Crystal structure of the SPRY domain of human SPSB2 in the apo state

The apo structure of the SPRY domain of human SPRY domain-containing SOCS box protein 2 (SPSB2) determined at 1.9 Å resolution shows that its inducible nitric oxide synthase (iNOS)-binding site is highly preformed and that the C-terminal His₆ tag binds to a shallow pocket adjacent to the iNOS-binding site, which may help in structure-based and fragment-based SPSB2 inhibitor design.

Abstract

The SPRY domain-containing SOCS box protein 2 (SPSB2) is one of four mammalian SPSB proteins that are characterized by a C-terminal SOCS box and a central SPRY/B30.2 domain. SPSB2 interacts with inducible nitric oxide synthase (iNOS) via the SPRY domain and polyubiquitinates iNOS, resulting in its proteasomal degradation. Inhibitors that can disrupt SPSB2–iNOS inter­action and augment NO production may serve as novel anti-infective and anticancer agents. The previously determined murine SPSB2 structure may not reflect the true apo conformation of the iNOS-binding site. Here, the crystal structure of human SPSB2 SPRY domain in the apo state is reported at a resolution of 1.9 Å. Comparison of the apo and ligand-bound structures reveals that the iNOS-binding site is highly preformed and that major conformational changes do not occur upon ligand binding. Moreover, the C-terminal His₆ tag of the recombinant protein binds to a shallow pocket adjacent to the iNOS-binding site on a crystallographically related SPSB2 molecule. These findings may help in structure-based and fragment-based SPSB2 inhibitor design in the future.

1. Introduction  

The SPRY domain-containing SOCS (suppressor of cytokine signalling) box protein 2 (SPSB2; also named SSB-2 previously) is one of four mammalian SPSB proteins (SPSB1–SPSB4) that are characterized by a C-terminal SOCS box and a central protein–protein interaction module known as the SPRY/B30.2 domain (D’Cruz, Babon et al., 2013 ▸; Perfetto et al., 2013 ▸). Previously, Kuang and coworkers found that SPSB2 is a negative regulator of inducible nitric oxide synthase (iNOS) that modulates the lifetime of iNOS and controls cellular nitric oxide (NO) production (Kuang et al., 2010 ▸). SPSB2 utilizes the SPRY domain to interact with an N-terminal region of iNOS and uses the SOCS box to recruit an E3 ubiquitin ligase complex, resulting in the polyubiquitination and proteasomal degradation of iNOS (Kuang et al., 2010 ▸). Subsequently, it has been shown that SPSB1 and SPSB4 also polyubiquitinate iNOS and regulate the proteasomal degradation of iNOS (Lewis et al., 2011 ▸; Matsumoto et al., 2011 ▸; Nishiya et al., 2011 ▸; Wang, Luo et al., 2018 ▸). SPSB2-deficient macrophages showed significantly increased iNOS activity and NO production, and were more active in killing Leishmania major parasites; therefore, inhibitors that can disrupt the endogenous SPSB2–iNOS interaction and augment NO production may serve as novel anti-infective and anticancer agents (Kuang et al., 2010 ▸). Indeed, since this original discovery, a range of inhibitors aiming to disrupt SPSB2–iNOS interaction have been developed by different groups (Yap et al., 2014 ▸, 2016 ▸; Leung et al., 2014 ▸; Harjani et al., 2016 ▸; You et al., 2017 ▸; Sadek et al., 2018 ▸).

With respect to structural studies, solution (Yao et al., 2005 ▸; Masters et al., 2006 ▸) and crystal (Kuang et al., 2009 ▸) structures of the murine SPSB2 SPRY domain alone, as well as crystal structures of the human SPSB2 SPRY domain in complex with a peptide from the Drosophila DEAD-box RNA helicase VASA (Filippakopoulos et al., 2010 ▸), with a rationally designed RGD-containing cyclic peptide inhibitor (You et al., 2017 ▸) and with cyclic heptapeptides (Sadek et al., 2018 ▸) have been reported. Crystal structures have also been reported for GUSTAVUS (a Drosophila orthologue of SPSB proteins) in complex with elongins B and C (Woo, Imm et al., 2006 ▸) or with a VASA peptide (Woo, Suh et al., 2006 ▸). SPRY domains are found in many different proteins, including some tripartite-motif (TRIM) proteins and some RING-finger (RNF) proteins. SPRY domain structures have been determined for TRIM21 (Keeble et al., 2008 ▸), TRIM72 (Park et al., 2010 ▸), TRIM5α (Biris et al., 2012 ▸; Yang et al., 2012 ▸), TRIM25 (D’Cruz, Kershaw et al., 2013 ▸; Koliopoulos et al., 2018 ▸) and TRIM20 (Weinert et al., 2015 ▸), and solution NMR has been used to investigate the SPRY domains of TRIM25 (Kong et al., 2015 ▸) and RNF135 (Zhang et al., 2019 ▸).

It is worth noting that in the crystal structure of ligand-free murine SPSB2 the iNOS-binding site was occupied by residues from crystallographically related molecules (Kuang et al., 2009 ▸), raising the question as to whether this structure reflects the true apo conformation of the iNOS-binding site. On the other hand, an apo crystal structure of human SPSB2, without a ligand bound to its iNOS-binding site, has not been determined. Therefore, the extent to which the iNOS-binding site undergoes conformational changes upon binding to iNOS or related inhibitors remains unclear. Here, we report the crystal structure of the human SPSB2 SPRY domain in the apo state at a resolution of 1.9 Å. Comparison of the apo and ligand-bound structures revealed that the iNOS-binding site is highly preformed and that major conformational changes do not occur upon ligand binding. Interestingly, the C-terminal His₆ tag of the recombinant protein binds to a shallow pocket adjacent to the iNOS-binding site on a crystallographically related SPSB2 molecule. These findings may help in structure-based and fragment-based SPSB2 inhibitor design in the future.

2. Materials and methods  

2.1. Macromolecule production  

Recombinant SPRY domain (residues 22–220) of human SPSB2 (UniProtKB accession No. Q99619) was expressed and purified using a method similar to that described previously (You et al., 2017 ▸). Briefly, a cDNA fragment encoding the SPRY domain with a C-terminal His₆ tag was constructed into the pET-15b vector. The recombinant protein was expressed in Escherichia coli BL21 (DE3) cells grown in Luria–Bertani medium. The protein was expressed at 18°C for 16 h after induction with 0.1 mM isopropyl β-d-1-thiogalactopyranoside when the OD₆₀₀ reached 0.6–0.8. The cells were harvested by centrifugation and were lysed using a cell disruptor (JNBIO). The protein was initially purified using an Ni–NTA column (GE Healthcare) and was further purified by gel filtration using a HiLoad 16/600 Superdex 75 column (GE Healthcare) equilibrated against buffer consisting of 50 mM Tris–HCl pH 7.5, 150 mM NaCl, 5 mM DTT. The protein was concentrated using Amicon centrifugal filter units (Merck Millipore). Macromolecule-production information is summarized in Table 1 ▸.

Table 1. Macromolecule-production information.

Source organism	Homo sapiens
DNA source	Synthetic gene
Forward primer	TTTTTTCCATGGGCGACCTCTCCTGTCCC
Reverse primer	TTTTTTCTCGAGTTATCAATGGTGATGATGGTGATGAGAACCCCTTTCGCCCAGGTAG
Cloning vector	pET-15b
Expression vector	pET-15b
Expression host	E. coli BL21 (DE3)
Complete amino-acid sequence of the construct produced†	MGDLSCPEGLEELLSAPPPDLGAQRRHGWNPKDCSENIEVKEGGLYFERRPVAQSTDGARGKRGYSRGLHAWEISWPLEQRGTHAVVGVATALAPLQTDHYAALLGSNSESWGWDIGRGKLYHQSKGPGAPQYPAGTQGEQLEVPERLLVVLDMEEGTLGYAIGGTYLGPAFRGLKGRTLYPAVSAVWGQCQVRIRYLGERGSHHHHHH

^†The underlined sequences are the N-terminal cloning artifacts and the C-terminal His₆ tag.

2.2. Crystallization  

Purified protein samples at concentrations of 4, 6 and 12 mg ml⁻¹ were centrifuged at 15 700g for 10 min to clarify the solution prior to setting up crystallization trials. Crystallization trials were carried out using the hanging-drop vapour-diffusion method at 16°C. Initial crystallization trials were performed using commercial sparse-matrix screens (PEG/Ion, PEG/Ion 2 and Index from Hampton Research). The crystals used for data collection and structure determination were obtained from a solution consisting of equal volumes (1 µl) of 4 mg ml⁻¹ protein solution and reservoir solution consisting of 0.2 M ammonium acetate, 0.1 M Bis-Tris pH 5.5, 25%(w/v) polyethylene glycol 3350. Crystallization information is summarized in Table 2 ▸.

Table 2. Crystallization.

Method	Hanging-drop vapour diffusion
Plate type	24-well plates
Temperature (K)	289
Protein concentration (mg ml−1)	4
Buffer composition of protein solution	50 mM Tris–HCl pH 7.5, 150 mM NaCl, 5 mM DTT
Composition of reservoir solution	0.2 M ammonium acetate, 0.1 M Bis-Tris pH 5.5, 25%(w/v) polyethylene glycol 3350
Volume and ratio of drop (µl)	1:1
Volume of reservoir (ml)	0.5

2.3. Data collection and processing  

Prior to data collection, crystals were transferred into a cryoprotectant solution consisting of the reservoir solution supplemented with 25%(v/v) glycerol and were immediately flash-cooled in liquid nitrogen. A 1.9 Å resolution X-ray diffraction data set was collected from a single crystal on beamline BL17U (Wang, Zhang et al., 2018 ▸) at the Shanghai Synchrotron Radiation Facility (SSRF), Shanghai, People’s Republic of China at a wavelength of 0.9792 Å at 100 K. The data were processed using the XDS package (Kabsch, 2010 ▸) and the xia2 system (Winter, 2010 ▸) with the CCP4 package (Winn et al., 2011 ▸). Scaling was performed with AIMLESS (Evans, 2011 ▸). Detailed statistics for data collection and processing are shown in Table 3 ▸.

Table 3. Data collection and processing.

Values in parentheses are for the outer shell.

Diffraction source	BL17U, SSRF
Wavelength (Å)	0.9792
Temperature (K)	100
Detector	EIGER 16M
Crystal-to-detector distance (mm)	250
Rotation range per image (°)	1
Total rotation range (°)	360
Exposure time per image (s)	0.2
Space group	P1211
a, b, c (Å)	32.45, 75.10, 79.20
α, β, γ (°)	90, 101.38, 90
Mosaicity (°)	0.3
Resolution range (Å)	34.49–1.90 (1.95–1.90)
Total No. of reflections	96555 (6709)
No. of unique reflections	28566 (2029)
Completeness (%)	97.3 (95.1)
Multiplicity	3.4 (3.3)
〈I/σ(I)〉	7.6 (2.9)
R meas	0.184 (0.813)
Overall B factor from Wilson plot (Å2)	13.5

2.4. Structure solution and refinement  

The structure was determined by molecular replacement with MOLREP (Vagin & Teplyakov, 2010 ▸) using the crystal structure of SPSB2 in complex with a VASA peptide (PDB entry 3emw; Filippakopoulos et al., 2010 ▸). The structural model was built in Coot (Emsley & Cowtan, 2004 ▸) and refined using REFMAC5 (Murshudov et al., 2011 ▸). The coordinates have been deposited in the Protein Data Bank with PDB code 6jkj. Structure-solution and refinement statistics are provided in Table 4 ▸. Structural figures were prepared using the PyMOL molecular-graphics system (Schrödinger).

Table 4. Structure solution and refinement.

Values in parentheses are for the outer shell.

PDB code	6jkj
Resolution range (Å)	34.49–1.90 (1.95–1.90)
Completeness (%)	97.1 (94.9)
σ Cutoff	None
No. of reflections, working set	27047 (1920)
No. of reflections, test set	1502 (104)
Final R cryst	0.217 (0.421)
Final R free	0.270 (0.449)
No. of non-H atoms
Protein	2933
Water	132
Total	3065
R.m.s. deviations
Bonds (Å)	0.009
Angles (°)	1.555
Average B factors (Å2)
Protein	20.98
Water	24.28
Clashscore	4
Ramachandran plot
Most favoured (%)	97.1
Allowed (%)	2.9
Outliers (%)	0

3. Results and discussion  

3.1. Apo structure of the human SPSB2 SPRY domain  

The recombinant human SPSB2 SPRY domain contained SPSB2 residues 22–220 plus two cloning-artifact residues (MG) at the N-terminus and two linker residues followed by a His₆ tag (GSHHHHHH) at the C-terminus. The protein crystallized in space group P12₁1 and the structure was determined at 1.9 Å resolution (PDB entry 6jkj). There are two molecules in the asymmetric unit, with a Matthews co­efficient of 2.07 ų Da⁻¹ and a solvent content of 40.6% (Matthews, 1968 ▸). A total of 195 and 199 of the 209 residues were modelled for chains A and B, respectively. Because of a lack of continuous density, the four N-terminal residues, the 145–148 region and seven C-terminal residues of chain A were not modelled, whereas in chain B the six N-terminal residues, the 146–148 region and Gly158 were not modelled. Notably, the C-terminal His₆ tag is disordered in chain A but is structured in chain B and binds to a crystallographically related molecule (see below). Apart from this there are no significant structural differences between chain A and chain B; they can be superimposed very well with a root-mean-square deviation (r.m.s.d.) of 0.14 Å for 154 C^α atoms. Overall, the structure exhibits good geometry, with 97.1% of the residues in the most favoured region of the Ramachandran plot, 2.9% in the allowed region and no outliers. Consistent with previously reported crystal structures (Kuang et al., 2009 ▸; Filippakopoulos et al., 2010 ▸), the SPRY domain of SPSB2 consists of two short helices packed against a bent β-sandwich that is formed by two β-sheets wrapped around a long loop extending from the central strands of the concave β-sheet (Fig. 1 ▸). The central region of this loop participates in iNOS binding, and Tyr120 on the tip of the loop is indispensable (Kuang et al., 2010 ▸). The opposite side of the domain close to the N-terminal helix is known to form the support for the C-terminal SOCS box (not included in this construct) that interacts with elongins B and C to recruit an ubiquitin ligase complex (Woo, Imm et al., 2006 ▸).

Figure 1.

The crystal structure of the human SPSB2 SPRY domain in the apo state. The structure reported here (PDB entry 6jkj, chain A) is shown as a cartoon model; α-helices, β-strands and loops are coloured red, yellow and green, respectively. The N- and C-termini and the iNOS-binding site are labelled. The 145–148 region that is missing in the structure is indicated by a black dashed line.

The SPRY domains of human and murine SPSB2 share 92% amino-acid sequence identity (data not shown). The apo structure of the human SPSB2 SPRY domain closely matches that of murine SPSB2, with an r.m.s.d. of 0.34 Å for 163 C^α atoms (Fig. 2 ▸ a). Nevertheless, large structural differences are seen in the 153–163 region (upper right in Fig. 2 ▸ a). It has been shown that the corresponding regions in SPRY/B30.2 domains from different proteins have large-amplitude motions in solution and adopt different conformations (Yao et al., 2006 ▸). To reveal the possible conformational changes that the iNOS-binding site undergoes upon ligand binding, the apo structure reported here was superimposed on the structure of the human SPSB2 SPRY domain in complex with a cyclic peptide inhibitor, cR8, which was designed to mimic the bound conformations of iNOS residues (You et al., 2017 ▸). The apo and inhibitor-bound structures aligned very well, including the 153–163 region, with an r.m.s.d. of 0.16 Å for 163 C^α atoms (Fig. 2 ▸ b). Surprisingly, when the iNOS-binding residues in these three structures were analyzed in detail, we found that the side chains of several key iNOS-binding residues possess different conformations in the human and murine ligand-free structures (Fig. 2 ▸ c). Specifically, the loop residues Ala205–Gln211 adopt different positions and the side chains of Trp207 and Gln209 have different orientations (Fig. 2 ▸ c). On the other hand, the side-chain conformations of iNOS-binding residues are essentially identical in both the apo and inhibitor-bound human structures, including Arg68, Ala72, Thr102, Tyr120, Val206, Trp207 and Gln209 (Fig. 2 ▸ d). Therefore, these results indicate that the iNOS-binding site is highly preformed, at least in the crystals, and major conformational changes do not occur upon ligand binding. Our findings that the human apo structure superimposes better with the human ligand-bound structure than with the murine ‘apo’ structure were somewhat unexpected, but can be reconciled by the fact that the iNOS-binding site in the murine ‘apo’ structure was actually occupied by some amide side chains from crystallographically related molecules (Kuang et al., 2009 ▸). Thus, we conclude that the apo structure of the human SPSB2 SPRY domain reported here is a better model to represent its apo state.

Figure 2.

Comparison of apo and ligand-bound structures of the SPSB2 SPRY domain. The apo structure of human SPSB2 reported here (green; PDB entry 6jkj; chain A) is superimposed with (a, c) murine ligand-free SPSB2 (gold; PDB entry 3ek9; Kuang et al., 2009 ▸) and (b, d) human ligand-bound SPSB2 (magenta; PDB entry 5xn3; You et al., 2017 ▸). The iNOS-binding site is boxed in (a) and (b) and is expanded in (c) and (d), with the side chains that participate in iNOS binding shown and labelled. Murine SPSB2 residues 145, 149, 154 and 158 are labelled in (a) to indicate the 146–148 and 155–157 regions that are missing from the structure.

3.2. Binding of the His₆ tag to the crystallographically related molecule  

Interestingly, in the structure reported here the C-terminal His₆ tag (including two preceding linker residues) of chain B binds to a crystallographically related SPSB2 molecule (denoted chain B′), while in chain A this C-terminal tail is disordered (Fig. 3 ▸ a). The C-terminal tail of chain B shows continuous main-chain electron density and clear side-chain electron density (Supplementary Fig. S1). When chain B′ is superimposed on the SPSB2–cR8 complex structure (You et al., 2017 ▸) it is evident that the His₆ tag of chain B binds to a region on SPSB2 that is adjacent to the iNOS-binding site occupied by the inhibitor cR8 (Fig. 3 ▸ b). Because this structural information may aid in the structure-based design of SPSB2 inhibitors, we have analyzed the interactions in detail. Hence, the hexahistidine residues are numbered His1–His6 and the two linker residues connecting the SPRY domain and the hexahistidines are numbered Gly−2 and Ser−1 accordingly. The C-terminal His₆ tag of chain B forms a short 3₁₀-helix-like structure at Ser−1 to His3 and has an extended conformation at His4 to His6 (Fig. 3 ▸ b). The His₆ tag interacts with SPSB2 residues located in three loops, i.e. loop Cys51–Ile57, loop Arg68–Asp76 and Gln116–Asp118 in the aforementioned central long loop (Fig. 3 ▸ b), and occupies a shallow pocket on SPSB2 that is negatively charged with significant shape complementarity (Fig. 3 ▸ c). Since the crystal was obtained using a slightly acidic reservoir solution (pH 5.5), the hexahistidine residues may have been protonated and attracted to this negatively charged pocket. Some hydrogen bonds are formed between the C-terminal His₆ tag and the crystallo­graphically related SPSB2 molecule (Fig. 3 ▸ d). The carbonyl group of Ser−1 forms hydrogen bonds to the side chain of Thr117 and the main-chain N atom of Asp118, while the side chain of Ser−1 forms a hydrogen bond to the side chain of Asp118. The main-chain N atom of His2 forms a hydrogen bond to the side chain of Gln116, whereas the main-chain N atoms of both His3 and His4 form hydrogen bonds to the side chain of Asp118. In addition, the His4 carbonyl group and the side chain of His6 form hydrogen bonds to the side chain of Asn56 and to Pro70, respectively. Nevertheless, the interaction is likely to be weak, firstly because the recombinant protein eluted as a monomer of ∼23 kDa on a gel-filtration column (data not shown) and secondly because only half of the molecules in the crystal are involved in binding the His₆ tag. We then proceeded to use isothermal titration calorimetry (ITC) to measure the binding affinity. An octameric peptide (Ac-GSHHHHHH) was synthesized by Sangon Biotech, Shanghai, People’s Republic of China. The purity of the peptide was ∼95% and its observed mass (1026.90 Da) was consistent with its theoretical mass (1026.98 Da) (Supplementary Fig. S2). ITC measurements were carried out in buffers at pH 7.5 and 5.5, both containing 150 mM NaCl to mimic the physiological ionic strength. Solutions of 10 µM protein in the cell were titrated by the injection of 500–1000 µM peptide. We found that the binding affinity was too low to be determined under these conditions (Supplementary Fig. S3).

Figure 3.

Binding of the C-terminal His₆ tag to the crystallographically related SPSB2 molecule. (a) Cartoon representation of the structure (PDB entry 6jkj) showing that the His₆ tag (including two preceding linker residues) of chain B binds to the crystallographically related chain B′. The His₆ tag of chain A is missing from the structure. Chains A and B are coloured green and cyan, respectively; the His₆ tag of chain B is coloured magenta. Chains A′ and B′ are coloured grey. (b) The His₆ tag binds to a region adjacent to the iNOS-binding site. Chain B′ (grey) is superimposed with the structure of the SPSB2–cR8 complex (gold; PDB entry 5xn3). The main chain and side chains of the His₆ tag of chain B are coloured magenta and light pink, respectively. (c) The His₆ tag binds to a shallow pocket on SPSB2. Chain B′ is shown as a surface model with electrostatic potentials calculated by PyMOL. (d) Stereoview of stick models showing hydrogen bonding between the His₆ tag and SPSB2. SPSB2 (chain B′) residues are coloured grey; N and O atoms are coloured blue and red, respectively. Hydrogen bonds are indicated by yellow dashed lines.

Since inhibitors that can disrupt intracellular SPSB2–iNOS interaction have potential as novel anti-infective and anti­cancer agents (Kuang et al., 2010 ▸), some molecules of this type have recently been developed by different groups through structure-based (Yap et al., 2014 ▸, 2016 ▸; Harjani et al., 2016 ▸; You et al., 2017 ▸; Sadek et al., 2018 ▸) and fragment-based (Leung et al., 2014 ▸; Norton et al., 2016 ▸) approaches. One strategy to further improve their binding affinities is the linking of the molecule to a compound that can bind to an adjacent binding site on the target protein using a suitable linker (Shuker et al., 1996 ▸; Hajduk & Greer, 2007 ▸). The novel structure of the His₆ tag bound to a binding pocket adjacent to the iNOS-binding site on human SPSB2 may be therefore of value for developing potent SPSB2 inhibitors in the future.

Supplementary Material

PDB reference: human SPSB2 in the apo state, 6jkj

Funding Statement

This work was funded by National Natural Science Foundation of China grants 81571539 and 31270817. Fundamental Research Funds for the Central Universities grant 21617443.