Abstract Cryo-electron tomography (cryo-ET) visualizes the 3D spatial distribution of macromolecules at nanometer resolution inside native cells. While this label-free cryogenic imaging technology produces data containing rich structural information, automated identification of macromolecules inside cellular tomograms is challenged by noise and reconstruction artifacts, as well as the presence of many molecular species in the crowded volumes. Here, we present a computational procedure that uses artificial neural networks to simultaneously localize with a multi-class strategy several macromolecular species in cellular cryo-electron tomograms. Once trained, the inference stage of DeepFinder is significantly faster than template matching, and performs better than other competitive deep learning methods at identifying macromolecules of various sizes in both synthetic and experimental datasets. On cellular cryo-ET data, DeepFinder localized membrane-bound and cytosolic ribosomes (~3.2 MDa), Rubisco (~540 kDa soluble complex), and photosystem II (~550 kDa membrane complex) with comparable accuracy to expert-supervised ground truth annotations. Furthermore, we show that DeepFinder is flexible and can be combined with template matching to localize the missing macromolecules not found by one or the other method. The DeepFinder algorithm is therefore very promising for the semi-automated analysis of a wide range of molecular targets in cellular tomograms, including macromolecules with weights of 500-600 kDa and membrane proteins.
In mammalian cells, one-third of all polypeptides is transported into or through the ER-membrane via the Sec61-channel. While the Sec61-complex facilitates the transport of all polypeptides with amino-terminal signal peptides (SP) or SP-equivalent transmembrane helices (TMH), the translocating chain-associated membrane protein (now termed TRAM1) was proposed to support transport of a subset of precursors. To identify possible determinants of TRAM1 substrate specificity, we systematically identified TRAM1-dependent precursors by analyzing cellular protein abundance changes upon TRAM1 depletion in HeLa cells using quantitative label-free proteomics. In contrast to previous analysis after TRAP depletion, SP and TMH analysis of TRAM1 clients did not reveal any distinguishing features that could explain its putative substrate specificity. To further address the TRAM1 mechanism, live-cell calcium imaging was carried out after TRAM1 depletion in HeLa cells. In additional contrast to previous analysis after TRAP depletion, TRAM1 depletion did not affect calcium leakage from the ER. Thus, TRAM1 does not appear to act as SP- or TMH-receptor on the ER-membrane’s cytosolic face and does not appear to affect the open probability of the Sec61-channel. It may rather play a supportive role in protein transport, such as making the phospholipid bilayer conducive for accepting SP and TMH in the vicinity of the lateral gate of the Sec61-channel. Abbreviations: ER, endoplasmic reticulum; OST, oligosaccharyltransferase; RAMP, ribosome-associated membrane protein; SP, signal peptide; SR, SRP-receptor; SRP, signal recognition particle; TMH, signal peptide-equivalent transmembrane helix; TRAM, translocating chain-associated membrane protein; TRAP, translocon-associated protein.
