Just like a bull in a china shop, formaldehyde (FA) was thought to damage almost everything it bumped into. But new research shows that it’s a lot more targeted, and it really likes proteins involved in DNA methylation.
Sure, FA can preserve brains in jars, but it’s most often used in consumer products like glues and cosmetics, and of course, it’s a handy reagent in the lab. The human body makes a fair amount of this toxic, cancer-causing substance, too, and that piqued the interest of Christopher Chang’s lab (University of California, Berkeley).
Based on in vitro studies, they and others had an inkling that FA was more selective than people had suspected. To get a more precise understanding, the Chang lab performed a series of elegant and methodical experiments—unlike an awkward bull. Here’s what they learned:
- With isotopic tandem orthogonal proteolysis-activity-based protein profiling on mouse liver lysates, they found privileged FA-sensitive cysteine sites on proteins involved in FA-dependent pathways, one-carbon metabolism, and methylation
- MAT1A was one of the interacting proteins identified in the profiling experiment, and it is involved in making S-adenosylmethionine (SAM), a methyl donor that methylates DNA and histones
- FA mostly reacted with a few cysteines on MAT1A, including Cys120, which were modified by excess FA and not by other aldehydes
- FA was specific for MAT1A compared to the MAT2A isoform; the effects were dose dependent and affected its catalytic activity more than its binding ability
In the next agile experiments, the nimble team knocked out MAT1A and MAT2A with CRISPR-Cas9 in two different sets of HepG2 cells. They supplemented the media with ethionine so that when methylation occurred, methyl group donation by SAM would result in formation of S-adenosylethionine (SAE)—which would build up—instead of S-adenosylhomocysteine (SAH). The media was also one-carbon restricted so that methyl units would be depleted. Sure enough, cells with only MAT1A had decreased SAE production with increasing FA, meaning less methylation was occurring.
In a mouse model that chronically overproduced FA, the team found:
- A decreased SAM/SAH ratio compared to wild-type
- Histones K4 and K79 had less mono- and dimethylation
- Certain sites in the Mat1a promoter were less methylated in the mouse model than in wild-type
- Immunoblotting showed much higher levels of MAT1A compared to wild-type, but MAT2A levels were the same as wild-type, so it wasn’t just that FA was making MAT1A switch isoforms
- mRNA trends were similar, suggesting that something was going on at the level of transcription, where they also identified transcription factors involved in this process
So, it seems that the bull in this china shop prefers to mess with proteins and processes associated with methylation, and the team’s working model is a feedback loop in which FA reduces MAT1A activity, reducing methylation of the promoter region of Mat1a itself, which causes more MAT1A to be made to compensate for this.
“We have discovered that formaldehyde has the capacity to modify the epigenetic landscape of our cells, which might contribute to the well-documented carcinogenic properties of formaldehyde,” says Lucas Pontel, one of the co-authors, who is at Josep Carreras Institute.
It’s no bull—you can see for yourself at Science, November 2023.