Full Thesis is on Research Gate DOI:10.13140/RG.2.2.14708.63368
CHAPTER 5
WHOLE-MOUNT IN SITU HYBRIDIZATION IN RICE
5.1 Introduction
Studying gene expression patterns is an essential component in the characterization of gene function. Where and when a gene is expressed will limit its spatial-temporal function and provide information on the extent to which a gene can play a role in plant responses to chemical or environmental treatments. A number of strategies have been regularly employed in order to investigate gene expression patterns. These include the use of transgenic reporter constructs, gene chip technology, immunohistomchemical analysis and nucleic acid, (particularly mRNA) hybridization. In reporter constructs, detectable protein products such as β-glucuronidase (Jefferson et. al, 1987), green fluorescent protein (Shimomura et al., 1962) and luciferase (Strehler and McElroy, 1949) are often placed under transcriptional control of specific promoters of the studied genes. A more advanced gene chip analysis using microarray technology such as Affymetrix GeneChips is also a valuable tool, especially in obtaining genome-wide expression profiling (Auer et al., 2009). However these methods are often considered insufficient in providing accurate identification of gene expression and generally need to be validated against other methods involving detection of proteins and mRNA (Hejátko et al., 2006). In particular, although historically most methods of gene expression analysis involved extraction of RNA and protein from whole organs, these methods lose important spatial aspects of gene expression. This is especially so if the biological target of investigation is essentially cellular in nature, as is the case for stomata. Since a protein product is often the end result of result gene expression, protein patterns can be localized and characterized using immunohistomchemical techniques (Komar and Long, 2013). mRNA expression can also be visualised to cellular resolution using in situ hybridization techniques (Javelle et al., 2011).
Classically, mRNA localization uses thinly sectioned biological samples which have been embedded in paraffin wax to preserve the structure. However this step adds more time to the entire procedure as well as being laborious and needs sufficient sectioning and mounting skills. Furthermore data comes in fragments of serial sections, sometimes providing challenges in visualising patterns in 3D. As a result an idea to analyse whole tissues arose, originally performed on wild-type Drosophila fly embryos (Tautz and Pfeifle, 1989) and hence termed whole-mount in situ hybridization (WISH). This method has been used successfully on tissues from animals such as chicken embryos (Darnell et al., 2010), zebra fish embryos (Christine and Bernard, 2008), mouse embryos (Neufeld et al., 2013) and freshwater planarian, a type of flatworm (Rybak-Wolf and Solana, 2014). However WISH performed on plants has not been as successful with relatively few reports on its regular application. For plants, WISH has been most extensively optimized for Arabidopsis (Engler et al., 1998; Brewer et al., 2006; Traas, 2008) but its use has been restricted to small, transparent tissues such as roots and embryos. This narrow application of WISH in plants is probably due to several technical problems, particularly the presence of cell walls that hinder reagent permeability, relatively deeper target tissue in more mature organs, as well as naturally low mRNA levels in certain cells. Consequently the choice of tissue is restricted to simple ones and does appear as useful for more complex, denser tissues such as floral and shoot apical meristems (SAM).
To my current knowledge, there has not been any WISH protocols reported for rice so far. In this thesis the aim was to characterise the development of stomata in early leaf development. Analyses of Arabidopsis and grasses, including rice, have identified a series of genes involved in stomatal patterning and differentiation. With the longer term view of analysing the patterns of expression of these genes, this chapter sets out to establish a WISH protocol for rice leaves. As well as establishing a protocol for rice, a WISH method for this plant might also enable progress to be made in the analysis of spatial gene expression in other monocots.
5.1.1 The principle of WISH
The principle of WISH is shown in Fig. 5.1. An RNA probe (riboprobe or simply probe) is labelled to produce detectable signals should the hybridization with the targeted cellular mRNA (Fig. 5.1A) take place. Digoxigenin (DIG, a steroid found exclusively in Digitalis purpurea plant) was the probe label used in this study, although compared to radiolabelled probes it can lack similar sensitivity. However it has the advantage of convenience and has been widely used in various in situ hybridisation protocols.
An antisense labelled probe (Fig. 5.1B) hybridizes within the tissue of interest which has been treated to both maintain structure integrity but to allow probe permeability. If hybridization takes place, subsequent addition of an anti-digoxigenin antibody which has an alkaline phosphatase (AP) enzyme attached to it leads to a complex in which the position of the AP indicates the localisation of the target mRNA. Visualization is achieved by enzymatic reaction of AP with a colourless substrate mixture containing BCIP (5-bromo-4-chloro-3-indolyl-phosphate), and NBT (4-nitro blue tetrazolium). The reaction leads to a resultant dark blue-purple precipitate (Fig. 5.1C). Fig. 5.1D shows an example of localized dark blue precipitation in certain cells using an eFF1a (elongation factor) probe. WISH is faster than conventional in situ hybridization since it does not require tissue sectioning. Nevertheless, setting up the procedure involves extensive optimisation, which is described in the first part of this chapter. Apart from probe preparation, other critical steps in WISH include tissue fixation-permeation, hybridization, signal detection and imaging, as well as general experimental setup and selection of target tissue.
In theory, WISH would be a powerful tool to study stomatal development in rice given their appearance on the leaf surface. Thus deeper tissue penetration should not be an issue for the probes.
Figure 5.1 Overview of steps for in situ hybridization method. Tissue is treated to preserve cell structure with enhanced permeability to allow hybridization with the targeted mRNA in the cytoplasm (A). DIG-labelled riboprobe binds to the complementary mRNA (B) and allowed to react with an anti-DIG antibody that carries the alkaline phosphatase enzyme. Addition of the substrate 5-bromo-4-chloro-3-indolyl-phosphate results in a localized blue-purple precipitation (C). (D) An example of a localized detection of blue-purple precipitate in a developing P4 leaf. |
Ideally I wanted to make probes for genes related to stomatal development in rice such as OsSPCH, OsMUTE and OsFAMA but due to time constraints all the probes used in this study were obtained from colleagues (Julia Van Campen and Supatthra Narawatthana). These probes (described in Table 1) belong to genes involved in different developmental processes which have been studied in rice or other orthologous plants, or whose exact functions are yet to be determined. These genes will be used to develop the WISH protocol for rice tissues. For example, the expression of histone H4 can be used as a marker for S phase of the cell cycle, thus should be expressed in some cells early in leaf development (Meshi et al., 2000; Nelson et al., 2002) whereas DL (DROOPING LEAF), which is known to promote midrib formation, should be expressed only within the middle region of the leaf (Ohmori et al., 2011). If these genes show expected expression patterns, this provides confidence in the patterns observed for other target genes at a later stage.
5.2 Aim
1. Develop a whole mount in situ hybridisation method for the analysis of gene expression in rice leaves with the ultimate aim of characterising stomatal gene expression during early rice leaf development.
5.3 Methodology
In general in situ hybridization is sensitive to contaminants especially ribonucleases (RNases that catalyse RNA degradation) which are ubiquitous in laboratory environments. Therefore the work area was regularly cleaned using RNaseZap® (Ambion Inc.). All solutions were prepared using DEPC-treated (Diethyl pyrocarbonate) water (0.1% v/v in UHP water, shaken vigorously and incubated at 37°C before being autoclaved to inactivate DEPC). All apparatus and equipment were also treated to ensure that they were RNase free by either baking (150° for 3h) or rinsing in DEPC-treated water followed by autoclaving. All chemicals used were of the highest grade and purity possible.
The initial attempt to perform WISH on rice in this study was fundamentally based on the Arabidopsis protocol by Traas (2008) who adapted it from Ludevit et al. (1992) and Zachgo et al. (2000). The authors cautioned that this method is only reliable in a limited number of plant parts such as root meristems, embryos, and very young primordia. Moreover interpretation of results must be carried out with extra care. Since a number of optimisation steps were required to develop the final WISH method, this section is reported in three subsections; (i) performing WISH on rice using a protocol adapted for Arabidopsis, (ii) modifications of certain steps to prevent non-specific binding of either probes or antibodies, (iii) optimization of cutting of transverse sections to allow better penetration of reagents and probes into the tissue and easier interpretation of the data and (iv) the final method used for gene examining expression patterns for different probes.
All plants were grown under high light conditions using methodology described in section 2.1. In subsection (i) of this experiment, primordia at the P3 and P4 stages, which were relatively soft and small to work, with were used. Microdissection using 25G and 30G needles was performed to obtain tissue primordia. For subsection (ii) only P4 primordia were used. When the optimised method had been achieved, the plant material used was a 1 cm section of tissue from the base of the rice culm containing all layers of different plastochron stages. This particular cutting (in stub form) had leaf no. 7 as the youngest and leaf no. 3 as the oldest leaf layer.
All riboprobes (probes) in this study were obtained from Julia Van Campen prepared using modified methods of Narawatthana, (2013) and Jane Langdale (Oxford lab, personal communication). Probe details are described in Table 5.1. Briefly, total RNA was extracted from rice leaf primordia using a TRIzol method (optimized from Narawatthana, (2013) and Van Campen, unpublished). This was used to generate cDNA and subsequently used as template in PCR amplifications. The purified PCR products were ligated to the KpnI/SacI double digested pBluescript II SK (-) plasmid which contains the promoter for T3 or T7 RNA polymerase. Primers used in this study were designed using Primer3. The recombinant plasmids were used for synthesizing sense (T7) and antisense (T3) riboprobes using the respective RNA polymerase. The digested and linearized plasmids were used for in vitro transcription where probes were labelled with DIG. Probe hydrolysis was carried out for probes > 500 bp since short probe fragments facilitate tissue penetration.
5.3.1 Unmodified WISH method for Arabidopsis performed on rice
Initially the WISH method developed for Arabidopsis was performed on rice to determine which aspects of the methodology worked, and which did not work on rice. Plants tissues were fixed in 4% (v/v) formaldehyde and 10% (v/v) dimethyl sulfoxide in a glass vial containing 0.1M PBS solution for 30 min. Tissues were left in fixative overnight at 4ºC. Samples were then retrieved and dehydrated through an ethanol series: 30%, 60%, 70%, 85%, 95% and 100% for 30 min in each concentration. Samples were transferred to 1.5mL Eppendorf tubes and kept in 100% ethanol at -20 ºC until needed. Samples were then retrieved and washed twice with 100% ethanol, then xylene for 30 min and for four washes in 100% ethanol, followed by twice in methanol. Samples were transferred to 50% methanol (v/v) and 50% PBTF (0.01M PBS; 0.1% v/v Tween-20; 4% formaldehyde) solution and incubated for 5 min. The solution was changed and samples were post-fixed with 100% PBTF for 25 min then washed five times in PBT (0.01M PBS; 0.1% v/v Tween-20) for 10 min per wash. Samples were then incubated in PBTK (4% w/v Proteinase-K; PBT) for 10 min and washed four times in PBT for 5 min per wash. Samples were post-fixed again in PBTF for 25 min and washed four times in PBT for 5 min per wash. Samples were then left in PBT for 60 min.
Table 5.1:
Riboprobes used for the development of the whole mount in-situ hybridization method in rice. Further gene details can be found by entering locus ID at www.rice.plantbiology.msu.edu. All antisense probes were generated using T3 RNA polymerase while T7 RNA polymerase was used for sense probes. cDNAs were cloned into the pBS vector and linearised with Kpn1 or Xho1 for T3 probes or SacI or Xba1 for T7 probes.
Riboprobe | Locus ID | Riboprobe length (bp) | Annotation |
eEF1a | LOC_Os03g08010 | 118 | Elongation factor Tu, putative, expressed |
H4 (Histone-4) | LOC_Os10g39410 | 301 | Core histone H2A/H2B/H3/H4 domain containing protein, putative, expressed |
DL (Drooping Leaf) | LOC_Os03g11600 | 436 | YABBY domain containing protein, putative, expressed |
DWF7 (Delta-7-Sterol-C5) | LOC_Os01g04260 | 452 | Fatty acid hydroxylase, putative, expressed |
THF1 (Thylakoid Formation-1) | LOC_Os07g37250 | 343 | THYLAKOID FORMATION1, chloroplast precursor, putative, expressed |
CAB (Chlorophyll A-B) | LOC_Os01g41710 | 378 | Chlorophyll A-B binding protein, putative, expressed |
COV1 | LOC_Os02g16880 | 360 | Protein of unknown function DUF502 domain containing protein, expressed |
CUL1 (Cullin-1) | LOC_Os05g05700 | 458 | Cullin, putative, expressed |
MON4 (MONOPTEROS) | LOC_Os04g56850 | 343 | Auxin response factor, putative, expressed |
FACKEL | LOC_Os09g39220 | 610 | Delta14-sterol reductase, putative, expressed |
CDKb2 | LOC_Os08g40170 | 650 | Cyclin-dependent kinase B2-1, putative, expressed |
Equal volume of PBT and WMHB (hybridisation buffer containing 50% v/v formamide; 0.5M SSC buffer; 5% w/v heparin salt; 10% w/v single-stranded salmon testes DNA; 0.1% v/v Tween-20) was used in the last wash. Samples were rinsed twice in 100% WMHB and pre-hybridized in WMHB for 3 h at 55 ºC in a water bath. The buffer used for pre-hybridization was then replaced with fresh pre-warmed WMHB and 2μL of denatured DIG-labelled probe was added. Hybridisation was carried out for 18 h at 55 ºC. Samples were gradually washed in PBT solution mixed with WMHB at 25%, 50% and 100% concentration for 20 min each. Samples were further washed in 100% PBT for 5 min. The PBT solution was replaced with new PBT solution containing antibody (diluted 1:2000) and incubated for 3 h. The antibody solution was discarded and samples were rinsed in PBT three times for 5 min each. Samples were then equilibrated for 5 min in detection buffer (0.1M Tris-HCl pH 9.5; 0.1M NaCl; 0.05M MgCl2; water). Samples were reacted in BCIP/NBT mix (0.0075% w/v BCIP; 0.015% w/v NBT; detection buffer) in the dark at room temperature for 36 h. The enzyme reaction was stopped in TE buffer (0.1M Tris-HCl pH 9.5; 0.001M EDTA; water). Samples were dehydrated through an ethanol series (30%, 60%, 70%, 85%, 95% and 100%, 90 sec each), followed by Histoclear twice for 90 sec. Samples were mounted on a microscope slide and observed using OLYMPUS BX51 compound light microscope.
Fig. 5.2 shows results using the unmodified Arabidopsis method on the P3 and P4 rice leaf primordia. The probe used for this trial was eEF1A, the eukaryotic elongation factor gene, as it is constitutively expressed in developing leaves. In general the unmodified Arabidopsis method did work as there was positive signal (dark purple staining) detected for the eEF1A antisense probe for both P3 (Fig. 5.2 A) and P4 (Fig. 5.2 B) primordia, whereas the eEF1A sense probe gave no expression except for some purplish staining only in limited areas (Fig. 5.2 C). eEF1A abundance in cells reflects its primary roles in the synthesis of all cellular proteins, especially essential in delivering aminoacyl-tRNA (aa-tRNA) to the elongating ribosome, thus its presence is essential for cell viability (Cottrelle et al., 1985; Sasikumar et al., 2012). It is a good choice for an endogenous control in this study since the expression in plants should be relatively high in meristematic regions which are active in protein synthesis and lower in older, metabolically less active regions (Pokalsky et al., 1989).
Despite the overwhelming positive staining observed for the antisense probe, as expected for a prolonged staining time, it was interesting to note how strongly zoned the expression pattern was in P4 primordia (Fig. 5.2 B). The staining only occurred in the mesophyll regions of the leaf rather than in the veins or bulliform cells (BC) regions. BC or motor cells are highly vacuolated and thin-walled cells (Jane and Chiang, 1991) that are responsible for leaves’ ability to roll up thus conserving water content during water stress through reduced transpiration rate (Hsiao et al. 1984). Since BC contain large vacuoles, it was expected to see limited antisense probe signal, as reported here. Even though this method seemed to work with the antisense-probe, the result for sense probe still showed some staining in some leaf areas, especially the x-shaped silica bodies (Fig. 5.2C), thus indicating the possibility of non-specific binding of probes or the antibody used. Due to the promising results, attempts were made to optimise WISH by focusing on the blocking steps to prevent non-specific binding of probes and antibody.
5.3.2 Optimization of the blocking reagent to prevent non-specific hybridization and binding.
Non-specific binding could occur due to both probes and/or antibodies. To test the idea that the antibody was the issue, the original unmodified method as previously described was used and the P4 stage primordia was the tissue of choice. In this experiment one sample was left to hybridize without the presence of eEF1A-sense probe but with antibody addition, while the other sample was left to hybridize with the eEF1A-sense probe but without any antibody addition. The sense probe was used as a control to further confirm the finding.
Fig. 5.3 confirms this suspicion as the leaf treated with antibody only produced a similar signal pattern as before (Fig. 5.3 C) where the staining tended to take place in some leaf areas and x-shaped silica bodies (Fig. 5.3 A and 5.3 B). WISH on leaf tissue without any antibody addition resulted in a completely clear tissue (Fig. 5.3 C), thus showing that the WMHB blocking solution used was good enough to prevent non-specific probe interactions.
The next step involved the introduction of antibody blocking reagents to the protocol by introducing bovine serum albumin (BSA), a common blocking agent that would bind to charged sites (Burry, 2009). The BSA solution (1% w/v BSA; 0.3% v/v Triton X-100; 0.1M Tris-HCl and 0.15M NaCl) alone was used as a Post-WMHB solution in all the steps mentioned in sections 5.3.4.3 and 5.3.4.4. Results using the same protocol with BSA solution addition and P4 tissue as earlier, showed that although background staining still present in some leaf areas, staining on x-shaped silica bodies was absent (Fig. 5.4 A). It was hypothesized that if the antibody’s affinity could be specified only for the defined epitopes, this should remove the unwanted background staining.
Figure 5.2
eEF1A mRNA expression pattern in rice using the unmodified WISH protocol for Arabidopsis. Strong antisense signal (dark purple staining following 36 h staining time) are visible in the young P3 (A) and P4 (B) leaf primordia. In general there is no signal detected for the sense probe (C) except in some interveinal gap (IG) areas (circled) and x-shaped silica bodies located on the veins (arrowheads). For any given IG area there is a vein (V), mesophyll cells (M) and bulliform cells (B) underneath the epidermis.
To further attempt to reduce the background staining acetone powder (Ace-Pow) containing preadsorbed antibody on crushed leaf tissues was prepared and a small amount added to the antibody solution. Acetone powder was prepared by grinding leaves of young rice seedlings and acetone in liquid N2. This solution was centrifuged (10000 RPM for 2 min), decanted and left to air dry to obtain the pellet (powder). The presence of the powder in the antibody solution further reduced non-specific binding (Fig. 5.4 B). Then both BSA solution and Ace-Pow were combined in the next attempt which resulted in even greater reduction in the background staining (Fig. 5.4 C). Finally, to complement BSA as a blocking agent, sheep serum (10% v/v, Sigma) was added into the mixture (which now became the Post-WMHB solution used throughout the experiment) and together with Ace-Pow this completely prevented the background staining (Fig. 5.4 D).
5.3.3 Cutting of transverse samples for better riboprobe penetration into tissues and interpretation of results
WISH was performed using the optimized blocking reagents from the previous subsections on P3 leaf primordia (Fig. 5.5 A) but now using a probe for a histone-4 (H4) gene. This probe was used in all further experiments as the eEF1A probe ran out. H4 expression is linked to the S phase of the cell cycle, leading to localised and high levels of signal in tissue containing dividing cells, expected to be present in P3 leaves. In this experiment the staining time was reduced to about 30-45 min (checked for signals every 15 min) before stopping the reaction. In several experiments, there was no staining observed for the H4 sense probe (Fig. 5.5 B), thus giving confidence in the blocking reagents used (as optimized in the previous subsections). The results consistently showed that the expression of H4 antisense probe occurred in the deeper tissue layer and was concentrated near the bottom of the P3 primordium (Fig. 5.5 C).
Therefore, to improve visualisation of the H4 antisense probe deeper in the tissue a whole leaf cluster (in stub form, one cm from the culm base) was sampled for transverse sectioning as shown in Fig. 5.6. This allowed the expression pattern and localization could be better understood in leaves of different ages and stages. This way of sampling and sectioning of tissues was thus adopted for studying gene expression patterns related to leaf development throughout the rest of this study.
5.3.4 Optimised WISH for rice
Following the modifications to the Arabidopsis protocol described above, the final methodology developed for rice is described below:
5.3.4.1 Tissue selection, fixation and permeabilization
Rice leaf primordia of either P3 or P4 stage or clusters of leaf layers cut from about 1 cm from the culm base (Fig. 5.6 A) were used for this study (the former only for preliminary attempts whilst virtually all the results reported here used the latter sample source). Plants tissues that were used for some probes were slightly different in terms of age but the outermost layer (the oldest) in the leaf cluster was always leaf no. 3 (Fig. 5.6 B). Samples were cleaned, trimmed and cut using a single edged blade to produce approximately 1 cm tall stubs. Samples were fixed in Eppendorf tube containing 1000 µl of fixative (4% v/v methanol-free formaldehyde, Sigma; 10% v/v dimethyl sulfoxide, Sigma; 0.1M PBS) overnight in a 37°C orbital shaker at 50 rpm. This is important to ensure thorough and better fixative penetration in all leaf layers. Samples were rinsed 2 x for 5 min each in PBS (pH 7.4) on a rocking table. From this step onwards all subsequent steps were done on a rocking table (at room temperature, 20-25°C) to ensure thorough reagent penetration into the tissues. Samples were dehydrated in an ethanol series of 30%, 70%, 95% and 2x 100% for 5 min each. If needed, samples could be stored in 100% ethanol for several weeks in -20°C freezer. Upon retrieval samples were brought out and left to equilibrate to room temperature before proceeding with the subsequent steps.
Figure 5.3
Staining pattern in P4 leaf tissue using the unmodified WISH protocol except by incorporating either eEF1A-sense probe or antibody only. The presence of antibody alone causes positive purple staining in some leaf regions (A) and when magnified (B) reveals staining in some interveinal gap areas and x-shaped silica bodies (arrowheads). When no antibody was added to the hybridized eEF1A-sense probe, no signal was obtained (C). V indicates vein.
Figure 5.4:
Background staining pattern in P4 leaf tissue following the introduction of different mixtures of antibody blocking reagents in specific WISH steps. Bovine serum albumin (BSA) solution only prevents x-shaped silica bodies staining on the veins (v) but not in interveinal gap regions (A). Incubation of antibody in acetone powder (Ace-Pow) substantially reduce background staining (B, circled) and the combination of both BSA and Ace-Pow further refines the result (C, circled). The coupling of sheep serum and BSA together with Ace-Pow synergistically prevents background staining, thus producing a completely clear tissue (D).
5.3.4.2 Post-fixation and pre-hybridization treatments
Samples were post-fixed in absolute methanol (Fluka), 2 x for 2 min each, to further permeabilize the cells and precipitate remaining protein molecules. The solution was then changed to 1:1 methanol and PBTF (4% v/v methanol-free formaldehyde; 0.1M PBS; 0.1% v/v Tween-20) for 5 min and then to 100% PBTF for 30 min. Samples were rinsed in PBT (0.01M PBS; 0.1% v/v Tween-20) 4 x for 10 min each. Samples were incubated in PBTK (4% w/v Proteinase-K, Sigma; PBT) for 15 min and rinsed in PBT 4 x for 10 min each. Digestion with Proteinase-K is crucial for ensuring successful hybridization through cell lysis and inactivation of nucleases that might otherwise degrade the mRNA of interest. However, insufficient digestion will result in a reduced hybridization signal but if too long tissue integrity is compromised. This all depends on the physical nature of the tissue; the larger samples need stronger and longer Proteinase-K treatment. Samples were post-fixed again in PBTF for 30 min and rinsed in PBT 4x for 5 min each. Samples were prepared for hybridization by washing in half-strength WISH buffer (WMHB) (50% v/v formamide, Sigma; 0.5M SSC buffer; 5% w/v heparin salt, Sigma; 10% w/v single-stranded salmon testes DNA, Sigma; 0.1% v/v Tween-20) 2 x for 2 min each. Formamide is a solvent used to lower the melting point and annealing temperature of nucleic acid strands during hybridization (Betty et al., 1969) together with high salt concentration provided by SSC to reduce stringency thus promoting hybridization (Schildkraut, 1965). The rate and quality of hybridization is further achieved by incorporating anionic macromolecules such as single-stranded salmon sperm DNA that prevents non-specific probe interactions (Bancroft and Marilyn, 2008) while heparin salt in general reduces background for DIG-labelled probes (James et al., 1994). Samples were then prehybridized in new full strength WMHB for 3 h at 55°C in a water bath. This is important to equilibrate samples with the WMHB solutions thus blocking sites of non-specific interactions before the addition of probes.
Figure 5.5: mRNA expression pattern of the rice Histone-4 gene in P3 leaf primordia. WISH was performed on the P3 tissue was dissected out. SEM micrograph (A) shows the whole structure of P3. The sense probe yielded no signals (B) and looks clear under magnification. The antisense probe produced signals (C) concentrated near the bottom of primordium and when magnified reveals dotted/speckled staining pattern. |
Figure 5.6:
Longitudinal section of a typical rice culm at the base (A), exposing leaf layers of different plastochron (P) stages and shoot apical meristem (SAM). Dashed line indicates a typical cutting level (about 1cm from the culm base) used for the whole mount in-situ to reveal the most number of leaf layers when being viewed from the top. The magnified image is the resultant stub cut (Image by Ahmad Nazrin Zakaria). When the stub is viewed from the top (B), it always shows leaf no. 3 (L3) as the outermost and oldest leaf layer.
5.3.4.3 Hybridization and antibody incubations
Samples were replaced with pre-warmed (55°C) WMHB. 2 µL of 1/10 dilution DIG-labelled probe was added per reaction. Samples were hybridized for 20 h at 55° in a water bath. Samples were then thoroughly washed using fresh WMHB 5 x for 25 min each. Post-WMHB solution (10% v/v sheep serum, Sigma; 1% w/v bovine serum albumin, Sigma; 0.08M Tris-HCl, Sigma; 0.12M NaCl; 0.24% v/v Triton X-100, Sigma) was gradually introduced to the samples (Post-WMHB: WMHB) at 25%, 50% and 100% for 20 min each. Samples were further washed in Post-WMHB 3 x for 5 min each. Post-WMHB solution is a mix containing proteins (sheep and bovine serums) that coat any membranes/surfaces in the samples that have high affinity for proteins so that the antibody used to bind with the DIG-labelled probes will not bind everywhere non-specifically thus reducing background staining (Otto, 1993). Samples were then incubated for at least 3 h on ice in an antibody solution (0.1% v/v Anti-Digoxigenin-AP Fab fragments, Roche; Acetone powder from young leaves; Post-WMHB) which had been prepared one day earlier and kept in dark at 4°C.
5.3.4.4 Colorimetric signal detection and imaging
The antibody solution was discarded and samples were rinsed in Post-WMHB 3 x for 5 min each. Samples were equilibrated for 5 min in detection buffer (0.1M Tris-HCl pH 9.5; 0.1M NaCl; 0.05M MgCl2; water). Samples were reacted in BCIP/NBT mix (0.0075% w/v BCIP, Roche; 0.015% w/v NBT, Roche; detection buffer) in the dark at room temperature until sufficient colour development was detected for antisense probes. The mechanism for colour development has been described in section 5.1.1. This step varied depending on the probes used so during incubation period the signal was checked every 10-15 min until the staining produced was deemed satisfactory. Successfully hybridized mRNA was detected as dark blue-purple staining within cells. To a certain extent background staining could be present but normally a lighter blue-purple colour. Reactions were stopped in TE buffer (0.1M Tris-HCl pH 9.5; 0.001M EDTA; water) 2 x for 10 min each. Samples were carefully manipulated using fine tweezers and placed onto glass slides and constantly hydrated as necessary using the TE buffer while being viewed and photographed under a OLYMPUS BX51 compound light microscope. Images were captured and processed using a OLYMPUS DP71 camera and built-in CELL-A software.
5.4 Results
Once the WISH protocol for rice had been optimised it was used to explore the expression patterns of the genes listed in Table 5.1.
The H4 gene is linked to the S phase of the cell cycle. The H4 antisense probe was clearly observed as dotted purple stain in the younger leaf (leaf no. 5 - 7) cluster and young axillary leaf cluster (Fig. 5.7 A) indicating tissue containing actively dividing cells. The same localized expression pattern was absent in older leaves (L3 and L4) which had reached maturity. Signals were completely absent when the sense probe was used (Fig. 5.7 B) again showing the effectiveness of the blocking reagents used in this method. Furthermore, it only took 45 min for the staining to be clearly perceived, thus a great time saving compared to the original method. However, artefact staining was also present, although it appeared as brownish stains and usually occur in lignified dermal layers. Nevertheless this view of the transversely cut leaf cluster in general gave a clear view of expression localization, thus was employed as a way of cutting for the subsequent attempts with different probes, described below.
Five of the nine other genes probes examined (CDKb2, DL, MON4, COV1 and DWF7) showed positive results with staining only observed for the antisense probes (Fig. 5.8 – 5.12). Interestingly the probe for CYCLIN-DEPENDENT-KINASE (CDKb2) gene had a similar expression pattern as in H4 probe earlier which was dotted in appearance and only in the youngest leaf layers (Fig. 5.8 A, L6 and L7). In a younger L5, expression was only observed in vascular region but no expression at all was seen in older leaf layers (L3 and L4) or samples hybridized with the sense probe (Fig. 5.8 B). The antisense probe for the DL (DROOPING LEAF) gene was specifically expressed in dark vertical bands within the middle region of the leaf that would become the midrib in mature leaf (Fig. 5.9 A). Furthermore the expression only happened in the two youngest leaf blades (L6 and L7) with no expression in older leaf layers (L5-L3, which were basically leaf sheaths thus had no midribs) or in the sense probes (Fig. 5.9 B).
Figure 5.7:
mRNA expression pattern of rice Histone-4 gene in a transversely cut culm containing a cluster of leaf layers. The outermost and oldest layer is leaf no. 3 (L3) and the leaf layers gets progressively smaller and younger towards the cluster centre. (A) shows positive expression pattern (dotted purple stain) for antisense probe in relatively younger leaf layers (L5-7) and the inset image is the magnified centre region of the cluster. No signal is detected anywhere for the sense probe in (B). Axillary leaf cluster (AX) that will give rise to a new tiller are also present. Artefact staining (non-purple) is present in lignified dermal layers (arrowheads).
The antisense probe for the MONOPTEROS (MON4) gene also showed a specific expression localized in vascular bundle regions (Fig. 5.10 A), consistent with a role in vascular differentiation and patterning, especially in developing and relatively younger leaf layers (L5-L7). Such expression was absent in more mature leaf sheath layers (L3 and L4) and in the sense probe samples (Fig. 5.10 B). The expression of the CONTINUOUS VASCULAR RING (COV1) gene was highest in the youngest leaf sheath layer (L4) and it persisted in the oldest L3 sheath layer where staining was observed in the septum region between lacuna in which vascular bundles could be found in the adaxial part (Fig. 5.11 A). Staining was much paler in the youngest leaves (L5 and L6) and completely absent for the sense probe (Fig. 5.11 B). The expression of the DELTA-7-STEROL-C5-6-DESATURASE (DWF7) gene also showed the same pattern as in COV1 where it occurred highly in the youngest leaf sheath layer (L4) particularly around lacuna edges, in the septum regions of mature L3 sheath layer but much paler in the youngest L6 and L7 (Fig. 5.12 A) and no expression in samples hybridized with the sense probe (Fig. 5.12 B).
The expression of the CHLOROPHYLL A-B BINDING PROTEIN (CAB) was only high in the mesophyll areas of developing leaf layers (L5 and L6) but in areas that would give rise to lacuna and vascular bundles, the purplish signal was much paler (Fig. 5.13 A). The expression was also weaker in the youngest L7 samples and the more mature leaf sheath layers (L3 and L4) that contained fully formed lacuna and vascular bundles, whereas sense probe yielded no signal (Fig. 5.13 B). CULLIN-1 (CUL1) also showed some positive expression for the antisense probe, especially around the lacuna edges (Fig. 5.14 A). Even though the staining intensity was not as prominent as with the other probes reported before, it was higher than that observed with the sense probe which produced essentially clear tissue (Fig. 5.14 B).
Despite the promising results produced by the optimised WISH on the majority of the probes tested, certain probes yielded unsatisfactory results. They were probes for THYLAKOID FORMATION FACTOR (THF1) and DELTA-14-STEROL REDUCTASE (FACKEL) genes. Both THF1 (Fig. 5.15 A and B) and FACKEL (Fig. 5.16 A and B) produced signal not only with the antisense probes but with the sense probes as well.
Figure 5.8: mRNA expression pattern of rice Cyclin-dependent kinase B2-1 (CDKb2) gene. The outermost and oldest layer is leaf no. 3 (L3) and the leaf layer gets progressively smaller and younger towards the cluster centre. (A) shows positive expression pattern (dotted purplish-crimson stain) for antisense probe in the youngest leaf layers (L6 and 7) and in the axillary leaf clusters (AX). Expression also localized within vascular bundle regions (circled). No equivalent staining is detected anywhere for the sense probe in (B). Artefact staining (non-purple) is present in lignified dermal layers (arrowheads). |
Figure 5.9:
mRNA expression pattern of rice Drooping Leaf (DL) gene. The outermost and oldest layer is leaf no. 3 (L3) and the leaf layer gets progressively smaller and younger towards the cluster centre. (A) shows positive expression pattern (dark vertical bands) for antisense probe in the youngest leaf layers (L6 and 7) which localized within presumptive midrib regions. No equivalent staining is detected anywhere for the sense probe in (B). Axillary leaf cluster (AX) that will give rise to a new tiller are also present. Artefact staining (non-purple) is present in lignified dermal layers (arrowheads).
Figure 5.10:
mRNA expression pattern of rice MONOPTEROS (MON4) gene. The outermost and oldest layer is leaf no. 3 (L3) and the leaf layer gets progressively smaller and younger towards the cluster centre. (A) shows positive expression localization for antisense probe in the three youngest leaf layers (L5 - 7) within vascular bundle regions (circled) and in axillary leaf cluster (AX). No equivalent staining is detected anywhere for the sense probe in (B). Artefact staining (non-purple) is present in lignified dermal layers (arrowheads).
Figure 5.11:
mRNA expression pattern of rice CONTINUOUS VASCULAR RING (COV1) gene. The outermost and oldest layer is leaf no. 3 (L3) and the leaf layers gets progressively smaller and younger towards the cluster centre. (A) shows high positive expression (purple stain) for antisense probe in the youngest leaf sheath layer (L4) and also localized in the septum regions between lacuna (circled) in the older sheath layer (L3). No equivalent staining is detected anywhere for the sense probe in (B). Artefact staining (non-purple) is present in lignified dermal layers (arrowheads).
Figure 5.12
mRNA expression pattern of rice DELTA-7-STEROL-C5 (DWF7) gene. The outermost and oldest layer is leaf no. 3 (L3) and the leaf layers get progressively smaller and younger towards the cluster centre. (A) shows high positive expression (purple stain) for antisense probe in the youngest leaf sheath layer (L4) especially at the edge of lacuna (circled in L4) and also localized in the septum regions between lacuna in the older sheath layer (circled in L3). No equivalent staining is detected anywhere for the sense probe in (B). (AX) is the axillary leaf cluster and non-purple artefact staining is present in lignified dermal layers (arrowheads).
Figure 5.13:
mRNA expression pattern of rice CHLOROPHYLL A-B (CAB) gene. The outermost and oldest layer is leaf no. 3 (L3) and the leaf layer gets progressively smaller and younger towards the cluster centre. (A) shows high positive expression (purple stain) for antisense probe in the mesophyll areas of developing leaf layers (L5 and L6). Expression is much paler in the presumptive lacuna (circled) and vascular bundle (VB) areas, and this is true for axillary cluster (AX) region as well. No equivalent staining is detected in the youngest L7, the more mature leaf sheath layers (L3 and L4) that contained fully formed lacuna and VB and also for the sense probe in (B). Non-purple artefact staining is present in lignified dermal layers (arrowheads).
Figure 5.14:
mRNA expression pattern of rice CULLIN-1 (CUL1) gene. The outermost and oldest layer is leaf no. 3 (L3) and the leaf layer gets progressively smaller and younger towards the cluster centre. (A) shows positive expression (slight dark stain) for antisense probe in the youngest leaf sheath layer (L4) especially at the edge of lacuna (circled) . No equivalent staining is detected anywhere for the sense probe in (B). (AX) is the axillary leaf cluster and non-purple artefact staining is present in lignified dermal layers (arrowheads).
Figure 5.15:
False-positive mRNA expression pattern of rice THYLAKOID FORMATION FACTOR (THF1) gene. The outermost and oldest layer is leaf no. 3 (L3) and the leaf layer gets progressively smaller and younger towards the cluster centre. (A) shows positive staining (purple) for antisense especially in the youngest leaves (L6 and L7) . Similar purple staining is detected in L5 and L6 for the sense probe in (B). (AX) is the axillary leaf cluster and non-purple artefact staining is present in lignified dermal layers (arrowheads).
Figure 5.16:
False-positive mRNA expression pattern of rice DELTA-14-STEROL REDUCTASE (FACKEL) gene. The outermost and oldest layer is leaf no. 3 (L3) and the leaf layer gets progressively smaller and younger towards the cluster centre. (A) shows positive staining (purple) for antisense especially in the youngest leaves (L5-L7) . Similar purple staining pattern is detected in L5-L7 for the sense probe in (B). (AX) is the axillary leaf cluster and non-purple artefact staining is present in lignified dermal layers (arrowheads).
5.5 Discussion
The aim of the results reported in this chapter was to establish a whole mount in situ hybridisation (WISH) procedure for the analysis of gene expression in rice leaves, particularly for future studies of stomatal development in the epidermis. Due to time restrictions, it was not possible to extend the analysis to stomata-related genes but for the first time a WISH protocol that worked in a reliable fashion for a number of genes was developed for rice.
WISH experiments using Histone-4 revealed an interesting finding. Using P3 leaf primordia, the signal for the antisense probe was concentrated near the base of the primordium and seemed to be in the deeper tissue layer (Fig. 5.5 C). This pattern of signal fits with the expected localisation of cell division towards the leaf base and to Histone-H4 being expressed only during the S phase of the cell cycle, thus leading to a spotty pattern (Nelson et al., 2002). In addition, transverse cutting of an intact leaf cluster allowed better tissue penetration of the reagents and probes allowing gene expression patterns deeper in the tissue to be more easily visualised since chronologically each primordium is at a unique differentiation or maturity stage, each having particular tissues or structures (Fig. 5.6 B). For example the experiment using the H4 antisense probe showed that Histone-4 expression was zoned in the youngest leaf layers in the cluster (Fig. 5.7 A).
In traditional in situ hybridization methods thin tissue sections (2-dimensional) are taken while the WISH method uses whole tissues (3-dimensional) allowing better interpretation of gene expression patterns as well as taking a significantly shorter amount of time to perform (3 days vs. one week). Comparison of the traditional in situ hybridisation of the H4 gene probe to leaf cross-sections (Van Campen, unpublished) and WISH, showed that the expression patterns were highly comparable. This suggests that the pattern obtained using WISH reflected the endogenous pattern of Histone-H4 expression. Thus for all other gene probes reported in this chapter, the optimized WISH on transverse cut leaf clusters (WISH-TC) was employed.
Experiments using WISH-TC with the probes for other genes of various functions yielded convincing and reliable results for seven out of the nine tested (limited staining when hybridized with the sense probes, signal observed with anti-sense probe). When compared with the control H4 expression (Fig. 5.7 A), cyclin-dependent kinase (CDK) had the most similar expression localization and appearance in which the spotty staining patterns were detected in the youngest leaf layers and in the axillary leaf cluster (Fig. 5.8 A). This finding is similar to the previously reported work using Arabidopsis (Segers et al., 1996; Andersen et al., 2008). In these reports Segers et al. (1996) summarize that the second type of CDK gene (CDKb2, as used in this study) shows expression in the cell cycle (thus can be a candidate endogenous control besides H4) while Andersen et al. (2008) highlight the general importance of CDK genes in ensuring normal cell cycle progression and for meristem organization, with high CDKb2 genes expression in the SAM during the cell cycle essential to prevent meristematic defects.
WISH-TC has also proven to be useful in detecting gene expression which is tightly confined in a very specific leaf area, as in DL (DROOPING LEAF) expression (Fig. 5.9 A). DL expression appears as dark vertical bands in a presumptive midrib location in the leaf blade. This result is similar to the expression reported by Ohmori et al., (2011) who also worked with rice. DL plays two important roles in rice, namely leaf midrib formation and flower carpel specification (Yamaguchi et al., 2004). Since DL is thought to promote a larger midrib, it plays a role in the formation of erect leaves, thus can be a target in rice leaf engineering. Leaf erectness is undoubtedly a useful agronomic trait in rice because erect leaves perceive more sunlight even for lower leaves and require smaller growing area than those with floppier leaves. In fact it has been proven to increase biomass production since erect rice plants can be fitted into a more densely spaced crop (Sakamoto et al., 2006).
Positive expression was also detected using probes for genes that have roles in leaf vasculature. MONOPTEROS (MON4) is a type of auxin response factor (ARF) and in this study its expression was highly localized in actively developing vascular bundle systems (Fig. 5.10 A) and this agrees with the localization found abundantly in emerging veins of Arabidopsis (Wenzel et al., 2007). ARF transcription factors bind on promoter regions of genes and regulate expression with respect to auxin level. Auxin itself is a phytohormone whose role is essential in vascular patterning since its high level is detected in procambial cells in Arabidopsis (Mattson et al., 2003) whilst local application promotes vascular strand formation (Sachs, 1981). Meanwhile CONTINUOUS VASCULAR RING (COV1) product is a putative integral membrane protein and is likely to play a role in the maintenance or the initiation of a defined vascular bundle (Parker et al., 2003). WISH-TC supports this hypothesis by showing even COV1 expression in the younger leaf sheaf layer (vascular bundles yet to develop) while in the mature leaf sheath layer expression is concentrated only in the septum space between lacunas where a continuous ring of vascular cambium is located (Fig. 5.11 A). To a certain extent, the expression of DELTA-7-STEROL-C5-DESATURASE (DWF7) is similar to COV1 (Fig. 5.12). Comparison of these results with published data show that in Arabidopsis DWF7 protein localization (linked to oleosin formation) is also detected in leaf vascular tissue (Silvestro, 2013). It is interesting that DWF7 expression is concentrated around the edges of the lacuna. If oleosin is present in rice it could reflect a role in desiccation in this region (Hsieh and Huang, 2007).
Genes associated with photosynthesis were also tested using WISH-TC, namely CHLOROPHYLL A-B BINDING PROTEIN (CAB). The observed expression pattern Fig. 5.13) agrees with Vainstein et al. (1989) who reported that CAB expression is abundant in both bundle sheath cells and mesophyll cells of maize leaves. The lack of expression in the mature leaf sheath layers is consistent with the idea that CAB levels decline in relatively older leaves as reported in barley leaves (Humbeck and Krupinska, 2003). A higher expression pattern for this probe might have been expected, but since the protein is encoded for a gene family more work is needed to investigate whether different members of the family show differential expression during leaf development. Final positive expression, although not as pronounced as other probes discussed earlier, was obtained using a probe for a CULLIN-1 (CUL1) gene. This was similar to DWF7 expression in that was concentrated at the edges of lacuna (Fig. 5.9). Plants produce three primary types of CULLIN and CUL1 is one of them that assembles the CULLIN-RING E3 ubiquitin ligase complex which controls many aspects of plant adaptation and development (Hua and Vierstra, 2011). The functional interpretation of this expression pattern requires more work but is an example of how WISH can reveal interesting patterns.
Despite the reliability and versatility of the WISH-TC method presented so far, it is still not entirely perfect since it can give data which are difficult to interpret (e.g. with the THYLAKOID FORMATION FACTOR (THF1) and DELTA-14-STEROL REDUCTASE (FACKEL) probes (Fig. 5.15 and 5.16). Further optimization steps may be necessary so that the existing WISH-TC work as well with these probes. For example it may be necessary to reduce the probe length, especially for FACKEL (610 bp, Table 5.1), since it is a general practice to have probe length <500 bp to ensure good tissue penetration. Staining time and ambient temperature are factors that need to be taken into account as well since the staining times are rather long for these probes (almost 5hours for THF1 and 30hours for FACKEL). Since the colour development is an enzymatic reaction, performing it in a constant temperature environment (23-25°C) in an incubator rather than on the bench will give a more moderate staining speed thus more power at stopping the reaction once sufficient signal has been detected. It may also be good to further enhance the blocking power, for example by adding 5% non-fat dry milk powder. As pointed out earlier, no single protein mixture can be a universal blocker for all probes. Finally in order to overcome one persistent shortcoming in WISH-TC (non-purple artefact staining of the lignified dermal layers) a number of approaches could be used to solubilize the lignin, such as microwave-assisted alkali treatment and liquid ammonia (Janker-Obermeier et al., 2012; Strassberger et al., 2015). Nevertheless these additional steps for further optimization of WISH should be done with great care and one at a time so that the general robustness of the method on rice is not compromised.
The optimized WISH method presented in this study has been shown to work on nine out of eleven probes for genes of various roles and functions. Taken as a whole, the optimization of WISH for rice is important since it opens the door to the more rapid analysis of the spatial control of gene expression in leaves compared to, for example, traditional in situ hybridisation on leaf sections. A main challenge in rice (and plant) biology is to understand the function of the thousands of genes expressed in the plant at any one time. Determining when and where a gene is expressed provides important information on the potential function of that gene. Developing an atlas of rice gene expression using WISH is something that could be explored in the future. The application of WISH on the relatively complex and challenging leaf tissue paves the way to adapt this method for use on other organs in rice, such floral parts and seed development. However, it is clear from the analysis reported here that it is possible that different tissues might show different problems in terms of background, so further optimisation of the procedure will be possible. It is also clear that the success of the approach is probe dependent. Although the results showing here are encouraging, the probes selected for analysis represent genes which are relatively highly expressed (Van Campen, unpublished data). The method needs to be further tested using probes for genes which are expressed at lower levels to judge the sensitivity of the approach. The protocol reported here was developed with the aim of investigating the expression of genes involved in stomatal differentiation. Unfortunately time limitations mean that these experiments have not yet been performed. Since stomata only form in a few cells in one layer of tissue (the epidermis) during a limited time of leaf development, it will be interesting to see if the sensitivity of WISH is sufficient to visualise these expression patterns.
Despite this caution, the WISH method described here provides encouraging results that it can be used to provide important data on gene expression patterns during early leaf development in rice.
The control of stomatal properties in rice (Oryza sativa L.) and their influence on photosynthetic performance
June 2017
DOI:
Thesis for: Doctor of Philosophy
Advisor: Prof. Andrew J. Fleming
Attribution 4.0 International — CC BY 4.0 - Creative Commons
Comments