Spatial quantitative analysis of fluorescently labeled nuclear structures: problems, methods, pitfalls.

The vast majority of microscopic data in biology of the cell nucleus is currently collected using fluorescence microscopy, and most of these data are subsequently subjected to quantitative analysis. The analysis process unites a number of steps, from image acquisition to statistics, and at each of these steps decisions must be made that may crucially affect the conclusions of the whole study. This often presents a really serious problem because the researcher is typically a biologist, while the decisions to be taken require expertise in the fields of physics, computer image analysis, and statistics. The researcher has to choose between multiple options for data collection, numerous programs for preprocessing and processing of images, and a number of statistical approaches. Written for biologists, this article discusses some of the typical problems and errors that should be avoided. The article was prepared by a team uniting expertise in biology, microscopy, image analysis, and statistics. It considers the options a researcher has at the stages of data acquisition (choice of the microscope and acquisition settings), preprocessing (filtering, intensity normalization, deconvolution), image processing (radial distribution, clustering, co-localization, shape and orientation of objects), and statistical analysis.

Graphical representation and multicomponent analysis of single-frequency fluorescence lifetime imaging microscopy data.

Graphical representation of fluorescence lifetime imaging microscopy data demonstrates that a mixture of two components with single exponential decays can be resolved by single frequency measurements. We derive a method based on linear fitting that allows the calculation of the fluorescence lifetimes of the two components. We show that introduction of proper error-weighting results in a non-linear method that is mathematically identical to a global analysis algorithm that was recently derived. The graphical approach was applied to cellular data obtained from a lifetime-based phosphorylation assay for the epidermal growth factor receptor and yielded results similar to those obtained by a global analysis algorithm.

Evaluation of global analysis algorithms for single frequency fluorescence lifetime imaging microscopy data.

Global analysis of fluorescence lifetime image microscopy (FLIM) data can be used to obtain an accurate fit of multi-exponential fluorescence decays. In particular, it can be used to fit a bi-exponential decay to single frequency FLIM data, which is not possible with conventional fitting techniques. Bi-exponential fluorescence decay models can be used to analyse quantitatively single frequency FLIM data from samples that exhibit fluorescence resonance energy transfer (FRET). Global analysis algorithms simultaneously fit multiple measurements acquired under different experimental conditions to achieve higher accuracy. To demonstrate that bi-exponential models can indeed be fitted to single frequency data, we derive an analytical solution for the special case of two measurements and use this solution to illustrate the properties of global analysis algorithms. We also derive a novel global analysis algorithm that is optimized for single frequency FLIM data, and demonstrate that it is superior to earlier algorithms in terms of computational requirements.

Improved spatial discrimination of protein reaction states in cells by global analysis and deconvolution of fluorescence lifetime imaging microscopy data.

The deconvolution of fluorescence lifetime imaging microscopy (FLIM) data that were processed with global analysis techniques is described. Global analysis of FLIM data enables the determination of relative numbers of molecules in different protein reaction states on a pixel-by-pixel basis in cells. The three-dimensional fluorescence distributions of each protein state can then be calculated and deconvolved. High-resolution maps of the relative concentrations of each state are then obtained from the deconvolved images. We applied these techniques to quantitatively image the phosphorylation state of ErbB1 receptors tagged with green fluorescent protein in MCF7 cells.

Ezrin is a downstream effector of trafficking PKC-integrin complexes involved in the control of cell motility.

Protein kinase C (PKC) alpha has been implicated in beta1 integrin-mediated cell migration. Stable expression of PKCalpha is shown here to enhance wound closure. This PKC-driven migratory response directly correlates with increased C-terminal threonine phosphorylation of ezrin/moesin/radixin (ERM) at the wound edge. Both the wound migratory response and ERM phosphorylation are dependent upon the catalytic function of PKC and are susceptible to inhibition by phosphatidylinositol 3-kinase blockade. Upon phorbol 12,13-dibutyrate stimulation, green fluorescent protein-PKCalpha and beta1 integrins co-sediment with ERM proteins in low-density sucrose gradient fractions that are enriched in transferrin receptors. Using fluorescence lifetime imaging microscopy, PKCalpha is shown to form a molecular complex with ezrin, and the PKC-co-precipitated endogenous ERM is hyperphosphorylated at the C-terminal threonine residue, i.e. activated. Electron microscopy showed an enrichment of both proteins in plasma membrane protrusions. Finally, overexpression of the C-terminal threonine phosphorylation site mutant of ezrin has a dominant inhibitory effect on PKCalpha-induced cell migration. We provide the first evidence that PKCalpha or a PKCalpha-associated serine/threonine kinase can phosphorylate the ERM C-terminal threonine residue within a kinase-ezrin molecular complex in vivo.

Imaging biochemistry inside cells.

Proteins provide the building blocks for multicomponent molecular units, or pathways, from which higher cellular functions emerge. These units consist of either assemblies of physically interacting proteins or dispersed biochemical activities connected by rapidly diffusing second messengers, metabolic intermediates, ions or other proteins. It will probably remain within the realm of genetics to identify the ensemble of proteins that constitute these functional units and to establish the first-order connectivity. The dynamics of interactions within these protein machines can be assessed in living cells by the application of fluorescence spectroscopy on a microscopic level, using fluorescent proteins that are introduced within these functional units. Fluorescence is sensitive, specific and non-invasive, and the spectroscopic properties of a fluorescent probe can be analysed to obtain information on its molecular environment. The development and use of sensors based on the genetically encoded variants of green-fluorescent proteins has facilitated the observation of 'live' biochemistry on a microscopic level, with the advantage of preserving the cellular context of biochemical connectivity, compartmentalization and spatial organization. Protein activities and interactions can be imaged and localized within a single cell, allowing correlation with phenomena such as the cell cycle, migration and morphogenesis.

Quantitative imaging of lateral ErbB1 receptor signal propagation in the plasma membrane.

Evidence for a new signaling mechanism consisting of ligand-independent lateral propagation of receptor activation in the plasma membrane is presented. We visualized the phosphorylation of green fluorescent protein (GFP)-tagged ErbB1 (ErbB1-GFP) receptors in cells focally stimulated with epidermal growth factor (EGF) covalently attached to beads. This was achieved by quantitative imaging of protein reaction states in cells by fluorescence resonance energy transfer (FRET) with global analysis of fluorescence lifetime imaging microscopy (FLIM) data. The rapid and extensive propagation of receptor phosphorylation over the entire cell after focal stimulation demonstrates a signaling wave at the plasma membrane resulting in full activation of all receptors.

Global analysis of fluorescence lifetime imaging microscopy data.

Global analysis techniques are described for frequency domain fluorescence lifetime imaging microscopy (FLIM) data. These algorithms exploit the prior knowledge that only a limited number of fluorescent molecule species whose lifetimes do not vary spatially are present in the sample. Two approaches to implementing the lifetime invariance constraint are described. In the lifetime invariant fit method, each image in the lifetime image sequence is spatially averaged to obtain an improved signal-to-noise ratio. The lifetime estimations from these averaged data are used to recover the fractional contribution to the steady-state fluorescence on a pixel-by-pixel basis for each species. The second, superior, approach uses a global analysis technique that simultaneously fits the fractional contributions in all pixels and the spatially invariant lifetimes. In frequency domain FLIM the maximum number of lifetimes that can be fit with the global analysis method is twice the number of lifetimes that can be fit with conventional approaches. As a result, it is possible to discern two lifetimes with a single-frequency FLIM setup. The algorithms were tested on simulated data and then applied to separate the cellular distributions of coexpressed green fluorescent proteins in living cells.

Three-dimensional spectral imaging by hadamard transform spectroscopy in a programmable array microscope.

We report the acquisition and deconvolution of three-dimensional spectrally resolved images in a programmable array microscope implementing a Hadamard transform fluorescence spectroscopy system with adjustable spectral resolution. A stack of 16 two-dimensional spectral images was collected at 400 nm intervals along the optical axis. The specimen consisted of a polytene chromosome spread from Drosophila melanogaster doubly labelled for the Polyhomeotic protein by indirect immunofluorescence labelling with Alexa594 and for DNA with YOYO-1. The resulting four-dimensional data set consisted of the xyz spatial dimensions (898 x 255 x 16) with a 26-point spectrum at each spatial location. The total exposure time to the sample was 34 min. The system requires the acquisition of multiple images, and thus works best with fluorophores that are resistant to photobleaching. Image deconvolution reduced the amount of out-of-focus blur by up to a factor of 8, resulting in a dramatic improvement in the visualization of the chromosome backbone and localization of the specific Polyhomeotic domains.

Multiple frequency fluorescence lifetime imaging microscopy.

The experimental configuration and the computational algorithms for performing multiple frequency fluorescence lifetime imaging microscopy (mfFLIM) are described. The mfFLIM experimental set-up enables the simultaneous homodyne detection of fluorescence emission modulated at a set of harmonic frequencies. This was achieved in practice by using monochromatic laser light as an excitation source modulated at a harmonic set of frequencies. A minimum of four frequencies were obtained by the use of two standing wave acousto-optic modulators placed in series. Homodyne detection at each of these frequencies was performed simultaneously by mixing with matching harmonics present in the gain characteristics of a microchannel plate (MCP) image intensifier. These harmonics arise as a natural consequence of applying a high frequency sinusoidal voltage to the photocathode of the device, which switches the flow of photoelectrons 'on' and 'off' as the sinus voltage swings from negative to positive. By changing the bias of the sinus it was possible to control the duration of the 'on' state of the intensifier relative to its 'off' state, enabling the amplitude of the higher harmonic content in the gain to be controlled. Relative modulation depths of 400% are theoretically possible from this form of square-pulse modulation. A phase-dependent integrated image is formed by the sum of the mixed frequencies on the phosphor of the MCP. Sampling this signal over a full period of the fundamental harmonic enables each harmonic to be resolved, provided that the Nyquist sampling criterion is satisfied for the highest harmonic component in the signal. At each frequency both the phase and modulation parameters can be estimated from a Fourier analysis of the data. These parameters enable the fractional populations and fluorescence lifetimes of individual components of a complex fluorescence decay to be resolved on a pixel-by-pixel basis using a non-linear fit to the dispersion relationships. The fitting algorithms were tested on a simulated data set and were successful in disentangling two populations having 1 ns and 4 ns fluorescence lifetimes. Spatial invariance of the lifetimes was exploited to improve the accuracy significantly. Multiple frequency fluorescence lifetime imaging microscopy was then successfully applied to resolve the fluorescence lifetimes and fluorescence intensity contributions in a rhodamine dye mixture in solution, and green fluorescent protein variants co-expressed in live cells.

An optical sectioning programmable array microscope implemented with a digital micromirror device.

The defining feature of a programmable array microscope (PAM) is the presence of a spatial light modulator in the image plane. A spatial light modulator used singly or as a matched pair for both illumination and detection can be used to generate an optical section. Under most conditions, the basic optical properties of an optically sectioning PAM are similar to those of rotating Nipkow discs. The method of pattern generation, however, is fundamentally different and allows arbitrary illumination patterns to be generated under programmable control, and sectioning strategies to be changed rapidly in response to specific experimental conditions. We report the features of a PAM incorporating a digital micromirror device, including the axial sectioning response to fluorescent thin films and the imaging of biological specimens. Three axial sectioning strategies were compared: line scans, dot lattice scans and pseudo-random sequence scans. The three strategies varied widely in light throughput, sectioning strength and robustness when used on real biological samples. The axial response to thin fluorescent films demonstrated a consistent decrease in the full width at half maximum (FWHM), accompanied by an increase in offset, as the unit cells defining the patterns grew smaller. Experimental axial response curves represent the sum of the response from a given point of illumination and cross-talk from neighbouring points. Cross-talk is minimized in the plane of best focus and when measured together with the single point response produces a decrease in FWHM. In patterns having constant throughput, there appears to be tradeoff between the FWHM and the size of the offset. The PAM was compared to a confocal laser scanning microscope using biological samples. The PAM demonstrated higher signal levels and dynamic range despite a shorter acquisition time. It also revealed more structures in x-z sections and less intensity drop-off with scanning depth.

A comparison of image restoration approaches applied to three-dimensional confocal and wide-field fluorescence microscopy.

We have compared different image restoration approaches for fluorescence microscopy. The most widely used algorithms were classified with a Bayesian theory according to the assumed noise model and the type of regularization imposed. We considered both Gaussian and Poisson models for the noise in combination with Tikhonov regularization, entropy regularization, Good's roughness and without regularization (maximum likelihood estimation). Simulations of fluorescence confocal imaging were used to examine the different noise models and regularization approaches using the mean squared error criterion. The assumption of a Gaussian noise model yielded only slightly higher errors than the Poisson model. Good's roughness was the best choice for the regularization. Furthermore, we compared simulated confocal and wide-field data. In general, restored confocal data are superior to restored wide-field data, but given sufficient higher signal level for the wide-field data the restoration result may rival confocal data in quality. Finally, a visual comparison of experimental confocal and wide-field data is presented.

Homologous association of the Bithorax-Complex during embryogenesis: consequences for transvection in Drosophila melanogaster.

Transvection is the phenomenon by which the expression of a gene can be controlled by its homologous counterpart in trans, presumably due to pairing of alleles in diploid interphase cells. Transvection or trans-sensing phenomena have been reported for several loci in Drosophila, the most thoroughly studied of which is the Bithorax-Complex (BX-C). It is not known how early trans-sensing occurs nor the extent or duration of the underlying physical interactions. We have investigated the physical proximity of homologous genes of the BX-C during Drosophila melanogaster embryogenesis by applying fluorescent in situ hybridization techniques together with high-resolution confocal light microscopy and digital image processing. The association of homologous alleles of the BX-C starts in nuclear division cycle 13, reaches a plateau of 70% in postgastrulating embryos, and is not perturbed by the transcriptional state of the genes throughout embryogenesis. Pairing frequencies never reach 100%, indicating that the homologous associations are in equilibrium with a dissociated state. We determined the effects of translocations and a zeste protein null mutation, both of which strongly diminish transvection phenotypes, on the extent of diploid homologue pairing. Although translocating one allele of the BX-C from the right arm of chromosome 3 to the left arm of chromosome 3 or to the X chromosome abolished trans-regulation of the Ultrabithorax gene, pairing of homologous alleles surprisingly was reduced only to 20-30%. A zeste protein null mutation neither delayed the onset of pairing nor led to unpairing of the homologous alleles. These data are discussed in the light of different models for trans-regulation. We examined the onset of pairing of the chromosome 4 as well as of loci near the centromere of chromosome 3 and near the telomere of 3R in order to test models for the mechanism of homologue pairing.

Improved restoration from multiple images of a single object: application to fluorescence microscopy.

We present an approach for the combined restoration of multiple different images of a single object. A linear Tikhonov filter adapted for this purpose is derived in detail. Nonlinear constrained algorithms can also be adapted, and we illustrate this possibility for an iterative constrained Tikhonov algorithm. Both the linear and the iterative constrained Tikhonov algorithms were used to analyze performance in fluorescence confocal imaging by use of simulated and experimental data. One can improve the quality of restored confocal images significantly if the signal that normally is rejected by the detection pinhole of a confocal laser scanning microscope is also recorded on a separate detector such that the two recorded signals are used together for image restoration according to the proposed algorithms.

Imaging the intracellular trafficking and state of the AB5 quaternary structure of cholera toxin.

The subcellular localization and corresponding quaternary state of fluorescent labelled cholera toxin were determined at different time points after exposure to living cells by a novel form of fluorescence confocal microscopy. The compartmentalization and locus of separation of the pentameric B subunits (CTB) from the A subunit (CTA) of the toxin were evaluated on a pixel-by-pixel (voxel-by-voxel) basis by measuring the fluorescence resonance energy transfer (FRET) between CTB labelled with the sulfoindocyanine dye Cy3 and an antibody against CTA labelled with Cy5. The FRET efficiency was determined by a new technique based on the release of quenching of the Cy3 donor after photodestruction of the Cy5 acceptor in a region of interest within the cell. The results demonstrate vesicular transport of the holotoxin from the plasma membrane to the Golgi compartment with subsequent separation of the CTA and CTB subunits. The CTA subunit is redirected to the plasma membrane by retrograde transport via the endoplasmic reticulum whereas the CTB subunit persists in the Golgi compartment.