Study of an FBXO7 patient mutation reveals Fbxo7 and PI31 co-regulate proteasomes and mitochondria

  1. Sara Al Rawi
  2. Lorna Simpson
  3. Neil Q. McDonald
  4. Veronika Chernuha
  5. Orly Elpeleg
  6. Massimo Zeviani
  7. Roger A. Barker
  8. Ronen Spiegel
  9. Heike Laman

Reviewed by

Read the full article See related articles

Listed in

Log in to save this article

Abstract

Mutations in FBXO7 have been discovered associated with an atypical parkinsonism. We report here a new homozygous missense mutation in a paediatric patient that causes an L250P substitution in the dimerization domain of Fbxo7. This alteration selectively ablates the Fbxo7-PI31 interaction and causes a significant reduction in Fbxo7 and PI31 levels in patient cells. Consistent with their association with proteasomes, L250P patient fibroblasts have reduced proteasome activity and proteasome subunits. We also show PI31 interacts directly with the MiD49/51 fission adaptor proteins, and unexpectedly, PI31 acts as an adaptor enabling SCF Fbxo7 ligase to ubiquitinate MiD49. Thus, the L250P mutation changes the function of Fbxo7 by altering its substrate repertoire. Although MiD49/51 expression was reduced in L250P patient cells, there was no effect on the mitochondrial network. However, patient cells had higher levels of ROS and reduced viability under stress. Our study shows that Fbxo7 and PI31 affect each other’s functions in regulating both proteasomal and mitochondrial function and demonstrate a new function for PI31, as an adaptor for the SCF Fbxo7 E3 ubiquitin ligase.

Article activity feed

  1. Review Commons

    Note: This rebuttal was posted by the corresponding author to Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Reply to the reviewers

    The authors do not wish to provide a response at this time.

  2. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #3

    Evidence, reproducibility and clarity

    In the present paper, authors study the effect of the L250P mutation on Fbxo7, leading to a severe infantile onset motor impairment.

    They show by co-IP that the mutation selectively ablates the interaction of Fbxo7 with the protesomal adaptor PI31. It induced a reduction in endogenous Fbxo7, which had a reduced half-life, and concomitantly of PI31 levels, without affecting its half-life, and a reduction in the expression of some subunits of the proteasome, and altogether, a reduction in its activity. To identify substrates of SCF Fbxo7 reliant on PI31 they used some databases and proteins arrays and identified MiD49 and MiD51, involved in mitochondrial fission machinery. Authors show PI31 acts as an adaptor and allows SCFfbxo7 ligase to ubiquitinate MiD49, therefore L250P mutation alters its substrate repertoire. It is shown Fbxo7 stabilizes Mid49 and Mid51. Importantly, reduced levels of Fbxo7 (KD) mimic the effect of the mutation.

    Despite the affectation of MiD49, mitochondrial network appeared unaffected. However they observed that Fbxo7 L250P mutation led to a general alteration of the mitochondrial function: reduction of the mitochondrial mass, leading to reduced oxygen consumption, lower mitophagy and biogenesis and increased ROS production. Data are well presented, findings are convincing and complete and discussion seems appropriate.

    I just have some minor comments:

    According to the data showing that Fbxo7 KD mimics the effects of the L250P mutation, it appears that the altered stabilization of Fbxo7 is a key event in the process and results observed. How do the authors explain the reduction of Fbxo7 half-life induced by the mutation?

    I would find more appropriate that to estimate the mitochondrial mass, the relative area or volume of Mitotracker green fluorescence is quantified, rather than its average intensity.

    It is not accurate to say TMRE is only able to enter mitochondria and fluorescence when there is an intact membrane potential. Rather, its accumulation is dependent on the mitochondrial membrane potential.

    Significance

    Authors provide a comprehensive study of the effects of the novel mutation L250P in Fbxo7 gene, leading to infantile onset PD. Findings are new and describe the mechanism this mutation changes the substrate repertoire of SCFFbxo7 ligase and affect mitochondrial function. The paper could be of interest to both clinical and basic researchers focused on PD and mitochondrial function.

  3. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #2

    Evidence, reproducibility and clarity

    Summary

    In this article, Sara Al Rawi and colleagues hypothesize that the homozygous L250P FBXO7 mutation identified in an infant with reduced facial and limb movements and axial hypotonia affects the Fbxo7-PI31 interaction domain. The authors used patient fibroblasts carrying the pathogenic mutation to show reduced expression of Fbxo7 and PI31 on those cells, and reduced proteasome levels and activity. Moreover, their data suggest that the L250P mutation affects the mitochondrial function and promotes increased levels of reactive oxygen species (ROS).

    Major comments:

    • The only thing that be argued is that Supplementary Table 1 containing the list of whole exome sequencing (WES) hits is missing. Please add it to the manuscript to check that the FBXO7 variant is the most plausible one considering the clinical phenotype of the affected individual.

    Minor comments:

    There are a few minor details that I would modify to improve the general readability of the manuscript:

    • The authors state that parents of the affected individual are "related", but do not specify the degree of consanguinity that they have. It was only in my second read of the manuscript that I realized that parents were consanguineous, and that the presence of a homozygous mutation made sense. It would help to state the degree of consanguinity between parents so that the reader can understand that a homozygous mutation is plausible.
    • In the clinical description of the patient, I would add the interpretation of a "developmental quotient (DQ) = 40", such as "a score <75 indicates a developmental delay", since some basic scientists may not be used to interpret those scores and cannot identify right away the degree of clinical impairment.
    • Many researchers are used to interpret the potential pathogenicity of a "candidate variant" using the CADD score. You can add this score to the paragraph where you mention the predicted pathogenicity of other in silico tests.
    • In the section "Fbxo7 stabilizes MiD49/51 protein levels", the beginning of the second paragraph: "To test whether knock-down of Fbxo7 levels phenocopies the effect of the L250P point mutation, [...]" is hard to understand. I would rewrite this first sentence in a different way to make it easier to read.

    Significance

    The manuscript is well written, and the results are coherent and supported by impressively detailed methods. The results of this manuscript are important for the advance of the movement disorders field. It shows how the novel L250P FBXO7 mutation in homozygous status can cause a very early onset (neonatal) movement disorder. Additionally, the functional analyses performed in patients' fibroblasts characterize very well the effect of this mutation on the cellular machinery showing proteasomal dysfunction, mitochondrial dysregulation, and ubiquitination alteration.

    The content of the manuscript is interesting for a broad spectrum of individuals: from pediatric movement disorders specialists to basic researchers interested in proteasomal and mitochondrial dysfunction.

  4. Review Commons

    Note: This preprint has been reviewed by subject experts for Review Commons. Content has not been altered except for formatting.

    Learn more at Review Commons

    Referee #1

    Evidence, reproducibility and clarity

    This study reports a human pediatric patient carrying a mutation in the Fbxo7/PARK15 gene. Several previous reports demonstrated that loss of Fbxo7/PARK15 gene function causes early-onset parkinsonian-pyramidal syndrome, but the precise underlying molecular mechanism remains to be elucidated. Based on studies with patient fibroblasts the authors suggest that Fbxo7 and its conserved binding partner PI31 regulate proteasomes and mitochondria, and that PI31 is an adaptor for the SCF Fbxo7 E3 ubiquitin ligase. The observations presented here are potentially interesting, but at this point they lack sufficient experimental evidence to support the main conclusions.

    A general weakness of this study is that experiments focus on patient fibroblasts, and not neurons. Fbxo7/PARK15 patients as well as mouse mutants have neurological phenotypes, but no overt defects in skin or connective tissues. Therefore, it is not clear how the reported observations relate to the clinical symptoms. This is of particular concern since a conflicting report (Kraus et al., 2023) has found no change in basal mitophagy in Fbxo7 KO iNeurons.

    The authors show that the proteasome regulatory protein PI31 is cleaved in Fbxo7 mutant cells. The cleavage and inactivation of PI31 upon inactivation of Fbxo7 was originally reported in Bader et al., 2011 (Cell 145, 371-82), but this paper is not cited. On the other hand the authors are citing a yet unpublished biorxiv preprint by Sanchez-Martinez et al. (https://www.biorxiv.org/content/10.1101/2022.10.10.511602v3) on the Drosophila Fbxo7 ortholog but another highly relevant preprint showing that transgenic expression of PI31 can extensively compensate for the inactivation of Fbxo7 in mice is not included (https://www.biorxiv.org/content/10.1101/2020.05.05.078832v1). This lack of consistency in citing prior works is concerning and should imperatively be rectified to provide a transparent and more accurate account of the novelty of the presented findings. Therefore, the bibliography of the manuscript is incomplete and relevant citations are missing.

    The idea that PI31 is an adaptor for SCF-FBXO7 ubiquitination has not sufficient experimental support. The authors use the correlation of the L250P mutation not interacting with PI31 and not ubiquitinating MiD49 to propose that PI31 is an adaptor needed for MiD49 (and TOMM22 and Rpl23) ubiquitination by FBXO7 - this is an over-interpretation; this claim should be toned down or supported by further investigations.

    FBXO7 can ubiquitinate MiD49 in vitro, but in vivo it appears to protect MiD49 from ubiquitination. Moreover, reduced FBXO7 levels (either by L250P mutation or knock down) result in decreased MiD49 and MiD51. There is no mechanistic explanation for these seemingly contradictory findings.

    Mitochondrial homeostasis of patient fibroblasts appears aberrant. In particular, there seems a reduction of mitochondrial mass, reduced basal mitophagy and reduced respiration, which contrasts the report by Krauss et al. (2023). A potential explanation for these disparate observation is suggested by the decreased transcription of two mitochondrial transcription factors, PGC1α and PPARγ. How FBXO7 inactivation leads to this decrease in PGC1α and PPARγ or a decrease in basal autophagy is not clear.

    In sum, there are some potentially interesting preliminary observations, but the study is not convincing because the results are over-interpreted and the analysis is not sufficiently rigorous.

    Specific comments:

    Figure 1 panel C has no loading control for total lysates.

    Figure 2 panel B. The authors state that this experiment suggests direct interaction of PI31 with MiD49 (in the discussion they drop "suggests"). GST-pulldown with bacterially expressed GST-PI31 with MiD49/51 in vitro transcribed in rabbit reticulocyte lysate, which is less complex than HEK293 lysate, but not a purified system (for example, they contain proteasomes and other UPS components). This does not prove direct interaction. They show a Coomassie gel of the purified GST and GST-PI31, but not the reticulocyte lysate.

    Figure 2 panels I and J. In panel I they show a decrease of Mi49 following FBXO7 KD, a main point of the paper that MiD49 is a FBXO7/PI31 substrate. However, in panel J for time zero of their hydrogen peroxide treatment time course it appears that Mi49 from the FBXO7 KD is as abundant if not greater than the Mi49 for the control time zero. Why? Even in the graph in panel K they start at the same level, though that probably is due to the way they normalize the data, which is not stated clearly.

    The legend for figure 3 panel C states that the figure shows cells from control and the patient imaged under basal conditions or following treatment with 2-deoxyglucose, yet in the figure there are only two panels!?

    For the proteasome activity assay, in the text they state that they use Suc-LLVY-AMC, which releases a fluorescent signal when proteolytically cleaved, but in the Materials and Methods they say that they use the Proteasome-Glo Chymotrypsin-like assay (Promega G8621), which is a two-step luminescent assay. These are not the same assays.

    Finally, this ms is very similar to a bioxrchive paper from the Laman-lab, but there are some notable omissions:

    1. In their bioxrchive paper they show a very clear decrease in LMP7(beta5i) protein in figure 1 panel D. This would go along with the decrease in other proteasome subunits, but this result is not mentioned in this manuscript. Why?
    2. The bioXrchive figure 1 panel E was replaced here with a considerably lower quality Western blot (Figure 1 panel F, using tubulin instead of GAPDH as the loading control). Again, the reason is not clear.

    Significance

    A better understanding of how mutations in Fbxo7/PARK15 cause juvenile onset neuronal degeneration would be very important and significant. Unfortunately, the current study is not sufficiently rigorous while the results are over-interpreted which will confuse readers.

  5. Version published to 10.1101/2021.12.22.473884v1 on bioRxiv