Protein annotation is a task that describes protein X in terms of topic Y. Until now, most of protein annotation work has been done manually by human annotators. However, as the number of biomedical papers grows ever rapidly, manual annotation becomes difficult, and there is increasing need to automate the protein annotation process. Recently, Information Extraction (IE) has been used to solve this problem. Typically, IE requires pre-defined relations and hand-crafted IE rules or annotated corpora, and these requirements are difficult to satisfy in real world domains such as the biomedical domain. In this paper, we describe an IE system which requires only sentences labeled relevant or not to a given topic by domain experts.