From masochistic enzymology to mechanistic physiology and disease

The pioneering work of Eugene Kennedy in the 1950s established the choline pathway for phosphatidylcholine (PC) biosynthesis. However, the regulation of PC biosynthesis was poorly understood at that time. When I started my lab at the University of British Columbia in the 1970s, this was the focus of my research. This article provides my reflections on these studies that began with enzymology and the use of cultured mammalian cells, and progressed to utilize the techniques of molecular biology and gene-targeted mice. The research in my lab and others demonstrated that the regulated and rate-limiting step in the choline pathway for PC biosynthesis was catalyzed by CTP:phosphocholine cytidylyltransferase. This enzyme is regulated by its movement from a soluble form (largely in the nucleus) to a membrane-associated form where the enzyme becomes activated. Gene targeting in mice subsequently demonstrated that this gene is essential for development of mouse embryos. The other mammalian pathway for PC biosynthesis is catalyzed by phosphatidylethanolamine N-methyltransferase (PEMT) that converts phosphatidylethanolamine to PC. Understanding of the regulation and function of the integral membrane protein PEMT was improved when the enzyme was purified (a masochistic endeavor) in 1987, leading to the cloning of the Pemt cDNA. Generation of knock-out mice that lacked PEMT showed that they were protected from atherosclerosis, diet-induced obesity, and insulin resistance. The protection from atherosclerosis appears to be due to decreased secretion of lipoproteins from the liver. We continue to investigate the mechanism(s) by which Pemt−/− mice are protected from weight gain and insulin resistance.

I am extremely fortunate to have had such excellent mentors throughout my scientific career. Without these dedicated teachers and colleagues, I am certain that I would not have become a biochemistry professor, nor would I have learned how to become an effective mentor myself.
I became interested in chemistry because of my teacher, Malachi Pancoast, in 11th grade at Springfield High School in Springfield, Pennsylvania. He exhibited intense enthusiasm for chemistry and challenged his students to understand and appreciate this subject. He was a caring person, as shown by his sending birthday cards each year to his students. Even after he retired from teaching, he continued his tradition of sending birthday cards until his death in 2010; his obituary estimated that he sent about 1,000 birthday greetings each year.
As an undergraduate, I attended Dickinson College in Carlisle, Pennsylvania and majored in chemistry. A central reason I went to Dickinson was the opportunity to take non-science courses as well. Among those courses that were memorable were Philosophy of Religion, Art History, Abnormal Psychology, and American Literature. In the 3rd year, I took Organic Chemistry with Dr. M. Benton Naff. He explained the subject well and stimulated our critical thinking. Many of the students (especially the pre-meds) did not appreciate him because he demanded such a high standard. I grew to really enjoy organic chemistry and even did a small research project with Dr. Naff. I do remember that he sometimes became angry when students arrived late to class; consequently, he locked the classroom door at 10 a.m. when the class was due to start. Of course, this occurred in an era before lectures were posted on-line so taking notes in class was important.
One evening Dr. Naff organized a talk on cholesterol biosynthesis by Dr. Charles (Chuck) Sweeley ( Fig. 1) from the University of Pittsburgh. At this time (1963), biochemistry was not taught at the undergraduate level in most colleges and universities in the United States (biochemistry was taught in medical schools). Dr. Sweeley's lecture was truly an eye-opener for me and led to my application to the University of Pittsburgh for PhD studies. Another motivating factor lurking in the background was the increase in American involvement, particularly for young men of my age, in the Vietnam war. As history has now recorded, this involvement was a tragic mistake resulting in millions of unnecessary deaths. In the spring of 1964, a letter arrived offering me a position as a Ph. D. student in Dr. Sweeley's laboratory. I was astounded to learn that the positon came tuition-free and with a stipend of $225 per month! I accepted and moved to Pittsburgh in September.

Why do glycolipids accumulate in Fabry's disease?
I was Prof. Sweeley's second graduate student, and he mentored me in great depth from the start of my Ph. D. studies until graduation in 1968. At least once a day, he came into the lab to ask how my research was going and make suggestions. He was a smoker at that time and often leaned over the lab bench where I was using organic solvents such as ether; luckily, the lab did not burst into flames. When the Dean of the School died of lung The author declares that he has no conflicts of interest with the contents of this article. 1 To whom correspondence should be addressed. Tel.: 780-492-8286; E-mail: dvance@ualberta.ca.
cro REFLECTIONS cancer, Chuck quit smoking the next day and never, to my knowledge, smoked again.
In 1963, before I arrived, Dr. Sweeley with his colleague Bernard Klionsky had elucidated the structure of the glycosphingolipid, trihexosyl ceramide, that accumulates abnormally in the inherited, lysosomal storage Fabry's disease and leads to death (from kidney failure or heart disease) in mid-life in males who inherit the abnormal gene (1). Dr. Tamio Yamakawa and his colleague Shizue Suzuki had determined the structure of the related glycosphingolipid, globoside, which accounts for ϳ10% of lipids in red cells (2). Thus, the hypothesis presented to me for my Ph. D. studies was: because ϳ1% of red cells are degraded each day in humans, the catabolism of globoside could account for much of the accumulating trihexosyl ceramide observed in Fabry patients. The first step in testing this idea was to develop a method for separation of the various glycosphingolipids from human red cells and plasma. Chuck was a skilled analytical biochemist. Thus, under his supervision, I developed a thin-layer chromatographic procedure to separate the main neutral glycolipids from red cells and plasma. We submitted an abstract to the Federation of American Societies for Experimental Biology (FASEB) meeting in Chicago in 1967. Luckily, the abstract was selected for an oral presentation, as poster presentations were not in vogue in 1967. Of course, I had to make the slides a couple of weeks in advance because personal computers and PowerPoint were decades away. My talk was scheduled as the first talk of the meeting, and Eugene Kennedy was the chair of the session. To say that Chuck was hyperexcited would be an understatement. The talk went well, and Chuck's blood pressure normalized. The resulting paper (3), describing the method, is the most cited paper of my career.
To test the hypothesis that red cell globoside was a major source of ceramide trihexoside in humans (knock-out mice were still decades away), we required a method to track the amounts of deuterium-labeled glucose incorporated into the glycosphingolipids. Chuck, Ron Bentley (in the lab next door), and others had recently developed a method for gas chromatographic analysis of carbohydrates, which I also used. Chuck then pioneered a gas chromatography-mass spectrometry method to analyze the amount of deuterium-labeled carbohydrate in glycosphingolipids from plasma and red cells using one of the first machines for this technique, the LKB 9000. With a detection method in hand, we infused 35 g of 6,6-dideutero glucose into a control patient and a Fabry patient over 12 h. Despite the huge amount of labeled glucose administered, we did not detect enough label in globoside to support the hypothesis that red cell globoside was a precursor of plasma trihexosyl ceramide. Nevertheless, we did see ϳ2% plasma lactosyl ceramide and glucosyl ceramide labeled with 6,6-dideutero glucose, which exchanged onto the red cells (4).
I completed my Ph. D. in 1968 and stayed in Pittsburgh for postdoctoral studies while my wife, Jean (who changed her name to Jean Vance), completed her Ph. D. training with Ronald Bentley. I was very fortunate to be able to join the laboratory of David Feingold in the Department of Microbiology where I learned enzymology. In his lab, I developed a technique for N-terminal sequencing of proteins and peptides using Edman degradation (5). We used methyl-thiohydantoin to react with the N-terminal amino acid and detected the product by gasliquid chromatography. The novelty of our procedure was that after the first cycle of the Edman reaction, we removed one aliquot for gas chromatographic analysis. We repeated the Edman reaction in the remaining sample, and then removed another aliquot on which we performed gas chromatography. Thus, after the first round, we generated a phenylalanine derivative, and after the second round, the gas-chromatographic profile showed phenylalanine and aspartate. We, therefore, were able to obtain the N-terminal sequence for up to 4 amino acids of a peptide without difficulty. During this time, I also took the opportunity to broaden my horizons by attending the medical student lectures in microbiology and virology.

An unexpected regulatory mechanism
In 1970, Jean and I moved to Massachusetts where I became a postdoctoral fellow with Konrad Bloch at Harvard University. When I joined his lab, I had thought I would work on purification of one of the membrane-bound enzymes involved in cholesterol biosynthesis. Such endeavors are certainly masochistic. However, his lab had discovered that a polysaccharide stimulated the fatty-acid synthase from Mycobacterium phlei (later identified as Mycobacterium smegmatis) and was trying to identify the major carbohydrate of the polysaccharide. My expertise in gas chromatography of carbohydrates (from Chuck Sweeley's lab) allowed me to quickly identify the carbohydrate as 3-O-methyl mannose. Thus, 6 weeks after arrival in the lab, I was an author of a paper in Proceedings of the National Academy of Sciences (PNAS) (6).

REFLECTIONS: Mechanistic physiology and disease
Dr. Bloch ("KB" as we all called him) then asked me to determine the mechanism by which this polysaccharide stimulated the activity of the fatty-acid synthase in M. phlei. This was a challenging project in which I became totally immersed. Almost 2 years later, in the last 2 weeks that I was in Bloch's lab, we solved the riddle. It turned out that the polysaccharide was not actually "stimulating" the reaction but rather relieving feedback inhibition by long-chain acyl-CoAs (24-carbon chain) that were made by the fatty-acid synthase. The polysaccharide bound the acyl-CoAs, promoting diffusion of the "sticky" acyl-CoA product away from the active site of the enzyme (7).
Konrad Bloch (Fig. 2) was an extraordinary human being and scientist in so many ways. Before I joined his lab, I naively worried that because he had a Nobel Prize (awarded in 1964), he might not be so interested in his research. I was totally wrong! His intense curiosity and the science drove him, and this did not change after the Nobel Prize. Indeed, his interest and curiosity in science continued for the rest of his life. We kept in touch with him and his very supportive wife, Lore, on many occasions after I left his laboratory. One time I asked him whether there was any part of his career that he regretted. As I recall, he said, "My only regret is that I wasted so much effort and time of many trainees in my lab by asking them to purify membrane proteins/ enzymes" -and to think that I went to his lab to learn how to purify integral membrane enzymes! While I was studying the regulation of fatty-acid synthase in Bloch's lab, there were many other first-rate labs also working on fatty-acid synthase and its regulation. Thus, it became abundantly clear that, when I started my own lab, it would not be wise to work on this topic. However, I realized that, although the choline pathway for phosphatidylcholine (PC) 2 biosynthesis ( Fig. 3) had been established in the 1950s through the groundbreaking work of Eugene Kennedy (Fig. 4) and col-leagues, in 1972, almost nothing was known about its regulation. This was a fundamental question in biochemistry that was wide open.
Because I had audited a virology course in Pittsburgh, I was aware of certain viruses that were surrounded by a lipid envelope that was obtained from the host cell. Thus, I hypothesized that because the virus was "stealing" lipids from the host cells, PC synthesis in the host cells might be stimulated after viral infection. For that reason, I wanted to test this idea in a virology laboratory. I was fortunate that Derek C. Burke, a British virologist at the University of Warwick in England, was willing to allow me to spend a year in his lab to test this hypothesis. The work went well, and I learned how to work with animal viruses and cultured cells. The unexpected result was that replication of the lipid-enveloped Semliki Forest virus did not stimulate PC biosynthesis (8), whereas replication of the non-enveloped polio virus did stimulate PC biosynthesis. The mechanism for the latter effect appeared to be that the concentration of cytoplasmic CTP increased and accelerated the rate-limiting step in PC biosynthesis catalyzed by CTP:phosphocholine cytidylyltransferase, which converts phosphocholine to CDP-choline (9) (Fig. 3). This result was an important first step in establish-

REFLECTIONS: Mechanistic physiology and disease
ing myself in the field, and gave me more confidence in envisioning my own laboratory.
In the fall of 1972, I sent job applications to many universities. In particular, the University of British Columbia (UBC) had advertised a tenure track job for someone to teach a graduate course on lipids. I interviewed and was delighted when they offered me the position. As a typical American, I knew very little about Canada. The Department of Biochemistry at UBC provided a very good opportunity for me to start my lab, and Vancouver was a very attractive place to live. Moreover, because I was married to a British citizen, Canada seemed appropriate. I always expected my first job would be on the East Coast of the United States; that one advertisement changed the direction of my life.

The UBC years
The Ph. D. project of my first graduate student, Chris Richardson, naturally built on my studies in Warwick. He was to study the mechanism of assembly of the lipid envelope of Semliki Forest virus. He demonstrated that the virus obtained its envelope by budding from the plasma membrane of infected cells (10). Because none of the three enzymes in the choline pathway for PC biosynthesis had been purified, we started to purify the soluble enzymes choline kinase and CTP:phosphocholine cytidylyltransferase (CT). We wisely left the integral membrane enzyme, cholinephosphotransferase, in the "to do bucket." We made good progress on purification of choline kinase (11) and CT (12). CT was eventually purified to a single band on an SDS gel by Paul Weinhold's lab (13), and choline kinase was purified by Kozo Ishidate's lab in Japan (14). To my knowledge, no one has ever purified cholinephosphotransferase.

Regulation of CTP:phosphocholine cytidylyltransferase
In 1978, Pat Choy, a postdoctoral fellow in my lab, discovered that CT was activated by lipids (15). Research from Claudia Kent's lab (16) and our own lab (17) showed that the lipid activation in vitro reflected the observation that CT was inactive in cytosol and activated only upon translocation to cellular membranes (reviewed in Ref. 18). In the last 30 years, the mechanism underlying CT translocation has been exquisitely elucidated by research in Rosemary Cornell's lab (19).
The primary focus of our lab subsequently became an understanding of the mechanism(s) of regulation of the choline pathway for PC biosynthesis in mammalian cells. Initial results demonstrated that the rate-limiting step in the pathway was catalyzed by CT (Fig. 3). When choline is taken up by cells, it is rapidly phosphorylated by choline kinase. The subsequent conversion of phosphorylcholine (P-choline) to CDP-choline then determines the rate of PC biosynthesis because the levels of CDP-choline are very low and it rapidly reacts with diacylglycerol to form PC (20). It is a somewhat unique situation that the first step in a biosynthetic pathway (choline kinase in this case) is apparently not regulated or rate-limiting for the pathway. Subsequent research has demonstrated that the rate of the CT reaction increases when levels of PC in the membranes decrease, when the cells are supplemented with fatty acids, or when the enzyme becomes dephosphorylated (17,21,23,24), in addition to when the concentration of CTP increases.
Under normal metabolic conditions in mammalian cells and tissues, there is an excess pool of CT that is not associated with membranes; translocation of CT to membranes activates the enzyme. Thus, the store of inactive CT allows cells to activate CT rapidly without increasing gene transcription or protein synthesis. Although transcriptional regulation of CT might not be important in short-term regulation of CT activity, we investigated mechanisms by which transcription of CT is regulated over longer periods (reviewed in Ref. 25). The results from over a decade of research in my lab and the lab of Hiroyuki Sugimoto demonstrated that the transcription factors Sp1, Rb, TEF4, Ets-1, and E2F enhance the expression of CT, whereas Net represses CT expression. In contrast, transcription factors (sterol regulatory element-binding proteins (SREBPs), liver X receptors (LXRs), and peroxisome proliferatoractivated receptors (PPARs)) involved in the metabolism of cholesterol and fatty acids are not major players in the regulation of CT. Perhaps the explanation for this difference between the regulation of cholesterol/fatty acids and phospholipid synthesis is that instead of being linked to energy metabolism, transcriptional regulation of CT expression is linked to the cell cycle, as well as cell growth and differentiation.
A surprising finding from Claudia Kent's lab was that CT is primarily located in the nucleus of cells (26). Much more research on this topic has been performed, and it is now clear that CT␣, which contains a nuclear localization signal, is primarily nuclear, whereas CT␤ (encoded by a distinct gene) lacks a nuclear localization signal and is cytoplasmic (reviewed in Ref. 27). The reason why CT␣ is primarily localized to the nucleus has not yet been established and is a challenging area of research.

PEMT brings us to masochistic enzymology
Although the choline pathway for PC biosynthesis is present in all nucleated mammalian cells, an alternative pathway in which phosphatidylethanolamine (PE) is converted to PC by PEMT is present in the liver, but only at very low levels (Ͻ1% of the activity in liver) in a few other tissues. PEMT catalyzes three

REFLECTIONS: Mechanistic physiology and disease
reactions in which PE is sequentially methylated to PC with S-adenosylmethionine as the methyl donor (Fig. 5). Although PEMT is an integral membrane protein of microsomes, Wolfgang Schneider, a postdoc in my lab, made a heroic effort to purify PEMT from rat liver (28). He obtained a partial purification and solubilization of the protein. Subsequently, in the mid-1980s, Neale Ridgway joined my lab as a Ph. D. student and, after tremendous efforts, was successful in purifying PEMT to homogeneity (29). A major problem was that the detergentsolubilized enzyme was quite unstable in the Tris buffer that we were using. One day Neale had the idea of trying a phosphate buffer instead. The solubilized enzyme was perfectly stable in phosphate buffer, allowing Neale to complete the purification. The term "masochistic enzymology" refers to the daunting, almost impossible, job of purifying membrane proteins. Of course, in 2017, almost no one purifies proteins from tissues due to the ease of cloning the cDNAs, allowing expression of large amounts of defined proteins.

PC and lipoproteins
In 1982, I thought it was time to learn about lipoproteins because PC is a major component of these cholesterol-carrying particles that are involved in atherosclerosis. The PC component of lipoproteins had, until this time, been largely ignored. Thus, I took a sabbatical year in Dan Steinberg's lab at the University of California at San Diego. I worked with David Weinstein, a young faculty member, to develop protocols for culture of primary rat hepatocytes and isolation of the lipoproteins secreted from these cells (30). Using this method, we initiated studies on the role of PC biosynthesis in the secretion of very low-density lipoproteins (VLDLs). At this time, Zemin Yao joined my lab as a Ph. D. student, one of the first students to come out of China after the "Cultural Revolution" ended. He demonstrated that PC biosynthesis was required for normal secretion of VLDLs from rat hepatocytes (31). In contrast, the formation of high-density lipoproteins was not dependent on PC biosynthesis. In more recent years, with the advent of genetargeted mice, we have been able to demonstrate that both the choline pathway and the PEMT pathway for PC biosynthesis are required for normal secretion of VLDLs from mouse hepatocytes (32) (33). Why one PC biosynthetic pathway does not substitute for the other in VLDL secretion is not clear at this time.

Important diversions
In 1978, I was appointed Associate Dean for Research in the Faculty of Medicine at UBC. It was a half-time appointment, and I wanted to try administration. Bill Webber was the Dean of Medicine at that time, and he was committed to ensuring that my research could continue during my time in the Dean's Office. Consequently, he made certain that the 3-year appointment really did constitute a maximum of 50% time commitment. Although it was interesting to learn the challenges of faculty administration, I decided that I did not wish to pursue administration at that level. Instead, I thought that being Chair of the Biochemistry Department at UBC would better fit my interests and provide more time for research. I was given that opportunity in 1983. Although I enjoyed some aspects of being Chair, such as recruiting and helping new faculty members become established, I concluded without any doubt that I preferred to put my energies into full-time research and teaching.
Other diversions were more closely related to research. I was appalled at the level of lipid biochemistry presented in biochemistry textbooks. Even the wonderful textbooks by Lehninger and Stryer gave at best a cursory treatment of lipids with no information on lipoproteins. I knew there was much more that students should learn about lipids at the introductory level. Thus, when Geoff Zubay asked me to contribute the lipid chapters to his textbooks Biochemistry (four editions) and Principles of Biochemistry, I agreed with enthusiasm. I'm proud to say that lipid biochemistry is now much better represented at the introductory level. Moreover, the treatment of lipids in textbooks at the advanced level was poor to nonexistent. Thus, Jean Vance and I coedited an advanced textbook called Biochemistry of Lipids, Lipoproteins and Membranes (46). The first edition was published in 1985, and we continued to serve as editors through the 5th edition that was published in 2008. It was gratifying to see that this book appeared to have an impact on graduate student and postdoctoral training throughout the world. When we traveled to other countries, we always saw it being widely used. The sixth edition of Biochemistry of Lipids, Lipoproteins and Membranes, now with Neale Ridgway and Roger McLeod as editors, was published in 2016.
Finally, as a research-active professor, one is obliged to assist in the review of research manuscripts and grants submitted by other scientists. The system depends on these external reviews, and virtually all scientists accept that duty and perform it at a high level. Perhaps because I consider this to be such an important responsibility, I dedicated more time to review than I could have, serving on many journal editorial boards (including JBC) and as Executive Editor of the lipid section and Editor-in-Chief of Biochimica Biophysica Acta (BBA) for 14 years total.

REFLECTIONS: Mechanistic physiology and disease
departments. Wolfgang Schneider was already a Professor of Biochemistry at the University of Alberta, and Mark Poznansky was a Professor of Physiology; these two scientists were initial members of the Lipid Group.
When I told colleagues at UBC about the offer, they could hardly believe that anyone from Vancouver would consider moving to Edmonton, Alberta because the winter in Edmonton is long and very cold. The snow usually comes in early November and does not evaporate (I don't mean melt) until March. On the plus side, the sky is often brilliant blue and the total amount of snow is very little with only a few centimeters falling at a time. Moreover, Jean and I thought the opportunity to interact on a daily basis with a group of other lipid researchers would be very stimulating. After 31 years in Edmonton, we have never regretted that decision.

Molecular biology meets lipids
The move to the University of Alberta in 1986 also came at the time when molecular biological techniques were starting to have a major impact on lipid research. After the purification of PEMT in 1987, we were able to obtain some protein sequence. At that time, a molecular biologist, Zheng Cui, joined the lab and cloned and expressed the cDNA encoding PEMT (34). Surprisingly, using an excellent antibody raised against a PEMT dodecapeptide, we discovered that the enzyme was highly enriched in mitochondria-associated membranes (34), a fraction of endoplasmic reticulum that transiently and reversibly contacts mitochondrial outer membranes, as described by Jean Vance in 1990 (35).
We were curious about the function of PEMT because PC is also made by the choline/CT pathway. From the cDNA sequence of PEMT, we were in a position to undertake the arduous task of generating a knock-out mouse in which the Pemt gene was no longer expressed. Fortunately, a new graduate student, Chris Walkey, accepted the challenge. In close collaboration with our colleague, Lou Agellon, Chris characterized the mouse Pemt gene (36) and disrupted it in mice (37). In 2017, gene targeting is almost routine; in contrast, in the mid-1990s, this was not a trivial task.
The Pemt Ϫ/Ϫ mice grew normally and did not exhibit obvious abnormalities. In the knock-out mice, the amount of the active form of hepatic CT was 60% higher than in Pemt ϩ/ϩ mice, which probably explained why the levels of PC were not lower in the livers of Pemt Ϫ/Ϫ mice. This result was, at first, disappointing but, upon reflection, not unexpected because the livers still contained abundant CT activity, which apparently compensated for the lack of PEMT. Chris then fed the Pemt Ϫ/Ϫ mice a choline-deficient diet so that availability of the substrate for the choline pathway for PC synthesis was reduced. After 3 days, the Pemt Ϫ/Ϫ mice that lacked choline in their diet developed end-stage liver failure (38), and the experiment had to be stopped (Fig. 6). The conclusion was that PEMT expression in the liver had been maintained during evolution to provide a source of PC when dietary choline was insufficient. The requirement for choline is very specific because dimethylethanolamine, which is converted into phosphatidyldimethylethanolamine, a phospholipid with physical properties similar to those of PC, did not prevent liver failure in Pemt Ϫ/Ϫ mice fed a choline-deficient diet (39).

Transition to mechanistic physiology and disease
The development of gene-targeting technology has profoundly altered our understanding of the function of enzymes and proteins in biology. Previously, one could possibly have developed an inhibitor of PEMT that might have revealed the role of PEMT in evolutionary biology. However, there is always concern about incomplete inhibition and off-target effects of inhibitors, whereas gene-targeting technology is very specific and unambiguous. Thus, in 1997, we were in the enviable position of exploring and understanding the role of PEMT in mouse physiology and disease.
As our first project, we wanted to test a hypothesis that Pemt Ϫ/Ϫ mice might be protected from atherosclerosis because hepatic VLDL secretion is attenuated by elimination of PEMT. We used mice that lacked the LDL receptor because when these mice are fed a high-cholesterol diet, they develop severe atherosclerosis. When PEMT was eliminated from mice lacking the LDL receptor (and the mice were fed a high-cholesterol diet), atherosclerosis in the artery (aortic root) was strikingly decreased by 80% (40) (Fig. 7). This result was very surprising because no one had predicted that PEMT and PC biosynthesis might be directly involved in atherosclerosis. Thus, these observations indicated that inhibition of PEMT might be a novel approach to reducing cardiovascular disease.
As we were feeding high-cholesterol diets to Pemt Ϫ/Ϫ mice, we designed a curiosity-driven experiment testing what seemed like an even more far-fetched hypothesis: that there might be a link between PEMT and obesity/diabetes. A postdoctoral fellow in the lab, René Jacobs, fed Pemt ϩ/ϩ and Pemt Ϫ/Ϫ mice a high-fat diet (60% fat as lard). As predicted, the Pemt ϩ/ϩ mice gained weight and developed insulin resistance (which can lead to type 2 diabetes). However, the Pemt Ϫ/Ϫ mice were protected against weight gain and insulin resistance (41) (Fig. 8). The diet we used contained normal concentrations of choline (1.3 g of choline/kg of diet), so the mice were not technically cholinedeficient. We hypothesized, however, that because the PEMT reaction generates PC that can be degraded to choline, perhaps the Pemt Ϫ/Ϫ mice were choline-insufficient. Supplementation of the diet with choline eliminated the protective phenotype of

REFLECTIONS: Mechanistic physiology and disease
the Pemt Ϫ/Ϫ mice; thus, the choline-supplemented Pemt Ϫ/Ϫ mice gained weight and became insulin-resistant. These observations indicated that Pemt Ϫ/Ϫ mice are choline-insufficient when fed a high-fat diet. Despite these exciting findings, at this time in 2017, we still have established neither the mechanism by which Pemt Ϫ/Ϫ mice are protected from obesity and insulin resistance, nor the mechanism by which excess choline reverses this phenotype. We hope to have the answer(s) in the next year or two -stay tuned.
Recently, we've returned to one of our previous results, that Pemt Ϫ/Ϫ mice fed a choline-deficient diet experience liver failure. An enterprising graduate student, Zhaoyu Li, took on the challenge of understanding the mechanism leading to this liver failure. In a variety of experimental approaches, we demonstrated that liver failure in choline-deficient Pemt Ϫ/Ϫ mice is due to loss of plasma membrane integrity caused by a decrease in the molar ratio of PC to PE (42). In livers of mice in which PEMT was eliminated, the amount of hepatic PC was decreased, whereas the amount of PE, the substrate for the PEMT reaction, was increased. PC is a bilayer-forming lipid, whereas PE is an non-bilayer-forming lipid. A decrease of PC in the plasma membrane of hepatocytes, as well as an increase in PE, promotes loose packing of the membrane and, consequently, impairs membrane integrity, causing leakage. Thus, these and other studies suggest that the PC/PE ratio is a key regulator of cell membrane integrity and that a low PC/PE ratio plays a role in the progression of steatosis (fatty liver) to steatohepatitis (fatty liver disease) (reviewed in Ref. 43). This conclusion is supported by studies showing that the molar ratio of PC to PE is a predictor of non-alcoholic fatty liver disease (NAFLD) and survival of mice after partial hepatectomy (44), suggesting that the lack of PEMT leads to non-alcoholic fatty liver disease. At this point, there are no examples of humans that lack PEMT.
Similarly, mitochondrial function can be impaired by alterations in the PC/PE ratio. Using Pemt Ϫ/Ϫ mice, Jelske van der Veen, a postdoctoral fellow in the lab, revealed a strong positive correlation between the mitochondrial PC/PE molar ratio and cellular ATP levels (45). In addition, Jean Vance demonstrated that in cells with diminished activity of the mitochondrial enzyme phosphatidylserine decarboxylase, mitochondrial PE levels are reduced, the PC/PE ratio is increased, and ATP synthesis is inhibited (22). On the other hand, in PEMT-deficient mice, as the level of mitochondrial PE increases, pyruvate is diverted from decarboxylation and into the tricarboxylic acid cycle for ATP production. As a result, glucose production from pyruvate is diminished, possibly contributing to the protection of Pemt Ϫ/Ϫ mice from insulin resistance. Moreover, the sensitivity of the response of cellular functions to the PC/PE ratio might explain why the biosynthesis of PC and PE is so tightly regulated and why the rate of phospholipid synthesis is programmed to change within seconds.
When I started investigating regulation of PC biosynthesis, I certainly did not predict where we would be in 2017. It has been gratifying to be part of the process of elucidating control mechanisms in PC biosynthesis. It was challenging and very stimulating to see the field evolve into an understanding of the functions of PC biosynthesis in animal physiology and disease. There is still much to do. How our understanding of PC biosynthesis and function will evolve during the next 45 years is not predictable, but certainly an exciting prospect.

Appreciation
I have indicated the crucial roles that my mentors played in my career. A successful career is also dependent on many other factors. I have had a stable and supportive family life in spades. In particular, Jean Vance has been immensely important in so many ways, both personally and scientifically. Good fortune in recruiting outstanding postdoctoral fellows and graduate students from the very start of my appointment at UBC has provided an exciting scientific environment and many long-term friendships. Stable, continuous grant funding from 1973 through 2019 from the Canadian government has been a crucial factor in my research productivity. Important funding has also been awarded by the Alberta Heritage Foundation for Medical

REFLECTIONS: Mechanistic physiology and disease
Research, private foundations (in particular the Heart Foundations of Canada, the UK, Alberta, and British Columbia, as well as the American Heart Association), and pharmaceutical companies. UBC and the University of Alberta have provided me with excellent lab space and a reasonable teaching/administrative load. My research progress would have been severely hampered without the fantastic technical assistance I received throughout my career. In particular, I would like to recognize Susanne Lingrell who has worked with me at the University of Alberta for more than 30 years. Similarly, Harry Paddon was the key technical assistant at UBC. In addition, I have been fortunate to have had excellent collaborators from the University of Alberta and from countries all over the world. "No man is an island" (from the English poet John Donne in 1624), and this is particularly true in scientific research. I am exceedingly grateful.
I am also grateful for the invitation to contribute this Reflections article to the Journal of Biological Chemistry. Throughout my career, JBC has been, and continues to be, the optimum venue in which to publish my biochemical research. This will be the 88th article that I have published in this esteemed journal.