Pulse-chase SILAC-based analyses reveal selective oversynthesis and rapid turnover of mitochondrial protein components of respiratory complexes

Abstract

Mammalian mitochondria assemble four complexes of the respiratory chain (RCI, RCIII, RCIV, and RCV) by combining 13 polypeptides synthesized within mitochondria on mitochondrial ribosomes (mitoribosomes) with over 70 polypeptides encoded in nuclear DNA, translated on cytoplasmic ribosomes, and imported into mitochondria. We have previously observed that mitoribosome assembly is inefficient because some mitoribosomal proteins are produced in excess, but whether this is the case for other mitochondrial assemblies such as the RCs is unclear. We report here that pulse-chase stable isotope labeling with amino acids in cell culture (SILAC) is a valuable technique to study RC assembly because it can reveal considerable differences in the assembly rates and efficiencies of the different complexes. The SILAC analyses of HeLa cells indicated that assembly of RCV, comprising F₁/F_o-ATPase, is rapid with little excess subunit synthesis, but that assembly of RCI (i.e. NADH dehydrogenase) is far less efficient, with dramatic oversynthesis of numerous proteins, particularly in the matrix-exposed N and Q domains. Unassembled subunits were generally degraded within 3 h. We also observed differential assembly kinetics for individual complexes that were immunoprecipitated with complex-specific antibodies. Immunoprecipitation with an antibody that recognizes the ND1 subunit of RCI co-precipitated a number of proteins implicated in FeS cluster assembly and newly synthesized ubiquinol-cytochrome c reductase Rieske iron-sulfur polypeptide 1 (UQCRFS1), the Rieske FeS protein in RCIII, reflecting some coordination between RCI and RCIII assemblies. We propose that pulse-chase SILAC labeling is a useful tool for studying rates of protein complex assembly and degradation.

Keywords: NADH dehydrogenase; mitochondria; mitochondrial biogenesis; mitochondrial respiratory chain complex; oxidative phosphorylation system; protein assembly; protein dynamics; protein synthesis; protein turnover; proteomics; respiratory complex assembly; stable isotope labeling with amino acids in cell culture (SILAC).

Figure 1.

Kinetics of synthesis and import of mitochondrial RC proteins. Total HeLa cell mitochondrial proteins were isolated after SILAC pulses of 3, 4, 6, or 12 h and analyzed by fragmentation and LC-MS/MS. Average peptide H:L ratios are plotted as a function of labeling time. A, representative reference proteins exhibit similar rates of accumulation of newly synthesized protein with some variability surrounding an average shown by the red line. The dashed line shows the expected rate of accumulation of newly synthesized proteins, assuming exponential growth with a generation time of 24 h and negligible protein turnover. B, subunits of Complex V, F₁F_o-ATPase, on average (purple line) show synthesis kinetics similar to the reference proteins (red line). C, comparison of the synthesis rates of subunits of all RC. NDUFS6 is not included in this analysis, because it is synthesized and incorporated at even higher rates, over twice that of the average RCI subunit. Higher rates of accumulation reflect progressively greater oversynthesis. D, ribbon model of RCI based on RSCB Protein Data Bank model 5LDW with individual subunits colorized according to their extent of oversynthesis in mitochondria from cells pulse-labeled for 3 and 4 h relative to the H:L ratios predicted by the model. Summary H:L ratio data are from Table S1. Molecular models were generated using PyMOL.

Figure 2.

Protein H:L ratios in mitochondria decrease during chase intervals of 3, 7, and 10 h as excess oversynthesized protein is degraded. A and B, peptide H:L ratios of reference proteins and RCV subunits, respectively, during the chase incubation. C, comparison of the decreasing H:L ratios of peptides in subunits of each respiratory complex during the chase. Shown are averages (bars) and S.D. (error bars) for the various groups of proteins. The initial large error bars reflect diversity in the levels of oversynthesis of individual proteins after pulse labeling, whereas the smaller error bars after a 10-h chase indicate that this variation is decreased as excess protein copies are degraded.

Figure 3.

Comparison between the H:L ratios following 6-h pulse labeling in the total mitochondrial fraction (red bars) and those in immunoprecipitated respiratory complexes (blue bars). Note the change in scale on the y axis between panels, reflecting much higher levels of synthesis and import of some proteins than others. The horizontal dotted lines indicate that ∼19% of complexes are expected to be newly synthesized during a 6-h pulse interval. Error bars are not shown on these figures, but the primary data available in Table S3 indicate an average error of ±5% of the mean values for 45 RC proteins with good peptide coverage in both the total mitochondrial and IP samples. A and B, H:L ratios are shown for peptides in proteins identified in both total mitochondrial preparations (red) and samples immunoprecipitated with complex-specific antibodies (blue) in complex V and complex IV, respectively. C and D, H:L ratio results for respiratory complex I proteins similar to those in A and B for the matrix-exposed N and Q domains in C and the proximal (PP) and distal (PD) membrane domains in D.

Figure 4.

Newly synthesized UQCRFS1 and a module of LYR-domain proteins are selectively immunoprecipitated with RCI. A, IP of RCI from lysates of mitochondria from cells labeled for 6 h recovered several proteins that are not structural subunits of RCI (Table S3). For proteins represented by statistically significant numbers of peptides (>20) the chart shows the mean values ± S.E. (error bars) for the total mitochondrial pool (red) and for the fraction of protein recovered by RCI IP (blue). UQCRFS1 stands out as a member of another RC, RCIII, with a significant overrepresentation of newly synthesized protein. B, the network of protein interactions among NDUFAB1, its associated LYR domain proteins in RCI (NDUFA6 and NDUFB9), UQCRSF1, NFS1, and other assembly factors recovered in RCI IP samples visualized using STRING (http://string-db.org).³