Full text of DOI:10.1110/ps.9.11.2161

Protein Science ~2000!, 9:2161–2169. Cambridge University Press. Printed in the USA.
Copyright © 2000 The Protein Society
Solution structure of DinI provides insight
into its mode of RecA inactivation
BENJAMIN E. RAMIREZ,1 OLEG N. VOLOSHIN,2 R. DANIEL CAMERINI-OTERO,2
and AD BAX1,3
1Laboratory of Chemical Physics, National Institutes of Diabetes and Digestive and Kidney Diseases,
National Institutes of Health, Bethesda, Maryland 20892-0520
2Genetics and Biochemistry Branch, National Institutes of Diabetes and Digestive and Kidney Diseases,
National Institutes of Health, Bethesda, Maryland 20892-0520
~Received May 30, 2000; Final Revision August 15, 2000; Accepted August 24, 2000!
Abstract
The Escherichia coli RecA protein triggers both DNA repair and mutagenesis in a process known as the SOS response.
The 81-residue E. coli protein DinI inhibits activity of RecA in vivo. The solution structure of DinI has been determined
by multidimensional triple resonance NMR spectroscopy, using restraints derived from two sets of residual dipolar
couplings, obtained in bicelle and phage media, supplemented with J couplings and a moderate number of NOE
restraints. DinI has an a0b fold comprised of a three-stranded b-sheet and two a-helices. The b-sheet topology is
unusual: the central strand is flanked by a parallel and an antiparallel strand and the sheet is remarkably flat. The
structure of DinI shows that six negatively charged Glu and Asp residues on DinI’s kinked C-terminal a-helix form an
extended, negatively charged ridge. We propose that this ridge mimics the electrostatic character of the DNA phos-
podiester backbone, thereby enabling DinI to compete with single-stranded DNA for RecA binding. Biochemical data
confirm that DinI is able to displace ssDNA from RecA.
Keywords: bicelle; DinI; dipolar coupling; liquid crystal; NMR; Pf1; RecA
Maintaining a stable genome is essential for the survival of an
organism. For this reason, it is critical that an organism possesses
some mechanism that repairs DNA damage. DNA damage prevents
replication by DNA polymerases, resulting in the formation of
single-stranded DNA ~ssDNA!. In Escherichia coli, binding of the
RecA protein to ssDNA induces a DNA repair process known as
the SOS response ~Friedberg et al., 1995!. Once RecA binds to
ssDNA, the ssDNA-RecA cofilament acts as a coprotease, promot-
ing the autocleavage of the LexA repressor. This cleavage event
inactivates LexA, resulting in the expression of the 30 or so pro-
teins that are the constituents of the SOS response. Faithful repair,
that is, repair without introducing mutations, is a result of recom-
binational repair mediated by RecA and other proteins in the SOS
response ~Cox, 2000!. If these error free processes fail to repair the
damage, additional SOS proteins activate so-called translesion DNA
synthesis, which enables DNA replication to proceed even at dam-
aged sites ~Walker, 1998!. This error-prone process is referred to as
SOS mutagenesis, because mutations arise by the incorporation of
random bases. SOS mutagenesis is also induced by the coprotease
activity of the ssDNA-RecA cofilament, because one of the pro-
teins involved in this process, UmuD, is active only after RecA-
mediated autocleavage ~Shinagawa et al., 1988!. Clearly, the
coprotease activity of ssDNA–RecA cofilament plays a pivotal role
in inducing DNA repair and mutagenesis.
One of the SOS proteins synthesized upon inactivation of LexA
is a small 81-residue protein known as DinI, which inhibits the
coprotease activity of RecA ~Yasuda et al., 1998!. When DinI is
overexpressed, stimulation of the SOS response by mitomycin C
and autocleavage of LexA and UmuD are blunted in vivo ~Yasuda
et al., 1998!. The mechanism of how DinI down regulates the SOS
response was not determined. However, recent biochemical data
indicate that DinI can prevent ssDNA binding to RecA and can
even displace ssDNA from ssDNA–RecA cofilaments ~O.N. Volo-
shin, B.E. Ramirez, A. Bax, & R.D. Camerini-Otero, in prep.!.
We have determined the solution structure of DinI in an effort to
understand the mechanism through which it inhibits RecA copro-
tease activity. Although conventional multidimensional NMR is
adequate to define the structure of a small protein such as DinI at
an adequate level of resolution, the present structure determination
relied primarily on structural restraints derived from residual di-
polar couplings, measured in two liquid crystalline media. Dipolar
couplings have been shown to significantly improve the accuracy
of structures determined by NMR, as exemplified by an increase of
the number of residues that lie in the most-favored regions of the
Ramachandran plot ~Tjandra et al., 1997!, improved agreement
with structure-predicted chemical shift changes ~Ottiger et al., 1997!,
Reprint requests to: Ad Bax, NIH, Building 5, Rm. 126, Bethesda, Mary-
land 20892; e-mail: bax@nih.gov.
2161

2162
B.E. Ramirez et al.
and closer agreement with crystal structures in cases where these
are available ~Clore et al., 1999!. Residual dipolar couplings result
from the weak alignment with the magnetic field of a protein
dissolved in an aqueous, dilute liquid crystalline ~LC! medium.
Such LC media consist of large, magnetically oriented particles
~Tjandra & Bax, 1997a!. Dipolar couplings are a function of the
orientation of the internuclear vector relative to the molecular align-
ment tensor, and hence, provide structural information that is dis-
tinctly different from other NMR data such as nuclear Overhauser
effects ~NOEs!, coupling constants, or chemical shifts.
A single alignment tensor restricts the internuclear vector ori-
entation to one of two cones that are related by inversion symme-
try. This degeneracy can be greatly reduced by measuring dipolar
couplings for a second alignment tensor orientation ~Ramirez &
Bax, 1998!. With two independent alignment tensors, the number
of possible internuclear orientations are limited to the points where
the corresponding cones intersect. The most commonly used LC
media in biomolecular NMR are either planar phospholipid mi-
celle disks, commonly known as bicelles ~Sanders & Schwonek,
1992; Tjandra & Bax, 1997a; Vold et al., 1997!, and filamentous
phage ~Clore et al., 1998b; Hansen et al., 1998!. In the medium
containing nearly neutral bicelles, protein alignment is dominated
by steric interactions. Instead, in the strongly negatively charged
filamentous phage medium electrostatic interactions frequently dom-
inate ~Zweckstetter & Bax, 2000!. Consequently, proteins will
align differently in bicelle and phage LC media, yielding indepen-
dent alignment tensors ~Clore et al., 1998b!. The DinI structure
was calculated from these combined sets of dipolar couplings,
measured in bicelle and phage media, supplemented by a moderate
number of NOE and J-coupling derived torsion angle restraints.
1
1
Fig. 1. Histogram of DNH
,
1DCaHa
,
2DC9HN
,
1DC9N, and DCaC9 residual
1
2
1
1
dipolar couplings. The DCaHa, DC9HN, DC9N, and DCaC9 couplings were
normalized to match those of the DNH couplings, using scaling factors of
0.50, 3.2, 8.2, and 5.0, respectively ~Ottiger & Bax, 1998b!. A: Residual
dipolar couplings obtained in a LC solution consisting of 5% bicelle LC
~30:10:1 DTPC:DHPC:CTAB! and 100 mM NaCl. B: Residual dipolar
couplings obtained in a LC solution consisting of ;8 mg0mL Pf1 phage
1
1
1
1
and 150 mM NaCl. For the Pf1 sample, only the DNH, DCaHa, and DCaC9
couplings were measured for reasons discussed in the text.
Results and discussion
The structure of DinI was determined from heteronuclear multi-
dimensional NMR data acquired from uniformly labeled 15N and
13C015N samples. At 1 mM, the average backbone amide 15N
transverse relaxation time T2 equals ;85 ms, considerably shorter
than the ;150 ms expected for a monomeric 9 kD globular pro-
tein. Adding salt markedly decreases tC, which is further reduced
by decreasing the protein concentration. As a compromise, all
experiments were performed at 0.6 mM DinI and 100 mM NaCl
~T2 ϭ ;130 ms!, unless specified otherwise. A sample at pH 6.6
and a temperature of 25 8C yielded the best dispersion and most
uniform peak intensity in the 15N-1H heteronuclear single quantum
coherence ~HSQC! spectrum. At 35 8C, a number of resonances in
the 15N-HSQC spectrum broadened significantly.
LC, and phage LC media are compared in Figure 2. Initial values
for the Da and R were again determined from the normalized
distribution of these observed residual dipolar couplings ~Fig. 1B!,
yielding Da ϭ 21.0 Hz and R ϭ 0.19. Subsequent refinement of
these values by conducting a grid search yielded Da ϭ 22.0 Hz and
R ϭ 0.09. The alignment tensors determined in the phage LC and
bicelle LC are quite different, as judged by the lack of correlation
between DNH residual dipolar couplings observed in these two
media ~Fig. 3!.
A total of 1,786 experimental NMR restraints were used in a
simulated annealing algorithm to determine the 3D structure of
1
1DNH, DNC9, DC9HN, DCaHa, and DCaC9 residual dipolar cou-
plings were measured for DinI dissolved in bicelle LC media.
Approximate values for the axial ~Da! and rhombic ~Dr! compo-
nents of DinI’s alignment tensor in the bicelle medium were de-
termined from the normalized distribution of all the observed residual
dipolar couplings ~Fig. 1A! ~Clore et al., 1998a!. Initial values
1
2
1
1
1
were Da ϭ Ϫ10.5 Hz ~normalized for DNH! and R ϭ 0.38 ~R ϭ
Dr0Da!. Da and R were subsequently refined by conducting a grid
search to find values that yield the lowest energy structures. Final
values were Da ϭ Ϫ10.0 Hz and R ϭ 0.28.
1
1
A second set of DNH,
1DCaHa, and DCaC9 residual dipolar
Fig. 2. Small sections taken from the 3D ~HA!CA~CO!NH spectra for
residues A17–L21, showing 13Ca-$1Ha% doublets in the isotropic ~left!,
bicelle-aligned ~center! and phage-aligned ~right! 3D constant-time ~HA!CA-
~CO!NH spectra for residues A17–A22. These residues form the beginning
of DinI’s first a-helix, and the residual dipolar couplings exhibit the peri-
odicity characteristic of a regular a-helix.
couplings was recorded for DinI dissolved in the Pf1 filamentous
phage LC medium. 13Ca-$1Ha% doublets obtained from the three-
dimensional ~3D! constant time ~HA!CA~CO!NH triple resonance
experiment for a helical segment of DinI in the isotropic, bicelle

Solution structure of DinI
2163
1
Fig. 3. Correlation between the DNH residual dipolar couplings recorded
in bicelle and phage LC media. The lack of correlation illustrates the
different orientations of the alignment tensor in the phage LC. In the bicelle
alignment tensor frame, the phage alignment tensor is characterized by
Euler angles a ϭ 1158; b ϭ 898; g ϭ 208.
DinI. These restraints included 333 residual dipolar couplings ~1DNH
,
2
1
1
1DNC9, DC9HN, DCaHa, and DCaC9! obtained from the bicelle LC
sample and 192 residual dipolar couplings ~1DNH 1DCaHa, and
,
1DCaC9! obtained from the phage LC sample. There are 1,114 NOE
restraints, including 59 ambiguous NOEs, and only 140 of the
unique restraints are long range. No extensive effort was made to
collect long-range NOE restraints. Rather, we relied on the dipolar
restraints to provide both accuracy and precision. Nevertheless, we
Fig. 4. Backbone superposition of the final 15 DinI structures calculated
by simulated annealing.
3
3
also measured JC9Cg JNCg for all the Ile, Thr, Val, and aromatic
residues to ensure selection of the correct side-chain conformation.
These experiments are relatively rapid and are also straightforward
to interpret. Unique x1 angles were found for all residues, except
N-terminal strand is flanked by both an antiparallel and a parallel
strand, and is remarkably flat. The central, N-terminal strand is
connected to the first a-helix via a segment that contains the se-
quence SPLP. The number of dipolar restraints per residue in this
region is much smaller because most of the backbone dipolar
coupling measurements rely on detection of the HN signal, which
is absent for Pro residues. Therefore, more careful analysis of
NOEs in this region was essential for ensuring the correct local
geometry in this region. Strong HaiϪ1 Ϫ Proi-Hd NOEs for all three
prolyl residues in DinI indicated trans peptide bonds. This infor-
mation can be difficult to extract unambiguously from dipolar
restraints, particularly in view of the fact that the number of mea-
surable backbone dipolar couplings for Pro residues is much reduced.
Strand b1 is followed by a long a-helix ~A17–A32! that is
capped by F33. Most of the F33 phenyl ring is exposed to solvent,
although it also shows some long-range NOEs to residues such as
M1 and I63. The connection between this a-helix and strand b2
~V40–Y44! is well defined, with no evidence of conformational
exchange contributions. The outside strands of the three-stranded
b-sheet are connected by a three-alanine linker, A45–A46–A47.
The amide of A46 is not observable in the 15N-1H HSQC spectrum
as a result of conformational exchange. The third b-strand ~N48–
I53! is well-defined and rigid, and is connected to the C-terminal
helix ~K57–F78! by a short three-residue sequence: G54–A55–
T56. The amide of G54 is not observed in the 15N-1H HSQC
spectrum, also as a result of conformational exchange.
3
3
T6, Y31, H39, and V79, which exhibit JC9Cg and JNCg values
3
indicative of x1 averaging. JC9Cg values for F33 could not be
determined because it is followed by a Pro residue, but its very
3
small JNCg coupling rules out a x1 ϭ 1808 rotamer.
A superposition of the 15 final simulated annealing structures is
shown in Figure 4. Although the precision of the backbone is very
high ~root-mean-square deviation ~RMSD! ϭ 0.17 Å!, the RMSD
is considerably higher when all non-H atoms are included ~0.84 Å!.
It is well known that the precision at which an NMR structure is
determined frequently is considerably higher than its accuracy.
However, we believe the quality of the DinI structure to be quite
good as it yields favorable scores when cross-validated using
Q-factor analysis ~see below!. Moreover, when evaluated by the
program PROCHECK ~Laskowski et al., 1993!, the overall quality
indicators also score relatively high. For example, even if the
potential for driving backbone torsion angles to the regions most
populated in the Protein Data Bank ~PDB! ~Kuszewski et al.,
1997! is turned off, about 90% of the residues have backbone
torsion angles in the most favored region of the Ramachandran
map, with no residues in the generously allowed or disallowed
regions. Table 1 provides a summary of the structural statistics.
Description of the structure
DinI possesses an a-b fold ~Fig. 5!, where two a-helices ~residues
17–32 and 57–78! pack against a three-stranded b-sheet ~residues
1–8, 40–44, 48–53!. The b-sheet is unusual in that the central,
The C-terminal a-helix displays a pronounced kink of ;708 at
residues E72–S73–A74, which is the position where it crosses over

2164
B.E. Ramirez et al.
Table 1. Structural statisticsa
RMSDs from experimental distance restraints ~Å!b
All ~1114!
0.056 6 0.001
0.063 6 0.001
0.038 6 0.002
0.061 6 0.002
0.051 6 0.003
Intraresidue ~592!
Interresidue sequential ~|i Ϫ j| ϭ 1! ~278!
Interresidue short range ~1 Ͻ |i Ϫ j| Ͻ 5! ~104!
Interresidue long range ~|i Ϫ j| Ͼ 5! ~140!
RMSDs from experimental dihedral restraints
~deg! ~148!b
0.65 6 0.03
RMSDs from residual dipolar couplings ~Hz!
1DNH ~bicelle! ~69!
1.14 6 0.01
1.99 6 0.02
0.46 6 0.01
0.32 6 0.01
0.89 6 0.01
1.28 6 0.01
2.55 6 0.04
0.97 6 0.01
1DCaHa ~bicelle! ~70!
1DCaC9 ~bicelle! ~69!
1DNC9 ~bicelle! ~61!
2DC9HN ~bicelle! ~64!
1DNH ~phage! ~58!
1DCaHa ~phage! ~65!
1DCaC9 ~phage! ~69!
Deviations from idealized covalent geometry
Bonds ~Å!
Angles ~deg!
Impropers ~deg!
0.004 6 0.0001
0.93 6 0.01
0.68 6 0.01
ELJ ~kcal0mol!c
Ϫ355 6 7.5
Coordinate precision ~Å!d
Backbone nonhydrogen atoms
All nonhydrogen atoms
0.17 6 0.03
0.84 6 0.06
PROCHECK quality indicators
Residues in most-favored region of Ramachandran plot 97089%e
No. of bad contacts
202
aStatistics are for the final 15 simulated annealing structures. The num-
ber of terms for the various restraints is given in parentheses. The final
values for the force constants of the various terms in the simulated anneal-
ing target function are as follows: 1,000 kcal molϪ1 ÅϪ2 for bond lengths,
500 kcal molϪ1 radϪ2 for angles and improper torsions ~which serve to
maintain planarity and chirality!, 4 kcal molϪ1 ÅϪ4 for the quartic Van der
Waals repulsion term ~with the van der Waals radii set to 0.8 times their
value in CHARMM PARAM 19020 parameters!, 30 kcal molϪ1 ÅϪ2 for
experimental distance restraints ~interproton distances!, 100 kcal molϪ1
radϪ2 for torsion angle restraints, 1.0 kcal molϪ1 HzϪ2 for the bicelle
1
derived DNH dipolar coupling restraints, 0.25 kcal molϪ1 HzϪ2 for the
1
1DCaHa dipolar coupling restraints, 2.5 kcal molϪ1 HzϪ2 for the DCaC9
1
dipolar coupling restraints, 6.4 kcal molϪ1 HzϪ2 for the DC9N dipolar
2
coupling restraints, 1.8 kcal molϪ1 HzϪ2 for the DC9HN dipolar coupling
restraints. These force constants were divided by two for dipolar coupling
restraints measured in phage, to account for the stronger alignment.
bNone of the structures exhibited interproton distance violations greater
than 0.5 Å or dihedral angle violations greater than 58. Torsion angle
restraints included 76 f, 51 c, and 21 x1 angles.
cThe Lennard–Jones van der Waals energy was calculated with the
CHARMM PARAM 19020 parameters and was not included in the simu-
lated annealing target function.
dDefined as the average RMS difference ~residues 2–78! between the
final 15 simulated annealing structures and the mean coordinates.
eEighty-nine percent for a separate series of calculations without inclu-
sion of the Ramachandran term in the structure calculation. There were
no f0c angles in the generously allowed or disallowed region of the
Ramachandran plot.
Fig. 5. A: Backbone ribbon diagram of DinI, with helices colored blue,
b-strands red, and loops, turns and the flexible C-terminus in green. B:
Map of the electrostatic surface potential. The electrostatic potential is
colored from red ~negative charge! to blue ~positive charge!. These figures
were generated using the programs MOLMOL ~Koradi et al., 1996! and
GRASP ~Nicholls et al., 1991!.
the first a-helix. The presence of this kink is supported not only by
to cluster in a narrow range. Residues 75–78 have a characteristic
3
1
1
secondary Ca shifts and JHNHa couplings, but also by the DNH
1DNH coupling that differs significantly from the DNH values of
couplings: because all the N—H bonds in a helix point approxi-
mately in the same direction, DNH values for a straight helix tend
residues 57–71. Short-range NOEs confirm the helical character
for residues 75–78, and long-range NOEs between F770W78 and
1

Solution structure of DinI
2165
residues A17, A20, and L21 in the first a-helix confirm the ori-
entation of this helical section. Backbone amides just prior to the
kink region ~T70–E72! are strongly attenuated by intermediate
rate conformational exchange, and consequently, only exhibit very
weak NOEs and decreased accuracy for the dipolar couplings, in
particular for DNC9, DC9HN, and DCaC9. The dipolar couplings
tightly define the orientation of the C-terminal helical segment
~residues 74–78!, however. No long-range NOEs are observed for
the V79 methyl groups, in agreement with its x1 averaging noted
above, and the backbone 15N T2 of the C-terminal residues S80 and
E81 indicates extensive internal dynamics for these terminal two
residues.
protein is slightly more structured, i.e., 13Ca chemical shifts for
several a-helical residues ~A17, A20, Q68, and A74! shifted fur-
ther downfield, and small upfield shifts were observed for several
b-sheet 13Ca resonances ~T6, R43, and N48!. The largest chemical
shift changes occur for residues L67–D75 that include the kink in
the C-terminal a-helix. This suggests that these chemical shift
changes are caused by DHPC binding to DinI. This was confirmed
by titrating a DinI sample with DHPC. At a DHPC concentration
of 4 mM, the 15N-HSQC spectrum was superimposable with the
15N-HSQC spectrum recorded in the aligned bicelle medium. In-
terestingly, this concentration provides yet another, independent
measure for the concentration of free DHPC in a bicelle LC solu-
tion, previously inferred from very different types of data ~Ottiger
& Bax, 1998a; Struppe & Vold, 1998!. The DHPC titration exper-
iments indicated a very weak Kd of ;2 mM. Titration of CT
13C-HSQC spectra with DHPC allowed mapping of the resonances
previously assigned in the bicelle-free medium. The assignment of
the N and HN resonances in the aligned state was also confirmed
by the Ca and C9 chemical shifts, measured in the 3D 1H-
coupled~F1! ~H!CA~CO!NH and 3D 13Ca-coupled ~F1! HNCO
experiments, respectively, which were needed for deriving the DCaHa
and DC9Ca couplings. Although small changes in chemical shifts
were evident for a number of these resonances, Ca and C9 reso-
nances generally were much less affected than amide resonances.
The residues of DinI most affected by the presence of DHPC
were delineated by calculating the average chemical shift change
that occurred in the presence of 10 mM DHPC ~Fig. 6!. The
structure of DinI indicates the presence of two primarily hydro-
phobic pockets where DHPC may bind. The indole ring of W71
acts a barrier, separating the two pockets; L13, L21, W77, and F78
line the inside of one of the pockets, while V5, I7, L50, and L67
line the other one. Weak NOEs between DHPC and side chains of
L13, I18, L21, W71, and W77 of DinI are present in NOE spec-
troscopy ~NOESY! spectra, recorded in the presence of 10 mM
DHPC. The low intensity of the intermolecular NOEs and the
weak binding constant suggest that binding of DHPC to DinI is
1
2
1
Similarity to other structures
Using the DALI program ~Holm & Sander, 1993!, we searched the
PDB for proteins with folds similar to DinI. High similarity scores
were found for segments of only two proteins: ~1! residues P1-
L102 of 5 carboxymethyl hydroxy muconate isomerase ~1otg, Z ϭ
5.3, RMSD ϭ 2.6 Å, 6% identity! ~Subramanya et al., 1996!.
However, compared to DinI, the order of the strand and helical
segments is not identical. ~2! The P2-S98 region of the cytokine
d-dopachrome tautomerase ~1dpt, Z ϭ 5.0, RMSD ϭ 2.8 Å, 17%
identity! ~Sugimoto et al., 1999!. In contrast to DinI, neither pro-
tein is active in any processes involving DNA metabolism. Their
remote structural similarity to DinI therefore is believed to be
coincidental.
Interaction between DinI and the liquid crystal
1
Significant changes in the H-15N HSQC spectrum were observed
between samples without and with bicelles. In fact, several reso-
nances sharpened in the presence of bicelles ~e.g., A15, L67, T70,
and E72! but there was no evidence of a second set of resonances
nor that the protein had denatured. To the contrary, the secondary
Ca shifts of the protein in the bicelle medium indicated that the
Fig. 6. ~A! Worm and ~B! space-filling models of DinI, color-coded to indicate chemical shift differences for each residue upon
addition of 10 mM DHPC. Blue indicates residues with little to no change in chemical shifts while red denotes residues with the largest
changes in chemical shifts. All assigned chemical shifts were used to calculate the chemical shift difference. The chemical shift changes
Dd of different nuclei were normalized and averaged on a per-residue basis ~Grzesiek et al., 1997!. Specifically, the average chemical
shift difference for a given residue is calculated as @ ~Dd !20N#102, where the summation extends over all N nuclei i for which shifts
͚
i
i
were measured in a given residue, and 15N and 13C Ddi were scaled down by factors of 5 and 2, respectively, prior to calculating
@
~Dd !20N#102
.
͚
i
i

2166
B.E. Ramirez et al.
nonspecific, and therefore, no further effort was made to charac-
terize the DinI–DHPC complex in more detail. Nevertheless, it is
conceivable that the pockets where DHPC binds are important for
other, yet unknown intermolecular interactions with other proteins
or small molecules, which may modulate DinI–RecA affinity.
The relatively small magnitude of the chemical shift changes
suggests that the DinI structure remains largely unchanged in the
presence of DHPC. This is confirmed by residual dipolar couplings
measured in the phage LC medium. No significant chemical shift
changes relative to the isotropic sample are observed for DinI in
the presence of Pf1. Alignment of DinI in the phage medium is
dominated by electrostatic interactions, but no significant binding
of DinI to the phage appears to take place. If such binding were to
occur, the protein 15N T2 in the bound state would be extremely
short, and exchange with the free state would result in a shortening
of the time-averaged 15N relaxation rate. However, the 15N T2
values are unchanged relative to those observed in isotropic solution.
At a phage concentration of 8 mg0mL and 150 mM NaCl, DinI
is aligned stronger than desirable. This overalignment poses sev-
eral problems: ~1! measurement of residual dipolar couplings be-
comes less accurate because of the additional splittings that arise
from 1H-1H dipolar couplings and concomitant lower signal to
noise, ~2! the transfer efficiency during INEPT segments of the
fitted to the structure calculated using bicelle dipolar couplings, in
addition to the NOE and torsion restraints. The high correlation
coefficient ~R ϭ 0.97! and the absence of any areas in the protein
with more pronounced differences between predicted and observed
dipolar couplings confirm that DHPC binding does not signifi-
cantly alter the DinI structure. Conversely, these data indicate that
DinI’s electrostatic interaction with the phage does not distort its
structure.
Structure validation
The quality of the DinI NMR structure can be assessed by a quality
or Q-factor based on residual dipolar couplings ~Cornilescu et al.,
1998; Clore & Garrett, 1999; Drohat et al., 1999!. The Q-factor, a
measure of the agreement between dipolar couplings predicted by
the NMR structure and the measured dipolar couplings, is defined
as ~Cornilescu et al., 1998; Drohat et al., 1999!
Q ϭ rms~Dmeas Ϫ Dpred!0rms~Dmeas!.
The Q-factor is somewhat analogous to the crystallographic free
R-factor ~Brünger, 1992!. Cross-validation using residual dipolar
coupling Q-factors was done as follows: A subset of residual di-
1
1
NMR pulse sequences is diminished for residues where DNH and
polar couplings either in the same class ~e.g., DNH! or a mixture
1
1
1DCaHa approach JNH and JCaHa, and ~3! the presence of un-
resolved dipolar couplings to remote nuclei results in resonance
broadening. These problems precluded the reliable measurement
from the various classes is chosen randomly. This subset is gen-
erally ;10% of the total number of observed residual dipolar
couplings. If for a selected coupling both phage and bicelle values
are available, the subset includes both entries. Structures are cal-
culated with all the experimental restraints, but excluding the di-
polar restraints of the subset. From the resulting structures, dipolar
couplings are predicted for those in the subset. The average Q-factor
for the DinI NMR structure, calculated from five different ran-
domly selected subsets of residual dipolar couplings, is 18%. This
is slightly better than, for example, the agreement between the
1.8 Å X-ray crystal structure of ubiquitin ~Vijay-Kumar et al.,
1987! and its dipolar couplings ~Ottiger & Bax, 1998b!, but not as
high as what can be obtained by exhaustive analysis of all NOEs,
side-chain dipolar and J couplings ~J.L. Marquardt, unpubl. obs.!.
Interestingly, the Q-factor increases several percent when the term
driving backbone angles to the most populated regions of the Ra-
machandran map ~Kuszewski et al., 1997! is excluded from the
X-PLOR target function, indicating that this term actually im-
proves the quality of the structure, and is not merely cosmetic.
1
2
of DC9N and DC9HN couplings in the phage medium. At Pf1 con-
centrations below 8 mg0mL and using 150 mM NaCl, phase sep-
aration occurs over a period of several days, which disappears
upon thorough mixing. So, in the presence of 150 mM NaCl, a
phage concentration of 8 mg0mL is near the lower limit where a
stable, aligned phase can be maintained. The electrostatic nature of
DinI alignment by Pf1 phage was confirmed by titrating a 12
mg0mL Pf1 sample containing 0.6 mM DinI with NaCl. The de-
gree of DinI alignment decreases ;30% when increasing @NaCl#
from 150 to 375 mM.
The structure calculated using only the bicelle dipolar couplings
is found to be in very good agreement with the dipolar couplings
measured in the phage medium. Figure 7 shows a correlation be-
1
tween the DCaHa couplings measured in phage and those best
Interaction with RecA
The DinI–RecA interaction inhibits the recombinase and copro-
tease activities of RecA, resulting in the termination of recombi-
national DNA repair and translesion DNA replication by the SOS
response. DinI was shown to be a downregulator of the SOS re-
sponse, preventing the ssDNA–RecA cofilament’s promotion of
the autocleavage of LexA and UmuD ~Yasuda et al., 1998!. Recent
biochemical studies indicate that DinI can displace ssDNA from
RecA ~O.N. Voloshin, B.E. Ramirez, A. Bax, & R.D. Camerini-
Otero, in prep.!. Examination of the electrostatic potential surface
of DinI indicates the presence of an extended, negatively charged
ridge that is formed by residues located in the C-terminal helix of
DinI, specifically: E65, E69, E72, D75, and D76 ~Fig. 5B!. The
distribution of negative charges on this helix resembles the ar-
rangement of negative charges in a single strand of DNA. We
propose that the C-terminal helix of DinI is designed to mimic the
Fig. 7. Correlation between DCaHa values measured in phage medium and
predicted by the structure calculated with bicelle ~but no phage! dipolar
coupling restraints, together with all other experimental restraints. The
correlation coefficient R equals 97%.

Solution structure of DinI
2167
phosphodiester backbone of DNA, allowing DinI to compete with
ssDNA for the RecA binding site. The absence of a structure of
DinI complexed to RecA precludes a direct assessment of the
functional significance of the kink in the C-terminal helix in the
proposed binding model. However, because the spatial arrange-
ment of the negative charges along this ridge is influenced by
this kink, we expect it to be an essential feature for enhancing the
complementarity between DinI and RecA. Moreover, mutagenesis
data confirm the importance of E72, located in the kink, for DinI’s
ability to inactivate RecA in vivo ~O.N. Voloshin, unpubl. obs.!.
elsewhere ~O.N. Voloshin, B.E. Ramirez, A. Bax, & R.D. Camerini-
Otero, in prep.!. Samples for NMR consisted of 0.6 mM DinI,
20 mM phosphate, 100 mM NaCl, 0.1% azide, and pH 6.6 ~95%
H2O05% D2O or 99.9% D2O!, in a 250 mL Shigemi microcell.
NMR spectroscopy
All NMR experiments were performed at 25 8C, except where
indicated. Data were collected on Bruker DMX500, DMX600, and
DRX800 spectrometers equipped with self-shielded three-axis gra-
dient, triple-resonance probes. All NMR data were processed with
the NMRPipe package and analyzed with NMRDraw ~Delaglio
et al., 1995! and PIPP ~Garrett et al., 1991!. Sequential backbone
assignments were determined from the following 3D experiments:
CBCA~CO!NH, HNCACB, HNHA, 15N-separated total correla-
tion spectroscopy ~TOCSY!, and 15N-separated NOESY ~Bax &
Conclusions
The structure of DinI in combination with numerous biochemical
experiments ~O.N. Voloshin, B.E. Ramirez, A. Bax, & R.D.
Camerini-Otero, in prep.! suggests that regulation of RecA by DinI
occurs through the ability of DinI to compete with ssDNA for
RecA binding. The importance of the C-terminal helix of DinI in
this process is highlighted by experiments that show that this helix
alone is able to compete with ssDNA for binding to a peptide
fragment derived from the ssDNA-binding L2 loop of RecA ~O.N.
Voloshin, B.E. Ramirez, A. Bax, & R.D. Camerini-Otero, in prep.!.
The distribution of negative charges on the surface of DinI’s
C-terminal kinked helix mimics the electrostatic character of a
ssDNA phosphodiester backbone. We, therefore, believe that this
distribution of negative charge is the primary factor that causes
DinI to compete with ssDNA for the binding site on RecA. Thus,
DinI appears to be yet another example of a protein that acts by
mimicry of DNA. In a recent report, Liu et al. found that
dTAF~II!230 can bind to TATA-box binding protein ~TBP! through
a surface area that mimics the widened major groove of the TATA-
box found in the TBP–DNA complex ~Liu et al., 1998!. In another
example, Selmer et al. report on a protein that mimics tRNA ~Selmer
et al., 1999!. In both these cases, the similarity lies more in the
shape than in the electrostatic properties, however. Because RecA
interacts primarily with the backbone phosphate groups of DNA,
the charge distribution of DinI is the primary feature that enables
it to mimic DNA.
1
Grzesiek, 1993!. Aliphatic 13C and H side-chain resonances were
determined from a 3D ~H!CCH-COSY experiment ~Fesik et al.,
1990; Kay et al., 1990; Gehring & Ekiel, 1998!. Aromatic protons
were assigned from 2D homonuclear COSY, TOCSY ~43.5 ms
mixing time!, and NOESY ~100 ms mixing time! experiments,
recorded on a uniformly labeled 15N sample in 99% D2O at
3
800 MHz. In addition to JHNHa measurement for obtaining f
constraints ~Vuister & Bax, 1993!, 3JC9Cg and 3JNCg in aromatic and
Cg-methyl carrying residues were measured by quantitative J cor-
relation spectroscopy for obtaining x1 angles ~Grzesiek et al., 1993;
Vuister et al., 1993; Hu et al., 1997!. Interproton distance restraints
were obtained from a 3D 15N-separated NOE ~100 ms mixing
time! and a 3D constant-time ~CT! 13C-separated NOE ~CT ϭ
28 ms; mixing time ϭ 100 ms! experiment. Interproton distance
restraints involving aromatic protons were obtained from the 2D
homonuclear NOESY spectrum, recorded at 800 MHz.
Residual dipolar couplings were obtained from the difference in
J splittings measured in spectra of aligned and isotropic ~in water
at 25 8C! samples. Both bicelle and filamentous phage were used
as alignment media. The bicelle medium contained 50 mg0mL
lipids in the molar ratio of DTPC:DHPC:CTAB ϭ 30:10:1 ~DTPC
ditridecanoylphosphocholine; DHPC, dihexanoylphosphocholine;
CTAB, cetyltrimethylammonium bromide!. All NMR experiments
on the bicelle sample were recorded at 28 8C because 25 8C was
found to be too close to the temperature where the liquid crystal is
no longer stable. The phage LC sample contained 8 mg0mL Pf1
phage. All NMR experiments on the phage LC sample were also
recorded at 28 8C to slightly reduce the degree of alignment and
Very few homologs of DinI have been identified so far, and none
have been reported in eukaryotes. However, it is conceivable that
the type of charge mimicry found in DinI also occurs in other
proteins that regulate activities of nonspecific nucleic acid binding
proteins, both in prokaryotes and in eukaryotes.
1
improve protein linewidth. DNH residual dipolar couplings were
obtained from 1H-coupled ~F1! IPAP $15N,1H%-HSQC experiments
Materials and methods
1
2
~Ottiger et al., 1998!. DNC9 and DC9HN residual dipolar couplings
Sample preparation
were obtained from 1H,13C9-coupled ~F1! IPAP $15N,1H%-HSQC
experiments ~Wang et al., 1998!. DCaHa residual dipolar cou-
1
The DNA encoding DinI was cloned into the pET21a vector using
standard molecular biology techniques. The protein was expressed
from the pET21a construct in the E. coli BL21~DE3! strain. Start-
ing from a single colony taken from a freshly streaked agar plate,
a 25 mL minimal medium starter culture containing 15N-NH4Cl
and 13C-glucose as the sole sources of nitrogen and carbon, re-
plings were obtained from 3D CT-~H!CA~CO!NH experiments
without 1H-decoupling in the F1 dimension ~Tjandra & Bax, 1997b!.
1DCaC9 residual dipolar couplings were obtained from 3D 13Ca-
coupled HNCO experiments. Because the 13C9 linewidth is dom-
inated by chemical shift anisotropy, these HNCO spectra were
recorded at the lowest field strength ~500 MHz 1H frequency!,
yielding narrower linewidths but, nevertheless, adequate resolution.
spectively, was grown until it appeared slightly turbid ~OD600
;
1.0!, and then refrigerated overnight. The following day a 500 mL
minimal medium containing 15N-NH4Cl and 13C-glucose culture
was inoculated with 5 mL of the starter culture. The culture was
grown at 37 8C to an OD600 of ;0.8, at which point protein ex-
pression was induced with 1 mM IPTG. The cells were harvested
after 4 h. Details regarding the purification procedure are described
Structure calculations
Distance calibration of the interproton distance constraints ob-
tained from the 3D 15N-separated NOESY experiment was done as
follows: All non-Gly residues in the allowed negative f region

2168
B.E. Ramirez et al.
have HNiϪ1-Hai distances of 2.9 6 0.2 Å. Other interproton distances
were scaled relative to this average 2.9 Å distance using an em-
pirical 10r4 dependence of the intensity, which in a very approx-
imate manner accounts for the effects of spin diffusion ~Güntert
et al., 1991!. Distance ranges were set to 625% of the target
distance, except for degenerate peaks. For degenerate peaks, no
lower distance bound was used, and 10r6 averaging was employed
in simulated annealing calculations ~Clore et al., 1986!. Distance
calibration of the interproton distance restraints obtained from the
3D 13C-separated NOESY-CT-HSQC experiment was done as fol-
lows: For NOEs to methyl groups ~detected during t3!, the average
NOE intensity from the adjacent vicinal methine proton to the
dipolar restraints, together with NOE and experimental dihedral
angle restraints.
Data have been deposited in the PDB ~accession number 1F0A!
Acknowledgments
We thank Marius Clore, Georg Kontaxis, and John Marquardt for useful
discussions and Frank Delaglio, Dan Garrett, and Markus Zweckstetter for
software support. B.E.R. is supported by a postdoctoral fellowship from
the National Science Foundation ~DBI-9807412!.
References
methyl group in Ile, Thr, Ala, and Val residues was used as a
reference @~ rϪ6
)
ϭ 2.1 Å#. For nonmethyl 1H signals, the
Ϫ106
͚
i
i
Bax A, Grzesiek S. 1993. Methodological advances in protein NMR. Acc Chem
Res 26:131–138.
Brünger AT. 1992. Free R-value—A novel statistical quantity for assessing the
accuracy of crystal-structures. Nature 355:472–475.
Brünger AT. 1993. X-PLOR manual version 3.1. New Haven, Connecticut: Yale
University.
Clore GM, Brünger AT, Karplus M, Gronenborn AM. 1986. Application of
molecular-dynamics with interproton distance restraints to 3-dimensional
protein-structure determination—A model study of crambin. J Mol Biol
191:523–551.
average intraresidue HNHa cross peak to diagonal peak intensity
ratio is used as a 2.9 Å reference distance, again using an empirical
10r4 distance dependence, and relatively tight 625% tolerances.
No lower bound distance was given for degenerate peaks and
residues with extensive conformational averaging ~as evidenced by
very intense 1H-13C correlations in the CT-HSQC spectrum!. Inter-
proton distance restraints derived from the 2D homonuclear NOESY
experiment were grouped into five overlapping classes 2.0–2.7,
2.3–3.2, 2.7–3.7, 3.0–4.5, and 3.3–5.5 Å, corresponding to very
strong, strong, medium, weak, and very weak NOEs, respectively.
Degenerate peaks were also classified using these intensity clas-
sifications, except no lower distance bound was employed.
Clore GM, Garrett DS. 1999. R-factor, free R, and complete cross-validation for
dipolar coupling refinement of NMR structures. J Am Chem Soc 121:9008–
9012.
Clore GM, Gronenborn AM, Bax A. 1998a. A robust method for determining the
magnitude of the fully asymmetric alignment tensor of oriented macromol-
ecules in the absence of structural information. J Magn Reson 133:216–221.
Clore GM, Starich MR, Bewley CA, Cai ML, Kuszewski J. 1999. Impact of
residual dipolar couplings on the accuracy of NMR structures determined
from a minimal number of NOE restraints. J Am Chem Soc 121:6513–6514.
Clore GM, Starich MR, Gronenborn AM. 1998b. Measurement of residual
dipolar couplings of macromolecules aligned in the nematic phase of a
colloidal suspension of rod-shaped viruses. J Am Chem Soc 120:10571–10572.
Cornilescu G, Delaglio F, Bax A. 1999. Protein backbone angle restraints from
searching a database for chemical shift and sequence homology. J Biomol
NMR 13:289–302.
Cornilescu G, Marquardt JL, Ottiger M, Bax A. 1998. Validation of protein
structure from anisotropic carbonyl chemical shifts in a dilute liquid crys-
talline phase. J Am Chem Soc 120:6836–6837.
Cox MM. 2000. Recombinational DNA repair in bacteria and the RecA protein.
Prog Nucleic Acid Res Mol Biol 63:311–366.
Delaglio F, Grzesiek S, Vuister GW, Zhu G, Pfeifer J, Bax A. 1995. Nmrpipe—A
multidimensional spectral processing system based on unix pipes. J Biomol
NMR 6:277–293.
Drohat AC, Tjandra N, Baldisseri DM, Weber DJ. 1999. The use of dipolar
couplings for determining the solution structure of rat apo-S100B ~beta
beta!. Protein Sci 8:800–809.
Fesik SW, Eaton HL, Olejniczak ET, Zuiderweg ERP, McIntosh LP, Dahlquist
FW. 1990. 2D and 3D NMR-spectroscopy employing C-13-C-13 magneti-
zation transfer by isotropic mixing—Spin system-identification in large pro-
teins. J Am Chem Soc 112:886–888.
Friedberg EC, Walker GC, Siede W. 1995. DNA repair and mutagenesis. Wash-
ington, DC: ASM Press.
Garrett DS, Powers R, Gronenborn AM, Clore GM. 1991. A common-sense
approach to peak picking in 2-dimensional, 3-dimensional, and 4-dimensional
pectra using automatic computer-analysis of contour diagrams. J Magn
Reson 95:214–220.
Gehring K, Ekiel I. 1998. H~C!CH-COSY and ~H!CCH-COSY experiments for
C-13-labeled proteins in H2O solution. J Magn Reson 135:185–193.
Grzesiek S, Bax A, Hu JS, Kaufman J, Palmer I, Stahl SJ, Tjandra N, Wingfield
PT. 1997. Refined solution structure and backbone dynamics of HIV-1 Nef.
Protein Sci 6:1248–1263.
Grzesiek S, Vuister GW, Bax A. 1993. A simple and sensitive experiment for
measurement of Jcc couplings between backbone carbonyl and methyl car-
bons in isotopically enriched proteins. J Biomol NMR 3:487–493.
Güntert P, Braun W, Wüthrich K. 1991. Efficient computation of 3-dimensional
protein structures in solution from nuclear-magnetic-resonance data using
the program Diana and the supporting programs Caliba, Habas and Glomsa.
J Mol Biol 217:517–530.
Hansen MR, Mueller L, Pardi A. 1998. Tunable alignment of macromolecules
by filamentous phage yields dipolar coupling interactions. Nat Struct Biol
5:1065–1074.
Structures were calculated by simulated annealing using the
program X-PLOR ~Brünger, 1993! adapted to incorporate residual
dipolar coupling restraints ~Tjandra et al., 1997! and a conforma-
tional database ~Kuszewski et al., 1997!. The minimized target
function comprises harmonic potentials for covalent geometry
~bonds, angles, and improper torsions! and residual dipolar cou-
pling restraints, quadratic square-well potentials for interproton
distance and dihedral angle restraints, and a quartic van der Waals
repulsion term for nonbonded contacts. Force constants are listed
in the footnote to Table 1. Initial folds were calculated in Cartesian
coordinate space starting from a completely extended structure
using only NOE restraints and backbone dihedral angle restraints
derived from the program TALOS ~Cornilescu et al., 1999!. TA-
LOS restraints were available for 60 residues. All 20 residues for
which no unambiguous TALOS restraints were obtained fall out-
side regions of regular secondary structure. Also, x1 angle re-
straints were derived for most Val, Ile, Thr, and aromatic residues
from 3JC9Cg and 3JNCg values. The 10 lowest energy structures were
used as input for a second simulated annealing calculation that
included residual dipolar coupling restraints measured in the bi-
celle medium, together with all NOE restraints. TALOS backbone
dihedral angle restraints were not used in this second stage. In-
3
stead, 42 f dihedral angle restraints were derived from JHNHa
1
values, aided by the finding that for all non-Gly residues JCaHa Ն
139 Hz, indicating f Ͻ 0 ~Vuister et al., 1992!. For non-Gly
3
residues without an unambiguous JHNHa-derived f restraint but
1
with JCaHa Ն 139 Hz, only the less restrictive f Ͻ 0 restraint was
1
used. In fact, the smallest JCaHa measured was 140 Hz for T6.
1
1
These JCaHa values are a byproduct of the DCaHa measurement,
but nevertheless greatly improve convergence during the early
stages of the structure calculation. A total of 51 loose c restraints
were derived from the corresponding intraresidue and sequential
Ha-HN NOE intensity ratios. The 10 lowest energy structures from
the initial round of calculations were used as starting structures in
a final refinement stage that included both the bicelle and phage

Solution structure of DinI
2169
Holm L, Sander C. 1993. Protein–structure comparison by alignment of distance
matrices. J Mol Biol 233:123–138.
cleavage of UmuD protein and Sos mutagenesis. Proc Natl Acad Sci USA
85:1806–1810.
Hu JS, Grzesiek S, Bax A. 1997. Two-dimensional NMR methods for deter-
mining ~chi 1! angles of aromatic residues in proteins from three-bond
J~C9C gamma! and J~NC gamma! couplings. J Am Chem Soc 119:1803–1804.
Kay LE, Ikura M, Bax A. 1990. Proton proton correlation via carbon carbon
couplings—A 3-dimensional NMR approach for the assignment of aliphatic
resonances in proteins labeled with C-13. J Am Chem Soc 112:888–889.
Koradi R, Billeter M, Wüthrich K. 1996. MOLMOL: A program for display and
analysis of macromolecular structures. J Mol Graphics 14:51.
Kuszewski J, Gronenborn AM, Clore GM. 1997. Improvements and extensions
in the conformational database potential for the refinement of NMR and
X-ray structures of proteins and nucleic acids. J Magn Reson 125:171–177.
Laskowski RA, Macarthur MW, Moss DS, Thornton JM. 1993. Procheck—A
program to check the stereochemical quality of protein structures. J Appl
Crystallogr 26:283–291.
Liu DJ, Ishima R, Tong KI, Bagby S, Kokubo T, Muhandiram DR, Kay LE,
Nakatani Y, Ikura M. 1998. Solution structure of a TBP-TAF~II!230 com-
plex: Protein mimicry of the minor groove surface of the TATA box un-
wound by TBP. Cell 94:573–583.
Nicholls A, Sharp KA, Honig B. 1991. Protein folding and association—
Insights from the interfacial and thermodynamic properties of hydrocar-
bons. Proteins Struct Funct Genet 11:281–296.
Struppe J, Vold RR. 1998. Dilute bicellar solutions for structural NMR work.
J Magn Reson 135:541–546.
Subramanya HS, Roper DI, Dauter Z, Dodson EJ, Davies GJ, Wilson KS,
Wigley DB. 1996. Enzymatic ketonization of 2-hydroxymuconate: Speci-
ficity and mechanism investigated by the crystal structures of two isomer-
ases. Biochemistry 35:792–802.
Sugimoto H, Taniguchi M, Nakagawa A, Tanaka I, Suzuki M, Nishihira J. 1999.
Crystal structure of human D-dopachrome tautomerase, a homologue of
macrophage migration inhibitory factor, at 1.54 angstrom resolution. Bio-
chemistry 38:3268–3279.
Tjandra N, Bax A. 1997a. Direct measurement of distances and angles in bio-
molecules by NMR in a dilute liquid crystalline medium. Science 278:1111–
1114.
Tjandra N, Bax A. 1997b. Large variations in C-13~alpha! chemical shift an-
isotropy in proteins correlate with secondary structure. J Am Chem Soc
119:9576–9577.
Tjandra N, Omichinski JG, Gronenborn AM, Clore GM, Bax A. 1997. Use of
dipolar H-1–N-15 and H-1–C-13 couplings in the structure determination of
magnetically oriented macromolecules in solution. Nat Struct Biol 4:732–738.
Vijay-Kumar S, Bugg CE, Cook WJ. 1987. Structure of ubiquitin refined at 1.8
A resolution. J Mol Biol 194:531–544.
Ottiger M, Bax A. 1998a. Characterization of magnetically oriented phospho-
lipid micelles for measurement of dipolar couplings in macromolecules.
J Biomol NMR 12:361–372.
Vold RR, Prosser RS, Deese AJ. 1997. Isotropic solutions of phospholipid bi-
celles: A new membrane mimetic for high-resolution NMR studies of poly-
peptides. J Biomol NMR 9:329–335.
Ottiger M, Bax A. 1998b. Determination of relative N-H-N N-C9, C-alpha-C9,
and C~alpha!-H-alpha effective bond lengths in a protein by NMR in a dilute
liquid crystalline phase. J Am Chem Soc 120:12334–12341.
Ottiger M, Delaglio F, Bax A. 1998. Measurement of J and dipolar couplings
from simplified two-dimensional NMR spectra. J Magn Reson 131:373–
378.
Ottiger M, Tjandra N, Bax A. 1997. Magnetic field dependent amide N-15
chemical shifts in a protein–DNA complex resulting from magnetic ordering
in solution. J Am Chem Soc 119:9825–9830.
Ramirez BE, Bax A. 1998. Modulation of the alignment tensor of macromol-
ecules dissolved in a dilute liquid crystalline medium. J Am Chem Soc
120:9106–9107.
Sanders CR, Schwonek JP. 1992. Characterization of magnetically orientable
bilayers in mixtures of dihexanoylphosphatidylcholine and dimyristoylphos-
phatidylcholine by solid-state NMR. Biochemistry 31:8898–8905.
Selmer M, Al-Karadaghi S, Hirokawa G, Kaji A, Liljas A. 1999. Crystal struc-
ture of Thermotoga maritima ribosome recycling factor: A tRNA mimic.
Science 286:2349–2352.
Vuister GW, Bax A. 1993. Quantitative J correlation—A new approach for
measuring homonuclear 3-bond J~H~N!H~Alpha! coupling-constants in N-15-
enriched proteins. J Am Chem Soc 115:7772–7777.
Vuister GW, Delaglio F, Bax A. 1992. An empirical correlation between 1J~Ca-
Ha! and protein backbone conformation. J Am Chem Soc 114:9674–9675.
Vuister GW, Wang AC, Bax A. 1993. Measurement of 3-bond nitrogen carbon-J
couplings in proteins uniformly enriched in N-15 and C-13. J Am Chem Soc
115:5334–5335.
Walker GC. 1998. Skiing the black diamond slope: Progress on the biochemistry
of translesion DNA synthesis. Proc Natl Acad Sci USA 95:10348–10350.
Wang YX, Marquardt JL, Wingfield P, Stahl SJ, Lee-Huang S, Torchia D, Bax
A. 1998. Simultaneous measurement of H-1–N-15, H-1–C-139, and N-15–
C-139 dipolar couplings in a perdeuterated 30 kDa protein dissolved in a
dilute liquid crystalline phase. J Am Chem Soc 120:7385–7386.
Yasuda T, Morimatsu K, Horii T, Nagata T, Ohmori H. 1998. Inhibition of
Escherichia coli RecA coprotease activities by DinI. EMBO J 17:3207–3216.
Zweckstetter M, Bax A. 2000. Prediction of sterically induced alignment in a
dilute liquid crystalline phase: Aid to structure determination by NMR. J Am
Chem Soc 122:3791–3792.
Shinagawa H, Iwasaki H, Kato T, Nakata A. 1988. RecA protein-dependent