Cell-type-specific interacting proteins collaborate to regulate the timing of Cyclin B protein expression in male meiotic prophase

ABSTRACT During meiosis, germ cell and stage-specific components impose additional layers of regulation on the core cell cycle machinery to set up an extended G2 period termed meiotic prophase. In Drosophila males, meiotic prophase lasts 3.5 days, during which spermatocytes upregulate over 1800 genes and grow 25-fold. Previous work has shown that the cell cycle regulator Cyclin B (CycB) is subject to translational repression in immature spermatocytes, mediated by the RNA-binding protein Rbp4 and its partner Fest. Here, we show that the spermatocyte-specific protein Lut is required for translational repression of cycB in an 8-h window just before spermatocytes are fully mature. In males mutant for rbp4 or lut, spermatocytes enter and exit meiotic division 6-8 h earlier than in wild type. In addition, spermatocyte-specific isoforms of Syncrip (Syp) are required for expression of CycB protein in mature spermatocytes and normal entry into the meiotic divisions. Lut and Syp interact with Fest independent of RNA. Thus, a set of spermatocyte-specific regulators choreograph the timing of expression of CycB protein during male meiotic prophase.

The work in this paper is thorough and carefully done.The results are new believable, and interesting.But three major items need to be addressed before the manuscript would be suitable for publication, and some clarifications to the writing or figures are needed to make it easier to comprehend.

Comments for the author
Major points to address 1.
There is only a small difference in timing (~8%) between meiosis exit in mutants relative to controls or each other.Can the authors provide information to assure the reader that the synchronization has the precision to detect such a small timing difference?Are the strains sufficiently identical to remove concern about genetic background effects on the timing?Related, the inference of timing of CycB expression in Figure 1 G-I would be more convincing if the figure showed the cell-types where expression is detected (i.e. by morphology) or if the location/spacing of CycB expression was quantified.2.
The co-immunoprecipitations are done on extracts.It seems possible that the proteins interact in vivo within different complexes that are incorrectly amalgamated in Fig. 2c, the title of the paper, and elsewhere.Perhaps some Rbp4 interacts with the 3' UTR of CycB mRNA, and other Rbp4 is in a complex with Lut.Can the authors provide evidence that the proteins always co-localize in the cell with each other and with CycB mRNA? 3.
The authors need to establish that Cyclin B expression is sufficient to drive the meiotic program -whether through citations or experiments.
Minor points to address 1.
Aspects of the data/writing that could be better described/explained a.
The band in the negative control in Figure 2B seems problematic for their interpretation.b.
Figure 3 A-D -there is an apparent expansion of CycB staining in spermatogonia at the later time points c.
Line 189-190 -more explanation of what this implies, full name references d.
Fig 4E-F "at least one cyst" is a reasonable way to quantify this data but please, at least qualitatively, describe how similarly the cysts behaved -they would be expected to be synchronized.e.
More explanation/justification is needed tfor why expression of Syp in heads in the syp mutant is described as surprising.

2.
Suggestions for tightening up the writing: a.
Refer to the timings in a consistent way."PHS" is a great way to indicate that you are using and referring to the bam-HS system, but you only use this on occasion.I would be more consistent i.e. in Figure 1D,E, Figure 3, 6 etc. b.
Line 141 -The topic sentence of this paragraph could better distinguish the experiments in this paragraph from those directly above.i.e. you are now using the hs-Bam system to narrow in on the timing of the interactions c.
Line 142 -Suggest mentioning/contrasting the CycB RNA expression pattern here.d.
The genes in the proposed seemed to arise out of the blue -please describe better why they were tested/studied e.
Paragraph 189-201 vs 202-207 -Slightly different language is used to describe the same thing.If you mean that they are the same, you can write this more concisely.If you mean that they are different, please clarify how.f.
The study of the genes in this paper seems to come out of left field.I would appreciate more explanation of how you arrived at these genes.

3.
Figure 2F : "10% input" across the top left vs bottom right -I'm assuming the top left is a typo? 4.
The lut CRISPR mutant has not been demonstrated to be a null.You can refer to it as a "likely null" or give it an allele number.The sample shown in Figure 9K looks poorly fixed, which can lead to higher background staining.I'm concerned about the quantification of expression levels from samples such as this.

7.
In Fig8A it looks like syp is upregulated in rbp4-/-.Could this indicate competitive binding with Fest?

Advance summary and potential significance to field
This manuscript describes the characterization of several RNA binding proteins that physically interact to control the translation of cyclin B within the developing germline of Drosophila males.Cyclin B is translationally repressed in immature Drosophila spermatocytes and previous work from the lab showed that Rbp4 and its partner Fest can bind to elements within the 3'UTR of cyclin B mRNA.The work presented here extends these earlier findings and shows that another spermatocyte-specific protein, Lut, also participates in the repression of cyclin B during an 8 hour window of time, just before spermatocytes are mature.In addition, spermatocyte-specific isoforms of Syncrip regulate the normal expression of cyclin B and entry into meiosis.Lastly, the authors provide evidence that Lut and Syp interact with Fest in an RNA-independent manner.This study will be of general interest to the readership of Development.Understanding how different RNA binding proteins coordinate to regulate the stage and tissue-specific translation of mRNAs is of great interest to the field.This study provides further insights into the regulated translation of cyclin B mRNA as germ cells transition into meiosis.However, several points should be addressed prior to publication.

Comments for the author
The study has many moving parts with 10 figures and 2 supplemental figures.The transition from the Lut data (Figs 1-4) to the Syp data (Figs 5-9) should be improved.Simplifying and consolidating the figures may result in a clearer and more cohesive presentation of the overall story.
One of the primary weaknesses of the paper is that the authors do not provide strong evidence that these factors directly bind to the 3'UTR of cyclin B in vivo.Beyond the biotinylated pulldowns shown for Syp, the authors should perform RNA-immunoprecipitation for Rbp4, Fest, Lut, and Syp in the presence of crosslinking agents to show that these factors indeed interact with endogenous cyclin B mRNA.
Additional structure/function analysis would bolster the claim that Rbp4, Fest, Lut and Syp form a complex in which Fest acts as a scaffold.Specific mutations in Fest that disrupt its interactions with other complex members should be made and tested for functional significance.Does the HA-lut transgene first presented in Figure 1 rescue the lut mutant phenotype?If so, these data should be presented.Are lut males completely sterile?A brief description of the overall fertility of these mutants should be included in the paragraph describing the generation of the null mutant (lines 119-126).
In Figure 2A, treatment with RNAse appears to increase the interaction between Lut and Fest.Do the authors have an explanation for this?
The IP presented in Figure 2B is relatively weak (IP shows lower levels of HA-Fest relative to 10% input).
The authors could consider moving this panel into supplemental data or removing it altogether.The authors state, "In lut mutants, however, round spermatids typically had four nuclei and one large mitochondrial derivative per cell (Fig. 4H), suggesting complete failure of cytokinesis after meiosis I and II.In contrast, the rbp4 mutant showed evidence of much milder cytokinesis defects (Fig. 4I, arrows)."Given that cyclin B protein expression accumulates at earlier times in the rbp4 mutant relative to lut, can the authors provide a further explanation of how the lut phenotype relates to misexpression of Cyclin B protein?The models presented in Figure 10 suggest that Lut represses cyclin B translation independent of rbp4.Do the authors mean to suggest that the physical interactions between rbp4 and lut (bridged by Fest) described in Fig 2 are not functionally significant?Further discussion along these lines is warranted.

Author response to reviewers' comments
We thank the reviewers for their useful suggestions.In response we have made major updates to the paper, adding new results.These include easyCLIP data showing direct binding of Syp to the cycB RNA in vivo; smFISH against cycB RNA to match all of the timepoints and genotypes used for anti-CycB immunostaining in Figure 3; smFISH against cycB for Figure 2 to match the 72h and 104h PHS anti-CycB immunostaining; and a repeat of smFISH (cycB and loopin-1) for Figure 9.In addition, we made substantial revisions to the text to address reviewers' concerns.Major changes to the text of the manuscript include detailed explanations of how we came to be interested in lut and syp; additional details in the Materials and Methods about both the hs-Bam time-course and the ability of the HA-Lut transgene to rescue the lut mutant; and removal of speculative models from figures and text showing Lut and Fest recruited to the cycB RNA via Rbp4/Syp.In addition, in the absence of data showing co-sedimentation through a velocity gradient or comigration through a sizing column, we have removed the term "complex" from the title and throughout the manuscript.Although Lut, Fest, Rbp4, and Syp proteins clearly interact, we cannot prove that they form a monolithic unified complex at all stages of their action.
See below for point-by-point responses (blue text) to reviewers' comments and questions (black text).In the manuscript itself, updates / changes to the text requested by the reviewers are in red; sections re-written for clarity or moved to methods are in blue.

-Rev1
The work in this paper is thorough and carefully done.The results are new, believable, and interesting.But three major items need to be addressed before the manuscript would be suitable for publication, and some clarifications to the writing or figures are needed to make it easier to comprehend.

Reviewer 1 Comments for the Author:
Major points to address 1.There is only a small difference in timing (~8%) between meiosis exit in mutants relative to controls or each other.Can the authors provide information to assure the reader that the synchronization has the precision to detect such a small timing difference?Are the strains sufficiently identical to remove concern about genetic background effects on the timing?Related, the inference of timing of CycB expression in Figure 1 G-I would be more convincing if the figure showed the cell-types where expression is detected (i.e. by morphology) or if the location/spacing of CycB expression was quantified.
The lut hs-Bam time-course flies differ from the control (vs.wt hs-Bam time-course flies) on chr 2 (location of lut), and the rbp4 hs-Bam time-course flies differ from the control (vs.wt hs-Bam time-course flies) on chr. 3 (the location of rbp4).This means that the 8% difference in the timing of entry into the meiotic divisions in both lut and rbp4 (vs.wt) is likely not due to a shared difference in genetic background between those two strains and the control strain in the hs-Bam time-course.
For Figure 1G-I, cell morphology is sacrificed to the IF conditions, especially the fixation step in methanol, so we cannot provide insight into germ cell stage in those particular samples.Instead we addressed the point by immunostaining for CycB protein in the heat-shock time-course to get granular detail on the timing of expression (Figure 3) -much more than what we would achieve by measuring the distance from CycB-expressing cysts to the tip of the testis or the edge of the spermatogonial zone.
Measuring the distance between CycB-positive cysts and the apical tip cannot be used because if a strain has elongating spermatids present, they push the early germ cell stages up against the tip, while in strains without robust spermatid elongation, the early germ cell stages spread away from the tip much more than in wild type.
2. The co-immunoprecipitations are done on extracts.It seems possible that the proteins interact in vivo within different complexes that are incorrectly amalgamated in Fig. 2c, the title of the paper, and elsewhere.Perhaps some Rbp4 interacts with the 3' UTR of CycB mRNA, and other Rbp4 is in a complex with Lut.Can the authors provide evidence that the proteins always colocalize in the cell with each other and with CycB mRNA?
We have removed all speculation that Lut and Fest are recruited to the cycB 3'UTR via Rbp4.This includes deleting the model images from Figures 2 and 8.In addition, since we cannot provide data demonstrating that the proteins all co-migrate in a uniform peak in velocity sedimentation or sizing columns, we have removed "complex" from the title and throughout the manuscript.
3. The authors need to establish that Cyclin B expression is sufficient to drive the meiotic program -whether through citations or experiments.
In the original version, we described premature CycB expression as "permissive but not sufficient" for driving early cell cycle entry.We have re-written the discussion to make this clearer.We do not claim that expression of CycB is sufficient to drive entry into the meiotic divisions.Indeed our data from rbp4 mutants indicate that CycB expression is clearly not sufficient in young and midstage spermatocytes.CycB expression is not sufficient in the classical literature on control of the mitotic cell cycle, where activation of Cdk1 by removal of inhibitory phosphorylation is also required.
Minor points to address 1. Aspects of the data/writing that could be better described/explained a.The band in the negative control in Figure 2B seems problematic for their interpretation.
We have removed this panel (a reciprocal IP) from Figure 2, as reviewer #2 didn't think it added much and wanted it removed (or moved to SOM).
b. Figure 3 A-D -there is an apparent expansion of CycB staining in spermatogonia at the later time points We have added an explanation that this is the result of an expansion of the spermatogonial population itself due to the bam genetic background, once Bam protein is no longer being provided by heat shock (described in Materials and Methods).F "at least one cyst" is a reasonable way to quantify this data but please, at least qualitatively, describe how similarly the cysts behaved -they would be expected to be synchronized.
Under our heat-shock time-course conditions, the differentiating spermatocyte cysts are metasynchronous, not totally synchronous; this is especially noticeable for short-duration events like the meiotic divisions.At 48h PHS, all spermatocytes are polar; at 72h PHS all spermatocytes have moved beyond the polar stage.However, at 108h PHS, some cysts might be in meiosis I, others may be in meiosis II, and yet others may contain early round spermatids.This intra-testis asynchrony was observed in all three genotypes scored for Figure 4.For this reason we chose to score the leading edge of the distribution of cysts entering and exiting meiotic division for Figure 4. We have added this info to the Materials & Methods.
e.More explanation/justification is needed for why expression of Syp in heads in the syp mutant is described as surprising.
We cut that sentence as the signal from one western blot was not sufficiently quantitative to make that conclusion.
2. Suggestions for tightening up the writing: a. Refer to the timings in a consistent way."PHS" is a great way to indicate that you are using and referring to the bam-HS system, but you only use this on occasion.I would be more consistent i.e. in Figure 1D,E, Figure 3, 6, etc.
Thank you.We now utilize "PHS" throughout the manuscript.
b. Line 141 -The topic sentence of this paragraph could better distinguish the experiments in this paragraph from those directly above.i.e. you are now using the hs-Bam system to narrow in on the timing of the interactions We have rewritten the topic sentence to make the purpose of the experiment clearer.
c. Line 142 -Suggest mentioning/contrasting the CycB RNA expression pattern here.
We have added smFISH images showing cycB RNA for 72h and 104h PHS to Figure 2, underneath the corresponding anti-CycB panels.
d.The genes in the proposed seemed to arise out of the blue -please describe better why they were tested/studied f.The study of the genes in this paper seems to come out of left field.I would appreciate more explanation of how you arrived at these genes.
We have added content at the start of both the Lut and Syp sections to explain how these genes first caught our attention.Lut was identified as an interesting candidate from IP-MS analysis of proteins that interacted with Rbp4 and Fest.Syp was identified in an RNAi screen of predicted RNA-binding proteins whose transcripts were upregulated in spermatocytes.
e. Paragraph 189-201 vs 202-207 -Slightly different language is used to describe the same thing.If you mean that they are the same, you can write this more concisely.If you mean that they are different, please clarify how.
The first paragraph describes the expression from the four promoter options; the second paragraph, from the four C-terminal forms.We could collapse this into one paragraph but in doing so we think it risks making the complex locus confusing.
3. Figure 2F : "10% input" across the top left vs bottom right -I'm assuming the top left is a typo?
Thank you.We deleted the errant label.
4. The lut CRISPR mutant has not been demonstrated to be a null.You can refer to it as a "likely null" or give it an allele number.
We have changed "null" to "loss-of-function" in the results.We will refer to it as " lut 1 " when talking about the allele, and continue using "lut" as shorthand for lut 1 /Df, as noted.We have removed a stray arrow from panel A', changed the arrows in D to arrowheads and rotated them so that the reader's eye is not tempted to correlate them with the arrowheads in D', and updated the figure legend so that it says more precisely what arrowheads and arrows are pointing to (nucleoli, and the outside edge of relevant nuclei, respectively).
6.The sample shown in Figure 9K looks poorly fixed, which can lead to higher background staining.I'm concerned about the quantification of expression levels from samples such as this.
We have updated Figure 9 with new cycB smFISH images, plus smFISH against loopin-1 as an internal control.At the time of this experiment, we assumed loopin-1 would be a negative control; Syp easyCLIP data obtained much later indicated that the loopin-1 RNA is itself bound by Syp (Figure S3).That result made our finding that loopin-1 RNA levels were reduced in the syp mutant (Figure 9) less surprising.
7. In Fig8A it looks like syp is upregulated in rbp4-/-.Could this indicate competitive binding with Fest?
For this experiment we used more rbp4 testes (77 pairs) than wt (55 pairs), as rbp4 testes appear somewhat smaller than wt by eye.This may account in part for why there is more eYFP-Syp in the input and in the IP lanes.We therefore cannot conclude that Syp protein is upregulated in rbp4 mutant testes.
***** Reviewer 2 Advance Summary and Potential Significance to Field: However, several points should be addressed prior to publication.
Reviewer 2 Comments for the Author: The study has many moving parts with 10 figures and 2 supplemental figures.The transition from the Lut data (Figs 1-4) to the Syp data (Figs 5-9) should be improved.Simplifying and consolidating the figures may result in a clearer and more cohesive presentation of the overall story.
One of the primary weaknesses of the paper is that the authors do not provide strong evidence that these factors directly bind to the 3'UTR of cyclin B in vivo.Beyond the biotinylated pulldowns shown for Syp, the authors should perform RNA-immunoprecipitation for Rbp4, Fest, Lut, and Syp in the presence of crosslinking agents to show that these factors indeed interact with endogenous cyclin B mRNA.
We have added UV cross-linking immunoprecipitation (CLIP) data showing that Syp protein expressed in spermatocytes directly binds the cycB RNA (and others) in vivo.In a pilot CLIP experiment with each of the four proteins, Syp and Rbp4 showed robust RNA-binding (see Figure S3A for Syp), Lut showed faint RNA-binding, and Fest did not bind RNA.Because there were other RNA-positive bands above Rbp4-eYFP in the membrane for the pilot experiment, we did not expect to get a clean result from an Rbp4-eYFP easyCLIP library.
However, based on the promising preliminary result with Syp, we did easyCLIP for eYFP-Syp followed by sequencing the resulting cDNA libaries and found that Syp bound to both the 5' and 3'UTRs of the cycB RNA (Figure 5Q).Syp bound the 5' and/or 3'UTRs of a number of other RNAs as well.However, the binding appeared selective: compared to a set of spermatocyte-expressed genes with similar expression levels to cycB, cycB bound more Syp than most of the RNAs (see histogram in Figure S3C).
Additional structure/function analysis would bolster the claim that Rbp4, Fest, Lut and Syp form a complex in which Fest acts as a scaffold.Specific mutations in Fest that disrupt its interactions with other complex members should be made and tested for functional significance.
Absent co-sedimentation experiments or co-migration in sizing columns, we cannot prove that Rbp4, Fest, Lut, and Syp are in one complex.The small amounts of material from hand-dissected testis make such biochemical assays extremely difficult.In light of this, we have removed "complex" from the title and throughout the paper, also removing the ball-and-stick diagrams from the figures.
In terms of structure/function studies of Fest, we built transgenic flies carrying an HA-tagged version of Fest that contained only the conserved C-terminal half of the protein (Gly264-Y513), in hopes of testing whether this chunk of Fest was sufficient to bind one or more of Rbp4, Lut, and Syp.The protein appeared to be expressed normally by immunostaining but frustratingly did not show up at all in our IP-WB experiment, suggesting that it may have aggregated with insoluble material post-lysis.
We feel that more analysis (especially point mutations) is beyond the scope of this paper.Does the HA-lut transgene first presented in Figure 1 rescue the lut mutant phenotype?If so, these data should be presented.Are lut males completely sterile?A brief description of the overall fertility of these mutants should be included in the paragraph describing the generation of the null mutant (lines 119-126).
As noted in our response to Reviewer #1, more pairs of rbp4 testes (77 rbp4 vs. 55 wt) were used in this experiment.This may have contributed to the excess of eYFP-Syp protein in 8A.Beyond that, we do not know whether Syp expression is negatively regulated by Rbp4.Because the mutant testes have different ratios of cell types compared to wild-type testes, one cannot draw conclusions about relative levels of expression of a protein in testes from one genotype compared to another from bulk tissue assays like western blots.
We opted to use more fest testes in Figure 8C to get the Rbp4-HA levels to roughly match each other for this experiment, at the expense of having equal levels of Syp (the same goes for Figure 8D and HA-Lut).Note that fest mutant germ cells are quite odd compared to rbp4 and lut, suggesting that Fest likely has a role independent of Rbp4 and Lut that is related to microtubule growth/organization.Does Lut influence the levels and/or interactions of Fest and Syp?
We performed co-immunoprecipitation (co-IP) assays for Fest and Syp in wt vs. lut mutant testis extracts and found that HA-Fest co-IPed with eYFP-Syp in anti-GFP immunoprecipitations at roughly similar levels in the two genetic backgrounds.However, because eYFP-Syp did not immunoprecipitate well in the lut background, the result is not quite publication-quality, and we would prefer to leave it out of the paper.
Data presented in figure 9 suggest that Rbp4, Lut, and Syp are needed for the stability of cyclin B mRNA.Why these levels appear to increase in the lut syp double mutants (Fig 9K ) relative to the single mutants is unclear.Attempting to quantify mRNA levels using mean fluorescence is problematic without internal controls.Additional methods should be used to make this point because the presented images do not appear to match the quantification provided in We have updated Figure 9 with a repeat of the smFISH that includes loopin-1 as an internal control.Levels of loopin-1 were lower in the syp mutant; this result was unexpected at the time but was not surprising after we discovered later by easyCLIP that the loopin-1 RNA was a target of Syp binding (Figure S3).
The models presented in Figure 10 suggest that Lut represses cyclin B translation independent of rbp4.Do the authors mean to suggest that the physical interactions between rbp4 and lut (bridged by Fest) described in Fig 2 are not functionally significant?Further discussion along these lines is warranted.
As noted earlier, we have not been able to prove that Lut and Fest are complexed with Rbp4 (or with Syp, for that matter) on the cycB RNA, and we have re-worded the text conservatively in accordance with that.
If we had shown that Lut and Fest associate with the cycB RNA, we could speculate that Rbp4 plays two roles: a Lut-independent early role (54-94h PHS) and possibly a Lut-aiding late role (94h-102h PHS), in which Rbp4 contributes to the full recruitment of Lut (via Fest) to the cycB RNA.Our previous ball-and-stick model (now removed from the manuscript) had Fest associating with the cycB 3'UTR via both Rbp4 and Syp; in our experience, the Rbp4-Fest interaction is stronger than Syp-Fest, so Fest (and Lut) might in theory not be recruited as efficiently to the cycB 3'UTR in an rbp4 mutant as in wt.
Again, we have not been able to show this, so this speculation is for reviewers' eyes only.

Second decision letter
MS ID#: DEVELOP/2023/201709 t follow what the arrows are supposed to show in Fig 7A'-D' 6.

Further
biochemical experiments are needed to support the model presented in Fig 2C.It remains unclear whether all three proteins co-assemble into a single complex or whether sub-complexes form.The quality of the image presented in Fig 2E does not match other images throughout the paper.The authors should provide another image that convincingly shows Cyclin B expression at 104 hrs after hs-bam induction.In Fig 2F-the "10% input" at the top of the panel may be misplaced.For Figure 3-The authors should provide data describing the levels of cyclin B mRNA in the controls and mutants at all the timepoints.Some of the control data are presented in figure 6. Knowing the RNA levels in the mutants in Fig 3 would strengthen the interpretation of these data.
In panels Fig8B and Fig 8D,  the levels of Syp appear to increase in the absence of rbp4 and fest respectively.Do the authors think each of these proteins influence the levels of one another?Does Lut influence the levels and/or interactions of Fest and Syp? Data presented in figure9suggest that Rbp4, Lut, and Syp are needed for the stability of cyclin B mRNA.Why these levels appear to increase in the lut syp double mutants (Fig 9K)relative to the single mutants is unclear.Attempting to quantify mRNA levels using mean fluorescence is problematic without internal controls.Additional methods should be used to make this point because the presented images do not appear to match the quantification provided in Fig 9M.Statistics are needed in Fig 9M as well.
c. Line 189-190 -more explanation of what this implies, full name, references We added the full name and references for SYNCRIP (lines 207-8 of the revised manuscript) d.Fig 4E- 5. I don't follow what the arrows are supposed to show in Fig 7A'-D' Fig 9M.Statistics are needed in Fig 9M as well.