A major advance in our understanding of heredity and genetics is recognizing that organism characters are produced by networks and not by individual "gene products." The idea of single-gene determination of phenotypic traits has yielded to systems biology.
As my friend Adam Wilkins expressed it writing about the features of multicellular organisms: "A mutation that affects a developmental process does so by affecting either a gene whose product acts as an upstream controlling element, an intermediary connecting link, or as a downstream output of the network that governs the trait's development." Multimolecular networks carry out every kind of vital activity. These include metabolism, biosynthesis, damage repair, sensing, signaling, cell division, cell differentiation, tissue formation and multicellular development.
In the early days of molecular biology, theorists like Francis Crick imagined that we could have a cellular division of labor where certain kinds of molecules (nucleic acids or DNA and RNA) provided the information and encoded the proteins, which did all the work. This was the basic idea behind his famous "Central Dogma of Molecular Biology" and underlies genetic determinism, the philosophy that "your genes are your destiny."
We now have a more sophisticated view and realize that cell networks involve different interacting molecular partners: proteins, DNA, RNA and many other biochemical classes. These partners fit together, modify each other's shapes (and sometimes even chemical composition), and combine to build higher-order structural complexes. Think of it as molecular Legos.
All the interactions between the various components determine what job each network does and how it processes regulatory information essential to making the task come out right for survival, growth and reproduction. It is a form of molecular circuitry, analogous to electronics, where different pieces are put together in defined arrangements for specific tasks.
Each functional network has its own architecture. The components are usually encoded by many different parts of the genome, and some of them are often particular DNA sequences located at different places. In order to execute most vital processes, for instance, dozens of proteins have to locate hundreds of discrete sites in the genome to produce the various molecules needed.
Switching from gene-based to network-based thinking raises many questions about how complex interactive networks change in the course of evolution. We know that certain types of networks have been copied and reused in modified form to take on new tasks in the course of evolution. Multicellular organisms emerged with organized body plans that have become increasingly complicated. Orchids have proliferated countless exotic and beautiful flowers. An eye development network has been adapted to control wing patterns in butterflies.
How do genomes encoding complex networks evolve to provide novel functions in any reasonable period of time? According to the conventional view of evolution as a random succession of independent, localized, slight changes, each providing its own selective advantage, the evolutionary process appears hopelessly slow and subject to taking too many steps backwards before it takes one forward.
The ideal solution would be rapid ways of copying and modifying the DNA encoding the network components on a wholesale basis. Although such processes were ruled out a priori by the founders of conventional evolutionary theory, they are just what molecular biologists studying DNA change and genome sequences have discovered: duplication and natural genetic engineering processes for amplifying and distributing DNA encoding network components throughout the genome.
In other words, natural genetic engineering is a way to build molecular circuits, Lego-like, rapidly. Let us see how this works and some of what has been documented so far.
The first step is often, but not always, duplication of the entire genome. This is one of the evolutionary benefits of sexual reproduction because rare couplings, either within a species or between different species, produce individuals that have undergone whole genome doubling (WGD). WGD has been documented in yeasts, other fungi, protists, and even vertebrates. WGD is extremely common in plants and is now recognized as a major factor in Darwin's "abominable mystery," the rapid diversification of seed plants.
One advantage of WGD is that it automatically duplicates all cell networks, thus having one copy to do its original job and leaving another free to adopt a new task.
Further steps have to include rewiring the proteins, DNA sites and RNA molecules in new ways. For proteins, as we discussed in a previous blog, the rewiring is facilitated by the modular organization of proteins into domains. Since molecular interactions are domain-specific, domain swapping can change the connectivity of a protein to other proteins or to DNA.
DNA recognition sites are often carried on a particular class of natural genetic engineering tools called "mobile elements," which can duplicate themselves and move in one step to new locations in the genome. Many network-specific interaction sites have been localized to various mobile element types, and their role in functional networks has been documented, including one needed for pregnancy. The capacity of mobile elements to coordinate distant genetic loci was first described 55 years ago by Barbara McClintock.
We know less about regulatory RNAs because our recognition of their importance is so recent. But the available evidence identifies mobile elements as a major source of these so-called "non-coding" RNA molecules. Many mobile elements are highly variable and can modify and then disperse regulatory RNA sequences rapidly.
In this blog, we have seen that molecular biology has taught us deep lessons about two aspects of genome function:
(1)Genomes influence organism characteristics by encoding networks in a distributed fashion, not by the "one gene - one trait" model envisaged by early 20th Century genetics pioneers;
(2)Genomes change structure by rapid duplication and natural genetic engineering events that can happen at multiple locations (or even over the entire genome) in a short period of time.
What I find gratifying is that both these 21st-century ways of thinking about the genome fit so well in elaborating an evolutionary scenario. Although the scenario contains many gaps, the DNA record gives confidence that it points us in the right direction.