Information

How to measure the length of mitochondria from z stack fluorescent microscopy image?

How to measure the length of mitochondria from z stack fluorescent microscopy image?


We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

I have been working on yeast cells to analyse the effect of DNA damaging agents on mitochondrial structure. I have imaged my culture treated with MMS for a period of 6 hours and while observing the images, I suspect there is fragmentation of mitochondria in later point of time. So, I am wondering if I can get measurements of mitochondria using z stack images.


ImageJ is usually the standard software to measure cell characteristics, a little bit of a learning curve but there is a large suite of analysis methods and image adjustments/filters.


Challenges of Using Expansion Microscopy for Super-resolved Imaging of Cellular Organelles

In Expansion Microscopy (ExM) subcellular structures are imaged in isotropically expanded fixed samples, consequently allowing to resolve enlarged subcellular structures, otherwise hidden to standard microscopy. Upon comparison of the expansion factors of different cellular compartments in cells within the same gel, we found significant differences in expansion factors of a factor of above 2.


Introduction

Live-cell microscopy has been accessible for decades, as is evident from a movie that was taken with 16-mm film over 50 years ago of a neutrophil chasing a bacterium (David Rogers, Vanderbilt University, http://www.biochemweb.org/neutrophil.shtml). The technique now spans all fields of the life sciences and extends to the physical sciences as well. In recent years, technological advances, including sensor sensitivity, computing power, brighter and more-stable fluorescent proteins (FPs), and new fluorescent probes for cellular compartments, have given researchers the tools to study complex biological processes in great detail (Goldman and Spector, 2005). However, expertise in the optimization of image-acquisition conditions for various microscopy platforms is required to harness the full potential that live-cell microscopy offers.

As with any measuring device, it is best to minimize any perturbations by optimizing the system so that it is minimally invasive. As part of their normal life cycle, most tissues and cells are never exposed to light, and it is known that ultraviolet (UV) light damages DNA, focused infrared (IR) light can cause localized heating, and fluorescence excitation causes phototoxicity to tissues and cells (Pattison and Davies, 2006). The main cause of phototoxicity in living cells is the oxygen-dependent reaction of free-radical species, which are generated during the excitation of fluorescent proteins or dye molecules with surrounding cellular components. Thus, for live-cell imaging, it is best to reduce the amount of excitation light by optimizing the efficiency of the light path through the microscope, and by using detectors that are optimized to detect most of the fluorescence emission. Low concentrations of fluorescent probes also need to be used to avoid causing nonspecific changes to the biological processes of interest.

With live-cell microscopy, there must be a compromise between acquiring beautiful images and collecting data that provide a high enough signal-to-noise ratio (S/N) to make meaningful quantitative measurements of a living specimen. Therefore, the focus of this Commentary is to discuss how to keep cells or tissue alive and healthy during image acquisition, to provide guidelines for different types of samples to delineate the different imaging modalities that are most appropriate, and to provide general and imaging-platform-specific recommendations for instrument components and settings.

A certain level of knowledge about transmitted-light microscopy and fluorescence microscopy is assumed. For the beginner, there are good articles on light microscopy (Murphy, 2001), fluorescence microscopy (Brown, 2007 Lichtman and Conchello, 2005 Murphy, 2001 North, 2006 Wolf, 2007) and confocal microscopy (Hibbs, 2004 Mueller, 2005 Pawley, 2006) that provide the necessary background. There are also many excellent publications on live-cell imaging that provide a lot of detailed and valuable information (Dailey et al., 2006 Day and Schaufele, 2005 Goldman and Spector, 2005 Haraguchi, 2002 Wang et al., 2008) (http://cshprotocols.cshlp.org/cgi/collection/live_cell_imaging). Image analysis is also a crucial component of live-cell imaging, but is beyond the scope of this article. Therefore, the reader is referred to other papers for more details on performing accurate colocalization measurements (Bolte and Cordelieres, 2006 Comeau et al., 2006 Kraus et al., 2007) and accurately quantifying fluorescence signals (Brown, 2007 Cardullo and Hinchcliffe, 2007 Murray, 2007 Swedlow, 2007 Wolf et al., 2007). The focus here is on imaging mammalian cells however, most aspects of this discussion carry over to the imaging of any living organism or tissue.


Visual Methods For Evaluating Colocalization

Colocalization of two probes may be subjectively identified by the appearance of structures whose color reflects the combined contribution of both probes when the images of each probe are superimposed (or “merged”). So, for example, colocalization of fluorescein and rhodamine can be apparent in structures that appear yellow, because of the combined contributions of green and red fluorescence, respectively. Figure 1A shows a projected image volume of Madin Darby Canine Kidney (MDCK) cells incubated in a combination of Texas Red-labeled transferrin and Alexa 488-labeled transferrin. Since both probes are internalized via the same transferrin receptors, they would be expected to codistribute in endosomes following internalization, as is apparent in the constant yellow color of endosomes. In contrast, internalized Texas Red-dextran and Alexa 488-transferrin distribute to two distinct compartments, which appear red and green, respectively, in Fig. 1E .

Colocalization analysis of endocytic probes. A: maximum projection of an image volume of Madin Darby Canine Kidney (MDCK) cells following incubation with transferrin conjugated to Texas Red (B) and Alexa 488 (C) Arrows indicate endosomes containing both probes. D: scatterplot of red and green pixel intensities of the two cells shown in A. E: maximum projection of an image volume of MDCK cells following incubation in Texas Red-dextran (F) and Alexa 488-transferrin (G). Arrows indicate examples of lysosomes containing Texas Red-dextran. H: scatterplot of red and green pixel intensities of the image shown in E.

Superposition of fluorescence images is certainly the most prevalent method for evaluating colocalization, and tools for displaying multiple-channel fluorescence images as merged color images are implemented in all biological image analysis software. However, results can be ambiguous. The problem is that an intermediate color, indicating colocalization, is obtained only if the intensities of the two probes are similar. The insets in Fig. 1A show how small changes in the relative intensity of two probes can completely alter the combined color of the endosomes and thus the perception of probe colocalization. For this reason, the overall degree of colocalization throughout a sample may be visually apparent only under very specific labeling conditions, when the fluorescence of the two probes occurs in a fixed and nearly equal proportion. In general, the most reliable method for visually comparing the relative distribution of two probes is a side-by-side comparison of the two images, with arrows provided as landmarks (compare the colocalization of the probes in Fig. 1 , B and C, with that in Fig. 1 , F and G).

The results of fluorescence colocalization studies can also be represented graphically in scatterplots where the intensity of one color is plotted against the intensity of the second color for each pixel, similar to the output provided for flow cytometry data. Under the conditions of proportional codistribution, such as in the data shown in Fig. 1A , the points of the scatterplot cluster around a straight line, whose slope reflects the ratio of the fluorescence of the two probes ( Fig. 1D ). In contrast, the lack of colocalization of dextran and transferrin in the image shown in Fig. 1E is reflected by the distribution of points into two separate groups, each showing varying signal levels of one probe with little or no signal from the other probe ( Fig. 1H ). The ability to produce and export scatterplots is common to nearly all biological image analysis software packages.

Scatterplots can provide additional insights into colocalization studies. First, they can be used to identify populations of distinct compartments. Our laboratory has used scatterplots to identify two populations of endosomes in MDCK epithelial cells, one at the apex that is enriched in internalized IgA and lacking internalized transferrin, and the other in lower portions of the cell that contains both IgA and transferrin ( Fig. 2A ) (9, 46). These two compartments are readily distinguished in scatterplots in which two different linear relationships are obtained, with the slopes reflecting the distinctive ratios of internalized IgA and transferrin in each type of compartment ( Fig. 2B ). The scatterplot obtained from images of cells treated with brefeldinA was used to support the observation that brefeldinA induced a fusion of these different compartments, resulting in a population of endosomes with a single, intermediate ratio of IgA to transferrin ( Fig. 2C ).

Colocalization analysis of endocytic probes. A: three-dimensional stereopair image of a maximum projected image volume MDCK cells following incubation with Oregon Green-IgA and Texas Red-transferrin (Tf). B: scatterplot of red and green pixel intensities of MDCK cells following incubation with Oregon Green-IgA and Texas Red-transferrin. C: scatterplot of individual red and green pixel intensities of brefeldinA-treated MDCK cells following incubation with Oregon Green-IgA and Texas Red-transferrin. D: Pearson's correlation coefficients (PCCs) of images of internalized Texas Red-transferrin and green fluorescent protein (GFP)-Rab10 or GFP-Rab11a in MDCK cells. E: PCCs of images of internalized Cy5-IgA and GFP-Rab10, GFP-Rab11a, or GFP-Rab10-Q68L in MDCK cells. F: PCCs of images of Texas Red-transferrin and Cy5-transferrin internalized by MDCK cells before and after rotating Cy5-Tf image by 90 degrees. AC were reproduced with permission from the Company of Biologists (http://jcs.biologists.org), Wang EX et al. 𠇋refeldin A rapidly disrupts plasma membrane polarity by blocking polar sorting in common endosomes of MDCK cells.” J Cell Sci 114: 3309�, 2001. DF were reproduced and adapted with permission from The American Society for Cell Biology, from Babbey CM et al. “Rab10 regulates membrane transport through early endosomes of polarized Madin-Darby Canine Kidney cells.” Molec Biol Cell 17: 3156�, 2006. (www.molbiolcell.org) (4).

The visual techniques described above are useful for exploring the relative distribution of different molecules in cells. Superposition of images is useful for providing a spatial sense of colocalization, identifying regions of the cell or compartments where molecules colocalize. Scatterplots are useful for detecting the presence of different populations of compartments. They also provide a qualitative indication of the degree of colocalization. However, these representations are generally not helpful for comparing the degree of colocalization in different experimental conditions nor for determining whether the amount of colocalization exceeds random coincidence. In the next sections we will describe several approaches that can be used to quantify colocalization. These methods are simple to employ and have been implemented in a variety of image processing software packages. However, there are numerous subtleties and assumptions in each that must be understood before they can be productively applied to biological images.


Measuring cell fluorescence using ImageJ¶

Repeat this step for the other cells in the field of view that you want to measure.

NB: Size is not important. If you want to be super accurate here take 3+ selections from around the cell.

  • Once you have finished, select all the data in the Results window and copy and paste into a new spreadsheet (or similar program)
  • Use this formula to calculate the corrected total cell fluorescence (CTCF).

CTCF = Integrated Density – (Area of selected cell X Mean fluorescence of background readings)

Notice that rounded up mitotic cells appear to have a much higher level of staining due to its smaller size concentrating the staining in a smaller space. If you used the raw integrated density you would have data suggesting that the flattened cell has less staining then the rounded up one, when in reality they have a similar level of fluorescence.

This method is based, with permission, on an original protocol from QBI, The University of Queensland, Australia.


STAGE TWO: ENSEMBLE LEVEL MEASUREMENTS

Cell Cycle Example

We have divided image analysis into an image processing stage where the individual cells are parameterized, and a second stage where ensemble averages are made and the cells are characterized biologically. We have mentioned the advantages of doing this from a processing point of view, the second stage being much less CPU intensive than the first. Another advantage is procedural we have left the difficult technical task of reducing the cell image to a set of numbers and can now concentrate on the biology. This stage of the analysis is often repeated and tuned numerous times during the course of an analysis, with input from various collaborators, so it is useful not to have to reanalyze all of the images every time someone has a good idea about what to measure. What is measured at the ensemble level depends on the experiment being performed and is thus often unique. For example, if one were performing a cell cycle time-course experiment on polar protein abundance and localization with synchronized cells, it would be reasonable to plot the peak polar signal averaged over an ensemble of high quality cells for each time point in the cell cycle. An example of this type of analysis is shown in Figure਄ for a Caulobacter crescentus cell cycle experiment. The values shown are from ensemble averages over many cells and the error bars reflect the error in those measurements. Although it took several hundred seconds to parameterize all of the imaged cells during the first stage of this analysis, the second stage, involving selection of high quality data, determining the mean of the peak fluorescent signals in the three channels and plotting the data, takes only seconds.

Polar localized protein concentration as a function of cell cycle. (A) An example of cell cycle dependent protein localization in Caulobacter. (B) An example single-cell image shows the correspondence with the visible changes in localized protein concentration (C) Automated analysis detects time variation in polar fluorescence intensities of the CpaE-CFP, PleC-YFP, and DivJ-RFP reporter fusions.

Mutant Screening Example

A large-scale mutant screen provides a more challenging example. In this case it is necessary to process images from thousands of mutants, and then classify them on the basis of several criteria such as polar-localized signal amplitude and location, cytoplasmic-delocalized signal amplitude, localized protein polarity, as well as a host-cell shape parameters. As an example we show in Figureਅ the localized fluorescence amplitude signals from a triply labeled strain of Caulobacter crescentus used in a high throughput screen for mutants that mislocalized one or more of the reporter proteins. Each data point is an ensemble average over a number of cells whose peak signal is in the high intensity tail of the peak signal distribution. The data shows the distinct difference between reporter signals resulting from transposon disruptions in the PodJ and TacA genes and the unmutagenized control strains. In Figureਅ we can simultaneously observe three-dimensional phenotypic data, labeled by genotype. However, incorporating more fluorescence based phenotypic information requires analysis tools capable of dealing with higher-dimensional data. Because we know the genotype of every mutant analyzed for each fluorescent phenotype, we can use hierarchical clustering as shown in Figureਆ , taking advantage of all the fluorescent data available in the screen. The cluster shown in Figureਆ represents a robust cluster of mutants that emerged from a much larger dendrogram of over 800 mutants. The genes identified in the cluster correspond to the elements of a pathway controlling the relocalization of the sensor histidine kinase DivJ to the stalked pole at the swarmer to stalked transition (Biondi et al. 2006 Radhakrishnan et al. 2008 Christen and Fero 2009). Thus, it is possible to completely automate a fluorescence microscopy based screen that leads to functional gene module identification.

Ensemble measurement of polar localized protein based on individual cell measurements. The polar fluorescence intensities of the CpaE-CFP, PleC-YFP, and DivJ-RFP reporter fusions are plotted for the nonmutagenized control strain (black) as well as gene disruptions in two separate open reading frames coding for the PodJ (blue) and TacA (red) proteins.

Example of Genotype-Phenotype microscopy based analysis. A collection of automatically determined cell metrics is used to identify a functional cluster indicating a gene module coordinating temporal and spatial protein localization. Heat map representation of the calculated z-score values from localized protein abundance (L), delocalized protein (D), the fraction of cells identified as being biopolar (B), and the fraction similarly identified as monopolar (M) are shown for a highly significant gene cluster designated by the dendrogram.

Shape Characterization

The shape of a mutant or non–wild-type cell is often a critical part of a fluorescence microscopy experiment. Although one of the easiest things to pick out by eye, it is surprisingly difficult to parameterize mutant shapes, because it is hard to anticipate all the shapes one might encounter. The membrane parameterization techniques we have described do a reasonably good job following a complex membrane profile but the chances of missing an interesting phenotype increase as the cells begin to differ markedly from what was expected by the algorithm developer. Thus, mutant shape characterization can benefit from more global analysis techniques. In these approaches, the cell is not assumed to have a particular form that yields to parameterization in the way that we have described above. Rather, the cells are interpreted as observations that can be compared either to each other or to a control set and tested for similarity, or to a set of alternative models for classification. This process lends itself to the use of pattern recognition and supervised or unsupervised statistical learning techniques. Supervised techniques require a training data set and produce a model that can then be applied to a test set for purposes of classification. The canonical examples of these approaches are classification and regression trees and various types of neural nets (Hastie et al. 2009). In an unsupervised approach, no assumption about what is expected is made and the algorithm is expected to bootstrap its way to finding unique classes based on the data itself (Heidmann 2005). Image data is not often directly amenable to these approaches and is usually first reduced in complexity by the computation of numerical representations of the image. After a cell has been thus re-interpreted it can be classified via one of the supervised or unsupervised approaches. Good examples of morphology analysis are found in Pincus and Jones (Pincus and Theriot 2007 Jones et al. 2009).


DISCUSSION

BMVC-12C-P, o-BMVC-12C-P, o-BMVC-6C-P and o-BMVC play dual roles as fluorescent probes and G4 stabilizers in vitro. BMVC-12C-P and o-BMVC-12C-P have been shown to induce cancer cell death however, o-BMVC-6C-P and o-BMVC have not. The former two are fluorescent anti-cancer agents, while the latter two are fluorescent G4 ligands. Our investigation using fluorescence revealed the mechanism that leads to the differences in cytotoxicity and the major target in mitochondria. Briefly, o-BMVC-6C-P presents the same G4 stabilization characteristics as o-BMVC-12C-P does and also targets mitochondria however, it presents weak fluorescence and has no cytotoxic effects on cancer cells. This implies that the effects of the interactions between o-BMVC-6C-P and other biomolecules are limited. Our fluorescence study of isolated mitochondria suggests that mitochondrial uptake (associated with lipophilicity ( 28) and regulated by the length of alkyl chain) is the mechanism primarily responsible for differences in cytotoxicity between these G4 ligands. Specifically, the strong fluorescence of o-BMVC-12C-P in mitochondria is due to ( 1) the interaction with mtDNA and ( 2) the fact that a sufficient quantity of o-BMVC-12C-P can suppress mtDNA gene expression and eventually induce cell death. This study demonstrated the importance of fluorescent anti-cancer agents in therapeutic applications and also helped to elucidate the underlying mechanisms that could be exploited through the development of effective anti-tumor agents.

Obtaining proof for the existence of G4 in living cells has been a long-term challenge ( 32) however, recent studies on the visualization of exogenous and endogenous G4s in human cells firmly support the existence of G4s in live cells ( 8, 23, 33–35) and fixed cells ( 36–38). In addition, the formation of G4s has been implicated in many biological processes, which has led to the suggestion that the topology of G4 may provide a novel therapeutic target. Our results from fluorescence studies and bioactivity assays provide convincing evidence to support the existence of G4 in isolated mitochondria as well as in the mitochondria of live cells.

The fact that there are ∼200 PQF sequences of mtDNA ( 16, 17) underlines the importance of identifying all G4s that may be involved in G4 ligand-induced cell death and elucidating how the G4 ligand affects mtDNA function. In this study, BMVC-12C-P and o-BMVC-12C-P were shown to stabilize the mtDNA G4s however, the results of RT-PCR provide evidence of mtDNA gene-dependent suppression with more pronounced suppression of ND3 and COXI. PCR stop assays suggest that G4s formed by mt6363 (in ND3) and mt9438 (in COX I) are at least partially the targets of BMVC-12C-P and o-BMVC-12C-P. This makes it possible to use RT-PCR in conjunction with PCR stop assays for the identification of mtDNA G4s involved in G4 ligand induced cell death.

Previous reports have shown a high degree of correlation between the PQF sequences of mtDNA and mtDNA deletion, which could disrupt genomic stability in mitochondria ( 16, 17). It has also been reported that a DNA/RNA hybrid G4 of a conserved sequence block II (CSB II) may play a regulatory role in transcriptional termination ( 38–41). Thus, these studies indicate the existence of G4s and the potential roles played by mtDNA G4s in mitochondrial function. The fact that numerous mitochondrial diseases are related to G4 DNA-forming sequences means that a rational drug design capable of manipulating or even disrupting G4 structures could be an important avenue for the further development of treatment modalities. We believe that mtDNA G4 is a promising target for mitochondrial medicine.

In summary, our findings illustrate the importance of fluorescence as a tool for exposing targets in live cells, detecting G4 in mitochondria, unraveling the anti-cancer mechanisms of agent in cells and guiding drug development. Specifically, this work provides very strong evidence to support the existence of G4s in the mitochondria of live cells and presents the first example of a G4 anti-cancer agent that targets the mtDNA of cancer cells and eventually causes cell death.

We thank Dr. Shing-Jong Huang and Mrs. Shou-Ling Huang (Instrumentation Center, National Taiwan University) for assistance in obtaining the NMR data using the Bruker AVIII 500 and 800MHz NMR spectrometers.


RESULTS

Tethering ER and mitochondria in Drosophila using an engineered linker

The association between mitochondria and the ER has been widely documented in different experimental models. The study of this association has been the focus of intensive research, using both cellular and animal models. However, the diversity of the reported functions of these contacts suggests that their role is still unclear (reviewed in Csordás et al., 2018).

Here, we used a genetic approach to explore the effects of tethering the ER and mitochondria in fruit flies (Drosophila melanogaster) using an artificial linker. To increase the physical coupling between both organelles, we used a previously reported linker that induces enhanced proximity (6 nm) between mitochondria and the ER in mammals (Csordás et al., 2006). We first engineered a codon-optimised version of this synthetic linker for expression in Drosophila. This synthetic linker (from here on referred to as ‘linker’) consists of monomeric red fluorescent protein (RFP) with a mitochondrial targeting sequence at the N terminus and an ER targeting sequence at the C terminus (Fig. 1A). Transgenic flies expressing this construct were generated by PhiC31 integrase-mediated transgenesis (see Materials and Methods) and the expression of the linker was assayed by immunofluorescence. Confocal analysis of the larval wing discs revealed punctate staining of the linker (Fig. 1B).

In vivo expression analysis of an engineered tether between mitochondria and the ER. (A) A schematic illustration of the linker tethering mitochondria (blue) and the ER (grey). The magnified panel on the left details the relative position of the two tags (green) targeting both mitochondria and the ER and bridged by monomeric RFP. (B) Analysis of the expression of the linker targeted to the posterior compartment of the larval wing disc. Genotype: w UAS-linker/+ hhGal4, UAS-GFP/+.

In vivo expression analysis of an engineered tether between mitochondria and the ER. (A) A schematic illustration of the linker tethering mitochondria (blue) and the ER (grey). The magnified panel on the left details the relative position of the two tags (green) targeting both mitochondria and the ER and bridged by monomeric RFP. (B) Analysis of the expression of the linker targeted to the posterior compartment of the larval wing disc. Genotype: w UAS-linker/+ hhGal4, UAS-GFP/+.

Linker expression increases MERCs and decreases mitochondrial length

Next, to determine if linker expression increases mitochondria–ER tethering, we quantified mitochondria–ER contacts in the adult brains as previously described (Celardo et al., 2016). Ultrastructural analysis by transmission electron microscopy (TEM) revealed a significant increase in mitochondria–ER contacts in flies expressing the linker compared to controls (Fig. 2A and B). We next compared the length of the interface between mitochondria and the ER and the distance between the two organelles in flies expressing the linker to the controls. This showed that linker expression increases the average length of the interface (Fig. 2C) and decreases the distance between mitochondria and the ER (Fig. 2D). The points of contact between the ER and mitochondria correlate with the localisation of dynamin-related protein 1 (DRP1), a protein involved in mitochondrial fission in vertebrate cells. Such contact sites have been proposed to be involved in mitochondrial division (reviewed in Rowland and Voeltz, 2012). Confocal analysis of mitochondria labelled with a fluorescent tag (mito-GFP) in the mechanosensory neurons of the ventral nerve cord showed that the expression of the linker resulted in decreased mitochondrial length (Fig. 2E and F). Mitochondrial fragmentation in Drosophila has been proposed to drive the removal of defective mitochondria by mitophagy (Lieber et al., 2019). To determine whether linker expression affects mitochondrial removal through fragmentation, we next assessed mitochondrial density by measuring the levels of the mitochondrial matrix enzyme citrate synthase, an indirect measure of mitochondrial density in flies (Magwere et al., 2006). Citrate synthase levels in flies expressing the linker were not altered compared to those in controls (Fig. 2G), indicating that the mitochondrial fragmentation observed upon linker expression was not accompanied by a loss of mitochondrial mass. Mitochondria generate the majority of ATP as a source of cellular energy. We measured total levels of ATP in adult flies expressing the linker but did not find significant changes in ATP levels compared to those in controls (Fig. 2H). We also assessed the functional status of mitochondria by measuring the activity of respiratory complexes in flies expressing the linker and found no significant changes (Fig. 2I and J). Together, these results show that forcing contacts between mitochondria and the ER using an artificial linker induces mitochondrial fragmentation without altering global mitochondrial density or activity. Enhancing MERCs led to a fragmented mitochondrial network. In flies this fragmentation is associated with the transcriptional activation of the mitochondrial unfolded protein response (UPR mt ) (de Castro et al., 2012) that can enhance lifespan (Jensen et al., 2017). To determine if enhancing MERCs is associated with a transcriptional stress response, we employed microarray technology with an in silico workflow (Fig. 3A). We failed to detect alterations in transcripts associated with mitochondrial fission or fusion (data not shown) but instead identified transcriptional alterations associated with resistance to high oxygen environments in flies (Fig. 3B).

Forcing contacts between mitochondria and the ER induces mitochondrial fragmentation. (A,B) The quantification of mitochondria–ER contacts in adult fly brains (asterisks, chi-square two-tailed, 95% confidence intervals) (B) and representative electron microscopy images (A). The yellow arrows show mitochondria in contact with the ER (arrows). ER, endoplasmic reticulum m, mitochondria. (C) Analysis of the length of the interface between mitochondria and the ER in adult brains. The mean is shown above the individual measurements. (D) Analysis of the distance between mitochondria and the ER at MERCs. Both the minimum and maximum observed distance at each individual MERC was measured. The quantification of distance is shown as a combined violin and box plot. The mean distance is also shown. P-value, two-tailed unpaired t-test, compared to control. (E,F) The expression of the linker results in mitochondrial fragmentation. Confocal analysis of mitoGFP in the larval mechanosensory axons. Representative confocal images (E) and the quantification of mitochondrial length (F). The quantification of mitochondrial length is shown as a combined violin and box plot (P-value, two-tailed unpaired t-test, compared to control). (G) The expression of the linker does not affect overall mitochondrial mass. Mitochondrial mass assessed by measuring the activity of the mitochondrial matrix enzyme citrate synthase in adults (mean±s.d., ns, P > 0.05, two-tailed unpaired t-test, compared to control). (H–J) The expression of the linker does not alter ATP levels (H) or mitochondrial function (I,J). Mitochondrial function was assayed by measuring the activity of NADH: cytochrome c reductase (I) and succinate: cytochrome c reductase (J) (mean±s.d., ns, P > 0.05, two-tailed unpaired t-test, compared to control). Genotypes (A–D): control: w elavGal4/+ +, linker: w elavGal4/UAS-linker +, (E,F): control: w elavGal4/+ UAS-mitoGFP/+, linker: w elavGal4/UAS-linker UAS-mitoGFP/+, (G–J): control: w + daGal4/+, linker: w UAS-linker/+ daGal4/+.

Forcing contacts between mitochondria and the ER induces mitochondrial fragmentation. (A,B) The quantification of mitochondria–ER contacts in adult fly brains (asterisks, chi-square two-tailed, 95% confidence intervals) (B) and representative electron microscopy images (A). The yellow arrows show mitochondria in contact with the ER (arrows). ER, endoplasmic reticulum m, mitochondria. (C) Analysis of the length of the interface between mitochondria and the ER in adult brains. The mean is shown above the individual measurements. (D) Analysis of the distance between mitochondria and the ER at MERCs. Both the minimum and maximum observed distance at each individual MERC was measured. The quantification of distance is shown as a combined violin and box plot. The mean distance is also shown. P-value, two-tailed unpaired t-test, compared to control. (E,F) The expression of the linker results in mitochondrial fragmentation. Confocal analysis of mitoGFP in the larval mechanosensory axons. Representative confocal images (E) and the quantification of mitochondrial length (F). The quantification of mitochondrial length is shown as a combined violin and box plot (P-value, two-tailed unpaired t-test, compared to control). (G) The expression of the linker does not affect overall mitochondrial mass. Mitochondrial mass assessed by measuring the activity of the mitochondrial matrix enzyme citrate synthase in adults (mean±s.d., ns, P > 0.05, two-tailed unpaired t-test, compared to control). (H–J) The expression of the linker does not alter ATP levels (H) or mitochondrial function (I,J). Mitochondrial function was assayed by measuring the activity of NADH: cytochrome c reductase (I) and succinate: cytochrome c reductase (J) (mean±s.d., ns, P > 0.05, two-tailed unpaired t-test, compared to control). Genotypes (A–D): control: w elavGal4/+ +, linker: w elavGal4/UAS-linker +, (E,F): control: w elavGal4/+ UAS-mitoGFP/+, linker: w elavGal4/UAS-linker UAS-mitoGFP/+, (G–J): control: w + daGal4/+, linker: w UAS-linker/+ daGal4/+.

A transcriptional signature associated to high oxygen tolerance is induced by enhancing MERCs. (A) Workflow used for the characterisation of transcripts involved in tolerance to hyperoxia in adult flies expressing the linker. (B) Transcripts with altered expression in flies expressing the linker. Red and blue correspond to significant (FDR ≤ 5%) upregulated and downregulated transcripts, respectively. The original signature for hyperoxia tolerance transcripts was identified by Zhao and colleagues (Zhao et al., 2010).

A transcriptional signature associated to high oxygen tolerance is induced by enhancing MERCs. (A) Workflow used for the characterisation of transcripts involved in tolerance to hyperoxia in adult flies expressing the linker. (B) Transcripts with altered expression in flies expressing the linker. Red and blue correspond to significant (FDR ≤ 5%) upregulated and downregulated transcripts, respectively. The original signature for hyperoxia tolerance transcripts was identified by Zhao and colleagues (Zhao et al., 2010).

Enhancing MERCs causes increased locomotor activity

We have previously shown that increased mitochondrial fission in flies with decreased expression of Drosophila mitofusin (dMfn) causes locomotor defects in adults flies (Garrido-Maraver et al., 2019). To determine whether mitochondrial fission caused by linker expression affects motor activity, we measured locomotion in both larvae and adults by expressing the synthetic linker ubiquitously. First, we analysed locomotor activity in larvae using a motility tracking assay. We found a significant increase in the crawling speed (Fig. 4A) and total displacement (Fig. 4B) of larvae expressing the linker. Next, analysis of locomotor activity in adult flies for a period of up to 20 days showed a generalised increase in the activity of flies expressing the linker (Fig. 4C and D).

Enhancing MERCs increases locomotor activity. (A,B) Analysis of larval crawling in control and linker-expressing animals. Larval mean speed (A) and displacement (B) were measured (mean±s.d. asterisk, two-tailed unpaired t-test). (C) Total activity in adult flies over a period of 20 days was measured using the Trikinetics system. (mean±s.d. asterisks, two-tailed unpaired t-test compared to control). (D) Actogram showing that the linker expressing flies have a higher average activity per hour than control over a period of 480 h. Genotypes: control: w + daGal4/+, linker: w UAS-linker/+ daGal4/+.

Enhancing MERCs increases locomotor activity. (A,B) Analysis of larval crawling in control and linker-expressing animals. Larval mean speed (A) and displacement (B) were measured (mean±s.d. asterisk, two-tailed unpaired t-test). (C) Total activity in adult flies over a period of 20 days was measured using the Trikinetics system. (mean±s.d. asterisks, two-tailed unpaired t-test compared to control). (D) Actogram showing that the linker expressing flies have a higher average activity per hour than control over a period of 480 h. Genotypes: control: w + daGal4/+, linker: w UAS-linker/+ daGal4/+.

Calcium levels are increased at mitochondria–ER contacts

MERCs play a key regulatory role in several cellular functions, including the transfer of Ca 2+ and lipids between the ER and the mitochondria (reviewed in Rowland and Voeltz, 2012). Additionally, in hepatocytes, Ca 2+ transport from the ER to mitochondria at contact sites elevates mitochondrial Ca 2+ and increases mitochondrial ROS (Arruda et al., 2014). To measure Ca 2+ levels in the mitochondria of flies expressing the linker, we used mitycam, a mitochondria-targeted fluorescent calcium reporter for Drosophila (Terhzaz et al., 2006). As our artificial linker is tagged with RFP, we compared the mitycam signal in areas of RFP fluorescence (which correspond to mitochondria–ER contacts) to that in areas in which no RFP fluorescence was detected (which correspond to mitochondria distant from the ER). This revealed that mitochondria close to the ER contain higher levels of Ca 2+ when compared to those in mitochondria distant from the ER (Fig. 5A and B).

Increased calcium levels in mitochondria at MERCs. (A,B) Analysis of mitochondrial calcium uptake in neurons at the ventral nerve cord in larvae expressing the linker. The signal of the mitycam reporter (mitycam signal) was analysed close to and away from the linker signal (RFP). (A) The quantification of the mitycam fluorescence [mean±s.d. asterisks, two-tailed unpaired t-test compared to control, n=region of interest (ROI) measured]. (B) Representative confocal images of the mitycam signal and the linker in larval brain cells. Genotypes (A,B): w elavGal4,UAS-mitycam/ UAS-linker.

Increased calcium levels in mitochondria at MERCs. (A,B) Analysis of mitochondrial calcium uptake in neurons at the ventral nerve cord in larvae expressing the linker. The signal of the mitycam reporter (mitycam signal) was analysed close to and away from the linker signal (RFP). (A) The quantification of the mitycam fluorescence [mean±s.d. asterisks, two-tailed unpaired t-test compared to control, n=region of interest (ROI) measured]. (B) Representative confocal images of the mitycam signal and the linker in larval brain cells. Genotypes (A,B): w elavGal4,UAS-mitycam/ UAS-linker.

Artificially increasing MERCs increases ROS levels and enhances longevity

We next tested whether the linker expression affects the levels of mitochondrial ROS in adult flies. We performed this analysis using CellRox, a fluorescent ROS indicator. Our analysis showed an age-dependent increase in the levels of ROS (Fig. 6A and B) in adult flies expressing the linker. It has been shown that increased ROS production in flies extends lifespan (Scialò et al., 2016). We therefore asked whether the effects on ROS levels induced by linker expression can affect the lifespan of adult flies. Our analysis showed that enhancing MERCs through linker expression extends lifespan (Fig. 6C).

Enhancing MERCs increases ROS levels in aged flies and increases lifespan. (A,B) Analysis of mitochondrial ROS levels in young (3-day-old) and aged (30-day-old) controls and linker-expressing flies. Representative confocal images of 30-day-old brains stained with CellRox green are shown (A). The quantification of CellRox green signal (mean±s.d., asterisks, two-tailed unpaired t-test, compared to control) (B). (C) Lifespan analysis was performed over a period of 70 days (asterisks, log-rank Mantel-Cox test). Genotypes (A–C): control: w + daGal4/+, linker: w UAS-linker/+ daGal4/+.

Enhancing MERCs increases ROS levels in aged flies and increases lifespan. (A,B) Analysis of mitochondrial ROS levels in young (3-day-old) and aged (30-day-old) controls and linker-expressing flies. Representative confocal images of 30-day-old brains stained with CellRox green are shown (A). The quantification of CellRox green signal (mean±s.d., asterisks, two-tailed unpaired t-test, compared to control) (B). (C) Lifespan analysis was performed over a period of 70 days (asterisks, log-rank Mantel-Cox test). Genotypes (A–C): control: w + daGal4/+, linker: w UAS-linker/+ daGal4/+.

Increasing contacts alleviates alterations in locomotion and lifespan in flies expressing Αβ42 arc

Signalling at mitochondria–ER contact points has been linked to several neurodegenerative disorders including Alzheimer's disease (AD) (Schon and Area-Gomez, 2013). In AD, an abnormal toxic protein in the brain, amyloid-β (Aβ), accumulates and causes neuronal cell death. Drosophila can be used to model AD through the expression of arctic mutant (Glu22Gly) Aβ peptides (Αβ42 arc) in fly neurons. This causes progressive locomotor defects and the premature death of the flies (Crowther et al., 2005). We therefore tested whether forcing mitochondria–ER contacts by linker expression affects the toxic consequences of the expression of Αβ42 arc in flies. We observed that linker expression suppressed climbing defects in flies expressing Αβ42 arc (Fig. 7A). Lifespan analysis confirmed that the decreased lifespan of flies expressing Αβ42 arc was suppressed by linker expression (Fig. 7B). We conclude that forcing mitochondria–ER contacts by the expression of the artificial linker confers protection in this Drosophila model of AD.


Authors Biography & Contact Information

Bio: Robert Berdan is a professional nature photographer living in Calgary, AB specializing in nature, wildlife and science photography. Robert retired from CellNeurobiology research to pursue photography full time many years ago. Robert offers photo guiding and private instruction in all aspects of nature photography, Adobe Photoshop training, photomicrography and macro-photography. Portrait of Robert by Dr. Sharif Galal showing some examples of Robert's science research in the background.


Discussion

Over the past decade, a wide variety of high-performance fluorophores have been developed [49,50]. These reagents exhibit a broad range of physical and spectral properties [51], are capable of targeting proteins or peptides in living or fixed cells [40], and can also be used as indicators of biological dynamics [52]. Combining two or more fluorescent probes offers significantly a higher level of information [25,53,54], but may also lead to signal crossover [9]. Current spectral unmixing tools solve this problem to some extent, but their applicability is usually limited. In this paper, we suggested and experimentally examined an approach by using k-means clustering based unsupervised machine learning as a more flexible alternative to separating mixed images blindly.

There are two major issues with current unmixing tools available to biologists which have highly restricted the spectral resolutions that can be achieved by fluorescence microscopy especially the 2PLSM. Firstly, unmixing methods based on linear inversion calculations, such as linear unmixing [9,11,12,37,55], spectral deconvolution [25,46] and similarity unmixing [26], rely heavily on the cumbersome pre-measurements of emission spectra either through separately recording the spectra of all fluorochromes [26] or manually selecting ROIs with pure labels in the image [9]. Background and autofluorescence, if present, also need to be defined spectrally and treated as additional spectra [11,55], which are even harder to measure or estimate. LUMoS, as it does not directly calculate the abundances of fluorophores, is a completely “blind” unmixing process, and is therefore, much easier to implement and free from those restrictions of acquisition conditions. When background and autofluorescence are present in the sample, additional clusters could be added, and those undesired signals could be separated and removed (Fig 6). Secondly, linear unmixing, Non-negative Matrix Factorization (NMF) [20,56], deconvolution, and Principle Component Analysis (PCA) [57] all require determined (Nfluorophores = Nchannels) or over-determined (Nfluorophores<Nchannels) image acquisition systems, greatly restricting the total number of fluorophores that can be imaged by the hardware design. Although Independent Component Analysis (ICA) does not intrinsically require less fluorophores than detectors, its success for spectral unmixing in fluorescence microscopy has been limited to relatively few independent sources which are usually same or fewer than the number of detectors [58–60]. As LUMoS can be set to create an arbitrary number of clusters for an image, it can be used in under-determined situations (Nfluorophores>Nchannels) for expanding the capabilities of an imaging system (Figs 4 and 5). Moreover, as the readout noise increases with the number of detection channels used [37,58], LUMoS can achieve the high quality unmixing results with as few channels as possible to minimize the readout noise.

Similar but more complicated clustering based methods have been introduced and developed in the field of satellite imaging [29,61,62]. Remote sensing image unmixing is similar to fluorescence image unmixing in many ways, and many unmixing ideas commonly used for microscopy imaging were initially introduced in remote sensing [37]. The ultimate goal of both imaging modalities’ unmixing is to decompose the spectral signature of mixed signals into a set of endmembers and corresponding abundances [38,63]. However, the uniqueness of fluorescence microscopy makes its spectral unmixing task different from remote sensing. First and foremost, the number and type of fluorophores (endmembers) are known in advance in microscopy, which offers a great advantage and simplicity of using clustering algorithms such as k-means for fluorescence image unmixing. Most of the time, the first step of remote sensing image unmixing is to determine endmember [38,64], and many of the advanced unmixing algorithms have been focused on how to better estimate the number and characteristics of endmembers, such as adaptive possibilistic clustering [62] and neural network autoencoder [65]. Second, due to the chemical mixtures of landscape objects, the abundance of one pixel from a satellite image normally comprises fractions of each endmembers, thus remote sensing image unmixing methods output abundances for each pixel as fractions of different chemical components [38,63]. However, in fluorescence microscopy, biologists usually assume a distinct labeling of a structure by one specific fluorophore, unless colocalized labeling was designed. The goal of fluorescence image unmixing is more towards unambiguously distinguishing each labeled structure rather than decomposing each pixel into many different chemical components. Therefore, using classification based hard clustering, such as LUMoS, by assuming one pixel per fluorophore is more appropriate in the field of fluorescence imaging and the results of which are more interpretable for biologists. Third, remote sensing images have hundreds of spectral bands which is usually much more than the number of endmembers, making linear algebra based unmixing methods, such as linear unmixing, NMF, and deconvolution, better suited [38,63,64]. Because fluorescence microscopes have much fewer detectors (usually ≤4), many unmixing methods applied for remote sensing are insufficient for fluorescence imaging with potentially more fluorophores than detectors. In considerations of those features of fluorescence imaging, we applied k-means clustering as a simple, easy-to-use, and flexible method for microscopy image unmixing.

The implications of k-means clustering are usually limited by the difficulties in choosing an optimal number of clusters, “k” [32,66]. However, in the case of fluorescence microscopy, “k” is known and determined by the number of fluorochromes used, making k-means clustering a well-suited method for spectral unmixing. Usually, the “k” is set to be the total number of fluorophores plus one (considering the background noise) (examples in Figs 3–5). When special circumstances happen, options are available to optimize the “k” to tailor LUMoS for different cases. For example, when there are known colocalization labeling or autofluorescence structures (Fig 6), additional clusters could be added by considering colocalization and autofluorescence as distinct “fluorophores”. When applying LUMoS, carefully examining the image data to better determine “k” in advance may improve the unmixing results.

There are also limitations of our algorithm, especially when unique circumstances are associated with the imaging data. As demonstrated in the simulations, our approach may cease to be useful when it misclassifies a significant portion of the pixels belonging to a fluorophore of interest. This can occur when there are relatively unbalanced structure sizes, significantly overlapping emission spectra, and a low SNR. Additionally, although considering the information of nearby pixels by using a median filter, LUMoS still does not take any spatial information at biological structure level into account so its clustering ability is limited to classifying individual pixels rather than whole structures as some other methods attempt [67,68], and may fail when two fluorophores have very similar signatures (S4 Fig). In one paper [67], total variation regularization was combined with sparse regression to consider spatial-contextual information during remote sensing image unmixing. Sparse regression (more commonly used for remote sensing data) requires a known spectral library which is hard to obtain for biological microscopy, and is not required by LUMoS. Another spatial-spectral unmixing algorithm was proposed and successfully applied for biological microscopy imaging by using dictionary learning to separate spectrally close but morphologically different structures [68]. However, single-stained reference images were required to learn the morphological information and generate the dictionary. These reference images can be time-consuming to collect and sample specific. Future improvements to LUMoS may introduce a spatial regularizer [67] or a morphology dictionary [68] to further enhance the robustness of the algorithm, while still maintaining the advantage of the blindness of k-means. LUMoS specifically assumes the abundance of each fluorophores is binary at pixel level, which produces unambiguous classification of individual fluorophores. If there is colocalization at structure scale, for example one structure labeled with more than one fluorophore, the colocalization group can be treated as an additional cluster to be separated and analyzed (Fig 6). However, implicit in our unmixing algorithm is the assumption that a pixel represents an exclusive single label without considering nano-scale colocalization due to the imaging spatial resolution limitations. This assumption is valid for spatially well-dispersed fluorescent structures relative to the imaging resolution, but may not hold when two labeled structures are contacting or too close to each other. We expect future improvements by adding the options of fuzzy clustering [69,70] or overlapping k-means [71] to extend the flexibility of LUMoS when there are nano-scale colocalization considerations.

In conclusion, we presented a blind and flexible tool for fluorescence image spectral unmixing—LUMoS. Both experimental and synthetic results demonstrated its ability to robustly separate mixed fluorophores in terms of the quality of results and ability to converge in a variety of scenarios. The LUMoS method has also greatly expanded the fluorophore options beyond the number limit of detectors and excitation lasers. These qualities make LUMoS a simple, general, and reliable spectral unmixing approach to quickly apply to any fluorescence images. Last but not least, an optimal strategy for spectral unmixing should always combine image processing algorithms with careful dye selections and rigorous image acquisitions. LUMoS can be coupled with spectral imaging or other hardware designs to yield excellent multi-color imaging results, and will offer new avenues for understanding the complex biological organizations.



Comments:

  1. Dagonet

    What a graceful question

  2. Jeriel

    no-no-no-no-no time for me to communicate with you here, I'll go dunu grass

  3. Grokazahn

    Who knows



Write a message