The evolution of structured illumination microscopy in studies of HIV
Introduction
Structured illumination microscopy (SIM) uses the interference of patterned light to enhance resolution. The use of light patterns to improve resolution was first proposed in 1963 by Lukosz and Marchand [1]. The practical implementation of this theory was pioneered by Tony Wilson in the 1990’s by developing use of light patterns for improving the resolution of images by optical sectioning [2]. Commercial applications of this work using were the Apotome, Vivatome (Carl Zeiss), DSD (Oxford Instruments) and Carv (BD Biosciences). Use of patterned light to improve bioimaging was further developed by Sedat, Gustafsson and Heintzmann in the 2000’s [3], [4], [5]. Gustafsson et al. carried out the pioneering work needed to enhance resolution beyond the Abbe limit first in the x,y plane (laterally) and later axially [4], [5].
To improve resolution improvement in 3D biological samples; Gustaffson placed a complex grid in the excitation light path of an incoherent light source (N.B in reality this was a laser source which was scrambled to make the light incoherent) [4] (Fig. 1A). This generates a diffraction pattern on the sample. The interaction of resulting patterned light and the biological sample generated high frequency interference, termed Moiré effect. These interference patterns have a periodicity much greater than that of the original pattern and contain high frequency information describing fine structures in the sample (Fig. 1B). If the grating is rotated and translated to give three different Moiré patterns (Fig. 1C), the high spatial frequency information (fine detail) can then be algorithmically extracted and shifted to their true location in the frequency domain [4]. This computation yields an image reconstructed from Moiré patterns which has a twofold improvement in spatial resolution compared to the raw data (Fig. 1D). In practise more Moiré patterns are needed to acquire SIM sample data. This is because a single raw data image is a sum of several different high spatial frequency information components. For data restoration these information components need to be separated. This can be done by acquiring additional data sets with different phases of light [5]. For 3D SIM the sample was illuminated with three separate beams and Moiré patterns generated from three rotations and five phase shifts were needed to computationally determine the high resolution structure [5], [6] (Fig. 1C). 3D SIM allowed biological samples, to be imaged at a 1.8 fold resolution improvement over the Abbe limit axially and laterally [6], [7] (Fig. 1C). Proof of principle work with 3D SIM imaging in the cell nucleus revealed chromatin was more fibrous than observed by confocal microscopy and holes in DAPI staining at the nuclear periphery which could not be resolved by light microscopy, showing DNA was excluded from nuclear pore complexes [6].
SIM methodology demonstrated by Gustafsson and Schmellerah was further developed for applications in research and commercialisation by Prof. J.H. Sedat, Prof. M. Gustafsson and more recently Prof. R. Heintzman [3], [4], [5]. The first commercially available SIM system was brought to the market by Applied Precision in 2008 and termed OMX [7]. Wicker and Heintzmann further developed the Gustafsson model in 2010 by using five rotations and five phases and a projected pattern of light generated by a spatial light modulator rather than a grid. This gave an improvement of the isotropy of resolution which meant in a perfect sample a resolution 100 nm could be achieved ([3] and K. Wicker, “Increasing resolution and light efficiency in light microscopy,” PhD thesis, King’s College London, U.K. (2010)). Following this other have published algorithms which improve image analysis for SIM [3], [12], [13], [14] and describe how to analyse and interpret SIM data (http://www.micron.ox.ac.uk/software/SIMCheck.php).
Human immunodeficiency virus was responsible for 1.6 million deaths worldwide in 2012 (http://apps.who.int/gho/data/node.main.623) with the majority of these being concentrated in Africa, India and Thailand. In order to infect cells Human immunodeficiency virus (HIV) binds to cell surface expressed CD4 and a chemokine co-receptor, either CCR5 or CXCR4, to induce fusion at the plasma membrane. Once fusion has occurred the viral capsid (CA), which contains the viral genome, is delivered into the cell cytoplasm. HIV viral particles are smaller than the 200 nm resolution limit of standard light microscopy methods, such as confocal imaging [8], [9]. Therefore the precise details of the events between CA delivery and localization to the nuclear pore are elusive and it is difficult to study, using light microscopy, how and where the virus acts in the cell. Recent studies using atomic force microscopy of HIV particles sets viral size as varying between 80 and 146 nm [10]. After fusion at the cellular membrane HIV viral CA (CA) are released into the cellular cytoplasm. To study early events it is useful to carry out studies in 3D since viral CA can be in any cellular location between the plasma membrane and the nuclear membrane, so the whole cell needs to be imaged to capture all of the CA [11]. Further it is preferable to carry out studies using instrumentation which can observe viral particles inside the whole cell at a resolution similar to that of the capsid size of between 110 and 146 nm. 3D SIM, with its improved resolution and ability to acquire a cell in 3D presents an excellent methodology for this [8].
In this study we use imaging of HIV CA post-entry of virus to document the evolution of the SIM technique. The equipment used is representative of what was available at the time however our industrial partners are continually bringing new products to market. We would advise interested parties to evaluate what is currently available on the market to inform decisions about equipment. Here we present a technical summary comparing the instrumentation which has been available over the past 5 years for these studies.
Section snippets
Sample preparation
The physical principles behind SIM assume that the sample is flat or fairly flat and that the only cause of diffraction of light will be from the sample caused by labelling of the epitope of interest. This means that not every biological sample is appropriate for SIM imaging, for instance tissue samples may not perform as well without clearing and careful sectioning. For an optimal resolution enhancement the sample must ideally be mounted on high precision coverslips (0.17 mm +/− 0.05 mm, e.g.
Data reconstruction and analysis
All of the raw data were acquired in the manufacturers proprietary format and reconstructed using the software provided by the manufacturer. Each of the systems used were able to acquire a widefield image as well as a SIM image and the software for each allowed acquisition of this. The Elyra system was part of a multimodal SIM/STORM/confocal system so a point scanned confocal image could also be obtained for comparative studies. One of the major points of difficulty of with SIM systems is that
Conclusions
Structured illumination microscopy is a tool which is of great use to the field of HIV research. It allows visualisation of viral particles smaller than have been detected previously and allows more accurate information about spatial positioning and size of the molecules to be identified. As modern implementations of SIM have more laser lines it would be possible to expand this study to include more markers of HIV as well as the nuclear staining to facilitate a more confident interpretation of
Acknowledgements
We thank Dr. Daniel Metcalf, Nikon Instruments and Dr. Nicolas Sergeant, Carl Zeiss Microscopy for their excellent help, advice and discussion in the preparation of this manuscript. The Nikon Imaging Centre Kings College London for provision of scientific advice. Professor Jason Swedlow and Dr. Ian Dobbie for advice. This work was funded by: QMUL Molecular Cellular Medicine research theme, EuroBioimaging and a MRC Next generation Optical Microscopy grant (ESRIC).
References (14)
Biophys. J.
(2008)- et al.
Opt. Acta
(1963) - et al.
Opt. Lett.
(1997) - et al.
Opt. Express
(2013) J. Microsc.
(2000)Science
(2008)Cold Spring Harb. Protoc.
(2011)