TECHNICAL
TIPS
389
389
391
396
397
hrrinda rd zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONM
An ine x pe nsive . hom e -m a de
zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
& H w be t i Scbumz
Doubkinsim
~On-embryosfOr- zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGF
ofoverfappfng~ oflpne= prrJaog
In S&Jhybridization of whole-mount embryos is an essential tool to study the expression of developmentally regulated genes
in mousel and other ver+r?brates.Furthermore, it is often desirable to compare the expression pattern of two d&rent genes
in the same embryo distinguishing the co-expression areas (if any) A number of papers have described the use of two in
situ probes of various speciesw, using either biotin or fluorescein p&s
in combination with digoxigenin probes. The major
difficulty in such experiments is to obtaii sensitive and clearIy distinguishable signals during the detection of the two probes.
labelled antibodies against digotigenin, fluorescein or biotin, does not usually provide efmugb senThe use of fluorescently
sitivity to detect antIsense RNA probes. On the other hand, the use of conventisnal alkaline @~ ~ Phatase or peroxidase sub
suates to amplify the signal presents problems when attempting to define accurately the domains of coexpression of two
different genes, where the two staining colours overlap. Such areas of overlap normally appear darker than the areas of
single staining; they are not visualized as a dlfferent &our, nor as two colours visible at the same time, nor as two &ours
visible under different conditions (e.g. under different llhunlnarlons). We describe here a scheme that allows visualization
of the expression and co-expression of two genes using whole-mount hybrldizatlon or a combination of whole-mount
hybrldlltion
and section staining that m a k e s t he areas of co-expression clearly distinguishabk. Our protocol is similar to
described procedure@
for hybridization of fluorescein (FIX)- and digoxigenin (DIG)_labeIled RWA probes, which are
detected using antibody-alkaline phosphatase conjugates, and combines a lluorochmme with the more usual nitro blue tetrazolium chloride/5-bromo-4-chloro-3-indolyl-phosphate,
toruidme salt (NBT/BCIP) staining6. MBT/BClP Gioehringer
Mannhelm) are chromogenic substrates and m
(Enzyme Labelled Fluorescence mRNA in situ hybridization kit, Molecular
Probes) is a fluorescent substrate visible under UV light [a$-diamlno2-phenylindole
(DAPI) filter]. FITC-, DIGlabelled and
unlabelled nucleotldes, as well as anti-FlTC and anti-DIG antlhodies conlugated to alkaline phosphatase, are from RoelUnger
Mannheim. Figure 1 shows a hindliib bud of a El15 d mouse embryo hybridized with a Hardll p&&s
(DIG-labelled
and RLF stained) and Hox&~ (Ref. 8; PI‘TC-labelled and NBT/UCIP stained) as described below, using transmitted microscope light, W light, and a combination of the two illuminations. Single illuminations show both staining colours alternately,
whereas the double illumination shows both staining colour reactions simultaneous!y. allowing proper visualization of the
areas where Har d11 ancl Hoxdl 3 are co-expressed.
zyxwvuts
zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
mc.
@J -OBER
1996 VOL. 12 No. IO
385
TECHNICAL
The mcthtxlo1o~
zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPON
Tms
with the modifica-
tion5 to the standard protocols a& as follow> (a step-by-step protocol is available
0n rcquc~t).
Pcuiflcaelon
of probes
WA pt.&es xe q’nthesizc..
ustiastanj &mlpmocols
and puritied bv pn+itation zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLK
’ with zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGFEDCBA
0 .1 dum t 4 S: lithium &l&e
and 3
for 30 min (X2).
We do not have any problem with 12-u-l-PDIG (dig~xogc~-lZ-uridine-5’-triphm
phate) nuclttiides, but we fmd it is
very importxtt to check in a 1%
a-r*-TBE
gel that unbound 12-UT-F’-
: volum~3 ethanol at -20°C
FIX
nudeor i des hem
zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIHGF
a.5 a gr een spot
under LV light> do not remain in the
probe solution. Unbuund IZ-UTP-FtTC
nuclczick?;
increase the backgmnd
and .so impair the ~&ant
pattern
signal
hybridization
UdffrstplVbL?
ZZ$IZ!Z
we follow the standard procedures on
whole-mount mouse embryo+5 hybridizing the two probes si mul t ~ne0usl y. T he
main modifications to standard protocols
during pre-hybridization and hybridization
step are thar we LLLC
a I 3 X SSC {pH 5 with
citric acid) hybridization buffe r9 instead of
the typically 5X SSC. This stringent
hybridization sirn~lities the post-hybridiza t ion washing steps. which are reduced to
two quick rinses and two washes with the
same hyinidi7ation buffer, and a wash with
1: I hybridization buffer:TBST <TrisHCI
saline buffer. 0.1% Tween 20, pH 7.6) or
MABTfnnlric acid buffer, 0.1% Tween 20.
pH 7.5). 30 min arch w a sh a t hybddiia don
t rm pe rx ure .
We gt? i bener results using
FITC-lahell~il prohe for the first developmrnt and incubating with a 1/4ooO-l/8000
dilution of anti-FITC anubody conjugated to
Fw;me 1. Double in si#tc hvbridization of mouse embrvo hindliib at 11.5 d.
hybridized simukaneously to‘ an antisense Hoxdl .3 p&
(FITC-labelled and
NBTIECLP stained) and an antin5e
Hard21
probe7~* (DIG-lab&d
and ELF
stained). (a) Whole-mount staining of HoxdU before sectioning.Hard13 has been
detected with anti-FITC antibody conjugated to alkaline phosphatase and stained
with NBT/BCIP. The Had11 hybridization has not yet been detected with the antiDIG antibody. (b) Har d11 st ai n@ on a 15 w sagittal cryostat section of (aI at the
same magnification seen under W light. Har d11 probe, hybridized on wholtmount embryo together with HoxdI3 pr obe. has been detected after sectioning with
an anti-DIG antibody conjugated to alkaline phosphatase using ELF as a substrate.
(c) Magnification of 6) seen under transmittedlight. Only Hoxdlj hybridization in
the mesenchyme can be seen. (d) Same field as (c) seen under W light. Only
zyxwvutsrqponmlkjihgfedcbaZYXWV
zyxwvutsrqponmlkjihgf
Hoxdl l
hybridization
in the mesenchyme
can
be seen.
Note that the
Har d23
hybridization area is smaller than and localized within the distal domain of the
Hard1 1 area of expression. le) Same field as (c> and Cd)seen under uansmitted and
W light. Note that Hard11 (in green) and Hoxdl.3 Cm blue-purple) can be seen at
the same time in the area of co-expression. and Hard11 can be also seen in the rest
of its expression domain (green-yellow). Reciprocal in SUU hybridization in the
opposite hindlims bud of the same embryo (not shown) shows the same pattern of
expression and co-expr&on
but with the staining colours exchanged. Bars:
zyxwvutsrqponmlkjihgfedcbaZ
(a. bP 0.4 mm; (c-e)=
alkaline phosphatase.
0.1 mm.
The fi
probe has to be developed
with NBVBCIP
substrate.
Emhyos older than 8.5 d should be .sectioned to allow detection of the fluorescent substrate. Appropriate sections can then
be c hose n fGr detection of the second probe in ams of expected co-expression. We typical$ do so after taking pictures of the
whole-mount staining. DC not re-foe the embwos; cryoprotect in 30% sucrose and embed in OCT (Tissue Tek OCT compound,
Miles). After sectioning. detection of the fti probe by NBT/BCIP staining can be re-started again if the tkt prc#e staining is not
visible enough by washing in staining reaction buffer (So rnin) and adding fresh iVBT/E+CIPsubstrate. When developed to the
desired extent, kill the activity of the first phosphatase by incubatmg either with glycine-HCI buffer pH 2.2 (for 1 hl or methanol
(Zr: 15 min). both work well. Incubate with the anti-DIG antibody (l/2000-1/8ooo), wash extensively with TEST or hMBT and
then prwced according to the instructions given by ELF supplier. Generally KM-150 ml of ELF working substrate per slide, kept
in rl humidified c ha m be r in t he da rk . is e nough. Good second colour reaction is usually obtained in 1-24 h. PeriodicAlly check
the extent of the colour reaction under Uv light usi?g DAPl filter.Wash for 15 min. Do not wash extensively after the mlour reaction as this cxn remove that fluorescent preapltate. The sections c;n be counterstained with Hoechst 3342, DAPl or pmpidium
lodldc (Hoechst 33342 is provided in the ELF staining kit). For long-term storage, post-fut the staining with 2% formaldehyde
Xl mg ml-t MA in phosphate-buffered saline PBS for 30 min and mount in Citifluor or ELFmounting medium.
lf. after the fint probe development, sedioning is required !typially for embryos dder than 8.5-9.5d), it is useful to pefiom
;I duubk i#rs&r hybridimtion and its reciprocal, in order to obtain whole mount and overlapping pattern images of both analysec
g~nea. Thr \kuali;llttiun uf the expression and co-expression of two different genes that this methodology provides allows
ccwxprehsion domains to be defined accumtely (Fig. 1). The method does not, however, allow co-expression to be identifiec
within an individual cell due to the nature of the ELF substnte precipitate, which form crystals instead of a homogeneous stain
1°Kprcclpitate. 1’he dark blu<bpurple NBT/BCIP precipitate c%nmask weak Eti staining (e.g. when using to detect weak probes)
hur we have not found this to be a problem with our probes. If masking should occur, we suggest using the N5T/ElCIP to detec
the wc-aker prolx Finally. we have obtained the same results performing the double in sitar hybridmtion on cryostat as well a!
on paraffin scaions WC have followed the standard protocols for in si t u hybridization on sections’” with the modifications statec
aln)vc. Thcw modifications wil! prob_Jbly prove applicable IO a wide variety of experimental systems.
TIG OCTOBER 1996 VOL. 12 No. 10
386
TECHNICAL
TIPS
zyxwvutsrqponmlkjihgfedcbaZYXW
zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPONMLKJIH
fhrorescent frr situ hybridizationro zyxwvutsrqponmlkjihgfedcbaZYXWVU
zebrafish
embryos
networks
that reguiate devrlopment.
rt i\ ofp:n zyxwvutsrqponmlkjihgfedcbaZYXWV
use ful 10 know whether
individual
cells XC’ \lmul-
Double
To undcrsrand
the genetic
aneously
expressing
more
zolour in situ hybridization
than one gene ~zpcctcd
10 be invoh cd in J particul;lr iJr<,cck\. This que%hcm an be 4vert
hy t~\r*
using chmmogrnic
suh>tr.ltes’.l
in ti\\uc sealc,ns
or ~1hole-mount
e m b ryo % . If the domxns
of
this p rc hlc m b y
the stxining o f o ne c hro m o g e n a n m a 4 rlxtrd ~htorher. We have overcome
using fluo re sc e nt sub stra te s fo r a lka line p ho sp llntnse - c c )nj~l~te d
antilxxlies
WC have rested rhrclr difTert,nt Fast Red w hsmtc.s:
t 3 (V e T or
La hor~l~~ri~sI .
(1 1 f,M
Ikd
tahlera~ (Raehringer
.Clannhclm).
and
C1) Vectorm Red alkaline
phosphatase
subsrr~c
13) Sigma Fa@’
Fast Red TR/ Na p htho l A S- .LiX < Sig m a ). Bc h < uh< tr.rle ha ke e n WC C I in c o m hin:ltio n with the c nzym c la tx4c d
fluo re sc e nc e sub stra e (ELF’.
Mo le c ula r Pro b e s). O ne p ro b e is la !+ 4!c d with tluorcs~?in.
and the other with digo:ouigenin. e nd
they are visualized
hy srqurntial
inculrJtion
in alkaline phosphJta.w4
nnlugarccl anubcxlies
to ffunrr.xrin
and digr>xigenin.
rcspectively. The first antibody
is staned
with one of the three Fa.51 Keti formulatlon~
and the second
with rhe ELF 5uhsrnrc
The
1.(,-di:lminn-2-ph~nylindole)
lilter ~‘t.5. re spc c tike ly
signals a re vie we d b y fluo re sc e nc e m ic ro sc o py with rho d a m ine an’1 DAPI
i% ~o le c ub r Pro b e s sup p he s :I va ie ty o f k1t.s c o nta ining ELF T” 51 h:;irXe . WC ilx
the !4 iht l3 t f?
h~miulatton
in the ELF- hP
immunohistochemistry
kit (E-M’100).
mxle up in Tris-HCI ijuffer at pfl 8.2. VrcrmrTw Red i.c
The different Fast Red suh\tmtcs
hzve d&rent
chardaenstics.
m a d e up from three solutions
provided
in the kit. St.linmg is quite r~pld. appc:mng
within a few minutes to l-1 IX Prolonged
incubation
tends not to inter&y
the signal. With zehr&sh
cmhryos.
the yolk ztainb ~~lknv hut background
m rhc embryo rcmainb
lo w. The Fa t Red from &whring e r Ma nnhe im is in the fo rm o f k1hle t.s. The srrlut~ m p ro d uc ts 3 pre c ipita te a fte r p rrrlo ng c d I~C LIb a tio n. The signal de ve lo ps !e szzquickly
than Vecto?
Red. hut producrs
:I mart’ intense red precipitate.
B;tckgounds
can TV
q uite o ra ng e in the yo lk a nd in the e m hvo . With Sig m a FC JSF’ Fa st Re d . ta hlc ts a rc ‘ ~ p p p fie d fo r b o th the Td s b uffe r a nd the Fat
Red substrate. The intensity
of the signal and the speed of development
is slmllar tcr Hochriilger
FJ .s~ ‘~ Re d.
l%ckgmunrl%
are 3
Red AX> produce5
le.%\
little less than with the Roehringer
Fnht T’ I Red hut greater than with Vccrt)r*” Red. %gma FM T’I
AI three .suhstrJteh .fre less sensitive
th:m the mrosf frequently
precipitate
in the staining solution than the Btwhrin,+ 11-r suhstrztr.
used substrate
combination
of nitro blue tetnzolium
chlundei5-hrom~-chlort~3-rn~lolyl-phosph;lt~.
toluidme-salt
tNUT/BCI?.
Boehringer
Mannheim).
However,
we have detected
s1gnal.s wirh a Iarge numhr
of diffcrenr zebmfish
probe.<. With the R.CA+X
fluorescence
becomes
3 problem.
For strung prolxa\ t11c*hcst hignal-to-noise
rztira arc ohrzined
wirh
signals the background
Vector’” Red.
The alkaline phosph;ltare
an he inactivated
by hc~t treating dt hi°C for 30 min \\i!h the V&o?
Red precipitate.
bui bofh
of the other Fast Red precipitates
3re lost If heat treated. Therefore.
the ;rlkalme phc>\phatase
actlviry ir; inactivated
hv incuharinp
in 100 rnbl glycine pH 2.2. 0.1% Tween-20
for 30 min. i?xce&ve
.stSning \\iith the Fat Red .should IX avoided
beau.sc
:I bar-y
red precipitate
an panially quench
the Ew”
signal.
We always perform the ELF staining 35 the second rtzraion 3s It 1~ more .senhlti\c rllan the Fast Red 4>~tr~tc~ and can be left
to develop
for up to 24 h. The inactivation
step is VCT): imponant
3s falure to complercly
inactivate the fint alkaline phrrsphsa*e
will g ive 3 we a k sig na l with rhc ELF. which appears
to be co-ltx~lized
with the t:Jst Red precipitatr.
Thus. we \ug_gcst perforaiing the fo llo wing c o ntro ls.
For some embryos.
in place of incubation
in the .scconcl antibody
(-rep 3 below). cc:ntinne to incubate in blocking solution for the same period and then ctxxmur
w~rh the W:!shc\ a though antiixxty
had lwen added. Suh.sequen~
sta ining with ELF sho uld b e ne g a tive o r g ive a n e vrn, no n- lo a liz~ d
ha c kg ro und . (1) Perfo.m reciproccal q wrim e nts
in whic h
A, l&!led
with Huo rc we m a nti vi.wa lizd
with Fa t Re d : p ro k
the visua lia tio n
m e tho d s fo r the p ro b e s a re re vwe d : (i! probe
R, labelted
with digoxigentn
and visualized
wrh ELF: (ii) p ro b e IS. la b e lie d with tluo rc sc e in a nd visualized
with Fast Ked; probe
ELF signal 15 rmly detected
for one probe when FJ 3 .r Red is
A, labelled with digoxigenin
and visualized
with ELF If ctrltxdircd
used to de te c t the strongest
signa! then the inactivation
of the alkaline phn@ar;lx~
1~1s probably
incomplere.
(3) H~hndizitions
should be carried out with single p ro b e s a nd \&u&&on
with fisr Red. ELF and V+T/BCIP sepxxrely
to dertxnime
direshalth
of detection
for different domains
of e xp re ssIo n o f the tm nsc np t.
- xp re ssio n
2
o ve rla p . ho we ve r.
zyxwvutsrqponmlkj
C
LII are
Fast
(1)
387