New Hep C Protease Structure

[Guest guest]

New Hep C Protease Structure

Hep C Vaccine wanted

Over 150 million people are infected with hepatitis C virus, there is

no vaccine, and current therapies are not always effective. More

efficient antivirals are much sought after, so the report of the

crystal structure of the NS2 autoprotease of hepatitis C virus is a

major advance. The structure, which reveals NS2 as a dimeric cysteine

protease, will help in elucidating NS2's role in the viral life

cycle, and it may aid the design of new drugs.

Letter

Nature 442, 831-835(17 August 2006) | doi:10.1038/nature04975;

Received 8 May 2006; Accepted 9 June 2006; Published online 23 July

2006

Structure of the catalytic domain of the hepatitis C virus NS2-3

protease

Ivo C. Lorenz1,2, ph Marcotrigiano1,2, G. Dentzer1 and

M. Rice1

1. Laboratory of Virology and Infectious Disease, Center for the

Study of Hepatitis C, The Rockefeller University, 1230 York Avenue,

New York, New York 10021, USA

2. *These authors contributed equally to this work

Hepatitis C virus is a major global health problem affecting an

estimated 170 million people worldwide1. Chronic infection is common

and can lead to cirrhosis and liver cancer. There is no vaccine

available and current therapies have met with limited success2. The

viral RNA genome encodes a polyprotein that includes two proteases

essential for virus replication3,4. The NS2-3 protease mediates a

single cleavage at the NS2/NS3 junction, whereas the NS3-4A protease

cleaves at four downstream sites in the polyprotein. NS3-4A is

characterized as a serine protease with a chymotrypsin-like fold5,6,

but the enzymatic mechanism of the NS2-3 protease remains

unresolved7,8,9. Here we report the crystal structure of the

catalytic domain of the NS2-3 protease at 2.3 Å resolution. The

structure reveals a dimeric cysteine protease with two composite

active sites. For each active site, the catalytic histidine and

glutamate residues are contributed by one monomer, and the

nucleophilic cysteine by the other. The carboxy-terminal residues

remain coordinated in the two active sites, predicting an inactive

post-cleavage form. Proteolysis through formation of a composite

active site occurs in the context of the viral polyprotein expressed

in mammalian cells. These features offer unexpected insights into

polyprotein processing by hepatitis C virus and new opportunities for

antiviral drug design.

Crystallization of the catalytic domain of the hepatitis C virus

(HCV) NS2-3 protease (NS2pro, consisting of residues 94-217 of NS2;

Fig. 1a) using native and selenomethionine-containing protein yielded

two crystal forms with the same space group (P21, Supplementary Table

1). The asymmetric units of the native and selenomethionine-

containing protein contained twelve and six NS2pro molecules

organized into six and three tightly packed dimers, respectively

(Supplementary Fig. 1).

The NS2pro monomer consists of two subdomains connected by an

extended linker (Fig. 1b and Supplementary Fig. 2a). The amino-

terminal subdomain contains two antiparallel -helices (H1 and H2)

followed by several turns and loops that contact both H1 and H2. The

polypeptide chain continues into an extended region before entering a

four-stranded, antiparallel -sheet in the C-terminal subdomain. The

last -strand continues to the C terminus of NS2. Figure 1c, d

represents views of the NS2pro dimer, which resembles a 'butterfly'

with two-fold symmetry along the vertical axis (Fig. 1c). The N-

terminal subdomain of one molecule interacts with the C-terminal

subdomain of the other molecule and vice versa. The two extended

linkers cross over in the middle of each molecule and each contribute

a -strand (b1) to the antiparallel -sheet in the C-terminal subdomain

of the other molecule. The N termini of the two monomers lie

relatively close to each other, whereas the solvent-exposed C termini

are positioned on opposite sides of the molecule.

Critical residues for NS2-3 proteolytic activity7,8, His 143 and Glu

163, are located in the loop region following helix H2 in the N-

terminal subdomain, whereas another critical residue, Cys 184, lies

at the end of the linker arm in the b1-b2 loop of the C-terminal

subdomain. At the dimer interface, the histidine and glutamate

residues from one monomer are close to the cysteine from the other

chain (Fig. 2a). The arrangement of these three residues is

suggestive of a composite cysteine protease active site (Fig. 2b and

Supplementary Fig. 2b).

The NS2pro structure represents a novel protein fold (DALI server10 Z

scores of less than 3.0). However, superimposing His 143, Glu 163 and

Cys 184 of NS2pro with the active sites from cysteine proteases such

as papain11 (Fig. 2c) and poliovirus 3C protease12 (Fig. 2d)

demonstrated a similar spatial distribution. The orientation of the

catalytic cysteine residues from NS2pro and 3Cpro is similar to the

catalytic serine residues of Sindbis virus capsid13 (Fig. 2e) and the

cellular protease subtilisin14 (Fig. 2f). Thus, like poliovirus

3Cpro, HCV NS2-3 is a cysteine protease with a serine protease active

site geometry15. To the best of our knowledge, NS2pro represents the

first example of a cysteine or serine protease that forms a dimer

containing a pair of composite active sites. However, these features

are reminiscent of retroviral aspartic proteases, which consist of

dimers with a single active site at the dimerization interface16.

Other proteases such as caspases17 require dimerization for activity,

but they do not contain composite active sites.

Interestingly, Pro 164, which is entirely conserved in all HCV

sequences, has a cis-peptide conformation (Fig. 2b and Supplementary

Fig. 2c). Pro 164 may bend the peptide backbone of the catalytic Glu

163 to establish the correct geometry of the glutamate side chain for

catalysis. In addition, the cis-proline may contribute to dimer

stabilization as the linker connecting the two subdomains follows Pro

164.

The backbone carboxylic acid of the C-terminal residue of NS2pro, Leu

217, remains coordinated in the active site via contacts to the side

chains of His 143 and Cys 184, and to the backbone nitrogen of Cys

184 (Fig. 2b). Comparison of the NS2pro structure with protease-

inhibitor complexes indicates that the C terminus of NS2 may exert an

inhibitory function after cleavage8. Sindbis virus capsid protein

mediates a single autoproteolytic cleavage at its C terminus, which

remains bound to the active site inhibiting further catalysis13. The

C-terminal residues of Sindbis virus capsid (Trp 264) and HCV NS2

(Leu 217) are in the same orientation relative to the catalytic triad

(Fig. 2e). Comparing the structure of NS2pro with subtilisin bound to

the inhibitor Eglin-C14 demonstrates that the C terminus of NS2pro

occupies a position equivalent to the inhibitor (Fig. 2f). We propose

a model in which each NS2pro molecule catalyses a single NS2-3

cleavage event, allowing tightly regulated processing. However, we

cannot rule out the possibility that NS2 mediates proteolysis of

other viral or cellular proteins if the C-terminal -strand (b5) is

displaced from the active site.

Processing at the NS2/NS3 junction requires the NS3 protease domain

and is stimulated by the addition of exogenous zinc7,8,18,19.

However, because NS2 contains a complete cysteine-protease active

site with the C terminus positioned for catalysis, the function of

NS3 remains undetermined. The crystal structure of NS2pro represents

the post-cleavage form, which may differ from the NS2-3 precursor.

NS3 may interact with the highly conserved surface surrounding the

NS2 active site (Supplementary Fig. 3), contributing to a functional

catalytic environment and correct positioning of the scissile bond.

The backbone nitrogen of Cys 184 contacts the carboxylic acid of Leu

217 and may serve as part of the oxyanion hole to stabilize the

transition state during catalysis. A residue in uncleaved NS2-3

(either a backbone nitrogen of NS2 oriented differently in the pre-

cleavage form or a residue within the NS3 serine protease domain) may

also contribute to the oxyanion hole. The zinc requirement may be due

to its structural function in NS3 (refs 5, 6) rather than a role in

NS2-3 catalysis. Thus, limiting zinc could indirectly inhibit NS2-3

cleavage by affecting NS3 folding20. Consistent with this idea, the

NS2-3 protease is inhibited by mutations in zinc-coordinating

residues of NS3 (refs 7, 8).

Molecular surface analysis and biochemical data support the NS2pro

dimer model. NS2pro shows a high degree of amino-acid sequence

conservation at the interface between the two monomers (Supplementary

Fig. 3). The cleavage rate of purified NS2-3 is concentration

dependent, indicating that the active form of the protease is

oligomeric19. Analytical ultracentrifugation of NS2pro yielded a

single, monodisperse species with a molecular weight of 39 kDa that

most likely corresponds to a dimer with bound detergent (data not

shown). Moreover, cross-linking of NS2pro in solution with

disuccinimidyl suberate (DSS) led to the identification of a dimeric

species (Supplementary Fig. 4).

A series of experiments in mammalian cells was designed to test

whether NS2pro can form dimers with a functional composite active

site in vivo. HCV full-length polyproteins containing either a H143A

or a C184A mutation in the NS2 active site are defective in NS2-3

processing7,8. However, if a composite active site can form, co-

expression of the two mutant polyproteins should result in partial

NS2-3 cleavage (Fig. 3a). Indeed, when HCV polyproteins with NS2

containing either a H143A or C184A mutation were co-expressed, NS2

and NS3 cleavage products were detected (Fig. 3b), indicating the

formation of a functional composite active site. Moreover, it is

possible to predict from the crystal structure which mutant

polypeptide in the mixing experiment is cleaved, because the C-

terminal Leu 217 and the catalytic Cys 184 originate from the same

chain. Thus, mixing of the two mutants should lead to cleavage of NS2-

3(H143A), whereas NS2-3(C184A) is predicted to remain unprocessed. By

expressing NS2-3 proteins with a Flag or haemagglutinin (HA) tag

fused to the N terminus of NS2, it is possible to distinguish Flag-

NS2 from HA-NS2 by immunoprecipitation using epitope-specific

antibodies and by different electrophoretic mobilities of the various

polypeptides (Fig. 3c). Mixing of Flag-NS2-3(H143A) with HA-NS2-3

(C184A) resulted in cleavage of Flag-NS2-3, whereas HA-NS2-3 remained

unprocessed (Fig. 3c, second lane from the right). When HA-NS2-3

(H143A) was mixed with Flag-NS2-3(C184A), only HA-NS2-3 was cleaved

(Fig. 3c, first lane from the right). Mixing wild-type Flag-NS2-3

with double-mutant HA-NS2-3(H143A/C184A) or vice versa yielded only

cleaved wild-type NS2 (Fig. 3d), whereas the double-mutant

polypeptide remained unprocessed because neither of the composite

active sites is functional when a wild-type and a double-mutant NS2-3

dimerize. Finally, when cells are co-transfected with wild-type Flag-

NS2-3 and HA-NS2-3 and lysed with a mild detergent, Flag-NS2 and HA-

NS2 can be co-precipitated using either an anti-Flag or an anti-HA

antibody (Fig. 3e). These data strongly support the NS2pro crystal

structure and prove that NS2 can form dimers with composite

functional active sites.

Our data change the current view of HCV polyprotein processing and

raise interesting regulatory possibilities. Previously, NS2-3

cleavage was thought to occur as a unimolecular reaction in cis. The

apparent requirement for dimerization to form an active NS2-3

protease suggests that NS2-3 cleavage and therefore formation of the

active RNA replicase may be dependent on the concentration of NS2.

Thus, a requirement for NS2-3 dimerization and subsequent proteolytic

processing could delay the initiation of RNA replication, which may

allow the virus to accumulate sufficient amounts of active NS3-4A

protease to antagonize the induction of type I interferons by the

host21.

The cleaved NS2 dimer and higher-order oligomers may have additional

roles in the viral life cycle, such as a function in membrane-

associated virus assembly22. The solvent-accessible surface of the

NS2pro dimer, coloured according to electrostatic potential (Fig. 4a,

, showed a high proportion of neutral and basic regions. The

surface of the molecule near helices H1 and H2 is mainly hydrophobic,

with basic residues lying underneath. The crystal structure of NS2pro

contained several molecules of n-octyl--glucoside and n-decyl--

maltoside interacting with those two helices (Fig. 4d, e and

Supplementary Fig. 2d). We propose a model in which NS2pro interacts

peripherally with cellular membranes (Fig. 4c, f). The N termini of

two NS2pro monomers would lie close to the membrane (Fig. 4f). In the

full-length protein, dimerization of NS2 may cause the N-terminal

transmembrane domains of two monomers to form a 'bundle' of

transmembrane segments that may serve an important function during

virion morphogenesis.

The NS2pro structure presented here will allow further studies to

elucidate the role of NS2 dimerization and other functions of NS2 in

the viral life cycle. In addition, the structure establishes a

foundation for the design of small-molecule inhibitors directed

against the well-defined active site cleft and other conserved

features of the protein.

Methods

Protein preparation

NS2pro (NS2 residues 94-217), a truncated form of NS2 shown to be

proteolytically active in the context of an NS2-3 precursor that

includes the NS3 protease domain18,19, was expressed in Escherichia

coli. Lysis and subsequent purification of the protein by immobilized-

metal affinity chromatography, ion exchange chromatography and gel

filtration was done in the presence of detergent. The final

concentration on a cation exchange column yielded highly pure protein

at 6-9 mg ml-1.

Crystal growth and freezing

Crystals of NS2pro were grown using hanging-drop vapour diffusion at

4 °C. The well contained a solution of 0.1 M Tris, pH 8.5, 0.8 M

ammonium acetate, 0.25 M lithium chloride and 12% PEG 3350. Crystals

were frozen in well solution supplemented with 5.4 mM decyl maltoside

using stepwise addition of glycerol to a final concentration of 25%.

Data collection and structure determination

Data were collected at beamlines X9A and X29 at the Brookhaven

National Laboratory's National Synchrotron Light Source. Phases were

calculated from selenomethionine-containing protein by

multiwavelength anomalous diffraction (MAD). After data indexing and

scaling with DENZO/SCALEPACK23, 19 of the 24 selenium sites were

found using SnB24. An interpretable electron density map was obtained

using MLPHARE25, followed by density modification and phase

combination using SOLOMON and DM25. Several rounds of iterative model

building and refinement were done using O26 and CNS27. The 2.9 Å

resolution structure was used as a search model to obtain phases for

a native data set at 2.28 Å resolution using molecular replacement.

The final model contained 176 solvent molecules, 12 detergent

molecules, and 12 molecules of NS2pro. A summary of the refinement

statistics is shown in Supplementary Table 1.

Expression in mammalian cells

NS2 containing wild-type or mutant active site residues was expressed

in mammalian cells in the context of a full-length viral polyprotein

or as NS2-3 precursor with an N-terminal Flag or HA tag. Cells

metabolically labelled with 35S were lysed and immunoprecipitation of

NS2 performed using antibodies against NS2 or the Flag and HA tags.

Proteins were separated by SDS-polyacrylamide gel electrophoresis and

visualized by autoradiography.