Professional Documents
Culture Documents
Poland
2
(Received
ABSTRACT
We show results of our high force (up to 1.8 nN) AFM force
spectroscopy measurements of a double stranded DNA. We have found that
the force spectra of torsionally constrained molecules display a small
plateau occurring at a force of approximately 1nN. This transition, not
reported before, is absent in molecules with rotational freedom. Based on
all-atom molecular dynamics simulations, we suggest that this plateau is a
result of reducing the diameter of a double helix through extreme
stretching. The simulation suggests that the molecule is forced into a form
resembling an underwound P-DNA, with bases protruding outside of
backbones. These results broaden our understanding of fundamental aspects
of DNA nanomechanics.
PACS numbers 87.15.La 87.14.gk 87.64.Dz
A simple, and yet a very effective method of nonspecific molecule picking with an
untreated AFM tip allowed us to achieve a stretching force close to 2nN. We used the pUC18
plasmid (Fermentas, SD0051), linearized with the EcoRI enzyme (Fermentas, ER0271). To
check a possible influence of the DNA base pair sequence and a molecule length on the
observed effect, a lambda phage (Fermentas, SD0011) was studied as well using the same
protocol. SMD simulations have been performed using NAMD 2.8 code [20] and all-atom
CHARMM27 force field [21]. Simulations mimicked topologies of our AFM force
spectroscopy experiment - dragged atoms had constraints in all directions except the helical
axis, thus preventing rotation during stretching. A control set of simulations was also made,
with the pulling carbon atoms unconstrained and a free rotation around the helical axis [22].
In the majority of cases molecules picked with an AFM tip detached and were lost at
stretching forces below 0.5 nN. Nevertheless, patient molecule fishing allowed us to record
48 force curves (each resulting from different molecule) above this limit, with detachment
force beyond 1 nN (a maximum observed stretching force was 1.8 nN). Force curves obtained
showed characteristics either of torsionally constrained (rounded onset of overstretching
transition at 110pN, no hysteresis upon relaxation) or unconstrained DNA stretching (sharp
onset of overstretching at 65pN, significant hysteresis upon relaxation). Apart from the well
known overstretching and melting transitions also a third, never previously reported plateau
was present, but only if the stretched molecule was torsionally constrained (Fig. 1). This
feature appeared at very high extensions, when the molecule was stretched approximately 2.3
times its contour length and for forces approaching 1 nN. As can be seen in Fig. 1, this
plateau is very small when compared to the overstretching and melting transitions.
Force spectrograms of molecules that did not detach during stretching show that the
transition is reversible (Fig. 2(a)) and is present both in the stretching and relaxing curves
similarly to the overstretching and melting. This plateau was observed in each constrained
DNA force curve, obtained in independent experiments on different samples and in the
molecules of different length (Fig. 2(b)), provided the stretching force was large enough.
Repetitive stretching of the same molecule also marked its presence in each force
spectrogram (Fig. 3(a)). Additionally, when a portion of a stretched molecule detached and
was stretched again the transition appeared each time (Fig. 3(b)). This third plateau cannot
thus be considered just a minor detachment event or an instrumental artifact. In our opinion it
is an integral feature of DNA mechanics and reflects a forced conformational change within
the highly stretched double helix itself.
This force value does not seem to depend on the length of the stretched molecule. The
extension resulting from this transition is, on the other hand, clearly proportional to the
contour length (Fig. 4(a) inset). A rough estimate with a linear fit yields a 0.015 nm extension
per single base pair resulting from this plateau.
The third plateau never appeared when the stretched molecule was unconstrained.
This can be clearly seen on Fig 5, where normalization was made for force curves showing
characteristics of resulting from molecules with a free rotation. This implies that the origin of
this transition is inherently connected with the DNA torsional constrain.
In order to explain the mechanism behind this plateau we performed SMD simulations
for both constrained and unconstrained dsDNA. Those simulations show that when the DNA
cannot unwind, the stretched double helix reduces its diameter, forcing both strands into
contact. As a result, stacking is disrupted and bases start protruding outwards (Fig. 6). The
molecule adopts a shape similar to the previously reported P-DNA [24]. However, the
structure obtained through extreme stretching is underwound in comparison to the P-DNA
created by the extreme overtwisting with magnetic tweezers. Conversion of SMD force vs
extension curves relative to the contour length allowed for comparison with AFM force
spectroscopy data. We can see that the plateau coincides well with the transition to the
underwound P-DNA. This leads to conclusion that this origin of the third plateau is a sound
hypothesis.
SMD with a free rotation of stretched molecule leads to formation of the Zip-DNA
form, with an interchangeable base stacking, reported previously [25]. This explains the lack
of the third plateau in unconstrained molecules, as stretching induced helix diameter
reduction cannot take place in that case. Additionally, we can expect the force necessary to
cause the diameter reduction would depend on the arrangement of the DNA helix. On the
other hand, in AFM experiment with nonspecific molecule attachment a variation in pulling
angle or in separation of DNA strands is very likely. Those conditions can explain the fact
that the plateau did appear within a certain range of forces rather than for a specific one.
We did a rough estimate to test this model further. A height of a single helical pitch is
equal to
, with l being the helix length and r the helix radius. With 38 bp S-
DNA present after overstretching a helix length would be 26.6 nm. Based on AFM imaging
of stretched DNA [26] we can assume that before the transition the helix radius is
approximately 0.6 nm. Reducing the helix radius by 0.25 nm, which would correspond to the
average size (0.5 nm) of a single pyrimidine and purine base would result in elongation of
helical pitch by 0.18 nm and 0.0047 nm per single base pair. The extension of 0.015 nm per
single base pair obtained from the linear fit from the inset on Fig 4 is approximately three
times smaller.
constrained DNA has P-DNA form [12], thus 1/5 of base pairs could not contribute to the
extension through the diameter reduction. Additionally, further bases protruding outside
backbones appear beyond overstretching transition [27] due to melting. Thus, when
comparing the extension per single base pair obtained from our experiment with a proposed
model, the actual value should not be calculated over all base pairs, but adjusted accordingly
at the very least by a factor of 0.8. Considering this, we can find a good agreement between
our results and this simple geometrical estimate.
In summary, we report for the first time the presence of a mechanical transition in a
highly stretched DNA with a torsional constraint. Based on SMD simulations we postulate
that the plateau observed in force spectra is caused by the double helix diameter reduction of
the strained molecule, leading to formation of structure similar to the underwound P-DNA.
Our results give a new insight into the understanding of DNA mechanics. The DNA proves to
be a remarkably resilient molecule. As long as the double helix does not unwind, it is not
completely denatured until a force approaching 1 nN is reached. Even after such extreme
stretching, this molecule may easily return to its relaxed conformation. Additionally, if the
hypothesis of the forced transition to an underwound P-DNA is correct, then a new
mechanism, besidest the full separation of strands or overtwisting, to expose nucleobases
seems to be possible. Thus, strong mechanical deformation could serve as a means to provide
an access to the genetic sequence. A similar mechanism of opening binding sites in globular
proteins through the mechanical unfolding has been postulated [28].
ACKNOWLEDGEMENTS
We thank Prof. Piotr E. Marszalek for stimulating our work in single molecule force
spectroscopy and Prof. Richard Lavery for DNA pdb files.
Janusz Strzelecki acknowledges grant Krok w przyszo (I and III edition) from Marszaek
of Kujawsko-Pomorskie Voivodeship and EU.
[1]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
[15]
[16]
T. Hugel, M. Rief, M. Seitz, H. E. Gaub, and R. R. Netz, Physical Review Letters 94,
48301 (2005).
[17]
A. A. Guerrero, M. C. Gamero, V. Trachana, A. Ftterer, C. Pacios-Bras, N. P. DazConcha, J. C. Cigudosa, C. Martnez-A, and K. H. M. Van Wely, Proceedings of the
National Academy of Sciences 107, 4159 (2010).
[18]
N. J. Ganem and D. Pellman, The Journal of Cell Biology 199, 871 (2012).
[19]
[20]
[21]
[22]
See Supplemental Material for the AFM force spectroscopy experiment procedure and
SMD simulations details.
[23]
M. Rabbi and P. E. Marszalek, Cold Spring Harbor Protocols 2007, pdb. prot4900
(2007).
[24]
[25]
[26]
[27]
[28]
FIGURE CAPTIONS
FIG. 2. A third plateau observed in a highly stretched constrained DNA. The transition
underlying this event is reversible, as it appears both during the stretching (upper black) and
the relaxation (lower red) of a single DNA molecule (a). A magnified fragment of a force
curve, showing the third plateau in detail is displayed in the inset. This plateau was observed
each time a force exceeding 800 pN was achieved, as can be seen in three representative
curves obtained for different molecules on different samples, with significant variation in
contour length (b).
FIG. 3 The third plateau is present in repeatedly stretched DNA molecule. The plateau
was observed in a series of multiple stretching of the same DNA molecule (a). Two curves
(upper black stretching, lower red relaxation) that show this transition both before and after a
minor detachment of a stretched molecule are also shown (b).
FIG. 4 Normalization of 10 force curves showing the third plateau and a histogram of
transition forces. Curves were normalized for the force of 1100 pN [9 pUC18 (black) and 1
lambda phage (bright green)]. A full overlap can be seen in overstretching and melting
transition areas (a). A dependence of plateau length on the molecule contour length
10
[expressed in number of base pairs (bp)] is shown in the inset [pUC18 (black square) and 1
lambda phage (bright green diamonds)]. A linear fit to this data (red solid line) gives an
estimate of 0.015 nm increase in length per base pair due to transition. This plateau appears
for a certain range of forces (820 -1020 pN) (b) instead of a specific one. A molecule length
does not seem to influence the value of those forces.
FIG. 5. Lack of the third plateau in torsionally relaxed DNA molecules. The figure shows
a set of 6 curves resulting from the stretching of unconstrained molecules, with the extension
normalized for force of 1100 pN. The plateau which was present in the constrained DNA is
not observed here.
FIG. 6. SMD simulations of highly stretched DNA show a possible mechanism of the
transition. The figure shows curves acquired from two 10 ns (grey) and one 100 ns (pink)
SMD simulations of poly(dA-dT) poly(dA-dT) and two 10 ns (green) and 100ns (blue)
poly(dG-dC) poly(dG-dC). The rotation of molecule during stretching was prevented. An
experimental force curve (black) is included as a reference. The extension is calculated in
reference to the initial B-DNA form length. Snapshots from different phases of the
simulation, marked by orange dots are also shown. A constrained stretched molecule is
forced to reduce its diameter as the extension increases and the base pairing and stacking is
disrupted. At the extension coincident with the plateau a structure resembling an underwound
P-DNA is formed, with base pairs protruding outside of backbones. The transition can thus be
attributed to the transition into this new form.
11
FIG. 1
Strzelecki
12
FIG. 2
Strzelecki
13
FIG. 3
Strzelecki
14
FIG. 4
Strzelecki
15
FIG. 5
Strzelecki
16
FIG. 6
Strzelecki
17