Structure and Content-Guided Video Synthesis with Diffusion Models