Physical Interaction between Cyclin-Dependent Kinase 5 (CDK5) and Clock Factors Affects the Circadian Rhythmicity in Peripheral Oscillators

Round 1

Reviewer 1 Report

This is an interesting paper presenting a new player, a Cyclin-dependent kinase 5, that may play its role in the modulation of the biological clock period in peripheral cells.

There is one crucial and are some minor comments and questions below.

Main point: Supplementary materials are needed for a better understanding of the results. Now supplemental file just replicates main figures and provides legends, supplementary figures themselves are not provided.

Minor points:

1a Figure 1. When NIH3T3 cells are transfected with any other Sh-RNAs one can observe that period lengthening is accompanied by a lower amplitude of bioluminescence, suggesting an uncoupling of otherwise synchronized molecular elements that generate bioluminescence. Since CDK5 plays an important role in the regulation of synaptic plasticity, the lack of its function may alter inter-cellular synchronicity.

1b shRNA-b is interesting since it is the only case when lower bioluminescence amplitude is accompanied by a period equal to 24 hours. Sh-RNA-C and shRNA-D not only lengthen the period but as associated with the quick fading of the circadian signal of bioluminescence. Did the authors check the statistical correlation between the amplitude of oscillation and tau deviation from 24 hours?

2. It could be interesting to find out whether this phenomenon relates to the rhythmicity of the cell, or rather inter-cellular communication/synchronization.
3. How do authors explain the principal difference in the circadian period, they have found between Cdk5 loss-of0unction in SCN (period shortening) and fibroblasts (period lengthening)? Could it be a matter of alternative splicing of the key element of substrate or its target in neurons vs fibroblasts, or selective interference with “third player substance” that might be tissue-specific and only present in SCN or fibroblasts, or its presence if time-windowed, that may explain the difference in the effects upon circadian period?

Author Response

Main point: Supplementary materials are needed for a better understanding of the results. Now supplemental file just replicates main figures and provides legends, supplementary figures themselves are not provided.

There must have been an unrecognized problem while uploading the manuscript. Therefore, we now provide the previously missing supplemental figures.

Minor points:

1a Figure 1. When NIH3T3 cells are transfected with any other sh-RNAs, one can observe that period lengthening is accompanied by a lower amplitude of bioluminescence, suggesting an uncoupling of otherwise synchronized molecular elements that generate bioluminescence. Since CDK5 plays an important role in the regulation of synaptic plasticity, the lack of its function may alter inter-cellular synchronicity.

This is an interesting question. Albeit, NI H3T3 cells do not rely on intercellular coupling. Also, one must keep in mind that we measure the availability of luciferase (which is driven by the rhythmic promoter) and the availability of intracellular ATP with the luciferase assay. At the same time, the over-expression of shRNA may interfere with the proper functioning of the miRNA machinery in the cell. Consequently, without co-transfection of a red-shifted luciferase construct with a promoter unaffected by the knock-down, it is difficult to make a statement about the amplitude (unfortunately, we do not have an adequate machine to measure this). We realized that the amplitude was slightly reduced with an indirect (roscovitine) or direct (Cdk5 ko cell lines) method. However, amplitude reduction is not a reliable sign of a dysfunctional clock at the cellular level, which was the central question for us. In fact, as also discussed by the reviewer, the lower amplitude can be associated with some systemic alteration like decoupled clock between different cells, which means the single-cell still has a functional clock but with a different phase. Therefore we focused on the period length, a direct parameter connected to the circadian clock dysfunction at the cellular level.

1b shRNA-b is interesting since it is the only case when lower bioluminescence amplitude is accompanied by a period equal to 24 hours. Sh-RNA-C and shRNA-D not only lengthen the period but as associated with the quick fading of the circadian signal of bioluminescence. Did the authors check the statistical correlation between the amplitude of oscillation and tau deviation from 24 hours?

We previously characterized the efficiency of the knock-down constructs to reduce the expression of Cdk5 and, therefore, the amount of protein in tissue culture cells (Brenna et al., 2019). In particular, shRNA-b did not reduce the amount of CDK5 protein compared to untransfected and scramble-construct transfected cells. Hence, the phenotype could be a transitory phenotype from the scramble-construct and shRNA-a to shRNA-c and shRNA-d.

2. It could be interesting to find out whether this phenomenon relates to the rhythmicity of the cell or rather inter-cellular communication/synchronization.

Due to the co-transfection of the luciferase reporter and the shRNA-expressing construct, we measure luciferase activity only in cells transfected with both constructs. Consequently, the observed phenotype is intracellular. Therefore, we think that transient co-transfection experiments cannot address the question of intercellular coupling.

3. How do authors explain the principal difference in the circadian period, they have found between Cdk5 loss-of-function in SCN (period shortening) and fibroblasts (period lengthening)? Could it be a matter of alternative splicing of the key element of substrate or its target in neurons vs fibroblasts, or selective interference with “third player substance” that might be tissue-specific and only present in SCN or fibroblasts, or its presence if time-windowed, that may explain the difference in the effects upon circadian period?

After weighing all the evidence provided in this paper, we think the principal difference between the SCN and the periphery is the availability of CLOCK. Supposedly, in Cdk5-deficient cells, due to stabilizing and importing large amounts of CLOCK into the nucleus, ARNTL1 (BMAL1) becomes the rate-limiting factor. Hence, the excess of CLOCK squelches other regulatory elements, also explaining the amplitude reduction. A similar phenomenon was described for the ClockD19 mutant (Antoch et al., 1997; Gekakis et al., 1998; Oishi et al., 2000), which exerts a similarly long period phenotype.

Author Response File: Author Response.pdf

Reviewer 2 Report

Authors investigated the role of CDK5 in regulating peripheral clock in tissue culture cells. They claim that lack of CDK5 prolongs the period length of the circadian oscillator. This study is well done. However, there are issues to be fixed before accepting the manuscript.

1. Supplementary figures are missing in the current version of the manuscript.
2. To understand the difference between amplitude, phase, etc., it is recommended that authors analyze differential rhythmicity for figure 5 (e.g., CircaCompare (Parsons et al., 2020)). I would also recommend that authors provide a detailed protocol for circadian analysis.
3. Please check nomenclature of genes and protein names across the manuscript.

Author Response

Authors investigated the role of CDK5 in regulating peripheral clock in tissue culture cells. They claim that lack of CDK5 prolongs the period length of the circadian oscillator. This study is well done. However, there are issues to be fixed before accepting the manuscript.

1. Supplementary figures are missing in the current version of the manuscript.

There must have been an unrecognized problem while uploading the manuscript. Therefore, we now provide the previously missing supplemental figures.

2. To understand the difference between amplitude, phase, etc., it is recommended that authors analyze differential rhythmicity for figure 5 (e.g., CircaCompare (Parsons et al., 2020)). I would also recommend that authors provide a detailed protocol for circadian analysis.

We added a detailed description of how we performed the bioluminescence analysis. First, the period length was determined with the fast Fourier transform option of the LumiCycle analysis software. From the LumiCycle analysis, it is clear that the Cdk5-deficient cells rapidly shift from the controls. Second, we analyzed the mRNA expression of clock genes in fig.5A, using CircaCompare. The new results are in figure S3B. Third, we created an additional table, to sum up the difference between the two genotypes in terms of amplitude and phase shift for all the clock genes measured in our manuscript. We replaced our previous ChIP analysis using PRISM with the one made using CircaCompare. Finally, we cited the paper in our manuscript.

3. Please check the nomenclature of genes and protein names across the manuscript.

We checked and adapted the nomenclature for the mouse.

Author Response File: Author Response.pdf

Round 2

Reviewer 2 Report

overall manuscript looks good!