1Department of Molecular Genetics and Cell Biology, The University of Chicago 1103 E. 57th Street, Chicago IL 60637 USA
2Department of Organismic and Evolutionary Biology, Harvard University, 22 Divinity Ave., Cambridge MA 02138
*Correspondence should be addressed to:
DuPont Biosolutions Enterprise,
Research Center, N.W. 62n Ave., Johnston, Iowa 50131-1004
Keywords: Gene Expression, mRNA, in situ Hybridization, Arabidopsis, Asteraceae, probe labeling, multiple probe arraying, PCR
The article describes a non-radioactive in situ hybridization procedure suitable for plant tissues. The flexibility, sensitivity, and signal specificity for detecting abundant and rare mRNAs are shown in studies with various plant tissue types including cotyledons of germinated seeds, foliage leaves of Flaveria trinervia (Asteraceae), and flowers of Arabidopsis (Brassicaceae). Introduction of touchdown hybridization temperature control, multiple section and multiple probe arraying, and PCR based DNA template generation in the procedure significantly simplified optimization process.
In situ hybridization (ISH) is a powerful tool in locating expression of a new gene and in detecting temporal and spatial gene expression patterns. Since the first development of the method (1, 2) and its application to animal development using radioactively labeled probe (3, 4), several non-radioactive methods have been developed (5, 6, 7, 8). Compared with radioactive methods, nonradioactive methods show many advantages, such as shorter signal development time, higher histological resolution and higher detection sensitivity, lower cost of probe labeling and longer shelf life of labeled probes, and free of radioactive hazards to environment and human health. Recently nonradiactive, high-throughput in situ hybridization has been proposed to monitor expression of a large number of genes in a multiple tissue array on a single glass slide in animals (9, 10).
Of all the non-radioactive labeling and detection methods developed, Digoxigenin systems are currently the most successful for animal and plant tissues. Partly because Digoxigenin, which was extracted from plants of the genus Digitalis, is not known to be synthesized in animal and other plants, many background problems that plague other nonradioactive methods are avoided. In high plants, Digoxigenin system is widely used for in situ hybridization (7, 8). However, common problems that affect other non-radioactive in situ hybridization systems, such as, high background noise, lack of consistency in signal specificity in different tissue types, and across different species remain. Since many factors such as, probe quality, plant tissue fixation, prehybridization tissue treatment, hybridization condition and hybridization buffer compositions, and stringency of post-hybridization washing, all affect the outcome, optimization of these parameters is often a time-consuming, and frustrating process.
With forthcoming of the post-genomic era, development of a in situ hybridization procedure that require minimal optimization is highly in demand for automation and for high throughput multiple tissue arraying. Here we report an in situ hybridization procedure suitable for different plant tissue types. The protocol was developed by modifying of various radioactive and non-radioactive procedures (3, 11, 12, 13) and the introduction of several new features, such as, touchdown hybridization temperature control, multiple section array, and PCR-based DNA template generation. We have applied this procedure to detecting gene expression patterns in cotyledons of germinating seeds, foliage leaves, and flowers of plant species Arabidopsis (Brassicaceae) and Flaveria trinervia (Asteraceae) (14, 15). Our experiments show that the procedure is capable of detecting both highly abundant mRNA transcripts, such as rbcS, and low abundance mRNAs such as Agamous, with only minimal modification and optimization. This makes the method suitable for high throughput multiple-tissue type and multiple-gene in situ hybridization.
MATERIALS AND METHODS
Preparation of plant tissue sections
For Flaveria trinervia (Asteraceace), expanded young foliage leaves were collected fresh and cleaned with dH20 to remove dirt and dust, and cut into 5 mm to 20 mm long fragments; cotyledons were collected from light-grown seedlings of 7 day old after germination. For Arabidopsis thaliana, flowering shoot segments that contain flowers of various developmental stages were collected. To avoid RNA degradation, all plant tissue samples were collected fresh and were soaked into 4% paraformaldehyde fixation solution immediately after removal from plants. At room temperature, bench top vacuum was applied to facilitate infiltration of fixation solution into plant tissues. Vacuum was increased slowly until air bubbles appeared on tissue surfaces, vacuum pressure was held constant for 30 minutes, and the released slowly. This process was repeated if tissues still floated on the surface of fixation solution. Tissues were transferred to fresh 4% paraformaldehyde and were fixed for 24 hours at 4 °C. Fixation solution was replaced with 30%, 50%, and 70% EtOH each for 15 to 30 minutes and the tissues were stored at 4 °C in 70% EtOH in a tightly covered vial or bottle.
Preparation of 4% formaldehyde solution:
Take 80 ml of PBS, adjust to pH 11 with NaOH, heat to 60 °C, add 4 gm of paraformaldehyde powder (Sigma), stir until dissolved (2-3 minutes). Cool the solution on ice to room temperature, and readjust to pH 7.0 with H2SO4. Keep the solution bottle at 4 °C refrigerator or in a ice box before use. PBS buffer preparation: for 1 liter of 10 x PBS buffer stock, use NaCl 76 gm, Na2HPO4.2H2O 12.46 gm, and NaH2PO4.2H2O, 4. 68 gm, make up to 1 liter with ddH2O.
Tissue infiltration and embedding:
Transfer the fixed tissues from 70% EtOH storage solution into 70%, 85%, 95% ethanol for 30- 60 minutes each, and into 100% ethanol for 2 x 30 minutes. Then transfer the tissues to a premix of 25% TBA+75% EtOH for 60 minutes at room temperature (TBA: Tert-butyl alcohol, Sigma). All the steps below are done in a 60 °C oven: transfer the tissues to 50% TBA+50% ethanol, and then 100% TBA for 60 minutes each. Transfer the tissues to a mixture of 50% TBA + 50% Paraplast Plus (Oxford Lab) in a sealed beaker or bottle, infiltrate for 6 hours or overnight (melting paraffin at 60 °C before mixing with TBA). Transfer tissues into 100% paraffin and change paraffin every 12 hours, for at least three changes before embedding.
Uncoated and precleaned microscopic glass slides (Fisher Scientific ) are coated with Poly-D-lysine (Sigma) in coating solution (50-100 µg/ml poly-D lysine (Sigma P1149) in 10 mM Tris-HCl, pH 8.1) for 20 minutes and dry in a 38 °C oven. Trim paraffin block and cut sections of 5 to 8µm thick with a newly-sharpened microtome knife.
Lay ribbons of paraffin sections on the water surface in a 42°C water bath for 30 minutes and then break the ribbon into section triplet and load to a slide by slipping the slide under the water surface. Alternatively, place a slide on a prewarmed slide warmer at 42°C, and apply several drops of DEPC treated water to the slide. Float a ribbon of tissue sections ("shiny" side down) on top of the water drop and allow to sit for a few minutes (we use paintbrushes to handle the ribbons). During this time the ribbon should flatten out. Drain off water using a Kimwipe (excess water will cause bubbles). After loading to a slide, keep the slide on a slide warmer in the vertical position for 30 minutes to dry (otherwise air bubbles might form); then incubate the slide in flat position overnight on the slide warmer at 42°C so that tissue adheres to the slide. Glass slides with paraffin sections can be stored within a desiccant jar for several years at 4 °C.
We usually lay sections on a slide in a 3 x 5 or 3 x 8 array, i.e. with 3 rows and 5 to 8 columns on a single slide, depending on the size of a paraffin section. In this format, sections from different tissue types can be arranged on the same slide for more reliable comparative studies.
Preparation of DNA templates for in vitro transcription
DNA templates used for in vitro transcription to generate sense and antisense RNA probes are from a cDNA fragment inserted into a plasmid vector with RNA polymerase binding sites on both flanks of the insert. For rbcS probes, we use a pBlueScript II SK (+) clone with 400 bp of cDNA covering exon II and exon III of the RbcS-R1 gene from Flaveria ramosissima (14). For PEPCase gene (Ppc) probes, we use a pBluescript I KS (+) plasmid containing a 1.82 kb cDNA fragment from the 3' region of the gene (3 kb) that encodes a C4 isoform of PEPCase from Flaveria trinervia (16). The template for Agamous was a 1.7 Kb Arabidopsis Agamous cDNA fragment with the MADS box being deleted, cloned in pGEM -T vector (17).
Inserts from the above cDNA clones were amplified by PCR using a universal primer pair, M13-20 and M13-reverse,situated outside of RNA polymerase binding sites. The PCR approach allows one to rapidly generate a large amount of DNA template that has a insert and RNA polymerase binding sites. This is easier than the conventional and tedious procedure that includes bacteria culture, plasmid DNA preparation, restriction digestion, fragment separation by gel electrophoresis, elution, and purification. For 50 µl of PCR reaction: ddH20 (DEPC), 33.5 µl; 10 X PCR buffer, 5.0 µl; dNTP mix (2.5 mM each), 4.0 µl; diluted cDNA clone (20 ng/µl), 5.0 µl; M13-20 primer (200 ng/µl), 1.0 µl; M13-reverse primer (200 ng/µl), 1.0 µl; Taq DNA polymerase (Boehringer Mannheim, 5 units/µl), 0.5 µl; MgCl2 (50 mM), 2.5 µl. The annealing condition was 56 °C, 1 minute; Extension was 72 °C, 1-2 minutes. 25-35 cycles. All above components were mixed on ice and loaded into a thermal cycler (MJ PT-100). The 50 µl PCR reaction was purified with QIAquick-spin PCR product purification columns (QIAGEN Inc.) and the DNA was eluted into 50 µl of ddH2O (DEPC).
Preparation of sense and antisense riboprobes
In vitro transcription:
From prepared DNA template, digoxigenin labeled sense and antisense RNA probes were generated by in vitro transcription using T7 and T3 RNA polymerase for rbcS and PEPCase genes (Boehringer Mannheim) and T7 and SP6 RNA polymerase for Agamous gene (Promega). 20 µl in vitro transcription reaction was set up as follows: DNA template, 1 µg; 10x in vitro transcription buffer, 2 µl; NTP Labeling Mix, 2 µl (Boehringer Mannheim, 10 mM ATP, 10 mM CTP, 10 mM GTP, 6.5 mM UTP, and 3.5 mM DIG-UTP); RNase inhibitor (Boehringer Mannheim), 1unit/µl; RNA polymerase, 2 µl; ddH2O to 20µl. Incubate the reaction at 37(C for 2-3 hours. 2 µl of DNase I, RNase-free (Boehringer Mannheim) was added and incubated for 20 minutes at 37(C. The reaction can be stored at -20(C before hydrolysis and purification.
To check yield and quality of in vitro transcription reaction, 2 µl from the 20 µl total was taken and run in 0.5 x TBE buffer and 1% agarose gel for 20 minutes at 150V (fresh TBE buffer was used and the gel box and comb was washed with 0.2N NaOH for 20 minutes and rinsed thoroughly with dH20 before use).
to improve the pemeability of RNA probe into tissues during in situ hybridization, labeled RNA probes longer than 150 bp were hydrolyzed into pieces between 75bp and 150bp by adding one equal volume of hydrolysis buffer (0.12 M Na2CO3, 0.08 M NaHCO3, pH10.2) to the purified probes, mixed, then incubated at 60(C for t minutes. Here, t =Lo-Lf/k•Lo•Lf, Lo=initial length of probe, Lf=final length of probe (150bp), k=0.11 kb-1 min-1 (Cox et al., 1984). The hydrolized probe was purified using the following procedure: add 1 µl glycogen solution (Boehringer Mannheim), mix, add 1/10 volume of 10% acetic acid (v/v), mix, then add 1/10 volume of 3M sodium acetate (pH 4.5) and 3 volumes of 100% ethanol and mix; precipitate at -70(C for 20 minutes, spin at 13, 000 x g for 15 minutes at RT, wash pellet briefly with 70% EtOH, air dry; resuspend pellet in 20-100 µl ddH2O (DEPC), aliquot and store the probe in -80(C freezer. The probe was quantified using dot blotting and colormetic detection method (see 7).
Deparaffining sections: slides with paraffin sections were soaked in xylene in a glass jar for 3 x 10 minutes and air-dried in a fume hood. Section rehydration: slides were transferred through the following ethanol series: 100%, 95%, 80%, 70%, 50%, 30%, ddH2O, ddH2O, 2 minutes for each step. HCl treatment: slides were incubated in 0.2 M HCl for 20 minutes at RT, then, rinsed in ddH2O, 2 X 2 minutes, and excess H2O was blotted. Protease K treatment: slides were incubated in protease K solution (100 mM Tris.HCl pH 8.0, 50 mM EDTA, 2 mg ml-1 protease K) at 37°C for 30 minutes, rinsed briefly in PBS at RT. Note: the protease K concentration and the incubation time need to be optimized empirically to get the best signal without much loss of tissue and cell morphology. Refixation: slides were soaked in freshly made 4% formaldehyde solution in a fume hood for 20 minutes at RT, and rinsed in PBS for 2 X 5 minutes at RT. Acetylation: slides were placed into 300 ml of 0.1 M triethanolamine pH 8.0 (Preparation: triethanolamine (sigma) 3.9 ml, conc. HCl 1.2 ml, ddH2O 294.9 ml) in a deep slide tank. 300 (l acetic anhyride was added with gentle stirring for 10 minutes at RT, then rinsed in PBS for 2 X 5 minutes, and in 2 X SSC for 5 minutes at RT. The slides can either go directly to prehybridization without going through an ethanol dehydration series, or they can be stored for up to one week in a wet slide jar at 4°C.
Prehybridization and Hybridization
Slides were incubated in 2 X SSC for 10 minutes at RT, laid flat wet on a slide supporter inside a humid chamber, such as a sandwich box that contains paper towel soaked with 5 X SSC. 1-2 (l prehybridization medium was applied onto each single section, and the slides were incubated at 50°C for 1-2 hours in an hybridization oven . No coverslip is needed during prehybridization and hybridization. In this way, each section, instead of each slide, is an experiment unit, allowing many different treatments, different probes, and probe concentrations etc., to be compared on a single slide. Preparation of 1000 µl prehybridization medium: 10 x in situ salts, 125 µl; deionized formamide, 500 µl; 50% dextran sulfate (Pharmathia), 250 µl; 20 mg/ml tRNA (Boehringer Mannheim), 25 µl; 5 mg/ml polyA RNA (Boehringer Mannheim), 60 µl; and ddH2O (DEPC), 40 µl. Preparation of 10 x in situ salt: 3.0 M NaCl, 0.1 M Tris HCl pH 6.8, 50 mM Na2HPO4, 50 mM NaH2PO4, 50 mM EDTA, in DEPC-treated ddH2O. In above solution, all components except Tris HCl, were DEPC-treated.
1:2 to 1:4 mix of probe and prehybridization medium was prepared and the mix was incubated at 80°C for 2 minutes, and placed in ice immediately; 1-2 µl probe mix was applied directly onto each section by mixing gently with the prehybridization buffer. Incubate slides at 50 °C overnight. The final probe concentration should be about 0.5-2.0 ng /µl. Several probes from different genes were usually applied to a single slide for more accurate comparison. To ensure each probe can hybridize at its own optimum annealing temperature, a touchdown or a stepdown hybridization temperature control scheme was used, that is, the hybridization starts at 54 °C and ends at 44 °C by step-wise decrease of 2 °C /2 hours.
Hybridization solution was removed from slides (shaking off to a paper towel) and the slides were soaked in prewarmed (42 °C) 2 X SSC for 5 minutes in a 42°C water bath or oven, then, in 1 X SSC for 10 minutes at 42°C. RNase A treatment: slides were incubated at 37°C for 20 minutes in RNase A buffer (5 mg/ml RNase A in NTE buffer (500 mM NaCl, 10 mM Tris HCl pH 8.0, 1 mM EDTA)). The slides were washed in the following series: 1 X SSC, 3 X 5 minutes at RT; 1 X SSC, 30 minutes at 37°C; 0.5 X SSC, 5 minutes at RT; 0.5 X SSC, 60-90 minutes at 42°C; 0.5 X SSC, 10 minutes at RT. Slides were drained well between every buffer change to avoid RNase carryover.
Slides were incubated in buffer I for 10 minutes at RT, then in buffer II for 30 minutes. 200-400 (l diluted antibody (1:500 dilution of anti-DIG antibody conjugate (Boehringer Mannheim) using buffer II ) was applied on each slide and the slide was incubated in a humid chamber, such as the one used for hybridization for 1-2 hours at RT. Buffer I preparation: 150 mM NaCl, 100 mM Tris-HCl pH 7.5, filter through a 0.2 µm or 0.45 µm filter, add 0.3% Triton X 100, mix well with vigorous stirring. Buffer II preparation: add 2 gm of blocking agent (Boehringer Mannheim) to 100 ml buffer I without Triton X, stir and heat to 60°C to dissolve (need about an hour); aliquot and store buffer II in -20°C freezer for future use.
Slides were washed in buffer I for 3 x 10 minutes at RT, then, in buffer III for 2 x 10 minutes at RT (buffer III preparation: 100 mM Tris-HCl, 100 mM NaCl, 50 mM MgCl2 , adjust pH to 9.5 before adding the MgCl2, filter through a 0.2 (m or 0.45 (m filter before use). 400-500 (l fresh color substrate solution was applied to each slide. Color buffer preparation: in 1 ml of buffer III, add 4.5 (l NBT (Boehringer Mannheim), 3.5 (l X-phosphate (Boehringer Mannheim), and mix. Let color develop in dark in a humid chamber and monitor color development occasionally after 30 minutes in color buffer. In our experimental condition, we detected signal after 30 minutes with highly abundant mRNAs and after 1-2 days for rare mRNAs. To detect extremely weak signal, replace old color buffer with new one after 2 days of color reaction. color reaction was stopped with 400 (l buffer for 10 minutes ( buffer IV:100 mM Tris HCl pH 8.0, 100 mM EDTA). Slides were mounted permanently with Merkoglas (EM Science) or any other suitable mounting medium. Sections were photographed in darkfield with either a Photolab microscope (Nikon) on Kodak Ectachrome 64T or with digital video camera. For image scanning, slides were mounted temporally in buffer IV.
RESULTS AND DISCUSSION
AGAMOUS mRNAs accumulate in whorl 3 and whorl 4 of Arabidopsis flowers
Figure 1 shows the AGUMOUS transcript accumulation pattern in Arabidopsis flowers of late stage 6 to stage 7. Agamous mRNAs accumulates predominantly in whorl 3 (stamens) and whorl 4 (carpels), and their accumulation is very low or not detected in whorl 1 (sepal) and whorl 2 (petal). Our results confirm the previously published expression pattern of AG genes in Arabidopsis ( 17, 18, 19).
Pep mRNA and rbcS mRNA accumulation is cell type-specific in cotyledons and foliage leaves of Flaveria trinervia (C4)
Pep and rbcS encode two key carboxylases in the C4 photosynthetic pathway: phosphoenolpyruvate carboxylase (PEPCase) and ribulose-1, 5, -bisphosphate carboxylase/oxygenase (Rubisco). The accumulation of the two enzymes and their corresponding mRNAs have been found to be cell type-specific in the C4 plants, maize and Amaranthus (11, 13). Using mRNA in situ hybridization we found that their mRNA accumulation is also cell type-specific in Flaveria trinervia, a dicot C4 plant (14). Pep mRNAs accumulated in mesophyll (M) cells, but not in bundle sheath (BS) cells of foliage leaves (Figure 2A, 2B) and cotyledons (14). RbcS mRNAs accumulated in BS cells but not in M cells of foliage leaves (Figure 2C). We also found that rbcS gene expression is BS cell-specific in cotyledons of 7 day and older seedlings grown in the light condition (Figure 2D).
Compared to the expression level of rbcS and Pep gene both encode metabolic enzymes, the expression level of AGAMOUS gene, a transcriptional factor, is very low in flower tissues. However, using our in situ hybridization procedure, by extending color reaction time from 3 hours to 3 days, we were be able to detect the expression pattern of AGAMOUS gene without increasing non-specific background noise.
Tissue fixation, paraffin infiltration, embedding, and sectioning:
We tested two different fixation solutions, FAA and 4% paraformaldehyde, and found the 4% paraformaldehyde always gave better tissue morphology and cell content preservation. The 4% paraformaldehyde solution should be prepared fresh in a bottle with tight cap to minimize the loss of formadehyde gas produced during the dissolving and heating process. We found that for cotyledons, leaves, flower, and shoot meristems increasing fixation time at 4(C from 24 hours to 76 hours significantly improved tissue morphology and cell content preservation. We also found that increasing infiltration time to 3 days with two changes into new paraffin every day in a 60(C - 62(C oven gave better paraffin infiltration,. To avoid air bubble in paraffin block, heat paraffin to 64(C before pouring into mold and orientate tissue in the mold with prewarmed forceps as quick as possible. We have compared the in situ hybridization quality among sections from 3 to 10 (m and find that 5 to 7 (m sections show the best overall quality for leaves and flowers, for meristem tissues which have small and dense cells, sections of 3 to 6 (m show better resolution.
Riboprobe labeling and purification:
Probe yield and quality are critical for success of in situ hybridization. The PCR approach we used can minimize the length of unwanted flanking plasmid DNA sequence and reduce both the time and effort in DNA template preparation. The high yield and high purity DNA template generated using PCR and Qiagen spin column purification significantly increases the riboprobe yield of in vitro transcription.
Glycogen purification is a simple but very efficient procedure for recovering labeled and hydrolyzed RNA probes. We also used siliconized microtube to further reduce yield loss during probe purification. We found probe hydrolysis not only helped to improve probe permeability into plant tissue, but also significantly reduced non-specific background.
The goal of section pretreatment is to expose more endogenous RNA to the tissue surface without significantly damaging the tissue morphology, which is achieved by optimizing the two key steps of pretreatment: the HCl treatment and proteinase K treatment. Longer incubation of slides in HCl and proteinase K can increase the amount of exposed endogenous mRNA, but at the same time can also increase the degree of damage to tissue morphology. The concentration and the length of treatment need determined empirically. We found the conditions we used are suitable for cotyledon, leaves, and floral tissues in the two species.
Hybridization and post-hybridization washing:
We found that including polyA RNA and tRNA in prehybridization and hybridization medium is important for reducing non-specific hybridization. The ratio of probe and hybridization medium, called dilution ratio, should be strictly controlled since the final salt concentration on slide is critical for proper probe/target mRNA annealing. In our experiment, if the probe was suspended in ddH2O or low salt buffer after purification, we use 1:2 dilution (1 part probe: 2 parts hybridization medium) which generate strong and specific in situ hybridization signals at 50 (C. When the yield of labeled probe is low, the probe should be directly resuspended into hybridization medium so that any change in dilution ratio will not affect the final salt concentration. When more than one probes from different genes are included in the same in situ hybridization reaction, each of them might require different annealing temperature, a touchdown hybridization temperature control should be used to ensure proper annealing for each probe. RNase A treatment and high stringency washing are two key steps in removing non-specific riboprobe hybridization during post-hybridization washing.
Immunological detection and color detection:
We have tested the effects of different block agents on reducing background noise and find that the Boehringer Mannheim blocking agent is better than other commercial blocking agents we have used for membrane hybridization. Thoroughly washing off non-specific anti-DIG antibody after immunological detection is also important for reducing background.
To minimize the disturbance of color reaction buffer on the slide during the color reaction, monitor color development by quick eye-spotting instead of by checking under microscope and only use microscope when color development close to finish. Promptly stop color reaction with stop buffer (Buffer IV) and always keep the slides in a tightly sealed box at 4 (C refrigerator to reduce color substrate oxidation which produces non-specific background noise. Minimize the exposure to light during color reaction because light damages color substrate. Dehydration through ethanol series, an essential step before permanently mounting slide with hydrophobic mounting medium, will cause loss of color signal. Therefore we prefer mounting slides temporally with buffer IV or ddH2O and take photographs. The temporally mounted slides can be kept in a humid chamber (the one used for hybridization) for several weeks for future photographing.
We would like to thank Dr. Peter Westhoff for providing the Ppc cDNA clone and Dr. Zhongchi Liu for the Arabidopsis Agamous cDNA clone. Thanks for Dr. Jane Langdale, Dr. David Jackson, and Dr. J. L Wang, Dr. Toby Kellogg, and Dr. William Alverson,for many helpful discussions. Weber Ameral, Lena Hileman, Lulu Leroux, and Aida Pascual have provided valuable technical assistance. The work was supported by National Science Foundation and Department of Energy grants to LJM, an Alfred P. Sloan Foundation grant to DB, and a Hutchins Plant Biology Predoctoral Fellowship to GS. GS recieved fellowship support through a National Institutes of Health Training Grant.