Computational analysis of the LRRK2 interactome

Abstract

LRRK2 was identified in 2004 as the causative protein product of the Parkinson's disease locus designated PARK8. In the decade since then, genetic studies have revealed at least 6 dominant mutations in LRRK2 linked to Parkinson's disease, alongside one associated with cancer. It is now well established that coding changes in LRRK2 are one of the most common causes of Parkinson's. Genome-wide association studies (GWAs) have, more recently, reported single nucleotide polymorphisms (SNPs) around the LRRK2 locus to be associated with risk of developing sporadic Parkinson's disease and inflammatory bowel disorder. The functional research that has followed these genetic breakthroughs has generated an extensive literature regarding LRRK2 pathophysiology; however, there is still no consensus as to the biological function of LRRK2. To provide insight into the aspects of cell biology that are consistently related to LRRK2 activity, we analysed the plethora of candidate LRRK2 interactors available through the BioGRID and IntAct data repositories. We then performed GO terms enrichment for the LRRK2 interactome. We found that, in two different enrichment portals, the LRRK2 interactome was associated with terms referring to transport, cellular organization, vesicles and the cytoskeleton. We also verified that 21 of the LRRK2 interactors are genetically linked to risk for Parkinson's disease or inflammatory bowel disorder. The implications of these findings are discussed, with particular regard to potential novel areas of investigation.

Keywords: GO terms enrichment; Interactome; LRRK2; Parkinson’s disease; Protein–protein interactions.

Figure 1. LRRK2 heterologous interactions as reported in the merged dataset.

LRRK2 in the middle of the graph is linked to candidate partners (black dots) through 422 annotations as described in the merged dataset. Some partners have one connection only; others have multiple connections based on the number of annotations. Different colours represent different methods used to infer the interaction; note that since IntAct and BioGRID use different classifications for the interaction detection method, a simplified and harmonized version has been applied to this figure to help the reader. In particular: (i) Affinity Capture stands for—Affinity Capture-Western and Affinity Capture-MS (in BioGRID)—Anti Bait Coimmunoprecipitation, Anti Tag Coimmunoprecipitation, Coimmunoprecipitation, Pull Down and Affinity Chromatography Technology (in IntAct); (ii) Biochemical Activity stands for—Biochemical Activity (in BioGRID)—Protein Kinase Assay (in IntAct); (iii) Co-localization stands for—Co-localization (in BioGRID)—Confocal microscopy and Fluorescence Microscopy (in IntAct); (iv) Reconstituted Complex stands for—Fluorescence Polarization Spectroscopy and Surface Plasmon Resonance (in IntAct).

Figure 2. Number of annotations capturing each of the LRRK2 interactors.

Only the interactors reported in 2 or more annotations, and used in the enrichment analysis, are included in the figure, all other interactors were identified in just a single experiment. (A) High occurrence LRRK2 interactors, >1 publication, >1 method; (B) Medium-high occurrence LRRK2 interactors, >1 publication, 1 method; and (C) Medium-low occurrence LRRK2 interactors, 1 publication, >1 method.

Figure 3. Families of LRRK2 interactors.

Each of the 62 proteins in the LRRK2 filtered interactome was associated to a family (if available) based on the UniProtKB “family&domains” classification. Proteins belonging to the same family were plotted accordingly to the total number of publications (y-axis); each family was then classified based on the authorships as non-independent (different family members described by the same research group) or independent (different family members described by at least two different research groups).

Figure 4. Dendrogram of “GO biological process” terms enriched for LRRK2 interactors in WebGestalt.

Hierarchical levels of the dendrogram are alternatively represented in blue and yellow; red text indicates the top 10 GO terms. The table lists details of the enriched terms: GO term and ID, number of proteins in the GO term category (Ref.), number of LRRK2 interactors associated with the GO term (Genes), p-value adjusted for multiple testing.

Figure 5. Dendrogram of “GO cell component” terms enriched for LRRK2 interactors in WebGestalt.

Hierarchical levels of the dendrogram are alternatively represented in blue and yellow; red text indicates the top 10 GO terms. The table lists details of the enriched terms: GO term and ID, number of proteins in the GO term category (Ref.), number of LRRK2 interactors associated with the GO term (Genes), p-value adjusted for multiple testing.

Figure 6. LRRK2 interactors enrichment for “GO cell component” and “GO biological process” terms.

The LRRK2 filtered interactome is shown after WebGestalt enrichment analysis. The enriched GO terms were grouped in: transport/localization, cell organization, regulation of kinase activity, cytosol, vesicles, cytoskeleton and cell projections. Every protein in the filtered interactome was connected to the GO term groups it has participated toward the enrichment. In the table the 7 groups are listed in columns (A–G) and rows (1–7). The percentage of proteins from the LRRK2 interactome that contributed toward the enrichment of the GO term group listed in the row and that were also reported in the GO term group reported in the column was calculated (intersection between the group in the row and the group in the column). In the last row of the table the percentage of the proteins that contributed toward the enrichment of each GO term group was calculated against the total number of LRRK2 interactors (i.e., 62 proteins). Cells discussed in the text are highlighted.

Figure 7. “GO biological process” terms enriched for LRRK2 interactors in GO supported by Panther.

Pie charts showing the composition of every significance level of enrichment. The legend shows the 12 GO term groups, although in the charts the group name was reported only for those that reached the 10% contribution toward the enrichment.

Figure 8. Annotations for LRRK2 self-interaction.

LRRK2 in the middle of the graph is linked to the first author of publications describing self-interaction. Different colours represent different methods used to infer the interaction; note that since IntAct and BioGRID use different classifications for the interaction detection method, a simplified and harmonized version has been applied to this figure to help the reader. In particular: (i) Affinity Capture stands for—Affinity Capture-Western (in BioGRID)—Anti Tag Coimmunoprecipitation, and Pull Down (in IntAct); (ii) Biochemical Activity stands for—Biochemical Activity (in BioGRID)—Protein Kinase Assay (in IntAct).

Figure 9. Profile of LRRK2 self-interaction.

Each residue described as involved in LRRK2 self-interaction is reported with a dot, the dimension of the dot is proportional to the number of publications annotating that specific residue. A profile covering the entire length of the LRRK2 protein is shown in blue, in the bottom half of the image; the y value of the profile represents the number of publications in which the fragment in the profile was reported as associated with LRRK2 self-interaction.