Egg cells can reset the genetic material of differentiated (somatic) cells, allowing them to become reprogrammed into new cell types. That ability is exploited in both reproductive and therapeutic cloning, though both processes are plagued by inefficiency. Now, scientists at the University of Cambridge in the UK have tracked a mechanism through which nuclei from specialized cells 'remember' their previous existence even as they differentiate into the multiple cell types of a cloned embryo. This may help explain why cloning is so inefficient. Reporting in Nature Cell Biology, Ray Kit Ng and John Gurdon show how this memory is preserved1.

As embryonic cells differentiate, their individual fates are marked in their genomes. So-called epigenetic processes designate swaths of genes for either silencing or expression; these designations are reversible and can change as cells differentiate. In a process called methylation, chemical tags added to particular nucleotides keep genes quiet. In another process, gene activity is controlled by how strands of DNA are associated with proteins.

Working in frogs, the researchers had previously noticed that cloned embryos made with nuclei from endoderm cells continued to express high levels of an endoderm gene even in cells not destined to become endoderm. To investigate the mechanism, the researchers blocked methylation, but the expression was unchanged. The researchers decided to explore the issue using another, better characterized gene called MyoD, which is active in muscle and known not to be methylated. They created cloned frog embryos by transferring nuclei expressing the gene into frog eggs. In 9 of 17 embryos, cells not destined to become muscle still expressed the gene, indicating that a mechanism other than methylation was responsible. Remarkably, the effects persisted over 24 cell divisions and after nuclei had been transferred into 2 series of enucleated eggs.

The researchers turned their attention to histones, the protein complexes that spool DNA. In cells that will go on to form muscle in normal embryos, MyoD is associated with a particular histone protein variant called H3.3, which is associated with active genes. Ng and Gurdon found that in embryos created by nuclear transfer, the cells expressing MyoD were also associated with H3.3. Further, increasing the amountsof H3.3 in two-cell embryos increased the number of cells expressing MyoD in the wrong lineages. However, when the researchers forced cells to make an inactive mutant version of H3.3, expression fell. This is strong evidence that DNA 'remembers' how it was wrapped around histones through multiple rounds of DNA replication. The work also shows how the process can be manipulated so that genes within a transferred nucleus 'forget' their previous epigenetic state. Though the process likely varies by gene and species, this data could illuminate techniques to reprogram specialized human cells. It also indicates that normal cell development is governed by epigenetic memory.