Recombinant DNA technology
Recombinant DNA technology is a basic technic to prepare expression vectors to produce various recombinant proteins for special purposes e.g. drug targets, biotherapeutics, subunit vaccines, etc. We can use the advantage of codon optimization during the design of the expression vectors. We have sets of promoters, signal sequences, fusion tags, identifier tags that we can use in proper combinations to design and reach the goals of our clients.
Protein expression technologies
There is particular challenge to recover the active conformation of a protein after completing the expression. TargetEx has multi-year experience working with ‘troublesome’ proteins and the art is in choosing the right expression system and designing the appropriate expression construct. For the production of recombinant proteins, we have a well-equipped molecular biology laboratory, with bacterial and yeast protein expression units, mammalian and insect cell culture laboratories.
Protein purification technologies
Protein purification is one of the key expertise of the company. We successfully purified different proteins for various purposes: R&D, X-ray or NMR study, vaccine, biosimilar product, diagnostic enzymes or for in house developments. TargetEx also offers services for affinity-tagged and non-tagged native proteins from small scale (evaluation the protein quality) to laboratory scale expression. Up-scalable purification process development (up to 1 g) could meet the strict quality requirements of the clients. Extensive biophysical characterization of the expressed and/or refolded proteins is also provided. Purification is available for process related or product related impurities.
Protein refolding technology
RefoldAll™ is a proprietary technology of TargetEx developed to provide solutions for the recombinant protein expression situations where the expression proceeds with low yield or the protein has low solubility and/or high aggregation characteristics. The service includes:
- Subcloning of the target sequence into a bacterial expression vector
- Production of inclusion bodies
- Purification of inclusion bodies
- Renaturation screen and optimization
- Purification of the properly folded protein
- Protein activity measurement and/or structural study of the folded protein
Case Study: active MASP-1 and MASP-2 fragment
The structure of MASP-2 CCP2-SP (Harmat et al., 2004)
Mannan-binding lectin-associated serine protease MASP-1 and MASP-2 are modular serine proteases and form complexes with mannan-binding lectin, the recognition molecule of the lectin pathway of the complement system. Both enzymes autoactivate and cleave synthetic oligopeptide substrates. In a competing oligopeptide substrate library assay, MASP-1 showed extreme Arg selectivity, whereas MASP-2 exhibited a less restricted, trypsin-like specificity. MASP-2 cleaves C2 and C4 at high rates. MASP-1 and MASP-2 fragment react with C1-inhibitor, which completely blocks the enzymatic action of the enzymes.
The catalytic region of MASP-1 and MASP-2, consisting of the two complement control protein modules and the serine protease domain (CCP1-CCP2-SP) was expressed in E. coli cells. Since the recombinant proteins accumulated as inclusion bodies inside the bacterial cells, renaturation procedures were needed to restore the native structure using our RefoldAll™ technology. The renatured recombinant proteins were purified by ion exchange chromatography. The pure MASP-2 CCP2-SP protein fragment was then crystallized and its structure analyzed. Synchroton X-ray diffraction yielded < 2Å resolution.
Technologies for protein analytics:
- The recombinant proteins are analyzed using a line of biochemical tools. The recombinant protein is identified using western blot and/or mass spectrometry. Protein purity is analyzed by SDS-PAGE using a sensitive dye. As SDS-PAGE is only sensitive to protein impurities we developed and use fluorescent methods to quantify RNA, DNA and endotoxin content in the protein preparations.
- For proteins used in RNA or DNA based applications the absence of RNAse and DNAse is essential for high quality results, therefore we also developed methods to guarantee the RNAse and DNAse free protein preparations.
- The activities of recombinant proteins are measured using enzymatic assays when possible. We developed and applied a whole range of activity assays including those for proteinases, kinases, polymerases, oxidases, etc. The readout is based on fluorometric, luminometric or photometric methods using 96 or 384 microplates. We also developed ELISA based assays for the measurement of protein-protein binding.
- The recombinantly expressed and purified proteins are subjected to biophysical analytical processes beside the conventional biochemical methods. The secondary structure element composition of the protein is predicted from far-UV circular dichroism (CD) measurement and compared to that obtained from crystal structure (or in the absence of it, homology model) to verify the existence of the correctly folded state of the protein.
- Differential scanning calorimetry (DSC) is a tool to verify the existence of the correct tertiary structure of the protein by monitoring the thermal denaturation of the globular protein. Differential scanning fluorescence (DSF) is a very similar tool which requires an external fluorophore, but is faster and much higher throughput than DSC. The DSF technique is useful for high throughput ligand binding studies as well as a fast, preliminary tool, on the basis of ligands stabilizing the native state of the protein.
- Ligand binding can also be quantified using surface-based methods: surface plasma resonance (SPR) and quartz crystal microbalance (QCM). Both techniques make possible quantifying the association (kon) and dissociation (koff) rates of ligand binding beside measuring the affinity values (Kd). Isothermal titration calorimetry (ITC) is useful for quantifying the thermodynamic parameters of binding (enthalpy and entropy) in solution. These tools are complementing each other for a full picture of protein-ligand binding.
- To gain insight into ligand-antibody interactions we developed a phage-display/SPR/bioinformatics-based workflow for the analysis and determination of the epitope/epitope binding region of the antigen/antibody interaction (Hajdú et al. „Monoclonal antibody proteomics: Use of antibody mimotope displaying phages and the relevant synthetic peptides for mAb scouting.” Immunology letters 160.2 (2014): 172-177.).
Computer-aided drug design (CADD) approaches complement the in vitro/in vivo biological screening technologies in order to facilitate the discovery process reducing time and efforts and increasing the success rate. Analysis of the biologically active chemical space, and prediction of the pharmaceutically relevant properties enable efficient ligand-based –and protein-target-based virtual screening approaches. As a result of the virtual screening process target-focused libraries are generated.
Target-focused library generation
Target-focused compound libraries are designed to interact with an individual protein target or target-families. The design of target-focused libraries generally utilizes structural information about the target or target-families of interest or the structure of active ligands available for the targets. Focused library screening often results in significant fold increase in the hit rate compared with random screening of commercial libraries.
Since multi-million small molecules are available from commercial sources by now the generation of the target-focused libraries has become a standard approach in early phase drug discovery. Ligand-based or protein structure-based virtual screening approaches are applied.
Ligand-based virtual screening approaches:
Ligand-based approaches utilize various chemoinformatics methods (2D/3D similarity search, diversity and property-based filtering, bioisosteric replacements, fragment-based approaches, 3D pharmacophore model building) together with measured and calculated data of the active molecules to identify potential active candidates.
2D similarity search with diversity, property-based filtering and bioisosteric replacements
The key concept of the 2D ligand-based virtual screening approaches is the Similarity Property Principle, which states that similar molecules should have similar biological properties. If determining the similarity between the biologically active reference compound and each molecule in a database, followed by ranking the database molecules according to the similarities it would lead to potentially active, target-focused libraries.
Random vs. Focused libraries
The similarity search uses 2D ﬁngerprints, i.e., binary strings encoding the presence or absence of a substructure within the molecules. Simple 2D ﬁngerprints are often applied, when numerous reference compounds and multimillion compound databases are available.
We extensively applied 2D ligand-based virtual screening in many projects complemented with drug-likeness or target-family based physico-chemical parameter space filtering, clustering, diversity filtering, application of bioisosteric replacements and 3D modelling and docking.
The structural similarity most frequently expressed in Tanimoto coefficient. The performance of the 2D similarity search relies on the diversity of the seeds and the searchable chemical space (vendor libraries), the type of fingerprints as well as the applied similarity cut-off values.
The major physico-chemical properties (Molecular weight, cLogP, H-bond donors, H-bond acceptors), are included in the Lipinski’s Rule of 5, while rotatable bonds and topological polar surface (tPSA) area are noted as Veber Rules. If most of the calculated parameters of a novel drug candidate fall into the pre-defined ranges, the concerning molecule is drug-like and has a high probability for oral absorption. Similarly, a multiparameter optimization (MPO) score was empirically defined for CNS compatibility including favorable ranges of physico-chemical parameters. The above rules provide general cut-off values and ranges for library filtering and represent a parameter “window” favorable for the specific target area. (Flachner B, et al. Med. Chem. Res. 23(3) 1234-1247 (2014).).
Target-family based physico-chemical parameter space filtering
Several target families have a distinct property range, therefore, defining a target-specific parameter space is often more accurate for focused library filtering than the universal Lipinski’s Rule of 5 based filtering. We successfully applied property-based filtering for identifying PDE5 inhibitors (Tömöri T et al. Molecular Diversity. 16(1), 59-72 (2012).)
Clustering and diversity-based filtering.
Clustering or scaffold analysis, and diversity selection are the methods of choice to reduce the number of compounds for biological screening. Diversity selection uses 2D molecular fingerprints, however, in that case the most dissimilar compounds are selected to represent the entire chemical space of the original library. Clustering and diversity selection are routinely applied in many in-house drug discovery projects.
Application of bioisosteric replacements
Bioisosterism is based on the assumption that certain substructures or fragments of biologically active compounds are interchangeable without losing the biological activities. Such structural motifs possess similar steric and electronic features, including shape, volume and charge distribution as well as physicochemical properties (e.g., hydrophobicity). Such bioisosteric analogs could extend the chemical space of the reference/entry compounds for similarity or substructure search, which increases the probability to find potential active compounds. We applied bioisosteric replacements in several projects: Szaszkó M, et al. Molecular Diversity, 21(1), 175-186. (2017), and Hajdú, et al. I., (2018). Bioorg. Med. Chem. Letters, 28(18), 3113-3118.)
Fragments are small molecules typically with a molecular mass between 150–250 Da and rather hydrophilic, yet represent an extended chemical space. Although fragment hits are weak binders, they are very “atom efficient” representing the basic (enthalpy based) interactions with the proteins. Fragments could be gained from the disconnection of bioactive drug-like compounds if available and the resulting fragments would serve as starting points for de novo fragment-based design using 2D similarity search. After in vitro screening the fragment hits could be optimized by fragment merging, linking, or growing that could lead to potential drug-like candidates. We applied the fragment-based approach for identifying potential glutaminyl cyclase inhibitors. (Szaszkó M, et al. Molecular Diversity, 21(1), 175-186. (2017),
3D similarity approaches
The possible binding features of the small molecules can be assessed by their 3D conformational flexibility and shape. Applying flexible alignment analysis and molecular dynamics the statistically average conformations generated allows to compare the 3D similarity between two compounds. The 3D similarity measures are expressed in 3D Tanimoto coefficient (T3D). We combined 2D/3D similarity selection methods, applied a 2D/3D fusion score and a reasonable cut-off value to identify novel PDE4 inhibitors. (Dobi K, et al. Molecules, 19(06), 7008-7039 (2014).
3D pharmacophore model building and screening
Pharmacophore models can be generated based on the 3D structure of the available seed (reference) compounds. By analyzing their structures six common pharmacophoric features (hydrogen-bond donor and acceptor, acidic and basic centers, as well as hydrophobic centers and aromatic rings) can be identified and arranged in 3D space. Pharmacophore screening offers finding de novo compounds that share the same features, but are based on a different molecular architecture than the seed compounds. We successfully applied pharmacophore screening in combination with 2D similarity search to identify potential 5-HT6 antagonists from large commercial repositories (Dobi K, et al. Chemical biology & drug design. 86(4), 864-80. (2015).)
When the 3D structure of the target protein is available (preferably with bound ligands) or can be easily deduced, 3D modelling and docking procedures can be involved in the focused library generation. It offers a de novo approach to select compounds from either discovery or focused libraries. This virtual screening approach involves docking of the candidate small molecules into a protein target followed by the calculation of a scoring function and free energy of the binding interaction, which estimates the probability if the small molecule is able to bind the protein with high affinity, or not. Our group applied this method in combination with 2D similarity search to identify potential PDE5 and C1s inhibitors, as well as fragment merging for glutaminyl cyclase inhibitor discovery. (Tömöri T, et al. Molecular Diversity. 16(1), 59-72 (2012).; Szaszkó M et al., Molecular Diversity, 21(1), 175-186. (2017),). Szilágyi, K. et al. Molecules 2019, 24(20), 3641; doi:10.3390/molecules24203641)