To study engagement and learning, we developed an asymmetric collaborative version of an existing VR-experience about environmental sustainability, enabling student teams of three, with only one wearing a headset at any time, to collaborate in three different virtual reality subgames. We compared this collaborative version to a non-collaborative version in a between-subjects experimental study (n = 20), finding that student teams using the collaborative version interacted to a larger extent while performing the subgames. The students in these teams self-reported a higher level of engagement, and observations suggested that interactions extended beyond necessary instructions. We did not find any significant differences when it comes to learning measured by scores on pre- and post-knowledge tests, but quantitative analysis of responses indicates that the VR-experience affected students overall and a qualitative analysis of an open question suggested a particular effect on the collaborative teams. Logged performance data indicated that collaborative teams managed to solve the tasks involved in the subgames at least as efficiently as the teams using the non-collaborative version.