Alternative transcription start sites contribute to acute-stress–induced transcriptome response in human skeletal muscle

More than half of human protein-coding genes have an alternative transcription start site (TSS). We aimed to investigate the contribution of alternative TSSs to the acute-stress–induced transcriptome response in human tissue (skeletal muscle) using the cap analysis of gene expression approach. TSSs were examined at baseline and during recovery after acute stress (a cycling exercise). We identified 44,680 CAGE TSS clusters (including 3,764 first defined) belonging to 12,268 genes and annotated for the first time 290 TSSs belonging to 163 genes. The transcriptome dynamically changes during the first hours after acute stress; the change in the expression of 10% of genes was associated with the activation of alternative TSSs, indicating differential TSSs usage. The majority of the alternative TSSs do not increase proteome complexity suggesting that the function of thousands of alternative TSSs is associated with the fine regulation of mRNA isoform expression from a gene due to the transcription factor-specific activation of various alternative TSSs. We identified individual muscle promoter regions for each TSS using muscle open chromatin data (ATAC-seq and DNase-seq). Then, using the positional weight matrix approach we predicted time course activation of “classic” transcription factors involved in response of skeletal muscle to contractile activity, as well as diversity of less/un-investigated factors. Transcriptome response induced by acute stress related to activation of the alternative TSSs indicates that differential TSSs usage is an essential mechanism of fine regulation of gene response to stress stimulus. A comprehensive resource of accurate TSSs and individual promoter regions for each TSS in muscle was created. This resource together with the positional weight matrix approach can be used to accurate prediction of TFs in any gene(s) of interest involved in the response to various stimuli, interventions or pathological conditions in human skeletal muscle. Ten amateur endurance-trained athletes (long distance runners, cyclists and cross country skiers, median age 32 years [interquartile range, 30–36 years]; weight 75 kg [71–78 kg]; V’O2max/kg [maximal pulmonary O2 consumption rate] 58 ml/min/kg [54–60 ml/min/kg of body mass]) were involved in our study. Each subject carried out an intermittent exercise (60 min, [3 min at intensity 50% of lactate threshold [LT4 , power at blood lactate 4 mmol/l] + 2 min, 100% LT4] x 12) on a cycle ergometer 2 h after a standardized breakfast (3582 kJ; 22 g protein, 154 g carbohydrates and 16 g fat). Subjects ate a standardized lunch (3714 kJ; 45 g protein, 183 g carbohydrates and 27 g fat) 1 h 15 min after an intermittent exercise. Biopsy samples were taken under local anesthesia (2 mL 2% lidocaine) using a Bergstrome needle with aspiration from the m. vastus lateralis prior to, 2 min, 1 h, 3 h, and 6 h after an intermittent exercise (1st, 2d, and 3d from the one leg, 4st and 5st from another leg). For each subject samples from time points: prior to, 1 h, 3 h, and 6 h after an exercise were analyzed in one run. Additionally, we analyzed all samples from time point 2 min after an exercise in a separate run. To improve the quality of identification of CAGE TSSs, all data were used for calculation. However, differential expressions of CAGE TSSs and genes as well as differential TSSs usage were evaluated for time points: prior to, 1 h, 3 h, and 6 h after an exercise only.

超过一半的人类蛋白质编码基因存在可变转录起始位点（Transcription Start Site, TSS）。本研究旨在利用基因表达帽式分析（Cap Analysis of Gene Expression, CAGE）方法，探究可变TSS对人体骨骼肌组织急性应激诱导的转录组应答的调控贡献。研究于基线状态下以及急性应激（骑行运动）后的恢复阶段对TSS进行检测。

本研究共鉴定出隶属于12268个基因的44680个CAGE TSS簇（其中3764个为首次定义），并首次注释了隶属于163个基因的290个TSS。急性应激后的最初数小时内，转录组发生动态变化；10%的基因的表达变化与可变TSS的激活相关，提示存在差异性TSS使用模式。大多数可变TSS并不会增加蛋白质组的复杂性，这表明数千个可变TSS的功能与通过转录因子特异性激活不同可变TSS，从而精细调控单个基因的mRNA同工型表达密切相关。

本研究利用骨骼肌开放染色质数据（ATAC-seq与DNase-seq），为每个TSS鉴定出专属的肌肉启动子区域。随后借助位置权重矩阵（Position Weight Matrix, PWM）方法，我们预测了参与骨骼肌收缩应答的"经典"转录因子的时序激活模式，同时也鉴定出多种尚未被充分研究的转录因子。

急性应激诱导的转录组应答与可变TSS的激活相关，这表明差异性TSS使用是精细调控基因对应激刺激应答的关键机制。本研究构建了一套涵盖骨骼肌中精准TSS及每个TSS专属启动子区域的综合数据集资源。该资源结合位置权重矩阵方法，可用于精准预测人类骨骼肌中参与各类刺激、干预措施或病理状态应答的任意目标基因的转录因子（Transcription Factor, TF）。

本研究共纳入10名经过耐力训练的业余运动员（包括长跑运动员、自行车选手与越野滑雪运动员，年龄中位数为32岁[四分位间距：30~36岁]；体重中位数为75kg[四分位间距：71~78kg]；每公斤体重最大摄氧量（V’O2max/kg，即最大肺部氧消耗速率）中位数为58ml/min/kg[四分位间距：54~60ml/min/kg体质量]）。所有受试者于标准化早餐（总能量3582kJ，含蛋白质22g、碳水化合物154g、脂肪16g）后2小时，在自行车功率计上完成间歇性运动方案：共60分钟，即[3分钟以血乳酸4mmol/L对应功率（乳酸阈值LT4）50%的强度运动 + 2分钟以100%LT4强度运动]重复12组。受试者于间歇性运动结束后1小时15分钟，食用标准化午餐（总能量3714kJ，含蛋白质45g、碳水化合物183g、脂肪27g）。

分别于间歇性运动前、运动后2分钟、1小时、3小时及6小时，在局部麻醉（2mL 2%利多卡因）下，使用Bergstrome抽吸针从股外侧肌采集活检样本（第1、2、3次采样取自单侧下肢，第4、5次采样取自对侧下肢）。对于每名受试者，运动前、运动后1小时、3小时及6小时的活检样本将在同一批次实验中完成检测；此外，所有运动后2分钟的样本将在单独的批次实验中完成检测。

为提升CAGE TSS鉴定的质量，本研究使用全部数据进行计算；但仅针对运动前、运动后1小时、3小时及6小时这四个时间点，对CAGE TSS与基因的差异表达以及差异性TSS使用模式进行了统计学评估。