Type I restriction endonucleases show no turnover following DNA cleavage and as a consequence a study of the exact mechanism of DNA cleavage has been difficult and protracted in part due to the limiting amounts of enzyme available. However, despite these problems a model for DNA cleavage exists based on observations with the classical EcoKI system. The basis of this model is the observation that DNA is not cleaved at specific locations, as occurs with Type II restriction endonucleases, but is cut into fragments of random size. The original proposal was that the ATPase activity of the endonuclease led to translocation of the DNA, for random lengths of time, and then cleavage occurred producing random-sized fragments. However, by synchronizing the translocation process by addition of ATP, Studier and Bandyopadhyay (1988) were able to demonstrate that cleavage occurs at a position midway between two adjacent sites and proposed a cooperative cleavage event involving the interactions of two molecules following DNA translocation. This model was proposed following work with linear DNA, and we have recently refined this model for cleavage of covalently closed circular plasmid DNA.
Digestion of linear DNA by Type I restriction endonucleases is generally activated following the head-on collision of two translocating enzymes (Studier & Bandyopadhyay 1988). However, as reported by Szczelkun (2002), the resulting distributions of cleavage loci along the DNA vary with different enzymes; in some cases, cleavage is located in a discrete region midway between a pair of recognition sites while in other cases cleavage is broadly distributed and occurs at nearly every intervening locus. Statistical models for DNA translocation, collision, and cleavage are described that can account for these observations and that are generally applicable to other DNA-based motor proteins. If translocation is processive (stepping forward is significantly more likely than DNA dissociation), then the linear distribution of an ensemble of proteins can be described simply using a Poisson relationship. The pattern of cleavage sites resulting from collision between two processive Type I enzymes over a distance d can then be described by a binomial distribution with a standard deviation 0.5d1/2. Alternatively, if translocation is non-processive (stepping forward or dissociating become equally likely events), the linear distribution is described by a continuum of populated states and is thus extended. Comparisons of model data to the kinetics of DNA translocation and cleavage discount the non-processive model. Instead, the observed differences between enzymes are due to asynchronous events that occur upon collision. Therefore, Type I restriction enzymes can be described as having processive DNA translocation but, in some cases, non-processive DNA cleavage.
In a recent paper Jindrova et al (2005), have presented a more extensive analysis of the process of DNA cleavage. DNA ends produced by EcoKI, EcoAI and EcoR124I, members of the Type IA, IB and IC families, respectively, were characterised by cloning and sequencing the restriction products from reactions with plasmid DNA substrate containing a single recognition site for each enzyme. They showed that all three enzymes cut this substrate randomly with no preference for a particular base composition surrounding the cleavage site, producing both 5'- and 3'-overhangs of varying lengths. EcoAI preferentially generated 3'-overhangs of 2–3 nt, whereas EcoKI and EcoR124I displayed some preference for the formation of 5'-overhangs of a length of 6–7 and 3–5 nt, respectively. A mutant EcoAI endonuclease assembled from wild-type and nuclease-deficient restriction subunits generated a high proportion of nicked circular DNA, whereas the wild-type enzyme catalyzed efficient cleavage of both DNA strands. Therefore, it can be concluded that Type I restriction enzymes require two restriction subunits to introduce DNA double-strand breaks, each HsdR subunit nicking one strand of the DNA.
The catalytic nature of Type I enzymes was investigated more thoroughly by Bianco et al (2009) who showed turnover of the enzyme EcoR124I following dimerisation, cleavage and subsequent isolation of that enzyme from DNA - the isolated enzyme was able to carry out further rounds of cleavage. This dimerisation has also been observed with the enzyme EcoKI using Atomic Force Microscopy (Neaves et al 2009), which involves the translocating form of the enzyme (Berge et al 2000). However, both groups have shown that dimerisation is not essential for cleavage, which suggests that the nature and reason for this dimerisation is yet to be fully understood. One interesting observation with EcoKI is that single molecule studies of translocation, using a Magnetic Tweezer setup, were unsuccessful suggesting that dimerisation may be important to this function.
Last modified on
21 September 2011
© Dr Keith Firman
Author Dr Keith Firman.