- Email: [email protected]
Crystal structure of the nitrile-specifier protein NSP1 from Arabidopsis thaliana
Accepted Manuscript Crystal structure of the nitrile-specifier protein NSP1 from Arabidopsis thaliana Weiwei Zhang, Yu Zhou, Kan Wang, Yanan Dong, Wenhe Wang, Yue Feng PII:
S0006-291X(17)30873-2
DOI:
10.1016/j.bbrc.2017.05.027
Reference:
YBBRC 37742
To appear in:
Biochemical and Biophysical Research Communications
Received Date: 2 May 2017 Accepted Date: 3 May 2017
Please cite this article as: W. Zhang, Y. Zhou, K. Wang, Y. Dong, W. Wang, Y. Feng, Crystal structure of the nitrile-specifier protein NSP1 from Arabidopsis thaliana, Biochemical and Biophysical Research Communications (2017), doi: 10.1016/j.bbrc.2017.05.027. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
G13 G27
T131 G149 D132 F150
V133 V151
TE D
M AN U
SC
H135 D153
RI PT
AtNSP1-N PDB: 1XXR
R292
AC C
EP
V287 G320 F337 G321
F271
E386 V370 D390 W432
H394
ACCEPTED MANUSCRIPT
Crystal Structure of the nitrile-specifier protein NSP1 from Arabidopsis thaliana Weiwei Zhanga,1, Yu Zhoub,1, Kan Wangc,1, Yanan Donga, Wenhe Wanga and Yue
a
RI PT
Fenga*
Beijing Key Lab of Bioprocess, Beijing University of Chemical Technology, Beijing
100029, PR China
National Institute of Biological Sciences, Beijing, No. 7 Science Park Road,
SC
b
Zhongguancun Life Science Park, Beijing 102206, PR China
Department of Anesthesiology, China-Japan Friendship Hospital, Beijing 100029, PR
M AN U
c
China. 1
These authors contributed equally to this work.
*
Abstract
TE D
Correspondence email: [email protected]
As important components of the glucosinolate-myrosinase system, specifier proteins mediate plant defense against herbivory and pathogen attacks. After tissue disruption,
EP
glucosinolates are hydrolyzed by myrosinases to instable aglucones, which will rearrange to form defensive isothiocyanates. Nevertheless, this reaction could be
AC C
redirected to form other products by specifier proteins. Up to now, identified specifier proteins include epithiospecifier proteins (ESPs), thiocyanate forming proteins (TFPs), and nitrile-specifier proteins (NSPs). Recently, the structures of ESP and TFP have been reported. However, both the structure and the catalytic mechanism of NSPs remain enigmatic. Here, we solved the crystal structure of the NSP1 protein from Arabidopsis thaliana (AtNSP1). Structural comparisons with ESP and TFP proteins revealed several structural features of AtNSP1 different from those of the two proteins. Subsequent molecular docking studies showed that the R292 residue in AtNSP1 displayed a conformation different from those of the corresponding residues in ESP
ACCEPTED MANUSCRIPT and TFP proteins, which might account for the product specificity and catalytic mechanism of AtNSP1. Taken together, the present study provides important insights into the molecular mechanisms underlying the different product spectrums between NSPs and the other two types of specifier proteins, and shed light on the future studies
RI PT
of the detailed mechanisms of other specifier proteins.
Keywords: nitrile-specifier protein, glucosinolate, Arabidopsis thaliana, crystal structure, active site
SC
1. Introduction
Glucosinolates are important plant secondary metabolites present in Brassicaceae
M AN U
plants such as the model plant Arabidopsis thaliana (A. thaliana)[1]. They are a group of thioglucosides, acting as components of an activated chemical defense system in plants [2]. The system is widespread in all the plants of the Brassicales order and involved in numerous plant-insect and plant-pathogen interactions [2, 3]. Upon tissue disruption, the system is activated, causing the release of glucosinolates and their
TE D
hydrolytic enzymes, the so called myrosinases, which are in separate storage compartments under normal conditions. The cleavage of the thioglucosidic bond of the glucosinolates by myrosinases produces aglucone intermediates, which are
EP
unstable and spontaneously rearrange to toxic isothiocyanates through a Lossen rearrangement [4-6]. Isothiocyanates are toxic to organisms including bacteria, fungi, nematodes and insects, and therefore act as defend compounds to protect plants from
AC C
herbivory and pathogen attack [7, 8]. In addition, isothiocyanates such as sulforaphane has also attracted a lot of interest as cancer-preventive components in the human diet. However, the outcome of glucosinolate hydrolysis actually depends on the side chain structure of the glucosinolate and the presence of supplementary proteins called specifier proteins [4, 9, 10]. In the presence of specifier proteins, the production of isothiocyanates can be diverted to the generation of nitriles, epithionitriles, or thiocyanates. Specifier proteins are believed to be enzymes, acting on the unstable aglucones, the hydrolysis products of glucosinolates [4, 9-11]. The nature of the generated product depends on the particular specifier protein, the side
ACCEPTED MANUSCRIPT chain structure of the glucosinolate, as well as the reaction conditions [12]. According to their product spectrum, specifier proteins are classified into epithiospecifier proteins (ESPs), thiocyanate forming proteins (TFPs), and nitrile-specifier proteins (NSPs). First isolated from Crambe abyssinica and named by
RI PT
Tookey [13], ESPs were the best characterized specifier proteins for their activity in promoting epithionitrile formation in the presence of allylglucosinolates and myrosinases. Epithionitrile formation requires the presence of a terminal double bond in the side chain of the aglucone substrate. TFPs are named by their ability to promote
SC
thiocyanate formation [14]. Two TFPs have been characterized, which are from Thlaspi arvense (TaTFP) and Lepidium sativum (LsTFP). Surprisingly, they are
M AN U
different in their substrate/product specificity. TaTFP and LsTFP can form thiocyanate only upon allylglucosinolate and benzylglucosinolate hydrolysis, respectively [4, 9, 11]. NSPs divert the myrosinase-catalyzed hydrolysis of a given glucosinolate from the formation of isothiocyanates to that of simple nitriles. Importantly, NSPs are responsible for constitutive and herbivore-induced simple nitrile formation in
TE D
Arabidopsis ecotype Columbia-0 rosette leaves [1]. Furthermore, Wittstock et al showed that simple nitrile formation after tissue disruption in A. thaliana depended almost entirely on NSP2 in seeds and mainly on NSP1 in seedlings [15]. Recently, we
EP
and others reported the crystal structures of AtESP and TaTFP respectively, which provided insights into the catalytic mechanisms and distinct product spectrums of the two proteins [8, 11]. However, both the structure and catalytic mechanism of NSPs
AC C
have not yet been resolved. Moreover, specifier proteins are well characterized by their four to five Kelch repeats, which form β-propeller tertiary structures [8, 11]. Nevertheless, apart from the conserved Kelch repeats, most NSP proteins also harbor one or two N-terminal jacalin-like lectin domains (Fig. 1) [1, 16, 17]. In contrast to ESP and TFP proteins, NSPs do not possess epithiospecifier activity although they present high sequence similarity to AtESP and TaTFP in the conserved Kelch repeats (Supplementary Fig. S1). Therefore, it is very interesting to investigate the mechanism underlying the product spectrum of NSP proteins. In the present study, we expressed and purified
ACCEPTED MANUSCRIPT AtNSP1 in Escherichia coli, and solved its crystal structure. Structural comparisons with AtESP and TaTFP revealed that AtNSP1 displayed several different structural features. Further molecular docking studies of AtNSP1 and AtESP with the allyglucosinolate substrate identified that the active site residue R292 of AtNSP1
RI PT
displayed an orientation different from the corresponding residues in AtESP and TaTFP, which might account for the product specificity and catalytic mechanism of AtNSP1. The present study will shed light on the molecular basis of the different product spectrums of specifier proteins and provide insights into the evolution of
SC
metabolic diversity in a complex plant defense system.
M AN U
2. Materials and methods 2.1. Materials
The pfu polymerase, the two restriction enzymes Nde I and Xho I and T4 DNA ligase were all purchased from New England Biolabs (NEB). Ni-NTA beads were purchased from QIAGEN and the Superdex-200 column was purchased from GE Healthcare.
TE D
2.2. Cloning, expression, purification and crystallization
The gene encoding AtNSP1 (At3g16400), was amplified by PCR from the cDNA library of A. thaliana. The amplified gene was inserted into pET22b (Novagen,
EP
Madison, WI), resulting in a plasmid encoding a His6 tag fusion (AAALEHHHHHH) at the C terminus of the recombinant AtNSP1. The protein was purified as previously reported [8] and concentrated to 25 mg/ml before use.
AC C
Crystallization screening was performed by the sitting-drop vapor-diffusion method at 291 K. 1 µL protein solution (25 mg/mL) was mixed with an equal volume of reservoir solution in 48-well plates and the drops were equilibrated against 80 µL reservoir solution. Crystals only appeared from one single condition, SaltRx 2 condition No. 45 [4.0 M ammonium acetate, 0.1 M BIS-TRIS propane pH 7.0]. Crystals of the best diffraction quality appeared in about 1 week from the condition containing 4.6 M Ammonium acetate, 0.1 M BIS-TRIS propane pH 7.0 and 0.5 M Ammonium chloride, equilibrated against 200 µL reservoir solution. These crystals were used for data collection.
ACCEPTED MANUSCRIPT 2.3. Data collection, processing and structure determination Crystals were cryoprotected in the reservoir solutions supplemented with 10 % (w/v) glycerol, flash cooled in the liquid nitrogen, and stored at cryogenic temperatures for data collection. X-ray diffraction data were collected on beamline BL17U of the
RI PT
Shanghai Synchrotron Radiation Facility (SSRF) using a Quantum 315r CCD detector (Area Detector Systems Corporation). All data were indexed, integrated and scaled with HKL-2000 [18]. Further processing was carried out using programs from the CCP4 suite [19]. Data collection statistics are summarized in Table 1. The molecular
SC
replacement was performed with PHASER [20] using the structures of Myrosinase Binding Protein At3g16450.1 from Arabidopsis thaliana (PDB: 2JZ4) and AtESP
M AN U
(PDB: 5GQ0) as models. The final model rebuilding was performed using COOT [21] and the protein structure was refined with PHENIX [22] using NCS and stereochemistry information as restraints. Structural figures were generated in PyMOL [23]. 2.4. Molecular docking
TE D
A Fe2+ was manually positioned with soft distance restraints between the proposed Fe2+-coordinating residues E386, D390 and H394 in AtNSP1, and E260, D264 and H268 in AtESP [24] at a distance of 2.1–2.4 Å. The system was then submitted to
EP
energy minimization and 20-ns molecular dynamic (MD) simulation routines. All MD simulations were performed using the Desmond software package [25] and the OPLS-AA 2005 force field [26]. A 10 Å buffered orthorhombic boundary system was
AC C
built with TIP3P water and then neutralized by counter ions. Each system was minimized using 5000 steps of steepest descent algorithm, followed by L-BGFS algorithm. Temperature was gradually increased to 300 K. A 50 kcal·mol−1·Å−2 harmonic position restraint was applied to all heavy atoms of the protein and Fe2+ atom during system equilibration. The restraints were gradually removed and the production run was processed in MTK-NPT (1 bar, 300 K) ensemble for 20 ns. The M-SHAKE algorithm [27] was applied to constrain all bonds involving hydrogen atoms with a time step of 2 fs. Molecular docking of AtNSP1 and AtESP was performed with UCSF DOCK 3.6
ACCEPTED MANUSCRIPT program [28]. For the ligand, 200 binding poses were generated and the submitted to a MM-GB/SA refinement and rescoring procedure [29], where the side chain of binding-site residues were sampled along with the docked ligand applying Protein Local Optimization Program (PLOP) [30]. The docked complex structure was
RI PT
minimized first, followed by the side chain prediction of the binding-site residues within 7 Å of the ligand, and the ligand was minimized with the fixed protein structure. 3. Results and discussion
SC
3.1. Expression, purification, crystallization and data collection
AtNSP1 was successfully cloned and expressed in E. coli and purified to homogeneity
M AN U
with a C-terminal His6 tag. The 52.9 kDa molecular weight of the monomeric His6 AtNSP1 predicted from the sequence was confirmed by 15 % SDS–PAGE (Supplementary Fig. S2). The retention volume of the main peak in the gel-filtration chromatography profile of the target protein indicated that it is a monomer in solution (Supplementary Fig. S2). The crystals of AtNSP1 with best diffraction power were
TE D
crystallized from 4.6 M Ammonium acetate, 0.1 M BIS-TRIS propane pH 7.0 and 0.5 M Ammonium chloride (Supplementary Fig. S3). The crystal of AtNSP1 diffracted to a maximum resolution of 3.02 Å (Table 1) and
EP
belonged to the space group P3121, with unit-cell parameters a = b = 120.902, c = 123.006 Å. An initial model for the AtNSP1 structure was obtained by molecular replacement via PHASER [20]. There is one molecule found in the asymmetric unit.
AC C
The correctness of the solution of PHASER with space group P3121 was clearly indicated by final highest Z-scores and log-likelihood gain (LLG) values for the molecule (RFZ=6.4, TFZ=12.2, LLG=207). The electron density was clear for the majority of the molecules, allowing subsequent refinement to be conveniently conducted. Several rounds of refinement in the resolution range 50.00–3.02 Å were conducted using restrained refinement by REFMAC5 and manual refinement by Coot [31]. 3.2. Overall structure and the N-terminal jacalin-like lectin domain of AtNSP1 Both the Protein Interfaces, Surfaces and Assemblies (PISA) service [32] and gel
ACCEPTED MANUSCRIPT filtration results indicated that AtNSP1 exists as a monomer in solution, consistent with the results of Gumz et al [11]. The structure clearly revealed the two domains of AtNSP1 (Fig. 2A). Similar as TaTFP and AtESP, NSP1 also exhibited a six-blade (β1-β6) β-propeller fold in its active domain (Fig. 2B), with a core root mean square
RI PT
deviation (RMSD) of 0.94 Å to the TaTFP structure among 319 Cα atoms and 0.91 Å to the AtESP structure among 317 Cα atoms (Fig. 2B). To keep consistency with the TaTFP and AtESP studies [5, 8], we used the same nomenclature to name the anti-parallel β strands within each blade (βx1-βx4), and the loops connecting βx2-βx3
SC
in each blade and adjacent blades in AtNSP1 (Supplementary Fig. S1).
Dali search with the the N-terminal jacalin-like lectin domain of AtNSP1 (designated
M AN U
as AtNSP1-N hereafter) returned several entries of lectin domains [33-35] with high structural similarities (Supplementary Fig. S4A-C). Lectins are multivalent carbohydrate-binding proteins of non-immune origin [33]. Jacalin is one of the two lectins from jackfruit (Artocarpus integrifolia) seeds, with galactose specificities. The so called jacalin-like lectin domain, however, binds mannose specifically. Similar as
TE D
other jacalin-like lectin domain, AtNSP1-N exhibited a classical β-prism fold, in which three antiparallel β-sheets, each consisting of four β-strands, shared an interface. Superimposition of the structure of the N-terminal domain of AtNSP1 with those of
EP
other lectin proteins revealed an RMSD from 1.55 to 1.61 Å among 137 Cα atoms (Supplementary Fig. S4A-C). The structure of AtNSP1-N differs from those of other lectin domains in three aspects. First, while most lectin domain proteins fold as
AC C
tetramers, AtNSP1 is a monomer in solution. Second, more than half of the loops in AtNSP1-N show different conformations in the other lectin proteins (Supplementary Fig. S4A-C). Third, the active site residues involved in sugar binding in other lectin domains are not fully conserved in AtNSP1-N (Supplementary Fig. S4E). Interestingly, AtNSP1-N showed both high amino acid sequence identity (over 50 %) and structural similarities to the jacalin domains of Arabidopsis myrosinase-binding proteins (MyroBPs), such as the product of gene At3g16450.1 (Supplementary Fig. S4D). It is worth investigating whether the N-terminal domains of NSP proteins bind mannose or other sugar or sugar derivatives, and whether this binding will affect the
ACCEPTED MANUSCRIPT activity of these proteins. 3.3. Active site structure Based on the structural and biochemical studies into AtESP and TaTFP proteins [5, 8, 11], the active sites of specifier proteins have been located at the bottom of the
RI PT
β-propeller structure. Structural alignments between AtNSP1, AtESP and TaTFP revealed that the loops 3L2 and 4L2 (connecting β32-β33 and β42-β43, respectively) are shorter in AtNSP1 than those in AtESP and TaTFP (Fig. 2B and Supplementary Fig. S1). Interestingly, Gumz et al. have observed a significant conformational change
SC
in loop 4L2, when they docked the aglucone substrate to the TaTFP structure [5]. This loop is also the secondary structural element showing the most remarkable difference
M AN U
between the structures of AtESP and TaTFP. Therefore, the shorter length of loop 4L2 in AtNSP1 might partially account for its different product spectrum. Furthermore, we compared the putative active sites of the three proteins. In spite of different product spectrums, AtNSP1 and AtESP did not show obvious distinctions in the position of their active site residues (Fig. 3A), except that two residues T129 and
TE D
P263 of AtESP were replaced by S270 and M389 in AtNSP1. More differences were observed between the structures of AtNSP1 and TaTFP (Fig. 3B). M186 and C250, which have been proposed to stabilize the substrate in TaTFP, were replaced by G321
EP
and V370 in AtNSP1, respectively. Moreover, R94 in TaTFP, which has been proposed to be involved in substrate sulfate binding, displayed apparent conformational differences from the corresponding residue R237 in AtNSP1. Further
AC C
comparison of the three structures together revealed two remarkable features of AtNSP1 different from those of AtESP and TaTFP. One is that the active site residue M389 is replaced by P263 and R269 in AtESP, and TaTFP, respectively. The other is that although R157, which was formerly proved to be important in both AtESP and TaTFP activity, showed almost the same conformation in these two proteins, the corresponding residue R292 of AtNSP1 displayed a different conformation (Fig. 3). 3.4. Molecular docking studies into AtNSP1 and AtESP Since AtNSP1 had an active site structure more similar to that of AtESP but a different product specificity when using allylglucosinolate as a substrate, we
ACCEPTED MANUSCRIPT performed molecular docking studies into AtNSP1 and AtESP with Fe2+ and allylglucosinolate. First, Fe2+ ions were manually positioned with soft distance restraints between the previously proposed Fe2+-coordinating residues E386, D390 and H394 in AtNSP1, and E260, D264, and H268 in AtESP [24] at a distance of
RI PT
2.1–2.4 Å, respectively. Both systems were then submitted to energy minimization and molecular dynamic (MD) simulation routines. After that, molecular dockings of AtNSP1 and AtESP were performed and the docked complex structures were energy minimized first, followed by side chain prediction of the binding-site residues within
SC
7 Å of the ligand, and the ligand was minimized with the fixed protein structure.
The docking process did not induce obvious conformational changes in the two
M AN U
proteins (data not shown), different from the docking studies into TaTFP by GumZ et al [11]. The docking results indicated that both the allylglucosinolate and Fe2+ were bound
at
similar
sites
of
AtESP
and
AtNSP1
(Fig.
4).
In
the
AtNSP1-allylglucosinolate structure, the Fe2+ ion was strongly coordinated by the sidechains of E386 and D390, and a water molecule. Importantly, the distance
TE D
between the Fe2+ and the NE2 atom of His394 was 3.9 Å, indicating very weak interaction between them. Together with R292, the sidechain of H394 formed four hydrogen bonds with the sulfate group of the substrate. Moreover, the rest of the
EP
aglucone body of the substrate formed hydrophobic interactions with F271, G320, G321, F337, V370 and W432.
In the AtESP-allylglucosinolate structure, the Fe2+ ion was coordinated by the
AC C
sidechains of E260, D264, and H268, which were used in the initial positioning of the Fe2+ ion in the docking studies. The sulfate group of the substrate formed three hydrogen bonds with the sidechain nitrogen atom of R157 and the mainchain nitrogen atom of G185. The substrate was also stabilized by the hydrophobic interactions from F130, G186, F202, V244 and W303. Interestingly, R94, which was previously shown to be involved in substrate sulfate binding in the docking results of TaTFP, has not been observed in substrate binding in either of the docked complex structures of AtNSP1 and AtESP. This is also consistent with the different conformations of TaTFPR94 from those of AtNSP1R237 and AtESPR94. Comparison of the two complex
ACCEPTED MANUSCRIPT structures revealed that R292 residue of AtNSP1 displays an orientation different from that of the corresponding R157 residue of AtESP. In addition, AtNSP1R292 also formed hydrogen bonds with the docked substrate with partially different atoms from those of AtESPR157. These indicated that this residue might account for the distinct
RI PT
product specificity of these two proteins. 4. Discussion
In spite of high sequence identities, AtNSP1 and AtESP could lead to different products, simple nitrile and epithionitrile, respectively, when using allylglucosinolate
SC
as substrates. It has been proposed that during the production of simple nitrile by NSP proteins, the negatively charged sulfur derived from the thioglucosidic bond could be
M AN U
abstracted from the aglucone by the Fe2+ ion, and the sulfur atom might be released as elemental sulfur at the end of the reaction due to the lack of a suitable acceptor [4]. However, in the presence of ESP proteins, the sulfur abstracted by the Fe2+ ion would further attack the terminal double bond resulting in formation of the thiirane ring through a currently unknown mechanism. We propose that R157, which displayed
TE D
similar conformations in AtESP and TaTFP but different conformation in AtNSP1, might play a role in the formation of epithionitrile. However, as aglucones are unstable chemicals, it is quite difficult to obtain the enzyme-substrate complex
EP
structures of specifier proteins. Therefore, future molecular dynamic simulations, together with mutagenesis studies based on the AtNSP1 structure reported here could further provide insights into the detailed mechanism underlying the different product
AC C
spectrums of specifier proteins.
Acknowledgements
We would like to thank the staff at beamline BL17U of the Shanghai Synchrotron Radiation Facility for their assistance with data collection. This work was supported by the National Nature Science Foundation of China (31400635 and 31670766) and the Fundamental Research Funds for the Central Universities (JD1709 and XK1701).
ACCEPTED MANUSCRIPT References
AC C
EP
TE D
M AN U
SC
RI PT
[1] M. Burow, A. Losansky, R. Muller, A. Plock, D.J. Kliebenstein, U. Wittstock, The genetic basis of constitutive and herbivore-induced ESP-independent nitrile formation in Arabidopsis, Plant Physiol, 149 (2009) 561-574. [2] U. Wittstock, B.A. Halkier, Glucosinolate research in the Arabidopsis era, Trends Plant Sci, 7 (2002) 263-270. [3] B.A. Halkier, J. Gershenzon, Biology and biochemistry of glucosinolates, Annu Rev Plant Biol, 57 (2006) 303-333. [4] U. Wittstock, M. Burow, Tipping the scales--specifier proteins in glucosinolate hydrolysis, IUBMB Life, 59 (2007) 744-751. [5] W. Brandt, A. Backenkohler, E. Schulze, A. Plock, T. Herberg, E. Roese, U. Wittstock, Molecular models and mutational analyses of plant specifier proteins suggest active site residues and reaction mechanism, Plant Mol Biol, 84 (2014) 173-188. [6] K.D. Wittstock U, Lambrix V, Reichelt M, Gershenzon J, Glucosinolate hydrolysis and its impact on generalist and specialist insect herbivores., In: Romeo JT (ed) Integrative phytochemistry: from ethnobotany to molecular ecology. Else-vier, Amsterdam, pp 101-125, (2003). [7] L. Rask, E. Andreasson, B. Ekbom, S. Eriksson, B. Pontoppidan, J. Meijer, Myrosinase: gene family evolution and herbivore defense in Brassicaceae, Plant Mol Biol, 42 (2000) 93-113. [8] W. Zhang, W. Wang, Z. Liu, Y. Xie, H. Wang, Y. Mu, Y. Huang, Y. Feng, Crystal structure of the Epithiospecifier Protein, ESP from Arabidopsis thaliana provides insights into its product specificity, Biochem Biophys Res Commun, 478 (2016) 746-751. [9] J.C. Kuchernig, M. Burow, U. Wittstock, Evolution of specifier proteins in glucosinolate-containing plants, BMC Evol Biol, 12 (2012) 127. [10] V. Lambrix, M. Reichelt, T. Mitchell-Olds, D.J. Kliebenstein, J. Gershenzon, The Arabidopsis epithiospecifier protein promotes the hydrolysis of glucosinolates to nitriles and influences Trichoplusia ni herbivory, Plant Cell, 13 (2001) 2793-2807. [11] F. Gumz, J. Krausze, D. Eisenschmidt, A. Backenkohler, L. Barleben, W. Brandt, U. Wittstock, The crystal structure of the thiocyanate-forming protein from Thlaspi arvense, a kelch protein involved in glucosinolate breakdown, Plant Mol Biol, 89 (2015) 67-81. [12] A.M. Bones, J.T. Rossiter, The enzymic and chemically induced decomposition of glucosinolates, Phytochemistry, 67 (2006) 1053-1067. [13] H.L. Tookey, Crambe thioglucoside glucohydrolase (EC 3.2.3.1): separation of a protein required for epithiobutane formation, Can J Biochem, 51 (1973) 1654-1660. [14] M. Burow, A. Bergner, J. Gershenzon, U. Wittstock, Glucosinolate hydrolysis in Lepidium sativum--identification of the thiocyanate-forming protein, Plant Mol Biol, 63 (2007) 49-61. [15] U. Wittstock, K. Meier, F. Dorr, B.M. Ravindran, NSP-Dependent Simple Nitrile Formation Dominates upon Breakdown of Major Aliphatic Glucosinolates in Roots, Seeds, and Seedlings of Arabidopsis thaliana Columbia-0, Front Plant Sci, 7 (2016)
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
1821. [16] R. Kissen, A.M. Bones, Nitrile-specifier proteins involved in glucosinolate hydrolysis in Arabidopsis thaliana, J Biol Chem, 284 (2009) 12057-12070. [17] X.Y. Kong, R. Kissen, A.M. Bones, Characterization of recombinant nitrile-specifier proteins (NSPs) of Arabidopsis thaliana: dependency on Fe(II) ions and the effect of glucosinolate substrate and reaction conditions, Phytochemistry, 84 (2012) 7-17. [18] Z. Otwinowski, W. Minor, Processing of X-ray diffraction data collected in oscillation mode, Methods Enzymol., 276 (1997) 307-326. [19] The CCP4 suite: programs for protein crystallography, Acta Crystallogr D Biol Crystallogr, 50 (1994) 760-763. [20] A.J. McCoy, R.W. Grosse-Kunstleve, P.D. Adams, M.D. Winn, L.C. Storoni, R.J. Read, Phaser crystallographic software, J Appl Crystallogr, 40 (2007) 658-674. [21] P. Emsley, K. Cowtan, Coot: model-building tools for molecular graphics, Acta Crystallogr. D, 60 (2004) 2126-2132. [22] P.D. Adams, R.W. Grosse-Kunstleve, L.W. Hung, T.R. Ioerger, A.J. McCoy, N.W. Moriarty, R.J. Read, J.C. Sacchettini, N.K. Sauter, T.C. Terwilliger, PHENIX: building new software for automated crystallographic structure determination, Acta Crystallogr D Biol Crystallogr, 58 (2002) 1948-1954. [23] W.L. DeLano, The PyMOL molecular graphics system, Nature Publishing Group, a division of Macmillan Publishers Limited. All Rights Reserved., 2002. [24] W. Brandt, A. Backenköhler, E. Schulze, A. Plock, T. Herberg, E. Roese, U. Wittstock, Molecular models and mutational analyses of plant specifier proteins suggest active site residues and reaction mechanism, Plant molecular biology, 84 (2014) 173-188. [25] K.J. Bowers, E. Chow, H. Xu, R.O. Dror, M.P. Eastwood, B.A. Gregersen, J.L. Klepeis, I. Kolossvary, M.A. Moraes, F.D. Sacerdoti, Scalable algorithms for molecular dynamics simulations on commodity clusters, Proceedings of the 2006 ACM/IEEE conference on Supercomputing, ACM, 2006, pp. 84. [26] J.L. Banks, H.S. Beard, Y. Cao, A.E. Cho, W. Damm, R. Farid, A.K. Felts, T.A. Halgren, D.T. Mainz, J.R. Maple, Integrated modeling program, applied chemical theory (IMPACT), Journal of computational chemistry, 26 (2005) 1752-1780. [27] V. Kräutler, W.F. Van Gunsteren, P.H. Hünenberger, A fast SHAKE algorithm to solve distance constraint equations for small molecules in molecular dynamics simulations, Journal of computational chemistry, 22 (2001) 501-508. [28] S.-M. Peng, Y. Zhou, N. Huang, Improving the accuracy of pose prediction in molecular docking via structural filtering and conformational clustering, Chinese Chemical Letters, 24 (2013) 1001-1004. [29] N. Huang, C. Kalyanaraman, K. Bernacki, M.P. Jacobson, Molecular mechanics methods for predicting protein–ligand binding, Physical Chemistry Chemical Physics, 8 (2006) 5166-5177. [30] M.P. Jacobson, G.A. Kaminski, R.A. Friesner, C.S. Rapp, Force field validation using protein side chain prediction, Journal of Physical Chemistry B-Condensed Phase, 106 (2002) 11673-11680.
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
[31] P. Emsley, K. Cowtan, Coot: model-building tools for molecular graphics, Acta Crystallogr D Biol Crystallogr, 60 (2004) 2126-2132. [32] E. Krissinel, K. Henrick, Inference of macromolecular assemblies from crystalline state, J Mol Biol, 372 (2007) 774-797. [33] A.A. Jeyaprakash, A. Srivastav, A. Surolia, M. Vijayan, Structural basis for the carbohydrate specificities of artocarpin: variation in the length of a loop as a strategy for generating ligand specificity, J Mol Biol, 338 (2004) 757-770. [34] A. Rabijns, A. Barre, E.J. Van Damme, W.J. Peumans, C.J. De Ranter, P. Rouge, Structural analysis of the jacalin-related lectin MornigaM from the black mulberry (Morus nigra) in complex with mannose, FEBS J, 272 (2005) 3725-3732. [35] M. Gabrielsen, P.S. Abdul-Rahman, S. Othman, O.H. Hashim, R.J. Cogdell, Structures and binding specificity of galactose- and mannose-binding lectins from champedak: differences from jackfruit lectins, Acta Crystallogr F Struct Biol Commun, 70 (2014) 709-716.
ACCEPTED MANUSCRIPT Table 1 Data collection and refinement statistics of AtNSP1 AtNSP1 Data collection P3121
Unit Cell (Å)
120.902, 120.902, 123.006
Wavelength (Å)
0.9796
Resolution (Å)
3.022 (3.12-3.022)
Rmerge %
15.7 (89.9)
I/sigma
19.3 (3.5)
Completeness (%)
99.9 (100.0)
M AN U
SC
RI PT
Space Group
Redundancy
12.0 (12.4)
Wilson B factor (Å2)
66.38
Refinement R factor
0.2270 0.2672
No. atoms Overall B factors:
EP
RMSD bond lengths
TE D
Rfree
RMSD bond angles
3542 protein atoms 66.8 0.012 1.44
AC C
Ramachandran plot statistics (%) In preferred regions
92
In allowed regions
7.36
Outliers
0.64
PDB code
5GQT
Values in parentheses are for the highest resolution shell. Rmerge=ΣhΣi|Ih,i-Ih|/ΣhΣiIh,i, where Ih is the mean intensity of the i observations of symmetry related reflections of h. R=Σ|Fobs-Fcalc|/ΣFobs, where Fcalc is the calculated protein structure factor from the atomic model (Rfree was calculated with 5% of the reflections selected).
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
Figures
TE D
Fig. 1 Domain architecture of AtNSP1, AtESP, and TaTFP
(A) Domain architectures of AtNSP1, AtESP, and TaTFP proteins. (B) Sequence alignment of the jacalin-like lectin domains among AtNSP1, NSP homologs from
EP
Arabidopsis thaliana, and jacalin-like lectin domains with known 3D structures. 2JZ4 and 1VBO are the PDB codes of jacalin-like lectin domain proteins. 2JZ4 and 2JZ4* indicate the first and second jacalin domain of the protein. AtNSP4 and AtNSP4*
AC C
indicate the first and second jacalin domain of AtNSP4.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 2 Overall structure of AtNSP1
TE D
(A) Two perpendicular views of the AtNSP1 structure. The N-terminal jacalin-like lectin domain and C-terminal active domain are colored in cyan and magenta, respectively. (B) Two perpendicular views of the structural alignments of the active
EP
domains of AtNSP1 (magenta), AtESP (blue) and TaTFP (orange). The positions of
AC C
3L2 and 4L2 loops are indicated.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 3 Comparison of the active sites of AtNSP, AtESP and TaTFP (A) Superposition of the proposed active site residues in AtNSP1 (colored in magenta) and AtESP (colored in cyan). The residues are shown in sticks. (B) Superposition of the proposed active site residues in AtNSP1 (colored in magenta) and TaTFP (colored
AC C
EP
TE D
in orange). The residues are shown in sticks.
SC
RI PT
ACCEPTED MANUSCRIPT
M AN U
Fig. 4 Molecular docking studies of AtNSP1 and TaTFP with Fe2+ ions and allyglucosinolates
(A) The active site structure of AtNSP1 with docked Fe2+ (silver sphere) and the allyglucosinolate (yellow stick). The polar interactions are shown as red dashed lines. The water molecule is shown as a red sphere.
TE D
(B) The active site structure of AtESP with docked Fe2+ (silver sphere) and the allyglucosinolate (green stick). The polar interactions and the water molecule are shown as in (A).
EP
(C) Superimposition of the two structures in (A) and (B). Only the polar interactions
AC C
in (A) are shown.
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
The crystal structure of AtNSP1 was solved at 3.0 Å resolution. The AtNSP1 structure showed an N-terminal jacalin-like lectin domain. Molecular docking studies revealed the different conformation of R157 in AtNSP1. The study suggested the different product spectrum mechanism of specifier proteins.
S0006-291X(17)30873-2
DOI:
10.1016/j.bbrc.2017.05.027
Reference:
YBBRC 37742
To appear in:
Biochemical and Biophysical Research Communications
Received Date: 2 May 2017 Accepted Date: 3 May 2017
Please cite this article as: W. Zhang, Y. Zhou, K. Wang, Y. Dong, W. Wang, Y. Feng, Crystal structure of the nitrile-specifier protein NSP1 from Arabidopsis thaliana, Biochemical and Biophysical Research Communications (2017), doi: 10.1016/j.bbrc.2017.05.027. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
ACCEPTED MANUSCRIPT
G13 G27
T131 G149 D132 F150
V133 V151
TE D
M AN U
SC
H135 D153
RI PT
AtNSP1-N PDB: 1XXR
R292
AC C
EP
V287 G320 F337 G321
F271
E386 V370 D390 W432
H394
ACCEPTED MANUSCRIPT
Crystal Structure of the nitrile-specifier protein NSP1 from Arabidopsis thaliana Weiwei Zhanga,1, Yu Zhoub,1, Kan Wangc,1, Yanan Donga, Wenhe Wanga and Yue
a
RI PT
Fenga*
Beijing Key Lab of Bioprocess, Beijing University of Chemical Technology, Beijing
100029, PR China
National Institute of Biological Sciences, Beijing, No. 7 Science Park Road,
SC
b
Zhongguancun Life Science Park, Beijing 102206, PR China
Department of Anesthesiology, China-Japan Friendship Hospital, Beijing 100029, PR
M AN U
c
China. 1
These authors contributed equally to this work.
*
Abstract
TE D
Correspondence email: [email protected]
As important components of the glucosinolate-myrosinase system, specifier proteins mediate plant defense against herbivory and pathogen attacks. After tissue disruption,
EP
glucosinolates are hydrolyzed by myrosinases to instable aglucones, which will rearrange to form defensive isothiocyanates. Nevertheless, this reaction could be
AC C
redirected to form other products by specifier proteins. Up to now, identified specifier proteins include epithiospecifier proteins (ESPs), thiocyanate forming proteins (TFPs), and nitrile-specifier proteins (NSPs). Recently, the structures of ESP and TFP have been reported. However, both the structure and the catalytic mechanism of NSPs remain enigmatic. Here, we solved the crystal structure of the NSP1 protein from Arabidopsis thaliana (AtNSP1). Structural comparisons with ESP and TFP proteins revealed several structural features of AtNSP1 different from those of the two proteins. Subsequent molecular docking studies showed that the R292 residue in AtNSP1 displayed a conformation different from those of the corresponding residues in ESP
ACCEPTED MANUSCRIPT and TFP proteins, which might account for the product specificity and catalytic mechanism of AtNSP1. Taken together, the present study provides important insights into the molecular mechanisms underlying the different product spectrums between NSPs and the other two types of specifier proteins, and shed light on the future studies
RI PT
of the detailed mechanisms of other specifier proteins.
Keywords: nitrile-specifier protein, glucosinolate, Arabidopsis thaliana, crystal structure, active site
SC
1. Introduction
Glucosinolates are important plant secondary metabolites present in Brassicaceae
M AN U
plants such as the model plant Arabidopsis thaliana (A. thaliana)[1]. They are a group of thioglucosides, acting as components of an activated chemical defense system in plants [2]. The system is widespread in all the plants of the Brassicales order and involved in numerous plant-insect and plant-pathogen interactions [2, 3]. Upon tissue disruption, the system is activated, causing the release of glucosinolates and their
TE D
hydrolytic enzymes, the so called myrosinases, which are in separate storage compartments under normal conditions. The cleavage of the thioglucosidic bond of the glucosinolates by myrosinases produces aglucone intermediates, which are
EP
unstable and spontaneously rearrange to toxic isothiocyanates through a Lossen rearrangement [4-6]. Isothiocyanates are toxic to organisms including bacteria, fungi, nematodes and insects, and therefore act as defend compounds to protect plants from
AC C
herbivory and pathogen attack [7, 8]. In addition, isothiocyanates such as sulforaphane has also attracted a lot of interest as cancer-preventive components in the human diet. However, the outcome of glucosinolate hydrolysis actually depends on the side chain structure of the glucosinolate and the presence of supplementary proteins called specifier proteins [4, 9, 10]. In the presence of specifier proteins, the production of isothiocyanates can be diverted to the generation of nitriles, epithionitriles, or thiocyanates. Specifier proteins are believed to be enzymes, acting on the unstable aglucones, the hydrolysis products of glucosinolates [4, 9-11]. The nature of the generated product depends on the particular specifier protein, the side
ACCEPTED MANUSCRIPT chain structure of the glucosinolate, as well as the reaction conditions [12]. According to their product spectrum, specifier proteins are classified into epithiospecifier proteins (ESPs), thiocyanate forming proteins (TFPs), and nitrile-specifier proteins (NSPs). First isolated from Crambe abyssinica and named by
RI PT
Tookey [13], ESPs were the best characterized specifier proteins for their activity in promoting epithionitrile formation in the presence of allylglucosinolates and myrosinases. Epithionitrile formation requires the presence of a terminal double bond in the side chain of the aglucone substrate. TFPs are named by their ability to promote
SC
thiocyanate formation [14]. Two TFPs have been characterized, which are from Thlaspi arvense (TaTFP) and Lepidium sativum (LsTFP). Surprisingly, they are
M AN U
different in their substrate/product specificity. TaTFP and LsTFP can form thiocyanate only upon allylglucosinolate and benzylglucosinolate hydrolysis, respectively [4, 9, 11]. NSPs divert the myrosinase-catalyzed hydrolysis of a given glucosinolate from the formation of isothiocyanates to that of simple nitriles. Importantly, NSPs are responsible for constitutive and herbivore-induced simple nitrile formation in
TE D
Arabidopsis ecotype Columbia-0 rosette leaves [1]. Furthermore, Wittstock et al showed that simple nitrile formation after tissue disruption in A. thaliana depended almost entirely on NSP2 in seeds and mainly on NSP1 in seedlings [15]. Recently, we
EP
and others reported the crystal structures of AtESP and TaTFP respectively, which provided insights into the catalytic mechanisms and distinct product spectrums of the two proteins [8, 11]. However, both the structure and catalytic mechanism of NSPs
AC C
have not yet been resolved. Moreover, specifier proteins are well characterized by their four to five Kelch repeats, which form β-propeller tertiary structures [8, 11]. Nevertheless, apart from the conserved Kelch repeats, most NSP proteins also harbor one or two N-terminal jacalin-like lectin domains (Fig. 1) [1, 16, 17]. In contrast to ESP and TFP proteins, NSPs do not possess epithiospecifier activity although they present high sequence similarity to AtESP and TaTFP in the conserved Kelch repeats (Supplementary Fig. S1). Therefore, it is very interesting to investigate the mechanism underlying the product spectrum of NSP proteins. In the present study, we expressed and purified
ACCEPTED MANUSCRIPT AtNSP1 in Escherichia coli, and solved its crystal structure. Structural comparisons with AtESP and TaTFP revealed that AtNSP1 displayed several different structural features. Further molecular docking studies of AtNSP1 and AtESP with the allyglucosinolate substrate identified that the active site residue R292 of AtNSP1
RI PT
displayed an orientation different from the corresponding residues in AtESP and TaTFP, which might account for the product specificity and catalytic mechanism of AtNSP1. The present study will shed light on the molecular basis of the different product spectrums of specifier proteins and provide insights into the evolution of
SC
metabolic diversity in a complex plant defense system.
M AN U
2. Materials and methods 2.1. Materials
The pfu polymerase, the two restriction enzymes Nde I and Xho I and T4 DNA ligase were all purchased from New England Biolabs (NEB). Ni-NTA beads were purchased from QIAGEN and the Superdex-200 column was purchased from GE Healthcare.
TE D
2.2. Cloning, expression, purification and crystallization
The gene encoding AtNSP1 (At3g16400), was amplified by PCR from the cDNA library of A. thaliana. The amplified gene was inserted into pET22b (Novagen,
EP
Madison, WI), resulting in a plasmid encoding a His6 tag fusion (AAALEHHHHHH) at the C terminus of the recombinant AtNSP1. The protein was purified as previously reported [8] and concentrated to 25 mg/ml before use.
AC C
Crystallization screening was performed by the sitting-drop vapor-diffusion method at 291 K. 1 µL protein solution (25 mg/mL) was mixed with an equal volume of reservoir solution in 48-well plates and the drops were equilibrated against 80 µL reservoir solution. Crystals only appeared from one single condition, SaltRx 2 condition No. 45 [4.0 M ammonium acetate, 0.1 M BIS-TRIS propane pH 7.0]. Crystals of the best diffraction quality appeared in about 1 week from the condition containing 4.6 M Ammonium acetate, 0.1 M BIS-TRIS propane pH 7.0 and 0.5 M Ammonium chloride, equilibrated against 200 µL reservoir solution. These crystals were used for data collection.
ACCEPTED MANUSCRIPT 2.3. Data collection, processing and structure determination Crystals were cryoprotected in the reservoir solutions supplemented with 10 % (w/v) glycerol, flash cooled in the liquid nitrogen, and stored at cryogenic temperatures for data collection. X-ray diffraction data were collected on beamline BL17U of the
RI PT
Shanghai Synchrotron Radiation Facility (SSRF) using a Quantum 315r CCD detector (Area Detector Systems Corporation). All data were indexed, integrated and scaled with HKL-2000 [18]. Further processing was carried out using programs from the CCP4 suite [19]. Data collection statistics are summarized in Table 1. The molecular
SC
replacement was performed with PHASER [20] using the structures of Myrosinase Binding Protein At3g16450.1 from Arabidopsis thaliana (PDB: 2JZ4) and AtESP
M AN U
(PDB: 5GQ0) as models. The final model rebuilding was performed using COOT [21] and the protein structure was refined with PHENIX [22] using NCS and stereochemistry information as restraints. Structural figures were generated in PyMOL [23]. 2.4. Molecular docking
TE D
A Fe2+ was manually positioned with soft distance restraints between the proposed Fe2+-coordinating residues E386, D390 and H394 in AtNSP1, and E260, D264 and H268 in AtESP [24] at a distance of 2.1–2.4 Å. The system was then submitted to
EP
energy minimization and 20-ns molecular dynamic (MD) simulation routines. All MD simulations were performed using the Desmond software package [25] and the OPLS-AA 2005 force field [26]. A 10 Å buffered orthorhombic boundary system was
AC C
built with TIP3P water and then neutralized by counter ions. Each system was minimized using 5000 steps of steepest descent algorithm, followed by L-BGFS algorithm. Temperature was gradually increased to 300 K. A 50 kcal·mol−1·Å−2 harmonic position restraint was applied to all heavy atoms of the protein and Fe2+ atom during system equilibration. The restraints were gradually removed and the production run was processed in MTK-NPT (1 bar, 300 K) ensemble for 20 ns. The M-SHAKE algorithm [27] was applied to constrain all bonds involving hydrogen atoms with a time step of 2 fs. Molecular docking of AtNSP1 and AtESP was performed with UCSF DOCK 3.6
ACCEPTED MANUSCRIPT program [28]. For the ligand, 200 binding poses were generated and the submitted to a MM-GB/SA refinement and rescoring procedure [29], where the side chain of binding-site residues were sampled along with the docked ligand applying Protein Local Optimization Program (PLOP) [30]. The docked complex structure was
RI PT
minimized first, followed by the side chain prediction of the binding-site residues within 7 Å of the ligand, and the ligand was minimized with the fixed protein structure. 3. Results and discussion
SC
3.1. Expression, purification, crystallization and data collection
AtNSP1 was successfully cloned and expressed in E. coli and purified to homogeneity
M AN U
with a C-terminal His6 tag. The 52.9 kDa molecular weight of the monomeric His6 AtNSP1 predicted from the sequence was confirmed by 15 % SDS–PAGE (Supplementary Fig. S2). The retention volume of the main peak in the gel-filtration chromatography profile of the target protein indicated that it is a monomer in solution (Supplementary Fig. S2). The crystals of AtNSP1 with best diffraction power were
TE D
crystallized from 4.6 M Ammonium acetate, 0.1 M BIS-TRIS propane pH 7.0 and 0.5 M Ammonium chloride (Supplementary Fig. S3). The crystal of AtNSP1 diffracted to a maximum resolution of 3.02 Å (Table 1) and
EP
belonged to the space group P3121, with unit-cell parameters a = b = 120.902, c = 123.006 Å. An initial model for the AtNSP1 structure was obtained by molecular replacement via PHASER [20]. There is one molecule found in the asymmetric unit.
AC C
The correctness of the solution of PHASER with space group P3121 was clearly indicated by final highest Z-scores and log-likelihood gain (LLG) values for the molecule (RFZ=6.4, TFZ=12.2, LLG=207). The electron density was clear for the majority of the molecules, allowing subsequent refinement to be conveniently conducted. Several rounds of refinement in the resolution range 50.00–3.02 Å were conducted using restrained refinement by REFMAC5 and manual refinement by Coot [31]. 3.2. Overall structure and the N-terminal jacalin-like lectin domain of AtNSP1 Both the Protein Interfaces, Surfaces and Assemblies (PISA) service [32] and gel
ACCEPTED MANUSCRIPT filtration results indicated that AtNSP1 exists as a monomer in solution, consistent with the results of Gumz et al [11]. The structure clearly revealed the two domains of AtNSP1 (Fig. 2A). Similar as TaTFP and AtESP, NSP1 also exhibited a six-blade (β1-β6) β-propeller fold in its active domain (Fig. 2B), with a core root mean square
RI PT
deviation (RMSD) of 0.94 Å to the TaTFP structure among 319 Cα atoms and 0.91 Å to the AtESP structure among 317 Cα atoms (Fig. 2B). To keep consistency with the TaTFP and AtESP studies [5, 8], we used the same nomenclature to name the anti-parallel β strands within each blade (βx1-βx4), and the loops connecting βx2-βx3
SC
in each blade and adjacent blades in AtNSP1 (Supplementary Fig. S1).
Dali search with the the N-terminal jacalin-like lectin domain of AtNSP1 (designated
M AN U
as AtNSP1-N hereafter) returned several entries of lectin domains [33-35] with high structural similarities (Supplementary Fig. S4A-C). Lectins are multivalent carbohydrate-binding proteins of non-immune origin [33]. Jacalin is one of the two lectins from jackfruit (Artocarpus integrifolia) seeds, with galactose specificities. The so called jacalin-like lectin domain, however, binds mannose specifically. Similar as
TE D
other jacalin-like lectin domain, AtNSP1-N exhibited a classical β-prism fold, in which three antiparallel β-sheets, each consisting of four β-strands, shared an interface. Superimposition of the structure of the N-terminal domain of AtNSP1 with those of
EP
other lectin proteins revealed an RMSD from 1.55 to 1.61 Å among 137 Cα atoms (Supplementary Fig. S4A-C). The structure of AtNSP1-N differs from those of other lectin domains in three aspects. First, while most lectin domain proteins fold as
AC C
tetramers, AtNSP1 is a monomer in solution. Second, more than half of the loops in AtNSP1-N show different conformations in the other lectin proteins (Supplementary Fig. S4A-C). Third, the active site residues involved in sugar binding in other lectin domains are not fully conserved in AtNSP1-N (Supplementary Fig. S4E). Interestingly, AtNSP1-N showed both high amino acid sequence identity (over 50 %) and structural similarities to the jacalin domains of Arabidopsis myrosinase-binding proteins (MyroBPs), such as the product of gene At3g16450.1 (Supplementary Fig. S4D). It is worth investigating whether the N-terminal domains of NSP proteins bind mannose or other sugar or sugar derivatives, and whether this binding will affect the
ACCEPTED MANUSCRIPT activity of these proteins. 3.3. Active site structure Based on the structural and biochemical studies into AtESP and TaTFP proteins [5, 8, 11], the active sites of specifier proteins have been located at the bottom of the
RI PT
β-propeller structure. Structural alignments between AtNSP1, AtESP and TaTFP revealed that the loops 3L2 and 4L2 (connecting β32-β33 and β42-β43, respectively) are shorter in AtNSP1 than those in AtESP and TaTFP (Fig. 2B and Supplementary Fig. S1). Interestingly, Gumz et al. have observed a significant conformational change
SC
in loop 4L2, when they docked the aglucone substrate to the TaTFP structure [5]. This loop is also the secondary structural element showing the most remarkable difference
M AN U
between the structures of AtESP and TaTFP. Therefore, the shorter length of loop 4L2 in AtNSP1 might partially account for its different product spectrum. Furthermore, we compared the putative active sites of the three proteins. In spite of different product spectrums, AtNSP1 and AtESP did not show obvious distinctions in the position of their active site residues (Fig. 3A), except that two residues T129 and
TE D
P263 of AtESP were replaced by S270 and M389 in AtNSP1. More differences were observed between the structures of AtNSP1 and TaTFP (Fig. 3B). M186 and C250, which have been proposed to stabilize the substrate in TaTFP, were replaced by G321
EP
and V370 in AtNSP1, respectively. Moreover, R94 in TaTFP, which has been proposed to be involved in substrate sulfate binding, displayed apparent conformational differences from the corresponding residue R237 in AtNSP1. Further
AC C
comparison of the three structures together revealed two remarkable features of AtNSP1 different from those of AtESP and TaTFP. One is that the active site residue M389 is replaced by P263 and R269 in AtESP, and TaTFP, respectively. The other is that although R157, which was formerly proved to be important in both AtESP and TaTFP activity, showed almost the same conformation in these two proteins, the corresponding residue R292 of AtNSP1 displayed a different conformation (Fig. 3). 3.4. Molecular docking studies into AtNSP1 and AtESP Since AtNSP1 had an active site structure more similar to that of AtESP but a different product specificity when using allylglucosinolate as a substrate, we
ACCEPTED MANUSCRIPT performed molecular docking studies into AtNSP1 and AtESP with Fe2+ and allylglucosinolate. First, Fe2+ ions were manually positioned with soft distance restraints between the previously proposed Fe2+-coordinating residues E386, D390 and H394 in AtNSP1, and E260, D264, and H268 in AtESP [24] at a distance of
RI PT
2.1–2.4 Å, respectively. Both systems were then submitted to energy minimization and molecular dynamic (MD) simulation routines. After that, molecular dockings of AtNSP1 and AtESP were performed and the docked complex structures were energy minimized first, followed by side chain prediction of the binding-site residues within
SC
7 Å of the ligand, and the ligand was minimized with the fixed protein structure.
The docking process did not induce obvious conformational changes in the two
M AN U
proteins (data not shown), different from the docking studies into TaTFP by GumZ et al [11]. The docking results indicated that both the allylglucosinolate and Fe2+ were bound
at
similar
sites
of
AtESP
and
AtNSP1
(Fig.
4).
In
the
AtNSP1-allylglucosinolate structure, the Fe2+ ion was strongly coordinated by the sidechains of E386 and D390, and a water molecule. Importantly, the distance
TE D
between the Fe2+ and the NE2 atom of His394 was 3.9 Å, indicating very weak interaction between them. Together with R292, the sidechain of H394 formed four hydrogen bonds with the sulfate group of the substrate. Moreover, the rest of the
EP
aglucone body of the substrate formed hydrophobic interactions with F271, G320, G321, F337, V370 and W432.
In the AtESP-allylglucosinolate structure, the Fe2+ ion was coordinated by the
AC C
sidechains of E260, D264, and H268, which were used in the initial positioning of the Fe2+ ion in the docking studies. The sulfate group of the substrate formed three hydrogen bonds with the sidechain nitrogen atom of R157 and the mainchain nitrogen atom of G185. The substrate was also stabilized by the hydrophobic interactions from F130, G186, F202, V244 and W303. Interestingly, R94, which was previously shown to be involved in substrate sulfate binding in the docking results of TaTFP, has not been observed in substrate binding in either of the docked complex structures of AtNSP1 and AtESP. This is also consistent with the different conformations of TaTFPR94 from those of AtNSP1R237 and AtESPR94. Comparison of the two complex
ACCEPTED MANUSCRIPT structures revealed that R292 residue of AtNSP1 displays an orientation different from that of the corresponding R157 residue of AtESP. In addition, AtNSP1R292 also formed hydrogen bonds with the docked substrate with partially different atoms from those of AtESPR157. These indicated that this residue might account for the distinct
RI PT
product specificity of these two proteins. 4. Discussion
In spite of high sequence identities, AtNSP1 and AtESP could lead to different products, simple nitrile and epithionitrile, respectively, when using allylglucosinolate
SC
as substrates. It has been proposed that during the production of simple nitrile by NSP proteins, the negatively charged sulfur derived from the thioglucosidic bond could be
M AN U
abstracted from the aglucone by the Fe2+ ion, and the sulfur atom might be released as elemental sulfur at the end of the reaction due to the lack of a suitable acceptor [4]. However, in the presence of ESP proteins, the sulfur abstracted by the Fe2+ ion would further attack the terminal double bond resulting in formation of the thiirane ring through a currently unknown mechanism. We propose that R157, which displayed
TE D
similar conformations in AtESP and TaTFP but different conformation in AtNSP1, might play a role in the formation of epithionitrile. However, as aglucones are unstable chemicals, it is quite difficult to obtain the enzyme-substrate complex
EP
structures of specifier proteins. Therefore, future molecular dynamic simulations, together with mutagenesis studies based on the AtNSP1 structure reported here could further provide insights into the detailed mechanism underlying the different product
AC C
spectrums of specifier proteins.
Acknowledgements
We would like to thank the staff at beamline BL17U of the Shanghai Synchrotron Radiation Facility for their assistance with data collection. This work was supported by the National Nature Science Foundation of China (31400635 and 31670766) and the Fundamental Research Funds for the Central Universities (JD1709 and XK1701).
ACCEPTED MANUSCRIPT References
AC C
EP
TE D
M AN U
SC
RI PT
[1] M. Burow, A. Losansky, R. Muller, A. Plock, D.J. Kliebenstein, U. Wittstock, The genetic basis of constitutive and herbivore-induced ESP-independent nitrile formation in Arabidopsis, Plant Physiol, 149 (2009) 561-574. [2] U. Wittstock, B.A. Halkier, Glucosinolate research in the Arabidopsis era, Trends Plant Sci, 7 (2002) 263-270. [3] B.A. Halkier, J. Gershenzon, Biology and biochemistry of glucosinolates, Annu Rev Plant Biol, 57 (2006) 303-333. [4] U. Wittstock, M. Burow, Tipping the scales--specifier proteins in glucosinolate hydrolysis, IUBMB Life, 59 (2007) 744-751. [5] W. Brandt, A. Backenkohler, E. Schulze, A. Plock, T. Herberg, E. Roese, U. Wittstock, Molecular models and mutational analyses of plant specifier proteins suggest active site residues and reaction mechanism, Plant Mol Biol, 84 (2014) 173-188. [6] K.D. Wittstock U, Lambrix V, Reichelt M, Gershenzon J, Glucosinolate hydrolysis and its impact on generalist and specialist insect herbivores., In: Romeo JT (ed) Integrative phytochemistry: from ethnobotany to molecular ecology. Else-vier, Amsterdam, pp 101-125, (2003). [7] L. Rask, E. Andreasson, B. Ekbom, S. Eriksson, B. Pontoppidan, J. Meijer, Myrosinase: gene family evolution and herbivore defense in Brassicaceae, Plant Mol Biol, 42 (2000) 93-113. [8] W. Zhang, W. Wang, Z. Liu, Y. Xie, H. Wang, Y. Mu, Y. Huang, Y. Feng, Crystal structure of the Epithiospecifier Protein, ESP from Arabidopsis thaliana provides insights into its product specificity, Biochem Biophys Res Commun, 478 (2016) 746-751. [9] J.C. Kuchernig, M. Burow, U. Wittstock, Evolution of specifier proteins in glucosinolate-containing plants, BMC Evol Biol, 12 (2012) 127. [10] V. Lambrix, M. Reichelt, T. Mitchell-Olds, D.J. Kliebenstein, J. Gershenzon, The Arabidopsis epithiospecifier protein promotes the hydrolysis of glucosinolates to nitriles and influences Trichoplusia ni herbivory, Plant Cell, 13 (2001) 2793-2807. [11] F. Gumz, J. Krausze, D. Eisenschmidt, A. Backenkohler, L. Barleben, W. Brandt, U. Wittstock, The crystal structure of the thiocyanate-forming protein from Thlaspi arvense, a kelch protein involved in glucosinolate breakdown, Plant Mol Biol, 89 (2015) 67-81. [12] A.M. Bones, J.T. Rossiter, The enzymic and chemically induced decomposition of glucosinolates, Phytochemistry, 67 (2006) 1053-1067. [13] H.L. Tookey, Crambe thioglucoside glucohydrolase (EC 3.2.3.1): separation of a protein required for epithiobutane formation, Can J Biochem, 51 (1973) 1654-1660. [14] M. Burow, A. Bergner, J. Gershenzon, U. Wittstock, Glucosinolate hydrolysis in Lepidium sativum--identification of the thiocyanate-forming protein, Plant Mol Biol, 63 (2007) 49-61. [15] U. Wittstock, K. Meier, F. Dorr, B.M. Ravindran, NSP-Dependent Simple Nitrile Formation Dominates upon Breakdown of Major Aliphatic Glucosinolates in Roots, Seeds, and Seedlings of Arabidopsis thaliana Columbia-0, Front Plant Sci, 7 (2016)
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
1821. [16] R. Kissen, A.M. Bones, Nitrile-specifier proteins involved in glucosinolate hydrolysis in Arabidopsis thaliana, J Biol Chem, 284 (2009) 12057-12070. [17] X.Y. Kong, R. Kissen, A.M. Bones, Characterization of recombinant nitrile-specifier proteins (NSPs) of Arabidopsis thaliana: dependency on Fe(II) ions and the effect of glucosinolate substrate and reaction conditions, Phytochemistry, 84 (2012) 7-17. [18] Z. Otwinowski, W. Minor, Processing of X-ray diffraction data collected in oscillation mode, Methods Enzymol., 276 (1997) 307-326. [19] The CCP4 suite: programs for protein crystallography, Acta Crystallogr D Biol Crystallogr, 50 (1994) 760-763. [20] A.J. McCoy, R.W. Grosse-Kunstleve, P.D. Adams, M.D. Winn, L.C. Storoni, R.J. Read, Phaser crystallographic software, J Appl Crystallogr, 40 (2007) 658-674. [21] P. Emsley, K. Cowtan, Coot: model-building tools for molecular graphics, Acta Crystallogr. D, 60 (2004) 2126-2132. [22] P.D. Adams, R.W. Grosse-Kunstleve, L.W. Hung, T.R. Ioerger, A.J. McCoy, N.W. Moriarty, R.J. Read, J.C. Sacchettini, N.K. Sauter, T.C. Terwilliger, PHENIX: building new software for automated crystallographic structure determination, Acta Crystallogr D Biol Crystallogr, 58 (2002) 1948-1954. [23] W.L. DeLano, The PyMOL molecular graphics system, Nature Publishing Group, a division of Macmillan Publishers Limited. All Rights Reserved., 2002. [24] W. Brandt, A. Backenköhler, E. Schulze, A. Plock, T. Herberg, E. Roese, U. Wittstock, Molecular models and mutational analyses of plant specifier proteins suggest active site residues and reaction mechanism, Plant molecular biology, 84 (2014) 173-188. [25] K.J. Bowers, E. Chow, H. Xu, R.O. Dror, M.P. Eastwood, B.A. Gregersen, J.L. Klepeis, I. Kolossvary, M.A. Moraes, F.D. Sacerdoti, Scalable algorithms for molecular dynamics simulations on commodity clusters, Proceedings of the 2006 ACM/IEEE conference on Supercomputing, ACM, 2006, pp. 84. [26] J.L. Banks, H.S. Beard, Y. Cao, A.E. Cho, W. Damm, R. Farid, A.K. Felts, T.A. Halgren, D.T. Mainz, J.R. Maple, Integrated modeling program, applied chemical theory (IMPACT), Journal of computational chemistry, 26 (2005) 1752-1780. [27] V. Kräutler, W.F. Van Gunsteren, P.H. Hünenberger, A fast SHAKE algorithm to solve distance constraint equations for small molecules in molecular dynamics simulations, Journal of computational chemistry, 22 (2001) 501-508. [28] S.-M. Peng, Y. Zhou, N. Huang, Improving the accuracy of pose prediction in molecular docking via structural filtering and conformational clustering, Chinese Chemical Letters, 24 (2013) 1001-1004. [29] N. Huang, C. Kalyanaraman, K. Bernacki, M.P. Jacobson, Molecular mechanics methods for predicting protein–ligand binding, Physical Chemistry Chemical Physics, 8 (2006) 5166-5177. [30] M.P. Jacobson, G.A. Kaminski, R.A. Friesner, C.S. Rapp, Force field validation using protein side chain prediction, Journal of Physical Chemistry B-Condensed Phase, 106 (2002) 11673-11680.
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
[31] P. Emsley, K. Cowtan, Coot: model-building tools for molecular graphics, Acta Crystallogr D Biol Crystallogr, 60 (2004) 2126-2132. [32] E. Krissinel, K. Henrick, Inference of macromolecular assemblies from crystalline state, J Mol Biol, 372 (2007) 774-797. [33] A.A. Jeyaprakash, A. Srivastav, A. Surolia, M. Vijayan, Structural basis for the carbohydrate specificities of artocarpin: variation in the length of a loop as a strategy for generating ligand specificity, J Mol Biol, 338 (2004) 757-770. [34] A. Rabijns, A. Barre, E.J. Van Damme, W.J. Peumans, C.J. De Ranter, P. Rouge, Structural analysis of the jacalin-related lectin MornigaM from the black mulberry (Morus nigra) in complex with mannose, FEBS J, 272 (2005) 3725-3732. [35] M. Gabrielsen, P.S. Abdul-Rahman, S. Othman, O.H. Hashim, R.J. Cogdell, Structures and binding specificity of galactose- and mannose-binding lectins from champedak: differences from jackfruit lectins, Acta Crystallogr F Struct Biol Commun, 70 (2014) 709-716.
ACCEPTED MANUSCRIPT Table 1 Data collection and refinement statistics of AtNSP1 AtNSP1 Data collection P3121
Unit Cell (Å)
120.902, 120.902, 123.006
Wavelength (Å)
0.9796
Resolution (Å)
3.022 (3.12-3.022)
Rmerge %
15.7 (89.9)
I/sigma
19.3 (3.5)
Completeness (%)
99.9 (100.0)
M AN U
SC
RI PT
Space Group
Redundancy
12.0 (12.4)
Wilson B factor (Å2)
66.38
Refinement R factor
0.2270 0.2672
No. atoms Overall B factors:
EP
RMSD bond lengths
TE D
Rfree
RMSD bond angles
3542 protein atoms 66.8 0.012 1.44
AC C
Ramachandran plot statistics (%) In preferred regions
92
In allowed regions
7.36
Outliers
0.64
PDB code
5GQT
Values in parentheses are for the highest resolution shell. Rmerge=ΣhΣi|Ih,i-Ih|/ΣhΣiIh,i, where Ih is the mean intensity of the i observations of symmetry related reflections of h. R=Σ|Fobs-Fcalc|/ΣFobs, where Fcalc is the calculated protein structure factor from the atomic model (Rfree was calculated with 5% of the reflections selected).
ACCEPTED MANUSCRIPT
M AN U
SC
RI PT
Figures
TE D
Fig. 1 Domain architecture of AtNSP1, AtESP, and TaTFP
(A) Domain architectures of AtNSP1, AtESP, and TaTFP proteins. (B) Sequence alignment of the jacalin-like lectin domains among AtNSP1, NSP homologs from
EP
Arabidopsis thaliana, and jacalin-like lectin domains with known 3D structures. 2JZ4 and 1VBO are the PDB codes of jacalin-like lectin domain proteins. 2JZ4 and 2JZ4* indicate the first and second jacalin domain of the protein. AtNSP4 and AtNSP4*
AC C
indicate the first and second jacalin domain of AtNSP4.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 2 Overall structure of AtNSP1
TE D
(A) Two perpendicular views of the AtNSP1 structure. The N-terminal jacalin-like lectin domain and C-terminal active domain are colored in cyan and magenta, respectively. (B) Two perpendicular views of the structural alignments of the active
EP
domains of AtNSP1 (magenta), AtESP (blue) and TaTFP (orange). The positions of
AC C
3L2 and 4L2 loops are indicated.
M AN U
SC
RI PT
ACCEPTED MANUSCRIPT
Fig. 3 Comparison of the active sites of AtNSP, AtESP and TaTFP (A) Superposition of the proposed active site residues in AtNSP1 (colored in magenta) and AtESP (colored in cyan). The residues are shown in sticks. (B) Superposition of the proposed active site residues in AtNSP1 (colored in magenta) and TaTFP (colored
AC C
EP
TE D
in orange). The residues are shown in sticks.
SC
RI PT
ACCEPTED MANUSCRIPT
M AN U
Fig. 4 Molecular docking studies of AtNSP1 and TaTFP with Fe2+ ions and allyglucosinolates
(A) The active site structure of AtNSP1 with docked Fe2+ (silver sphere) and the allyglucosinolate (yellow stick). The polar interactions are shown as red dashed lines. The water molecule is shown as a red sphere.
TE D
(B) The active site structure of AtESP with docked Fe2+ (silver sphere) and the allyglucosinolate (green stick). The polar interactions and the water molecule are shown as in (A).
EP
(C) Superimposition of the two structures in (A) and (B). Only the polar interactions
AC C
in (A) are shown.
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
The crystal structure of AtNSP1 was solved at 3.0 Å resolution. The AtNSP1 structure showed an N-terminal jacalin-like lectin domain. Molecular docking studies revealed the different conformation of R157 in AtNSP1. The study suggested the different product spectrum mechanism of specifier proteins.