MINIREVIEW Alternative binding proteins: Affibody binding proteins developed from a small three-helix bundle scaffold Per-A ˚ ke Nygren Department of Molecular Biotechnology, School of Biotechnology, Royal Institute of Technology (KTH), Stockholm, Sweden In several areas of life science, including biosepara- tion, diagnostics, imaging and therapy, the availability of reagents capable of the selective recognition of a predefined analyte or target structure is critical. Until the mid - 1990’s, such reagents were traditionally immunoglobulin-based, consisting of either polyclonal sera or monoclonal antibodies generated towards the target analyte through immunization of laboratory animals. However, recent advances in methodology for the generation and handling of large and complex libraries of various biomolecules [1], have led to a development of new means for obtaining affinity reagents of different types. Antibodies, as well as non-immunoglobulin-based affinity reagents of differ- ent classes (both protein and nucleic acid-based), can now routinely be identified via functional selection or screening from libraries of different types for their ability to bind to a desired target structure under defined conditions [2,3]. These novel reagents, typi- cally built from frameworks that are smaller and structurally less complex than immunoglobulins, are expected to offer advantages related to both produc- tion and application issues [4]. The topic of this review is the so-called affibody binding proteins, an example of the new emerging classes of protein-based affinity reagents based on frameworks other than the Keywords affibody binding proteins; affinity chromatography; combinatorial protein engineering; in vivo imaging; peptide synthesis; phage display; protein chips; protein–protein interactions; selection; viral retargeting Correspondence P A ˚ . Nygren, Department of Molecular Biotechnology, School of Biotechnology, Roslagstullsbacken 21, AlbaNova University Center, Royal Institute of Technology (KTH), SE-106 91 Stockholm, Sweden Fax: +46 85537 8481 Tel: +46 85537 8328 E-mail: perake@biotech.kth.se (Received 16 November 2007, revised 11 February 2008, accepted 2 April 2008) doi:10.1111/j.1742-4658.2008.06438.x In recent years, classical antibody-based affinity reagents have been challenged by novel types of binding proteins developed by combinatorial protein engineering principles. One of these classes of binding proteins of non-Ig origin are the so-called affibody binding proteins, functionally selected from libraries of a small (6 kDa), non-cysteine three-helix bundle domain used as a scaffold. During the first 10 years since they were first described, high-affinity affibody binding proteins have been selected towards a large number of targets for use in a variety of applications, such as bioseparation, diagnostics, functional inhibition, viral targeting and in vivo tumor imaging ⁄ therapy. The small size offers the possibility to produce functional affibody binding proteins also by chemical synthesis production routes, which has been found to be advantageous for the site-specific introduction of various labels and radionuclide chelators. Abbreviations ABD, albumin-binding domain; Ad5, adenovirus type 5; Ab, amyloid-beta; DOTA, tetraazacyclododecane tetraacetic acid; DTPA, diethylenetriane pentaacetic acid; FRET, fluorescence energy resonance transfer; gp120, glycoprotein 120; SPA, staphylococcal protein A; SPPS, solid-phase peptide synthesis; T m , melting temperature. 2668 FEBS Journal 275 (2008) 2668–2676 ª 2008 The Author Journal compilation ª 2008 FEBS immunoglobulin fold. This class of small (6 kDa) affinity proteins was first described 10 years ago [5] and since then affibody molecules have been devel- oped towards many targets and have been used in several different affinity technology applications. The history of affibody proteins starts with staphylo- coccal protein A (SPA), which is a familiar reagent to most researchers working with immunotechnology or related topics. Various native and recombinant vari- ants of this natively cell wall-anchored immunoglobu- lin-binding [Fc binding (IgG) and Fab binding (VHIII)] bacterial receptin have, for several decades, been widely used for the detection and purification of antibodies from different sources. In addition, various portions of the SPA gene, containing a region encod- ing five highly homologous Ig-binding domains, have also been frequently used as affinity gene fusion part- ners for the production and selective purification (IgG-affinity chromatography) of recombinant fusion proteins [6]. The framework (or scaffold) upon which affibody binding proteins are based is the so-called Z domain, an engineered variant of a consensus SPA domain (domain B). During the engineering process, a few changes were introduced to facilitate directional head-to-tail polymerization at the gene fragment level and to increase the chemical stability of the protein towards hydroxylamine (via a Gly to Ala substitution simultaneously resulting in a loss of the Fab-binding capacity) [7,8]. This three-helix bundle Z protein is small (58 amino acids), composed of a single polypep- tide subunit capable of rapid and independent folding, devoid of cysteine residues, has a native capability of high-affinity interaction with binding partner proteins via a defined set of surface-located residues, and can be easily expressed in soluble and proteolytically stable forms in various host cells on its own or in fusion with other protein partners [6]. Inspiration from advances in protein library technology, and coinciding frustra- tion caused by the unsuccessful prokaryotic production of several cloned antibody-derived constructs in the home laboratory, led to the decision to explore if the advantageous features of the Z domain could be employed for the generation of a novel class of affinity reagents. The driving hypothesis was that the native binding specificity for Fc of IgG, via amino acid sub- stitutions at the involved face of the molecule, could be switched to any other protein target of interest while simultaneously retaining most, if not all, the ben- eficial features of the ancestral Z-protein scaffold. Examples from the selection, characterization and use of target binding affibody proteins in different appli- cations show that these assumptions have now been verified in practice. Library design and selections Using combinatorial protein engineering principles, so-called naive (or unbiased) libraries of candidate affibody binding proteins have been constructed through the genetic randomization of 13 surface- located positions of the Z-protein-domain scaffold. The choice of positions to include was based on anal- ysis of an available X-ray crystallography structure of the co-complex between the homologous B-domain of SPA and human IgG [9]. The positions chosen were all located to helices 1 (seven positions) and 2 (six positions) of the three-helix bundle structure and include most of the positions involved in the interac- tion with the Fc region of human IgGl (hIgG1), but also a few additional surface-located positions at the same face of the Z domain [10] (Fig. 1A–C). Using NN(G ⁄ T) degenerate codons at those position (as in the libraries denoted Zlib-1 [10] and Zlib2002 [11], including 32 codons in total and all 20 amino acids), the theoretical diversity obtained is 32 13 (3.7 · 10 19 ) genetic variants encoding 20 13 (8.2 · 10 16 ) different protein variants. Owing to limitations related to the efficiency of cell transformation during the construc- tion of libraries, only a fraction (typically 10 7 –10 9 clones) of the theoretical complexity has hitherto been searched in any selection experiment using a naive library of affibody proteins. However, in efforts to increase the affinities obtained for ‘first generation’ affibody binding proteins selected from naive libraries, secondary libraries involving a more in-depth search of a local volume of the entire structure space have been constructed and used for selections (‘affinity maturation’). Here, either all previously randomized positions in an entire helix segment (‘helix shuffling’), or a number of positions distributed between both helices of a ‘first generation’ affibody molecule, have been genetically rerandomized followed by new rounds of stringent selections [12–14]. Hitherto, all reported de novo selections of affibody binding proteins have been performed using phage display technology based on standard phagemid vectors and selection protocols. Alternative selection systems are also being investigated, including b-lactamase-based protein fragment complementation assay [15], staphylococcal display [16,17], microbead display [18] and ribosomal display (S. Grimm and P A. Nygren, unpublished results). Affinities obtained for different targets are similar to those obtained for naive libraries of antibody fragments or alternative scaffolds, with dissociation constants (K D ) ranging from micromolar to low picomolar, depend- ing on library complexity and target structure P A ˚ . Nygren Affibody binding proteins FEBS Journal 275 (2008) 2668–2676 ª 2008 The Author Journal compilation ª 2008 FEBS 2669 characteristics. The highest affinity reported to date for an affibody molecule is 22 pm for an affinity- matured variant directed to the breast cancer marker ErbB2 (Her2) [14]. Structural and biophysical analyses of affibody proteins An interesting question is if binding sites and fine-detail epitopes (generally preferred by affibody binding pro- teins developed via in vitro selection) differ compared with epitopes seen by other binding protein classes, including immunoglobulins generated via immuniza- tions or library technology. The interaction surface made up by the two randomized alpha-helices could be regarded to differ significantly from the corresponding surface in immunoglobulins, which is primarily made up from contributions from six variable peptide loops. However, in some instances efforts to roughly map the epitopes via binding competition studies using mono- clonal antibodies have shown the existence of at least overlapping epitopes for these two different classes of affinity reagents, also generated via different routes [13,19–21]. Data from the structures of two different af- fibody protein::target protein complexes (Z SPA-1 ::Z WT , anti-Z Taq ::Z Taq ), from both NMR and X-ray crystral- lography experiments, has provided detailed informa- tion on the mode of binding and the interface involved in the interaction [22–24]. Analyses of these two com- plexes, both involving so-called ‘anti-idiotypic affibody proteins’ (i.e. an affibody binding protein developed towards the ancestral SPA protein or a second affibody binding protein, used as targets during selection) [25,26] revealed, for example, that these affibody bind- ing proteins bind their respective targets in a perpendic- ular orientation relative each other (Fig. 1D). Furthermore, the interaction interfaces, involving up to 11 of the 13 randomized positions, were found to be predominantly of non-polar nature and to comprise some 800–900 A ˚ 2 per subunit, typical of protein–pro- tein interactions [27]. In addition, substantial structural re-arrangements of side chains have been observed to occur upon binding (induced-fit mechanism), as seen from comparisons with structures of the free proteins. The interactions appear to be predominantly enthalphy driven, with favourable desolvation contributions AB CD Fig. 1. Affibody binding proteins: library design and target binding. Illustration of the three-helix bundle affibody protein scaffold Z in band representation (green), with the 13 positions randomized during affibody protein library constructions highlighted (yellow) (A). Side views (B) and top views (C), respec- tively, of the affibody protein scaffold show- ing, in sphere representation (yellow), the 13 surface-located positions in helices 1 and 2 employed for library constructions. (D) The structure of the complex between an affi- body protein selected for binding to Taq DNA polymerase (Z Taq ; white) and its anti- idiotypic affibody protein (anti-Z Taq ; green with randomized positions in yellow). RSCB Protein Data Bank (PDB) entries 1Q2N (Z domain) and 2B87 (anti-idiotypic complex) were used to generate the images in PYMOL software. See the text for details. Affibody binding proteins P A ˚ . Nygren 2670 FEBS Journal 275 (2008) 2668–2676 ª 2008 The Author Journal compilation ª 2008 FEBS counter-balanced by conformational entropy changes. Recently, NMR technology was used to determine the solution structure of a peculiar affibody protein::target complex involving the amyloid-b (Ab) peptide and two copies of a phage-selected anti-Ab affibody molecule [28]. The structure revealed that the beta-hairpin-struc- tured Ab peptide in the complex was embraced and sol- ubilized by the two affibody proteins, which, in turn, were covalently linked together via a disulfide bridge. In addition, a stretch of amino acids within the N-ter- minal thirds of both copies of the affibody variant, nor- mally of alpha-helical secondary structure (helix 1), were in the complex found to adopt beta-strand confor- mations involved in hydrogen bonds with the Ab pep- tide (Fig. 2). However, the selection targets used for obtaining these complexes may not be fully representa- tive of a large target protein, potentially containing sev- eral candidate epitopes, and general conclusions regarding epitope preferences will have to await future determinations of affibody protein::target co-complex structures. Biophysical studies have shown that different affi- body proteins are, as expected, differently affected by the introduced substitutions. Whereas the two affibody proteins anti-Z Taq and Z Taq (T m values between 55 and 57 °C) both show Z WT -like structures (T m  78 °C) as free proteins in solution [24,29] the low-affinity SPA-binding Z SPA-1 affibody protein associated with a relatively low thermal stability (T m  40 °C) shows ‘molten globule’-like characteristics and only upon interaction with the Z WT binding partner is capable of adapting a compact three-helix bundle configuration [22]. Interestingly, a conformationally stabilized variant of this Z SPA-1 affibody protein, obtained by the intro- duction of an intramolecular S–S bridge involving positions outside the binding region, showed a 10-fold higher affinity for the Z WT target [30]. Thermodynamic studies of the interaction by isothermal titration calo- rimetry showed that the enhanced interaction strength seen for the mutant compared with the wild-type Z SPA-1 was the net result of enthalpy rather than entropy effects (involving a net zero balance between changes in desolvation entropy and conformational entropy terms for the mutant compared with the wild-type [30]). Examples of selected affibody binding proteins and their applications As mentioned earlier in this article, the availability of reagents capable of the selective recognition of biomol- ecules is crucial in many areas of life science. Since the first description of affibody proteins, the suitability of these ligands in many different and diverse application areas has been investigated, involving different presen- tation formats, labels and routes of production. Because of to the small size of affibody binding pro- teins it has been shown to be possible, in addition to conventional recombinant production, also to use solid-phase peptide synthesis (SPPS) for the production of full-length affibody proteins. This route of produc- tion has facilitated the site-specific introduction of various labels for detection or immobilization purposes as well the incorporation of different chelators for the later attachment of radionuclides for in vivo imaging purposes (see below). Not surprisingly, considering the close relationship to SPA, affinity chromatography was one of the first applications to be investigated. Two of the first affi- body proteins to be selected were binders with affini- ties in the micromolar range to Taq DNA polymerase and human apolipoprotein A-1 [5,31]. After head- to-tail dimerization, both affibody proteins were successfully used as resin-coupled ligands in affinity chromatography. Interestingly, a high stability towards alkaline column sanitation schemes (including repeated pulses with 0.5 m NaOH) were demonstrated for these molecule #1 molecule #2 Intermolecular S-S bridge Aβ (1-40) peptide Fig. 2. Solution structure of Alzheimer¢s amyloid-b (Ab) peptide bound by a phage-selected anti-Ab peptide affibody homodimer. The picture shows the structure of a peculiar affibody protein::tar- get protein complex determined by NMR, involving one molecule of the Alzheimer¢sAb peptide (green) and two molecules of a phage-selected anti-Ab peptide affibody variant (blue and yellow), covalently linked via a disulfide bridge. The PDB entry 2OTK [28] was used to generate the image in PYMOL software. See the text for details. P A ˚ . Nygren Affibody binding proteins FEBS Journal 275 (2008) 2668–2676 ª 2008 The Author Journal compilation ª 2008 FEBS 2671 affinity ligands; an important property when it comes to column cleaning and sanitation [31]. The usefulness of various types of bioseparation applications, ranging from depletion of proteomics samples to industrial separation, and involving different ‘feed stocks’ (fer- mentation medium, bacterial lysates, human plasma or serum), has also been shown for other affibody pro- teins. This includes variants directed to recombinant human factor VIII [13], human IgA [32], human Ab peptides [11] and human transferrin [33]. Via gene-fusion principles, affibody proteins have also been utilized as affinity moieties in the context of larger fusion proteins. For example, the constant regions of human IgG have been recruited for con- struction of antibody-like fusion constructs, where the Fab arms were replaced with affibody proteins [34]. Fusion of affibody molecules to the Escherichia coli b-galactosidase enzyme has enabled the intracellular high-level production of soluble recombinant enzyme conjugates that are useful for diagnostics [35]. Further- more, genetic fusion of IgA- or IgE-specific affibody proteins to the surface-anchoring regions of SPA was shown to result in whole-cell-based reagents for analyte detection when constructs were expressed in Staphylococcus carnosus [36]. Various formats of affibody proteins have also been used as capture ligands on microarrays, using different principles for their attachment to the surface. Here, either monomeric affibody proteins, produced by SPPS and containing site-specifically introduced biotin or cysteine groups for directed immobilization via differ- ent coupling chemistries [37], or recombinant head-to- tail multimeric versions of affibody proteins, have been evaluated [38]. For example, six different divalent con- structs were spotted side-by-side and shown to permit detection of the fluorescently labeled analytes down to low picomolar levels [38]. In one study, the non-mam- malian origin of affibody proteins was shown to be an advantage for diagnostic applications involving sand- wich assays. Exchanging one of the reagents in a clas- sical two-antibody sandwich assay for an affibody protein of prokaryotic origin was shown to prevent problems with false-positive background signals result- ing from the presence in serum of so-called heterophilic antibodies, equipped with binding specificities capable of bridging capture and detection antibody reagents in a non-analyte-dependent manner [39]. For use as ‘one-stop’ reagents in analyses of receptor levels on cell surfaces by flow cytometry, immunofluorescence and immunohistochemistry, efficient maleimide-directed labelling of affibody proteins by fluorophore groups (e.g. Oregon Green 488) or horseradish peroxidase has been reported [40]. For the detection of unlabeled analytes, two differ- ent principles based on fluorescence energy resonance transfer (FRET) and affibody binding proteins have been developed. In an intramolecular format, based on the site-specific dual incorporation of one donor and one acceptor fluorophore into the same affibody binding protein at sites flanking the target binding area, the basis for signalling is a change in the inter- action between the two fluorophores upon target ana- lyte binding [41,42]. An alternative format, based on FRET between donor and acceptor fluorophores pres- ent on two different affibody proteins that are parts of an anti-idiotypic affibody protein pair whose inter- action (and therefore also the intermolecular FRET efficiency) is influenced by the presence of different concentrations of an analyte, has also been described [43]. The capability of affibody proteins to interfere, via binding, with the functions of other proteins have also been investigated. In one study a human CD28- specific affibody fusion protein was shown to be capable of blocking the CD28–CD80 costimulatory signalling between two cell lines [44]. Furthermore, additional examples of functional blocking appli- cations are the use of a DNA polymerase-binding affibody protein, capable of inactivating the enzyme at low temperatures to obtain so-called hot-start PCR conditions [45], and the expression of a receptor- specific affibody protein equipped with an endo- plasmic reticulum retention peptide, showing promise as a means for interfering with the export of the receptor to the cell surface (E. Vernet, A. Konrad, E. Lundberg, P A. Nygren and T. Gra ¨ slund, unpub- lished results). Functional insertion of affibody-based affinity moie- ties in foreign proteins have also been described for work with adenovirus type 5 (Ad5) fibers. In one study, it was shown to be possible to incorporate an HIV glycoprotein 120 (gp120)-specific affibody protein into a truncated version of the Ad5 fiber, which were then found to acquire the capability of selectively recognizing gp120 proteins [20]. Because the use of affibody protein-modified Ad5 vectors in humans would be hampered if high titers of neutralizing anti- (affibody scaffold) Ig were present, a panel of normal human sera was, in one study, used to investigate this issue [46]. Encouragingly, the results showed that only two out of 50 samples of serum tested showed any significant infection-inhibiting ability, indicating that pre-existing anti-(affibody scaffold) Ig are relatively rare. A head-to-tail dimeric version of an anti-Her2 affibody protein has been inserted into the H1 loop of the knob structure in Ad5 fibers [47]. Virus particles Affibody binding proteins P A ˚ . Nygren 2672 FEBS Journal 275 (2008) 2668–2676 ª 2008 The Author Journal compilation ª 2008 FEBS containing such fibers were demonstrated to infect cells via Her2 receptors rather than via the normal Coxsackie B virus and Ad receptor route. These results, as well as the results from a similar study carried out using Her2-specific affibody binders [48], thus represent important steps towards the possibility of developing retargeted Ad viruses for directed gene therapy applications. Many investigations have been performed to evaluate the potential of affibody proteins for use as target specific probes in the in vivo imaging of tumors via single photon emission computed tomog- raphy (SPECT) or positron emission tomography (PET) [49,50]. Some hallmarks of an ideal in vivo imaging probe include high tumor retention and tissue penetration, rapid blood clearance kinetics, low non- tumor organ uptake and rapid and site-specific label- ling with different radionuclides depending on the preferred modality. In this respect, the 6 kDa size of a monovalent affibody protein makes it smaller than the smallest antibody-derived fragments (V domains are 15 kDa [51]), and thus provides both the potential for rapid clearance through the kidneys as well as a route for chemical synthesis of the reagent facilitating the incorporation of labels. In addition, the modular nat- ure of the affibody scaffold facilitates the construction of multimeric constructs and ⁄ or fusions to other domains for tuning the biodistribution properties if desired (see below). The in vivo imaging studies using affibody proteins have so far primarily been focused on the Her2 (ErbB2) target, a receptor belonging to the tyrosine kinase class, which is overexpressed in breast and urinary bladder carcinomas and is a well-known target for immunother- apy via the monoclonal antibody trastuzumab (Hercep- tin) [52]. Evaluated monovalent and divalent affibody protein constructs have been based on first-generation (Z HER2:4 ) or second-generation (Z HER2:342 ) Her2-specific affibody proteins with monovalent affinities (K D )of50 and 22 pm, respectively [14,53]. In preclinical studies, involving, for example, cell-binding studies in vitro and in vivo experiments in laboratory animals containing tumor xenografts, both recombinant affibody constructs and synthetic variants have been explored. Different radionuclides (e.g. 76 Br, 125 I, 111 In, 114m I, 99m Tc and 211 At) have been attached via different principles, involving various linkers or chelators. For example, CHX-A¢¢- or benzyl-diethylenetriamine pentaacetic acid (DTPA), tetraazacyclododecane tetraacetic acid (DOTA), mercaptoacetyl-glycyl-glycyl-glycyl, or merca- ptoacetyl-seryl-seryl-seryl entities have been introduced into affibody proteins via non-directed or directed cova- lent coupling or through SPPS, depending on the affi- body construct and the radionuclide used [54–59]. Data obtained for an 111 In-benzyl-DTPA-Z HER2:342 con- struct, produced via coupling of the chelator to a recom- binant affibody protein moiety and injected in mice bearing Her2-expressing SKOV-3 xenografts, showed a rapid clearance of non-tumor-bound label from the blood (but not from the kidney). This resulted in high tumor to blood ratios ( 100, 4 h postinjection) and high contrast images where the tumor could be clearly visualized [59]. In a later study, a synthetically produced DOTA-Z HER2:342 construct, subsequently labeled with 111 In, also showed high tumor to blood ratios. The DOTA chelator used here could also be labeled with other radionuclides, including the positron emitter 68 Ga for PET and the b-emitter 177 Lu, of interest for solid tumor cancer treatment. In a first-time in-human clinical study, microdoses (< 100 lg) of both 68 Ga-labelled and 111 In-labeled DOTA-Z HER2:342 material was injected into patients with recurrent breast cancer. Using SPECT, small Her2-positive metastases were reported to be clearly detectable by in vivo imaging only 1 h postinjection, showing the potential for use of this class of reagents in hu mans [60]. A radiolabeled 177 Lu-CHX-A’- DTPA-ABD-(Z HER2:342 ) 2 construct, containing a serum albumin-binding domain (ABD) fused to a divalent version of the Z HER2:342 affibody protein has been evalu- ated for possible use in therapy [61]. The ABD moiety was included for its capability, demonstrated in a previ- ous study, to increase the residence time in serum of proteins (produced as gene fusions to ABD) via its bind- ing to serum albumin, a protein with very slow clearance rates and which is present at high concentrations in the blood [62]. In mice with SKOV-3 microxenografts (high HER2 expression), tumor formation was totally prevented after receiving the compound, in contrast to animals receiving a control affibody protein construct. The inclusion of the ABD moiety also dramatically reduced the renal uptake as compared with an unfused control construct [61]. Recently, the selection and initial in vitro evaluations on the selectivity and cellular uptake of a first-generation affibody binding protein generated towards the EGF receptor (Her1 or ErbB1) have also been reported [21,63,64]. Taken together, the results obtained from the extensive use of affibody binding proteins in a variety of applications during the first 10 years from when they were first described suggests that they will continue to be valuable alternatives to antibodies in several biorecognition-based applications in the future. Acknowledgements Dr Amelie Eriksson Karlstro ¨ m is gratefully acknowl- edged for critical reading of the manuscript. P A ˚ . Nygren Affibody binding proteins FEBS Journal 275 (2008) 2668–2676 ª 2008 The Author Journal compilation ª 2008 FEBS 2673 References 1 Lin H & Cornish VW (2002) Screening and selection methods for large-scale analysis of protein function. Angew Chem Int Ed Engl 41, 4402–4425. 2 Nygren PA & Skerra A (2004) Binding proteins from alternative scaffolds. J Immunol Methods 290, 3–28. 3 Binz HK, Amstutz P & Pluckthun A (2005) Engineering novel binding proteins from nonimmunoglobulin domains. Nat Biotechnol 23, 1257–1268. 4 Binz HK, Amstutz P & Pluckthun A (2005) Engineering novel binding proteins from nonimmunoglobulin domains. Nat Biotechnol 23, 1257–1268. 5 Nord K, Gunneriusson E, Ringdahl J, Stahl S, Uhlen M & Nygren PA (1997) Binding proteins selected from combinatorial libraries of an alpha-helical bacterial receptor domain. Nat Biotechnol 15, 772–777. 6 Stahl S & Nygren PA (1997) The use of gene fusions to protein A and protein G in immunology and biotech- nology. Pathol Biol (Paris) 45 , 66–76. 7 Nilsson B, Moks T, Jansson B, Abrahmse ´ n L, Elmblad A, Holmgren E, Henrichson C, Jones TA & Uhle ´ nM (1987) A synthetic IgG-binding domain based on staph- ylococcal protein A. Protein Eng 1, 107–113. 8 Jansson B, Uhlen M & Nygren PA (1998) All individual domains of staphylococcal protein A show Fab binding. FEMS Immunol Med Microbiol 20, 69–78. 9 Deisenhofer J (1981) Crystallographic refinement and atomic models of a human Fc fragment and its complex with fragment B of protein A from Staphylococcus aureus at 2.9- and 2.8-A resolution. Biochemistry 20, 2361–2370. 10 Nord K, Nilsson J, Nilsson B, Uhlen M & Nygren PA (1995) A combinatorial library of an alpha-helical bac- terial receptor domain. Protein Eng 8, 601–608. 11 Gronwall C, Jonsson A, Lindstrom S, Gunneriusson E, Stahl S & Herne N (2007) Selection and characteriza- tion of Affibody ligands binding to Alzheimer amyloid beta peptides. J Biotechnol 128, 162–183 [Epub 27 Sep 2006]. 12 Gunneriusson E, Nord K, Uhlen M & Nygren P (1999) Affinity maturation of a Taq DNA polymerase specific affibody by helix shuffling. Protein Eng 12, 873–878. 13 Nord K, Nord O, Uhlen M, Kelley B, Ljungqvist C & Nygren PA (2001) Recombinant human factor VIII-spe- cific affinity ligands selected from phage-displayed com- binatorial libraries of protein A. Eur J Biochem 268 , 4269–4277. 14 Orlova A, Magnusson M, Eriksson TL, Nilsson M, Larsson B, Ho ¨ ide ´ n-Guthenberg I, Widstro ¨ m C, Carls- son J, Tolmachev V, Sta ˚ hl S & Nilsson FY (2006) Tumor imaging using a picomolar affinity HER2 bind- ing affibody molecule. Cancer Res 66, 4339–4348. 15 Nord O, Gustrin A & Nygren PA (2005) Fluorescent detection of beta-lactamase activity in living Escherichia coli cells via esterase supplementation. FEMS Microbiol Lett 242, 73–79. 16 Lofblom J, Wernerus H & Stahl S (2005) Fine affinity discrimination by normalized fluorescence activated cell sorting in staphylococcal surface display. FEMS Micro- biol Lett 248 , 189–198. 17 Kronqvist N, Lofblom J, Jonsson A, Wernerus H & Stahl S (2008) A novel affinity protein selection system based on staphylococcal cell surface display and flow cytometry. Protein Eng Des Sel 21, 247–255. 18 Nord O, Uhlen M & Nygren PA (2003) Microbead dis- play of proteins by cell-free expression of anchored DNA. J Biotechnol 106, 1–13. 19 Hansson M, Ringdahl J, Robert A, Power U, Goetsch L, Nguyen TN, Uhle ´ n M, Sta ˚ hl S & Nygren PA (1999) An in vitro selected binding protein (affibody) shows conformation-dependent recognition of the respiratory syncytial virus (RSV) G protein. Immunotechnology 4, 237–252. 20 Wikman M, Rowcliffe E, Friedman M, Henning P, Lindholm L, Olofsson S & Stahl S (2006) Selection and characterization of an HIV-1 gp120-binding affibody ligand. Biotechnol Appl Biochem 45, 93–105. 21 Nordberg E, Friedman M, Go ¨ string L, Adams GP, Brismar H, Nilsson FY, Sta ˚ hl S, Glimelius B & Carlsson J (2007) Cellular studies of binding, inter- nalization and retention of a radiolabeled EGFR- binding affibody molecule. Nucl Med Biol 34, 609–618. 22 Wahlberg E, Lendel C, Helgstrand M, Allard P, Dinc- bas-Renqvist V, Hedqvist A, Berglund H, Nygren PA &Ha ¨ rd T (2003) An affibody in complex with a target protein: structure and coupled folding. Proc Natl Acad Sci U S A 100, 3185–3190 [Epub 19 Feb 2003]. 23 Hogbom M, Eklund M, Nygren PA & Nordlund P (2003) Structural basis for recognition by an in vitro evolved affibody. Proc Natl Acad Sci U S A 100, 3191– 3196 [Epub 25 Feb 2003]. 24 Lendel C, Dogan J & Hard T (2006) Structural basis for molecular recognition in an affibody:affibody com- plex. J Mol Biol 359, 1293–1304 [Epub 6 May 2006]. 25 Eklund M, Axelsson L, Uhlen M & Nygren PA (2002) Anti-idiotypic protein domains selected from protein A-based affibody libraries. Proteins 48, 454–462. 26 Eklund M, Sandstrom K, Teeri TT & Nygren PA (2004) Site-specific and reversible anchoring of active proteins onto cellulose using a cellulosome-like com- plex. J Biotechnol 109, 277–286. 27 Lo Conte L, Chothia C & Janin J (1999) The atomic structure of protein-protein recognition sites. J Mol Biol 285, 2177–2198. 28 Hoyer W, Gronwall C, Jonsson A, Stahl S & Hard T (2008) Stabilization of a b-hairpin in monomeric Alzhei- mers amyloid-b peptide inhibits amyloid formation. Proc Natl Acad Sci U S A 105, 5099–5104. Affibody binding proteins P A ˚ . Nygren 2674 FEBS Journal 275 (2008) 2668–2676 ª 2008 The Author Journal compilation ª 2008 FEBS 29 Dogan J, Lendel C & Hard T (2006) Thermodynamics of folding and binding in an affibody:affibody complex. J Mol Biol 359, 1305–1315 [Epub 6 May 2006]. 30 Wahlberg E & Hard T (2006) Conformational stabiliza- tion of an engineered binding protein. J Am Chem Soc 128, 7651–7660. 31 Nord K, Gunneriusson E, Uhlen M & Nygren PA (2000) Ligands selected from combinatorial libraries of protein A for use in affinity capture of apolipoprotein A-1M and taq DNA polymerase. J Biotechnol 80, 45–54. 32 Ronnmark J, Gronlund H, Uhlen M & Nygren PA (2002) Human immunoglobulin A (IgA)-specific ligands from combinatorial engineering of protein A. Eur J Biochem 269, 2647–2655. 33 Gronwall C, Sjoberg A, Ramstrom M, Hoiden- Guthenberg I, Hober S, Jonasson P & Stahl S (2007) Affibody-mediated transferrin depletion for proteomics applications. Biotechnol J 18, 18. 34 Ronnmark J, Hansson M, Nguyen T, Uhlen M, Robert A, Stahl S & Nygren PA (2002) Construction and char- acterization of affibody-Fc chimeras produced in Escherichia coli. J Immunol Methods 261, 199–211. 35 Ronnmark J, Kampf C, Asplund A, Hoiden-Guthen- berg I, Wester K, Ponten F, Uhlen M & Nygren PA (2003) Affibody-beta-galactosidase immunoconjugates produced as soluble fusion proteins in the Escherichia coli cytosol. J Immunol Methods 281, 149–160. 36 Gunneriusson E, Samuelson P, Ringdahl J, Gronlund H, Nygren PA & Stahl S (1999) Staphylococcal surface display of immunoglobulin A (IgA)- and IgE-specific in vitro-selected binding proteins (affibodies) based on Staphylococcus aureus protein A. Appl Environ Micro- biol 65, 4134–4140. 37 Renberg B, Shiroyama I, Engfeldt T, Nygren PK & Karlstrom AE (2005) Affibody protein capture micro- arrays: synthesis and evaluation of random and directed immobilization of affibody molecules. Anal Biochem 341, 334–343. 38 Renberg B, Nordin J, Merca A, Uhlen M, Feldwisch J, Nygren PA & Karlstrom AE (2007) Affibody molecules in protein capture microarrays: evaluation of multido- main ligands and different detection formats. J Prote- ome Res 6, 171–179. 39 Andersson M, Ronnmark J, Arestrom I, Nygren PA & Ahlborg N (2003) Inclusion of a non-immunoglobulin binding protein in two-site ELISA for quantification of human serum proteins without interference by hetero- philic serum antibodies. J Immunol Methods 283, 225– 234. 40 Lundberg E, Hoiden-Guthenberg I, Larsson B, Uhlen M & Graslund T (2007) Site-specifically conjugated anti-HER2 Affibody molecules as one-step reagents for target expression analyses on cells and xenograft sam- ples. J Immunol Methods 319, 53–63 [Epub 21 Nov 2006]. 41 Karlstrom A & Nygren PA (2001) Dual labeling of a binding protein allows for specific fluorescence detection of native protein. Anal Biochem 295, 22–30. 42 Engfeldt T, Renberg B, Brumer H, Nygren PA & Karl- strom AE (2005) Chemical synthesis of triple-labelled three-helix bundle binding proteins for specific fluores- cent detection of unlabelled protein. Chembiochem 6, 1043–1050. 43 Renberg B, Nygren PK, Eklund M & Karlstrom AE (2004) Fluorescence resonance energy transfer-based detection of analytes using antiidiotypic affinity protein pairs. Anal Biochem 334, 72–80. 44 Sandstrom K, Xu Z, Forsberg G & Nygren PA (2003) Inhibition of the CD28-CD80 co-stimulation signal by a CD28-binding affibody ligand developed by combinato- rial protein engineering. Protein Eng 16, 691–697. 45 Finnzymes. Phusion Hot Start High-Fidelity DNA poly- merase (http://www.finnzymes.fi/pcr/phusion_hot_start_ high_fidelity_dna_polymerase.html). 46 Myhre S, Henning P, Granio O, Tylo ¨ AS, Nygren PA, Olofsson S, Boulanger P, Lindholm L & Hong SS (2007) Decreased immune reactivity towards a knobless, affibody-targeted adenovirus type 5 vector. Gene Ther 14, 376–381. 47 Magnusson MK, Henning P, Myhre S, Wikman M, Uil TG, Friedman M, Andersson KM, Hong SS, Hoeben RC, Habib NA, Sta ˚ hl S, Boulanger P & Lindholm L (2007) Adenovirus 5 vector genetically re-targeted by an Affibody molecule with specificity for tumor antigen HER2 ⁄ neu. Cancer Gene Ther 14, 468–479 [Epub 2 Feb 2007]. 48 Belousova N, Mikheeva G, Gelovani J & Krasnykh V (2008) Modification of adenovirus capsid with a designed protein ligand yields a gene vector targeted to a major molecular marker of cancer. J Virol 82, 630– 637. 49 Nilsson FY & Tolmachev V (2007) Affibody molecules: new protein domains for molecular imaging and tar- geted tumor therapy. Curr Opin Drug Discov Devel 10, 167–175. 50 Tolmachev V, Orlova A, Nilsson FY, Feldwisch J, Wennborg A & Abrahmsen L (2007) Affibody mole- cules: potential for in vivo imaging of molecular targets for cancer therapy. Expert Opin Biol Ther 7, 555–568. 51 Holliger P & Hudson PJ (2005) Engineered antibody fragments and the rise of single domains. Nat Biotech- nol 23, 1126–1136. 52 Nahta R & Esteva FJ (2007) Trastuzumab: triumphs and tribulations. Oncogene 26, 3637–3643. 53 Wikman M, Steffen AC, Gunneriusson E, Tolmachev V, Adams GP, Carlsson J & Stahl S (2004) Selection and characterization of HER2 ⁄ neu-binding affibody P A ˚ . Nygren Affibody binding proteins FEBS Journal 275 (2008) 2668–2676 ª 2008 The Author Journal compilation ª 2008 FEBS 2675 ligands. Protein Eng Des Sel 17, 455–462 [Epub 18 Jun 2004]. 54 Orlova A, Rosik D, Sandstrom M, Lundqvist H, Einarsson L & Tolmachev V (2007) Evaluation of [(111 ⁄ 114m)In]CHX-A’’-DTPA-Z(HER2:342), an Affi- body ligand conjugate for targeting of HER2-expressing malignant tumors. Q J Nucl Med Mol Imaging 1,1. 55 Mume E, Orlova A, Larsson B, Nilsson AS, Nilsson FY, Sjoberg S & Tolmachev V (2005) Evaluation of ((4-hydroxyphenyl)ethyl)maleimide for site-specific radiobromination of anti-HER2 affibody. Bioconjug Chem 16, 1547–1555. 56 Tran T, Engfeldt T, Orlova A, Widstrom C, Bruskin A, Tolmachev V & Karlstrom AE (2007) In vivo evalua- tion of cysteine-based chelators for attachment of 99mTc to tumor-targeting Affibody molecules. Biocon- jug Chem 18, 549–558 [Epub 2 Mar 2007]. 57 Engfeldt T, Tran T, Orlova A, Widstro ¨ m C, Feldwisch J, Abrahmsen L, Wennborg A, Karlstro ¨ m AE & Tol- machev V (2007) (99m)Tc-chelator engineering to improve tumour targeting properties of a HER2-specific affibody molecule. Eur J Nucl Med Mol Imaging 13, 13. 58 Steffen AC, Almqvist Y, Chyan MK, Lundqvist H, Tolmachev V, Wilbur DS & Carlsson J (2007) Biodistri- bution of 211At labeled HER-2 binding affibody mole- cules in mice. Oncol Rep 17, 1141–1147. 59 Tolmachev V, Nilsson FY, Widstrom C, Andersson K, Rosik D, Gedda L, Wennborg A & Orlova A (2006) 111In-benzyl-DTPA-ZHER2:342, an affibody-based conjugate for in vivo imaging of HER2 expression in malignant tumors. J Nucl Med 47, 846–853. 60 Baum R, Orlova A, Tolmachev V & Feldwisch J (2006) Receptor PET ⁄ CT and SPECT using an affibody mole- cule for targeting and molecular imaging of HER2 posi- tive cancer in animal xenografts and human breast cancer patients. J Nucl Med 47(suppl. 1), 108P. 61 Tolmachev V, Orlova A, Pehrson R, Galli J, Baastrup B, Andersson K, Sandstro ¨ m M, Rosik D, Carlsson J, Lundqvist H, Wennborg A & Nilsson FY (2007) Radio- nuclide therapy of HER2-positive microxenografts using a 177Lu-labeled HER2-specific Affibody molecule. Cancer Res 67, 2773–2782. 62 Nygren P, Uhle ´ n M, Flodby P, Andersson R & Wigzell H (1991) In vivo stabilization of a human recombinant CD4 derivative by fusion to a serum-albumin-binding receptor. Vaccines 91, 363–368. 63 Friedman M, Nordberg E, Hoiden-Guthenberg I, Brismar H, Adams GP, Nilsson FY, Carlsson J & Stahl S (2007) Phage display selection of Affibody molecules with specific binding to the extracellular domain of the epidermal growth factor receptor. Protein Eng Des Sel 20, 189–199 [Epub 23 Apr 2007]. 64 Friedman M, Orlova A, Johansson E, Eriksson TL, Hoiden-Guthenberg I, Tolmachev V, Nilsson FY & Stahl S (2008) Directed evolution to low nanomolar affinity of a tumor-targeting epidermal growth factor receptor 1-binding affibody molecule. J Mol Biol 376, 1388–1402. Affibody binding proteins P A ˚ . Nygren 2676 FEBS Journal 275 (2008) 2668–2676 ª 2008 The Author Journal compilation ª 2008 FEBS . MINIREVIEW Alternative binding proteins: Affibody binding proteins developed from a small three-helix bundle scaffold Per -A ˚ ke Nygren Department of Molecular. introduction of various labels and radionuclide chelators. Abbreviations ABD, albumin -binding domain; Ad5, adenovirus type 5; Ab, amyloid-beta; DOTA, tetraazacyclododecane