Crystal structures of glycogen-debranching enzyme mutants in complex with oligosaccharides

Five crystal structures of glycogen-debranching enzyme mutants in complex with oligosaccharides are reported.

Abstract

Debranching is a critical step in the mobilization of the important energy store glycogen. In eukaryotes, including fungi and animals, the highly conserved glycogen-debranching enzyme (GDE) debranches glycogen by a glucanotransferase (GT) reaction followed by a glucosidase (GC) reaction. Previous work indicated that these reactions are catalyzed by two active sites located more than 50 Å apart and provided insights into their catalytic mechanisms and substrate recognition. Here, five crystal structures of GDE in complex with oligosaccharides with 4–9 glucose residues are presented. The data suggest that the glycogen main chain plays a critical role in binding to the GT and GC active sites of GDE and that a minimum of five main-chain residues are required for optimal binding.

1. Introduction  

Glycogen is a polymer of glucose and serves as an important energy store in many organisms, including humans. It is highly branched and it has been estimated that branch points occur in glycogen every 12 glucose residues (Roach et al., 2012 ▸). In glycogen, the glucose residues are connected by α-1,4-glycosidic bonds and by α-1,6-glycosidic bonds at branch points (Adeva-Andany et al., 2016 ▸; Chikwana et al., 2013 ▸). During glycogen mobilization, glycogen phosphorylase hydrolyses the last α-1,4-glycosidic bond at the nonreducing end of glycogen chains and releases one glucose residue a time. The glycogen phosphorylase reaction stops upon reaching glycogen branch points, making glycogen debranching a critical reaction in mobilizing glycogen. In eukaryotes, including humans, this reaction is catalyzed by the highly conserved glycogen-debranching enzyme (GDE). To highlight its physiological importance, a deficiency of GDE in humans causes glycogen storage disease type III (GSDIII), which can lead to severe symptoms including growth retardation, hypertrophic cardiomyopathy, hypoglycemia, hepatomegaly and progressive skeletal myopathy (Lucchiari et al., 2007 ▸).

The glycogen phosphorylase reaction stops when the glycogen branched chain is shortened to four glucose residues. GDE recognizes this substrate and converts it into an unbranched glycogen chain by a two-step reaction, allowing its further degradation by glycogen phosphorylase (Roach et al., 2012 ▸; Roach, 2002 ▸). GDE possesses α-1,4-glucanotransferase (GT) and α-1,6-glucosidase (GC) activities (Nakayama et al., 2001 ▸; Liu et al., 1991 ▸), which catalyze the first and second steps, respectively. In the first step, the GT activity removes three glucose residues from the branch and transfers them to the nonreducing end of a neighboring glycogen chain. In the second step, the GC activity hydrolyzes the α-1,6-glycosidic linkage between the branch and the glycogen main chain, releasing the last glucose residue in the branch (Cori & Larner, 1951 ▸; Walker & Whelan, 1960 ▸; Brown & Illingworth, 1962 ▸; Abdullah & Whelan, 1963 ▸; Brown et al., 1963 ▸).

In previous work (Zhai et al., 2016 ▸), we determined the crystal structure of Candida glabrata GDE (CgGDE), which revealed that CgGDE is composed of four domains. The N- and C- terminal domains are located at either end of the protein and are separated by two middle domains (M1 and M2; Fig. 1 ▸ a). Our structure-guided mutagenesis and biochemical studies indicated that the highly conserved residue pairs Glu564/Asn535 and Asp1241/Glu1492 located more than 50 Å apart in the N- and C-terminal domains (GT and GC domains) catalyze the GT and GC reactions, respectively. We further determined the structure of CgGDE in complex with maltopentaose, which contains five glucose residues, and found that maltopentaose molecules bind to multiple sites in CgGDE, which most likely mediate its inter­action with glycogen (Fig. 1 ▸ a). Two maltopentaose molecules, M and B, were found in the GT active site, which most likely mimic binding of the glycogen main chain and branch to this site, respectively (Figs. 1 ▸ b and 1 ▸ d). The M and B molecules wrap around Trp470 and Trp496, respectively, in the GT active site and form extensive interactions with additional residues in the neighborhood. One maltopentaose molecule was found in the GC active site, and is most likely to mimic the binding of the glycogen main chain to this site. It forms extensive interactions with Arg1123 and Tyr1407 and additional residues in this region (Figs. 1 ▸ c and 1 ▸ e). Here, we present five structures of CgGDE mutants in complex with oligosaccharides with 4–9 glucose residues. These structures suggest that the glycogen main chain plays a critical role in the binding of glycogen to the GT and GC active sites and that a minimum of five main-chain residues are required for optimal binding to these sites, providing additional insights into substrate recognition by GDE.

Figure 1.

Structure of CgGDE and its oligosaccharide-binding sites. (a) Structure of CgGDE (PDB entry 5d0f). Subdomains A, B and C in the GT domain, the GC domain and domains M1 and M2 in CgGDE are colored as indicated in the bar diagram. The previously identified oligosaccharide-binding sites are referred to as sites 1–5 and are indicated. (b, c) Oligosaccharide binding to the GT (b) and GC (c) active sites in CgGDE. The asterisk in (b) indicate the possible connection between the glycogen main chain and branch. The asterisk in (c) indicates the possible location of the branch. Catalytic residues and additional important residues at the binding sites are highlighted. (d, e) As in (b) and (c) with CgGDE presented in surface representation. Structural figures were prepared with PyMOL (http://www.pymol.org).

2. Materials and methods  

2.1. Protein expression and purification  

CgGDE mutants were generated, expressed and purified following a previously published protocol (Zhai et al., 2016 ▸). Briefly, the CgGDE gene was amplified from the C. glabrata genome and inserted into pET-26b vector (Novagen; Table 1 ▸). CgGDE mutations were generated with the QuikChange kit (Agilent Technologies) and verified by DNA sequencing. The plasmid was transformed into Escherichia coli BL21 Rosetta (DE3) cells, which were cultured in LB medium supplemented with 34 mg l⁻¹ kanamycin and 25 mg l⁻¹ chloramphenicol and induced with 0.3 mM isopropyl β-d-1-thiogalactopyranoside (Bio Basic) at 16°C for 12 h. Harvested cells were lysed using an AH-2010 homogenizer (ATS Engineering) in a buffer consisting of 20 mM Tris pH 7.5, 300 mM sodium chloride, 2 mM β-mercaptoethanol. The lysate was clarified by centrifugation at 15 000g for 30 min and soluble CgGDE was purified using nickel–nitrilotriacetic acid (Qiagen), ion-exchange (HiTrap Q HP, GE Healthcare) and size-exclusion (Sephacryl S300 HR, GE Healthcare) columns. Purified CgGDE was concentrated to 20 mg ml⁻¹ in a buffer consisting of 20 mM Tris–HCl pH 7.5, 200 mM NaCl, 2 mM DTT, 5% glycerol, flash-cooled in liquid nitrogen and stored at −80°C.

Table 1. Macromolecule-production information.

Source organism	C. glabrata
Cloning vector	pET-26b
Expression vector	pET-26b
Expression host	E. coli BL21 Rosetta (DE3) cells
Complete amino-acid sequence of the construct produced	MSAHRTLLLRLSDSGEPVTSCSYGQGVLTLPSLPLPQGKKLGDMPVYTVKLAIPAGSPVTRDGLIWTNCPPDFSTQFDREKFYKKIIKTSFHEDDHIDLDIYVPGTYCFYLSFKNDKDELETTRKFYFVVLPILSVNDKFIPLNSIAMQSVVSKWMGPTIKDWEKVFARVASKKYNMIHFTPLQHRGESNSPYSIYDQLEFDPTVFKSEKEVADMVERLRTEHNILSLTDIVFNHTANNSQWLLDHPEAGYNHKTSPHLISAIELDKKLLDFSEQMEALGYPVDLKTVDDLIKVMDGIKEHVIGELKLWEFYVVDVKQTVSELREKWGNSKSWSDDNIPSKDDSTNLAQFVRDNATEPGFGSLGERGSNKINIDKFAAILKKLHSEDYNNGIEELATKILNDINLPFYKEYDDDINEVLEQLFNRIKYLRIDDHGPKQGPITKKLPLSEPYFTRFKAKDGEEYALANNGWIWDGNPLVDFASSQSKAYLRREVIVWGDCVKLRYGKGPSDSPYLWERMSKYVEMNARIFNGFRIDNCHSTPLHVGQYFLDVARRVNPNLYVVAELFSGSEAMDCLFVERLGISSLIREAMQAWSEEELSRLVHRHGGRPIGSYKFVPLDDFPYPADVKIDEEYCAYNPDDHSVKCVSEIMIPKTLTATPPHALFMDCTHDNETPNQKRTVEDTLPNAALVAFCSSAIGSVYGYDEVFPQLLDLVQEKRTYSCAENTGISKVKTLLNNMREEIASEAVDIEDSEMHVHHDGQYITFHRTNAKNGKGWYLVARTKFHSSGDQMLPRIKLSQTKATFKAAFSLERTGDAPISDEIIEGIPTKLRELTGFDIGFDENTKETSILLPQDFPQGSIVIFETQQLGIDDSLDHFIRSGAIKATEKLSLESINYVLYRAEQEEYDYSEGRSGAYDIPDYGKPVYCGLQGWVSILRKIIFYNDLAHPLSNNLRNGHWAVDYVVNRLDLYKDKEGVAEVQEWLRSRMERIKQLPSYLVPSFFALVVGIMYGCCRLRAMQLMSDNVGKSTVFVQSLAMTSIQMVSAMKSTSILPDQNIAAMAAGLPHFSTNYMRCWGRDVFISLRGLLLTTGRYEEAKEHILAFAKTLKHGLIPNLLDAGRNPRYNARDAAWFFVQAIQDYVTIVPGGVSLLQEKVTRRFPLDDEYIPYDDPKAFSYSSTIEEIIYEILNRHAGGIKYREANAGPNLDRVMKDEGFNVEVNVDWETGLIHGGSQFNCGTWMDKMGESEKANSVGVPGTPRDGAAVEINGLLKSCLRFVLQLSKDGKFKYTEVTKPDGSKISLSSWNDLLQENFERCFYVPKNKEDDNKFEIDATIINRRGIYKDLYRSGKPYEDYQFRPNFTIAMVVAPELFTPDYAAGAIELADQVLRGPVGMRTLDPSDYNYRPYYNNGEDSDDFATSKGRNYHQGPEWVWCYGYFIRAYHYFNFLTNPKCQVEGSAKKLKPSSYLYRKLYSRLLKHREWIENSPWAGLAELTNKDGEVCNDSSPTQAWSTGCLLDLFYDLWISYEE

2.2. Crystallization  

To produce crystals of the CgGDE–oligosaccharide complexes, CgGDE and oligosaccharides (Sigma–Aldrich) were mixed in a 1:1000 molar ratio and incubated on ice for an hour before sitting-drop vapor-diffusion crystallization experiments. The reservoir solution consisted of 10% polyethylene glycol 5000 monomethyl ether, 0.1 M HEPES pH 7.0, 5% Tacsimate pH 7.0. Before data collection, crystals were equilibrated in reservoir solution supplemented with 25% ethylene glycol for 30 s, flash-cooled and stored in liquid nitrogen.

2.3. Data collection and processing  

Diffraction data were collected using an ADSC Q315 charge-coupled device detector on beamline BL17U at the Shanghai Synchrotron Radiation Facility at 100 K. Diffraction data for the D535N mutant were indexed and integrated with MOSFLM (Battye et al., 2011 ▸) and scaled with SCALA (Evans, 2006 ▸). Diffraction data for the W470A mutant were indexed, integrated and scaled with HKL-2000 (Otwinowski & Minor, 1997 ▸). Diffraction data for the W470A/E564Q mutant were indexed and integrated with XDS (Kabsch, 2010 ▸) and scaled with XSCALE (Kabsch, 2010 ▸). Data-collection statistics are summarized in Table 2 ▸.

Table 2. Data collection and processing.

Values in parentheses are for the outer shell.

	D535N with G4	W470A with G5	W470A with G6	W470A with G7	W470A/E564Q with G9
Wavelength (Å)	0.97939	0.91769	0.91769	0.97908	0.97914
Space group	C2	C2	C2	C2	C2221
a, b, c (Å)	160.25, 198.94, 134.59	156.52, 200.32, 134.28	156.54, 199.83, 133.88	156.34, 199.27, 133.74	160.66, 206.28, 258.05
α, β, γ (°)	90.00, 104.87, 90.00	90.00, 100.84, 90.00	90.00, 100.79, 90.00	90.00, 100.77, 90.00	90.00, 90.00, 90.00
Resolution range (Å)	25.75–3.10 (3.27–3.10)	38.77–3.10 (3.15–3.10)	38.44–3.10 (3.15–3.10)	40.48–3.20 (3.26–3.20)	46.15–3.40 (3.49–3.40)
No. of unique reflections	73585 (10713)	73503 (3642)	73954 (3680)	66972 (3348)	58913 (4334)
Completeness (%)	99.8 (99.9)	99.6 (99.7)	99.8 (99.9)	99.9 (100.0)	99.6 (100.0)
Multiplicity	3.7 (3.8)	3.6 (3.6)	3.8 (3.7)	3.8 (3.7)	7.3 (7.5)
〈I/σ(I)〉	6.70 (1.30)	18.58 (1.03)	15.74 (0.98)	13.43 (1.12)	8.31 (1.44)
CC1/2	0.991 (0.564)	(0.470)†	(0.427)†	(0.624)†	0.998 (0.588)
R r.i.m.	0.141 (1.055)	0.053 (1.073)	0.067 (1.046)	0.068 (0.873)	0.125 (1.081)

^† HKL-2000 did not report the overall CC_1/2, only CC_1/2 values for each resolution shell.

2.4. Structure solution and refinement  

The structures were determined by molecular replacement with Phaser (McCoy et al., 2007 ▸) using the crystal structure of native CgGDE (PDB entry 5d06; Zhai et al., 2016 ▸) as the search model. Inspection/modification of the structures and refinement were carried out with Coot (Emsley et al., 2010 ▸) and Phenix (Liebschner et al., 2019 ▸), respectively. The geometry of the oligosaccharides in the structures was analyzed with Privateer (Agirre et al., 2015 ▸). Structure-refinement statistics are summarized in Table 3 ▸.

Table 3. Structure refinement.

Values in parentheses are for the outer shell.

	D535N with G4	W470A with G5	W470A with G6	W470A with G7	W470A/E564Q with G9
Resolution range (Å)	25.75–3.10 (3.14–3.10)	38.77–3.10 (3.14–3.10)	38.44–3.10 (3.14–3.10)	40.48–3.20 (3.25–3.20)	46.15–3.40 (3.46–3.40)
No. of reflections	73414 (2571)	71917 (2399)	72918 (2509)	66074 (2599)	58885 (2648)
Final R work	0.259	0.2593	0.2432	0.2184	0.2773
Final R free	0.319	0.3199	0.2997	0.2863	0.3467
No. of atoms
Protein	24556	24538	24538	24538	24538
Ligand	23	112	158	159	92
R.m.s. deviations
Bonds (Å)	0.011	0.012	0.011	0.010	0.011
Angles (°)	1.257	1.438	1.204	1.226	1.508
Average B factors (Å2)
Protein	86.589	112.284	100.184	99.211	114.005
Protomer A	85.758	111.509	100.228	97.309	110.581
Protomer B	87.420	113.058	100.139	101.113	117.429
Ligand	146.459	162.374	156.546	137.811	139.005
B factor estimated from Wilson plot (Å2)	61.63	72.92	71.50	70.32	73.94
No. of glucose residues	2	18	20	19	17
No. of glucose residue outliers	0	5	5	3	2

3. Results and discussion  

3.1. Structure determination  

In this study, to further probe the mechanism of glycogen binding to GDE, we co-crystallized CgGDE with oligosaccharides containing four, five, six, seven and nine glucose residues (referred to here as G4, G5, G6, G7 and G9). In our previous study, to prevent the conversion of maltopentaose to other oligosaccharides by the GT reaction, we co-crystallized the GT-inactive E564Q mutant with maltopentaose. Similarly, in this study we used the GT-inactive mutant D535N in the co-crystallization experiments. To probe oligosaccharide binding to the GT active site, we also studied the W470A mutant, which is expected to disrupt the oligosaccharide M binding site and greatly inhibited the GT activity (Zhai et al., 2016 ▸). To completely remove the GT activity of the W470A mutant, we also included a W470A/E564Q double mutant in the structural studies. The structures were determined using molecular replacement and were refined to resolutions of 3.1–3.4 Å (Table 3 ▸).

All of the crystals contained two very similar protomers in the asymmetric unit, which can be superimposed with root-mean-square deviations (r.m.s.d.s) for related C^α atoms of between 0.4 and 1.4 Å. They are very similar to the previously determined structures of CgGDE and can be superimposed onto these structures with r.m.s.d.s for related C^α atoms of between 0.6 and 1.2 Å. Therefore, the D535N, W470A and W470A/E564Q mutations and the oligosaccharides that we used in this study did not cause significant structural changes in CgGDE.

3.2. Oligosaccharide binding to CgGDE  

We have previously shown that CgGDE contains multiple glycogen-binding sites, including its GT and GC active sites and a site in its M2 domain, as well as additional sites in its GT and GC domains (Fig. 1 ▸ a; Zhai et al., 2016 ▸). In the structures presented in this study, strong difference electron densities appeared at many of these locations after molecular-replacement calculations. Oligosaccharide molecules can be fitted into these densities (Figs. 2 ▸ a–2 ▸ e), indicating that the oligosaccharides that we used in co-crystallization experiments bind at many of these sites.

Figure 2.

Oligosaccharide binding to CgGDE. (a–e) Electron-density maps of oligosaccharides bound to sites 1–5 in the CgGDE structures presented in this study: the D535N mutant complexed with G4 (a), the W470A mutant complexed with G5 (b), the W470A mutant complexed with G6 (c), the W470A mutant complexed with G7 (d) and the W470A/E564Q mutant complexed with G9 (e). The upper panels show difference electron-density maps before adding oligosaccharides to the model, contoured at 2σ. The lower panels show the final electron-density map contoured at 1σ.

In the structures of CgGDE co-crystallized with G5–G7 presented in this study, as well as in protomer A of the structure of CgGDE co-crystallized with G9, the strong electron density in the GC active site can be modeled as a G5 molecule (Figs. 2 ▸ b–2 ▸ e). In protomer B of the structure of CgGDE co-crystallized with G9, the strong electron density in the GC active site can be modeled as a G7 molecule (Fig. 2 ▸ e). The structures of the modeled G5 molecules and residues 2–6 of the modeled G7 molecule are very similar to the structure of the modeled G5 molecule in the GC active site in the previously reported CgGDE structure (Fig. 3 ▸; Zhai et al., 2016 ▸). Residues 1 and 7 in the modeled G7 form few interactions with CgGDE. No density for oligosaccharides could be found at the GC active site in the structure of the D535N mutant co-crystallized with G4. Since the D535N mutation does not affect the GC active site, the structures suggest that G4 does not bind to the GC active site.

Figure 3.

Comparison of oligosaccharides bound to the GC active site. Protomers A (a) and B (b) in the CgGDE structures presented in this study and the previously reported structure of the E564Q mutant complexed with G5 (PDB entry 5d0f; gray C atoms) are superimposed and the oligosaccharide bound to the GC active site is shown. The C atoms of the oligosaccharide in the structures of the W470A mutant complexed with G5, the W470A mutant complexed with G6, the W470A mutant complexed with G7 and the W470A/E564Q mutant complexed with G9 are colored red, cyan, green, orange and blue, respectively. The glucose residues are numbered from the nonreducing end to the reducing end. The different conformations adopted by residue 3 in the oligosaccharide bound to protomer A in the structure of the W470A/E564Q mutant in complex with G9 may be caused by crystal-packing interactions with the nonreducing end of this oligosaccharide.

Interestingly, in all of the structures presented in this study no densities for oligosaccharides can be found in the GT active site. This suggests that either the oligosaccharides we used in our co-crystallization experiments do not bind at this site or the mutations that we introduced prevented oligosaccharide binding to this site.

Although the mutations that we introduced are not expected to disrupt oligosaccharide binding at sites 3–5, some of these sites in our structures did not appear to have bound oligosaccharides. No densities for oligosaccharides were observed at site 3 in the structures of the D535N mutant in complex with G4 and of the W470A/E564Q mutant in complex with G9, at site 4 in protomer A in the structure of the D535N mutant in complex with G4 or at site 5 in all of the structures presented in this study (Figs. 2 ▸ a–2 ▸ e). Together, these data suggest that oligosaccharide binding to sites 3–5 is somewhat weaker than binding to the GC active site.

GDE contains multiple glycogen-binding sites, making it difficult to probe glycogen binding to an individual site using conventional binding experiments. The structural studies presented in this study complement our previous study on GDE by providing additional insights into the glycogen–GDE interaction at the GC and GT active sites. The structures presented in this study indicate that oligosaccharides with five, six, seven or nine glucose residues can bind at the GC active site, but that G4, which contains four glucose residues, cannot. In addition, the structures indicate that G5–G9 form similar interactions with the GC active site. Our previous study suggested that G5 bound at the GC active site is most likely to mimic the binding of the glycogen main chain. Together, the previous and present studies suggest that the glycogen main chain plays a critical role in the binding of glycogen to the GC active site of CgGDE and that five main-chain residues are required for optimal binding.

A similar analysis provides insights into glycogen binding to the GT active site. In our previously reported structure of CgGDE in complex with G5, all five glucose residues are resolved for the M oligosaccharide, which probably mimics the binding of the glycogen main chain to the GT active site. In the structure of the D535N mutant in complex with G4 presented here no density for the M oligosaccharide could be observed. The D535N mutation is not located close to the oligosaccharide M binding site (Fig. 1 ▸ b). The mutation did not cause significant structural changes in the overall CgGDE structure or in the structure of the oligosaccharide M binding site. Together, the previous and present studies suggest that a minimal of five glycogen main-chain residues are required for optimal binding to the GT active site of CgGDE. It is not surprising that the M oligosaccharide is not observed in the structures of CgGDE with the W470A mutation, since the mutation is expected to disrupt glycogen main chain binding (Fig. 1 ▸ b; Zhai et al., 2016 ▸). Interestingly, the B oligosaccharide that probably mimics the binding of the glycogen branch to the GT active site is not visible in any of the structures presented in this study. Our previous study suggests that the oligosaccharide B binding site may not accommodate oligosaccharides with more than five glucose residues, but it does not explain why G4 or G5 do not bind to this site in the present study. In fact, in our previously reported structure of the E654Q mutant in complex with G5, the B oligosaccharide was clearly resolved (Zhai et al., 2016 ▸). The absence of oligosaccharide B seems to correlate with the absence of oligosaccharide M. Together, these data suggest that glycogen main chain binding could facilitate glycogen branch binding. Trp470 and Trp496 at the center of the oligosaccharide M and B binding sites, respectively, form van der Waals interactions (Fig. 1 ▸ b). It is possible that binding of the glycogen main chain to the oligosaccharide M binding site may restructure the oligosaccharide B binding site through Trp470, Trp496 and additional residues in the neighborhood to facilitate branch binding. Our previously reported CgGDE structures and the structures presented in this study were determined at medium resolution. The proposed restructuring may be too subtle to be captured by these structures.

In summary, our study provides additional insights into glycogen binding to the active sites of CgGDE. The structures suggest that the glycogen main chain plays critical roles in the binding of glycogen to the GC and GT active sites of CgGDE and that five main-chain residues are required for optimal binding. The GC and GT active sites of GDE are highly conserved (Zhai et al., 2016 ▸), and other GDEs probably also interact with glycogen with similar mechanisms. The structures presented here could possibly pave the way to further probe the mechanism of the interaction between GDE and its substrate glycogen.

Supplementary Material

PDB reference: glycogen-debranching enzyme, D535N mutant, complex with maltotetraose, 7ekw

PDB reference: W470A mutant, complex with maltopentaose, 7eim

PDB reference: W470A mutant, complex with maltohexaose, 7ejp

PDB reference: W470A mutant, complex with maltoheptaose, 7ejt

PDB reference: W470A/E564Q mutant, complex with maltononaose, 7ekx

Funding Statement

This work was funded by Natural Science Foundation of China grant 31870769 to Song Xiang; National Natural Science Foundation of China grant 32071205 to Song Xiang.