Volume: 18, Issue: 4

Original Research

G9a/GLP-sensitivity of H3K9me2 Demarcates Two Types of Genomic Compartments

Zixiang Yan, Luzhang Ji, Xiangru Huo, Qianfeng Wang, Yuwen Zhang, Bo Wen

In the nucleus, chromatin is folded into hierarchical architecture that is tightly linked to various nuclear functions. However, the underlying molecular mechanisms that confer these architectures remain incompletely understood. Here, we investigated the functional roles of H3 lysine 9 dimethylation (H3K9me2), one of the abundant histone modifications, in three-dimensional (3D) genome organization. Unlike in mouse embryonic stem cells, inhibition of methyltransferases G9a and GLP in differentiated cells eliminated H3K9me2 predominantly at A-type (active) genomic compartments, and the level of residual H3K9me2 modifications was strongly associated with B-type (inactive) genomic compartments. Furthermore, chemical inhibition of G9a/GLP in mouse hepatocytes led to decreased chromatin-nuclear lamina interactions mainly at G9a/GLP-sensitive regions, increased degree of genomic compartmentalization, and up-regulation of hundreds of genes that were associated with alterations of the 3D chromatin. Collectively, our data demonstrated essential roles of H3K9me2 in 3D genome organization.

Page 359-370

Original Research

m6A Regulates Liver Metabolic Disorders and Hepatogenous Diabetes

Yuhuan Li, Qingyang Zhang, Guanshen Cui, Fang Zhao, Xin Tian, Bao-Fa Sun, Ying Yang, Wei Li

N6-methyladenosine (m6A) is one of the most abundant modifications on mRNAs and plays important roles in various biological processes. The formation of m6A is catalyzed by a methyltransferase complex (MTC) containing a key factor methyltransferase-like 3 (Mettl3). However, the functions of Mettl3 and m6A modification in hepatic lipid and glucose metabolism remain unclear. Here, we showed that both Mettl3 expression and m6A level increased in the livers of mice with high fat diet (HFD)-induced metabolic disorders. Overexpression of Mettl3 aggravated HFD-induced liver metabolic disorders and insulin resistance. In contrast, hepatocyte-specific knockout of Mettl3 significantly alleviated HFD-induced metabolic disorders by slowing weight gain, reducing lipid accumulation, and improving insulin sensitivity. Mechanistically, Mettl3 depletion-mediated m6A loss caused extended RNA half-lives of metabolism-related genes, which consequently protected mice against HFD-induced metabolic syndrome. Our findings reveal a critical role of Mettl3-mediated m6A in HFD-induced metabolic disorders and hepatogenous diabetes.
N6-甲基腺嘌呤（N6-methyladenosine, m6A）是mRNA上最普遍存在的甲基化修饰之一，由甲基转移酶3（Methyltransferase-like 3, Mettl3)等甲基转移酶复合物（Methyltransferase Complex, MTC）催化形成，参与调控多种RNA代谢过程及生物学功能。然而，Mettl3和m6A修饰在肝脏的脂质和葡萄糖代谢中的作用尚不清楚。本研究报道了在高脂饮食（High Fat Diet, HFD）诱导的代谢紊乱的小鼠肝脏中Mettl3和m6A水平升高，肝脏条件性Mettl3过表达加重了HFD诱导的肝脏代谢紊乱和胰岛素抵抗，肝脏条件性敲除Mettl3则显著减轻了HFD诱导的代谢综合征，主要表现为体重增加减慢、脂质累积减少以及胰岛素敏感性的改善。进一步机制研究表明，Mettl3敲除介导的m6A缺失使得代谢相关基因mRNA的半衰期延长，保护小鼠不受HFD诱导的代谢综合征的影响。该发现揭示了Mettl3介导的m6A在HFD引发的肝脏代谢紊乱和肝源性糖尿病中的重要调控作用。

Page 371-383

Original Research

The mRNA Binding Proteome of Proliferating and Differentiated Muscle Cells

Monika Hiller, Miriam Geissler, George Janssen, Peter van Veelen, Annemieke Aartsma-Rus, Pietro Spitali

Muscle formation is a coordinated process driven by extensive gene expression changes where single cells fuse together to form multinucleated muscle fibers. Newly synthesized mRNAs are then regulated by RNA binding proteins (RBPs), affecting post-transcriptional transcript metabolism. Here, we determined how large-scale gene expression changes affect the catalog of RBPs by studying proliferating and differentiated muscle cells in healthy and dystrophic conditions. Transcriptomic analysis showed that the expression of more than 7000 genes was affected during myogenesis. We identified 769 RBPs, of which 294 were muscle-specific and 49 were uniquely shared with cardiomyocytes. A subset of 32 RBPs (half of which were muscle-specific) was found to be preferentially associated with target mRNAs in either myoblasts (MBs) or myotubes (MTs). A large proportion of catalytic proteins were bound to mRNAs even though they lack classical RNA binding domains. Finally, we showed how the identification of cell-specific RBPs enabled the identification of biomarkers that can separate healthy individuals from dystrophic patients. Our data show how interactome data can shed light on new basic RNA biology as well as provide cell-specific data that can be used for diagnostic purposes.

Page 384-394

Original Research

Large-scale Proteomic and Phosphoproteomic Analyses of Maize Seedling Leaves During De-etiolation

Zhi-Fang Gao, Zhuo Shen, Qing Chao, Zhen Yan, Xuan-Liang Ge, Tiancong Lu, Haiyan Zheng, Chun-Rong Qian, Bai-Chen Wang

De-etiolation consists of a series of developmental and physiological changes that a plant undergoes in response to light. During this process light, an important environmental signal, triggers the inhibition of mesocotyl elongation and the production of photosynthetically active chloroplasts, and etiolated leaves transition from the “sink” stage to the “source” stage. De-etiolation has been extensively studied in maize (Zea mays L.). However, little is known about how this transition is regulated. In this study, we described a quantitative proteomic and phosphoproteomic atlas of the de-etiolation process in maize. We identified 16,420 proteins in proteome, among which 14,168 proteins were quantified. In addition, 8746 phosphorylation sites within 3110 proteins were identified. From the combined proteomic and phosphoproteomic data, we identified a total of 17,436 proteins. Only 7.0% (998/14,168) of proteins significantly changed in abundance during de-etiolation. In contrast, 26.6% of phosphorylated proteins exhibited significant changes in phosphorylation level; these included proteins involved in gene expression and homeostatic pathways and rate-limiting enzymes involved in photosynthetic light and carbon reactions. Based on phosphoproteomic analysis, 34.0% (1057/3110) of phosphorylated proteins identified in this study contained more than 2 phosphorylation sites, and 37 proteins contained more than 16 phosphorylation sites, indicating that multi-phosphorylation is ubiquitous during the de-etiolation process. Our results suggest that plants might preferentially regulate the level of posttranslational modifications (PTMs) rather than protein abundance for adapting to changing environments. The study of PTMs could thus better reveal the regulation of de-etiolation.
黄化苗的复绿过程是植物为响应光照经历的一系列发育和生理过程。在整个过程中，植物需要快速响应光信号并完成光形态建成，使黄化叶片复绿并实现由“库”阶段向“源”阶段的转换。在玉米中，复绿过程已得到很充分的研究，但是对“库-源”转换的调控研究较少。在本研究中，我们描述了玉米复绿过程的定量蛋白质组和磷酸化蛋白质组图谱，提供了迄今为止最全面的玉米复绿叶片中蛋白质丰度和磷酸化的动态分析。我们结合2D-HPLC-MS/MS测序和3D-HPLC-MS/MS深度测序，共鉴定到16,420个蛋白，其中14168实现定量分析。另外，通过磷酸化富集结合HPLC-MS/MS测序分析，鉴定到3110个蛋白上的8746个磷酸化位点，其中1507个蛋白存在三个以上的磷酸化位点，表明蛋白的多磷酸化修饰在复绿过程中普遍存在。对蛋白质组和磷酸化修饰组学数据深入挖掘发现，仅有7%（998）的蛋白在复绿过程中蛋白量发生显示变化，而超过26%的磷酸化蛋白的磷酸化水平在复绿过程中受到调控。大量磷酸化蛋白的多个位点的磷酸化水平在黄化苗复绿过程中发生动态而复杂的变化，暗示植物可能通过精细调控蛋白的复杂磷酸化修饰响应光环境变化。我们的结果表明，相较于调控蛋白量的变化，调节翻译后修饰变化是植物应对周围环境因素变化时更为有效的手段。

Page 397-414

Original Research

Transcriptomic and Proteomic Analysis of Mannitol-metabolism-associated Genes in Saccharina japonica

Shan Chi, Guoliang Wang, Tao Liu, Xumin Wang, Cui Liu, Yuemei Jin, Hongxin Yin, Xin Xu, Jun Yu

As a carbon-storage compound and osmoprotectant in brown algae, mannitol is synthesized and then accumulated at high levels in Saccharina japonica (Sja); however, the underlying control mechanisms have not been studied. Our analysis of genomic and transcriptomic data from Sja shows that mannitol metabolism is a cyclic pathway composed of four distinct steps. A mannitol-1-phosphate dehydrogenase (M1PDH2) and two mannitol-1-phosphatases (M1Pase1 and MIPase2) work together or in combination to exhibit full enzymatic properties. Based on comprehensive transcriptomic data from different tissues, generations, and sexes as well as under different stress conditions, coupled with droplet digital PCR (ddPCR) and proteomic confirmation, we suggest that SjaM1Pase1 plays a major role in mannitol biosynthesis and that the basic mannitol anabolism and the carbohydrate pool dynamics are responsible for carbon storage and anti-stress mechanism. Our proteomic data indicate that mannitol metabolism remains constant during diurnal cycle in Sja. In addition, we discover that mannitol-metabolism-associated (MMA) genes show differential expression between the multicellular filamentous (gametophyte) and large parenchymal thallus (sporophyte) generations and respond differentially to environmental stresses, such as hyposaline and hyperthermia conditions. Our results indicate that the ecophysiological significance of such differentially expressed genes may be attributable to the evolution of heteromorphic generations (filamentous and thallus) and environmental adaptation of Laminariales.
甘露醇（mannitol）作为褐藻中的储碳化合物和渗透保护剂，在海带（Sja）中合成并大量积累但其潜在的控制机制尚不清晰。本研究对海带基因组和转录组数据的分析表明，甘露醇代谢是由四个不同步骤组成的循环途径。一种甘露醇-1-磷酸脱氢酶（M1PDH2）和两种甘露醇-1-磷酸酶（M1Pase）共同作用呈现出完整的酶学特性。根据不同组织、不同世代、不同性别以及不同胁迫条件下的转录组数据，结合微滴式数字PCR（ddPCR）和蛋白质组数据验证，发现SjaM1Pase1在甘露醇生物合成过程中起主要作用，而基础甘露醇合成代谢和碳水化合物池动态变化可能负责碳储存和抗压力机制。蛋白质组学数据表明，甘露醇代谢在海带的昼夜周期中保持不变。此外，研究还发现甘露醇代谢相关基因（MMA）在多细胞丝状体（配子体）和大型薄壁叶状体（孢子体）世代之间呈现表达差异，并对环境胁迫（如低盐和高温条件）有不同的响应。本研究结果表明，这些差异表达基因的生理生态意义可能与海带异形世代（丝状体和叶状体）的进化和海带目藻类的环境适应有关。

Page 415-429

Original Research

Tissue-specific Gene Expression Changes Are Associated with Aging in Mice

Akash Srivastava, Emanuel Barth, Maria A. Ermolaeva, Madlen Guenther, Christiane Frahm, Manja Marz, Otto W. Witte

Aging is a complex process that can be characterized by functional and cognitive decline in an individual. Aging can be assessed based on the functional capacity of vital organs and their intricate interactions with one another. Thus, the nature of aging can be described by focusing on a specific organ and an individual itself. However, to fully understand the complexity of aging, one must investigate not only a single tissue or biological process but also its complex interplay and interdependencies with other biological processes. Here, using RNA-seq, we monitored changes in the transcriptome during aging in four tissues (including brain, blood, skin and liver) in mice at 9 months, 15 months, and 24 months, with a final evaluation at the very old age of 30 months. We identified several genes and processes that were differentially regulated during aging in both tissue-dependent and tissue-independent manners. Most importantly, we found that the electron transport chain (ETC) of mitochondria was similarly affected at the transcriptome level in the four tissues during the aging process. We also identified the liver as the tissue showing the largest variety of differentially expressed genes (DEGs) over time. Lcn2 (Lipocalin-2) was found to be similarly regulated among all tissues, and its effect on longevity and survival was validated using its orthologue in Caenorhabditis elegans. Our study demonstrated that the molecular processes of aging are relatively subtle in their progress, and the aging process of every tissue depends on the tissue’s specialized function and environment. Hence, individual gene or process alone cannot be described as the key of aging in the whole organism.

Page 430-442

Original Research

Global Analysis of Gene Expression Profiles Provides Novel Insights into the Development and Evolution of the Large Crustacean Eriocheir sinensis

Jun Wang, Xiaowen Chen, Funan He, Xiao Song, Shu Huang, Wucheng Yue, Yipei Chen, Zhixi Su, Chenghui Wang

Chinese mitten crab (Eriocheir sinensis) is an important aquaculture species in Crustacea. Functional analysis, although essential, has been hindered due to the lack of sufficient genomic or transcriptomic resources. In this study, transcriptome sequencing was conducted on 59 samples representing diverse developmental stages (fertilized eggs, zoea, megalopa, three sub-stages of larvae, juvenile crabs, and adult crabs) and different tissues (eyestalk, hepatopancreas, and muscle from juvenile crabs, and eyestalk, hepatopancreas, muscle, heart, stomach, gill, thoracic ganglia, intestine, ovary, and testis from adult crabs) of E. sinensis. A comprehensive reference transcriptome was assembled, including 19,023 protein-coding genes. Hierarchical clustering based on 128 differentially expressed cuticle-related genes revealed two distinct expression patterns during the early larval developmental stages, demonstrating the distinct roles of these genes in “crab-like” cuticle formation during metamorphosis and cuticle calcification after molting. Phylogenetic analysis of 1406 one-to-one orthologous gene families identified from seven arthropod species and Caenorhabditis elegans strongly supported the hypothesis that Malacostraca and Branchiopoda do not form a monophyletic group. Furthermore, Branchiopoda is more phylogenetically closely related to Hexapoda, and the clade of Hexapoda and Branchiopoda and the clade of Malacostraca belong to the Pancrustacea. This study offers a high-quality transcriptome resource for E. sinensis and demonstrates the evolutionary relationships of major arthropod groups. The differentially expressed genes identified in this study facilitate further investigation of the cuticle-related gene expression networks which are likely associated with “crab-like” cuticle formation during metamorphosis and cuticle calcification after molting.

Page 443-454

Original Research

The Wolfiporia cocos Genome and Transcriptome Shed Light on the Formation of Its Edible and Medicinal Sclerotium

Hongmei Luo, Jun Qian, Zhichao Xu, Wanjing Liu, Lei Xu, Ying Li, Jiang Xu, Jianhong Zhang, Xiaolan Xu, Chang Liu, Liu He, Jianqin Li, Chao Sun, Francis Martin, Jingyuan Song, Shilin Chen

Wolfiporia cocos (F. A. Wolf) has been praised as a food delicacy and medicine for centuries in China. Here, we present the genome and transcriptome of the Chinese strain CGMCC5.78 of W. cocos. High-confidence functional prediction was made for 9277 genes among the 10,908 total predicted gene models in the W. cocos genome. Up to 2838 differentially expressed genes (DEGs) were identified to be related to sclerotial development by comparing the transcriptomes of mycelial and sclerotial tissues. These DEGs are involved in mating processes, differentiation of fruiting body tissues, and metabolic pathways. A number of genes encoding enzymes and regulatory factors related to polysaccharide and triterpenoid production were strikingly regulated. A potential triterpenoid gene cluster including the signature lanosterol synthase (LSS) gene and its modified components were annotated. In addition, five nonribosomal peptide synthase (NRPS)-like gene clusters, eight polyketide synthase (PKS) gene clusters, and 15 terpene gene clusters were discovered in the genome. The differential expression of the velevt family proteins, transcription factors, carbohydrate-active enzymes, and signaling components indicated their essential roles in the regulation of fungal development and secondary metabolism in W. cocos. These genomic and transcriptomic resources will be valuable for further investigations of the molecular mechanisms controlling sclerotial formation and for its improved medicinal applications.

Page 455-467

Original Research

Denoising Autoencoder, A Deep Learning Algorithm, Aids the Identification of A Novel Molecular Signature of Lung Adenocarcinoma

Jun Wang, Xueying Xie, Junchao Shi, Wenjun He, Qi Chen, Liang Chen, Wanjun Gu, Tong Zhou

Precise biomarker development is a key step in disease management. However, most of the published biomarkers were derived from a relatively small number of samples with supervised approaches. Recent advances in unsupervised machine learning promise to leverage very large datasets for making better predictions of disease biomarkers. Denoising autoencoder (DA) is one of the unsupervised deep learning algorithms, which is a stochastic version of autoencoder techniques. The principle of DA is to force the hidden layer of autoencoder to capture more robust features by reconstructing a clean input from a corrupted one. Here, a DA model was applied to analyze integrated transcriptomic data from 13 published lung cancer studies, which consisted of 1916 human lung tissue samples. Using DA, we discovered a molecular signature composed of multiple genes for lung adenocarcinoma (ADC). In independent validation cohorts, the proposed molecular signature is proved to be an effective classifier for lung cancer histological subtypes. Also, this signature successfully predicts clinical outcome in lung ADC, which is independent of traditional prognostic factors. More importantly, this signature exhibits a superior prognostic power compared with the other published prognostic genes. Our study suggests that unsupervised learning is helpful for biomarker development in the era of precision medicine.

Page 468-480

Application Note

mrMLM v4.0.2: An R Platform for Multi-locus Genome-wide Association Studies

Ya-Wen Zhang, Cox Lwaka Tamba, Yang-Jun Wen, Pei Li, Wen-Long Ren, Yuan-Li Ni, Jun Gao, Yuan-Ming Zhang

Previous studies have reported that some important loci are missed in single-locus genome-wide association studies (GWAS), especially because of the large phenotypic error in field experiments. To solve this issue, multi-locus GWAS methods have been recommended. However, only a few software packages for multi-locus GWAS are available. Therefore, we developed an R software named mrMLM v4.0.2. This software integrates mrMLM, FASTmrMLM, FASTmrEMMA, pLARmEB, pKWmEB, and ISIS EM-BLASSO methods developed by our lab. There are four components in mrMLM v4.0.2, including dataset input, parameter setting, software running, and result output. The fread function in data.table is used to quickly read datasets, especially big datasets, and the doParallel package is used to conduct parallel computation using multiple CPUs. In addition, the graphical user interface software mrMLM.GUI v4.0.2, built upon Shiny, is also available. To confirm the correctness of the aforementioned programs, all the methods in mrMLM v4.0.2 and three widely-used methods were used to analyze real and simulated datasets. The results confirm the superior performance of mrMLM v4.0.2 to other methods currently available. False positive rates are effectively controlled, albeit with a less stringent significance threshold. mrMLM v4.0.2 is publicly available at BioCode (https://bigd.big.ac.cn/biocode/tools/BT007077) or R (https://cran.r-project.org/web/packages/mrMLM.GUI/index.html) as an open-source software.
蛋白质琥珀酰化是一种生物化学反应，其中琥珀基(-CO-CH2-CH2-CO-)连接到蛋白质分子的赖氨酸残基上。赖氨酸琥珀酰化在活细胞中起着重要的调节作用。然而，赖氨酸琥珀酰化的底物位点特异性难以在实验中确定，导致这一领域的研究受到了限制。为了促进这对这种蛋白翻译后修饰的理解，一些工具已经被开发用于计算、鉴定琥珀酰赖氨酸位点。在本研究中，我们开发了一种基于氨基酸组成的赖氨酸琥珀酰化位点底物特异性研究方法。利用从公共资源收集的、经实验验证的赖氨酸琥珀酰化位点信息，使用两个样品标识程序发现了琥珀酰化和非琥珀酰化位点之间的位置特异性氨基酸组成的显着差异。这些发现使得能够采用有效的机器学习方法，即支持向量机，以训练不仅具有氨基酸组成，而且具有k间隔的氨基酸对组成的预测模型。使用十倍交叉验证方法选择最佳模型后，利用从已发表的研究文章中手动提取的独立数据集进行验证，结果显示所选模型的性能明显优于现有工具。最后，选择的模型用于开发基于Web的工具SuccSite，以辅助蛋白质琥珀酰化的研究。该网站使用两种蛋白质作为案例研究，以证明琥珀酰化位点的有效预测。我们将通过集成更多实验数据集来定期更新SuccSite。SuccSite可从http://csb.cse.yzu.edu.tw/SuccSite/免费访问。

Page 481-487

Corrigendum

Corrigendum to “Large-scale Identification and Time-course Quantification of Ubiquitylation Events During Maize Seedling De-etiolation” [Genomics Proteomics Bioinformatics 17 (6) (2019) 603–622]

Yue-Feng Wang, Qing Chao, Zhe Li, Tian-Cong Lu, Hai-Yan Zheng, Cai-Feng Zhao, Zhuo Shen, Xiao-Hui Li, Bai-Chen Wang

Page 488