Notice: Wiley Online Library will be unavailable on Saturday 27th February from 09:00-14:00 GMT / 04:00-09:00 EST / 17:00-22:00 SGT for essential maintenance. Apologies for the inconvenience.
We have generated a novel genetic system to visualize cell lineages in living tissues at high resolution. Heat shock was used to trigger the excision of a specific transposon and activation of a fluorescent marker gene. A histone-YFP marker was used to allow identification of cell lineages and easy counting of cells. Constitutive expression of a green fluorescent membrane protein was used to provide a precise outline of all surrounding cells. Marked lineages can be induced from specific cells within the organism by targeted laser irradiation, and the fate of the marked cells can be followed non-invasively. We have used the system to map cell lineages originating from the initials of primary and lateral roots in Arabidopsis. The lineage marking technique enabled us to measure the differential contribution of primary root pericycle cell files to developing lateral root primordia. The majority of cells in an emerging lateral root primordium derive from the central file of pericycle founder cells while off-centre founder cells contribute only a minor proliferation of tissue near the base of the root. The system shows great promise for the detailed study of cell division during morphogenesis.
If you can't find a tool you're looking for, please click the link at the top of the page to "Go to old article view". Alternatively, view our Knowledge Base articles for additional help. Your feedback is important to us, so please let us know if you have comments or ideas for improvement.
Multicellular organisms develop via a coordinated pattern of cell division and cell differentiation. In plants such as Arabidopsis thaliana, cell division during embryogenesis is highly regular, resulting in a close correspondence between cell lineage and fate. However, in the adult plant, cell divisions are often less regular and it is clear that constant positional signals are exchanged between cells and their neighbours to mediate development (Sena et al., 2004). To understand development it is important to have both genealogical information and an understanding of the influence of neighbouring cells.
The use of inducible promoters has further enhanced the utility of recombinant systems; it allows the expression of the transgene to be induced at the desired time point in development. The most popular type of inducible promoter is the heat shock promoter from soybean in plants (Kilby et al., 1995) and from Drosophila in animals (Halfon et al., 1997; Stringham and Candido, 1993). By placing the Ac transposase under the control of a heat shock promoter from soybean, Ac-mediated transposition was induced at high temperatures (Balcells et al., 1994).
This paper describes a novel transposon-based system in Arabidopsis which uses heat shock-induced transposase activity to switch on expression of the nuclear-localized yellow fluorescent protein (YFP) reporter gene. This reporter gene is normally transcriptionally silent due to the presence of a dissociator (Ds) element inserted between the promoter and reporter. Heat shock switches on expression of the Ac transposase which in turn excises the Ds element causing the reporter to be expressed in that particular cell and in all cells that derive from it (Figure 1).
Cell lineage patterns arising within the tissue initial zone of the primary root apical meristem (RAM) were analysed. Arabidopsis primary root tissue development is very regular and has been described from sectioned material. The origination of each of the tissues is well described (Baum et al., 2002; Dolan et al., 1993; Scheres et al., 2002; Wenzel and Rost, 2001). Dolan et al. (1993) describe a model of main root meristem organization in Arabidopsis. In the model, initials of the different root tissues are located surrounding the central cells of the quiescent centre; epidermis and lateral root cap (LRC) share a common initial as do cortex and endodermis. Pericycle, vascular tissue and columella files all derive from single initials. These findings were supported by clonal sector analysis in fixed tissues (Dolan et al., 1994). Our live-cell lineage analysis supports the findings of these earlier reports and demonstrates the utility of the system.
Unlike the main RAM which is fairly easy to observe, lateral root initiation is difficult to study, especially in living roots as the new cell walls are very thin and occur within the primary root where stains, e.g. propidium iodide, do not penetrate. The best images of early developmental stages come from studies of cleared whole-mounts (Malamy and Benfey, 1997) and sectioned materials (Dubrovsky et al., 2000, 2001). Lateral roots initiate in the pericycle layer of primary roots. Pericycle cell-cycle marking studies (Beeckman et al., 2001; Himanen et al., 2002) and cell type-specific marking (Malamy and Benfey, 1997), each utilizing GUS staining, showed the earliest indications of lateral root initiation. In addition, Malamy and Benfey (1997) produced a cell lineage map of the median longitudinal view of the emerging lateral root primordium (LRP). The relative contribution of different pericycle cell files to the LRP is not clear, however various studies placed the number of LRP founder cells at 6–11 (Casimiro et al., 2003; Dubrovsky et al., 2000; Laskowski et al., 1995). Pericycle cells in longitudinal files in contact with the xylem differentiate as LRP founder cells. Generally, two files of pericycle cells are in direct contact with a xylem file but a minimum of three files of the pericycle are usually involved in LRP formation (Casimiro et al., 2003). Xylem-contacting pericycle cells show the first signs of differentiation as LRP founders but adjacent files also contribute to the LRP. We have used our lineage marking system combined with a live-cell outlining marker to demonstrate uneven contribution of the LRP pericycle founder files to the emerging lateral root.
In addition, we show that the lineage marking system is laser-inducible to allow targeting of specific cell types as lineage progenitors.
Heat shock induction can be used successfully to generate marked live-cell lineages
Arabidopsis plants carrying the Ds1 construct (Ds-H2B-YFP) were genetically crossed with hs-Ac plants. F1 seedlings from the crosses were analysed for the ability to induce activation of the H2B-YFP gene after heat shock. One of the tested lines (P6/7 × hs-Ac) showed a high level of transposition, the others had low to no transposition. Subsequent analyses were carried out with the progeny of this line (approximately 60% of these seedlings had one or more transposed sectors). Seedlings subjected to heat shock at 42°C showed a minimum requirement of 20 min (data not shown) before any transposition was observed. Routinely, heat shock was applied to seedlings on media plates for 30 min.
Using the lineage marking system to analyse clonal relationships in the primary root
Before commencing the analysis, we checked that the reporter gene, H2B-YFP, was capable of being expressed at high levels in every cell type in the primary RAM. We analysed Arabidopsis seedlings carrying a 35S-H2B-YFP construct (Boisnard-Lorig et al., 2001). Nuclei were fluorescent in every cell in every file of the RAM (not shown).
In hs-Ac/Ds-H2B-YFP lines, cell lineages demonstrating the origin of every cell type in the RAM were observed. Examples of tissue initiation are shown in Figure 2(a–d). In each case, a progenitor cell and its derivatives were marked with fluorescence. The LRC and epidermis derive from a single initial cell (Figure 2a,b). Subsequent divisions of this cell and its daughter cells resulted, on occasion, in marked epidermal cell files >60 cells long (42 marked epidermal cells are shown in Figure 2a). The LRC/epidermis initial first divides periclinally (Figure 2b, inner outlined group). Subsequent anticlinal divisions within the two resultant layers produce the LRC to the outside and the epidermis to the inside (Figure 2b, larger outlined group). The LRC/epidermis initial cell is regenerated as the distal (towards root tip) cell in the inner layer. Root cortex and endodermal layers also derive from a single initial cell, the basal (away from root tip) daughter of which divides periclinally to form the endodermis towards the inside and the cortex towards the outside (Figure 2c). Uniseriate files of root cap columella cells (Figure 2d) and stele cells (not shown) derive from a single initial cell in each file. Fluorescence was rarely observed in the central cells of the quiescent centre and no consistent lineage patterns were observed after such an event. Often, a low level of YFP fluorescence was observed in the columella cells at the root tip. This fluorescence was very dim compared to that resulting from Ds1 transposition.
Analysis of lateral root lineages
Use of the EGFP-LTI6b plasma membrane marker enhanced visualization of cellular architecture within initiating LRP where traditional stains (e.g. propidium iodide) do not penetrate. Heat shock-induced transposition of Ds1 resulted in lineage marking within LRP. Of particular interest was the situation in which a single file of pericycle cells was marked in the primary root (Figure 3a–d). Two outcomes resulted, either almost all cells within the LRP were marked (Figure 3a,b), or marking was confined to a basal group of cells displaced from the centre axis of the lateral root (Figure 3c,d). The first visible sign of LRP formation occurred in three axially adjacent pericycle files when several transverse divisions in single or in axial pairs of pericycle cells are followed by tangential periclinal divisions of an inner pair of derivatives resulting in the inner layer (IL) and outer layer (OL) of LRP cells (Figure 3a). When the marked parent pericycle file was central in the three axially adjacent files that contribute to LRP formation, virtually the entire primordium was marked (Figure 3b). Figure 3(c), however, illustrates the situation in which the marked parent pericycle file is not central. In the median longitudinal section (seen in xy-plane), IL and OL cells have formed but cell marking is confined to an adjacent longitudinal file (seen in zy-plane) that has not proliferated to the same extent. Proliferation of cells within the marked file results in only a basal group of marked LRP cells (Figure 3d). Occasionally, small, spontaneously arising marked sectors that did not proliferate into the LRP were also observed (Figure 3e).
Analysis of cell lineages in other tissues
The cell lineage marking system was observed to work in aerial parts of Arabidopsis seedlings as well as in roots. Lineage sectors were marked in hypocotyls, stems, cotyledons and leaves. Figure 3(f) shows a stomatal lineage formed by successive asymmetric divisions of a single, marked meristemoid cell and a final symmetric division to produce a guard cell pair.
Laser induction of the live-cell lineage marker
Laser induction of lineage marking is possible. Figure 4 shows a time series. A control image of the main root tip expressing EGFP-LTI6b was taken immediately before laser-induced heat shock (Figure 4a). No H2B-YFP was detected in the root tip. The vascular initials were targeted and heat shocked (Figure 4a, outlined region). Figure 4(b) shows the root tip after pulsed-laser heat shocking. The EGFP-LTI6b has photo bleached in the targeted region. At 6 h after heat shock, Ds1 transposition had occurred and triggered the production of H2B-YFP to mark a four-cell lineage in one axial file of vascular tissue and a single marked cell in a non-adjacent file (Figure 4c). The most distal (youngest) cells in the marked file remained in the cell division zone of the meristem as the root grew and, at 29 h post-induction, the original marked file had increased via transverse cell divisions to include 14 fluorescent nuclei (Figure 4d, 8 of 14 nuclei shown in the original file). The original single marked cell did not divide and is no longer visible in Figure 4(d). During the same period, one of the earlier marked cells had divided longitudinally to produce a second, adjacent file of marked cells. Four of the fluorescent nuclei in the marked adjacent file are visible in Figure 4(d).
Heat shock-induced marking of cell lineages
We describe a new system to mark and follow cell lineages- in living tissues. Use of a live visual reporter gene circumvents the problem of destructive assays and allows one to follow the lineage of a given cell through time without the dilution of marker signal that occurs in other live-cell marking systems, e.g. photo-activation of fluorescent proteins or fluorescent tracer injection. It also overcomes the limitations of the availability or generation of periclinal chimeras and the low frequency of chimerism and associated cell damage generated by X-ray irradiation.
There are several advantages to using a heat shock promoter for triggering the formation of marked clones. The promoters are characterized by quick response times coupled with high induction levels (Lyznik et al., 1995), and heat shock treatment can be adjusted to titrate the frequency of sectoring, and to control the timing of clone formation during development. In addition, localized induction of heat shock using a laser can be used to follow the lineage of a particular cell over several days.
Lineage marking of the Arabidopsis primary RAM
The origin of tissues in the Arabidopsis RAM has been described from sectioned root material (Dolan et al., 1993; Wenzel and Rost, 2001). Live-cell lineages resulting from Ds1 transposition in tissue initial cells generally support the observations made using sectioned material. The LRC and epidermis arise from asymmetric T-division of a single initial (Dolan et al., 1993; Wenzel and Rost, 2001) to produce a commonly observed wedge-shaped lineage in median section view. Marked epidermal cell files >60 cells long represent at least four to five divisions of the LRC/epidermal initial cell and many daughter cell divisions with no reduction in the brightness of marking over time. Perpetual marking is not possible in other live-cell lineage analysis systems as the marker is diluted by subsequent rounds of cell division. Cortex and endodermal layers were observed to arise both from a common initial and from separate, adjacent initials. This observation is due to alternation in the orientation of the cell division plane in the initial. A previous report (Baum et al., 2002) described the transition to separate initials as the root aged. Columella, pericycle and vascular tissue cell files were all observed to arise from Ds1 transposition in single initials. Ds1 transposition to mark the central cells of the quiescent centre was infrequently observed. Very often, the majority of cells in the columella display a dim YFP fluorescence that may have resulted from low levels of leaky H2B-YFP expression amplified by endoreduplication in these cells (Kidner et al., 2000). Nuclear fluorescence in marked columella lineages was an order of magnitude brighter.
Roots were sometimes observed in which cells were marked in adjacent tissue layers. This may have resulted from intrusive growth from a marked tissue into an adjacent tissue layer with a subsequent switch in cell fate, so that the intruding cells matched other cells in the same file (Kidner et al., 2000). Observations of this type can also arise from independent, closely adjacent transposition events, although only about 25% of primary root tips observed had more than one cell lineage marked. Our live-cell lineage marking system will allow direct observation to establish the frequency of invasive growth and positional fate switching during normal growth.
Lateral root initiation
Unlike the primary root primordium which is enclosed within the seed, newly initiated lateral root primordia (LRP) are ideal organs for study of de novo patterning as they are accessible to microscope and laser. Unfortunately, LRP initiation is difficult to study in living roots because of the internal location of the first pericycle divisions in the parent root. Stages of LRP development have been described in cleared whole-mounts using histochemical markers of reporter gene activity obtained by wide-field microscopy (Dubrovsky et al., 2001; Malamy and Benfey, 1997). However, due to limited resolution, it was only possible to map the very earliest cell divisions in the non-emergent LRP in median longitudinal view. Our method for outlining cells with EGFP-LTI6b while marking cell lineages allows clear visualization of cell arrangements throughout the developing primordia, as well as direct visualization of actual division patterns. We have confirmed the order of cell divisions described previously (Malamy and Benfey, 1997) and have provided new insight regarding the contribution of adjacent pericycle files of the parent root to the developing lateral root. Cells in three axial files of the pericycle in contact with the primary root protoxylem undergo a series of transverse divisions to produce the lateral root founder cell population. The number of cells in this founder population is variously put at 6–11 (Casimiro et al., 2003; Laskowski et al., 1995). This number will remain contentious as it depends on how founder cells are defined. Most cells in an emerging LRP, however, derive from the centre file of parent root pericycle founder cells. When the central file of pericycle progenitors is marked by Ds1 transposition, all cells in the LRP apical meristem region are marked and, ultimately, all cells in the lateral root excepting a few cells at the base are marked (Figures 3b and 5a). When one of the two flanking pericycle files is lineage marked initially, only a small population of cells at the base of the lateral root is ultimately marked (Figures 3d and 5b). This developmental pattern can be studied in a single parent root with a marked pericycle file because while the marked file is central in one LRP it will be flanking in others. The possible function of the small proliferation of cells at the base of the lateral root that derive from an off-centre progenitor file was not determined.
Laser-induced live-cell lineage marking
Confocal microscope systems allow precise targeting of laser output at individual cells. Laser induction of heat shock promoters has been used previously in Caenorhabditis elegans (Harris et al., 1996; Stringham and Candido, 1993) and in Drosophila (Halfon et al., 1997) but not for live-cell marking. We have used repeated scanning of the 488 nm line from an argon laser to activate the heat shock promoter in specific regions of the Arabidopsis RAM. In this example, when the vascular tissue initials were targeted two separate cells were activated (Figure 4a–d). One of these was in the cell division zone (closer to the root tip) and a 14-cell lineage developed in a single file over the next 29 h. During the 6–29 h after heat shock, a longitudinal division occurred in one of the marked cells resulting in an adjacent file of marked cells. Thus, radial proliferation of the vascular tissue begins within the cell division zone of the root. The second originally activated cell did not divide.
Development of the heat shock protocol required optimization of the laser treatment. Continuous scanning of the root tip resulted in high levels of Ds1 transposition. Precise targeting was possible only if the heat generated from continuous scanning did not heat the media and surrounding cells. To this end, discontinuous scanning was used so that heat would not build up and propagate to non-targeted tissues. The best results were obtained by scanning 50–100 times and allowing 2.5–5 sec to elapse between scans. With this level of scanning, not all nuclei within the region-of-interest were activated. Single-cell targeting should be possible, i.e. for activation of a central cell or a specific initial, if scan number is increased and the inter-scan time lapse is sufficient. Attenuation of laser output or the use of longer wavelengths as in multiphoton laser scanning microscopy might help to control the precision of targeting even further.
This live-cell marking system is ideal for tracking cell division patterns during plant growth. In addition to confirmation of established lineage patterns in the main root, we have observed the unequal contribution of adjacent pericycle files into the LRP. The situation we describe is that of the entire radial organization of the emerging LRP deriving from a few cells in a single file of pericycle progenitors. Lateral root tissues do not derive, as we might have expected, equitably from three adjacent pericycle files. This lineage marking technique will allow further study of LRP development, in particular, in traverse section view so that the origin of radial symmetry from a single file of progenitors can be described.
Lineage patterns are also easily observed in stems and leaves and this system will greatly simplify the study of processes such as stomatal ontogeny. Lineage marking techniques are vital for the study of whole organism development. The components of this genetic system are not even limited to plants. The maize Activator (Ac) element belongs to the hAT family of transposons that have been characterized in animals including humans (Kempken and Windhofer, 2001). Heat shock promoters are known for a number of different systems (Balcells et al., 1994; Kilby et al., 1995) and modified fluorescent proteins are used widely. In principle, construction of a similar lineage marking system in other plant species and organisms would be straightforward.
Plant material and growth conditions
Arabidopsis thaliana plants, ecotype Landsberg erecta were used in the experiments described. Seeds for plants (hs-Ac) carrying the soybean heat shock promoter (Gmhsp 17.3-B) upstream of the maize Ac transposase gene (Balcells et al., 1994; Schoffl et al., 1984) were a generous gift from George Coupland's laboratory (Max-Planck-Institute for Plant Breeding Research, Köln, Germany). A gene fusion between an Arabidopsis histone 2b and yellow fluorescent protein (H2B-YFP) was constructed in our laboratory and used to generate transgenic Arabidopsis lines (Boisnard-Lorig et al., 2001). Seeds of Arabidopsis line N84726 (see Plasmids used) were obtained from the Nottingham Arabidopsis Stock Centre. Seeds were surface sterilized in 20% (v/v) Parazone (commercial bleach) for 15 min, washed twice with water and plated on media plates containing 1/2 strength MS basal salts + Gamborg's B5 vitamins (Sigma M-0404), 0.5 g l−1 MES, 0.7% agar (Sigma A-1296), pH 5.6 and sealed with gas permeable tape. Seedlings were grown under continuous light at 18–22°C. If required, seedlings were transferred to soil and grown in the greenhouse. Seeds were harvested and allowed to dry at room temperature for at least a week prior to use.
The Ds1 transposable element was PCR amplified from the Adh1-Fm335 allele in maize (generous gift from Liz Dennis, CSIRO Plant Industry, Canberra, Australia) using the following oligonucleotides: forward 5′-GGCAGATCTGGGACTGATAGGGATGAAAACGG-3′ and reverse 5′-GGCAGATCTCCTCAGTCCCTAGGGATGAAAGTGGTAATCCG-3′ (Sutton et al., 1984). The amplified Ds1 sequence was cloned into the vector pBIN35S-H2B-YFP (Boisnard-Lorig et al., 2001) between the CaMV 35S promoter and the H2B-YFP reporter gene, using the BamHI site downstream of the 35S promoter to make Ds-H2B-YFP. The cell membrane marker EGFP-LTI6b was PCR-amplified from genomic DNA of line N84726 (Cutler et al., 2000) using the following oligonucleotides: forward 5′-GGCGGAATCCAACAATGGTGAGCAAGGGCGAG-3′ and reverse 5′- GGCGAGCTCTCAAAAGGTGATGATATA-3′ and the PCR fragment was cloned into the pGEM-T-Easy vector (Promega, Madison, WI, USA), sequenced and cloned downstream (BamHI and SacI sites) of the CaMV 35S promoter into the pBIB binary vector carrying a hygromycin resistance selection gene to yield pBIB-35S-EGFP-LTI6b. All binary vectors were electroporated into the Agrobacterium tumefaciens strain GV3101 (Koncz and Schell, 1986). Arabidopsis plants were transformed with these vectors using the floral dip method (Clough and Bent, 1998). Primary transformants were selected on agar media containing kanamycin (50 μg ml−1) or hygromycin (40 μg ml−1). Plants from the T2 generation, characterized by resistance to selection agent, were used for crossing.
Heat shock conditions
hs-Ac/Ds-H2B-YFP seedlings growing on agar media were heat shocked at 42°C for 30 min unless stated otherwise. After heat shock, seedlings were transferred to 18–22°C.
Analysis of lineage patterns
Seedlings were analysed initially for YFP fluorescence using a Leica MZFLIII epi-fluorescence stereomicroscope (Filter sets - GFPPlus, excitation 480/40 nm, barrier 510 nm and YFP, excitation 510/20 nm, barrier 560/40 nm). Further analysis of marked lineages was done using a Leica TCS-SP confocal microscope (Leica, Milton Keynes, UK). In seedlings without the EGFP-LTI6b membrane marker, cells were outlined by staining with 5 μg ml−1 propidium iodide (Sigma, St. Louis, MO, USA) for 1.5 min, rinsed and mounted in water. To prevent seedling movement during high magnification confocal microscopy, seedlings were mounted between a microscope slide and a no. 0 coverslip in warm (35°C) 0.7% low melting point agarose, which was then allowed to set. EGFP was excited using the 488 nm line of an argon laser and propidium iodide and YFP were excited using the 514 nm line. Fluorescence emission was collected between 505–530 nm for EGFP, 527–553 nm for YFP, and 606–635 nm for propidium iodide. Specimens containing both EGFP and YFP fluorescence were imaged by sequential frame scanning.
Laser induction of Ds1 transposition
Seedlings were grown in Nunc coverglass-bottomed two-cell chambers (Fisher Scientific, Loughborough, UK) containing agar media as described above. Chambers were placed inside larger Petri dishes to prevent desiccation and seedlings grown in continuous light at 18–22°C in a growth room. At 3–5 days post-germination, roots had grown through the agarose to the bottom of the chamber and then continued to grow along the coverglass. Observation and heat shock treatment were performed with a Zeiss LSM 510 confocal microscope system equipped with an Axiovert 100M inverted microscope (Zeiss, Welwyn Garden City, UK). Prior to heat shock, primary root tips (n = 25) were imaged as a control using EGFP and YFP settings similar to those described above. Heating of specific cells within the root tip was accomplished by repeated scanning of the 488 nm argon ion laser line within a region-of-interest drawn over the meristem initials. Precise activation of the H2B-YFP marker within the region-of-interest occurred when a 100× 1.4 NA oil-immersion objective lens was used and the 25 mW laser was set to maximum output power. One hundred scans were performed at 2.5 sec intervals using the time-series function of the Zeiss LSM 510 software. Immediately after heat shock, the laser power was reduced and EGFP and YFP fluorescence was imaged. Subsequent images of the heat-shocked root tip were taken periodically during the next several minutes after adjustment of the microscope stage to compensate for root growth. Chambers containing heat-shocked seedlings were then returned to the growth room. Root tips of heat-shocked plants were imaged periodically during the following 72 h.
We are grateful to George Coupland and Liz Dennis for providing us with plasmids and Chris Hawes (Oxford Brookes University) for use of the Zeiss LSM 510 confocal microscope. This work was supported by grants from The Gatsby Charitable Foundation, Biotechnology and Biological Sciences Research Council, Human Frontiers Science Program, Ceres Inc. and AgrEvo. L.L. was supported by an EMBO Fellowship (ALTF 110-1999).