The role of nuclear organization in trans-splicing based expression of heat shock protein 90 in Giardia lamblia

Hsp90 gene of G. lamblia has a split nature comprising two ORFs separated by 777 kb on chromosome 5. The ORFs of the split gene on chromosome 5 undergo transcription to generate independent pre-mRNAs that join by a unique trans-splicing reaction that remains partially understood. The canonical cis-acting nucleotide elements such as 5’SS-GU, 3’SS-AG, polypyrimidine tract and branch point adenine are present in the independent pre-mRNAs and therefore trans-splicing of Hsp90 must be assisted by spliceosomes in vivo. Using an approach of RNA-protein pull down, we show that an RNA helicase selectively interacts with HspN pre-mRNA. Our experiments involving high resolution chromosome conformation capture technology as well as DNA FISH show that the trans-spliced genes of Giardia are in three-dimensional spatial proximity in the nucleus. Altogether our study provides a glimpse into the in vivo mechanisms involving protein factors as well as chromatin structure to facilitate the unique inter-molecular post-transcriptional stitching of split genes in G. lamblia.


Response:
Our RNA-pull down experiment is only a qualitative experiment to identify proteins interacting with the HspN pre-mRNA. We have not attempted any comparative quantitation at this stage. Also, we would like to point out that lanes 5 and 6 represent negative controls. Lane 5 represents proteins binding to Hsp90FL RNA containing only the exonic region where we do not find the protein of interest, namely RNA helicase, whereas lane 6 represents the non-specific binding of proteins to beads alone which are devoid of any RNA. This is mentioned in the results section Page-12, line-300-302 (in the revised article with changes highlighted).
Concern: They do not provide evidence for binding of the helicase in vivo. There is also no evidence that the helicase plays any role in splicing, or even that it is in the nucleus. RNA Helicases are not generally sequence-specific and are often quite abundant. This is a very weak part of the paper and without additional functional evidence it (and all relevant Figures and tables) should be removed. As far as I can see all the authors have done is to identify a protein from a random part of a polyacrylamide gel.
Response: With all due respect, we do not think that RNA helicase identified in this study is from 'a random part of a polyacrylamide gel' as indicated by the reviewer for the following reasons-1. The band excised for MS analysis appears reproducibly in three independent experiments of HspN RNA pull down. 2. The same band does not appear in a) beads alone control (this would be the random part of a polyacrylamide gel as indicated by the reviewer) or b) HspC RNA pull down lane.
In connection with the reviewer's comment "There is also no evidence that the helicase plays any role in splicing, or even that it is in the nucleus", we would like to clarify that helicases have often been implicated in splicing mechanisms [1][2][3][4]. In fact, the spliceosomal complex involved in facilitating RNA splicing is also known to contain RNA helicases as an integral part of the complex.
Since the time of this manuscript's submission, a study from Balan B. et al., [5] reported its (RNA helicase) identification in the Giardia proteome and suggested its possible role in posttranscriptional regulation. Identification of this protein in our pull-down assay provides direct experimental evidence for its involvement and strengthens its possible role in post-transcriptional regulation. This manuscript has not addressed the in vivo binding and its role in the splicing reaction, which will be a subject matter of our subsequent manuscript.
Concern: (Figure 2) I think I vaguely understand what the authors are showing here but the Figure is really confusing and so is the Legend. What is an "EcoRI locus"? What are the numbers above and below the map? What are the products obtained with "randomly ligated" DNA? What are the "other forward primers"? What are "different loci chosen for this study"? A much clearer diagram is required showing locations and directions of all primers used. Also, to verify the results the PCR products must be sequenced.

1) What is an EcoRI locus?
An EcoRI locus/loci refers to the EcoRI recognition site(s) (indicated by vertical broken lines) in the vicinity and between the ORFs being investigated to determine their physical proximity in Giardia nucleus. The distance of the different recognition sites from the start codon of HspN has also been elaborated in Suppl. Figures S8, S11 and S14.

2) What are the numbers above and below the map?
The numbers below the schematic (-100, 0, 100, 200 series) in Figure 2A, 4A and 4D represents the precise scale of the chromosomes which indicate the position of the ORFs under investigation.
In chromosome 5, the scale also indicates the position of HspN ORF relative to HspC, Dhcβ C-2 and Dhcβ C-3.
The labels G minus 4, G minus 3alt, G minus 1, G internal, G plus 1, G plus 2, G plus 3, G plus 4 in Figure 4D represents the different EcoRI recognition sites chosen to investigate physical proximity between HspN on chromosome 5 and Dhcɣ C-1 on chromosome 3.
Explanation for random ligation achieved in the case of uncross-linked control gDNA has been incorporated as Supplementary Figure S6. Random ligation is evident from the random trend of the intensities of the hybrids in the middle panel of Figure 2B, 4B and 4E.

4) What are the "other forward primers"?
"Other forward primers" refers to forward primers except for 2F (constant primer) which is constant as shown in all the lanes of Figure 2B, Figure 4B and Figure 4E.

5) What are "different loci chosen for this study"?
"The different loci chosen for this study" refers to the EcoRI recognition sites represented as broken vertical lines in the schematics in Figure 2A, Figure 4A and 4D.
Detailed explanations for the above points have also been incorporated in the Supplementary  Figures S6, S7, S8, S9, S11, S12, S14 and S15, indicated explicitly in the revised manuscript text. Figures 2 and 4 have also been simplified to convey the results more clearly.
As suggested by the reviewer a much clearer diagram showing the location and directions of all primers used in 3C has been illustrated in the revised supplementary information (Suppl. Figures  S9, S12 and S15). The sequencing result of the enriched hybrid in each 3C experiment have now been included in the main Figures 2C and 4C and 4F.

Concern:
In panel C of Figure 2, with only 3 measurements you cannot legitimately use the standard error of the mean or even the standard deviation. Please show the individual measurements instead.

Response:
As suggested by the reviewer we have now modified the previous representation of crosslinking frequency to include the graphs for individual measurements instead. The relative crosslinking frequency showing interaction between 1) HspN with HspC, 2) HspN with Dhcβ C-2 and 3) HspN with Dhcɣ C-1 have now been shown individually in supplementary information (Suppl. Figures S10, S13 and S16) along with the agarose gel images after semi-quantitative PCR. Figure 3): This shows that probes that hybridize to HspN and HspC ORFs colocalize. That is interesting, but I am not sure about the negative control gene. Is this gene on the same chromosome? If not, it is not valid as a control. A better control would be to try two different genes that are located in between the HspN and HspC ORFs on the same chromosome. Another control that is essential is to do the same experiment after inhibiting transcription, to see whether the mutual association depends on RNA.

Response:
The query raised by the reviewer regarding negative control gene being on the same chromosome has been addressed in the revised version. We have now performed DNA-FISH for two different genes that are located in between the HspN and HspC ORFs on the same chromosome to serve as a negative control. We observe that the two different genes (Enolase gene-GL50803_11118 and Protein 21.1-GL50803_24412) between HspN and HspC on the same chromosome do not colocalize and therefore now serves as true negative control for our DNA-FISH experiment. The different panels from this experiment have now been incorporated in Figure  3.
The suggestion made by the reviewer is relevant and we would pursue such regulatory aspects of the long-range interaction in our subsequent manuscript that will interrogate the spatio-temporal localization of different pre-mRNAs from split genes. Figure 4 is truly incomprehensible. I gave up trying to understand it. If the conclusion is that the two separate trans-spliced loci are all together in one place, this actually seems strange because I would not expect it. (This is not the equivalent of spliced leader trans-splicing, and most trypanosomes genes are also not located near the spliced leader array.) In any case, the results need to be be confirmed by sequencing and by in situ hybridisation, again with additional appropriate controls.

Concern:
Any discussion is premature because this needs more experimental work in order to support the conclusions.
Response: To make Figure 4 more understandable, clarification about the different loci used for the study has been incorporated in the text of revised manuscript. The map showing the linear genomic distance between HspN and Dhcβ C-2 on the same chromosome has been shown and the direction of the primers used for the study has been clearly indicated in supplementary information (Suppl. Figure S11 and S12). The map showing the relative distances between the chosen EcoRI loci on chromosome 3 along with the direction of flanking primers to examine physical proximity between HspN on chromosome 5 and Dhcɣ C-1 on chromosome 3, have been elaborated in supplementary information (Suppl. Figure S14 and S15).
Though agreeing with the fact that this trans-splicing process is very different from SL-transsplicing, we stand by our findings from 3C experiments that suggest nuclear proximity of transspliced genes. It actually makes perfect sense that the genes coding for the trans-spliced exons be in proximity to facilitate juxta positioning of corresponding pre-mRNAs to expedite the transsplicing reaction. However, to address the reviewer's queries, we have further performed hybrid sequencing and in situ hybridization with appropriate controls. Our experiments clearly bear out that the trans-spliced genes are in proximity within the nucleus.
The sequencing result of the enriched hybrids in each 3C experiment to investigate the physical proximity of HspN with HspC, Dhcβ C-2 and Dhcɣ C-1 has been included in the manuscript Figures  2 and 4. Additionally, the DNA FISH result validating the interaction between trans-spliced genes-HspN (on chromosome 5) with the dynein genes, Dhcβ C-2 (chromosome 5), Dhcβ C-3 (chromosome 5) and DhcƔ C-1(chromosome 3) along with the appropriate negative control has now been incorporated in the revised version of the manuscript as Figure 5.

Reviewer #2:
Concern: For the most part, conclusions are supported, but additional support for the DEAD/DEXH-box RNA helicase is needed.

Response:
Our results demonstrating binding of RBPs to HspN pre-mRNA through RNA-protein pull down followed by MS based identification are convincing. We have also shown that RNA helicase selectively interacts with HspN pre-mRNA and not seen in the controls. Understanding detailed role and mechanism by which the putative helicase may facilitate Hsp90 trans-splicing will be the subject matter of our future studies.
Concern: As stated before, the limitations are not clearly described on this issue alone.

Response:
As suggested by the reviewer, limitations concerning the results pertaining to RNA helicase have now been incorporated in the revised version of the manuscript (Page-13, Lines 337-339-in the revised article with changes highlighted). These include-1) the possibility that the RNA helicase may be of general nature, also interacting with other RNAs in the nucleus. 2) further biochemical experiments are required to understand the mechanistic aspects of this protein.

Concern:
The authors do discuss the importance and relevance of their study, including public health issues.
Response: As per the reviewer's suggestion, the importance and relevance of this study, including public health issues -"Our attempt to understand the in vivo mechanism of Hsp90 trans-splicing could indeed be useful to further expand our understanding of the molecular mechanisms of this elegant process. In addition, the results from our study could be leveraged to design better anti-giardial drugs to combat Giardiasis, which remains to be a public health concern in developing and under-developed countries." is now discussed in the study, on page-22, line-588-592 (in the revised article with changes highlighted).