Expert Opinion on Biological Therapy
ISSN: 1471-2598 (Print) 1744-7682 (Online) Journal homepage: http://www.tandfonline.com/loi/iebt20
Cambridge Healthtech Institute’s 5th Annual ‘Molecular Display:The Chemistry Set for Proteins and Small Molecules’ Conference Andrew M Coley & Joanne L Casey To cite this article: Andrew M Coley & Joanne L Casey (2003) Cambridge Healthtech Institute’s 5th Annual ‘Molecular Display:The Chemistry Set for Proteins and Small Molecules’ Conference, Expert Opinion on Biological Therapy, 3:5, 855-858, DOI: 10.1517/14712598.3.5.855 To link to this article: http://dx.doi.org/10.1517/14712598.3.5.855
Published online: 03 Mar 2005.
Submit your article to this journal
Article views: 6
View related articles
Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=iebt20 Download by: [Florida State University]
Date: 05 November 2015, At: 21:02
Meeting Highlights Peptides, Proteins & Antisense
1. Keynote presentations 2. New technologies and target selection
Cambridge Healthtech Institute’s 5th Annual ‘Molecular Display: The Chemistry Set for Proteins and Small Molecules’ Conference 12 – 13 May 2003, Cambridge, MA, USA
3. Use of phage display to discover small molecules 4. Applications in proteomics
Downloaded by [Florida State University] at 21:02 05 November 2015
5. Applications in novel therapeutics and diagnostics 6. Expert opinion
Andrew M Coley† & Joanne L Casey CRC for Diagnostics, Biochemistry Department, La Trobe University, Victoria, 3086, Australia This meeting covered recent advances in the molecular display of peptides, proteins and nucleotides, including selection and mutational technologies. The scientific organisers assembled an impressive array of ‘molecular display’ heavyweights. It promised to be a stimulating meeting and the events of the following 2 days did not disappoint. The majority of the presentations were concerned with the development of novel display technologies and processes. Antibodies currently represent > 30% of the biopharmaceutical market, but are likely to be superseded by more efficient display frameworks which avoid their inherent drawbacks. In order to generate such novel therapeutics and diagnostics, high affinity reagents must be selected and/or generated from hitherto unexplored nucleic acid sequences and displayed on suitable frameworks. This meeting was concerned with the identification, generation and validation of novel sequences and framework molecules. The keynote addresses were followed by four themed sessions entitled New technologies and target selection, The discovery of small molecules using phage display, Applications in proteomics, and Novel therapeutics and diagnostics. There was a panel discussion after each session. Keywords: affinity, phage display, protein scaffold, protein structure Expert Opin. Biol. Ther. (2003) 3(5):855-858 1.
For reprint orders, please contact:
[email protected]
Ashley Publications www.ashley-pub.com
Keynote presentations
D Witrup opened the meeting by chairing the first session, which contained the keynote addresses. Appropriately, the first presentation was by A Plückthun. He introduced his talk by discussing the requirements of the new technologies and the approach his lab has undertaken in order to explore the boundaries of molecular display. His presentation was divided into three main areas; initially he discussed the novel technology he has embraced, namely ribosome display, and provided an excellent explanation of the philosophy, principles and mechanism of this display format. He then went on to discuss the potential applications of ribosome display with respect to selection of binding molecules by off-rate, stability and enzymatic turnover. He also touched on how he had noticed substantial improvements in antibody reactivity when applying their selection to the ribosome display format. Novel scaffold discovery using ankyrin repeats was discussed and it was suggested these molecules had good crystal structures with stabilising β-turns, good thermal stability and demonstrated low nM Kds. With this in mind, he described a protocol with which his lab has developed some intracellular enzyme inhibitors with subnanomolar affinities. Lastly, he touched on attempts to develop novel proteins thus far undiscovered
2003 © Ashley Publications Ltd ISSN 1471-2598
855
Downloaded by [Florida State University] at 21:02 05 November 2015
Cambridge Healthtech Institute’s 5th Annual ‘Molecular Display: The Chemistry Set for Proteins and Small Molecules’ Conference
and/or unused in nature. His approach has been to take a mixture of protein structures (α-helices, β-strands and turns), randomise them with respect to each other, then select for structural integrity and stability. This screening was achieved by assessing protease resistance and hydrophobic interaction. J Szostak (Howard Hughes Medical Institute) introduced his presentation entitled ‘Evolution of novel proteins by mRNA display’ by comparing the limits of sequence space with physical space. He concluded that the total diversity in a random polypeptide library consisting of 100 amino acids would be 20100 (or 10130) individual clones. If such a library were constructed, it would occupy a volume larger than the known universe: a sobering thought when libraries are generally constructed in ‘Eppendorf ’ tubes. Of course, the takehome message from his astronomical calculations is that we cannot hope to cover all the bases in terms of random sequence followed by ‘dumb’ selection, but have to evolve the selected polypeptides as we select them in order to take account of this spatial inadequacy of the experiment. As the title of his presentation suggested Szostak’s display technology of choice is mRNA display, he ably explained the mechanisms and idiosyncrasies associated with mRNA display and discussed its applications. An mRNA display library of 80 random amino acids, which has a theoretical diversity of 8 × 1098 possible members, was synthesised. He sampled 6 × 1012 of these, which is still a large library, larger than any commercial peptide display library in terms of diversity and size of individual polypeptide dimensions. As a proof of principle, he selected the library on streptavidin and demonstrated selection of an ‘HPQ’ motif, which had already been described as binding streptavidin, and then went on to describe his work on selecting polypeptides that bind to ATP. After nine rounds of selection, 5% of clones bound to ATP, demonstrating that functional proteins can be selected from random peptide libraries using mRNA display. Furthermore, nine additional rounds of selection incorporating three rounds of mutation resulted in 35% of polypeptides binding to ATP that were all derived from a single clone. When the best performer was analysed further, it was shown to have a Kd of 100 nM for ATP and significantly lower affinity for ADP, AMP, CTP, UTP and GTP. Unfortunately, this polypeptide was a poor performer in biological assays, as it was prone to aggregation. The question of whether it is possible to evolve a protein for stability while maintaining ATP binding characteristics was then raised. He chose to address this by carrying out selection of the ATP binders in the presence of guanidine–HCl, and demonstrated some stability improvements in the ATP-binding polypeptides. Szostak then presented sequence alignments and comparisons of the ATP-binding followed by the stabilised ATP-binding polypeptides. 2.
New technologies and target selection
R Balint (KaloBios, Inc.) described a process termed ‘functional epitope-guided selection’ or FEGS. The selection 856
process can be carried out in the bacterial periplasm and is based on the receptor and ligand fused to either β-lactamase (BL) or β-lactamase inhibitory protein (BLIP). In an uninhibited state, the receptor interacts with the ligand and this then facilitates the BL/BLIP interaction, rendering the enzyme inactive. The net result is that the growth of Escherichia coli in this state is suppressed in the presence of ampicillin. Alternatively, in the presence of a periplasmic polypeptide capable of competing, the receptor–ligand interaction results in a reduction in BL inhibition, due to the relative loss in BL/BLIP association and subsequent growth of the bacterium in media containing ampicillin. Balint went on to describe his proof of concept experiment using CD40– BLIP as the receptor and CD40L–BL as the ligand. He also mentioned that when expression of the antibody fragment was reduced, a higher diversity of inhibitory antibodies was achieved. With proof in tact, he then described applications of the technology to, first, humanisation of antibody fragments, then affinity maturation using N-terminally displayed peptide inhibitors of the antibody/antigen reaction, a process he called CDRO affinity maturation. Finally, Balint also briefly mentioned further complex applications, based on similar technology, where decoy enzyme activators were used in the periplasmic competition assay. M Sleigh (EvoGenix Pty, Ltd) then changed the subject in her presentation entitled ‘High diversity variant libraries for rapid protein optimisation’. The technology behind the diversity generation centres on the ability to generate random mutations in displayed peptides and proteins. The ability to generate supposedly random mutations in DNA has been around for several years. These methods rely on either increasing the inherent error rate of Taq polymerase or the use of chemical mutagens. Neither of these technologies are ideal, as they are not truly random, due to base preferences and/or problems with secondary structure of the DNA template and/ or intellectual property considerations. We were told the EvoGenix solution to these problems was to use an RNA replicase of viral origin to introduce random mutations into the target sequence. This technology particularly suits ribosome and mRNA display technologies, as RNA is the encoding molecule. Dr Sleigh showed data supporting her assertions that RNA replicases are far superior to DNA polymerases and chemical mutagens in generating mutations of a much more random nature. Her data demonstrated that random mutations generated in the β-lactamase gene, followed by selection of E. coli mutants for greater tolerance to β-lactam antibiotics, resulted in an impressive 1 × 104 increase in resistance to the β-lactam antibiotic cefataxime concentration after just two rounds of mutagenesis. S Koide from the University of Chicago then introduced the inherent problems with the more established molecular display platforms in a presentation entitled ‘Monobodies: binding proteins based on the fibronectin type III domain.’ He suggested the requirement for alternative scaffolds that present peptide-based affinity reagents. Perhaps not surprisingly, in
Expert Opin. Biol. Ther. (2003) 3(5)
Downloaded by [Florida State University] at 21:02 05 November 2015
Coley & Casey
view of the title of his presentation, he identified a fibronectin type III domain as a candidate monobody. The single fibronectin domain (FNfn10) binds αvβ3 integrin poorly, but when a phage-displayed version with randomised ‘B-C’ and ‘F-G’ loops was panned on αvβ3, the resulting domain bound to αvβ3 with a Kd of 0.5 nM, an improvement of > 103-fold compared with the wild type. He then demonstrated that the soluble form of this αvβ3-binding FNfn10 was useful in cell binding assays and was heat stable. As FNfn10 has no intramolecular disulfide bonds, it was postulated that it might function as an intracellular monobody. To this end he demonstrated a potential use for this scaffold in applications in a yeast two-hybrid system. A Belcher (MIT) then changed tack somewhat and described her application of peptide phage display to materials science in her excellent presentation entitled ‘Phage-based electronic and magnetic materials.’ She clearly knew she was among a very small number of materials scientists (if not the only one) in the room and, therefore, explained in graphic, but nonetheless basic terms the theory behind her ideas. These revolved around the fact that inorganic compounds use proteins to control their properties. With this in mind, she panned a phage peptide library on inorganic compounds used in the electronics industry. She showed the selection of phage with an affinity for sulfides of zinc and cadmium. She also showed that when the peptides mediating this affinity were displayed on the gene VIII product, they could facilitate the growth of the phage–galinium complex, and that the complexes were stable, extremely highly ordered and efficiently conducted electricity. This raised the proposition that they could be laid down as a wafer and etched to leave microwires in the same way as a conventional silicone chip is constructed. Some impressive fluorescent images were presented to suggest that this was indeed possible and indicated that this phenomenon was of high interest to the electronics industry.
Use of phage display to discover small molecules 3.
may act as a nucleus for folding of the protein around the small molecule, thereby resulting in a much stronger interaction. Therefore, the amino acid residues responsible for initial contact may not be the same amino acids responsible for the eventual high affinity binding of the small molecule. He went on to conclude that the peptides isolated in this study were responsible for the initial low affinity contact with ATP. This explains the observed homology of these peptides to ATP binding proteins, but their lack of homology to the ATP binding sites contained within the proteins. S Cwirla (XenoPort, Inc.) described a novel mechanism of using phage as nanoparticles. The approach was to chemically conjugate compounds from a combinatorial library to members of a phage library. Selection involved internalisation of the compound and hence the phage. Identification of the compound mediating the internalisation was a relatively trivial case of hybridisation of the selected phage (DNA) to an array containing the phage library that was constructed prior to conjugation, followed by correlation of the hybridising clone to the conjugated compound. He showed efficient chemical conjugation and selectability, followed by a proof of principle involving the selection of a 960-member folate analogue library. H Pederson (Nuevolution A/S) then gave a presentation describing technology involving the display of small molecules on DNA scaffolds, and went on to describe the principles and processes involved behind the technology and the potential advantages in terms of lead compound discovery. This was followed by E Lazerides (Targeted Molecules Corporation), who described his company’s proprietary technology, which he called an ‘in vivo discovery engine’. The basic ideas presented were LIVO and kSAR, where a phage peptide library was administered to a mouse (intravenous). The mouse was sacrificed and target organ/s excised and the phage eluted. The peptide/s identified as ‘targeting’ were then mutated to improve required qualities, and the chemical topology of the functionally optimised peptide was compared with a library of small molecule structures. 4.
L Makowski (Department of Chemistry, University of California Irvine) presented work centred on the identification of new small molecule-binding consensus sequences using peptide phage display. He chose ATP-binding proteins as his proof of principle. A previously well-characterised ATP-binding motif is the octapeptide ‘P-loop’ (GXXXXGKS/T). Makowski described how he tried to identify P-loops in a commercially available random peptide phage library by panning on immobilised ATP. Five hundred ATP-binding peptides were identified, but no P-loops were among their number. He went on to point out that there are 35 different classes of ATP-binding protein and 75% of his ATP-binding peptides had homology to at least one of these proteins, but, interestingly, homology was not to the ATP-binding sites. He concluded that when proteins encounter a small molecule, they interact in a relatively low affinity, low specificity fashion and the small molecule
Applications in proteomics
H Lin (Department of Chemistry, Columbia University) kicked off the session with a presentation entitled ‘Chemical complementation: a genetic assay for protein evolution and proteomics’. His process involves the yeast three-hybrid system, where a cleavable substrate with methotrexate and dexamethazone moieties at either pole allows the association of transcriptional activators of the lacZ gene, and hence the eventual formation of a blue colony when cultured under the appropriate conditions. When the yeast are transformed with a gene encoding an enzyme which cleaves the substrate, thereby dissociating the methotrexate and dexamethazone, the transcriptional activators are less likely to associate, thus producing a white colony. After an informative presentation by Y Chen (Phylos, Inc.), who described mRNA display and its applications to human antibody libraries, E Patz (Duke
Expert Opin. Biol. Ther. (2003) 3(5)
857
Downloaded by [Florida State University] at 21:02 05 November 2015
Cambridge Healthtech Institute’s 5th Annual ‘Molecular Display: The Chemistry Set for Proteins and Small Molecules’ Conference
University Medical Centre) requested more tools with which to carry out molecular imaging of his oncology patients. He has approached this problem with a combined proteomics and phage display technique. Biopsies were fractionated and profiled by mass spectrometry and a protein was identified as a putative lung cancer marker. Phage peptide libraries were panned on this recombinant protein and tumour cells. The resulting peptides demonstrated a high affinity and allowed effective imaging of both primary and secondary tumours. M Tainsky (Wayne State University) then continued the cancer diagnostic theme in his presentation. He has panned phage cDNA derived from ovarian cancer cell libraries on ovarian cancer patients’ antibodies and used the resulting clones in an array format with test sera from patients and controls. He demonstrated this technique had potentially good prognostic and diagnostic value.
C Rusconi (Duke University) gave a presentation on DNA and RNA aptamers and pointed out that amino acid polymers were not the only medium for the selection of high affinity reagents, but single stranded nucleic acid aptamers demonstrated selectability and high affinity. His group have managed to select a DNA-based antidote to clinically used blood clotting inhibitors. S Klussmann (Noxxon Parma AG) then turned us all inside-out with a description of his ‘Spiegelmers’. Spiegelmers are enantio aptamers, that is to say they are the mirror image of conventional aptamers. In order to develop them, the mirror image of the target is constructed and a conventional aptamer is selected on this enantio target. It, therefore, stands to reason that when the enantio, or mirror image, aptamer is constructed, it will bind to the right-way-round version of the target. Klussmann took great pains to convince us that this is the case.
Applications in novel therapeutics and diagnostics
6.
5.
After a talk by A Sato (Dyax Corp), who described how Dyax were exploiting phage display with their multitude of libraries (both peptide- and antibody-based), D Sidhu (Genentech) took to the stage and reviewed his group’s work on shotgun alanine scanning in the study of protein–protein interactions. Genentech has an interest in the biology of human growth hormone (hGH); it is, therefore, appropriate that he should use hGH in his studies. He described the philosophy behind the shotgun alanine scanning and homologue scanning techniques he has pioneered. He then went on to demonstrate how he has rationally designed a higher affinity hGH on the basis of his results from the scanning techniques. This approach was applied to a recombinant anti-Her2 antibody, where increased affinity was achieved, and a camelid VH domain where a 17 amino acid addition was constructed in the complementarity determining region 3 loop and shown to be stable. This stability was shown by shotgun alanine scanning to be conferred by consensus sequences at either end of the 17 amino acid loop. He also described some work on PDZ domains, where they had established which amino acids were responsible for the affinity of the domain for its ligands, again work carried out using the shotgun scanning technique.
858
Expert opinion
The meeting provided the opportunity to discuss emerging concepts in the field of molecular display. It was clear that a majority of the driving force behind the development of novel display technologies is the eventual supersedence of antibodies by new display molecules able to target clinically relevant markers. The presentations and subsequent discussions during this meeting examined the steps forward that have been taken by the major players in the field with respect to the next generation of recombinant therapeutics. It was also clear that novel molecular display technologies have an important role to play in the elucidation of receptor–target interactions. It can be concluded from this excellent meeting that in the near future it will be possible to generate a custom binder to almost any target molecule, without the disadvantages of immunogenicity and tissue penetration associated with antibodies. Affiliation Andrew M Coley† & Joanne L Casey †Author for correspondence CRC for Diagnostics, Biochemistry Department, La Trobe University, Victoria, 3086, Australia Tel: +61 (03) 9479 1158; Fax: +61 (03) 9479 2467; E-mail:
[email protected]
Expert Opin. Biol. Ther. (2003) 3(5)
