Targeted covalent inhibitors (TCIs) first form a noncovalent complex with the target protein and then form a covalent bond with a specific amino acid residue near the binding pocket. Their development focuses on increasing target occupancy and duration of action while controlling warhead reactivity and proteome-wide off-target risk.

Keywords: targeted covalent inhibitors; covalent drugs; covalent warheads; cysteine; reversible covalent inhibition; irreversible covalent inhibition; covalent kinetics; ABPP; chemical proteomics; drug design

1 Basic Concepts of Covalent Drugs and Targeted Covalent Inhibitors

1.1 Covalent Drugs and Targeted Covalent Inhibitors

Covalent drugs are compounds that produce pharmacological effects by forming covalent adducts with biological macromolecules, whereas targeted covalent inhibitors are a subgroup developed primarily to inhibit the function of a specific target protein. Some compounds contain electrophilic groups that can directly participate in covalent reactions, while others require metabolic conversion or activation in a specific chemical environment to generate reactive intermediates. Not all covalent drugs are conventional inhibitors, and some compounds can alter protein conformation, ligand affinity, or biophysical properties through reversible covalent binding.

1.2 Structural Components of Targeted Covalent Inhibitors

Targeted covalent inhibitors generally consist of a recognition scaffold, a linker region, and a covalent warhead. The recognition scaffold positions the compound in the target-protein pocket through hydrogen bonding, hydrophobic interactions, electrostatic interactions, and shape complementarity; the linker region determines the distance, orientation, and conformational freedom between the covalent warhead and the target residue; and the covalent warhead forms a covalent bond with cysteine, serine, threonine, or another reactive residue. An ideal covalent inhibitor has limited reactivity in its free state, and the warhead reacts only after formation of an appropriate noncovalent complex, when a high local effective concentration and correct spatial orientation are achieved.

1.3 From Empirical Covalent Drugs to Rational Design

Early covalent drugs such as Aspirin, Penicillin, Fosfomycin, and Omeprazole were not discovered according to modern targeted covalent drug-design principles, and their covalent mechanisms were gradually elucidated after their introduction into clinical use. With advances in protein structure determination, mass spectrometry, covalent kinetics, and chemical proteomics, covalent drug development has shifted toward rational design based on target-protein structure, reactive-residue distribution, and covalent-warhead properties and has expanded to protein kinases, proteases, proteasomes, metabolic enzymes, small GTPases, and other targets.

2 Reaction Mechanisms and Kinetic Basis of Covalent Inhibition

2.1 Two-Step Reaction Mechanism

Targeted covalent inhibition generally involves two consecutive steps: noncovalent recognition and covalent bond formation:

E + I ⇌ E·I → E–I

Here, E represents the target protein, I represents the inhibitor, E·I represents the noncovalent complex, and E–I represents the covalent complex. The first step determines whether the compound can enter the target-protein pocket and adopt the correct binding conformation, whereas the second step determines the rate and stability of covalent modification. Increasing warhead electrophilicity alone cannot substitute for high-quality noncovalent recognition, and highly reactive compounds lacking targeted positioning are more likely to cause extensive nonspecific protein modification.

2.2 Irreversible and Reversible Covalent Inhibition

(1) Irreversible Covalent Inhibition

Irreversible covalent inhibitors form relatively stable covalent bonds with their target proteins, and the inhibitory effect generally does not rapidly disappear as the free-drug concentration declines before the target protein is degraded or resynthesized. Acrylamides, chloroacetamides, epoxides, epoxyketones, and certain sulfonyl fluoride groups can be used for irreversible covalent design. Their duration of action is jointly affected by covalent-bond stability, target-protein turnover, and cellular compensatory mechanisms.

(2) Reversible Covalent Inhibition

The adducts formed by reversible covalent inhibitors can dissociate through a reverse reaction. Common warheads include nitriles, boronic acids, aldehydes, ketones, and certain activated carbonyl groups. This strategy can strengthen binding and prolong target occupancy while reducing the probability of permanent modification of nontarget proteins, although selectivity still depends on adduct-dissociation rate, tissue exposure, metabolic stability, and the range of proteome reactivity.

2.3 Kinetic Parameters of Covalent Inhibition

The potency of irreversible covalent inhibitors should not be evaluated solely using a single IC₅₀ value. K_I mainly reflects formation of the noncovalent complex before covalent bond formation, k_inact represents the maximum rate of the covalent inactivation step, and k_inact/K_I integrates target-recognition capacity and covalent bond-forming efficiency. Because IC₅₀ varies with preincubation time, enzyme concentration, substrate concentration, and assay conditions, compounds should be compared under standardized conditions and evaluated together with target occupancy and the recovery of activity after washout.

Table 1 Major Differences Between Noncovalent Inhibitors and Targeted Covalent Inhibitors

3 Selection of Targetable Amino Acid Residues

Figure 1 Structures of representative amino acid residues targeted by covalent inhibitors

3.1 Cysteine, Serine, and Threonine

(1) Cysteine

The thiol group of the cysteine side chain is strongly nucleophilic and is the most widely targeted residue in covalent drug design. Its actual reactivity is affected by local pKa, hydrogen-bonding networks, solvent accessibility, surrounding charge, and protein conformation. Cysteines located near a functional pocket, capable of adopting a suitable spatial orientation toward the warhead, and less conserved among homologous proteins are more suitable as selective covalent sites.

(2) Serine

The hydroxyl group of the serine side chain has lower intrinsic nucleophilicity than the cysteine thiol, but it can be activated by neighboring acid-base residues when located in the catalytic centers of serine proteases, penicillin-binding proteins, and other hydrolases. β-Lactams, boronic acids, phosphates, and activated carbonyl groups can be used as warheads for targeting catalytic serine residues.

(3) Threonine

The hydroxyl group of the threonine side chain can also participate in covalent reactions, and the N-terminal threonine of proteasome catalytic subunits is a representative covalent drug target. This residue contains both an N-terminal amino group and a side-chain hydroxyl group and can form reversible or irreversible covalent adducts with boronic acids, boronate esters, epoxyketones, and related warheads.

3.2 Lysine, Tyrosine, and Other Reactive Residues

(1) Lysine and Protein N-Terminal Amino Groups

The ε-amino group of lysine side chains and the α-amino group at protein N termini can react with aldehydes, activated esters, sulfonyl fluorides, fluorosulfates, and related electrophilic groups. Lysine is generally protonated, and its reactivity is highly dependent on the local hydrophobic environment, neighboring residues, and hydrogen-bonding network. The chemical environment of a protein N-terminal amino group differs from that of a lysine side chain and can sometimes provide greater positional selectivity.

(2) Tyrosine and Histidine

The phenolic hydroxyl group of tyrosine can display increased nucleophilicity in a locally basic environment and react with environmentally sensitive electrophilic groups. The imidazole group of histidine can participate in nucleophilic reactions and metal coordination, but covalent drug development targeting histidine is less mature than that targeting cysteine or serine and generally requires precise positioning within the binding pocket to ensure selectivity.

(3) Aspartate and Glutamate

The side-chain carboxyl groups of aspartate and glutamate can form covalent adducts in specific catalytic environments or after chemical activation. However, because carboxylate residues are widely distributed in proteins and are generally negatively charged, achieving selective design is difficult.

(4) Methionine and N-Terminal Proline

The thioether group of the methionine side chain can be selectively modified using certain oxidative reagents or hypervalent iodine reagents and is currently used mainly in chemical-probe and protein-labeling studies. The proline side chain itself does not contain an independent nucleophilic group, but when proline is located at the protein N terminus, its cyclic secondary amine can serve as a covalent reaction site.

Table 2 Representative Amino Acid Residues Targetable by Covalent Ligands

4 Selection and Optimization of Covalent Warheads

4.1 Principles for Covalent Warhead Selection

The covalent warhead determines which residue a candidate compound can react with, the rate of covalent bond formation, adduct stability, and reaction reversibility. Warhead selection should comprehensively consider the target residue, local microenvironment, desired duration of action, chemical stability, metabolic pathways, and proteome selectivity. When the same warhead is attached to different recognition scaffolds or different positions, its actual reactivity, target occupancy, and off-target profile may differ substantially. Therefore, a warhead should not be evaluated independently of the complete molecule.

4.2 Control of Warhead Reactivity

Insufficient warhead reactivity may lead to slow intracellular bond formation and inadequate target occupancy, whereas excessive reactivity may cause extensive reactions with glutathione, albumin, and nontarget proteins. Structural optimization can control bond-forming efficiency by adjusting electron-withdrawing groups, steric hindrance, linker length, conformational restriction, and warhead orientation, so that the covalent reaction depends as much as possible on the high local effective concentration provided by the target-protein pocket rather than on high intrinsic compound reactivity.

Table 3 Common Covalent Warheads and Their Design Characteristics

5 Advantages and Development Limitations of Targeted Covalent Inhibitors

5.1 Major Advantages

(1) Increased Target Occupancy

After covalent bond formation, a compound can maintain high target occupancy at a relatively low free concentration and is suitable for enzymes, receptors, and signaling proteins requiring sustained inhibition.

(2) Prolonged Duration of Action

The pharmacological effect of an irreversible covalent inhibitor can outlast its plasma exposure. Recovery of pharmacological activity depends more strongly on target-protein degradation and resynthesis than on dissociation of the drug from the binding pocket.

(3) Use of Distinctive Residues to Expand the Selectivity Window

When a reactive residue near the target-protein pocket is absent, conformationally distinct, or differently positioned in homologous proteins, covalent bond formation can be used to improve selectivity for the target protein.

(4) Enhanced Binding to Shallow or Transient Pockets

Covalent bond formation can compensate for insufficient noncovalent binding energy and provide opportunities to develop ligands for certain shallow pockets, transient pockets, and mutation-specific pockets, although the target must still possess a recognizable structure and an appropriate residue.

5.2 Major Limitations

(1) Nonspecific Covalent Modification

Compounds with excessive reactivity or insufficient spatial positioning may modify glutathione, plasma proteins, and nontarget proteins, increasing cytotoxicity, organ toxicity, and pharmacokinetic uncertainty.

(2) Reactive Metabolites

Some candidate compounds may generate more reactive electrophilic intermediates during metabolism and form protein adducts unrelated to the intended target.

(3) Immune-Related Risks

Covalent adducts formed between drugs and proteins may generate new antigenic epitopes, but the actual risk depends on adduct abundance, tissue distribution, carrier-protein type, exposure duration, and individual immune background.

(4) Resistance

Mutation of the target residue can directly prevent covalent bond formation, while binding-pocket conformational changes, target-protein amplification, drug efflux, rapid target-protein turnover, and bypass signaling activation may also reduce drug effects.

6 Discovery Methods for Covalent Hit Compounds

6.1 High-Throughput Screening

High-throughput screening is suitable for identifying candidates from collections of known covalent inhibitors, electrophilic compounds, and bioactive small molecules containing potential covalent warheads. Because covalent inhibition usually exhibits clear time dependence, screening systems should control preincubation time, substrate concentration, enzyme concentration, and reducing-agent conditions and should avoid evaluating compound potency solely on the basis of an IC₅₀ value measured at a single time point.

6.2 Fragment-Based Drug Discovery

Fragment-based drug discovery (FBDD) uses low-molecular-weight, structurally simple fragments to cover a broad chemical space. Covalent fragments can use weak noncovalent recognition to position a warhead near a specific residue and can then be optimized through structural extension, fragment merging, and conformational restriction to increase binding. The discovery of BTK covalent inhibitors such as TAK-020 demonstrates that covalent fragments can serve as starting points for highly selective lead compounds, although fragment hits still require modification-site confirmation, kinetic analysis, and proteome-selectivity evaluation.

6.3 DNA-Encoded Libraries

DNA-encoded libraries (DELs) can screen tens of millions to billions of chemical structures using relatively small amounts of experimental material. Covalent DELs require synthetic reactions and warhead structures that are compatible with DNA tags, and competition screening, mutant-protein screening, no-target controls, and nonreactive analogs are required to distinguish genuine targeted binding from nonspecific chemical reactions.

6.4 Activity-Based Protein Profiling

Activity-based protein profiling (ABPP) uses chemical probes containing a reactive group and a reporter group to label proteins with specific catalytic activities or nucleophilic properties. In competitive ABPP, cells, tissues, or proteomes are first treated with a candidate covalent compound and then exposed to a broad-spectrum probe. Decreased probe labeling is used to identify potential targets and off-target proteins. This approach can therefore be used for unknown-target discovery and for evaluating intracellular target occupancy and proteome selectivity.

7 Evaluation Methods for Covalent Inhibitors

7.1 Kinetic and Biochemical Activity Evaluation

Biochemical evaluation of covalent inhibitors should determine whether inhibition increases with prolonged preincubation and should use data from multiple concentrations and time points to fit K_I and k_inact. Recovery of enzymatic activity after washout, dilution, or dialysis can help distinguish reversible noncovalent inhibition, reversible covalent inhibition, and irreversible covalent inhibition. For substrate-competitive covalent inhibitors, substrate concentration should also be controlled to avoid misinterpreting substrate-competition effects as differences in covalent bond-forming efficiency.

7.2 Mass Spectrometric and Structural Analysis

Intact-protein mass spectrometry can determine whether the target protein exhibits the expected mass increase and can identify the number of adducts formed per protein molecule; protein digestion followed by LC-MS/MS can localize the specific modified residue and determine whether multiple sites are modified; and X-ray crystallography, cryo-electron microscopy, and nuclear magnetic resonance can further confirm the binding mode of the recognition scaffold, warhead orientation, target-residue conformation, and bond-forming geometry. For reversible covalent adducts, denaturation, reduction, digestion, and chromatographic conditions must be optimized to prevent adduct dissociation during sample processing.

7.3 Target Occupancy and Proteome Selectivity

Intracellular target occupancy can be evaluated using competitive probes, click-chemistry probes, targeted mass spectrometry, and cellular thermal shift analysis and should be analyzed together with downstream signaling inhibition and cellular phenotypes. Competitive ABPP, quantitative chemical proteomics, and adduct proteomics can systematically identify nontarget proteins, whereas glutathione reactivity, plasma stability, liver microsomal metabolism, and reactive-metabolite trapping can be used to assess intrinsic warhead reactivity and metabolism-related risks.

Table 4 Major Evaluation Methods for Covalent Inhibitors

8 Mechanisms of Representative Covalent Drugs

8.1 Aspirin and β-Lactam Drugs

Aspirin acetylates a serine residue in the cyclooxygenase active site; in COX-1, this modification blocks arachidonic acid access and persistently inhibits prostaglandin and thromboxane synthesis, whereas acetylated COX-2 retains altered oxygenase activity and can generate products such as 15R-HETE. β-Lactam antibiotics mimic bacterial peptidoglycan substrates and form stable acyl-enzyme intermediates with active-site serine residues in penicillin-binding proteins, thereby blocking peptidoglycan crosslinking and cell-wall synthesis.

8.2 Covalent Inhibitors of BTK and EGFR

Ibrutinib first binds the ATP-binding pocket of BTK and then forms a covalent bond with Cys481 through its acrylamide warhead. Afatinib, Dacomitinib, Neratinib, and Osimertinib use cysteines located near the ATP-binding pockets of ERBB-family kinases to achieve irreversible inhibition. The selectivity of these compounds depends not only on the target cysteine but also on the binding preference of the recognition scaffold for different kinase conformations and mutation states, whereas mutation of the target residue may reduce covalent bond-forming efficiency and generate resistance.

8.3 Covalent Proteasome Inhibitors

Proteasome catalytic subunits use N-terminal threonine as a nucleophilic residue. The boronic acid group of Bortezomib forms a reversible covalent tetrahedral adduct with this threonine, whereas the epoxyketone group of Carfilzomib forms a more stable irreversible adduct with the N-terminal threonine. These compounds demonstrate that different warheads can be used to regulate covalent reaction rate, stability, and reversibility against the same target residue.

8.4 KRAS G12C Covalent Inhibitors

The KRAS G12C mutation introduces a cysteine at position 12 and provides a specific site for mutation-selective covalent drug design. Sotorasib and Adagrasib bind the Switch II pocket of KRAS G12C and form an irreversible covalent bond with Cys12, biasing KRAS toward an inactive state. KRAS reactivation, enhanced upstream receptor signaling, bypass-pathway activation, and secondary mutations can still reduce the durability of these drugs.

8.5 Reversible Covalent Inhibitors of Viral Proteases

Boceprevir forms a reversible covalent adduct between its ketoamide warhead and the catalytic serine of the HCV NS3/4A protease, whereas Nirmatrelvir uses a nitrile group to form a reversible covalent adduct with Cys145 of the coronavirus main protease Mpro. These designs exploit the high nucleophilicity of protease catalytic residues to strengthen binding while retaining a degree of dissociation capacity.

9 Products Related to Covalent Drugs and Covalent Inhibitor Research

References

[1] Zheng L, et al. Development of Covalent Inhibitors: Principle, Design and Application in Cancer. MedComm – Oncology. 2023;2:e56.

[2] Huang F, et al. Covalent Warheads Targeting Cysteine Residue: The Promising Approach in Drug Development. Molecules. 2022;27(22):7728.

[3] Bon M, et al. Fragment-Based Drug Discovery—The Importance of High-Quality Molecule Libraries. Mol Oncol. 2022;16(21):3761-3777.

[4] Sabat M, et al. Discovery of the Bruton's Tyrosine Kinase Inhibitor Clinical Candidate TAK-020 (S)-5-(1-((1-Acryloylpyrrolidin-3-yl)oxy)isoquinolin-3-yl)-2,4-dihydro-3H-1,2,4-triazol-3-one, by Fragment-Based Drug Design. J Med Chem. 2021 Sep 9;64(17):12893-12902.

For more related articles, please see below:

[1] Small Molecule Inhibitors Selection Guide