www.acsnano.org
Iron-Sequestering Nanocompartments as Multiplexed Electron Microscopy Gene Reporters Downloaded via UNIV OF SOUTHERN INDIANA on July 21, 2019 at 15:19:14 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
Felix Sigmund,†,‡,§,¶ Susanne Pettinger,†,‡,§,¶ Massimo Kube,∥,¶ Fabian Schneider,∥ Martina Schifferer,⊥,# Steffen Schneider,⊗,∇ Maria V. Efremova,†,‡,§,^ Jesús Pujol-Martí,& Michaela Aichler,○ Axel Walch,○ Thomas Misgeld,⊥,# Hendrik Dietz,∥ and Gil G. Westmeyer*,†,‡,§ †
Department of Nuclear Medicine, TUM School of Medicine, Technical University of Munich, 81675 Munich, Germany Institute of Biological and Medical Imaging, Helmholtz Zentrum München, 85764 Neuherberg, Germany § Institute of Developmental Genetics, Helmholtz Zentrum München, 85764 Neuherberg, Germany ∥ Laboratory for Biomolecular Design, Department of Physics, Technical University of Munich, 85748 Garching, Germany ⊥ Institute of Neuronal Cell Biology, TUM School of Medicine, Technical University of Munich, 80802 Munich, Germany # German Center for Neurodegenerative Diseases (DZNE), 81377 Munich, Germany ⊗ Computational Neuroengineering, Department of Electrical and Computer Engineering, Technical University of Munich, 80333 Munich, Germany ∇ Tübingen AI Center, University of Tübingen, 72076 Tübingen, Germany ^ Laboratory of Chemical Design of Bionanomaterials for Medical Applications, Department of Chemistry, Lomonosov Moscow State University, 119991 Moscow, Russian Federation & Department “Circuits - Computation - Models”, Max Planck Institute of Neurobiology, 82152 Martinsried, Germany ○ Research Unit Analytical Pathology, Helmholtz Zentrum München, 85764 Neuherberg, Germany ‡
S Supporting Information *
ABSTRACT: Multicolored gene reporters for light microscopy are indispensable for biomedical research, but equivalent genetic tools for electron microscopy (EM) are still rare despite the increasing importance of nanometer resolution for reverse engineering of molecular machinery and reliable mapping of cellular circuits. We here introduce the fully genetic encapsulin/cargo system of Quasibacillus thermotolerans (Qt), which in combination with the recently characterized encapsulin system from Myxococcus xanthus (Mx) enables multiplexed gene reporter imaging via conventional transmission electron microscopy (TEM) in mammalian cells. Cryo-electron reconstructions revealed that the Qt encapsulin shell self-assembles to nanospheres with T = 4 icosahedral symmetry and a diameter of ∼43 nm harboring two putative pore regions at the 5-fold and 3-fold axes. We also found that upon heterologous expression in mammalian cells, the native cargo is autotargeted to the inner surface of the shell and exhibits ferroxidase activity leading to efficient intraluminal iron biomineralization, which enhances cellular TEM contrast. We furthermore demonstrate that the two differently sized encapsulins of Qt and Mx do not intermix and can be robustly differentiated by conventional TEM via a deep learning classifier to enable automated multiplexed EM gene reporter imaging. KEYWORDS: encapsulin, electron microscopy, cryo-electron microscopy, gene reporter, multiplexing, compartmentalization, iron biomineralization
E
lectron microscopy (EM) has become a gold standard for deciphering mechanistic details of cellular processes1−3 and for uncovering the network architecture © XXXX American Chemical Society
Received: April 23, 2019 Accepted: June 7, 2019 Published: June 7, 2019 A
DOI: 10.1021/acsnano.9b03140 ACS Nano XXXX, XXX, XXX−XXX
Article
Cite This: ACS Nano XXXX, XXX, XXX−XXX
Article
ACS Nano of cell circuits.4−9 However, compared to the versatile arsenal of fluorescent proteins and other optical reporters,10,11 equivalent genetic tools for EM are still sparse and not easily multiplexable. Semigenetic EM reporter systems are based on the genetically controlled precipitation of exogenous chemicals,12−16 which may reduce their spatial precision and tissue penetration, thus limiting their applicability in high-throughput analyses of thicker tissue blocks. In addition, fixation and permeabilization can affect the ultrastructure of samples, which also constrains gold immunolabeling techniques17−20 that may furthermore be complicated by limited epitope accessibility. As for a fully genetic EM marker, the canonical iron storage protein ferritin has been tested in E. coli, yeast, and mammalian cells21−23 for generating EM contrast via its electron-dense iron core.24 However, due to the small size of ferritin (∼12 nm outer diameter, 4−7.5 nm iron core21), it is often difficult to differentiate individual ferritin particles by EM from wellcontrasted cellular structures such as ribosomes. In our search for iron-sequestering protein complexes that are large enough to be reliably distinguished from endogenous cellular objects by EM, we turned to the large family of encapsulins naturally occurring in bacteria and archaea.25−28 This class of proteinaceous spherical nanocompartments has so far been described as icosahedral structures with either T = 1 or T = 3 symmetry (∼18 or ∼32 nm diameter, respectively),27 which can encapsulate cargo proteins with a wide range of functions, e.g., in oxidative stress response, anaerobic ammonium oxidation, or iron biomineralization.27,29,30 It has also been shown that foreign cargos such as (split-)fluorescent proteins or enzymes can be genetically targeted into the encapsulin lumen in bacterial and mammalian hosts.31−35 The encapsulin system of Myxococcus xanthus29 (MxEnc) was shown to encapsulate the ferritin-like cargo proteins MxBCD and mineralize up to ∼30000 iron atoms per nanocompartment, which is about one order of magnitude more than the amount ferritins can store.29 In addition, we have recently proven heterologous expression and self-assembly of MxEnc in mammalian cells and mouse brains and demonstrated that encapsulation of native and engineered cargo proteins still occurs in mammalian cells. This made it possible to locally catalyze iron biomineralization inside the MxEnc shell and detect encapsulin-expressing tissues by MRI and cellular cryoelectron microscopy (cryo-EM).35 To enable multiplexed gene reporter imaging via conventional TEM in mammalian cells, we herein introduce the encapsulin system of Quasibacillus thermotolerans as selfassembling nanospheres (QtEnc) with T = 4 icosahedral symmetry and an ∼43 nm diameter that do not intermix with nanocompartments from Mx (∼32 nm diameter) upon simultaneous expression in mammalian cells. The native cargo (QtIMEF) autotargets to the inner surface of QtEnc and exhibits ferroxidase activity leading to efficient intraluminal iron sequestration resulting in robust EM contrast. These features of Qt nanocompartments allowed their robust differentiation from Mx encapsulins via fully automated image classification with a deep learning model.
(QtEnc, MxEnc) and the native cargos (QtIMEF, MxBCD as a P2A construct) (Figure 1a,b; Supplementary Table 1).
Figure 1. Encapsulins of different bacterial species express, assemble, and load cargo proteins in mammalian cells. (a) Schematic depiction of the heterologous expression and selfassembly of encapsulins in mammalian cells. (b) Scheme of the mammalian expression constructs encoding the shell and cargo proteins of Quasibacillus thermotolerans (QtEnc+QtIMEF) and Myxococcus xanthus (MxEnc+MxBCD). See Supplementary Table 1 for details. (c) Coomassie-stained Blue Native PAGE (BN CM) gel loaded with cell lysates of HEK293T cells expressing QtEnc +QtIMEF and MxEnc+MxBCD. (d) Coomassie-stained SDSPAGE (SDS CM) gel of both encapsulin/cargo systems purified via their FLAG epitope showing bands for the respective shell and coprecipitated cargo proteins. The dye running front is visible above the 10 kDa marker band on both lanes (*).
Blue Native PAGE (BN-PAGE) analysis of lysates from HEK293T cells coexpressing MxEnc+MxBCD showed two bands corresponding to the T = 1 and T = 3 assemblies, as we have previously shown.35 In contrast, expression of QtEnc +QtIMEF yielded a single band running slightly higher than the band corresponding to the T = 3 assembly of MxEnc (Figure 1c). To further investigate the size differences between QtEnc and MxEnc, we purified QtEnc nanoshells from HEK293T cells via their external FLAG-epitope and performed dynamic light scattering (DLS) yielding diameters of 39 ± 14 nm for QtEnc and 40 ± 2 nm for QtEnc+QtIMEF (Supplementary Figure 1). To confirm that autotargeting of QtIMEF to QtEnc still occurs in mammalian cells, we purified encapsulins from HEK293T cells via their external FLAG-epitope and evaluated their protein contents via SDS−PAGE. We detected two bands corresponding to the shell protein and the cargo (QtEnc: 32 kDa, QtIMEF: 23 kDa)27 (Figure 1d). We then used densitometric SDS gel analysis to determine the ratio of shell to cargo for both encapsulin systems and estimated that the MxEnc shell accounted for 47 ± 4% of total protein and the three cargo proteins for 53 ± 4% (MxB: 86 ± 3 molecules; MxC: 93 ± 9 molecules; MxD: 50 ± 15 molecules; n = 3), whereas for QtEnc 51 ± 1% of total protein belonged to the shell and 49 ± 1% (231 ± 5 molecules; n = 3) to the QtIMEF cargo (Figure 1d; Supplementary Table 2).
RESULTS AND DISCUSSION Expression, Assembly, and Cargo Loading of QtEnc in Mammalian Cells. To express the encapsulin genes from Qt in comparison to Mx in mammalian cells, we used mammalian expression constructs for the shell proteins B
DOI: 10.1021/acsnano.9b03140 ACS Nano XXXX, XXX, XXX−XXX
Article
ACS Nano
Figure 2. Cryo-EM reconstruction of QtEnc reveals icosahedral T = 4 symmetry. (a) Segmented electron density of QtEnc+QtIMEF purified from mammalian cells. The four different monomer conformations are colored according to their interconnectivity and position. The 5-fold centers are on opposite sides of a 3-fold center with two monomers in between, which indicates a T = 4 icosahedral symmetry of the shell. Boxes show zoomed-in views of the 5-fold, 3-fold, and 2-fold symmetry centers. The resolution of the map is 6 Å; scale bars represent 2 nm. (b) Cutaway view through the maximum diameter of QtEnc (∼43 nm) showing the shell (radially color-coded) and coexpressed QtIMEF cargo (violet) at different electron densities. A gap of 2.5 nm is apparent between cargo and shell.
Figure 3. QtEnc+QtIMEF biomineralizes iron more efficiently than MxEnc+MxBCD. (a, b) DAB-enhanced Prussian Blue-stained BN-PAGE (BN DAB PB) analysis of whole-cell lysates. HEK293T cells expressing MxEnc+MxBCD or QtEnc+QtIMEF were supplemented with different concentrations of ferrous ammonium sulfate (FAS) for 36 h with or without cotransfection of the iron-transporter Zip14. The upper panels show bands corresponding to the assembled encapsulin constructs (Enc), whereas the lower panels correspond to endogenous ferritin (Ft). Brown precipitates indicate iron-loaded protein complexes. In panel a, the DAB-enhancement reaction was stopped after 2 h, whereas in b the reaction was stopped after 30 min. (c) Total iron contents normalized to total protein concentration (OD280) in cell lysates of HEK293T cells (three independent biological replicates) expressing QtEnc+QtIMEF or MxEnc+MxBCD with or without Zip14 after supplementation of different FAS concentrations for 36 h. Stars indicate p values