Interfering with the ERC1–LL5β interaction disrupts plasma membrane–Associated platforms and affects tumor cell motility

Cell migration requires a complex array of molecular events to promote protrusion at the front of motile cells. The scaffold protein LL5β interacts with the scaffold ERC1, and recruits it at plasma membrane–associated platforms that form at the front of migrating tumor cells. LL5 and ERC1 proteins support protrusion during migration as shown by the finding that depletion of either endogenous protein impairs tumor cell motility and invasion. In this study we have tested the hypothesis that interfering with the interaction between LL5β and ERC1 may be used to interfere with the function of the endogenous proteins to inhibit tumor cell migration. For this, we identified ERC1(270–370) and LL5β(381–510) as minimal fragments required for the direct interaction between the two proteins. The biochemical characterization demonstrated that the specific regions of the two proteins, including predicted intrinsically disordered regions, are implicated in a reversible, high affinity direct heterotypic interaction. NMR spectroscopy further confirmed the disordered nature of the two fragments and also support the occurrence of interaction between them. We tested if the LL5β protein fragment interferes with the formation of the complex between the two full-length proteins. Coimmunoprecipitation experiments showed that LL5β(381–510) hampers the formation of the complex in cells. Moreover, expression of either fragment is able to specifically delocalize endogenous ERC1 from the edge of migrating MDA-MB-231 tumor cells. Coimmunoprecipitation experiments show that the ERC1-binding fragment of LL5β interacts with endogenous ERC1 and interferes with the binding of endogenous ERC1 to full length LL5β. Expression of LL5β(381–510) affects tumor cell motility with a reduction in the density of invadopodia and inhibits transwell invasion. These results provide a proof of principle that interfering with heterotypic intermolecular interactions between components of plasma membrane–associated platforms forming at the front of tumor cells may represent a new approach to inhibit cell invasion.


Editor Comments:
Thank you for submitting your manuscript to PLOS One. Your work has now been evaluated by two referees and both of them showed great interest in the story. There are, however, a few concerns that need to be addressed before we can move forward. I hope that you find these comments helpful to improve the manuscript.
We thank the Editor and the two Reviewers for the useful suggestions that have helped us to improve our study. Following are our answers to the points raised by the two Reviewers (reported below in italics). In the file "Revised Manuscript with Track Changes" changes have been underlined and highlighted in yellow. -Insert a sentence describing the differences between lamellipodia and invadopodia.

Reviewer's Responses to
We have now introduced the information in the Introduction, at pages 4, 5, and 6.

I would include an overview about the already published results (or search in databases) about the expression level changes of ERC1 and LL5β in cancers (especially it would be most relevant in breast cancer, as the MDA-MB-231 is a human breast adenocarcinoma cell line used in this study). This would fit at the very end of the results part.
This is a nice suggestion allowing us to better explain the importance of studying these proteins with respect to cancer. In fact, fusions of either LL5b or ERC1 with different genes including tyrosine kinase receptors (e.g. ERC1-RET) [Nakata et al. 1999] have been identified in invasive breast cancer and other tumors (cBioPortal, TCGA). Moreover, tumor cells rely on several molecular pathways that are essential to their survival and function, such as invasion. ERC1/LL5 proteins are part of the PMAPs machinery that regulates tumor cell invasion. Understanding the biology and regulation of this protein network may be relevant not only when these proteins are altered in patients, but also for all tumors in which these proteins are necessary for invasion and the formation of metastases: an example of non-oncogene addiction genes [Luo et al, 2009], that are becoming attractive drug targets for combined cancer therapies [Petroni et al, 2022]. This part has been added to the conclusions, at page 31.

Need reference for this sentence: Line 364: "In COS7 cells the endogenous LL5 protein is represented mostly by LL5alfa"
Since no data have been published yet on the relative expression of LL5a and b in COS7 cells, we have now added a new panel S1A Fig, and  The approach suggested by the Reviewer is interesting. On the other hand it represents an addition requiring quite an extensive effort for this lab, since we have to set conditions for a technique not in use in this lab. We will consider this approach for a future extension of the present work.

Anyway, it is not clear if the MDA-MB-231 cell express or not the endogenous LL5β full length protein?
The expression of the endogenous LL5b protein in MDA-MB-231 cells was carefully shown by using isoform-specific antibodies in our previous publication: Astro et al. 2014 Liprin-α1, ERC1 and LL5 define polarized and dynamic structures that are implicated in cell migration. J Cell Sci 127:3862 (see Figure 1 of this publication).
We have now made this clear by mentioning this reference at page 17. The exact molecular function of ERC1 in cell migration is not known yet. We hypothesize that ERC1 is needed to assemble the PMAPs that promote protrusion and the turnover of focal adhesions, as previously shown (Astro et al, 2016). We had mentioned this in the conclusions. Of note, while the fragments affect the localization of endogenous ERC1 at PMAPs, we found that they did not affect the localization of the other endogenous PMAP components liprin-a1 and LL5: this is shown in supplementary figures S2 Fig and S3 Fig.

The formation of invadopodia is multistep process. The function of ERC1 in forming the invadopodia is not mentioned. Which molecules do ERC1 participate in the formation of invasive protrusions? How signal is the pathway induced by ERC1? The author needs to describe. In addition, the author should perform the experiments showing the invadopodial signal transduction including ERC1 and inactivating by adding fragments.
The sequence of molecular events linking ERC1 to the formation of invadopodia is still unknown: this topic is interesting but goes beyond the scope of this manuscript.
3. In quantitative evaluation of invadopoaia, the author used invadopodial number par area. As far as I know, usually, the gelatin degeneration area per cell is correlate the invasive ability. Why the author used the unit " invadopodial number per" ?
We have used the density of invadopodia to correct for the different projected cell areas. We have now added also the quantification of the total number of invadopodia per cell (new Fig 6E).
4. If full length-LL5beta can make dimer/oligomer and the fragment can not, the GFP (with fulllength/fragment LL5beta) density of the ERC1 accumulating legion may be different. Therefore, the correlation of GFP and ERC1 intensity should be measured and compared.
In Figure 4, while full length LL5b accumulates at the ERC1-positive PMAPs, the fragment LL5b(381-510) is diffuse; in addition this fragment induces the almost complete redistribution of ERC1 from PMAPs. This result suggest that the LL5beta fragment, but not the full length LL5beta, displaces endogenous ERC1 from PMAPs. Of note, the LL5beta fragment does not displace the endogenous LL5beta and Liprin-alpha1 (S3 Fig). 3. The reference to  and  in Figure 6 is misleading and should be corrected to LL5β-GFP (). Correction made.