Best of BMB in 2021

Protein folding in silico

After winning the CASP14 protein structure prediction competition in November 2020 by accurately predicting numerous protein structures that had been solved experimentally and kept secret, the program AlphaFold was published in the journal Nature in July. The program’s developers at the Google affiliate Deep Mind also launched a protein structure database in partnership with the European Bioinformatics Institute. In August, an academic group based at the University of Washington published a related approach called RoseTTAfold in the journal Science.

Both programs depend on machine learning to extract rules from an enormous number of previously solved structures and predict new ones.

Biologists expect to use this computational advance to save time in solving future structures, predict protein–protein interactions, and perhaps design drugs. Michael Airola, a structural biologist at Stony Brook University who works on challenging proteins that bind to lipids, said the program offers testable hypotheses his lab has found useful.

Vanderbilt University biochemist Fred Guengerich said, “It is modeling and the proof is in the predictability.” Still, he wrote, his lab used an AlphaFold model in a recent paper “in the absence of a real structure available.”

Structure prediction isn’t just for proteins, either. Karin Musier–Forsyth, a biochemist at Ohio State University, pointed out another deep learning effort “somewhat analogous to the AlphaFold breakthrough, but for RNA,” which also was published in Science in August.

Glycosyltransferases find a new target

Another molecule unexpectedly has joined the crowded milieu on the cell surface — although evidence suggests it may have been there all along.

While testing to see whether RNA in the cytoplasm might be modified with a small, reversible sugar group, researchers in chemist Carolyn Bertozzi’s lab at Stanford University found large, RNase-sensitive glycosylated molecules on the surface of several immortalized cell lines. The find suggested that the surface glycome includes RNA molecules with complex carbohydrate modifications only seen on proteins until now. The modified RNAs are from a noncoding family that genetic studies have linked to autoimmune disorders; the team found that they bind to a receptor class called sialic acid binding immunoglobulin type lectins, or siglecs. The research, preprinted in 2019, was published in the journal Cell in June.

Bertozzi told ASBMB Today contributor Ankita Arora, “Once again, we are humbled by how little we know about biology.”

The finding satisfied an initially skeptical Bertozzi and eventually peer reviewers, but some scientists remain cautious. University of Georgia glycobiologist Gerald Hart wrote, “I hope Carolyn’s findings are correct, but until they are repeated by others, a healthy dose of skepticism is warranted. … The field missing a large N-linked type glycan seems hard to fathom, but we all need to keep an open mind.”

Mass in vivo transfection

Development of a new type of vaccine, an advance that was realized in 2020, began to have an impact on daily life in 2021. Two mRNA-based vaccines that confer robust immunity to SARS-CoV-2, along with a more conventional peptide antigen vaccine, became available to healthcare workers in late December 2020 and were, within the U.S., widely available for adults by May.

Many researchers joked about receiving their transfections, a term for an often-used lab technique that delivers nucleic acids in liposomes that can reach the cytoplasm through endocytosis.

While public health practitioners still are grappling with vaccine hesitancy and misinformation, researchers in the pharmaceutical industry regard the shots’ efficacy as a key proof of concept. Companies such as Moderna have begun to work on other therapies, including more vaccines, that can be delivered in mRNA form.

‘Nothing is undruggable’

Some goals take a long time to reach. Scientists have known since the early 1980s that the Ras family of GTPases, which activate growth signals, can be powerful oncogenes. Ras mutations are involved in about one-third of cancers.

Yet for decades, the search for a way to block cancer-related signaling by one member of the Ras family, a mutant KRas called G12C, without harming healthy tissues failed miserably. Because KRas has a smooth surface and few ligand binding pockets, medicinal chemists began to think it might be impossible to design an inhibitor to block its activity. But this year, two small-molecule inhibitors targeting KRas G12C were approved to treat certain cancers. University of California, San Francisco, chemist Charles Craik, who nominated the advance, wrote, “These breakthrough drugs have changed the course of management of KRas driven cancers and open the way for additional approaches of combination therapies to benefit patient care.”

In other feats of small-molecule design, targeted protein degradation has expanded beyond proteome-targeting chimeras, or PROTACs, to include small molecules that tee their targets up for lysosomal or macroautophagy-based degradation, according to a March review article in RSC Chemical Biology. Meanwhile, researchers also have reported a small-molecule drug candidate that can block the interaction between protease PCSK9, which regulates low-density lipoprotein receptor levels, and its target, the LDL receptor. “Nothing is undruggable,” remarked Jay Bradner, the president of Novartis’ research arm, while announcing the publication in the journal Cell Chemical Biology in September.

A daisy chain of knots, loops, hairpins and petaloids

Single-stranded RNA is never as straight as cartoons make it look. Instead, the molecule tends to base pair over short stretches, forming loops, hairpins and other secondary structures that can impart function. Several groups have reported the secondary structure of the SARS-CoV-2 single-stranded RNA genome both within virions and in host cells. Together, the studies show that the genome’s secondary structure involves numerous stem-and-loop elements concatenated into what one research team called “petaloid structures” in their June study in the journal Nature Communications. The genome further folds into an approximately spherical tertiary structure to fit into virus capsids.

A study published late in December 2020 in Nucleic Acids Research reported on a number of structural elements that might present druggable pockets, crevices on the face of the viral genome that small molecules could slip into to act, perhaps, as therapeutics. Musier–Forsyth, who nominated these studies, called them “quite impressive.”

Three cheers for cryo-EM

While the technique has been around for a few years, cryo-electron microscopy continues to dazzle researchers with its speed and clarity in determining structures. Craik wrote, “Major technological efforts … facilitated achieving the goal of determining protein structures at atomic resolution and started a ‘resolution revolution,’ which forever changed the landscape of structural biology.”

Several scientists pointed to its impact, especially advances in understanding protein complexes. As an example, Craik cited the “breakneck speed” with which structures of the SARS-CoV-2 spike protein bound to human Ace2 receptor were published in 2020, just months after the virus initially was isolated.

The technique also is offering new insights into long-standing questions. Binks Wattenberg, a biochemist at Virginia Commonwealth University, nominated a pair of papers that appeared in March in Nature Structural and Molecular Biology on the structure of a serine palmitoyltransferase complex. “This structure reveals so much about how this critical enzyme is regulated by accessory subunits,” Wattenberg wrote, adding that it also “illustrates the accumulating power of cryo-electron microscopy for determining the structures of multi-subunit membrane proteins.”

This year, researchers at the Rockefeller University reported in September in the journal Science on a cryo-EM structure of three stages of the giant ribonucleoprotein complex called the small subunit processome, a precursor to the mature ribosome. Ribosome assembly is a complex and intricate process, with pieces coming together a few at a time into a massive molecular machine, each of whose two subunits includes 30 or more proteins built on an RNA core. Many chaperones that help to build the ribosome fall away before the mature ribosome is complete.

The series of structures the researchers found helps to unravel the steps in ribosome assembly and the structure of intermediates. Yale biophysicist Susan Baserga, who nominated this finding, wrote, “Twenty years after my laboratory purified and named (the SSU processome) in yeast, we have the human structure. Very exciting!”

Getting therapeutic proteins into the brain

The blood–brain barrier is a major hurdle for treating certain diseases. Tight junctions between blood vessel endothelial cells play a protective role, shielding the brain from pathogens and from circulating molecules that are useful for whole-body physiology but toxic to neurons. Still, the barrier can make it extremely difficult for researchers to deliver drugs into the brain.

Matthew Gentry, a biochemist at the University of Kentucky, nominated an approach to evade the blood–brain barrier that was published in the journal Cell by researchers at Denali Therapeutics and several universities in September. The approach takes advantage of the brain’s mechanism for importing necessary iron. The vascular endothelia have a receptor for the iron carrier protein transferrin. Denali used a protein chimera made of the transferrin receptor binding domain and a therapeutic cargo protein to deliver a replacement for the lysosomal protein progranulin in mice. The chimeric protein restored ordinary lipid levels and showed some tissue-level restoration of health in the animals’ brains. Progranulin deficiency can cause frontotemporal dementia in humans, so the results are of clinical interest. The company has announced that it is testing candidate drugs to treat other lysosomal storage disorders as well.

Genomic data from innumerable patients

Broad Institute proteomics director Steven Carr wrote, “I think the advances made by the NCI-CPTAC consortium deserve special mention.”

The Clinical Proteomic Tumor Analysis Consortium, or CPTAC, sponsored by the National Cancer Institute, is a massive collaboration aimed at proteogenomic analysis of a variety of common cancer types. The combination of DNA sequences, proteomes, post-translational modifications and RNA-seq data for individual patients, often matched to adjacent healthy tissue, gives deep insight into how cancer develops. The project was launched in 2014, and its data had been used in at least 740 publications in 2021 when this issue of ASBMB Today went to press.

This year, proteogenomics studies using the data have given detailed portraits of what goes wrong in pancreatic ductal adenocarcinoma, lung squamous cell carcinoma and glioblastoma. Each study includes data on 100 tumors or more and identifies patterns of potential molecular vulnerability that promise to help future patients. Other studies aim to guide treatment or cancer staging, and still others offer mechanistic insights into disease phenotypes such as treatment resistance. Carr wrote, “The information that’s been provided … has been pretty amazing, and has highlighted (how) proteomics adds to and goes beyond genomic methods.”

Lipids and virology

A virus would not make it far without its host cell. Add to that the fact that humans tend to change more slowly than viruses, and it becomes clear why understanding how host cell factors contribute to viral pathogenesis can help to develop broad-spectrum antivirals.

Looking back on this year, University of Wisconsin metabolic biochemist James Ntambi wrote, “The close link of lipid metabolism to virology noted this year is beginning to excite me.”

In addition to the lipid nanoparticles that encapsulate mRNA vaccines, Ntambi highlighted a September article in the journal Nature Metabolism that reported that fatty acid synthase, which produces palmitate through a series of steps, is a crucial enzyme for SARS-CoV-2 replication. The virus spread less completely through a cell culture with fatty acid synthase inhibited, and the same inhibitor — approved as a treatment for obesity — also reduced viral titers in mice. The study’s authors pointed out that palmitoylation is an important modification for the function of the viral spike protein.

Ntambi wrote, “What this says is that the lipid biosynthetic pathway could be seen as a potential antiviral strategy for COVID-19 treatment.”

Proximity proteomics

Protein–protein interactions drive a lot of biology but can be difficult to capture. They are often fleeting, may involve only glancing contact and sometimes affect only a modified subset of the two binding partners. Al Burlingame, a professor at the University of California, San Francisco, nominated proximity labeling coupled to proteomics for highlighting at year’s end. These methods use radicals, affinity tags or other often complementary processes to modify molecules close to an introduced bait enzyme.

Using proximity proteomics, researchers can conduct subcellular proteomics and other narrowly focused investigations. Some of the studies published and preprinted this year include studies of lipid raft signaling, mitochondrial RNA anchor, interactions between surface proteins at the synapse, and assembly of the kinetochore.

In a March article in Chemical Society Reviews, scholars led by David MacMillan and Tom Muir wrote that the development of proximity proteomics illustrates how “the union of chemistry and biology can present powerful tools that can impact human health.”