This study shows, in the context of an intact vertebrate organism, that Pwp2h is critical for the production of mature 18S rRNA, an integral component of the 40S ribosomal subunit. In zebrafish, as in yeast, Pwp2h depletion results in reduced levels of the immediate precursor to mature 18S rRNA and a concomitant decrease in the production of mature 18S rRNA and assembly of 40S ribosomal subunits. Thus the role of Pwp2h in the 90S pre-ribosomal particle or small subunit processome is conserved from yeast to vertebrates.
In our pwp2h-deficient model, titania (ttis450), the growth of the endodermal organs, eyes, brain and craniofacial structures is severely arrested and autophagy is markedly up-regulated. To the best of our knowledge, this is the first time that a link between impaired ribosome biogenesis and autophagy has been demonstrated. We further show that elevated rates of autophagy support the survival of intestinal epithelial cells and increase the lifespan of ttis450 larvae, thereby demonstrating that autophagy is a survival mechanism invoked in response to ribosomal stress. In our zebrafish model, autophagy induction does not depend on inhibition of the Tor pathway or activation of Tp53.
The death of ttis450
larvae at 8–9 dpf demonstrates that pwp2h
encodes a protein that is indispensable for life. However, the development of ttis450
larvae until 72 hpf is supported by the deposition of maternal, wild-type pwp2h
mRNA (and/or protein) into oocytes by their heterozygous mother. At 72 hpf, the tissues in which pwp2h
is most highly expressed are the intestinal epithelium, pharyngeal arches, liver, dorsal midbrain, cerebellum, dorsal hindbrain, retinal epithelium and pancreas. These tissues are also the most rapidly proliferating tissues in WT larvae at 72 hpf 
and the most severely affected tissues in ttis450
larvae. Thus the tissue-specific phenotype of ttis450
larvae may be explained by maternally-derived WT pwp2h
mRNA being exhausted first in developing organs containing highly proliferative cells.
In WT zebrafish larvae there is a transient spike in Torc1 activity (as measured by p-RPS6) at around 72 hpf that is coincident with the activation of anabolic pathways required for cell growth and proliferation during the endoderm to intestine transition 
. Torc1 is thought to play a role in developing organisms as an organ size checkpoint, potentiating growth signals that promote the rapid expansion of organs until they reach a genetically programmed cell size 
. Therefore the persistent and robust activity of Torc1 we observe in the intestinal epithelium and liver of ttis450
larvae at 96 hpf may be a consequence of these organs being markedly smaller than their WT counterparts at this stage.
The gross phenotype of ttis450
is highly reminiscent of another zebrafish mutant, nil per os (npo)
, in which the morphogenesis of the intestinal epithelium is also arrested. In npo
the failure of the primitive gut endoderm to transform into a monolayer of polarized and differentiated epithelium is caused by a mutation in rbm19
, a gene encoding a protein with six RNA recognition motifs that is also thought to play a role in ribosome biogenesis 
. The same authors showed that essentially the same hypoplastic intestinal phenotype was recapitulated by exposure of WT zebrafish larvae to the Torc1 inhibitor, rapamycin 
, which presumably stimulated autophagy. It would be interesting to determine whether the growth arrest of the digestive organs in the npo
mutant is also accompanied by autophagy.
The degree of activation of the Tor pathway is thought to be one of the major factors governing autophagy. However, Tor inhibition is not the mechanism responsible for autophagy in ttis450
larvae and recent work suggests that autophagy regulation is a very complex process involving the integration of signals from many diverse signalling pathways 
. Indeed, proteomic analysis of binding partners of components of the autophagy machinery suggests that several hundred molecules participate in the regulation of the human autophagy network 
. While much recent attention has been focused on the direct phosphorylation of Ulk1/Atg1 by AMPK, acting either cooperatively or independently of Tor to exert autophagy control 
, there are many reports of other kinases capable of controlling autophagy by a variety of Tor-independent mechanisms 
. The dissociation of the key BH3 domain-containing autophagy protein, Beclin 1 (mammalian orthologue of yeast Atg6) from its inhibitors Bcl2 and Bcl-XL as a result of phosphorylation of one or other components is also a critical determinant in the induction of autophagy 
. In the case of ttis450
larvae, it is plausible that autophagy induction may involve a targeted pathway, selective for ribosomes 
, which by analogy with mitophagy 
, is invoked to digest damaged cargo such as non-functional organelles.
Somewhat surprisingly, we also ruled out involvement of Tp53 in the induction of autophagy in ttis450 larvae, even though Tp53 protein is active in ttis450 larvae at 96 hpf. However, we believe the increased expression of Tp53 target genes such as p21 and cyclinG1 may be responsible, at least in part, for the reduction in the number of cells in the S phase of the cell cycle we observed at this time-point. To explain this, we surmise that as ribosome biogenesis is progressively impaired, the ttis450 larvae mount a two-stage response to Pwp2h depletion. Initially, the cells undergo a Tp53-mediated cell cycle arrest. However, as the synthesis of new proteins, including Tp53 and its targets, is progressively impaired, the cells invoke autophagy to prolong their survival.
The notion of the existence of a second type of programmed cell death, distinct from apoptosis, which emanates from catastrophic levels of autophagy, is a hotly debated topic 
. Using TEM, we did not see any evidence of cell death in the IECs of ttis450
larvae, even at 7–8 dpf just before the larvae die, affirming that the levels of autophagy induced in the IECs of ttis450
larvae prolong cell survival rather than trigger cell death. We proved this by disrupting the formation of the early autophagosome by inhibiting the translation of atg5
mRNA. This resulted in the death of IECs in ttis450
larvae and a markedly reduced lifespan.
As mentioned previously, ttis450
larvae exhibit impaired development of the craniofacial cartilages, exocrine pancreas and brain, tissues that are often clinically abnormal in patients with certain human ribosomopathies, including Diamond Blackfan anaemia and Schwachman Diamond syndrome 
. Recently, two new zebrafish models of dyskeratosis congenita (DC) based on mutations in components of the H/ACA RNP complex were described 
. Like ttis450
, these mutants display impaired production of 18S rRNA and induction of Tp53 target genes, consistent with previous studies demonstrating that defects in ribosome biogenesis induce Tp53 activation and cell cycle arrest 
. Moreover, hematopoietic stem cells in these mutants were depleted via a Tp53-dependent mechanism, providing a plausible explanation for why DC patients are susceptible to bone marrow failure 
. In one of these mutants, the gut and craniofacial structures were also reported to be underdeveloped and, as observed in ttis450
, these defects persisted on a Tp53 mutant background 
. We speculate that the p53-independent features of this model of DC may be caused by elevated rates of autophagy. If so, and these findings are confirmed in human DC, it will be important to determine whether elevated autophagic activity contributes to prolonged cell survival prior to considering clinical interventions to limit this process.
There is currently a great deal of interest in the development of novel therapeutics that target the cancerous translation apparatus through the combined inhibition of ribosome biogenesis, translation initiation and translation elongation 
. To avoid inadvertently prolonging cancer cell survival, these approaches could benefit from a detailed understanding of the mechanisms and cellular contexts that induce autophagy in response to ribosomal stress. While such insights may be forthcoming from studies performed on cell lines, it is likely that complementary experiments carried out in the context of an entire vertebrate organism, such as the zebrafish model introduced here, may also be fruitful.