Characterizing protein–protein interactions (PPIs) could help researchers uncover pathways of disease progression, thereby aiding diagnostic and preventative strategies. Hosp et al. (2015) took this approach, using quantitative interaction proteomics to develop interactome maps for a number of neurodegenerative diseases (NDDs).1 Creating PPI networks and comparing wild type with disease-specific protein isoforms, the researchers hoped to gain insight on functional changes important in pathogenesis.
The research team used a number of techniques to explore disease-specific interactomes:
- Quantitative affinity purification with mass spectrometry–based proteomics analysis (qAP-MS)
- Stable isotope labeling of amino acids in cell culture (SILAC)
- Transient transfection of human HEK293T cells in vitro to express myc-tagged expression vectors for cross-linking and coprecipitation studies
- Drosophila melanogaster models for in vivo examination of protein function
- Reporter assays
Focusing on proteins such as Ataxin-1 known to be associated with NDDs such as Alzheimer’s disease (AD), Parkinson’s disease, Huntington’s disease and spinocerebellar ataxia type 1, Hosp et al. examined interactomes associated with both wild type and disease-associated variants. The researchers initially performed qAP-MS on SILAC-labeled cell lysates, examining the samples by liquid chromatography–tandem mass spectrometry (LC-MS/MS) on LTQ Orbitrap, LTQ Orbitrap XL and LTQ Orbitrap Velos mass spectrometers (all Thermo Scientific). On initial analysis, they found that most of the binding proteins identified were non-specific. However, with closer interrogation of the data, the research team could narrow down to identify specific interaction partners.
With further refinement of the workflow, Hosp et al. completed qAP-MS screens for five disease-specific proteins associated with the four NDDs. They compared the interactomics of wild type proteins with those found for the disease-associated variants. This involved 72 pull-downs of 12 bait-tagged proteins, running at least two forward and two reverse experiments. The team achieved good consistency with this approach, calculating a correlation coefficient of R = 0.83 overall.
Setting validation criteria of at least two times enrichment over controls and showing inverted SILAC ratios in reverse experiments, the team identified 373 proteins in the interactome networks. Although most PPIs were unique to a single disease, 10 showed shared interaction partners with at least three out of the four NDDs under investigation. Hosp et al. validated the data with western immunoblotting and co-immunoprecipitation, and found that it repeated findings in other studies.
Turning to in vivo studies using Drosophila, the researchers used shRNA knockdowns in a tissue-specific manner to examine the effect of protein interaction partners on the NDD phenotype. Hosp et al. found enhancement of the NDD phenotype as shown by changes to the rough-eye phenotype.
To examine the interactome data in association with human disease, the researchers conducted a genome-wide association study for AD. They examined single nucleotide polymorphisms associated with AD and found significant associations with the interactomics data collected. Furthermore, comparisons between wild type and disease-associated proteins showed that interactome maps generated were functionally relevant in AD.
The researchers then explored the functional relevance of the data. Noting that mitochondrial dysfunction is an early marker of AD, they focused on leucine-rich pentatricopeptide repeat motif-containing protein (LRPPRC), an essential regulator of the mitochondrial respiratory chain. They found that LRPPRC preferentially associated with the disease-specific form of Amyloid beta precursor protein (APPsw) found in early onset AD. Furthermore, immunohistochemistry showed that the two proteins co-localize within cells in AD brain tissue sections but not in AD plaques, thus suggesting intracellular association. The team also found that APPsw downregulated LRPPRC mRNA and protein levels, and reduced mitochondrial function.
The researchers acknowledge that the initial interactome data come from overexpression studies in a human cell line and therefore may not be completely applicable to in vivo disease development. However, the supplementary studies do suggest a role for protein partners. Furthermore, Hosp et al. are confident that quantitative interaction proteomics can aid clinical research in determining the role played by disease-specific variants in pathogenesis of NDD phenotypes.
1. Hosp, F., et al. (2015) “Quantitative interaction proteomics of neurodegenerative disease proteins,” Cell Reports, 11(7) (pp.1134–46), doi: 10.1016/j.celrep.2015.04.030.