Discussion
We set out to investigate the possibility that ribosomal rRNA
inheritance is a major contributing factor to establishing
neuron-specific translation capacity. Using cell type-specific
EC-tagging and standard EU-tagging, we found that nascent rRNA synthesis
is very limited in neurons. We also found that progenitor rRNA, most
likely in pre-rRNA form, associates with mitotic chromosomes and is
passed to progeny at cytokinesis. The functional importance of rRNA
inheritance was revealed by RNAi experiments targeting RNA polymerase I
in either progenitors or neurons. Knockdown of RNA polymerase I in
neurons had no effect on neurodevelopment or protein synthesis while
knockdown in progenitors caused a severe neurodevelopment defect and
significantly reduced protein synthesis in neurons. Our results support
a model in which neurons utilize inherited rRNA to meet their protein
synthesis needs.
While we primarily focused on rRNA inheritance in neurons, our data also
suggest that rRNA is passed from neuroblasts to INPs and GMCs. In 4-hour
EU-labeling experiments, tagged rRNA is strongly detected in these
differentiated progenitors and this signal may reflect both nascent and
inherited rRNA. The average neuroblast cell cycle is about 1.4 hours
[27] so multiple INPs and GMCs are produced during a 4-hour EU
feeding. Since we could not reliably detect tagged RNA using shorter
labeling times, it is unlikely that the strong signal in INPs and GMCs
is solely derived from rRNA synthesized in neuroblasts during interphase
(no rRNA synthesis occurs during mitosis [28]). Based on this
relationship between labeling time and neuroblast cell cycle length, we
predict that at least some nascent rRNA synthesis occurs in INPs and
GMCs. This raises the question of how rRNA transcription is activated in
progenitors that do not express Myc. One possibility is that sufficient
rRNA polymerase I components, cofactors, and rRNA processing enzymes are
inherited from the neuroblast. These inherited factors may support
nucleolus assembly and rRNA synthesis in INPs and GMCs. Nucleolar
proteins and pre-rRNA associate with mitotic chromosomes in mammalian
cells and are thought to direct nucleolus formation in daughter cells
[28]. A similar mechanism could direct nucleolus formation and
activity in neuroblast progeny. In this model, the absence of Myc in
GMCs would limit production of nucleolar factors and RNA polymerase I
components so that little if any of the machinery necessary to trigger
rRNA synthesis is inherited by neurons.
This work contributes to a growing body of evidence that the protein
synthesis landscape of neurons is distinct from neural progenitors.
While proliferating cells have a high demand for growth-supporting
translation, the ribosomal properties of neurons might differ for
reasons beyond their post-mitotic status. One possibility is that
decreased ribosome concentration establishes neuron-specific translation
programs. For example, a decrease in ribosome concentration can
selectively limit translation of mRNAs whose translation is difficult to
initiate [4]. Our findings may also relate to changes in ribosome
requirements that occur throughout the lifetime of a neuron. Immature
neurons require ribosomes in their axon growth cones to support
pathfinding and synapse formation, but ribosomes are essentially absent
from the axons of mature neurons. A ubiquitin-proteosome mechanism
removes ribosomes from mature axon terminals [14] but this system
could be overwhelmed if mature neurons produced large amounts of
ribosomes. Inherited rRNA in newly born neurons may help establish
ribosome levels appropriate for growth and synapse formation while weak
rRNA synthesis in mature neurons may help ensure ribosomes do not
accumulate outside of dendrites and the soma. Similar evidence of
dynamic changes in rRNA levels comes from EU imaging in cultured
hippocampal neurons and larval zebrafish brains [32]. This study
identified a general decrease in RNA synthesis upon neuronal
depolarization and a restriction of RNA synthesis to the neurogenic
regions of the brain. Finally, reduced ribosome abundance may serve a
protective function: lower levels of protein synthesis decrease the
likelihood of generating misfolded proteins [1] that may be toxic to
neurons.
Here we show that rRNA synthesis is restricted along the differentiation
axis and that inherited rRNAs are sufficient for neurodevelopment and
protein synthesis. Inheritance of ribosomal proteins remains to be
investigated. We have previously shown that ribosomal protein mRNAs have
decreased stability in neurons compared to somatic cells [33].
Ribosomal protein abundance in progenitors and neurons may also be
regulated via mRNA decay and coordination of rRNA and ribosomal protein
levels during differentiation is an important avenue for future studies.
It will also be interesting to investigate differences in ribosome
production among individual neuroblast lineages. A recent study on the
effects of nucleolar stress found that neuroblasts that produce mushroom
body neurons are less sensitive to nucleolar stress and have greater
reserves of nucleolar proteins than other neuroblasts [34]. Studies
of ribosome synthesis and inheritance in Drosophila will help
identify conserved mechanisms of neural translation and may contribute
to our understanding of ribosomopathies that cause multiple human
diseases [35].