MINIREVIEW
The capsid protein of human immunodeficiency
virus: designing inhibitors of capsid assembly
Jose
´
L. Neira
1,2
1 Instituto de Biologı
´
a Molecular y Celular, Universidad Miguel Herna
´
ndez, Alicante, Spain
2 Biocomputation and Complex Systems Physics Institute, Zaragoza, Spain
Introduction
HIV-1, the agent responsible for AIDS, belongs to the
retrovirus family. Its genome is formed by two copies
of a single-stranded RNA, which are encased within a
lipoprotein shell, together with several replicative and
accessory proteins. The virus genome is a model of
economic packaging because the virus uses complex
processing (both in the production and cleavage of
mRNAs and in the final viral polypeptides) to
generate several proteins that can work in the host
cell. Perhaps the most important example of this eco-
nomic architecture is the production of Gag and Gag-
Pol fusion proteins.
The GagPol fusion protein, formed by a ribosomal
frame shift event, is cleaved into the Gag and Pol
polypeptides by the viral protease. Further cleavage
of the Pol chain yields the viral integrase (p31), a
new protease (p10), the reverse transcriptase (p50)

capsid critically depends on CA–CA interactions, including CTD interac-
tion with itself and with the CA N-terminal domain, NTD. This minireview
reports on the search and the design of peptides and small organic com-
pounds that are able to interact with the CTD and ⁄ or CA of HIV-1. Such
molecules aim to disrupt and ⁄ or alter the oligomerization capability of
CTD. The different peptides designed so far interact with CTD mainly via
hydrophobic contacts with residues close or belonging to the interface
between the dimerization helices. A CTD-binding organic compound also
establishes hydrophobic contacts with regions involved in the interface
between the NTD and CTD. These results open new venues for the devel-
opment of new antiviral drugs that are able to interact with CA and ⁄ or its
domains, hampering HIV-1 assembly and infection.
Abbreviations
CA, capsid protein of HIV-1 (p24); CAC1, a CTD-binding designed peptide; CAI, a CTD-binding phage display peptide; CAP-1, N-(3-chloro-4-
methylphenyl)-N¢-{2-[({5-[(dimethylamino)-methyl]-2-furyl}-methyl)-sulfanyl] ethyl}-urea); CTD, C-terminal domain of CA, comprising residues
146–231 of the intact protein; CTDW184A ⁄ M185A, a double mutant of CTD with Ala substitutions at the positions Trp184 and Met185;
HSQC, heteronuclear single quantum coherence; MA, matrix protein of HIV-1; NC, nucleocapsid protein; NTD, N-terminal domain of CA,
comprising residues 1–145 of the intact protein; NYAD-1, an improved designed version of CAI.
6110 FEBS Journal 276 (2009) 6110–6117 ª 2009 The Author Journal compilation ª 2009 FEBS
Gag polyprotein, yields the matrix protein (MA),
the capsid protein (CA), the nucleocapsid protein
(NC) and p6 proteins, as well as two spacer pep-
tides.
Currently, most of the available drugs in the market
against AIDS target the reverse transcriptase and pro-
tease enzymes, produced by Pol. In general, the com-
pounds involved in the treatment of HIV infection
belong to one of the following types: (a) non-nucleo-
tide reverse transcriptase inhibitors; (b) nucleotide
reverse transcriptase inhibitors; and (c) protease inhibi-

peptides and one organic compound have been found
to inhibit the C-terminal domain of CA (CTD) dimer-
ization or, alternatively, its interactions with the N-ter-
minal domain of CA (NTD). This minireview focuses
on the structure of the complexes formed by CTD and
those molecules, as well as what can be learnt from
such studies. The use of drugs such as Bevirimat
(Panacos Pharmaceuticals, Watertown, MA, USA)
(although it is in Phase II trials) is not discussed
because its mechanism of action is not based on inhibi-
tion of CA self-assembly. Bevirimat is a novel HIV-1
maturation inhibitor that inhibits specifically the final
rate-limiting step in Gag processing; the drug prevents
the release of mature capsid protein from its precursor,
resulting in the production of immature, non-infectious
virus particles [15].
Before describing how those designed molecules
work and inhibit dimer formation, it is necessary to
describe in some detail the domain architecture of CA
of HIV-1. The CA protein is formed by two indepen-
dently folded domains, NTD and CTD, separated by a
flexible linker [8,16–18]; see also Fig. 1B in Mascare-
nhas and Musier-Forsyth [7]. The NTD (residues
1–146 in the numbering of the whole intact protein) is
composed of five coiled-coil a-helices (helices 1–5 of
CA), with two additional short a-helices (helices 6 and
7) following an extended proline-rich loop [16–18]. The
CTD domain (residues 147–231) is a dimer both in
solution and in the crystal form [8,19]. Each CTD
monomer is composed of a short 3

CTD. This observation has facilitated the initial search
for inhibitors of HIV-1 assembly by using simplified
experimental setups in vitro; if a small molecule is able
to dissociate CTD dimers in solution, then it could be
considered a potential inhibitor of CA assembly and
HIV-1 infectivity (dependent on that hypothesis always
being tested by direct experiments). Furthermore, the
same observation validates the view that the structures of
CTD small-molecule complexes can be physiologically
relevant, and thus they could provide important guide-
lines for the rational design of better assembly inhibitors,
as well as physiologically acceptable antiviral drugs.
Inhibition of CTD self-assembly by
peptides
Protein assembly processes can be considered as good
targets for antivirals because they depend only on recur-
ring weak interprotein interactions, and the disruption
of only a few of these interactions may be sufficient to
suppress infectivity. In principle, destabilization of the
CA dimeric species and, subsequently, HIV capsid
assembly can be carried out by using isolated CTD as a
competitive inhibitor [20,21]. Alternatively, virion
assembly can be hampered by the use of small peptides
or organic molecules which bind to: (a) CTD (or CA)
monomers and occupy the dimerization interface or
(b) nearby sites to the dimerization interface, thus weak-
ening the binding to the other monomer. Below, the
structural implications of peptides that are able to bind
to CTD are reviewed, as well as how they work based
on the reported structural studies.

mediated by the peptide because the protein eluted at
larger volumes in the presence of equimolar peptide
concentrations.
A quantitative value of the affinity constant of the
CAC1–CTD complex was determined by fluorescence
(using an anthraniloyl-labeled peptide), affinity chro-
matography and isothermal titration calorimetry. The
three techniques yielded similar values for the apparent
dissociation constant of the complex, in the order of
50 lm, which is only three- to five-fold higher than the
dissociation constant of dimeric CTD (10–18 lm) [8].
Because CAC1 is random-coil when isolated in solu-
tion, as shown by 1D-NMR experiments, it is impor-
tant to note that the above determined affinity
constant must take into account the entropy penalty to
fold the peptide. In view of this fact, and the previous
observation that not every residue energetically
involved in the CTD interface was contained within
CAC1 [23], the comparable affinity of the peptide–
CTD complex and the CTD–CTD dimer was unex-
pected. However, later studies have provided a possible
explanation for this observation. As in the CAC1 pep-
tide, a-helix 9 becomes only permanently structured
when the CTD monomers dimerize [25,26], and then a
substantial entropy penalty must be paid on the forma-
tion of both the peptide–CTD and the CTD–CTD
complexes [6]. As expected, CAC1 is able to efficienty
inhibit the in vitro assembly of the HIV-1 capsid
(R. Bocanegra and M. G. Mateu, unpublished results).
To map the region of CTD where CAC1 actually

an improved solubility, but also with a larger helicity
to assess whether the latter improves the affinity for
CTD.
A phage display peptide as an inhibitor of CA
assembly
A CTD-binding, 12-mer peptide, CAI, was identified
by Stitch et al. [27] using phage display techniques.
The sequence of CAI is: ITFEDLLDYYGP, which
does not bear any resemblance to the sequence of the
dimerization helix of CTD (see above). CAI is the first
reported peptide able to inhibit the assembly of the
mature HIV-1 capsid in vitro. However, because of the
lack of permeability of cells to CAI, the peptide was
unable to inhibit HIV-1 infection ex vivo. The CAI
peptide isolated in solution is random-coil, but it
adopts an a-helical conformation upon binding to
CTD, as shown by the X-ray structure of the nonmu-
tated-CTD–CAI complex [28] (Fig. 2A, main panel).
The aromatic moiety of Phe3, together with Leu6 and
Ile1, are the residues of CAI with the largest number
of protein contacts (Fig. 2A, inset); these data suggest
that the CTD–CAI interaction occurs mainly through
hydrophobic contacts.
The X-ray structure of the CAI–CTD complex is a
five helix bundle [28] (Fig. 2A, main panel). In this
bundle, the dimerization helix of CTD (a-helix 9)
reorients to maximize interactions with CAI; this
movement results in a change of the buried surface at
the dimeric interface when compared with the buried
interface in the dimeric CTD. The affinity constant of

in binding to CAC1 and vice versa (see above); these
data suggest that the modes of binding of both peptides
are quite similar, if not identical.
Bartonova et al. [30] have recently shown that resi-
dues Tyr169, Asn183, Glu187 and Leu211, all of which
form the binding pocket site of CAI, are necessary for
the quaternary structure integrity of CTD [30]. When
residues Tyr169, Asn183, Glu187 and Leu 211 were
replaced with alanine, the resulting mutants were
competent for immature capsid-like particle assembly
in vitro and budding of immature-like capsids. The
results obtained by Bartonova et al. [30] pinpoint an
indirect (allosteric) effect of CAI binding on the assem-
bly of the immature capsid lattice, but a direct binding
effect in the conserved CTD binding pocket during
mature assembly. On the other hand, mutations at
Tyr169 and Leu211 do not yield mature-like particles
and give rise to non-infectious virions with nonregular
mature cores. The X-ray structures of these two
mutants differ from that of wild-type, and those of the
assembly-competent mutants E187A and N183A: the
conformations of the two assembly-incompetent
mutants (Y169A and L211A) in the absence of CAI
are the same as the structures of any mutant when
complexed to the peptide. Interestingly, Glu187
(together with Ser178, Glu180 and Gln192) is one of
the residues that decreases the association constant of
CTD, as previously demonstrated in an alanine
scanning of the dimerization interface [23].
Barklis et al. [31] have shown that CAI does not

peptides and organic molecules. (A) Main
panel: structure of CTD (green) in complex
with the peptide CAI (purple) (accession no.
2buo) [28,30]; the helices are named after
the elements of structure in the wild-type
CTD in Fig. 1. Inset: the interface region is
shown with residues Phe3 (CAI) and Glu180
and Asn183 (CTD) indicated (the labels are
in different colours: blue for protein residues
and black for the peptide amino acids); for
the sake of clarity, only these two residues
of the CTD are shown. (B) Main panel:
structure of CTD (green) in complex with
the peptide NYAD-1 (cyan) (accession no.
2k1c) [33]; the helices are named after the
elements of structure in the wild-type CTD
in Fig. 1. Inset: the interface region is
shown with residues Phe3 (CAI) and
Asn183 (CTD) indicated (the labels are in dif-
ferent colours); for clarity, only Asn183 is
shown. (C) The structure of NTD in complex
with the small organic molecule CAP-1
(accession no. 2pxr) [36]; the helices are in
red, the b-sheets in yellow and the loops in
green. The circle at the bottom of the figure
indicates the region where the binding of
CAP-1 occurs. The figures were produced
using
PYMOL () [39].
CA small-molecule interactions J. L. Neira

the solution structure of the monomeric double mutant
CTDW184A ⁄ M185A in complex with NYAD-1
(Fig. 2B, main panel). In the structure, the a-helix 9 is
fully formed, with the kink present at the Thr188 [33],
in contrast to what happens with the structure of the
isolated monomeric double mutant [34]; thus, NYAD-
1 makes the dimerization helix of CTD adopt a more
native-like structure (if we assume native-like structure
to comprise that shown by each of the monomers in
the wild-type dimeric CTD). The peptide binds to a
hydrophobic pocket formed by residues of the four
helices, and it adopts a helical conformation (as CAI)
where its hydrophobic side chains (especially Phe3)
make extensive contacts with specific hydrophobic
patches of the double monomeric mutant (Fig. 2B,
inset). When a comparison of the NMR structure of
the complex NYAD-1-CTDW184A ⁄ M185A with the
X-ray structure of CAI-CTDW184A ⁄ M185A is carried
out, the differences are mostly observed in the
movement of a-helix 9 (the dimerization helix in the
wild-type nonmutated CTD); this helix is pushed away
much further from its original wild-type position in the
design of Stitch et al. [27] (6 A
˚
) than in that of
Zhang et al. [32] (3 A
˚
). It is tempting to suggest that
this short movement is also responsible for the larger
affinity constant of NYAD-1. Although the ability of

side chains and ⁄ or backbones of several residues in the
protein. Thus, most of the contacts of CAP-1 with CTD
are hydrophobic, similar to the peptides (see above).
Scheme 1. The CAP-1 and CAP-2 comp-
unds designed by Tang et al. [35] and Kelly
et al. [36].
J. L. Neira CA small-molecule interactions
FEBS Journal 276 (2009) 6110–6117 ª 2009 The Author Journal compilation ª 2009 FEBS 6115
Helices 4 and 7 of NTD form the interface required
to create a stable hexagonal CA lattice of the virion
capsid [37,38], and these helices pack against a-helices
8, 9 (the dimerization helix) and 11 of the CTD
domain. Thus, CAP-1 does not inhibit capsid assem-
bly through direct binding to CTD but rather because
it hampers the proper docking of both NTD and
CTD.
A second organic compound, CAP-2 (Scheme 1),
was shown also to bind the NTD of CA, but it is
cytotoxic and it was not investigated further [35].
Conclusions
In summary, it appears that even closely designed pep-
tides (such as CAI and NYAD-1) display subtle, but
key structural differences that account for the different
inhibition properties exhibited in vitro and in vivo.Itis
also clear that there are multiple ways to disrupt (or
better destabilize) the dimeric CTD because all the
peptide designs suggest slightly different modes of
binding. These modes of binding appear to be related
to: (a) the high flexibility of the a-helix-8- a-helix-9
interface (Fig. 1) and (b) a high number of hydropho-

´
a Hurtado-
Go
´
mez are acknowledged for their careful experimental
work. May Garcı
´
a, Marı
´
a del Carmen Fuster, Javier
Casanova and Olga Ruiz de los Pan
˜
os are deeply
thanked for providing excellent technical assistance.
References
1 Clerq EDe (2002) New developments in anti-
HIV chemotherapy. Biochim Biophys Acta 1587,
258–275.
2 Lucas GM, Chaisson EE & Moore RD (1999) Highly
active antiretroviral therapy in a large urban clinic: risk
factors for virologic failure and adverse drug reactions.
Ann Intern Med 131, 81–87.
3 Yerly S, Kaiser L, Race E, Bru JP, Clavel F & Perrin L
(1999) Transmission of antiretroviral-drug-resistant
HIV-1 variants. Lancet 354, 729–733.
4 Moore JP & Stevenson M (2000) New targets for inhib-
itors of HIV-1 replication. Nat Rev Mol Cell Biol 1,
41–49.
5 Neira JL (2009) Biophysical and structural studies on
the capsid protein of the human immunodeficiency virus

Trend Biochem Sci 29, 618–625.
14 Zutshi R, Brickner M & Chmielewski J (1998) Inhibit-
ing the assembly of protein–protein interfaces. Curr
Opin Struct Biol 2, 62–66.
CA small-molecule interactions J. L. Neira
6116 FEBS Journal 276 (2009) 6110–6117 ª 2009 The Author Journal compilation ª 2009 FEBS
15 Martin DE, Salzwedel K & Allaway GP (2008) Beviri-
mat: a novel maturation inhibitor for the treatment of
HIV-1 infection. Antivir Chem Chemother 19, 107–113.
16 Momany C, Kovari LC, Prongay AJ, Keller W, Gitti
RK, Lee BM, Gorbalenya AE, Tong L, McClure J,
Ehrlich LS et al. (1996) Crystal structure of dimeric
HIV-1 capsid protein. Nat Struct Biol 3, 763–770.
17 Gitti RK, Lee BM, Walker J, Summers MF, Yoo S &
Sundquist WI (1996) Structure of the amino-terminal
core domain of the HIV-1 capsid protein. Science 273,
231–235.
18 Gamble TR, Vajdos FF, Yoo S, Worthylake DK,
Houseweart M, Sundquist WI & Hill CP (1996) Crys-
tal structure of human cyclophilin A bound to the
amino-terminal domain of HIV-1 capsid. Cell 87,
1285–1294.
19 Worthylake DK, Wang H, Yoo S, Sundquist WI & Hill
CP (1999) Structures of the HIV-1 capsid protein
dimerization domain at 2.6 A
˚
resolution. Acta Crystal-
logr D 55, 85–92.
20 del Alamo M & Mateu MG (2005) Electrostatic repul-
sion, compensatory mutations, and long-range non-

of the capsid protein. Biophys J 93, 1264–1276.
26 Alcaraz LA, del A
´
lamo M, Mateu MG & Neira JL
(2008) Structural mobility of the monomeric Cterminal
domain of the HIV-1 capsid protein. FEBS J 275,
3299–3311.
27 Stitch J, Humbert M, Findlow S, Bodem J, Muller B,
Dietricht U, Werner M & Kra
¨
usslich H-G (2005) A
peptide inhibitor of HIV-1 assembly in vitro. Nat Struct
Mol Biol 12, 671–677.
28 Ternois F, Sticht J, Duquerroy S, Kra
¨
usslich H-G &
Rey FA (2005) The HIV-1 capsid protein C-terminal
domain in complex with a virus assembly inhibitor. Nat
Struct Mol Biol 12, 678–682.
29 Kovalevski BJ, Kennedy R, Khorchid A, Kleiman L,
Matsuo H & Musier-Forsyth K (2007) Critical role of
helix 4 of HIV-1 capsid C-terminal domain in interac-
tions with human lysyl-tRNA Synthetase. Proc Natl
Acad Sci USA 282, 32274–32279.
30 Bartonova V, Igonet S, Sticht J, Glass B, Haberman A,
Vaney MC, Sehr P, Lewis J, Rey FA & Kra
¨
usslich H-
G (2008) Residues in the HIV-1 capsid assembly inhibi-
tor binding site are essential for maintaining the assem-

37 Sundquist WI & Hill CP (2007) How to assemble a cap-
sid. Cell 131, 17–19.
38 Ganser-Pornillos BK, Cheng A & Yeager M (2007)
Structure of a full-length HIV-1 CA: a model for the
mature capsid lattice. Cell 131, 70–79.
39 DeLano WL (2002) The PyMOL Molecular Graphics
System. DeLano Scientific LLC, San Carlos, CA.
J. L. Neira CA small-molecule interactions
FEBS Journal 276 (2009) 6110–6117 ª 2009 The Author Journal compilation ª 2009 FEBS 6117