A fundamental goal in chemical biology and drug development is to develop improved methods to identify small molecules to engage biological targets, particularly “undruggable” proteins. This is a major focus of the Kodadek Laboratory. Some of the efforts in which we are engaged are summarized below.
Rapid Identification of Protein-Binding Fragments
Fragment-based methods are an efficient way to develop high affinity protein ligands. In this approach, thousands of low molecular weight molecules (usually < 300 Daltons) are somehow screened for binding to a protein target. These fragments are then either grown or linked to eventually generate higher affinity drug/probe molecule leads (Figure 1).
Figure 1. Fragment-based drug development. The red shapes represent low molecular weight small molecules and the blue shape is the protein target.
Because these interactions will be weak (KD – high µM to low mM), only a small number of biophysical assays are capable of detecting them reliably enough to be used in screening, such as NMR or SPR, which are tedious, expensive, and require specialized infrastructure. We are in the process of developing a much simpler and higher throughput method for the discovery of protein-binding fragments.
Figure 2. Binding of bead-displayed, low molecular weight fragments to tetrameric Streptavidin (SA). Left: cartoon highlighting the avidity-driven binding to a SA-POI (protein of interest) fusion protein. Center: A SA-binding fragment (A) and a non-binding control (B). Right: Flow cytometry plot showing the amount of fluorescence captured by beads displaying fragment A or fragment B after exposure to fluorescently labeled SA after a one-hour incubation and thorough washing.
The system relies on the fact that bead-displayed molecules can bind multimeric proteins tightly due to avidity effects even when their intrinsic affinity for the protein is weak. For example, the low molecular weight Fragment A shown in Fig. 2 binds to the tetrameric protein Streptavidin (SA) weakly (KD = 5 mM) in solution but binds SA stably when displayed on tiny (10 µm) TentaGel beads. When using fluorescently labeled SA, this interaction can be detected readily in a 384 well format using a common fluorescent plate reader. We are constructing novel fragment libraries on 10 µm TentaGel beads and screening them against a number of multimeric proteins, of which there are a large number in the human proteome. We are also extending the methodology to monomeric protein targets by constructing trimeric and tetrameric fusion protein derivatives of these proteins for use in assays (Fig. 2).
DNA-Encoded libraries for Fragment Expansion
We create novel DNA-encoded libraries (DELs) on TentaGel beads by split and pool solid-phase synthesis.1,2 These DELs are screened by incubation with a fluorescently labeled protein, followed by passing the beads through a fluorescence activated cell sorter (FACS) instrument gated to collect the bead that capture high levels of the fluorescent target (Fig. 3). The DNA encoding tags on these beads are amplified and deep sequenced, providing the predicted structures of the candidate protein ligands.
Figure 3. One bead one compound (OBOC) DNA-encoded libraries. A. About 1% of the sites on 10 µm TentaGel beads are modified to carry a “DNA headpiece” to which a forward PCR primer site is ligated. B. Split and pool synthesis is carried out. At each split, a chemical step and the ligation of an encoding oligonucleotide is performed. This is continued until the library is finished, at which time a bead-specific barcode (See C & D) and a reverse PCR primer site are added to the DNA chain. The result is a DNA-encoded OBOC library in which each bead displays many copies of a single small molecule. E. These bead libraries are screened by incubation with target and off-target proteins labeled with different colored fluors. Beads that display ligands that bind the target, but not the off-target, are collected using a FACS. The encoding tags on the “red but not green” beads are amplified and deep sequenced to reveal the structures of the predicted hits. Because bead screening has a high false positive rate, only compounds that appear in the hit pool on multiple beads (assessable due to the bead-specific bar code) are considered for further validation. These “redundant hits” are much more likely to be bona fide ligands than “singletons” in the hit pool.
Recently, we have established methods to make DELs of novel, cell permeable macrocycles in which the ring is closed by capture of a reversibly formed imine. This is done either by reducing the imine with NaCNBH3 or creating an imidazopyridinium heterocycle through reaction with a 2-formyl pyridine (Fig. 4). These synthetic workflows allow the facile incorporation of protein-binding fragments into the DEL as shown in Fig. 4. The resultant “biased DELs” can then be screened for high affinity ligands to the protein of interest.
Figure 4. Macrocycle libraries for fragment to lead discovery.
Literature Cited
1. Paciaroni, N.G., Ndungu, J.M., and Kodadek, T. (2020). An aldehyde explosion strategy for the synthesis of structurally complex DNA-encoded libraries. Chem. Comm. 56, 4656-4659.
2. Koesema, E., Roy, A., Paciaroni, N.G., Coito, C., Tokmina-Roszyk, M., and Kodadek, T. (2022). Synthesis and Screening of a DNA-Encoded Library of Non-Peptidic Macrocycles. Angew Chem Int Ed Engl, e202116999. 10.1002/anie.202116999.