Acarbose is an α-glucosidase inhibitor used clinically to treat type 2 diabetes. This compound was first isolated from bacteria in the 1970s and approved to be used as an anti-diabetic drug in the US and Europe in the 1990s. As type 2 diabetes mellitus cases around the world have been steadily increasing, the need of acarbose and other antidiabetic drugs has also increased. However, despite its significant therapeutic importance, how this bioactive compound is made in bacteria was not clearly understood. Over the past four decades, several research groups have attempted to decipher the biosynthetic pathway to acarbose, but only the first half of the pathway was known. Elucidating the biosynthetic pathway to this highly polar pseudooligosaccharide appeared to be very challenging. Therefore, when I was suggested to work on this project by my postdoctoral mentor, Prof. Taifo Mahmud, I experienced emotions of honor and horror at the same time.

I felt extremely honored because I knew that this project has historical meaning to our laboratory, as my mentor was one of the authors that reported the first biosynthetic step of the acarbose pathway more than 20 years ago.  He was a postdoctoral research associate with Prof. Heinz G. Floss at the University of Washington, who was at the time collaborating with Prof. Wolfgang Piepersberg at the University of Wuppertal on the acarbose biosynthesis project. The next two steps of the pathway were reported independently by the Piepersberg group in the early 2000. After almost two decades of hiatus, two other steps were recently reported by the group of Profs. Linquan Bai and Zixin Deng at Shanghai Jiaotong University. Our lab at Oregon State University was mainly interested in elucidating the later steps of acarbose biosynthesis, as it may involve unique enzymes such as the C-N bond forming enzyme pseudoglycosyltransferase. Several of our previous lab members have worked on and off on this project with moderate successes, and the baton was then passed on to me.

The horror part of the story is that all of the intermediates involved in acarbose biosynthesis are expected to be highly polar, not commercially available, and somewhat unstable. In addition, I learned that some of the enzymes were not expressed well in Escherichia coli. Fortunately, I have experience in organic synthesis and have previously worked with highly polar compounds, and together with two other chemists in the lab, we were able to synthesize some of the compounds needed for the study. Some other compounds were prepared enzymatically or a gift from Prof. Floss. To obtain soluble recombinant enzymes, I tried numerous conditions for the expression of recombinant proteins in E. coli but had to give up some of them as they did not give soluble proteins. But, I was finally able to produce the challenging recombinant proteins in Streptomyces and characterized their biochemical functions and their involvements in acarbose biosynthesis.

In our recent paper published in Nature Communications (click here), we report the complete biosynthetic pathway to acarbose. Aside from enzymes that are necessary to produce sedoheptulose 7-phosphate, dTDP-4-amino-4,6-dideoxyglucose, and maltose, the main pathway to acarbose consists of ten enzymes, of which five enzymes were characterized in this study including a pseudoglycosyltransferase, AcbS, and its homologue AcbI. Interestingly, despite their high sequence similarity, these two enzymes have very distinct catalytic activities. AcbI catalyzes a canonical glycosylation reaction, whereas AcbS catalyzes a non-glycosidic C-N bond formation. In this study, we were able to reconstitute the acarbose pathway from valienol 7-phosphate all the way to the final product acarbose. It was a joyful experience working on this challenging project and I realized that the tacit knowledge and expertise related to the study of aminocyclitol biosynthesis available in the lab has significantly contributed to our success in elucidating the biosynthetic pathway to this high value pharmaceutical.