Genus envy

July 01, 2013

In 1997, a breakthrough was made in rare/orphan disease research. An evolutionarily conserved gene called NPC1 was shown to be responsible for Niemann-Pick disease type C, a degenerative lysosomal storage disorder that affects ~ 1 in 150,000 people on Earth, half of whom manifest symptoms as children. The discovery of NPC1 should have unleashed a torrent of follow up studies in simple model organisms like yeast, worms and flies, all of which have an ancestral version of NPC1. Instead, what followed was a trickle, with clunky rodent models getting all the basic research attention. Is that partly why 16 years later we still don’t have a cure for NPC?


It was once axiomatic to say that model organisms illuminate cellular bits that have been conserved by evolution over the eons. If you need a reminder of just how much our humanity is shared with other life on Earth, check out this cool evolutionary conservation app (h/t @carlzimmer & @javierherroro7). Despite this overwhelming evidence of commonality, the biomedical establishment operates with a mindset of human exceptionalism. According to this mindset: 1) any organism simpler than a mouse or a rat is not relevant to drug discovery; 2) technological advances in human cell in vitro culture and genetic manipulation obviate the need for non-human models. I believe this view is both conceptually flawed and economically inefficient. The basic understanding we so desperately need to cure NPC and thousands of rare/orphan diseases like it will only come from painting meticulous physiological portraits of human disease on a canvas of simple model organisms, starting with our far-removed unicellular cousins.


Here I present Saccharomyces cerevisiae, which goes by several aliases: budding yeast; brewer’s yeast; baker’s yeast. As you can tell from the monikers, we and yeast go way back. Thousands of years ago the lucky bastard who first stumbled upon a natural fermentation put brew and brew together, and our fates have been entwined since. The use of fungi as model organisms in experimental biology dates back to the 1930s and 1940s to the seminal “one-gene, one-enzyme” auxotrophy studies of George Beadle and Edward Tatum on the bread mold Neurospora, as shown in this hand-illustrated slide by Tatum:


beadle and tatum


The genome of S. cerevisiae (hereafter yeast) weighs in at 12 Mb, or megabases, and boasts around 5,000 genes. Depending on how the calculation is done, 20% – 30% of yeast genes have a statistically significant match to a human gene at the DNA level. For scale, the human genome is 3000 Mb, or 200 times larger than the yeast genome, and features ~20,000 genes. Yet most biomedical researchers appear to treat that 20% – 30% as though it were 1%, or ignore it altogether. Have they simply forgotten the literature, or is it the hex of human exceptionalism?


It’s not as though that conserved bloc of genes is chopped liver in terms of cellular functions. Obviously included in this tally are enzymes involved in central metabolism, e.g., glycolysis, or the breakdown of the sugar glucose into chemical energy. But non-metabolism genes and the proteins they encode are also part of the mix. There’s actin and tubulin, two proteins that comprise the dynamic scaffolding, or cytoskeleton, of cells; histone, a protein that wraps DNA double helices in a regulatory embrace; clathrin, the triskelia-shaped protein that forms Bucky Ball coats around lipid droplets called vesicles. And it’s not just the pipes and dry wall that’s shared. Even complex enzymes like kinases are conserved from yeast to humans, including one of my favorites TOR, which stands for Target Of Rapamycin, an ancient nutrient sensor.


The full force of evolutionary conservation is no more persuasively felt than in gene-replacement experiments. If DNA sequence alignment indicates that two genes are related in organisms separated by over a 1 billion years of evolution, how do we know that this DNA sequence similarity translates into functional interchangeability? Swap the modern version for the ancient one, and see if the cell or organism behaves normally. It’s a concept from genetics called complementation. It must have been in those heady days that the expression “the awesome power of yeast genetics” was born:


awesome power


Once I got a taste for yeast in my first-year graduate school laboratory rotations, there was no turning back. In my graduate and postdoctoral research over the last decade, I’ve been trying to connect basic discoveries made in yeast to human diseases, and now my focus is rare/orphan disease. The presence of a yeast gene (called NCR1) that is 34% identical to the human NPC1 gene should be sufficient justification for a comparison. However, I admit that if one were judging relevance solely based on 1:1 correspondences between yeast cells and human beings, the case for yeast would not appear strong at first blush.


NPC is known first and foremost for excessive cholesterol accumulation, though a number of cell membrane lipids also accumulate aberrantly, including sphingolipids. Turns out yeast cells don’t even produce cholesterol — strike one, says the skeptic. Well, yeast synthesize a structurally related sterol molecule called ergosterol. Ergosterol and cholesterol are homologous in the same way whale flippers are homologous to our arms. NPC affects different organ systems, but obviously yeast don’t have brains — so strike two? Well, yeast cell growth defects caused by specific mutations are approximations of tissue-specific cell death in multicellular organisms.


So what have we learned about Niemann-Pick type C from yeast cells?


Let’s start by PubMedding the search terms ‘yeast Niemann Pick.’ This query returns 33 scholarly articles. When you actually read these papers, the cure to NPC isn’t hiding in plain sight, but there are scores of leads crying out for integration with more complex model organisms, validation in patient-derived cells, or further genetic dissection in yeast or other unicellular eukaryotes. Seminal work has been down by two groups. Stephen Sturley‘s lab first used the powerful test of complementation to show that the yeast version of NPC1 could reverse the aberrant cholesterol accumulation of mutant mammalian cells lacking NPC1, as shown below in a reproduction of Figure 2 of Malathi et al:


sturley fig2


In the top right control panel, aberrant cholesterol accumulation is marked by the tiny white arrow inside the cell in the red oval. In the bottom left panel featuring mutant NPC1 mammalian cells that express the yeast version of NPC1, no cholesterol accumulation is observed inside the cell in the orange oval. Yet as hard as they tried, the Sturley Lab couldn’t find other obvious cellular defects associated with cholesterol metabolism, but they did find evidence of aberrant accumulation of sphingolipids, a major component of certain cell membranes. So how are cholesterol homeostasis and sphingolipid homeostasis coordinated by NPC1?


J. Wylie Nichols’s lab tried to address that question and more by building on the foundation laid by Sturley Lab. In a 2005 paper, the Nichols Lab showed that disease-causing NPC1 mutations involve amino acids (marked by asterisks) that are identical or highly conserved between the yeast version (NCR1) and human NPC1:




The amino-acid position circled in orange above corresponds to one of the most common NPC disease alleles, I1061T — the isoleucine at position 1061 is mutated to a threonine. What’s significant about the Nichols approach is they mutated the yeast version of NPC1 to match the disease-causing alleles in human NPC1 patients. Yeast NCR1 mutants don’t have any strong growth defects on their own, though there are simple explanations like the media and strain background that can suppress modest growth defects. But these “NPC patient-matched” yeast mutants do exhibit resistance to a class of lipid poisons. Resistance to lipid poisons in yeast is not the same as Niemann Pick disease in humans, but it’s enough to enable more vista-opening genetic screens, the kinds of screens that will reveal how NPC1 actually works.


Studying yeast alone is not going to cure NPC, but if you take evolutionary conservation at face value, the awesome power of yeast genetics is a modest downpayment on a cure. In the next installment, I’ll connect the above yeast work to the handful of studies of NPC in worms and flies. Due to their greater similarity to humans, these multicellular model organisms reveal much more about NPC1 functions, and offer ways to modify or even reverse disease progression.


  • Matt Rich

    I think that one of major stepping stones here is figuring out complementation for all those yeast genes that seem like they should complement a human knockout (or vice versa, although the yeast-to-human argument is a stronger one), and a couple people (one of them being me) have found that there doesn’t seem to be a straightforward relationship between aspects like homology and whether a yeast gene can complement a human gene. Both myself (during my rotation with Maitreya Dunham) and a Ed Marcotte’s lab have worked toward figuring some that out in essentially the same way — taking essential yeast genes that have human homologs (in my case 1:1:1 homologs yeast:mouse:human, of which there are, since I can’t find my slides, maybe 200), knocking them down conditionally, and then trying to rescue viability using expression from the human homologs. IIRC, there was very little correlation between things like homology, # of interactors, etc… that could predict how well a given gene would complement, which makes it seem like an important study is to simply blast through all of them and see who works. Of course, we were both trying at “humanizing” yeast — I would think that it’s safe to assume that the other way ’round isn’t going to be the same, and making that assay high throughput is going to be a lot harder.

    • Ethan O. Perlstein

      Thanks for the comment, Matt! It’s interesting what you say about the 1:1:1 homologs. What was the relationship between homolog pair sequence identity and strength of complementation?

      I’m really interested in cases where disease alleles affect identical or highly conserved amino-acid positions in pairs of disease gene homologs. Also, I don’t think complementation is enough. If the goal is personalized model organisms, we’re going to have to dial in not only the disease-causing mutations but also the disease-modifying mutations. Monogenic diseases vary in ages of onset and severity; modifier loci are thought to explain this variation, as they do in genetic screens using simple model organisms.

      • Matt Rich

        IIRC (and Marcotte said more about this when he was here, compared to what little I accomplished during that rotation) there wasn’t really any correlation between homology and complementation, which makes things complicated. I’ve been waiting for that paper to come out, but haven’t seen anything yet.

        And yeah, complementation isn’t enough — but it’s a good qualifier for determining the genes where you can safely make human::yeast comparisons. It also has to be the first step in the disease pipeline you’re describing, I’d say.

  • Pedro Mendes

    Nice piece, I enjoyed it. But the use of yeast as a model organism goes further back: the birth of biochemistry, with Edward Buchner and others following him, was all done in yeast in the 1890’s

  • Ash

    So it’s interesting. On one hand they don’t rely enough on yeasts and fruit flies. On the other hand they often rely too much on rodents.

    • Ethan O. Perlstein


  • David

    I think that it would strengthen your case if you could write about a clear case of yeast research leading to a new treatment for human patients. And, I don’t mean in some abstract sense of yeast “shedding light on” or “providing insight into” a human disease. What’s the most direct link that you can come up with between yeast research and clinical treatment? (Excluding, of course, anti-fungal drug development, as that doesn’t seem to fall within your experimental scope.) I’d be really interested to learn what the answer is!