Automotive model-based systems engineering needs to be dapted to the industry specific needs, in particular by implementing appropriate means of representing and operating with variability. We rely on existing modeling techniques as an opportunity to provide a description of variability adapted to a systems en- gineering model. However, we also need to take into account requirements related to backwards compatibility with current practices, given the industry experience in mass customization. We propose to adopt the product line paradigm in model-based systems engineering by extending the orthogonal variability model, and adapting it to our specific needs. This brings us to an expression closer to a description of constraints, related to both orthogonal variability, and to SysML system models. We introduce our approach through a discussion on the different aspects that need to be covered for expressing variability in systems engineering. We explore these aspects by observing an automotive case study, and relate them to a list of contextual requirements for variability management.