Heated and uncivil discussions are regrettably common in social media, potentially leading to harmful atmospheres. In the present work, we explore how relatively simple user interface mechanisms might nudge users to choose a more civil tone by promoting an emotion regulation process called affect labeling-i.e., naming one's emotions. In particular, we focus on labeling emotions in response to online news articles, as this is a ubiquitous form of online multimedia. Based on systematic design exploration, we present four concrete mechanisms as alternatives for affect labeling news articles: an emotion wheel, an intensity slider, emotion analysis done through machine learning, and free text labeling. In a qualitative evaluation study, 18 interviewees discussed the designs in focus groups in order to identify design aspects that might lead to high acceptability and usability. The study findings help reflect how emotion regulation could be facilitated by UI mechanisms, which design aspects both encourage affect labeling and are acceptable by users, and how affect labeling can meaningfully influence online interactions.