Study shows freezing protein samples before X-ray analysis distorts structure

A new study by researchers at St. Jude Children’s Research Hospital and the University of California-Irvine shows that the conformation of proteins can be significantly distorted by the practice of freezing with liquid nitrogen preparatory to analysis by X-ray crystallography. 

Freezing a sample to liquid nitrogen temperatures of 77K (–370ºF) is standard procedure to limit the damage X-rays can do to the protein. However, as with most substances, freezing shrinks the protein and can cause the large molecule to change its shape, or conformation. 

The results can especially affect drug research, where knowing the proper shape of the protein helps chemists figure out what substances will likely interact with them to achieve a desired effect.

Currently about 95% of the protein structures stored in public databases have been determined using cryogenic (very low temperature) technology.

“We need to rethink how we collect, analyze and utilize structural information when we set out to discover bioactive molecules,” corresponding author Marcus Fischer told the St. Jude’s press office. “You can view temperature as an experimental knob we can turn to explore hidden protein conformations.” 

The work of the six-member team was published Aug. 18 in Chemical Science, the peer-reviewed journal of the Royal Society of Chemistry.

Protein structure

Proteins are long chains of the 22 amino acids used in living organisms. Their sequence is determined by the genome. The chain, known as a polypeptide, is considered the primary level of structure. It can consist of tens, hundreds or thousands of individual amino acids.

Due to the electrical and chemical bonds among their atoms, the chains commonly reassemble into two basic shapes, a spiral or helix, or a sheet of parallel chains. This is known as the secondary structure.

The tertiary structure arises as these secondary structures bend and twist together in a wide variety of ways, in a process called protein folding. 

Quaternary structure describes the overall shape formed when there are two or more individual polypeptide chains. 

Taken together, these levels of structure are known as the protein conformation – the interatomic angles, twists, bends and folds that describe its overall shape. While the specific sequence of amino acids determines the conformation, there is more to it. 

Many proteins can possess more than one conformation state, depending on the ambient temperature, pressure, pH, salinity, voltage and other external factors.

Conformation at room temperature

In a 2014 study, Fischer and colleagues at the University of California-San Francisco had shown that room temperature X-ray crystallography could help pharmacologists to better understand the energies involved in ligand binding (the attachment of substances to biomolecules). Generally, the lower-energy state is the one preferred by nature.

The results of the crystallography are fed into computational modeling that aid researchers especially in drug research.

In this new study, Fischer and the six-member team examined a well-investigated protein structure, the T4 lysozyme L99A cavity. 

A lysozyme is a type of enzyme that can help the immune system attack and destroy invading bacteria or virus. Mutations in the usually well-packed lysozyme can cause the appearance of an internal cavity which can allow foreign substances, known as ligands, to bind to it.

L99A refers to a common mutation that replaces the amino acid leucine with alanine, creating such a cavity. 

“Despite decades of work on this protein,” the authors write, “shifting to RT [room temperature] reveals new global and local structural changes.” 

The changes included “uncovering an apo helix conformation that is hidden at cryo but relevant for ligand binding, and altered side chain and ligand conformations,” they write.

“Our results suggest caution when consulting cryogenic structural data alone," the authors conclude. “Temperature artifacts can conceal errors and prevent successful computational predictions, which can mislead the development and application of computational methods in discovering bioactive molecules.” 

The study was authored by Shanshan Y. C. Bradford, Léa El Khoury, Yunhui Ge, Meghan Osato, David Mobley, and  Marcus Fischer.