What Do Advances in Structural Proteomics Mean for Drug Development?

During drug development, understanding how a compound binds to its target(s) is an important step in lead selection and optimization. A thorough dissection of drug–protein interactions helps to investigate structure-activity relationships and elucidate mechanisms of binding, which set key milestones in compound development.

But this important step is also fraught with challenges. This is because the current gold-standard techniques for interrogating drug-target interactions, including nuclear magnetic resonance (NMR), X-ray crystallography, cryo-electron microscopy (EM) and hydrogen-deuterium exchange (HDX), all have significant limitations.

These techniques, while instrumental for investigating protein structures and binding events, are resource-intensive, costly to run and ineffective for large or membrane-bound proteins. For example, NMR has a size limit of around 30 kilodaltons (kDa),¹ which is far below the size of most typical drug targets.

In most cases, proteins must be recombinantly purified and under some circumstances truncated or labeled in order to be compatible with these methods. This leads to laborious sample preparation and may introduce artifacts as the target protein is removed from its native environment.

In drug development, where a detailed understanding of the target and its biology is paramount, any incomplete or misleading information in the early stages can have major implications later down the road.

Innovations in structural proteomics

Thankfully, advances in structural proteomics techniques – which aim to characterize the structural properties of the proteome in situ – are providing a window to monitor these previously inaccessible proteins in a cellular milieu.

During my time at ETH Zurich, I worked on the development of a novel technique combining Limited Proteolysis (LiP) and quantitative mass spectrometry (MS) in the laboratory of Professor Paola Picotti.² LiP reveals protein structural changes through the exposure of the proteome to non-specific proteases and, by means of MS, peptides derived from these cleavage events are identified and quantified.

Thanks to further developments in quantitative MS pioneered by our team at Biognosys, an automated and integrated workflow was developed which enables the unbiased identification of small molecule drug targets in complex proteomes.³ Since 2018, LiP-MS has powered our next-generation proteomics platform TrueTarget™, supporting a wide range of companies from early-stage biotech to large pharma.

Importantly, TrueTarget can study proteins in a near-native environment, for example in cellular or tissue lysate. This offers a more representative view of in vivo protein structure compared with in vitro studies. It also allows for better reconstitution of cellular components such as ionic strength, chaperones and post-translational modifications, which are critical to protein function and drug efficacy.

Fundamentals and Advances in Mass Spectrometry Proteomics

After many years in the shadow of genomics, proteomics is now ready to step into the spotlight. Download this eBook to learn more about the fundamentals  of proteomics, advances in mass spectrometry proteomics and transformative proteomics insights from discovery to clinic. 

Read eBook

Sponsored Content

A new way to identify drug targets across the proteome

Indeed, one of the major areas for the application of TrueTarget is drug target deconvolution. The platform offers many advantages here, as it can directly identify the protein target(s) of a compound across the entire proteome without the need for prior labeling or modification.³

Using machine learning, we can also identify features suggestive of drug binding, which can be integrated into a single LiP score to identify and rank the most likely targets and their binding sites.

In a collaboration with AstraZeneca, our team applied this workflow to profile a CDK9 inhibitor.  The results of this study were presented at the Cancer Proteomics and Screening Technologies mini symposium at AACR 2021 and published in ACS Chemical Biology as a cover story in January 2022.⁴

Using the LiP score, we could clearly separate CDKs from other parts of the proteome. And, thanks to the dose-response curve, we could see that CDK9 bound more strongly than other CDKs to the inhibitor.

We could also make use of the peptide-level resolution of LiP-MS to investigate the binding site of the inhibitor across four targets, providing valuable information on the selectivity, the strength of target engagement and likely mechanism of action of the compound.

Introducing HR-LiP, a new method to dissect drug–protein interactions

Despite the capabilities of LiP-MS to pinpoint potential binding regions in an unbiased target deconvolution experiment, there are still challenges to achieving high protein sequence coverage for a portion of the proteome. This typically includes low-abundance proteins, large protein complexes and membrane proteins, most of which are also difficult to purify. 

To address this challenge, in collaboration with Cedilla Therapeutics we developed a novel workflow - High Resolution Limited Proteolysis (HR-LiP) – with a modified protocol designed to increase sequence coverage for a target of interest.

This further enhances the TrueTarget platform with additional capabilities for target validation. The protocol starts with overexpressing the protein of interest in any cell line of interest. Native cell lysates are then incubated with drug compounds at several concentrations. Following LiP and MS analysis, we use the resulting dose-response curves to identify binding events.

Last year at AACR, we presented examples of this workflow in action, investigating the binding of several inhibitors of two critical targets in oncology: gefitinib and afatinib for epidermal growth factor receptor (EGFR) and JQ1 for bromodomain-containing protein 4 (BRD4).⁵

Despite their importance in cancer research, both proteins have been mainly studied using recombinantly purified fragments due to their size. EGFR is a large (170 kDa) transmembrane protein and BRD4 is a 200 kDa multi-domain protein.

With the HR-LiP technology, however, we were able to achieve a sequence coverage of > 80% for both proteins, which were expressed functionally and in full length without the need for tagging or modifications. We also revealed new insights into the binding sites and mechanisms of action of the inhibitors that were not detectable with other methods.

The future of structure-based drug design?

Advances in structural proteomics, exemplified by new techniques like HR-LiP, allow us to monitor drug binding on full-length proteins in a cellular environment, at ~10 amino acid resolution. This creates new opportunities to study allosteric interactions and other types of binding that may induce conformational changes affecting protein stability.

Given its ease of implementation, without the need for recombinantly purified proteins, HR-LiP on our TrueTarget platform allows us to map drug–protein interactions for almost any target protein or complex of interest. This combination of high throughput and high resolution has huge benefits for target validation and lead optimization stages of drug development.

These advances are increasing time and cost-efficiency and provide a fit-for-purpose solution to prob drug–target interactions, accelerating and de-risking small molecule drug development. Further developments of the technology could also support the characterization of protein–protein and protein–antibody interactions, making mass spectrometry structural proteomics an indispensable part of the drug discovery pipeline.

References:

1. Gauto DF, Estrozi LF, Schwieters CD, et al. Integrated NMR and cryo-EM atomic-resolution structure determination of a half-megadalton enzyme complex. Nat Commun. 2019;10(1). doi: 10.1038/s41467-019-10490-9

2. Feng Y, De Franceschi G, Kahraman A, et al. Global analysis of protein structural changes in complex proteomes. Nat Biotechnol. 2014;32(10):1036-1044. doi: 10.1038/nbt.2999

3. Piazza I, Beaton N, Bruderer R, et al. A machine learning-based chemoproteomic approach to identify drug targets and binding sites in complex proteomes. Nat Commun. 2020;11(1):4200. doi: 10.1038/s41467-020-18071-x

4. Hendricks JA, Beaton N, Chernobrovkin A, et al. Mechanistic insights into a CDK9 inhibitor via orthogonal proteomics methods. ACS Chem Biol. 2021;17(1):54-67. doi: 10.1021/acschembio.1c00488

5. Nigel B, Adhikari J, Bruderer R, et al. 2136 - Prediction of small molecule-protein binding events for BRD4 and EGFR inhibitors using HR-LiP, a novel structural proteomics approach. Paper presented at AACR 2022; April 11, 2022; New Orleans.  https://www.abstractsonline.com/pp8/#!/10517/presentation/13180. Accessed March 09, 2023.

About the author

Yuehan Feng, PhD, is the director of Scientific Alliances at Biognosys, a global proteomics company that offers large-scale proteomics solutions based on proprietary mass spectrometry technology.