SEARCH

SEARCH BY CITATION

Keywords:

  • Image Cytometry and Sorting;
  • confocal microscopy;
  • optical sectioning;
  • photochromic fluorescence resonance energy transfer;
  • photogelation

Abstract

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUDING REMARKS
  6. Acknowledgements
  7. LITERATURE CITED

Background

Innovative thinking and experimentation were the hallmarks of Mack Fulwyler's approach to research. This report summarizes some of the ideas and their early realizations that he pursued in the field of imaging cytometry, work that was not published before his untimely death, although he composed the initial draft of this report.

Methods

Included are related experiments implemented in the programmable array microscope (PAM) devised for patterned illumination and detection, the instrument that Mack Fulwyler employed during a sabbatical leave in Göttingen in 1998. Despite being the originator of instrumentation for flow cytometry and sorting, Mack Fulwyler was intensely interested in imaging systems, recognizing their ability to resolve cellular details obscured by the whole cell signals generally acquired in flow. At one point, these interests merged with those of two other authors (I.T.Y. and T.M.J.), leading to the Image Cytometry and Sorting (ICAS) strategy and project. A major goal was uncomplicated rare cell detection and isolation using a sequential process of cellular labeling via suitable probes, whole field imaging, and selective area-restricted photoinduced reactions designed to encapsulate and/or chemically or physically tag cells in a manner permitting subsequent fractionation by bulk techniques.

Results and Conclusion

This publication features photoinduced polymerization, photodecaging, photoactivation, and photochromic conversion reactions carried out by Fulwyler and/or the other authors with the PAM, employing operator designated patterns and locations in various samples. Photopolymerization of polyethylene glycol-diacrylate to a gel-like structure allowing the specific selection of objects (cells) for further analysis and processing techniques was the approach explored personally by Mack Fulwyler in relation to the ICAS concept. © 2005 International Society for Analytical Cytology

Programmable array microscopes (PAMs) are a class of instruments in which a spatial light modulator (SLM) is placed in the primary image plane to create patterns of illumination and/or detection. The evolution of PAMs has encompassed a variety of imaging concepts and technical realizations: confocal microscopy (1–9), imaging spectroscopy (10–12), deconvolution of conjugate and nonconjugate images (9), optical sectioning integrated with imaging spectroscopy (13), and lifetime imaging combined in hybrid systems with optical sectioning (14) or imaging spectroscopy (15). In these implementations, micromirror devices (1–9, 16), nematic liquid crystals (10), and ferroelectric liquid crystals (7) have served as SLMs for control of the patterns of illumination and detection. Characteristic of PAMs are the absence of macroscopic moving parts and software control of the (arbitrary) patterns of illumination and detection. Thus, the SLM may be programmed for very diverse requirements, such as in endoscopy (5, 17), profilometry (18, 19), point spread function engineering (20), adaptive correction of aberrations (21), and photodirected oligonucleotide synthesis (22).

More recent PAM-based studies have involved novel implementations of fluorescence recovery after photobleaching in diffusion measurements (23), patterned photochemistry facilitating patterned cell adhesion to substrates (24), fringe generation and projection (8), and three-dimensional absorption tomography (25). In these applications, the PAM does not merely constitute an imaging system but also functions as an extremely versatile “workbench” with which diverse light-induced manipulations of the sample can be accomplished.

The prototyping Image Cytometry and Sorting (ICAS)-related PAM experiments performed by Mack Fulwyler at the Max Planck Institute for Biophysical Chemistry in 1998, well before the publication of the related studies cited above, are included in the studies featured in this work: (a) patterned photodecaging, equivalent to an “inverse” fluorescence recovery after photobleaching experiment, to observe translocation of high-molecular-weight dextran in the cytoplasm; (b) photoinitiated generation of a gel for encapsulating cells; and (c) reversible patterning of a photochromic sample.

MATERIALS AND METHODS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUDING REMARKS
  6. Acknowledgements
  7. LITERATURE CITED

Programmable Array Microscope: Workbench

Experiments were carried out in a PAM (4) with a dual-sided relay system (9). Two configurations of the PAM system were used (Figs. 1 and 2). The first employed only a single side of the PAM, whereas the second approach featured optical paths, one used for imaging and the other for generating patterned ultraviolet (UV) light (Fig. 2). The PAM consisted of a previously described add-on module installed on a Nikon E-600 microscope with a modified tube lens (4). A digital micromirror device (DMD) was used as the programmable array unit (SLM) to define regions of patterned UV illumination for inducing the photogelation, photodecaging, and photochromic reactions. The DMD was mounted in the primary image plane and the light from “on” elements was relayed alternately from a 250-W super high pressure mercury arc lamp system (Lumatec, Munich, Germany) to the object plane of the microscope or from the object plane to a CCD camera (KX-2, Apogee Instruments, Auburn, CA). Both halves of the optical relay incorporated 1:1 all-reflective achromatic imaging optics (ASE Optics, Honeyoye Falls, NY, USA) based on an Offner triplet design. Further optical details are given elsewhere (9). An automated xy-motion control system (Märzhäuser, Wetzlar, Germany) allowed samples to be reproducibly repositioned after photochemical gelation, removal from the microscope, rinsing, and reinstallation in the microscope.

thumbnail image

Figure 1. PAM workbench configured for spatially selective photodecaging. In this configuration, only a single side of the system was used. UV light was introduced using illumination source A. Illumination source B was used for dual-probe (fluorescein and Alexa546) excitation and detection. The DMD provided configurable illumination patterns to ROIs. After being deflected by the DMD, light is refocused into the microscope object plane, where it is spatially restricted to initiate selective photochemical reactions. This system has the advantage of allowing the ROI to be viewed in a convenient and versatile manner. To achieve this end, fluorescence or transmission light images are recorded after traversing the DMD and being relayed in an image-preserving manner to a CCD camera.

Download figure to PowerPoint

thumbnail image

Figure 2. PAM workbench configured for selective gelation. For these experiments both sides of the PAM were employed. The “gelation” (or, more generally, photoconversion) path was used to avoid the formation of a gel “corral” surrounding the field of interest. The DMD device was driven using electronics provided by Texas Instruments and was larger than the 640- × 480-pixel region that could be actively driven by the default electronics. The remaining portion was “tied” to the off position, giving rise to a gel border surrounding the entire field. Functions of the excitation sources were optically sectioned fluorescence (PAM; excitation source A), nonconfocal fluorescence or photoreaction (excitation source B), and patterned photoreaction (PAM; excitation source C).

Download figure to PowerPoint

Cells

3T3 mouse fibroblast (3T3 Balb/x) cells were grown in Dulbecco's Modified Eagle's Medium (GibcoBRL, Grand Island, NY, USA) containing 10% fetal bovine serum and antibiotics. The cells were passaged 1 day before the experiments and grown on Cellocate coverslips (Eppendorf, Hamburg, Germany) to less than 30% confluency for microinjection. Chinese hamster ovary cells and Chinese hamster ovary cells stably transfected with an erbB1-eGFP fusion protein (26) were used in photogelation experiments; erbB1 is the epidermal growth factor receptor. The cells were mixed in various ratios and cultured in microwell slides that contained Dulbecco's Modified Eagle's Medium.

Photodecaging

Cells were microinjected with caged fluorescein dextran (molecular mass 70 kDa; Molecular Probes, Eugene, OR, USA; Invitrogen, La Jolla, CA, USA) and an Alexa546-conjugated polyclonal F(ab)2' (Molecular Probes). The Alexa546-F(ab)2' was coinjected to allow rapid identification of injected cells and to optimize the plane of focus for subsequent decaging and observation. The concentrations of injected probes were 60 μM for the caged dextrans and approximately 0.3 mg/ml for the antibody. After injection, the cells were allowed to recover by incubation in regular culture medium for at least 1 h, after which they were mounted in a closed customized perfusion chamber with a volume of ∼700 μl using Hepes buffered saline as the imaging medium.

Images based on Alexa546 immunoglobulin G (IgG) immunofluorescence were recorded using a Nikon 40×, 0.6 numerical aperture, Plan Fluor objective, a 546 ± 10 nm excitation filter, a 575-nm dichroic filter, and a 580-nm longpass filter combination (Nikon). The Alexa546 images served to document the entire field and the locations of the illuminated (for photoreaction) regions of interest (ROIs). After establishing the location of a set of injected cells, an ROI was programmed on the DMD, defining both the region from which the fluorescence originated and reached the camera and the region that was exposed to illumination from the UV lamp. The fluorescein signal before decaging and after each decaging step was acquired selectively with the filter set: 450-490-nm excitation, 505-nm dichroic, and 520-560-nm emission.

Photogelation

Gel components

Initiation of photopolymerization of polyethylene glycol diacrylate (PEG-DA) was achieved with a solution of 2,2-dimethoxy-2-phenyl acetophenone dissolved in N-vinylpyrrolidone (kind gift of Jeffrey Hubbell) (27). A 20% (w/v) stock solution of PEG-DA was diluted to 1% in Hepes buffered saline immediately before use and 1.6 μl of the photoinitiator per milliliter of solution was added. Photopolymerization was induced by exposure to the 365-nm line of the 250-W super high pressure mercury arc lamp system; this light was introduced through the second input port of the PAM system.

Cells

A mixture of fluorescent and nonfluorescent cells was prepared by mixing erB1-eGFP expressing cells with hybridoma cells in a ratio of 1:10. The cell suspension was centrifuged, the supernatant was removed, and the cells were resuspended in 500 μl of polymer solution. After the polymerization step, the coverslip was removed and the slide immersed in phosphate buffered saline for 1 h before washing in a vigorous stream of phosphate buffered saline to remove unpolymerized reagent. The sample was then inserted into the microscope for imaging.

Photochromic Conversion

Disodium-6-amino-2-{5-[3-(4-{4-[3,3,4,4,5,5-hexafluoro-2-(2-meth-oxy-benzo[b]thiophen-3-yl)-cyclopent-1-enyl]-3,5-dimethyl-thiophen-2-yl}-benzoylamino)-propionylamino]-pentyl}-1,3-dioxo-2,3-dihydro-1H-benzo[de]isoquinoline-5,8-disulfonate was synthesized as described previously (28). This model compound consists of a Lucifer Yellow donor coupled to a diarylethene acceptor. Half a milligram was dissolved in 50 μl of methanol and applied on a RPC-18 TLC plate cut to the size of a microscope slide. Irradiation with the mercury arc lamp for conversion from the open (colorless) acceptor form to the closed (colored) acceptor form was at 365 nm and photoreversion was achieved by exposure to 546-nm light. The donor fluorescence was monitored by excitation at 450 to 490 nm. In the closed form, the donor was quenched by the process denoted as photochromic fluorescence resonance energy transfer (pcFRET) (28, 29). In pcFRET systems, the nonabsorbing open form of the acceptor is not FRET competent; thus, the donor exhibits its unquenched natural emission intensity.

RESULTS AND DISCUSSION

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUDING REMARKS
  6. Acknowledgements
  7. LITERATURE CITED

Single-Cell Activation

The series of images presented in Figure 3 documents selective photodecaging of caged fluorescein dextran (70 kDa) in single cells. The 21 panels of this figure are displayed in seven rows and three columns. Column 1 depicts the ROIs that were active at each decaging step, and columns 2 and 3 document the fluorescein and Alexa546 channels, respectively. Row 1 corresponds to images before illumination of the ROI with UV light; the fluorescence intensities were negligible. Row 2 shows fluorescence from a single cell that had received UV exposure. The entire cell fluoresced, indicating that at an exposure longer than 120 s, the labeled dextran was freely mobile within the cell. An additional 120 s of illumination led to increased intensity (row 3). In row 4, a new, smaller ROI was defined. As a consequence, the affected cell was seen to fluoresce less brightly, although the intensity increased upon dilation of the ROI (row 5). In row 6, a fourth ROI was defined, which resulted in the observation of fluorescence from an additional cell; in row 7, the entire field was exposed to UV light.

thumbnail image

Figure 3. Selective photodecaging of single cells. A field consisting of three cells was observed during the selective photoactivation of caged 70-kDa fluorescein dextran microinjected with Alexa546-F(ab)2'. The ROI subjected to UV illumination is shown in the left column of images, and the middle and right columns correspond to images based on whole field fluorescein and Alexa546 signals, respectively. The ROI image was generated using the filters of the Alexa546 channel. The latter provides visualization of the microinjected cells and serves to indicate the location of the ROI. Images in row 1 represent a field of three cells with little or no signal in the fluorescein channel (intensity rescaled for better visibility). An ROI created to illuminate a segment of a cell (row 2) resulted in a bright fluorescence throughout the cytoplasm of the cell, but in contrast to the Alexa546 channel no fluorescence was observed in the nucleus, which was inaccessible to the high-molecular-weight dextran. The subsequent rows illustrate a sequence of photoactivation steps continuing until all microinjected cells were visualized. See text for further details.

Download figure to PowerPoint

Intracellular Movement of Photoactivated Dextrans

The series of images in Figure 4 documents the translocation of photodecaged fluorescein-labeled dextrans within a single cell. This figure consists of nine images. Column 1 depicts a single cell observed in the Alexa546 (IgG) channel (row 1), the position of the ROI used for UV exposure for photodecaging (row 2), and the appearance of the entire cell after 120 s of exposure to UV light (row 3). Column 2 documents the distribution of fluorescein-labeled dextrans inside the cell immediately after 10 s of UV exposure (row 1), 1 min after exposure (row 2), and the difference between the two images (row 3). Column 3 shows similar data acquired after a 20-s UV exposure. In both cases, the decrease of the signal within the ROI was accompanied by an increase in the remainder of the cell (with the exception of the nucleus).

thumbnail image

Figure 4. Movement of fluorescein-labeled dextrans in a single cell. A time series was recorded showing the redistribution of photoactivated (decaged) 70-kDa fluorescein dextran from a small region of a cell. In the first experiment (middle column) the ROI was exposed to UV light for 10 s. An image was taken immediately after exposure and a second image was taken 1 min later. The difference image (bottom) shows transport of dextran across the cell on time scale of a few minutes. The right column presents a similar experiment in which a 20-s decaging step was employed. The left column of images depicts reference conditions: Alexa546-IgG channel (top; all other images in the figure were taken with the fluorescein channel); ROI (middle); and fully decaged field (bottom).

Download figure to PowerPoint

Selective Gelation and Adhesion of Beads

Fluorescent beads were introduced into the PEG-DA solution and a series of photogelation experiments were carried out to determine the ability of the gel to entrap objects (Fig. 5). Efficient photogelation occurred by using a low magnification, low numerical aperture objective (2×), and an objective with higher magnification (20×) and numerical aperture (0.75). These experiments demonstrated that beads could be selectively trapped by gelation, although some residual motion of the beads within the gel was observed. The slight expansion of the gelation ROI was attributable to diffusion of the initiator. Similar experiments using a 1:10 mixture of fluorescent to nonfluorescent cells demonstrated the ability to selectively capture “marked” (fluorescent) cells within a mixed population (data not shown).

thumbnail image

Figure 5. Photoinitiated gelation to entrap selected beads in a field. An initial fluorescence image of the field revealed 19 beads (left column, top image). An illumination ROI encompassing nominally five beads (left column, middle image) resulted in three beads remaining entrapped after the non-gelled material was washed off (left column, bottom image). Although it was evident that gelation had taken place in the ROI, considerable movement of beads inside the gel was observed. A similar behavior was observed in other experiments (middle and right columns; the ROI delineations have been omitted). In the second series, the beads were seen to move slightly between the whole field and the ROI images and further movement occurred during the rinse procedure. The lower right image shows the shape of the gelled region. The beads are overexposed but scattering off the gel clearly demarcates the edges.

Download figure to PowerPoint

Two-cell isolation strategies related to the ICAS strategy were envisioned based on these results. In the first, a rare cell having a desired characteristic could be selectively attached to a microscope slide via the photogelation reaction. Alternatively, unwanted cells could be attached to a slide and the rare cell washed off. The imaging requirements for such systems have also been addressed by the construction of a fast ICAS scanner (30).

Generation of Patterns by Photochromic Photolithography

A field of a sample in the nonabsorbing FRET-incompetent open acceptor form was irradiated with UV light in a square line pattern loaded into the DMD. The resulting fluorescence image (Fig. 6) showed the same distribution, albeit with inverted intensity (“on” pixels [RIGHTWARDS ARROW] dark lines) due to conversion to the closed acceptor form that quenched donor emission via resonance energy transfer (the principle of pcFRET) (28), i.e., conversion from the open to the closed form constituted a photophysical mask. A major virtue of the photochromic process is its reversibility, allowing arbitrary patterns to be generated, erased, or reformed at will. Although in Figure 6 only spectroscopic properties were exploited for the generation of a pattern, photochromic conversions are much more versatile, i.e., can be induced between molecular forms differing dramatically in mechanical (bending or twisting, gel-sol transitions) or other physicochemical properties such as charge distribution (31).

thumbnail image

Figure 6. Patterning by photochromic conversion (pcFRET). A TLC plate upon which the model donor-acceptor photochromic pair Lucifer Yellow (donor) and a diarylethene (acceptor) had been deposited was exposed to a rectangular line pattern of UV light by programming the DMD. Photoconversion to the closed, absorbing form of the acceptor quenched the Lucifer Yellow, thereby generating a negative (quenched) line pattern. Two images of different fields and with different scales are shown. The pattern could be erased by exposure to green (546 nm) light, restoring the original unstructured fluorescence surface (data not shown).

Download figure to PowerPoint

CONCLUDING REMARKS

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUDING REMARKS
  6. Acknowledgements
  7. LITERATURE CITED

The experiments featured in this report demonstrate the great flexibility and potential of structured illumination/imaging systems as light-based tools applied to photoreactions for generating and/or abolishing physical structures. The classic distinction between analytical and preparative devices is blurred in such situations. These techniques not only have great potential for studies of cellular state and function, particularly for exploiting spectroscopic modalities such as FRET, but also for technologic applications such as photolithography and photodeposition.

The challenges to this emerging technology include (a) increasing the efficiency and speed of photoreaction, (b) devising reagents less dependent on the use of near-UV irradiation, and (c) implementing automated sample handling procedures. Had Mack Fulwyler been allowed to extend his productive life to a natural end, there can be no doubt that he would have contributed prodigiously to further development of structured illumination and PAMs for medical analysis and diagnosis, just to mention one area of great personal interest and concern to him. It remains for us, his survivors, to take up the challenge.

Acknowledgements

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUDING REMARKS
  6. Acknowledgements
  7. LITERATURE CITED

The PEG-DA and photoinitiator were generously provided by Dr. Jeffrey Hubbell, Institute for Biotechnology and Bioengineering, Lausanne, Switzerland. We thank Stefan Hoeppner for assistance in the preparation of this report. This study was supported in part by the Volkswagen Foundation for Development and Application of Photochromic Compounds (grant I/77 897 to E.A.J.-E. and T.M.J.), the pcFRET (E.A.J.-E. and T.M.J.), the Agencia Nacional de Promoción de la Ciencia y Tecnología (E.A.J.E.), the Secretaría de Ciencia, Tecnología e Innovación Productiva (E.A.J.-E.), Universidad de Buenos Aires (E.A.J.-E.), the German-Argentine DLR-BMBF-SECyT program (E.A.J.-E. and T.M.J.), and NATO for support of the ICAS project (grant RG.85/0324 to M.F., I.T.Y., and T.M.J.).

LITERATURE CITED

  1. Top of page
  2. Abstract
  3. MATERIALS AND METHODS
  4. RESULTS AND DISCUSSION
  5. CONCLUDING REMARKS
  6. Acknowledgements
  7. LITERATURE CITED