Crystallization and crystallographic studies of a novel chickpea 11S globulin

The first structural report of CL-AI from chickpea seeds reveals a trimeric arrangement and structural features that are conserved among 11S globulins.

Abstract

Chickpea is a crop that is known as a source of high-quality proteins. CL-AI, which belongs to the 11S globulin and cupin superfamily, was initially identified in chickpea seeds. CL-AI has recently been shown to inhibit various types of α-amylases. To determine its molecular mechanism, the crystal structure of CL-AI was solved at a final resolution of 2.2 Å. Structural analysis indicated that each asymmetric unit contains three molecules with threefold symmetry and a head-to-tail association, and each molecule is divided into an α-chain and a β-chain. CL-AI has high structural similarity to other 11S globulins and canonical metal-dependent enzyme-related cupin proteins, whereas its stimilarity to α-amylase inhibitor from Phaseolus vulgaris is quite low. The structure presented here will provide insight into the function of CL-AI.

1. Introduction

Chickpea (Cicer arietinum L.) is the third most important pulse crop in the world, and mainly grows in arid and semi-arid regions. Chickpeas are a nutrient-dense food and provide a rich content of protein, dietary fiber, folate and certain dietary minerals such as iron and phosphorus (El-Adawy, 2002 ▸; Jukanti et al., 2012 ▸). According to classification by the World Food and Agriculture Organization, chickpeas are rich in essential amino acids such as lysine, isoleucine and tryptophan (Milán-Carrillo et al., 2007 ▸).

The main proteins in chickpea seeds belong to the storage protein family; these proteins are hydrolyzed during germination and nourish the early stages of seedling growth (Czubinski et al., 2015 ▸). 11S globulins are the major storage globulins in chickpea seeds. In legumes, they constitute about half of the total seed proteins (Utsumi, 1992 ▸). 11S globulins are synthesized in the endoplasmic reticulum, where they undergo signal peptide cleavage. They are then sorted into different organelles for further processing. Finally, they are packed and stored in protein storage vacuoles until seed germination (Tandang-Silvas et al., 2010 ▸).

The cupin superfamily of proteins possess remarkable functional diversity, with representatives found in archaea, eubacteria and eukaryota (Dunwell et al., 2004 ▸). The identification of the cupin superfamily was originally based on the recognition that the wheat protein germin shared a nine-amino-acid sequence with another protein, spherulin, produced by the slime mold Physarum polycephalun during starvation (Dunwell & Gane, 1998 ▸). Knowledge of the three-dimensional structures of these proteins led to the collective name ‘cupin’ on the basis of their β-barrel cores (‘cupa’ is the Latin term for a small barrel or cask; Dunwell, 1998 ▸). The cupin domain was originally described as two conserved motifs, each composed of two β-strands (Dunwell et al., 2001 ▸). Motif 1 was designated as G(X)₅HXH(X)_3,4E(X)₆G and motif 2 as G(X)₅PXG(X)₂H(X)₃N (Dunwell et al., 2004 ▸). The two histidine residues and the glutamate residue in motif 1, together with the histidine residue in motif 2, act as ligands to bind the active-site metal ion in the cupin domain. It is now clear that the primary sequence of these two motifs is much less conserved than was first suggested (Dunwell et al., 2004 ▸; Hao et al., 2009 ▸). The cupin superfamily is now extremely diverse and extends to catalytically inactive seed-storage and sugar-binding metal-independent proteins as well as metal-dependent enzymes based on structure similarity.

CL-AI (an α-amylase inhibitor from C. arietinum L.), which was initially purified from chickpea seeds, is classified into the 11S globulin family of the cupin superfamily. Ma and co­workers demonstrated that the α chain and β chain of CL-AI, and CL-AI in its single-stranded state, could inhibit various types of α-amylases. They deemed that the cupin domains of CL-AI were involved in its inhibitory activity (Hao et al., 2009 ▸; Wang et al., 2018 ▸). According to our knowledge, this is the first report of α-amylase inhibitory activity of an 11S globulin. To explain the structural features and inhibition mode of CL-AI, it was necessary to determine its crystal structure.

Here, we determined the structure of CL-AI at 2.2 Å resolution, which provides a structural basis for understanding the inhibitory activity of CL-AI.

2. Materials and methods

2.1. Macromolecule production

cDNA of CL-AI was cloned into the pSUMO3 expression vector (Addgene) and overexpressed as described previously (Lin et al., 2015 ▸). Briefly, the target gene was induced by 0.2 mM isopropyl β-d-1-thiogalactopyranoside (IPTG) when the cell density reached an optical density at 600 nm (OD₆₀₀) of approximately 0.6 in 2×YT medium. After growth for 16 h at 298 K, the cells were collected and resuspended in lysis buffer consisting of 10 mM Tris–HCl pH 7.5, 0.5 M NaCl, 20 mM imidazole. After disruption by sonication, cell debris was removed by centrifugation at 20 000g for 30 min at 280 K. The supernatant was applied onto a HisTrap Ni-Sepharose column (GE Healthcare) and washed three times with 150 ml lysis buffer. Subsequently, the target protein was eluted from the column using a 20–200 mM imidazole gradient in lysis buffer. Fractions containing the fusion protein were pooled and then digested with SENP2 protease at a molar ratio of 1:100 to remove the SUMO3 tag from the fusion protein. A HiTrap SP column (GE Healthcare) was used for further purification. Elution was performed with a linear gradient from 0.1 to 0.5 M NaCl. The fractions containing the target protein were concentrated and subjected to gel-filtration chromatography (HiLoad Superdex 200 16/600, GE Healthcare) in a buffer consisting of 10 mM Tris–HCl pH 7.4, 0.1 M NaCl. Macromolecule-production information is summarized in Table 1 ▸.

Table 1. Macromolecule-production information.

Source organism	Cicer arietinum
Expression vector	pSUMO3
Expression host	Escherichia coli

2.2. Crystallization

CL-AI was concentrated to about 10 mg ml⁻¹ for crystallization. The PEG/Ion, PEG/Ion 2 and Crystal Screen kits from Hampton Research and the Wizard 1 and Wizard 2 kits from Emerald BioStructures were used for crystal screening at 298 K using a Mosquito robot (TTP Labtech). A Rock Imager (Formulatrix) was used to observe the growth process of the crystals. Crystals that had been scaled up in 24-well sitting-drop plates were used for diffraction. Uniform cube-shaped crystals (about 100 × 100 × 100 µm) were obtained after one week in a condition consisting of 20% polyethylene glycol monomethyl ether 2000 (PEG MME 2000), 0.1 M Tris pH 7.0, which differed from the crystallization conditions for homologous proteins, using the sitting-drop vapor-diffusion method. Crystallization information is summarized in Table 2 ▸.

Table 2. Crystallization.

Method	Vapor diffusion, sitting drop
Temperature (K)	298
Protein concentration (mg ml−1)	10
Buffer composition of protein solution	10 mM Tris pH 7.4, 150 mM NaCl
Composition of reservoir solution	20% PEG MME 2000, 0.1 M Tris pH 7.0

2.3. Data collection and processing

Crystals were cryoprotected in 20% PEG MME 2000, 0.1 M Tris pH 7.0, 20%(v/v) glycerol before plunging them into liquid nitrogen. Diffraction data were collected on the BL17U beamline at Shanghai Synchrotron Radiation Facility (SSRF) with a beam size of around 80 µm (horizontal) × 45 µm (vertical) FWHM and a flux of around 4.1 × 10¹² photons s⁻¹ at 11.5 keV (Wang et al., 2017 ▸). The data were indexed and processed with iMosflm and scaled with AIMLESS in the CCP4 suite (Winn et al., 2011 ▸). Data-collection and processing statistics are summarized in Table 3 ▸.

Table 3. Data collection and processing.

Diffraction source	Beamline BL17U1, SSRF
Wavelength (Å)	0.9792
Temperature (K)	100
Detector	ADSC Q315 CCD
Rotation range per image (°)	1
Total rotation range (°)	360
Exposure time per image (s)	0.5
Space group	I121
a, b, c (Å)	97.37, 100.61, 131.12
α, β, γ (°)	90, 96.10, 90
Resolution range (Å)	79.65–2.20 (2.25–2.20)
Total No. of reflections	232183 (16373)
No. of unique reflections	63495 (4458)
Completeness (%)	99.5 (100)
Multiplicity	1.9
〈I/σ(I)〉	11 (3.4)
R r.i.m.	0.096 (0.608)
CC1/2	0.996 (0.832)
Overall B factor from Wilson plot (Å2)	35.915

2.4. Structure solution and refinement

The structure was solved by Phaser (McCoy et al., 2005 ▸) using a monomer from the soybean proglycinin A1aB1b homotrimer (PDB entry 1fxz; Adachi et al., 2001 ▸) as the model. The model was subsequently manually built using Coot (Emsley et al., 2010 ▸) and refined using Phenix (Liebschner et al., 2019 ▸). Figures were produced using PyMOL (http://www.pymol.org). The atomic coordinates and structure factors have been deposited in the Protein Data Bank as entry 5gyl. Sequence alignment was performed with MultAlin and ESPript3 (Robert & Gouet, 2014 ▸; Corpet, 1988 ▸). Refinement statistics are summarized in Table 4 ▸.

Table 4. Structure solution and refinement.

Resolution range (Å)	41.0–2.20 (2.23–2.20)
No. of reflections, working set	62994 (2440)
No. of reflections, test set	3050 (122)
Final R cryst	0.205 (0.5136)
Final R free	0.225 (0.5252)
R.m.s. deviations
Bond lengths (Å)	0.004
Angles (°)	0.692
Average B factor (Å2)	43.80
Ramachandran plot
Most favored (%)	97.05
Allowed (%)	2.95

3. Results and discussion

To determine the structure of CL-AI, we expressed recombinant CL-AI in Escherichia coli cells. The main peak in the elution profile (at a buffer volume of ∼65 ml) of CL-AI corresponds to a molecular weight of 160 kDa. Assuming that the overall fold of recombinant CL-AI resembles the globular form of the protein standards, the molecular weight calculated from the size-exclusion chromatography experiment implies that CL-AI exists as a trimer in solution (Fig. 1 ▸ a). CL-AI detected by reduced SDS–PAGE has an apparent molecular weight of between 55 and 70 kDa, which is slightly higher than that predicted from the amino-acid sequence (55 kDa). In addition, there is a small band at the position corresponding to 160 kDa that is consistent with the molecular weight of the CL-AI trimer (Fig. 1 ▸ b). We attribute these two phenomena to protein samples that were not fully reduced or denatured before electrophoresis. Finally, we obtained cubic crystals of CL-AI that diffracted to 2.2 Å resolution (Fig. 1 ▸ c).

Figure 1.

Purification and crystallization of CL-AI. (a) Gel-filtration chromatography of CL-AI using a HiLoad Superdex 200 16/600 column. (b) Fractions A10–B13 from gel filtration were stained by Coomassie Blue on reduced SDS–PAGE. Lane M contains the protein molecular-weight marker labeled in kDa (Thermo Fisher). Fractions A12–B15 were pooled and concentrated to 10 mg ml⁻¹ for crystallization. (c) Cube-shaped crystals were obtained in the condition 20% PEG MME 2000, 0.1 M Tris pH 7.0 by the sitting-drop vapor-diffusion method. The scale bar at the bottom right represents 300 µm.

Each asymmetric unit contains three molecules with threefold symmetry. The three CL-AI molecules are non­covalently linked in a head-to-tail association, as shown in Fig. 2 ▸(a). Each CL-AI molecule is divided into an α-chain and a β-chain. 26 β-strands and six α-helices fold into two highly conserved jelly-roll barrel domains and two extended helix domains (Figs. 2 ▸ b and 2 ▸ c). The root-mean-square deviations (r.m.s.d.s) among the three monomers range between 0.065 and 0.077 Å for the C^α atoms of all residues. The α-chain and β-chain of CL-AI are linked by two disulfide bridges, which are structural characteristics of 11S globulins.

Figure 2.

Overall structure and schematic diagram of CL-AI. (a) The three molecules in the asymmetric unit are shown in green, cyan and magenta. (b) Each molecule can be divided into an α-chain and a β-chain. The extended helix is colored cyan and the β-barrel core (cupin domain) is colored magenta. Cysteines involved in disulfide bonds are shown as red sticks. (c) Rectangles with arrows represent β-sheets and ellipses represent α-helices.

We analyzed the CL-AI structure from the following three aspects. Firstly, we compared the sequence homology of CL-AI with those of other 11S globulins. The complete sequence of CL-AI was found to show sequence identities of 77% to pea prolegumin and 58% to soybean proglycinin A1aB1b. The structures of CL-AI, soybean proglycinin A1aB1b and pea pro­legumin are highly superimposable, with r.m.s.d. values ranging between 0.65 and 0.77 Å based on main-chain atoms (Supplementary Fig. S1). The r.m.s.d. values are larger at sites near the variable regions. Whether other 11S globulins have α-amylase inhibitor activity still needs to be explored.

Secondly, oxalate decarboxylase was found to have the highest structure similarity to CL-AI among classic enzyme-related cupin proteins, with a Z-score of 32.8 and a sequence identity of 13% using a structure-based search with the DALI (distance matrix alignment) server (Holm, 2020 ▸; Fig. 3 ▸ a). The r.m.s.d. values between oxalate decarboxylase and CL-AI are 3.2 Å based on all C^α atoms in β-barrel cores. The sequence alignment between CL-AI and oxalate decarboxylase showed that the conserved metal-binding histidine and glutamate residues located in the β-barrel cores of oxalate decarboxylase have been lost in the α-chain and β-chain of CL-AI. These results indicated that they belong to different branches of the cupin superfamily from an evolutionary point of view.

Figure 3.

Structure superposition and sequence alignment between oxalate decarboxylase (PDB entry 1uw8) and CL-AI (PDB entry 5gyl). (a) Oxalate decarboxylase is colored wheat and CL-AI is colored green. The variable regions of oxalate decarboxylase and CL-AI have been removed for clarity. The metal-binding histidine and glutamate in the active site of oxalate decarboxylase are shown as orange sticks and manganese is shown as an orange sphere. Key residues for CL-AI inhibitory activity are shown as cyan sticks. (b) Conserved amino-acid residues are highlighted with a red background. Homologous residues in oxalate decarboxylase and CL-AI are colored red. Orange triangles represent the conserved His/Glu of oxalate decarboxylase, and cyan ovals represent the key residues for CL-AI inhibitory activity. Solid lines indicate β-sheets and dotted lines represent α-helices in CL-AI.

Finally, the DALI search result showed an α-amylase inhibitor (α-AI) from Phaseolus vulgaris to be the closest match to the α-chain of CL-AI, with a Z-score of 2.3, and the β-chain, with a Z-score of 2.5. Both α-AI isolated from P. vulgaris and CL-AI contain barrel domains that consist of several β-sheets. However, α-AI does not easily superpose with CL-AI. In addition, sequence homology between α-AI and CL-AI is very low and difficult to detect (Supplementary Fig. S2). The two hairpin loops of α-AI, which insert into the active site of the α-amylase, play a key role in its inhibitory activity (Nahoum et al., 1999 ▸). Referring to previous reports, Gly96 in motif 1 and Pro144 and Asn153 in motif 2 of the α-chain of CL-AI that play a crucial role in inhibitory activity are located on loops between β-sheets (Wang et al., 2018 ▸). We term the two faces along the axis through the α-chain and β-chain the inner face and the outer face for easier discussion (Fig. 3 ▸ a). Residues on the outer face have space to bind to other proteins, whereas residues on the inner face can only interact with the other two CL-AIs within a trimer. From the perspective of amino-acid distribution, Gly96 and Asn153 are located on the outer face and Pro144 is located on the inner face. Thus, it is possible that CL-AI binds the active site of α-amylase with its outer face loops to perform inhibitory activity, similar to the α-amylase inhibitor from P. vulgaris. Residues on the inner face are likely to affect the inhibitory activity by reducing the stability of the overall structure.

In summary, we report the structure of CL-AI, a novel 11S globulin from chickpea, at 2.2 Å resolution, which lays the foundation for subsequent structural and functional studies.

Supplementary Material

PDB reference: Cicer arietinum 11S gloubulin, 5gyl