LRRK1 functions as a scaffold for PTP1B-mediated EGFR sorting into ILVs at the ER–endosome contact site

ABSTRACT Proper control of epidermal growth factor receptor (EGFR) signaling is important for maintaining cellular homeostasis. Given that EGFR signaling occurs at the plasma membrane and endosomes following internalization, endosomal trafficking of EGFR spatiotemporally regulates EGFR signaling. In this process, leucine-rich repeat kinase 1 (LRRK1) has multiple roles in kinase activity-dependent transport of EGFR-containing endosomes and kinase-independent sorting of EGFR into the intraluminal vesicles (ILVs) of multivesicular bodies. Active, phosphorylated EGFR inactivates the LRRK1 kinase activity by phosphorylating Y944. In this study, we demonstrate that LRRK1 facilitates EGFR dephosphorylation by PTP1B (also known as PTPN1), an endoplasmic reticulum (ER)-localized protein tyrosine phosphatase, at the ER–endosome contact site, after which EGFR is sorted into the ILVs of endosomes. LRRK1 is required for the PTP1B–EGFR interaction in response to EGF stimulation, resulting in the downregulation of EGFR signaling. Furthermore, PTP1B activates LRRK1 by dephosphorylating pY944 on the contact site, which promotes the transport of EGFR-containing endosomes to the perinuclear region. These findings provide evidence that the ER–endosome contact site functions as a hub for LRRK1-dependent signaling that regulates EGFR trafficking.

In the present study Hanafusa and colleagues demonstrated that LRRK1 is required for EGFR dephosphorylation by ER-localized PTP1B, which occurs at ER-endosome contact sites, followed by sorting of dephosphorylated/inactivated EGFR into the intraluminal vesicles of endosomes. LRRK1 does not participate in EGF-induced formation of the ER-endosome contact sites, but does function as a scaffold that mediates the interaction of PTP1B with phosphorylated EGFR, leading to downregulation of EGF signaling. This is a well-written paper and provides a new insight into the function of LRRK1, and thus is a very good candidate for publication in JCS after the following points are adequately answered.

Comments for the author
Major points: First, it appears that the amount of phosphorylated EGFR in siLRRK1-treated cells (right lane in the "EGFR, Total lysates" panel) was a bit smaller than that in siControl cells (3rd lane from the right), raising the possibility that no binding of EGFR to Flag-PTP1B(C215S) in siLRRK1-treated cells is due to a smaller amount of phospho-EGFR produced. Regarding this, why was an anti-pY1068-EGFR antibody not used for immunoblotting? Because the anti-EGFR antibody gave fuzzy bands, likely reflecting the formation of phosphorylated forms, use of an anti-pY1068-EGFR antibody may give a clearer result. Use of the latter antibody could also provide direct evidence for sustained activation of EGFR in LRRK1depleted cells (Fig. 2). Second, does Flag-PTP1B(C215S) bind to LRRK1 with a similar affinity to Flag-PTB1B wild-type ( Fig. 6A)? This should be examined because Flag-PTP1B(C215S) was used for binding experiments (Fig. 6B).  It is better to show that expression of an siRNA-resistant LRRK1 construct can compensate for the phenotype of LRRK1-depleted cells. Alternatively, another siRNA should be tested to exclude the possible off-target effects.

Minor points:
Describe how siRNA-resistant Flag-PTP1B constructs were made.

Reviewer 2
Advance summary and potential significance to field In this paper the authors build on their previous observations (confirmed in the current study) that LRRK1 regulates the sorting of EGFR onto intraluminal vesicles (ILVs) of MVBs. They show that LRRK1 regulates the PTP1B-mediated EGFR dephosphorylation that occurs at membrane contact sites between EGFR-containing endosomes and the ER. They further show that LRRK1 is not required for contact site formation using a very elegant split GFP assay that they have validated using a known membrane contact site regulator, annexin 1. They then present evidence that LRRK1 acts as a scaffold to enable EGFR:PTP1B interaction.
One of the most interesting parts of this study is the idea that LRRK1 has 2 roles in EGFR trafficking that are regulated by its phosphorylation. They had previously shown that EGFR phosphorylates LRRK1 on a site that inhibits LRRK1 kinase activity and that LRRK1 kinase activity though not involved in sorting of EGFR onto ILVs is required for movement of EGFR-containing endosomes to the perinuclear region.
Here they show that LRRK1 is a substrate of PTP1B. A potential order of events is presented whereby EGFR phosphorylates and inactivates LRRK1 and LRRK1 acts as a scaffold at membrane contact sites to allow PTP1B and EGFR to interact and drive sorting of EGFR onto ILVs. PTP1B then dephosphorylates LRRK1 allowing the LRRk1 kinase activity to promote movement of EGFR containing endosomes to the perinuclear region. This is a very interesting and novel model and definitely of potential interest to readers of J Cell Science.

Comments for the author
Some very nice approaches have been used in this study. However quantitation of effects on EGFR and LRRK1 phosphorylation is lacking in some places as detailed below, as well as additional points. Fig. 1A: Endosomes appear larger in the controls than the LRRK1 and PTP1B siRNA-treated cells. Is this the case? If anything I would expect that inhibition of ILV formation would lead to enlarged MVBs. The images should be shown at the same magnification and if the changes in endosome size are consistent then they should be commented upon. Figure 2 shows the effects of LRRK1 depletion on Erk phosphorylation. The effects of LRRK1 depletion on phosphorylation of EGFR (compared with total EGFR) should also be shown. This is important because, although accumulation of phosphorylated EGFR upon LRRK1 depletion is shown in Figure 5, this is not quantified and the role of LRRK1 in regulating PTP1B-mediated EGFR dephosphorylation is central to the paper.
Fig 5B shows that when LRRK1 or PTP1B are depleted there are more punctae costaining for GFP and phospho-EGFR suggesting that phosphorylated EGFR is accumulating at membrane contact sites. Whilst this is a very elegant assay, only single cells are shown. Could the number of GFP/phospho-EGFR positive punctae/cell under the different conditions be quantitated? Figure 6 B is investigating the effect of siLRRK1 on EGFR:PTP1B interaction using a catalytically inactive form of PTP1B which prolongs what is normally a very transient interaction between the phosphatase and its substrate. The authors state that the interaction between PTP1B was dependent on EGFR stimulation and reduced by LRRK2 depletion. However, I can see a band at approximately the right molecular weight in all lanes where EGFR has been transfected. I can see a higher molecular weight band in the control siRNA cells treated with EGF that has been indicated by an arrow but I think all of the lanes are showing EGFR and the 10 minute EGFR band has reduced mobility due to phosphorylation. The differences between the lanes apart from the mobility, appear quite subtle so I think that some sort of quantitation with statistics is required.
The text at the top of page 13 is somewhat confusing. It states that the nonphosphorylatable form of LRRK1 results in "EGFR accumulation of mixed endosomes in the perinuclear region". What are 'mixed endosomes'? Also, "These findings suggest that LRRK1 kinase activity remains inactive until the completion of EGFR sorting and requires activation to initiate the endosomal trafficking of EGFR." What is meant by 'initiation of endosomal trafficking'? It might be clearer to say 'transport to the perinuclear region'.
Figure 7 nicely shows that LRRK1 can be dephosphorylated by PTP1B in vitro and that overexpression of a cytosolic form of PTP1B dephosphorylates LRRK1. However the role of endogenous PTP1B in LRRK1 dephosphorylation should be investigated by depleting PTP1B and investigating effects on LRRK1 phosphorylation.
That LRRK1 is a substrate of PTP1B is important and this should be included in the abstract, together with sentence summarising the multiple roles of LRRK1 in regulating EGFR traffic.

First revision
Author response to reviewers' comments First, it appears that the amount of phosphorylated EGFR in siLRRK1-treated cells (right lane in the "EGFR, Total lysates" panel) was a bit smaller than that in siControl cells (3rd lane from the right), raising the possibility that no binding of EGFR to Flag-PTP1B(C215S) in siLRRK1-treated cells is due to a smaller amount of phospho-EGFR produced. Regarding this, why was an anti-pY1068-EGFR antibody not used for immunoblotting? Because the anti-EGFR antibody gave fuzzy bands, likely reflecting the formation of phosphorylated forms, use of an anti-pY1068-EGFR antibody may give a clearer result. Use of the latter antibody could also provide direct evidence for sustained activation of EGFR in LRRK1-depleted cells (Fig. 2).
As suggested, we have performed immunoblotting using the anti-pY1068-EGFR antibody. However, in the PTP1B(C215S) immunoprecipitates, we could not detect the clear bands of pY1068-EGFR. Therefore, we quantified the amount of EGFR immunoprecipitated with PTP1B(C215S) and compared the relative levels of EGF-induced PTP1B-EGFR interaction between control and LRRK1 siRNA-treated cells. We confirm that LRRK1 knockdown decreases the PTP1B-EGFR interaction in response to EGF stimulation (Fig. 6C).

Fig. 7C: Is pY944-LRRK1 present at ER-endosome contact sites?
The localization of pY944-LRRK1 seems to be different from that of pY-EGFR (Fig. 5).
It is better to show that expression of an siRNA-resistant LRRK1 construct can compensate for the phenotype of LRRK1-depleted cells. Alternatively, another siRNA should be tested to exclude the possible off-target effects.
We have already demonstrated that the expression of siRNA-resistant LRRK1 constructs rescues the defect of EGFR ILV sorting in LRRK1-depleted cells (Hanafusa et al., 2011).

Minor points:
Describe how siRNA-resistant Flag-PTP1B constructs were made.
In this study, we have not used the siRNA-resistant Flag-PTP1B construct. As you pointed out, we described the fluorescence probe Alexa-Fluor 405 in the "MATERIALS and METHODS" section. We have corrected this description. As suggested, we have performed the quadruple staining (Fig. 5B). As pointed out, endosomes appeared larger in the controls than in the LRRK1 and PTP1B siRNA-treated cells. In cells depleted of LRRK1 or PTP1B, endosomes were distributed around the plasma membrane ( Fig. 1A-C). These results suggest that depletion of LRRK1 or PTP1B may delay fusion between endosomes. We mention this point (p. 8, lines 21-24) As suggested, we showed the images at the same magnification. Figure 5, this is not quantified and the role of LRRK1 in regulating PTP1B-mediated EGFR dephosphorylation is central to the paper.

Figure 2 shows the effects of LRRK1 depletion on Erk phosphorylation. The effects of LRRK1 depletion on phosphorylation of EGFR (compared with total EGFR) should also be shown. This is important because, although accumulation of phosphorylated EGFR upon LRRK1 depletion is shown in
As suggested, we showed that LRRK1 depletion led to the accumulation of phosphorylated EGFR (Fig. S3). As suggested, we showed quantification of the number of GFP puncta co-localized with pY1068-EGFR (Fig. 5C).

Figure 6 B is investigating the effect of siLRRK1 on EGFR:PTP1B interaction using a catalytically inactive form of PTP1B which prolongs what is normally a very transient interaction between the phosphatase and its substrate. The authors state that the interaction between PTP1B was dependent on EGFR stimulation and reduced by LRRK2 depletion. However, I can see a band at approximately the right molecular weight in all lanes where EGFR has been transfected. I can see a higher molecular weight band in the control siRNA cells treated with EGF that has been
indicated by an arrow but I think all of the lanes are showing EGFR and the 10 minute EGFR band has reduced mobility due to phosphorylation. The differences between the lanes, apart from the mobility, appear quite subtle so I think that some sort of quantitation with statistics is required.
As suggested, we quantified the amount of EGFR immunoprecipitated with PTP1B(C215S) and compared the relative levels of EGF-induced PTP1B-EGFR interaction between control and LRRK1 siRNA-treated cells. We confirm that LRRK1 knockdown decreases the PTP1B-EGFR interaction in response to EGF stimulation (Fig. 6C).
The text at the top of page 13 is somewhat confusing. It states that the nonphosphorylatable form of LRRK1 results in "EGFR accumulation of mixed endosomes in the perinuclear region". What are 'mixed endosomes'? Also, "These findings suggest that LRRK1 kinase activity remains inactive until the completion of EGFR sorting and requires activation to initiate the endosomal trafficking of EGFR." What is meant by 'initiation of endosomal trafficking'? It might be clearer to say 'transport to the perinuclear region'.
We have corrected the description.

Figure 7 nicely shows that LRRK1 can be dephosphorylated by PTP1B in vitro and that overexpression of a cytosolic form of PTP1B dephosphorylates LRRK1. However, the role of endogenous PTP1B in LRRK1 dephosphorylation should be investigated by depleting PTP1B and investigating effects on LRRK1 phosphorylation.
As suggested, we showed that the Tyr-944 phosphorylation of LRRK1 disappeared 30 min after EGF stimulation and that when PTP1B was depleted with siRNA, the pY944-LRRK1 signal remained (Fig. 7B).
That LRRK1 is a substrate of PTP1B is important and this should be included in the abstract, together with sentence summarising the multiple roles of LRRK1 in regulating EGFR traffic.
As suggested, we have rewritten the abstract. I am happy to tell you that your manuscript has been accepted for publication in Journal of Cell Science, pending standard ethics checks. As reviewer #1 has some comments about the abstract, please change it in proofs.

Advance summary and potential significance to field
In the present study Hanafusa and colleagues demonstrated that LRRK1 is required for EGFR dephosphorylation by ER-localized PTP1B, which occurs at ER-endosome contact sites, followed by sorting of dephosphorylated/inactivated EGFR into the intraluminal vesicles of endosomes. LRRK1 does not participate in EGF-induced formation of the ER-endosome contact sites, but does function as a scaffold that mediates the interaction of PTP1B with phosphorylated EGFR, leading to downregulation of EGF signaling. This is a well-written paper and provides a new insight into the function of LRRK1, and in the revised version the authors have adequately addressed my concerns.
Comments for the author I have a couple of minor comments to the abstract. Line 8 in the abstract: It may be better to state "Active, phosphorylated EGFR", not to merely state "Active EGFR". 3rd line from the bottom in the abstract: "the transport of EGFR-containing endosomes" should be changed to "the transport of EGFR-containing endosomes to the perinuclear region"

Reviewer 2
Advance summary and potential significance to field In this paper the authors build on their previous observations (confirmed in the current study) that LRRK1 regulates the sorting of EGFR onto intraluminal vesicles (ILVs) of MVBs. They convincingly show that LRRK1 regulates the PTP1B-mediated EGFR dephosphorylation that occurs at membrane contact sites between EGFR-containing endosomes and the ER. They further show that LRRK1 is not required for contact site formation using a very elegant split GFP assay that they have validated using a known membrane contact site regulator, annexin 1. They then present evidence that LRRK1 acts as a scaffold to enable EGFR:PTP1B interaction.
One of the most interesting parts of this study is the idea that LRRK1 has 2 roles in EGFR trafficking that are regulated by its phosphorylation. They had previously shown that EGFR phosphorylates LRRK1 on a site that inhibits LRRK1 kinase activity and that LRRK1 kinase activity though not involved in sorting of EGFR onto ILVs is required for movement of EGFR-containing endosomes to the perinuclear region. Here they show that LRRK1 is a substrate of PTP1B. A potential order of events is presented whereby EGFR phosphorylates and inactivates LRRK1 and LRRK1 acts as a scaffold at membrane contact sites to allow PTP1B and EGFR to interact and drive sorting of EGFR onto ILVs. PTP1B then dephosphorylates LRRK1 allowing the LRRk1 kinase activity to promote movement of EGFR containing endosomes to the perinuclear region. This is a very interesting and novel model and would definitely be of potential interest to readers of J Cell Science.

Comments for the author
The authors have satisfied by concerns.