Form and function of Poly(A) tails

Abstract

Central to mRNA metabolism is the poly(A)-tail, a stretch of adenosine nucleotides at the mRNA 3' end. In this dissertation, I investigate the role of the tail in the dynamics of mRNA decay, and describe the predominant mechanisms of decay for thousands of mammalian mRNAs. Next, I examine the effects of microRNAs, which influence mRNA decay and perturb tail length dynamics. Finally, I describe a physiological context in which the tail helps to control translation: neurons of the mouse brain. mRNA decay is tightly regulated in eukaryotes, determining the steady-state abundances and rates of accumulation of mRNAs. Despite this central role, the dynamics of decay have been described for only a handful of mRNAs. We determine these dynamics for thousands of endogenous mRNAs. Nascent mRNAs have reproducible and heterogeneous tail lengths just after they escape the nucleus.

Once in the cytoplasm, most mRNAs are substrates for deadenylation, the rates of which vary by over 1000-fold, a range sufficiently large to capture the variation in mRNA decay rates. Surprisingly, once their tails become short, mRNAs decay at rates that also span a 1000-fold range. Moreover, these rates are coupled to their deadenylation rates, suggesting a concerted process of remodeling the mRNA--protein complex during decay. MicroRNAs (miRNAs) are small RNAs that influence decay of mRNA targets by recruiting deadenylases. Despite this recruitment, we observe no changes to steady-state tail length for miRNA targets. Resolving this paradox, we find that miRNAs not only deadenylate their targets but also increase the decay rate of short-tailed target molecules. By enhancing both rates, miRNAs do not alter the distribution of tail lengths of target mRNAs but enhance the rate at which mRNAs traverse these lengths. Neurons have unique requirements for translational control.

We perform ribosome profiling in primary neuronal cultures and brain tissues from a mouse. mRNA poly(A)-tail lengths explain some (~5%) of the large variance in translational efficiency we observe, as does coding-sequence length, expression level, and codon composition. For some mRNAs, neuronal stimulation modifies tail lengths, and for a subset, transcription cannot explain these changes. A linear model that uses known determinants to predict translational efficiency explains only a portion (30- 40%) of its variance, indicating the need for additional investigation of mechanisms of translation in neurons.