Characterisation of an interacting network of proteins that regulate endosomal tubulation and the HSP phenotype

Applicant: Evan Read pHD

This is a collaborative project between the Reid group in Cambridge and the Beetz group in Jena.  With Dr Beetz, we have identified and begun to characterise a gene that modifies the severity of spastin-HSP.  Spastin-HSP patients who have abnormality of this gene have a significantly younger age at onset of hereditary spastic paraplegia.  Our preliminary data show that the modifying gene encodes a protein that functions in the regulation of endosomal tubulation, a cellular process in which we have recently shown spastin to be involved.  Using an integrated set of experiments in cultured cell lines and in vivo models, we aim to understand how this disease-modifying protein functions, and how its function relates to that of spastin.  We will specifically test the hypothesis that additive effects of abnormality of both proteins on endosomal tubule fission causes the disease have an earlier age at onset. If this were the case, it would open up a new therapeutic candidate pathway that could be targeted with the aim of increasing age at onset of spastin-HSP, so decreasing the disease burden. In addition, we will also explore whether the upregulated BMP signalling seen in cellular models of spastin HSP can be explained by abnormal traffic of BMP receptors through endosomal tubular compartments regulated by spastin and this new modifying protein.

 

 

The team

left to right: Rachel Allison, Evan Reid, Timothy Newton and James Connell,

Final report

 

TWS final report 04/02/15

 

Spastin and Dpy30 have an epistatic effect in HSP

The most frequent cause of HSP is an autosomal dominant mutation in SPAST (spastin).   Families with SPAST mutations show significant variations in the age of onset of HSP symptoms, but the underlying molecular explanation for this is unknown.  In this respect we are investigating DPY30 (Dpy-30), a gene that is situated ‘head to head’ with spastin.  We have found in collaboration with Christian Beetz (University of Jena) that there is an epistatic effect between DPY30 and SPAST. Around 10-15% of spastin mutations involve large deletions, which can extend into neighbouring regions of the genome.  When these deletions include the promoter region or exon 1 of DPY30 a significantly earlier age of onset of HSP symptoms is observed.  We have shown that an increase in endosomal tubulation occurs when cells are depleted of spastin, and this phenotype is also observed when cells are depleted of Dpy-30.  

We aim to test the hypothesis that increased disease severity in patients haplo-insufficient for spastin and Dpy-30 is caused by additive effects on the same cell biological pathway.  We propose that this pathway may involve endosomal tubulation and/or abnormal lysosomal biogenesis.

Work Carried out and Underway

Automation of data collection for endosomal tubulation:

The increased endosomal tubulation seen when cells are depleted of spastin has also been observed in other proteins involved in endosomal fission such as IST1.  Further study of this process and its relation to genes mutated in HSP was impeded by the necessity of identifying and counting endosomal tubules by hand to quantify a deficiency in tubule fission.  This process was a time consuming and rate limiting step, in need of automation.  We have therefore developed an image analysis protocol using imageJ and R that selectively counts tubules present in cells more reproducibly and rapidly than manual counting.  This has been shown to work for Ist1, Spastin and Dpy30 knockouts.  We are currently writing this up as a  methods paper.

Characterising a functional relationship between SPAST and DPY30

As patients with a genomic deletion involving SAPST and DPY30 have an early age of onset, and both genes have a role in regulating endosomal tubulation, we investigated whether an additive effect could be observed in endosomal tubulation when both genes were knocked down by siRNA in mammalian cells.  Initial experiments in HeLa cells suggest no additive effect can be seen when both genes were depleted.  A likely explanation for this is that spastin knockdown (which does not fully reflect the disease state of haplo-insufficiency) has saturated the amount of endosomal tubulation possible in this assay.

To address this issue, we aimed to generate cell models partially lacking both spastin and DPY30, which would be more reflective of the true disease state seen in patients.  We initially proposed that by titrating siRNA levels we may achieve a partial knock-down of SPAST or DPY30, allowing an additive effect to be observed.  However, this is a technically challenging process, as it is difficult to reproducibly generate consistent partial knockdowns in these genes to levels seen in heterozygous patients.  We therefore opted to use a powerful new technique, CRISPR-Cas based gene editing, to knock-out alleles of spastin and DPY30 in human fibroblasts and HeLa cells. We have carried out initial experiments with this technique and pilot data indicates that we have generated clones with null spastin and dyp-30 alleles individually.  Subsequent generation of cell lines doubly haplo-insufficient for Spastin and Dpy30 will provide an excellent system to assess any additive effects accurately, and we will do this in the near future once the appropriate lines are made. 

We also aim to use the SPAST and DPY30 depleted fibroblasts generated by this work to create, via induce pluripotent stem cell techniques, human neuronal cell lines, with precise human disease mutations.  These can be used to investigate in human neuronal cells whether haplo-insufficiency of both genes has an additive effect phenotypes such as endosomal tubulation and axonal swelling. 

In complementary studies, we also plan to use CRISPR-based gene editing on cells from a spastin mouse knock-in model generated within the lab. The dominant-negative N384K mutation in this model, equivalent to human N386K, renders spastin ATPase defective.  By using CRISPR to remove one or both functional alleles of DPY30 in heterozygous or homozygous knock-in cells, we will determine the additive effects on endosomal tubulation. 

Dpy30 is not recruited to endosomes

As DPY30 has a putative interaction with the endosomal ESCRT-complex protein CHMP2B, and as Dpy30 regulates endosomal tubulation, we aimed to test whether Dpy30 is recruited to endosomes.  We examined this with a dominant negative form of a protein, Vps4 (Vps4 E235Q) which traps the ESCRTIII complex and associated proteins on endosomes.  No co-localisation was observed between Vps4EQ-GFP and Dpy-30, suggesting that no such recruitment via the ESCRT-III complex takes place.

To confirm the putative interaction between Dpy30 and CHMP2B, we have optimised antibodies for both proteins for immunoblotting and immunoprecipiation. These experiments are underway.

Protein domains required for regulation of endosomal tubulation by dpy30

We proposed to carry out siRNA-rescue studies to investigate the functional requirements for dpy30 to regulate endosomal tubulation.  Four rescue constructs were generated.  Dpy30 contains a ‘dpy30 like domain’ at the C-terminus and an unstructured N-terminal tail.  Within the C-terminal half of the protein two leucine residues, 65 and 66, mediate the interaction with ASH2L, another part of the WRAD complex.  Therefore in addition to a siRNA resistant construct, three more mutant forms were generated:

1)      No N-terminal unstructured tail.

2)      No C-terminal Dpy30 domain.

3)      A full-length construct, which incorporates L65E and L66E substitutions, preventing binding to ASH2L.

These constructs were used to create stable cell lines.  However, only cells containing the wildtype and L65E,L66E constructs survived.  This suggests truncated forms of Dpy30 are toxic to cells, possibly due to misfolding and ER dysfunction.  Pilot experiments to optimise siRNA rescue indicated that the relatively low increase in endosomal tubulation seen in a Dpy30 knockdown in this system, combined with the low expression of Dpy30 constructs would result in very low effect sizes.  We therefore plan to use the CRISPR edited cell lines lacking Dpy30 to generate new cell lines expressing the rescue constructs.  The complete knockout seen in CRISPR cell lines will increase signal to noise ratios and allow rescue experiments to be carried out.

To determine whether the role DPY30 has in the regulation of endosomal tubulation requires the function of the WRAD complex the depletion of complex members ASH2L and RBBP5 was carried out in HeLa cells.  Three siRNAs were used to target each gene.  Increases in endosomal tubulation were observed with three siRNAs to ASH2L and two siRNAs to RBBP5, with one siRNA to Ash2L causing a statistically significant increase.

Identifying lysosomal enlargement as a novel phenotype of spastin and dpy30 depletion

Further work carried out by other lab members on the effect of spastin depletion in HeLa cells with siRNA and in MEFS from our knock-in mouse has identified the enlargement of lysosomes, when stained with LAMP1 antibodies.  This could suggest that cellular components that should be recycled by the endosome are instead trafficked to the lysosome for degradation, a potential explanation for pathogenesis in HSP.  Preliminary data suggests this phenotype can also be observed in cells depleted of Dpy30.  Using the CRISPR gene edited cell lines described above, we will analyse whether haplo-insufficiency of spastin and Dyp-30 have additive effects on this phenotype.

BMP receptor trafficking effects

We have proposed that the abnormal fission of endosomal tubules could lead to the incorrect trafficking of key receptors required for axonal function, such as the BMPR receptor.  We have shown previously that BMP signalling is elevated when spastin is depleted.  Studying this process was impeded by the lack of good quality antibodies for the main mammalian BMP receptors, type Ia, Ib and II.  We have therefore generated stable HeLa cell lines expressing either HA tagged BMPR Ia or myc-tagged BMPR II.  These cells can now be used to determine the effect on receptor localisation that depletion of spastin or Dpy-30 causes, and any additive effect that a double knockdown or double haploinsufficiency may cause.  Preliminary experiments in cells lacking spastin have shown that the BMPRs are re-distributed, supporting the hypothesis that spastin regulates the endosomal membrane traffic of these receptors.

Further funding progress

The work carried out as part of this project allowed us to generate sufficient pilot data to secure further funding for the project, in the form of a two year grant from the US Spastic Paraplegia Foundation.  We are grateful to the TWS for providing us with the necessary pilot funding to enable us to obtain this larger grant.