Recognition of glycan and protein substrates by N-acetylglucosaminyltransferase-V
Introduction
Glycosylation is the most common post-translational modifications in eukaryotic cells, and it has been proposed that more than 50% of proteins are glycosylated. Glycosylation plays pivotal roles for protein folding, trafficking, and signal transduction [1,2]. Alterations in glycan structures regulate the functions of proteins and cells and are widely used as markers for certain cell types and diseases [3,4]. Among several types of glycans, N-glycans are highly conserved and abundant in eukaryotes [5]. The structures of N-glycans are diverse and specific to proteins. The number of N-acetylglucosamine (GlcNAc) branches is a major factor to produce the structural diversity of N-glycans [6,7]. The presence or absence of a certain GlcNAc branch results in functional alterations of specific target proteins, leading to dramatic changes in disease pathology [4,8]. For example, loss of the bisecting GlcNAc branch in mice greatly improves Alzheimer's disease pathology by relocating amyloid-β-producing enzyme to lysosomes [9,10], and Mgat4a-deficient mice lacking another GlcNAc branch exhibit type 2 diabetes-like phenotypes caused by impaired cell surface localization of glucose transporter-2 in pancreatic beta-cells [11,12]. These findings indicate that modifications of key target proteins with certain GlcNAc branches are critically important for physiological functions, but it is unclear how each branch is expressed on specific glycoproteins.
N-Acetylglucosaminyltransferase-V (GnT-V; encoded by MGAT5) produces the β1,6-linked GlcNAc branch on α1,6-mannose residue in N-glycans using UDP-GlcNAc as a high energy sugar donor (Fig. 1A) [[13], [14], [15]]. Many studies have shown that this β1,6-branch promotes tumorigenesis and cancer malignancy [16,17]. Expression of the MGAT5 gene is driven by the oncogenic Ras-Raf-Ets pathway [18,19] and aberrantly elevated in various types of cancers [20,21]. Close relationships between expression of the β1,6-branch and a poor prognosis of cancer have been reported [20,21]. Furthermore, studies of Mgat5 knockout (KO) mice and cells have revealed that tumorigenesis, tumor growth and metastasis are strongly suppressed by Mgat5 KO [22,23], highlighting the potential of GnT-V as a drug target for cancer therapy. Notably, not all N-glycosylated proteins are modified by GnT-V and these cancer-related functions of GnT-V are mediated by specific target glycoproteins such as growth factor receptors, matriptase, and cadherins [[24], [25], [26], [27], [28]]. To delineate the regulation of the biosynthesis and functions of β1,6-branched glycans on specific proteins, it is necessary to understand the detailed mechanisms of how GnT-V selectively acts on target proteins.
GnT-V localizes to the Golgi and a type-II transmembrane protein with a single pass transmembrane domain [13,14] as is the case with other typical Golgi-localized glycosyltransferases [29]. Enzymology studies have characterized the specificity of GnT-V for glycan substrates, revealing that GnT-V strictly acts on α1,6-mannose, but not on α1–3 mannose [14,15,30], and the presence of bisecting GlcNAc completely prevents GnT-V activity [6,7,14,30]. In addition, GnT-V does not require divalent cations such as Mn2+ for its catalysis, which is different from other GlcNAc transferases with the GT-A fold containing the DXD motif. Although these in vitro studies revealed important aspects of GnT-V-mediated reactions, it remains elusive how GnT-V acts on selective proteins. A previous study has shown that GnT-V modifies only a few native glycoproteins, but acts on much more kinds of denatured glycoproteins [31], suggesting that the structures of glycoprotein substrates largely affect the actions of GnT-V.
To deeply understand the reaction mechanisms catalyzed by GnT-V, we recently solved the crystal structures of a catalytic domain of GnT-V with or without a substrate analog in which donor and acceptor derivatives were conjugated through a linker (bisubstrate-type inhibitor [32]) [33]. The catalytic domain of GnT-V forms a GT-B fold, which is consistent with its non-requirement for metals because glycosyltransferases with a GT-B fold usually lack the DXD motif and do not require a metal cation. The N-glycan-binding pocket of GnT-V is localized in a deep position of the catalytic domain. GlcNAc on α1,6-mannose is tightly sandwiched by two aromatic residues in the catalytic cavity, whereas GlcNAc on α1,3-mannose or bisecting GlcNAc are not accommodated by this pocket because of steric hindrance [33]. These observations are well consistent with the substrate preference observed in previous biochemical studies, and clarified how GnT-V recognizes the acceptor site of N-glycans. We also created a docking model of GnT-V and a biantennary N-glycan and found a possibility that several amino acid residues outside of the catalytic pocket may interact with the N-glycan core and surrounding polypeptide of the acceptor glycoprotein [33]. This possibility was supported by the finding that GnT-V exhibits higher activity for an integral N-glycan structure linked to Asn than a shorter N-glycan substrate without the core GlcNAc and Asn [33]. These findings suggest that GnT-V actively recognizes substrate proteins.
Here, we attempted to further understand the recognition mechanisms of glycoprotein substrates by GnT-V. Based on the docking model, we selected four candidate residues outside of the catalytic pocket, which are potentially involved in the interaction with the N-glycan core and polypeptide moiety. Using in vitro and cell-based approaches, we found that mutants of these residues show altered and reduced activity for glycoprotein substrates, suggesting that GnT-V acts on its target glycoproteins through these residues.
Section snippets
Antibodies and lectins
An anti-glyceraldehyde-3-phosphate dehydrogenase (GAPDH) mAb (MAB374) was purchased from Millipore, and an anti-Golgin97 mAb (13192) was obtained from Cell Signaling Technology. The anti-GnT-V mAb was kindly provided by Dr. Naoyuki Taniguchi (Osaka International Cancer Institute) [34]. Biotinylated Sambucus sieboldiana lectin (SSA) (J218) was purchased from J-Chemical, FITC-labeled L4-PHA was purchased from J-Oil mills, and leuko-agglutinating phytohemagglutinin (L4-PHA) (J212) and biotinylated
Candidate residues involved in recognition of glycoprotein substrates
In a previous study, we solved the crystal structure of the GnT-V catalytic domain complexed with a bisubstrate-type inhibitor containing a part of the acceptor N-glycan (GlcNAcβ1-2Manα1–6Man) (Fig. 1A, red box). Based on this complex structure, we generated the docking model of GnT-V and an integral N-glycan containing α1,3- and α1,6-Man arms with GlcNAc, chitobiose, core-fucose, and Asn (Fig. 1B, C, D, Supplementary Fig. 1, 2). The trisaccharide including the acceptor site Man to which
Discussion
In this study, we prepared four point mutants of GnT-V. Based on our docking model (Fig. 1B, C, D, Supplementary Fig. 1, 2), we chose four candidate residues that presumably interact with non-acceptor sites (outside of the acceptor trisaccharide, GlcNAcβ1-2Manα1–6Man) of glycoprotein substrates. Indeed, mutation of two sites (V354N and K361A) resulted in particularly reduced activity and an altered substrate preference.
The point mutant V354N largely reduced the activity of GnT-V (Fig. 2C) with
Funding
This work was partially supported by a Grant-in-Aid for Scientific Research (C) to YK [17K07356], a Grant-in-Aid for Scientific Research (B) to YK [20H03207], the Leading Initiative for Excellent Young Researchers (LEADER) project (YK) from the Japan Society for the Promotion of Science (JSPS), a CREST grant [(18070267) to YK] from JST, grants from the Takeda Science Foundation to YK, and a grant from the Uehara Memorial Foundation to YK.
Author contributions
Y.K. conceived and supervised the project. T.H., R.F.O. and M.Y. performed the experiments. M.N. and S.K.M. generated the structural models. T.H., M.N., S.K.M. and Y.K. wrote the manuscript. All authors commented on the manuscript and approved submission.
Tetsuya Hirata: Investigation, Writing – Original Draft. Masamichi Nagae: Investigation, Writing – Review & Editing, Visualization. Reina F. Osuka: Investigation. Sushil K. Mishra: Investigation, Writing – Review & Editing. Mayumi Yamada:
Declaration of Competing Interest
The authors declare that they have no conflicts of interest with the contents of this article.
Acknowledgments
We thank Ms. Yuko Tokoro and Chizuko Yonekawa for technical assistance. We also thank Mitchell Arico from Edanz Group (www.edanzediting.com/ac) for editing a draft of this manuscript.
References (47)
- et al.
Glycosylation in cellular mechanisms of health and disease
Cell
(2006) - et al.
Disease-associated glycans on cell surface proteins
Mol. Asp. Med.
(2016) - et al.
The biosynthesis of highly branched N-glycans: studies on the sequential pathway and functional role of N-acetylglucosaminyltransferases I, II, III, IV, V and VI
Biochimie
(1988) - et al.
N-glycan and Alzheimer’s disease
Biochim. Biophys. Acta Gen. Subj.
(2017) - et al.
Dietary and genetic control of glucose transporter 2 glycosylation promotes insulin secretion in suppressing diabetes
Cell
(2005) - et al.
Purification and characterization of rat kidney UDP-N-acetylglucosamine: alpha-6-D-mannoside beta-1,6-N-acetylglucosaminyltransferase
J. Biol. Chem.
(1992) - et al.
A mouse lymphoma cell line resistant to the leukoagglutinating lectin from Phaseolus vulgaris is deficient in UDP-GlcNAc: alpha-D-mannoside beta 1,6 N-acetylglucosaminyltransferase
J. Biol. Chem.
(1982) - et al.
Glycans and cancer: role of N-glycans in cancer biomarker, progression and metastasis, and therapeutics
Adv. Cancer Res.
(2015) - et al.
Transcriptional regulation of the N-acetylglucosaminyltransferase V gene in human bile duct carcinoma cells (HuCC-T1) is mediated by Ets-1
J. Biol. Chem.
(1996) - et al.
Transcriptional regulation of N-acetylglucosaminyltransferase V by the src oncogene
J. Biol. Chem.
(1997)