HomeInVivo Biosystems BlogDisease ModelingA View From the Bench: Roadblocks and Frustrations When Expectation Meets Reality.

A View From the Bench: Roadblocks and Frustrations When Expectation Meets Reality.

In an effort to create better models for epilepsy, we’ve been working on a grant-funded project to humanize important synaptic machinery by replacing endogenous coding sequences with human coding sequences.  Syntaxin is a key synaptic protein, encoded by the gene STX1A, that helps synaptic vesicles dock, fuse with the membrane, and release neurotransmitters.  Syntaxin has a homolog in C. elegans known as UNC-64.

Before humanizing a strain with STX1A, we first sought to create an unc-64 knock-out. The knock-out strain phenotype is an important control group to make sure that the humanized STX1A is rescuing some of the innate functionality of the homologous gene. We have a ton of experience making knock-outs and after a brief glance at the literature, we had no reason to expect that we should have any trouble with this one. After all - in mice, the STX1A null mutants develop normally, and the reference allele on wormbase.org is non-lethal and exhibits common movement phenotypes (i.e., sluggish, unc, fainter) (Fujiwara et al. 2017). We moved ahead with confidence that we would be able to generate a homozygous null mutant. But in this case...well, let’s just say we ran up against some roadblocks and frustrations felt by many researchers when expectations don’t match with reality.

Due to our high success rates with CRISPR, we were confident we would find a homozygous mutation quickly. When we screened 39 of our F1 generation, we found only heterozygous (het) or wildtype (WT) animals. While disappointing, this is not completely abnormal, so we proceeded to the F2 generation with the expectation that basic mendelian inheritance (1 out of every 4 animals with a homozygous mutation) would prevail. After screening 69 of our F2 generation we still only had WT and Het animals!  By this time, we were frustrated, but through our practical knowledge gained by creating thousands of animal models we know that mendelian genetics can and will betray you.  However, getting homozygotes generally works in larger populations: if you don’t find it the first  (or second) time, test more animals. So we more than doubled the number of F3 animals screened, but still no homozygous mutant was found.

Ok, this was getting annoying. After revisiting our original research we thought that we might be selecting away from the mutation due to preferentially picking crawling animals. So we picked F4s based on an expected dumpy/movement phenotype. We still could not find any homozygous mutants!  With hundreds of animals tested (342, not that I was counting), patience had run out.  Having a thorough and rigorous screening method allowed us to be confident that we were not missing the mutation. I began to ask myself for another explanation. 

One of our founders has a saying,  “One month in the lab can save you a whole day in the library!” he’ll quip, highlighting the importance to looking at the primary literature. I decided to take the advice. After a little digging, I found an older paper that stated the following: “A complete loss-of-function mutation in unc-64 results in a worm that completes embryogenesis, but arrested development shortly thereafter as a paralyzed L1 larvae, presumably as a consequence of neuronal dysfunction,” (Saifee, Wei, and Nonet 1998). Ah-ha! Wow. This seemed to be exactly what was happening for us.  I went back to Wormbase to dig into the reference allele that we had checked for the phenotype for earlier, and sure enough, it was not a null mutation. Next time, I’ll be more trusting in our screening process to give us the right answer…but I’ve also learned that digging into the primary literature can save your sanity. 

Once we had identified the true null phenotype, we moved ahead with our humanization. This process went smoothly (a much needed reprieve) and we were able to identify homozygous mutants within our first screening generation. When compared to our experience trying to generate the null model this easy process demonstrated to us that our humanized gene will function in the C elegans. We had successfully rescued the null phenotype! This gives us the ability to work with a human avatar and further explore disease models for epilepsy.


About the Author: Jenn Lawson (R&D Tech III)

Jenn is a R&D Lab Tech III at InVivo Biosystems, specializing in C elegans transgenics. She received her Bachelor of Science from the University of Utah. Her background includes over 16 years of professional research experience in both academic and biotech labs. Prior to joining InVivo Biosystems she worked with and contributed to numerous published scientific studies using mouse models. She is happy to now be working with an animal model that does not bite. Jenn loves to read and is fascinated by human behavior and evolutionary psychology.

Kat McCormick - Director of R&D at NemaMetrix

About the Editor: Kat McCormick, Ph.D (Director of Business Development)

Kat has a BA from Bryn Mawr College, where she studied leech electrophysiology, and a Ph.D. from University of Oregon, where she studied the neuronal basis of navigation under Dr. Shawn Lockery.  She joined InVivo Biosystems in 2014 and led the R&D group for 5 years before transitioning to a role in Business Development. Outside of work, she enjoys reading, cooking, and spending time with her 3 year old and new baby.

Share this article
Scroll to Top