NMR structural characterization of HIV-1 virus protein U
cytoplasmic domain in the presence of
dodecylphosphatidylcholine micelles
Marc Wittlich
1,2
, Bernd W. Koenig
1,2
, Matthias Stoldt
1,2
, Holger Schmidt
1,2,
* and Dieter Willbold
1,2
1 Institut fu
¨
r Strukturbiologie und Biophysik (ISB-3), Forschungszentrum Ju
¨
lich, Germany
2 Institut fu
¨
r Physikalische Biologie, Heinrich-Heine-Universita
¨
tDu
¨
sseldorf, Germany
Introduction
VpU (virus protein U) is an 81 amino acid transmem-
brane protein encoded by HIV-1 and some simian
immunodeficiency virus strains, e.g. SIV
CPZ
. VpU is

accession code 15513
(Received 31 May 2009, revised 2
September 2009, accepted 7 September
2009)
doi:10.1111/j.1742-4658.2009.07363.x
The HIV-1 encoded virus protein U (VpU) is required for efficient viral
release from human host cells and for induction of CD4 degradation in the
endoplasmic reticulum. The cytoplasmic domain of the membrane protein
VpU (VpUcyt) is essential for the latter activity. The structure and dynam-
ics of VpUcyt were characterized in the presence of membrane simulating
dodecylphosphatidylcholine (DPC) micelles by high-resolution liquid state
NMR. VpUcyt is unstructured in aqueous buffer. The addition of DPC
micelles induces a well-defined membrane proximal a-helix (residues I39–
E48) and an additional helical segment (residues L64–R70). A tight loop
(L73–V78) is observed close to the C-terminus, whereas the interhelical lin-
ker (R49–E63) remains highly flexible. A 3D structure of VpUcyt in the
presence of DPC micelles was calculated from a large set of proton–proton
distance constraints. The topology of micelle-associated VpUcyt was
derived from paramagnetic relaxation enhancement of protein nuclear spins
after the introduction of paramagnetic probes into the interior of the
micelle or the aqueous buffer. Qualitative analysis of secondary chemical
shift and paramagnetic relaxation enhancement data in conjunction with
dynamic information from heteronuclear NOEs and structural insight from
homonuclear NOE-based distance constraints indicated that micelle-associ-
ated VpUcyt retains a substantial degree of structural flexibility.
Abbreviations
DHPC, dihexanoyl phosphatidylcholine; DPC, dodecylphosphatidylcholine; DPC-d38, perdeuterated DPC; HSQC, heteronuclear single
quantum coherence; PRE, paramagnetic relaxation enhancement; TFE, trifluoroethanol; VpU, virus protein U; VpUcyt, C-terminal,
cytoplasmic domain of VpU (residues 39–81) plus N-terminal Gly-Ser dipeptide; b-TrCP, b-transducin repeat-containing protein; TASK,
TWIK-related acid-sensitive K

idues (Y27–K38, notation according to strain HV1S1)
connects the transmembrane part (I6–V26) and the
extremely acidic cytoplasmic domain (I39–L81). The
transmembrane domain consists of a well-characterized
and defined a-helix (referred to as helix 1) [17–20]. The
structure of the cytoplasmic domain was investigated
by various groups under diverse solution conditions
and at different levels of sophistication. VpU-derived
peptides were studied in native buffer [21], in trifluoro-
ethanol (TFE) solution [22,23], under high salt condi-
tions [24], in the presence of detergent micelles of
dihexanoyl phosphatidylcholine (DHPC) [25] or dod-
ecylphosphatidylcholine (DPC) [26], and associated to
phospholipid membranes [1,27,28]. There is consensus
on the formation of two cytoplasmic helices (helices 2
and 3) in various solvent conditions, but the extension
of helix 3 varies substantially [23–26]. The observation
of additional structural elements and, possibly, a ter-
tiary fold of the cytoplasmic domain of VpU remains
highly debated. To date, the most detailed descriptions
of the soluble VpU region are based on proton–proton
distances derived from solution NMR. Unfortunately,
these studies have been conducted in 50% TFE [22,23]
or in buffer containing 500 mm sodium sulfate [24],
conditions that might induce artificial conformations.
TFE stabilizes the secondary structure and supports the
formation of a-helices [29]. Furthermore, TFE may
weaken the tertiary structure by destabilizing hydro-
phobic interactions [30–33]. Very high ionic strength
appeared to induce a tertiary fold in the cytoplasmic

nounced change in the CD spectrum, whereas increasing
the detergent concentration further enhanced the helical
character only moderately and a clear saturation of the
effect was observed. In particular, the amount of unor-
dered secondary structure elements could be estimated
to be clearly more than 80% in the absence of DPC.
Upon the addition of 100 mm DPC, a substantial
Fig. 1. CD spectra of VpUcyt (53 lM) in sodium phosphate buffer
with and without membrane-mimicking DPC micelles.
M. Wittlich et al. NMR structure of micelle-associated HIV-1 VpUcyt
FEBS Journal 276 (2009) 6560–6575 ª 2009 The Authors Journal compilation ª 2009 FEBS 6561
fraction of  30% a-helical secondary structure and
 16% turn formed, whereas only  40% of unordered
conformations remained.
NMR spectroscopy and resonance assignment
The experimental conditions for the NMR study of
VpUcyt in the presence of membrane-mimicking
micelles were carefully optimized. First, various combi-
nations of buffer composition, choice of detergent and
temperature were tested. High-quality
15
N-
1
H-hetero-
nuclear single quantum coherence (HSQC) spectra
with enhanced spectral dispersion and the expected
number of cross-peaks of VpUcyt were obtained with
perdeuterated DPC (DPC-d38) at 30 °C. A series of
HSQC spectra of 1 mm VpUcyt was recorded with
varying amounts of DPC-d38 in the sample (from 0 to

13
C-labelled VpUcyt. Resonance assignment tables
were deposited at the BMRB (accession code: 15513).
An overlay of
15
N-
1
H-HSQC spectra of VpUcyt
recorded in detergent-free buffer (black) and in the
presence of 100 mm DPC-d38 (red) is shown in
Fig. 2A. The corresponding chemical shift changes in
backbone amide
1
H and
15
N spins of VpUcyt, together
with a weighted average of the absolute changes, are
presented as a function of the amino acid sequence
position in Fig. 2B–D. Continuous stretches with
prominent chemical shift changes were observed for
residues 39–51 and 64–78. In contrast, amide chemical
shifts of residues 52–63 were virtually unaffected by
the presence of DPC micelles.
Local helix propensity derived from chemical
shifts
The difference between the observed chemical shifts of a
protein and the corresponding amino acid residue-spe-
cific random coil chemical shifts is referred to as second-
ary chemical shift. In particular,
13

the addition of DPC are shown as a function of sequence position.
A measure of the total chemical shift change {D
total
d =[(Dd
1
H)
2
+ (0.1 · Dd
15
N)
2
]
1 ⁄ 2
} [88] is presented in (D). No amide corre-
lation of L42 was observed in the detergent-free sample.
NMR structure of micelle-associated HIV-1 VpUcyt M. Wittlich et al.
6562 FEBS Journal 276 (2009) 6560–6575 ª 2009 The Authors Journal compilation ª 2009 FEBS
a-helix is indicated by downfield shifts of
13
C
a
and
13
CO and upfield shifts of
1
H
a
resonances with average
changes of 2.6, 1.7 and 0.37 p.p.m., respectively,
whereas b-sheet conformation is indicated by shifts in

values expected for a regular helix. This procedure
provided fractional helicities of  80% (Dd
13
C
a
) for
helix 2 and 40% (Dd
1
H
a
) for helix 3.
NOE-derived secondary structure and tertiary fold
of VpUcyt in the presence of DPC micelles
2D
15
N-edited NOESY spectra of VpUcyt recorded
with and without DPC micelles in the sample were
very different (Fig. 4). The number of cross-peaks was
rather limited in DPC-free buffer, but strongly
increased upon the addition of micelles. In particular,
a substantial number of d
NN
(i,i + 1) cross-peaks
emerged near the diagonal, indicating helical segments.
Extensive signal overlap in 2D NOESY spectra in
combination with rather broad lines was overcome by
Fig. 3. Secondary chemical shifts of VpUcyt
observed in the absence (left) and presence
(right) of detergent micelles (100 m
M DPC-

M
DPC-d38 using identical acquisition and processing parameters.
M. Wittlich et al. NMR structure of micelle-associated HIV-1 VpUcyt
FEBS Journal 276 (2009) 6560–6575 ª 2009 The Authors Journal compilation ª 2009 FEBS 6563
the acquisition of heteronuclear-edited 3D NOESY
experiments of VpUcyt in DPC-containing buffer.
Secondary structure-specific short- and medium-
range NOEs are summarized in Fig. 5. Strong
d
NN
(i,i + 1) cross-peaks in conjunction with less
intense d
aN
(i,i + 1) peaks and continuous stretches of
d
aN
(i,i + 3), d
ab
(i,i + 3), and perhaps d
aN
(i,i +4) or
d
aN
(i,i + 2) peaks are indicative of helices. The two
helices deduced from the chemical shift data (Fig. 3)
are also clearly discernable in the NOE diagram. Helix
2 exhibits all classes of NOE cross-peaks expected.
Helix 3 displays several d
aN
(i,i + 2) peaks in addition

(i,i + 1) cross-peaks, a feature that is incompatible
with a rigid regular helical structure [39].
Calculation of the VpUcyt structure in micelle solu-
tion employed 604 upper distance limits derived from
unambiguous NOESY cross-peaks. The set of
1
H-
1
H
distances consisted of 147 intraresidue, 223 sequential,
219 medium-range (2 £ |i ) j| £ 5), and 15 long-range
(|i ) j| > 5) constraints. All 15 long-range connectivi-
ties were encoded by weak NOEs that gave rise to upper
distance limits of 0.55 nm (Table 1). Stripes from a
13
C-
resolved NOESY experiment exemplifying long-range
NOEs of VpUcyt in the presence of DPC micelles are
shown in Fig. 6. Long-range NOEs are crucial for delin-
eating the tertiary fold of VpUcyt. Multiple long-range
Fig. 5. Summary of
1
H-
1
H connectivities of VpUcyt in DPC micelle solution derived from 3D NOESY spectra. The amino acid sequence of
VpUcyt is shown at the top. Capital letters denote residues 39–81 of VpU. The N-terminal Gly-Ser dyad in lower case remains on VpUcyt
after thrombin cleavage of the fusion protein. In case of sequential HN(i) ⁄ HN(i + 1) and H
a
(i) ⁄ HN(i + 1) cross-peaks, the height of the boxes
is proportional to the estimated NOE intensities. Observation of additional classes of NOE interactions is visualized in the six rows below.

c
4.430 2.457; 2.461
A50-CH
3
b
-P75-CH
2
d
1.509 1.846; 1.940
G54-CH
2
a
-W76-H
a
4.070 4.777
H72-H
d
-V78-H
a
7.332 4.207
H72-H
d
-V78-CH
3
c
7.332 1.017
H72-H
e
-V78-CH
3

c
-E69-CH
2
b
; D44-H
a
-R70-CH
2
c
).
This subset of NOEs indicated an approximately anti-
parallel arrangement of helices 2 and 3. Other long-
range NOEs were observed between the interhelical
linker and C-terminal residues (A50-CH
3
b
-P75-CH
2
d
;
G54-CH
2
a
-W76-H
a
) and between C-terminal residues
defining a tight loop (H72-H
d
-V78-CH
3

3
c
and E69 CH
2
b
should be accompanied by a
detectable NOE between I43 CH
3
d
and E69 CH
2
c
. How-
ever, the corresponding cross-peak exists, but is highly
ambiguous due to spectral overlap. The assignment of
all observed long-range NOEs was carefully checked.
Only unambiguously identified long-range NOEs were
used for structure calculation. This conservative
approach explains the limited number of long-range
NOEs that were employed for structure calculation as
listed in Table 1.
A set of 100 VpUcyt conformers was generated by the
program cyana starting from randomized conforma-
tions and using the 604 distance constraints as the only
experimental input. The 20 structures with the lowest
energy were selected for statistical analysis (Table 2).
The entire set of 604 distance restraints was reasonably
well satisfied in all 20 conformers; the maximum dis-
tance violation amounted to 0.021 nm. The geometric
quality of the calculated structures was acceptable; 89%

exemplary long-range
1
H-
1
H NOEs.
Table 2. Analysis of the 20 lowest energy VpUcyt structures in
DPC micelles.
Experimental restraints
Total NOE restraints 604
Intraresidue 147
Sequential 223
Medium range (2 £ |i ) j| £ 5) 219
Long range (|i ) j| > 5) 15
CYANA structural statistics
Target function (nm
2
) 0.0025 ± 0.0006
Sum of NOE violations
a
> 0.015 nm 0.25 nm
Maximum NOE violation in the ensemble 0.021 nm
rmsd to mean structure (nm) (backbone only ⁄ all heavy atoms)
VpUcyt (39–80) 0.100 ⁄ 0.149
Helix 2 (39–48) 0.027 ⁄ 0.081
Interhelical linker (49–63) 0.074 ⁄ 0.141
Helix 3 (64–70) 0.030 ⁄ 0.091
Tight loop (73–78) 0.016 ⁄ 0.046
Helix 3 and loop (64–78) 0.049 ⁄ 0.088
Ramachandran analysis
Most favoured 58%

investigate whether the reduced structural definition of
the interhelical and C-terminal regions was due to
increased mobility of the respective residues. Hetero-
nuclear
1
H-
15
N NOE data reflect local variations in
protein backbone dynamics on the pico- to nano-
second time scale. Positive
1
H-
15
N NOE values close
to 0.8 are expected in the absence of fast internal
motions of protein backbone N-H bond vectors [42].
Rapid internal motion will reduce the NOE, which
may even become negative for highly mobile residues
exhibiting large amplitude motions on a sub-nano-
second time scale [38].
Figure 8 shows
1
H-
15
N NOEs of VpUcyt backbone
amides in the presence and absence of DPC micelles.
Small and rather consistent
1
H-
15

B
Fig. 8.
1
H-
15
N-hetero-NOE values of backbone amides of VpUcyt in
the absence (A) and presence (B) of 100 m
M DPC-d38. Intensities
of R45 and V68 could not be determined due to heavy signal over-
lap; P75 lacks a backbone amide group. Helical regions and a tight
loop of VpUcyt are denoted by grey stripes and an open rectangle,
respectively.
Helix 3
Loop
Linke
r
Helix 2
N
C
Fig. 7. Backbone line representation of the 20 lowest energy con-
formers of VpUcyt calculated from distance restraints in 100 m
M
DPC-d38 solution (left). The overlay is based on minimizing the
rmsd between amino acid residues 39–78. The ribbon diagram of a
low-energy conformer of VpUcyt is shown on the right. The tight
loop (residues 73–78) is shown as a green worm. Side chains of
the Ser53 and Ser57 in the interhelical linker, forming a highly con-
served phosphorylation motif, are visualized in ball-and-stick format.
NMR structure of micelle-associated HIV-1 VpUcyt M. Wittlich et al.
6566 FEBS Journal 276 (2009) 6560–6575 ª 2009 The Authors Journal compilation ª 2009 FEBS

data indicated a location of helices 2 and 3, as well as
of the residues between helix 3 and the loop, in the
micelle–water interface region. The highly anionic
interhelical linker was solvent exposed and fully acces-
sible to the Mn
2+
ions. Residues at both ends of the
interhelical linker were superficially associated with the
micelle interface and remained easily accessible by
water-soluble Mn
2+
ions. The strongly charged C-ter-
minal end of VpU (D
77
-VDD-L
81
) was partially pro-
tected from quenching by the doxyl probe and the
protection level increased towards the C-terminus. Fur-
thermore, resonances of these last five residues became
almost undetectable in the presence of Mn
2+
. The sol-
vent-exposed C-terminus of VpUcyt probably pointed
away from the surface of the micelle. The three amide
cross-peaks originating from the hydrophobic cluster
L
73
-AP-W
76

B
Fig. 9. PRE data of VpUcyt in DPC micelles. Paramagnetic probes
16-doxylstearic acid in the interior of the micelle (A) and Mn
2+
in
the aqueous buffer (B) selectively attenuate distinct regions of VpU-
cyt, reflecting the topology and the dynamics of the micelle-associ-
ated protein. Helical regions and a tight loop of VpUcyt are denoted
by grey stripes and an open rectangle, respectively.
M. Wittlich et al. NMR structure of micelle-associated HIV-1 VpUcyt
FEBS Journal 276 (2009) 6560–6575 ª 2009 The Authors Journal compilation ª 2009 FEBS 6567
(Fig. 7). The number of residues in helical regions of
the NMR-derived structures is in good agreement with
the  30% a-helical secondary structure content esti-
mated from the CD spectra of VpUcyt in 100 mm
DPC. Helix 2 of VpUcyt in DPC micelles is slightly
shorter at the C-terminal end in comparison with the
corresponding helix in earlier studies on VpU peptides
in 50% TFE (helix 2 spans residues 37–51) [22,23], in
DHPC micelles (helix 2 spans residues 30–49) [25], or
in high salt buffer (helix 2 spans residues 40–50) [24].
The N-terminal start of helix 2 cannot be compared
due to the different lengths of the VpU peptides
studied.
The length and sequence position of helix 3 differ
appreciably between studies. It is shortest in the pres-
ence of DPC micelles (residues 64–70), slightly
extended in high salt buffer (residues 60–68) [24], but
approximately twice as long in DHPC micelles (resi-
dues 58–70) [25] and in 50% TFE (residues 57–72)

NOESY data recorded on a protein undergoing
dynamic exchange contain contributions from different
conformations. A faithful reconstruction of the confor-
mational ensemble that gives rise to the observed spec-
trum is not straightforward. The single tertiary fold of
VpUcyt derived from the experimental NOE data
should therefore be considered as a ‘limit’ structure. A
limit structure does not necessarily represent the time
and population-weighted mean structure of a protein,
but may contain structural motifs from several, more
or less different conformations in dynamic exchange.
The observed secondary structure elements and the ter-
tiary contacts may be present to a different extent in
each individual conformation. Interestingly, the com-
plete set of upper distance constraints extracted from
NOESY experiments on VpUcyt in the presence of
DPC micelles is simultaneously satisfied in the con-
verged low-energy VpUcyt conformers presented in
Fig. 7. We conclude that the tertiary fold described
here is feasible and might be adopted by a substantial
fraction of micelle-bound VpUcyt.
Topology of micelle-bound VpUcyt
The position of VpUcyt relative to the micelle–water
interface was uncovered by selective PRE of protein
nuclear spins. The paramagnetic agents employed were
confined either to the hydrophobic interior of the
Fig. 10. Surface representations of VpUcyt with amino acids colour
coded based on PRE data (top). The green colour indicates residues
that are mainly affected by 16-doxylstearic acid, suggesting spacial
proximity to the interior of the micelle. The red colour indicates res-

probe, are coloured in red. Intermediate behaviour is
indicated by shades of light green, yellow and orange,
reflecting increasing water accessibility in this order.
Opposite faces of the same VpUcyt structure are dis-
played in the upper row of Fig. 10. The orientation of
the presented molecule was manually adjusted to
reflect the PRE data in the following way: green-col-
oured regions of the protein that appear to be close to
the core of the micelle but distant from water are
pointing downwards; red-coloured elements that
should be highly water exposed are positioned as close
as possible to the upper edge of the drawing area. The
vertical arrow represents the normal vector of the
micelle–water interface. Each surface plot and the cor-
responding ribbon representation shown underneath
depict the same orientation of VpUcyt.
The question arises, can the NOE-derived tertiary
fold of VpUcyt be reconciled with the residue-specific
PRE data in Fig. 9? The orientation of VpUcyt rela-
tive to the micelle normal shown in Fig. 10 is com-
patible with many, but not all, of the PRE data.
Helices 2 and 3, as well as the amino acids located
between helix 3 and the tight loop, partially dive into
the micelle and are largely shielded from water
(green). In contrast, central residues of the interhelical
linker extend away from the detergent–water interface
(red). Both ends of the interhelical linker and the last
three residues of VpUcyt exhibit intermediate quench-
ing characteristics (orange) and may occupy a region
close to both the interior of the micelle and the aque-

Viral proteins such as HIV-1 Vpr and Tat, together
with many others, are often referred to as fully or
partially flexible, intrinsically unstructured, or natively
unfolded proteins. Under standard solution condi-
tions, such proteins show a high degree of conforma-
tional disorder and flexibility. These proteins
frequently possess propensities for various secondary
structure elements that are adopted only temporarily
and ⁄ or in a fraction of the protein population.
Recent data suggest that even proteins that adopt a
well-defined structure by conventional standards may
exhibit minor populations of additional conforma-
tions. Some of those transiently formed conforma-
tions may be perfectly suited for a selected protein
ligand interaction. The distinguished protein confor-
mation is then recognized by the binding partner.
Directed withdrawal of a particular conformational
subpopulation from the equilibrium is counteracted
by a continuous readjustment of the conformational
ensemble [44]. This scenario of ‘conformational selec-
tion’ was recently proposed as an alternative to the
traditional ‘induced fit’ model of protein interactions
[44].
Viral proteins often target numerous cellular factors.
A diversified set of protein conformational subpopula-
tions is required for productive interaction with multi-
ple targets in the frame of the ‘conformational
selection’ model. Proteins referred to as ‘intrinsically
unstructured’ or ‘natively unfolded’ may therefore be
well adapted for interaction with diverse partners. This

NMR. Our data show that VpUcyt becomes at least
partially structured in the presence of membrane-mim-
icking DPC micelles. In full-length VpU, the cytoplas-
mic domain is indifferently anchored to a lipid
membrane. We propose that the structure of micelle-
associated VpUcyt is a reasonable approximation of
the physiologically relevant membrane-attached cyto-
plasmic region of VpU. This view is supported by the
experimentally confirmed location of VpUcyt at the
micelle–water interface. The polar headgroup of DPC
is chemically identical to that of the large fraction of
phospholipids in biological membranes that feature a
phosphatidylcholine headgroup. The observed struc-
ture of micelle-bound VpUcyt is obviously very differ-
ent from the completely unordered VpUcyt in plain
buffer. The association of peptides with lipid mem-
branes often has a pronounced influence on peptide
structure and may be a crucial prerequisite for produc-
tive interaction of a peptide or protein domain with its
membrane receptor [61].
VpU-induced proteasomal degradation of newly syn-
thesized CD4 in the endoplasmic reticulum requires
post-translational phosphorylation of VpU residues
S53 and S57 by casein kinase type II [62]. These two
serines are located in the interhelical linker, which
retains a high degree of structural flexibility upon the
addition of DPC micelles. Chemical shifts of residues
in the linker region are almost unaffected (Figs 2 and
3) and their backbone mobility remains high in the
presence of membrane-mimicking micelles (Fig. 8B).

side chain in a folded protein domain. It remains an
open question: does the low-energy structure calculated
from the measured NMR data represent the active
CD4-binding conformation of VpU? Only an eagerly
awaited complex structure of the interacting domains
will finally answer this question.
Materials and methods
VpUcyt expression and purification
The C-terminal cytoplasmic domain of VpU(residues 39–
81; VpU residue numbering refers to HIV-1 strain HV1S1,
Swiss-Prot accession number P19554) was produced as a
recombinant protein fusion with an N-terminal glutathione
S-transferase affinity tag in Escherichia coli. Thrombin
cleavage of the fusion releases the soluble 45-residue VpU-
cyt polypeptide comprising VpU(39–81) and a preceding
Gly-Ser dipeptide. The amino acid sequence of VpUcyt is
given in Fig. 5. The detailed protocol for cloning, expres-
sion and purification of uniformly
15
N- and
13
C-labelled
VpUcyt has been published previously [26].
NMR spectroscopy
Lyophilized VpUcyt was dispensed at a concentration of
1mm in  330 lL NMR buffer [20 mm sodium phosphate,
pH 6.2, 100 mm NaCl, 0.02% (w ⁄ v) NaN
3
, 10% (v ⁄ v)
2

NMR spectra were recorded with an HCN-cold probe with
cryogenically cooled
1
H-coil and preamplifier circuitry. The
WATERGATE sequence was used for suppression of the
water signal [64]. Proton and
13
C chemical shifts were
referenced directly to sodium 3-(trimethylsilyl)propane-1-
sulfonate that had been added as an internal standard,
whereas
15
N chemical shifts were referenced indirectly to
sodium 3-(trimethylsilyl)propane-1-sulfonate. NMR data
were processed with vnmrj (Varian) or nmrpipe [65] and
analysed with cara [66].
Resonance assignment
Sequential assignment of the
1
H,
13
C and
15
N resonances of
VpUcyt was accomplished using a combination of 2D and
3D NMR spectra:
1
H-
15
N-HSQC [67,68],

Oor
2
H
2
O, respectively. To
check for potential spin diffusion, a series of NOESY spec-
tra were recorded with mixing times from 100 to 400 ms.
The intensity of the observed cross-peaks increased linearly
with mixing time. Apart from this intensity increase, the
NOESY cross-peak pattern looked qualitatively almost
identical. No indication of spin diffusion was found in any
of the spectra. In particular, every single pair of proton res-
onances (A and B) that showed a long-range NOE was
checked for the possible existence of a shared relaxation
partner C with strong cross-peaks between C and both
spins A and B. No such common relaxation partner was
found for any of the long-range NOEs in Table 1. Finally,
the NOESY spectra with a mixing time of 200 ms were
used for complete analysis.
A list of upper distance constraints for structure calcula-
tion was derived from the NOESY data using the auto-
mated NOESY analysis software radar developed by the
Wu
¨
thrich group at ETH in Zu
¨
rich (.
biol.ethz.ch/groups/wuthrich_group/software). radar com-
bines the previously described components atnos for
automated NOESY cross-peak picking and NOE signal

generated list. Newly identified unique NOEs were manu-
ally assigned, calibrated and converted into additional
upper distance constraints.
The 3D structure of micelle-associated VpUcyt was cal-
culated on the basis of the final set of proton–proton dis-
tance constraints using cyana, a software program that
combines simulated annealing with molecular dynamics in
torsion angle space [80]. The calculated protein conforma-
tions were screened for secondary structure using the pro-
gram dssp [81]. The software molmol [82] was employed
for structure visualization. In addition, molmol generates a
Ramachandran plot for assessment of the geometric quality
of protein conformers. The coordinates of the 20 lowest
energy structures have been deposited in the RCSB Protein
Data Bank under accession code 2K7Y.
Secondary chemical shifts
The analysis of secondary chemical shifts presented here is
based on random coil values determined by Schwarzinger
et al. [83,84] using Ac-GGXGG-NH2 peptides in 8 m urea
and additional corrections for sequence effects.
Heteronuclear NOEs
Heteronuclear
1
H-
15
N NOEs were derived from 2D spectra
recorded with the NOE-TROSY pulse sequence [85]. Spectra
were acquired at 18.8 T with or without
1
H saturation dur-

Mn
2+
. The paramagnetic doxyl moiety is confined to the
hydrophobic interior of the micelle. It predominantly
broadens NMR signals of nuclei buried in the micelle. The
water-soluble Mn
2+
ions preferentially affect NMR signals
of solvent-exposed spins [86]. The effect of a paramagnetic
probe on individual amino acids is quantified in terms of
the percentage of NMR signal retention, which is derived
from comparison of HSQC cross-peak intensities of
VpUcyt in micellar solution observed in the absence or
presence of the paramagnetic agent.
The effect of different concentrations of either 16-doxylstea-
ric acid (between 0.2 and 8 mm)orMn
2+
(between 0.1 and
3mm) on the NMR signals of VpUcyt (1 mm) in NMR buffer
containing 100 mm DPC-d38 was tested in an initial screen.
The results reported here were obtained at the optimal con-
centrations of 0.1 mm Mn
2+
or 5 mm 16-doxylstearic acid.
CD spectroscopy
Samples contained 53 lm VpUcyt and various amounts (0,
5, 10, 100 mm) of DPC-d38 in 20 mm sodium phosphate
buffer (pH 6.2) complemented with 100 mm KF. KF was
used for ionic strength adjustment instead of NaCl in order
to avoid the strong absorption of chloride ions at low

Correlation of the structural and functional domains in
the membrane protein Vpu from HIV-1. Proc Natl
Acad Sci USA 96, 14336–14341.
2 Schubert U & Strebel K (1994) Differential activities
of the human immunodeficiency virus type 1-encoded
Vpu protein are regulated by phosphorylation and
occur in different cellular compartments. J Virol 68,
2260–2271.
3 Margottin F, Bour SP, Durand H, Selig L, Benichou S,
Richard V, Thomas D, Strebel K & Benarous R (1998) A
novel human WD protein, h-beta TrCp, that interacts
with HIV-1 Vpu connects CD4 to the ER degradation
pathway through an F-box motif. Mol Cell 1, 565–574.
4 Butticaz C, Michielin O, Wyniger J, Telenti A & Ro-
thenberger S (2007) Silencing of both beta-TrCP1 and
HOS (beta-TrCP2) is required to suppress human
immunodeficiency virus type 1 Vpu-mediated CD4
down-modulation. J Virol 81, 1502–1505.
5 Schubert U, Henklein P, Boldyreff B, Wingender E,
Strebel K & Porstmann T (1994) The human immuno-
deficiency virus type 1 encoded Vpu protein is phos-
phorylated by casein kinase-2 (CK-2) at positions Ser52
and Ser56 within a predicted alpha-helix-turn-alpha-
helix-motif. J Mol Biol 236, 16–25.
6 Estrabaud E, Le Rouzic E, Lopez-Verges S, Morel M,
Belaidouni N, Benarous R, Transy C, Berlioz-Torrent
C & Margottin-Goguet F (2007) Regulated degradation
of the HIV-1 Vpu protein through a betaTrCP-indepen-
dent pathway limits the release of viral particles. PLoS
Pathog 3, e104.

Gag is clathrin-independent. Virol 369, 299–308.
15 Zhang L, Huang Y, Yuan H, Tuttleton S & Ho DD
(1997) Genetic characterization of vif, vpr, and vpu
sequences from long-term survivors of human immuno-
deficiency virus type 1 infection. Virol 228, 340–349.
16 Stephens EB, McCormick C, Pacyniak E, Griffin D,
Pinson DM, Sun F, Nothnick W, Wong SW,
Gunderson R, Berman NE et al. (2002) Deletion of the
vpu sequences prior to the env in a simian-human
immunodeficiency virus results in enhanced Env
precursor synthesis but is less pathogenic for pig-tailed
macaques. Virol 293, 252–261.
17 Park SH, Mrse AA, Nevzorov AA, Mesleh MF,
Oblatt-Montal M, Montal M & Opella SJ (2003)
Three-dimensional structure of the channel-forming
trans-membrane domain of virus protein ‘‘u’’ (Vpu)
from HIV-1. J Mol Biol 333, 409–424.
18 Park SH, De Angelis AA, Nevzorov AA, Wu CH &
Opella SJ (2006) Three-dimensional structure of the
transmembrane domain of Vpu from HIV-1 in aligned
phospholipid bicelles. Biophys J 91, 3032–3042.
19 Park SH & Opella SJ (2007) Conformational changes
induced by a single amino acid substitution in the
trans-membrane domain of Vpu: implications for
HIV-1 susceptibility to channel blocking drugs. Protein
Sci 16, 2205–2215.
20 Sharpe S, Yau WM & Tycko R (2006) Structure and
dynamics of the HIV-1 Vpu transmembrane domain
revealed by solid-state NMR with magic-angle spinning.
Biochemistry 45, 918–933.

consequences of phosphorylation of two serine residues
in the cytoplasmic domain of HIV-1 VpU. J Pept Sci
14, 804–810.
27 Henklein P, Kinder R, Schubert U & Bechinger B
(2000) Membrane interactions and alignment of struc-
tures within the HIV-1 Vpu cytoplasmic domain: effect
of phosphorylation of serines 52 and 56. FEBS Lett
482, 220–224.
28 Kochendoerfer GG, Jones DH, Lee S, Oblatt-Montal
M, Opella SJ & Montal M (2004) Functional character-
ization and NMR spectroscopy on full-length Vpu from
HIV-1 prepared by total chemical synthesis. J Am Chem
Soc 126, 2439–2446.
29 So
¨
nnichsen FD, Van E-JE, Hodges RS & Sykes BD
(1992) Effect of trifluoroethanol on protein secondary
structure: an NMR and CD study using a synthetic
actin peptide. Biochemistry 31, 8790–8798.
30 Klaus W, Dieckmann T, Wray V, Schomburg D,
Wingender E & Mayer H (1991) Investigation of the
solution structure of the human parathyroid hormone
fragment (1–34) by 1H NMR spectroscopy, distance
geometry, and molecular dynamics calculations.
Biochemistry 30, 6936–6942.
31 Alexandrescu AT, Ng YL & Dobson CM (1994) Charac-
terization of a trifluoroethanol-induced partially folded
state of alpha-lactalbumin. J Mol Biol 235, 587–599.
32 Sticht H, Willbold D, Ejchart A, Rosin Arbesfeld R,
Yaniv A, Gazit A & Ro

39 Koenig BW, Ferretti JA & Gawrisch K (1999) Site-spe-
cific deuterium order parameters and membrane-bound
behavior of a peptide fragment from the intracellular
domain of HIV-1 gp41. Biochemistry 38, 6327–6334.
40 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.
41 Laskowski RA, Rullmannn JA, MacArthur MW,
Kaptein R & Thornton JM (1996) AQUA and
PROCHECK-NMR: programs for checking the quality
of protein structures solved by NMR. J Biomol NMR
8, 477–486.
42 Kay LE, Torchia DA & Bax A (1989) Backbone
dynamics of proteins as studied by 15N inverse detected
heteronuclear NMR spectroscopy: application to staph-
ylococcal nuclease. Biochemistry 28, 8972–8979.
43 Gonzalez ME & Carrasco L (2003) Viroporins. FEBS
Lett 552, 28–34.
44 Lange OF, Lakomek NA, Fares C, Schroder GF, Wal-
ter KF, Becker S, Meiler J, Grubmuller H, Griesinger
C & de Groot BL (2008) Recognition dynamics up to
microseconds revealed from an RDC-derived ubiquitin
ensemble in solution. Science 320, 1471–1475.
45 Willbold D, Kruger U, Frank R, Rosin-Arbesfeld R,
Gazit A, Yaniv A & Rosch P (1993) Sequence-specific
resonance assignments of the 1H-NMR spectra of a
synthetic, biologically active EIAV Tat protein. Bio-
chemistry 32, 8439–8445.
46 Willbold D, Rosin-Arbesfeld R, Sticht H, Frank R &

54 Engler A, Stangler T & Willbold D (2002) Structure of
human immunodeficiency virus type 1 Vpr(34–51)
peptide in micelle containing aqueous solution. Eur J
Biochem 269, 3264–3269.
55 Bruns K, Fossen T, Wray V, Henklein P, Tessmer U &
Schubert U (2003) Structural characterization of the
HIV-1 Vpr N terminus: evidence of cis ⁄ trans-proline
isomerism. J Biol Chem 278, 43188–43201.
56 Morellet N, Bouaziz S, Petitjean P & Roques BP (2003)
NMR structure of the HIV-1 regulatory protein VPR.
J Mol Biol 327, 215–227.
57 Seewald MJ, Metzger AU, Willbold D, Rosch P &
Sticht H (1998) Structural model of the HIV-1 Tat(46–
58)-TAR complex. J Biomol Struct Dyn 16, 683–692.
58 Anand K, Schulte A, Vogel-Bachmayr K, Scheffzek K
& Geyer M (2008) Structural insights into the cyclin
T1-Tat-TAR RNA transcription activation complex
from EIAV. Nat Struct Mol Biol 15, 1287–1292.
59 Schubert U, Bour S, Ferrer-Montiel AV, Montal M,
Maldarell F & Strebel K (1996) The two biological
activities of human immunodeficiency virus type 1 Vpu
protein involve two separable structural domains.
J Virol 70, 809–819.
60 Paul M, Mazumder S, Raja N & Jabbar MA (1998)
Mutational analysis of the human immunodeficiency
virus type 1 Vpu transmembrane domain that promotes
the enhanced release of virus-like particles from the
plasma membrane of mammalian cells. J Virol 72,
1270–1279.
61 Sargent DF & Schwyzer R (1986) Membrane lipid

69 Kay LE, Keifer P & Saarinen T (1992) Pure absorption
gradient enhanced heteronuclear single quantum corre-
lation spectroscopy with improved sensitivity. JAm
Chem Soc 114, 10663–10665.
70 Wittekind M & Mueller L (1993) HNCACB, a
high-sensitivity 3D NMR experiment to correlate
amide-proton and nitrogen resonances with the a- and
ß-carbon resonances in proteins. J Magn Reson 101B,
201–205.
71 Ikura M, Kay LE & Bax A (1990) A novel approach
for sequential assignment of 1H, 13C, and 15N spectra
of proteins: heteronuclear triple-resonance three-dimen-
sional NMR spectroscopy. Application to calmodulin.
Biochemistry 29, 4659–4667.
72 Vuister GW & Bax A (1993) Quantitative J correlation:
a new approach for measuring homonuclear three-bond
J(HN-Ha) coupling constants in
15
N-enriched proteins.
J Am Chem Soc 115, 7772–7777.
73 Grzesiek S, Anglister J & Bax A (1993) Correlation of
backbone amide and aliphatic side-chain resonances in
13
C ⁄
15
N-enriched proteins by isotropic mixing of
13
C
magnetization. J Mag Reson 101B, 114–119.
74 Bax A, Clore M, Driscoll PC, Gronenborn AM, Ikura M

82 Koradi R, Billeter M & Wuthrich K (1996) MOLMOL:
a program for display and analysis of macromolecular
structures. J Mol Graph 14, 51–55.
83 Schwarzinger S, Kroon GJ, Foss TR, Wright PE &
Dyson HJ (2000) Random coil chemical shifts in acidic
8 M urea: implementation of random coil shift data in
NMRView. J Biomol NMR 18, 43–48.
84 Schwarzinger S, Kroon GJ, Foss TR, Chung J, Wright
PE & Dyson HJ (2001) Sequence-dependent correction
of random coil NMR chemical shifts. J Am Chem Soc
123, 2970–2978.
85 Zhu G, Xia Y, Nicholson LK & Sze KH (2000) Protein
dynamics measurements by TROSY-based NMR exper-
iments. J Magn Reson 143, 423–426.
86 Damberg P, Jarvet J & Graslund A (2001) Micellar
systems as solvents in peptide and protein structure
determination. Methods Enzymol 339, 271–285.
87 Sreerama N & Woody RW (2000) Estimation of protein
secondary structure from circular dichroism spectra:
comparison of CONTIN, SELCON, and CDSSTR
methods with an expanded reference set. Anal Biochem
287, 252–260.
88 Schweimer K, Hoffmann S, Bauer F, Friedrich U,
Kardinal C, Feller SM, Biesinger B & Sticht H (2002)
Structural investigation of the binding of a herpesviral
protein to the SH3 domain of tyrosine kinase Lck.
Biochemistry 41, 5120–5130.
M. Wittlich et al. NMR structure of micelle-associated HIV-1 VpUcyt
FEBS Journal 276 (2009) 6560–6575 ª 2009 The Authors Journal compilation ª 2009 FEBS 6575