Acta Cryst. (2013). D69, 409–419doi:10.1107/S0907444912049311
Acta Crystallographica Section D
Effect of Rap1 binding on DNA distortion and
potassium permanganate hypersensitivity
Yann-Vaı ¨ Le Bihan,aBe ´atrice
Matot,a‡ Olivier Pietrement,b
Marie-Jose `phe Giraud-Panis,c
Sylvaine Gasparini,aEric Le
Marie-He ´le `ne Le Dua*
aCEA/DSV/IBiTec-S/SB2SM, Laboratoire de
Biologie Structurale et Radiobiologie, CNRS
UMR 8221, University Paris-Sud, Ba ˆtiment 144,
CE Saclay, 91191 Gif-sur-Yvette, France,
bMaintenance des Ge ´nomes, Microscopies
Mole ´culaires et Bionanosciences, UMR 8126
CNRS and University Paris Sud, 94805 Villejuif,
France,cInstitute for Research on Cancer and
Aging, Nice (IRCAN), University of Nice, CNRS
UMR 7284, INSERM U1081, 28 Faculte ´ de
Me ´decine, University of Nice, Nice, France, and
dLBPA, UMR 8113 du CNRS, ENS Cachan,
61 Avenue du Pre ´sident Wilson, 94235 Cachan,
‡ Present address: Laboratoire RMN AIM-CEA,
Institut de Myologie – G. H. Pitie-Salpetriere,
83 Boulevard de l’Hopital, 75651 Paris CEDEX
# 2013 International Union of Crystallography
Printed in Singapore – all rights reserved
Repressor activator protein 1 (Rap1) is an essential factor
involved in transcription and telomere stability in the budding
yeast Saccharomyces cerevisiae. Its interaction with DNA
suggesting local DNA melting and/or distortion. In this study,
various Rap1–DNA crystal forms were obtained using
specifically designed crystal screens. Analysis of the DNA
conformation showed that its distortion was not sufficient to
explain the permanganate reactivity. However, anomalous
data collected at the Mn edge using a Rap1–DNA crystal
soaked in potassium permanganate solution indicated that
the DNA conformation in the crystal was compatible with
interaction with permanganate ions. Sequence-conservation
analysis revealed that double-Myb-containing Rap1 proteins
all carry a fully conserved Arg580 at a position that may
favour interaction with permanganate ions, although it is not
involved in the hypersensitive cytosine distortion. Permanga-
nate reactivity assays with wild-type Rap1 and the Rap1[R580A]
mutant demonstrated that Arg580 is essential for hypersensi-
tivity. AFM experiments showed that wild-type Rap1 and the
Rap1[R580A] mutant interact with DNA over 16 successive
binding sites, leading to local DNA stiffening but not to
accumulation of the observed local distortion. Therefore,
Rap1 may cause permanganate hypersensitivity of DNA by
forming a pocket between the reactive cytosine and Arg580,
driving the permanganate ion towards the C5–C6 bond of the
Received 3 August 2012
Accepted 30 November 2012
PDB Reference: Rap1–DNA
Telomeres are specialized nucleoprotein complexes at the end
of linear eukaryotic chromosomes that control numerous
DNA processes. They protect the linear chromosome extre-
mity against DNA-repair machinery, regulate the transcription
of nearby genes, control the terminal replication of chromo-
somal DNA and localize chromosome ends within the nuclear
space (Greider & Blackburn, 1989; Gilson & Ge ´li, 2007; de
Lange, 2009). Alteration of telomere structure criticallyaffects
these processes, leading to genome instability and the devel-
opment of many diseases, including cancer (Blackburn, 2000).
Telomeric DNA consists of repeated short G-rich sequences
with a single-stranded tail on the 30-oriented strand. Telomeric
DNA tracts vary from a few hundred base pairs in lower
eukaryotes (ciliates and yeast) to thousands of base pairs in
higher eukaryotes (plants and vertebrates) (Blackburn, 1994;
Rhodes & Giraldo, 1995). In Saccharomyces cerevisiae, the
telomere nucleoproteic assembly includes Rap1, its functional
partners Rif1, Rif2, Sir3 and Sir4, and the single-stranded
DNA-binding protein Cdc13 and its functional partners Stn1
and Ten1 (Giraud-Panis et al., 2010). Rap1 directly binds
double-stranded DNA through a central double-Myb domain
(DBD), recruits the Rif and Sir proteins through a C-terminal
globular domain (RCT) and contains a single BRCT domain
located in its N-terminal part (Morse, 2000). The remaining
40% of the peptide chain forms unstructured regions that
confer high flexibility to the molecule. Rap1 tightly wraps
DNA through its DBD, leading to orientation of the N- and
C-terminal regions on each side of the DNA axis. The
N-terminal moiety remains flexible, whereas the DBD-RCT
partly gains stability upon DNA binding (Matot et al., 2012).
Rap1 binds the 13 bp double-stranded DNA consensus
sequence ACACCCRYACAYM, which includes the half-sites
ACACC or ACAYC at positions 1–6 and ACATYor ACAYM
at positions 9–13 (Ko ¨nig et al., 1996; Graham & Chambers,
1994; Conrad et al., 1990). Footprinting experiments coupled
with bandshift assays have shown that Rap1 binds successive
telomeric repeats simultaneously, with an average frequency
of one molecule per 18 bp (Gilson et al., 1993). Each Rap1
binding site is associated with a specific KMnO4hypersensitive
site. This ability of Rap1 to induce KMnO4hypersensitivity at
the specific position ACACCCACACACC suggests that its
interaction may affect the DNA conformation (Gilson et al.,
1993). Potassium permanganate is commonly used as a probe
to characterize DNA–protein interaction in vitro and in vivo
(Spicuglia et al., 2004; Sclavi, 2008). It oxidizes the C5–C6
bond of pyrimidine bases, primarily thymidine (T), with a
preference for nucleotides located on single-stranded DNA
(Borowiec et al., 1987). Thus, potassium permanganate is
routinely used to detect regions of induced DNA melting.
However, in some examples, such as that involving Epstein–
Barr virus origin binding protein (EBNA1; Bochkarev et al.,
1998) and that involving Rap1 protein, the crystal structure
raises further questions regarding this reactivity, as the DNA
distortion remains close to the standard B-DNA conformation
(Ko ¨nig et al., 1996; Matot et al., 2012). It is unknown whether
or not the crystal packing constrains the DNA conformation
and minimizes the observed distortion. Therefore, it remains
unclear whether permanganate hypersensitivity is a conse-
quence of DNA melting or of DNA distortion. Also, a direct
although no such interaction has ever been documented. For
instance, the successive Rap1 binding sites at telomeres may
lead to the accumulation of small distortions that could affect
the conformation of telomeric DNA.
In order to better understand the effect of Rap1 binding on
DNA conformation and potassium permanganate hypersen-
sitivity, we compared the DNA conformation in various Rap1-
DBD–DNA crystal forms. The minor DNA distortion
observed encouraged us to check the compatibility of our
crystal structure with permanganate interaction by analysing
the anomalous signal of manganese observed for crystals
soaked in potassium permanganate solutions. An analysis of
residue conservation in double-Myb-containing Rap1 mole-
cules incited us to study the effect of a point mutation on
potassium permanganate reactivity. These assays revealed that
the protein itself participates in this phenomenon. Finally, we
performed AFM analysis to visualize the effect of successive
Rap1 binding on DNA conformation and report that Rap1
binding is associated with straightening of the DNA molecule
but not with accumulation of distortion.
2. Materials and methods
The cloning, production and purification of His6-tagged
Rap1[358–827], Rap1[1–827]and Rap1[1–827]mutant Rap1[R580A]
were performed as described previously (Matot et al., 2012).
Based on the crystallization conditions of protein–DNA
complexes described in 32 different publications (Suppe-
mentary Table S11) and on the use of spermine and cobalt
hexammine as described in Nowakowski et al. (1999), we
designed a 48-experiment screen through an incomplete
factorial approach as implemented in SAmBA (http://
www.igs.cnrs-mrs.fr/samba/; Audic et al., 1997; Supplementary
Table S2). Crystallization experiments were performed using
the Rap1[358–827]construct in complex with various double-
stranded DNA oligonucleotides of between 19 and 31 bp in
length with blunt or cohesive ends (Eurogentec). Crystal-
lization trials were performed at PF6 Pasteur Institute using
a MatrixMaker robot to prepare our specific crystallization
screen and a Cartesian robot to complete the crystallization
experiments. Initial crystallization hits were optimized using
sitting-drop vapour-diffusion experiments in our laboratory.
The most suitable crystals were obtained at room temperature
with a Rap1[358–827]–DNA complex concentration of between
3.3 and 5.5 mg ml?1in a solution consisting of 100 mM Tris–
HCl pH 8.0, 20% PEG MME 550, 100 mM CaCl2, 5% glycerol.
Two different crystal forms were obtained using two different
oligonucleotides, named SG19 and SG24 after their respective
crystal space groups (Table 1; Matot et al., 2012). For the
KMnO4data set, crystals of KMnO4were directly introduced
into a crystallization drop containing SG19 complex crystals
2.2. Data collection and structure determination
Native diffraction data were collected on the PX1 beamline
at the SOLEIL synchrotron and were reduced with XDS
(Kabsch, 2010). The structures were solved by molecular
replacement with Phaser (McCoy et al., 2007) using PDB entry
1ign as a model probe (Ko ¨nig et al., 1996). Electron density
and SDS–PAGE analysis of dissolved crystals revealed that
the Rap1[353–827]construct was partly degraded and that in
addition to the 31 bp oligonucleotide only the double-Myb
moiety and part of the linker between the DBD and the
C-terminus was present in the crystals. Both structures were
refined using BUSTER (Bricogne et al., 2009). Diffraction data
for SG19 complex crystals soaked in the presence of KMnO4
Le Bihan et al.
Acta Cryst. (2013). D69, 409–419
1Supplementary material has been deposited in the IUCr electronic archive
(Reference: CB5017). Services for accessing this material are described at the
back of the journal.
were collected on the PX1 beamline at the Mn K edge
wavelength (1.89 A˚), integrated with MOSFLM and reduced
with SCALA from the CCP4 package (Winn et al., 2011). Two
runs of BUSTER were performed to improve the phases
before calculating the anomalous difference map. Because of
the poor diffraction of the KMnO4-soaked crystals [4.1 A˚
resolution, I/?(I) = 7.0], the integration step was critical to
extract the anomalous signal. We performed successive data-
integration runs with the SCALEIT software (Howell &
Smith, 1992) using different levels of outlier rejection cutoff.
For each integrated data set, we calculated the anomalous
difference electron-density maps using calculated phases from
the refined structure of the Rap1–DNA complex. The quality
of the anomalous signal was evaluated by inspection of the
anomalous difference peaks using the presence of small peaks
typical of the phosphorus anomalous signal along the DNA-
molecule backbone as a quality reference. Statistics of data
collection, refinement and model validation are shown in
2.3. Conservation analysis
We first used PSI-BLAST (Altschul et al., 1990) to search
for homologues of the Rap1[360–601] sequence and then
selected sequences with 80% homology coverage using
Blammer. The 12 sequences identified were aligned with
Kalign (Lassmann & Sonnhammer, 2005) and the resulting
alignment, together with the SG19 coordinate file, were
processed with ConSurf (Ashkenazy et al., 2010) for conser-
vation analysis. Residue conservation was visualized using
WebLogo for the sequence (Crooks et al., 2004) and using
PyMOL for the three-dimensional structure (DeLano, 2002).
2.4. Permanganate-hypersensitivity assay
KMnO4footprinting experiments were performed on short
double-stranded oligonucleotides of telomeric sequence using
either wild-type Rap1 (wtRap1) or the Rap1[R580A]mutant.
Adjustments of the KMnO4hypersensitivity assay were made
with respect to previously published assays (Gilson et al., 1993;
Idrissi et al., 1998), mostly because shorter oligonucleotides
were used. Two 21-base single-stranded oligonucleotides,
TGTGTGGGTGTGCGG-30, were purchased from Euro-
gentec. The C-rich strand was radiolabelled at its 50-terminus
using [?-32P]dATP and T4-PNK, purified using Bio-Spin 6
columns (Bio-Rad) and annealed with a 1.2-fold molar excess
of the complementary G-rich strand. Protein–DNA complexes
were formed by mixing 2 nM labelled double-stranded oligo-
nucleotide with 40 nM of either wtRap1 or Rap1[R580A]in a
binding buffer consisting of 20 mM Tris pH 7.5, 150 mM NaCl.
Binding reactions were incubated for 10 min on ice followed
by 10 min at room temperature (RT). KMnO4reactions were
achieved by the addition of 18 mM KMnO4followed by 1 min
incubation at RT. The reactions were stopped by the addition
of 470 mM ?-mercaptoethanol, 0.5 M sodium acetate pH 5.2,
0.03 mg ml?1calf thymus DNA. Proteins were removed by
extraction with phenol/chloroform/isoamyl alcohol (50:49:1)
and the DNA was precipitated with ethanol and dried.
Oxidized oligonucleotides were cleaved by dissolution in 1 M
Acta Cryst. (2013). D69, 409–419Le Bihan et al.
Statistics of data collection, refinement and model validation.
Values in parentheses are for the last shell.
Data set SG19 (Matot et al., 2012)SG24 SG19 + KMnO4
Diffraction limits (A˚)
Unit-cell parameters (A˚,?) a = 40.6, b = 102.9,
a = 63.8, b = 122.6,
c = 149.4,
? = ? = ? = 900
Phaser, one solution
with 75% threshold
a = 40.6, b = 103.2,
c = 115.7,
? = ? = ? = 90
c = 116.8,
? = ? = ? = 90
Phaser, one solution
with 75% threshold
No. of unique reflections
Refinement resolution (A˚) 2.95
Figure of merit
No. of residues
No. of bases
No. of water molecules
R.m.s.d., bonds (A˚)
R.m.s.d., angles (?)
piperidine and incubation at 363 K for 25 min. Samples were
dried and subjected to two cycles of dissolution in water and
lyophilization. The resulting DNA fragments were dissolved in
formamide denaturing buffer, heated for 5 min at 363 K and
subjected to denaturing electrophoretic migration in a gel
consisting of 20% acrylamide/bisacrylamide (19:1), 7 M urea
and 1? TBE, with 0.06 M sodium acetate pH 5.2 added to the
lower reservoir of the electrophoretic apparatus to improve
the resolution of the very short fragments (Hsu et al., 2006).
After electrophoresis, the gels were dried, exposed for auto-
radiography and viewed using a Typhoon PhosphorImager
(GE Healthcare). A G+A-specific DNA-sequencing reaction
was performed on the radiolabelled C-rich strand and loaded
onto each gel as a reference to localize the KMnO4-sensitive
sites (Maxam & Gilbert, 1977). The intensities of the bands in
the gels were quantified using the ImageQuant 5.0 software.
The following equation, adapted from calculation of fractional
saturation of binding sites used in DNase I footprinting assays
(Brenowitz et al., 2001), was used to normalize band intensities
and gives the normalized KMnO4sensitivity (NKS),
where D refers to the integrated optical density of a band, n
refers to any lane corresponding to a protein–DNA sample, r
refers to the reference lane (control sample with no protein
added to the reaction mixture), site refers to the band for
which the KMnO4sensitivity is to be determined and std
refers to a band for which the integrated optical density does
not change for any samples (with or without KMnO4or with
or without protein).
2.5. ITC experiments
ITC titrations were performed at 283 K in a VP-ITC
calorimeter (GE Healthcare). Before measurement, all solu-
tions were degassed under vacuum. The two partners were
prepared in the same buffer consisting of 50 mM Tris pH 7.5,
150 mM NaCl, 10 mM ?-mercaptoethanol. Wild-type Rap1
and Rap1[R580A](4 mM) were titrated against a 21 bp telomeric
DNA fragment (40 mM) at 283 K using automatic 10 ml
injections. The thermodynamic parameters ?H (enthalpy
change), n (stoichiometry) and Ka(association constant) were
obtained by nonlinear least-squares fitting of the experimental
data using the Origin software provided with the instrument.
Standard equations were used to calculate the free energy of
binding (?G) and the entropy (?S). Because degradation was
observed in the crystal obtained using Rap1[358–827], we
performed degradation experiments using the same proteins
as in the permanganate assay and ITC experiments. After 4 h
at 277 K, no degradation of the proteins was detected by SDS–
PAGE, consistent with an effect on the KMnO4reactivity or
on DNA affinity that was exclusively related to the R580A
2.6. AFM sample preparation
16 Rap1 sites (50-ACACCCRYACACC-30) were inserted by
iterative cloning into the pUC19 plasmid between the BamHI
and BglII restriction sites. A 735 bp fragment with the 16 Rap1
sites at its centre was obtained by PCR amplification with
(forward) and 50-TCTCCCCGCGCGTTGGCCGATTCA-30
(reverse). The PCR product was purified on an anion-
exchange MiniQ column (GE Healthcare) with a SMART
system (Amersham Biosciences), precipitated with ethanol
and suspended in TE buffer (10 mM Tris pH 7.5, 1 mM
EDTA). The homogeneity of the length of the DNA mole-
cules was verified by gel electrophoresis and by TEM
(Dupaigne et al., 2008) and the concentration was determined
by measurement of the absorption at 260 nm using a Gene-
Quant 1300 spectrophotometer (GE Healthcare). The Rap1–
DNA complexes were prepared with a 735 bp DNA fragment
at 2 nM concentration (i.e. 32 nM of Rap1 binding sites) and
Rap1 protein at 64 nM concentration (i.e. two proteins per
DNA-binding site) in 10 mM Tris pH 7.6, 100 mM NaCl
buffer. The solutions were incubated for 15 min at 303 K. 5 ml
of DNA or DNA–Rap1 solution was deposited on a freshly
cleaved mica surface (muscovite mica V-1 quality, EMS) pre-
treated with 20 mM spermidine for 1 min (Hamon et al., 2007),
incubated for 2 min and rinsed with 25 ml 0.2%(w/v) uranyl
acetate solution (Re ´vet & Fourcade, 1998). The surface was
blotted and dried.
2.7. AFM imaging
Imaging was carried out in intermittent contact mode with
a MultiMode system operating with a NanoScope V controller
(Bruker). We used silicon AC160TS cantilevers (Olympus)
with resonance frequencies of about 300 kHz. All images were
collected at a scan frequency of 1 Hz and a resolution of
2048 ? 2048 pixels. The images were analysed with the
NanoScope software and a third-order polynomial function
was used to remove the background.
2.8. Analysis of DNA curvature
The local DNA curvature of DNA molecules in our model
system is characterized by the ratio, S/D, of the curvilinear
distance (S) to the linear distance (D) (Muzard et al., 1990)
between points A and B, which are located 79 nm (i.e. 232 bp)
from each DNA end and delimit the Rap1 binding-site region
(see Figs. 7a and 7b). This ratio is close to 1 for a straight
region and greater than 1 for a curved region. For each DNA
molecule, we determined the position of points A and B using
the ImageJ software (NIH) and measured the distances S and
D. As a control, the total length of all DNA molecules
analysed was measured and any molecule with an incorrect
length was excluded from the statistical analysis.
Le Bihan et al.
Acta Cryst. (2013). D69, 409–419
3. Results and discussion
3.1. Structure analysis
We previously solved the X-ray structure of
Rap1[353–602]in complex with a double-stranded DNA
oligonucleotide that contains three telomeric hemi-
sites (Matot et al., 2012). Using our specifically
designed crystallization screening, we obtained a
second crystal form that diffracted to 3.0 A˚resolution
using different overhang extremities of the oligonu-
cleotide (Table 1). The two crystal forms are called
SG19 and SG24 after their respective crystal space
groups. Superimposition of our structures with the
available Rap1–DNA complex structure (PDB entry
1ign; Ko ¨nig et al., 1996) shows r.m.s. deviations of
1.77 A˚for SG19 and 1.51 A˚for SG24. In both struc-
tures Rap1[353–602]binds exclusively to two hemi-sites
separated by three nucleotides; the additional DNA
region, containing one hemi-site and five nucleotides,
remains free (Figs. 1a and 1b).
Our structures provide useful new information
regarding the protein, the DNA and the protein–
DNA and protein interdomain interactions. Three
protein regions were missing in the previous 1ign
structure: loops 482–512, 565–571 and 579–586. In
SG19 and SG24 the loop corresponding to 482–512 is
better defined, although residues 480–498 and 483–
501, respectively, were not visible in the electron-
density map. We observed no secondary-structure
elements in this loop, which adopts slightly different
conformations in the two crystal forms, with no
particular interaction network. The second loop
(residues 565–571) is well defined in SG19 and lacks
one residue in SG24 (Supplementary Fig. S1a). This
region lengthens helix 4 of Myb2 by one helical turn
and shows strong backbone-to-backbone interactions
between Glu564 O and Arg569 N and between
Lys565 O and Asn568 N (Supplementary Fig. S1b).
These interactions orient the C-terminal part of Myb2
towards the DNA molecule. Beyond this turn, the
loop follows helix 4 in an antiparallel manner and is
stabilized by hydrogen bonds between the Glu572
backbone and Tyr560 OH and Tyr557 OH. Hydro-
phobic interactions between Met574 and Leu577 on
one side and Tyr557, Tyr552 and Phe548 on the other
further stabilize the loop (Supplementary Fig. S1c).
This hydrophobic network is associated with tight
interaction between the backbone N atoms of
Asn576, Leu577 and Thr578 and the DNA phosphate
of Cyt20 (chain C, SG24 numbering; see below).
Arg580 plunges into the DNA major groove; its
guanidinium group is involved in a bidentate inter-
action with guanidine 8, forms a stacking interaction
with thymidine 7 on the G-rich strand and is parallel
to, and located 6.7 A˚from, the KMnO4hypersensitive
cytosine 20 on the C-rich strand (see below). The
following Arg583 interacts more loosely (at a distance
Acta Cryst. (2013). D69, 409–419Le Bihan et al.
(a) Overall structure of Rap1–SG24 with the DNA envelope coloured according to
B factor; the Rap1[360–602]cartoon is shown in yellow and the bending axis is shown
as a dark line. (b) Accessibility of C-rich (red) and G-rich (yellow) strands. (c)
Diagram of the interaction of RAP1 with phosphate (mauve), sugar (pink) or base
(blue). Residues in bold are engaged in multiple interactions and underlined
residues do not interact with DNA in the 1ign structure. (d–f) DNA-conformation
parameters of SG19 (blue), SG24 (red) and 1IGN (yellow), as measured with the
Curves+ software (Lavery et al., 2009). (d) Histogram of DNA bending. (e) Curve of
major-groove width. (f) Curve of minor-groove width. The pink arrow indicates the
hypersensitive Cyt20. The C-rich strand sequence is provided as a reference in each
of 3.5 A˚) with the phosphates of thymidine 7 and guanidine 8
on the G-rich strand. The next region that deeply penetrates
into the DNA major groove includes Pro589 and Gly590, the
backbone of which interacts directly with cytosines 16 and 17,
and Asn591, the side chain of which interacts with the phos-
phate of cytosine 16. The following residues, 592–597, form a
hook that grips the C-rich strand of the DNA major groove
(Asn593 and Ser594) and Myb1 helix 2 (Tyr592 and Lys597).
In total, the buried surface of the loop 568–595 is around
670 A˚2and that of each Myb domain (regions 360–411 and
446–567) corresponds to 920 A˚2. The C-terminal 591–597
extremity of this wrapping loop is important for the functional
integrity of Rap1, whereas Arg580 located at the other
extremity is not (Matot et al., 2012).
3.2. DNA conformation analysis
The DNA molecule is well defined in both the SG19 and the
SG24 complex crystals, although the use of BUSTER was
critical to build the regions free of protein, which displayed
elevated thermal motion and have two missing nucleotides in
the C-rich strand of SG24 (Fig. 1a). The DNA length is 110 A˚,
of which 60 A˚is covered by Rap1 (Fig. 1b). The mapping of
protein–DNA interactions enlarges the binding site by two
residues in the 30direction and three residues in the 50direc-
tion on the G-rich strand beyond that described for the
reference sequence (Fig. 1c), which is in agreement with
previous observations (Del Vescovo et al., 2004). Bending,
distortion and groove-size parameters were calculated with
the Curves+ program (Lavery et al., 2009) using both the full-
length DNA of SG19 and SG24 and the Rap1 binding site
(ACACCCACACACC) of PDB entry 1ign. The bending of
the molecule is more pronounced in SG19 and SG24 than in
PDB entry 1ign, in particular in the region ACACCCACA-
ACACACC which corresponds to the path of the C-terminal
loop (Fig. 1d). In this region the major groove is widened
(Fig. 1e), whereas the minor groove is narrowed close to
cytosine 20 (SG24 numbering; Fig. 1f), which corresponds to
the potassium permanganate-hypersensitive site (ACACC-
CACACACC; Vignais & Sentenac, 1989; Gilson et al., 1993).
distortion of the permanganate-hypersensitive Cyt20 occurs
through displacement towards the major groove (Fig. 2b) and
changes in both its intra-base-pair propeller parameter
(Fig. 2c) and its inter-base-pair slide parameter (Fig. 2d). The
DNA backbone phosphates of Cyt20 and Ade21 form
hydrogen bonds to the side chains of Arg544, Lys575, Asn576
and Thr578, with Arg544 being stabilized through a stacking
interaction that involves Phe449 and Phe548 (Fig. 2e). The
constraints induced by this interaction network are associated
Le Bihan et al.
Acta Cryst. (2013). D69, 409–419
DNA-conformation analysis using the Curves+ software (Lavery et al., 2009). (a) Axis nomenclature; (b) displacement along the minor–major groove
axis; (c) intra-base-pair DNA propeller; (d) inter-base-pair DNA slide. The arrows show the distorted base hypersensitive to KMnO4. (e) Protein/DNA–
backbone interaction network at Cyt20–Ade21. Residues are coloured according to conservation scores from ConSurf as indicated on the right
(Ashkenazy et al., 2010).
with a distortion of the DNA backbone, in particular at the
" (C40—C30—O30—P) and ? (C30—O30—P—O50) dihedral
angles, for which the values are ?97?and 169?, respectively,
instead of the 174?and ?88?observed in standard B-DNA
(Lavery et al., 2009). Two C atoms of the Cyt20 sugar are
involved in the distorted dihedral angles, which therefore
propagates the distortion up to the base. Potassium perman-
ganate can oxidize pyrimidine bases at the C5—C6 bond, with
a preference for thymine, allowing the detection of unpaired
or distorted bases (Rubin & Schmid, 1980). It is extensively
used to identify protein-binding sites on double-stranded
DNA, but very few structural data are available which show
minor DNA distortion at the hypersensitive site (Bochkarev et
al., 1998). Although we observe a clear distortion at Cyt20 in
our crystal structures, its amplitude is small, such that the
accessibility of the base in the crystal structures might not be
compatible with permanganate interaction.
3.3. Permanganate interaction with the Rap1–DNA complex
A truncated form of Rap1 that includes residues 353–598
is sufficient to induce KMnO4reactivity (Gilson et al., 1993).
Therefore, our crystallized complex should allow interaction
with KMnO4if the crystal packing does not constrain the
DNA conformation. To test whether this was the case, we
collected data at the Mn K edge wavelength using an SG19
crystal soaked in the presence of KMnO4. The anomalous
difference electron-density map revealed a 3.1? level peak
located close to the C5—C6 bond of the Cyt20 ring that can be
associated with a manganese signal (Fig. 3a). The low reso-
lution of our data is associated with a ratio of the number
of reflections (4148) versus the number of parameters to refine
(3133) that prevents reliable crystallographic refinement and
analysis. However, the presence of an anomalous peak
confirms that the DNA conformation in the crystal is
compatible with an interaction between one MnO4
the C5—C6 bond of Cyt20. Therefore, the small amplitude of
the DNA distortion observed in the native crystals suggests
that additional determinants may cause the specific hyper-
sensitivity of the cytosine in the sequence ACACCCACA-
CACC (Cyt20 in the SG24 structure). The position of Arg580
plugged into the DNA major groove and facing the Cyt20
Acta Cryst. (2013). D69, 409–419 Le Bihan et al.
Conservation analysis: residue frequency in 12 double-Myb-containing
Rap1 molecules using the WebLogo representation (Crooks et al., 2004).
The aliphatic residues Gly, Pro, Ala, Val, Leu and Ile and the thiol-
containing residues Met and Cys are shown in black, the acidic residues
Asp and Glu in red, the basic residues His, Lys and Arg in blue, the
aromatic residues Phe, Tyr and Trp in purple and the polar residues Asn,
Gln, Thr and Ser in green. The ?-helices present in the structure are
shown at the top using the same numbering as in Fig. 1(a). Residues
labelled with stars correspond to conserved residues involved in inter-
Myb loop stabilization (black) and in KMnO4-hypersensitive cytosine
stabilization (pink) and the conserved Arg580 (purple).
KMnO4-hypersensitive site. (a) 2.5? contour-level anomalous electron-density map at the manganese edge (orange) around Cyt20 (magenta). (b) 2Fo?
Fcelectron-density map at the 1? contour level (blue) of SG19 around Cyt20 (magenta). Residues are coloured according to conservation scores from
ConSurf as in Fig. 2(e). (c) Interaction model of the permanganate ion with the guanidinium group of Arg580 and the Cyt20 C5—C6 bond.
base at a distance of 6.7 A˚may implicate this residue as a
potential partner (Fig. 3b). Indeed, the position and charge of
the three N atoms of the guanidinium group of Arg580 are
compatible with an interaction with the three double-bonded
permanganate O atoms. In agreement with our anomalous
data, we manually docked a permanganate ion at this location,
which positioned the remaining permanganate O atom with an
appropriate orientation and distance (2.5 A˚) to react with the
C5—C6 bond of Cyt20 (Fig. 3c).
3.4. Conservation analysis
Residue conservation during evolution highlights positions
that are maintained owing to functional or structural con-
straints. Conservation analysis of Rap1 was performed using
ConSurf (Ashkenazy et al., 2010) based on the SG19 coordi-
nate file and 11 sequences of double-Myb containing proteins
related to Rap1 (Supplementary Table S3). In addition to the
highly conserved Myb domains, other regions display
conserved features. Residues involved in the stabilization of
the inter-Myb loop are fully conserved among double-Myb
Rap1 molecules (Fig. 4), which is in agreement with the
conserved positioning of the Myb domains independent of the
spacing in the DNA-binding site (Ko ¨nig et al., 1996; Taylor et
al., 2000; Matot et al., 2012). Similarly, residues Phe449,
Arg544, Phe548, Lys575, Asn576 and Thr578, which are
involved in Cyt20 interactions and are associated with its
distortion, are fully conserved (Figs. 2e and 4).
This suggests that this distortion is structurally
and/or functionally important. Finally, Arg580,
which plugs the DNA major groove and faces
Cyt20, is also fully conserved (Figs. 3b and 4).
3.5. The role of Arg580 in permanganate-
To investigate the molecular events asso-
ciated with the KMnO4 hypersensitivity of
Cyt20, we modified a residue that may favour
interaction with KMnO4but is not involved in
nucleotide distortion. Arg580 does not directly
interact with Cyt20, with any other nucleotide
of the C-rich strand or with residues involved
in Cyt20 distortion (Fig. 3b); this residue is
disordered in the 1ign structure, although the
distortion of Cyt20 in this structure is similar to
those in our structures (Figs. 2b, 2c and 2d).
Arg580 is appropriately oriented to favour
interaction with MnO4
tion in a loop suggests that its mutation to Ala
should not affect the protein structure or its
constructed an Arg580-to-Ala mutant of Rap1
(Rap1[R580A]) and compared the permanganate
reactivity of telomeric DNA in the presence of
this mutant with that in the presence of wild-
type Rap1. Cyt20 in the presence of wild-type
Rap1 was clearly reactive, whereas the signal
was almost completely absent in the presence
of the Rap1[R580A]mutant (Figs. 5a and 5b).
Permanganate reactivity does not necessarily
correlate with the affinity of Rap1 for different
DNA sequences (Gilson et al., 1993). However,
although the assays were performed in the
presence of saturating concentrations of Rap1,
we have checked that the loss of reactivity was
not a consequence of Rap1 affinity since we
modified the protein and not the DNA
sequence. ITC experiments using Rap1[R580A]
led to an association constant of (2.02 ? 0.11)
? 108M?1(Fig. 5c), which is close to the
?(Fig. 3c) and its loca-
Le Bihan et al.
Acta Cryst. (2013). D69, 409–419
Role of Arg580 of Rap1 in the hypersensitivity of DNA to KMnO4. (a) KMnO4
footprinting with no protein (lane 2), with wild-type Rap1 (lane 3; wtRap1), with
Rap1[R580A](lane 4) and with neither protein nor KMnO4(lane 5). Lane 1, G+A reference
sequencing reaction. The arrow shows the position of the known KMnO4-hypersensitive
cytosine and the star shows the band used for normalization. (b) Normalized quantification
results from two independent KMnO4-footprinting experiments. (c) ITC experimental
curve of Rap1[R580A]with the same double-stranded oligonucleotide as used for KMnO4
footprinting. (d) Thermodynamic parameters for the binding of wild-type Rap1 and
Rap1[R580A]to DNA as measured by ITC.
(2.5 ? 0.44) ? 108M?1reported previously for wild-type Rap1
with the same DNA oligonucleotide (Fig. 5d; Matot et al.,
2012). Therefore, the observed loss of KMnO4reactivity in the
presence of Rap1[R580A]compared with wild-type Rap1 does
not result from weaker binding of Rap1[R580A]. Thus, Arg580 is
a major determinant of the hypersensitivity of Cyt20 induced
by Rap1 binding.
3.6. The effect of Rap1 on telomeric
The number of Rap1 molecules
bound to DNA at telomeres varies from
one to two molecules at short telomeres
to 16 successive molecules at long
telomeres. At long telomeres, the weak
distortion of DNA associated with
Rap1 binding may either be compen-
sated through spacing between Rap1
sites or be amplified along the DNA
molecule, leading to a change in the
overall DNA conformation. To visualize
these possible long-range effects, we
performed an AFM experiment using
linear double-stranded DNA containing
16 Rap1 sites in complex with Rap1.
The length of the DNA molecule was
chosen in order to amplify any confor-
(Figs. 6a–6e) and Rap1[R580A](Fig. 6f)
covered the region containing the 16
sites in the DNA, leading to local stif-
fening; bare DNA molecules did not
(Fig. 6g). We measured the ratio of the
curvilinear (S) to the end-to-end (D)
lengths of the region containing the 16
Rap1 sites (Muzard et al., 1990). Avalue
of 1 for the S/D ratio reflects a perfectly
straight region and higher values reflect
greater curvature (Figs. 7a and 7b).
The mean value of the S/D ratio was
Acta Cryst. (2013). D69, 409–419 Le Bihan et al.
AFM experiments. (a) Wild-type Rap1–DNA complexes. (b–e) Enlarged images of wild-type Rap1–
DNA complexes showing local stiffening. (f) Rap1[R580A]–DNA complexes. (g) Bare DNA
molecules with 16 Rap1 sites located at the centre of the fragment. The DNA and protein
concentrations correspond to a ratio of two proteins per binding site. The scale bars represent
200 nm in (a), (f) and (g) and 100 nm in (b–e).
Analysis of DNA curvature. (a) Schematic representation of DNA molecules with a central region containing 16 Rap1 binding sites delimited by points
A and B both located 79 nm from the nearest end. (b) AFM images of bare DNA molecules and DNA in the presence of Rap1 with illustration of
curvilinear (S) and direct (D) distances. Scale bars represent 50 nm. (c) Frequency distribution of the S/D ratio for bare DNA (n = 137 molecules) and
DNA molecules in complex with wild-type Rap1 (n = 128 molecules) or Rap1[R580A](n = 109 molecules). (d) Box–whiskers graph of the frequency
distribution of the S/D ratio for bare DNA and for DNA incomplex with wild-type Rap1 or with Rap1[R580A]showing the median value (dashed line) and
the 25th and 75th percentiles (rectangle). Data points lower than the 10th percentile and higher than the 90th percentile are represented as circles.
1.32 ? 0.30 for bare DNA and 1.13 ? 0.08 and 1.16 ? 0.10 for
the wild-type Rap1 and Rap1[R580A]complexes, respectively
(the frequency distribution is presented in Fig. 7c). The
distributions were significantly different (P value lower than
10?4; Mann–Withney U test), confirming that binding of either
wild-type or mutant Rap1 straightened the DNA molecule
(Fig. 7d). We therefore conclude that the small distortion of
DNA upon Rap1 binding observed in the crystal structures is
not related to a particular effect in a telomeric context.
We report a structural analysis of the telomeric factor Rap1 in
complex with DNA and assess the hypersensitivity to potas-
sium permanganate associated with formation of this complex.
Potassium permanganate is routinely used to reveal DNA
melting or distortion in DNA–protein complexes (Spicuglia
et al., 2004; Sclavi, 2008). We show for the first time that an
arginine residue facilitates and probably induces this hyper-
sensitivity. We show that the limited DNA distortion at the
hypersensitive site observed in our crystal structures is
compatible with an interaction with the manganese ion
(Fig. 3a), suggesting that the DNA molecule probably adopts a
similar conformation in solution. The fact that the sensitivity
of Rap1–DNA complexes to KMnO4is associated with DNA
distortion and not DNA melting is in agreement with the
similar reactivity observed between linear and negatively
supercoiled telomeric DNA (Gilson et al., 1994). The Arg580
residue is fully conserved among all double-Myb proteins
related to Rap1, suggesting that it has important functional
and/or structural roles. Arg580 plunges into the DNA major
groove and faces the hypersensitive cytosine, leaving sufficient
space for a permanganate ion as proposed in our manual
docking (Fig. 3c). We show that an Arg580-to-Ala mutation
abolishes the permanganate reactivity of bound DNA,
although the affinity of the mutant (Rap1[R580A]) for DNA
remains similar to that of the wild-type protein (Fig. 5). Also,
similar DNA stiffening is caused by the binding of wild-type
Rap1 and Rap1[R580A](Fig. 7). Therefore, we propose that
the pocket formed between Cyt20 and Arg580, which can
accommodate a negatively charged tetrahedral ion, may be
associated with a functional role distinct from DNA distortion.
The residues that drive loop 568–580 towards the DNA major
groove are also highly conserved (Fig. 4 and Supplementary
Fig. S1c). This suggests that Arg580 may contribute to the
wrapping of Rap1 around the DNA, which is in agreement
with general involvement of arginine/major-groove hydrogen
bonds in the mechanistic signature of protein–DNA inter-
actions (Garvie & Wolberger, 2001). Finally, our AFM
experiments using linear DNA including 16 successive Rap1
sites revealed no accumulation of apparent distortion upon
binding of multiple Rap1 molecules (Figs. 6 and 7) that could
explain the changes in topology triggered by Rap1 binding
(Gilson et al., 1994).
We acknowledge Dr Ste ´phane Marcand and Dr Rachel
Lescasse from the Laboratoire Te ´lome `re et Re ´paration du
Chromosome at the Commissariat a ` l’Energie Atomique,
Direction des Sciences du Vivant, Institut de Radiobiologie
Cellulaire et Mole ´culaire for their help with DNA preparation
for AFM experiments. We are grateful to Dr Ahmed Haouz
and Patrick Weber from PF6 Pasteur Institute for their help
and suggestions with crystallization experiments and to the
PX1 beamline group at the SOLEIL synchrotron for their
assistance during diffraction data acquisition. This work was
supported by the Commissariat a ` l’Energie Atomique, by
the Centre National pour la Recherche Scientifique, by the
Institut National de la Sante ´ et de la Recherche Me ´dicale, by
the Agence Nationale de la Recherche (grant ANR-06-
BLAN-0076) and by the Ligue contre le cancer.
Altschul, S. F., Gish, W., Miller, W., Myers, E. W. & Lipman, D. J.
(1990). J. Mol. Biol. 215, 403–410.
Aono, S., Hartsch, T. & Schulze-Gahmen, U. (2003). Acta Cryst. D59,
Ashkenazy, H., Erez, E., Martz, E., Pupko, T. & Ben-Tal, N. (2010).
Nucleic Acids Res. 38, W529–W533.
Audic, S., Lopez, F., Claverie, J.-M., Poirot, O. & Abergel, C. (1997).
Proteins, 29, 252–257.
Blackburn, E. H. (1994). Cell, 77, 621–623.
Blackburn, E. H. (2000). Nature (London), 408, 53–56.
Bochkarev, A., Bochkareva, E., Frappier, L. & Edwards, A. M.
(1998). J. Mol. Biol. 284, 1273–1278.
Borowiec, J. A., Zhang, L., Sasse-Dwight, S. & Gralla, J. D. (1987). J.
Mol. Biol. 196, 101–111.
Brenowitz, M., Senear, D. & Kingston, R. (2001). Curr. Protoc. Mol.
Biol., Unit 12.14. doi:10.1002/0471142727.mb1204s07.
Bricogne, G., Blanc, E., Brandl, M., Flensburg, C., Keller, P., Paciorek,
W., Roversi, P., Sharff, A., Smart, O., Vonrhein, C. & Womack, T.
(2009). BUSTER. http://www.globalphasing.com.
Conrad, M. N., Wright, J. H., Wolf, A. J. & Zakian, V. A. (1990). Cell,
Crooks, G. E., Hon, G., Chandonia, J. M. & Brenner, S. E. (2004).
Genome Res. 14, 1188–1190.
De Lano, W. L. (2002). PyMOL. http://www.pymol.org.
Del Vescovo, V., De Sanctis, V., Bianchi, A., Shore, D., Di Mauro, E.
& Negri, R. (2004). J. Mol. Biol. 338, 877–893.
Dupaigne, P., Lavelle, C., Justome, A., Lafosse, S., Mirambeau, G.,
Lipinski, M., Pie ´trement, O. & Le Cam, E. (2008). PLoS One, 3,
Escalante, C. R., Shen, L., Escalante, M. C., Brass, A. L., Edwards, T.
A., Singh, H. & Aggarwal, A. K. (2002). J. Struct. Biol. 139, 55–
Garvie, C. W. & Wolberger, C. (2001). Mol. Cell, 8, 937–946.
Genis, C., Scone, P., Kasahara, H. & Nam, H.-J. (2008). Acta Cryst.
Gilson, E. & Ge ´li, V. (2007). Nature Rev. Mol. Cell Biol. 8, 825–838.
Gilson, E., Mu ¨ller, T., Sogo, J., Laroche, T. & Gasser, S. M. (1994).
Nucleic Acids Res. 22, 5310–5320.
Gilson, E., Roberge, M., Giraldo, R., Rhodes, D. & Gasser, S. M.
(1993). J. Mol. Biol. 231, 293–310.
Giraud-Panis, M. J., Pisano, S., Poulet, A., Le Du, M.-H. & Gilson, E.
(2010). FEBS Lett. 584, 3785–3799.
Graham, I. R.& Chambers, A. (1994). Nucleic Acids Res. 22, 124–130.
Greider, C. W. & Blackburn, E. H. (1989). Nature (London), 337,
Hamon, L., Pastre ´, D., Dupaigne, P., Le Breton, C., Le Cam, E. &
Pie ´trement, O. (2007). Nucleic Acids Res. 35, e58.
Hirsch, J. A., Wah, D. A., Dorner, L. F., Schildkraut, I. & Aggarwal,
A. K. (1997). FEBS Lett. 403, 136–138.
Horton, N. C., Dorner, L. F., Schildkraut, I. & Perona, J. J. (1999).
Acta Cryst. D55, 1943–1945.
Le Bihan et al.
Acta Cryst. (2013). D69, 409–419
Hovde, S., Brooks, A., Strong, K. & Geiger, J. H. (2002). Acta Cryst. Download full-text
Howell, P. L. & Smith, G. D. (1992). J. Appl. Cryst. 25, 81–86.
Hsu, L. M., Cobb, I. M., Ozmore, J. R., Khoo, M., Nahm, G., Xia, L.,
Bao, Y. & Ahn, C. (2006). Biochemistry, 45, 8841–8854.
Idrissi, F. Z., Ferna ´ndez-Larrea, J. B. & Pin ˜a, B. (1998). J. Mol. Biol.
Imasaki, T., Hashimoto, H., Shimizu, T., Kato, M., Tsuda, J., Kita, K.
& Sato, M. (2004). Acta Cryst. D60, 1165–1166.
Kabsch, W. (2010). Acta Cryst. D66, 125–132.
Kapetaniou, E. G., Kotsifaki, D., Providaki, M., Rina, M., Bouriotis,
V. & Kokkinidis, M. (2007). Acta Cryst. F63, 12–14.
Ko ¨nig, P., Giraldo, R., Chapman, L. & Rhodes, D. (1996). Cell, 85,
Lange, T. de (2009). Science, 326, 948–952.
Larivie `re, L., Kurzeck, J., Aschke-Sonnenborn, U., Ru ¨ger, W. &
More ´ra, S. (2002). Acta Cryst. D58, 1484–1486.
Lassmann, T. & Sonnhammer, E. L. L. (2005). BMC Bioinformatics,
Lavery, R., Moakher, M., Maddocks, J. H., Petkeviciute, D. &
Zakrzewska, K. (2009). Nucleic Acids Res. 37, 5917–5929.
Lu, P., Li, Y., Gorman, A. & Chi, Y.-I. (2006). Acta Cryst. F62,
Matot, B., Le Bihan, Y. V., Lescasse, R., Pe ´rez, J., Miron, S., David, G.,
Castaing, B., Weber, P., Raynal, B., Zinn-Justin, S., Gasparini, S. &
Le Du, M.-H. (2012). Nucleic Acids Res. 40, 3197–3207.
Maxam, A. M. & Gilbert, W. (1977). Proc. Natl Acad. Sci. USA, 74,
McCoy, A. J., Grosse-Kunstleve, R. W., Adams, P. D., Winn, M. D.,
Storoni, L. C. & Read, R. J. (2007). J. Appl. Cryst. 40, 658–674.
Morse, R. H. (2000). Trends Genet. 16, 51–53.
Muzard, G., The ´veny, B. & Re ´vet, B. (1990). EMBO J. 9, 1289–
Ng, C. K. L., Palasingam, P., Venkatachalam, R., Baburajendran, N.,
Cheng, J., Jauch, R. & Kolatkar, P. R. (2008). Acta Cryst. F64, 1184–
Nowakowski, J., Shim, P. J., Joyce, G. F. & Stout, C. D. (1999). Acta
Cryst. D55, 1885–1892.
O’Loughlin, T. J., Xu, Q., Kucera, R. B., Dorner, L. F., Sweeney, S.,
Schildkraut, I. & Guo, H.-C. (2000). Acta Cryst. D56, 1652–1655.
Pereira, J. H., Ha, S. C. & Kim, S.-H. (2008). Acta Cryst. F64, 175–
Pereira de Je ´sus, K., Serre, L., Hervouet, N., Bouckson-Castaing, V.,
Zelwer, C. & Castaing, B. (2002). Acta Cryst. D58, 679–682.
Pio, F., Ni, C.-Z., Mitchell, R. S., Knight, J., McKercher, S., Klemsz,
M., Lombardo, A., Maki, R. A. & Ely, K. R. (1995). J. Biol. Chem.
Redondo, P., Prieto, J., Ramos, E., Blanco, F. J. & Montoya, G. (2007).
Acta Cryst. F63, 1017–1020.
Re ´vet, B. & Fourcade, A. (1998). Nucleic Acids Res. 26, 2092–2097.
Rhodes, D. & Giraldo, R. (1995). Curr. Opin. Struct. Biol. 5, 311–322.
Richardson, J. M., Finnegan, D. J. & Walkinshaw, M. D. (2007). Acta
Cryst. F63, 434–437.
Rubin, C. M. & Schmid, C. W. (1980). Nucleic Acids Res. 8, 4613–
Sam, M. D., Abbani, M. A., Cascio, D., Johnson, R. C. & Clubb, R. T.
(2006). Acta Cryst. F62, 825–828.
Sam, M. D., Cascio, D., Johnson, R. & Clubb, R. T. (2003). Acta Cryst.
Schwartz, T., Shafer, K., Lowenhaupt, K., Hanlon, G., Herbert, A. &
Rich, A. (1999). Acta Cryst. D55, 1362–1364.
Sclavi, B. (2008). Biochem. Soc. Trans. 36, 745–748.
Shanmuganatham, K. K., Ravichandran, M., Howe, M. M. & Park,
H.-W. (2007). Acta Cryst. F63, 620–623.
Spicuglia, S., Kumar, S., Chasson, L., Payet-Bornet, D. & Ferrier, P.
(2004). J. Biochem. Biophys. Methods, 59, 189–194.
Suwa, Y.,Nakamura, T., Toma, S.,Ikemizu, S.,Kai, H. & Yamagata, Y.
(2008). Acta Cryst. F64, 171–174.
Tahirov, T. H., Inoue-Bungo, T., Sasaki, M., Fujikawa, A., Kimura, K.,
Sato, K., Adachi, S., Kamiya, N. & Ogata, K. (2001). Acta Cryst.
Tamulaitiene, G., Grazulis, S., Janulaitis, A., Janowski, R., Bujacz, G.
& Jaskolski, M. (2004). Biochim. Biophys. Acta, 1698, 251–254.
Tan, S., Hunziker, Y., Pellegrini, L. & Richmond, T. J. (2000). J. Mol.
Biol. 297, 947–959.
Taylor, H., O’Reilly, M., Leslie, A. & Rhodes, D. (2000). J. Mol. Biol.
Textor, L. C., Wilmanns, M. & Holton, S. J. (2007). Acta Cryst. F63,
Vassylyeva, M. N., Svetlov, V., Klyuyev, S., Devedjiev, Y. D.,
Artsimovitch, I. & Vassylyev, D. G. (2006). Acta Cryst. F62, 1027–
Viadiu, H., Kucera, R., Schildkraut, I. & Aggarwal, A. K. (2000). J.
Struct. Biol. 130, 81–85.
Viadiu, H., Vanamee, E´. S., Jacobson, E. M., Schildkraut, I. &
Aggarwal, A. K. (2003). Acta Cryst. D59, 1493–1495.
Vignais, M. L. & Sentenac, A. (1989). J. Biol. Chem. 264, 8463–8466.
Vivian, J. P., Porter, C., Wilce, J. A. & Wilce, M. C. J. (2006). Acta
Cryst. F62, 1104–1107.
Vivian, J. P., Wilce, J. A., Hastings, A. F. & Wilce, M. C. J. (2001). Acta
Cryst. D57, 421–424.
Watanabe, S., Kita, A., Kobayashi, K., Takahashi, Y. & Miki, K.
(2006). Acta Cryst. F62, 1275–1277.
Winn, M. D. et al. (2011). Acta Cryst. D67, 235–242.
Acta Cryst. (2013). D69, 409–419 Le Bihan et al.