endonuclease–cognate DNA, which was the first complex to indicate the potential role of induced fitof the
DNA in catalysis.32–34The enzyme is dimeric and engages a palindromic DNA target sequence 5-GAATTC-
3 . At the dyad centre of symmetry, there is a ‘kink’ deformation in which the DNA is unwound by roughly
25° and the central ApT base step becomes partially unstacked. More modest bends occur at other steps, and
the cumulative effect of the bends and kink is to open the major grooveand improve its accessibility to direct
protein–base contacts. The kink appears to be required to orient the phosphate backbone for hydrolytic attack.
By use of base analogues, Bernard Connolly and Linda Jen-Jacobson^34 found that removal of the N-6amino
group of the A residue at the central ApT step increased enzyme affinity by 4.0 kJ mol^1 , despite the loss
of a hydrogen bond between the protein and the base. This removal results in relaxation of any steric hin-
drance in the kinked conformation, and may also cause a re-distribution of partial charges on the purine
ring. How does the enzyme achieve its specificity for this particular sequence, despite the DNAs being
energetically costly to deform? The explanation may lie with the kink at the centre, since the deformation
of alternative base steps may have even greater energy penalties. Consequently, other DNA sequences
make comparatively poorer substrates.
The principle that recognition occurs through DNA distortion is also borne out by the structure of the
EcoRV endonuclease in complex with target DNA. Like EcoRI, EcoRV is dimeric, and it binds to a similar
palindromic target site 5-GATATC-3. Also like EcoRI, the EcoRV–DNA complex reveals some striking
deviations from uniform helical structure. The bases at the central TpA step of the target are unstacked and the
major groove is compressed. This deformation brings the phosphate group near the dyad into the immediate
vicinity of the catalytic residues. The bases at the TpA step are recognized viahydrophobic contacts to the
thymine methyl groups; but their mechanical properties are also an important aspect of recognition. The TpA
step is easily distorted into a high-roll conformation and the specificity in EcoRV relies on this distinctive
property. These deviations produce an extensive surface-area contact between the protein and DNA, and
this complex has one of the largest water-filled cavities of comparable protein–DNA complexes.^8
While EcoRI binds to both target and non-target DNA with nearly equal affinity, EcoRV discriminates
against non-target DNA at both the binding and the catalytic events.^23 Nevertheless, both binding and cataly-
sis involve indirect readout. Studies with base analogues show that several functional groups in the DNA
that are not contacted in the crystal structure nonetheless reduce enzymatic activity when they are deleted,
principally by decreasing the catalytic cleavage rate (kcat).^35 Other studies with base analogues indicate that
direct and indirect readouts contribute roughly equally in the discrimination of specific versusnon-specific
targets, which corresponds to about 59 kJ mol^1 binding energy.^23
A comparison of the structure of EcoRV in the free form and in complex with DNA shows that the protein
also undergoes some structural remodelling upon specific complex formation. Thus, both protein and DNA
interactively change structure to form the specific complex.
In contrast to EcoRI and EcoRV, the related restriction enzyme HincII has apparently less stringent
specificity, since it recognizes the sequences 5-GTPyPuAC-3, where Py ( pyrimidine) can be either C
or T and Pu ( purine) is either G or A. As for the other enzymes, recognition involves indirect readout.
Here, the DNA conformation is to some extent ‘self-recognized’via interactions of the central purines, which
stack even though they are on opposite strands.^20 This self-recognition viacross-strand purine–purine stacking
probably explains the ability of the enzyme to distinguish Py–Pu from the other possible base step combin-
ations at the central position of the recognition sequence.
10.5.2 DNA-Repair Endonucleases
The DNA of all organisms is modified continually by the spontaneous deamination of cytosine residues,
or by the consequences of photoactivation or by the actions of harmful chemicals (Chapter 8). In every
human cell, roughly 10^4 DNA bases are repaired every day.
The major DNA-repair endonuclease APE1 cleaves DNA at sites where bases are missing. The crystal
structure defines its active site and has helped to identify the way missing bases are recognized (Figure
10.13). Just as in the case of the bacterial restriction enzymes, the mechanism of action involves an acti-
vated hydroxyl group, which is directed to attack the scissile phosphate. Most enzymes have poor affinity
Protein–Nucleic Acid Interactions 405