One Question, Every Target:

The Evolution of Small-Molecule Drug Discovery

Promega Corporation

Publication Date: July 2026

Introduction

A drop hits still water. The first ring moves outward from the point of contact, then another follows, then another, each one carrying energy further from the center, each one expanding the circle of what the water touches. From a distance, the history of target-based small molecule drug discovery looks like these expanding ripples: One question, set in motion decades ago, expanding outward in rings that haven't stopped.

The question, at its core, has never changed: does this compound actually engage its target, in a living cell, in a way that matters? For a time, the tools and the targets aligned: methods were established, evidence was actionable and programs moved forward with confidence. That alignment held only as long as the tools could keep pace. As biology became better understood and researchers set their sights on harder, less tractable targets, the same question demanded more evidence to answer.

This is a story about expansion. Each new generation of targets added to the field's work without erasing what preceded it. The programs from the earliest years are still running. The work from the middle years is still being carried forward. The frontier exists at the outermost edge of a field that contains all of it simultaneously, and we will always be looking just beyond where the ripple has reached.

The Question Takes Shape

Target engagement, cellular context and the limits of biochemical assays

In the first generation of target-based drug discovery, the field built its foundation on proteins it could approach: albuterol, captopril, imatinib. A generation of therapies were built on the simple premise that if you could find the protein driving the disease, you could make something that fits its binding site and modulates its activity. The targets shared something critical: structurally defined binding sites shaped for small molecules. An assay ecosystem grew up around them. The tools defined what was knowable, and what was knowable was precise enough to move programs forward with confidence (1). For a long time, it held.

What that era also revealed, through accumulating cellular evidence across programs, was that the logic had a boundary. A compound that perfectly inhibited a purified enzyme in a test tube was not necessarily doing the same thing inside a living cell. The cellular environment changed everything: competing proteins, metabolic context, membrane dynamics, the sheer crowded complexity that biochemical assays were designed to exclude.

The lesson wasn't that biochemical assays were wrong. They were right about what they measured. The limitation was that what they measured stopped being sufficient when the biology moved past the horizon they were designed to see.

smdd-foundation-story-established-circles

Biochemical assays didn't become obsolete when their limits were recognized. They're still the workhorse of early drug discovery, used every day because they're fast, scalable and precise about what they measure. But the gap between what the assay shows and what the cell does never fully closed. It still shows up in unexplained selectivity differences, compounds that look clean in biochemical panels and behave unexpectedly in cells (2).

The cell isn't a noisy version of the biochemical assay—it's a different experiment.

The goal at the center of target-based drug discovery has never changed. Reach the right target, in a biologically relevant context, and engage it selectively and long enough to produce the desired effect. What researchers kept discovering, at every stage, was that confirming that level of specificity demanded more: different tools, higher standards of evidence. Matt Robers, a senior scientist at Promega who has spent years building cellular target engagement tools, frames the shift precisely: "The cell isn't a noisy version of the biochemical assay—it's a different experiment. Inside a cell, your target is embedded in biomolecular complexes, often localized to specific compartments. A compound that shows modest activity against an isolated protein but potently stabilizes an intracellular complex in cells isn't a false positive; it's telling you the complex is the real drug target."

The question hadn't changed, but what it took to answer it had.

The Question Expands

GPCR drug discovery, biased agonism and multi-dimensional target engagement

Drug discovery programs kept evolving as compounds grew more selective, cellular assays became more sophisticated and understanding of the biochemical-to-cellular gap matured. Researchers were also discovering, across target classes, that some of the biology they had been successfully drugging was more complex than the tools had been designed to capture. One of the clearest examples came from GPCR signaling. As researchers probed GPCR biology more deeply through the 1990s and into the 2000s, they began to recognize that the same receptor could activate fundamentally different downstream pathways depending on which ligand was bound, and that each pathway could produce distinct physiological outcomes (3). A compound could look right by every assay in the established playbook, a potent binder with strong G-protein activation and clean selectivity, and still behave in ways those measures didn't predict.

Confirming a compound was active was no longer enough. Researchers had to resolve which pathways it engaged, and in what balance.

GLP-1 receptor agonists had been in development for years before tirzepatide. Earlier drugs in the class were optimized around G-protein activation, the standard measurable signal, and they reached approval on that basis. Then tirzepatide arrived with a clinical profile that outperformed earlier GLP-1R agonists. When researchers later characterized how tirzepatide engaged the GLP-1 receptor, they found it favored G-protein signaling over β-arrestin recruitment. Measuring cAMP, the standard proxy for G-protein activation, wouldn't reveal that difference on its own. Biased signaling has been proposed as one contributor to the drug's profile, a question still being actively investigated (4–6).

Drug discovery researchers hadn't been measuring the wrong thing. They had been measuring what their tools were designed to measure. What the characterization of tirzepatide raised was a question the standard assay wasn't designed to answer: not whether the receptor was active, but which pathway it preferentially engaged. That raised a more demanding standard. Confirming a compound was active was no longer enough. Researchers had to resolve which pathways it engaged, and in what balance.

That distinction is increasingly shaping how next-generation GLP-1 receptor drugs are evaluated. Knowing a compound engaged its target in a living cell was only the starting point. Researchers also had to understand which aspect of engagement, through which pathway, toward which outcome.

Measurement had shifted from a go/no-go gate to a design tool.

The Question Changes Form

KRAS, covalent inhibitors and targeted protein degradation

For thirty years, KRAS was the most consequential undruggable target in oncology, mutated in roughly one in five human cancers and structurally resistant to every conventional approach. Oncogenic mutations impair the protein's ability to return to its inactive state, keeping downstream signaling switched on. KRAS lacks the well-defined binding pocket that small molecules typically exploit, and competing with the high intracellular concentrations of GTP that keep it active isn't a tractable strategy. Then, in 2013, the Shokat laboratory identified a hidden allosteric pocket in the G12C mutant. A new class of covalent inhibitors found it, bonded permanently to the mutant cysteine, and locked the protein before it could cycle back to active. Sotorasib received FDA approval in 2021. Eight years from cryptic pocket to approved drug (7–9).

What KRAS and targets like it had in common was this: the biology was well-characterized. Researchers knew the protein mattered, knew why and had a working sense of what a successful drug would need to accomplish. The barrier was chemistry. Drug hunters needed new ways to reach targets they could already see clearly.

Confirming that a covalent drug was doing its job required different tools than measuring enzyme inhibition. Researchers needed to verify that the compound was reaching and modifying the intended protein in living cells, that covalent occupancy was achieved at therapeutically relevant concentrations and sustained long enough to matter. Answering the question now required a different standard: not just how potent the compound is, but how completely does it engage, and how durably, in the relevant cellular context.

Covalent inhibition had extended drug discovery research into targets once considered structurally inaccessible. The problem it couldn't solve was what to do about proteins where binding, even permanent binding, wasn't enough. Researchers needed to work with proteins that function through scaffolding or interactions, where occupying a pocket is insufficient because the disease is driven by the protein's presence instead of enzymatic activity.

One response was to stop trying to inhibit these proteins and start trying to destroy them. Targeted protein degradation redirected the cell's own disposal machinery toward targets it would never naturally recognize. Rather than occupying a target continuously, degraders recruit an E3 ubiquitin ligase to tag the protein of interest for proteasomal destruction. This event-driven mechanism is catalytic in nature: a single degrader molecule can induce the degradation of multiple target proteins, achieving profound and sustained target knockdown at concentrations insufficient for conventional inhibitors (10). The protein isn't inhibited. It's eliminated.

The field needed tools to interrogate each mechanistic event in living cells, and to observe those processes kinetically.

This different mechanism demanded an entirely new measurement strategy. Does the ternary complex form between drug, E3 ligase and target? Does the target get tagged for degradation efficiently? How fast does the target degrade, at what concentration and in which cell type? How quickly does it recover?

Kristin Riching, a senior research scientist at Promega who has spent years developing tools to characterize these processes, describes the core challenge: "Every step of the mechanism can fail independently, and failure at any step can look identical if the only readout is target level at a fixed timepoint. The field needed tools to interrogate each mechanistic event in living cells, and to observe those processes kinetically."

These were not refinements of existing assays. They required new frameworks built around dynamic cellular processes, and they reinforced a theme researchers kept rediscovering: as modalities diversified, so did the measurements needed to understand them.

The Question Without a Method

AI drug discovery, undruggable targets and the frontier measurement challenge

Until recently, even the hardest targets arrived with the biology understood. Advances in genomics, proteomics and artificial intelligence have changed that, nominating targets at a pace and scale the field has never seen, in biology that hasn't been characterized.

What those tools can't provide is a method.

Drug discovery researchers have long had a list of targets they desperately want to reach but haven't found a reliable way to approach. Although the MYC protein has been implicated in roughly seventy percent of human cancers for more than four decades, it has remained a largely intractable therapeutic target (11). Alpha-synuclein drives Parkinson's disease pathology and has resisted drug development for over twenty-five year (12). Increasingly, the targets arriving through AI-driven approaches come with less biological context than even MYC or alpha-synuclein, proteins the field has studied for decades. Researchers may understand the association with disease without yet understanding the mechanism, the structure or what a meaningful drug would need to accomplish.

Many of the targets arriving through these approaches don't resemble anything the established playbook was built for. Intrinsically disordered proteins have no stable three-dimensional structure to design against. Transcription factors regulate gene expression through large, featureless surfaces rather than defined binding pockets. Scaffolding proteins drive disease through their organizational role rather than any catalytic activity that could be blocked. These are proteins researchers have long recognized as biologically important and pharmacologically intractable (13).

The measurement challenge they pose is qualitatively different from what came before. With biased signaling, researchers at least knew what receptor they were working with and had assay infrastructure to build on. With degraders, the target was often well-characterized even if the modality was new. With these frontier targets, researchers may be starting with a protein that has no tool compounds, no structural data in a relevant conformation, no validated biomarker of engagement and no established cellular model in which to measure effect. The question isn't which assay to run. It's how to design a measurement strategy from scratch for a target class that has never been successfully drugged.

Compelling targets are often prioritized despite limited tractability—the genetic evidence is strong, but there are no logical starting points for assay design.

Robers, who works directly on these frontier measurement challenges, describes what the field is navigating: "Compelling targets are often prioritized despite limited tractability—the genetic evidence is strong, but there are no logical starting points for assay design. What's emerging in response is a pivot toward function-agnostic target engagement methods that can exploit higher-order protein complexes."

By working with targets in their native cellular context, researchers can access interaction surfaces and cooperative binding phenomena that purified protein systems simply don't reveal."

At the frontier, the work starts from scratch. The target is nominated but not fully understood. The measurement framework doesn't exist. The question the researcher needs to ask—is my compound actually engaging this target, in a living cell, in a way that matters—is the same question the field has always been asking. The difference is that there is no established method for answering it, and in some cases no established understanding of what a meaningful answer would even look like.

Whether these targets will yield, no one knows. What the field does know is that it has been wrong about intractability before.

Conclusions

This is not a story about inevitable failure. The targets that resisted every approach weren't dead ends. They were the moments that forced the field to invent what came next. Instead, it is a story about expansion. Programs keep running, modalities join rather than replace and the whole field continues growing, every ring still in motion. Small molecule drug discovery doesn't move in a line. It expands in all directions from a central question that never stops sending energy outward.

The field has always moved forward the same way: by discovering the edge of what its tools could see, naming it and building what came next. Every researcher who has ever stood at that edge—uncertain, without a playbook, asking a question the current methods couldn't fully answer—has been at the outermost edge of something that never stops moving.

That pattern isn't just history. It's a lens for wherever you are. Every program carries a measurement gap that only reveals itself when the targets move beyond what the tools were built to capture. The kinase researcher whose biochemical selectivity doesn't translate to cellular selectivity is living inside one version of it. The degrader team whose ternary complex data doesn't translate to cellular degradation is living inside another. The frontier researcher who can't yet define what a meaningful answer would look like is living inside the hardest version yet. It doesn't tell you where your gap is. It tells you that one exists, and that naming it is how the field has always begun to solve it.

Drug discovery keeps moving. Your targets stay in reach.