Here’s the puzzle at the heart of today’s interview: why would evolution dial down an enzyme that brains rely on? In a new PNAS study, the authors trace two human-specific tweaks to adenylosuccinate lyase (ADSL)—a key node in purine metabolism. First, a protein change (A429V) that subtly destabilizes the enzyme; later, regulatory variants that lower its expression, especially in the brain. Together, these “double hits” raise the levels of ADSL’s substrates and—at least in a humanized mouse model—shift behavior: female mice, when water is scarce, get to it faster than their wild-type peers. The work hints that metabolism can tune cognition and decision-making under pressure, and that sex-specific effects may be part of the story.

The implications are provocative. The ADSL variants are now near-universal in modern humans, suggesting they were favored by selection—but for what? Energy budgeting in cortex? Sharper learning or planning under resource constraints? A trade-off akin to other dose-dependent human adaptations? The authors are cautious: mouse behavior doesn’t map neatly onto human psychology, and the mechanistic link between altered purine flux and cognition remains to be nailed down. Still, early population genetics and CSF metabolite signals point to lingering, if modest, effects in people today.

In the exchange that follows, the team walks us through the evolutionary timeline, the sex differences, the behavioral readouts, and why the next breakthroughs will likely come from large, carefully phenotyped human cohorts.

Your study shows that ADSL activity was deliberately reduced twice during modern human evolution—once through a protein-coding change and again via regulatory changes. What evolutionary advantage do you think this reduction provided?

It’s curious that the changes that reduced the ADSL activity were so strongly selected. The mouse results suggest that the amino acid change affects brain metabolism and behavior. It influences how efficiently the mice compete for water when they are thirsty, particular in females. However, what human behaviors this might translate to is hard to infer from a mouse model.

The A429V substitution spread before the Out-of-Africa migrations, while the regulatory haplotype rose later. How do these genetic timelines align with what we know about early Homo sapiens’ cultural or environmental changes?

The evidence of positive selection on the ADSL gene, together with the mouse results, suggests that the amino acid change may have conferred an evolutionary advantage to our ancestors, probably in specific tasks or environments. However, the behavioral impact of the regulatory haplotype, as well as how the two changes may have jointly influenced human behavior, remains unclear. This research will rely heavily on large-scale behavioral and genetic data from humans, larger than those available today. We hope to clarify the relationship in the near future.

Since severe ADSL deficiency causes neurological impairments, how do you explain the paradox that mild reduction may have been beneficial rather than harmful in early humans?

Such dose-effect or trade-off is not uncommon in humans, for example the malaria resistance conferred by sickle hemoglobin. In this study, we examined only the effects of the change on brain metabolism and on mouse competitive behavior for limited water resources. However, the influence on other physiological or behavioral traits, and the direction of such effects, remains to be explored. Further investigations are needed to achieve a more complete understanding of the functional consequences of this amino acid change.

Why do you think the brain and liver were particularly affected by substrate accumulation, and what might this reveal about the role of purine metabolism in human cognition?

Prof. Izumi Fukunaga: The substrate accumulation likely results from the low enzyme expression level in the brain and liver, but we do not know if this accumulation caused the behavioral change, or symptomatic. It reveals the importance of purine metabolism, but that not all cognitive functions are equally affected.

Only female mice showed stronger competition for water when carrying the modern human ADSL variant. What might explain this sex-specific effect, and do you suspect parallels in human biology?

One possibility is that the amino acid change has a larger metabolic impact in female than in male mice. Other possibilities include sex-dependent brain development or hormonal regulation of the metabolic pathways. More clues about the underlying mechanism will further reveal why this effect appears in female mice.

I suspect sex-specific effects may also exist in humans, although it is difficult to infer which aspects might be differentially affected based solely on the current mouse model.

The water-access task involves learning, planning, and decision-making. Do you believe ADSL changes influenced specific cognitive domains, or did they fine-tune general adaptability under resource pressure?

Prof. Izumi Fukunaga: Competing for water involves many cognitive functions, such as learning, planning, and decision making. General adaptability involves all these cognitive functions, so it will be intriguing to dissect exactly which aspects are particularly affected by the reduced ADSL activity.

Your analyses show clear signs of positive selection on the ADSL locus. Do you think this selection was driven by metabolic needs, cognitive advantages, or a combination of factors?

The precise selective pressures remain unclear. Given ADSL’s central role in purine biosynthesis and the brain’s especial sensitivity, both metabolic needs and brain-related factors, such as potential cognitive advantages, might have contributed to its positive selection.

With >97% of humans carrying these variants, do you think ADSL-related changes still play a role in shaping differences in cognition or behavior today, or have they become neutral remnants?

Their near-fixation doesn’t necessarily imply that they have become functional neutral. Subtle effects could still persist. Our phenotypic association studies suggest a weak and indirect link between the haplotype and human intelligence. This could imply that reduced ADSL activity continues to influence brain function, albeit modestly. 

Could insights from these evolutionary changes help us better understand or even treat pathological ADSL deficiencies in modern patients?

This work helps us better understand why the brain is predominantly affected by ADSL deficiency, potentially due to the gene’s relatively low expression there. This contrasts with the common assumption that the most pronounced effects of a gene manifest in tissues where it is most highly expressed.

Effective treatment of pathological ADSL deficiencies depends on a deeper understanding of the disease’s pathogenic origins. While several hypotheses have been proposed, including SAICAr toxicity, no direct evidence has yet established a causal relationship.

What are your plans for future work—do you intend to study ADSL changes in organoids, human populations, or other model systems to further unravel their role in human-specific traits?

Understanding its role in human populations is more interesting. However, this research heavily relies on large-scale behavioral and genetic data from humans, larger than those available today. We hope to realize it in the near future.