PSPC1-interchanged interactions with PTK6 and β-catenin synergize oncogenic subcellular translocations and tumor progression

Hepatocellular carcinoma (HCC) is one of the most lethal cancers worldwide due to metastasis. Paraspeckle component 1 (PSPC1) upregulation has been identified as an HCC pro-metastatic activator associated with poor patient prognosis, but with a lack of targeting strategy. Here, we report that PSPC1, a nuclear substrate of PTK6, sequesters PTK6 in the nucleus and loses its metastasis driving capability. Conversely, PSPC1 upregulation or PSPC1-Y523F mutation promotes epithelial-mesenchymal transition, stemness, and metastasis via cytoplasmic translocation of active PTK6 and nuclear translocation of β-catenin, which interacts with PSPC1 to augment Wnt3a autocrine signaling. The aberrant nucleocytoplasmic shuttling of active PTK6/β-catenin is reversed by expressing the PSPC1 C-terminal interacting domain (PSPC1-CT131), thereby suppressing PSPC1/PTK6/β-catenin-activated metastasis to prolong the survival of HCC orthotopic mice. Thus, PSPC1 is the contextual determinant of the oncogenic switch of PTK6/β-catenin subcellular localizations, and PSPC1-CT131 functions as a dual inhibitor of PSPC1 and PTK6 with potential for improving cancer therapy.

: You cannot definitively identify proteins on a Coomassie stained gel without immunoblotting. This panel does not demonstrate that PTK6 phosphorylates Y523 as indicated in the text. It shows that PSPC1 Y523 is required for PSPC1-PTK6 association. The impact of the Y523F mutation on PTK6 mediated PSPC1 phosphorylation should be shown. Supplemental 6c shows that PSPC1 PY523 associates with PTK6, not that PTK6 phosphorylates PSPC1 at this residue. Fig. 2b. Are these really spheroids and not clusters? The fluorescent cells do not appear to be part of the spheroids as indicated in the legend. It would be interesting to include staining for PSPC1, PTK6, and P-PTK6 in Fig. 2b.
In Fig. 6f, CT131 increases activation of nuclear PTK6 but not cytoplasmic PTK6 and there does not appear to be a shift in total PTK6. Immunofluorescence staining for active P-PTK6 and total PTK6 should be shown.
There are several technical issues. Problems with the language are present throughout: a few examples, "A prometastatic activator upregulated-PSPC1 in cancers might emerge as the lethal metastasis," "Indigenous knockout," "we performed divergent PSPC1/PTK6 constructs," etc. Typos are also present throughout such as r-catenin (γ-catenin). Inclusion of size markers on the westerns is recommended. Quantitation of protein expression changes highlighted in the westerns would be helpful. Lang et al. reported interesting aspects of PSPC1 functions. The authors first identified PTK6, which has tumor suppressive functions, as one of the major PSPC1-interacting proteins. They examined the importance of this interaction in cancer by using cell-based and mouse cancer models as well as cancer clinical data. Then they have shown regulatory roles of PSPC1 in the PTK6/beta-catenin axis. However, the drawback of this manuscript is to contain numerous grammatical errors and typos. In addition, one of the major conclusions is not fully supported by the data shown in the manuscript as described below.
Major points 1. The authors should consider consulting a language editing service. There are numerous grammatical errors, typos, and uncommon usage of terms in the manuscript. Thus, there are many ambiguous sentences, which often disturb reading. Even in the abstract, multiple sentences are not clear, and thus it would be possible not to convey the authors' thoughts to readers. 2. The authors claim that PSPC1-CT131 is a dual inhibitor targeting oncogenic PSPC1 and PTK6, and the interaction between the CT-131 and endogenous PSPC1 is shown in the model of Figure 8. But, no biochemical data for this interaction is shown in the manuscript. Previous studies have shown that PSPC1 dimerize and oligomerize with DBHS family proteins via the NOPS domain and the coiled-coil domain, respectively, but C-terminal domain is not reported for contributing the interaction. The PSPC1-CT131 alone actually binds PSPC1 (and Y523F mutant)? 3. Related to the above point, how does PSPC1-CT131 inhibit endogenous PSPC1 functions? In Figure 6h, PSPC1 expression was reduced by expression of EGFP-PSPC1-CT131. What is an interpretation for this result? Is this possible to be related to the mechanism of PSPC1 inhibition by the CT-131?
Minor points 1. Fig. 1c: The labels of immunoblotting should be corrected to HA or FLAG as used in Figure S1b and S1c. 2. In Figure S2a, in vitro kinase assay using PTK6 and PSPC1 WT is performed. The PSPC1-Y523F mutant is actually not phosphorylated in this assay? 3. In middle cartoon of Figure 8, cytoplasmic PTK6 should be phosphorylated according to the data shown in Figure 3c? 4. Methods: In "Antibodies" section, information of cell lines and plasmids are included. Please fix this point. 5. Methods: Transfection methods are described in "PSPC1 site-directed mutagenesis" section. The transfection methods should be separated or combined with other sections. 6. In Figure 5b & c and Figure 6 c & d, no labels are found on the x-axis. I guess that these are same as in Figure 5d and Figure 6e, respectively. But, I think that this labeling is uncommon. Please fix this point.
Reviewer #3 (Remarks to the Author): General comment: PTK6 kinase plays an opposite role in tumor progression which is dependent on its subcellular localization and transportation between nucleus and cytoplasm. PSPC1 protein, a substrate of PTK6 kinase closed to splicing speckles and recently has been reported by Dr. Jou's group to regulate TGF-β signal pathway for pro-metastatic switches of cancer cells. In this study, Lang et al. showed that PSPC1-Y523F cause oncogenic subcellular translocations of cytoplasmic PTK6 and nuclear β-catenin as well as Wnt3a autocrine loop for promoting liver cancer progression. PSPC1-CT131 which is the proline-rich C-terminal domain of PSPC1 containing NLS (nuclear localization signal), interacts with PTK6 in nucleus to inhibit PSPC1/PTK6 oncogenic functions. The study is interesting and provides mechanistic insights for understanding the role of subcellular localization of PSPC1/PTK6/β-catenin in regulating liver cancer progression. However, some key experiments supporting individual claims may not be sufficient to reveal causal and concrete regulatory mechanisms. Several issues should be clarified.
Comment 1: The SH2 domain interacts with phosphorylated tyrosine residues and SH3 domain binds to proline-rich region (Yeatman T. J., 2004). Does proline-rich region of PSPC1 directly bind to SH3 domain of PTK6 (Supplementary Figure 1b, c)? Comment 2: The author claims that PSPC1 upregulation and PSPC1-Y523F released PTK6 nuclear sequestration; however, the existing evidence of this study for demonstrating the above results are not enough. In particular, the interaction between PSPC1 and PTK6 should be confirmed in different cellular fractions and the immunofluorescent images which indicate PSPC1-Y523F instead of PSPC1-wild type released PTK6 nuclear sequestration should be provided.  Figure 7c, for clearly confirming the localization and expression of EGFP, EGFP images should be shown as an independent panel. Moreover, both N-cadherin and Ecadherin are cell-cell adhesion proteins, their localization should be at the cell-cell contacts.
Comment 7: The evidence for supporting PTK6/PSPC1 to regulate Wnt3a autocrine may not be sufficient. Detection of Wnt3a should in animal experiments and clinical samples is recommended.
Minor points: 1. For Figure 1h, 1k, 2d, 2f and 6d, the labeling of y-axis should be corrected to "Invasive cells per field". 2. In Supplementary Figure 1b, the author established a GFP-PSPC1-CT construct to confirm that Proline-rich C-terminal domain is a major binding site on PSPC1 for PTK6 binding. However, the full-length of PSPC1 is cloned in a HA vector. Both PSPC1-CT and full-length of PSPC1 should be individually cloned in the same vector. 3. For Figure 3b immunoprecipitation assay, the input of β-catenin, PSPC1 and PTK6 as well as internal control such as β-actin should be shown. 4. For Supplementary 4a, the groups of cell morphology in 2D culture are not labeled. 5. The spheroid images of Figure 2b are not clear.

Point-by-point response to the reviewer's comments:
Reviewers' comments: All of the responses are underlined.

Reviewer #1 (Remarks to the Author):
Paraspeckle component 1 (PSPC1) is shown to associate with PTK6 in HCC cell lines. Expression of PSPC1 Y523F abolishes interaction.
Overall it is not clear if PSPC1 and its phosphorylation is playing any direct role or if its overexpression is simply modulating PTK6 activity. Brauer, Zheng et al. 2010 showed that targeting active PTK6 to the nucleus is growth inhibiting. Here, overexpression of a nuclear Cterminal proline rich fragment of PSPC1 (CT-131) that presumably binds the PTK6 SH3 domain and relieves PTK6 autoinhibition in the nucleus, has therapeutic activity. This would allow PTK6 to phosphorylate other PTK6 nuclear targets such as parafibromin (Hatakeyama and colleagues, 2017 and 2018) or even beta-catenin (Palka-Hamblin et al. 2010). It is interesting that CT-131 has therapeutic activities but its mechanism of action is not clear.

Comments:
1. There is no discussion of recent papers showing that parafibromin, a beta-catenin regulator, is a PTK6 substrate, and how this might impact the interpretation of the studies (Tang et al., 2018 andKikuchi et al., 2017).

Responses:
As suggested, we added a paragraph and four references in the "Discussion" section at Page 23. As following: "Our study of PSPC1 as a substrate to sequester PTK6 tyrosine kinase in nucleus provides additional supports for the crucial role of nuclear tyrosine phosphorylation in modulating selective cofactors such as Sam68, β-catenin, PSPC1 and parafibromin of transcription factor complexes for activation of specific target genes to pursue the designated biological functions 1 -4 . For instance, parafibromin is a nuclear substrate of tyrosine phosphorylation by PTK6 and then dephosphorylation by SHP2 to cooperate with different transcription factor complexes such as phosphor-parafibromin/YAP/TEAD and de-phosphoparafibromin/TAZ/TEAD/β-catenin/TCF respectively for activation of different target genes 4 , 5 " 2. HGF is used "for induction of cancerous EMT microenvironment," but discussion of HGF activation of PTK6 (BRK) and cell migration through MET is not included (Lange and colleagues, 2010(Lange and colleagues, , 2012(Lange and colleagues, , 2013.

Responses:
Thank you for suggestions. We cited these papers and described in

Responses:
As suggested, we performed immunofluorescence images of active PTK6 and total form of PTK6 in Fig. 2f and immunoblotting data of total and active PTK6 for nuclear, cytosol and membrane fractions in Fig. 3a and 3c in Huh-7 and SK-hep1 cells.
5. Fig. 1a: You cannot definitively identify proteins on a Coomassie stained gel without immunoblotting.

Responses:
We moved original Fig. 1a to Supplementary Fig 1a and Fig. 2d). Moreover, we also performed PTK6 ADP-Glo™ Kinase Assay to measure the kinase ability of PTK6 in PSPC1 Y523F mutation and found that PTK6 could phosphorylate PSPC1 at Y523 residue in Supplemental Figure 6b.
8. Fig. 2b. Are these really spheroids and not clusters? The fluorescent cells do not appear to be part of the spheroids as indicated in the legend. It would be interesting to include staining for PSPC1, PTK6, and P-PTK6 in Fig. 2b.

Responses:
Thank you for the suggestions, we re-performed spheroids culture in phase image in the indicated PSPC1/PTK6 transfectants in SK-hep1 cells and displayed new figures in Fig. 2i. As suggested, we performed immunofluorescence staining for PSPC1, PTK6 and p-PTK6 in Fig. 2f.
9. In Fig. 6f, CT131 increases activation of nuclear PTK6 but not cytoplasmic PTK6 and there does not appear to be a shift in total PTK6. Immunofluorescence staining for active P-PTK6 and total PTK6 should be shown.

Responses:
As suggested, we performed immunofluorescence staining for active P- Responses: a. We send our manuscript for English editing via Springer Nature

Reviewer #2 (Remarks to the Author):
Lang et al. reported interesting aspects of PSPC1 functions. The authors first identified PTK6, which has tumor suppressive functions, as one of the major PSPC1-interacting proteins. They examined the importance of this interaction in cancer by using cell-based and mouse cancer models as well as cancer clinical data. Then they have shown regulatory roles of PSPC1 in the PTK6/beta-catenin axis. However, the drawback of this manuscript is to contain numerous grammatical errors and typos. In addition, one of the major conclusions is not fully supported by the data shown in the manuscript as described below.

Major points
1. The authors should consider consulting a language editing service.
There are numerous grammatical errors, typos, and uncommon usage of terms in the manuscript. Thus, there are many ambiguous sentences, which often disturb reading. Even in the abstract, multiple sentences are not clear, and thus it would be possible not to convey the authors' thoughts to readers.

Responses:
Thanks, we send our manuscript for English editing via Springer Nature Author Services as shown by the enclosed Certificate.
2. The authors claim that PSPC1-CT131 is a dual inhibitor targeting oncogenic PSPC1 and PTK6, and the interaction between the CT-131 and endogenous PSPC1 is shown in the model of Figure 8. But, no biochemical data for this interaction is shown in the manuscript.  Fig. 2c and Supplementary Fig. 7a). We also performed that ectopic expression of CT131 could co-localize with endogenous PSPC1 and associated with PSPC1 in Mahlavue cells Figures 7d and 7f). b. We further performed co-IP assay to examine the interaction of PSPC1-CT131 could both bind to wild type and Y523F mutant of

Responses:
In Fig. 6h, we did observe PSPC1-CT131-treatment could interact and reduce expression of endogenous PSPC1 in a dose dependent manner.

This observation is validated by immunofluorescence experiment in
Supplementary Fig. 7d. The inhibitory effects of PSPC1-CT131 might hijack the endogenous PSPC1 to increase PSPC1 protein degradation and/or reduce PSPC1 transcription. More detailed experiments will be required for future optimization of PSPC1-CT131 as a therapeutic agent.

Minor points
1. Fig. 1c: The labels of immunoblotting should be corrected to HA or FLAG as used in Figure S1b and S1c. (Input section)

Responses:
We corrected those figures. Thanks! 2. In Figure S2a, in vitro kinase assay using PTK6 and PSPC1 WT is performed. The PSPC1-Y523F mutant is actually not phosphorylated in this assay?

Responses:
We further performed the PSPC1-WT and PSPC1-Y523F mutant examined by in vitro kinase assay and confirmed that PSPC1-Y523F were actually not phosphorylated by PTK6 in Supplementary Fig. 2d.
3. In middle cartoon of Figure 8, cytoplasmic PTK6 should be phosphorylated according to the data shown in Figure 3c?

Responses:
Corrected. Thanks! 4. Methods: In "Antibodies" section, information of cell lines and plasmids are included. Please fix this point.

Responses:
Corrected. Thanks! 5. Methods: Transfection methods are described in "PSPC1 sitedirected mutagenesis" section. The transfection methods should be separated or combined with other sections.

Reviewer #3 (Remarks to the Author):
General comment: PTK6 kinase plays an opposite role in tumor progression which is dependent on its subcellular localization and transportation between nucleus and cytoplasm. PSPC1 protein, a substrate of PTK6 kinase closed to splicing speckles and recently has been reported by Dr. Jou's group to regulate TGF-β signal pathway for pro-metastatic switches of cancer cells. In this study, Lang et al. showed that PSPC1-Y523F cause oncogenic subcellular translocations of cytoplasmic PTK6 and nuclear β-catenin as well as Wnt3a autocrine loop for promoting liver cancer progression. PSPC1-CT131 which is the proline-rich C-terminal domain of PSPC1 containing NLS (nuclear localization signal), interacts with PTK6 in nucleus to inhibit PSPC1/PTK6 oncogenic functions. The study is interesting and provides mechanistic insights for understanding the role of subcellular localization of PSPC1/PTK6/β-catenin in regulating liver cancer progression. However, some key experiments supporting individual claims may not be sufficient to reveal causal and concrete regulator y mechanisms. Several issues should be clarified.

Responses:
We performed additional co-IP assays for PTK6-SH3 domain with PTK6 deletion constructs and PSPC1 proline-rich region deletion construct. Figures 1f, g). We validated that proline-rich region of PSPC1 directly bind to SH3 domain of PTK6.

Comment 2:
The author claims that PSPC1 upregulation and PSPC1-Y523F released PTK6 nuclear sequestration; however, the existing evidence of this study for demonstrating the above results are not enough.
In particular, the interaction between PSPC1 and PTK6 should be confirmed in different cellular fractions and the immunofluorescent images which indicate PSPC1-Y523F instead of PSPC1-wild type released PTK6 nuclear sequestration should be provided.

Responses:
Thank you for the suggestions. We performed PSPC1 and PTK6 interaction in different cellular fraction (nuclear and cytosol) in SK-hep1 cells (Figs. 3d and 3e). We also performed immunofluorescence experiments to indicate PSPC1-Y523F instead of PSPC-wild-type released PTK6 nuclear sequestration as shown in Fig. 2f.
Comment 3: For demonstrating the interaction between PSPC1 and PTK6 and their subcellular translocation, almost experiments in this study were performed by ectopic gene expression system. To obtain more convincing evidence that PSPC1 interacts with PTK6 naturally, endogenous gene expression experiments should be performed.

Responses:
We performed endogenous co-IP assays and showed that PSPC1 could associate with PTK6 endogenously in Huh-7 cells in Fig. 1a and 1b. Comment 4: Which kinase(s) phosphorylates PTK6?

Responses:
The hepatocyte growth factor (HGF) and its specific receptor c-Met tyrosine kinase regulate cancer cell migration, thereby conferring an aggressive phenotype 2 , 6 . Our results demonstrated that HGF-treated Huh

Responses:
As suggested, we added the independent panel images of EGFP in

Responses:
As suggested, we detected Wnt3a expression by using IHC and examined the expression percentage of Wnt3a of human HCC tumor samples (Figs. 5a and 5e) and tumors isolated from xenograft and metastatic tumors ( Supplementary Figures 5a and 5b).

Responses:
Thanks! We have modified accordingly. Figure 1b, the author established a GFP-PSPC1-CT construct to confirm that Proline-rich C-terminal domain is a major binding site on PSPC1 for PTK6 binding. However, the full-length of PSPC1 is cloned in a HA vector. Both PSPC1-CT and full-length of PSPC1 should be individually cloned in the same vector.

Responses:
Actually, we established HA-tagged and GFP-tagged constructs for both full-length and PSPC1-CT131 of PSPC1. We obtained similar results from either HA-or GFP-tag. When examining the effects of PSPC1-CT131 on the PSPC1 interaction and expressions, we will use different tag labeling to avoid cross-reaction. Figure 3b immunoprecipitation assay, the input of β-catenin, PSPC1 and PTK6 as well as internal control such as β-actin should be shown.

Responses:
We added the input (TCL) of β-catenin, PSPC1 and PTK6 as well as the internal control of β-actin in Fig. 3b. 4. For Supplementary 4a, the groups of cell morphology in 2D culture are not labeled.

Responses:
Thanks! We have corrected. Figure 2b are not clear.

Responses:
Thanks! We replaced the spheroid images in Fig. 2i.  Fig. 1a), and PSF associates with PTK6 through SH3 domainpolyproline interaction and is also a PTK6 substrate. Potential repercussions of disrupting PTK6-PSF association by overexpressing PSPC1 or PSPC1-CT131 are not discussed. The impact of PSPC1 overexpression and knockdown on endogenous PTK6 localization and activity is not demonstrated. In addition, PTK6 is not overexpressed alone, but only in combination with PSPC1 constructs, so it is difficult to differentiate distinct functions for PTK6 and PSPC1.  The SK-Hep-1 cell line used in several experiments (Fig. 2f, 2g, 3c etc.) appears to express endogenous PTK6, at levels not so different from the SNU387 "high" PTK6 expresser in Fig. S1 h, but we do not see any endogenous PTK6 by IF or western. Several liver cancer cell lines coexpress both PSPC1 and PTK6. The data would be stronger if the impact of PSPC1 on endogenous PTK6 localization/activation was examined.

References in the
The Fig. 2g legend states that PTK6 suppressed PSPC1-potentiated EMT. HA-PSPC1 is not expressed without PTK6 in this figure, so one cannot tell. With coexpressed wild type PSPC1, total and active PTK6 are nuclear (Fig. 2f), and nuclear PTK6 has been shown to inhibit beta-catenin transcription ( SK-hep1 transfectant orthotopic tumor studies with cells expressing PTK6 alone have not been included, so it is impossible to ascertain the role of PTK6 without overexpressed PSPC1.
Immunofluorescence for total and p-PTK6 should be shown in Fig. 6b, not in supplemental data.
In Fig. 6f similar levels of total PTK6 are present in the cytoplasmic and nuclear fractions in cells expressing either the PSPC1-CT131 or MutNLS-CT, not supporting the authors' assertion that expression of PSPC1-CT, but not MutNLS-CT altered subcellular localization of PTK6. While total protein levels do not appear to change, PSPC1-CT131 expression does result in increased active p-PTK6 in the nucleus (Fig. 6c nucleus). Do PSPC1 and PSPC1-CT131 sequester PTK6 in the nucleus or transport a pool of p-PTK6 to the nucleus? Or activate PTK6 in the nucleus through binding its SH3 domain relieving intramolecular inhibition?
Overexpression of CT131 would compete with other interactions of the PTK6 SH3 domain, such as its ability to interact with PSF. Potential contributions of PSF which binds both PSPC1 and PTK6 and is coimmunoprecipitated with PSPC1 in Fig. 1a are not discussed.
More detail should be provided in the figure legend of the model in presented in Fig. 8. It should be noted that many functions attributed to PSCP1 in the figure have been previously demonstrated for PTK6, including regulation of beta-catenin, cell migration and the EMT.
There are still some grammatical/spelling mistakes throughout the manuscript. In Fig. 2f legend, the meaning of "in divergent" is unclear? Fig. S1 h. Mahlavu or Mahalvu?
Reviewer #2 (Remarks to the Author): After the author's revision, I still notice multiple issues such as lack of controls in the experiments.
In addition, I have to claim that the mechanisms of actions of PSPC1-CT131 remains still ambiguous even though it is an integral part of this manuscript. Our comments on the authors' responses are shown below (start from arrows). . Reviewer #2 (Remarks to the Author): Lang et al. reported interesting aspects of PSPC1 functions. The authors first identified PTK6, which has tumor suppressive functions, as one of the major PSPC1-interacting proteins. They examined the importance of this interaction in cancer by using cell-based and mouse cancer models as well as cancer clinical data. Then they have shown regulatory roles of PSPC1 in the PTK6/beta-catenin axis. However, the drawback of this manuscript is to contain numerous grammatical errors and typos. In addition, one of the major conclusions is not fully supported by the data shown in the manuscript as described below.
Major points 1. The authors should consider consulting a language editing service. There are numerous grammatical errors, typos, and uncommon usage of terms in the manuscript. Thus, there are many ambiguous sentences, which often disturb reading. Even in the abstract, multiple sentences are not clear, and thus it would be possible not to convey the authors' thoughts to readers.
Responses: Thanks, we send our manuscript for English editing via Springer Nature Author Services as shown by the enclosed Certificate.
-> OK, but if you added anything after the English editing, they should be checked again by the editing service.
2. The authors claim that PSPC1-CT131 is a dual inhibitor targeting oncogenic PSPC1 and PTK6, and the interaction between the CT-131 and endogenous PSPC1 is shown in the model of Figure 8. But, no biochemical data for this interaction is shown in the manuscript. Previous studies have shown that PSPC1 dimerize and oligomerize with DBHS family proteins via the NOPS domain and the coiled-coil domain, respectively, but C-terminal domain is not reported for contributing the interaction. The PSPC1-CT131 alone actually binds PSPC1 (and Y523F mutant)?
Responses: For answering these questions, we added multiple new figures and modified our description in Pages 17-18. a. Our results so far demonstrated that the C-terminal domain of PSPC1 (PSPC1-CT131) served as a molecular docking target to PSPC1 and PTK6 (Fig. 6a) to modulate oncogenic subcellular translocations and facilitate tumor progression in HCC. With unique proline-rich interacting domain of PSPC1-CT131, we hypothesized that PSPC1-CT131 might simultaneously target PSPC1 and SH3 domain29 of PTK6 (Supplementary Fig. 2c and Supplementary Fig. 7a). We also performed that ectopic expression of CT131 could co-localize with endogenous PSPC1 and associated with PSPC1 in Mahlavue cells ( Supplementary Figures 7d and 7f). b. We further performed co-IP assay to examine the interaction of PSPC1-CT131 could both bind to wild type and Y523F mutant of PSPC1 in SK-hep1 in Supplemental Figures 7g.
-> In Supplementary Figure 7d, since no merged images are shown, it is hard to judge if they are co-localized. It would be better to show co-localization by using line profiles. In Supplementary   Figure 7f, the input data of the IP should be shown. In Supplementary Figure 7g, negative controls should be included. The authors claim that GFP-PSPC1-CT binds both PSPC1 WT and Y523F, however, it cannot rule out the possibility that the beads and/or antibodies used bind PSPC1 WT and Y523F. Thus, negative controls are essential.
3. Related to the above point, how does endogenous PSPC1 functions? In Figure 6h, PSPC1 expression was reduced by expression of EGFP-PSPC1-CT131. What is an interpretation for this result? Is this possible to be related to the mechanism of PSPC1 inhibition by the CT-131? Responses: In Fig. 6h, we did observe PSPC1-CT131-treatment could interact and reduce expression of endogenous PSPC1 in a dose dependent manner. This observation is validated by immunofluorescence experiment in Supplementary Fig. 7d. The inhibitory effects of PSPC1-CT131 might hijack the endogenous PSPC1 to increase PSPC1 protein degradation and/or reduce PSPC1 transcription. More detailed experiments will be required for future optimization of PSPC1-CT131 as a therapeutic agent.
-> The authors observed reduction of PSPC1 by the expression of PSPC1-CT, although the mechanisms are unclear. The authors claim that this observation is validated by the immunofluorescence data. However, I am still wondering how the PSPC1 antibody did not crossreact with PSPC1-CT131? Which part is the epitope of the antibody in PSPC1? It would be also important to show quantification data of fluorescent intensities of PSPC1 using quantified multiple images and performing statistical analyses. The authors mentioned several possible mechanisms for the actions of PSPC1-CT131. I think the authors should provide some experimental data to support their hypothesis. At present, it is even extremely unclear if the mechanism is direct or indirect. Thus, the model shown in Figure 8c would contain too much analogy.
Minor points 1. Fig. 1c: The labels of immunoblotting should be corrected to HA or FLAG as used in Figure S1b and S1c. (Input section)

Responses:
We corrected those figures. Thanks! -> The label looks still strange. In the previous version, the IPs were performed with STK6. However, in the revised version it is replaced by HA. If HA antibody was used for the IP, it does not fit the immunoblot data. The authors should precisely describe the data and carefully prepare the figures. Figure S2a, in vitro kinase assay using PTK6 and PSPC1 WT is performed. The PSPC1-Y523F mutant is actually not phosphorylated in this assay?

Responses:
We further performed the PSPC1-WT and PSPC1-Y523F mutant examined by in vitro kinase assay and confirmed that PSPC1-Y523F were actually not phosphorylated by PTK6 in Supplementary Fig.  2d.
-> There is no negative control data for this kinase assay. PTK(-) controls should be required. In addition, phosphatase treatment would be beneficial to show that the signals are phosphorylated proteins.
3. In middle cartoon of Figure 8, cytoplasmic PTK6 should be phosphorylated according to the data shown in Figure 3c?

Reviewer #1 (Remarks to the Author):
The authors have addressed many of the concerns raised in the previous review.
However, several questions remain. The splicing regulator PSF binds both PTK6 (Lukong, Richard, 2009) and PSPC1 (co-immunoprecipitation in Fig. 1a   The SK-Hep-1 cell line used in several experiments (Fig. 2f, 2g, 3c etc.) appears to express endogenous PTK6, at levels not so different from the SNU387 "high" PTK6 expresser in Fig. S1 h, but we do not see any endogenous PTK6 by IF or western.
Several liver cancer cell lines coexpress both PSPC1 and PTK6. The data would be stronger if the impact of PSPC1 on endogenous PTK6 localization/activation was examined.

Response:
To illustrate the impact of PSPC1 and PSPC1-Y523F on endogenous PTK6 localization/activation, we expressed PSPC1 constructs into SNU-387 cells to survey its impact on endogenous PTK6 localizations by western blotting analysis. Immunofluorescence for total and p-PTK6 should be shown in Fig. 6b, not in supplemental data.

Response:
As suggested, we moved the results that "PSPC1-CT131, but not MutNLS-CT131, co-localized PSPC1 and p-PTK6 in nucleus via immunofluorescent analysis" to

Response:
Since PSPC1 and PSPC1-CT131 mainly expressed in nucleus, the possibility to act as a chaperon to transport p-PTK6 to the nucleus is unlikely. Therefore, we suggest that PSPC1 or PSPC1-CT131 could sequester active p-PTK6 in the nucleus through binding its SH3 domain.
Overexpression of CT131 would compete with other interactions of the PTK6 SH3 domain, such as its ability to interact with PSF. Potential contributions of PSF which binds both PSPC1 and PTK6 and is coimmunoprecipitated with PSPC1 in Fig. 1a are not discussed. More detail should be provided in the figure legend of the model in presented in Fig. 8.

Response
It should be noted that many functions attributed to PSCP1 in the figure have been previously demonstrated for PTK6, including regulation of beta-catenin, cell migration and the EMT.

Response:
We modified the legend of the proposed model in Figure 8 as suggested.
There are still some grammatical/spelling mistakes throughout the manuscript. In Fig.   2f legend, the meaning of "in divergent" is unclear? Fig. S1 h. Mahlavu or Mahalvu?
Response: a. We corrected the legend statement to "Immunofluorescence for the detection of subcellular localizations of PSPC1, p-PTK6 and PTK6 in SK-Hep1" in c. We send the manuscript to a senior professor for Scientific Editing. We also send the manuscript to the AJE for the English Editing (certificate enclosed).

Reviewer #2 (Remarks to the Author):
After the author's revision, I still notice multiple issues such as lack of controls in the experiments. In addition, I have to claim that the mechanisms of actions of PSPC1-CT131 remains still ambiguous even though it is an integral part of this manuscript.
Our comments on the authors' responses are shown below (start from arrows).
Reviewer #2 (Remarks to the Author): Lang et al. reported interesting aspects of PSPC1 functions. The authors first identified PTK6, which has tumor suppressive functions, as one of the major PSPC1-interacting proteins. They examined the importance of this interaction in cancer by using cell-based and mouse cancer models as well as cancer clinical data. Then they have shown regulatory roles of PSPC1 in the PTK6/beta-catenin axis. However, the drawback of this manuscript is to contain numerous grammatical errors and typos. In addition, one of the major conclusions is not fully supported by the data shown in the manuscript as described below.
Major points 1. The authors should consider consulting a language editing service. There are numerous grammatical errors, typos, and uncommon usage of terms in the manuscript.
Thus, there are many ambiguous sentences, which often disturb reading. Even in the abstract, multiple sentences are not clear, and thus it would be possible not to convey the authors' thoughts to readers.

Responses:
Thanks, we send our manuscript for English editing via Springer Nature Author Services as shown by the enclosed Certificate.
-> OK, but if you added anything after the English editing, they should be checked again by the editing service.

Response:
We have send the manuscript to a senior PI in our organization for Scientific  Fig. 2c and Supplementary Fig. 7a). We also performed that ectopic expression of CT131 could co-localize with endogenous PSPC1 and associated with PSPC1 in Mahlavue cells ( Supplementary Figures 7d and 7f).
b. We further performed co-IP assay to examine the interaction of PSPC1-CT131 could both bind to wild type and Y523F mutant of PSPC1 in SK-hep1 in Supplemental Responses: In Fig. 6h, we did observe PSPC1-CT131-treatment could interact and reduce expression of endogenous PSPC1 in a dose dependent manner. This observation is validated by immunofluorescence experiment in Supplementary Fig. 7d. The inhibitory effects of PSPC1-CT131 might hijack the endogenous PSPC1 to increase PSPC1 protein degradation and/or reduce PSPC1 transcription. More detailed experiments will be required for future optimization of PSPC1-CT131 as a therapeutic agent.
-> The authors observed reduction of PSPC1 by the expression of PSPC1-CT, although the mechanisms are unclear. The authors claim that this observation is validated by the immunofluorescence data. However, I am still wondering how the PSPC1 antibody did not cross-react with PSPC1-CT131? Which part is the epitope of the antibody in PSPC1?
It would be also important to show quantification data of fluorescent intensities of PSPC1 using quantified multiple images and performing statistical analyses. I think the authors should provide some experimental data to support their hypothesis.

Response
At present, it is even extremely unclear if the mechanism is direct or indirect. Thus, the model shown in Figure 8c would contain too much analogy.

Response:
In Minor points 1. Fig. 1c: The labels of immunoblotting should be corrected to HA or FLAG as used in Figure S1b and S1c. (Input section) Responses: We corrected those figures. Thanks! -> The label looks still strange. In the previous version, the IPs were performed with STK6. However, in the revised version it is replaced by HA. If HA antibody was used for the IP, it does not fit the immunoblot data. The authors should precisely describe the data and carefully prepare the figures.

Response:
We corrected the labelling of subfigures in Fig. 1c. Figure S2a, in vitro kinase assay using PTK6 and PSPC1 WT is performed. The PSPC1-Y523F mutant is actually not phosphorylated in this assay?

Responses:
We further performed the PSPC1-WT and PSPC1-Y523F mutant examined by in vitro kinase assay and confirmed that PSPC1-Y523F were actually not phosphorylated by PTK6 in Supplementary Fig. 2d.
-> There is no negative control data for this kinase assay. PTK(-) controls should be required. In addition, phosphatase treatment would be beneficial to show that the signals are phosphorylated proteins.

Response:
Thanks for your suggestions. We further performed the PTK6 negative controls and the phosphatase treatment followed by in vitro kinase assay analysis as shown in Supplementary Fig. 2d.

Methods:
In "Antibodies" section, information of cell lines and plasmids are included.
Please fix this point.

Responses:
Corrected. Thanks! -> The title of the "Proteins tested by antibodies and characteristics of the corresponding antibodies" in supplementary table should be edited. And in the table, the boundaries of the target antibodies are unclear, especially between PSPC1 and PTK6. Lines should be put.

Response:
We corrected and modified the Supplementary Table 4. Thanks! 5. Methods: Transfection methods are described in "PSPC1 site-directed mutagenesis" section. The transfection methods should be separated or combined with other sections.

Responses:
Corrected. Thanks! -> Transfection methods are described in the method section "Cell culture and transfection". The title should be changed (e.g., to "Cell culture, plasmids, and transfection". The reference numbers in the method section should be shown with superscript font style.

Response:
We corrected the method section. Thanks!

SK-hep1
Reply to Reviewer#1 suggestion: Q: Figure 1d: How are PTK6 WT and KM localized within the cell without the overexpression of PSPC1? It would be interesting to also include the membrane fraction.
Subcellular distribution of wild-type (WT) and kinase-dead (KM) mutant of flag-tagged PTK6 in SK-hep1 cells by Western blotting analysis. Sp1, α-tubulin and Na+/K+-ATPase were used as internal controls for nuclear, cytoplasmic and membrane fractions, respectively. Overall the manuscript is improved. The data clearly demonstrate that PSPC1 and PTK6 interact and regulate each other. It is interesting that there is a correlation between PSPC PY523 , nuclear PTK6 and better patient outcome. A few issues with data remain.
It would be useful to include the rebuttal Fig R2 in the Supplemental figures. Figure 2i is described as spheroids. It is difficult to conclude that any spheroids are shown in the figure. Images look like cell clusters and clumps.
Some of the immunoblotting data are difficult to interpret. In Fig. 2h, multiple bands are present for several of the proteins, including PTK6. In Fig. 3B, total PTK6 in the IP blot appears to be running much smaller, as it is far from the IgG band, in contrast to several figures where it is very close to the IgG band. Size markers are not provided for the Fig. 3B TCL blots.
Reviewer #2 (Remarks to the Author): The authors addressed the major and minor points of my original revision. Therefore, I think that the manuscript is now suitable for publication.