 by Cuatrecasas. Since then, affinity chromatography has matured as a technique with acknowledged contributions in small and preparative-scale, as well as in large-scale purification of biomolecules, and has also suffered major refinements in order to comply with exigencies from regulatory agencies (Lowe, 2001; Roque et al., 2004a). The type of affinity ligand utilised has also evolved substantially: Groupspecific adsorbents, such as coenzymes, lectins, boronates and nucleic acids, were the first to be used as affinity ligands in the early 1970s, whilst textile dyes became popular in the late 1970s–1980s and inspired the introduction of designer dyes or biomimetic textile dyes (Lowe et al., 1992) and, later, the evolution into de novo designed ligands (Lowe et al., 2001). In these terms, affinity chromatography has actually been used in its more general sense of battractionQ: Not only the natural ligand involved in a biological interaction with the target protein is considered but also unnatural ligands that interact with protein’s binding site by non-biospecific interactions. The first group of ligands is referred to as bbiospecific ligandsQ, whereas the second group of ligands is termed bpseudobiospecific ligandsQ (Fig. 1).

ARTICLE IN PRESS 4

A.C.A. Roque, C.R. Lowe / Biotechnology Advances xx (2005) xxx–xxx

Fig. 1. Classification of affinity ligands for the purification of proteins by affinity chromatography. Biospecific ligands are naturally occurring molecules with affinity for the target protein-binding site. Pseudobiospecific ligands can be biological or non-biological molecules that interact with the target protein but do not occur naturally in the biological systems. Biomimetic ligands are a new class of pseudobiospecific ligands that are tailor-made to a specific target protein in order to increase specificity and the overall operational performance of the affinity adsorbents. HCIC — hydrophobic charge induction chromatography; IMAC — immobilized metal affinity chromatography.

Pseudobiospecific ligands may be synthetic (dyes) or biological (peptides, oligonucleotide aptamers) and some belong to the youngest class of biomimetic ligands. These ligands are tailor-made molecules that mimic the natural biological recognition between a target protein and a natural ligand. Biospecific ligand affinity chromatography has some inherent drawbacks that are difficult to overcome. The ligand has to be produced, usually at high cost, and biases its full characterization. On the other hand, the coupling yield is often low and the immobilization may modify the ligand. There is also a high probability of ligand leaching during harsh elution steps (Lowe, 2001). The drawbacks of pseudobiospecific affinity chromatography can be more easily surmounted, but they are dependent on the biological or synthetic source of the ligand. Still, combinatorial methods allied with molecular modelling and genetic engineering have contributed enormously to the development of highly resistant pseudobiospecific ligands. Affinity chromatography can be utilised in different stages of a proteomics analysis (Lee and Lee, 2004). A general scheme of the possible strategies to incorporate affinity chromatography in proteomics is depicted in Fig. 2. Affinity chromatography can be utilised as a depletion method, where certain proteins or a specific group of proteins, especially if present in high concentrations, can be removed to enhance the probability to visualize low abundance proteins in 2D-PAGE gels. This method can be

ARTICLE IN PRESS A.C.A. Roque, C.R. Lowe / Biotechnology Advances xx (2005) xxx–xxx

5

Fig. 2. Affinity chromatography and its possible applications in proteomic analysis.

problematic by diluting the sample during the washing step and also by losing lowabundance proteins if they present affinity for the immobilized ligands. Another option is the concentration/extraction method, where affinity chromatography can be helpful

ARTICLE IN PRESS 6

A.C.A. Roque, C.R. Lowe / Biotechnology Advances xx (2005) xxx–xxx

in the isolation of a small group of proteins before proceeding to analysis by 2DPAGE or by MS after proteolytic digestion. The concentration mode proves particularly useful for increasing the amount of low-abundance proteins or for the study of subproteomes. In the latter case, analysis of the phosphoproteome, for example, can be accomplished by immunoaffinity chromatography or by IMAC (Kalume et al., 2003) as these steps help in the enrichment of both phosphorylated proteins and peptides. Some comment should be made on the fact that multiple affinity columns might be required since only a single-protein or group of proteins can be enriched at any time when employing affinity chromatography. In general, when affinity chromatography is positioned prior to 2D-PAGE to separate proteins, the affinity ligands immobilized are a specific antigen, group-specific adsorbents (e.g. lectin to capture glycoproteins) or IMAC-ligands. When affinity chromatography is placed after enzyme digestion and peptides are being separated, then it is more common additionally to exploit ligands related to previous peptide chemical modifications (Jensen, 2004; Roque and Lowe, in press). In a more general approach, proteins and peptides can be modified or tagged to facilitate downstream separation (Scheich et al., 2003; Jensen, 2004).

3. Rational design in affinity chromatography The concept of the bbiomimeticQ ligand was initially introduced as an upgrade of textile dyes, i.e.bdyestuffsQ designed to better mimic the structure and binding of natural biological ligands (Lowe et al., 1992). Due to developments in computational technology, combinatorial libraries and high-throughput screening techniques, this concept has been extended to include not only synthetic biomimetic dyes and triazine non-dye ligands (designed de novo ligands), but also peptides and minimised protein domains (Roque et al., 2004a). A good example of biomimetic affinity ligands is the group of protein A ligands applied in the purification of antibodies, which includes a protein A mimetic multimeric peptide ligand (PAM-TG19318) (Fassina et al., 1996) and the de novo designed artificial protein A (ApA) (Li et al., 1998; Teng et al., 2000). The availability of crystallographic structures of proteins and complexes, together with the development of computer-based molecular modelling techniques, has prompted the bintelligentQ design and synthesis of affinity ligands with improved characteristics over their natural counterparts. These improvements include chemically defined and characterized ligands which are easy to synthesize with moderate to high specificity, the possibility of using mild elution conditions, the high stability of the ligand to sterilization-in-place and cleaning-in-place procedures, the high yield of ligand utilisation, and the low cost and high scalability (Roque et al., 2004a). Lowe and co-workers have been pioneers in de novo ligand design and synthesis of triazine scaffolded affinity ligands for the purification of different target biomolecules (Lowe et al., 1992, 2001; Lowe, 2001). The methodology represents an integrated procedure by combining structure-based design with combinatorial chemistry and integrating solid-phase synthesis of ligands with screening in parallel for affinity

ARTICLE IN PRESS A.C.A. Roque, C.R. Lowe / Biotechnology Advances xx (2005) xxx–xxx

7

chromatography of proteins (Fig. 3). Three main design strategies have been followed: (i) the template for design is a natural partner involved in binding to the target protein (Li et al., 1998); (ii) the template for design is a molecule that binds to the target site of the protein (Teng et al., 1999; Filippusson et al., 2000); and (iii) direct mimicking of natural biological recognition interactions (Palanisamy et al., 1999, 2000) (Table 1). Although it is possible to rationalise several factors in the ligand design, numerous indefinite factors are introduced upon immobilization on the solid support: the matrix and the coupling chemistry in addition to the ligand play an important role in binding to the target molecule (Lowe, 2001). However, performing the synthesis and screening on the solid-support helps to minimise these effects when considering the final application of the ligand as an affinity adsorbent.

Fig. 3. General strategy followed for the finding of a lead ligand following the design, synthesis and screening of a solid-phase combinatorial library of de novo designed ligands. Different strategies can be applied at the molecular modelling stage of de novo design of synthetic ligands (DESIGN), followed by the synthesis of a nxn combinatorial library of immobilized ligands (SYNTHESIS) further assessed for affinity and specificity for the target proteins (SCREENING). Screening procedures can involve, for example, FITC-based systems (a) or affinity chromatography (b). Potential lead ligands selected at this stage are synthesized in solution-phase, characterized by NMR and MS and further immobilized onto agarose beads to confirm affinity for target protein. A lead ligand is chosen and subsequent studies were performed to optimise its performance. A second generation of ligands might be designed, synthesized and tested.

8

Table 1 A summary of de novo designed biomimetic ligands based on a triazine-scaffold for the affinity purification of proteins

Kallikrein

IgG

IgG Human recombinant factor VIIa

Sugar moieties on glycoproteins

Recombinant insulin precursor MI3

Human a1-antitrypsin Prion protein

References

Solution-phase synthesis

(Burton, 1992)

Solid-phase combinatorial chemistry (12 ligands)

(Filippusson et al., 2000)

Solid-phase combinatorial chemistry (169 ligands)

(Roque et al., 2004b, 2005a, 2005b)

Solution-phase synthesis

(Li et al., 1998)

Solid-phase combinatorial chemistry (88 ligands) Solid-phase combinatorial chemistry; solution-phase synthesis of a sub-library

(Teng et al., 1999, 2000)

Solid-phase combinatorial chemistry (80 ligands)

(Palanisamy et al., 1999, 2000; Gupta and Lowe, 2004)

Solid-phase combinatorial chemistry (64 ligands); solution-phase synthesis of a sub-library Solid-phase combinatorial chemistry

(Sproule et al., 2000)

Solid-phase combinatorial chemistry (49 ligands)

(Renou et al., 2004)

Study of the target protein per se and selection of appropriate binding sites Study of the target protein per se and selection of appropriate binding sites

(Morrill et al., 2002, 2003)

(Lowe et al., 2001)

SpA — staphylococcal protein A; PpL — protein L from Peptostreptococcus magnus; IgG — immunoglobulin G; Gla — g-carboxyglutamic acid.

ARTICLE IN PRESS

IgG (Fab fragment)

Synthesis

Complex kallikrein/kininogen; ligands designed to mimic Phe–Arg dipeptide on the substrate Complex elastase/turkey ovomucoid inhibitor; ligands designed to mimic the natural inhibitor Complex PpL–IgG; ligands designed towards mimicking the natural PpL binding interfaces to IgG Complex SpA–IgG; ligands designed to mimic the Phe132–Tyr133 dipeptide on SpA Refinement of SpA mimicking ligand (Li et al., 1998) Complex tissue factor/factor VIIa; ligands designed to bind to the Gla-domain in Factor VIIa Protein–carbohydrate complexes: identification and mimicking of key residues determining monosaccharide specificity Study of the target protein per se and selection of appropriate binding sites

A.C.A. Roque, C.R. Lowe / Biotechnology Advances xx (2005) xxx–xxx

Elastase

Design strategy

ARTICLE IN PRESS A.C.A. Roque, C.R. Lowe / Biotechnology Advances xx (2005) xxx–xxx

9

4. Synthetic de novo affinity ligands in proteomics From a proteomics point of view, de novo designed synthetic ligands represent a good alternative to biological ligands and an opportunity to attain new classes of ligands for specific target proteins, as the power of designed ligands lies on the btailor-madeQ approach. As already presented in Table 1, the concepts of de novo designed synthetic ligands has already lead to the development of a series of triazine scaffolded molecules presenting affinity to different proteins or families of proteins. The utilisation of affinity chromatography as a depletion method has been recently reported to eliminate the immunoglobulin content from crude samples to be further submitted to 2D-PAGE or MS techniques (Lee and Lee, 2004). Similarly, affinity chromatography has also been utilised recently to separate, in a concentration mode, glycoproteins and glycosylated peptides resulting from proteomic analysis (Hashim et al., 2001; Hirabayashi et al., 2001). In the former, both general synthetic ligands (e.g. Cibacron blue F-3GA) and immunoadsorbents (e.g. bovine IgG-specific IgG immobilized on Sepharose) were utilised; in the latter, lectins were utilised as the ligand molecule. This section will focus particularly on the work accomplished in our group regarding the design, synthesis, evaluation and application of (i) immunoglobulin-binding ligands and (ii) glycoprotein-binding ligands (Fig. 4). Although displaying high selectivity, adsorbents based on proteins A/L and carbohydrate-binding systems suffer from high costs of production and purification, low binding capacities, limited life cycles and low scale-up potential, which is attributable to the biological nature of the ligands. The biomimetic ligands described below are fully

Fig. 4. De novo designed affinity ligands for mimicking proteins A and L. Crystallographic structure of the complex between protein A and Fc fragment of IgG (A) and of the complex between protein L and Fab fragments of IgG (D). Structure of the ligands artificial protein A (B) and L (E) and respective superimposition at the proteins A and L binding sites (C) and (F).

ARTICLE IN PRESS 10

A.C.A. Roque, C.R. Lowe / Biotechnology Advances xx (2005) xxx–xxx

synthetic in nature and can circumvent problems associated with biological ligands, while maintaining the affinity and specificity for the target proteins. 4.1. The artificial protein A Staphylococcus aureus protein A (SpA) was used as a template for the design of a synthetic ligand to bind specifically to immunoglobulin G (Li et al., 1998). The dipeptide motif Phe132–Tyr133 was crucial for the complex SpA–IgG (Deisenhofer, 1981) and was thus used as a starting point for the design of a series of biomimetic ligands. The most suitable ligand was able to bind IgG selectively from human plasma (98% purity after elution) and competitively to SpA, with affinity constants between 105–106 M 1 (Li et al., 1998). The lead compound was further refined by the synthesis of an IgG-binding combinatorial solid-phase ligand library comprising 88 adsorbents. The ligand library was assessed for its ability to bind pure human IgG and selected ligands further tested for their specificity to purify IgG from human plasma (Teng et al., 1999). Ligand 22/8 (ApA— artificial protein A) proved to be the most efficient for binding hIgG, (K a = 1.4  105 M 1) and for the separation of IgG from human plasma with recoveries of 67–69% and purities of 98–99% (Teng et al., 2000). This adsorbent also proved effective in the purification of antibodies from different species (chicken, cow, rabbit, pig, horse, rat, goat, sheep and mouse) and from different human classes (IgA and IgM) and IgG subclasses, including IgG3. Stability studies of ligand 22/8 in 1M NaOH over N140 h proved that it can withstand general procedures of cleaning-in-place and sterilisation, which represents an important advance over natural SpA (Teng et al., 2000). 4.2. The artificial protein L The lack of existence of a synthetic affinity ligand with a universal affinity for the scFv, Fab or F(abV)2 small antibody fragments prompted the design, synthesis and evaluation of a protein L mimic. Until then, affinity ligands binding specifically to the Fab fragments of immunoglobulins comprised anti-antibodies raised against the immunoglobulin of interest, molecules interacting with CDR regions (antigen or epitope/mimotope peptides) or microbial binding proteins (protein L). However, anti-antibodies represent an expensive and labour intensive option (Huse et al., 2002); ligands interacting with CDR regions are usually unique to each antibody and, therefore, do not represent a universal ligand (Murray et al., 1998). Protein L from Peptostreptococcus magnus was discovered in 1985 and is an immunoglobulin light chain binding protein particularly suitable for the purification of scFv, Fab and F(abV)2 biomolecules (Housden et al., 2003). Protein L binds with high affinity (K d of 1n M) to a large number of immunoglobulins with n1, n3 and n4 light chains (but not to n2 and E subgroups) and thus recognises 50% of human and more than 75% of murine immunoglobulins (Stura et al., 2002). A solid phase combinatorial triazine-scaffolded library of affinity ligands was designed de novo, synthesized and screened for the discovery of a protein L mimic. The thirteen compounds included in the combinatorial library were selected on the basis of the structure of the eleven different amino acid residues from the protein L domain recognised as being involved in the interaction with the n-light chains (Wikstrom et al., 1995; Graille et al.,

ARTICLE IN PRESS A.C.A. Roque, C.R. Lowe / Biotechnology Advances xx (2005) xxx–xxx

11

2001). The solid-phase library was synthesized by a bmix-and-splitQ procedure on Sepharose beads as a solid support. The members of the library were screened by an iterative procedure for their binding to human IgG and specificity towards the Fab moiety (Roque et al., 2005b). Using an FITC-based screening protocol, a total of 110 ligands were shown to bind N20% of loaded human IgG and these were subsequently assessed for selectivity for human Fab (Roque et al., 2004b). A total of 73 ligands displayed N20% binding to human Fab and, from these, 43 ligands showed b20% binding of loaded human Fc. Ligands 8/7, 5/9 and 12/3 were identified as putative lead ligands for mimicking the protein L interaction with the Fab moiety of immunoglobulins (Roque et al., 2005b). Ligand 8/7 was selected as the lead ligand and was further studied (Roque et al., 2005a). This ligand has shown to inhibit the interaction of PpL with IgG and Fab in competitive ELISA assays. Ligand 8/7 adsorbents bind 0.32 mg/g of human IgG and 0.28 mg/g of Fab with estimated K a of 104 M 1, while a commercial PpL adsorbent binds 0.47 and 0.36 mg/ g, respectively (mg/g of resin). Binding to Fc was negligible to both resins. Ligand 8/7adsorbent binds to immunoglobulins from different classes and sources that PpL also interacts with. Additionally, ligand 8/7 binds to IgG1 with n and E isotypes (92% and 100% of loaded protein) and polyclonal IgG from sheep, cow, goat and chicken. These properties were reflected in the isolation of immunoglobulins from crude samples, under non-optimised conditions, achieving purification factors of 7-fold and purities up to 95% (Roque et al., 2005a). The artificial protein A, ligand 22/8 (triazine scaffold substituted with 3-aminophenol and 4-amino-1-naphthol), possessed a more hydrophobic nature compared to ligand 8/7. This reflects the type of amino acid residues that are involved in the interaction of proteins A and L with immunoglobulins. In protein A, the Phe132–Tyr133 motif was shown to be relevant, whereas in protein L, a series of predominantly hydrophilic amino acid residues and a hydrogen-bond based interaction formed the basis of the complex with the light chains. 4.3. The artificial glycoprotein-binding ligands Group-specific immobilized lectins or biospecific antibodies traditionally carry out isolation of the glycosylated form of proteins and peptides. These adsorbents, however, present the usual disadvantages of biological ligands and can contaminate the final product by leaching. Lectins have proven to be toxic and damaging to mammalian cells in culture. Immobilized boronates are a family of synthetic ligands able to purify glycoproteins but require alkaline conditions to bind to sugars, which can be detrimental to the protein structure; other supramolecular structures have shown to have affinity for saccharides, but only in non-aqueous media (Gupta and Lowe, 2004). Rational design and solid-phase combinatorial chemistry were initially applied to the development of affinity adsorbents for glycoproteins and, from these preliminary studies, the triazine-based ligand 8/10 (histamine/tryptamine) presented affinity for the carbohydrate moiety of glucose oxidase and other mannosylated proteins (Palanisamy et al., 1999). In a follow-up study, analogues of histamine and tryptamine were investigated in order to find a ligand more suitable for facile solution-phase synthesis. Ligand 18/18, a 5aminoindan bis-substituted triazine molecule, has been shown to bind selectively to

ARTICLE IN PRESS 12

A.C.A. Roque, C.R. Lowe / Biotechnology Advances xx (2005) xxx–xxx

glucose oxidase with a K a in the order of 4.3  105 M 1 (Palanisamy et al., 2000). This biomimetic ligand displayed affinity for mannose, glucose and fucose moieties and the ability to bind to glycoproteins but was unable to bind selectively to the different saccharides (Palanisamy et al., 2000). Rational design and combinatorial chemistry were further employed on the discovery of ligand 11/11, a benzylamine bis-substituted triazine molecule, showing the following monosaccharide specificity mannoside N glucoside N galactoside, and affinity constants for binding to glycoproteins in the order of 104 M 1 (Gupta and Lowe, 2004). This biomimetic ligand also enhanced the understanding of saccharide selectivity, showing that the resolving power might be based on the formation of hydrogen bonds between the equatorial 3- and 4-hydroxyl groups from sugars and the planar, polar nitrogen groups of the triazine ligand. Ligand 11/11, like other triazine-based biomimetic ligands, has shown to be chemically stable and easy to synthesize, to possess non-fissile bonds and to withstand sterilization and cleaning in situ (Gupta and Lowe, 2004).

5. Concluding remarks Although it might be difficult to apply affinity chromatography as a stand-alone one-dimensional assay in proteomics, there is no doubt that the technique is very useful not only for the separation of specific groups of proteins but also for the exploration of post-translational modifications and the study of protein–protein interactions. The application of affinity chromatography for the enrichment or depletion of proteins, where group-specific ligands can be applied, is a powerful way of increasing resolution and sensitivity in subsequent 2D-PAGE or MS analysis (Lee and Lee, 2004). Synthetic biomimetic ligands have been shown to circumvent many of the disadvantages associated with natural biological ligands for the downstream processing of therapeutic proteins. It is also time to consider these molecules as an alternative to group-specific ligands utilised in proteomics analysis. In particular, the recently developed artificial proteins A and L and artificial lectin ligands represent a promising group of de novo designed affinity ligands to complement current proteomic analysis techniques.

References Burton NP. Design of novel affinity adsorbents for the purification of trypsin-like proteases. J Mol Recognit 1992; 5:55 – 68. Deisenhofer J. Crystallographic refinement and atomic levels of a human Fc fragment and its complex with fragment B of protein A from Staphylococcus aureus at 29- and 28- angstroms resolution. Biochemistry 1981; 20:2361. Fassina G, Verdoliva A, Odierna MR, Ruvo M, Cassini G. Protein a mimetic peptide ligand for affinity purification of antibodies. J Mol Recognit 1996;9:564 – 9. Filippusson H, Erlendsson LS, Lowe CR. Design, synthesis and evaluation of biomimetic affinity ligands for elastases. J Mol Recognit 2000;13:370 – 81. Fountoulakis M, Vlahou A. Proteomic approaches in the search for disease biomarkers. J Chromatogr A 2005; 814:11 – 9.

ARTICLE IN PRESS A.C.A. Roque, C.R. Lowe / Biotechnology Advances xx (2005) xxx–xxx

13

Graille M, Stura EA, Housden NG, Beckingham JA, Bottomley SP, Beale D, et al. Complex between Peptostreptococcus magnus protein L and a human antibody reveals structural convergence in the interaction modes of Fab binding proteins. Structure 2001;9:679 – 87. Gupta G, Lowe CR. An artificial receptor for glycoproteins. J Mol Recognit 2004;17:218 – 35. Hashim OH, Ahmad F, Shuib AS. The application of Artocarpus integer seed lectin-M in the detection and isolation of selective human serum acute-phase proteins and immunoglobulins. Immunol Invest 2001;30: 131 – 41. Hirabayashi J, Arata Y, Kasai K-I. Glycome project: concept, strategy and preliminary application to Caenorhabditis elegans. Proteomics 2001;1:295 – 303. Housden NG, Harrison S, Roberts SE, Beckingham JA, Graille M, Stura EA, et al. Immunoglobulin-binding domains: protein L from Peptostreptococcus magnus. Biochem Soc Trans 2003;31:716 – 8. Huse K, Bohme HJ, Scholz GH. Purification of antibodies by affinity chromatography. J Biochem Biophys Methods 2002;51:217 – 31. Jacobs JM, Monroe ME, Qian W-J, Shen Y, Anderson GA, Smith RD. Ultra-sensitive, high throughput and quantitative proteomics measurements. Int J Mass Spectrom 2005;240:195 – 212. Janson J-C. Large-scale affinity purification—state of the art and future prospects. Trends Biotechnol 1984;2: 31 – 8. Jensen ON. Modification-specific proteomics: characterization of post-translational modifications by mass spectrometry. Curr Opin Chem Biol 2004;8:33 – 41. Kalume DE, Molina H, Pandey A. Tackling the phosphoproteome: tools and strategies. Curr Opin Chem Biol 2003;7:64 – 9. Lee W-C, Lee KH. Applications of affinity chromatography in proteomics. Anal Biochem 2004;324:1 – 10. Li RX, Dowd V, Stewart DJ, Burton SJ, Lowe CR. Design, synthesis, and application of a protein A mimetic. Nat Biotechnol 1998;16:190 – 5. Loo RRO, Hayes R, Yang Y, Hung F, Ramachandran P, Kim N, et al. Top-down, bottom-up, and side-to-side proteomics with virtual 2-D gels. Int J Mass Spectrom 2005;240:317 – 25. Lowe CR. Combinatorial approaches to affinity chromatography. Curr Opin Chem Biol 2001;5:248 – 56. Lowe CR, Burton SJ, Burton NP, Alderton WK, Pitts JM, Thomas JA. Designer dyes—biomimetic ligands for the purification of pharmaceutical proteins by affinity-chromatography. Trends Biotechnol 1992;10:442 – 8. Lowe CR, Lowe AR, Gupta G. New developments in affinity chromatography with potential application in the production of biopharmaceuticals. J Biochem Biophys Methods 2001;49:561 – 74. Morrill PR, Gupta G, Sproule K, Winzor D, Christensen J, Mollerup I, et al. Rational combinatorial chemistrybased selection, synthesis and evaluation of an affinity adsorbent for recombinant human clotting factor VII. J Chromatogr B Biomed Sci Appl 2002;774:1 – 15. Morrill PR, Millington RB, Lowe CR. Imaging surface plasmon resonance system for screening affinity ligands. J Chromatogr B Biomed Sci Appl 2003;793:229 – 51. Murray A, Spencer DIR, Missailidis S, Denton G, Price MR. Design of ligands for the purification of anti-MUC1 antibodies by peptide epitope affinity chromatography. J Pept Res 1998;52:375 – 83. Newton RP, Brenton AG, Smith CJ, Dudley E. Plant proteome analysis by mass spectrometry: principles, problems, pitfalls and recent developments. Phytochemistry 2004;65:1449 – 85. Palanisamy UD, Hussain A, Iqbal S, Sproule K, Lowe CR. Design, synthesis and characterisation of affinity ligands for glycoproteins. J Mol Recognit 1999;12:57 – 66. Palanisamy UD, Winzor DJ, Lowe CR. Synthesis and evaluation of affinity adsorbents for glycoproteins: an artificial lectin. J Chromatogr B Biomed Sci Appl 2000;746:265 – 81. Patton WF. Detection technologies in proteome analysis. J Chromatogr B Biomed Sci Appl 2002;771:3 – 31. Renou ENS, Gupta G, Young DS, Dear DV, Lowe CR. The design, synthesis and evaluation of affinity ligands for prion proteins. J Mol Recognit 2004;17:248 – 61. Roque ACA, Lowe CR. Lessons from nature: on the molecular recognition elements of the phosphoprotein binding-domains. Biotechnol Bioeng in press. Roque ACA, Lowe CR, Taipa MA. Antibodies and genetically engineered related molecules: production and purification. Biotechnol Prog 2004a;20:639 – 54. Roque ACA, Taipa MA, Lowe CR. A new method for the screening of solid-phase combinatorial libraries for affinity chromatography. J Mol Recognit 2004b;17:262 – 7.

ARTICLE IN PRESS 14

A.C.A. Roque, C.R. Lowe / Biotechnology Advances xx (2005) xxx–xxx

Roque ACA, Taipa MA, Lowe CR. An artificial protein L for the purification of immunoglobulins and Fab fragments by affinity chromatography. J Chromatogr A 2005a;1064:157 – 67. Roque ACA, Taipa MA, Lowe CR. Synthesis and screening of a rationally designed combinatorial library of affinity ligands mimicking protein L from Peptostreptococcus magnus. J Mol Recognit 2005b;18:213–24. Scheich C, Sievert V, Bussow K. An automated method for high-throughput protein purification applied to a comparison of His-tag and GST-tag affinity chromatography. BMC Biotechnol 2003;3:12. Shi Y, Xiang R, Horvath C, Wilkins JA. The role of liquid chromatography in proteomics. J Chromatogr A 2004;1053:27 – 36. Sproule K, Morrill P, Pearson JC, Burton SJ, Hejnaes KR, Valore H, et al. New strategy for the design of ligands for the purification of pharmaceutical proteins by affinity chromatography. J Chromatogr B Biomed Sci Appl 2000;740:17 – 33. Stura EA, Graille M, Housden NG, Gore MG. Protein L mutants for the crystallization of antibody fragments. Acta Crystallogr D Biol Crystallogr 2002;58:1744 – 8. Teng SF, Sproule K, Hussain A, Lowe CR. A strategy for the generation of biomimetic ligands for affinity chromatography. Combinatorial synthesis and biological evaluation of an IgG binding ligand. J Mol Recognit 1999;12:67 – 75. Teng SF, Sproule K, Husain A, Lowe CR. Affinity chromatography on immobilized bbiomimeticQ ligands synthesis, immobilization and chromatographic assessment of an immunoglobulin G-binding ligand. J Chromatogr B Biomed Sci Appl 2000;740:1 – 15. Wikstrom M, Sjobring U, Drakenberg T, Forsen S, Bjorck L. Mapping of the immunoglobulin light chain— binding site of protein-L. J Mol Biol 1995;250:128 – 33.