Presented here is a protocol for laser-capture microdissection (LCM) of plant tissues. LCM is a microscopic technique for isolating areas of tissue in a contamination-free manner. The procedure includes tissue fixation, paraffin embedding, sectioning, LCM and RNA extraction. RNA is used in the downstream tissue-specific, temporally resolved analysis of transcriptomes.
The development of a complex multicellular organism is governed by distinct cell types that have different transcriptional profiles. To identify transcriptional regulatory networks that govern developmental processes it is necessary to measure the spatial and temporal gene expression profiles of these individual cell types. Therefore, insight into the spatio-temporal control of gene expression is essential to gain understanding of how biological and developmental processes are regulated. Here, we describe a laser-capture microdissection (LCM) method to isolate small number of cells from three barley embryo organs over a time-course during germination followed by transcript profiling. The method consists of tissue fixation, tissue processing, paraffin embedding, sectioning, LCM and RNA extraction followed by real-time PCR or RNA-seq. This method has enabled us to obtain spatial and temporal profiles of seed organ transcriptomes from varying numbers of cells (tens to hundreds), providing much greater tissue-specificity than typical bulk-tissue analyses. From these data we were able to define and compare transcriptional regulatory networks as well as predict candidate regulatory transcription factors for individual tissues. The method should be applicable to other plant tissues with minimal optimization.
Plant development and growth involve the coordinated action of transcriptional regulatory networks within different cells that exist in a complex cellular environment. To understand the activity of these regulatory networks, we require the knowledge of spatial and temporal gene expression within different cell types across developmental stages. However, analyses of gene expression are more commonly conducted in whole organs or bulk tissue samples due to the technical challenge of isolating and analyzing small numbers of cells. The method we describe here has allowed obtaining spatial and temporal tissue-specific transcriptome analysis by coupling LCM with RNA-seq.
LCM was developed two decades ago by Emmert-Buck and colleagues1. The technique enabled researchers to precisely isolate single-cells or clusters of cells from their environment using direct microscopic visualization and manipulation with a narrow beam laser1. Since then the method has been widely used in cancer biology and pathology2,3. Many plant research groups have also adapted LCM for the use with different plant species and different tissue types4,5,6,7,8,9,10,11. Recently, several papers have also used LCM on eudicot and monocot seeds to study embryo, endosperms and other seed structures during seed development and germination10,12,13. Most of the other commonly used single-cell isolation methods such as micro-pipetting, cell sorting, magnetic separation and microfluidic platforms depend on the enzymatic digestion or mechanical homogenization to dissociate cells. This may perturb gene expression, introducing technical artefacts that confound data interpretation14,15. These methods also require previous knowledge of marker genes for each cell type to relate the dissociated cells to their spatial location and true cell-type. A further group of techniques depends on affinity-based isolation of subcellular structures instead of whole cells, for example INTACT (Isolation of Nuclei Tagged in Cell Types) and TRAP (Translating Ribosome Affinity Purification)16,17. However, affinity labeling and purification of nuclei or ribosomes are technically challenging in plant species that do not have well-established transformation protocols. LCM takes advantage of quick tissue fixation to preserve transcript levels and conventional histological identification by direct visualization of cells within their normal tissue/organ context, which allows discrete cells to be isolated in a short period of time18,19.
The protocol presented here is an optimized method for the isolation of specific cells or cell types from the tissue sections of cereal seeds, which can be applied to most of the cells that can be histologically identified. LCM provides a contact-free method of cell isolation, greatly reducing contamination and increasing integrity of recovered RNA. Furthermore, the method illustrates the power of LCM on large-scale genome wide studies starting with small quantities of biological materials. We also describe linear amplification of RNA for generating sufficient input material for downstream transcript/transcriptome analyses.
There are ten main steps in this LCM RNA-seq protocol for spatial and temporal tissue-specific transcriptomes, including fixation of tissue samples, dehydration, paraffin infiltration, embedding, sectioning, LCM, RNA extraction, RNA amplification, RNA quantification and qRT-PCR and/or RNA-seq (Figure 1).
Figure 1: Flowchart of LCM followed by RNA-seq or qRT-PCR. LCM is a spatially precise and contact-free technique to collect cells from fixed tissue sections using a laser beam under microscopic visualization. The process starts with fixation of tissue samples, followed by dehydration using a gradient series of ethanol and xylene, and finished with paraffin infiltration. The process can be fully automated by using a tissue processor. Once the tissue is infiltrated with paraffin, it is embedded in a mold with molten paraffin using an embedding station. Sectioning is carried out using microtome set to the desired thickness. Slides are prepared and LCM conducted immediately before RNA is to be extracted from captured cells. RNA extraction is followed directly by two rounds of RNA amplification prior to qRT-PCR and/or RNA-seq. Please click here to view a larger version of this figure.
As the final product is RNA, take care to avoid contaminating the work with RNases. Wearing gloves is a must. Use diethyl pyrocarbonate (DEPC) -treated water, buffers, etc. Autoclave buffers and bake glassware before use.
1. Tissue fixation
2. Tissue processing
3. Paraffin embedding
4. Preparation of polyethylene naphthalate (PEN) membrane slides
5. Sectioning
6. Laser-capture microdissection
7. RNA extraction
8. RNA amplification
9. RNA quantification
10. qRT-PCR and/or RNA-seq
We generated spatial and temporal tissue-specific transcriptomes from barley seeds during germination using our LCM RNA-seq protocol10. The study was carried out by applying LCM RNA-seq to small number of cells from three embryo organs (plumule, radicle tip, scutellum) every 8 h over a 48 h time course during germination (0-48 h, 7 time points) (Figure 2A,B).
Figure 2: Cells collected from barley seeds by LCM. (A) Longitudinal cross-section of a barley seed and, inset, an enlarged cartoon indicating the regions from which cells were captured. Blue = plumule, magenta = radicle, green = scutellum. (B) LCM of plumule, radicle, and scutellum. Top panel, tissue before laser pressure catapulting (LPC). Bottom panel, tissue after LPC. Cells were captured from 5 consecutive longitudinal sections of each sample with three biological replicates per sample. Each replicate consisted of approximately 200 cells (~0.05 – 0.15 mm2 total area). Bar = 100 µm. E = endosperm, C = crushed cell layer, SE = scutellar epithelium, S = scutellum. Figure reproduced from previous data in The Plant Journal with permission10. Please click here to view a larger version of this figure.
The fixation and embedding are critical steps as poorly fixed tissue sample can result in tissue damage and loss of RNA quantity and quality. These steps must be optimized and adjusted for different species and tissue types19,20,21. Here we used vacuum infiltration to improve the penetration of the fixatives into the seed tissues and cells. Dehydration and paraffin infiltration are conducted gradually to avoid damage to the tissue samples. We have accelerated the whole dehydration and paraffin infiltration process by using an automated tissue processor with agitation and vacuum applied at each step. This highly reduces the risk of RNA degradation occurring during the prolonged tissue processing steps conducted by a human operative. Before sectioning, it is crucial to treat the membrane slide to ensure paraffin adhesion and to make sure it is RNase-free. If the membrane slides are not prepared properly the sample section will not adhere to the membrane, which will affect the LCM catapulting step and decrease tissue transfer to the adhesive caps because the sample will separate from the membrane (see Figure 1).
Another critical step in the LCM RNA-seq protocol is LCM parameter optimization. There are five main parameters to be optimized: (1) cutting speed; (2) cutting energy; (3) cutting focus; (4) LPC energy; and (5) LPC focus. It is important to adjust the cutting parameters correctly in order to cut the selected area precisely and dislodge from surrounding tissue without burning the edge of selected area (Figure 3A,B). The proper LPC energy and focus are essential to reliably catapult the selected area from the slide to the special adhesive caps of collection tubes, and always inspect captured samples in the caps to ensure enough material is collected (Figure 3C). The efficiency of LCM enables 1000 cells to be captured within one hour, which greatly reduces the likelihood of RNA degradation during sample processing. The use of collection tubes with adhesive caps allows “dry” collection without any capturing liquid, greatly reducing possible RNases activities. It is important to keep the captured samples on ice and do the RNA extraction immediately after collecting all the replicates.
RNA extraction and amplification are carried out using commercially available kits by following the manufacturers’ instructions. Quantification of total RNA before RNA amplification using automated electrophoresis system will detect distinct electrophoretic bands and fluorescent peaks of 18S and 28S ribosomal subunits in good quality RNA samples (Figure 3D). Additional peaks corresponding to chloroplast and mitochondrial ribosomal RNA will be found before 18S and 28S peaks. Due to multiple RNA peaks in plant species, the RIN (RNA Integrity Number) values are typically lower than 9 (the expected RIN value from mammalian tissues) and do not accurately reflect the RNA integrity22. It is appropriate to continue to RNA amplification when clear 18S and 28S peaks are visible without sign of RNA degradation. The amount of RNA recovered in our barley seed experiment varied between three different tissue types (plumule, radicle and scutellum) and, also between early and late stages of germination, despite approximately equal numbers of cells being harvested. The amounts from barley seeds ranged from 100 pg to 20 ng from approximate 200 cells (0.5 to 100 pg per cell). The minimum amount of input RNA recommended for RNA amplification was 100 pg. The amount of RNA extracted depends on efficiency of extraction from different harvested cells and total transcript abundance in the particular cell-type at that stage of development19,21. It is useful to consider this when planning an LCM RNA-seq experiment, though such prior information will not be available in all cases.
Successfully synthesized aRNA will exhibit a unimodal, symmetrical size distribution from 100 to 1000 nucleotides with a peak around 300 nucleotides after two rounds of amplification (Figure 3E). In our barley seed experiment, we obtained 500 pg to 2 µg of aRNA after two rounds of RNA amplification10. Comparison between amplified and unamplified RNA has demonstrated that RNA amplification is reproducible, linear and does not introduce a systematic bias to transcript data23,24,25. The output aRNA can be used for qRT-PCR and/or RNA-seq for spatial and temporal tissue-specific gene expression analysis. Validation after RNA-seq must be conducted to corroborate results from analysis of aRNA. This can be done by examining expression of cell-type marker genes from the RNA-seq data using RNA in situ hybridization.
Figure 3: Representative result of LCM. (A) Tissue after cutting with proper cutting energy and focus. A clear fine line can be seen where the cut was made. The selected area will be dislodged from surrounding materials. Bar = 100 µm. (B) Tissue after cutting with incorrect cutting energy and focus. The selected area is still connected to surrounding materials. Bar = 100 µm. (C) Examining captured cells in special adhesive caps of collection tubes. Bar = 100 µm. (D) An example of total RNA profile before RNA amplification (left, gel-like image; right, electropherogram) with distinct electrophoretic bands and fluorescent peaks of 18S and 28S ribosomal subunits. Unlabeled peaks correspond to additional ribosomal RNA including chloroplast and mitochondrial ribosomes. (E) An example of a typical aRNA profile after RNA amplification (left, gel-like image; right, electropherogram). A distribution of sizes from 100 to 1000 nucleotides, with a peak around 300 nucleotides, was detected. Please click here to view a larger version of this figure.
RNA-seq analyses were performed on all samples in biological triplicates. A multidimensional scaling (MDS) plot of genes expressed in the different tissues over 48 h of germination illustrates greater similarity between samples of a single tissue than between samples from the same time point but different tissues (Figure 4A). We next set out to determine how extensively gene expression was differentially regulated in individual tissues over time during germination. The number of differentially expressed genes (DEGs) increased progressively over the course of germination in each tissue, relative to that tissue’s 0 h timepoint (Figure 4B). Twenty-five percent (910) of plumule, 34% (1876) of radicle and 41% (2562) of scutellum DEGs were found to be exclusively differentially expressed in that tissue (i.e., these genes were not differentially expressed in the other tissues, Figure 4C). Taken together, these results demonstrate how LCM RNA-seq can be used to understand temporal and spatial gene expression during plant development.
Figure 4: Analysis of RNA libraries of LCM samples showing distinct separation of different tissues over 48 h of barley germination. (A) Multidimensional scaling (MDS) plot of genes expressed in the different tissues over 48 h of germination. Each point represents 1 sample, and the distance between 2 points reflects the leading log fold change (FC) of the corresponding RNA samples. The x, y, and z axis represent first, second, and third dimensional logFC. P = plumule, R = radicle, S = scutellum, 0-48h = h of germination, R1-3 = Replicate 1-3. (B) Number of significant up-regulated (red) and down-regulated (blue) differentially expressed genes (DEGs) for six time points compared to time point 0 h for three tissue types. (C) Venn diagram showing number of DEGs in three tissue types. Genes in overlapping sets show the differential expression in two or three tissue types. Please click here to view a larger version of this figure.
Many tissue-specific gene expression studies have been limited by hand dissection of samples, which is time-consuming, labor intensive, has a high risk of contamination and can only utilize samples that a human operative is sufficiently dexterous to harvest. LCM is a precise and contact-free technique to collect cells from fixed tissue sections using a mechanically operated laser beam under microscopic visualization.
Good sample preparation is critical for LCM. The process relies upon proper fixing and embedding of samples to maintain a good balance between tissue morphology and integrity of RNA. The protocol presented here provides optimized fixing and embedding steps for barley seeds which includes longer fixing time and vacuum infiltration of fixatives. Analysis of different tissues or species will require optimization of the conditions, as these typically differ between tissues. Another important aspect of this protocol is the RNA amplification step which ensures sufficient amounts of RNA are generated for RNA-seq library construction. We find that RNA amplification results in high-quality libraries, but alternatively low input library preparation methods might be employed25. Histological identification of target cells is a critical step in our method and may provide challenges for rare or poorly characterized cell-types. Target cell recognition may be improved by using traditional histological stains or immunofluorescence stains as part of the preprocessing step or, alternatively, by using transgenic lines that express fluorescently labeled cell-specific markers26,27.
The LCM technique was first developed to be used on animal-derived samples for cell-type-specific transcript analysis1. However, a broad range of genomic, proteomic, and epigenetic analyses have subsequently been developed. Adapting these methods for application in plant biological studies is an ongoing challenge18,19,20,28. Schad et al. demonstrated that proteins and metabolites from Arabidopsis vascular bundles could be analyzed by using LCM with 2-D gel electrophoresis and mass spectrometry29,30. Recently, Latrasse et al. carried out chromatin immunoprecipitation (ChIP)-qPCR analysis on LCM-dissected stamens and carpels using LCM, establishing that tissue homogeneity improves the resolution of cell-specific epigenetic events31. Turco et al. also coupled LCM with bisulfite sequencing to study DNA methylation events during vascularization in sorghum, highlighting the tissue-specific difference between vascular and nonvascular tissues32. Interestingly, several studies have applied LCM to explore plant-microbe and plant-pathogen interactions33,34,35. The ability of LCM to visualize and capture cells at different infection stages provided information about the spatial and temporal responses of infected cells.
Our LCM protocol coupled with RNA-seq presented here will facilitate the spatio-temporal global profiling of gene expression of distinct cell types. With improved purification methods for chromatin and proteins along with enhanced mass spectrometry instruments, LCM will be an emerging tool to facilitate cell-type-specific epigenomic and proteomic studies in plants.
The authors have nothing to disclose.
This work was supported by the Australian Research Council Centre of Excellence in Plant Energy Biology (CE140100008) to JW. M.G.L was supported by a La Trobe University starting grant. We thank the La Trobe Genomics Platform for their support in high-throughput sequencing and data analysis. We thank Associate Professor Matthew Tucker for expert advice on establishing LCM in our lab.
Acetic acid 100 % ACS/R. | AnalaR NORMAPUR (BioStrategies) | VWRC20104.323 | |
AdhesiveCap 200 opaque | Zeiss | 415190-9181-000 | |
Clear base moulds 8 X 10 | Leica | 3803015 | |
Diethyl pyrocarbonate | Sigma-Aldrich | 40718-25ML | |
High Sensitivity RNA ScreenTape | Agilent | 5067-5579 | |
Lowprofile disp.blades DB80LS | Leica | 14035843489 | |
MembraneSlide 1.0 PEN | Zeiss | 415190-9041-000 | |
MessageAmp II aRNA Amplification Kit | Ambion (ThermoFisher) | AMB17515 | |
On-Column DNase I Digestion Set | Sigma-Aldrich | DNASE70 | |
Ovation RNA-Seq System V2 | NuGen (Integrated Science) | 7102-08 | |
Paraffin (Surgipath Paraplast) | Leica | 39601006 | |
PicoPure RNA Isolation Kit | ABI (ThermoFisher) | KIT0214 | |
RNaseZap RNase Decontamination Solution | Ambion (ThermoFisher) | AM9780 | |
Xylene | AnalaR NORMAPUR (BioStrategies) | VWRC28975.360 | |
Leica Benchtop Tissue Processor | Leica Biosystems | TP1020 | |
Leica Heated Paraffin Embedding Module | Leica Biosystems | EG1150H | |
Leica Cold Plate | Leica Biosystems | EG1150C | |
Safemate Class 2 Biological Safety Cabinets | LAF Technologies Pty Ltd | Safemate 1.5 | |
Leica Fully Automated Rotary Microtome | Leica Biosystems | RM2265 | with PALMRobo v 4.6 software |
Zeiss PALM MicroBeam LCM system | Zeiss miscroscopy | ||
TapeStation | Agilent | TapeStation 2200 |