Cell-cycle quiescence maintains Caenorhabditis elegans germline stem cells independent of GLP-1/Notch

Essential revisions:

The experiments are very carefully done and clearly presented. This work lays an important foundation for future molecular genetic studies to address mechanism. The manuscript has a few shortcomings, however, which can for the most part be addressed textually. The manuscript should do a better job of incorporating two recent advances in this field. The first, from one of the reviewers, Dr. Greenstein (Starich et al., 2014) in which it was shown that soma germline gap junctions are as important as the GLP-1/Notch pathway for controlling germline stem cells. In the absence of these junctions, germ cells can neither proliferate nor differentiate, even if the GLP-1/Notch pathway is constitutively activated. These gap junctions enable the transit of small molecules, ions, and metabolites between germline and soma, thereby providing a potential mechanism for nutritional inputs to the germline.

Good suggestion. We have included this information in our expanded Introduction.

The second is a major paper from Tim Schedl's lab (Fox and Schedl, 2015), which substantially refines our understanding of germline stem cell behavior in C. elegans. After, this publication, it seems questionable (or controversial at the very least) whether it is correct to refer to a proximal population of "transit-amplifying cells" in this system (as in Figure 1A here). A more complete consideration of Fox and Schedl (2015) raises an alternative interpretation of the authors' glp-1/Notch temperature-shift experiment (Major Point 1, below).

Fox and Schedl (2015) present a careful and detailed analysis of germline stem cell behavior after loss of GLP-1/Notch. They report that when GLP-1/Notch signaling is removed, germ cells undergo one (or sometimes two) divisions before entering meiotic prophase. This finding is fully consistent with our previous understanding of cells in the proximal progenitor zone being in a transitional yet mitotically dividing state, and this finding is also consistent with the results of our current manuscript.

The observations of Fox and Schedl (2015) lead them to hypothesize that germ cells in mitotic S-phase or G2 are not able to enter meiotic prophase without passing through a cell division. Their hypothesis raised an alternative interpretation of our glp-1/Notch temperature-shift experiments, but their hypothesis was directly refuted by our cdk-1 RNAi experiments. The reviewers raise questions about our cdk-1 RNAi experiments and ask for further confirmation of these experiments using HIM-3 staining (Major Point 1, below). We performed HIM-3 staining, as requested, and found that HIM-3 staining supported our original conclusions. Thus, our results do not support the interpretation suggested by Fox and Schedl (2015).

The reviewer also uses Fox and Schedl (2015) to question our use of the term ‘transit-amplifying cells’ to refer to cells in the proximal progenitor zone. In response to this criticism, we sought a suitable replacement term and selected ‘transient progenitors’ as a possibility. This term follows the suggestion of Fox and Schedl (2015) to use ‘progenitors’, with the addition of ‘transient’ to convey the idea of a transitional state.

In selecting a term to refer to germ cells in the proximal progenitor zone, we sought to convey two ideas, neither of which is controversial. First, we wished to convey the idea that cells in the proximal progenitor zone are in the mitotic cell cycle. This idea is supported by many studies, including Fox and Schedl (2015). Second, we wished to convey the idea that cells in the proximal progenitor zone are in a transitional state. This idea is supported by three key observations: (i) cells in the proximal progenitor zone have higher levels of the meiosis-associated protein GLD-1 than cells in the distal progenitor zone (many studies); (ii) cells in the proximal progenitor zone have lower levels of the stem cell regulators sygl-1 and lst-1 than cells in the distal progenitor zone (Kershner et al., 2014); and (iii) cells in the proximal progenitor zone enter the meiotic cell cycle when their movement out of the progenitor zone is restricted, whereas cells in the distal progenitor zone arrest in the mitotic cell cycle and do not enter meiosis with that treatment (Cinquin et al., 2010). Based on these multiple lines of evidence, we sought terminology that would convey the idea that cells in the proximal progenitor zone are in a transitional yet mitotically dividing state. The term ‘transit-amplifying cells’ terms has long been used in the stem cell literature to mean just that – progeny of stem cells that continue to divide mitotically but that have limited division potential and have begun to differentiate.

We have now replaced ‘transit-amplifying cells’ with ‘transient progenitors’ throughout the revised manuscript. Yet we have mixed feelings about this replacement. The reviewers argue that ‘amplifying’ may mislead readers about the number of divisions occurring in the proximal progenitor zone. This problem was addressed in our original manuscript by informing readers that the number of divisions in this zone is one or two (and this information is retained in our revised manuscript). Moreover, the term ‘transit-amplifying cells’ has been used throughout the vertebrate literature to refer to cells with similar properties to cells in the proximal progenitor zone. Does the confusion caused by ‘amplifying’ outweigh the confusion caused by adding a new term to the literature? We ask the reviewers to consider the pros and cons of the change in terminology and the need for a new term. Our preference is to use the term ‘transit-amplifying cells’, but we agree to replace it with ‘transient progenitors’ if the reviewers remain unconvinced.

The authors should consider the following specific points:

1) The authors interpret the result that starvation in glp-1(ts) does not cause cells to enter meiosis at the non-permissive temperature as meaning that "starved animals maintain germline stem cells independent of GLP-1/Notch." Given the results of Fox and Schedl (2015) that the majority of glp-1(ts) germ cells complete a single division before entering meiosis, it seems quite possible this is a consequence of the G2 arrest of cells after starvation. Upon moving back to 15C and re-feeding, germ cells complete their mitotic division and return to a situation in which GLP-1/Notch signaling is re-established. The authors interpret a cdk-1(RNAi) experiment as evidence that a G2 block is insufficient to prevent meiotic entry. Inspection of the Methods section associated with this experiment (subsection “Temperature shifts”, third paragraph) does not provide sufficient information as to the reliability of the method to score meiotic entry in these gonads by DNA morphology. The authors should consider evaluating meiotic entry by analyzing the recruitment of meiosis-specific chromosomal proteins to chromatin (e.g., HIM-3, etc.). At the very least, readers should be able to see for themselves what these gonads look like. The authors seem to be claiming that GLP-1/Notch signaling becomes non-essential during starvation conditions, not that it is turned off. If the latter was the case, then reporters for sygl-1 and lst-1 might be revealing.

As requested, we have now performed HIM-3 staining in glp-1(q224ts) animals treated with cdk-1 RNAi and shifted to the restrictive temperature. We observed that HIM-3 was recruited to chromosomes following the temperature shift (Figure 5–figure supplement 1). Thus, cell-cycle arrest is insufficient to inhibit meiotic entry. This result strengthens our conclusion that the ability of quiescence to maintain germline stem cells in absence of GLP-1/Notch is not simply an effect of cell-cycle arrest.

We also note that Cinquin et al. (2010) reported entry of cell cycle-arrested germ cells into meiotic prophase without DNA replication and without mitotic division:

“In unablated controls, the distal germ cells did not enter meiosis, as assayed by absence of abundant GLD-1 (100%, n = 4; Figure 2B), absence of crescent-shaped DNA (100%, n = 10; Figure 2B), and absence of chromosomal HIM-3 (100%, n = 6). By contrast, when the DTC was ablated, the distal germ cells accumulated abundant GLD-1 (100%, n=5; Figure 2C) and entered meiotic prophase, as assayed by both crescent-shaped DNA (100%, n = 9) and HIM-3 localization (100%, n = 4; Figure 2D). We do not understand the mechanism by which germ cells escape cell cycle arrest after DTC ablation, but note that some nuclei with crescent-shaped DNA and chromosomal HIM-3 had not undergoneDNAsynthesis after DTCablation (see Methods, Figure 2E). Therefore, these cells appear to have progressed from their cell cycle arrest to meiotic prophase without DNA replication.”

2) The Introduction needs some rewriting. For example, readers would benefit from a more thorough review of the C. elegans literature and that of Drosophila, as pertains to how nutrition affects germ cells.

Good suggestion. We have expanded the Introduction to include a more thorough review of how nutrition affects germ cells in larval C. elegans and in Drosophila.

3) In the Introduction, or in the Results section (near the end of the subsection “During starvation, germ cells progress slowly through S phase and arrest in G2”), point out that PGCs arrest in G2 during embryogenesis and remain arrested in G2 when the newly hatched larvae are starved, only undergoing proliferation upon feeding (Fukuyama et al., 2006).

Another good suggestion. We have added this information to our expanded Introduction.

4) Figure 7A and the subsection “Starvation-induced quiescence is distinct from cell-cycle arrest induced by DNA-damage”: that the "phosphor-inhibited CDK-1" antibody reports only P-CDK-1 and does not cross-react with unphosphorylated CDK-1 in the immunostaining experiment has not been validated here or in Craig et al. (2012). The authors should present a validation of this reagent.

The phospho-specificity of our phospho-CDK-1 antibody was validated in C. elegans embryos by Rahman et al. PNAS 2014, which we cite in our Methods section. We have now additionally validated the phospho-specificity of this antibody in the C. elegans germline, by showing that staining disappears in germlines treated with lambda protein phosphatase (Figure 7–figure supplement 1).