By Panagiotis L. Kastritis, PhD

To function correctly, cellular proteins depend not only on their complex tertiary and quaternary structures, but also on their proximity to and interactions with other molecules participating in the same metabolic or enzymatic pathway. Characterizing such topography and interactions is vital to a full understanding of cell function and dysfunction.

But visualizing and characterizing native protein-protein interactions remains a challenge, not least because of the complexity and size of protein structures within the cell. To date, most insights into protein function have come from analytical techniques that destroy higher order protein structure and disrupt the spatial arrangements that are vital for effective function.

Fortunately, a new era for cell and structural biology has dawned. Novel sample preparation techniques coupled with the advent of cryogenic electron microscopy (cryo-EM) and advances in complementary techniques, such as mass spectrometry (MS), mean it is now possible to visualize and interrogate large protein complexes in a near-native state. Already, technological and computational advances have produced an explosion of data on protein structure and function.

New integrative methods are enabling a holistic approach to structural biology. Consequently, transient scaffolds and subunits that used to be undetectable can now be revealed, improving our knowledge of the cell’s ultrastructure and offering new avenues for drug discovery.

A new perspective on protein architecture

Cell architecture and protein stoichiometry are vital for correct cell functioning, Yet the mechanical and chemical methods commonly used to isolate cellular proteins disrupt a myriad of finely controlled interactions between protein subunits and complexes. Research to explore catalytic pathways is often reductionist—identifying individual constituents before adding purified components together, mostly after heterologous expression, to test and monitor interactions.

A novel, more holistic approach provides a new perspective. It uses cell extracts to keep protein complexes intact and in natural proximity to other proteins, allowing scientists to study higher order protein architecture and organization. The extracts can be investigated using a range of cutting-edge visualization, analytical, and functional assays, many of which allow proteins to be studied in a native state.

Data generated using innovative techniques, such as cryo-EM to visualize proteins and crosslinking MS (XL-MS) to identify and characterize them, is complemented by information from functional assays and advanced computational biology to reveal previously unseen details of intricate protein-protein interactions within the cell. This integrated approach, further fueled by technological and digital automation, is pushing forward the frontiers of cell biology and classical structural biology.

Novel structural and nonstructural elements revealed

The benefits of this integrative structural biology approach can be illustrated by the pyruvate dehydrogenase complex (PDHc), which is a giant, 10-megadalton enzymatic assembly, ubiquitous in eukaryotic cells, that converts pyruvate to acetyl coenzyme A. Multiple PDHc components have been characterized in isolation and localized to the mitochondrial inner membrane–matrix interface, but the complex’s quaternary structure has remained elusive due to its sheer size, heterogeneity, and plasticity.
Paradoxically, previous studies have identified structural components overlaying the pyruvate binding site, raising the question of how the substrate enters and effectively binds at the site.

Recently, investigators have taken a more holistic approach to better understand the structural organization of PDHc (Figure 1). They enriched cell extracts from a thermophilic fungus and employed a range of analytical assays, including MS and cryo-EM, to reveal, for the first time, the active structure of the ubiquitous PDHc.1

Figure 1. Diverse biochemical and biophysical methods were combined to probe a native cell extract of the thermophilic fungus Chaetomium thermophilium and investigate the structural organization of a-keto acid dehydrogenase complexes. This integrative structural biology approach enabled researchers to elucidate the spatial relationships between the protein components in the dynamic assembly that is the pyruvate dehydrogenase complex (PDHc). They also, for the first time, identified enzyme clusters that form a transient catalytic nanocompartment that is known as the “pyruvate dehydrogenase factory.” (Cell Rep. 2021; 34: 108727.)

Large protein complexes, isolated from cell lysates using size-exclusion chromatography (SEC), were visualized at the atomic level using cryo-EM and their composition determined using MS techniques. Chemical XL-MS, which uses chemicals to stabilize protein-protein interactions ahead of MS analysis, helped pinpoint interactions in the protein assemblies.

Linking the resulting MS data to molecular signatures identified in cryo-EM micrographs has revealed a multitude of structures, including the fatty acid synthase metabolon, double- and single-membrane structures, liposomes with encapsulated biomolecules, and other higher order complexes.2

This integrated approach has provided new insights into the three-dimensional structure of the PDHc binding site and proved for the first time that it is accessible to the pyruvate substrate. This approach also unveiled an asymmetric protein configuration directly involved in pyruvate oxidation.1

The big reveal came when a giant nanocompartment was visualized for the first time—a transient catalytic chamber called the “pyruvate dehydrogenase factory.” It is likely to be the first of many similar structures identified using these methods (Figure 2).

Figure 2. Asymmetric reconstruction of the full pyruvate dehydrogenase complex (PDHc) from native cell extracts. Left panel: Previous studies showed that PDHc has a component that is stably attached to the E2p core, covering the binding site of the substrate. Right panel: Application of the integrative methodology discussed in this article has provided a detailed model of the PDH factory nanocompartment. (Cell Rep. 2021; 34: 108727.)

Changing the interface for cell and structural biology

Technologies, such as MS and cryo-EM, continue to advance at pace, together with functional and biophysical assays. Added to that, automation and machine learning tools have increased throughput and efficiency and have led to an explosion in datasets that allow investigators to take a much broader view of intricate protein interactions. Generating large datasets makes it possible to mine data distributions rather than single data points. Moreover, large datasets provide investigators with new opportunities to interrogate cellular processes and identify points for intervention.

Additionally, machine learning–inspired algorithms are enhancing image processing workflows to accelerate analyses of thousands of protein interactions.3 This high-resolution, high-throughput approach is revealing distinct structural signatures. These signatures can be correlated with proteomic data and cryo-EM maps to characterize previously unidentified protein communities at high resolution.

It is this integrated approach that is providing new insights into molecular organization at near atomic scale and revealing novel protein-protein interactions that are vital for correct cell function.4 Undoubtedly, these insights will take drug discovery beyond an enzyme’s active site to other specific points of protein interaction.

New horizons for drug discovery

There is still a great deal to learn about the cell’s proteome and its exact architecture. Bringing together a combination of biological and biophysical assays into an integrative model, using cell extracts and purified molecules, provides enormous potential for solving the most challenging questions of cell structure and function.

The integrative approach offers a bridge to combine data on purified proteins and subunits with visual and analytical data on larger complexes in their native state, at an unprecedented resolution.

For the first time, transient protein structures that are vital for enzyme activity can be visualized and investigated. And automation and machine learning are transforming the scale and rate at which new insights can be acquired, as witnessed recently with the AlphaFold and RoseTTAFold projects for the prediction of three-dimensional structures of proteins and their interactions.

All this is providing inspiration for new avenues of enquiry. But to push the technologies to the limits, investigators must work together and develop a coherent model that can be applied across laboratories and used by investigators in different disciplines.

Within a decade, the pieces of the cellular puzzle should be in place. A full inventory of cellular protein structures is in sight, and integrated data on how the cell functions and is organized is on the horizon. Critically, the wide implementation of an integrated analytical approach, using state-of-the-art and emerging technologies, is necessary to underpin this explosion in knowledge and accelerate and expand the scope of drug discovery in years to come.

1. Kyrilis FL, Semchonok DA, Skalidis I, et al. Integrative structure of a 10-megadalton eukaryotic pyruvate dehydrogenase complex from native cell extracts. Cell Rep. 2021; 34: 108727. DOI: 10.1016/j.celrep.2021.108727.
2. Skalidis I, Tüting C, Kastritis PL. Unstructured regions of large enzymatic complexes control the availability of metabolites with signaling functions. Cell Commun. Signal. 2020; 18(1): 136. DOI: 10.1186/s12964-020-00631-9
3.  Kyrilis FL, Belapure J, Kastritis PL. Detecting Protein Communities in Native Cell Extracts by Machine Learning: A Structural Biologist’s Perspective. Front. Mol. Biosci. 2021; 8: 660542. DOI: 10.3389/fmolb.2021.660542.
4. Tüting C, Kyrilis FL, Müller J, et al. Cryo-EM snapshots of a native lysate provide structural insights into a metabolon-embedded transacetylase reaction. Nat. Commun. 2021; 12: 6933. DOI: 10.1038/s41467-021-27287-4.

Panagiotis L. Kastritis, PhD (, is junior professor of cryo-electron microscopy and computational structural biology at the Interdisciplinary Research Center HALOmem, Charles Tanford Protein Center Institute of Biochemistry and Biotechnology, Martin Luther University Halle-Wittenberg, Germany.