A direct-to-biology high-throughput chemistry approach to reactive fragment screening

Methods for rapid identification of chemical tools are essential for the validation of emerging targets and to provide medicinal chemistry starting points for the development of new medicines. Here, we report a screening platform that combines ‘direct-to-biology’ high-throughput chemistry (D2B-HTC) with photoreactive fragments. The platform enabled the rapid synthesis of >1000 PhotoAffinity Bits (HTC-PhABits) in 384-well plates in 24 h and their subsequent screening as crude reaction products with a protein target without purification. Screening the HTC-PhABit library with carbonic anhydrase I (CAI) afforded 7 hits (0.7% hit rate), which were found to covalently crosslink in the Zn2+ binding pocket. A powerful advantage of the D2B-HTC screening platform is the ability to rapidly perform iterative design–make–test cycles, accelerating the development and optimisation of chemical tools and medicinal chemistry starting points with little investment of resource.


HTC-PhABit
HTC-PhABit Similarity  Table S5. Smiles of the HTC-PhABits within the heat spot containing the sulfonamide hits. Numbering unique to the heatmap.

Solvents, reagents and consumables
Solvents were anhydrous and reagents purchased from commercial suppliers were used as received.
Protein stock solutions used: Low pH: Sample was eluted using a gradient shown in Table S6 with a flow rate of 1.0 mL/min.  High pH: Sample was eluted using a gradient shown in Table S7 with a flow rate of 1.0 mL/min. Solvent A (10 mM ammonium bicarbonate in water adjusted to pH 10 with ammonia solution) and solvent B (acetonitrile). Alternate-scan Positive and Negative Electrospray on a Waters SQD2 instrument, with a scan range of 100-1000 Da and a scan frequency of 5 Hz. A HpH method was used and the sample was eluted using a gradient shown in Table S7 with a flow rate of 1.0 mL/min. Solvent A (10 mM ammonium bicarbonate in water adjusted to pH 10 with ammonia solution) and solvent B (acetonitrile).

Mass directed automated preparative HPLC (MDAP)
Mass directed Autoprep HPLC was carried out on a Waters® Xselect instrument equipped with a CSH High pH: C18 column (150 mm × 30 mm, 5 μm packing diameter, 40.0 mL/min flow rate) using a gradient elution at ambient temperature using mobile phases of water with 10 mM ammonium bicarbonate adjusted to pH 10 with ammonia solution (solvent A) and acetonitrile (solvent B).
The gradient of acetonitrile required to elute product was determined by the LC-MS retention time.
The methods were selected dependent on the retention time of desired material and are shown below (Table S8) and an exemplar gradient for Method B (Table S9). Low pH: Gradient elution using mobile phases of water with 0.1% v/v solution of formic acid in water (solvent A) and 0.1% v/v solution of formic acid in acetonitrile (solvent B) (Table S10).
High pH: Gradient elution using mobile phases of water with 10 mM ammonium bicarbonate adjusted to pH 10 with ammonia solution (solvent A) and acetonitrile (solvent B) (Table S11). Da with a scan rate of 1.03 s in positive mode. The following MS parameters were used: capillary voltage limit -4200 V; desolvation temperature -340 °C; drying gas flow -8.0 L/min. the summed scans were deconvoluted (using a maximum entropy algorithm) over a m/z range with an expected mass range dependent on the protein. The deconvoluted spectra were exported as csv files and analysed using R Studio software (Version 3.6.3) to generate pdf files of the spectra. The median of the protein only controls were subtracted from the sample spectra to remove baseline signal. The peak height for unmodified protein and labelled protein were recorded and used to calculate percentage photocrosslinking using the equation:

Centrifuge
Plates were centrifuged using a Sorvall Legend RT (401198833) model at 1000 rpm for 1 minute.  150<MW<250, BioByte cLogP<9 and only 1 aliphatic amine group. 2 The structures were tagged based on 6 aliphatic amine types, differentiating based on primary/secondary, hindered/non-hindered and cyclic/non-cyclic. The amines were also tagged by the calculated pKa using ChemAxon pKa. 3 All amines having acidic pKa<7 and amines with secondary hindered linear aliphatic amines were excluded. A ChemAxon LibMCS clustering was performed on the remaining ca. 4000 molecules. 4 Since the aim of the current selection was to choose compounds which can have analogues, all compounds were excluded which did not belong to a cluster with at least 4 members. The remaining ca.   (Table S15). The crosslinking vs concentration for each HTC-PhABit was plotted on GraphPad Prism 5.0.4, each data point was recorded in triplicate and the mean and standard deviation was calculated and plotted. In some instances, for comparison, crosslinking yields were normalised to allow comparison.

Concentration-response with purified PhABits
The five purified sulfonamide hits ( The crosslinking vs concentration for each HTC-PhABit was plotted on GraphPad Prism 5.0.4, each data point was recorded in triplicate and the mean and standard deviation was calculated and plotted. In some instances, for comparison, crosslinking yields were normalised to allow comparison.