An account of solvent accessibility in protein-RNA recognition
www.nature.com/scientificreports
OPEN
An account of solvent accessibility
in protein-RNA recognition
Sunandan Mukherjee
Received: 1 March 2018
Accepted: 21 June 2018
Published: xx xx xxxx
& Ranjit Prasad Bahadur
Protein–RNA recognition often induces conformational changes in binding partners. Consequently,
the solvent accessible surface area (SASA) buried in contact estimated from the co-crystal structures
may differ from that calculated using their unbound forms. To evaluate the change in accessibility upon
binding, we compare SASA of 126 protein-RNA complexes between bound and unbound forms. We
observe, in majority of cases the interface of both the binding partners gain accessibility upon binding,
which is often associated with either large domain movements or secondary structural transitions
in RNA-binding proteins (RBPs), and binding-induced conformational changes in RNAs. At the noninterface region, majority of RNAs lose accessibility upon binding, however, no such preference is
observed for RBPs. Side chains of RBPs have major contribution in change in accessibility. In case
of flexible binding, we find a moderate correlation between the binding free energy and change in
accessibility at the interface. Finally, we introduce a parameter, the ratio of gain to loss of accessibility
upon binding, which can be used to identify the native solution among the flexible docking models. Our
findings provide fundamental insights into the relationship between flexibility and solvent accessibility,
and advance our understanding on binding induced folding in protein-RNA recognition.
Protein-RNA recognition is essential for gene expression and its regulations. The initial contact between a
RNA-binding protein (RBP) and a RNA, often termed as encounter complex, triggers subsequent conformational changes in order to form a stable and functional association1,2. These conformational changes can either
be of small scale including side chain movements of amino acid residues or base flipping of nucleotides, or be
of large scale movements such as reorientation of polypeptide domains or change in secondary structures of
RNA. Moreover, secondary structural transitions can also induce major conformational changes in both the
binding partners3. It has been observed that the conformational changes upon binding are often associated with
significant changes in solvent accessibility in the binding partners. Lee and Richard, in 19714, first coined the
term “accessible surface area” to quantify the area of protein surface. Later, Chothia5,6 described the correlation
between accessible surface area and hydrophobic energy in protein folding. According to his study, the gain in
ΔG per squared Angstrom decrease in solvent accessible surface area (SASA) of proteins is 25 cal/mol. Recent
studies have shown that relative solvent accessibility can be used to predict the extent of conformational changes
associated with protein-protein recognition7. Besides, it has also been found that the bound conformations of
macromolecules have larger SASA than their unbound states8. Moreover, the intrinsic flexibility of proteins can
also be measured by their buried and accessible surface area9. Recently, Barik et al.10 showed that the change in
SASA upon binding can be used as a parameter to predict the binding hotspots at protein–RNA interfaces. The
prerequisite to study the change in solvent accessibility upon protein-RNA binding is the atomic structures of the
complexes and their corresponding unbound forms of the binding partners. The growing interests to decipher the
3-dimensional structures of protein-RNA complexes and their unbound structures facilitated the development of
protein-RNA docking benchmarks11–13.
In this study, we evaluate the change in SASA values calculated from the bound complex and their corresponding unbound components of protein-RNA complexes taken from the docking benchmark version 213. We
find, in majority of the cases the interface of both the binding partners gain accessibility in order to provide more
surface area to promote the stable interactions. However, majority of RNA non-interface region lose accessibility,
while, no such significant bias is observed at the non-interface region of RBPs. The change in interface accessibility is significantly contributed by the side chains, however, a moderate correlation between the change in accessibility and the backbone conformation is also observed. Interestingly, large change in accessibility is observed
when the binding is more flexible including large domain movements and secondary structural transition of
Computational Structural Biology Laboratory, Department of Biotechnology, Indian Institute of Technology
Kharagpur, Kharagpur, 721302, India. Correspondence and requests for materials should be addressed to R.P.B.
(email: )
SCIenTIfIC RePortS | (2018) 8:10546 | DOI:10.1038/s41598-018-28373-2
1
www.nature.com/scientificreports/
tRNA (A)
rRNA (B)
duplex RNA (C)
single-stranded RNA
(D)
No of complexes
28
5
40
53
126
PURU
10
3
4
4
21
PURB
16
2
32
45
95
PBRU
2
0
4
4
10
2625
1733
2329
1891
2187
BPB
1259 (±566)
839 (±276)
1153 (±681)
863 (±580)
1040 (±626)
APB
23783 (±11457)
8282 (±2342)
18598 (±11530)
16912 (±11836)
18559 (±11903)
BRB
1375 (±736)
965 (±198)
1007 (±284)
1048 (±604)
1132 (±572)
ARB
8028 (±3363)
6553 (±295)
7701 (±6721)
4786 (±4042)
6943 (±4922)
Interface
−132.4 (±174.3)
37.5 (±62.8)
−145.9 (±213.7)
−111.5 (±192.1)
−120.5 (±195.2)
Exposed
−174.3 (±167.9)
−23.1 (±20.9)
−167.3 (±209.4)
−183.4 (±169.5)
−172.0 (±184.3)
Buried
43.5 (±35.5)
77.8 (±46.9)
88.8 (±74.0)
87.4 (±75.6)
77.2 (±68.0)
Non-interface
4.5 (±35.3)
33.3 (±20.7)
−10.3 (±31.5)
9.8 (±46.6)
3.4 (±40.1)
Exposed
−25.2 (±16.3)
−5.5 (±NA)
−31.0 (±19.1)
−19.3 (±27.5)
−24.6 (±24.1)
Buried
23.0 (±27.7)
59.2 (±16.7)
26.5 (±24.1)
33.5 (±49.0)
30.4 (±33.5)
Interface
−141.0 (±104.3)
−73.9 (±195.2)
16.1 (±135.9)
−174.6 (±67.0)
−92.5 (±143.7)
Exposed
−158.9 (±94.3)
−184.4 (±143.3)
−68.0 (±32.1)
−174.6 (±67.0)
−144.1 (±91.8)
Buried
20.0 (NA)
146.9 (NA)
142.3 (±134.6)
0.0 (NA)
122.7 (±119.1)
Non-interface
16.5 (±40.1)
32.6 (±18.3)
77.9 (±112.2)
26.1 (±41.5)
40.3 (±75.8)
Exposed
−29.7 (±34.0)
0.0 (±NA)
−20.3 (±12.2)
−30.2 (±6.1)
−25.6 (±21.9)
Buried
36.3 (±22.3)
32.6 (±18.3)
143.3 (±100.8)
44.8 (±29.9)
67.3 (±73.5)
All
Average B (Å2)
BB
Average ASA (Å2)
Average δAP (Å2)a
Average δAR (Å2)b
Table 1. Statistics on change in accessibility upon binding in protein–RNA complexes. aValues are calculated on
116 PURU and PURB complexes. bValues are calculated on 31 PURU and PBRU complexes. Standard deviations are
provided in parentheses.
RBPs upon binding. We find a moderate correlation between the change in accessibility and binding free energy
when the interface undergoes significant change in conformation upon binding. Analysis of secondary structural (...truncated)
