Molecular Brain Research, 12 (1992) 7-22 © 1992 Elsevier Science Publishers B.V. All rights reserved. 0169-328X/92/$03.50 BRESM 70347
Isolation and characterization of a library of cDNA clones that are preferentially expressed in the embryonic telencephalon Matthew H. Porteus*, A. Elizabeth J. Brice*, Alessandro Bulfone*, Ted B. Usdin**, Roland D. Ciaranello and John L.R. Rubenstein* Nancy Prizker Laboratory of Developmental and Molecular Neurobiology, Department of Psychiatry and Behavioral Sciences, B002, Stanford Medical School, Stanford University, Stanford, CA 94305 (U.S.A.) (Accepted 11 June 1991)
Key words: Telencephalon; cDNA cloning; Development; Subtractive hybridization; Striatum; Cerebral cortex; Differential gene expression
In order to isolate genes involved in development of the mammalian telencephalon we employed an efficient cDNA library procedure. By subtracting an adult mouse telencephalic cDNA library from an embryonic day 15 (El5) mouse telencephalic cDNA library we generated two subtracted libraries (ES1 and ES2). We estimate that ES1 contains between 200 and 600 different cDNA clones, which approximates the number of genes that are preferentially expressed in the El5 telencephalon, compared to the adult telencephalon. Northern analysis of 20 different cDNA clones shows that 14 of these are expressed at least 5-fold more in the El5 telencephalon than the adult telencephalon. Limited sequencing of the 14 differentially expressed clones reveals that 10 have no significant identity to sequences in GenBank and EMBL databases, whereas the other 4 have significant sequence identity to vimentin, historic 3.3, topoisomerase I and the B2 repeat element. In situ hybridization using one of the differentially expressed cDNAs, TES-1, demonstrates that it is transiently expressed in the anlage of the basal ganglia. In situ hybridization with another differ¢ntially expressed cDNA clone, TES-4, shows that it is specifically expressed in differentiating cells of the neural axis with a distinctive rostral-caudal temporal pattern. These findings, and the methods that we have developed, provide a framework for future investigations of the genetic control of telencephalon development.
INTRODUCTION The telencephalon is the rostral end of the neural tube and forms as a pair of evaginations from the forebrain 29' 3~.47. The forebrain differs from the more caudal central nervous system in several respects. The convex shape of the forebrain neural groove is different from the concave shape of the rest of the neural tube 3a. The forebrain neuroectoderm, unlike the more caudal neuroectoderm, has no imderlying notochord, a structure which is important in neural induction 4°. Instead, the forebrain neuroectoderm overlies the prechordal plate, and does not have a floor plate 3. These differences suggest that some aspects of forebrain development may involve mechanisms that are not used in the rest of the nervous system. Early in development, the walls of the entire neural tube are made up of undifferentiated pseudostratilied neuroepithelial cells. A t specific times, cells leave the mitotic cycle and migrate from the ventricular zone to-
wards the pial surface. Depending upon the location of the ventricular zone, the cells contribute to the formation of different structures. For instance, in the telencephalon, cells from the dorsolateral walls form the neocortex, those from the medial walls form the archicortex, and those from the ventral walls form the striatum 29. Each subregion of the telencephalon is organized in a different way. For instance, the mature neocortex is a laminated structure consisting of six cell layers 3°'42 while the mature striatum is a mosaic of patch cells in a sea of matrix cells 17'19'20'51. The cells of the different layers and compartments differ in many ways including their afferent and efferent connections, their repertoire of neurotransmitters and neurotransmitter receptors, and the time in ontogeny at which they become post-mitotic. The molecular mechanisms which regulate how the ventricular zone gives rise to these specialized structures and defined cell types are not known. Genes which have increased expression during particular developmental
* Present address: Department of Psjchiatry, Langley Porter Psychiatric Institute, University of California at San Francisco, 401 Parnassus Ave., San Francisco, CA 94143, U.S.A. ** Present address: Laboratory of Cell Biology, NIMH, Bethesda, MD 20892, u.S.A. Correspondence: J.L.R. Rubenstein. Present address: Department of Psychiatry, Langley Porter Psychiatric Institute, University of California at San Francisco, 401 Parnassus Ave., San Francisco, CA 94143, U.S.A.
events are likely to be involved in these events and should provide molecular information about how these developmental processes are regulated. To date, the characterization of developmentally regulated genes encoding proteins that mediate cell-cell interactions or that regulate transcription have begun to elucidate the molecular mechanisms of insect neural and vertebrate CNS development14.21.25. It is likely that telencephalic development is controlled in large part by the expression of genes whose products perform critical regulatory roles in differentiation. The use of mutations to identify these classes of genes has been successful in invertebrates 2s'37. While several mutants in mammalian CNS development have been described 6 and the use of insertional mutagenesis promises to create more 1'5s these approaches have, as of yet, led to the identification of only a small number of genes important in neurodevelopment 24'3s. Like a genetic screen, nucleic acid subtractive hybridization can be used to focus a search for genes that affect certain developmental events t2'22. This technique has also been used by neurobiologists to identify mammalian brain-specific or neurodevelopmentally regulated genes 9'16'31'43'5°, and in Drosophila melanog~ter it has been used to generate a collection of head-specific cDNA clones 39. In order to systematically identify both rare and abundant m R N A species that are expressed in specific temporal and spatial patterns within the telencephalon, we developed methods for performing subtractive hybridization between directional cDNA libraries in single-stranded phagemid vectors (see Fig. 1 and reference 44). We have now applied these methods to perform subtractive hybridizations between cDNA libraries prepared from embryonic day 15 (El5) and adult mouse telencephalons. By El5 of mouse telencephalic development, the cerebral cortex and stHatum are beginning to differentiate. Cells undergoing their final mitotic division at this time are destined to become the deep layers of the cerebral cortex 4'32 and the patch cells of the striatum 51. Embryonic structures such as the ventricular zone, the subventricular zone, intermediate zone and the subplate are still present 2. Genes which are preferentially expressed in the El5 telencephalon, compared to the adult telencephalon, may play important roles in the function of these transient embryonic structures, and in the differentiation of the newly post-mitotic cells. In this paper we describe the use of subtractive hybridization to construct two cDNA libraries which are highly enriched for cDNAs corresponding to genes that are preferentially expressed in the embryonic telencephalon. From these libraries we have isolated and characterized 14 different c D N A clones corresponding to differentially expressed genes. One cDNA clone is derived from a gene that is expressed
throughout the nervous system in differentiating cells, and one is transiently expressed in the developing striamm. MATERIALS AND METHODS Molecular biology General molecular biology techniques were used as described in refs. 11 and 45, except where noted. Isolation of poly(A +) RNA Timed pregnant Balb C mice were sacrificed by cervical dislocation. The day on which a vaginal plug was found was defined as embryonic day 1 (El). Embryos were dissected in ice-cold Dulbecco's modified Eagle's medium (DMEM). Brains were removed and telencephalic vesicles dissected from the rest of the nervous system tissue. The telencephalons were frozen in liquid nitrogen and stored at -70 °C. Adult Balb C mice (males and females) were sacrificed by decapitation and the telencephalons were removed by dissection. Total RNA was purified using guanidine isothiocyanate and cesium chloride centrifugation. Poly(A+) RNA was purified using one cycle of oligo(dT) cellulose (Collaborative Research Type III). 1.4 g of El5 telencephalon (from 80 embryos) yielded 2.5 mg total RNA and 60/tg of poly(A +) RNA. 10 g of adult telencephalons (from 31 animals) gave 6.0 mg of total RNA and 300 /~g of poly(A +) RNA. Construction of directional cDNA libraries in phagemid vectors eDNA libraries were constructed with inserts in opposite orientations in two phagemid vectors, E61 and E145, using an adaptorprimer method44. Each was transformed into two E. coli strains, MC1061 and XL-1. We used MC1061 because of its high transformation efficiency by electroporation, and XL-1 because it can be used for phagemid rescue. Size-selected libraries were made by cutting phagemid DNA from the MC1061 iiblaries with the restriction endonuclease Nod, and purifying phagemids with inserts greater than 1 kB by low melt agarose gel electrophoresis. The linear phagemids were cyclized by ligation, and transformed into XL-1. See Table I for a list of these libraries. Conversion of cDNA fibraries to single-stranded phagemids Single-stranded phagemid phagemlds' were rescued from thes3 cDNA libraries in XL-1 bacteria using the bacteriophage R408" . Single-stranded DNA was purified using potassium iodide density gradient centrifugation, magnesium-phenol extraction, and restriction endonuclease digestion44. Preparation of biotinylated driver Single-stranded phagemid DNA, which was to be used as driver in the subtractive hybridizations, was biotinylated44. Before using biotinylated driver in subtractive hybridization, it was verified to transform bacteria at >10e lower efficiency than unbiotinylated target and that it could be quantitatively extracted using the streptavidin-phenol/chloroform method. Subtractive hybridization The three subtractions described in the paper were performed using slight modifications of the methods described in Rubenstein et a144. The first subtractive hybridization led to the production of the Embryo Subtracted 1 Library (ES1). The hybridization reaction contained 20 #g of biotinylated driver (anti-sense adult singlestranded cDNA library) and 0.5 ~ug of target (sense El5 singlestranded cDNA library). Two marker DNAs were also included: 5 ng of pACYC184 and "--0.5#g of R408. 2/~g each of three oligonucleotides (A: 5" dGCGGCCGCITITITITIT 3"; B: 5" dATCAAGCTrATCGCCCT 3"; C: dT 19-24) were added to the reaction. These oligonucleotides are complementary to the polylinker and poly(A) domains of the target phagemids, and therefore can prevent d•er-target hybridization in these regions of the phagemids. The DNA was dissolved in lib (40% deionized forma-
mide, 500 mM NaCi, 50 mM Na HEPES, 2 mM Na EDTA, pH 7.6), covered with light-weight mineral oil (Sigma), heated to 90 °C for 2 min, and incubated at 52 °C for 7 days. Biotinylateddriver and driver-targetheteroduplexes were removed using streptavidin and phenol/chloroformextraction~'4s, The targetphagemids leftin solutionwere taken through a second round of subtractivehybridization with 20 l~g of biotinylated anti-senseadult single-stranded c D N A libraryas driver,exactly as in the firstround, After the two rounds of hybridization,we estimate that a CoT of 1,000 mol nucleotides/s/literwas reached. The second subtractive hybridization produced the Adult Subtracted Library (AS). The hybridizationreaction contained I0 ~g of driver (biotinylatedsense adult single-strandedc D N A library) and 0.7/~g of target (anti-sense adult single-strandedc D N A library). It also contained two markers; 2 ng of E28 (a kanamycin resistantplasmid, see Rubenstein et al.44)and -0.3/~g of R408 DNA. The reaction contained 2/~g each of 3 blocking oligonucleotides (D: 5" d A A A A A A A A A A O C G G C C O C 3", E: 5" dCOATAAGCTI'GATI" 3", and F: dA 19-24). The D N A was heated in liB to 90 °C for 2 min and then incubated for 7 days at 52 °C to a CoT of approximately 300 mol nucleotides/s/liter. At this CoT, most of the highly and moderately abundant cDNA clones should have hybridized, thereby enriching for the rarer cDNA clones 52. The third subtractive hybridization produced the Embryo Subtracted 2 Library (ES2). The hybridization reaction contained 17 /~g of biotinylated single-stranded driver (4/zg of anti-seuse adult cDNA library, 9/~g of size-selected anti-sense adult cDNA library, and 4/~g anti-sense adult from the AS cDNA library) and 3/tg of single-stranded target (E15 sense size-selected library). The target also contained two marker plasmids (4 ng of pACYC184 and - 1 /tg of R408) and 2/~g each of blocking oligonucleofides A, B and C. The reaction went through two rounds of subtractive hybridization exactly as described for ES1 to a CoT of approximately 1000 mol nucleotides/s/iiter.
Conversion of single-stranded phagemids to double-strand Single-stranded phagemids were converted to double-stranded form as previously described 44 using oligonucleotide A as the primer. Oligonucleotide A spans the polylinker-cDNA junction of target libraries in the sense orientation. This oligonucleotide was used for ESI and ES2. Oligonucleotide D was used as the primer for AS. The reaction was performed as described in Rubenstein et al. 44. Because these primers span the polylinker-cDNA junction, they select against target phagemids without inserts, target phagemids with inserts in the wrong orientation, and driver phagemids being converted to double-stranded form. The use ,af Taq DNA ; Jlymerase allowed high-temperature annealing of primer to template, thereby increasing the specificity of the single-stranded to double-stranded conversion process. Bacterial transformation and antibiotic resistance avsay The products of the Taq D N A polymerase reaction were introduced into MC1061 by electroporation 15. Det ernnmn " " g the effi ciency of subtraction by antibiotic resistance was done as in Rubenstein et al.44. Colony screening, and Southern, Northern, and slot blot analysis Colony screening, southern analysis, and northern analysis were performed using standard molecular biology procedures 45. cDNA probes were synthesized in two ways. The first used a standard reverse transcription protocol 45 with 1/tg of poly(A +) RNA, and an oligo(dTt5 ) primer to generate a probe of specific activity of 2 x 107 cpm//~g; this was used in the colony screening experiments shown in Fig. 3. The second protocol involved using random primed DNA synthesis kit (Pharmacia) to label 50 ng of double-stranded cDNA to a specific activity of -109 cpm//~g. This cDNA was originally produced for the construction of the cDNA libraries 44. These probes were used for the southern (data not shown) and slot blot analysis in Fig. 5. Probes for the northern analyses (Fig. 4) were
generated by random-primed labelling of agarose-gel purified cDNA inserts from XhoI-HindIII digested phagemlds. The probes use~ ,:~ backscreen ES1 to determine the number of different cDNA clones were produced in the following manner. Inserts from the cDNA clones were cut out from the vector by HindIII-Xhol digestion and purified by agarose gel electrophoresis. Oligonucleotide A, which is specific to the polylinker-cDNA junction, was used ~,s primer for the DNA polymerase. It was annealed to the cDNA inserts in 50 mM Tris-HCl pH 7.5, 10 mM MgCI2, 1 mM dithiothreitol, 300/~M dATP, dGTP, and d'ITP, by heating to 95 °C for 3 min and then slowly cooling to 37 °C. 50/tCi[32p]dCTP (3000 Ci/mM) and Klenow fragment (2.5 units) were then added, and the reaction was incubated for 45 min at 37 °C. Probes of a specific activity of 8 x l0 s cpm//~g were generated. We found that labelling with a specific primer dramatically reduced background hybridization in colony screening compared to random priming, probably by reducing the labelDng of trace amounts of vector sequences that contaminate the purified cDNA inserts. Quantitation of hybridization was done by scanning the autoradiograms using a LKB scanning densitometer.
DNA sequencing DNA sequencing was performed using the Sequenase 2.0 kit (U.S. Biochemical) on double-stranded template using T3 or T7 oligonucleotide primers according to manufacturer's specifications. Analysis of the DNA sequences were performed using the Genetics Computer Group sequence analysis software, and the GenBank and EMBL databases. In situ hybridization Frozen sections of whole embryos or brains were cut and desiccated according to Herkenham and Pert 23 and fixed, hybridized, and washed according to Mueller et al.35. In brief, tissue was frozen in 2-methyl-butane on dry ice, cut in 12/~m sections at -20 °C, mounted on chrom-alum coated slides, kept on dry ice, desiccated overnight at -2 °C, and stored at -70 °C. Sections were fixed in DEPC (diethylpyrocarbonate)-treated PBS (150 mM NaCl, 10 mM sodium phosphate pH 7.4) containing 4% paraformaldehyde (fresh) for 20 min, rinsed 3 times in PBS for 5 min, dehydrated through graded ethanol, and air-dried. Sections were then incubated in 1 /~g/ml proteinase K in PKB (100 mM Tris-HCI pH 8.0, 50 mM EDTA) at 37 °C for 30 min and post-fixed as above. Sections were acetylated by rinsing in 0.1 M triethanolamine, blotting dry, and incubating in 200 ml of 0.1 M triethanolamine containing 500/~l of acetic anhydride for 10 min at 20 °C. They were then rinsed in DEPC-treated water, air-dried, and stored at 4 °C. Hybridization was as follows: 106 cpm of riboprobe was heated in 40/A of ISHB (50% deionized formamide, 5 x SSC (1 x SSC = 0.015 M sodium citrate, 0.15 M NaCI pH 7.0), 10 mM/~-mercaptoethanol (BME), 10% dextran sulfate, 2 x Denl'lardt's, 250/~g/ml yeast tRNA, 500 /~g/ml single-stranded sh~ared herring sperm DNA) to 65 °C for 5 min, and cooled to 20 °C. 40 ltl of ISHB-riboprobe was layered on top of each tissue section. The samples were then covered with a parafilm coverslip, and incubated in a humidified chamber at 42 °C for 18 h. The slides were twice washed for 30 rain at 20 °C in 2 x SSC; 10 raM BME; then in 2 x SSC, 1 mM EDTA containing 20 /~g/ml RNase A and 1 unit/aft RNase T1 (both from Pharmacia) at 20 °C for 40 min; then twice in 50% formamide, 2 x SSC, 1 mM EDTA, 10 mM BME at :.~0°C for 30 min; then briefly washed at 20 °C in 0.2 x SSC, and dehydrated in 75% and 95% ethanol. The sections were autoradiographed for 1-2 days at 20 °C using Hyperfilm-/~lax (Amersham). Finally the sections were dipped in NTB2 emulsion (Kodak) and exposed for 6 days. The emulsion was developed using Kodak D-19, and the tissue was counterstained with Cresyl violet. The samples were examined using a Nikon Optiphot-2 microscope, and photographed using Ektachrome ASA 200 film. The anti-sense cY,TdqA probe for in situ hybridization was made according to manufacturers specifications using the Stratagene riboprobe kit. Briefly, 1 ~g of proteinase K-treated, HindIII-digested
10 TABLE I
Catalogue of cDNA libraries The characteristics of each cDNA library used in this paper are summarized in the table. The first column lists the names of each library. The second column states which vector was used in the given library. E61 and E145 are phagemid vectors based on pBluescript described in Rubenstein et al."u. The third column lists the source of the poly(A +) RNA from which the eDNA was synthesized. The fourth column lists the orientation of the cDNA inserts in the phagemid vector; in the sense orientation the sense strand of the cDNA is recovered when performing the M13 rescue procedure. The fifth column lists the bacterial strain that the fibrary was made in. The sixth column lists the number of independent recombinants in each library. The final column lists the average insert size in the entire eDNA library, which was estimated from the size distribution of the cDNA inserts.
Name
Vector
Tissue source
Orientation
Bacterial strain
Independent recombinants
Averageinsert size (kb)
E159 El60 E162 E163 El70 El71 E173 E174 E212 E213 ES1 ES2 AS
E61 E61 E145 E145 E61 E61 E145 E145 E145 E61 E61 E61 E145
Adult Telen. Adult Telen. Adult Telen. Adult Telen. E15 Telen. El5 Telen. El5 Telen. E15 Telen. Adult Telen. El5 Telen. E15 Telen. El5 Telen. Adult Telen.
Sense Sense Anti-sense Anti-sense Sense Sense Anti-sense Anti-sense Anti-sense Sense Sense Sense Anti-sense
XL-1 MC-1061 XL-1 MC-106I XL-1 MC-1061 XL-1 MC-1061 XL-1 XL-1 MC-1061 MC-1061 XL-1
2x 4x 2x 4× 1.6 x 8x 1.6 x 8x 1x 3x 2.5 X 5x 5×
1 1 1 1 1 1 1 1 1.5 1.5 0.6 1 0.6
phagemid was incubated with 10 units of T3 RNA polymerase, 50 /~Ci[35S]rUTP (1270-Ci/mM), in 40 mM Tris-HCl pH 8.0, 8 mM MgCI2, 2 mM spermidine, 50 mM NaCI, 30 mM DTI', 400/~M rATE rCTP, rGTP, 25 units RNase Block at 37 °C for 30 min. The sense cRNA probe was made using Nod-digested DNA template and T7 RNA polymerase. RESULTS
Construction of adult and El5 telencephalic directional cDNA libraries We made directional cDNA libraries using an adaptor-primer procedure in a modified pBluescript vector from poly(A +) R N A from El5 and adult telencephalons as previously described 44. Libraries were made from both tissues with cDNA inserts in both orientations and were amplified in XL-1 Blue and MC-1061 bacteria. The original libraries had average insert sizes of about 1 kilobase (kb). From these libraries we made size-selected libraries with inserts > 1 kb. The characteristics of each library are summarized in Table I. From each of the XL-1 libraries, single-stranded phagemids were prepared by R408 helper phage rescue s3. Anti-sense libraries generated single-stranded anti-sense phagemids and sense libraries generated single-stranded sense phagemids. After helper phage rescue, the. average insert size of the libraries decreased to approximately 0.6 kb. The decrease is probably the result of smaller molecules converting to single-stranded phagemids more effÉciently than larger molecules since we have found no evidence of deletions.
106 107 106 10 7
106 106 106 106 106 105 10 3
104 104
Subtractive hybridization between adult and El5 libraries: production o f ES1 To isolate clones preferentially expressed in the E l 5 telencephalon, we used cDNA library subtraction. The general method is shown in Fig. 1 and is described in detail in Rubenstein et al. 44. All of the subtractions desc~'ibed in this paper consist of two parallel reactions: subtracted (S) and control (C). The S reaction contains both driver and target D N A , while the C reaction contains only target D N A . The abbreviations S and C will stand for the 'subtracted' and 'control' reactions throughout the paper. In the first subtraction, the driver was the biotinylated adult single-stranded c D N A library with inserts in the anti-sense orientation, and the target was the E l 5 single-stranded c D N A library with inserts in the sense orientation. The target contained two marker D N A molecules to measure the efficiency of the subtraction: 6.4 kb single-stranded R408 helper phage and the 4.2 kb double-strando, d tetracycline resistant plasmid, pACYC184. The target was processed through two rounds of subtractire hybridization to a final CoT of about 1,000. We chose a CoT of 1000 mol nucleotides/s/liter based upon the hybridization results of Van Ness and Halm 52 who showed that under these conditions >90% of infrequent copy brain R N A would hybridize in a R N A - c D N A reaction. After the subtractive hybridization, we converted the remaining single-stranded target molecule.~ to doublestranded form (double-stranded phagemids transform bacteria 300- to 1000-fold more efficiently than singlestranded phagemids 44) using Thermus aquaticus (Taq)
11 D N A polymerase. The products of the Taq D N A polymerase reaction were transformed into bacteria and thereby generated the 'Embryo Subtracted 1 Library' (ES1) (Table I). ES1 contains 2500 independent recombinants with an average insert size of 600-700 bp. Of 60 clones analyzed, all contained inserts, suggesting that most or all of the 2500 independent recombinants contain inserts. The control library had a complexity of 25,000 independent recombinants and had an average insert size of 600-700 bp. B
'-0o0o0! @ B
~ Hylxidbudlon Reection B
• B OrganicP h a s e
Go @ o I
S~d,n
I Phenol:CHCIs ~ ~ , , Extraction
AqueousPhase
B
O)! (E) O' Convert to
Double Stranded
Trsnsform F..¢oli Subtracted Ubrary
Fig. 1. Library subtraction scheme. To isolate genes, such as Gene Z, from t,e E15 target fibrary that are preferentially or exclusively expressed during development, we used the cDNA fibrary subtraction scheme depicted. Biotinylated (B) single-stranded driver phagemids are mixed with single-stranded target phagemids and two marker DNA molecules: single-stranded R408 and doublestranded pACYC184. Adult (A) cDNA clones are cloned in one orientation, depicted by an up arrow, and El5 (E) cDNA clones are cloned in the opposite orientation, depicted by a down arrow. cDNAs in the target fibrary, such as actin, will hybridize to their complements in the driver library to form a heteroduplex, cDNAs that have no complement will remain single-stranded. After hybridization all driver molecules and heteroduplexes are removed from aqueous solution by adding streptavidin which specifically binds biotin and then extracting with phenol/chloroform. Target molecules which have not hybridized to the driver DNA, sl~ch as Gene Z, as well as the marker molecules R408 and pACYC184, remain in aqueous solution. The unsubtracted target molecules are converted to double-stranded form using a strand-specific primer and Taq DNA polymerase. The product of the Taq DNA polymerase reaction is transformed into E. coli by electroporation and a subtracted library of preferentially expressed cDNA clones is thereby created.
Analysis of ES1 We analyzed the efficiency and specificity of the ES1 subtraction in 5 ways: agarose gel electrophoresis, southern analysis, a marker plasmid assay, a colony screen assay, and slot blot analysis. Comparing ethidium bromide fluorescence of the single-stranded DNAs from the ES1-S and ES1-C reactions separated by agarose gel electrophoresis showed that most of the phagemids had been extracted from the ES1-S reaction (Fig. 2). The R408 D N A , which should not be extracted in this procedure, was only slightly reduced ~n the ES1-S reaction. We attribute the small decrease to non-specific trapping of R408 D N A during the extraction of biotinylated driver-target duplexes. Thus, while the R408 band is only slightly reduced, the amount of target phagemids is greatly reduced, suggesting that ES1 is enriched for sequences that are preferentially expressed in the E l 5 telencephalon. This result qualitatively demonstrated that the subtraction worked and we proceeded to analyze ES1 further. Second, we analyzed ES1-S and ES1-C by southern analysis (data not shown). Qualitatively, radioactive adult telencephalic c D N A hybridizes much less to ES1-S D N A than to ES1-C D N A . Therefore, abundant sequences found in the adult telencephalon have been removed from ES1. ES1-S D N A hybridizes more to the probe from E l 5 c D N A than to the adult probe. Thus, ES1 contains cDNA clones that hybridize more to a probe derived from E l 5 telencephalic poly(A +) R N A than from a probe derived from adult telencephalic poly(A +) RNA. Third, to determine the efficiency of the subtraction, we measured the ratio of the pACYC184 marker plasmid (tetracycline resistant colonies) to target molecules (ampicillin resistant colonies) between ES1-S and ES1-C samples; this is the marker plasmid assay 44. After sub-
TARGET: DRIVER:
R408
EMBRYO (x A D U L T
e ADULT ADULT C S
[[
Fig. 2. Analysis of ES1 and AS by agarose gel electrophoresis. After one round of subtractive hybridization, one-third of the control (C) and subtraction (S) reactions from ES1 (target: sense embryo; driver: anti-sense (a)-adult) and AS (target: anti-sense (a)-adult; driver: sense-adult) were electrophoresed on an agarose gel and stained with ethidium bromide. The location of R408, a single strand marker molecule is shown by an arrow. Target molecules that remain in solution after the subtractive hybridization are the fluorescent smear below the R408 band.
12 traction the marker plasmid was enriched 70-fold in the ES1-S sample relative to the ES1-C sample (Table II). As another indicator of specificity, the number of marker tetracycline colonies remained constant between the ES1-S and ES1-C reactions (Table II). Fourth, we screened colonies from the S and C reactions of ES1 with adult telencephalic cDNA; this is the colony screen assay. The cDNA probe hybridized to 60 out of 270 colonies from the ES1-C sample but only 2 out of 240 colonies from the ES1-S sample (Fig~ 3). These results suggest that ES1 is enriched by approximately 30-fold for genes that are preferentially expressed in the El5 telencephalon. Fifth, we measured the efficiency of the subtraction by screening slot blots containing ES1-S and ES1-C D N A with adult or E15 telencephalic cDNA; this is the slot blot analysis (Fig. 5 and Table II). Scanning densitometry shows that the amount of hybridization to the ES1-S sample is approximately 40-fold less than to the ES1-C sample with the adult cDNA probe. Based on the marker plasmid assay, the colony screen assay, and the slot blot analysis we estimate ES1 to be enriched for genes preferentially expressed in the E15 telencephalon by 30- to 70-fold. ES1 contains 2500 independent recombinants. To estimate how many of these clones represented different cDNAs, we made radioactive probes from cDNA inserts from eleven random clones and screened ES1 colonies. One of the clones, TES-4, is represented approximately
100 times. The other ten cones are represented on average 4 times. Thus, if we exclude the TES-4 clones from the calculation, there are approximately 600 different clones in ES1 (2400/4), and if we include T E S - 4 there are approximately 200 different cDNA clones in ES1 (2500/13). Subtractive hybridization between sense adult and antisense adult libraries: production o f A S
Rare cDNAs are less efficiently subtracted in any subtractive hybridization because of kinetic constraints 52. In order to increase the probability of subtracting shared but rare cDNAs, we created a library of less abundant adult cDNAs by subtractive hybridization to use as supplemental driver in subsequent subtractive hybridizations. In this subtraction, the driver was the singlestranded sense adult c D N A library and the target was the single-stranded anti-sense adult c D N A library. The complexity of this subtracted library, the 'anti-sense Adult Subtracted Library' (AS), was 45,000 independent recombinants with an average insert ~ize of 600-700 bp. We analyzed AS by agarose gel elecuophoresis (Fig. 2), the marker plasmid assay (Table II), and the colony screen assay (Fig. 3). Agarose gel electrophoresis shows that the amount of single-stranded c D N A phagemid had decreased in the AS-S sample compared to the AS-C sample but the R408 marker for specificity remained approximately constant. The marker plasmid assay estimates that this subtraction enriched for less abundant
TABLE II Efficiency of subtractive hybridizations
The table summarizes the analyses of the 3 different subtractive hybridizations described in the paper: ES1, ES2, and AS. S refers to the subtracted sample and C to the control sample from each subtractive hybridization. CoT is the molar concentration of driver nucleotides (Co), multiplied by the length of hybridization in seconds (T). Total Amp R Colonies refers to how many ampiciilin resistant colonies (target phagemids) were recovered after subtractive hybridization. Analogously, Total TetR Colonies refers to the number of tetracyclineresistant marker colonies (pACYC184 plasmid) recovered after subtractive hybridization. In the case of AS, the marker colonies were kanamycin resistant (E28) and not tetracycline resistant. The AmpRPl'etR ratio is determined by dividing the total number of ampiciUin resistant colonies by the total number of tetracyctine resistant colonies. Transformation refers to the estimate of subtractive hybridization efficiency using the bacterial transformation assay and was determined by dividing the AmpR/TetR ratio from the C sample by the AmpR/'I'etR ratio from the S sample. Colony screen refers to the estimate of subtraction efficiency based on the colony screen assay (see text) and Slot blot refers to the estimate of subtraction efficiency based on the slot blot analysis (see text). Subtraction
CoT
ES1 -S -C
1000
ES2 -S -C
1000
AS
350
-S
-C
Total Amp R colonies
Total TetR colonies
2.4 x 103 1.9 x 105
1.7 x 104 1.9 x 104
0.14 10
5.3 x 104 7.2 x 105
1.2 x 104 1.6 x !04
4.4 45
104 1.1 × 106
7.5 × 103 5.7 x 103
6
4.5 ×
AmpRITetR
193
Transformarion
Colony screen
Slot blot
70
30
40
10
-
10
30
4
13
SUBTRACTED & •
CONTROL ES1 LIBRARY
ES1 LIBRARY o
& &
e
e
~
e g
B
7
O
A
~
IY
Fig. 3. Analysis of ESI and AS by colony hybridization to adult telencephafic cDNA probe. 240 colonies from ES1-S, 270 colonies from ES1-C, 100 colonies from AS-S, and 240 colonies from AS-C were screened with an adult telencephalic cDNA probe. Each dark dot represents a bacterial colony which harbors a cDNA phagemid that hybridizes with the probe. The triangles depict orientation markers on each filter. The filters were hybridized for 3 days in 20 ml of hybridization mix, and were washed twice at room temperature with 2 x SSC and twice at 65 °C for 30 rain in 0.1 x SSC and 0.1% SDS. The autoradiograms were exposed for 20 h.
adult cDNAs by 30-fold. The colony screen assay showed that the adult telencephalic probe hybridized to 110 colonies out of 270 from the AS-C sample but only to 10 out of 100 from the AS-S sample in the colony screen assay; this result suggests a 4-fold estimate for enrichment. Thus, we estimate AS to be enriched for less
abundant cDNA clones by 4- to 30-fold.
Subtractive hybridization of size-selected El5 target using driver that contains the adult subtracted library: production of ES2 The cDNA inserts in ES1 were relatively small. To
14 well as the non-size-selected anti-sense adult c D N A to ensure that any t a r g e t inserts < 1 kb w o u l d be sub-
TES-1 TES-2 TES-3 TES-4 TES:6 E
A
E
A
E
A
E
A
E
A
tracted. This subtraction g e n e r a t e d the ' E m b r y o Subtracted 2 Library' (ES2) (Table II). ES2 contains 50,000 independent r e c o m b i n a n t s with an average insert size of
~i!:i+:.~;.~+++
1 kb. We obtained a larger n u m b e r of i n d e p e n d e n t recombinants in ES2 t h a n ES1 because we used larger
28S -- I
~
--- 28S
18S
- - 18S
A
!:+:i+~:i:i:
cyclo-
III ill
,.i,,
..
Fig. 4. Northern analysis. Inserts from TES-1, TES-2, TES-3, TES-4, and TES-6 were isolated "y NotI-HindIII digestion, labelled by random priming, and used to probe northern blots consisting of 10/~g of total RNA from E15 and adult telencephalons. In each blot, E marks the lanes with E15 RNA and A marks the lanes with adult RNA. The migration of the 28S and 18S ribosomal markers are marked with arrows. The left arrows mark the ribosomal RNA migration for the three blots on the left, while the fight arrows mark the ribosomal RNA migration for the two blots on the right. The blots were hybridized for 48 h in 10 ml of hybridization mix, and washed twice at room temperature for ten min in 2 x SSC and 0.1% SDS, and twice at 60 °C for 30 min in 0.1 x SSC and 0.1% SDS. The TES-1 blot was au.~oradiographed for 12 h, the TES-2, TES-4, and TES-6 blots were autoradiographed for 18 h, and the TES-3 blot was autoradiographed for 7 days. No hybridization could be detected in the adult lane with TES-4 even after a 7 day exposure. To establish that equal amounts of undegraded RNA was loaded onto each lane, the same blots were later probed with cyclophilin (cyclo), a non-developmentally regulated gene z~.
I
1Q
I
1"
II
2*
II
2*
III
S
III
N
3*
IV
3"
11
7
11
7
12
8
12
8
13
4*
13
4•
14
9
14
9
S
6*
S
6"
C
10
C
10
S'
Cyclo
S'
Cyclo
C'
pBS
C'
pB$
E
create a subtracted library with larger inserts we used the anti-sense adult and sense E15 c D N A libraries with inserts > 1 kb as the driver and target respectively. We also added A S c D N A to the driver in o r d e r to increase the efficiency of subtracting less a b u n d a n t E15 c D N A s , as
Fig. 5. Slot blot analysis of ES1, ES2 and individual clones. 1 ~g of DNA from ES1-S (S), ES1-C (C), ES2-S (S') and ES2-C (C') and from individual cDNA clones was blotted onto duplicate Duralon membranes either by slot blot (the lower two panels), or by oval blot (the upper two panels). One blot was probed with a random primed El5 telencephalic cDNA probe and the other was probed with a random primed adult telencephalic cDNA probe. The blots were hybridized for five days in 10 ml of hybridization mix, and washed using the conditions described in the legend to Fig. 4. The blots were autoradiographed for 2 days. Clones from ES1 have Arabic letter names, whereas clones from ES2 have Roman numeral names. More information regarding many of the clones is given in Table lII. Cyclo is the abbreviation for the cDNA clone encoding cyclophilin and pBS is the abbreviation for the phagemid vector (the E61 form of pBS) used for all of the TES clones. The cDNA clones marked by an asterisk are those that are also shown in Fig. 4.
~.~
D
5
A
Q
I t
I
2*
V
3*
6*
3'
6"
Xll
Vl
Xll
Vl
Xlll
VII
XlII
VII
VIII
XlV
VIII
IX
XV
IX
XVl
X
XVl
X
XVll
Xl
XVll
Xi
16
!5
16
15
4*
Cyclo
41
Cyclo
18
pBS
18
pBS
XIV XV
41 .~
15 TABLE III
Summary of TES clones We have performed northern analysis on 20 eDNA clones encoding different mRNA tra nscri_'pts. Tb_e characteristics of each clone are summarized in this table. Clones from ES1 are given Arabic numerals and clones from ES2 are given roman numerals. Subtraction (SUB) refers to the subtracted library from which the cDNA clone was isolated. INSERT is the size in kb of the cDNA insert. E/A refers to the level of differential expression based on the northern analysis, in which the ratio of hybridization to the major mRNA species is compared between the El5 and the adult RNAs. D refers to a northern in which there is clear evidence that the transcript is differentially ~xpressed, but the background was too high to carefully quantitate. N.B. means that no bands were seen on the northern. Major is the size of the most abundant RNA transcript in the E15 telencephalon and Minor is the size(s) of the less abundant RNA transcripts. The scale for the relative abundance (Abund.) is as follows: + - 0.5%. cDNA clones that have been partially sequenced (Sequence) are marked. TES-11 has a partial B2 sequence, TES-13 is homologous to the 3" end of a human topoisomerase 1 cDNA, TES-V encodes the vimentin coding sequence, and TES-XIV encodes the histone 3.3 coding sequence, cDNA clones which have been examined by in situ hybridization are marked in the in situ column.
Name
Sub
Insert
E/A
Major
Minor
Abund.
Sequence
TES-1 "FES-2 TES-3 'IT.S-4 TES-6 TES-7 TES-8 TES-10 TES-11 TES-12 TES-13 TES-14 TES-V TES-VI TES-X TES-XII TES-XIII TES-XIV TES-XV TES-XVI TES-XVII
ES1 ES1 ES1 ES1 ES1 ES1 ES1 ES1 ES1 ES1 ESl ES1 ES2 ES2 ES2 ES2 ES2 ES2 ES2 ES2 ES2
1.7 1.3 1.2 1.9 1.3 1.2 1.5 1.1 0.5 0.6 0.4 0.8 2.0 1.0 3.1 1.5 1.1 1.0 1.8 1.6 0.5
25 25 5 >100 1 1 >10 5 D 6 6 N.B. 20 5 1 10 N.B. 10 1 20 >50
2.8 3.5 5.5 8.0 2.0 1.6 9.0 1.6 3.8 2.2 3.9
1.6, 4.4, 6.5 5.5 3.3 4.2 5.5, 6.5, 12, i4
++ ++ + ++ + + + + + ++ ++ + +++ +++ + ++ + +++
+ + + +
2.0
4.5 1.5 1.4 3.2, 5.5
3.9 8.0 1.2 15 1.6 9.0
amounts of target initially and because we used glycogen as a carrier in the ethanol precipitation steps instead of linear polyacrylamide (glycogen minimally inhibits transformation by electroporation, while linear polyacrylamide inhibits by >10-fold44). We analyzed ES2 by southern analysis (data not shown), the marker plasmid assay (Table II), and slot blot analysis (Fig. 5). As in ES1, adult c D N A probe hybridizes much less to ES2-S D N A than to ES2-C D N A . ES2-S D N A shows significant hybridization to the E l 5 cDNA probe suggesting that ES2 contains clones of relatively high abundance that are preferentially expressed in the E15 telencephalon. In the marker plasmid assay, the marker is enriched 10-fold in the ES2-S sample relative to the ES2-C sample. Scanning densitometry of the differential slot blot analysis shows that the adult probe hybridized 10-fold less to ES2-S D N A than to ES2-C DNA. The E l 5 probe hybridized approximately equally to both S and C D N A . In sum, ES2 is enriched approximately 10-fold for genes that are preferentially expressed in the E15 telencephalon.
1.8, 2.5, 5.0
Isolation and characterization of a library of cDNA clones that are preferentially expressed in the embryonic telencephalon Matthew H. Porteus*, A. Elizabeth J. Brice*, Alessandro Bulfone*, Ted B. Usdin**, Roland D. Ciaranello and John L.R. Rubenstein* Nancy Prizker Laboratory of Developmental and Molecular Neurobiology, Department of Psychiatry and Behavioral Sciences, B002, Stanford Medical School, Stanford University, Stanford, CA 94305 (U.S.A.) (Accepted 11 June 1991)
Key words: Telencephalon; cDNA cloning; Development; Subtractive hybridization; Striatum; Cerebral cortex; Differential gene expression
In order to isolate genes involved in development of the mammalian telencephalon we employed an efficient cDNA library procedure. By subtracting an adult mouse telencephalic cDNA library from an embryonic day 15 (El5) mouse telencephalic cDNA library we generated two subtracted libraries (ES1 and ES2). We estimate that ES1 contains between 200 and 600 different cDNA clones, which approximates the number of genes that are preferentially expressed in the El5 telencephalon, compared to the adult telencephalon. Northern analysis of 20 different cDNA clones shows that 14 of these are expressed at least 5-fold more in the El5 telencephalon than the adult telencephalon. Limited sequencing of the 14 differentially expressed clones reveals that 10 have no significant identity to sequences in GenBank and EMBL databases, whereas the other 4 have significant sequence identity to vimentin, historic 3.3, topoisomerase I and the B2 repeat element. In situ hybridization using one of the differentially expressed cDNAs, TES-1, demonstrates that it is transiently expressed in the anlage of the basal ganglia. In situ hybridization with another differ¢ntially expressed cDNA clone, TES-4, shows that it is specifically expressed in differentiating cells of the neural axis with a distinctive rostral-caudal temporal pattern. These findings, and the methods that we have developed, provide a framework for future investigations of the genetic control of telencephalon development.
INTRODUCTION The telencephalon is the rostral end of the neural tube and forms as a pair of evaginations from the forebrain 29' 3~.47. The forebrain differs from the more caudal central nervous system in several respects. The convex shape of the forebrain neural groove is different from the concave shape of the rest of the neural tube 3a. The forebrain neuroectoderm, unlike the more caudal neuroectoderm, has no imderlying notochord, a structure which is important in neural induction 4°. Instead, the forebrain neuroectoderm overlies the prechordal plate, and does not have a floor plate 3. These differences suggest that some aspects of forebrain development may involve mechanisms that are not used in the rest of the nervous system. Early in development, the walls of the entire neural tube are made up of undifferentiated pseudostratilied neuroepithelial cells. A t specific times, cells leave the mitotic cycle and migrate from the ventricular zone to-
wards the pial surface. Depending upon the location of the ventricular zone, the cells contribute to the formation of different structures. For instance, in the telencephalon, cells from the dorsolateral walls form the neocortex, those from the medial walls form the archicortex, and those from the ventral walls form the striatum 29. Each subregion of the telencephalon is organized in a different way. For instance, the mature neocortex is a laminated structure consisting of six cell layers 3°'42 while the mature striatum is a mosaic of patch cells in a sea of matrix cells 17'19'20'51. The cells of the different layers and compartments differ in many ways including their afferent and efferent connections, their repertoire of neurotransmitters and neurotransmitter receptors, and the time in ontogeny at which they become post-mitotic. The molecular mechanisms which regulate how the ventricular zone gives rise to these specialized structures and defined cell types are not known. Genes which have increased expression during particular developmental
* Present address: Department of Psjchiatry, Langley Porter Psychiatric Institute, University of California at San Francisco, 401 Parnassus Ave., San Francisco, CA 94143, U.S.A. ** Present address: Laboratory of Cell Biology, NIMH, Bethesda, MD 20892, u.S.A. Correspondence: J.L.R. Rubenstein. Present address: Department of Psychiatry, Langley Porter Psychiatric Institute, University of California at San Francisco, 401 Parnassus Ave., San Francisco, CA 94143, U.S.A.
events are likely to be involved in these events and should provide molecular information about how these developmental processes are regulated. To date, the characterization of developmentally regulated genes encoding proteins that mediate cell-cell interactions or that regulate transcription have begun to elucidate the molecular mechanisms of insect neural and vertebrate CNS development14.21.25. It is likely that telencephalic development is controlled in large part by the expression of genes whose products perform critical regulatory roles in differentiation. The use of mutations to identify these classes of genes has been successful in invertebrates 2s'37. While several mutants in mammalian CNS development have been described 6 and the use of insertional mutagenesis promises to create more 1'5s these approaches have, as of yet, led to the identification of only a small number of genes important in neurodevelopment 24'3s. Like a genetic screen, nucleic acid subtractive hybridization can be used to focus a search for genes that affect certain developmental events t2'22. This technique has also been used by neurobiologists to identify mammalian brain-specific or neurodevelopmentally regulated genes 9'16'31'43'5°, and in Drosophila melanog~ter it has been used to generate a collection of head-specific cDNA clones 39. In order to systematically identify both rare and abundant m R N A species that are expressed in specific temporal and spatial patterns within the telencephalon, we developed methods for performing subtractive hybridization between directional cDNA libraries in single-stranded phagemid vectors (see Fig. 1 and reference 44). We have now applied these methods to perform subtractive hybridizations between cDNA libraries prepared from embryonic day 15 (El5) and adult mouse telencephalons. By El5 of mouse telencephalic development, the cerebral cortex and stHatum are beginning to differentiate. Cells undergoing their final mitotic division at this time are destined to become the deep layers of the cerebral cortex 4'32 and the patch cells of the striatum 51. Embryonic structures such as the ventricular zone, the subventricular zone, intermediate zone and the subplate are still present 2. Genes which are preferentially expressed in the El5 telencephalon, compared to the adult telencephalon, may play important roles in the function of these transient embryonic structures, and in the differentiation of the newly post-mitotic cells. In this paper we describe the use of subtractive hybridization to construct two cDNA libraries which are highly enriched for cDNAs corresponding to genes that are preferentially expressed in the embryonic telencephalon. From these libraries we have isolated and characterized 14 different c D N A clones corresponding to differentially expressed genes. One cDNA clone is derived from a gene that is expressed
throughout the nervous system in differentiating cells, and one is transiently expressed in the developing striamm. MATERIALS AND METHODS Molecular biology General molecular biology techniques were used as described in refs. 11 and 45, except where noted. Isolation of poly(A +) RNA Timed pregnant Balb C mice were sacrificed by cervical dislocation. The day on which a vaginal plug was found was defined as embryonic day 1 (El). Embryos were dissected in ice-cold Dulbecco's modified Eagle's medium (DMEM). Brains were removed and telencephalic vesicles dissected from the rest of the nervous system tissue. The telencephalons were frozen in liquid nitrogen and stored at -70 °C. Adult Balb C mice (males and females) were sacrificed by decapitation and the telencephalons were removed by dissection. Total RNA was purified using guanidine isothiocyanate and cesium chloride centrifugation. Poly(A+) RNA was purified using one cycle of oligo(dT) cellulose (Collaborative Research Type III). 1.4 g of El5 telencephalon (from 80 embryos) yielded 2.5 mg total RNA and 60/tg of poly(A +) RNA. 10 g of adult telencephalons (from 31 animals) gave 6.0 mg of total RNA and 300 /~g of poly(A +) RNA. Construction of directional cDNA libraries in phagemid vectors eDNA libraries were constructed with inserts in opposite orientations in two phagemid vectors, E61 and E145, using an adaptorprimer method44. Each was transformed into two E. coli strains, MC1061 and XL-1. We used MC1061 because of its high transformation efficiency by electroporation, and XL-1 because it can be used for phagemid rescue. Size-selected libraries were made by cutting phagemid DNA from the MC1061 iiblaries with the restriction endonuclease Nod, and purifying phagemids with inserts greater than 1 kB by low melt agarose gel electrophoresis. The linear phagemids were cyclized by ligation, and transformed into XL-1. See Table I for a list of these libraries. Conversion of cDNA fibraries to single-stranded phagemids Single-stranded phagemid phagemlds' were rescued from thes3 cDNA libraries in XL-1 bacteria using the bacteriophage R408" . Single-stranded DNA was purified using potassium iodide density gradient centrifugation, magnesium-phenol extraction, and restriction endonuclease digestion44. Preparation of biotinylated driver Single-stranded phagemid DNA, which was to be used as driver in the subtractive hybridizations, was biotinylated44. Before using biotinylated driver in subtractive hybridization, it was verified to transform bacteria at >10e lower efficiency than unbiotinylated target and that it could be quantitatively extracted using the streptavidin-phenol/chloroform method. Subtractive hybridization The three subtractions described in the paper were performed using slight modifications of the methods described in Rubenstein et a144. The first subtractive hybridization led to the production of the Embryo Subtracted 1 Library (ES1). The hybridization reaction contained 20 #g of biotinylated driver (anti-sense adult singlestranded cDNA library) and 0.5 ~ug of target (sense El5 singlestranded cDNA library). Two marker DNAs were also included: 5 ng of pACYC184 and "--0.5#g of R408. 2/~g each of three oligonucleotides (A: 5" dGCGGCCGCITITITITIT 3"; B: 5" dATCAAGCTrATCGCCCT 3"; C: dT 19-24) were added to the reaction. These oligonucleotides are complementary to the polylinker and poly(A) domains of the target phagemids, and therefore can prevent d•er-target hybridization in these regions of the phagemids. The DNA was dissolved in lib (40% deionized forma-
mide, 500 mM NaCi, 50 mM Na HEPES, 2 mM Na EDTA, pH 7.6), covered with light-weight mineral oil (Sigma), heated to 90 °C for 2 min, and incubated at 52 °C for 7 days. Biotinylateddriver and driver-targetheteroduplexes were removed using streptavidin and phenol/chloroformextraction~'4s, The targetphagemids leftin solutionwere taken through a second round of subtractivehybridization with 20 l~g of biotinylated anti-senseadult single-stranded c D N A libraryas driver,exactly as in the firstround, After the two rounds of hybridization,we estimate that a CoT of 1,000 mol nucleotides/s/literwas reached. The second subtractive hybridization produced the Adult Subtracted Library (AS). The hybridizationreaction contained I0 ~g of driver (biotinylatedsense adult single-strandedc D N A library) and 0.7/~g of target (anti-sense adult single-strandedc D N A library). It also contained two markers; 2 ng of E28 (a kanamycin resistantplasmid, see Rubenstein et al.44)and -0.3/~g of R408 DNA. The reaction contained 2/~g each of 3 blocking oligonucleotides (D: 5" d A A A A A A A A A A O C G G C C O C 3", E: 5" dCOATAAGCTI'GATI" 3", and F: dA 19-24). The D N A was heated in liB to 90 °C for 2 min and then incubated for 7 days at 52 °C to a CoT of approximately 300 mol nucleotides/s/liter. At this CoT, most of the highly and moderately abundant cDNA clones should have hybridized, thereby enriching for the rarer cDNA clones 52. The third subtractive hybridization produced the Embryo Subtracted 2 Library (ES2). The hybridization reaction contained 17 /~g of biotinylated single-stranded driver (4/zg of anti-seuse adult cDNA library, 9/~g of size-selected anti-sense adult cDNA library, and 4/~g anti-sense adult from the AS cDNA library) and 3/tg of single-stranded target (E15 sense size-selected library). The target also contained two marker plasmids (4 ng of pACYC184 and - 1 /tg of R408) and 2/~g each of blocking oligonucleofides A, B and C. The reaction went through two rounds of subtractive hybridization exactly as described for ES1 to a CoT of approximately 1000 mol nucleotides/s/iiter.
Conversion of single-stranded phagemids to double-strand Single-stranded phagemids were converted to double-stranded form as previously described 44 using oligonucleotide A as the primer. Oligonucleotide A spans the polylinker-cDNA junction of target libraries in the sense orientation. This oligonucleotide was used for ESI and ES2. Oligonucleotide D was used as the primer for AS. The reaction was performed as described in Rubenstein et al. 44. Because these primers span the polylinker-cDNA junction, they select against target phagemids without inserts, target phagemids with inserts in the wrong orientation, and driver phagemids being converted to double-stranded form. The use ,af Taq DNA ; Jlymerase allowed high-temperature annealing of primer to template, thereby increasing the specificity of the single-stranded to double-stranded conversion process. Bacterial transformation and antibiotic resistance avsay The products of the Taq D N A polymerase reaction were introduced into MC1061 by electroporation 15. Det ernnmn " " g the effi ciency of subtraction by antibiotic resistance was done as in Rubenstein et al.44. Colony screening, and Southern, Northern, and slot blot analysis Colony screening, southern analysis, and northern analysis were performed using standard molecular biology procedures 45. cDNA probes were synthesized in two ways. The first used a standard reverse transcription protocol 45 with 1/tg of poly(A +) RNA, and an oligo(dTt5 ) primer to generate a probe of specific activity of 2 x 107 cpm//~g; this was used in the colony screening experiments shown in Fig. 3. The second protocol involved using random primed DNA synthesis kit (Pharmacia) to label 50 ng of double-stranded cDNA to a specific activity of -109 cpm//~g. This cDNA was originally produced for the construction of the cDNA libraries 44. These probes were used for the southern (data not shown) and slot blot analysis in Fig. 5. Probes for the northern analyses (Fig. 4) were
generated by random-primed labelling of agarose-gel purified cDNA inserts from XhoI-HindIII digested phagemlds. The probes use~ ,:~ backscreen ES1 to determine the number of different cDNA clones were produced in the following manner. Inserts from the cDNA clones were cut out from the vector by HindIII-Xhol digestion and purified by agarose gel electrophoresis. Oligonucleotide A, which is specific to the polylinker-cDNA junction, was used ~,s primer for the DNA polymerase. It was annealed to the cDNA inserts in 50 mM Tris-HCl pH 7.5, 10 mM MgCI2, 1 mM dithiothreitol, 300/~M dATP, dGTP, and d'ITP, by heating to 95 °C for 3 min and then slowly cooling to 37 °C. 50/tCi[32p]dCTP (3000 Ci/mM) and Klenow fragment (2.5 units) were then added, and the reaction was incubated for 45 min at 37 °C. Probes of a specific activity of 8 x l0 s cpm//~g were generated. We found that labelling with a specific primer dramatically reduced background hybridization in colony screening compared to random priming, probably by reducing the labelDng of trace amounts of vector sequences that contaminate the purified cDNA inserts. Quantitation of hybridization was done by scanning the autoradiograms using a LKB scanning densitometer.
DNA sequencing DNA sequencing was performed using the Sequenase 2.0 kit (U.S. Biochemical) on double-stranded template using T3 or T7 oligonucleotide primers according to manufacturer's specifications. Analysis of the DNA sequences were performed using the Genetics Computer Group sequence analysis software, and the GenBank and EMBL databases. In situ hybridization Frozen sections of whole embryos or brains were cut and desiccated according to Herkenham and Pert 23 and fixed, hybridized, and washed according to Mueller et al.35. In brief, tissue was frozen in 2-methyl-butane on dry ice, cut in 12/~m sections at -20 °C, mounted on chrom-alum coated slides, kept on dry ice, desiccated overnight at -2 °C, and stored at -70 °C. Sections were fixed in DEPC (diethylpyrocarbonate)-treated PBS (150 mM NaCl, 10 mM sodium phosphate pH 7.4) containing 4% paraformaldehyde (fresh) for 20 min, rinsed 3 times in PBS for 5 min, dehydrated through graded ethanol, and air-dried. Sections were then incubated in 1 /~g/ml proteinase K in PKB (100 mM Tris-HCI pH 8.0, 50 mM EDTA) at 37 °C for 30 min and post-fixed as above. Sections were acetylated by rinsing in 0.1 M triethanolamine, blotting dry, and incubating in 200 ml of 0.1 M triethanolamine containing 500/~l of acetic anhydride for 10 min at 20 °C. They were then rinsed in DEPC-treated water, air-dried, and stored at 4 °C. Hybridization was as follows: 106 cpm of riboprobe was heated in 40/A of ISHB (50% deionized formamide, 5 x SSC (1 x SSC = 0.015 M sodium citrate, 0.15 M NaCI pH 7.0), 10 mM/~-mercaptoethanol (BME), 10% dextran sulfate, 2 x Denl'lardt's, 250/~g/ml yeast tRNA, 500 /~g/ml single-stranded sh~ared herring sperm DNA) to 65 °C for 5 min, and cooled to 20 °C. 40 ltl of ISHB-riboprobe was layered on top of each tissue section. The samples were then covered with a parafilm coverslip, and incubated in a humidified chamber at 42 °C for 18 h. The slides were twice washed for 30 rain at 20 °C in 2 x SSC; 10 raM BME; then in 2 x SSC, 1 mM EDTA containing 20 /~g/ml RNase A and 1 unit/aft RNase T1 (both from Pharmacia) at 20 °C for 40 min; then twice in 50% formamide, 2 x SSC, 1 mM EDTA, 10 mM BME at :.~0°C for 30 min; then briefly washed at 20 °C in 0.2 x SSC, and dehydrated in 75% and 95% ethanol. The sections were autoradiographed for 1-2 days at 20 °C using Hyperfilm-/~lax (Amersham). Finally the sections were dipped in NTB2 emulsion (Kodak) and exposed for 6 days. The emulsion was developed using Kodak D-19, and the tissue was counterstained with Cresyl violet. The samples were examined using a Nikon Optiphot-2 microscope, and photographed using Ektachrome ASA 200 film. The anti-sense cY,TdqA probe for in situ hybridization was made according to manufacturers specifications using the Stratagene riboprobe kit. Briefly, 1 ~g of proteinase K-treated, HindIII-digested
10 TABLE I
Catalogue of cDNA libraries The characteristics of each cDNA library used in this paper are summarized in the table. The first column lists the names of each library. The second column states which vector was used in the given library. E61 and E145 are phagemid vectors based on pBluescript described in Rubenstein et al."u. The third column lists the source of the poly(A +) RNA from which the eDNA was synthesized. The fourth column lists the orientation of the cDNA inserts in the phagemid vector; in the sense orientation the sense strand of the cDNA is recovered when performing the M13 rescue procedure. The fifth column lists the bacterial strain that the fibrary was made in. The sixth column lists the number of independent recombinants in each library. The final column lists the average insert size in the entire eDNA library, which was estimated from the size distribution of the cDNA inserts.
Name
Vector
Tissue source
Orientation
Bacterial strain
Independent recombinants
Averageinsert size (kb)
E159 El60 E162 E163 El70 El71 E173 E174 E212 E213 ES1 ES2 AS
E61 E61 E145 E145 E61 E61 E145 E145 E145 E61 E61 E61 E145
Adult Telen. Adult Telen. Adult Telen. Adult Telen. E15 Telen. El5 Telen. El5 Telen. E15 Telen. Adult Telen. El5 Telen. E15 Telen. El5 Telen. Adult Telen.
Sense Sense Anti-sense Anti-sense Sense Sense Anti-sense Anti-sense Anti-sense Sense Sense Sense Anti-sense
XL-1 MC-1061 XL-1 MC-106I XL-1 MC-1061 XL-1 MC-1061 XL-1 XL-1 MC-1061 MC-1061 XL-1
2x 4x 2x 4× 1.6 x 8x 1.6 x 8x 1x 3x 2.5 X 5x 5×
1 1 1 1 1 1 1 1 1.5 1.5 0.6 1 0.6
phagemid was incubated with 10 units of T3 RNA polymerase, 50 /~Ci[35S]rUTP (1270-Ci/mM), in 40 mM Tris-HCl pH 8.0, 8 mM MgCI2, 2 mM spermidine, 50 mM NaCI, 30 mM DTI', 400/~M rATE rCTP, rGTP, 25 units RNase Block at 37 °C for 30 min. The sense cRNA probe was made using Nod-digested DNA template and T7 RNA polymerase. RESULTS
Construction of adult and El5 telencephalic directional cDNA libraries We made directional cDNA libraries using an adaptor-primer procedure in a modified pBluescript vector from poly(A +) R N A from El5 and adult telencephalons as previously described 44. Libraries were made from both tissues with cDNA inserts in both orientations and were amplified in XL-1 Blue and MC-1061 bacteria. The original libraries had average insert sizes of about 1 kilobase (kb). From these libraries we made size-selected libraries with inserts > 1 kb. The characteristics of each library are summarized in Table I. From each of the XL-1 libraries, single-stranded phagemids were prepared by R408 helper phage rescue s3. Anti-sense libraries generated single-stranded anti-sense phagemids and sense libraries generated single-stranded sense phagemids. After helper phage rescue, the. average insert size of the libraries decreased to approximately 0.6 kb. The decrease is probably the result of smaller molecules converting to single-stranded phagemids more effÉciently than larger molecules since we have found no evidence of deletions.
106 107 106 10 7
106 106 106 106 106 105 10 3
104 104
Subtractive hybridization between adult and El5 libraries: production o f ES1 To isolate clones preferentially expressed in the E l 5 telencephalon, we used cDNA library subtraction. The general method is shown in Fig. 1 and is described in detail in Rubenstein et al. 44. All of the subtractions desc~'ibed in this paper consist of two parallel reactions: subtracted (S) and control (C). The S reaction contains both driver and target D N A , while the C reaction contains only target D N A . The abbreviations S and C will stand for the 'subtracted' and 'control' reactions throughout the paper. In the first subtraction, the driver was the biotinylated adult single-stranded c D N A library with inserts in the anti-sense orientation, and the target was the E l 5 single-stranded c D N A library with inserts in the sense orientation. The target contained two marker D N A molecules to measure the efficiency of the subtraction: 6.4 kb single-stranded R408 helper phage and the 4.2 kb double-strando, d tetracycline resistant plasmid, pACYC184. The target was processed through two rounds of subtractire hybridization to a final CoT of about 1,000. We chose a CoT of 1000 mol nucleotides/s/liter based upon the hybridization results of Van Ness and Halm 52 who showed that under these conditions >90% of infrequent copy brain R N A would hybridize in a R N A - c D N A reaction. After the subtractive hybridization, we converted the remaining single-stranded target molecule.~ to doublestranded form (double-stranded phagemids transform bacteria 300- to 1000-fold more efficiently than singlestranded phagemids 44) using Thermus aquaticus (Taq)
11 D N A polymerase. The products of the Taq D N A polymerase reaction were transformed into bacteria and thereby generated the 'Embryo Subtracted 1 Library' (ES1) (Table I). ES1 contains 2500 independent recombinants with an average insert size of 600-700 bp. Of 60 clones analyzed, all contained inserts, suggesting that most or all of the 2500 independent recombinants contain inserts. The control library had a complexity of 25,000 independent recombinants and had an average insert size of 600-700 bp. B
'-0o0o0! @ B
~ Hylxidbudlon Reection B
• B OrganicP h a s e
Go @ o I
S~d,n
I Phenol:CHCIs ~ ~ , , Extraction
AqueousPhase
B
O)! (E) O' Convert to
Double Stranded
Trsnsform F..¢oli Subtracted Ubrary
Fig. 1. Library subtraction scheme. To isolate genes, such as Gene Z, from t,e E15 target fibrary that are preferentially or exclusively expressed during development, we used the cDNA fibrary subtraction scheme depicted. Biotinylated (B) single-stranded driver phagemids are mixed with single-stranded target phagemids and two marker DNA molecules: single-stranded R408 and doublestranded pACYC184. Adult (A) cDNA clones are cloned in one orientation, depicted by an up arrow, and El5 (E) cDNA clones are cloned in the opposite orientation, depicted by a down arrow. cDNAs in the target fibrary, such as actin, will hybridize to their complements in the driver library to form a heteroduplex, cDNAs that have no complement will remain single-stranded. After hybridization all driver molecules and heteroduplexes are removed from aqueous solution by adding streptavidin which specifically binds biotin and then extracting with phenol/chloroform. Target molecules which have not hybridized to the driver DNA, sl~ch as Gene Z, as well as the marker molecules R408 and pACYC184, remain in aqueous solution. The unsubtracted target molecules are converted to double-stranded form using a strand-specific primer and Taq DNA polymerase. The product of the Taq DNA polymerase reaction is transformed into E. coli by electroporation and a subtracted library of preferentially expressed cDNA clones is thereby created.
Analysis of ES1 We analyzed the efficiency and specificity of the ES1 subtraction in 5 ways: agarose gel electrophoresis, southern analysis, a marker plasmid assay, a colony screen assay, and slot blot analysis. Comparing ethidium bromide fluorescence of the single-stranded DNAs from the ES1-S and ES1-C reactions separated by agarose gel electrophoresis showed that most of the phagemids had been extracted from the ES1-S reaction (Fig. 2). The R408 D N A , which should not be extracted in this procedure, was only slightly reduced ~n the ES1-S reaction. We attribute the small decrease to non-specific trapping of R408 D N A during the extraction of biotinylated driver-target duplexes. Thus, while the R408 band is only slightly reduced, the amount of target phagemids is greatly reduced, suggesting that ES1 is enriched for sequences that are preferentially expressed in the E l 5 telencephalon. This result qualitatively demonstrated that the subtraction worked and we proceeded to analyze ES1 further. Second, we analyzed ES1-S and ES1-C by southern analysis (data not shown). Qualitatively, radioactive adult telencephalic c D N A hybridizes much less to ES1-S D N A than to ES1-C D N A . Therefore, abundant sequences found in the adult telencephalon have been removed from ES1. ES1-S D N A hybridizes more to the probe from E l 5 c D N A than to the adult probe. Thus, ES1 contains cDNA clones that hybridize more to a probe derived from E l 5 telencephalic poly(A +) R N A than from a probe derived from adult telencephalic poly(A +) RNA. Third, to determine the efficiency of the subtraction, we measured the ratio of the pACYC184 marker plasmid (tetracycline resistant colonies) to target molecules (ampicillin resistant colonies) between ES1-S and ES1-C samples; this is the marker plasmid assay 44. After sub-
TARGET: DRIVER:
R408
EMBRYO (x A D U L T
e ADULT ADULT C S
[[
Fig. 2. Analysis of ES1 and AS by agarose gel electrophoresis. After one round of subtractive hybridization, one-third of the control (C) and subtraction (S) reactions from ES1 (target: sense embryo; driver: anti-sense (a)-adult) and AS (target: anti-sense (a)-adult; driver: sense-adult) were electrophoresed on an agarose gel and stained with ethidium bromide. The location of R408, a single strand marker molecule is shown by an arrow. Target molecules that remain in solution after the subtractive hybridization are the fluorescent smear below the R408 band.
12 traction the marker plasmid was enriched 70-fold in the ES1-S sample relative to the ES1-C sample (Table II). As another indicator of specificity, the number of marker tetracycline colonies remained constant between the ES1-S and ES1-C reactions (Table II). Fourth, we screened colonies from the S and C reactions of ES1 with adult telencephalic cDNA; this is the colony screen assay. The cDNA probe hybridized to 60 out of 270 colonies from the ES1-C sample but only 2 out of 240 colonies from the ES1-S sample (Fig~ 3). These results suggest that ES1 is enriched by approximately 30-fold for genes that are preferentially expressed in the El5 telencephalon. Fifth, we measured the efficiency of the subtraction by screening slot blots containing ES1-S and ES1-C D N A with adult or E15 telencephalic cDNA; this is the slot blot analysis (Fig. 5 and Table II). Scanning densitometry shows that the amount of hybridization to the ES1-S sample is approximately 40-fold less than to the ES1-C sample with the adult cDNA probe. Based on the marker plasmid assay, the colony screen assay, and the slot blot analysis we estimate ES1 to be enriched for genes preferentially expressed in the E15 telencephalon by 30- to 70-fold. ES1 contains 2500 independent recombinants. To estimate how many of these clones represented different cDNAs, we made radioactive probes from cDNA inserts from eleven random clones and screened ES1 colonies. One of the clones, TES-4, is represented approximately
100 times. The other ten cones are represented on average 4 times. Thus, if we exclude the TES-4 clones from the calculation, there are approximately 600 different clones in ES1 (2400/4), and if we include T E S - 4 there are approximately 200 different cDNA clones in ES1 (2500/13). Subtractive hybridization between sense adult and antisense adult libraries: production o f A S
Rare cDNAs are less efficiently subtracted in any subtractive hybridization because of kinetic constraints 52. In order to increase the probability of subtracting shared but rare cDNAs, we created a library of less abundant adult cDNAs by subtractive hybridization to use as supplemental driver in subsequent subtractive hybridizations. In this subtraction, the driver was the singlestranded sense adult c D N A library and the target was the single-stranded anti-sense adult c D N A library. The complexity of this subtracted library, the 'anti-sense Adult Subtracted Library' (AS), was 45,000 independent recombinants with an average insert ~ize of 600-700 bp. We analyzed AS by agarose gel elecuophoresis (Fig. 2), the marker plasmid assay (Table II), and the colony screen assay (Fig. 3). Agarose gel electrophoresis shows that the amount of single-stranded c D N A phagemid had decreased in the AS-S sample compared to the AS-C sample but the R408 marker for specificity remained approximately constant. The marker plasmid assay estimates that this subtraction enriched for less abundant
TABLE II Efficiency of subtractive hybridizations
The table summarizes the analyses of the 3 different subtractive hybridizations described in the paper: ES1, ES2, and AS. S refers to the subtracted sample and C to the control sample from each subtractive hybridization. CoT is the molar concentration of driver nucleotides (Co), multiplied by the length of hybridization in seconds (T). Total Amp R Colonies refers to how many ampiciilin resistant colonies (target phagemids) were recovered after subtractive hybridization. Analogously, Total TetR Colonies refers to the number of tetracyclineresistant marker colonies (pACYC184 plasmid) recovered after subtractive hybridization. In the case of AS, the marker colonies were kanamycin resistant (E28) and not tetracycline resistant. The AmpRPl'etR ratio is determined by dividing the total number of ampiciUin resistant colonies by the total number of tetracyctine resistant colonies. Transformation refers to the estimate of subtractive hybridization efficiency using the bacterial transformation assay and was determined by dividing the AmpR/TetR ratio from the C sample by the AmpR/'I'etR ratio from the S sample. Colony screen refers to the estimate of subtraction efficiency based on the colony screen assay (see text) and Slot blot refers to the estimate of subtraction efficiency based on the slot blot analysis (see text). Subtraction
CoT
ES1 -S -C
1000
ES2 -S -C
1000
AS
350
-S
-C
Total Amp R colonies
Total TetR colonies
2.4 x 103 1.9 x 105
1.7 x 104 1.9 x 104
0.14 10
5.3 x 104 7.2 x 105
1.2 x 104 1.6 x !04
4.4 45
104 1.1 × 106
7.5 × 103 5.7 x 103
6
4.5 ×
AmpRITetR
193
Transformarion
Colony screen
Slot blot
70
30
40
10
-
10
30
4
13
SUBTRACTED & •
CONTROL ES1 LIBRARY
ES1 LIBRARY o
& &
e
e
~
e g
B
7
O
A
~
IY
Fig. 3. Analysis of ESI and AS by colony hybridization to adult telencephafic cDNA probe. 240 colonies from ES1-S, 270 colonies from ES1-C, 100 colonies from AS-S, and 240 colonies from AS-C were screened with an adult telencephalic cDNA probe. Each dark dot represents a bacterial colony which harbors a cDNA phagemid that hybridizes with the probe. The triangles depict orientation markers on each filter. The filters were hybridized for 3 days in 20 ml of hybridization mix, and were washed twice at room temperature with 2 x SSC and twice at 65 °C for 30 rain in 0.1 x SSC and 0.1% SDS. The autoradiograms were exposed for 20 h.
adult cDNAs by 30-fold. The colony screen assay showed that the adult telencephalic probe hybridized to 110 colonies out of 270 from the AS-C sample but only to 10 out of 100 from the AS-S sample in the colony screen assay; this result suggests a 4-fold estimate for enrichment. Thus, we estimate AS to be enriched for less
abundant cDNA clones by 4- to 30-fold.
Subtractive hybridization of size-selected El5 target using driver that contains the adult subtracted library: production of ES2 The cDNA inserts in ES1 were relatively small. To
14 well as the non-size-selected anti-sense adult c D N A to ensure that any t a r g e t inserts < 1 kb w o u l d be sub-
TES-1 TES-2 TES-3 TES-4 TES:6 E
A
E
A
E
A
E
A
E
A
tracted. This subtraction g e n e r a t e d the ' E m b r y o Subtracted 2 Library' (ES2) (Table II). ES2 contains 50,000 independent r e c o m b i n a n t s with an average insert size of
~i!:i+:.~;.~+++
1 kb. We obtained a larger n u m b e r of i n d e p e n d e n t recombinants in ES2 t h a n ES1 because we used larger
28S -- I
~
--- 28S
18S
- - 18S
A
!:+:i+~:i:i:
cyclo-
III ill
,.i,,
..
Fig. 4. Northern analysis. Inserts from TES-1, TES-2, TES-3, TES-4, and TES-6 were isolated "y NotI-HindIII digestion, labelled by random priming, and used to probe northern blots consisting of 10/~g of total RNA from E15 and adult telencephalons. In each blot, E marks the lanes with E15 RNA and A marks the lanes with adult RNA. The migration of the 28S and 18S ribosomal markers are marked with arrows. The left arrows mark the ribosomal RNA migration for the three blots on the left, while the fight arrows mark the ribosomal RNA migration for the two blots on the right. The blots were hybridized for 48 h in 10 ml of hybridization mix, and washed twice at room temperature for ten min in 2 x SSC and 0.1% SDS, and twice at 60 °C for 30 min in 0.1 x SSC and 0.1% SDS. The TES-1 blot was au.~oradiographed for 12 h, the TES-2, TES-4, and TES-6 blots were autoradiographed for 18 h, and the TES-3 blot was autoradiographed for 7 days. No hybridization could be detected in the adult lane with TES-4 even after a 7 day exposure. To establish that equal amounts of undegraded RNA was loaded onto each lane, the same blots were later probed with cyclophilin (cyclo), a non-developmentally regulated gene z~.
I
1Q
I
1"
II
2*
II
2*
III
S
III
N
3*
IV
3"
11
7
11
7
12
8
12
8
13
4*
13
4•
14
9
14
9
S
6*
S
6"
C
10
C
10
S'
Cyclo
S'
Cyclo
C'
pBS
C'
pB$
E
create a subtracted library with larger inserts we used the anti-sense adult and sense E15 c D N A libraries with inserts > 1 kb as the driver and target respectively. We also added A S c D N A to the driver in o r d e r to increase the efficiency of subtracting less a b u n d a n t E15 c D N A s , as
Fig. 5. Slot blot analysis of ES1, ES2 and individual clones. 1 ~g of DNA from ES1-S (S), ES1-C (C), ES2-S (S') and ES2-C (C') and from individual cDNA clones was blotted onto duplicate Duralon membranes either by slot blot (the lower two panels), or by oval blot (the upper two panels). One blot was probed with a random primed El5 telencephalic cDNA probe and the other was probed with a random primed adult telencephalic cDNA probe. The blots were hybridized for five days in 10 ml of hybridization mix, and washed using the conditions described in the legend to Fig. 4. The blots were autoradiographed for 2 days. Clones from ES1 have Arabic letter names, whereas clones from ES2 have Roman numeral names. More information regarding many of the clones is given in Table lII. Cyclo is the abbreviation for the cDNA clone encoding cyclophilin and pBS is the abbreviation for the phagemid vector (the E61 form of pBS) used for all of the TES clones. The cDNA clones marked by an asterisk are those that are also shown in Fig. 4.
~.~
D
5
A
Q
I t
I
2*
V
3*
6*
3'
6"
Xll
Vl
Xll
Vl
Xlll
VII
XlII
VII
VIII
XlV
VIII
IX
XV
IX
XVl
X
XVl
X
XVll
Xl
XVll
Xi
16
!5
16
15
4*
Cyclo
41
Cyclo
18
pBS
18
pBS
XIV XV
41 .~
15 TABLE III
Summary of TES clones We have performed northern analysis on 20 eDNA clones encoding different mRNA tra nscri_'pts. Tb_e characteristics of each clone are summarized in this table. Clones from ES1 are given Arabic numerals and clones from ES2 are given roman numerals. Subtraction (SUB) refers to the subtracted library from which the cDNA clone was isolated. INSERT is the size in kb of the cDNA insert. E/A refers to the level of differential expression based on the northern analysis, in which the ratio of hybridization to the major mRNA species is compared between the El5 and the adult RNAs. D refers to a northern in which there is clear evidence that the transcript is differentially ~xpressed, but the background was too high to carefully quantitate. N.B. means that no bands were seen on the northern. Major is the size of the most abundant RNA transcript in the E15 telencephalon and Minor is the size(s) of the less abundant RNA transcripts. The scale for the relative abundance (Abund.) is as follows: + - 0.5%. cDNA clones that have been partially sequenced (Sequence) are marked. TES-11 has a partial B2 sequence, TES-13 is homologous to the 3" end of a human topoisomerase 1 cDNA, TES-V encodes the vimentin coding sequence, and TES-XIV encodes the histone 3.3 coding sequence, cDNA clones which have been examined by in situ hybridization are marked in the in situ column.
Name
Sub
Insert
E/A
Major
Minor
Abund.
Sequence
TES-1 "FES-2 TES-3 'IT.S-4 TES-6 TES-7 TES-8 TES-10 TES-11 TES-12 TES-13 TES-14 TES-V TES-VI TES-X TES-XII TES-XIII TES-XIV TES-XV TES-XVI TES-XVII
ES1 ES1 ES1 ES1 ES1 ES1 ES1 ES1 ES1 ES1 ESl ES1 ES2 ES2 ES2 ES2 ES2 ES2 ES2 ES2 ES2
1.7 1.3 1.2 1.9 1.3 1.2 1.5 1.1 0.5 0.6 0.4 0.8 2.0 1.0 3.1 1.5 1.1 1.0 1.8 1.6 0.5
25 25 5 >100 1 1 >10 5 D 6 6 N.B. 20 5 1 10 N.B. 10 1 20 >50
2.8 3.5 5.5 8.0 2.0 1.6 9.0 1.6 3.8 2.2 3.9
1.6, 4.4, 6.5 5.5 3.3 4.2 5.5, 6.5, 12, i4
++ ++ + ++ + + + + + ++ ++ + +++ +++ + ++ + +++
+ + + +
2.0
4.5 1.5 1.4 3.2, 5.5
3.9 8.0 1.2 15 1.6 9.0
amounts of target initially and because we used glycogen as a carrier in the ethanol precipitation steps instead of linear polyacrylamide (glycogen minimally inhibits transformation by electroporation, while linear polyacrylamide inhibits by >10-fold44). We analyzed ES2 by southern analysis (data not shown), the marker plasmid assay (Table II), and slot blot analysis (Fig. 5). As in ES1, adult c D N A probe hybridizes much less to ES2-S D N A than to ES2-C D N A . ES2-S D N A shows significant hybridization to the E l 5 cDNA probe suggesting that ES2 contains clones of relatively high abundance that are preferentially expressed in the E15 telencephalon. In the marker plasmid assay, the marker is enriched 10-fold in the ES2-S sample relative to the ES2-C sample. Scanning densitometry of the differential slot blot analysis shows that the adult probe hybridized 10-fold less to ES2-S D N A than to ES2-C DNA. The E l 5 probe hybridized approximately equally to both S and C D N A . In sum, ES2 is enriched approximately 10-fold for genes that are preferentially expressed in the E15 telencephalon.
1.8, 2.5, 5.0
