EPFL researchers help complete the largest map yet of how human cells read DNA and reveal how chemical marks can change the meaning of the genetic instructions.

Every cell in the body contains essentially the same DNA; and yet, a brain cell behaves differently from a muscle cell or an immune cell. The difference lies largely in how each cell reads its genetic instructions.

Proteins called transcription factors bind specific DNA sequences and help control when and where genes are active. They direct processes ranging from embryonic development to immune function. When this regulation goes wrong, disease can result. But although the human genome contains around 1,600 transcription factors, the DNA-binding preferences of many have remained unknown.

An international collaboration led by Timothy Hughes at the University of Toronto has now filled many of these gaps in a study published in Nature. The researchers combined five experimental platforms with computational analyses, performing more than 4,800 experiments, identifying DNA-binding motifs for 177 transcription factors that were poorly characterized before, adding around 130 distinct motifs to the known vocabulary of human gene regulation.

The work has generated a “Codebook”: the most comprehensive catalogue of human transcription factor binding preferences assembled to date.

Bart Deplancke’s lab at EPFL played a central role in the development of the Codebook. “Our genomes contain the instructions for life, but understanding how cells actually read those instructions has remained one of biology’s major challenges,” says the professor, who is a co-corresponding author of the Nature study. “This work brings us much closer to a complete dictionary of the proteins that control gene expression.”

A second layer of information

But reading DNA is only part of the story. DNA also carries chemical modifications that influence how genes are regulated without changing the sequence itself. One of the best known is DNA methylation.

In a companion paper published in Nature Communications, Deplancke’s team developed meSMiLE-seq, a microfluidic method that compares transcription factor binding to methylated and unmethylated DNA in the same experiment.

The researchers used the method to study 114 transcription factors and obtained DNA-binding models for 48. Fourteen showed greater affinity for methylated DNA or recognized alternative methylation-dependent motifs, while 13 showed reduced affinity for methylated sequences. Comparisons with cellular data indicated that methylation helps direct some transcription factors to different locations in the genome.

“DNA methylation does not simply switch genes on or off,” says Antoni Gralak, the study’s first author. “It changes how the genome is interpreted. Our work shows that many transcription factors effectively read an additional layer of information that sits on top of the DNA sequence.”

DNA: a user’s guide with annotations

Together, the two studies expand the catalogue of DNA sequences recognized by human transcription factors and show how DNA methylation modifies some of these interactions. The Codebook establishes where transcription factors can bind based on DNA sequence, while the meSMiLE-seq study shows that this recognition is often further modified by chemical marks on the DNA itself.

“This richer understanding could improve how scientists interpret genetic variants associated with disease, explain why identical DNA sequences behave differently in different cell types, and ultimately help researchers better understand development, ageing and diseases such as cancer,” say Deplancke.

“Rather than reading DNA as a fixed instruction manual, the work suggests that cells interpret the genome more like a living document in which both the letters themselves and the chemical annotations written on top of them determine the final message,” he adds.

Other contributors

Altius Institute for Biomedical Sciences, USABC Children’s Hospital Research Institute, CanadaBiosoft.Ru LLC, RussiaCharles University, Czech RepublicCincinnati Children’s Hospital Medical Center, USAFederal Research Center for Information and Computational Technologies, RussiaGénome Québec Innovation Centre, CanadaInstitute of Biochemistry and Genetics, Ufa Federal Research Centre of the Russian Academy of Sciences, RussiaInstitute of Organic Chemistry and Biochemistry, Czech Academy of Sciences, Czech RepublicInstitute of Protein Research, Russian Academy of Sciences, RussiaLife Improvement by Future Technologies Center, RussiaLomonosov Moscow State University, RussiaMartin Luther University Halle-Wittenberg, GermanyMax Planck Institute of Biochemistry, GermanyMcGill University, CanadaMemorial Sloan Kettering Cancer Center, USAMoscow Center for Advanced Studies, RussiaRadboud University Medical Center, NetherlandsSirius University of Science and Technology, RussiaSwiss Institute of Bioinformatics, SwitzerlandTUD Dresden University of Technology, GermanyUniversité de Lausanne, SwitzerlandUniversity of British Columbia, CanadaUniversity of Cincinnati College of Medicine, USAUniversity of Southern California, USAUniversity of Toronto, CanadaVavilov Institute of General Genetics, Russian Academy of Sciences, RussiaVictor P. Dahdaleh Institute of Genomic Medicine, Canada