DNA bending caused by the binding of AraC to DNA
In this work we studied the structure of the AraC/DNA complex using AFM
to measure the DNA bend angle caused by the binding of AraC to thearaI1 -araI2 region of DNA. AFM is able to produce a direct
view of bending at the single molecule level without the need for
labeling the DNA or protein. Independently of the measurements, we built
a model using known AraC domain structures, biochemical contacts between
AraC and DNA, and modifications of the conformation of DNA in order to
be consistent with all dimensions. The measured DNA bend angle and the
one predicted in the model agree within the uncertainties. We believe
that the model shows the likely overall positions of the basic
components (DDs, DBDs, DNA) of AraC bound to DNA.
The measured bend angle distribution for AraC/DNA complexes has a peak
at 69° ± 25°. A possible reason for the broad distribution is that the
7-residue linker of the DBD bound to I2 is flexible, leading to a
range of angles. Experimental evidence indicates that the interdomain
linker is highly flexible and most likely not an alpha
helix.6 In vivo , AraC binds tightly to the two
direct repeat half-site sequences araI1-araI2. Experiments also
show that AraC can bind tightly to inverted
(araI1 -invaraI1 ) DNA half-sites, indicating the ability of
the linker to change its coiling conformation. Another possible reason
for the broad distribution is that the AraC/DNA complexes would land on
the mica substrate in different ways; a collapse of DNA onto mica could
appear in the AFM scan to have a range of bend angles.
All of our bend angle distributions have a peak at 0°. There could be
unbound DNA on the mica surface. Using a dissociation constant
Kd = 2 x 10-12 M for AraC – DNA
binding in the presence of arabinose,27 calculations
show that prior to deposition on mica, all of the DNA should be bound to
AraC. However in the final deposition and washing steps, it is possible
that some of the DNAs became unbound. In addition, AraC/DNA complexes
might have landed on the surface in an orientation where the DNA would
appear straight (Figure 7C). If that were the case, one would expect
that the structure would have a higher height where the DNA and protein
overlap. We did not observe this effect and cannot make a definitive
statement due to the presence of aggregates and the noise level in the
images.
We can speculate on the role of DNA bending. AraC is a transcription
regulatory protein, and like other DNA binding proteins, AraC causes DNA
to bend. Perez-Martin and Espinosa28 proposed that
increases in DNA curvature would facilitate RNA polymerase-promoter
interactions during the initiation of transcription. In the arabinose
operon in E. coli , upon the addition of arabinose, AraC binds toaraI1 and araI2 , and AraC and cyclic AMP receptor protein
(CRP) both help RNA polymerase (RNAP) bind to the pBADpromoter as well as speed the formation of open complex to initiate
transcription.29, 30 We conjecture that substantial
bending of the DNA binding sites of CRP, AraC, and RNAP would allow a
compact structure to form that could involve specific contacts among the
three proteins and improve protein-DNA interactions.15Bending of DNA may also assist melting of the double-stranded structure
in the region where transcription is to begin, like sharp bending of a
rope separates its strands.
A schematic diagram is shown in Figure 8 of CRP, AraC, RNAP and DNA. The
dashed lines show a potential path of DNA where CRP has induced a DNA
bend angle of 80°, AraC a bend of 69°, and RNAP a bend in between
30°-90°, as suggested by experimental data.31, 32, 33As can be seen, somewhat greater bending anywhere in the complex could
bring CRP and RNAP into contact. The bending of DNA by AraC appears to
facilitate the process of forming a compact structure.