Noise properties of reconstructed images in a kilo-voltage on-board imaging system with iterative reconstruction techniques: A phantom study

X-ray computed tomography (CT) images obtained with a kilo-voltage (kV) on-board imaging (OBI) system improve the accuracy of patient setup and treatment planning. The use of iterative reconstruction techniques (IRTs) for CT imaging can also reduce radiation dose compared to analytic reconstruction techniques. Despite these improvements, the image quality varies with IRTs, and the noise structure of reconstructed images can be distorted by IRTs. In this study, the noise properties and spatial resolution of the images reconstructed by IRTs were evaluated in terms of conventional noise metrics, high-order statistics, noise spectral density (NSD) and modulation transfer function (MTF) at different radiation doses. A kV OBI system mounted on a Varian Trilogy machine and a CATPHAN600 phantom were used to obtain projections, and the projections were reconstructed by Feldkamp (FDK), algebraic reconstruction technique (ART), maximum-likelihood expectation–maximization (MLEM) and total variation (TV) minimization algorithms. The reconstructed images were compared according to mean, standard deviation, skewness, kurtosis, NSD and MTF at different radiation doses. The results demonstrated that the noise properties and spatial resolution of reconstructed images depend on the type of IRT and the radiation dose. The noise structures are altered by IRTs and can be characterized by high-order statistics and NSD, as well as conventional noise metrics. In conclusion, high-order statistics and NSD should be considered in order to provide detailed information for the images reconstructed by IRTs. Also, trade-off among noise properties, spatial resolution and contrast is important to optimize image quality obtained using IRTs.     

Automatic particle pickup method using a neural network has high accuracy by applying an initial weight derived from eigenimages: a new reference free method for single-particle analysis

The single-particle analysis is a structure-determining method for electron microscope (EM) images which does not require crystal. In this method, the projections are picked up and averaged by the images of similar Euler angles to improve the signal to noise ratio, and then create a 3-D reconstruction. The selection of a large number of particles from the cryo-EM micrographs is a pre-requisite for obtaining a high resolution. To pickup a low-contrast cryo-EM protein image, we have recently found that a three-layer pyramidal-type neural network is successful in detecting such a faint image, which had been difficult to detect by other methods. The connection weights between the input and hidden layers, which work as a matching filter, have revealed that they reflect characters of the particle projections in the training data. The images stored in terms of the connection weights were complex, more similar to the eigenimages which are created by the principal component analysis of the learning images rather than to the averages of the particle projections. When we set the initial learning weights according to the eigenimages in advance, the learning period was able to be shortened to less than half the time of the NN whose initial weights had been set randomly. Further, the pickup accuracy increased from 90 to 98%, and a combination of the matching filters were found to work as an integrated matching filter there. The integrated filters were amazingly similar to averaged projections and can be used directly as references for further two-dimensional averaging. Therefore, this research also presents a brand-new reference-free method for single-particle analysis.     

Accelerated Optical Projection Tomography Applied to In Vivo Imaging of Zebrafish

Optical projection tomography (OPT) provides a non-invasive 3-D imaging modality that can be applied to longitudinal studies of live disease models, including in zebrafish. Current limitations include the requirement of a minimum number of angular projections for reconstruction of reasonable OPT images using filtered back projection (FBP), which is typically several hundred, leading to acquisition times of several minutes. It is highly desirable to decrease the number of required angular projections to decrease both the total acquisition time and the light dose to the sample. This is particularly important to enable longitudinal studies, which involve measurements of the same fish at different time points. In this work, we demonstrate that the use of an iterative algorithm to reconstruct sparsely sampled OPT data sets can provide useful 3-D images with 50 or fewer projections, thereby significantly decreasing the minimum acquisition time and light dose while maintaining image quality. A transgenic zebrafish embryo with fluorescent labelling of the vasculature was imaged to acquire densely sampled (800 projections) and under-sampled data sets of transmitted and fluorescence projection images. The under-sampled OPT data sets were reconstructed using an iterative total variation-based image reconstruction algorithm and compared against FBP reconstructions of the densely sampled data sets. To illustrate the potential for quantitative analysis following rapid OPT data acquisition, a Hessian-based method was applied to automatically segment the reconstructed images to select the vasculature network. Results showed that 3-D images of the zebrafish embryo and its vasculature of sufficient visual quality for quantitative analysis can be reconstructed using the iterative algorithm from only 32 projections—achieving up to 28 times improvement in imaging speed and leading to total acquisition times of a few seconds.     

Structural basis of the iron storage function of frataxin from single-particle reconstruction of the iron-loaded oligomer

The mitochondrial protein frataxin plays a central role in mitochondrial iron homeostasis, and frataxin deficiency is responsible for Friedreich ataxia, a neurodegenerative and cardiac disease that affects 1 in 40000 children. Here we present a single-particle reconstruction from cryoelectron microscopic images of iron-loaded 24-subunit oligomeric frataxin particles at 13 and 17 A resolution. Computer-aided classification of particle images showed heterogeneity in particle size, which was hypothesized to result from gradual accumulation of iron within the core structure. Thus, two reconstructions were created from two classes of particles with iron cores of different sizes. The reconstructions show the iron core of frataxin for the first time. Compared to the previous reconstruction of iron-free particles from negatively stained images, the higher resolution of the present reconstruction allowed a more reliable analysis of the overall three-dimensional structure of the 24-meric assembly. This was done after docking the X-ray structure of the frataxin trimer into the EM reconstruction. The structure revealed a close proximity of the suggested ferroxidation sites of different monomers to the site proposed to serve in iron nucleation and mineralization. The model also assigns a new role to the N-terminal helix of frataxin in controlling the channel at the 4-fold axis of the 24-subunit oligomer. The reconstructions show that, together with some common features, frataxin has several unique features which distinguish it from ferritin. These include the overall organization of the oligomers, the way they are stabilized, and the mechanisms of iron core nucleation.     

Higher order structures of Adalimumab,Infliximab and their complexes with TNFα revealed by electron microscopy

Adalimumab and Infliximab are recombinant IgG1 monoclonal antibodies (mAbs) that bind and neutralize human tumor necrosis factor alpha (TNFα). TNFα forms a stable homotrimer with unique surface‐exposed sites for Adalimumab, Infliximab, and TNF receptor binding. Here, we report the structures of Adalimumab‐TNFα and Infliximab‐TNFα complexes modeled from negative stain EM and cryo‐EM images. EM images reveal complex structures consisting of 1:1, 1:2, 2:2, and 3:2 complexes of Adalimumab‐TNFα and Infliximab‐TNFα. The 2:2 complex structures of Adalimumab‐TNFα and Infliximab‐TNFα show diamond‐shaped profiles and the 2D class averages reveal distinct orientations of the Fab domains, indicating different binding modes by Adalimumab and Infliximab to TNFα. After separation by size exclusion chromatography and analysis by negative stain EM, the 3:2 complexes of Adalimumab‐TNFα or Infliximab‐TNFα complexes are more complicated but retain features recognized in the 2:2 complexes. Preliminary cryo‐EM analysis of 3:2 Adalimumab‐TNFα complex generated a low‐resolution density consistent with a TNFα trimer bound with three Fab domains from three individual antibody molecules, while each antibody molecule binds to two molecules of TNFα trimer. The Fc domains are not visible in the reconstruction. These results show the two mAbs form structurally distinct complexes with TNFα.     

Single particle reconstructions of the transferrin-transferrin receptor complex obtained with different specimen preparation techniques

The outcome of three-dimensional (3D) reconstructions in single particle electron microscopy (EM) depends on a number of parameters. We have used the well-characterized structure of the transferrin (Tf)-transferrin receptor (TfR) complex to study how specimen preparation techniques influence the outcome of single particle EM reconstructions. The Tf-TfR complex is small (290kDa) and of low symmetry (2-fold). Angular reconstitution from images of vitrified specimens does not reliably converge on the correct structure. Random conical tilt reconstructions from negatively stained specimens are reliable, but show variable degrees of artifacts depending on the negative staining protocol. Alignment of class averages from vitrified specimens to a 3D negative stain reference model using FREALIGN largely eliminated artifacts in the resulting 3D maps, but not completely. Our results stress the need for critical evaluation of structures determined by single particle EM.     

Convolutional neural networks for reconstruction of undersampled optical projection tomography data applied to in vivo imaging of zebrafish

Optical projection tomography (OPT) is a 3D mesoscopic imaging modality that can utilize absorption or fluorescence contrast. 3D images can be rapidly reconstructed from tomographic data sets sampled with sufficient numbers of projection angles using the Radon transform, as is typically implemented with optically cleared samples of the mm‐to‐cm scale. For in vivo imaging, considerations of phototoxicity and the need to maintain animals under anesthesia typically preclude the acquisition of OPT data at a sufficient number of angles to avoid artifacts in the reconstructed images. For sparse samples, this can be addressed with iterative algorithms to reconstruct 3D images from undersampled OPT data, but the data processing times present a significant challenge for studies imaging multiple animals. We show here that convolutional neural networks (CNN) can be used in place of iterative algorithms to remove artifacts—reducing processing time for an undersampled in vivo zebrafish dataset from 77 to 15 minutes. We also show that using CNN produces reconstructions of equivalent quality to compressed sensing with 40% fewer projections. We further show that diverse training data classes, for example, ex vivo mouse tissue data, can be used for CNN‐based reconstructions of OPT data of other species including live zebrafish.      

BAmSA: Visualising transmembrane regions in protein complexes using biotinylated amphipols and electron microscopy

Membrane protein (MP) complexes play key roles in all living cells. Their structural characterisation is hampered by difficulties in purifying and crystallising them. Recent progress in electron microscopy (EM) have revolutionised the field, not only by providing higher-resolution structures for previously characterised MPs but also by yielding first glimpses into the structure of larger and more challenging complexes, such as bacterial secretion systems. However, the resolution of pioneering EM structures may be difficult and their interpretation requires clues regarding the overall organisation of the complexes. In this context, we present BAmSA, a new method for localising transmembrane (TM) regions in MP complexes, using a general procedure that allows tagging them without resorting to neither genetic nor chemical modification. Labels bound to TM regions can be visualised directly on raw negative-stain EM images, on class averages, or on three-dimensional reconstructions, providing a novel strategy to explore the organisation of MP complexes.     

Automatic CTF correction for single particles based upon multivariate statistical analysis of individual power spectra

Three-dimensional electron cryomicroscopy of randomly oriented single particles is a method that is suitable for the determination of three-dimensional structures of macromolecular complexes at molecular resolution. However, the electron-microscopical projection images are modulated by a contrast transfer function (CTF) that prevents the calculation of three-dimensional reconstructions of biological complexes at high resolution from uncorrected images. We describe here an automated method for the accurate determination and correction of the CTF parameters defocus, twofold astigmatism and amplitude-contrast proportion from single-particle images. At the same time, the method allows the frequency-dependent signal decrease (B factor) and the non-convoluted background signal to be estimated. The method involves the classification of the power spectra of single-particle images into groups with similar CTF parameters; this is done by multivariate statistical analysis (MSA) and hierarchically ascending classification (HAC). Averaging over several power spectra generates class averages with enhanced signal-to-noise ratios. The correct CTF parameters can be deduced from these class averages by applying an iterative correlation procedure with theoretical CTF functions; they are then used to correct the raw images. Furthermore, the method enables the tilt axis of the sample holder to be determined and allows the elimination of individual poor-quality images that show high drift or charging effects.     

Serial section scanning electron microscopy (S3EM) on silicon wafers for ultra-structural volume imaging of cells and tissues

High resolution, three-dimensional (3D) representations of cellular ultrastructure are essential for structure function studies in all areas of cell biology. While limited subcellular volumes have been routinely examined using serial section transmission electron microscopy (ssTEM), complete ultrastructural reconstructions of large volumes, entire cells or even tissue are difficult to achieve using ssTEM. Here, we introduce a novel approach combining serial sectioning of tissue with scanning electron microscopy (SEM) using a conductive silicon wafer as a support. Ribbons containing hundreds of 35 nm thick sections can be generated and imaged on the wafer at a lateral pixel resolution of 3.7 nm by recording the backscattered electrons with the in-lens detector of the SEM. The resulting electron micrographs are qualitatively comparable to those obtained by conventional TEM. S(3)EM images of the same region of interest in consecutive sections can be used for 3D reconstructions of large structures. We demonstrate the potential of this approach by reconstructing a 31.7 μm(3) volume of a calyx of Held presynaptic terminal. The approach introduced here, Serial Section SEM (S(3)EM), for the first time provides the possibility to obtain 3D ultrastructure of large volumes with high resolution and to selectively and repetitively home in on structures of interest. S(3)EM accelerates process duration, is amenable to full automation and can be implemented with standard instrumentation.     

Three-dimensional spectral signal-to-noise ratio for a class of reconstruction algorithms

A three-dimensional (3D) version of the spectral signal-to-noise ratio (SSNR)-based resolution measure is introduced. The measure is defined for a class of 3D reconstruction algorithms that use interpolation in Fourier space. The statistical properties of the SSNR are discussed and related to the properties of another resolution measure, the Fourier shell correlation (FSC). The new measure was tested on 3D structures calculated from a simulated set of quasi-evenly spaced 2D projections using a nearest-neighbor interpolation and a gridding algorithm. In the latter case, the results agree very well with the FSC-based estimate, with the exception of very high SSNR values. The main applicability of the 3D SSNR is tomography, where due to the small number of projections collected, FSC cannot be used. The new measure was applied to three sets of tomographic data. It was demonstrated that the measure is sufficiently sensitive to yield theoretically expected results. Therefore, the 3D SSNR opens up the possibility of evaluating the quality of tomographic reconstructions in an objective manner. The 3D distribution of SSNR is of major interest in single-particle analysis. It is shown that the new measure can be used to evaluate the anisotropy of 3D reconstructions. The distribution of SSNR is characterized by three anisotropy indices derived from principal axes of the 3D inertia covariance matrix of the SSNR. These indices are used to construct a 3D Fourier filter which, when applied to a 3D reconstruction of a macromolecule, maximizes the SNR in real space and minimizes real-space artifacts caused by uneven distribution of 2D projections.     

A fully automatic 3D reconstruction method using simulated annealing enables accurate posterioric angular assignment of protein projections

Single-particle analysis is a structure determining method using electron microscopic (EM) images, which does not require protein crystal. In this method, projections are picked up and used to reconstruct a three-dimensional (3D) structure. When the conical tilting method is not available, the particle images are usually classified and averaged to improve the signal-to-noise ratio. The Euler angles of these average images must be posteriorically assigned to create a primary 3D model. We developed a new, fully automatic unsupervised Euler angle assignment method, which does not require an initial 3D reference and which is applicable to asymmetric molecules. In this method, the Euler angle of each average image is initially set randomly and then automatically corrected in relation to those of the other averages by iterated optimizations using the Simulated Annealing (SA) algorithm. At each iteration, the 3D structure is reconstructed based on the current Euler angles and reprojected back in the average-input directions. A modified cross-correlation between each reprojection and its corresponding original average is then calculated. The correlations are summed as a total 3D echo-correlation score to evaluate the Euler angles at this iteration. Then, one of the projections is selected, its Euler angle is changed randomly, and the score is also calculated. Based on the score change, judgment of whether to accept or reject the new angle is made using the SA algorithm, which is introduced to overcome the local minimums. After a certain number of iterations of this process, the angles of all averages converge so as to create a reliable primary 3D model. This echo-correlated 3D reconstruction with simulated annealing also has potential for wide application to general 3D reconstruction from various types of 2D images.     

Single particle reconstruction of membrane proteins: A tool for understanding the 3D structure of disease-related macromolecules

Membrane proteins play important roles in cell functions such as neurotransmission, muscle contraction, and hormone secretion, but their structures are mostly undetermined. Several techniques have been developed to elucidate the structure of macromolecules; X-ray or electron crystallography, nuclear magnetic resonance spectroscopy, and high-resolution electron microscopy. Electron microscopy-based single particle reconstruction, a computer-aided structure determination method, reconstructs a three-dimensional (3D) structure from projections of monodispersed protein. A large number of particle images are picked up from EM films, aligned and classified to generate two-dimensional (2D) averages, and, using the Euler angle of each 2D average, reconstructed into a 3D structure. This method is challenging due to the necessity for close collaboration between classical biochemistry and innovative information technology, including parallel computing. However, recent progress in electron microscopy, mathematical algorithms, and computational ability has greatly increased the subjects that are considered to be primarily addressable using single particle reconstruction. Membrane proteins are one of these targets to which the single particle reconstruction is successfully applied for understanding of their structures. In this paper, we will introduce recently reconstructed channel-related proteins and discuss the applicability of this technique in understanding molecular structures and their roles in pathology.     

Three-Dimensional Histology Volume Reconstruction of Axonal Tract Tracing Data: Exploring Topographical Organization in Subcortical Projections from Rat Barrel Cortex

Topographical organization is a hallmark of the mammalian brain, and the spatial organization of axonal connections in different brain regions provides a structural framework accommodating specific patterns of neural activity. The presence, amount, and spatial distribution of axonal connections are typically studied in tract tracing experiments in which axons or neurons are labeled and examined in histological sections. Three-dimensional (3-D) reconstruction techniques are used to achieve more complete visualization and improved understanding of complex topographical relationships. 3-D reconstruction approaches based on manually or semi-automatically recorded spatial points representing axonal labeling have been successfully applied for investigation of smaller brain regions, but are not practically feasible for whole-brain analysis of multiple regions. We here reconstruct serial histological images from four whole brains (originally acquired for conventional microscopic analysis) into volumetric images that are spatially registered to a 3-D atlas template. The aims were firstly to evaluate the quality of the 3-D reconstructions and the usefulness of the approach, and secondly to investigate axonal projection patterns and topographical organization in rat corticostriatal and corticothalamic pathways. We demonstrate that even with the limitations of the original routine histological material, the 3-D reconstructed volumetric images allow efficient visualization of tracer injection sites and axonal labeling, facilitating detection of spatial distributions and across-case comparisons. Our results further show that clusters of S1 corticostriatal and corticothalamic projections are distributed within narrow, elongated or spherical subspaces extending across the entire striatum / thalamus. We conclude that histology volume reconstructions facilitate mapping of spatial distribution patterns and topographical organization. The reconstructed image volumes are shared via the Rodent Brain Workbench (www.rbwb.org).     

Quaternary organization of the Staphylothermus marinus phosphoenolpyruvate synthase: angular reconstitution from cryoelectron micrographs with molecular modeling

Digital electron images of frozen-hydrated preparations of the 2.25-MDa Staphylothermus marinus phosphoenolpyruvate synthase (EC 2.7.9.2) have been analyzed by single-particle classification and averaging and iterative quaternion-based angular reconstitution. Contrast transfer function correction of micrographs obtained at different defocus values was used to improve the informational quality of the projection averages. Three-dimensional reconstructions were obtained to roughly 3-nm spatial resolution, in which the 24 identical subunits were arranged to form an octahedral complex, although the amino-terminal nucleotide-binding domain was not resolved. An atomic model of the subunit was generated by homology modeling using as the reference the known X-ray crystallographic structure of the related enzyme pyruvate orthophosphate dikinase (EC 2.7.9.1) from Clostridium symbiosum (Protein Data Bank entry 1DIK). The S. marinus protein could be arranged into an assembly of 12 homodimers to match the three-dimensional reconstruction in terms of shape and size of the homodimers, as well as overall shape and size of the complex. The quaternary model indicated that active sites of three monomers were localized around cavities (or putative channels) centered at the threefold axes of rotational symmetry and that carboxyl-terminal alpha-helical segments of four monomers were localized at the fourfold axes of rotational symmetry where they could facilitate interdimer interaction. The quaternary arrangement also indicated numerous potential hydrophobic and electrostatic interactions at the interdimer interfaces that could contribute further to structural stability.     

Electron microscopic structural analysis of Photosystem I,Photosystem II,and the cytochromeb6/f complex from green plants and cyanobacteria

Electron microscopy (EM) in combination with image analysis is a powerful technique to study protein structure at low- and high resolution. Since electron micrographs of biological objects are very noisy, substantial improvement of image quality can be obtained by averaging individual projections. Crystallographic and noncrystallographic averaging methods are available and have been applied to study projections of the large protein complexes embedded in photosynthetic membranes from cyanobacteria and higher plants. Results of EM on monomeric and trimeric Photosystem I complexes, on monomeric and dimeric Photosystem II complexes, and on the monomeric cytochromeb6/f complex are discussed.     

Iterative reconstruction of cryo-electron tomograms using nonuniform fast Fourier transforms

Algorithms for three-dimensional (3D) reconstruction of objects based on their projections are essential in various biological and medical imaging modalities. In cryo-electron tomography (CET) a major challenge for reconstruction is the limited range of projection angles, which manifests itself as a “missing wedge” of data in Fourier space making the reconstruction problem ill-posed. Here, we apply an iterative reconstruction method that makes use of nonuniform fast Fourier transform (NUFFT) to the reconstruction of cryo-electron tomograms. According to several measures the reconstructions are superior to those obtained using conventional methods, most notably weighted backprojection. Most importantly, we show that it is possible to fill in partially the unsampled region in Fourier space with meaningful information without making assumptions about the data or applying prior knowledge. As a consequence, particles of known structure can be localized with higher confidence in cryotomograms and subtomogram averaging yields higher resolution densities.     

A Fast Image Alignment Approach for 2D Classification of Cryo-EM Images Using Spectral Clustering

Three-dimensional (3D) reconstruction in single-particle cryo-electron microscopy (cryo-EM) is a significant technique for recovering the 3D structure of proteins or other biological macromolecules from their two-dimensional (2D) noisy projection images taken from unknown random directions. Class averaging in single-particle cryo-EM is an important procedure for producing high-quality initial 3D structures, where image alignment is a fundamental step. In this paper, an efficient image alignment algorithm using 2D interpolation in the frequency domain of images is proposed to improve the estimation accuracy of alignment parameters of rotation angles and translational shifts between the two projection images, which can obtain subpixel and subangle accuracy. The proposed algorithm firstly uses the Fourier transform of two projection images to calculate a discrete cross-correlation matrix and then performs the 2D interpolation around the maximum value in the cross-correlation matrix. The alignment parameters are directly determined according to the position of the maximum value in the cross-correlation matrix after interpolation. Furthermore, the proposed image alignment algorithm and a spectral clustering algorithm are used to compute class averages for single-particle 3D reconstruction. The proposed image alignment algorithm is firstly tested on a Lena image and two cryo-EM datasets. Results show that the proposed image alignment algorithm can estimate the alignment parameters accurately and efficiently. The proposed method is also used to reconstruct preliminary 3D structures from a simulated cryo-EM dataset and a real cryo-EM dataset and to compare them with RELION. Experimental results show that the proposed method can obtain more high-quality class averages than RELION and can obtain higher reconstruction resolution than RELION even without iteration.     

The effect of overabundant projection directions on 3D reconstruction algorithms

The experimental process of collecting images from macromolecules in an electron microscope is such that it does not allow for prior specification of the angular distribution of the projection images. As a consequence, an uneven distribution of projection directions may occur. Concerns have been raised recently about the behavior of 3D reconstruction algorithms for the case of unevenly distributed projections. It has been illustrated on experimental data that in the case of a heavily uneven distribution of projection directions some algorithms tend to elongate the reconstructed volumes along the overloaded direction so much as to make a quantitative biological analysis impossible. In answer to these concerns we have developed a strategy for quantitative comparison and optimization of 3D reconstruction algorithms. We apply this strategy to quantitatively analyze algebraic reconstruction techniques (ART) with blobs, simultaneous iterative reconstruction techniques (SIRT) with voxels, and weighted backprojection (WBP). We show that the elongation artifacts that had been previously reported can be strongly reduced. With our specific choices for the free parameters of the three algorithms, WBP reconstructions tend to be inferior to those obtained with either SIRT or ART and the results obtained with ART are comparable to those with SIRT, but at a very small fraction of the computational cost of SIRT.     

An atomic model of the thin filament in the relaxed and Ca2+-activated states

Contraction of striated muscles is regulated by tropomyosin strands that run continuously along actin-containing thin filaments. Tropomyosin blocks myosin-binding sites on actin in resting muscle and unblocks them during Ca2+-activation. This steric effect controls myosin-crossbridge cycling on actin that drives contraction. Troponin, bound to the thin filaments, couples Ca2+-concentration changes to the movement of tropomyosin. Ca2+-free troponin is thought to trap tropomyosin in the myosin-blocking position, while this constraint is released after Ca2+-binding. Although the location and movements of tropomyosin are well known, the structural organization of troponin on thin filaments is not. Its mechanism of action therefore remains uncertain. To determine the organization of troponin on the thin filament, we have constructed atomic models of low and high-Ca2+ states based on crystal structures of actin, tropomyosin and the "core domain" of troponin, and constrained by distances between filament components and by their location in electron microscopy (EM) reconstructions. Alternative models were also built where troponin was systematically repositioned or reoriented on actin. The accuracy of the different models was evaluated by determining how well they corresponded to EM images. While the initial low and high-Ca2+ models fitted the data precisely, the alternatives did not, suggesting that the starting models best represented the correct structures. Thin filament reconstructions were generated from the EM data using these starting models as references. In addition to showing the core domain of troponin, the reconstructions showed additional detail not present in the starting models. We attribute this to an extension of TnI linking the troponin core domain to actin at low (but not at high) Ca2+, thereby trapping tropomyosin in the OFF-state. The bulk of the core domain of troponin appears not to move significantly on actin, regardless of Ca2+ level. Our observations suggest a simple model for muscle regulation in which troponin affects the charge balance on actin and hence tropomyosin position.