Selection of suitable detergents for obtaining an active dengue protease in its natural form

from E. coli

Lynette Sin Yee Liew1, Micelle Yueqi Lee1, Ying Lei wong1, Jinting Cheng2, Qingxin Li3* and CongBao

Kang1,*

1Experimental Therapeutics Centre, Agency for Science, Technology and Research (A*STAR)

31 Biopolis Way, Nanos, #03-01, Singapore, 138669

2Institute of Materials Research and Engineering (IMRE), 3 Research Link, 117602, Singapore

3Institute of Chemical & Engineering Sciences, Agency for Science, Technology and Research

(A*STAR)

1 Pesek Road, Jurong Island, Singapore, Singapore 627833

* Corresponding to CongBao Kang: cbkang@etc.a-star.edu.sg Qingxin Li: li_qingxin@ices.a-

star.edu.sg

1

Abstract

Dengue protease is a two-component enzyme and is an important drug target against dengue virus.

The protease activity and protein stability of dengue nonstructural protein 3 (NS3) require a co-

factor region of a four-span membrane protein NS2B. A natural form of dengue protease containing

full-length NS2B and NS3 protease domain NS2BFL-NS3pro is useful for dengue drug discovery. In

current study, detergents that can be used for protease purification were tested. Using a water

soluble protease construct, 39 detergents were selected for NS2B and NS2BFL-NS3pro purification.

The results show that 18 detergents were able to sustain the activity of the natural dengue protease

and 11 detergents could be used for NS2B purification. The results obtained in this study will be

useful for biochemical and biophysical studies on dengue protease.

Key words: dengue virus; dengue protease, membrane protein, detergent micelles

2

Introduction

Dengue virus (DENV) together with other Flaviviridae family members such as West Nile virus (WNV)

is an important human pathogen. DENV affects people all over the world especially in tropical and

sub-tropical regions. It is estimated that there are approximately 390 million human infections

annually and 96 million cases are with manifest symptoms [1]. DENV infection can cause dengue

fever (DF) and some patients can develop serious diseases such as dengue hemorrhagic fever (DHF)

or dengue shock syndrome (DSS) [2].

The genome of DENV is a single-strand, positive-sense RNA and encodes a polyprotein that can be

further processed into three structural proteins (capsid, premembrane, and envelope) and seven

non-structural (NS) proteins (NS1, NS2A, NS2B, NS3, NS4A, NS4B, and NS5) by host and viral

proteases [3]. The function of the structural proteins is to form the viral particles. The NS proteins

have diverse functions that are necessary for viral replication [3-5]. Host signal peptidase and a viral-

encoded protease (NS3) are important for releasing NS proteins by cleaving the junctions between

different NS proteins. NS3 is a multi-functional protein containing a N-terminal protease domain

(NS3pro) formed by approximately 180 amino acids and C-terminal domain harbouring RNA helicase,

and nucleotide triphosphatase (NTP) activities [6-8]. NS3 protease activity requires a co-factor region

from NS2B that is a membrane protein containing four transmembrane segments [9]. Due to the

importance of NS3 protease activity in viral maturation, dengue protease was a validated drug target

[10]. High-throughput drug screening and structure-based drug design have been used to develop

NS3 protease inhibitors, but there is no potent inhibitor available [11].

Structural studies revealed that that the active pocket of protease is not druggable because of the

charges in the active site [11-13]. Developing allosteric inhibitors or inhibitors breaking NS2B and

NS3 interaction may be a feasible way to have a potent protease inhibitor [14]. The conventional

3

construct is water soluble and contains a cofactor region from NS2B covalently linked with NS3pro

through a Gly4-Ser-Gly4 linker (linked protease) [9]. Further studies showed that removal of this

artificial linker (un-linked protease) can improve the dynamic nature of the protease [15]. It will be

helpful for protease inhibitor development when a natural protease is available for biochemical and

biophysical studies. As NS2B is a four-span membrane protein [16], a natural form of dengue

protease should contain at least full length NS2B (NS2BFL) and NS3 pro (NS2BFL-NS3pro).

Accumulated studies have demonstrated that this natural dengue protease could be obtained from

E. coli [17, 18]. Although different membrane-mimicking systems such as lipid bilayer, bicelles and

nanodisk are efficient systems for membrane protein folding [19-22], a detergent system that does

not inhibit protease activity will be helpful for biochemical and biophysical characterization of this

natural protease because it can be used to extract protein from the E. coli membrane before the

protein is reconstituted into different membrane systems or study NS2BFL and NS3pro interaction in

vitro. We have shown that lyso-myristoyl phosphatidylcholine (LMPC) can sustain dengue protease

activity [17]. We also expressed and purified NS2B into lyso-myristoyl phosphatidylglycerol (LMPG)

micelles for structural studies [23]. The protease activity was still low when the natural protease was

purified in LMPC or LMPG micelles [17]. Our previous detergent screening was conducted using the

following steps. The efficiency of detergents on extraction of dengue protease from E. coli

membrane was tested followed with measurements of protease activity of purified protease in

screened detergents. It is not surprising that screened detergents may not be suitable for enzymatic

activity because the screening was based on the efficiency of different detergents on extracting the

target protein from E. coli membrane. We then used a modified approach in detergent screening in

current study. The effect of detergents on the enzymatic activity of a water soluble protease was

first evaluated. Only detergents having no obviously inhibitory effect on protease activity were

selected for further protease purification. This method will allow us to identify suitable detergents

that can sustain dengue protease activity and can also be used to study the interaction between the

NS2BFL and NS3pro in vitro.

4

In this study, we use a water soluble protease to screen a detergent library containing over 80

detergents to identify candidates that do not inhibit protease activity. There are 39 detergents that

showed no inhibitory effect on water soluble protease. 11 detergents are shown to be suitable for

structural and functional studies of NS2B and 18 detergents are able to purify natural dengue

protease NS2BFL-NS3pro.

Materials and Methods

Expression and purification of un-linked protease

The un-linked protease containing the co-factor region of NS2B and protease domain of NS3

(NS3pro) was expressed and purified as previously described [15]. Briefly, plasmids harbouring NS2B

cofactor region and NS3pro were transformed into E. coli BL21 (DE3). Protease was induced at 18 C

overnight by adding isopropyl β-D-thiogalactopyranoside (IPTG) to 0.5 mM concentration when the

A600 reached 0.6. The cells were suspended in a buffer that contained 20 mM Tris-HCl, pH7.8, 300

mM NaCl and 2 mM β-mercaptoethanol. The un-linked protease was purified using a Ni-NTA column

and a size exclusion chromatography. The protease was buffer exchanged to a storage buffer that

contained 20 mM Tris-HCl, pH7.8, 50 mM NaCl and 1 mM DTT.

Protease activity using Bz-nKKR-AMC as a substrate

Protease activity assays were performed in 96-well plates. The enzyme activity was measured in an

assay buffer that contained 50 mM Tris-HCl, pH 8.0, 20% glycerol in a final volume of 100 µl as

previously described [24]. The protease specific, fluorophore-tagged substrate Bz-nKKR-AMC was

used in the assay. The substrate concentration in the protease buffer was 60 μM and protease

concentration was 20 nM. Detergent was added to the mixture to 1% final concentration. Substrate

cleavage was monitored after addition of protease at 37 C. The increase in fluorescence (excitation

380 nm, emission 450 nm) was continuously monitored on a Tecan Safire 2 microplate reader. The

reaction mixture containing no detergent was used as a control.

Expression of dengue 4 NS2B and the natural protease NS2BFL-NS3pro

The NS2B of DENV 4 was expressed in E. coli at 18 C for 18 h [16]. For the NS2BFL-NS3pro of DENV 4

(Fig. 1), the construct contained full length NS2B and N-terminal protease domain of NS3. The amino

acid sequence is shown in Fig. 1. The cDNA of NS2B-NS3pro was synthesized by Genscript and cloned

5

into NdeI and XhoI sites of pET29b. The resulting plasmid encodes a protein containing NS2BFL,

NS3pro fused with a C-terminal tag (LEHHHHHH) for purification. Plasmid was transformed into E.

coli (BL21DE3) cells. Protein was induced for 18 h by adding IPTG to 0.5 mM final concentration

when A600 of the cell culture reached 0.8. Cells were harvested by centrifugation at 10,000 ×g. Cells

were resuspended in a buffer containing 20 mM Tris-HCl, pH7.8, 300 mM NaCl and 2 mM β-

mercaptoethanol. Cells were broken in an ice bath by sonication. Sonication was performed using a

probe sonicator 4000 package (Misonix) at 5 seconds on, 5 seconds off, 40% power to break the

cells. The cell debris or inclusion bodies were removed by centrifugation at 3000 ×g, 4 C for 10 min.

The resulting supernatant was used for further detergent screening. The NS2B-NS3pro S135A

mutation was made by site-directed mutagenesis. The mutant protease was induced using the same

protocol as that of the wild type protease.

Effect of detergents on extracting NS2B and NS2BFL-NS3pro from E. coli membrane

The E. coli cell lysate containing NS2B or NS2BFL-NS3pro was mixed with different type of detergents

(2%) at 4 C for 1 h. The mixture was cleared by centrifugation at 20, 000 ×g, 4 C for 20 min and the

supernatant was loaded in a SDS-PAGE gel for analysis. To test effect of different detergent on

protein purification, 1 ml of cell lysate was mixed with different detergent (2% final concentration),

followed with centrifugation at 20, 000 ×g, 4 C for 20 min. The resulting supernatant was mixed with

50 l of Ni2+-NTA resin. The resin was washed with a washing buffer that contained 20 mM Tris-HCl,

pH7.8, 1% detergent, 20 mM imidazole, and 2 mM β-mercaptoethanol. Protein was eluted with 120

l of elution buffer containing 500 mM imidazole, pH6.5, 1% detergent and 2 mM β-

mercaptoethanol. Samples (6 l and 4 l for the purified sample) were separated SDS-PAGE followed

by Western blot using an anti-his antibody (Qiagen) and second antibody (Genscript). Histidine-tag

containing protein was visualized using Bio Rad VersaDoc MP 4000 Molecular Digital Imaging

System.

Purification of NS2BFL-NS3pro using screened detergents

6

Large-scale protein purification was conducted using a step similar to aforementioned one. E. coli

cells from 1 L M9 medium were collected and resuspended in a buffer containing 20 mM Tris-HCl,

pH7.8, 300 mM NaCl and 2 mM β-mercaptoethanol. Cells were broken by sonication and the cell

debris or inclusion bodies were removed by centrifugation at 3,000 ×g, 4 C for 10 min. The cell

membrane was obtained by ultra-centrifugation at 45, 000 rpm in a 70 Ti rotor for 2 h. The pellet

was resuspended in a buffer containing 20 mM Tris-HCl, pH7.8, 300 mM NaCl, 2 mM β-

mercaptoethanol and 1% detergent. The solution was further cleared by centrifugation at 20, 000 ×g

for 20 min. Protein was purified as previously described [23].

Results

Effect of detergents on water soluble dengue protease

The natural form of dengue protease construct (NS2BFL-NS3pro) contains an NS2BFL-a four-span

transmembrane protein and the NS3pro that is a water soluble protein (Fig. 1). To sustain its

function in vitro, a membrane system that does not affect NS3pro activity is needed to support NS2B

folding. Detergent micelles were used for extracting membrane proteins from cell membrane and

mimicking the membrane environment. Our previous study showed that this natural dengue

protease had detectable enzymatic activities in LMPC micelles, but the activity is still low and this

detergent may not be suitable for structural studies using X-ray crystallography and the NMR

spectrum of NS2BFL in LMPC also exhibited poor signals [17]. To further explore suitable detergent

micelles that can be used in dengue protease extraction and folding, we first tested the effect of

different detergents on the protease activity of a water soluble protease using Bz-nKKR-AMC as a

substrate [15](Fig. 1). We used the same detergent library as the one used in the previous study [17].

This library contains over 80 detergents that were commonly used in structural and functional

studies of membrane proteins. Different detergent has different critical micelle concentration (CMC)

7

that is the concentration to form micelles. The concentration of detergent used in the assay buffer

was 1% which is above the CMCs of most detergents. Most of the detergents showed no obvious

inhibitory effect on the protease activity (Fig. 2), mirroring the result of WNV protease [25]. Thirty

nine detergents that can sustain more than 85% of the protease activity were selected for further

studies.

Effect of detergent on NS2BFL and NS2BFL-NS3pro purification

Two factors need to be considered whether a detergent is suitable for a membrane protein

preparation. First is that the detergent can be used to extract a target protein from the E. coli

membrane. Second is that the detergent can be used for protein purification without reducing the

target protein yield. The effect of selected 39 detergents on extraction and purification of NS2BFL

and NS2BFL-NS3pro from DENV4 was tested. It has been noted that the detergent concentration in

extraction and purification buffers will affect the final results. It is time consuming to explore such

conditions. We then used 2% detergent when we extracted proteins and 1% detergent in the

washing step. Most of the selected detergents were able to exact both NS2BFL and NS2BFL-NS3pro

from the E. coli cell membrane (Figs. 3, 4). The extraction efficiency differed for different detergents.

Few detergents showed very low efficiency in extracting proteins (Figs. 3, 4). Among these 39

detergents, 17 of them can be used in NS2BFL purification, evidenced by the observation of an

intense purified NS2B band (Fig. 3, Table 1). The NS2BFL-NS3pro protease contains a protease

cleavage site between NS2BFL and NS3pro (Fig. 1). The activity of NS2BFL-NS3pro could be observed

by monitoring appearing of a 25kDa band in Western blot result using an anti-his antibody due to

production of NS3pro (a HHHHHH sequence at its C-terminus) by self-cleavage of NS2BFL-NS3pro

(Fig. 4). When the protease is active in the detergent, a band from NS3pro will be observed in the

western blot. The NS2BFL-NS3 is active in most detergents except for C1 (CYMAL-2). As C1 did not

affect the activity of the water soluble dengue protease, NS2BFL-NS3pro loss its activity in C1 might

be due to mis-folding of NS2BFL. Among these tested detergents, 18 of them were shown to be

8

suitable for NS2BFL-NS3pro purification based on two reasons. First is that the intensity of protein

bands corresponding to NS2BFL-NS3pro/NS3pro are thicker than that of NS2BFL-NS3pro before

purification. Second is that the intensity of NS3pro band is equal to or more intense than that of

NS2BFL-NS3pro after purification (Fig. 4, Table 1). The protease was most active in n-Dodecyl-N,N-

dimethylamine-N-oxide (LDAO), n-Octyl-β-D-glucopyranoside (OG), and n-Decyl-β-D-

maltopyranoside (DM) detergents (Fig. 4). It has been noted that the selected detergents have

different effects on NS2BFL purification from NS2BFL-NS3pro (Table 1). Based on our results, 11

overlapped detergents were suggested for exaction and purification of NS2BFL if purified NS2BFL is

used for structural and NS3 binding studies (Table 1). When 10 detergents excluding C8E4 and C12E9

were used to extract and purify a NS2BFL-NS3pro S135A mutant-a mutation in the protease active

site causing reduction of the protease enzymatic activity, the NS3pro band was reduced significantly

(Fig. 5), confirming that the NS3pro band observed (Fig. 4) was arisen from self-cleavage of the

natural protease NS2BFL-NS3pro.

Large scale protein purification was carried out using selected detergents. All these detergents

showed lower purification efficiency (data not shown) than LMPG-a detergent that can be used to

purify NS2B and NS2BFL-NS3pro in large quantity and high purity [23](Fig. 6). Among these selected

detergents, only DM was able to purify the target protein (Fig. 6), but several further purification

steps such as ion change chromatography is needed to obtain a much pure protein. We also tried to

purify NS2BFL-NS3pro using combination of LMPG and DM micelles, namely LMPG was used in

extraction and washing steps, DM was used to elute protein from the resin. A much purer protein

than ones using only DM in the purification was obtained (Fig. 6). Lacking of the NS2B and NS3pro

bands still suggest that the enzymatic activity was low under such conditions.

Discussion

9

Dengue protease is a validated target due to its importance in viral maturation. Water soluble

protease containing partial NS2B and NS3pro has been well characterized. Although water soluble

protease is active in vitro, its charged active site makes drug discovery targeting the protease active

site challenging [11]. The conventional linked protease was shown to have open and closed

conformations in vitro [12, 26, 27]. Interestingly, removal of the artificial linker improved protease

stability, which makes protease exist in mainly a closed conformation [15, 26]. Studies have shown

that it is possible to develop allosteric dengue protease inhibitors [28], which implies that a natural

protease will be useful for drug discovery.

Studies have shown that a natural protease containing full-length NS2B and NS3 protease could be

expressed and purified from E. coli [17, 18]. A suitable membrane system is required for the folding

of NS2BFL-NS3pro, which will be helpful for protease characterization. Our study showed that LMPC

micelles can sustain dengue protease activity [17]. LMPC is a detergent suitable for structural studies

of membrane proteins using solution NMR spectroscopy, but it may not be suitable for structural

study of membrane proteins using X-ray crystallography. Further detergent screening is still helpful

for dengue protease inhibitor development and biochemical characterization for the protease. In

this study, we showed that most detergents from a detergent library did not inhibit water soluble

protease activity using Bz-nKKR-AMC as a substrate (Fig. 2), which is similar to WNV protease [25].

When the selected 39 detergents were used for NS2BFL-NS3pro purification, 18 of them were

observed to be suitable for NS2BFL-NS3pro purification because of the higher protease activity (Fig.

3B). The NS3 protease activity was decreased in the presence of other detergents, which might be

due to the folding of NS2B or the interaction between substrate and the active site was blocked by

the detergents. Compared with our previous study [17], the current study is focus on screening

detergents that can sustain NS2BFL-NS3pro protease activity. The yield of purified NS2BFL-NS3pro

using newly screened detergents was much lower than the one using LMPC or LMPG. We found that

approximately 0.3 mg of NS2BFL-NS3pro could be purified from 1 L M9 culture when DM was used

in protein extraction purification. Although the protein was obtained with low yield, protease activity

10

was sustained (Fig. 6). Higher protein yield could be obtained by growing large quantity of E. coli

cells or extracting protein from E. coli membrane. Nevertheless, we clearly showed that these 18

detergents could sustain protease activity during protein extraction and purification steps (Fig. 3 and

Fig. 4).

Structural study of NS2BFL-NS3pro might be challenging. It is feasible to study the free form of full

length NS2B and to understand its interaction with NS3pro in vitro. Our recent study has confirmed

the membrane topology of NS2B in LMPG micelles [16, 17], but NS2BFL-NS3pro is not active in LMPG

micelles [17]. Lacking protease activity in LMPG micelles may not arise from the mis-folding of NS2B

or NS3 protease because our structural study indicated NS2B is folded in solution [16] and the water

soluble protease also sustain activities in LMPG micelles (Fig. 2). One of the possibilities might be

that LMPG micelles could affect protease and subtract interactions. Therefore, our detergent

screening for NS2B purification is useful for NS2B and NS3pro interactions studies in vitro. There

were 17 detergents were shown to be good for NS2B purification (Fig. 3A), and 11 of them were able

to keep NS2BL-NS3pro active. Therefore, these 11 detergents will be suitable for structural study of

full length NS2B using different biophysical methods such as X-ray and NMR. They are also suitable

for conducting in vitro studies to understand the molecular interaction between NS2B and NS3pro.

As NS2BFL-NS3pro is formed by a NS3pro region that is water soluble and NS2BFL that is a four-span

membrane protein playing an essential role in regulating NS3pro activity, these 18 detergents that

are suitable for NS2BFL-NS3pro purification might also be applicable to other multi-span membrane

protein purifications because they are able to sustain the correct folding of both NS2BFL and NS3pro

(Fig. 2, Fig. 3).

In summary, using a water soluble dengue protease construct, 39 detergents were shown to be able

to sustain over 85% protease activities. Among these selected detergents, 18 detergents were

shown to be able to sustain the activity of the natural from of dengue protease containing full length

NS2B and NS3pro, and 10 detergents were suitable for purifying dengue NS2B for structural studies.

11

Our results will be useful for structural and functional studies of the dengue protease, which is

helpful for developing protease inhibitors.

ACKNOWLEDGEMENTS

The authors would like to thank the financial support from A*STAR JCO grants (1331A028). Q.L. is

supported by the Biospecialties project (1526004161). We also appreciate the technical support

from Dr. Jeffrey Hill, Dr. Joma Joy, Qiwei Huang, and Ying Lei Wong from ETC, A*STAR.

Figure Legends

Figure 1. Construct used in this study. A. Diagram of the NS2BFL-NS3pro in micelles. Full length NS2B

is shown in green and NS3pro is shown in brown. The cofactor region is highlighted in blue. B. The

structure of dengue protease containing NS2B co-factor region, NS3pro and an inhibitor. The protein

12

structure of dengue protease (PDB id 3U1I) was shown. Co-factor region, NS3pro and Bz-NKKR-H

inhibitor are shown in blue, brown and yellow, respectively. C. The sequence of NS2BFL-NS3pro used

in this study. The NS2B is shown in back and co-factor region is shown blue. NS3pro is shown brown.

The protease cleavage site is highlighted with an arrow.

Figure 2 Effect of detergent on protease activity. The un-linked dengue protease containing co-factor

region of NS2B and NS3 protease domain was used in the assay. The assay mixture contained no

detergent was used as control. The relative activity of each detergent to the control sample was

plotted.

Figure 3. Extraction and purification of NS2B using different detergents. Effect of different detergents

on NS2B extraction and purification are shown. Protein was first extracted using different detergents

and followed by analysis using Western blot. The extracted fraction was further purified in the

presence of different detergents. M.W is molecular weight. The labels on the top of the gel are

representing different detergents that are listed in Table 1. The Western-blot was conducted using

an anti-his antibody. Arrow indicates NS2B protein band. B. Effect of different detergents on

NS2BFL-NS3pro. The labels on the top of the gel are same as those in A and shown in Table 1.

Figure 4. Extraction and purification of NS2BFL-NS3pro using different detergents. Effect of different

detergents on NS2BFL-NS3pro extraction and purification are shown. The label is same as that of Fig.

3. The Western-blot was conducted using an anti-his antibody. Arrow indicates NS2BFL-NS3pro and

NS3pro bands because the histidine tag is that the C-terminus of NS3pro. Detergents that suitable

for NS2BFL-NS3pro purification are labelled with “*”.

Figure 5. Effect of S135A mutation on the protease activity. Extraction and purification of S135

mutants using the screened detergent was conducted using the same protocol as the wild type. The

NS2BFL-NS3pro with a S135A mutation was prepared as the wild type and purified using the 10

13

detergents. The reduction of the NS3 band suggested that the protease activity was reduced after

mutation. The band from E. coli protein is indicated using an asterisk. Lane 1 to 10 indicate different

detergents used in the study. Lanes 11 is control sample (C) in which 8 M urea were used in all the

preparation buffers instead of detergent.

Figure 6. Purification of NS2BFL-NS3pro using LMPG and DM micelles. A. Extraction and purification

of NS2BFL-NS3pro using LMPG. B. Purification of NS2BFL-NS3pro using DM detergent. C. Purification

of NS2BFL-NS3pro using combination of LMPG and DM. LMPG was used to extract protein from the

cell membrane. In the washing step, the resin with protein was first washed with a washing buffer

containing 0.1% LMPG. The resin was then washed with a washing buffer containing 0.1% DM before

eluted with an elution buffer containing 0.1% DM and 300 mM imidazole at a pH of 6.5.

14

Table 1 Detergents used for purification of NS2B and NS2BFL-NS3pro

Column 1 is label that relates to the labels in Figure 3. Detergents list in this table are showing no

significant inhibitory effect on the protease activity of the un-linked protease. Detergents that can be

used to purify NS2B and NS2BFL-NS3pro are highlighted with an X. The detergents that can be used

for NS2B purification without affecting protease activity are highlighted with an asterisk in the first

column.

15

References

[1] S. Bhatt, P.W. Gething, O.J. Brady, J.P. Messina, A.W. Farlow, C.L. Moyes, J.M. Drake, J.S. Brownstein, A.G. Hoen, O. Sankoh, M.F. Myers, D.B. George, T. Jaenisch, G.R. Wint, C.P. Simmons, T.W. Scott, J.J. Farrar, S.I. Hay, The global distribution and burden of dengue. Nature 496 (2013) 504-507.[2] H. Lindback, J. Lindback, A. Tegnell, R. Janzon, S. Vene, K. Ekdahl, Dengue fever in travelers to the tropics, 1998 and 1999. Emerg Infect Dis 9 (2003) 438-442.[3] B.D. Lindenbach, H.-J. Thiel, C.M. Rice, Flaviviridae: The Virus and Their Replication, p. 1101-1152. In D.M. Knipe and P.M. Howley (ed), Fields virology, 5th., vol. 1. Lippincott William & Wilkins, Philadelphia, Pa. (2007).[4] J.L. Munoz-Jordan, G.G. Sanchez-Burgos, M. Laurent-Rolle, A. Garcia-Sastre, Inhibition of interferon signaling by dengue virus. Proc. Natl. Acad. Sci. USA 100 (2003) 14333-14338.[5] J. Morrison, S. Aguirre, A. Fernandez-Sesma, Innate immunity evasion by Dengue virus. Viruses 4 (2012) 397-413.[6] P. Warrener, J.K. Tamura, M.S. Collett, RNA-stimulated NTPase activity associated with yellow fever virus NS3 protein expressed in bacteria. Journal of Virology 67 (1993) 989-996.[7] G. Wengler, G. Czaya, P.M. Farber, J.H. Hegemann, In vitro synthesis of West Nile virus proteins indicates that the amino-terminal segment of the NS3 protein contains the active centre of the protease which cleaves the viral polyprotein after multiple basic amino acids. J Gen Virol 72 ( Pt 4) (1991) 851-858.[8] H. Li, S. Clum, S. You, K.E. Ebner, R. Padmanabhan, The serine protease and RNA-stimulated nucleoside triphosphatase and RNA helicase functional domains of dengue virus type 2 NS3 converge within a region of 20 amino acids. J. Virol. 73 (1999) 3108-3116.[9] D. Leung, K. Schroder, H. White, N.X. Fang, M.J. Stoermer, G. Abbenante, J.L. Martin, P.R. Young, D.P. Fairlie, Activity of recombinant dengue 2 virus NS3 protease in the presence of a truncated NS2B co-factor, small peptide substrates, and inhibitors. J Biol Chem 276 (2001) 45762-45771.[10] N. Mass, B. Selisko, H. Malet, F. Peyrane, C. Debarnot, E. Decroly, D. Benarroch, M.P. Egloff, J.C. Guillernot, K. Alvarez, B. Canard, Dengue virus: viral targets and antiviral drugs. Virologie (Montrouge) 11 (2007) 121-133.[11] A. Poulsen, C. Kang, T.H. Keller, Drug design for flavivirus proteases: what are we missing? Curr Pharm Des 20 (2014) 3422-3427.[12] C.G. Noble, C.C. Seh, A.T. Chao, P.Y. Shi, Ligand-bound structures of the dengue virus protease reveal the active conformation. J Virol 86 (2012) 438-446.[13] D. Luo, T. Xu, C. Hunke, G. Gruber, S.G. Vasudevan, J. Lescar, Crystal structure of the NS3 protease-helicase from dengue virus. J Virol 82 (2008) 173-183.[14] M. Yildiz, S. Ghosh, J.A. Bell, W. Sherman, J.A. Hardy, Allosteric inhibition of the NS2B-NS3 protease from dengue virus. ACS Chem Biol 8 (2013) 2744-2752.[15] Y.M. Kim, S. Gayen, C. Kang, J. Joy, Q. Huang, A.S. Chen, J.L. Wee, M.J. Ang, H.A. Lim, A.W. Hung, R. Li, C.G. Noble, T. Lee le, A. Yip, Q.Y. Wang, C.S. Chia, J. Hill, P.Y. Shi, T.H. Keller, NMR analysis of a novel enzymatically active unlinked dengue NS2B-NS3 protease complex. J Biol Chem 288 (2013) 12891-12900.[16] Y. Li, Q. Li, Y.L. Wong, L.S. Liew, C. Kang, Membrane topology of NS2B of dengue virus revealed by NMR spectroscopy. Biochim Biophys Acta 1848 (2015) 2244-2252.[17] Q. Huang, Q. Li, J. Joy, A.S. Chen, D. Ruiz-Carrillo, J. Hill, J. Lescar, C. Kang, Lyso-myristoyl phosphatidylcholine micelles sustain the activity of Dengue non-structural (NS) protein 3 protease domain fused with the full-length NS2B. Protein Expr Purif 92 (2013) 156-162.

16

[18] O. Choksupmanee, K. Hodge, G. Katzenmeier, S. Chimnaronk, Structural platform for the autolytic activity of an intact NS2B-NS3 protease complex from dengue virus. Biochemistry 51 (2012) 2840-2851.[19] A. Diller, C. Loudet, F. Aussenac, G. Raffard, S. Fournier, M. Laguerre, A. Grelard, S.J. Opella, F.M. Marassi, E.J. Dufourc, Bicelles: A natural 'molecular goniometer' for structural, dynamical and topological studies of molecules in membranes. Biochimie 91 (2009) 744-751.[20] A. Das, J. Zhao, G.C. Schatz, S.G. Sligar, R.P. Van Duyne, Screening of type I and II drug binding to human cytochrome P450-3A4 in nanodiscs by localized surface plasmon resonance spectroscopy. Anal Chem 81 (2009) 3754-3759.[21] M. Etzkorn, T. Raschle, F. Hagn, V. Gelev, A.J. Rice, T. Walz, G. Wagner, Cell-free Expressed Bacteriorhodopsin in Different Soluble Membrane Mimetics: Biophysical Properties and NMR Accessibility. Structure (2013).[22] C. Kang, Q. Li, Solution NMR study of integral membrane proteins. Curr Opin Chem Biol 15 (2011) 560-569.[23] Q. Huang, A.S. Chen, Q. Li, C. Kang, Expression, purification, and initial structural characterization of nonstructural protein 2B, an integral membrane protein of Dengue-2 virus, in detergent micelles. Protein Expr Purif 80 (2011) 169-175.[24] J. Li, S.P. Lim, D. Beer, V. Patel, D. Wen, C. Tumanut, D.C. Tully, J.A. Williams, J. Jiricek, J.P. Priestle, J.L. Harris, S.G. Vasudevan, Functional profiling of recombinant NS3 proteases from all four serotypes of dengue virus using tetrapeptide and octapeptide substrate libraries. J Biol Chem 280 (2005) 28766-28774.[25] Q. Huang, Q. Li, A.S. Chen, C. Kang, West Nile virus protease activity in detergent solutions and application for affinity tag removal. Anal Biochem 435 (2013) 44-46.[26] L. de la Cruz, W.N. Chen, B. Graham, G. Otting, Binding mode of the activity-modulating C-terminal segment of NS2B to NS3 in the dengue virus NS2B-NS3 protease. FEBS J 281 (2014) 1517-1533.[27] L. de la Cruz, T.H. Nguyen, K. Ozawa, J. Shin, B. Graham, T. Huber, G. Otting, Binding of low molecular weight inhibitors promotes large conformational changes in the dengue virus NS2B-NS3 protease: fold analysis by pseudocontact shifts. J Am Chem Soc 133 (2011) 19205-19215.[28] M. Yildiz, S. Ghosh, J.A. Bell, W. Sherman, J.A. Hardy, Allosteric Inhibition of the NS2B-NS3 Protease from Dengue Virus. ACS Chem Biol (2013).

Table 1 Detergents showed high efficiency in NS2B and NS2BFL-NS3pro purification

No Name NS2B NS2BFL-NS3proF12 n-Octyl-β-D-maltopyranoside  X

G12* Tetraethylene glycol monooctyl ether (C8E4)  X XA8* ANAPOE®-C12E9  X X

17

B10 CHAPSA12 ANAPOE®-X-100  XB4 ANZERGENT® 3-8

D10* n-Dodecyl-N,N-dimethylamine-N-oxide  X  XB11 CHAPSOG11 n-Tridecyl-β-D-maltopyranoside  XB8 Big CHAPF7* Hexaethylene glycol monooctyl ether  X  XA2* ANAPOE®-C12E10  X  XH7* dihexanoylphosphatidylcholine  X  XA6 ANAPOE®-C10E9B7 ANZERGENT® 3-14  XB5 ANZERGENT® 3-10D6 n-Decyl-β-D-thiomaltopyranosideD7 n-Dodecyl-β-D-thiomaltopyranoside

G10 n-Tetradecyl-β-D-maltopyranoside  XC1 CYMAL®-2D2 n-Dodecyl−β−D-Maltopyranoside  X

D12 2,6-Dimethyl-4-heptyl-b-D-maltopyranosideA7 ANAPOE®-C12E8  XC6* CYMAL®-6  X  XA3 ANAPOE®-35  XB1 ANAPOE®-X-114F5 n-Hexyl-β-D-glucopyranosideD5 n-Decyl-a-D-maltopyranoside  XC7* CYMAL®-7  X  XG5 Pentaethylene glycol monodecyl ether (C10E5)

F11* n-Nonyl-β-D-thiomaltopyranoside  X  XH4* n-Undecyl-β-D-thiomaltopyranoside  X  XD4* n-Decyl-β-D-maltopyranoside  X  XB3 ANAPOE®-X-405  XE12 FOS-CHOLINE®-ISO-9B12 CYMAL®-1G1 n-Octyl-β-D-glucopyranoside  XD3 n-Dodecyl−α−D-Maltopyranoside  XG9 Sucrose monododecanoate  XF6 n-Hexyl-β-D-maltopyranoside

18

Figure 1

19

Figure 2

20

Figure 3

Figure 4

21

22

Figure 5

23

Figure 6

24