New & Noteworthy

Moonlighting Proteins Do Double Duty

January 11, 2018

While most people only know her as the friendly waitress at the local diner, it is by night that Louise gets to live her passion. At night, she is a mime.


Like a waitress who moonlights as a mime, some proteins have more than one job too. (Max Pixel)

If she loses her voice, Louise will struggle as a waitress but do fine as a mime. It is a different story when she gets walloped with a nasty flu and can’t get out of bed. Now she can’t wait tables by day or be stuck in an imaginary box by night.

In a new study out in GENETICS, Espinosa-Cantu and coworkers get to explore which enzymes in our devoted pal Saccharomyces cerevisiae are like Louise in that they do two unrelated jobs. They set out to systematically identify enzymes that have both a catalytic job and a job unrelated to its catalytic function. In other words, they set out to find enzymes that moonlight at second jobs.

Other researchers have stumbled upon these moonlighting enzymes in the course of their regular studies. Something along the lines of stumbling upon Louise struggling against an imaginary wind at a local park.

The difference here is that Espinosa-Cantu and coworkers systematically looked for dual-function enzymes. They did this by comparing the effects of knocking out the entire protein vs. just knocking out the catalytic activity of the enzyme with a single point mutation. They compared Louise having laryngitis vs. her being bedridden with illness.

They found that 4 out of the 11 enzymes they looked at did indeed have additional jobs unrelated to their catalytic function. These enzymes were both mimes and waitresses.

They started out looking at enzymes involved in amino acid biosynthesis mostly because it is easy to test them for loss of function. The yeast can’t grow in the absence of the amino acid the enzyme helps synthesize!

Next, like every good yeast researcher should, they turned to the Saccharomyces Genome Database (SGD) to identify 86 genes that when knocked out, caused the mutant yeast strain to not grow in the absence of an amino acid. They winnowed this list to 18 by focusing only on those enzymes with well-characterized catalytic sites and that were annotated with a single molecular function.

Then, they needed to have versions of these enzymes that were essentially identical to wild type except that they could not catalyze their reaction. They focused on single amino acid mutants that wiped out catalytic activity. The authors got this to work in 15 of their candidate enzymes.

They were now ready for the big experiment. They started out with a strain deleted for one of the enzymes and then added back either a wild type version or a catalytically inactive mutant version on a low copy (CEN-ARS) plasmid. They also included a third strain that had just an empty plasmid (their knockout strain).

They compared the growth of these three strains under 19 different growth conditions and looked for conditions where the knockout strain had a larger effect on growth compared to the strain with the catalytic mutation. The idea here is that if an enzyme does more than one job, knocking it out should have a larger effect than just disabling its enzymatic activity.


If Penny the waitress from Big Bang Theory is sick, both the Cheesecake factory and her movie set will suffer. (Wikia)

Before analyzing their results, the authors wanted to make sure that their wild type strain, the knockout strain with the deleted gene added on a CEN-ARS plasmid, grew equivalently to a true wild type strain. This was only true in 11/15 of the strains and so their analysis focuses on these 11 enzymes.

Espinosa-Cantu and coworkers found no difference in growth rate with all three strains for 38.6% of the growth rate comparisons. Under these conditions, the enzyme isn’t doing much for the growth rate of the cells.

In 233 growth conditions, the enzyme did matter for growth and in 70% of these, there was no difference between the knockout and catalytically inactive strains. Presumably under these conditions the enzymes have a single job—their catalytic function.

But this left 30% of cases where there was a difference suggesting that under these conditions, the enzymes have more than one job. Most of these involved 4 of the 11 genes—ALT1, BAT2, ILV1, and ILV2.

They decided to focus in more detail on one of these, ILV1. To figure out what else Ilv1p might be doing besides its threonine deaminase job, they generated two separate strains of double knockouts with 3,878 non-essential genes. The first set of 3,878 strains had a deletion of ILV1 while the second set had a catalytically inactive form of ILV1.

When they looked at the cases where the ILV1 knock out strain was more severely affected compared to the catalytically inactive strain, they found that 187 out of 345 negative interactions had at least some non-catalytic component.  The non-catalytic interactions were enriched for gene regulation with chromatin modification being one of the most striking.

These interactions put the non-catalytic role of ILV1 in a category more similar to stress response, protein sorting and chromatin remodeling genes than to other metabolic genes. This suggests that Ilv1p is both involved in amino acid metabolism and stress response. Quite a different moonlighting gig compared to its day job!

So researchers need to be careful when drawing conclusions from their knockout experiments. Sometimes the effect you are seeing is unrelated to the function you are studying. Sometimes when a waitress falls ill, the mime performance is affected as well.

No this isn’t a show about ILV1.

by Barry Starr, Ph.D., Director of Outreach Activities, Stanford Genetics

Categories: Research Spotlight

Tags: protein moonlighting , pleiotropy , systems genetics