I think the question is pretty clear. Be sure to read the description of Turing machines used in your course.
Basically, a TM with a two-way infinite tape can be simulated by a TM with two tapes that are infinite only to the right. The portion of the original TM's tape from the initial square to the left is modeled by one tape, and the portion from the initial square to the right is modeled by the second tape. In turn, a two-tape machine can be simulated by a one-tape machine. This should be described in most books about the subject, for example, in Elements of the Theory of Computation by Lewis and Papadimitriou.
If you still have questions about the formulation of the problem, please specify which terms or concepts are not clear.