ABSTRACT
We report an efficient procedure for
in situ
hybridization with a multi-well format on
Caenorhabditis elegans
embryos for large scale screening of gene expression patterns in this organism.
Each hybridization well contains embryos at various stages throughout
embryogenesis. The validity of the method was confirmed through results with
control genes whose expression patterns have been reported;
glp-1
in very early embryos,
myo-2
in pharyngeal muscle and
unc-54
in body wall muscle. Several collagen genes and a pepsinogen gene were also
examined to establish a set of lineage-specific markers. As a pilot project, we examined
~
100 unique cDNA species classified by our cDNA project, finding that
~10% of the cDNA groups were expressed in specific cells and at specific stages.
The nematode
Caenorhabditis elegans
is one of the best organisms for studying the molecular mechanisms of
development, since an enormous amount of information has been accumulated with
respect to anatomy, development, genetics and the genome. The entire cell
lineage has been traced from zygote to adult, which consists of 959 somatic
cells (
1
). A fertilized egg asymmetrically divides to produce the somatic founder cell
AB and the germline founder cell P1. The P1 blastomere divides three times in
stem cell fashion, producing three somatic founder cells EMS, C and D and the
germline precursor cell P4. EMS further divides to produce two blastomeres E
and MS. These blastomeres generate a variety of tissues; hypodermis (derived
from the AB and C blastomeres), body nervous system (from AB), body wall muscle
(from AB, MS, C and D), pharynx (from AB and MS), intestine (from E), somatic
gonad (from MS) and germline (from P4). The fates of the blastomeres are
determined in the early phase of embryogenesis. The mechanisms of fate
determination by maternal genes have been under extensive investigation.
However, the mechanisms of the subsequent execution of fate are largely
unknown, since a large number of downstream genes have not yet been identified.
Thus, identifying genes that are expressed in specific cell lineages will
provide important clues to these mechanisms.
The most straightforward way to this end is to look at the expression patterns
of genes in this organism one by one. Several methodologies for
in situ
hybridization and promoter trapping have been reported (
2
-
6
). However, a much more efficient strategy for systematic analysis of patterns
of gene expression is needed, since a large number of genes identified in the
C.elegans
genome projects are awaiting analysis. The consortium of the Sanger Centre and
Washington University has sequenced >25% of the genome, from which >4000 genes
have been predicted (
7
,
8
). In our laboratory, we are carrying out a cDNA project from which ~4500 cDNA species, corresponding to 35% of the total number of genes, have
been identified (Y. Kohara
et al
., manuscript in preparation). Current progress suggests that all genes of this
organism, estimated at ~13 000, will be identified within a couple of years.
In this paper, we present an efficient procedure for
in situ
hybridization suitable for this end and the result of a pilot project for large
scale screening of gene expression patterns.
Plasmid pJC124 of
col-3
DNA (
9
) was supplied by J. Kramer, cDNA clone cm01b7 (
10
) was supplied by R. H. Waterston and other cDNA clones were from our stock (Y.
Kohara
et al
., manuscript in preparation; their sequences can be seen in DDBJ/GenBank or
ACEDB).
cDNAs in [lambda]ZAPII vectors were PCR amplified using vector primers BS619
(TGAATTGTAATACGACTCAC) and BS711 (TGCAGGAATTCGGCACGA).
clb-2
DNA was PCR amplified from genomic DNA using primers clb-2 03 (ACAACCTGGACTTCGTGGAG) and clb-2 02 (GCCAGAATCCGTGATTGGTG). The amplified cDNA were purified by
Sephacryl S-400 spun column chromatography. Digoxigenin (DIG)-labeled antisense DNA was made by linear PCR as described (
11
), in reaction mixtures (10 [mu]l) using DIG-dUTP, amplified cDNA (~50-200 ng) and anchored oligo(dT) primers. In the case of
clb-2
, a gene-specific primer was used. Unincorporated substrates were removed by
Sephadex G50-spun column chromatography. The eluates were subjected to partial
digestion by DNase I in the reaction mixture (25 [mu]l), containing 50 mM Tris-HCl, pH 7.5, 10 mM MgCl
2
, 400 [mu]g/ml phenol-extracted salmon testis DNA and 1 [mu]l 14 ng/ml DNase I (freshly diluted in 0.1 M NaCl) at 37oC for 30 min. The reaction mixtures were heated at 75oC for 5 min and then stored at -20oC. We used 5 and 2.5 [mu]l (and sometimes 1 and 0.5 [mu]l) of the mixtures as probes for
duplicate hybridizations.
Standard techniques for cultivation and handling of worms have been described (
12
). Worms of the wild-type N2 strain were harvested from a mixed stage population and digested
with alkaline hypochlorite. The resulting embryos were allowed to hatch to L1
larvae by incubating overnight in S-basal buffer (0.1 M NaCl, 50 mM KPO
4
, pH 6, 5 mg/l cholesterol). The L1 population was fed to young adults in liquid
culture. The worms were collected, digested with alkaline hypochlorite for ~10 min and then forced out through a 23 gauge needle onto nylon mesh (50 [mu]m). The embryos in the filtrate were washed four times with M9 buffer
(0.3% KH
2
PO
4
, 0.6% Na
2
HPO
4
, 0.5% NaCl, 1 mM MgSO
4
) and finally resuspended in 100 [mu]l M9 buffer in a siliconized microcentrifuge tube.
The suspension of embryos was added to an equal volume (100 [mu]l) of 15 mg/ml yatalase solution (an enzyme complex containing chitinase,
chitobiase and [beta]-1,3-glucanase activities; TAKARA Shuzo Co., Japan) in 0.3 M
mannitol, 50 mM HEPES, pH 7.2, 10 mM NaCl, 10 mM MgCl
2
and 2 mM DTT and was immediately vortexed for 70 s at room temperature.
Chitinase (Sigma C-6137) at 1 mg protein/ml also works, but we found that yatalase had better
reproducibility than chitinase. The embryos were washed three times with embryo
handling buffer (0.3 M mannitol, 50 mM HEPES, pH 7.2, 10 mM NaCl, 10 mM MgCl
2
, 0.04% EGTA, 2 mM NH
4
NO
3
, 0.1% gelatin and 2 mM DTT), once with basal EH buffer (embryo handling buffer
without EGTA, NH
4
NO
3
, gelatin and DTT) and then resuspended in basal EH buffer at a ratio of 100 [mu]l packed embryos/ml buffer at 4oC. It is recommended that the extent of devitellinization be monitored
at this point by observing the elongation of embryos of the 2-fold stage due to breakage of the vitelline membrane. To achieve
devitellinization in 95% of embryos after subsequent methanol treatment, it is
necessary for 20-30% embyros of the 2-fold stage to show elongation at this point.
Drops of basal EH buffer were placed in the wells of 8-well microscope slides (Flow Laboratories) that had been coated with poly-L-lysine. Embryo suspension (5 [mu]l) was delivered to each well and the embryos were left to
settle to the bottom for 8-10 min at 4oC.
Excess buffer was removed and the slides were immediately immersed in methanol
at -20oC for 5 min. The embryos were rehydrated by immersing the slides in
a series of mixtures at 4oC; in methanol for 5 min, in 70% methanol + 30% fixative [3.7% formaldehyde
in 0.08 M HEPES, pH 6.9, 1.6 mM MgSO
4
, 0.8 mM EGTA and 1* PBS (137 mM NaCl, 2.7 mM KCl, 4.3 mM Na
2
HPO
4
, 1.5 mM KH
2
PO
4
)] for 2 min, in 50% methanol + 50% fixative for 2 min, in 30% methanol + 70%
fixative for 2 min and finally in the fixative for 20 min. The slides can be
stored at -20oC in ethanol for at least 2 months after the following dehydration
treatment at room temperature; in 30% ethanol + 70% PBS for 5 min, in 50%
ethanol + 50% PBS for 5 min, in 70% ethanol + 30% PBS for 5 min and finally
twice in ethanol for 5 min each.
The slides were rehydrated at room temperature in the following series of
solutions; in 70% ethanol + 30% PBS containing 0.03% H
2
O
2
for 2 min, in 50% ethanol + 50% PBS for 5 min and finally in 30% ethanol + 70%
PBS for 5 min. The slides were immersed in PBT (PBS with 0.1% Tween-20) for 5 min. To cut the glycosidic bonds of the proteoglycans that
appear in late embryos, the slides were immersed in 0.2 N HCl for 20 min at
room temperature. After washing twice in PBT for 5 min each, the slides were
incubated in proteinase K solution (10 [mu]g/ml in PBT) at room temperature for 11 min. The digestion was stopped by
immersing the slides in 2 mg/ml glycine in PBT for 2 min. After washing twice
in PBT for 2 min each, the specimens were refixed by immersing the slides in
fixative at room temperature for 20 min. After washing twice for 5 min each in
PBT, the slides were immersed in 2 mg glycine/ml PBT at room temperature for 5
min and then washed once in PBT for 5 min.
The slides were immersed in the following series of buffers; in 50% basal
hybridization solution (hybridization solution without salmon testis DNA and
yeast tRNA) + 50% PBT for 10 min and then in basal hybridization solution for
10 min. Pre-hybridization was performed as follows. The slides were wiped off and
waterproof lines surrounding the sample wells were drawn with a PAP pen (Cosmo
Bio Co., Japan). The sample well region surrounded by the waterproof line was
covered with 150 [mu]l heat-denatured hybridization solution (50% deionized formamide, 5* SSC, pH 7.0, 100 [mu]g/ml heparin, 0.1% Tween-20, 100 [mu]g/ml sonicated salmon testis DNA and 100 [mu]g/ml yeast tRNA) and incubated at 48oC for 1 h in a moist chamber.
After pre-hybridization, the slides were placed in the proper positions on a sheet
of silicone rubber that was placed on the lower block of a 96-well dot blotting apparatus (SRC96D; Schleicher & Shuell). The upper block, which has 96 holes individually equipped with
O rings, was placed on top of the slides and quickly assembled in such a way
that the holes and the wells matched perfectly. To each sample well was added
50 [mu]l hybridization solution containing 5 or 2.5 [mu]l heat-denatured probe DNA, followed by layering with 100 [mu]l mineral oil to prevent evaporation. The top of the block was
sealed with sealing tape and hybridization was performed at 48oC overnight in a moist chamber. A similar method of multi-well hybridization was reported for FISH mapping on human chromosomes
(
13
).
After hybridization, 0.4 ml 50% basal hybridization solution + 50% PBT was
delivered into each hole of the apparatus to dilute the probes to minimize
cross contamination in subsequent handling. The mixtures in the holes were
discarded by inverting the apparatus. The apparatus was quickly disassembled
and the slides were washed twice in 50% basal hybridization solution + 50% PBT
for 10 min each at 48oC, four times in 0.8* PBS, 0.1% CHAPS (3-[(3- cholamidopropyl)-dimethylammonio]-1-propane-sulfonate; Sigma C3023) for 20
min each at 48oC and then in PBT twice at room temperature.
The slides were incubated twice in PBtr (PBS, 0.1% Triton-X 100, 0.1% BSA and 0.01% NaN
3
) at room temperature for 10 min each and then subjected to an alkaline
phosphatase-mediated color reaction as previously described (
3
). The color reactions were stopped by washing twice in PBS + 20 mM EDTA. The
embryo specimens were mounted in Mount-Quick Aqueous (Cosmo Bio) and observed on a Zeiss Axioplan microscope with
Nomarski optics. In cases in which staining with DAPI was also done, the
specimens were mounted in 90% glycerol, 1* TBS, 1%
n
-propylgallate.
During the first 30 min after fertilization, a
C.elegans
embryo forms a very tough eggshell that makes the embryo impermeable to
fixatives and hybridization probes. The eggshell can be broken physically or
enzymatically, but once the eggshell is broken, the embryo becomes very fragile
due to a rapid change in osmotic pressure. This is particularly serious in
early embryos, since the cells are large. Therefore, the main point of the
method is removing the eggshell while maintaining good morphology in the
embryo. After testing various procedures, we have established a protocol in
which the eggshell is removed enzymatially by treating the embryos with
chitinase in isotonic buffer. The processed embryos are stuck to poly-L-lysine-coated multi-well slides and then subjected to fixation procedures.
Currently, we use 8-well microscope slides whose intervals between the wells match the
standard 96-well format perfectly. The slides are assembled with a 96-well dot blot apparatus and are subjected to hybridization as
depicted in Figure
1
. The O rings, with which the holes of the dot blot apparatus are individually
equipped, prevent leakage of the hybridization solution, enabling us to perform
multiple hybridization reactions on a single slide. One apparatus accommodates
four slides, meaning that 32 different probes can be analyzed on the apparatus.
Furthermore, one great advantage with
C.elegans
is that each well of the slides contains a population of embryos at various
stages throughout embryogenesis.
Another point of the method is the nature of the hybridization probes. We use
digoxigenin (DIG)-labeled single-stranded DNA probes, which are made by linear PCR on cDNA inserts,
using anchored oligo(dT) primers to minimize the effect of a long poly(A)
stretch. We found that the method was sensitive to the size of the DNA probes;
the presence of probes >500 bases frequently produced a high background,
particularly on early embryos that are rich in yolk protein. Thus, we
established an efficient protocol which facilitates shortening of a large
number of probes to ~100-300 bases through partial digestion with DNase I. The concentrations
of probes in the hybridization reactions are also important parameters for the
signal-to-noise ratio of the results. We determined the optimal concentrations
for several test probes, but the concentrations must be optimized for
individual probes. This task would be quite cumbersome if we apply the method
to a large number of cDNA clones. To bypass the problem, we adopted a
duplication strategy in which hybridization was performed in two (or sometimes
four) wells with serially diluted concentrations of the probes, expecting that
one of the concentrations would give the best results.
To test the accuracy of the method, we applied it to the control genes,
glp-1
,
myo-2
and
unc-54
, whose expression patterns have been reported. The maternal gene
glp-1
plays an important role in fate determination of the anterior blastomere AB and
its mRNA is detected from oocyte to very early embryo (
2
). The
myo-2
gene encodes pharyngeal muscle-specific myosin and is expressed in pharyngeal cells (
14
). The
unc-54
gene encodes body wall muscle-specific myosin and is expressed in body muscle cells (
14
).
Figure
2
shows typical images of the results of
in situ
hybridizations with these probes. Background signals are sufficiently low. With
the
glp-1
probe, signals are seen in very early embryos (two to four cell embryos) (Fig.
2
A). The results coincide with those previously reported (
2
,
3
). The
myo-2
probe stained only the pharynx (Fig.
2
B) and the
unc-54
probe stained only body muscle cells (Fig.
2
C and D), which agrees with the results from immunostainings (
14
). The combined results show the validity of the method. Since the specimen
contains embryos at various stages throughout embryogenesis, we can learn about
the stage in which transcription of a zygotic gene starts. Transcription of
myo-2
seems to start in 2-fold embryos, but not in 1.5-fold embryos (Figs
2
B and
3
C). Transcription of
unc-54
starts in the posterior region in late gastrulation (Figs
2
C and
3
D).
Although the identification of cells is easy in embryos earlier than mid-gastrulation, it becomes harder in embryos later than gastrulation. More
tissue-specific markers, such as the
myo-2
and
unc-54
probes, are desirable, which will make it easy to interpret the results
produced by a large scale screening of gene expression patterns. Thus, we are
collecting such probes through
in situ
analysis of genes whose expression is expected to be tissue specific. These
markers will also be useful as differentiation markers.
Cuticle collagen genes were chosen as markers for hypodermis.
Caenorhabditis elegans
has ~100 cuticle collagen-related genes (
9
), of which some are unique in the genome. The
col-3
gene is one of the unique cuticle collagens (
9
). Figure
3
A shows that
col-3
is expressed in the main body syncytium of the hypodermis from late-stage embryos. Another cuticle collagen gene (cDNA CELK01595) showed a
different pattern of expression, starting from the posterio-dorsal region of embryos at late gastrulation and finally detected all
over the main body syncytium except for seam cells in late embryos (Fig.
3
B).
The
clb-2
gene encodes one of the [alpha](IV) collagens associated with the basement membrane that separates the
hypodermis from muscle (
15
). It has been speculated that either gut cells or muscle cells secrete the
product of
clb-2
(
16
). Figure
3
D and E clearly indicates that
clb-2
is expressed in body wall muscle cells, because the expression pattern is
essentially the same as that of
unc-54
.
A homolog to pepsinogen, cDNA clone cm01b7 (
10
), showed specific expression in gut cells from the 1.5-fold stage (Fig.
3
F) as expected, because pepsinogen is one of the most abundant products in the
stomach in vertebrates.
We applied the
in situ
method to ~100 cDNA species that had been classified in our cDNA project (Y. Kohara
et al
., manuscript in preparation). As shown in Figure
4
(1-12), 12 of the clones show specific patterns of expression. The patterns
are roughly categorized into several groups: (lanes 4-6, 8, 11 and 12) maternal expression; (lanes 2 and 9) both maternal and
zygotic expression; (lane 7) zygotic expression from early stage; (lanes 1, 3
and 10) zygotic expression from mid-stage. Other cDNAs showed ubiquitous distribution throughout embryogenesis
(data not shown) or no expression (like lane 13). This set of cDNAs is derived
from a cDNA library of a mixed stage population including larvae and adults as
well as embryos. Therefore, the cDNA groups that gave no signals may be
expressed post-embryonically.
The main points of our method are the removal of the eggshell by an enzymatic
procedure before fixation and the usage of a multi-well apparatus for hybridization. These make the method so efficient that
currently one person can perform 192 hybridizations (96 different probes) at
one time using six dot blot apparatuses. It takes a week to do the task,
including preparation of the probes. The quality of the results is high enough
to identify positive cells. Background signals are a common problem with
in situ
methods, but our duplication strategy, in which hybridization for a probe is
duplicated with different probe concentrations, makes the results very reliable
with minimum reduction of efficiency. It is also a great advantage to have each
hybridization provide information on mRNA distribution at all stages of
embryogenesis. Thus, with our plan it is quite feasible that within a year we
will survey the expression patterns of the set of 4500 cDNA species that we
have classified.
Accumulation of data on expression patterns will enable us to more finely
classify patterns with respect to cell lineage and developmental stage, which
will lead to identification of sets of genes that show the same expression
patterns. For example, the cDNA group CELK00231 showed a very similar pattern
to that of a pepsinogen homolog [Figures
3
F and
4
(3)]. The
clb-2
gene showed essentially the same pattern of expression as that of
unc-54
(Fig.
3
D and E), indicating that body muscle cells secrete [alpha](IV) collagen, the main component of basement membrane, which is encoded
by the gene.
The
in situ
screening currently being performed in this laboratory is finding many more
such examples (our unpublished results). Some of these genes might be under the
same regulation mechanisms. In
C.elegans
, the genomic sequences surrounding these genes are, in many cases, available
from the genomic sequencing project. These sequences will become immediate
targets for analysis of the regulatory regions by experimental means and/or
through informatics methods. The factors that regulate these genes will be
expressed in the same cell lineage but earlier than the expression of target
genes. Our screening is also revealing examples of sets of genes which are
expressed in a specific cell lineage but at different stages (our unpublished
results). If one of these genes has a similarity to a transcription factor, it
will be of interest to disrupt the gene to examine the transcription of other
genes that are normally expressed later in the cell lineage. For this end, a
system for transposon-mediated gene disruption is available in
C.elegans
(
17
).
For closer examination of the time and place of the start of expression,
multiple labeling detection using a confocal microscope should be considered.
We use the color reaction to detect hybridization signals for permanent storage
of the specimens, but the
in situ
method can be applied to a fluorescent detection system simply by changing the
mounting reagent.
In situ
analysis gives information only on the distribution of mRNA and other
mechanisms, including translation regulation, protein localization and protein
modification, also play important roles in development. However, since this
work revealed that >10% of the probes showed specific patterns of mRNA
distribution,
in situ
screening will provide a large amount of invaluable material for studying the
molecular mechanisms of development. Currently we are focusing on
embryogenesis, but we are planning to extend the screening to post-embryonic stages with several modifications, ultimately aiming at
understanding the entire life of the worm.
We thank Jim Kramer and Bob Waterston for cDNA clones and Geraldine Seydoux for
her
in situ
protocol. This work was supported by a Grant-in-Aid for Creative Basic Research on `Human Genome Analysis' from the
Ministry of Education, Science and Culture of Japan and by the RIKEN Project on
Human Genome Analysis.
REFERENCES
Return