The Networks of Genes Encoding Palmitoylated Proteins in Axonal and Synaptic Compartments Are Affected in PPT1 Overexpressing Neuronal-Like Cells

CLN1 disease (OMIM #256730) is an early childhood ceroid-lipofuscinosis associated with mutated CLN1, whose product Palmitoyl-Protein Thioesterase 1 (PPT1) is a lysosomal enzyme involved in the removal of palmitate residues from S-acylated proteins. In neurons, PPT1 expression is also linked to synaptic compartments. The aim of this study was to unravel molecular signatures connected to CLN1. We utilized SH-SY5Y neuroblastoma cells overexpressing wild type CLN1 (SH-p.wtCLN1) and five selected CLN1 patients’ mutations. The cellular distribution of wtPPT1 was consistent with regular processing of endogenous protein, partially detected inside Lysosomal Associated Membrane Protein 2 (LAMP2) positive vesicles, while the mutants displayed more diffuse cytoplasmic pattern. Transcriptomic profiling revealed 802 differentially expressed genes (DEGs) in SH-p.wtCLN1 (as compared to empty-vector transfected cells), whereas the number of DEGs detected in the two mutants (p.L222P and p.M57Nfs*45) was significantly lower. Bioinformatic scrutiny linked DEGs with neurite formation and neuronal transmission. Specifically, neuritogenesis and proliferation of neuronal processes were predicted to be hampered in the wtCLN1 overexpressing cell line, and these findings were corroborated by morphological investigations. Palmitoylation survey identified 113 palmitoylated protein-encoding genes in SH-p.wtCLN1, including 25 ones simultaneously assigned to axonal growth and synaptic compartments. A remarkable decrease in the expression of palmitoylated proteins, functionally related to axonal elongation (GAP43, CRMP1 and NEFM) and of the synaptic marker SNAP25, specifically in SH-p.wtCLN1 cells was confirmed by immunoblotting. Subsequent, bioinformatic network survey of DEGs assigned to the synaptic annotations linked 81 DEGs, including 23 ones encoding for palmitoylated proteins. Results obtained in this experimental setting outlined two affected functional modules (connected to the axonal and synaptic compartments), which can be associated with an altered gene dosage of wtCLN1. Moreover, these modules were interrelated with the pathological effects associated with loss of PPT1 function, similarly as observed in the Ppt1 knockout mice and patients with CLN1 disease.


De-glycosylation assay
The removal of N-glycosylated residues was performed on 10 μg of protein homogenates in presence of PNGase F (Roche Diagnostics), overnight at 37°C. Untreated and de-glycosylated protein samples were resolved by SDS-PAGE under denaturing conditions. Proteins were transferred to PVDF membranes and probed with anti-PPT1 (Sigma Aldrich), as described in Material and Methods section. A prolonged electrophoretic run was performed to better resolve the PPT1 isoforms.

Morphometric evaluation of axonal-like structures
Differentiated and non-treated, engineered cell lines were analyzed using immunofluorescence of a SMI31-R antibody (Covance), which immunolabelled phosphorylated heavy chain neurofilaments (pNF-H) and, to a lesser extent, medium chain neurofilaments (pNF-M). Images were acquired at 10x magnification by Axiovision (Carl Zeiss); at least 1,500 cells were analysed. ImageJ and NeuronJ plugin were used to manually trace SMI31R immunolabelled structures. Both primary processes (sprouting from the cell body) and secondary processes (branching from a primary process) were traced. Only primary processes longer than 30 µm were taken into account for quantitative, comparative analysis. To normalize the number of neurofilament processes against the cellular density, nuclei were automatically quantified by CellProfiler 2.1.1. Data were collected from at least four independent experiments and reported as mean ± s.e.m; statistical analyses were performed using unpaired t-test versus empty-vector cells in the same culture conditions, considering as significant a pvalue ≤0.05 and as highly significant a p-value ≤0.01.
The degree of colocalisation of PPT1 immunofluorescent signal on Lamp2 lysosomal staining was investigated in neuroblastoma cell lines which were transfected with either a wtCLN1 or with cDNAs harbouring missense mutations (p.L222P and p.V181L). The quantitative analysis confirmed a reduced colocalization of mutated PPT1 in lysosomal compartment as compared to wtPPT1, in accordance with the diffuse immunostaining described in Figure 1 (pictures e, f, h and i). Mean ± s.d.; One-way ANOVA followed by Bonferroni's post-test; ***p<0.001.
Supplementary Figure S2. Characterization of wild-type and missense mutation bearing SH-SY5Y PPT1 cell lines.
In protein homogenates isolated from wtPPT1 overexpressing cells, Western blot analysis with anti-PPT1 antibody revealed two bands of approximately 33 and 31 kDa, as detected in both parental (SH-SY5Y) and empty vector (SH-pcDNA3) cell lines. Following PNGase treatment to remove Nglycosylated residues, wtPPT1 migrated as a single band of approximately 28 kDa. Without PNGase treatment, both missense mutation-bearing cell lines (SH-p.L222P and SH-p.V181L) exhibited a band running at higher molecular weight (37 kDa). Upon PNGase treatment, the same cells exhibited a shift in migration of ~1-2 kDa in comparison to de-glycosylated wtPPT1 (marked by asterisks). "-" not treated sample, "+" PNGase treated samples. In this experimental setup PPT1 proteoforms migrated at molecular weights lower than reported in Figure 1, due to conditions of SDS-PAGE, allowing for better resolving each of the PPT1 forms. Network depicting the subsets of genes embedded in intersections of a Venn diagram from Figure 4A. Twenty-five genes, specifically expressed in SH-p.wtCLN1 (except PPT1), and linked to IPA attributes and GO terms in Figure 4B are portrayed in orange.

Supplementary Figure S5. Cell Component GO term enrichment analysis of palmitoylated protein encoding genes identified in SH-p.wtCLN1 cells.
Several GO terms referred to neuronal components of membrane processes (neuron projection, growth cone and axon) as well as to synaptic compartments (synapse, presynapse, postsynaptic specialization). Fold enrichment (bar) represents the number of input genes observed against the expected number of genes assigned to a specific GO term; a fold enrichment >1 (green line) indicates that the category is overrepresented in the given experimental setting. P-value (red dots), reported as -log10, is determined by the binomial statistic and represents the probability that the number of genes observed in a category occurred by chance; the threshold is set to 1.3 (red line), which corresponds to a p-value <0.05.
Supplementary Figure S6. Qualitative morphological features of CLN1 cell lines following neuronal differentiation in RA-NBM medium.
Morphological changes in CLN1 transfected cells lines were checked by phase contrast microscopy during neuronal differentiation. Exposure to RA-NBM medium induced a significant elongation of membrane processes in all cell lines, with more arborisations at the end of differentiation (9 DIV). SHp.wtCLN1 cells showed less arborized and stunted processes in comparison to mock-transfected cells and other CLN1 transfected cell lines. DIV, days in vitro; scale bar equals to 50 µm.