High-Throughput miRFluR Platform Identifies miRNA Regulating B3GLCT That Predict Peters’ Plus Syndrome Phenotype, Supporting the miRNA Proxy Hypothesis

MicroRNAs (miRNAs,
miRs) finely tune protein expression and target
networks of hundreds to thousands of genes that control specific biological
processes. They are critical regulators of glycosylation, one of the
most diverse and abundant post-translational modifications. In recent
work, miRs have been shown to predict the biological functions of
glycosylation enzymes, leading to the “miRNA proxy hypothesis”
which states, “if a miR drives a specific biological phenotype...,
the targets of that miR will drive the same biological phenotype.”
Testing of this powerful hypothesis is hampered by our lack of knowledge
about miR targets. Target prediction suffers from low accuracy and
a high false prediction rate. Herein, we develop a high-throughput
experimental platform to analyze miR–target interactions, miRFluR.
We utilize this system to analyze the interactions of the entire human
miRome with beta-3-glucosyltransferase (B3GLCT), a glycosylation enzyme
whose loss underpins the congenital disorder Peters’ Plus Syndrome.
Although this enzyme is predicted by multiple algorithms to be highly
targeted by miRs, we identify only 27 miRs that downregulate B3GLCT,
a >96% false positive rate for prediction. Functional enrichment
analysis
of these validated miRs predicts phenotypes associated with Peters’
Plus Syndrome, although B3GLCT is not in their known target network.
Thus, biological phenotypes driven by B3GLCT may be driven by the
target networks of miRs that regulate this enzyme, providing additional
evidence for the miRNA proxy hypothesis.