AffinityProteome: Science and Technology Background

The aim of AffinityProteome is to facilitate and accelerate developments in affinity proteomics for the biotechnology sector by linking producers of recombinant binding molecules with the developers of novel advanced technologies for their application. The binding molecules themselves will encompass both established single chain antibody fragments and alternative binders of particular promise (ankyrin repeat proteins, nucleic acid aptamers). The innovative technologies of protein microarrays, proximity ligation and intrabodies will address different areas of binder application.
The project will focus on development of reagents for protein kinases and their targets in cell signalling pathways, which are central to all cellular responses and which when disregulated can lead to disorders of cell growth, notably cancers, as well as autoimmunity and a range of other diseases. The specific targets are in the TGF-ß and MAP kinase pathways. The development of validated reagents for detection and potentially modulation of these pathways will be of immediate benefit to clinical research in this vital area and provide new leads for the pharmaceutical industry.


Recombinant binders and high-throughput selection systems

There are several reasons to favour the use of recombinant systems for large-scale commercial binder production. Avoiding the use of animals for binder generation through immunisation is an important consideration both from the ethical and practical standpoints. Recombinant products obtained through selection technologies have the advantages of readily available coding DNA sequence information, which in the future may well substitute for storing of the binding molecules themselves. This has the distinct advantage of resulting in objective data which, at least in principle, can be checked and controlled at any time and any place by resynthesizing the coding sequence of the binder and then expressing it. Additionally, the possibility of further engineering of recombinant binders as fusion proteins for downstream applications, including translation into diagnostics and therapeutics, makes them an attractive choice.
The selection technologies to be employed in AffinityProteome are phage display, ribosome display and SELEX aptamer selection. The choice of technology is guided by several factors, including the nature of the binder, library size and construction, robustness, automation, capacity for binder evolution, etc. In addition to automation and throughput, cost will be a major consideration in any large scale binder production programme, so that engineering processes to require much less target protein and use of miniaturised and parallelised systems to allow economically feasible production of many binders will be important elements. Hence, in both ribosome and phage display, particular emphasis will be given to reducing the experimental effort of antigen generation, panning, screening and expression.

Binder types

(a) The recombinant antibody fragments to be used are single chain (sc) scFv, scVH/K and scFab, for which complex fragment libraries will be produced for selection using ribosome display and phage display respectively. scFv fragments are the most widely used format in phage display, while scVH/K (VH domain linked to the complete kappa light chain) is designed for ribosome display. The scFab format is newly designed by TUBS for proteomics scale phage display by fusing the ORFs for Fd and VL domains using an inert linker. A familiar hurdle to be overcome for antibody fragment production is poor yield in bacterial expression systems, due to the tendency to aggregation in the reducing conditions of the cytoplasm and the need for periplasmic expression to achieve disulphide formation. This can be partly solved by using frameworks which are tested for expression in E. coli, or for screening purposes, by cell free expression systems.

(b) Three characteristic molecular features make DARPins (Designed Ankyrin Repeat Proteins) exciting alternatives for proteomics applications. First, their production yield from bacteria far exceeds that of most other classes of proteins tested, as they can be produced in the cytoplasm at 1 mg of binder from only about 5-10 ml E .coli culture. This allows a degree of parallelisation not previously possible. Secondly, the absence of disulphide bonds allows their correct folding in the cytoplasm of prokaryotes and eukaryotes alike. This permits their use as ‘intrabodies’ in intracellular assays. Thirdly, their favourable folding properties make fusion with many other proteins very convenient, as functional proteins can be produced directly, e.g. binders can be screened for cell binding using DARPin fusions with GFP and FACS analyses.

(c) Single stranded (ss) DNA aptamers, selected by the in vitro SELEX technology are a particularly interesting source of high affinity binders against many proteins, with some distinct advantages in production and amplification and similar properties of protein binding to those of antibodies. Rapid association and slow dissociation rates give them an unexpectedly high affinity (10 – 100 pM KD) and, since the dominant epitope is invariably conformational, they are ideally suited to assays for native proteins. Aptamers can be distributed simply as sequence information, from which the users can carry out the synthesis themselves. They are now being applied in plasma biomarker discovery and high throughput detection on aptamer arrays and data from SomaLogic show non-specific binding diminished to the point that single aptamers can be used in place of sandwich antibody pairs.

Technologies and applications

The advanced affinity-probe technologies to be exploited and developed further in the course of the project are highly suited to cellular studies of protein localisation, pathway analysis and protein-protein interactions. They include array based systems for protein detection and interaction analysis (BBT, DKFZ); the ultrasensitive in situ proximity ligation method (OLINK, UU), allowing the detection and enumeration of individual protein molecules and complexes in individual cells; and intracellular knockdown by binders expressed as intrabodies after transfection in living cells (UZH).

(a) Peptide and protein microarray are essential to the throughput required to characterise large numbers of binding reagents, as well as allowing identification of protein function and interactions on a genome-wide scale in vitro, complementing cell-based technologies. Binder or capture microarrays have an enormous potential of developing into tools that will allow the type of global characterisation of molecule mixtures at the protein level. Protein arrays are more complex, with major issues of how to produce the large number of proteins to be arrayed and how to maintain them in a functional state once immobilised. Solutions to both problems are found in the novel methods being developed at BBT and DKFZ for the generation of protein arrays from encoding DNA in situ and for the printing of protein arrays from DNA array templates (BBT). The arrays will be applied in reagent quality control (specificity, cross-reactivity) and analysis of protein interactions which will guide the use of the affinity reagents in vivo. Peptide arrays will be used to define epitope recognition specificity of binders. Selected binders will be immobilised as capture arrays to profile protein expression of signal transduction proteins in cell extracts. Binders specific for the normal and phosphorylated forms will allow one to determine the levels and activities of kinase-dependent pathways in a ratiometric microarray format.

Proximity ligation in situ. An ultrasensitive technique allowing individual interacting pairs of protein molecules to be visualised and enumerated in human cells and clinical specimens, is the combination of in situ proximity ligation assay (PLA) and Rolling Circle Amplification of DNA (RCA). The PLA technique, invented by the UU group and now being taken forward commercially by OLINK, is a powerful and generally applicable technique using binding reagents to detect minute amounts of proteins with excellent specificity, including single molecules in cells in situ.
In in situ PLA a fixed biological sample such as a tissue section is incubated with two primary antibodies recognising two epitopes either on a single protein molecule or on two interacting proteins (Fig. a). The primary antibodies are either directly conjugated with oligonucleotides, or targeted by specific secondary reagents conjugated with oligonucleotides (Fig. b). The oligonucleotides are then ligated to form a circle (Fig. c), initiating an RCA reaction and generating a long concatameric product anchored to the detected protein. This DNA product can be sensitively visualised by fluorescent hybridising probes in conventional or confocal fluorescence microscopy (Fig. d). The strength of the technology in visualising and enumerating protein interactions within individual cells and will be applied here to proteins of cell signalling pathways.

(c) Intracellularly acting protein-specific reagents (intrabodies) can be used both as labelled probes to detect proteins or to interfere with their function (protein knockout). The technology is a cost-effective approach to study endogenous proteins in cells, based on transient and stable expression technology. Intrabodies have been successfully applied, mainly in the scFv format, to inhibit the function of target proteins in specific cellular compartments. Their varied mode of action gives them great potential in different approaches to the treatment of human diseases, as well as in functional genomics for characterisation of novel gene products and subsequent validation as potential drug targets. They have immense potential in the process of drug development and may ultimately become therapeutic entities in their own right. Intrabodies will be of particular importance for detection of target protein level, modification and localisation in situ, all equally important cornerstones of cellular function which should be addressed for comprehensive proteome analysis. The particular focus will be the intracellular action of DARPin and scFv binders against kinase and other targets (UZH, BBT, UNIK). The major limitation of scFv fragments is that, due to the absence of disulphide bond formation in the cytoplasm, their stability is reduced, such that only a subset of all scFv fragments will function intracellularly. Therefore, it is a great advantage to use a framework such as the ankyrin repeat (DARPin) scaffold where all members can fold in the absence of disulphide bonds and which do not have any unpaired cysteines.

Signal transduction pathways as proteome targets

AffinityProteome will develop the application of binders in protein interactions, pathway analysis and network definition, focusing on signal transduction in the TGF-beta and MAPK pathways as model proteomic systems. Three types of binders are of particular significance in this application. (i) The first is highly specific for a given kinase. Such binders would be of great utility in the total quantitation of a kinase, e.g. by binder arrays, as well as in spatio-temporal studies. (ii) The second type of binder would discriminate between the phosphorylated and non-phosphorylated forms of the kinase. This could be used to observe the effect of upstream activators or inhibitors (e.g. from extracellular ligands). (iii) The third type of binder would have a biological effect when expressed inside the cell. Intrabody binders would greatly clarify the flux of information in kinase signalling and enormously improve our toolbox for quantifying these parameters, which in turn would be the input of systems biology.

The TGF-ß family of cytokines regulates many aspects of cell physiology, including cell growth, differentiation, motility and death. It is implicated both as a transforming factor and as an inhibitor of cell proliferation. It stimulates tumour progression by enhancing the malignant phenotype, metastasis and invasion. TGF-ß receptors signal to normal or cancer cells via the central Smad pathway. Smads interact with each other in phosphorylated or non-phosphorylated forms; heteromeric complexes accumulate in the nucleus where they bind directly to DNA (Smad-binding elements) and associate with transcription factors, co-activators or co-repressors, leading to transcriptional induction or repression of a diverse array of genes. This is a key pathway in mammalian cellular regulation and its derangements in cancer, providing many target opportunities for binder-based functional analysis and potential intervention.

The MAPK pathways are major signalling pathways regulating cell division and cell death, activated by receptors for growth factors and stress stimuli. At least 12 isoforms of MAPKs exist in mammalian cells, divided into 4 main groups: the ‘classical’ MAPKs (ERK1 and 2), JNKs, p38s, and atypical MAPKs such as ERKs 3, 5 and 8. MAPKs are activated on specific threonine and tyrosine residues by a dual specificity MAPK kinase (MKK), which is in turn activated by a MAPK kinase kinase (MKKK) in response to appropriate extracellular stimuli. Growth factor receptor signals are relayed by a three-tier hierarchical array of protein kinases (Raf – MEK1/2 – ERK1/2); activated ERK1/2 accumulate in the nucleus where they phosphorylate transcription factors and so promote or repress gene expression. Other emerging components of the pathway include scaffold proteins, such as KSR, which bring signalling proteins together in complexes for interaction. Specific protein phosphatases bind, dephosphorylate and inactivate the kinases within the cascade. Mammalian cells possess at least two stress-activated protein kinases, p38 and JNK, that are activated by similar protein kinase cascades via MKK3/6 and 4/7, respectively. Parallel pathways which are less well studied but of considerable current interest and for which reagents are particularly needed include the MEK5 – ERK5 pathway; molecules capable of disrupting such pathways would be particularly valuable in distinguishing their specific roles.
There is also scope for multiple interaction (cross-talk) between the TGF-ß and MAPK pathways, recently demonstrated both functionally and mechanistically, with clear implication in cancer. Such cross-talk is another important focus area for binder-based investigation.

The intention is to use about 30 proteins as targets. They will include the Smad proteins involved in TGF-ß signalling (Smad2, Smad3, Smad4, Smad7); TGF-ß receptors type I (ALK5), type II and type III (endoglin); and transcription factors known to interact with Smad proteins (Sp1, FoxO, S100C, p53, p107, E2F4). For the MAPK pathway, the targets include MAPKs, notably ERK1,2,3,5, JNK and p38; MKKs including MEK1,2, MKK 3, 4, 6 and 7; MKKKs such as MEKK1 and RAF; scaffold proteins such as KSR and ß-arrestin and some small GTPases such as Ras and Rac.