Scientists have uncovered a fascinating piece of the puzzle surrounding the origins of human language, suggesting that a variant of a protein found only in modern humans might have played a role in the development of our ability to speak. In a study published in Nature Communications, researchers discovered that replacing a single building block in the protein NOVA1 with its human-specific version altered the vocal sounds that mice make. Notably, this human-specific variant is absent in Neanderthals and Denisovans, our closest extinct human relatives.
“This gene is part of a sweeping evolutionary change in early modern humans and hints at potential ancient origins of spoken language,” said Robert B. Darnell, head of the Laboratory of Molecular Neuro-Oncology at The Rockefeller University and senior author of the study. “NOVA1 may be a bona fide human ‘language gene,’ though certainly it’s only one of many human-specific genetic changes.”
The origins of human language remain one of the most enduring mysteries in science. While our capacity for complex communication clearly sets us apart, the specific genetic and biological mechanisms that enabled this ability are still largely unknown. We know that our close relatives, such as Neanderthals, possessed some features that might have allowed for spoken language. Their throat and ear anatomy, for example, appears to have been capable of producing and perceiving speech sounds. They also shared with us a version of a gene associated with speech ability. However, modern humans are unique in having expanded brain regions that are absolutely essential for both producing and understanding language.
To explore the genetic underpinnings of spoken language, researchers at The Rockefeller University focused on the protein NOVA1, which is known to be important for brain development. They were intrigued by the fact that humans have a slightly different version of this protein compared to other animals, including our closest extinct relatives. The research team hypothesized that this human-specific alteration in NOVA1 could be connected to the development of spoken language in our species.
To determine if a single change in NOVA1 might have contributed to our unique language abilities, the scientists used a gene-editing technique to create a line of mice that carried the human form of NOVA1. In most mammals and ancient humans, NOVA1 contains a specific building block known as isoleucine at position 197. In modern humans, however, this building block is replaced by valine—a very similar molecule, but one that appears to be unique to our species.
The researchers replaced the mouse version of NOVA1 with the human variant by injecting editing tools into fertilized mouse eggs. They verified that only the desired change was made and that no unintended mutations had occurred. The resulting mice grew up normally, showed typical brain development, and were healthy and fertile, making them an ideal model for studying the effect of the human variant.
Once the humanized mice were bred, the research team set out to see if the change in the NOVA1 protein affected the animals’ behavior, especially the sounds they make. Young mice, when separated from their mothers, emit ultrasonic calls that help attract attention, and adult male mice produce vocalizations during courtship. The scientists recorded these sounds using specialized microphones and analyzed them with computer programs that measured features such as pitch, duration, and complexity.
In addition to the behavioral tests, the team examined the brains of the mice to determine if the protein change affected the processing of genetic information. They employed several laboratory techniques to identify the regions where NOVA1 attaches to other molecules in the brain and to assess how the change influenced the splicing—or cutting and joining—of certain genes during the production of messenger molecules. In simple terms, the researchers looked at whether the human-specific change in NOVA1 caused differences in the patterns of RNA messages that direct brain function, particularly messages linked to the production of sound.
The study revealed that the human-specific change in NOVA1 did not disturb the overall ability of the protein to bind to its RNA partners or affect general brain development. Both the humanized mice and their normal littermates had similar levels of NOVA1 in their brains, and most genes were expressed at similar levels.
However, a closer look at the details uncovered important differences in the way certain genes were processed. In the humanized mice, several changes were observed in the splicing patterns of genes—especially those known to be involved in the neural circuits for vocalization. In other words, while the basic function of NOVA1 remained intact, the human version appeared to fine-tune RNA messages in regions of the brain that help control the production of sound.
Behavioral tests of the mice’s vocalizations further supported these findings. When isolated, mouse pups from the humanized group produced calls with a different pattern compared to those from the control group. Although the total number of calls was similar, details of the sounds—such as pitch and the “shape” of the calls—were noticeably altered.
“All baby mice make ultrasonic squeaks to their moms, and language researchers categorize the varying squeaks as four ‘letters’—S, D, U, and M,” Darnell noted. “We found that when we ‘transliterated’ the squeaks made by mice with the human-specific I197V variant, they were different from those of the wild-type mice. Some of the ‘letters’ had changed.”
In one test, the researchers classified the mouse calls into different types based on changes in pitch. They found that the calls of pups carrying the human form of NOVA1 exhibited a shift toward a higher frequency in certain call types. Adult male mice, when exposed to female mice in estrus, also produced vocalizations that differed in subtle ways—for instance, some simple call types were slightly longer and had lower starting and ending pitches.
“They ‘talked’ differently to the female mice,” Darnell said. “One can imagine how such changes in vocalization could have a profound impact on evolution.”
Additionally, the more complex sounds that featured a variety of pitch jumps showed greater variation in frequency. These differences suggest that the human version of NOVA1 can influence vocal patterns in a way that may be connected to the evolution of more refined spoken language.
The researchers also explored whether this single change in NOVA1 could be linked to the broader picture of human evolution. They compared genetic data from modern humans, Neanderthals, and Denisovans. Their analysis confirmed that the change from isoleucine to valine at position 197 is found only in modern humans. In a large survey of more than 650,000 human genomes, almost all individuals carried the human version of NOVA1. The few exceptions had the ancestral version, suggesting that the human change spread rapidly through ancient populations, possibly because it offered an advantage in vocal communication.
“Our data show that an ancestral population of modern humans in Africa evolved the human variant I197V, which then became dominant, perhaps because it conferred advantages related to vocal communication,” Darnell said. “This population then left Africa and spread across the world.”
While the study presents exciting evidence that a single protein change may have contributed to our unique language abilities, the researchers are cautious in interpreting their findings. The work was done in mice, and although mice produce sounds that can be compared to human vocalizations in a broad sense, the complexity of human speech is far greater. It remains to be seen exactly how the altered RNA processing observed in the mice would translate into the rich and varied speech of our species. The experiments were designed to study the immediate impact of the human change on vocal behavior and RNA patterns in the brain, but further research is needed to understand how these molecular effects connect to the higher-level functions of language and communication.
The team plans to extend their work by investigating how the human version of NOVA1 might interact with other proteins and influence neural circuits in brain regions involved in speech and communication. They are also interested in examining whether similar genetic changes occur in other proteins that regulate brain development and behavior, as well as exploring the potential role of NOVA1 in human conditions that affect speech and language, such as developmental delays or certain neurodevelopmental disorders.
The study, “A humanized NOVA1 splicing factor alters mouse vocal communications,” was authored by Yoko Tajima, César D. M. Vargas, Keiichi Ito, Wei Wang, Ji-Dung Luo, Jiawei Xing, Nurdan Kuru, Luiz Carlos Machado, Adam Siepel, Thomas S. Carroll, Erich D. Jarvis, and Robert B. Darnell.
