SEARCH

SEARCH BY CITATION

Keywords:

  • photoactivation;
  • PAGFP;
  • PSCFP2;
  • KikGR;
  • Kaede;
  • embryo;
  • chick;
  • cell tracing;
  • confocal imaging

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Tracing the lineage or neighbor relationships of cells in a migratory population or deep within an embryo is difficult with current methods. The recent explosion of photoactivatable fluorescent proteins (PAFPs) offers a unique cell labeling tool kit, yet their in vivo performance in intact embryos and applicability have not been thoroughly explored. We report a comparison study of PAGFP, PSCFP2, KikGR, and Kaede analyzed in the avian embryo using confocal and 2-photon microscopy. PAFPs were introduced into the chick neural tube by electroporation and each photoconverted in the neural crest or cells in the neural tube with exposure to 405 nm light, but showed dramatic differences in photoefficiency and photostability when compared at the same 2% laser power. KikGR and Kaede photoconverted with ratios only slightly lower than in vitro results, but cells rapidly photobleached after reaching maximal photoefficiency. PSCFP2 had the lowest photoefficiency and photoconverted nearly 70 times slower than the other dual-color PAFPs tested, but was effective at single-cell marking, especially with 2-photon excitation at 760 nm. The dual-color PAFPs were more effective to monitor cell migratory behaviors, since non-photoconverted neighboring cells were fluorescently marked with a separate color. However, photoconverted cells were limited in all cases to be visually distinguishable for long periods, with PSCFP2 visible from background the longest (48 hr). Thus, photoactivation in embryos has the potential to selectively mark less accessible cells with laser accuracy and may provide an effective means to study cell–cell interactions and short-term cell lineage in developmental and stem cell biology. Developmental Dynamics 236:1583–1594, 2007. © 2007 Wiley-Liss, Inc.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The proper organization of structures within an embryo depends on subpopulations of cells to exchange local neighbors or migrate extensively to specific target locations. A major challenge in biology is to gain insight into cell behaviors and cell lineage with the goal to understand underlying molecular mechanisms of pattern formation. Yet, the current ability to accurately, fluorescently mark and trace cell behaviors and their progeny at selected times and locations within intact embryos remains difficult. Current cell-marking strategies in embryos have benefited from delivery techniques ranging from fluorescent chemical dye injection to single and multiple cell electroporation of a fluorescent protein (Bronner-Fraser and Fraser, 1988; Itasaki et al., 1999; Haas et al., 2001; reviewed in Stern and Fraser, 2001). Each of these techniques is very effective when the cell or tissue of interest is accessible to glass needle penetration and/or injection or manipulation, an aspect that limits selective cell marking of subgroups of migratory cells or cells deeper within embryos. Transgenic technology, including the generation of visually striking multicolor-labeled mice (Hadjantonakis et al., 2003), and nuclear-targeted fluorescent proteins offer significant advantages for more accurate cell tracking (Hadjantonakis and Papaioannou, 2004). However, it is still challenging to selectively label single and multiple cells at specific developmental times. Thus, there is a need to develop a fluorescent-labeling technique that allows for selective cell marking at various developmental times for use in a wide range of embryo models.

The recent work on naturally occurring and engineered green fluorescent protein (GFP) and GFP-like photoactivatable fluorescent proteins (PAFPs), including PAGFP (Patterson and Lippincott-Schwartz, 2002), PSCFP2 (Chudakov et al., 2004), KikGR (Tsutsui et al., 2005), and Kaede (Ando et al., 2002), offer the possibility to use UV and violet light to fluorescently mark cells. PAGFP, developed from wildtype GFP (Chalfie et al., 1994), has a 100-fold increase in fluorescence intensity after near UV excitation, and is widely used to study intracellular dynamics in a number of cultured cell types (Lippincott-Schwartz and Patterson, 2003). Kaede is a natural PAFP, found in the coral Trachyphyllia geoffroyi, that undergoes an irreversible photoconversion of the green fluorescent protein fluorophore to red upon irradiation with UV light (Ando et al., 2002). After photoactivation in HeLa cells, Kaede has a 40-fold increase in the ratio of photoconverted red-to-green (R/G) fluorescence signal (or ∼2,000-fold increase in R/G ratio when divided by the initial R/G ratio prior to photoconversion) (Ando et al., 2002). Based on the photoconversion of Kaede, KikGR was genetically engineered from the coral Favia favus to undergo a similar green-to-red photoconvertibility (Tsutsui et al., 2005). KikGR is reported to be more photoefficient and photostable than Kaede in HeLa cells expressing KikGR and suitable for use at 37°C (Tsutsui et al., 2005). Lastly, PSCFP2 is a monomeric PAFP that undergoes irreversible photoconversion from a cyan-to-green fluorescent form in response to near UV light (Chudakov et al., 2004). Photoconversion of PSCFP2 has been demonstrated with an increase in fluorescence contrast of more than 1,500-fold in the green-to-cyan fluorescence ratio in E. coli and HeLa cells (Chudakov et al., 2004). Thus, a tool kit of PAFPs has emerged (Miyawaki et al., 2003; Lukyanov et al., 2005). However, their in vivo performance in intact embryos and suitability to cell tracing and other questions in developmental biology is largely unknown.

Here, we performed a comparison study of the fundamental photophysical properties of PAGFP, PSCFP2, KikGR, and Kaede in an avian embryo model. We took advantage of the chick embryo as a model system because of the ease of delivery of PAFPs into cells, accessibility to intravital imaging, and relatively rapid development rate. We were able to adapt our previously developed protocol for confocal photoactivation of PAGFP in single cells in embryos (Stark and Kulesa, 2005) for each PAFP. We measured the time to photoactivation, photoefficiency, and photostability of each of the four PAFPs and compared the results to in vitro data. We also compared the accuracy to fluorescently mark single migratory and densely packed cells using single and two-photon microscopy. We determined the length of time over which we could distinctly identify photoactivated versus non-photoactivated cells in re-incubated chick embryos. We tested the applicability of each PAFP to specific questions, and provide examples that show their strengths and limitations.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Photoactivation of PAGFP, PSCFP2, KikGR, and Kaede Was Reproduced in Cells Within Intact Avian Embryos

Our initial goal was to reproduce the photoactivation of PAGFP, PSCFP2, KikGR, and Kaede in vivo within an intact embryo model. The early avian embryo allowed for ease of delivery of each photoactivatable fluorescent protein (PAFP). Each PAFP was introduced separately into the lumen of the chick neural tube by injection and electroporation (Fig. 1A,B) and embryos were reincubated until the fluorescence signal could be detected (Fig. 1C). Each PAFP was photoconverted in ovo using confocal laser scanning with 405-nm excitation in cells within the neural tube (∼20 μm deep) and migrating cells en route to the developing second branchial arch (<70 μm deep) (Fig. 1C). Changes in fluorescence signals within cells were evaluated (Fig. 1D–S). The cell labeling was confined to the neural tube and premigratory neural crest cells (NCCs). NCCs are highly migratory and their invasion into unlabeled surrounding tissue made it easier to clearly evaluate the photoactivation process within individual cells. No cell death from 405-nm exposure was observed as seen in our Tunel staining in previous studies (Stark and Kulesa, 2005). For each PAFP, we also performed a lambda scan within a defined wavelength range to measure the emission spectrum of the PAFP during photoactivation. As a means of comparing the fundamental photophysical properties of each PAFP, the initial experiments were performed with constant imaging parameters, including the same laser power setting (2% laser power or 44 uW measured at the sample).

thumbnail image

Figure 1. Photoactivation of PAGFP, PSCFP2, KikGR, and Kaede is reproducible in cells within intact avian embryos. AC: The technique for injection, delivery, and photoactivation of each photoactivatable fluorescent protein into chick embryos is shown. C: Neural crest cells are targeted for in ovo photoactivation emerging from the r4 region of the neural tube and migrating to the branchial arch (BA). The depth of the cell can vary from 20 μm at the neural tube to 70 μm in the BA. D–S: Photoactivation of each fluorescent protein was recorded and a line profile of the fluorescence intensities within the cell was generated (dotted line, arrowhead at end of profile). The photoactivated region was contained to the field of view of each image shown, including the cell and its local neighborhood. D,F: In a typical embryo, PAGFP expression is shown in a non-photoactivated cell (D; outline) and (F) corresponding GFP intensity (excited at 488 nm) along the line profile (Ch1). E, G: After photoactivation, GFP intensity within the cell increases (G; Ch1). H,I: In a typical embryo expressing PSCFP2, the spectral shift from cyan to green fluorescence is shown. J,K: Fluorescence intensity values before and after photoactivation with 405 nm light show the shift from cyan (J; Ch1, excitation at 405 nm) to green (K; Ch2, excitation at 488 nm) fluorescence within the cell. LS: KikGR and Kaede expressed in embryos show a photoconversion from green (N,O; Ch1) to red (R,S; Ch2) fluorescence and flip in intensity profiles after photoactivation and excitation at 488 and 561 nm, respectively. For each photoactivatable fluorescent protein tested, multiple cells in n = 12 embryos were analyzed and each cell is approximately 30–40 μm across.

Download figure to PowerPoint

For PAGFP, excitation with 405 nm light caused the photoactivation of a single color PAFP and increase of the GFP fluorescence signal within the cell (Fig. 1D–G). Evaluation of a typical photoactivated cell with 488 nm light after 405-nm excitation showed an increase in mean GFP fluorescence signal (Fig. 1F,G; Ch1). Lambda scanning showed that the emission spectrum reaches a maximum single peak within the region, confirming the single color fluorescent form of PAGFP photoactivation with 405 nm light (Fig. 2A). In contrast to a single color fluorescent form, analysis of the in ovo photoconversion of PSCFP2 showed that the fluorescence intensity ratio of mean cyan fluorescence to green fluorescence properly inverted after the photoconversion process (Fig. 1H–K). PSCFP2 underwent photoconversion in a cell, expressed in a decrease in cyan fluorescence (Fig. 1J,K; Ch1) and appearance of a 490-nm excitation peak with emission maximum at 511 nm (Fig. 1J,K; Ch2). The increase in mean GFP fluorescence was evaluated in a cell by exciting with 488 nm light after each 405-nm excitation scan (Fig. 1J,K; Ch2). Lambda scanning showed that the emission spectrum consists of two peaks—a peak at a low wavelength (cyan fluorescence) decreasing in magnitude, and a peak at a higher wavelength (green fluorescence) increasing in magnitude—confirming the dual color fluorescent forms of PSCFP2 photoconversion (Fig. 2D). For KikGR and Kaede, there was a distinct appearance of green-to-red fluorescence within a cell exposed to 405 nm light (Fig. 1L,M, P,Q). Evaluation of a Kaede- or KikGR-photoactivated cell with 488 and 543 nm light after 405-nm light exposure showed a decrease in mean green fluorescence (Fig. 1N,O, R,S; Ch1) and an increase in mean red fluorescence (Fig. 1N,O, R,S; Ch2). In both cases, there was a significant change in the red-to-green fluorescence ratio after photoconversion (Fig. 1N,O,R,S; Supplemental Movie S1). Lambda scanning confirmed the rapid increase in a higher wavelength peak (red fluorescence) and decrease of a lower wavelength peak (green fluorescence), confirming the dual-color fluorescent form of KikGR and Kaede (Fig. 2G,J). Thus, the single fluorescent form photoactivation and dual-fluorescent form photoconversion was recapitulated in cells within an intact avian embryo.

thumbnail image

Figure 2. Photoactivation with consistent parameters affects each PAFP differently. A: Lambda plus time scanning of PAGFP-labeled cells within a typical chick embryo shows the effects of continuous 405-nm laser light excitation on the emission spectra, that is, a monotonic increase in the intensity with no spectral shift. B: Over a long period (>14 min) of low 405-nm laser power (2%) exposure to a cell, the GFP fluorescence intensity (Ch1; excited at 488 nm and plotted as the average intensity for a region of interest around the photoactivated cell) reaches an asymptotic value (the length of each set of scans is approximately 6.3 sec). C: The time sequence of photoactivation (2%, 405nm laser power) of PAGFP in a typical embryo. D: The lambda scan in PSCFP2 labeled embryos shows the small CFP peak in the 400-nm range in which the emission spectra over time (continuous excitation with 405 nm light) shift towards the 500-nm range. E: The line graph plots the average intensity value for a region of interest created around the photoactivating cell (Ch1, dark blue line represents the cyan protein decrease; excitation with 405 nm light; Ch2, the light blue line represents the increase of green converted protein; excitation with 488 nm light). F: The time sequence of photoactivation (2%, 405-nm laser power) of PSCFP2 in a typical embryo. GL: KikGR- and Kaede-labeled embryos both show a shift of emission and excitation from green (H, K; Ch1) to red (H,K; Ch2), excited at 488 and 561 nm, respectively, and rapid photoconversion. I,L: The time sequence in I KikGR- and L Kaede-photoconverted cells shows the rapid change in fluorescent forms from green to red. For each photoactivatable fluorescent protein tested, multiple cells in n = 12 embryos were analyzed.

Download figure to PowerPoint

Photoactivation With Consistent Parameters Affected Each PAFP Differently

The temporal aspects of the photoactivation process were analyzed to compare the in vivo photophysical properties of PAGFP, PSCFP2, KikGR, and Kaede-labeled embryos (Fig. 2). To create a fair comparison, we maintained the 405-nm laser power consistently at 2% (44 uW at the sample), used the same objective (40×/NA = 0.75), and recorded continuous excitation laser scans at the same pixel dwell time (2.56 us per pixel at 512 × 512) (Fig. 2). In ovo photoactivation of PAFP-labeled cells with 405 nm light was found to affect each PAFP differently. The fluorescence intensity of GFP signal in PAGFP-photoactivated cells monotonically increased over time and reached an asymptotic value near 200 (Fig. 2B). The photoactivation process was very rapid and visible after only a single 405-nm excitation scan (∼6 sec), and increased over time to expose an entire photoactivated cell (Fig. 2C). Nearly complete photoactivation of PAGFP took a relatively long time to reach (Fig. 2B) as measured by a maximum fluorescence intensity of GFP signal (∼70 excitation scans at 405nm over ∼7.5 min). In contrast, the photoconversion of PSCFP2 was a relatively slower process (Fig. 2E,F). During the photoconversion of PSCFP2, the mean cyan fluorescence signal (Fig. 2E; Ch1) decreased monotonically as the mean green fluorescence signal (Fig. 2E; Ch2) increased monotonically. The cyan-to-green fluorescence ratio was equal to 1 (photoconversion) at approximately 8 min, after ∼75 excitation scans with 405 nm light (Fig. 2E; see intersection of curves for Ch1 and Ch2) and reached a value of ∼400 at 15 min (Fig. 2E). In contrast, both KikGR- and Kaede-labeled cells photoconverted rapidly (Fig. 2H,I,K,L). KikGR- and Kaede-labeled cells photoconverted after a single 405-nm excitation scan (∼6 sec) (Fig. 2I,L) and were visually confirmed as photoconverted by 19 sec (Fig. 2I,L). During the photoconversion of either KikGR or Kaede in a cell, the mean green fluorescence signal (Fig. 2H,K; Ch1) decreased exponentially as the mean red fluorescence signal (Fig. 2H,K; Ch2) increased monotonically, reached a maximum, then decreased over time. The maximum red-to-green fluorescence ratios reached a value of ∼2,100 at around 1 min (Fig. 2H, KikGR) and a value of ∼1,400 at around 25 sec (Fig. 2K, Kaede).

KikGR- and Kaede-Labeled Cells Showed Timely Photoefficiency, and PAGFP and PSCFP2 Were Most Photostable

The efficiency of photoactivation was measured between the PAFPs by maintaining the same imaging parameters as above (Fig. 3). The ratio of photoactivated to non-photoactivated fluorescence signal was calculated in embryos over time for each PAFP (Fig. 3A,B; Table 1). KikGR was found to be the most photoefficient PAFP, as measured by the time to reach maximum photoconversion (Fig. 3A,B). This was due to the rapid exponential decrease of green fluorescence and increase in red fluorescence during the photoconversion process at 2% laser power (Fig. 2H). The rate of efficiency of photoactivation for KikGR was significantly higher than for Kaede-labeled embryos (Fig. 3B). KikGR-labeled cells photoactivated at approximately an exponential rate nearly 200 times greater than Kaede-labeled cells (Fig. 3B). Kaede-labeled cells consistently reached a lower maximum photo-efficiency value than the KikGR-labeled cells (Fig. 3A). Together, this translated into KikGR-labeled cells reaching a maximum photoconversion nearly 3 times faster than Kaede-labeled cells (Table 1). Both KikGR- and Kaede-labeled cells photobleached rapidly (Fig. 3A). The KikGR-labeled cells photobleached at an exponential rate approximately 220 times more rapidly than Kaede-labeled cells (Fig. 3A). In sharp contrast, PSCFP2 had the lowest efficiency of photoactivation and increased steadily to an asymptotic value just above the photoconverted threshold (Fig. 3A). PSCFP2 photoconverted nearly 73 times slower and reached a maximum photoconversion approximately 1 hr later than KikGR-labeled cells (Table 1). Both PSCFP2- and PAGFP-labeled cells were very photostable in vivo (Fig. 3A,B). PSCFP2-labeled cells did not show signs of photobleaching until after ∼240, 405-nm excitation scans (Fig. 3A; Table 1).

thumbnail image

Figure 3. A comparison of the efficiency of photoactivation between the PAFPs and the duration of distinction of photoactivated cells from non-photoactivated neighbors. Photoactivatable fluorescent proteins respond differently to constant, low-power (2%) 405-nm laser light in the chick embryo. The graph in A compares the photoefficiency and photostability of the four photoactivatable fluorescent proteins over time to reach and maintain maximal photoconversion. A: Recording of the log (base10) ratio of the mean fluorescence intensity ratio (I/Io) of photoactivated to non-photoactivated fluorescence signal plotted against increasing exposure to 405-nm laser excitation, in typical embryos labeled separately with one of the four photoactivatable fluorescent proteins. The scan time was consistent at 6.29 sec per frame of 405-nm light exposure and a pixel dwell time of 2.56 us. The mean intensity was calculated in a region of interest created around the photoactivated cell in both the unphotoactivated and the photoactivated channels. When the ratio (I/Io) is greater than 1, the cell is considered photoactivated. B: A close-up look near the origin of A at the photoefficiency of the four photoactivatable fluorescent proteins during the first 15, 405-nm excitation scans. C: Comparison of the time duration over which a photoactivated cell was distinguishable from a non-photoactivated neighbor in re-incubated embryos. Single cells were photoactivated in a number of embryos (the number of embryos analyzed for each construct is shown in the graph), which were re-incubated and surveyed at t = 24 and 48 hr. Kaede is the only PAFP that was unable to be found at t = 24hr and the points marked are t = 15 hr. For each photoactivated cell, the mean intensity value was calculated within a region drawn around the cell. This value was divided by the mean fluorescence calculated in a neighboring non-photoactivated cell to create the ratio labeling the y-axis. Each embryo photoactivated and surveyed is represented by a triangle (PSCFP2), asterisk (Kaede), open circle (KikGR), and dark circle (PAGFP).

Download figure to PowerPoint

Table 1. Comparison of the Time to Reach Photoactivation Ratio >1 (Photoactivated/Non-Photoactivated Signal) and Maximum Photoconversion Calculated Based on Data From Figure 3
Fluorescent proteinTime (sec) to reach photoactivation ratio >1Number of 405 nm excitation scansTime (sec) to reach maximum photoconversionNumber of 405 nm excitation scans
KikGR6.3137.77
Kaede6.31107.0218
PAGFP6.31837.25134
PSCFP2459.54744,008.0242

Photoactivation of PSCFP2-Labeled Cells Provided the Longest Duration (48 hr) to Distinguish Fluorescence From Non-Photoactivated Neighbors

To test the duration over which we could distinguish photoactivated from non-photoactivated cells, we surveyed re-incubated embryos at distinct time points after photoactivation (Fig. 3C). The photoactivated to non-photoactivated fluorescence ratios were compared in cells for each PAFP at post-photoactivation (T = 0+), 24 and 48 hr (Fig. 3C). Photoactivation was targeted to cells within the neural tube and migratory neural crest cells in the cranial region of embryos. This made it easier to locate and measure the photoactivated to non-photoactivated fluorescence intensity in cells, since the pathways of the migratory cells have been well documented. Kaede-photoactivated cells were found to have the weakest fluorescence intensity ratio, distinguishable only up to approximately 14 hr (Fig. 3C). Shortly after 14 hr, the fluorescence intensity ratio decreased below a value of 1, making the photoactivated cell indistinguishable from a non-photoactivated neighbor. In a similar manner, PAGFP photoactivated cells were difficult to visually separate from non-photoactivated cells after 24 hr (Fig. 3C). Many PAGFP-photoactivated cells had a fluorescence intensity ratio near a value of 1 by approximately 24 hr (Fig. 3C). KikGR-photoactivated cells were detected at 24 hr, with a fluorescence intensity ratio greater than 1, but were not detected at 48 hr (Fig. 3C). In dramatic contrast, PSCFP2-photoactivated cells were visually distinct after 48 hr and had a fluorescence intensity ratio greater than 1 (Fig. 3C). Thus, of all the PAFP's, only PSCFP2-photoactivated cells maintained a photoactivated to non-photoactivated fluorescence intensity ratio greater than 1 for up to 48 hr.

KikGR Was Effective for Selective Marking of Subgroups of Migratory Cells and Monitoring Cell Behaviors Without Introducing a Second Contrast Agent

The first application test of the PAFPs was to evaluate photoactivation to selectively mark subgroups of migratory NCCs in an intact embryo (Fig. 4). In a typical embryo, selective labeling of subgroups of migratory cells, such as the front or back of migratory streams, can be difficult and that prevents insight into cell dynamics and target invasion (Fig. 4A). We found that the rapid photoconversion of KikGR and dual-color fluorescent form was most effective for this purpose (Fig. 4A,B). In the control scenario, in the absence of photoactivation, NCCs that expressed the KikGR protein showed a positive difference in the ratio of the green-to-red fluorescence, confirming no or very little photoconversion (Fig. 4C,D). Subgroups of NCCs were selectively marked by photoactivating four different subregions within the cell migratory streams (Fig. 4E,G,I,K; box). Analysis of the photoactivation showed that KikGR-photoactivated cells clearly underwent a photoconversion from a green-to-red fluorescent form (Fig. 4E–L). Interestingly, the region of transition (near the boundary of the region selected for photoactivation prior to laser scanning excitation) was not a precise boundary of photoactivated to non-photoactivated signal, but was more diffuse as measured by the ratio of green-to-red fluorescence (Fig. 4F,H,J,L). In contrast to selectively marking the antero-posterior regions of a cell migratory stream, the proximal or distal subpopulations of cells were distinctly labeled (Fig. 4I–L). Thus, targeted fluorescent marking of subpopulations of cells in vivo within a migratory stream can be successful in an intact embryo model.

thumbnail image

Figure 4. Application of KikGR for selective marking of subgroups of migratory cells amongst a population over short periods of time. A: KikGR was introduced into the neural tube of chick embryos (HH St 8–9) and expressed by migratory neural crest cells after embryo re-incubation. The box (approximately 300 × 100 μm) shows the pre-photoactivated region. B: Subgroups of neural crest cells were photoactivated (red cells in box after photoactivation). C: A typical confocal image of neural crest cells in a control embryo re-incubated for 24 hr after exposure to 488 nm light, but not photoactivated with 405-nm light exposure. D: A line profile (0–90 μm in length) through the neural crest cell population in C shows the green (CH1) and red (fluorescence) intensity levels in cells and tissue along the line. E-L: Selective marking of subgroups of neural crest cells within migratory streams in the embryo shows targeting of the (E) anterior, (G) posterior, (I) proximal, and (K) distal areas of the migratory stream in different embryos. F, H, J, L: The line profiles through the photoactivated cells show the length of the line and regions of photoconversion (where the red-to-green fluorescence ratio is greater than 1). The region of the chick hindbrain where the migratory stream emerges from is labeled as rhombomere 4 (r4).

Download figure to PowerPoint

PSCFP2 Is Effective for Single Cell Tracing Over Short Time Periods (∼48 hr) by Using Single Photon or 2-Photon for Accuracy in Dense Populations

The second application test of the PAFPs was to evaluate photoactivation to selectively mark single cells for cell tracing over time in an intact embryo (Fig. 5). In a typical embryo, the ability to accurately mark a cell within a particular region is often challenging due to the position of the cell (depth) and density of cell neighbors. However, those challenges can be answered with 2-photon photoactivation when there is an advantage of deeper tissue imaging and a fine plane of focus. In high cell density populations labeled with PAGFP or PSCFP2, single-cell photoactivations can occur at 760 nm (Tsutsui, 2005) to prevent out-of-focus photoactivation of neighboring cells (Fig. 6). Based on the comparison study of the length of time over during which fluorescence signal could be observed within photoactivated cells, we selected PSCFP2 (Fig. 5). We focused on selectively marking cells within the chick neural tube and NCCs (Fig. 5). In a typical embryo, we photoactivated individual NCCs shortly after the cells emerged from the neural tube and evaluated cell positions at 24 hr after re-incubation (Fig. 5A–C). A lipophilic dye, DiI, was injected into the neural tube to globally label the neural tube and NCCs (Fig. 5A–C). Photoactivated cells and their progeny were located distributed throughout the target site after 24 hr (Fig. 5D,E) and were clearly distinguishable as photoactivated cells (Fig. 5E). Calculations of the average green-to-cyan (G/C) ratio in photoactivated cells at T = 0 and T = 24 hr, measured by mean pixel intensity values within a region around each cell, showed the average G/C ratio to be much greater than 1, clearly distinguishing photoactivated cells from neighbors (Fig. 5E, insets).

thumbnail image

Figure 5. Application of PSCFP2 to label small numbers of cells for lineage tracing. A: After injection and electroporation of PSCFP2 and co-injection of DiI (red) to label the chick neural tube and migratory neural crest cells, (B) small numbers of cells expressing PSCFP2 (blue colored cells in box) were (C) photoactivated (green colored cells in box). The box shows the region of photoactivation in pre- and (C) post-photoactivated embryos. The top inset shows the approximate position of the photoactivated neural crest cells adjacent to the neural tube (labeled by r4 to designate rhombomere 4 of the chick hindbrain). The larger inset is a magnified image of the photoactivated area. D,E: Twenty-four hours after re-incubation of embryos, the photoactivated cells (green) were located with confocal imaging of the target site of neural crest cell invasion amongst the population of non-photoactivated cells (red). E: The inset small box on the upper left-hand side of the image shows the approximate position of the photoactivated cells at T = 0+hr. The lower inset shows the average green to cyan (G/C) ratio measured from the photoactivated cells (green colored) at T=0+ and T=24 hr. The scale bar = 100 μm and multiple cells in n = 12 embryos were photoactivated and analyzed. F: PSCFP2 was injected and electroporated in HH Stage 8–9 embryo and reincubated for (G) 12–16 hr. H: Small numbers of cells within the neural tube were photoactivated (green colored cells in the image and the insets). The non-photoactivated cells appear in blue. I,J: Forty-eight hours after re-incubation of embryos, the cells are located in transverse sections of the embryos showing the cross-section of the neural tube. The inset shows the positions of the photoactivated cells (green colored). The scale bar = 100 μm and multiple cells in n = 12 embryos were photoactivated and analyzed. The lumen of the neural tube is visible as a dark spot at the dorsal part of the tissue section.

Download figure to PowerPoint

thumbnail image

Figure 6. Photoactivation of PSCFP2 using single photon versus 2-photon of cells in the neural tube, and a KikGR comparison with photoactivation of neural crest cells. PSCFP2 photoactivation was done in the chick neural tube to compare out-of-focus photoactivation effects with 1-photon and 2-photon laser (Coherent Chameleon) with a plan apochromat 20× NA 0.80 (Carl Zeiss inc.). A: The hindbrain neural tube of an E2 chick labeled with PSCFP2 shows three different photoactivation regions. The top and bottom neural tube cells were photoactivated by 2-photon where the middle cell was photoactivated with a 405-nm laser. The triangle, asterisk, and square coordinate with B, C, D, F, and I. The green cells are photoactivated and the blue cells are non-photoactivated PSCFP2-labeled cells. The upper (triangle) 2-photon photoactivated cell (B) and the single photon photoactivated cell (C) in a XY projection of the neural tube shows the dense area around the targeted cells. D: An XZ projection of a 1-photon photoactivated cell shows three cells that were photoactivated. E: The brighter cell was the target of the 405-nm laser for a single cell photoactivation but out-of-focus photoactivation occurred and caused the cell above and below to photoactivate. This can be observed in the line profile as the smaller peak on the left and the extension off of the larger peak on the right. F: An XZ projection of the 2-photon photoactivated cell demonstrates the power to only photoactivate the plane of focus. G: A line profile through the image reveals only one peak of intensity for a single cell, where a large or multiple peaks would suggest multiple cell photoactivation. The white boxes inside D and F are the regions of interest where the photoactivation scans took place. H: The XY projection of both the 1- and 2-photon photoactivation region shows a profile of a single sharp peak of intensity for the top 2-photon cell and a multiple, broad peak for the 1-photon photoactivated cell. I: The XY projection reveals, with the brightness increased, the 1-photon region as a cluster (∼3) of cells consisting of the larger targeted cell and the weaker residual photoactivated cells. J: Neural tube cells and neural crest cells are used in the comparison studies and both cell types were tested for photoactivation response. Simultaneous photoactivation was done on neural crest cells and neural tube cells and the ratios were determined as previously described. Each cell was plotted according to cell type with n = 18 cells. Neural tube cells and neural crest cells show no difference in the photoactivation of KikGR.

Download figure to PowerPoint

Neural tube cells (NTCs) maintain the properties of the PAFPs the same as NCCs but behaviorally the cells are different (Fig. 6). NTCs are non-migratory cell(s) and this creates an advantage of locating the cells after longer time points to observe the PAFP (Fig. 5G,H). NCCs may migrate to deeper positions and visual accessibility may be lost even though the cell may still maintain photoactivated proteins. To confirm that PSCFP2-photoactivated cells could be visually distinguished at 48 hr, we performed the same experiment (with the exception of not co-injecting the lipophilic dye, DiI, to label host cells) (Fig. 5F–J). PSCFP-2 photoactivated NTCs were consistently located after 48 hr of embryo re-incubation and were visually and quantitatively confirmed as having a photoactivated to non-photoactivated ratio greater than 1 (Fig. 5I,J). Since the embryonic tissue growth is substantial over these early developmental stages, photoactivated cells were accurately located in sections taken near the original photoactivated site (Fig. 5I,J). Thus, PSCFP2 can be applied to NCCs and prove to be a good PAFP candidate for short-term single cell lineage analysis in an avian model.

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

In this report, we showed that the fundamental photophysical properties of four photoactivatable fluorescent proteins (PAFPs) are maintained in an embryo model, but have distinct advantages for specific developmental biology applications. Measurements showed that PAGFP, PSCFP2, KikGR, and Kaede photoconversion was reproduced with similar characteristics as reported in cultured cells with a couple of exceptions. For Kaede- and KikGR-labeled embryos, we showed that cells reached 90% of maximal photoefficiency in 63 and 20 sec, respectively, with a photoactivated red-to-green average ratio of 6.0 (Kaede) and 43.0 KikGR (Fig. 2). The times to reach maximal photoefficiency compared well with published photoactivation results in HeLa cells (Ando et al., 2002; Tsutsui et al., 2005), However, our R/G ratios were lower (6.0 vs. 41.0 for Kaede). This difference may be a result of our not reaching a saturating level during photoconversion, since we judged the red-to-green photoconversion by visual inspection only. We suggest that photoactivation in embryos should be monitored by both visual inspection and a quantitative readout to ensure maximal photoconversion. Our results of PSCFP2 photoconversion in embryos was difficult to compare with published data in L929 cells, since we maintained a low 2–30% laser power (vs. 50–100% laser power in Chudakov et al., 2004) for comparison of photophysical properties with other PAFPs. However, we noted only a 400-fold increase in photoconverted green-to-cyan ratio (vs. >2,000-fold in Chudakov et al., 2004).

The photoactivatable fluorescent proteins displayed several advantages over current in vivo cell labeling techniques to trace cell behaviors and short-term cell lineage. First, photoactivation of dual-color fluorescent forms (PSCFP2, KikGR, and Kaede) offered the flexibility to selectively mark subgroups of migratory cells and simultaneously monitor photoactivated and non-photoactivated neighbors. We showed that it was possible to accurately mark subgroups of neuronal precursor cells along their migratory route (Fig. 4), for example at the front and back ends of migratory streams. With current techniques, one would have to inject two distinct lipophilic dyes in a precise, targeted manner in order to mimic this result. However, excess lipophilic dye that persists in the microenvironment could lead to an inadvertent dual-color labeling of trailing cells that migrate through the injection region. Second, photoactivation in embryos offered an increased accuracy to detect short-term cell lineage of a single precursor cell. We showed that PSCFP2 was best suited to selectively mark single cells for up to 48 hr (Fig. 5) using both confocal and 2-photon photoactivation (Fig. 6; Supplemental Movie S2). We found the in ovo photoactivation process allowed us to target individual cells in specific locations in the avian embryo below the surface ectoderm more consistently and in a more timely manner than glass needle lipophilic injection or electroporation.

Our comparison study did reveal some disadvantages to the photoactivation of PAFPs as a cell-labeling technique. The rapid photobleaching of KikGR- and Kaede-labeled cells, even at a low laser power (1 or 2%, 405 nm), suggested there is a delicate balance between photoefficiency to photoactivate and photobleaching (Fig. 3). Interestingly, rapid photobleaching was not reported in previous results (Ando et al., 2002; Tsutsui et al., 2005), which may be due to their brief, intermittent 1-sec UV light exposure in HeLa cells and short 10-sec exposure in hippocampal primary culture versus our continuous exposure with 405 nm light. We were not able to recapitulate KikGR photoconversion with 2-photon excitation at 760 nm as reported in cultured HEK-293 cells (Tsutsui et al., 2005).

Secondly, the length of visually distinguishing a photoactivated cell (up to 48 hr) was low compared to typical cell lineage experiments that require more than 2 days of cell tracing in the embryo (for example, the lipophilc dye, DiI, has reportedly maintained fluorescence in motorneurons for up to 4 weeks in culture; Kuffler, 1990). There are at least two reasons why photoactivated cells were only visible up to 48 hr. The thickening and opacity of the tissue in the growing chick embryo may have contributed to less visually distinct photoactivated cells. Also, we did notice variability in the initial fluorescence ratios after photoactivation (Fig. 3C). This was due to incomplete photoconversion of the PAFP. There is the possibility of re-photoactivation of marked cells, however, this would require time-lapse monitoring of labeled cells and this is not typical of cell lineage tracing experiments that rely on re-incubation of embryos after initial cell marking.

In summary, the current set of PAFPs are effective in the avian embryo to accurately label migratory cells and trace short-term cell lineage in a less invasive manner. Our measurements of the photophysical properties of four PAFPs in an avian embryo should offer a guide to photoactivation applications. The consistent emergence of novel engineered, monomeric, and high contrast PAFPs, for example Dendra2 (Gurskaya et al., 2006) and Dronpa (Ando et al., 2004), should increase current limitations to cell lineage tracing and distinct cell labeling using UV, violet, and less toxic blue light excitation. The application of PAFPs to embryo model systems, including zebrafish (Aramaki and Hatta, 2006; Distel et al., 2006), Drosophila (Post et al., 2005), Xenopus (Wacker et al., 2007), and chick (Stark and Kulesa, 2005), and in brain slices (Miyawaki, 2005; Mutoh et al., 2006), will also help expand applications to study complex morphogenetic events.

EXPERIMENTAL PROCEDURES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Embryo Preparation

Fertilized white leghorn chicken eggs (supplied by Ozark Hatchery, Oeosho, MO) were incubated at 38°C in a humidified incubator until the Hamburger and Hamilton (HH) stages 8–9 (Hamburger and Hamilton, 1951). Eggs were rinsed with 70% ethanol and 3 ml of albumin was removed before windowing the eggshell. A solution of 10% India ink (Pelikan Fount; PLK 51822A143, www.mrart.com, Houston, TX) in Howard Ringer's solution was injected below the area opaca to visualize each embryo.

Constructs

We obtained PAGFP (kind gift from Prof. Jennifer Lippincott-Schwartz), Kaede (MBL Int'l, AM-V0012, Woburn, MA), KikGR (kind gift from Prof. Atsushi Miyawaki) and PSCFP2 (from Evrogen, PS-CFP2-N vector, no. FP802, Moscow, Russia). Each construct was grown up to 5 μg/μl.

Electroporation Delivery

PAFP's were injected separately into the lumen of the chick neural tube in embryos at HH stages 8–9. The lumen of the chick neural tube provided a cavity within the embryo that is accessible to glass needle injection and visual inspection of contrast agent delivery. After windowing the eggshell, a sharpened tungsten needle was used to open a small hole in the vitelline membrane above the neural tube. Constructs were microinjected directly into the lumen of the neural tube, filling the hindbrain region, using a pulled borosilicate glass needle (Sutter; BF100-50-10, Novato, CA) attached to a Picospritzer III (Parker Hannifin Corporation, Fairfield, NJ). The constructs were electroporated into the right side of the neural tube using platinum electrodes and an Electro Square Porator ECM 830 (BTX, a division of Genetronics, San Diego, CA) with 20 volts of current and five 45-ms pulses at 1-sec intervals. A few drops of sterile Ringer's solution was applied to the embryo prior to the eggs being sealed with adhesive tape and re-incubated at 38°C for 12–16 hr in a humidified incubator (Model 1550, G.Q.F. Manufacturing Co., Savannah, GA).

Photoactivation

The general photoactivation process, including cell target acquisition and selection for photoactivation, followed a previous protocol developed for PAGFP (Stark and Kulesa, 2005) and was adapted for each PAFP. For the comparison of PAFPs, we used constant imaging parameters, including laser power at 2% (measured at 44 uW at the sample and the back aperture using a hand-held optical power meter; Model 840, Newport) in some experiments. All photoactivation was performed in ovo on intact chick embryos. A teflon membrane (Fisher Scientific, LLC) was stretched across an acrylic ring (∼2.2 cm i.d.* 2.6 cm o.d. * 0.5 cm) and secured into a window in the eggshell with beeswax (Fisher Scientific, LLC) according to a previous protocol (Kulesa and Fraser, 2000). All imaging was performed on an upright LSM 5 PASCAL microscope or LSM 510 META NLO (Carl Zeiss, Inc., Thornwood, NY) using a Plan-NeoFluor 10×/0.3, 40×/0.75, or C-Apochromat 40× W/1.20 (Zeiss). Collected images were saved using the AIM Software (Zeiss) and processed (histogram stretch only) in Adobe Photoshop CS2 (Adobe Systems, Inc., San Jose, CA).

PAGFP

Cells expressing fluorescence were located by using 488-nm laser excitation, low laser power (1–10%) with a 10×/0.3 objective (Zeiss) at 512×512 pixels. If PAGFP-labeled cells were too dim to visualize, the detector gain and pinhole diameter were increased, and frame averaging (n = 4) was applied. If the GFP expression was still too dim to visualize, the 488-nm laser power was gently increased. When a region of the embryo of interest where cells were expressing PAGFP was located with a 500–530-nm band pass filter, the optical zoom feature was used to select a region around an individual cell or group of cells. For photoactivation with 405-nm laser excitation, the laser power was set according to the optical zoom. The higher the zoom, the less percent laser power was required. For PAGFP-labeled cells, we used 2–3% laser power at 35× zoom or greater and 4–5% below 35× zoom.

Kaede and KikGR

Kaede and KikGR PAFPs were both very sensitive to UV light exposure that caused inadvertent photobleaching and/or photoconversion. Thus, it was important to consider the amount of laser and Hg light exposed to the sample. Hg lamp light was not used on the sample to survey embryos prior to photoactivation. Cells expressing Kaede or KikGR were located by using 488-nm laser excitation, low laser power (1–2%) with a 10×/0.3 objective at 512×512 pixels. KikGR and Kaede images were collected in single track mode using the filter combination of 505–530 nm (green) and LP-560 nm (red) where the 543- or 561-nm laser was used to excite the photoconverted proteins.

PSCFP2

PSCFP2-expressing cells were located using 405-nm laser excitation at 1–3% laser power with a 10×/0.3 objective at 512×512 pixels. Importantly, the PSCFP2 signal was collected without using the digital zoom feature so that there was not an accidental photoconversion of the entire cell population in the field of view. Prior to photoactivation, excitation and emission of PSCFP2 fluorescence peaked at 402 and 468 nm, respectively. To photoactivate PSCFP2, an individual cell was optically zoomed in on, and excited with 405-nm laser light (or 760 nm; 2-photon) at a power setting of 2% (for the comparison study). For applications, 5–15% laser power was used when the optical zoom measured 35× or greater and 15–30% below 35×. Images were collected using separate tracks with the filter combinations of 420–480 nm (cyan) and 505–530 nm (green) where a 488-nm laser was used to excite the photoconverted proteins.

Static Imaging

After the desired reincubation period (24–48 hr), each embryo was prepared for confocal imaging as described in Teddy and Kulesa (2004). Briefly, the embryo was cut from the egg with a pair of scissors (Fine Science Tools, 14060-10) and placed in a dish of Ringer's solution. Since all of the fluorescent labeling was focused in the head region of the embryo, all of the undesired tissues were cut away including portions of the tail and entire heart. On a glass slide (VWR, 48312-024), a circle of vacuum grease (Dow corning, 79810-99) was placed to create a well and the embryo was transferred in the circle with the left side facing down. All liquid was removed from around the embryo and a no. 1, 25-mm circle coverslip (VWR, 48380-080) was carefully placed over the entire grease ring. Confocal imaging was performed using the same configuration settings per PAFP described above.

Acknowledgements

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The authors thank Prof. Jennifer Lippincott-Schwartz for the kind gift of PAGFP and Prof. Atsushi Miyawaki for his kind gift of KikGR and technical advice. The authors also thank Dr. Joel Schwartz for his critical reading of the manuscript and technical advice, Dr. Winfried Wiegraebe for his technical advice, and helpful discussions with members of the Stowers Institute Imaging Center. This work was funded by the generosity of Mr. and Mrs. Stowers and the Stowers Institute for Medical Research.

REFERENCES

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. RESULTS
  5. DISCUSSION
  6. EXPERIMENTAL PROCEDURES
  7. Acknowledgements
  8. REFERENCES
  9. Supporting Information

The Supplementary Material referred to in this article can be found at http://www.interscience.wiley.com/jpages/1058-8388/suppmat

FilenameFormatSizeDescription
jws-dvdy.21174.anim1.avi2650K Supplementary Movie S1: The photoactivation of Kaede A single cell labeled with Kaede is seen over time during the photoactivation with a 405nm laser. A line profile of the cell was taken to support the decrease of green intensity while the red intensity increased.
jws-dvdy.21174.anim2.AVI1536K Supplementary Movie S2: PSCFP2 photoactivated comparison of single-photon versus two-photon A 3D rendering of two photoactivated PSCFP2 cells where the top cell was photoactivated by 2-photon and the second cell was done with a 405nm laser. The left side of the movie represents how the 2-photon photoactivation is capable of targeting a single cell where a single-photon has an out of focus photoactivation effect. The right side demonstrates the density of the labeled neural tube cells and how it can present a challenge when trying to photoactivate single cells with a single-photon laser.

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.