Our genetic material is littered with parasitic DNA sequences, known as transposons, which promote their own propagation and transmission, rather than their host’s1. Their movements (transposition) within and between genomes have profound consequences for the genetic code — sometimes leading to diseases, but also driving genetic diversity and evolution2. Transposase proteins mediate this movement by executing all the required chemical reactions.
Strikingly, these proteins have been repeatedly repurposed throughout evolution to produce new biological functions that benefit the host. A prime example is the vertebrate immune system, in which transposase-like RAG proteins help to assemble new genes from three pools of interchangeable DNA parts (known as V, D and J gene segments). This process is called V(D)J recombination, and equips immune cells with a diverse set of sensors that can recognize many threats3. Writing in Nature, Liu et al.4 report a series of structures of a transposase that is an ancestor5 of RAG, casting light on the evolutionary history of these proteins.
Transposases must recognize several DNA sites, and then cut and join them in the proper order6. To understand this multi-step process, we need to visualize the structures of the molecular machinery involved at all stages, which is a major technical challenge. Liu et al. have now drawn on a powerful combination of two techniques — X-ray crystallography and single-particle cryo-electron microscopy — to picture several steps of transposition in remarkable detail, thereby providing a molecular ‘movie’ of the process.
The authors’ achievements build on many years of structural studies of transposases7–11 and RAG12–15, providing an increasingly complete view of their functions and helping to connect the dots between the ‘selfish’ DNA rearrangements of transposases and the essential functions that evolved from them. Transposases are now known to have a catalytic core unit and diverse extra parts that bind to DNA or control transposase function6. They usually act in pairs, with each dimer holding two segments of the transposon DNA .
The transposase studied by Liu et al. mediates the movement of a transposon called Transib, and was isolated from a moth (Helicoverpa zea). Appropriately, the shape of the transposase complex resembles that of a moth: each ‘wing’ comes from one of the two transposase molecules in the complex, and is formed mainly from a protein region called the zinc-binding (ZnB) domain. The wings provide many of the interactions with the DNA, which forms the ‘antennae’ (Fig. 1).
As with many transposons, Transib moves by a cut-and-paste process: its transposase cleaves it out of the genome, using a catalytic core present in many transposases, and inserts it elsewhere in the genome16. Liu and co-workers’ structures of five steps in Transib transposition now reveal the remarkable conformational changes in the protein during this process.
Perhaps most notably, the authors find that the ZnB wings move constantly: they unfurl when transposon DNA first arrives, and then close and open again during the rest of the process. This ‘flapping’ accompanies some impressive DNA acrobatics, which brings different DNA parts into the protein’s core for ordered cleavage and joining. Remarkably, the ZnB domains help to capture not only the transposon, but also the target molecules into which Transib will be inserted — first opening to make space for the molecules, and then closing to plug the region of the target DNA into which Transib will be integrated into the core. The structures also show that the end section of the protein (the carboxy-terminal tail; CTT) contains three short α-helices that form an accordion-like structure, which connects the moving wings to the complex’s ‘body’. Previously reported transposase structures6–11 have revealed similar overall features, but the movements in the Transib transposase are much more extensive.
RAG protein complexes consist of two transposase-like RAG1 proteins and two RAG2 proteins. In these complexes, ZnB is present in RAG1, but is more fixed than in the Transib transposase; and no part is present13 that is analogous to CTT. RAG2, which is essential in V(D)J recombination but absent in most transposases, sits above the wings of RAG1 and holds a large part of the DNA12 — much as ZnB and CTT do in the Transib transposase.
Unlike transposases, the RAG complex cuts and removes the distinct DNA sequences found between the V, D and J gene segments, tightly coordinating the process to ensure that different types of segment are subsequently connected. In cells, RAG also largely stops the removed DNA from being reinserted elsewhere in the genome, to prevent potentially harmful changes to the genetic code. But how did these functions evolve? A logical proposal implicates RAG2, but a recent study11 of ProtoRAG — a relative of RAG found in invertebrates that contains RAG2 but still acts as a transposase — shows that things are more complicated. Elements in both RAG1 and RAG2 help to coordinate DNA cleavage and prevent insertion.
Liu and colleagues’ findings cast fresh light on the role of RAG2, showing that it carries out many of the functions of ZnB, but increases the rigidity of the whole RAG complex, compared with that of the transposase complex. It binds the DNA at the V, D and J segments more tightly than ZnB binds at the transposon, and does not undergo such large conformational changes (which can require a lot of energy, and thus reduce efficiency). This increased rigidity and tight binding might help to ensure the strict molecular coordination required for V(D)J recombination. It might also prevent release of cleaved DNA segments and/or stop the wings from reopening to accept any other DNA molecules — thereby preventing removed DNA from being reinserted elsewhere in the genome. If the wings do not open, then any incoming DNA would have to bend itself to an angle of about 150° before entering the protein, which is not easily done.
Note that Liu et al. were not able to directly observe the structure of the transposase in complex with intact target DNA. It therefore remains to be seen whether target DNA first binds to the transposase in a relaxed form and is then forced into a severe 150° bend. The authors also did not observe a complex in which the transposase binds intact transposon DNA such that the catalytic core is close enough to the ends to cleave them; instead, the authors observed intact transposon DNA bound with its ends away from the catalytic centre. In RAG, a large twist in the DNA is needed to position its breakpoints accurately for the cuts14. A similar twist might occur in Transib, but other explanations are also possible.
Efforts are now needed to define the exact functions of RAG2. Curiously, the cell-free RAG complex can readily insert excised DNA into another DNA molecule17,18 (a target DNA). Structures of RAG with a bound target DNA must therefore be obtained — ideally, both with the intact target and after insertion. These structures will show whether the target DNA becomes as sharply bent as it does in the Transib transposase, and reveal how RAG2 affects the binding of target DNA and its insertion of excised DNA.
Other proteins might be needed to promote the function of RAG. This possibility has previously been investigated, but the availability of new structures and methods provides further opportunities for research. For example, large molecular assemblies can now be studied inside cells using a technique called electron tomography19, and molecular interactions can be probed with advanced mass-spectrometry methods20. Analysis of genomic data from different species will also be helpful in identifying ancestors of RAG proteins other than ProtoRAG and the Transib transposase, and thereby exploring their evolutionary history. Such research will help to explain how parasitic genetic elements can be repurposed for crucial biological functions.
Nature 575, 447-448 (2019)