Researchers from the Salk Institute for Biological Studies, led by Brendan Miller and Alan Saghatelian, compiled a catalog of microproteins - short chains of amino acids, the building blocks of protein - in the frontal cortex of the human brain. The work was published in the journal Nature Aging. It turned out that many such molecules are represented differently in people with and without symptoms of Alzheimer's disease.
Microproteins are shorter than 150 amino acids. Because of their small size, they are hard to detect with standard laboratory methods, and for a long time such sequences were not included in standard protein catalogs. They are often encoded within or next to large known genes, so in routine analysis activity in these regions is attributed to the "parent" gene.
First the team created its own database: it analyzed RNA, the working copies of genes, in frontal cortex tissue from 12 deceased people and added the discovered instructions to the standard human protein reference. Then, using mass spectrometry - a method that identifies proteins by the mass of their fragments - the researchers studied 610 brain samples and found 4,321 microproteins, more than 3,000 of which were not listed in validated catalogs.
To confirm that the findings were real, the scientists compared the data with ribosomal profiling, which shows where cellular "protein factories" are actively reading genetic instructions, and applied the deep learning tool PROSIT to filter out false fragments. After filtering, 1,067 previously undescribed microproteins remained with high confidence that they are present in the brain.
The researchers then compared samples from people without Alzheimer's disease and with its symptoms. In patients with symptoms, dozens of microproteins had altered levels. For example, the microprotein of the BCL3 gene was less abundant, and the microprotein of the RAB3C gene was more abundant. When introduced into lab-grown human microglia, the main immune cells of the brain, these molecules remained stable and occupied specific areas inside the cells.
It turned out that about half of the microproteins are regulated independently of their large "parent" genes: the RNA level of the microprotein changed in Alzheimer's disease, while the RNA of the main gene remained the same. This was confirmed in a separate set of brain samples. The likely cause was named as alternative splicing - a process in which a cell "cuts and splices" RNA regions differently, producing different protein products.
Particular attention was paid to the microprotein of the MKKS gene. Standard databases attribute to this gene a large protein associated with a rare developmental disorder, but in the adult human frontal cortex it was the short microprotein that dominated. In people with Alzheimer's symptoms, its level was lower. The activity of the microprotein coincided with the work of genes responsible for mitochondria, the energy stations of the cell. When, using CRISPR, a gene-editing technology, the region producing this microprotein was removed from laboratory microglia, the cells began to consume less oxygen and produce less energy.
In addition, the researchers found signs that microproteins are also produced from pseudogenes - altered DNA sequences that are usually considered inactive. One of them, originating from the pseudogene ACTBP7, when added to microglia changed the distribution of the cell's structural fibers.
The authors emphasize limitations. The absence of a protein in mass spectrometry data does not mean it is not in the brain: it may have been below the detection threshold. An altered level of a microprotein in Alzheimer's disease does not by itself prove that it causes the disease: such shifts may be a cell's response to damage or a regulatory failure. To distinguish causes from consequences, further experiments with individual microproteins in living systems are needed.