Protein Synthesis Enzymes Have Evolved Additional Jobs

Living things possess at least one type of AARS molecule for each of the 20 proteinogenic amino acids. For some amino acids, there are two varieties, with a separate enzyme for use in protein translation that takes place in the mitochondrion. All of these synthetases have a core segment that is involved in linking tRNAs and amino acids, and all but one harbor one or more additional accessory domains. Plus, by alternatively splicing their mRNAs or fragmenting the proteins post-translationally, cells can make more than 300 different protein variants from AARS genes. Some of these variants moonlight as inflammatory cytokines. Others orchestrate the formation of blood vessels. The AARSs for glutamic acid and proline are merged into a two-part protein; the linker between them seems to control immune activity and fat metabolism, and may even influence life span. Many AARSs have been linked to human diseases caused by defects not in protein assembly, but in these other, noncanonical functions.

I heard how skeptical the field was about those discoveries. I don’t blame them. I would be confused too.

 —Xiang-Lei Yang, Scripps Research Institute

Some researchers now view the enzymes as drug targets for cancer, immune disease, and other conditions. The company Schimmel cofounded, aTyr Pharma in San Diego, envisions the AARS proteins themselves as an entirely new class of drugs, distinct from small molecules or other biologics. The firm is currently running a clinical trial testing a piece of the histidine enzyme, HisRS, for treating inflammatory lung disease.

Alternative AARS functions have been known in lower organisms such as bacteria since the 1980s, but their activities aren’t extensive, says Schimmel. Then, starting in the ’90s, Schimmel and others began to uncover noncanonical functions in higher eukaryotes, starting with unexpected roles in angiogenesis. The discovery of new functions for these ancient proteins was “a big surprise,” says David Dignam, a biochemist at the University of Toledo. But given the diverse functions that researchers studying AARSs have uncovered, many of which touch on crucial disease pathways, Dignam says he thinks aTyr’s approach makes sense. “Arguing that you can make medicines based on this, I think, is very logical.” 

While other proteins have adopted secondary functions, the quantity and variety of side gigs found in the AARSs is remarkable, says Schimmel. And he doesn’t think it’s a coincidence. These particular synthetases have been present and available for evolution to modify since protein-based life began. Given their essential role in protein synthesis, they’re consistently produced, and unlikely to disappear from any viable genome. That makes them a stable substrate for new functional domains. Moreover, they possess specific amino acid binding sites, ready to interact with other proteins. 

“It’s lock and key,” says Schimmel. “Any protein that sticks out a nice side chain that matches a synthetase could eventually become a partner.”

Building and blocking blood vessels

Schimmel says he’s long been fascinated with AARSs’ original function: interpreting the genetic code. Back in the ’90s, Schimmel’s lab, then at MIT, was sequencing the AARS genes. “We were interested in developing small molecules that would target them and kill their activities in specific ways,” he says. For example, if the AARS of a pathogen was different enough from that in people, he reasoned, one could develop an antibiotic that shuts off protein synthesis in the infectious agent.

Schimmel’s then-postdoc Keisuke Wakasugi got curious about the sequence of the gene encoding TyrRS, the AARS for tyrosine. In humans, TyrRS includes an extra segment at the carboxyl end of the enzyme, a feature that isn’t present in prokaryotes or lower eukaryotes. The amino acid sequence for this part of the protein was similar to that for a human cytokine, EMAP II, which recruits circulating immune cells into tissues to promote inflammation. Wakasugi decided to test that carboxyl domain for cytokine-like activity.

“That’s a silly idea,” Schimmel recalls thinking. But Wakasugi went ahead, and sure enough, the TyrRS carboxyl domain acted just like EMAP II, inducing cultured phagocytes and leukocytes to migrate and release inflammatory signals. The full-length TyrRS, in contrast, didn’t influence the cells’ behavior. That hinted at the possibility that the carboxyl domain could be broken off the TyrRS for immune functions. No one in the lab would believe the finding at first, so Wakasugi repeated the experiments, with the same results. 

Although it would take more than a decade to show that such AARS fragments were truly present and relevant in a living animal, Wakasugi knew he was onto something. “Paul and I were very excited to discover a novel and unexpected function of human TyrRS,” recalls Wakasugi, now a biochemist at the University of Tokyo. “Throughout this project, I felt that we opened the door to a whole new research field.” 

As part of the same study, Wakasugi also investigated the amino-terminal, catalytic domain of TyrRS, wondering if it might also influence cell migration. It behaved in a manner reminiscent of the cytokine interleukin-8 (IL-8). Both the TyrRS amino-terminal fragment and IL-8 bind to the IL-8 receptor on certain leukocytes, causing them to migrate in culture.

Schimmel recruited Xiang-Lei Yang, a postdoc with expertise in structural biology, to join his lab at Scripps in La Jolla, California, to investigate how TyrRS might manage alternative functions. Yang zeroed in on a particular sequence of amino acids, glutamic acid–leucine–arginine, required for the synthetase fragment’s cytokine activity. The same sequence was also found in IL-8 and related cytokines. In crystal structures, she found that full-length TyrRS buried this motif, but it was exposed in the cytokine-like fragment. 

IL-8 was known to promote the formation and growth of blood vessels, so Wakasugi also tested his TyrRS amino-terminal fragment for angiogenic activity. When he injected a bit of gel containing the fragment into mice, blood vessels grew and suffused the gel. To explore that action further, Schimmel phoned his Scripps colleague Martin Friedlander, an ophthalmologist and cell and developmental biologist, and asked him to test the TyrRS fragment in his mouse models of eye vascularization. Friedlander agreed, but also asked for a control. So along with the human TyrRS fragment, Wakasugi provided a natural splice variant of the tryptophan enzyme, TrpRS, that lacks the glutamic acid–leucine–arginine motif. 

The results, Friedlander recalls, weren’t exactly what he expected. TrpRS, the supposed control, “had a much more potent effect,” says Friedlander, who is also president of the Lowy Medical Research Institute in La Jolla. But that effect was the opposite of TyrRS action: rather than promote angiogenesis, as Wakasugi had seen in the gel, the TrpRS fragment blocked it in mammalian cell culture, chicken embryos, and mouse eyes. “TyrRS and TrpRS may have evolved as opposing regulators of angiogenesis,” says Wakasugi.

Scientists were initially resistant to the idea that an AARS could have functions beyond protein synthesis. Yang recalls attending a conference, shortly after Wakasugi published his work on angiogenesis, where others were unaware that she was a Schimmel acolyte. Thus incognito, “I heard how skeptical the field was about those discoveries,” she recalls. “I don’t blame them. I would be confused too.”

MOONLIGHTING ENZYMES Aminoacyl-tRNA synthetases play a fundamental role in protein translation, linking transfer RNAs to their cognate amino acids. But in the hundreds of millions of years that they’ve existed, these synthetases (AARSs) have picked up several side jobs. One of these is to manage the development of vertebrate vasculature. © Thom Graves
	Multiple AARSs play roles in the development of the vertebrate circulatory system. During development, the serine enzyme SerRS downregulates the expression of vascular endothelial growth factor A (VEGF-A), preventing over-vascularization.
	In addition, a combo synthetase for glutamic acid and proline, GluProRS, links up with other proteins to form the interferon-γ activated inhibitor of translation (GAIT) complex to block VEGF-A translation.
	A piece of the tryptophan synthetase TrpRS also contributes to dampening angiogenesis by binding and blocking VE-cadherin receptors on endothelial cells so they can’t link together to form blood vessel lining.
	Meanwhile, a fragment of the tyrosine synthetase TyrRS appears to promote the growth of blood vessels by stimulating migration of those endothelial cells.
When these functions arose in evolution According to Scripps Research Institute biochemist Paul Schimmel, the addition of accessory domains that perform such tasks parallels major events in the evolution of circulation. The first blood vascular system, which lacked the endothelium present in modern vertebrates, probably arose in a common ancestor of vertebrates and arthropods around 700 million to 600 million years ago. Around this same time, TyrRS acquired a glutamic acid–lysine–arginine motif that today is thought to promote angiogenesis. Then, around 540 million to 510 million years ago, an ancestral vertebrate evolved a closed vascular system, with blood pumping through vessels lined by endothelium. At some point around that same time period half a billion years ago, the TrpRS picked up a WHEP domain, which today regulates its ability to block angiogenesis. In addition, SerRS acquired a domain unique to this enzyme, which now prevents over-vascularization in developing zebrafish, and likely other vertebrates. GluProRS’s role in angiogenesis, on the other hand, doesn’t seem to be so precisely timed to the evolution of vasculature. A linker protein tied together the AARSs for glutamic acid and proline enzymes around 800 million years ago, before circulatory systems existed. © THOM GRAVES
700–600 MYA A primitive circulatory system appears in the common ancestor of arthropods and vertebrates. Endothelium-lined blood vessels are not yet present. TyrRS acquires a domain that today promotes angiogenesis. 540–510 MYA The closed circulatory system, with blood vessels lined by endothelial cells, appears in an ancestral vertebrate. TrpRS acquires a domain that blocks angiogenesis in modern animals. SerRS acquires a domain that today regulates vascular development in zebrafish and likely other vertebrates.

Vasculature and beyond

While the TyrRS and TrpRS functions Wakasugi and colleagues had discovered were interesting, it wasn’t clear that the enzyme fragments genuinely performed these functions in vivo. Yang realized that to give herself and other scientists confidence about noncanonical functions of AARSs, she’d have to find evidence that they were present in animals.

The team still hasn’t done so for TrpRS or TyrRS, but Wakasugi found her opportunity with the serine enzyme, SerRS. Multiple published genetic screens in zebrafish had identified defects in vascular development when SerRS was mutated. But mutations that knocked out the enzyme’s ability to link tRNAs and amino acids did not cause such defects, indicating that something else was going on.

To figure out what, Yang turned to a sequence, christened UNE-S, that is found in vertebrate, but not invertebrate, SerRS. Yang’s team—she joined the Scripps faculty in 2005, and now shares a lab with Schimmel—quickly identified a nuclear localization sequence within UNE-S, and determined that mutations altering this signal caused the vascular defects in zebrafish embryos. In the nucleus, they found, SerRS seems to minimize the expression of vascular endothelial growth factor A (VEGFA). The study, published in 2012, was the first to illustrate an essential, natural role for an AARS accessory domain in a living animal. Shortly thereafter, the team reported that nuclear SerRS blocks VEGFA by competing and interfering with c-Myc, a transcription factor that normally promotes the gene’s expression.

Meanwhile, Schimmel’s and Yang’s groups continued to try to explain the noncanonical functions of TrpRS and TyrRS, even as they found more side gigs for these enzymes. Yang led studies on the TrpRS fragment’s structure and mechanism. She discovered that full-length TrpRS doesn’t influence angiogenesis because it’s capped by a WHEP domain—so called because this domain appears in aminoacyl tRNA synthetases for tryptophan (W), histidine (H), glutamic acid (E), and proline (P), as well as in the glycine and methionine enzymes. Yang’s team found that when uncapped by proteases in the extracellular space, TrpRS binds to a cellular receptor called VE-cadherin. Specifically, tryptophans in the receptor seemed to enter the TrpRS’s active site to create the bond. That’s why Wakasugi saw that only the fragment, not the full TrpRS, blocked angiogenesis. 

More recently, Schimmel has also been interested in plant-based amino acid–like compounds, such as resveratrol, the stuff in red wine that’s thought to counter oxidative stress. Resveratrol and tyrosine are similar in that both contain a phenolic ring, and that’s important for resveratrol’s ability to influence the expression of pro- and anti-oxidative genes. In 2015, Schimmel’s team reported that under conditions of stress, TyrRS moves into the nucleus of human cultured cells or living mice, where any resveratrol present fits neatly into TyrRS’s active site. This turns off the normal TyrRS catalytic activity to connect tyrosine molecules with the appropriate tRNAs. Instead, TyrRS stimulates the activation of PARP-1, an enzyme involved in DNA repair.

A few years later, the team found that an alternatively spliced version of TyrRS stimulates platelet proliferation in mice and cultured cells, and could potentially be used to treat people with a low platelet count.

Schimmel expects noncanonical AARS functions will keep the group busy for a long time. “We are barely scratching the surface of what is to be learned,” he says. “I am as excited, or even more excited, about these enzymes as I was when I started out decades ago.”

Managing inflammation and metabolism

As evidence of noncanonical functions for AARSs was trickling out of Schimmel’s lab, Paul Fox, a biochemist at the Cleveland Clinic’s Lerner Research Institute, was studying the control of inflammation in macrophages. Specifically, his team was investigating a complex generated when the cells were exposed to the cytokine interferon-γ. A protein complex called GAIT (for interferon-γ activated inhibitor of translation), generated within macrophages, binds to and blocks mRNAs related to inflammation. Inside the complex, the researchers found GluProRS, an enzyme that includes the AARSs for glutamic acid and proline.  

“We ran into it just absolutely by accident,” Fox recalls. “I didn’t think it was an interesting enzyme.” But he knew of Schimmel’s work, and he picked up the phone to call Scripps.

One minute into the call, Schimmel interrupted to welcome Fox to what Schimmel called the most exciting area of AARS research: the noncanonical functions. Schimmel also promised his assistance, Fox says. “He’s been a big supporter and a friend ever since.” With tools supplied by Sunghoon Kim, a former Schimmel lab postdoc now at Yonsei University in South Korea, Fox’s team discovered that interferon-γ causes GluProRS to become phosphorylated, abandon its post in translation, and join up with the other GAIT members to halt the production of inflammatory proteins.

It’s not clear why the glutamic acid and proline synthetases paired up approximately 800 million years ago, but Fox has a hypothesis, which he published in 2018. Proline is synthesized from glutamic acid, and at that period in evolution, emerging animal proteins began to include more proline. That may have led to a rise in the production of ProRS that sopped up all the available proline, requiring more to be made from glutamic acid. That might have resulted in a deficit in glutamic acid levels, impairing protein synthesis. “The solution to that was to fuse the two synthetases into a single gene, so they have to be made in the same amounts,” says Fox. “No one’s stealing from the other.”

The linker between the two synthetases is crucial for GAIT complex activity; it’s made of three WHEP domains that bind to target RNAs. Fox speculates that sometime after the linker appeared, the cell coopted it to regulate inflammation.

More recently, Fox’s team wondered if the GAIT pathway might function in cells besides macrophages. When the researchers looked at fat cells, they saw that insulin treatment caused GluProRS to become phosphorylated and leave the protein-synthesis machinery. But it didn’t join the other GAIT partners. Instead, it paired with a normally cytosolic protein called fatty acid transport protein 1 (FATP1). Together, the molecular duo went to the fat cell’s membrane, where the transporter brought fatty acids into the cell.

I am as excited, or even more excited, about these enzymes as I was when I started out decades ago.

 —Paul Schimmel, Scripps Research Institute

The researchers engineered a mouse lacking the phosphorylation site needed to free GluProRS to find FATP1. With less fatty acid–storage ability, the mice were lean, weighing about 15 percent to 20 percent less than control animals. Moreover, they lived nearly four months longer, giving them a lifespan that was increased by about 15 percent. A similar gain in people would correspond to a decade or more. “If we could target that phosphorylation site, maybe we could increase life-span,” says Fox. His lab is in the very early stages of looking for a small molecule to inhibit that phosphorylation event.

Drug development

In the various jobs that AARSs have taken on above and beyond their traditional role, Schimmel and colleagues see a theme: they keep cells and bodies stable. “They seem to be something that’s playing a modulatory role, restoring more of a homeostasis,” says Leslie Nangle, a former Schimmel lab grad student who is now senior director for research at aTyr Pharma. Many researchers think it’s risky to mess with such essential enzymes, says Kim, but he and Schimmel see potential in targeting AARSs for treating disease. Schimmel’s company aTyr, of which Kim and Yang are also cofounders, hopes to turn the enzymes themselves into biologic therapeutics. In addition, in Seoul, Kim directs the nonprofit drug discovery organization Biocon, where researchers are developing several small molecules that interact with AARSs, as well as biologics based on natural AARS variants. 

Biocon is currently testing molecules to treat cardiac fibrosis, alopecia areata (an autoimmune disease that causes hair loss), and inflammation. A fibrosis treatment now under Phase 1 study targets the site on the proline synthetase that links the amino acid to its tRNA. Fibrosis results from an accumulation of collagen, which is two-thirds proline. Biocon researchers have found that a drug can go after that active site, knocking down the canonical function by more than 90 percent in healthy cultured cells without greatly affecting the synthesis of other proteins or cell proliferation, says Kim. At first, he and his colleagues didn’t believe their results, but he’s come to see sense in them. “A normal cell is not necessarily doing high level protein synthesis all the time,” he says. “As long as it has a certain degree of residual activity going on, then a normal cell can be perfectly happy.” 

For cancer and other conditions, Biocon is developing small molecule candidates that avoid the tRNA–amino acid linking site or target the extracellular activities of secreted AARSs, meaning that protein synthesis should not be affected. Similarly, aTyr researchers expect that the firm’s therapeutics, based on AARS derivatives themselves, to be relatively safe. “Coming from a world of natural physiology, you start to feel better about it,” says aTyr CEO Sanjay Shukla.

In the developing mouse retina, fragments of the tryptophan synthetase, TrpRS, that are missing a restrictive protein cap (B, C) prevent vascularization of a secondary tissue layer (right). (The right images include the shadows of the adjoining primary layers, which are shown at left.)

PNAS, 99:178–83, 2002, © 2002 NATIONAL ACADEMY OF SCIENCES

Nangle and colleagues, alongside aTyr’s subsidiary Pangu Biopharma in Hong Kong, began by cataloging natural AARS splice variants and then screening them for interesting biological activities in a variety of human cell–based assays. They looked for effects on cell proliferation and protection, immunomodulation and inflammation, cell differentiation, transcriptional regulation, and cholesterol transport. “We figured there’s got to be some therapeutic benefit there,” says Schimmel. 

Currently, aTyr is pursuing an immuno-modulator based on the WHEP domain of the histidine enzyme HisRS. In human T cell cultures, full-length HisRS quieted activated cells and reduced cytokine production. In further experiments, aTyr researchers found that the WHEP domain hooks up with receptors on those immune cells to dampen activity. The company hopes that its modified version of the HisRS WHEP peptide, attached to a bit of antibody to help it last longer in the bloodstream, will have the same quieting effect in an inflammatory disease called sarcoidosis. This disease affects a variety of organs, most often the lungs, and can sometimes require lifelong treatment with immune-suppressing steroids. Those medications come with a list of misery-inducing and dangerous side effects ranging from insomnia to glaucoma to infection. 

aTyr presented results from several animal models of inflammatory lung disease at the American Thoracic Society meeting in 2017, 2018, and 2019, and those findings suggest the company’s candidate 1923 looks promising. For example, the cancer drug bleomycin can cause lung damage, but HisRS or its WHEP domain reduced inflammation and fibrosis.¹⁹ Rats treated with bleomycin breathe quickly to compensate for their damaged lungs, but those treated with 1923 recovered normal respiratory rates.

aTyr’s 1923 has already been through a Phase 1 trial for safety in healthy people without any red flags. Now, the company is running a Phase 1/2 study in people with sarcoidosis, looking to confirm safety, find the right dosage, and perhaps even see signs of efficacy. Patients enter the trial while taking steroids, and the aim is to taper down the steroid dosage during the study. Those receiving 1923, it’s hoped, will see their symptoms stay the same or improve, while those on placebo should see them worsen as the steroid doses are lowered. 

It’s a testament to the need for a new treatment that volunteers are willing to risk having their symptoms intensify if they end up in the placebo arm, says participating physician Daniel Culver, a pulmonologist at the Cleveland Clinic. “[Steroids are] very toxic,” says Culver, who notes that one of his patients calls his steroid prescription “the Devil’s drug” because it does almost as much harm as good. “People are very, very motivated to find something different.” 

The study plans to enroll 36 participants, but has been delayed by the COVID-19 crisis. With such a small sample size, Culver doesn’t expect a “home run,” but he says he hopes the data will be good enough to embark on a larger, Phase 3 study. aTyr is also planning a Phase 2, 30-person trial of 1923 for respiratory complications associated with COVID-19.

If aTyr succeeds, it will be the first instance of a therapeutic built from an AARS—but probably not the last. As Kim sees it, AARSs are ready and waiting to respond to anything that challenges homeostasis, from cancer to the novel coronavirus. “I rename the synthetases ‘Molecular 911.’” 

Correction (June 2, 2020): The original version of a table in this story stated that during HIV infection, the synthetase for lysine (LysRS) is packaged into new viral particles that use its UUU sequence to prime reverse transcription in newly infected cells. Rather, the viral particles use LysRS to deliver its cognate tRNA, which is used as a promoter for reverse transcription. The Scientist regrets the error.