Restriction endonucleases are bacterial enzymes that bind to double-stranded DNA at specific "target" sequences and cut the strands of DNA either within the target sequence or nearby. They are a part of the immune system of bacteria, preventing the accumulation and potentially toxic incorporation of foreign or viral DNA in the bacterium. These enzymes were initially discovered in the 1970s and their ability to recognize and cleave specific DNA sequences has revolutionized research and led to the creation of many diagnostic techniques in molecular biology, biochemistry, and medical genetics. Almost 50 years later, restriction endonucleases are still routinely used for genetic research in model organisms and in test tubes. Additionally, there is interest in engineering restriction enzymes to make targeted modifications to DNA for agricultural and clinical applications.
Previously, Betty Shen, a staff scientist in Dr. Barry Stoddard's Laboratory (Basic Sciences Division), determined the structure of the HNH restriction enzyme PacI in complex with DNA. PacI recognizes an 8 basepair target site consisting entirely of A:T basepairs (5’-TTAAT/TAA-3’, where a “/” indicates the site of cleavage) and cleaves it to produce 2-basepair 5' overhangs. In the structure of PacI binding to DNA, they discovered a dramatic and unusual rearrangement of the DNA basepairs in the target site. "Because we wished to study if that observation (which had never been seen before or since) was a function solely of the enzyme, or if its unusual A:T-rich target site sequence might play a role in its own unpairing and perturbation, we searched for an unrelated restriction enzyme that targets a similar sequence," said Stoddard. The restriction endonuclease SwaI, which is found in the bacteria Staphylococcus warneri, became their top choice. In their recent publication in Nucleic Acids Research, the scientists report several high-resolution (~2Å) structures of SwaI and compare them to structurally similar endonucleases as well as the PacI endonuclease, which has a completely different structure but recognizes a similar 8-basepair A:T rich sequence.
SwaI recognizes and cleaves the 8 basepair palindrome 5'-ATTT/AAAT-3' and the cleavage generates blunt ends. Betty Shen used x-ray crystallography to determine the structure of SwaI alone, in complex with uncut DNA with calcium ions, and bound to fully cut DNA with magnesium ions. The enzyme is composed of two identical nuclease subunits that contact each other primarily through a long C-terminal helix. In the enzyme alone structure, the two C-terminal helices form an antiparallel helix bundle reminicent of a coiled coil, leaving a large open chasm between the two nuclease domains. In the DNA bound structures, the C-terminal helices curved slightly and each of the nuclease domains went through a rigid-body rotation at the junction where the helix meets the nuclease core, resulting in the closure of the chasm with the two nuclease domains, wrapping the enzyme tightly around the dsDNA at the target side and causing the DNA to bend and unravel at the two central basepairs. When SwaI binds DNA, a flexible loop spanning 11 residues (aa24-35) becomes ordered and the arginine (R35) in that loop contacts DNA bases in the target site from the minor groove. The relatively few number of contacts between the enzyme and the DNA is notable--just 6 amino acids of SwaI are involved and 4 of the 6 contact multiple, neighboring DNA bases. Upon SwaI binding, the target site DNA bends sharply and the two central A:T basepairs become unusually positioned: the adenines "leapfrog" to form a stack with their purine rings while the unpaired thymidines remain positioned relatively closely to their 3' neighboring bases.
Searching for proteins with similar structure to SwaI using the DALI and FATCAT servers, the scientists found that the molecules exhibiting the highest structural homology are the endonucleases R.HincII and R.EcoRV. Both of these enzymes are formed of two identical nuclease subunits and there is a relatively small root-mean-square-deviation(RMSD) between the Cα backbone of SwaI and that of HincII and EcoRV, respectively (~3Å and ~4Å). Because HincII and EcoRV only share 14% and 9% protein sequence identity to SwaI, this is a striking similarity in structure despite the dearth of homology at the sequence level. Notably, both of these enzymes, similar to SwaI, also recognize palindromic target sites and create blunt-ended DNA products.
Image provided by Betty Shen
The unusual distortion of the target site DNA bases induced by SwaI is reminiscent to that observed by PacI. While every base in the palindromic target of PacI is reorganized from its normal basepairing arrangement in the PacI-DNA complex, only the central two A:T pairs are disrupted in the SwaI-DNA structure. Most notably, in both SwaI and PacI, a crucial arginine residue plunges into the minor groove and distrupts the Watson-Crick pairing of the target DNA. In both structures, alternative A:A pairing is used to stabilized the disrupted A:T pairs – adanine p-stacking in the former case and Hoogsteen baseparing in the latter except it is between two purine bases rather than purine-pyrimidine. The complete lack of similarity between the overall tertiary structure of PacI and SwaI suggests that A:T rich target sequences may be recognized by their unique structural features such as their intrinsically higher flexibility, which translates to a a lower energy requirement for disrupting the Watson-Crick pairing, or their ability to form unique purine:purine pairings.
The Stoddard Lab's study incites further study into whether endonucleases recognize specific preexisting DNA conformations or whether the endonuclease induces the changes.
Shen BW, Heiter DF, Lunnen KD, Wilson GG, Stoddard BL. 2016. "DNA recognition by the SwaI restriction endonuclease involves unusual distortion of an 8 base pair A:T-rich target." Nucleic Acids Research.
Funding was provided by the National Institutes of Health, the US Department of Health and Human Services, and Fred Hutchinson Cancer Research Center.