Original research
Super-resolution fluorescence imaging to study cardiac biophysics: α-actinin distribution and Z-disk topologies in optically thick cardiac tissue slices

https://doi.org/10.1016/j.pbiomolbio.2014.07.003Get rights and content

Abstract

A major motivation for the use of super-resolution imaging methods in the investigation of cardiac biophysics has been the insight from biophysical considerations and detailed mathematical modeling that the spatial structure and protein organisation at the scale of nanometres can have enormous implications for calcium signalling in cardiac muscle. We illustrate the use of dSTORM based super-resolution in optically thick (∼10 μm) tissue slices of rat ventricular tissue to visualize proteins at the cardiac Z-disk and compare those images with confocal (diffraction-limited) as well as electron microscopy (EM) data which still provides a benchmark in terms of resolution. α-actinin is an abundant protein target that effectively defines the Z-disk in striated muscle and provides a reference structure for other proteins at the Z-line and the transverse tubules. Using super-resolution imaging α-actinin labelling provides very detailed outlines of the contractile machinery which we have used to study the properties of Z-disks and the distribution of α-actinin itself. We determined the local diameters of the myo-fibrillar and non-myofibrillar space using α-actinin labelling. Comparison between confocal and super-resolution based myofibrillar masks suggested that super-resolution data was able to segment myofibrils accurately while confocal approaches were not always able to distinguish neighbouring myofibrillar bundles which resulted in overestimated diameters. The increased resolution of super-resolution methods provides qualitatively new information to improve our understanding of cardiac biophysics. Nevertheless, conventional diffraction-limited imaging still has an important role to play which we illustrate with correlative confocal and super-resolution data.

Introduction

Progress in cardiac biophysics has often been closely linked with methodological advances, in particular with improvements in fluorescence imaging techniques. The introduction of fluorescent Ca2+ indicators with suitable affinity and rapid kinetics (Grynkiewicz et al., 1985) has contributed greatly to the study of calcium fluxes. To date such studies are arguably most advanced in muscle, and perhaps specifically in cardiac muscle. Similarly, the insight that the view of the cardiac myocyte cytosol as a well-stirred compartment in which Ca2+ is essentially uniform is problematic had first been suggested on mostly theoretical grounds (Stern, 1992) but the crucial breakthrough was the discovery of microscopic Ca2+ release events (Cheng et al., 1993). The technical advancement that enabled this discovery was the introduction of confocal microscopy to biomedical research as well as the availability of indicators from the Fluo-3 family (Minta et al., 1989) which are distinguished by a very large modulation of fluorescence with increasing Ca2+ (albeit at the cost of not being ratiometric). The resulting understanding has now led to a fairly well-developed theory of local control (for a review, see e.g. (Cannell and Kong, 2011)) which emphasizes the role of gradients in Ca2+ throughout the cell and the existence of local signalling domains. Such local control enables the cardiac muscle cell to work as a graded system based on calcium-induced calcium release (Fabiato, 1983) that is both robust and exhibits high amplification of a trans-sarcomeric trigger Ca2+ influx to allow intricate control of the cardiac Ca2+ transient, the primary modulator of cardiac contractility at the cellular level.

The local control view of cardiac EC coupling recognises the primary importance of spatial organisation and reveals the important role of the dyadic junctions between the surface membrane (and its extensions, the transverse (t-) tubules) and the membranes of the terminal sarcoplasmic reticulum (SR). At junctions the membranes oppose each other and the gap is only ∼15 nm in width (Fawcett and McNutt, 1969) which is critical to confining the signalling between surface L-type Ca2+ channels and ryanodine receptors (RyRs) which are the SR Ca2+ release channels. In the junctional signalling space Ca2+ concentration reaches peak levels that are several orders of magnitude higher than in the bulk cytosol. Junctions may also be small in lateral extent, electron tomography suggests complex 3D arrangements with few tens of nanometre extent along the shorter axes (Hayashi et al., 2009). In addition, the t-tubules themselves are sub-resolution in diameter (∼50–250 nm in the rat (Soeller and Cannell, 1999)) and the clusters of RyRs that are thought to underlie the microscopic Ca2+ sparks may be as small as ∼30 nm along their shortest axis (Baddeley et al., 2009, Hayashi et al., 2009) (but large clusters can reach micrometer sizes). All these structures that are part of the cardiac Ca2+ local signalling apparatus are “nano-objects” and generally not resolved by diffraction-limited imaging approaches, such as confocal microscopy (with a “best” lateral resolution of ∼200–250 nm). Accordingly there has been considerable interest in utilizing higher resolution methods, ideally compatible with fluorescence stains and indicators due to their high contrast and specificity.

This gap in our imaging capabilities has recently been filled with the advent of so called optical super-resolution microscopy, of which there are a number of different ‘flavours’. STED (or stimulated emission depletion) microscopy, uses a saturable fluorescence depletion effect to selectively darken all molecules but those within the centre of a laser spot that is scanned through the sample similar to confocal microscopy (Hell and Wichmann, 1994). Single molecule localisation microscopy, by contrast, is a widefield technique which relies on darkening all but a few fluorophores at any given time (Betzig et al., 2006, Hess et al., 2006, Rust et al., 2006). This allows the few fluorophores that are “bright” to be observed as isolated, diffraction limited spots whose centre can be localised with an accuracy that is only limited by the number of photons that we can collect (and the brightness of the background). By detecting and localising subsets of fluorochromes through typically thousands of frames a composite image can be constructed and rendered into a format familiar to conventional fluorescence micrographs using algorithms to estimate local marker density (Baddeley et al., 2010). These methods are known under a variety of acronyms such as PALM, fPALM, STORM etc and achieve lateral resolutions of ∼20–30 nm. In one variation, known as dSTORM (direct STORM (Heilemann et al., 2008)), conventional fluorochromes become photoswitchable under suitable conditions and this approach is particularly practical as normal imaging and labelling protocols can be followed to a large extent. We have previously extended dSTORM for multi-colour 3D super-resolution imaging, a method that we termed d4STORM (Baddeley et al., 2011). Several studies employing super-resolution imaging for the study of cardiac myocytes have clearly demonstrated its utility for studies of cardiac biophysics which have recently been reviewed (Soeller and Baddeley, 2013). An important structural component of cardiac myocytes (and other striated muscle cells) are the boundaries of sarcomeres called Z-lines (as they appear as lines in 2D micrographs) or, taking the full 3D structure of myofibrills into account, Z-disks, as in 3D the sarcomere boundaries are thin disks coinciding with the cross-section of myofibrillar bundles. A number of protein involved in excitation-contraction (EC) coupling are located around the edges of Z-disks such as RyRs and other junctional proteins (Jayasinghe et al., 2009, Soeller et al., 2007) as well as the t-tubules that are largely found in close proximity to the edges of myofibrillar Z-disks. One protein that is greatly concentrated within Z-disks and serves as a distinctive Z-disk marker is α-actinin. α-actinin anchors thin filaments at the Z-line and interactions with the sarcomeric protein titin and actin in thin filaments are critical for the development and maintenance of normal sarcomeric structure (Sjöblom et al., 2008).

We have previously demonstrated the use of super-resolution microscopy to achieve improved contrast and resolution in optically thick tissue sections (Baddeley et al., 2011). In the present study we provide a more extended study of tissue-section based dSTORM imaging to resolve the details of myofibrillar Z-disks and the distribution of the protein α-actinin. Detailed comparison of super-resolution images with both diffraction-limited widefield and confocal data as well as electron micrographs reveals the resolution and associated contrast improvement but also emphasizes remaining limitations. Our results support the use of super-resolution microscopy for nanoscale investigations of structures and protein distributions that underlie the biophysics of cardiac muscle, in particular Ca2+ handling and production of force.

Section snippets

Tissue preparation

All use of rats and rat tissue was approved by the Auckland University ethics committee prior to experimentation. Healthy adult Wistar rats in the weight range of 250–300 g were sacrificed by lethal injection of sodium pentabarbitone followed by cervical dislocation. The heart was removed and mounted on a Langendorff perfusion apparatus and perfused with 2% paraformaldehyde in phosphate buffered saline (PBS). Upon removal from the perfusion setup, the atria were first removed and the

Results

α-actinin is an actin-binding protein that is present in high density at the boundaries between sarcomeres, the Z-disks. It therefore is a prominent marker of myofibrillar bundle structure when viewed in cross-section. This can be seen in Fig. 1 which shows transverse (cross-sectional) views of rat cardiac ventricular myocytes in three different imaging modalities. In a thin section electron micrograph myofibrils (and mitochondria) can be readily identified in cross-section (Fig. 1A). The

Super-resolution imaging of alpha-actinin in Z-disks of cardiac myocytes

In this study we have focused on the use of dSTORM based tissue imaging of α-actinin. dSTORM, as one of the single-molecule localisation based super-resolution approaches relies on the detection of fluorescence bursts from single (or small groups of) molecules. The achievement of sufficient contrast to reliably detect these flashes or “blinks” in optically thick tissue sections is a challenge but can be overcome in a number of ways. Here we used a highly inclined illumination geometry (HILO, (

Conclusion

The recent introduction of practical super-resolution imaging techniques has enabled the observation of features of cardiac myocytes that were previously difficult to assess with conventional techniques. This includes features on the scale between ∼30 nm and 200 nm where cardiac myocytes have a rich subcellular structure including RyR clusters, t-tubules and details of their sarcomeric makeup. Here we demonstrated this with data on α-actinin and the structure of Z-disks. There is a wealth of

Editors’ note

Please see also related communications in this issue by Wang et al. (2014) and Vostarek et al. (2014).

Acknowledgements

We thank Dr. Masa Hoshijima, UCSD, for providing a slice of a mouse myocyte EM tomogram. Support by the Marsden Fund, the Health Research Council, the Lottery Health Board and The Human Frontier Science Program are gratefully acknowledged.

References (56)

  • S.M. Manisastry et al.

    Myofibril assembly visualized by imaging N-RAP, alpha-actinin, and actin in living cardiomyocytes

    Exp. Cell Res.

    (2009)
  • A. Minta et al.

    Fluorescent indicators for cytosolic calcium based on rhodamine and fluorescein chromophores

    J. Biol. Chem.

    (1989)
  • L.D. Peachey et al.

    Helicoids in the T system and striations of frog skeletal muscle fibers seen by high voltage electron microscopy

    Biophys. J.

    (1978)
  • C. Soeller et al.

    Super-resolution imaging of EC coupling protein distribution in the heart

    J. Mol. Cell. Cardiol.

    (2013)
  • M.D. Stern

    Theory of excitation-contraction coupling in cardiac muscle

    Biophys. J.

    (1992)
  • F. Vostarek et al.

    Studying dynamic events in the developing myocardium

    Prog. Bio. Mol. Biol.

    (2014)
  • K. Wang et al.

    Living cardiac tissue slices: An organotypic pseudo two-dimensional model for cardiac biophysics research

    Prog. Bio. Mol. Biol.

    (2014)
  • J. Wong et al.

    Nanoscale distribution of ryanodine receptors and caveolin-3 in mouse ventricular myocytes: dilation of t-tubules near junctions

    Biophys. J.

    (2013)
  • D. Baddeley et al.

    Visualization of localization microscopy data

    Microsc. Microanal.

    (2010)
  • D. Baddeley et al.

    4D super-resolution microscopy with conventional fluorophores and single wavelength excitation in optically thick cells and tissues

    PLoS One

    (2011)
  • D. Baddeley et al.

    Optical single-channel resolution imaging of the ryanodine receptor distribution in rat cardiac myocytes

    Proc. Natl. Acad. Sci. U S A

    (2009)
  • E. Betzig et al.

    Imaging intracellular fluorescent proteins at nanometer resolution

    Science

    (2006)
  • M.B. Cannell et al.

    Image enhancement by deconvolution

  • M.B. Cannell et al.

    Local control in cardiac E-C coupling

    J. Mol. Cell. Cardiol.

    (2011)
  • H. Cheng et al.

    Loss of enigma homolog protein results in dilated cardiomyopathy

    Circ. Res.

    (2010)
  • H. Cheng et al.

    Calcium sparks: elementary events underlying excitation-contraction coupling in heart muscle

    Science

    (1993)
  • D.J. Crossman et al.

    Changes in the organization of excitation-contraction coupling structures in failing human heart

    PLoS One

    (2011)
  • A. Fabiato

    Calcium-induced release of calcium from the cardiac sarcoplasmic reticulum

    Am. J. Physiol.

    (1983)
  • Cited by (0)

    View full text