Showing papers by "Yoshiki Yamaguchi published in 2022"

Discovery of a lectin domain that regulates enzyme activity in mouse N-acetylglucosaminyltransferase-IVa (MGAT4A)

Abstract: N-Glycosylation is a common post-translational modification, and the number of GlcNAc branches in N-glycans impacts glycoprotein functions. N-Acetylglucosaminyltransferase-IVa (GnT-IVa, also designated as MGAT4A) forms a β1-4 GlcNAc branch on the α1-3 mannose arm in N-glycans. Downregulation or loss of GnT-IVa causes diabetic phenotypes by dysregulating glucose transporter-2 in pancreatic β-cells. Despite the physiological importance of GnT-IVa, its structure and catalytic mechanism are poorly understood. Here, we identify the lectin domain in mouse GnT-IVa's C-terminal region. The crystal structure of the lectin domain shows structural similarity to a bacterial GlcNAc-binding lectin. Comprehensive glycan binding assay using 157 glycans and solution NMR reveal that the GnT-IVa lectin domain selectively interacts with the product N-glycans having a β1-4 GlcNAc branch. Point mutation of the residue critical to sugar recognition impairs the enzymatic activity, suggesting that the lectin domain is a regulatory subunit for efficient catalytic reaction. Our findings provide insights into how branching structures of N-glycans are biosynthesized.

Functional validation of novel variants in B4GALNT1 associated with early‐onset complex hereditary spastic paraplegia with impaired ganglioside synthesis

Abstract: Childhood‐onset forms of hereditary spastic paraplegia are ultra‐rare diseases and often present with complex features. Next‐generation‐sequencing allows for an accurate diagnosis in many cases but the interpretation of novel variants remains challenging, particularly for missense mutations. Where sufficient knowledge of the protein function and/or downstream pathways exists, functional studies in patient‐derived cells can aid the interpretation of molecular findings. We here illustrate the case of a 13‐year‐old female who presented with global developmental delay and later mild intellectual disability, progressive spastic diplegia, spastic‐ataxic gait, dysarthria, urinary urgency, and loss of deep tendon reflexes of the lower extremities. Exome sequencing showed a novel splice‐site variant in trans with a novel missense variant in B4GALNT1 [NM_001478.5: c.532‐1G>C/c.1556G>C (p.Arg519Pro)]. Functional studies in patient‐derived fibroblasts and cell models of GM2 synthase deficiency confirmed a loss of B4GALNT1 function with no synthesis of GM2 and other downstream gangliosides. Collectively these results established the diagnosis of B4GALNT1‐associated HSP (SPG26). Our approach illustrates the importance of careful phenotyping and functional characterization of novel gene variants, particularly in the setting of ultra‐rare diseases, and expands the clinical and molecular spectrum of SPG26, a disorder of complex ganglioside biosynthesis.

Dual impacts of a glycan shield on the envelope glycoprotein B of HSV-1: evasion from human antibodies in vivo and neurovirulence

Abstract: ABSTRACT Identification of the mechanisms of viral evasion from human antibodies is crucial both for understanding viral pathogenesis and for designing effective vaccines. Here we show in cell cultures that an N-glycan shield on the herpes simplex virus 1 (HSV-1) envelope glycoprotein B (gB) mediated evasion from neutralization and antibody-dependent cellular cytotoxicity due to pooled γ-globulins derived from human blood. We also demonstrated that the presence of human γ-globulins in mice and immunity to HSV-1 induced by viral infection in mice significantly reduced replication in their eyes of a mutant virus lacking the glycosylation site but had little effect on the replication of its repaired virus. These results suggest that an N-glycan shield on a specific site of HSV-1 envelope gB mediated evasion from human antibodies in vivo and from HSV-1 immunity induced by viral infection in vivo. Notably, we also found that an N-glycan shield on a specific site of HSV-1 gB was significant for HSV-1 neurovirulence and replication in the central nervous system of naïve mice. Thus, we have identified a critical N-glycan shield on HSV-1 gB that has dual impacts, namely evasion from human antibodies in vivo and viral neurovirulence. IMPORTANCE Herpes simplex virus 1 (HSV-1) establishes lifelong latent and recurrent infections in humans. To produce recurrent infections that contribute to transmission of the virus to new human host(s), the virus must be able to evade the antibodies persisting in latently infected individuals. Here, we show that an N-glycan shield on the specific site of the envelope glycoprotein B (gB) of HSV-1 mediates evasion from pooled γ-globulins derived from human blood both in cell cultures and mice. Notably, the N-glycan shield on the specific site of gB was also significant for HSV-1 neurovirulence in naïve mice. Considering the clinical features of HSV-1 infection, these results suggest that the glycan shield not only facilitates recurrent HSV-1 infections in latently infected humans by evading antibodies but is also important for HSV-1 pathogenesis during the initial infection. Herpes simplex virus 1 (HSV-1) establishes lifelong latent and recurrent infections in humans. To produce recurrent infections that contribute to transmission of the virus to new human host(s), the virus must be able to evade the antibodies persisting in latently infected individuals. Here, we show that an N-glycan shield on the specific site of the envelope glycoprotein B (gB) of HSV-1 mediates evasion from pooled γ-globulins derived from human blood both in cell cultures and mice. Notably, the N-glycan shield on the specific site of gB was also significant for HSV-1 neurovirulence in naïve mice. Considering the clinical features of HSV-1 infection, these results suggest that the glycan shield not only facilitates recurrent HSV-1 infections in latently infected humans by evading antibodies but is also important for HSV-1 pathogenesis during the initial infection.

A Scalable Many-core Overlay Architecture on an HBM2-enabled Multi-Die FPGA

Abstract: The overlay architecture enables to raise the abstraction level of hardware design and enhances hardware-accelerated applications’ portability. In FPGAs, there is a growing awareness of the overlay structure as typified by many-core architecture. It works in theory; however, it is difficult in practice, because it is beset with serious design issues. For example, the size of FPGAs is bigger than before. It is exacerbating the issue of the place-and-route. Besides, a single FPGA is actually the sum of small-to-middle FPGAs by advancing packaging technology like silicon interposers. Thus, the tightly coupled many-core designs will face this covert issue that the wires among the regions are extremely restricted. This article proposes efficient essential processing elements, micro-architecture design, and the interconnect architecture toward a scalable many-core overlay design. In particular, our work proposes a novel compact buffering technique to reduce memory resource utilization in tightly connected overlays while preserving computational efficiency. This technique reduces the utilization of BlockRAM to nearly 50% while achieving a best-case computational efficiency of 91.93% in a three-dimensional Jacobi benchmark. Besides, the proposed enhancements led to around 2× and 3× improvement in performance and power efficiency, respectively. Moreover, the improved scalability allowed increasing compute resources and delivering around 4× better performance and power efficiency, as compared to the baseline Dynamically Re-programmable Architecture of Gather-scatter Overlay Nodes overlay.

Chemical and Chemo-Enzymatic Syntheses of Glycans Containing Ribitol Phosphate Scaffolding of Matriglycan.

Abstract: Ribitol phosphate modifications to the core M3 O-mannosyl glycan are important for the functional maturation of α-dystroglycan. Three sequentially extended partial structures of the core M3 O-mannosyl glycan including a tandem ribitol phosphate were regio- and stereo-selectively synthesized: Rbo5P-3GalNAcβ, Rbo5P-1Rbo5P-3GalNAcβ, and Xylβ1-4Rbo5P-1Rbo5P-3GalNAcβ (Rbo5P, d-ribitol-5-phosphate; GalNAc, N-acetyl-d-galactosamine; Xyl, d-xylose). Rbo5P-3GalNAcβ with p-nitrophenyl at the aglycon part served as a substrate for ribitol phosphate transferase (FKRP, fukutin-related protein), and its product was glycosylated by the actions of a series of glycosyltransferases, namely, ribitol xylosyltransferase 1 (RXYLT1), β1,4-glucuronyltransferase 1 (B4GAT1), and like-acetyl-glucosaminyltransferase (LARGE). Rbo5P-3GalNAcβ equipped with an alkyne-type aglycon was also active for FKRP. The molecular information obtained on FKRP suggests that Rbo5P-3GalNAcβ derivatives are the minimal units required as the acceptor glycan for Rbo5P transfer and may serve as a precursor for the elongation of the core M3 O-mannosyl glycan.

Accelerating Radiative Transfer Simulation on NVIDIA GPUs with OpenACC

Abstract: To accelerate multiphysics applications, making use of not only GPUs but also FPGAs has been emerging. Multiphysics applications are simulations involving multiple physical models and multiple simultaneous physical phenomena. Operations with different performance characteristics appear in the simulation, making the acceleration of simulation speed using only GPUs difficult. Therefore, we aim to improve the overall performance of the application by using FPGAs to accelerate operations with characteristics which cause lower GPU efficiency. However, the application is currently implemented through multilingual programming, where the computation kernel running on the GPU is written in CUDA and the computation kernel running on the FPGA is written in OpenCL. This method imposes a heavy burden on programmers; therefore, we are currently working on a programming environment that enables to use both accelerators in a GPU–FPGA equipped high-performance computing (HPC) cluster system with OpenACC. To this end, we port the entire code only with OpenACC from the CUDA-OpenCL mixture. On this basis, this study quantitatively investigates the performance of the OpenACC GPU implementation compared to the CUDA implementation for ARGOT, a radiative transfer simulation code for fundamental astrophysics which is a multiphysics application. We observe that the OpenACC implementation achieves performance and scalability comparable to the CUDA implementation on the Cygnus supercomputer equipped with NVIDIA V100 GPUs.

Synthesis of the matriglycan hexasaccharide, -3Xylα1-3GlcAβ1-trimer and its interaction with laminin.

Abstract: Matriglycan, a polysaccharide that is a pivotal part of the core M3 O-mannosyl glycan composed of the repeating disaccharide -3Xylα1-3GlcAβ1-, interacts with laminin to stabilize muscle tissue. We herein report the synthesis of matriglycan-repeating hexasaccharides equipped with an alkyne linker to form glycoconjugates. The key step in the formation of an α-linked xylosyl glycoside was resolved by solvent-specific separation from an anomeric mixture. Successful glycan elongation was regio- and stereoselectively performed to obtain (-3Xylα1-3GlcAβ1)3-O(C2H4O)3CH2CCH and the biotin conjugate. We also investigated interactions between matriglycan hexasaccharides and laminin-G-like domains 4 and 5 of laminin-α2 using saturation transfer difference-NMR. The dissociation constant obtained from bio-layer interferometry was estimated to be 7.5 × 10-8 M. These results indicate that a chemical approach may be applied to the reconstruction of muscle tissue.

Packed SIMD Vectorization of the DRAGON2-CB

Abstract: For over a half-century, computer architects have explored micro-architecture, instruction set architecture, and system architecture to offer a significant performance boost out of a computing chip. In the micro-architecture, multi-processing and multi-threading arose as fusing highly parallel processing and the growth of semiconductor manufacturing technology. It has caused a paradigm shift in computing chips and led to the many-core processor age, such as NVIDIA GPUs, Movidius Myriad, PEZY ZettaScaler, and the project Eyeriss based on a reconfigurable accelerator. Wherein packed SIMD (Single Instruction Multiple Data) vectorizations attract attention, especially from ML (machine learning) applications. It can achieve more energy-efficient computing by reducing computing precision, which is enough for ML applications to obtain the results with low-accuracy calculations. In other words, accuracy-flexible computing needs to allow splitting off one N-bit ALU (Arithmetic Logic Unit) or one N-bit FPU (Floating-Point Unit) into multiple $M$-bit units. For example, a double-precision (64-bit operands width) FPU can be split into two single-precision (32-bit operands width) FPUs, or four half-precision (16-bit operands width) FPUs. Consequently, instead of executing one original operation, a packed SIMD vectorization simultaneously enables executing two or four reduced-precision operations. This article proposes a packed SIMD vectorization approach, which considers the Dynamically Reprogrammable Architecture of Gather-scatter Overlay Nodes-Compact Buffering (DRAGON2-CB) many-core overlay architecture. In particular, this article presents a thorough comparative study between packed SIMD using dual single-precision and quad half-precision FPU-only many-core overlays compared to the non-vectorized double-precision version.

Characterization of Novel Antibodies that Recognize Sialylated Keratan Sulfate and Lacto-N-fucopentaose I on Human Induced Pluripotent Cells: Comparison with Existing Antibodies.

Abstract: In this study, we characterized a series of antibodies generated in C57BL/6 mice (Mus musculus) using the Tic (JCRB1331) human induced pluripotent cell (hiPSC) line as an antigen. This report describes the isolation and characterization of two new antibodies, R-6C (IgM) and R-13E (IgM), and their comparisons with two existing antibodies, R-10G (IgG1) and R-17F (IgG1). Their epitopes were studied by Western blotting after various glycosidase digestions, binding analyses using enzyme-linked immunosorbent assays (ELISAs) and microarrays with various synthetic oligosaccharides. The minimum epitope structures identified were: Siaα2-3Galβ1-3GlcNAc(6S)β1-3Galβ1-4GlcNAc(6S)β1 (R-6C), Fucα1-2Galβ1-3GlcNAcβ1-3Galβ1 (R-13E), Galβ1-4GlcNAc(6S)β1-3Galβ1-4GlcNAc(6S)β1 (R-10G), and Fucα1-2Galβ1-3GlcNAβ1-3Galβ1-4Glc (lacto-N-fucopentaose I) (R-17F) (shown in Fig. 11 using symbol nomenclature). Most glycoprotein epitopes are expressed as O-glycans on podocalyxin, a mucin-type glycoprotein. The common feature of these epitopes is the presence of an N-acetyllactosamine type 1 structure (Galβ1-3GlcNAc) at their nonreducing termini, followed by a type 2 structure (Galβ1-4GlcNAc); this arrangement comprises a type 1-type 2 motif. This motif is also shared by TRA-1-60, a traditional onco-fetal antigen. In contrast, the R-10G epitope has a type 2-type 2 motif. Among these antibodies, R-17F and R-13E exhibit cytotoxic activity toward hiPSCs. R-17F and R-13E exhibit extremely high similarity in the amino acid sequences in their complementarity-determining regions (CDRs), which is consistent with their highly similar glycan recognition. These antibodies are excellent tools for investigating the biological functions of glycoconjugates in hiPSCs/hESCs; they could be useful for the selection, isolation and selective killing of such undifferentiated pluripotent stem cells.

FPGA-based acceleration of stereo matching using OpenCL

Abstract: Stereo vision finds a wide range of applications for robot navigation, advanced driving support system, and autonomous driving in the automotive industry. The disparity map can be obtained through the implementation of stereo vision architecture using stereo matching. A stereo matching algorithm has recently been executed in FPGA. This study is aimed at assessing the stereo matching with the use of Stratix V FPGA and OpenCL framework. The latter refers to a parallel programming framework that enhances productivity by raising the code’s abstraction. Additionally, OpenCL allows for the processing of stereo matching using channel extensions. In the experiment, we partitioned the OpenCL kernel into three smaller kernels to examine the stereo matching on FPGA for computation. Such an approach enables streaming image pixels from the FPGA global memory. A line-buffer is employed to avoid the load-store dependencies caused by memory accesses when streaming the pixels to the window buffer inside the stereo matching kernel. We can achieve a rapid execution time, which is advantageous for real-time implementation, by streaming the image pixels through an OpenCL kernel partitioned using channel extension. The execution time to compute the disparity map using the stereo KITTI dataset with 1242x375 pixels resolution reaches 2.38 ms or 420 fps for 6x6 sliding window size, 2.44 ms or 409 fps for 7x7, and 2.52 ms or 396 fps for 8x8.

Non-classical glycosylation determines intracellular trafficking of APP and Aβ production

Abstract: A primary pathology of Alzheimer’s disease (AD) is Aβ deposition in brain parenchyma and blood vessels, the latter being called cerebral amyloid angiopathy (CAA). Parenchymal amyloid plaques presumably originate from neuronal Aβ precursor protein (APP), but vascular amyloid deposits’ origins remain unclear. Endothelial APP expression in APP-knock-in mice was recently shown to expand CAA pathology, highlighting endothelial APP’s importance. Furthermore, two types of endothelial APP—with and without O-glycans—have been biochemically identified, but only the former is cleaved for Aβ production, indicating the critical relationship between APP O-glycosylation and processing. Here, we analyzed APP glycosylation and its intracellular trafficking in neurons and endothelial cells. Although protein glycosylation is generally believed to precede cell surface trafficking, which was true for neuronal APP, we unexpectedly observed that APP lacking O-glycans is externalized to the endothelial cell surface and transported back to the Golgi apparatus, where it then acquires O-glycans. Knockdown of genes encoding enzymes initiating APP O-glycosylation significantly reduced Αβ production, suggesting this non-classical glycosylation pathway contributes to CAA pathology and is a novel therapeutic target.

O-Glycan-Dependent Interaction between MUC1 Glycopeptide and MY.1E12 Antibody by NMR, Molecular Dynamics and Docking Simulations

Abstract: Anti-mucin1 (MUC1) antibodies have been widely used for breast cancer diagnosis and treatment. This is based on the fact that MUC1 undergoes aberrant glycosylation upon cancer progression, and anti-MUC1 antibodies differentiate changes in glycan structure. MY.1E12 is a promising anti-MUC1 antibody with a distinct specificity toward MUC1 modified with an immature O-glycan (NeuAcα(2-3)Galβ(1-3)GalNAc) on a specific Thr. However, the structural basis for the interaction between MY.1E12 and MUC1 remains unclear. The aim of this study is to elucidate the mode of interaction between MY.1E12 and MUC1 O-glycopeptide by NMR, molecular dynamics (MD) and docking simulations. NMR titration using MUC1 O-glycopeptides suggests that the epitope is located within the O-linked glycan and near the O-glycosylation site. MD simulations of MUC1 glycopeptide showed that the O-glycosylation significantly limits the flexibility of the peptide backbone and side chain of the O-glycosylated Thr. Docking simulations using modeled MY.1E12 Fv and MUC1 O-glycopeptide, suggest that VH mainly contributes to the recognition of the MUC1 peptide portion while VL mainly binds to the O-glycan part. The VH/VL-shared recognition mode of this antibody may be used as a template for the rational design and development of anti-glycopeptide antibodies.

Implementation and Performance Evaluation of Memory System Using Addressable Cache for HPC Applications on HBM2 Equipped FPGAs

Abstract: When we apply field programmable gate arrays (FPGAs) as HPC accelerators, their memory bandwidth presents a significant challenge because it is not comparable to those of other HPC accelerators. In this paper, we propose a memory system for HBM2-equipped FPGAs and HPC applications that uses block RAMs as an addressable cache implemented between HBM2 and an application. This architecture enables data transfer between HBM2 and the cache bulk and allows an application to utilize fast random access on BRAMs. This study demonstrates the implementation and performance evaluation of our new memory system for HPC and HBM2 on an FPGA. Furthermore, we describe the API that can be used to control this system from the host. We implement RISC-V cores in an FPGA as controllers to realize fine-grain data transfer control and to prevent overheads derived from the PCI Express bus. The proposed system is implemented on eight memory channels and achieves 102.7 GB/s of the bandwidth. It overcomes the memory bandwidth of conventional FPGA boards with four channels of DDR4 memory despite using only 8 of 32 channels of the HBM2.

Analysis of Selenoprotein F Binding to UDP‐Glucose:Glycoprotein Glucosyltransferase (UGGT) by a Photoreactive Crosslinker

Abstract: In the endoplasmic reticulum glycoprotein quality control system, UDP‐glucose : glycoprotein glucosyltransferase (UGGT) functions as a folding sensor. Although it is known to form a heterodimer with selenoprotein F (SelenoF), the details of the complex formation remain obscure. A pulldown assay using co‐transfected SelenoF and truncated mutants of human UGGT1 (HUGT1) revealed that SelenoF binds to the TRXL2 domain of HUGT1. Additionally, a newly developed photoaffinity crosslinker was selectively introduced into cysteine residues of recombinant SelenoF to determine the spatial orientation of SelenoF to HUGT1. The crosslinking experiments showed that SelenoF formed a covalent bond with amino acids in the TRXL3 region and the interdomain between βS2 and GT24 of HUGT1 via the synthetic crosslinker. SelenoF might play a role in assessing and refining the disulfide bonds of misfolded glycoproteins in the hydrophobic cavity of HUGT1 as it binds to the highly flexible region of HUGT1 to reach its long hydrophobic cavity. Clarification of the SelenoF‐binding domain of UGGT and its relative position will help predict and reveal the function of SelenoF from a structural perspective.