Story of New Vitamin, Pyrroloquinoline Quinone (PQQ) Laboratory for Molecular Dynamics of Mental Disorders "I've discovered the function of that gene! It has PQQ-binding motifs," Dr. Kasahara quietly exclaimed as he burst into my office one night. Before that moment, I had never even heard the name of that substance. This was June 2001. When I moved to RIKEN Brain Science Institute, I believed that manic-depression was caused by an abnormality in the Ca2+ controlling mechanism of mitochondria. However, nothing was known about the mitochondrial Ca2+ transporting molecule (mCU). The subject required a large-scale study that I could not afford before coming to BSI. Wanting to investigate this I put a recruitment ad in Nature. Soon after, I received an application from Takaoki Kasahara, and I found his hypothesis unique and appealing. What Kasahara proposed was a novel functional peptides. Bacteria create peptides, making the cell membrane permeable and thus allowing Ca2+ and other ions to enter and kill harmful bacteria. As mitochondrion is a primitive bacterium that evolved as a parasite in eukaryotic cells, it would not be surprising if it has this kind of peptide. He wondered if peptides like that transport Ca2+. These kinds of peptides have specific enzymes for synthesis, whose structures are characterized by A-and-T domains which attract amino acids. He conjectured that, if a gene with these domains could be cloned, it could be a synthesizing enzyme of mCU. At this time, I was still working as a psychiatrist at Tokyo University Hospital while Kasahara was studying biological rhythm as a graduate student at the same university. We communicated by e-mail and, together shaped this hypothesis and a research method. I started at BSI in January 2001 and, when Kasahara joined in April, our project finally started. Kasahara worked extremely hard at an amazing pace, and by early June, he was already cloning a full-length cDNA of a gene with A-T domains. Contrary to our expectations, however, only one A-T domain was unable to create peptide. I was disappointed and advised Kasahara to publish the results so far, although not very interesting, in some journal. Although he was skeptical about the acceptance of such a meaningless paper, he reluctantly began to searching the database and found PQQ-binding motifs (see Figure 1). Kasahara had identified a homologous gene with an NADPH-binding motif instead, a coenzyme that has an opposite function to that of PQQ. Because this was an enzyme involved in synthesizing lysine of yeast, he was sure that this product of cDNA must be aminoadipic semialdehyde dehydrogenase (AASDH), which should be involved in lysine decomposition. But why is this important? Until recently, enzymes with which PQQ acts as a coenzyme had been identified only among bacteria. In mice, a lack of PQQ causes fragile skin and poor reproductive performance, so PQQ might be an essential nutrient but its function had yet to be revealed. Finding an enzyme that had PQQ as a coenzyme was part of work being done to establish that PQQ was a vitamin. Although it was not the subject of our study, I decided-as a medical doctor rather than as a psychiatrist-that the study should be completed because, if a vitamin that should be included in drip infusion had been overlooked, then many patients may have suffered from deficiency syndrome. However, the way ahead was difficult. Kasahara struggled for a year and a half to prove enzyme activity. Protein purification using E. coli was unsuccessful. After trying every available method, he finally purified the protein in insect cells. Although he did not successfully measure activity, he overcame a series of challenges, including managing the high reactivity of PQQ, synthesizing enzyme substrate, measuring the substrate specificity with a different method, investigating its post-translational modification using mass spectrometry, and cloning enzymes modifying AASDH. During this endeavor, he hit on the idea of detecting metabolite changes using mice lacking in PQQ. Although I was unsure if this method could be used to prove the function of enzyme, he persevered. He fed mice a full nutrient compliment for growth-except PQQ-mixed with many reagents, and observed decreases in AASDH metabolite using amino acid analysis. I was surprised to read the first draft of his article-there was no mention of protein purification, substrate specificity, or post-translational modification. I was surprised and asked him if he was sure about his article as it was. He replied, "Those have nothing to do with the subject." Trusting his instincts, we submitted the article to Nature. The reaction was overwhelming. Expert readers, who had surely been involved in PQQ research for a long time, highly evaluated this work, and his insight turned out to be correct. BSI's Research Resource Center, and its exclusive experts, provided us remarkable support with mass spectrometry, amino acid analysis, and mice breeding. Without the support of RIKEN BSI and its facilities, this work could never have been done. International reaction following the publication of a short one-page article published in April 2003, was greater than expected. It was both delightful and overwhelming. I was delighted to see Kasahara rewarded for his endeavor and acknowledged for his effort to prove that PQQ is a vitamin for mammals. Figure 1. Structure of aminoadipic semialdehyde dehydrogenase (AASDH) Murine AAS dehydrogenase is a protein of 1,100 amino acids. There are pairs of A- (pink) and T- (green) domains on the N-terminal side and the T domain undergoes the post-translational modification. On the C-terminal side, there are seven repeats of PQQ-binding motifs (blue). Takaoki kasahara and Tadafumi Kato Nature 422, 832 (2003)