- Email: [email protected]
Structural characterization reveals that viperin is a radical S-adenosyl-l -methionine (SAM) enzyme
Biochemical and Biophysical Research Communications 391 (2010) 1390–1395
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Structural characterization reveals that viperin is a radical S-adenosyl-L-methionine (SAM) enzyme Goyal Shaveta a, Jiahai Shi b, Vincent T.K. Chow c, Jianxing Song a,b,* a
Department of Biochemistry, Yong Loo Lin School of Medicine, 10 Kent Ridge Crescent, Singapore 119260, Singapore Department of Biological Sciences, Faculty of Science, 10 Kent Ridge Crescent, Singapore 119260, Singapore c Department of Microbiology, Yong Loo Lin School of Medicine, National University of Singapore, 5 Science Drive 2, Singapore 117597, Singapore b
a r t i c l e
i n f o
Article history: Received 8 December 2009 Available online 22 December 2009 Keywords: Viperin Interferon Radical S-adenosyl-L-methionine (SAM) enzyme Iron–sulfur cluster Circular dichroism spectroscopy NMR spectroscopy
a b s t r a c t Viperin is an interferon-inducible protein inhibiting many DNA and RNA viruses. It contains an N-terminal transmembrane helix, a highly conserved C-terminus and a middle region carrying a CX3CX2C motif, characteristic of radical S-adenosyl-L-methionine (SAM) enzymes. So far no structural characterization has been reported and reconstitution of the [4Fe–4S] cluster in viperin all failed. Here, by dissecting the 361-residue human viperin into 12 fragments, followed by extensive CD and NMR characterization, Viperin (45–361) was identified to be soluble and structured in buffers. Most importantly, we have successfully reconstituted the [4Fe–4S] cluster in Viperin (45–361), thus providing the first experimental evidence confirming that viperin is indeed a radical SAM enzyme. Furthermore, the C-terminus Viperin (214–361) which is insoluble in buffers but again can be solubilized in salt-free water appears to be only partially folded. Our results thus imply that the radical SAM enzyme activity may play a key role in the broad antiviral actions of viperin. Ó 2009 Elsevier Inc. All rights reserved.
Introduction Viperin (virus inhibitory protein, endoplasmic reticulum-associated, interferon-inducible) was originally cloned from interferon treated human macrophages [1,2]. Subsequent studies reveal that viperin is a highly inducible candidate gene in response to a wide spectrum of viruses and microbial products such as LPS and double-stranded RNA [3,4]. Viperin is highly conserved across both mammals and lower vertebrates, localized in the endoplasmic reticulum (ER), thus implying its extremely important role. Indeed, viperin has been demonstrated to inhibit a large array of DNA and RNA viruses, such as human cytomegalovirus [1,2], influenza, hepatitis C virus (HCV) and alphaviruses [5–7], as well as human immunodeficiency virus [8]. However, it remains largely elusive how viperin is capable of inhibiting such a diverse spectrum of viruses since it is very unlikely that it functions by directly binding to viral proteins. The human viperin gene encodes a protein of 361 amino acids, and based on the sequence alignment [9], it appears composed of three distinctive regions: an N-terminal variable transmembrane domain approximately from residues 1 to 44, followed by a middle region (45–213) containing a sequence motif found in the super* Corresponding author. Address: Department of Biological Sciences, Faculty of Science, 10 Kent Ridge Crescent, Singapore 119260, Singapore. Fax: +65 6779 2486. E-mail address: [email protected] (J. Song). 0006-291X/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2009.12.070
family of S-adenosyl methionine (SAM)-dependent radical enzymes, and a C-terminal region (214–361) highly conserved in viperins of all species examined (Fig. 1A). Recently the N-terminal region was experimentally demonstrated to form an amphipathic a-helix which is anchored into the endoplasmic reticulum (ER) membranes to inhibit protein secretion [10], and onto lipid droplets to inhibit HCV [11]. On the other hand, a CX3CX2C sequence motif was previously identified over residues 83–90 of the human viperin (Fig. 1B) [12], which is conserved in all radical SAM enzymes required for the synthesis of molybdopterin (Mao A) (Fig. 1B), heme D1 (NIRJ), and PQQ (PQQIII) [12–15]. These enzymes with more than 2800 putative members all share one common feature, an unconventional [4Fe–4S] cluster coordinated by three rather than four closely spaced cysteine residues in the motif (Fig. 1C). The unique Fe uncoordinated by a cysteine residue is thus available for binding to the ligand, S-adenosyl methionine (SAM) (Fig. 1D) [12–16]. As such, viperin has been extensively hypothesized to be a SAM enzyme and indeed, mutation of the cysteine residues in the motif has been found to eliminate its antiviral activity against HCV [9]. However, previous attempts to reconstitute the [4Fe–4S] cluster in viperin have all failed [10]. So far no structural characterization has been reported for viperin. In the present study, we aimed to structurally characterize the human viperin by dissecting it into a set of 12 fragments of different lengths, followed by detailed CD and NMR assessment. Our results provide the first insight into the structural properties of
G. Shaveta et al. / Biochemical and Biophysical Research Communications 391 (2010) 1390–1395
Fig. 1. The human viperin and structure of a SAM enzyme (MoaA). (A) The domain organization of the human viperin. The N-terminal 44 residues form a transmembrane helix. The region over residues 45–361 is speculated to contain a radical S-adenosyl methionine (SAM)-dependent enzyme with one conserved CX3CX2C motif over residues 83–90; and a C-terminal region over residues 214–361 which is highly conserved within viperins of all species. (B) The crystal structure of MoaA in complex with SAM (1TV8), with the region corresponding to the viperin C-terminus (214–361) colored in greencyan. (C) Conserved CX3CX2C motifs derived from the human viperin (hVip); cofactor for molybdopterin biosynthesis (MoaA); lysine 2,3aminomutase (LAM); anaerobic ribonucleotide reductase (ARR); biotin synthase (BioB); and lipoyl synthase (LipA). (D) The structure of the MoaA [4Fe–4S] cluster in complex with the ligand SAM. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this paper.)
viperin. Most importantly, we successfully identified a buffer-soluble and structured fragment Viperin (45–361), which only has the N-terminal transmembrane helix deleted. Markedly, after many failures under aerobic condition, we have finally succeeded in reconstituting the [4Fe–4S] cluster under anaerobic condition, thus experimentally revealing that viperin is indeed a radical SAM enzyme. Materials and methods Cloning, expression and purification of the human viperin fragments. Total RNA was extracted from human Monocytic cells using Qiagen RNeasy Mini kit and then the full-length viperin RNA was reverse-transcribed into cDNA by RT-PCR using One-step RT-PCR kit (Qiagen). Subsequently, by using different pairs of primers, the 361-residue human viperin was dissected into 12 fragments, which include nine with the N-terminus differentially truncated, spanning over residues 11–361, 22–361, 36–361, 43–361, 45– 361, 54–361, 60–361, 71–361 and 81–361, respectively, as well as N-terminal fragment (1–72); putative SAM-containing region (71–213) and highly conserved C-terminus (214–361). These fragments as well as the entire viperin were further subcloned into
1391
His-tagged expression vector pET32a. All DNA sequences were confirmed by automated DNA sequencing. The recombinant proteins were expressed in Escherichia coli BL21 cells. Briefly, the cells were cultured at 37 °C to an OD600 of 0.6 and then IPTG was added to a final concentration of 0.4 mM to induce expression of the recombinant proteins for 5 h at 30 °C. The insoluble fragments in inclusion body were purified by Ni2+affinity column under denaturing condition in the presence of 8 M urea, followed by further HPLC purification on a reverse-phase C4 column (Vydac), while the soluble fragments in supernatant were purified by Ni2+-affinity column under native condition, followed by FPLC purification using a gel filtration column (HiLoad 16/60 Superdex 200) as previously described [17]. The identities of all recombinant proteins described above were verified by MALDI-TOF mass spectrometry. In particular, the protein sequence of Viperin (45–361) was further confirmed by sequencing 75 overlapped trypsin-digested peptide segments with mass spectrometry on a Micromass Q-TOF 2 machine. For heteronuclear NMR experiments, 15N-labeled proteins were prepared following the similar expression and purification procedure except for growing E. coli cells in M9 medium. The (15NH4)2SO4 salt was used for 15N-isotope labeling as previously described [17,18]. Reconstitution of the [4Fe–4S] cluster in Viperin (45–361). Viperin (45–361) containing a putative SAM domain was successfully identified to be soluble. So we attempted to reconstitute the [4Fe–4S] cluster in Viperin (45–361) first by adding FeCl3 into the cell culture medium to reach a final concentration of 0.15 mM. Unfortunately initial experiments were performed under aerobic condition. As a result, the protein samples purified under aerobic condition was colorless, indicating no formation of the [4Fe–4S] cluster. Further reconstitution following the previous protocol [19] also failed under aerobic condition. Nevertheless, when we purified Viperin (45–361) from cells cultured in the medium containing FeCl3 under anaerobic condition in DW Scientific MACSMG-1000 anaerobic chamber, the obtained protein samples are yellowish, indicating the formation of the [4Fe–4S] cluster to some extent. As such, the reconstitution was further performed under anaerobic condition by incubating the purified Viperin (45–361) protein with 10 mM DTT for 30 min, followed by further incubation for 1 h in the presence of Na2S and (NH4)2Fe(SO4)2 at a 10 M excess as previously described [19]. The reconstituted protein was desalted by buffer-exchange in Amicon centrifuge tubes with a molecular weight cutoff of 10 kDa. UV–visible, circular dichroism (CD) and NMR spectroscopy. UV– visible spectra were collected on a Nicolet evolution 300 spectrophotometer with a wavelength ranging from 295 to 585 nm. All CD experiments were performed on a Jasco J-810 spectropolarimeter equipped with a thermal controller using 1 mm path length cuvettes. Viperin (45–361) was dissolved in 10 mM phosphate buffer (pH 6.5). However, the C-terminus Viperin (214–361) was insoluble in buffers and consequently was solubilized in salt-free water (pH 4.0) as we recently described [17,18]. The far-UV CD spectra were recorded at protein concentrations of 20 lM while the near-UV CD spectra were recorded at protein concentrations of 100 lM at 25 °C. The thermal unfolding was conducted as previously described [20]. Final CD spectra were obtained by adding and averaging data from three independent scans. Estimation of the secondary structure contents by deconvoluting far-UV CD spectra is performed with the CDPro software package (http://lamar.colostate.edu/;sreeram/CDPro). NMR samples were prepared by buffer-exchanging the proteins into phosphate buffers (pH 6.5) for Viperin (45–361) or by dissolving the powers into Milli-Q water for Viperin (214–361) and Viperin (71–361) at pH 4.0, with an addition of 50 lL of D2O into 450 lL samples for NMR spin-lock. One-dimensional (1D) 1H NMR and
1392
G. Shaveta et al. / Biochemical and Biophysical Research Communications 391 (2010) 1390–1395
two-dimensional (2D) 1H–15N HSQC experiments were acquired on an 800 MHz Bruker Advanced spectrometer equipped with pulse field gradient units at 298 K as previously described [17,18,20].
Results Structural characterization of the human viperin fragments In the present study, we have cloned the full-length viperin and its 12 dissected fragments into His-tagged expression vector. However, the entire viperin was found to be not expressed. For other fragments, only five were expressible, which include fragments (43–361), (45–361), (71–361), (81–361) and (214–361). Out of them, the fragments (71–361), (81–361) and (214–361) were only found in inclusion body and could not be refolded by fast dilution or dialysis against various buffers. Viperin (43–361) was found to be partly in supernatant but the purified protein was very unstable, which was prone to aggregation even at very low concentrations (<20 lM). Interestingly, the majority of the recombinant Viperin (45–361) protein existed in supernatant and consequently it was purified by Ni2+-affinity column under native condition, followed by FPLC purification after the thrombin cleavage to remove His-tag. Moreover, its protein sequence was also confirmed by trypsin digestion followed by sequencing with mass spectrometry. As presented in Fig. 2A, Although expressed in the medium without FeCl3 and purified under aerobic condition, the recombinant Viperin (45–361) protein has a far-UV CD spectrum with one positive signal at 197 nm, and two negative ones at 208 and 222 nm, respectively, indicating the presence of a large fraction of a-helix secondary structure [17,18]. Indeed, further analysis of the CD spectrum indicates that the unconstituted Viperin (45–361) already contains 65% a-helix; 19% b-strand/turn and 16% unordered secondary structures. However, the unconstituted protein was unstable and prone to aggregation. For example, during the thermal unfolding at a protein concentration of 20 lM, precipitation started at 30 °C and almost completed at 45 °C (Fig. 2A). On the other hand, Viperin (45–361) appears to have considerable tertiary packing as evident from a dramatic difference between its near-UV spectra in the absence and presence of 8 M urea (Fig. 2B). Furthermore, the unconstituted Viperin (45–361) seems undergoing conformational exchanges on lm–ms time scale or/and dynamic aggregation over some regions. Consequently no very up-field peak is observed located closed to or less than 0 ppm in its 1D NMR spectrum (not shown); and HSQC peaks can only be detected for a small portion of residues (Fig. 2C). For instance, only one peak is visible for five Trp side chains in Viperin (45–361). Reconstitution of the [4Fe–4S] cluster in Viperin (45–361) Since viperin has been hypothesized to be a radical SAM enzyme, we thus attempted to reconstitute the [4Fe–4S] cluster in Viperin (45–361). Initially experiments were conducted under aerobic condition and consequently the reconstitution failed. However, when experiments were performed under anaerobic condition, the Viperin (45–361) sample purified from E. coli cells cultured in the medium containing FeCl3 is yellowish (Fig. 3A). This implies that the sample may already gain a small amount of the [4Fe–4S] clusters. Therefore, we further conducted the reconstitution under anaerobic condition by incubating Viperin (45–361) with the buffer previously described [19]. Strikingly, the reconstituted sample becomes brownish (Fig. 3A), indicating that the [4Fe–4S] cluster has been successfully reconstituted in Viperin (45–361). This is further evident from the UV–visible spectroscopic characterization (Fig. 3B). The reconstituted sample has the absorption property characteris-
Fig. 2. Characterization of Viperin (45–361) purified under aerobic conditions. (A) Far-UV CD spectra of Viperin (45–361) at a concentration of 20 lM in 10 mM phosphate buffer (pH 6.5) at different temperatures. The spectrum at 15 °C is colored in blue and that at 95 °C is in red. Inlet: thermal unfolding curve as monitored at 222 nm. (B) Near-UV CD spectra of Viperin (45–361) at a concentration of 100 lM in 10 mM phosphate buffer (pH 6.5) at 25 °C in the absence (blue) and presence of 8 M urea (red). (C) 1H–15N heteronuclear single quantum correlation (HSQC) spectrum of Viperin (45–361) at a concentration of 100 lM in 10 mM phosphate buffer (pH 6.5) at 25 °C. The green arrow is used to indicate the peak resulting from the side chain of Trp residue. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this paper.)
tic of [4Fe–4S] clusters: a shoulder at 325 nm and a broad peak at 415 nm. By contrast, the sample purified under aerobic condition has no absorption at the two wavelengths. Interestingly, the unconstituted sample but purified under anaerobic condition also has the two peaks but their intensities are much smaller, indicating the presence of a small amount of residual [4Fe–4S] clusters only. Notably, as revealed by deconvoluting the far-UV CD spectrum (Fig. 3C), the reconstituted Viperin (45–361) has a higher content of ordered secondary structures than unconstituted one, with 70% a-helix; 25% b-strand/turn. This suggests that the reconstitution of the [4Fe–4S] cluster can induce further formation of the ordered secondary structures from unordered regions in Viperin (45–361). Furthermore, upon reconstitution, the thermal stability of Viperin (45– 361) increased. The reconstituted sample started to precipitate at 45 °C (Fig. 3A) while the unconstituted one began at 30 °C (Fig. 2A). In particular, the very up-field NMR peaks are observed for samples 2 and 3 (Fig. 3E), but not 1, indicating that the reconstitution also triggered the formation of the tight tertiary packing. As such, the very broad NH peaks may mostly result from the paramagnetic effect of Fe ion [23].
G. Shaveta et al. / Biochemical and Biophysical Research Communications 391 (2010) 1390–1395
1393
Fig. 3. Characterization of Viperin (45–361) after reconstitution. (A) NMR samples of Viperin (45–361) purified under aerobic (1); anaerobic condition (2); and after reconstitution (3). (B) UV–visible spectra (295–585 nm) of different Viperin (45–361) samples at protein concentrations of 30 lM, purified under aerobic (1, black); anaerobic conditions (2, blue) and after reconstitution (3, red). (C) Far-UV CD spectra of Viperin (45–361) after reconstitution at a concentration of 20 lM in 10 mM phosphate buffer (pH 6.5) at different temperatures. The spectrum at 15 °C is colored in blue and that at 95 °C is in red. Inlet: thermal unfolding curve as monitored at 222 nm. (D) Near-UV CD spectra of the reconstituted Viperin (45–361) at a concentration of 100 lM in 10 mM phosphate buffer (pH 6.5) in the absence (blue) and presence of 8 M urea (red) at 25 °C. (E) One-dimensional 1H NMR spectra of Viperin (45–361) at 25 °C and a concentration of 250 lM in 50 mM phosphate buffer (pH 6.5) containing 500 mM NaCl, 10% Glycerol, 10 mM b-ME, 3 mM DTT; purified under anaerobic conditions (upper panel) and after reconstitution (lower panel). The red arrows indicate the very up-field NMR resonance peaks. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this paper.)
Structural characterization of the buffer-insoluble Viperin (214–361) Since the C-terminal region is highly conserved within all viperin proteins but has no significant homology with other proteins, here we present the detailed characterization of the human Viperin (214–361) by CD and NMR spectroscopy. Viperin (214–361) was highly expressed in E. coli cells but all existed in inclusion body. As such, we purified the fragment first by Ni2+-affinity column un-
der denaturing condition, followed by HPLC purification on a C4 column. Intriguingly the lyophilized power of Viperin (214–361) was insoluble in buffers but again could be dissolved in salt-free water (pH 4.0) at a protein concentration even up to 1 mM. This result is completely consistent with our previous discovery that buffer-insoluble proteins can in fact be solubilized in salt-free water [17,18,21]. Recently our discovery was also confirmed by other groups [22]. As seen in Fig. 4A, Viperin (214–361) in salt-free water
1394
G. Shaveta et al. / Biochemical and Biophysical Research Communications 391 (2010) 1390–1395
has a far-UV CD spectrum with a positive signal at 192 nm and two negative signals at 207 and 220 nm, respectively. Deconvolution of the CD spectrum indicates that it contains 24% a-helix; 16% b-strand/turn and 60% unordered secondary structures. Furthermore, Viperin (214–361) showed no precipitation during thermal unfolding even up to 95 °C. Noticeably, as judged by the far-UV CD spectrum (Fig. 4A), even at 95 °C its structure is not completely random coil. On the other hand, the non-cooperative unfolding curve implies that Viperin (214–361) has no tight tertiary packing. Also it is worthwhile to note that introduction of salt significantly induces the aggregation of Viperin (214–361) even at a protein concentration of only 20 lM. At 12 mM NaCl, almost all protein precipitated (Fig. 4B). Consistent with CD results, the absence of very up-field peak in its 1D NMR spectrum (not shown) and small HSQC spectral dispersions (Fig. 4C) again indicate that Viperin (214–361) has no tight tertiary packing. Interestingly, for Viperin (214–361) HSQC peaks can be detected for all three Trp side chains (Fig. 4C).
Fig. 4. Characterization of Viperin (214–361) solubilized in salt-free water. (A) FarUV CD spectra of Viperin (214–361) at a protein concentration of 20 lM in salt-free water (pH 4.0) at different temperatures. The spectrum at 15 °C is colored in blue and that at 95 °C is in red. Inlet: thermal unfolding curve as monitored at 222 nm. (B) Far-UV CD spectra of Viperin (214–361) at a protein concentration of 20 lM at different concentrations of NaCl (0, 0.5, 1, 1.5, 2, 2.5, 3, 4, 5, 6, 8, 10 and 12 mM) at 25 °C. The spectrum in salt-free water (pH 4.0) without addition of NaCl is colored in blue while that with a final NaCl concentration of 12 mM is in red. The Viperin (214–361) forms visible aggregates at 12 mM NaCl. (C) 1H–15N HSQC spectrum of Viperin (214–361) at a protein concentration of 1 mM in salt-free water (pH 4.0) at 25 °C. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this paper.)
Discussion Interferons (IFNs) initiate the first line of defense against viral infection by altering the expression of hundreds of genes, out of which viperin is highly inducible by both type I and type II IFNs. The key role of viperin in innate immunity is strongly implicated by its evolutionary conservation across both mammals and lower vertebrates. Recently several studies have successfully revealed that the N-terminal region functions by forming an amphipathic a-helix to target different membranes [5,10,11]. However, the structural and biochemical properties remain completely unknown for other regions. Previously based on the presence of a characteristic CX3CX2C motif, it was extensively hypothesized that viperin might be a radical SAM enzyme. However, attempts to reconstitute the [4Fe–4S] cluster in viperin all failed. In the present study, by a systematic dissection of viperin into 12 fragments, the Viperin (45–361) fragment only with the N-terminal transmembrane helix deleted has been successfully identified to be soluble and structured in buffers. Most importantly, after intense attempts, we have succeeded in reconstituting the [4Fe–4S] cluster in Viperin (45–361) under anaerobic condition. This success thus provides the first experimental evidence confirming that viperin is indeed a radical SAM enzyme. Intriguingly, Viperin (45–361) was characterized by CD spectroscopy to have largely formed secondary structures and considerable tertiary packing even before the reconstitution of the [4Fe–4S] cluster. This implies that the unconstituted Viperin (45– 361) may already adopt the overall fold common to radical SAM enzymes, which contains a large content of a-helix, as shown by the MoaA structure (Fig. 1B). However, the unconstituted Viperin (45–361) appears less stable and is more prone to aggregation than the reconstituted one. In different radical SAM enzymes, the region corresponding to Viperin (214–316) is highly variable and recognized to be responsible for binding to specific substrates [9,12–16]. Here we demonstrate that despite having some helix secondary structure, the Viperin C-terminus over residues 214–361 is only soluble in saltfree water but not buffers. Furthermore, as revealed by CD analysis, a large portion of Viperin (214–361) appears to become unfolded upon being isolated from Viperin (45–361). Based on the model we recently proposed [18,21], the high insolubility of Viperin (214–361) in buffers is caused by a significant exposure of the hydrophobic side chains to bulk solvent, which likely results from the inappropriate dissection. As seen in Fig. 1B, if Viperin (45–361) adopts the SAM enzyme fold, the C-terminal residues over 214– 361 should have extensive packing interactions with residues over 45–361. Consequently, the dissection of Viperin (214–361) will result in a dramatic loss of these packing interactions, thus leading to the unfolding of some regions, as well as a significant exposure of hydrophobic side chains. Taken together, the results from the present study imply that most residues of Viperin (45–361) may be required to form the SAM enzyme fold, which is not further dissectible. This speculation is further supported by our detailed CD and NMR characterization of another fragment, Viperin (71– 361). Viperin (71–361) was also insoluble in buffers but again could be dissolved in salt-free water at high protein concentrations. Similarly, Viperin (71–361) is only partially folded in saltfree water, with some secondary structures but no tight tertiary packing (data not shown).
Conclusion Here, for the first time we have characterized the structural properties of different viperin fragments and most importantly
G. Shaveta et al. / Biochemical and Biophysical Research Communications 391 (2010) 1390–1395
obtained a buffer-soluble and structured fragment, Viperin (45– 361). This thus led to our successful reconstitution of the [4Fe– 4S] cluster under anaerobic condition, thus experimentally confirming that viperin is indeed a radical S-adenosyl-L-methionine (SAM) enzyme. Together with the previous mutagenesis result [9], our present study strongly implies that the radical SAM enzymatic activity of viperin may be extremely critical for its broad antiviral actions. As a consequence, identification of the chemical reaction that viperin catalyzes represents an immediate priority for uncovering the molecular mechanism by which viperin functions. Moreover, the chemical molecules formed in the catalytic reaction by viperin may bear key information for further design of antiviral drugs of significant therapeutic applications. Also, Viperin (45–361) offers a promising template for determining the three-dimensional structure by either NMR spectroscopy or crystallography. Acknowledgments This study is supported by the Ministry of Education (MOE) of Singapore Tier 1 Grant R-154-000-330-112 and Tier 2 Grant R154-000-388-112 to Jianxing Song. References [1] K.C. Chin, P. Cresswell, Viperin (cig5), an IFN-inducible antiviral protein directly induced by human cytomegalovirus, Proc. Natl. Acad. Sci. USA 98 (2001) 15125–15130. [2] H. Zhu, J.P. Cong, T. Shenk, Use of differential display analysis to assess the effect of human cytomegalovirus infection on the accumulation of cellular RNAs: induction of interferon-responsive RNAs, Proc. Natl. Acad. Sci. USA 94 (1997) 13985–13990. [3] L.I. Brodsky, A.S. Wahed, J. Li, et al., A novel unsupervised method to identify genes important in the anti-viral response: application to interferon/ribavirin in hepatitis C patients, PLoS ONE 2 (2007) e584. [4] J. Fink, F. Gu, L. Ling, et al., Host gene expression profiling of dengue virus infection in cell lines and patients, PLoS Negl. Trop. Dis. 1 (2007) e86. [5] X. Wang, E.R. Hinson, P. Cresswell, The interferon-inducible protein viperin inhibits influenza virus release by perturbing lipid rafts, Cell Host Microbe. 2 (2007) 96–105. [6] K.J. Helbig, D.T. Lau, L. Semendric, H.A. Harley, M.R. Beard, Analysis of ISG expression in chronic hepatitis C identifies viperin as a potential antiviral effector, Hepatology 42 (2005) 702–710.
1395
[7] Y. Zhang, C.W. Burke, K.D. Ryman, W.B. Klimstra, Identification and characterization of interferon-induced proteins that inhibit alphavirus replication, J. Virol. 81 (2007) 11246–11255. [8] M.A. Rivieccio, H.S. Suh, Y. Zhao, M.L. Zhao, K.C. Chin, S.C. Lee, C.F. Brosnan, TLR3 ligation activates an antiviral response in human fetal astrocytes: a role for viperin/cig5, J. Immunol. 177 (2006) 4735–4741. [9] D. Jiang, H. Guo, C. Xu, J. Chang, B. Gu, L. Wang, T.M. Block, J.T. Guo, Identification of three interferon-inducible cellular enzymes that inhibit the replication of hepatitis C virus, J. Virol. 82 (2008) 1665–1678. [10] E.R. Hinson, P. Cresswell, The N-terminal amphipathic alpha-helix of viperin mediates localization to the cytosolic face of the endoplasmic reticulum and inhibits protein secretion, J. Biol. Chem. 284 (2009) 4705–4712. [11] E.R. Hinson, P. Cresswell, The antiviral protein, viperin, localizes to lipid droplets via its N-terminal amphipathic {alpha}-helix, Proc. Natl. Acad. Sci. USA (2009) 20452–20457. [12] P. Boudinot, P. Massin, M. Blanco, S. Riffault, A. Benmansour, vig-1, a new fish gene induced by the rhabdovirus glycoprotein, has a virusinduced homologue in humans and shares conserved motifs with the MoaA family, J. Virol. 73 (1999) 1846–1852. [13] H.J. Sofia, G. Chen, B.G. Hetzler, J.F. Reyes-Spindola, N.E. Miller, S.A.M. Radical, A novel protein superfamily linking unresolved steps in familiar biosynthetic pathways with radical mechanisms: functional characterization using new analysis and information visualization methods, Nucleic Acids Res. 29 (2001) 1097–1106. [14] P. Hänzelmann, H. Schindelin, Crystal structure of the S-adenosylmethioninedependent enzyme MoaA and its implications for molybdenum cofactor deficiency in humans, Proc. Natl. Acad. Sci. USA 101 (2004) 12870–12875. [15] S.C. Wang, P.A. Frey, S-Adenosylmethionine as an oxidant: the radical SAM superfamily, Trends Biochem. Sci. 32 (2007) 101–110. [16] P.A. Frey, A.D. Hegeman, F.J. Ruzicka, The radical SAM Superfamily, Crit. Rev. Biochem. Mol. Biol. 43 (2008) 63–88. [17] M. Li, J. Song, Nogo-B receptor possesses an intrinsically unstructured ectodomain and a partially folded cytoplasmic domain, Biochem. Biophys. Res. Commun. 360 (2007) 128–134. [18] M. Li, J. Liu, X. Ran, M. Fang, J. Shi, H. Qin, J.M. Goh, J. Song, Resurrecting abandoned proteins with pure water: CD and NMR studies of protein fragments solubilized in salt-free water, Biophys. J. 91 (2008) 4201–4209. [19] C. Paraskevopoulou, S.A. Fairhurst, D.J. Lowe, P. Brick, S. Onesti, The elongator subunit Elp3 contains a Fe4S4 cluster and binds S-adenosylmethionine, Mol. Microbiol. 59 (2006) 795–806. [20] Z. Wei, J. Song, Molecular mechanism underlying the thermal stability and pH induced unfolding of CHABII, J. Mol. Biol. 348 (2005) 205–218. [21] J. Song, Insight into ‘‘insoluble proteins” with pure water, FEBS Lett. 583 (2009) 953–959. [22] K. Delak, C. Harcup, R. Lakshminarayanan, Z. Sun, Y. Fan, et al., The tooth enamel protein, porcine amelogenin, is an intrinsically disordered protein with an extended molecular configuration in the monomeric form, Biochemistry 48 (2009) 2272–2281. [23] M.J. Osborne, D. Crowe, S.L. Davy, C. Macdonald, G.R. Moore, NMR of paramagnetic proteins, Methods Mol. Biol. 60 (1997) 233–269.
Contents lists available at ScienceDirect
Biochemical and Biophysical Research Communications journal homepage: www.elsevier.com/locate/ybbrc
Structural characterization reveals that viperin is a radical S-adenosyl-L-methionine (SAM) enzyme Goyal Shaveta a, Jiahai Shi b, Vincent T.K. Chow c, Jianxing Song a,b,* a
Department of Biochemistry, Yong Loo Lin School of Medicine, 10 Kent Ridge Crescent, Singapore 119260, Singapore Department of Biological Sciences, Faculty of Science, 10 Kent Ridge Crescent, Singapore 119260, Singapore c Department of Microbiology, Yong Loo Lin School of Medicine, National University of Singapore, 5 Science Drive 2, Singapore 117597, Singapore b
a r t i c l e
i n f o
Article history: Received 8 December 2009 Available online 22 December 2009 Keywords: Viperin Interferon Radical S-adenosyl-L-methionine (SAM) enzyme Iron–sulfur cluster Circular dichroism spectroscopy NMR spectroscopy
a b s t r a c t Viperin is an interferon-inducible protein inhibiting many DNA and RNA viruses. It contains an N-terminal transmembrane helix, a highly conserved C-terminus and a middle region carrying a CX3CX2C motif, characteristic of radical S-adenosyl-L-methionine (SAM) enzymes. So far no structural characterization has been reported and reconstitution of the [4Fe–4S] cluster in viperin all failed. Here, by dissecting the 361-residue human viperin into 12 fragments, followed by extensive CD and NMR characterization, Viperin (45–361) was identified to be soluble and structured in buffers. Most importantly, we have successfully reconstituted the [4Fe–4S] cluster in Viperin (45–361), thus providing the first experimental evidence confirming that viperin is indeed a radical SAM enzyme. Furthermore, the C-terminus Viperin (214–361) which is insoluble in buffers but again can be solubilized in salt-free water appears to be only partially folded. Our results thus imply that the radical SAM enzyme activity may play a key role in the broad antiviral actions of viperin. Ó 2009 Elsevier Inc. All rights reserved.
Introduction Viperin (virus inhibitory protein, endoplasmic reticulum-associated, interferon-inducible) was originally cloned from interferon treated human macrophages [1,2]. Subsequent studies reveal that viperin is a highly inducible candidate gene in response to a wide spectrum of viruses and microbial products such as LPS and double-stranded RNA [3,4]. Viperin is highly conserved across both mammals and lower vertebrates, localized in the endoplasmic reticulum (ER), thus implying its extremely important role. Indeed, viperin has been demonstrated to inhibit a large array of DNA and RNA viruses, such as human cytomegalovirus [1,2], influenza, hepatitis C virus (HCV) and alphaviruses [5–7], as well as human immunodeficiency virus [8]. However, it remains largely elusive how viperin is capable of inhibiting such a diverse spectrum of viruses since it is very unlikely that it functions by directly binding to viral proteins. The human viperin gene encodes a protein of 361 amino acids, and based on the sequence alignment [9], it appears composed of three distinctive regions: an N-terminal variable transmembrane domain approximately from residues 1 to 44, followed by a middle region (45–213) containing a sequence motif found in the super* Corresponding author. Address: Department of Biological Sciences, Faculty of Science, 10 Kent Ridge Crescent, Singapore 119260, Singapore. Fax: +65 6779 2486. E-mail address: [email protected] (J. Song). 0006-291X/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.bbrc.2009.12.070
family of S-adenosyl methionine (SAM)-dependent radical enzymes, and a C-terminal region (214–361) highly conserved in viperins of all species examined (Fig. 1A). Recently the N-terminal region was experimentally demonstrated to form an amphipathic a-helix which is anchored into the endoplasmic reticulum (ER) membranes to inhibit protein secretion [10], and onto lipid droplets to inhibit HCV [11]. On the other hand, a CX3CX2C sequence motif was previously identified over residues 83–90 of the human viperin (Fig. 1B) [12], which is conserved in all radical SAM enzymes required for the synthesis of molybdopterin (Mao A) (Fig. 1B), heme D1 (NIRJ), and PQQ (PQQIII) [12–15]. These enzymes with more than 2800 putative members all share one common feature, an unconventional [4Fe–4S] cluster coordinated by three rather than four closely spaced cysteine residues in the motif (Fig. 1C). The unique Fe uncoordinated by a cysteine residue is thus available for binding to the ligand, S-adenosyl methionine (SAM) (Fig. 1D) [12–16]. As such, viperin has been extensively hypothesized to be a SAM enzyme and indeed, mutation of the cysteine residues in the motif has been found to eliminate its antiviral activity against HCV [9]. However, previous attempts to reconstitute the [4Fe–4S] cluster in viperin have all failed [10]. So far no structural characterization has been reported for viperin. In the present study, we aimed to structurally characterize the human viperin by dissecting it into a set of 12 fragments of different lengths, followed by detailed CD and NMR assessment. Our results provide the first insight into the structural properties of
G. Shaveta et al. / Biochemical and Biophysical Research Communications 391 (2010) 1390–1395
Fig. 1. The human viperin and structure of a SAM enzyme (MoaA). (A) The domain organization of the human viperin. The N-terminal 44 residues form a transmembrane helix. The region over residues 45–361 is speculated to contain a radical S-adenosyl methionine (SAM)-dependent enzyme with one conserved CX3CX2C motif over residues 83–90; and a C-terminal region over residues 214–361 which is highly conserved within viperins of all species. (B) The crystal structure of MoaA in complex with SAM (1TV8), with the region corresponding to the viperin C-terminus (214–361) colored in greencyan. (C) Conserved CX3CX2C motifs derived from the human viperin (hVip); cofactor for molybdopterin biosynthesis (MoaA); lysine 2,3aminomutase (LAM); anaerobic ribonucleotide reductase (ARR); biotin synthase (BioB); and lipoyl synthase (LipA). (D) The structure of the MoaA [4Fe–4S] cluster in complex with the ligand SAM. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this paper.)
viperin. Most importantly, we successfully identified a buffer-soluble and structured fragment Viperin (45–361), which only has the N-terminal transmembrane helix deleted. Markedly, after many failures under aerobic condition, we have finally succeeded in reconstituting the [4Fe–4S] cluster under anaerobic condition, thus experimentally revealing that viperin is indeed a radical SAM enzyme. Materials and methods Cloning, expression and purification of the human viperin fragments. Total RNA was extracted from human Monocytic cells using Qiagen RNeasy Mini kit and then the full-length viperin RNA was reverse-transcribed into cDNA by RT-PCR using One-step RT-PCR kit (Qiagen). Subsequently, by using different pairs of primers, the 361-residue human viperin was dissected into 12 fragments, which include nine with the N-terminus differentially truncated, spanning over residues 11–361, 22–361, 36–361, 43–361, 45– 361, 54–361, 60–361, 71–361 and 81–361, respectively, as well as N-terminal fragment (1–72); putative SAM-containing region (71–213) and highly conserved C-terminus (214–361). These fragments as well as the entire viperin were further subcloned into
1391
His-tagged expression vector pET32a. All DNA sequences were confirmed by automated DNA sequencing. The recombinant proteins were expressed in Escherichia coli BL21 cells. Briefly, the cells were cultured at 37 °C to an OD600 of 0.6 and then IPTG was added to a final concentration of 0.4 mM to induce expression of the recombinant proteins for 5 h at 30 °C. The insoluble fragments in inclusion body were purified by Ni2+affinity column under denaturing condition in the presence of 8 M urea, followed by further HPLC purification on a reverse-phase C4 column (Vydac), while the soluble fragments in supernatant were purified by Ni2+-affinity column under native condition, followed by FPLC purification using a gel filtration column (HiLoad 16/60 Superdex 200) as previously described [17]. The identities of all recombinant proteins described above were verified by MALDI-TOF mass spectrometry. In particular, the protein sequence of Viperin (45–361) was further confirmed by sequencing 75 overlapped trypsin-digested peptide segments with mass spectrometry on a Micromass Q-TOF 2 machine. For heteronuclear NMR experiments, 15N-labeled proteins were prepared following the similar expression and purification procedure except for growing E. coli cells in M9 medium. The (15NH4)2SO4 salt was used for 15N-isotope labeling as previously described [17,18]. Reconstitution of the [4Fe–4S] cluster in Viperin (45–361). Viperin (45–361) containing a putative SAM domain was successfully identified to be soluble. So we attempted to reconstitute the [4Fe–4S] cluster in Viperin (45–361) first by adding FeCl3 into the cell culture medium to reach a final concentration of 0.15 mM. Unfortunately initial experiments were performed under aerobic condition. As a result, the protein samples purified under aerobic condition was colorless, indicating no formation of the [4Fe–4S] cluster. Further reconstitution following the previous protocol [19] also failed under aerobic condition. Nevertheless, when we purified Viperin (45–361) from cells cultured in the medium containing FeCl3 under anaerobic condition in DW Scientific MACSMG-1000 anaerobic chamber, the obtained protein samples are yellowish, indicating the formation of the [4Fe–4S] cluster to some extent. As such, the reconstitution was further performed under anaerobic condition by incubating the purified Viperin (45–361) protein with 10 mM DTT for 30 min, followed by further incubation for 1 h in the presence of Na2S and (NH4)2Fe(SO4)2 at a 10 M excess as previously described [19]. The reconstituted protein was desalted by buffer-exchange in Amicon centrifuge tubes with a molecular weight cutoff of 10 kDa. UV–visible, circular dichroism (CD) and NMR spectroscopy. UV– visible spectra were collected on a Nicolet evolution 300 spectrophotometer with a wavelength ranging from 295 to 585 nm. All CD experiments were performed on a Jasco J-810 spectropolarimeter equipped with a thermal controller using 1 mm path length cuvettes. Viperin (45–361) was dissolved in 10 mM phosphate buffer (pH 6.5). However, the C-terminus Viperin (214–361) was insoluble in buffers and consequently was solubilized in salt-free water (pH 4.0) as we recently described [17,18]. The far-UV CD spectra were recorded at protein concentrations of 20 lM while the near-UV CD spectra were recorded at protein concentrations of 100 lM at 25 °C. The thermal unfolding was conducted as previously described [20]. Final CD spectra were obtained by adding and averaging data from three independent scans. Estimation of the secondary structure contents by deconvoluting far-UV CD spectra is performed with the CDPro software package (http://lamar.colostate.edu/;sreeram/CDPro). NMR samples were prepared by buffer-exchanging the proteins into phosphate buffers (pH 6.5) for Viperin (45–361) or by dissolving the powers into Milli-Q water for Viperin (214–361) and Viperin (71–361) at pH 4.0, with an addition of 50 lL of D2O into 450 lL samples for NMR spin-lock. One-dimensional (1D) 1H NMR and
1392
G. Shaveta et al. / Biochemical and Biophysical Research Communications 391 (2010) 1390–1395
two-dimensional (2D) 1H–15N HSQC experiments were acquired on an 800 MHz Bruker Advanced spectrometer equipped with pulse field gradient units at 298 K as previously described [17,18,20].
Results Structural characterization of the human viperin fragments In the present study, we have cloned the full-length viperin and its 12 dissected fragments into His-tagged expression vector. However, the entire viperin was found to be not expressed. For other fragments, only five were expressible, which include fragments (43–361), (45–361), (71–361), (81–361) and (214–361). Out of them, the fragments (71–361), (81–361) and (214–361) were only found in inclusion body and could not be refolded by fast dilution or dialysis against various buffers. Viperin (43–361) was found to be partly in supernatant but the purified protein was very unstable, which was prone to aggregation even at very low concentrations (<20 lM). Interestingly, the majority of the recombinant Viperin (45–361) protein existed in supernatant and consequently it was purified by Ni2+-affinity column under native condition, followed by FPLC purification after the thrombin cleavage to remove His-tag. Moreover, its protein sequence was also confirmed by trypsin digestion followed by sequencing with mass spectrometry. As presented in Fig. 2A, Although expressed in the medium without FeCl3 and purified under aerobic condition, the recombinant Viperin (45–361) protein has a far-UV CD spectrum with one positive signal at 197 nm, and two negative ones at 208 and 222 nm, respectively, indicating the presence of a large fraction of a-helix secondary structure [17,18]. Indeed, further analysis of the CD spectrum indicates that the unconstituted Viperin (45–361) already contains 65% a-helix; 19% b-strand/turn and 16% unordered secondary structures. However, the unconstituted protein was unstable and prone to aggregation. For example, during the thermal unfolding at a protein concentration of 20 lM, precipitation started at 30 °C and almost completed at 45 °C (Fig. 2A). On the other hand, Viperin (45–361) appears to have considerable tertiary packing as evident from a dramatic difference between its near-UV spectra in the absence and presence of 8 M urea (Fig. 2B). Furthermore, the unconstituted Viperin (45–361) seems undergoing conformational exchanges on lm–ms time scale or/and dynamic aggregation over some regions. Consequently no very up-field peak is observed located closed to or less than 0 ppm in its 1D NMR spectrum (not shown); and HSQC peaks can only be detected for a small portion of residues (Fig. 2C). For instance, only one peak is visible for five Trp side chains in Viperin (45–361). Reconstitution of the [4Fe–4S] cluster in Viperin (45–361) Since viperin has been hypothesized to be a radical SAM enzyme, we thus attempted to reconstitute the [4Fe–4S] cluster in Viperin (45–361). Initially experiments were conducted under aerobic condition and consequently the reconstitution failed. However, when experiments were performed under anaerobic condition, the Viperin (45–361) sample purified from E. coli cells cultured in the medium containing FeCl3 is yellowish (Fig. 3A). This implies that the sample may already gain a small amount of the [4Fe–4S] clusters. Therefore, we further conducted the reconstitution under anaerobic condition by incubating Viperin (45–361) with the buffer previously described [19]. Strikingly, the reconstituted sample becomes brownish (Fig. 3A), indicating that the [4Fe–4S] cluster has been successfully reconstituted in Viperin (45–361). This is further evident from the UV–visible spectroscopic characterization (Fig. 3B). The reconstituted sample has the absorption property characteris-
Fig. 2. Characterization of Viperin (45–361) purified under aerobic conditions. (A) Far-UV CD spectra of Viperin (45–361) at a concentration of 20 lM in 10 mM phosphate buffer (pH 6.5) at different temperatures. The spectrum at 15 °C is colored in blue and that at 95 °C is in red. Inlet: thermal unfolding curve as monitored at 222 nm. (B) Near-UV CD spectra of Viperin (45–361) at a concentration of 100 lM in 10 mM phosphate buffer (pH 6.5) at 25 °C in the absence (blue) and presence of 8 M urea (red). (C) 1H–15N heteronuclear single quantum correlation (HSQC) spectrum of Viperin (45–361) at a concentration of 100 lM in 10 mM phosphate buffer (pH 6.5) at 25 °C. The green arrow is used to indicate the peak resulting from the side chain of Trp residue. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this paper.)
tic of [4Fe–4S] clusters: a shoulder at 325 nm and a broad peak at 415 nm. By contrast, the sample purified under aerobic condition has no absorption at the two wavelengths. Interestingly, the unconstituted sample but purified under anaerobic condition also has the two peaks but their intensities are much smaller, indicating the presence of a small amount of residual [4Fe–4S] clusters only. Notably, as revealed by deconvoluting the far-UV CD spectrum (Fig. 3C), the reconstituted Viperin (45–361) has a higher content of ordered secondary structures than unconstituted one, with 70% a-helix; 25% b-strand/turn. This suggests that the reconstitution of the [4Fe–4S] cluster can induce further formation of the ordered secondary structures from unordered regions in Viperin (45–361). Furthermore, upon reconstitution, the thermal stability of Viperin (45– 361) increased. The reconstituted sample started to precipitate at 45 °C (Fig. 3A) while the unconstituted one began at 30 °C (Fig. 2A). In particular, the very up-field NMR peaks are observed for samples 2 and 3 (Fig. 3E), but not 1, indicating that the reconstitution also triggered the formation of the tight tertiary packing. As such, the very broad NH peaks may mostly result from the paramagnetic effect of Fe ion [23].
G. Shaveta et al. / Biochemical and Biophysical Research Communications 391 (2010) 1390–1395
1393
Fig. 3. Characterization of Viperin (45–361) after reconstitution. (A) NMR samples of Viperin (45–361) purified under aerobic (1); anaerobic condition (2); and after reconstitution (3). (B) UV–visible spectra (295–585 nm) of different Viperin (45–361) samples at protein concentrations of 30 lM, purified under aerobic (1, black); anaerobic conditions (2, blue) and after reconstitution (3, red). (C) Far-UV CD spectra of Viperin (45–361) after reconstitution at a concentration of 20 lM in 10 mM phosphate buffer (pH 6.5) at different temperatures. The spectrum at 15 °C is colored in blue and that at 95 °C is in red. Inlet: thermal unfolding curve as monitored at 222 nm. (D) Near-UV CD spectra of the reconstituted Viperin (45–361) at a concentration of 100 lM in 10 mM phosphate buffer (pH 6.5) in the absence (blue) and presence of 8 M urea (red) at 25 °C. (E) One-dimensional 1H NMR spectra of Viperin (45–361) at 25 °C and a concentration of 250 lM in 50 mM phosphate buffer (pH 6.5) containing 500 mM NaCl, 10% Glycerol, 10 mM b-ME, 3 mM DTT; purified under anaerobic conditions (upper panel) and after reconstitution (lower panel). The red arrows indicate the very up-field NMR resonance peaks. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this paper.)
Structural characterization of the buffer-insoluble Viperin (214–361) Since the C-terminal region is highly conserved within all viperin proteins but has no significant homology with other proteins, here we present the detailed characterization of the human Viperin (214–361) by CD and NMR spectroscopy. Viperin (214–361) was highly expressed in E. coli cells but all existed in inclusion body. As such, we purified the fragment first by Ni2+-affinity column un-
der denaturing condition, followed by HPLC purification on a C4 column. Intriguingly the lyophilized power of Viperin (214–361) was insoluble in buffers but again could be dissolved in salt-free water (pH 4.0) at a protein concentration even up to 1 mM. This result is completely consistent with our previous discovery that buffer-insoluble proteins can in fact be solubilized in salt-free water [17,18,21]. Recently our discovery was also confirmed by other groups [22]. As seen in Fig. 4A, Viperin (214–361) in salt-free water
1394
G. Shaveta et al. / Biochemical and Biophysical Research Communications 391 (2010) 1390–1395
has a far-UV CD spectrum with a positive signal at 192 nm and two negative signals at 207 and 220 nm, respectively. Deconvolution of the CD spectrum indicates that it contains 24% a-helix; 16% b-strand/turn and 60% unordered secondary structures. Furthermore, Viperin (214–361) showed no precipitation during thermal unfolding even up to 95 °C. Noticeably, as judged by the far-UV CD spectrum (Fig. 4A), even at 95 °C its structure is not completely random coil. On the other hand, the non-cooperative unfolding curve implies that Viperin (214–361) has no tight tertiary packing. Also it is worthwhile to note that introduction of salt significantly induces the aggregation of Viperin (214–361) even at a protein concentration of only 20 lM. At 12 mM NaCl, almost all protein precipitated (Fig. 4B). Consistent with CD results, the absence of very up-field peak in its 1D NMR spectrum (not shown) and small HSQC spectral dispersions (Fig. 4C) again indicate that Viperin (214–361) has no tight tertiary packing. Interestingly, for Viperin (214–361) HSQC peaks can be detected for all three Trp side chains (Fig. 4C).
Fig. 4. Characterization of Viperin (214–361) solubilized in salt-free water. (A) FarUV CD spectra of Viperin (214–361) at a protein concentration of 20 lM in salt-free water (pH 4.0) at different temperatures. The spectrum at 15 °C is colored in blue and that at 95 °C is in red. Inlet: thermal unfolding curve as monitored at 222 nm. (B) Far-UV CD spectra of Viperin (214–361) at a protein concentration of 20 lM at different concentrations of NaCl (0, 0.5, 1, 1.5, 2, 2.5, 3, 4, 5, 6, 8, 10 and 12 mM) at 25 °C. The spectrum in salt-free water (pH 4.0) without addition of NaCl is colored in blue while that with a final NaCl concentration of 12 mM is in red. The Viperin (214–361) forms visible aggregates at 12 mM NaCl. (C) 1H–15N HSQC spectrum of Viperin (214–361) at a protein concentration of 1 mM in salt-free water (pH 4.0) at 25 °C. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this paper.)
Discussion Interferons (IFNs) initiate the first line of defense against viral infection by altering the expression of hundreds of genes, out of which viperin is highly inducible by both type I and type II IFNs. The key role of viperin in innate immunity is strongly implicated by its evolutionary conservation across both mammals and lower vertebrates. Recently several studies have successfully revealed that the N-terminal region functions by forming an amphipathic a-helix to target different membranes [5,10,11]. However, the structural and biochemical properties remain completely unknown for other regions. Previously based on the presence of a characteristic CX3CX2C motif, it was extensively hypothesized that viperin might be a radical SAM enzyme. However, attempts to reconstitute the [4Fe–4S] cluster in viperin all failed. In the present study, by a systematic dissection of viperin into 12 fragments, the Viperin (45–361) fragment only with the N-terminal transmembrane helix deleted has been successfully identified to be soluble and structured in buffers. Most importantly, after intense attempts, we have succeeded in reconstituting the [4Fe–4S] cluster in Viperin (45–361) under anaerobic condition. This success thus provides the first experimental evidence confirming that viperin is indeed a radical SAM enzyme. Intriguingly, Viperin (45–361) was characterized by CD spectroscopy to have largely formed secondary structures and considerable tertiary packing even before the reconstitution of the [4Fe–4S] cluster. This implies that the unconstituted Viperin (45– 361) may already adopt the overall fold common to radical SAM enzymes, which contains a large content of a-helix, as shown by the MoaA structure (Fig. 1B). However, the unconstituted Viperin (45–361) appears less stable and is more prone to aggregation than the reconstituted one. In different radical SAM enzymes, the region corresponding to Viperin (214–316) is highly variable and recognized to be responsible for binding to specific substrates [9,12–16]. Here we demonstrate that despite having some helix secondary structure, the Viperin C-terminus over residues 214–361 is only soluble in saltfree water but not buffers. Furthermore, as revealed by CD analysis, a large portion of Viperin (214–361) appears to become unfolded upon being isolated from Viperin (45–361). Based on the model we recently proposed [18,21], the high insolubility of Viperin (214–361) in buffers is caused by a significant exposure of the hydrophobic side chains to bulk solvent, which likely results from the inappropriate dissection. As seen in Fig. 1B, if Viperin (45–361) adopts the SAM enzyme fold, the C-terminal residues over 214– 361 should have extensive packing interactions with residues over 45–361. Consequently, the dissection of Viperin (214–361) will result in a dramatic loss of these packing interactions, thus leading to the unfolding of some regions, as well as a significant exposure of hydrophobic side chains. Taken together, the results from the present study imply that most residues of Viperin (45–361) may be required to form the SAM enzyme fold, which is not further dissectible. This speculation is further supported by our detailed CD and NMR characterization of another fragment, Viperin (71– 361). Viperin (71–361) was also insoluble in buffers but again could be dissolved in salt-free water at high protein concentrations. Similarly, Viperin (71–361) is only partially folded in saltfree water, with some secondary structures but no tight tertiary packing (data not shown).
Conclusion Here, for the first time we have characterized the structural properties of different viperin fragments and most importantly
G. Shaveta et al. / Biochemical and Biophysical Research Communications 391 (2010) 1390–1395
obtained a buffer-soluble and structured fragment, Viperin (45– 361). This thus led to our successful reconstitution of the [4Fe– 4S] cluster under anaerobic condition, thus experimentally confirming that viperin is indeed a radical S-adenosyl-L-methionine (SAM) enzyme. Together with the previous mutagenesis result [9], our present study strongly implies that the radical SAM enzymatic activity of viperin may be extremely critical for its broad antiviral actions. As a consequence, identification of the chemical reaction that viperin catalyzes represents an immediate priority for uncovering the molecular mechanism by which viperin functions. Moreover, the chemical molecules formed in the catalytic reaction by viperin may bear key information for further design of antiviral drugs of significant therapeutic applications. Also, Viperin (45–361) offers a promising template for determining the three-dimensional structure by either NMR spectroscopy or crystallography. Acknowledgments This study is supported by the Ministry of Education (MOE) of Singapore Tier 1 Grant R-154-000-330-112 and Tier 2 Grant R154-000-388-112 to Jianxing Song. References [1] K.C. Chin, P. Cresswell, Viperin (cig5), an IFN-inducible antiviral protein directly induced by human cytomegalovirus, Proc. Natl. Acad. Sci. USA 98 (2001) 15125–15130. [2] H. Zhu, J.P. Cong, T. Shenk, Use of differential display analysis to assess the effect of human cytomegalovirus infection on the accumulation of cellular RNAs: induction of interferon-responsive RNAs, Proc. Natl. Acad. Sci. USA 94 (1997) 13985–13990. [3] L.I. Brodsky, A.S. Wahed, J. Li, et al., A novel unsupervised method to identify genes important in the anti-viral response: application to interferon/ribavirin in hepatitis C patients, PLoS ONE 2 (2007) e584. [4] J. Fink, F. Gu, L. Ling, et al., Host gene expression profiling of dengue virus infection in cell lines and patients, PLoS Negl. Trop. Dis. 1 (2007) e86. [5] X. Wang, E.R. Hinson, P. Cresswell, The interferon-inducible protein viperin inhibits influenza virus release by perturbing lipid rafts, Cell Host Microbe. 2 (2007) 96–105. [6] K.J. Helbig, D.T. Lau, L. Semendric, H.A. Harley, M.R. Beard, Analysis of ISG expression in chronic hepatitis C identifies viperin as a potential antiviral effector, Hepatology 42 (2005) 702–710.
1395
[7] Y. Zhang, C.W. Burke, K.D. Ryman, W.B. Klimstra, Identification and characterization of interferon-induced proteins that inhibit alphavirus replication, J. Virol. 81 (2007) 11246–11255. [8] M.A. Rivieccio, H.S. Suh, Y. Zhao, M.L. Zhao, K.C. Chin, S.C. Lee, C.F. Brosnan, TLR3 ligation activates an antiviral response in human fetal astrocytes: a role for viperin/cig5, J. Immunol. 177 (2006) 4735–4741. [9] D. Jiang, H. Guo, C. Xu, J. Chang, B. Gu, L. Wang, T.M. Block, J.T. Guo, Identification of three interferon-inducible cellular enzymes that inhibit the replication of hepatitis C virus, J. Virol. 82 (2008) 1665–1678. [10] E.R. Hinson, P. Cresswell, The N-terminal amphipathic alpha-helix of viperin mediates localization to the cytosolic face of the endoplasmic reticulum and inhibits protein secretion, J. Biol. Chem. 284 (2009) 4705–4712. [11] E.R. Hinson, P. Cresswell, The antiviral protein, viperin, localizes to lipid droplets via its N-terminal amphipathic {alpha}-helix, Proc. Natl. Acad. Sci. USA (2009) 20452–20457. [12] P. Boudinot, P. Massin, M. Blanco, S. Riffault, A. Benmansour, vig-1, a new fish gene induced by the rhabdovirus glycoprotein, has a virusinduced homologue in humans and shares conserved motifs with the MoaA family, J. Virol. 73 (1999) 1846–1852. [13] H.J. Sofia, G. Chen, B.G. Hetzler, J.F. Reyes-Spindola, N.E. Miller, S.A.M. Radical, A novel protein superfamily linking unresolved steps in familiar biosynthetic pathways with radical mechanisms: functional characterization using new analysis and information visualization methods, Nucleic Acids Res. 29 (2001) 1097–1106. [14] P. Hänzelmann, H. Schindelin, Crystal structure of the S-adenosylmethioninedependent enzyme MoaA and its implications for molybdenum cofactor deficiency in humans, Proc. Natl. Acad. Sci. USA 101 (2004) 12870–12875. [15] S.C. Wang, P.A. Frey, S-Adenosylmethionine as an oxidant: the radical SAM superfamily, Trends Biochem. Sci. 32 (2007) 101–110. [16] P.A. Frey, A.D. Hegeman, F.J. Ruzicka, The radical SAM Superfamily, Crit. Rev. Biochem. Mol. Biol. 43 (2008) 63–88. [17] M. Li, J. Song, Nogo-B receptor possesses an intrinsically unstructured ectodomain and a partially folded cytoplasmic domain, Biochem. Biophys. Res. Commun. 360 (2007) 128–134. [18] M. Li, J. Liu, X. Ran, M. Fang, J. Shi, H. Qin, J.M. Goh, J. Song, Resurrecting abandoned proteins with pure water: CD and NMR studies of protein fragments solubilized in salt-free water, Biophys. J. 91 (2008) 4201–4209. [19] C. Paraskevopoulou, S.A. Fairhurst, D.J. Lowe, P. Brick, S. Onesti, The elongator subunit Elp3 contains a Fe4S4 cluster and binds S-adenosylmethionine, Mol. Microbiol. 59 (2006) 795–806. [20] Z. Wei, J. Song, Molecular mechanism underlying the thermal stability and pH induced unfolding of CHABII, J. Mol. Biol. 348 (2005) 205–218. [21] J. Song, Insight into ‘‘insoluble proteins” with pure water, FEBS Lett. 583 (2009) 953–959. [22] K. Delak, C. Harcup, R. Lakshminarayanan, Z. Sun, Y. Fan, et al., The tooth enamel protein, porcine amelogenin, is an intrinsically disordered protein with an extended molecular configuration in the monomeric form, Biochemistry 48 (2009) 2272–2281. [23] M.J. Osborne, D. Crowe, S.L. Davy, C. Macdonald, G.R. Moore, NMR of paramagnetic proteins, Methods Mol. Biol. 60 (1997) 233–269.