Cheng Gong's team at the School of Medicine collaborated to discover a glucose-related metabolic small molecule that can effectively inhibit new coronary infections in severe cases

On May 9, a research paper entitled "A glucose-like metabolite deficient in diabetes inhibit cellular entry of SARS-CoV-2" was published in Nature Metabolism (2011) in collaboration with Professor Cheng Gong from Tsinghua University School of Medicine, Professor Zhang Renli from Shenzhen CDC, Researcher Zhao Guangyu from Military Medical Research Institute and Professor Li Liang from Southern University of Science and Technology School of Medicine.Nature Metabolism) in the journal.

Tsinghua University School of Medicine is the first completer. Professor Cheng Gong of Tsinghua University School of Medicine, Professor Zhang Renli of Shenzhen CDC, Researcher Zhao Guangyu of Military Medical Research Institute, and Professor Li Liang of Southern University of Science and Technology School of Medicine were the corresponding authors of the paper.

Liangqin Tong, a 2016 PhD student at Tsinghua University School of Medicine, Xiaoping Xiao, PhD, Tsinghua University School of Medicine, Min Li, PhD, Military Medical Research Institute, and Shisong Fang, PhD, Shenzhen CDC, are tied first authors.

Professor Peng-Hua Wang of the University of Connecticut School of Medicine, Dr. Xiao-Hui Liu of the Metabolism and Lipidomics Platform of Tsinghua University, Professor Wen-Jun Liu of the Institute of Microbiology, Chinese Academy of Sciences, Professor Jin-Cun Zhao of the Affiliated Hospital of Guangzhou Medical University, Researcher Hui Zhong of the Military Medical Research Institute, and Professor Long Yang of Tianjin University of Traditional Chinese Medicine are co-authors of the study.

The rate of severe illness and mortality were significantly higher in diabetic patients than in non-diabetic patients, suggesting that certain factors may exist in diabetic patients that affect their susceptibility to the new coronavirus. Cheng Gong's team at Tsinghua University School of Medicine and co-workers screened a human small molecule metabolite 1,5-Anhydro-D-glucitol (1,5-AG) with a structure similar to that of glucose, which could significantly inhibit NIV infection, and 1,5-AG was strictly negatively correlated with blood glucose concentration in humans. The significant deficiency of 1,5-AG in diabetic patients may be an important reason for the susceptibility of diabetic patients to severe illness and death after neo-coronavirus infection. Further studies showed that supplementation of 1,5-AG to diabetic mice was effective in blocking the onset of severe disease in neocon infection.

In this study, a metabolic small molecule 1,5-AG, which is highly relevant to diabetes, was screened from more than 200 metabolic small molecules in human serum. 1,5-AG is present in normal human serum at concentrations of 100-300 μM and has a steady-state distribution in various organs. 1,5-AG levels in diabetic patients are 5-15 times lower than those in healthy subjects. 1,5-AG exhibits significant anti-neo-coronavirus activity in human cells and 1,5-AG exhibited significant anti-neo-coronavirus ability in human cells and bronchial epithelial-like organs. Mechanistic studies showed that 1,5-AG could bind to the V952 and N955 sites of the HR1 structural domain of the S2 subunit of the neocoronavirus stinger protein and inhibit the infection of host cells by neocoronavirus by inhibiting the formation of S2 subunit 6-HB and affecting the process of virus-cell membrane fusion. Further studies showed that the site of action of 1,5-AG is highly conserved in coronaviruses, and 1,5-AG has significant antiviral effects against neo-coronavirus VOC mutant strains, SARS and MERS, and is a human metabolic small molecule that can broadly inhibit coronavirus infection.

A study of a type II diabetic db/db mouse model revealed that diabetic mice infected with neocoronavirus showed significant weight loss, significantly elevated pulmonary viral load, severe pulmonary pathological damage, and severe phenotypes of neocoronavirus infection. Supplementation of diabetic mice with 1,5-AG significantly inhibited the infection of neocoronavirus, reduced the viral load in the lung by 100-1000 fold, and substantially alleviated the lung tissue lesions. Analysis of the combined clinically relevant data revealed that serum levels of 1,5-AG were significantly lower in patients with severe neocoronavirus than in healthy subjects and non-severe neocoronavirus patients, suggesting that 1,5-AG can influence human susceptibility to neocoronavirus. This study revealed the key molecular mechanisms that make diabetic patients more susceptible to develop severe neo-coronavirus infection and found that supplementation with 1,5-AG helps diabetic patients resist neo-coronavirus infection (Figure 1).

It is worth mentioning that 1,5-AG is significantly decreased in the elderly and is one of the biomarkers of human aging. Whether the reduced 1,5-AG levels in the elderly are associated with their increased rate of severe disease after infection with Neocoron needs further study. In addition, 1,5-AG is widely found in our daily food, among which soybeans are rich in it. Moreover, 1,5-AG is the main active ingredient in several Chinese herbal medicines such as Polygala tenuifolia (Farsi). Based on this research result, it will be possible to develop a 'medicine-food' nutritional strategy to prevent severe pneumonia caused by neocrown infection in the future.

The above research was jointly funded by NSFC Basic Science Center, NSFC Distinguished Youth Fund, NSFC Key Project, Shenzhen Bay Laboratory, National Key R&D Program of Ministry of Science and Technology, Spring Wind Fund of Tsinghua University, Shenzhen Science and Technology Innovation Key Project, Shenzhen Three Projects, and Yunnan Expert Workstation.

Moisture adsorption-desorption full cycle generator! Can drive commercial electronics for long periods of time

Professor Liang-Ti Qu and Assistant Researcher Hu-Hu Cheng from Tsinghua University and Associate Researcher Feng Liu from Institute of Mechanics, Chinese Academy of Sciences have proposed a moisture adsorption-desorption generator (MADG) using three-dimensional (3D) porous ionizable components and surrounding package.The MADG not only exerts moisture adsorption power generation at high RH, but also imparts moisture desorption power generation at low RH based on ion diffusion, which is dominated by ion concentration difference and ion hydration energy dominated. In contrast to MEG with a single adsorption process, the full-cycle MADG integrates adsorption and desorption power generation into a closed-loop process, thus it provides reproducible power generation performance and converts versatile moisture-based energy into electrical energy. the MADG unit can generate high voltages of ~0.5 V and currents of ~100 μA at 100% RH (water absorption) and provide electrical output (~0.5 V) at 15 ± 5% RH (water desorption) The maximum output power density in the MADG is close to 120 mW/m², achieving an excellent trade-off between internal resistance and maximum output power density. In addition, the MADGs can directly provide enough power to drive commercial electronics and electrochemical processes for long periods of time, and to generate continuous full-cycle power based on dynamic relative humidity in real outdoor areas. The related work was published in the international journal Nature Communications under the title "Moisture adsorption-desorption full cycle power generation" (Nature Communications) on.

Research Background

As a recoverable resource, water is not only vital to life, but it is also the largest energy carrier, regulator and balancer on Earth. The ubiquitous hydrological cycle involves a transformation from liquid to gaseous water during evaporation (evaporation from the oceans) and the opposite transformation during condensation (precipitation in clouds), which provides a huge energy exchange (close to 60 × 10¹⁵ W or so per year). This potential energy evolves into many forms and is several orders of magnitude higher than the average power consumption of human activities, yet it is rarely exploited. Although wet gas power generation technologies have recently been developed to meet the energy needs of isolated off-grid areas, wet generators (MEG) tend to rely on a single adsorption process to provide electrical output, reflecting their unsustainable and non-repetitive power generation bottleneck. The alternation between high relative humidity (RH) and low RH throughout the day and the hydrological cycle is a common natural phenomenon. Due to the heavy dependence between high RH and power generation performance of MEG, power generation according to RH variations in the dynamic environment remains a huge challenge to be overcome at this stage.

Preparation and characterization of power generation materials

The schematic structure of the power generation device, its working principle and a photograph of the device are shown in Fig. 2. MADG consists of gold electrodes, a 3D ionizable porous assembly as the power generation material and an encapsulation layer, where the upper gold electrode has holes to allow the entry/removal of water. The power generation film consists of sodium alginate (SA), silica nanofibers (SiO₂) and reduced graphene oxide (rGO), called SAG film. SA acts as a dissociator of mobile Na+ ions. SiO₂ nanofibers facilitate the construction of hierarchical pore structures that promote water molecule and ion transport as well as mechanical stability in water. rGO is used to assemble the 3D conductive backbone and regulate the electrical resistance. electrical resistance. the SAG films were able to provide up to 210% water uptake at ~100% RH and 40 °C test conditions (Fig. 3). the Na element of the SAG films was exclusively in the form of mobile Na⁺ ions with 167% water uptake, indicating the efficient dissociation capability of the SAG films. Benefiting from the interconnected 3D skeleton structure and abundant mobile ions, the SAG films exert excellent ion transport capacity together with the nanostructure and provide an ionic conductivity of 0.11 S m-¹.

Power generation characteristics and performance optimization of MADG

The principle of MADG power generation is hydrated ion diffusion, which is driven by ion concentration difference for hygroscopic power generation and governed by ion hydration energy for desorption power generation. During high RH adsorption power generation, the moisture in the device gradually increases from top to bottom, leading to asymmetric moisture adsorption and Na+ ion dissociation, creating an ion concentration difference and resulting in power output. As the device saturates with adsorption, the ions will tend to be uniformly distributed in the final state. Subsequently, the saturated MADG is ready for moisture desorption for power generation. By directionally adsorbing water molecules from the atmosphere, the MADG can generate open-circuit voltages up to 0.5 V and short-circuit currents approaching 100 μA at 100% RH. The hydrated MADG spontaneously desorbs water molecules at low RH (15 ± 5% RH), generating an open-circuit voltage of about 0.5 V and a short-circuit current of about 50 μA. Compared to moisture generators based on a single adsorption process, the MADG is no longer limited by bottlenecks such as adsorption equilibrium and high dependence on high RH. The interconnected SAG film with 77% high porosity has an impressive water absorption capacity, water molecule diffusion coefficient, zeta potential and ionic conductivity, reflecting its excellent water molecule transport, ion dissociation and diffusion.

Generator system and MADG application

The authors propose ion diffusion desorption power generation based on ion hydration energy, which relies on two main points.

(1) The number of hydrated Na⁺ ions remains constant, and only the number of surrounding water molecules decreases with the decrease in water content.

(2) The ionic hydration energy of hydrated Na⁺ ions combined with more water molecules is lower than that of hydrated Na⁺ ions combined with fewer water molecules.

During spontaneous adsorption, the change in chemical potential energy, reasonably used as an energy source, can be converted into electrical energy and heat of adsorption. the MADG device is able to extract thermal energy from the surrounding environment by converting sensible heat into latent heat during desorption, thus generating electrical energy and a change in chemical potential energy. Therefore, the energy used for MADG power generation is green and renewable. In addition, the authors demonstrated the scaling performance of the MADG by connecting 21 units in series to produce voltages up to about 11 V, increasing linearly to an average of 0.53 V per unit. the current output of the 16 parallel units was boosted to about 1.3 mA. the power provided by the MADG was able to charge commercial capacitors of 0.47, 47 or 470 mF to 3 V using an integrated 6 × 3 array. the MADG unit is attached to a rotatable board that intelligently self-converts power generation and provides a continuous voltage output of ~0.5-0.6 V, illustrating its suitability in dynamic humidity environments.

Authors of a moisture adsorption-desorption generator (MADG) based on porous ionizable components that spontaneously adsorbs moisture at high relative humidity and desorbs it at low relative humidity to produce a cyclic electrical output. the MADG device produces a high voltage of ~0.5 V and a current of 100 μA at 100% relative humidity (RH), provides an electrical output at 15 ± 5% RH (~0.5 V and ~50 μA) and provide a maximum output power density close to 120 mW/m². This MADG device can conduct enough power to illuminate streetlights in outdoor applications and directly drive electrochemical processes. This work provides a closed-loop pathway for versatile moisture-based energy conversion.

Tsinghua team collaborates on continuous learning method for cryoelectron microscopy particle selection

On May 5, a team of Associate Professor Xue-Ming Li from the School of Life Sciences, Tsinghua University, a team of Professor Yuan Shen from the Department of Electrical Engineering, Tsinghua University, and a team of Professor Jiansheng Chen from the School of Computer and Communication Engineering, University of Science and Technology Beijing, jointly published in Nature Communications (Nature CommunicationsA research paper titled "EPicker is an exemplar-based continual learning approach for knowledge accumulation in cryoEM particle picking" has been published in the journal.

The paper reports the application of a paradigm-driven continuous learning approach to protein particle picking, extending the ability of the detection model to recognize biomolecules by continuously learning new knowledge during the particle picking process. The importance of developing a continuous learning approach is that it can enable artificial deep neural networks to have a human-like learning style and continuously learn new knowledge and skills in use, thus continuously enhancing their capabilities. EPicker can pick a wide range of biological objects such as protein particles, vesicles and fibers after training.

Xue-Ming Li, Associate Professor, School of Life Sciences, Tsinghua University; Yuan Shen, Professor, Department of Electrical Engineering, Tsinghua University; and Jiansheng Chen, Professor, School of Computer and Communication Engineering, University of Science and Technology Beijing are the co-corresponding authors of this paper. Xingyu Zhang, a 2019 master's student in the Department of Electrical Engineering, Tsinghua University, and Tianfang Zhao, a 2020 master's student in the Department of Electrical Engineering, Tsinghua University, are the co-first authors of the paper.

Research Background

In recent years, deep learning has gradually become a common method for particle selection in cryoelectron microscopy image processing processes. However, existing deep learning-based particle selection methods are unable to dynamically accumulate new knowledge into the model during new data training. That is, existing models are trained on new samples, and although they can obtain good performance on the latest data, they often cannot maintain their particle picking accuracy on the old data. In addition, existing methods are trained to produce generic models on specific datasets, which are expensive to store and compute when new training data have to be added, greatly limiting their recognition ability and accuracy on unseen data. Therefore, we need to improve the way and method of training existing deep learning networks. At the same time, existing cryoelectron microscopy facilities are generating large amounts of new data every day. It would be of great importance to develop a continuous learning technique that allows deep neural networks to continuously learn and accumulate new features in new data and continuously enhance the recognition of biological sample images during continuous application for the development of modern automated cryo-electron microscopy systems.

Innovation Research

To address the shortcomings of existing methods, the research team designed a particle selection algorithm based on continuous learning, which can continuously accumulate new particle selection knowledge during the training of the neural network and improve the particle selection ability of the generic model. The algorithm, by designing a two-way network structure (Figure 4) and incorporating knowledge distillation, history playback, regularization, and sparse labeling methods, continuously accumulates knowledge of new samples into the generic model without forgetting the old knowledge. This is a good solution to the problem that the model cannot pick old data samples after training on new data. Based on these algorithms, the research team has developed a new software system called EPicker. To further extend the applicability of the method, the team designed corresponding picking algorithms for a wide range of biological objects, including picking a variety of different biological objects such as vesicles and fibers, supporting both biased and unbiased particle picking methods to meet the different needs of users, and so on. The effectiveness and superiority of EPicker was verified by conducting extensive experiments on a representative and challenging dataset and comparing it with the more popular particle picking methods (Figure 5). The experimental results show that EPicker can obtain protein particle selection results with high accuracy, high recall and high generalization ability through an efficient and highly automated continuous learning process.

The above work has received funding support from the Key Research and Development Program of the Ministry of Science and Technology, the National Natural Science Foundation of China, the Beijing Center for High Precision Innovation in Structural Biology, the Beijing Center for Frontier Research in Biological Structures, the Joint Center for Life Sciences, and the Beijing National Research Center for Information Science and Technology.

Wang Yong's group in the Department of Geology reveals that climate effects of historical land use exacerbate global economic inequalities

Recently, Associate Professor Wang Yong's group in the Department of Earth System Science at Tsinghua University combined with the Sixth Coupled Model Comparison Program (CMIP6) multi-model, multi-ensemble climate simulation experiment and temperature-economic assessment model to study the impact of historical land use since 1850 on global surface mean temperature and day-by-day temperature variability through biogeophysical and biogeochemical processes, and further analyzed its impact on the global economy. The related results are titled "Contrasting Influences of Biogeophysical and Biogeochemical Impacts of Historical Land Use on Global Economic Inequality," was published on May 5 in Nature Communications (Nature Communications) Journal publication.

Shu Liu, a 2019 PhD student in the Department of Geosciences, Tsinghua University, is the first author of the paper, and Associate Professor Yong Wang, Department of Geosciences, Tsinghua University, is the corresponding author. Linyi Wei, a PhD student in the Department of Geosciences, Tsinghua University, Class of 2020, Bin Wang, a Distinguished Visiting Professor in the Department of Geosciences, Tsinghua University and a researcher at the Institute of Atmospheric Physics, Chinese Academy of Sciences, and Associate Professor Le Yu, Department of Geosciences, Tsinghua University, are co-authors of the paper.

Research Background

Climate change affects all aspects of agricultural production, energy supply, labor production, and human health, which in turn affects macroeconomic development. Previous observational studies have confirmed that per capita GDP growth is greatest when a country's average temperature is in the optimal temperature range; either too high or too low temperatures are unsuitable for economic development. It has been analyzed that greenhouse gas and anthropogenic aerosol emissions can change the annual average temperature and thus affect the global economy. In addition to GHG and anthropogenic aerosol emissions, land use/cover change (LULCC) is one of the important ways in which human activities affect climate change due to the growing human demand for food, clothing, shelter, and transportation, and the reclamation of large areas of natural vegetation for human land use such as cropland, pasture, and cities. Land use can produce both biogeophysical effects (altering surface albedo, Bowen ratio, etc.) and biogeochemical effects (causing greenhouse gas emissions, etc.), and its biogeophysical effects vary with the type of surface disturbed and latitude. Therefore, the impact of historical land use on global temperature and economy is still unclear.

Innovation Research

To address these issues, the research group designed a historical land use-temperature response-economic impact research route. Through CMIP6 climate simulation experiments, it was found that the biogeophysical effect of historical land use decreases the temperature in most of the global regions (especially in the middle and high latitudes of the Northern Hemisphere), while the biogeochemical effect causes a global-scale temperature increase. In most parts of the globe, the biogeochemical effects of land use dominate the combined effects of both, causing an increase in the annual mean temperature in these regions (Fig. 6a), so that economic development in developing countries with hot climates is negatively affected by them, but is favorable for developed countries with cold climates. Such differential economic impacts exacerbate global economic inequalities (Figure 6). In addition, it was found that the combined effect of land use increases the number of extreme temperature events in developing countries but decreases the number of extreme temperature events in developed countries, which further exacerbates global economic inequality (Figure 7).

The study assessed the economic impact caused by the climate effect of historical land use, which is an important guideline for the rational planning of future land use and the formulation of climate change mitigation policies.

The above research was supported by the Key Research and Development Program of the Ministry of Science and Technology and the National Natural Science Foundation of China.

Link to the paper.

[1] https://www.nature.com/articles/s42255-022-00567-z

[2] https://www.nature.com/articles/s41467-022-30156-3

[3] https://www.nature.com/articles/s41467-022-29994-y

[4] https://www.nature.com/articles/s41467-022-30145-6

--Tsinghua University News Network, Frontiers of Polymer Science