By Taryn Sumabat
This summer at Perlara, we ramped up our efforts to develop a Drosophila model of Niemann-Pick Type A (NPA). NPA is a rare lysosomal storage disease caused by mutations in the gene Smpd1, which encodes an enzyme called acid sphingomyelinase (ASM). Normally, ASM helps break down a fatty substance called sphingomyelin in our cells, but this function is impaired in patients with NPA. Consequently, sphingomyelin builds up to harmful levels in the cells of major organs like the liver, lungs, spleen, and brain, ultimately resulting in organ failure and death of patients before age three. Sadly, no treatment options currently exist for individuals with NPA. In PerlQuest partnership with Wylder Nation Foundation, Perlara is hoping to fix this, with a little help from our invertebrate friends.
To develop a model of NPA in Drosophila, we first sought to understand the function of its Smpd1 equivalent. The fly version of Smpd1 is a single gene called CG3376, which shares ~42% protein identity with human Smpd1. All Drosophila genes have a unique “CG” number yet those that have been characterized are usually referred to by their name. A gene without a name is often an indication that this gene is understudied. Indeed, very little research on CG3376 has been published, meaning that there remains much to uncover regarding the function of this gene. A study from 2002 found that CG3376 is expressed in the early fly embryo, specifically in structures that will form the pharynx, esophagus, and trachea, which are involved in digestive and respiratory functions. CG3376 may thus have important roles during early development, yet no reports of CG3376 mutants exist.
With no known CG3376 mutants available to the Drosophila research community, Perlara enlisted the help of Genetivision to make our own. Using CRISPR/Cas9, Genetivision was able to create a CG3376 genetic null, in which the entire coding sequence was deleted and replaced with a red fluorescent protein (dsRed).
We were excited to see what phenotypes would result from this null allele, but it turned out that these flies were inviable as homozygotes. NPA is a recessive disorder, meaning that an affected individual has inactivating mutations in both of his/her copies of Smpd1. Our CG3376 null seemed to also act as a recessive allele, given that it could only be maintained in heterozygotes, which had normal-looking appearances and behaviors. This suggests that CG3376 is an essential gene, as complete loss of CG3376 function results in lethality.
When during development are the homozygous-null flies dying? Lethality during a fly’s late larval or pupal stage of development is not too difficult to detect: dead larvae appear as blackened carcasses and dead pupae never come out of their pupal case. However, the CG3376 null flies did not show either of these signs. Thus, we reasoned that the homozygous flies might be dying as embryos or as very early larvae.
To test this hypothesis, we performed a quick phenotyping experiment using the Biosorter. First, we collected a couple thousand embryos from the CG3376 null strain. The following day, we ran these animals—which were by that point a mix of unhatched embryos and recently hatched larvae—through the Biosorter. The Biosorter is capable of separating animals marked with Green Fluorescent Protein (GFP) from non-GFP-expressing animals. Because our CG3376 heterozygotes had a GFP fluorescent marker that would not be present in CG3376 homozygotes, we wanted to see if the Biosorter would pick up any non-GFP animals. In fact, the Biosorter was able to recover newly-hatched larvae that lacked GFP!
When we examined these larvae more closely, we noticed a large number of them appeared dead, while those that were still alive moved abnormally. Moreover, we recovered some GFP-negative embryos that, given a closer look, had a ready-to-hatch larva wriggling around inside. Thus, the homozygous-null animals appear able to proceed through embryonic development but die soon after hatching, possibly due to the inability to move and/or feed properly. Undoubtedly, these results leave us with many questions about the biological function of CG3376. However, the early lethality also poses a significant challenge to our drug discovery efforts, as it would be incredibly tough to find a compound that could rapidly reverse this effect.
To navigate around this issue of early homozygous-lethality caused by the null allele, Perlara adopted an alternative approach of generating CG3376 hypomorphs (loss-of-function mutations with less severe phenotypic effects). Again, we worked with Genetivision to design a genome-editing strategy using CRISPR/Cas9. This time, a single guide RNA (gRNA) was used to direct Cas9 to make a targeted cut in the CG3376 DNA sequence.
If these breaks in the DNA are “repaired” through an efficient yet error-prone process called non-homologous end-joining, nucleotides can be inserted or deleted at the cleavage site. The resulting insertion/deletion (indel) mutations may give rise to an altered protein sequence, such as a premature stop codon. Our logic was that by targeting the DNA cleavage event to a region that is roughly 5/6 of the way through the coding sequence, the resulting polypeptides would still retain sufficient protein function to prevent the early lethality we saw with the null. In fact, we managed to recover 15 unique indels but to our surprise, they all appear to be homozygous-lethal! As with the null allele, we do not observe any signs of late larval or pupal lethality, suggesting that animals that are homozygous for these indel mutations may also die during early post-embryonic development (though we have not yet tested this using the Biosorter).
As summarized in the table below, most of the indels we recovered are expected to result in premature stop codons.
Many organisms utilize a surveillance mechanism called nonsense-mediated decay to eliminate mRNA transcripts that possess premature stop codons, thereby preventing the translation of these transcripts into their truncated protein products. This phenomenon may explain why some of our indels behave similarly to the null allele. However, we also obtained 3 protein deletions and 2 protein insertions that should not cause early stop codons. What could explain the surprising severity of these mutations? One possibility based on the published structure of human ASM is that these changes may affect certain secondary structural elements called alpha helices that normally support the protein’s proper form and function.
Given the homozygous-lethality of our CG3376 indels, we set up crosses between each unique combination of alleles to determine if any two allele combination is viable past the early larval stage. So far we have not been able to recover a viable combination, though our testing is not yet complete.
Despite the unexpected challenges of developing a Drosophila model of Niemann-Pick Type A, our work thus far offers us another case study in targeted mutagenesis that will be valuable for designing future disease models in flies. Not only that, it highlights the important role that CG3376 plays during Drosophila development. Clearly there is some interesting biology involving the Drosophila equivalent of acid sphingomyelinase that may inform our understanding of NPA. We have only just begun scratching the surface.