Many European countries witnessed massive protests after governments adopted measures to combat the COVID-19 pandemic. Unlike other protest movements, the participants are often very diverse in terms of their motivation and political ideology. Recent research in Germany, for instance, shows that large percentages of the protesters used to vote for the Green party but plan to vote for the right-wing populist party AfD in upcoming elections. Therefore, the protests provide an opportunity to study the following questions: When do MPs of established parties positively engage with anti-system protests, against the party line? We draw from existing research on MPs' reaction to exogenous shocks and their voters' related attitudes as well as social movement literature emphasizing protest as control of political decision-making. Combining these research strands with studies on the individual level on rebels in representative democracies, we explain reaction patterns between and diversity within different parties. We analyze German parties' and politicians' reactions using parliamentary debates as well as the MPs' Tweets for daily updated data on an individual level covering the period from the beginning of the pandemic until the end of 2020. To quantify these reactions, we develop a new measure for an automated sentiment analysis explicitly tailored to COVID-19 protests. It relies on established sentiment analysis procedures combined with machine learning algorithms to obtain more accurate results. Our findings have important implications for the representation of protest by parties, especially when there is uncertainty about the views of their own voters on the issue.