Box 2. Experimental methods for structural characterization of assemblies

A variety of methods are available for the experimental determination of macromolecular assembly structure (see Fig. 4)

X-ray crystallography is the most powerful method for structure determination because it is capable of providing an atomic structure of the whole assembly^{22, 64}. When suitable crystals and high-resolution crystallographic data are obtained, there is little need for other methods of structure characterization.

Nuclear magnetic resonance (NMR) spectroscopy allows determination of atomic structures of increasingly large subunits and even their complexes^{65, 66, 67, 68, 69}. Although NMR analysis is generally not as applicable as X-ray crystallography to protein structures with more than 300 amino acid residues, it can be applied to molecules in solution and is more suitable than X-ray crystallography to study their dynamics and interactions in solution.

Electron crystallography (two-dimensional electron microscopy or 2D EM) and single-particle analysis can reveal the shape and symmetry of an assembly, sometimes at near-atomic resolution, but more frequently at an intermediate resolution⁷⁰. Segmentation of the electron density may lead to an approximate configuration of subunits in a complex⁷¹. Proteins whose structures are already known can then be fitted into these density maps with an accuracy approaching one-tenth the resolution of the EM reconstruction^{47, 48, 49, 50}.

Electron tomography is based upon multiple tilted views of the same object⁵⁴. Although it can be used to study the structure of isolated macromolecular assemblies at relatively low resolution, its true potential lies in visualizing the assemblies in an unperturbed cellular context.

Immuno-electron microscopy can be used to determine an approximate position of a protein in the context of an assembly⁷². This task is achieved by using a construct of the protein of interest that binds to a gold-labelled antibody. The relative position of the gold particles is then identified by EM.

Chemical crosslinking with mass spectroscopy can be used to identify binary and higher-order protein contacts⁷³. The approach relies on bi- and tri-functional crosslinking reagents that covalently link proteins interacting with each other. Proteolytic digestion and subsequent mass spectroscopic identification of the crosslinked species reveal their composition. In addition, chemical crosslinking of specific residue types has recently been used to obtain intramolecular distance restraints⁷⁴.

Affinity purification with mass spectroscopy combines purification of protein complexes with identification of their individual components by mass spectroscopy (see reviews in this issue by Aebersold and Mann, page 198, and Fields and co-workers, page 208). During cell lysis, the whole assembly is partially broken into smaller complexes that are then isolated by a variety of methods, such as those relying on fusion proteins or antibodies as baits for affinity purification. Subunits in these smaller complexes are usually identified by a combination of gel electrophoresis and mass spectroscopy. Examples include the U1 subunit of the yeast and human spliceosome^{75, 76}, identification of proteins that interact with the GroEL complex⁷⁷, the sampling of protein interactions in the yeast nuclear-pore complex⁷², and a high-throughput identification of the hundreds of distinct protein complexes in budding yeast^{13, 78}.

Fluorescence resonance energy transfer (FRET) occurs when a higher-energy fluorophore stimulates emission by a lower-energy fluorophore that is within 60 Å of its inducer. It can be applied to monitor protein interactions if one protein is fused to a fluorescence donor and its potential partner to a fluorescence acceptor (see accompanying review by Fields and co-workers). Fluorescence donors and acceptors are usually spectral derivatives of the green fluorescence protein.

Site-directed mutagenesis and a variety of biochemical experiments (for example, footprinting) can reveal which subunits in a complex interact with each other and sometimes what face is involved in the interaction^{79, 80, 81}.

Yeast two-hybrid system detects binary protein interactions by activating expression of a reporter gene upon direct binding between the two tested proteins (see review by Fields and co-workers). The approach is based on the modularity of transcription factors that consist of a DNA-binding and an activation domain, each of them fused to two different genes encoding for the proteins whose interaction is tested. If the two expressed fusion proteins are in contact with each other, the two modules of the transcription factor are united, thereby inducing transcription of a set of reporter genes. Expression of reporter genes, in turn, is easily detected by a variety of tests, such as yeast colony colour and ability to grow in deficient media. The method is suitable for high-throughput applications (ref. 11; and see review by Fields and co-workers).

Protein arrays immobilize a variety of 'bait' proteins, such as antibodies and glutathione S-transferase, into an array on a specially treated surface; the array is then probed with sample proteins, resulting in a detection of binary interactions (see review by Fields and co-workers). Messenger RNA expression arrays immobilize stretches of mRNA and are used to measure the concentration of mRNA species in a sample as a function of tissue type, cell cycle and other environmental conditions^{82, 83}. Such data sets have been used to detect functionally linked proteins, which include proteins whose expression is co-regulated because they are members of the same assembly, are encoded on the same operon, or belong to the same biochemical pathway⁶.