Recombinant selection by microinjection: a simple cDNA cloning procedure for production of exclusively sense RNA transcripts

147

Gene. 83 (1989) 147-152 Elsevier GENE

03152

Short Communications Recombinant selection by microinjection: exclusively sense RNA transcripts (Recombinant

DNA;

transient

a simple cDNA cloning procedure for production of

complementation;

library partition;

adenine

phosphoribosyltransferase)

Martin Digweeda and Ursula Ginthert b u Institut ftir Humangenetik. Freie Universitiit Berlin, D-1000 Berlin 19 (F.R.G.); b Kernforschungszentrum Karlsruhe, Institut ftir Genetik und Toxikologie von Spaltstoffen, and Institut ftir Genetik der Universitiit Karlsruhe, D-7500 Karlsruhe 1 (F.R.G.) Tel. (07247)823293 Received

by T.A. Bickle: 22 March

Accepted:

1989

24 April 1989

SUMMARY

A new strategy for cDNA cloning is presented, designed particularly for identification of recombinants by functional analysis, after microinjection into somatic cells. First-strand synthesis is primed by the oligodeoxyribonucleotide:

After second-strand synthesis and blunting, double-stranded cDNA is formed, which carries restriction sites for Not1 and ApaI downstream from the coding sequence. The cDNA is ligated into a plasmid, between two promoters

for phage T7 and T3 RNA polymerases.

Following

transfection

and amplification

in Escherichiu

coli,

plasmids are extracted from the library or sublibraries. Linearisation with NotI, prior to in vitro transcription, cleaves the plasmids between the 3’-end of the coding sequence and the adjacent promoter and thus ensures that only sense RNAs, suitable for microinjection, are produced after addition of the RNA polymerases. Use of NotI, a rare cutter in the human genome, should ensure that the cDNA inserts are not damaged during linearisation. In the unlikely event that this does happen, a site for Apa1 is also available. The method is demonstrated for the human adenine phosphoribosyltransferase-encoding gene.

Correspondenceto: Dr. M. Digweed, Heubner

Institut

fur Humangenetik,

nt, nucleotide(s);

Weg 6, D-1000 Berlin 19 (F.R.G.)

Tel. (030)3203464;

phate-buffered

Fax (030) 3035 2900.

140 mM SDS,

Abbreviations:

APRT, adenine

gene encoding transferase;

APRT;

phosphoribosyltransferase;

HPRT,

Ig, immunoglobulin;

037X-11 19/89/$03.50

0 1989 Elsevier

hypoxanthine

uprt,

phosphoribosyl-

NaCl/200 kinase;

IVT, in vitro RNA transcript;

Science Publishers

B.V. (Biomedical

Division)

oligo, oligodeoxyribonucleotide; saline

NaCl/2.7 sodium

(8 mM

mM dodecyl

mM NaH,PO, tk, gene encoding

Na,HP0,/1.4

KCljO.5 mM sulfate;

PBS, phosmM

MgClJ0.9 SSPE

KH,PO,/

mM

CaCl,);

(20 x ),

pH 7.4/20 mM EDTA; Tk; TCA, trichloroacetic

3.6 M

Tk, thymidine acid.

148

produced

INTRODUCTION

Microinjection of poly(A) + RNA fractions into Tk - and HPRT - somatic cells, followed by appropriate functional mRNA

assays, led to the determination

by in vitro transcription.

the approach using APRT-

We describe here

we have developed and the results human fibroblasts as a model system.

of

size for these genes (Lin et al., 1982) before

they were

successfully

cloned.

length has been determined fibroblasts

from patients

Xeroderma

pigmentosum

Such experiments

Similarly,

mRNA

EXPERIMENTAL

AND

DISCUSSION

for the genes deficient in with the genetic (Legerski

et al.,

not only yield information

disease 1984).

(a) Microinjection

to

fraction containing

the mRNA for APRT

identify

a

poly(A)+RNA

on the

genes and their products, but also represent selection systems and starting points for the cloning of these

Poly(A)’ standard

RNA was isolated techniques

(Maniatis

from HeLa cells by et al.,

1982) and

genes. RNA transcribed in vitro from cDNA libraries, constructed in vectors with phage RNA polymerase promoters and injected into Xenopus oocytes, has led, via library partition, to isolation of cDNAs for murine IgGl induction factor (Noma et al., 1986), bovine substance K (Masu et al., 1987) and a Na+ /glucose cotransporter (Hediger et al., 1987). This strategy is obviously also attractive for identifying genes whose defects lead to mutant somatic cells with ciear phenotypic characteristics. However, so far, efforts have apparently not successfully passed the cDNA cloning barrier. The problems encountered may be related to difficulties in obtaining full-length cDNAs and the presence of antisense orientation of cloned cDNAs. Antisense RNA has been shown to inhibit translation of many cellular proteins, including Tk and actin (Izant and Weintraub, 1985). Thus functional analysis via microinjection requires oriented cDNA libraries,

fractionated by centrifugation through 15 % to 30 % sucrose gradients. Fibroblasts from a patient with APRT-deficiency (Doppler et al., 1981) were cultivated on coverslips, which were etched with a grid of 3 x 3 square fields, each field containing 50-100 cells. APRT, a purine salvage-enzyme, produces AMP from adenine and phosphoribosyl pyrophosphate: uprt fibroblasts cannot incorporate exogenous adenine. Injection with poly(A) + RNA transiently complemented the aprt- fibroblasts and allowed incorporation of [ 3H]adenine, which was then detected by autoradiography. Injection of poly(A) + RNA gradient fractions localised the complementing RNA to a fraction of 1000 (_+ 100) base length, as shown in Fig. 1. This agrees well with the size of human APRT mRNA on Northern blots (Hidaka et al., 1987).

since antisense RNAs may negate the effect of the particular mRNA sought; moreover, they will almost certainly compromise cellular performance in general. Several methods have been described for the construction of oriented cDNA libraries utilizing first-strand synthesis plasmid-primed either (Okayama and Berg, 1982) or linkers, to ensure correct insertion of cDNA (Palazzolo and Meye1987). We present here an alternative rowitz, approach, which is not dependent upon complex vector construction, avoids restriction digestion of the cDNA before completion of cloning and uses the simple one-tube blunt-ended cDNA synthesis procedure of Gubler (1988). Although the method does not yield oriented libraries as such, the use of a specific primer ensures that only sense RNAs can be

The poly(A)‘RNA fraction, enriched for the complementing mRNA, was used to construct a cDNA library using the strategy shown in Fig. 2. A portion of the RNA fraction (8 pg) was annealed to the ApaI-N&I-T,, primer, and first- and secondstrand cDNAs were synthesised using the protocol of Gubler (1988), except that Moloney murine leukemia virus reverse transcriptase was used with appropriate changes to the incubation temperature and buffer conditions. First-strand yield was 34% of the input RNA and alkaline gel analysis showed products with a length of 700-1000 nt (not shown). Double-stranded cDNA was blunted and ligated into SmaI-digested, dephosphorylated pT7T3-U19 plasmid DNA (Pharmacia, Sweden). After plating onto a GeneScreenPlus filter (NEN, F.R.G.) a

(h) cDNA cloning

149

OiijTTTTTBP /WI

/@?I

1

2

cT3

2 80.00 u" D

2

E

60.00

%

E 6 R

!IIL

18s V

28s

V

E

5s

40.00

m

0.00 2

6

10 14 mRNA fraction

Fig. 1. [3H]Adenine

incorporation

microinjected

HeLa

with

was carried

medical Instruments, tal capillary

and

Fibroblasts

by aprt-

poly(A) + RNA

U.K.) constricted

puller (Bachofer,

manipulators

18 number

22

F.R.G.).

ECET

were grown

transferred

[aH]adenine,

for 3 h further

extensively

with PBS,

(Eppendorf,

were

incubation.

at 4°C before developing.

medium

--AA.

incorporation poly(A)+ cells;

(panel

ly(A)‘RNA

of [3H]adenine.

RNA fraction

(Panel

of

acid (3: l),

sucrose

slides.

Slides

the were

for four days

above the cells indicate

A) Cells injected

with a

of 1000 ( f 100) nt; (panel B) uninjected

C) complementation gradient,

28S, 18s and 5S sedimented,

profile The fractions

.?

TT

the uprtr ceils

and exposed

Silver grains

‘w--T3 _

The cells were washed

on microscope

Sense-RNA

micro-

1.5 MBq/ml

fixed with methanol/acetic

mounted

P

F.R.G.).

essential

containing

with Ilford K2 film emulsion

l-7-

was performed

washed with 5% TCA and then with 80% and 100% ethanol; coverslips

with T7 and T3 RNA polymerases

ST:=

Electro-

with motor-driven

injector

medium

cT3

/

In vitro transcription

Microin-

(Clark

Injection

in Eagle’s minimal

to

fibroblasts

human

with 10% foetal calf serum. 20 h after injection,

coated

I

!

to 0.5 nm with a horizon-

microscope

the

4

26

fractions.

out using glass capillaries

under the Leitz Fluovert

were

A&

T7*

20.00

jection

Dig&with

3

V

for

a

HeLa

in which

in a parallel gradient,

poHeLa

Fig. 2. Schematic was annealed cDNAs

ofthe

cDNA cloning strategy.

were synthesized

and blunted. The cDNA (stippled box)

is ligated between two promoters vector plasmids

(open

Poly(A) + RNA

to the oligo primer (1)and first- and second-strand (blackened

box) (2) and libraries

are linearised

cDNAs,

so that

adjacent

promoter

orientations

whilst remaining

(4) for microinjection

boxes) of a plasmid as usual.

at the Not1 site provided

antisense

tation, to the remaining

formed

promoter

are removed

attached,

Isolated

by the inserted from the

in the sense-orien-

(3). Capped in vitro transcripts

are generated

as described

by Konarska

et al. (1984).

are indicated.

library of 58 000 independent recombinants in E. coli DHSsl was obtained. A replica filter of the library was made, at which point the filters were cut into nine equal-sized portions. The bacteria were washed off the nine fragments of the replica filter and frozen as glycerol stocks. Mini-prep analysis indicated that all recombinants had incorporated inserts with a Not1 site; of these, 52% also had an intact Apa1 site. Possibly second-strand synthesis did not extend past

the Not1 sequence in some recombinants or blunting with T4 DNA polymerase led to loss of the Apa1 site. (c) Transient complementation of in vitro transcripts

after microinjection

To examine complementation after injection of in vitro transcripts, capped RNA was transcribed from pT7T3 plasmids carrying cDNAs encoding human

Fig. 3. [3H]Adenine incorporation by aprt- human fibroblasts after microinjection with IVTs. Microinjection and labelling were as detailed in Fig. 1. An Xhoi fragment of APRT-OBS (Schneider, 1988) containing the entire APRT-encoding region and 40 bp and 300 bp of 5’ and 3’ flanking sequences, respectively, was ligated into the BumHI site of pT7T3-U19. The fk cDNA from pTKl1 (Bradshaw and Deininger, 1984), was recloned into the same vector. Capped transcripts (aprt-IVTs, tk-IVTs) were produced from these plasmids and dissolved at 1 mg/ml for injection (approx. 50000 molecules per injection), IVTs from the cDNA sublibraries were dissolved at 0.2 mg/ml. Fields of about 100 cells were injected; the photographs show several cells from each field at a magnification of 1000 x . (Panel A) Uninjected aprt- hbroblast; (panel B) aprt- fibroblasts injected with aprt-IVTs; (panels C-E) uprr- hbroblasts injected with apti-IVTs diluted with tk-IVTs to various ratios: (C) 1: 20; (D) I : 200; (E) I : 2000; (panels F-H) uprt- hbroblasts injected with IVTs from cDNA-sublibraries: (F) sublibrary 1; (G) sublibrary 4; (H) sublibrary 7.

IS1

APRT

and Tk. Injection

resulted when

of pure apt

in strong 3H labelling apt

transcripts

transcripts

5’ -aprt

NT

(Fig. 3B). However,

were diluted

with

anti-

tk tran-

scripts, the level and frequency of labelling were reduced (Fig. 3C,D). At a dilution in which the apt

sense

sense

HeLa

transcripts were at a frequency comparable to that in the HeLa poly(A) + RNA fraction (2 x 10 - 3), com-

and the relative

inefficiency

of in vitro

capping. We next injected

in vitro transcripts

RNA

sub-library

plementation was only barely detectable. This finding presumably reflects decreased stability of the in vitro transcripts

poly(A)-

7 WT

aprt

antisense

aprt

sense

IVT

IV1

from the

cDNA library constructed from the sucrose-gradient fraction. Plasmid DNA was extracted from 20 ml cultures of the bacteria from sublibraries 1 to 9. After digestion with NotI, linearised plasmids were purified on preparative agarose gels. As shown in Fig. 2, Not1 digestion of these recombinants, prior to in vitro transcription, prevents antisense RNA production, whilst using both RNA polymerases simultaneously ensures that all recombinants are represented amongst the transcription products. We consider this a great advantage over the method described by Masu et al. (1987) since there, antisense constructs are merely cleaved from the single SP6 promoter, thus reducing the effective size of the library. Whilst injection of T3/T7-transcripts from most of the nine sublibraries did not lead to complementation (e.g., sublibrary 1; Fig. 3F), in fields injected with transcripts from two sublibraries, 4 and 7, 7% of the fibroblasts were subsequently able to incorporate exogenous adenine (Fig. 3, G and H). In a further experiment, at a higher RNA concentration, 32% of the fibroblasts injected with transcripts from sublibrary 7, were complemented. We conclude that sublibraries 4 and 7 contain at least one full-length aprt-cDNA recombinant each. This was confirmed for sublibrary 7 by hybridising the in vitro transcripts to an antisense RNA transcript specific for 40 nt of 5’-flanking sequence and the first 18 nt of the coding sequence (Fig. 4). The same probe, synthesised as a sense-RNA, failed to hybridise to the sublibrary RNA (Fig. 4) indicating that antisense transcripts are not produced after in vitro transcription of the recombinants generated by the cDNA cloning procedure presented here. These results demonstrate that functional analysis after microinjection of somatic cells is a viable proposition for the selection of specific recombinants.

Fig. 4. Slot-blot from

cDNA

IVTs, antisense linearised

RNA-RNA

sublibrary

from cDNA

filters,

sublibrary

duplicate

blots

S’-specific

aprf RNA synthesised

orientation

from aBomHI-PstI

into pT7T3. Hybridisation at 60°C

of in vitro transcripts

poly(A))RNA,

sense

uprt

uprr IVTs and T3/T7 IVTs, from S pg of NorI-

plasmids

nitrocellulose

hybridization 7. HeLa

for 16 h (Zinn

HeLa poly(A)-RNA

7, were loaded

were

hybridised

onto

with

a

in vitro in sense or antisense

fragment

of APRT-0B8

recloned

was carried out in a formamide et al., 1983) with 200 &ml

as competitors,

system

yeast

filters were washed

and twice

for 30 min at 60°C with 0.05 x SSPE, O.OS% SDS.

We intend to use this method with other mutant human libroblasts with uncharacterised genetic defects. but for which a biological assay is available.

ACKNOWLEDGEMENTS

especially Dr. Manfred We wish to thank Schweiger and Dr. Rainer Schneider, Institute of Biochemistry, Leopold-Franzens-University, Innsbruck (Austria), for providing the plasmid APRT0B8 and the aprt- human tibroblasts. We also wish to thank Dr. Deininger, New Orleans, for the pTKl1 clone. Mrs. Gabriele Gutschmidt and Mrs. Susanne Rothe are thanked for their excellent technical assistance. This work was supported by grant Sp 144/10-02 from the Deutsche Forschungsgemeinschaft.

REFERENCES Bradshaw,

H.D. and Deininger,

gene: molecular

P.L.: Human

thymidine

cloning of a cDNA expressible

cells. Mol. Cell. Biol. 4 (1984) 2316-2320.

kinase

in mammalian

152

Doppler,

W., Hirsch-Kaufmann,

M.: Characterization deficiency

double-stranded cloning

co-transporter. Hidaka,

reaction

cDNA.

M.A., Coady,

pression

of blunt-ended

M.J., Ikeda, T.S. and Wright,

and cDNA Nature

sequencing

E.M.: Ex-

of the Na + /glucose

Invest.

80 (1987) 1409-1415.

Izant, J.G. and Weintraub, pression

RNA. Science

J. Chn.

and conditional

sup-

genes by antisense

229 (1985) 347-352. P.A.: Recognition

in splicing in vitro of mRNA

R.J., Brown,

D.L.: Transient

precursors.

of Cell

D.B.,

Peterson,

complementation

CA.

and Robberson,

of Xeroderma

of poly(A) + RNA.

pigmentoProc.

Natl.

Sci. USA 81 (1984) 5676-5679.

Lin, P.-F., Yamaizumi,

M., Murphey,

F.H.: Partial purification for human thymidine phoribosyl-transferase. 4290-4294.

K.,Tamaki,

of the mRNA

kinase and hypoxanthine/guanine

phos-

Proc. Natl. Acad. Sci. USA 79 (1982)

H., Harada,

S.: cDNA cloning ofbovine expression

Cloning.

Laboratory,

Cold

Y., Kuno, M. and

substance-K

system.

Nature

receptor

329

(1987)

836-838. Noma,

Y., Sideras,

P., Naito,

T., Bergstedt-Lindquist,

E., Tanabe,

S.,

T., Kinashi, T., Matsuda,

F., Yaoita, Y. and Honjo, T.: Cloning of cDNA encoding murine

IgGl

Okayama,

induction

Nature

factor by a novel strategy

cDNA.

M.J. and

Meyerowitz,

tion of RNA sequences.

E.M.: A family

Klonierung,

von DNAs

dukte eine Rolle wlhrend

K., DiMaio,

Charakterisierung

des Menschen, der DNA-Reparatur

Naturwissenschaftlichen

distinct regulatory

of lambda

SWAJ, allowing the amplifica-

Gene 52 (1987) 197-206.

R.: Molekulare

Sequenzierung

Zinn,

cloning of full-length

Mol. Cell. Biol. 110 (1982) 161-170.

phage cDNA cloning vectors,

sertation,

the

using SP6

319 (1986) 640-645

H. and Berg, P.: High-efficiency

Franzens-Universitat,

P.D., Egg, A. and Ruddle,

and characterization

J.: Molecular

NY, 1982.

oocyte

Schneider,

sum ceils by microinjection Acad.

through

Palazzolo,

R.A. and Sharp,

38 (1984) 731-736. Legerski,

Harbor,

promoter.

and endogenous

M.M., Padgett,

cap structures

phosphoribosyltransferase. H.: Constitutive

of exogenous

Spring

E.F. and Sambrook,

Manual. Cold Spring Harbor

Azuma, C., Severinson,

Y., Palella, T.D., O’Toole, T.E., Tarle, S.A. and Kelley, adenine

T., Fritsch,

A Laboratory

Nakanishi,

330 (1987) 379-381.

Human

Maniatis,

Masu, Y., Nakayama,

for the synthesis

Nucleic Acids Res. 16 (1988) 2726.

W.N.:

Konarska,

(APRT).

57 (1981) 404-410.

U.: A one-tube

Hediger,

F. and Schweiger, basis of a complete

ofthe adenine phosphoribosyltransferase

Hum. Genet. Gubler,

M., Schabel,

of the biochemical

Innsbruck,

D. and Maniatis, regions adjacent

gene. Cell 34 (1983) 865-879.

Fakultlt

deren

und Genpro-

spielen. Disder

Leopold-

1988. T.: Identification

of two

to the human p-interferon