Gene, 119 (1992) 247-251 0 1992 Elsevier Science

GENE

Publishers

B.V. All rights reserved.

247

0378-l 119/92/$05.00

06663

Short Communications

Cleavage and recognition pattern of a double-strand-specific endonuclease (I-Cre1) encoded by the chloroplast 23s rRNA intron of Chlamydomonas

reinhardtii (Group-I

intron;

ORF;

in vitro translation;

Andrew J. Thompson,

intron-encoded

enzyme;

staggered

Xiaoqin Yuan, Wieslaw Kudlicki”

cut; intron

mobility;

eukaryotic

organelles)

and David L. Herrin

Department of Botany, University of Texas at Austin, Austin, TX 78713, USA Received

by M. Belfort:

1 April 1992; Revised/Accepted:

5 May/7

May 1992; Received

at publishers:

11 June 1992

SUMMARY

Several group-I introns have been shown to specifically invade intron-minus alleles of the genes that contain them. This type of intron mobility is referred to as ‘intron homing’, and depends on restriction endonucleases (ENases) encoded by the mobile introns. The ENase cleaves the intron-minus allele near the site of intron insertion, thereby initiating gene conversion. The 23s (LSU) rRNA-encoding gene (LSU) of the chloroplast genome of Chlamydomonas reinhardtii contains a self-splicing group-I intron (CrLSU) that has a free-standing open reading frame (ORF) of 163 codons. Translation of CrLSU intron RNA in cell-free systems produces a polypeptide of approx. 18 kDa, the size expected for correct translation of the ORF. The in vitro-synthesized 18-kDa protein cleaves plasmid DNA that contains a portion of LSU where the intron normally resides, but lacking the intron itself. Cleavage by the intron-encoded enzyme (I-CreI) occurs 5 bp and 1 bp 3’ to the intron insertion site (in the 3’-exon) in the top (/) and bottom (,) strands, respectively, resulting in 4-nt single-stranded overhangs with 3’-OH termini. We also show that the recognition sequence of I-CreI spans the cleavage site and is 24 bp in length (5’-CAAAACGTC,GTGA/GACAGTTTGGT).

INTRODUCTION

Group-I introns are characterized by a semi-conserved structure and, in many cases, the ability to catalyze their own splicing from preRNA (Cech, 1987; 1988). Group-I

Correspondence to: Dr. D.L. Herrin, Texas at Austin,

Austin,

Botany

Department,

TX 78713, USA.Tel.

University

of

(512) 471-3843;

Fax (512) 471-3878. *Present address:

Department

of Texas at Austin, Abbreviations:

Austin,

bp, base

of Chemistry

and Biochemistry,

University

TX 78713, USA. Tel. (512) 471-4491.

pair(s);

C., Chlamydomonas;

cp, chloroplast;

CrLSU, intron in the LSU rRNA-encoding gene of the C. reinhardtii cp genome; ds, double strand(ed); DTT, dithiothreitol; ENase, restriction endonuclease(s); base(s) encoding

I-CreI, ENase encoded

or 1000 bp; LSU, large subunit LSU; nt, nucleotide(s);

open reading quence).

frame;

rRNA,

by the CrLSU

intron;

of the cp ribosome;

kb, kilo-

LSU,

oligo, oligodeoxyribonucleotide;

ribosomal

RNA;

SD, Shine-Dalgarno

gene ORF, (se-

introns are found in diverse organisms, in nuclear, organellar, and prokaryotic genomes (reviewed by Belfort, 1990; Dujon, 1989). It has been suggested that the wide distribution of these introns may be the result of transposition events (e.g., Dujon, 1989). Although transposition has not been experimentally demonstrated, mobility of group-I introns has: several group-I introns have been shown to invade intron-lacking alleles, converting them to intron-plus (see Belfort, 1990; Dujon, 1989; Perlman and Butow, 1989). This process, termed ‘intron homing’, is initiated by a ds ENase that is encoded by the mobile intron (Dujon, 1989; Perlman and Butow, 1989). The ENase produces a ds break in the intron-minus allele, near the site of intron insertion, and then, according to a current model, is converted to intron-plus via a ds break-repair (DSBR) mechanism (Belfort, 1990). The 23s rRNA gene (LSU) of the C. reinhardtii chloroplast contains a self-splicing group-I intron (CrLSU) (Her-

248 Ikb

pGEM235.

1

\/

y_bp

pGEM23S.E A

t!

region of C. reinhardtiicp DNA

Fig. 1. rRNA

The 16S, 7S, 3S, 23s and 5s rRNAs

and plasmid

are indicated;

constructs.

ile and ala refer to

isoleucine and alanine tRNA genes; filled boxes, coding regions; open box, intron

or intergenic

dodecapeptide HindHI.

spacers;

found

Methods.

previously

The

(Thompson

BamHI-Hind111

identical

and Herrin,

intron

ORF; ORFs.

of pGEM23S.l

was constructed

et al., 1989; Rochaix

+ HindHI-digested

vector pGEM3zf(

except for the absence Plasmid

spliced (Thompson

exon RNA was gel-purified transcribed

(Latham

pGEM23S.l

(Cetus Perkin-Elmer)

protocols.

+ ) (Prome-

with (lane 2) or without

pGEM23S.E

is

was transcribed

et al., 1990). The ligated-exon according

in RNA

and then am-

to the manufacturer’s

The oligos used for amplification

were as

follows: the 3’ oligo (oligo 14, 5’-ACCTATATAACGGC’ITGTCT-3’) is complementary

to the 3’-exon,

208-228

nt from the intron,

ATCCTTGA-3’)

is complementary

was cleaved sponding

with BamHI

and 3 nt

BamHI-Sac11

to the 5’ exon of pGEM23S.l

fragment

gel-purified, of pGEM23S.l.

is 3906

bp.

rin et al., 1990). This intron potentially encodes a freestanding ORF of 163 codons (Rochaix et al., 1985). The ORF does not encode a maturase essential for splicing of the intron in vivo (Thompson and Herrin, 1991), but appears to encode an endonuclease (I-CreI) that cleaves the LSU gene in the vicinity of the intron insertion site (Durrenberger and Rochaix, 1991). Suggestive evidence that CrLSU is a mobile intron has been obtained (Durrenberger and Rochaix, 1991), however, neither the cleavage nor the recognition site of I-CreI has been determined. Here we report on the in vitro translation of CrLSU intron excised from preRNA, demonstrating that it can act as a mRNA, and have characterized the ENase activity of the ORF product.

EXPERIMENTAL

conditions

of polypeptides 35 kDa, incubated translation

and Herrin,

subunit

reductase.

1991). The position size markers

using optimized and size (in kDa)

are indicated

of the type-1 protein

extract. and then,

on a 8M urea/l5;,”

et al., 1990), was self-spliced

used as molecular

catalytic

dihydrofolate

(lane 1) prior purification

gel (Latham

(Thompson

pGEM23S.l

on the right:

phosphatase;

Methods. For the translation,

20 kDa,

2 pg of RNA were

for 30 min at 27 “C in a 50 ~1 vol., the wheat germ extract and mixture were as described

tine (specific

activity, 40 Ci/mmol)

products

15% polyacrylamide

were analyzed gel, as described

and

[ “‘Clleu-

was used as the radioactive

aa. The

by Lax et al. (1986) by electrophoresis (Anderson

on a 0.1 “i, SDS-

et al., 1973), followed by

and

and ligated to the correpGEM23S.E

polyacrylamide

intron RNA in a wheat-germ

from HindHI-linearized

autoradiography.

and 9 nt of the LSU exon. The PCR product

+ SacII,

was transcribed

translation

3’ to a unique Sac11 site. The 5’ oligo (oligo 17, S’-GTACCCGGGGincludes 9 nt of the polylinker

of the CrLSU

RNA

1991), and the ligated-

using AMV reverse-transcriptase

Fig. 2. Translation

et al., 1985)

of the 88%bp intron and

and Herrin,

Taq polymerase

plified by PCR using

H,

described

1991). Briefly, it consists ofthe 1625-bp

(Lemieux

as follows.

vitro, the preRNA

Pl refers to a B, BamHI; was

WI) giving a total size of 4794 bp. Plasmid

to pGEM23S.l

was reverse

box, intron

group-1

construction

fragment

cloned into the BamHI ga, Madison,

hatched

in several

AND DISCUSSION

(a) In vitro translation of CrLSU intron RNA Fig. 1 shows a map of the cp rRNA gene region of C. reinhardtii; the 23s LSU gene and the CrLSU intron with

free-standing ORF are indicated. The nt sequence of CrLSU intron (Rochaix et al., 1985) predicts that an ORF of 163 codons is translated from an internal Met residue; Pl refers to a dodecapeptide that is found in a number of group-I ORFs. Previous work has shown that LSU preRNA and excised intron accumulate in vivo (Herrin et al., 1990). To determine if the intron-RNA can act as a mRNA, pGEM23S. 1 that contains the intron-containing portion of the LSU gene (and part of the spacer between 23s and 5s) on a BarnHI-Hind111 fragment (Fig. 1) cloned 3’ to the T7 promoter, was transcribed in vitro with T7 RNA polymerase and the RNA spliced as previously described (Thompson and Herrin, 1991). With these conditions, > 80% of the preRNA derived from pGEM23S. 1 became self-spliced (Thompson and Herrin, 1991). Also, a portion of the transcription product was purified by denaturing polyacrylamide gel electrophoresis (Latham et al., 1990) prior to self-splicing. The spliced RNAs were translated in a wheat-germ system (Lax et al., 1986) in the presence of [ i4C]leucine. SDS-polyacrylamide gel analysis of the products of translation showed a single major polypeptide of approx. 18 kDa, the size predicted for the product of the intron ORF (Fig. 2). Both RNA preparations translated to give a single approx. 18-kDa polypeptide, however, the

249

A

B

‘0

E”GA

A T C

Fig. 3. Assaying Plasmid

the CrLSU

pGEM23S.E

and approx.

lOO-ng aliquots

the ORF product,

intron

was linearized

(I-CreI)

were incubated

respectively

RNA

template.

with increasing

on

1 contained

After incubation,

the samples

(except lane 2) and then phenol-extracted

amounts

of

in the legend to incorporation).

mixture

an aliquot

pGEM23S.E substrate. Incubation with the translation Z-5) was in 50 mM Tris.HCl pH 9.0/25 mM MgC1,/2 h at 37°C.

activity.

2.6, 5.2, and 7.8 ng of ORF

[ “‘C]leucine

with a translation

Lane

for ENase

as described

contained

(based

lane 2, the DNA was incubated contain

product

In

that did not

of the linearized mixtures (lanes mM DTT, for 2

were treated

with RNase

and precipitated

on a 0.75 % agarose gel that contained

ethidium

of the ethidium-DNA

a photograph

shown.

Lane M contained

molecular

HincII,

which have been described

sizes of the six largest fragments

size markers, previously

fluorescence

PAT-1 digested

(Herrin

A

with ethanol.

The samples were electrophoresed bromide;

E’

with ScaI, which cuts in the vector,

which was synthesized

Fig. 2. For lanes 3-5, the reactions protein

ORF

T C

is with

et al., 1991); the

are, from the top, 5.4, 4.5, 3.2, 2.3, 2.0,

and 1.7 kb, respectively.

Fig. 4. Determination that anneals

(b) ENase activity of the intron ORF To determine if the CrLSU ORF product possesses specific ENase activity, a plasmid was constructed which contains the BarnHI-Hind111 fragment of the rRNA region

intron in the 3’-exon

(oligo 14, for

sequence

see legend to Fig. 1) was used to prime DNA synthesis

presence

of [“S]dATP.

the Sequenase

RNA that was not gel-purified prior to self-splicing (lane 1) gave a better yield. It is not clear why the RNA preparation that was not gel purified prior to splicing translates better; both preparations spliced at fairly high efficiency as determined by gel electrophoresis of the spliced RNA (not shown). One possibility is that contaminants from the polyacrylamide gel inhibited translation. Qualitatively similar results were also obtained using the reticulocyte and E. coli translation systems (data not shown). The E. coli system gave the highest rates of protein synthesis in response to the CrLSU intron RNA, although the reason for this is not clear, since the ORF is not preceded by a strong SD sequence (Rochaix et al., 1985). Further studies, however, were done primarily using the wheat-germ system, because it proved to be lower in endogenous DNase activity than the E. coli system, and was readily available. In conclusion, these results indicate that CrLSU intron RNA can act as a mRNA and that the major translation product is the size predicted from the 163-codon free-standing ORF.

site of I-CreI. (Panel A) An oligo

of the cleavage

208 nt from the CrLSU Plasmid

enzyme

pGEM23S.E

(version

2.0 from U.S. Biochemical,

OH), which was used as described

in the manufacturers

the dideoxynucleotide-termination

mix was omitted.

tion and ethanol precipitation,

in the

2 pg was the template

for

Cleveland,

protocols,

except

After phenol extrac-

10% of the s5S-labeled

DNA was cleaved

with 8 ng of I-CreI, exactly as described in the legend to Fig. 3. The cleaved DNA (lane E) was co-electrophoresed on a 6% polyacrylamide sequencing

gel with a sequence

the Sequenase terminated

protocols

with dideoxy

ladder that was generated

G, A, T, C, respectively).

(5’-TCGCTCAACGGATAAAAGTT-3’) the intron in the 5’-exon ence of [35S]dATP primer-extended

as described

in

using oligo 14 as primer (lanes G, A, T, C were (Panel

that anneals

was used to prime DNA synthesis

as described

for panel

A above.

B) Oligo 15

155- 175 nt from in the pres-

Cleavage

of the

DNA with the ENase (lane E) was similarly performed

as in panel A. The sequence ladder for oligo 15 (lanes G, A, T, C were terminated with dideoxy G, A, T, C, respectively) was also produced as described

by the manufacturer

[35S]dATP.

The primer-extended,

sequence

ladder

quencing

gel and autoradiographed.

age resulted

of Sequenase

(U.S.

ENase-cleaved

were coelectrophoresed

Biochemical) DNA

using

(lane E) and

on a 6% polyacrylamide

se-

The figure shows that ENase cleav-

in one major band for each 35S-labeled

strand.

where the intron normally resides, but lacking the intron (Fig. 1). The plasmid, pGEM23S.E, was linearized by restriction with ScaI, which cuts in the pGEM3zf( + ) vector (Promega, Madison, WI), and then incubated with the ORF-encoded protein. Fig. 3 shows that the linearized approx. 3.9-kb plasmid DNA (lane 1) was cleaved into two

250 fragments of approx. 2.1 and approx. 1.8 kb (lanes 3-5), and that the highest amount of ORF protein resulted in complete cleavage (lane 5). Incubation of the DNA with a control, minus-RNA translation mixture (lane 2) resulted in no cleavage of the piasmid DNA; the slightly altered mobility of the DNA in this lane is due to the large amount of endogenous wheat-germ RNA that was not removed. The sizes of the two digestion products are those predicted if ds cleavage occurs near the LSU exon-junction region in pGEM23S.E.

exact sites of cleavage, which are indicated in Fig. X, occurred 5 bp away from the site of intron insertion on the top strand and 1 bp away on the bottom strand, leaving 3’-OH overhangs of 4 nt. We verified that the ENase, which we propose to call I-CreI using the nomenclat~e of Dujon et al, (1989), leaves 3’-OH and 5’-PO4 ends by using T4 DNA ligase (Sambrook et al., 1989). Supercoiled pGEM23S.E was linearized with I-CreI, and following removal of the enzyme by phenol extraction and ethanol precipitation, the DNA was incubated with T4 DNA ligase. Agarose gel analysis showed that all of the linearized DNA had been religated (data not shown).

(c) The cleavage pattern of I-Cref The exact site of cleavage (on each strand) by the intron ENase was determined using oligo primers to direct synthesis of radiolabeled DNA that contained the region of cleavage. The primer-extended DNAs were then cleaved with the in vitro-synthesized ENase, and the products coelectrophoresed with a dideoxynucleotide sequence ladder generated with the same primers. Fig. 4 shows the results obtained with the two strand-specific primers; a single major band was obtained in the ENase-cleaved lane (E) for each primer. Control experiments showed there was no band at this position prior to cleavage (not shown). The @

GATC’GATC’

(d) The recognition sequence of 143~1 In order to delimit the recognition sequence of I-&I, we used the approach of Wenzlau et al. (1989). Dideoxynucleotide sequence ladders, generated as in Fig. 4, except using [32P]dATP in place of [ ?$]dATP, were cleaved with I-CreI and analyzed by denaturing gel electrophoresis. When the primed-DNA has been extended through the reco~ition sequence, the DNA is e~ciently cleaved and the sequence ladder disappears. Fig. 5A shows that comB) -

cleaved

G A T C’G

.

3’. Fig. 5. Delimiting sequencing

the recognition

reaction to generate

kit using [32P]dATP

A T C'

12345678

12345678

5’.

cleaved

. sequence

substrate

of 1-0~1.

(Panel A) Oligo

as label. Lanes G, A, T, C were terminated

cleaved with 8 ng of I-Cm1 as described

14 (see legend to Figs. 1 and 4A) and

for cleavage by the I-CreI ENase. The reactions with dideoxy

were performed

G, A, T, C, respectively.

in Fig. 3. The cleaved (lanes 5-8) and uncleaved

(lanes l-4)

I pg of pGEM23S.E

as described

in the Sequenase

were used in each (U.S. Biochemical)

5% of each of the 32P-labeled

DNAs were electrophoresed

substrates

was

on 6% polyacrylamide

sequencing

gels; the panel is a composite of two different gels electrophoresed under identical conditions. (Panel B) Oligo 15 (see legend to Fig. 4B) and 1 gtg of pGEM23S.E were used to generate a 32P-labeled sequence ladder of the bottom strand. The sequencing and cleavage reactions were performed as in panel A. Panel B is also a composite of two different 6% sequencing gels electrophoresed under identical conditions. Only the nt that extend beyond the cleavage site and are resistant to cleavage are identified to the right of each panel; no bands were visible in the upper portions of the autoradiographs that were excised. (C) The cleavage site and recognition sequence of I-CreI. The site of intron insertion is indicated by the downward arrow, the staggered lines indicate the sites of cleavage on each strand, and the horizontal bracket indicates the extent of the 24-bp recognition sequence.

251 plete cleavage of the top strand ladder occurs only after 12 nt upstream from the cleavage site have been synthesized. Fig. 5B shows that for the bottom strand complete cleavage occurred only after 9 nt from the cleavage site had been synthesized. These results are summarized in Fig. 5C; the recognition sequence of I-&I, as determined by this method, is 24 bp. The cleavage pattern of I-&I (close to the site of intron insertion and with 4-nt 3’-OH overhangs) contains features that have been found for other eukaryotic intron-encoded ENases (Colleaux et al., 1989; Marshall and Lemieux, 1991; Muscarella et al., 1990; Wenzlau et al., 1989); in contrast, the phage intron-encoded ENases cleave at some distance (> 10 bp) from the intron insertion site (see Belfort, 1990). The fact that the recognition sequence of I-&I spans the cleavage site, albeit asymmetrically, is also similar to the eukaryotic intron-encoded ENases. The size of the recognition sequence of I-CreI (24 bp), however, is somewhat longer than that of the mitochondrial enzymes I-&e1 and I-SceII, which are 18 bp (Colleaux et al., 1989; Wenzlau et al., 1989). However, we cannot rule out the possibility of end effects having some effect on our determination of the size of the recognition sequence of I-CreI, and that it may eventually prove to be a few nt smaller. Although, in this regard, it should be mentioned that we used the same approach as that used for the I-&e11 enzyme (Wenzlau et al., 1989). Finally, it should be noted that there is no obvious homology between the recognition sequence of I-CreI and the other known intron-encoded ENases.

REFERENCES Anderson,

C.W., Baum, P.R. and Gesteland,

ovirus 2-induced Belfort,

M.: Phage

Genet. Cech,

proteins.

R.F.: Processing

of aden-

J. Virol. 12 (1973) 241-252.

T4 introns:

self-splicing

and mobility.

Annu.

Rev.

24 (1990) 363-385.

T.R.:

The chemistry

of self-splicing

RNA

and RNA

enzymes.

Science 236 (1987) 1532-1539. Cech, T.R.: Conserved

sequences

ing an active site for RNA

and structures catalysis

-

of group I introns: a review.

Gene

build-

73 (1988)

259-271. Colleaux,

L., d’Aurio1, L., Galibert,

cleavage

site of the intron

F. and Dujon,

encoded

B.: Recognition

omega transposase.

Acad. Sci. USA 85 (1988) 6022-6026. Dujon, B.: Group-I introns as mobile genetic elements: anistic speculations

and

Proc. Nat].

facts and mech-

- a review. Gene 82 (1989) 91-113.

Dujon, B., Belfort, M., Butow, R.A., Jacq, C., Lemieux, C., Perlman, and Vogt, V.M.: Mobile introns: definition ofterms nomenclature. Durrenberger,

Gene 82 (1989) 115-118.

F. and Rochaix,

J.-D.:

Chloroplast

ribosomal

Chlamydomonas reinhardtii: In vitro self-splicing, activity and in vivo mobility. Herrin,

of

A.J. and Chen, Y.-F.: Self-splicing psbA introns.

self-splicing

Plant

of

Cell 3 (1991)

G.W.: RNA splicing in Chlomy-

D.L., Chen, Y.-F. and Schmidt,

domonas chloroplasts:

intron

DNA endonuclease

EMBO J. 10 (1991) 3495-3501.

D.L., Bao, Y., Thompson,

the Chlomydomonas chloroplast 1095-I 107. Herrin,

P.S.

and recommended

of 23s preRNA.J.

Biol. Chem. 265

(1990) 21134-21140. Latham,

J.A., Zaug,

cleavage

A.J. and Cech, T.R.:

of RNA by a group-I

Self-splicing

intervening

and enzymatic

sequence.

Methods

Enzy-

mol. 181 (1990) 558-569. Lax, S.R., Lauer, S.J., Browning, properties

of protein

wheat germ. Methods Lemieux, C., Boulanger, of the chloroplast

K.S. and Ravel, J.M.: Purification

synthesis

initiation

Enzymol.

and elongation

factors

118 (1986) 109-128.

J., Otis, C. and Turmel, M.: Nucleotide large subunit

and from

rRNA

sequence

Chlamydomonas

gene from

reinhardtii. Nucleic Acids Res. 17 (1989) 7997.

(e) Conclusion Identification of the CrLSU intron ORF as encoding an ENase specific for the exon-junction region of an intronminus LSU gene fragment provides further evidence that this intron is mobile. The ability to introduce the ENase cleavage and presumed homing site for CrLSU into the chloroplast of C. reinhardtii (Durrenberger and Rochaix, 1991) and the availability of an ORF-minus mutant (Thompson and Herrin, 1991), should allow a definitive demonstration of homing by this cp intron.

Marshall, clease

P. and Lemieux. encoded

rRNA-encoding Muscarella,

of the homing

endonu-

large subunit

gene of Chlamydomonas eugametos. Gene 164 (1991)

D.E., Ellison, E.L., Ruoff, B.M. and Vogt, V.M.: Character-

ization of I-go,

an intron-encoded

endonuclease

that mediates

Perlman,

P.S. and Butow, R.A.: Mobile introns

teins. Science 246 (1989) 1106-l Rochaix,

J.-D.,

Rahire,

and intron-encoded

J., Fritsch,

Laboratory

Manual.

M. and Michel,

Cold Spring Harbor, Thompson,

E.F.

pro-

109. F.: The chloroplast

intron of Chlamydomonas reinhardtii codes for a polypeptide maturases.

hom-

of Physarum

ing of a group I intron in the ribosomal DNA polycephalum. Mol. Cell. Biol. 10 (1990) 3386-3396.

Sambrook,

This research was supported by grants from the U.S. National Science Foundation (DMB89-05303) and the Robert A. Welch Foundation (F-l 164) to D.L.H. We thank Karen Browning for wheat-germ extract and for oligo synthesis.

pattern

in the chloroplast

241-245.

mitochondrial

ACKNOWLEDGEMENTS

C.: Cleavage

by the fifth intron

ribosomal related to

Nucleic Acids Res. 13 (1985) 975-984. and Maniatis,

T.: Molecular

2nd ed. Cold Spring Harbor

Cloning.

Laboratory

A

Press,

NY, 1989.

A.J. and Herrin,

D.L.: In vitro self-splicing

reactions

of the

Chlamydomonas reinhardtii chloroplast group-I intron Cr.LSU and in viva manipulation via gene replacement. Nucleic Acids Res. 19 (1991) 6611-6618. Wenzlau,

J.M., Saldanha,

R.J., Butow, R.A. and Perlman,

intron-encoded maturase is also an endonuclease mobility. Cell 56 (1989) 421-430.

P.S.: A latent

needed

for intron