Here we present a protocol for laser-assisted microdissection of specific plant cell types for transcriptional profiling. While the protocol is suitable for different species and cell types, the focus is on highly inaccessible cells of the female germline important for sexual and apomictic reproduction in the crucifer genus Boechera.
The understanding of developmental processes at the molecular level requires insights into transcriptional regulation, and thus the transcriptome, at the level of individual cell types. While the methods described here are generally applicable to a wide range of species and cell types, our research focuses on plant reproduction. Plant cultivation and seed production is of crucial importance for human and animal nutrition. A detailed understanding of the regulatory networks that govern the formation of the reproductive lineage (germline) and ultimately of seeds is a precondition for the targeted manipulation of plant reproduction. In particular, the engineering of apomixis (asexual reproduction through seeds) into crop plants promises great improvements, as it leads to the formation of clonal seeds that are genetically identical to the mother plant. Consequently, the cell types of the female germline are of major importance for the understanding and engineering of apomixis. However, as the corresponding cells are deeply embedded within the floral tissues, they are very difficult to access for experimental analyses, including cell-type specific transcriptomics. To overcome this limitation, sections of individual cells can be isolated by laser-assisted microdissection (LAM). While LAM in combination with transcriptional profiling allows the identification of genes and pathways active in any cell type with high specificity, establishing a suitable protocol can be challenging. Specifically, the quality of RNA obtained after LAM can be compromised, especially when small, single cells are targeted. To circumvent this problem, we have established a workflow for LAM that reproducibly results in high RNA quality that is well suitable for transcriptomics, as exemplified here by the isolation of cells of the female germline in apomictic Boechera. In this protocol, procedures are described for tissue preparation and LAM, also with regard to RNA extraction and quality control.
In transcriptional studies done at the tissue level, the transcriptomes of highly specialized but rare cell types are often masked by the more abundant surrounding cells. An example for such highly specialized cell types are the cells of the female reproductive lineage (germline) in plants. The female germline is specified within developing ovules, the precursors of seeds inside the gynoecium of the flower 1,2. The megaspore mother cell (MMC) is the first cell of the female germline. It undergoes meiosis to form a tetrad of reduced megaspores. Typically, only one of these megaspores survives and divides mitotically without cytokinesis, i.e., in a syncytium. These mitoses are followed by cellularization to form the mature gametophyte, which typically consists of four cell types: three antipodals, two synergid cells, the egg, and the central cell. The egg and central cells are the female gametes that get fertilized by two sperm cells during double fertilization to give rise to the embryo and endosperm of the developing seed 1,2. In the sexual model system Arabidopsis thaliana, only ~ 50 seeds develop per flower while about 50 – 80 seeds develop per flower in the closely related genus Boechera. Thus, the female germline consists of only a few highly specialized cell types, making it an excellent model to study developmental processes, such as cell specification and differentiation.
Moreover, insights into the gene regulatory processes governing plant reproduction can be of applied value. In plants, both sexual and asexual reproduction through seeds (apomixis) can occur. While sexual reproduction generates genetic diversity in a population, apomixis leads to the formation of clonal offspring that is genetically identical to the mother plant. Therefore, apomixis has great potential for applications in agriculture and seed production, as even complex maternal genotypes can be maintained unaltered over several generations 3,4,5. Because apomixis does not naturally occur in any major crop species, the engineering of apomixis in crops is of great interest 3,4,5. However, this long-term goal is difficult to achieve because the underlying genetic and molecular basis of apomixis is not understood in sufficient detail 6.
To gain insights into the transcriptional basis governing apomictic reproduction, cell type-specific transcriptional profiling using laser-assisted microdissection (LAM) and next generation sequencing (NGS) represents a very powerful approach 7,8. LAM has first been established for animal and biomedical research. In the past few years LAM has also been applied to plant biology 6,9,10. In contrast to other methods allowing profiling of individual cell and tissue types, LAM does not require the generation of marker lines 6,9,10. Therefore, it can be applied to any cell or tissue type without prior molecular knowledge. Another advantage of LAM is that it can be applied to any cell type as long as the cell can be recognized in dry sections based on position and/or structural features. LAM has the additional advantage that fixed tissues are used, which prevents changes of the transcriptional profile during processing.
The tissue of interest, e.g., floral tissue, is fixed in a non-crosslinking fixative prior to embedding in paraffin wax. Embedding in paraffin wax can be done manually, following established protocols 9,11. However, the use of an automated tissue processor for dehydration and infiltration with the wax generally results in higher reproducibility in terms of the conservation of RNA quality and tissue morphology. The alternative strategy of embedding tissues in resin has also been successfully used for cell type-specific analyses by LAM 8. However, the use of an automated tissue processor for embedding in wax is very time efficient, as many samples can be processed at once requiring a minimum of hands-on time. While typically no significant loss of RNA quality occurs during fixation and embedding, the preparation of thin sections with the microtome and, in particular, the mounting on the frameslides used for LAM remains a critical step for preservation of RNA quality. This has previously been noted and the use of a tape transfer system has been described to result in better RNA quality at this step 12. However, this adds an additional time-consuming step during preparation of the slides and also requires special equipment. The optimized protocol described below reproducibly produces RNA that is of sufficient quality for transcriptional profiling with GeneCHIPs and Next Generation Sequencing (NGS) approaches 7,11,13,14. In addition, with the laser microdissection microscope used, a high purity of the isolated cell types is routinely produced 7,11,13,14.
The genus Boechera is an excellent model system for studying the key steps of apomictic reproduction. In Boechera, a variety of different sexual and apomictic accessions have been identified and can be used for comparative analyses 15,16,17. In a comparison of cell type-specific transcriptomes of cells from the female germline from sexual Arabidopsis and apomictic Boechera, we identified genes and pathways that are differentially expressed, thereby identifying new aspects of the regulatory processes governing apomixis 7. In addition, this study verified the suitability of LAM for cell type-specific transcriptional analyses of small and rare cell types. We have already used this protocol for the analysis of different cell types in a variety of plant species, but species- and tissue-specific modifications to the protocol may be required in certain cases.
Note: This protocol describes tissue preparation, laser assisted microdissection, and RNA extraction for transcriptional profiling. Always use gloves throughout all steps of the protocol. Study and consider the safety instructions for each chemical used. In particular, keep in mind that Xylol is harmful and can penetrate gloves and that methanol is toxic. For all instruments used, please refer to user manuals accordingly.
1. Removal of RNAse Activity from Glassware and Other Equipment
2. Tissue Fixation
3. Tissue Embedding, Thin Sectioning, and Mounting on Slides for LAM
4. Laser-assisted Microdissection (LAM)
Note: The procedure for LAM will vary with the instrument used. Certain details of this protocol are adapted to a specific instrument and the technology of collecting the samples on a cap making use of electrostatic forces. In addition, details might vary even between different versions of the software. Please refer to manufacturer's instructions and user handbooks for a detailed description and specific instructions on instrument and software.
5. RNA Isolation and Quality Control
Note: RNA can be isolated by any method suitable for small amounts of RNA. In this protocol the use of an RNA isolation kit specified for small amounts of RNA is described. Follow the manufacturer's instructions.
Sample Preparation and LAM are done in Consecutive Steps
A number of consecutive steps are required to prepare RNA for transcriptional analysis from selected cell types by LAM (Figure 1). This starts with the harvest of the flowers and immediate fixation to ensure that the RNA population remains unchanged after harvest. The tissue is embedded, sectioned, and mounted on slides. This allows the isolation of the cells by LAM and the pooling of cell type-specific sections on one or several caps (Figure 1). The material can be frozen afterwards or directly processed by RNA extraction. Apart from the advantage of LAM to be applicable for cell type-specific analyses in both model and non-model species, the method remains rather time-consuming. At least one week of working time, sometimes more, is required for the steps from tissue harvest to RNA extraction (Figure 1). It needs to be taken into account that often sections harvested over several days of LAM are pooled in one biological sample and RNA extraction. For Boechera egg cells, the use of at least ~ 200 sections per sample is highly recommended, with up to ~ 100 sections isolated per day.
Isolation of the Cell Types of Interest Depends on their Visibility in Dry Sections
Isolation of the cell type-specific sections is the most time-consuming step of the protocol. So far, no automation of this step has been developed due to the complex nature of the samples. The identification of the cells of interest using a software tool is not feasible because the dry sections have low contrast and the section plane varies between the samples due to the fact that ovules are oriented at different angles within the tissue (Figure 2). Nevertheless, the unique morphology of the mature female gametophyte with the synergids, egg cell, and the central cell (antipodal cells degenerate before fertilization) allows the unambiguous identification of the egg cells by the researcher (Figure 2B, C; Figure 3). Indeed, in Boechera divaricarpa, the egg cell often shrinks a bit during preparation and thus separates from the surrounding tissues, a feature advantageous for the isolation of egg cell populations at high purity (Figure 2B, C; Figure 3).
Visual Inspection of the Cap after LAM is Recommended
Visual inspection of the cap surface after LAM allows the identification of any possible contaminations of the sample, e.g., by dust. In addition, it is helpful to ensure that the sections stick to the cap, as occasionally sections get lost during the procedure. Typically, many selected sections, e.g., egg cells, can be seen on one cap after several hours of LAM.
The Established Workflow Reproducibly Results in Good RNA Quality
From cell type-specific LAM isolations from female gametophytes, total RNA amounts are only in the low ng range. This limiting amount of RNA makes it necessary to use a representative control isolated from larger tissues of each slide as an approximation for RNA quality control after LAM. This is demonstrated by a comparison of cell type-specific RNA from B. divaricarpa egg cells as compared to controls from larger tissue areas isolated from the same slides, indicating similar RNA quality (Figure 4A, B). Using this method, a good to high RNA quality with RIN numbers ≥ 7 is reproducibly obtained. The RNA quality achieved was suitable for cell type-specific RNA-Seq, leading to the identification of 236 genes expressed in the apomictic egg cell but not in the apomictic central cell, synergid cell, or apomictic initial cell, nor in the cells or the mature sexual gametophyte of Arabidopsis thaliana (Figure 5) 7.
Figure 1: Scheme of the Steps of the Protocol, Indicating the Minimal Time Required per Step. Fixation of the flowers or tissues of interest is done overnight. The next day, embedding is started, which runs over night, resulting in embedded material at day 3 of the protocol. At day 3 microtome sections starting from embedded samples in wax blocks can be generated. The ribbons of the thin paraffin sections are mounted on slides and left to dry overnight. One to several days of LAM will be required to obtain enough material for one biological replicate. After RNA isolation and quality control, library preparation for RNA-Seq can be performed to prepare transcriptome analysis. Please click here to view a larger version of this figure.
Figure 2: Egg Cells are Clearly Identifiable in Thin Sections, Due to the Characteristic Morphology of the Female Gametophyte. (A) Schematic drawing of the mature female gametophyte (Polygonum type). (B – C) Thin sections through ovules harboring female gametophytes in Boechera divaricarpa. Egg cells are clearly visible. Due to the morphology of the female gametophyte often not all cell types (egg cell: e, synergids: s, central cell: cc) are visible in a single section. Scale bars = 50 µm. Please click here to view a larger version of this figure.
Figure 3. LAM of Egg Cells of Boechera divaricarpa. A and C. Section of the mature gametophyte of B. divaricarpa at 7 µm before and, B and D, after microdissection with the laser device (egg cell: e, synergids: s, central cell: cc). Scale bars = 50 µm. Please click here to view a larger version of this figure.
Figure 4. RNA Quality of Laser-Dissected Egg Cells and Control Sections. (A) RNA quality as analyzed from egg cell sections as compared to (B) larger tissue areas harvested from the same slides as controls. Please click here to view a larger version of this figure.
Figure 5. Identification of Genes Expressed Only in the Apomictic Egg Cell as Compared to the Cells of Apomictic and Sexual Mature Gametophytes or the Apomictic Initial Cell. Heatmap of 236 genes expressed in the apomictic egg cell but neither in the apomictic central cell, synergid cell, or apomictic initial cell of B. gunnisoniana nor the cells of the mature sexual gametophyte of A. thaliana7 Hierarchical clustering of samples and genes was based on Euclidean distance and hierarchical agglomerative clustering. Red denotes high expression and black low expression. Colors are scaled per row. apo: apomictic. Please click here to view a larger version of this figure.
The Protocol is Suitable for Different Cell and Tissue Types
LAM combined with transcriptome analyses by microarrays or RNA-Seq is a valuable tool to gain insights into the specific patterns of gene activity regulating developmental or physiological processes 7-11,13,14. However, the suitability of this method for any given cell type is critically dependent on structural issues. The cell needs to be clearly visible and unambiguously identifiable in the dry sections used for LAM. In Boechera divaricarpa, the morphology of the gametophyte allows an easy identification of the egg cell (Figure 2, Figure 3). In the end, the speed of isolation of a specific cell population also depends on the frequency at which the cell type can be found on one slide, as exchanging slides and scanning new slides are time-consuming steps. Given that dependent on the cell type, at least 200 – 250 cell sections should be pooled per sample, the suitability of the method for a certain study design depends largely on the time requirements for the LAM step.
In addition, depending on the morphology of the tissue, it might be advantageous in terms of structure to reduce the thickness of the sections to 6 µm. Similarly, thicker sections of ideally ≥ 8 µm typically result in higher RNA quantity and RNA of slightly higher quality from the same amount of tissue sections. In addition, the cells of interest should have at least a diameter of 8 – 10 µm to make isolation by LAM feasible.
In principle, the protocol described here can be adjusted to different cell and tissue types from distantly related species. However, the RNA quality can vary even between different cell types of the same species 7,10,13. This protocol has been adjusted based on small and critical cell types. Thus, it can now be applied to a broader range of samples without further adjustments. Slight modifications of the protocol might nevertheless be required, depending on the cell type and species under investigation. It is recommended to first use both alternative fixatives (see protocol 1.1) when setting up the method for a new species or cell type for a comparison of morphology and RNA quality.
In principal, the protocol can be adapted to and performed with different instruments suitable for laser microdissection 9. However, it needs to be noted that the instruments vary both in laser power and the minimal width of the laser beam that can be applied, as well as in the technology of sampling. Particularly for the isolation of small cells and tissue areas, lasers resulting in a narrow laser path are better suited. The laser beam of the instrument we use is rather fine (~ 1 µm broad) and allows dissection of small cells. The sampling technology of collecting the cells on a sticky cap surface by electrostatic adhesion is particularly advantageous for small cell types and the isolation of single cell sections. This minimizes the risk of sample loss by electrostatic effects.
The RNA Quality Obtained is Important for Further Transcriptional Analyses
One of the most critical points for successful RNA-Seq library preparation is RNA quality. While several providers of amplification kits optimized their technology for RNA of even lower integrity, the quality of the data obtained after transcriptome analysis by either microarrays or RNA-Seq typically increases with the quality of the input RNA. Using this established protocol, good and highly reproducible RNA quality for different developmental stages of the female reproductive tissues and the germline in Arabidopsis and Boechera was obtained (e.g., Figure 4). While the method is applicable also for more distantly related species (e.g., tomato, unpublished), it needs to be taken into account that small optimizations or modifications may be required depending on the species, tissue type, and even the laboratory environment as, for example, a high humidity might compromise RNA integrity during slide preparation and LAM.
A critical step during the protocol is the mounting of the sectioned wax stripes containing the samples to the slides. Here, in particular the contact to water is a critical step. While replacing water by EtOH does not result in any significant improvement of RNA integrity after isolation (unpublished), the replacement of water by methanol does. However, methanol should only be handled under the chemical hood and with great care. Depending on the sample type, as an alternative to the use of methanol, the drying times after mounting with water can be reduced to ~ 2 hr (see 3.3.3.1), resulting in equally good RNA quality. Performing a trial to test the RNA quality achieved by mounting either on water or methanol for the cell type and species of interest is recommended at the beginning of a new project.
LAM is a Powerful Technology for Transcriptional Analyses
In conclusion, we have successfully optimized and applied the combination of LAM and cell type-specific transcriptome analysis by microarrays and RNA-Seq to different cells of the sexual and apomictic germline lineage in Arabidopsis thaliana and in Boechera spp., thus allowing comparative transcriptional analyses (also see Figure 5) 7,11,13,14. The method described bears a number of advantages as compared to other technologies allowing cell type-specific transcriptional analysis. Importantly, depending on the recognizability of the cell types or tissues of interest in the section, it can be applied to rare cell types of both model and non-model species, such as the cells of the mature female gametophyte in Boechera spp., or it can even be used to isolate cellular domains, e.g., polar halves of cells 19. Similar applications would not be possible with other methods that rely on fluorescent activated sorting of cells or nuclei (FACS/FANS) 10.
A major drawback of the method is the time requirement. While the fixation and embedding does not require extended hands-on time and can simultaneously been done for a large amount of flowers, LAM as such is very time-consuming and, for a cell type-specific analysis, typically 3 days to 3 weeks of LAM (on average 5 hr per day at the laser microscope) needs to be planned. In this respect, it is helpful to do some pre-tests for the speed of isolation of the specific cell type targeted to properly design the study. In addition, even when using this optimized protocol, trials to test the RNA quality are highly recommended.
In summary, LAM is a powerful tool for the transcriptional analysis at the resolution of individual cell types or even subdomains of cells, which could not be achieved using alternative existing methods. In this respect, it bears tremendous potential to allow analyses, e.g., of developmental processes, with a very high temporal and spatial resolution. This was exemplified by the analyses of the transcriptomes of cells at all key steps of sexual and apomictic reproduction in Arabidopsis and Boechera7,11,13,14.
The authors have nothing to disclose.
We thank Timothy F. Sharbel (IPK Gatersleben) for providing Boechera divaricarpa seeds and Sharon Kessler (University of Oklahoma) for critical reading and proofreading. Work on cell type-specific transcriptome analyses to study gametophyte development and apomixis in UG´s laboratory is supported by the University of Zürich, by a fellowship of the “Deutsche Forschungsgemeinschaft” and the Marie Curie project IDEAGENA to AS, by grants from the “Staatssekretariat für Bildung und Forschung” in the framework of COST action FA0903 (to UG and AS) and the Swiss National Foundation (to UG).
Ethanol | VWR | 1,009,861,000 | absolute EMPROVE Ph Eur,BP,USP |
15 ml falcon centrifuge tubes | VWR | 62406-200 | |
2100 Bioanalyzer | Agilent | G2939AA | |
Acetic Acid | Applichem | A3686,2500 | 100% Molecular biology grade |
Ambion Nuclease free water | life technologies | AM9932 | |
ASP200 S | Leica | 14048043624 | tissue processor |
black cardboard | can be purchased in special paper shops | ||
DNA- and RNAse-free Frame Slides | Micro Dissect GmbH | 1,4 µm PET-membrane; can also be purchased from Leica | |
Dumond Forceps | Actimed | 0208-5SPSF-PS | |
ethanol lamp | |||
exsiccator | Sigma-Aldrich | Z354074-1EA | Nalgene Vaccuum Dessicator or similar equipment |
filter tips 10 µl | Axon Lab AG | AL60010 | can be replaced by similar tips |
filter tips 1000 µl | Axon Lab AG | AL60010 | can be replaced by similar tips |
filter tips 20 µl | Axon Lab AG | AL60020 | can be replaced by similar tips |
filter tips 200 µl | Axon Lab AG | AL60200 | can be replaced by similar tips |
forceps precision | VWE | 232-1221 | |
glass slide holder | Huber & Co.AG | 10.0540.01 | Färbekästen nach Hellendahl |
glass staining trough | Huber & Co.AG | 10.0570.01 | Färbekasten |
Heated Paraffin Embedding Module Leica | Leica | Leica EG 1150 H | blocking station, similar devises are suitable |
Heating and Drying Table | Medax | 15501 | other models and/or suppliers are suitable |
ice bucket | VWR ice bucket with lid | 10146-184 | similar buckets equally suitable |
light table | UVP An Analytical Jena Company | TW-26 | white light transluminator |
microscope slide | Thermo Scientific | 10143562CE | cut edges |
microtome blade | Thermo Scientific | FEAHS35 | S35 microtome blade disposable |
MMI Cell Cut Plus Instrument | MMI (Molecular Machines and Industries) | ||
Non-stick, RNAse free Microfuge tubes, 2ml | life technologies | AM12475 | |
Paraplast X-TRA | Roth | X882.2 | for histology |
PicoPure RNA Isolation Kit | life technologies | KIT0204 | Arcturus PicoPure RNA Isolation Kit |
plastic balancing trays | Semadeni AG | 2513 | |
plastic box | Semadeni AG | 2971 | Plastikdose PS |
plastic lid for heating plate | homemade | ||
preparation needle | VWR | 631-7159 | |
RNA 6000 Pico Kit | Agilent | 5067-1513 | |
RNAse free microfuge tubes | life technologies | AM12400 | |
RNAse ZAP Decontamination Solution | life technologies | AM9780 | |
Semi-automated Rotary Microtome | Leica | RM2245 | similar devises are equally suitable |
Tissue Loc Histo Screen Cassettes | Thermo Scientific | C-1000_AQ | similar cassettes of other suppliers are suitable |
Tubes with adhesive lid, without diffusor 500 µl | MMI (Molecular Machines and Industries) | 50204 | |
Xylol (Isomere) ROTIPURAN | VWR | 4436.2 | min. 99 %, p.a.,ACS, ISO SP |
process embedding cassettes | Leica | 14039440000 | Leica Jet Cassette I without lid |
Universal Oven | Memmert | UF55 | other models and/or suppliers are suitable |