- Email: [email protected]
Escherichia coli RecA protein modified with a nuclear location signal binds to chromosomes in living mammalian cells
EXPERIMENTAL
CELL RESEARCH
198,107-114
(1992)
Escherichia co/i RecA Protein Modified with a Nuclear Location Signal Binds to Chromosomes in Living Mammalian Cells MITSURU KIDO, YOSHIHIRO YONEDA,’ MAHITO NAKANISHI, TS~YOSHI UCHIDA,’ AND YOSHIO OKADA Institute for Molecular and Cellular Biology, I-3 Yam&a-Oku, Suita, Osaka 565, Japan
We tried to make a well-characterized bacterial protein function in mammalian cell nuclei. For this purporn we chose Emherichia coli RecA protein and fused its carboxy terminus to the nuclear location signal of SV40 large T-antigen by oligonucleotide-dependent modification of the gene. When injected into the cytoplasm, the modified RecA protein (T-RecA for the T-antigen signal) accumulated efficiently in the nuclei, whereas the wild-type RecA protein remained in the cytoplasm. The T-RecA protein retained its original in uiuo activity, judging from the finding that uv-sensitive bacteria (recA- E. cola became uv-resistant on transformation with the T-recA plasmid as well as the recA plasmid. For expression of the T-recA gene in mammalian cells, the 5’ region was replaced by the chicken Bactin promoter and Kozak’s initiation signal. A high level of expression was observed when Chinese hamster ovary (CHO-Kl) cells were transfected with this plasmid. Indirect immunofluorescence examination revealed that the T-RecA protein in nuclei of mammalian cells bound to chromatin. o 1992 Academic PMS, IUC.
MATERIALS
INTRODUCTION
It is very interesting to make an exogenous protein function in living mammalian cells, especially when the protein has well-characterized, biochemically important activity. Introduction of such proteins into mammalian cells should be helpful in understanding some functions of the cells. Many bacterial proteins have been isolated and characterized and so may be suitable for this purpose. We thought that an Escherichia coli protein might be functional in mammalian cells because studies showed that fibroblasts from a patient with xeroderma pigmentosum, an autosomal recessive human genetic disease of uv repair, showed little or no unscheduled DNA synthesis (UDS) after uv irradiation but gained a normal level 1 To whom correspondence and reprint requests dressed. ’ Dr. Tsuyoshi Uchida died on May 3,1989.
of UDS after concomitant treatment with bacteriophage T4 endonuclease V and HVJ (Sendai virus) [l, 21. We focused attention on the E. coli RecA protein which plays a major role in genetic recombination in E. coli [3,4]. In vitro studies indicated pleiotropic activities of this protein. A remarkable finding was a strand exchange reaction in vitro, which required continuous ATP hydrolysis. The recA mutations in E. cob not only block recombination completely, but also lead to pleiotropic effects, such as high sensitivity to uv and X-ray irradiation and enhanced DNA breakdown after uv irradiation. We developed a system in which functional bacterial RecA protein could be introduced efficiently into mammalian cell nuclei. We found that the RecA protein itself is unable to enter the nucleus. Therefore, we modified it to a nuclear protein with the nuclear localization signal peptide of SV40 large T antigen discovered by Kalderon et al. [5,6]. We demonstrated association of this modified RecA protein (T-RecA) with chromosomes by the indirect immunofluorescence method.
should
be ad-
107
AND METHODS
Cell lines. CHO-Kl cells (a Chinese hamster ovary cell line) were obtained as described [7]. HEL (human embryonic lung) cells were a gift from Dr. A. Tada [8]. Plasmids. pTM2 (Fig. 2) containing the recA gene was a generous gift from Dr. Hideyuki Ogawa (Department of Biology, Faculty of Science, Osaka University) [9, lo]. pAct(X) (see below and Fig. 4) has the chicken &actin promoter and is a derivative of pact-c-myb (a gift from Dr. Shunsuke Ishii, Tsukuba Life Science Center) [ll]. ~220.2 (Fig. 4) was obtained from Dr. Bill Sugden (University of Wisconsin) [ 121. Syntheses of oligonuckotides. Oligonucleotides were synthesized in automated DNA synthesizers Models GENET A-III (Nihon Zeon), 381A (Applied Biosystems), and 391 PCR-MATE (Applied Biosysterns). The preparations were then purified by high performance liquid chromatography or electrophoresis in 20% acrylamide gel containing 8 M urea. Oligonucleotide-directed mutagenesis. Oligonucleotide-directed mutagenesis was carried out as described by Zoller and Smith [13] and Kunkel et al [ 141. DNA sequencing for confirmation was done in principle as described by Sanger et al. [15] with use of a kit (Sequenase version 2.0, United States Biochemical Corp.). Constrzxtion of pT-RecA and pATR4. pT-RecA, an expression plasmid of E. coli, was constructed to modify the recA gene with the 0014~4827/92%3.00
Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
KID0
ET AL.
FIG. 1. Subcellular localizations of the RecA protein and the T-RecA protein microinjected into the cytoplasm of HEL cells. (A) RecA protein; (B) T-RecA protein. (a, c) Phase-contrast microscopy; (b, d) fluorescence microscopy. Cells were incubated for 30-60 min after microinjection and then fixed and treated with anti-RecA antibody followed by a FITC-conjugated second antibody.
nuclear location signal of the SV40 large T antigen (Pro-Lys-LysLys-Arg-Lys-Val-Glu-Asp-Pro) (Fig. 2). Briefly, the region of the recA gene downstream from the DdeI site near the 5’-end was replaced by synthetic nucleotides in which the carboxyl terminus of the RecA protein was fused with the nuclear location signal. For expression of the T-RecA protein in mammalian cells, pATR4 was constructed (Fig. 4). In pATR4, the T-recA gene was expressed by the chicken @-actin promoter and translation was initiated by Kozak’s sequence. The selection marker hph (hygromycin B phosphotransferase) was also introduced. pAct(X), a derivative of pact-c-myb, was used as a mammalian expression vector. For construction of pAct(X) the vector was replaced by pUC19, and a 2.0-kb NcoI-XbaI fragment (c-myb gene) was replaced by a XhoI linker. Purification of RecA and T-RecA proteins from E. coli. The E. coli C600 strain harboring pTM2 and pT-RecA (Fig. 2), respectively, was induced to overproduce RecA and T-RecA proteins by treatment with nalidixic acid (60 pglml) and the proteins were purified as described [16]. Both proteins could be purified by the same procedure. Microinjection. Samples of 3-5 mg/ml of RecA or T-RecA protein were dialyzed against modified phosphate-buffered saline (reverse PBS: 2.7 mM NaCl, 137 m&f KCl, 8.1 m&f NarHPO,, 1.5 m&f KH,PO,), filtered through a Millipore filter (filter type HV, pore size
0.45 pm), and microinjected into the cytoplasm of HEL cells with a micromanipulator (Narishige Scientific Instrument, Tokyo, Japan) as described [ 171. Transient and stable transformation of CHO cells. CHO cells were transfected with plasmid DNA (pATR4) to express the T-recA gene. For transient expression, the CaPO, method 1181was used. For stable transformation, electroporation [19] was used and then the cells were selected in selection medium containing 600-600 ag/ml hygromycin B (Boeringer-Mannheim) for 10 days, and stably transformed colonies were cloned (El-E39). The Indirect immunofluorescence stainings of RecA and T-RecA. first antibody against RecA protein was raised in rabbits by repeated intradermal injections of purified RecA protein with Freund’s adjuvant (Difco). The second antibody, fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit IgG, was purchased from Cappel Co. and filtered through a Millipore filter (type HV, 0.45 pm) before use. Cells were fixed with 3.7% formaldehyde in PBS at room temperature for 15 min, permeabilized with cooled 100% methanol for 5 min, and blocked by treatment with blocking solution (2 mg/ml ovalbumin, 50 mM glycine, and 2% goat serum in PBS) for at least 30 min. A volume of 50 ~1 of the first antibody diluted 200-fold with blocking solution was applied to each coverglass and the cells were kept at 4’C
RecA WITH
\
A NUCLEAR
LOCATION
SIGNAL
BINDS
109
CHROMOSOMES
red 3’and + nuclear transport s@nal 367
bp
NGCLNAIl glu gas ctt
gly g(rc ccg
val gta cat
ala irca Cgt
clu mm ctt
thr act tya
a80 sac ttg
olu ma ctt RecA
Val
0111 Aep
OTT CM
GAG CTC
GAT CTA
Pro
tbr b CCG taa GGC att
LOCATION
SIGNAL
asp
Ias
IYS
lays
Arc
wat eta
MO Trc
AM TIT
MA TIT
COT AM OCA TIT
I carboxyl
Lps
tarmlnus
ter taK (It.2
tcf.yatgtt
tc.¶t.scacaa actatgtgtt
Kggcgcatct CCCgCgtaCa
FIG. 2. Modification of the recA gene with the nuclear location signal of SV40 large T-antigen. (A) Construction of pT-recA; The recA gene was fused with the nuclear location signal. Twelve oligonucleotides were annealed to parts B, C, and D, and these parts were ligated to form part BCD (see below). The synthetic DNA region (BCD) was then ligated with an EcoRI-DdeI 205-bp fragment of the recA gene (part A). Finally part A-BCD was ligated with a 1.8-kb 5’-fragment of recA and vector pUC19. (B) DNA sequence of the synthetic DNA region (part BCD). The nuclear location signal of 10 amino acid residues was fused at the carboxyl terminus of the RecA protein. The 3’-flanking region for termination of mRNA was the same as in the original recA gene.
overnight. They were then washed six times with cooled PBS and incubated at room temperature for at least 1 h with 50 ~1 of the second antibody diluted 100-fold with blocking solution.. They were then washed six times with cooled PBS and mounted with Mowiol 488(Hoechst) containing 1,4-diazobicyclo-[2.2.2]-octane (DABCO) [20]. Survival curves of uv-irradiated E. coli. The bacterial strains used were K12 C600 (F-, thi-1, thr-1, leuB6, lacY1, tonA21, supE44, h-), and MV1184 [ara, A(lac-proAB), rpsL, thi, (@3OlacZAM15), A(srlrecA)306::TnlO(tet’), F’(traD36, proAB+, lacF, lacZAM15)]. Strain C600 is the recA+ wild type, while MV1184 is a recA- mutant type. Bacteria in the log phase of growth were suspended in 0.1 M MgSO, at a concentration of l-2 X 10’ cells/ml and irradiated with a 15-W Toshiba germicidal lamp at a distance of 45 cm. The dose in this condition is 1.9 J/mz/s (19 erg/mm2/s). Survival curves were obtained as described [21]. RESULTS
RecA protein with a nuclear location signal was eficiently transported into the nucleus. Chromatins are shielded by nuclear membranes in mammalian cells, so if the RecA protein has no ability to pass through these membranes it cannot interact with genes. We first examined whether the RecA protein could accumulate in the
nucleus. When the purified RecA protein was microinjetted into the cytoplasm of HEL cells, it remained in the cytoplasm and did not enter the nucleus, as shown by indirect immunofluorescence staining (Fig. 1A). In eukaryotic cells, all nuclear proteins are initially synthesized in the cytoplasm and then rapidly transported into the nucleus [22-241. The precise mechanism of nuclear transport is not fully understood, but active transport has been suggested [25]. Moreover, Kalderon et al. reported that a short amino acid sequence of the SV40 large T antigen, Pro-Lys-Lys-Lys-Arg-Lys-Val, specifies its nuclear translocation [5, 61. Conjugation with this signal peptide has also been shown to direct transport of exogenous protein into the nucleus [26]. On the basis of these reports, we added the nuclear location signal of the SV40 large T antigen to RecA protein by oligonucleotide-dependent modification of the recA gene (pT-RecA, Fig. 2). This modified RecA protein (T-RecA) was produced in E. coli and purified as described under Materials and Methods. Purified TRecA protein had a slightly larger molecular weight than the RecA protein, as judged by SDS-PAGE (data
110
&I
4 6 12 16 UV Irradiation (J/m2 ) FIG. 3. Demonstration of comparable activity of the T-RecA protein to that of the wild-type RecA protein by uv-survival curves. MV1184 is a recA- strain while C600 is a recA+ strain. MV1164 was transformed with two plasmids: one was pT-RecA (Fig. 2), expressing the T-RecA protein, and the other was pRecA, expressing the intact RecA protein. pRecA was constructed to adjust the copy number in E. coli to that for pT-RecA. Since the vector of pT-RecA was high copy pUC19, the vector of pTM2 (Fig. 2) was replaced by the same pUC19 to yield pRecA. For construction, a 3.2-kb BamHI fragment containing the wild-type recA gene from pTM2 was cloned into the BamHI site of pUC19.
not shown). When the purified T-RecA protein was microinjected into the cytoplasm of HEL cells, it accumulated in the nucleus, as shown by indirect immunofluorescence staining (Fig. 1B). The T-RecA protein was as active as the RecA protein. Modification sometimes leads to inactivation
of a protein. Therefore, we examined the activity of the T-RecA protein in E. coli in vivo. As the RecA protein plays central roles in recombination and repair, recAmutants of E. coli show high sensitivity to uv-light irradiation [3]. But transformation with a plasmid encoding the recA gene complements this high sensitivity. Therefore, we compared the survival curves of recA- E. coli transformed with pT-RecA and with pRecA to determine whether the T-RecA protein was inactivated during peptide modification. Figure 3 showed that a transformant of a recA- strain (MV1184) with pT-RecA became as uv-resistant as a transformant with pRecA. This finding indicates that the T-RecA protein was as active as the RecA protein in E. coli in vivo. T-RecA protein expressed in cells was bound to chromosomes. Injection with a microcapillary is suit-
able for determining the subcellular localization of proteins of interest, but it is not a suitable method for studying the effects of the T-RecA protein in mammalian cells because it is not possible to inject the protein into a
AL.
sufficiently large number of cells. Therefore, we cloned the T-recA gene into a mammalian expression vector. Figure 4 shows the procedure used. First, the T-recA gene was cloned into the vector pAct(X) with 5’-, 3’flanking bacterial sequences (pATRl), then these redundant sequences were deleted (pATR2, pATR3), and finally, as a selection marker, the hygromycin B phosphotransferase gene hy$ was added (pATR4). The bacterial Shine-Delgarno sequence was replaced by Kozak’s sequence during oligonucleotide-directed deletion (pATR2, Fig. 4B). As shown in Fig. 5, when the resulting plasmid pATR4 was transfected into CHO cells, the T-RecA protein was strongly expressed in the cells and efficiently transported into their nuclei. Figure 5A shows the transient expression achieved by the CaPO, method. Several percent of the total cells in samples examined at 24,48, and 96 h expressed the T-RecA protein. At 24 h after transfection, scattered cells expressed T-RecA, but by 96 h colonies of cells expressing T-RecA were observed (Fig. 5A). If the protein had harmful effects, stable transformants expressing T-RecA would not be obtained. But cells expressing T-RecA stably were obtained by hygromycin B selection. Figure 5B shows one clone (clone E13) in which 100% of the progeny cells expressed TRecA. Moreover, although expression of the T-RecA protein was high, cell growth seemed to be normal and stable. Figure 6 shows indirect immunofluorescence staining of T-RecA in stable transformants (clones El3 and E23) in the late mitotic phase. The cells were dividing into two daughter cells and condensed chromosomes were stained as a thick band in the center of the rounded cells. If the T-RecA protein acts on genes, it must interact with chromatin, so this result provided clear evidence for the interaction of the T-RecA protein with chromosomes. DISCUSSION
In this study we introduced E. coli RecA protein into mammalian cells. For its efficient transport into the nuclei, modification of its carboxyl terminus with a nuclear location signal was necessary. As this modification did not cause loss of activity of RecA, this technique should be useful in studies on the effects of other exogenous proteins in the nuclei of mammalian cells, especially proteins expected to have important activities on genes or chromatins. In the case of T4 endonuclease V, even on its introduction into the cytoplasm with HVJ (Sendai virus), it functioned in the nucleus. T4 endonuclease V is sufficiently small (about 20,000 Da) to allow its passive entry into nuclei. Although the molecular weight of the RecA protein is only 38,060 Da, the smallest active form
RecA WITH
A NUCLEAR
LOCATION
SIGNAL
BINDS
111
CHROMOSOMES
A
pAot(X)
4.lkb
T-rrcA !naoflbn pATR1
6.3kb
ongonuchot-ted
B dewoIl
d&on
Met.Ala.Ile.Asp.Glu.Asn ~crcCocAG/CCA,o~~.ATG.GCT.A~.GAC.~.MC 18
#9-aCtfro \ \
7.3kb
18
.er recA
intron
HT
coding
sequence
OligOllUCleOlide
S-flanking region
FIG, 4. Construction of plasmids for expression of T-RecA in mammalian cells. (A) The T-recA gene was cloned under the control of chicken &actin promoter. pAct(X) is an expression vector for mammalian cells (see Material8 and Methods). The T-recA gene (open box, recA, 5’ shaded, nuclear location signal) with 5’- and 3’-bacterial flanking sequences (closed boxes) was cloned into the XftoI site of pAct(X) (pATR1). Redundant bacterial sequences were then completely removed by oligonucleotide-directed deletion (pATR2, see below) and by digestions with XbaI and XhoI (pATR3). Finally a 2.0-kb NruI-Hind111 fragment of the hygromycin B phosphotransferase gene (hyga) was inserted as a selection marker (pATR4). Poly A (*) means poly(A) additional signal. (B) Details of oligonucleotide-directed deletion/insertion. For deletion of 5’-bacterial flanking sequences (0.9 kb, bold triangular line) from pATR1, the vector was replaced by pUCll8 and doublestranded DNA was converted to single-stranded DNA. The ssDNA was then hybridized with the 44-mer oligonucleotide. By this hybridization, the flanking region for deletion was looped out, and the @actin intron and recA coding sequence were hybridized adjacent to the eukaryotic initiation signal (Koxak’s sequence). Ordinary oligonucleotide-directed mutagenesis was then carried out as described [13,14]. In the 44-mer oligonucleotide the mark (/) indicates a splicing acceptor of the intron.
was reported to be a tetramer [9], and its apparent mo- plasmid that expressed the T-recA gene (Figs. 4 and 5). lecular weight estimated by gel filtration was much As shown in Fig. 1, direct microinjection was suitable for larger (M. Kido, unpublished observation). Many other injecting large amounts of protein into a few cells, cerprocaryotic proteins that interact with DNA show simi- tainly, and the erythrocyte-ghost fusion method was suitable for introducing small amounts of protein into lar aggregation. When the RecA protein was microinjetted into the cytoplasm, it remained in the cytoplasm many cells, but was not suitable for introducing proteins and did not enter the nucleus (Fig. 1A). But our results of large molecular weight. The third method was someshowed that after modification with a nuclear location what different from these two methods since it did not signal, even large molecular weight proteins such as involve direct injection of the protein. But as shown in RecA can be transported into nuclei (T-RecA, Figs. 2 Fig. 5, it was very effective for introducing large and 1B). amounts of protein into a large number of cells. RecA protein without any modification could be inReferring to data described here, one can choose an jected directly into nucleus. But nuclear microinjection appropriate method to introduce an exogenous protein usually damages cells, and the amount of material in- into living cells. If the protein is easily obtained, direct jected into the nucleus must be far less than the amount microinjection and erythrocyte-ghost-mediated injecinjected into the cytoplasm. Furthermore, it is impossi- tion are available. Direct microinjection of a protein ble to inject the protein into nuclei of a larger number of into the nucleus is not impossible, but modification of cells to study its effect. the protein by a nuclear location signal facilitates its After modification of the RecA protein with a signal nuclear transport. Moreover, if the gene has already peptide, our next problems were how to introduce it into been cloned, the third method used here is also availcells and what dose would be suitable. We tried three able. methods of injection: direct microinjection (Fig. l), inAfter efficient introduction of the T-RecA protein jection by erythrocyte-ghost fusion with HVJ (Sendai into nuclei, we demonstrated its direct association with virus) [27] (data not shown), and transfection with a chromosomes (Fig. 6). If the T-RecA protein in nuclei
KID0
ET AL.
FIG. 5. Transient and stable expressions of the T-RecA protein in CHO cells. Cells were transformed with pATR4 (Fig. 4) in which the T-recA gene was cloned into a mammalian expression vector. The CaPO, precipitation method and electroporation were used for transient and stable expression, respectively, as described under Materials and Methods. Cells were stained by the indirect immunofluorescence method with anti-RecA antibody. (A) Transient expression at 96 h, (B) stable expression (clone E13). (a, c, e) Phase-contrast microscopy; (b, d, f) fluorescence microscopy.
had been present in the nucleoplasm and had no affinity to chromatin, the figure in the mitotic phase would have been different: the condensed chromosomes in the center of mitotic cells would have shown no fluorescent
signal and the fluorescence would have been observed in the surrounding area, because there is no nuclear envelope in this phase. Therefore, Fig. 6 indicates that the T-RecA protein bound to mammalian chromosomes.
RecA WITH
FIG. 6. Indirect stable transformants
A NUCLEAR
LOCATION
SIGNAL
We thank Drs. K. Ueda, H. Harada, and K. Kohno for help in syntheses of oligonucleotides. We also thank Dr. K. Kohno for valuable suggestions. This work was supported by grants from the Ministry of Education, Science and Culture of Japan and the Toray Science Foundation.
K., Sekiguchi,
M., and Okada, Y. (1975) Proc. Natl.
Tanaka,
K., Hayakawa,
H., Sekiguchi,
9.
Ogawa, T., Wabiko, H., Tsurimoto, T., Horii, T., Masukata, H., and Ogawa, H. (1978) Cold Spring Harbor Symp. Quant. Biol. 43, 909-915.
10.
5. 6. 7.
Kalderon, D., Richardson, W. D., Markham, A. F., and Smith, A. E. (1964) Nature 311,33-38. Kalderon, D., Roberts, B. L., Richardson, W. D., and Smith, A. E. (1964) Cell 39,499~509. Kohno, K., Uchida, T., Mekada, E., and Okada, Y. (1985) Somatic Cell Mol. Genet. 11, 421-431.
Horii,
T., Ogawa, T., and Ogawa, H. (1986) Proc. Natl. Acad. Sci.
USA 77,313-317. 11. 12.
Nishina, Y., Nakagoshi, H., Imamoto, F., Gonda, T. J., and Ishii, S. (1989) Nuc.!& Acids Res. 17,107-117. Yates, J. L., Warren, N., and Sugden, B. (198.5) Nature 313, 812-815.
13.
Zoller, M. J., and Smith, M. (1983) in Methods in Enzymology (Wu, R., Grossman, L., and Moldave, K., Eds.), Vol. 106, pp. 468-500, Academic Press, San Diego.
14.
Kunkel, T. A., Roberts, J. D., and Zakour, R. A. (1967) in Methods in Enzymology (Wu, R., and Grossman, L., Eds.), Vol. 154, pp. 367-382, Academic Press, San Diego.
M., and Okada, Y. (1977)
Proc. Natl. Acad. Sci. USA 74,2958-2962. 3. Clark, A. J. (1973) Annu. Reu. Genet. ‘7,67-36. 4. Radding, C. M. (1982) Annu. Rev. Genet. 16,405-437.
F., Toba, M., and Tada, A. (1979) Arch. Virol. 60,347-
Taguchi, 351.
Acad. 5%. USA 72,4071-4075. 2.
phase of
8.
REFERENCES Tanaka,
CHROMOSOMES
immunofluorescence staining of T-RecA in the late mitotic phase. Pairs of daughter cells in the late mitotic (clones El3 and E23) are shown, (A, B) Phase-contrast microscopy; (a, b) fluorescence microscopy.
Thus our results indicate that this system should be useful for studies on the effects of the T-RecA protein on mammalian chromosomes. Moreover, the same procedure is available for introduction of other bacterial proteins into the mammalian nucleus.
1.
BINDS
15.
Sanger, F., Nicklen,
S., and Coulson,
A. R. (1977) Proc. Natl.
Acad. Sci. USA 74,5463-5467. 16.
Kuramitus,
S., Hamaguchi,
K., Ogawa, T., and Ogawa, H. (1981)
J. Biochem. 90,1033-1045. 17.
Yamaizumi,
M., Horwich,
A. L., and Ruddle, F. H. (1963) Mol.
Cell. Biol. 3, 511-522. 18. 19.
Chen, C., and Okayama, H. (1967) Mol. Cell. Btil. 7.2745-2752. Kido, M., Miwatani, H., Kohno, K., Uchida, T., and Okada, Y. Cell Struct. Funct., in press.
114
KID0
20. Harlow, E., and Lane, D. (1988) Antibodies: A Laboratory Manual, p. 418, Cold Spring Harbor Laboratory, New York. 21. Miller, J. H. (1972) Experiments in Molecular Genetics, pp. 121124, Cold Spring Harbor Laboratory, New York. 22. Yamaizumi, M., Uchida, T., Okada, Y., Furusawa, M., and Mitsui, Y. (1978) Nature 273,782-7&k 23. Sugawa, H., Imamoto, N., Wataya-Kaneda, M., and Uchida, T. (1985) Exp. Cell Res. 159,419-429. Received June lo,1991 Revised version received September 5, 1991
ET AL. 24. Tsuneoka, M., Imamoto-Sondoe, N., and Uchida, T. (1986) J. Biol. Chem.
261,1829-1834.
25. Imamoto-Sonobe, N., Yoneda, Y., Iwamoto, R., Sugawa, H., and Uchida, T. (1988) Proc. Natl. Ad. Sci. USA 85,3426-3430. 26. Yoneda, Y., Arioka, T., Imamoto-Sonobe, N., Sugawa, H., Shimonishi, Y., and Uchida, T. (1987) Exp. Cell Res. 170,439-452. 27. Furusawa, M., Yamaixumi, M., Nishimura, T., Uchida, T., and Okada, Y. (1976) Methods Cell Bid. 16,73-80.
CELL RESEARCH
198,107-114
(1992)
Escherichia co/i RecA Protein Modified with a Nuclear Location Signal Binds to Chromosomes in Living Mammalian Cells MITSURU KIDO, YOSHIHIRO YONEDA,’ MAHITO NAKANISHI, TS~YOSHI UCHIDA,’ AND YOSHIO OKADA Institute for Molecular and Cellular Biology, I-3 Yam&a-Oku, Suita, Osaka 565, Japan
We tried to make a well-characterized bacterial protein function in mammalian cell nuclei. For this purporn we chose Emherichia coli RecA protein and fused its carboxy terminus to the nuclear location signal of SV40 large T-antigen by oligonucleotide-dependent modification of the gene. When injected into the cytoplasm, the modified RecA protein (T-RecA for the T-antigen signal) accumulated efficiently in the nuclei, whereas the wild-type RecA protein remained in the cytoplasm. The T-RecA protein retained its original in uiuo activity, judging from the finding that uv-sensitive bacteria (recA- E. cola became uv-resistant on transformation with the T-recA plasmid as well as the recA plasmid. For expression of the T-recA gene in mammalian cells, the 5’ region was replaced by the chicken Bactin promoter and Kozak’s initiation signal. A high level of expression was observed when Chinese hamster ovary (CHO-Kl) cells were transfected with this plasmid. Indirect immunofluorescence examination revealed that the T-RecA protein in nuclei of mammalian cells bound to chromatin. o 1992 Academic PMS, IUC.
MATERIALS
INTRODUCTION
It is very interesting to make an exogenous protein function in living mammalian cells, especially when the protein has well-characterized, biochemically important activity. Introduction of such proteins into mammalian cells should be helpful in understanding some functions of the cells. Many bacterial proteins have been isolated and characterized and so may be suitable for this purpose. We thought that an Escherichia coli protein might be functional in mammalian cells because studies showed that fibroblasts from a patient with xeroderma pigmentosum, an autosomal recessive human genetic disease of uv repair, showed little or no unscheduled DNA synthesis (UDS) after uv irradiation but gained a normal level 1 To whom correspondence and reprint requests dressed. ’ Dr. Tsuyoshi Uchida died on May 3,1989.
of UDS after concomitant treatment with bacteriophage T4 endonuclease V and HVJ (Sendai virus) [l, 21. We focused attention on the E. coli RecA protein which plays a major role in genetic recombination in E. coli [3,4]. In vitro studies indicated pleiotropic activities of this protein. A remarkable finding was a strand exchange reaction in vitro, which required continuous ATP hydrolysis. The recA mutations in E. cob not only block recombination completely, but also lead to pleiotropic effects, such as high sensitivity to uv and X-ray irradiation and enhanced DNA breakdown after uv irradiation. We developed a system in which functional bacterial RecA protein could be introduced efficiently into mammalian cell nuclei. We found that the RecA protein itself is unable to enter the nucleus. Therefore, we modified it to a nuclear protein with the nuclear localization signal peptide of SV40 large T antigen discovered by Kalderon et al. [5,6]. We demonstrated association of this modified RecA protein (T-RecA) with chromosomes by the indirect immunofluorescence method.
should
be ad-
107
AND METHODS
Cell lines. CHO-Kl cells (a Chinese hamster ovary cell line) were obtained as described [7]. HEL (human embryonic lung) cells were a gift from Dr. A. Tada [8]. Plasmids. pTM2 (Fig. 2) containing the recA gene was a generous gift from Dr. Hideyuki Ogawa (Department of Biology, Faculty of Science, Osaka University) [9, lo]. pAct(X) (see below and Fig. 4) has the chicken &actin promoter and is a derivative of pact-c-myb (a gift from Dr. Shunsuke Ishii, Tsukuba Life Science Center) [ll]. ~220.2 (Fig. 4) was obtained from Dr. Bill Sugden (University of Wisconsin) [ 121. Syntheses of oligonuckotides. Oligonucleotides were synthesized in automated DNA synthesizers Models GENET A-III (Nihon Zeon), 381A (Applied Biosystems), and 391 PCR-MATE (Applied Biosysterns). The preparations were then purified by high performance liquid chromatography or electrophoresis in 20% acrylamide gel containing 8 M urea. Oligonucleotide-directed mutagenesis. Oligonucleotide-directed mutagenesis was carried out as described by Zoller and Smith [13] and Kunkel et al [ 141. DNA sequencing for confirmation was done in principle as described by Sanger et al. [15] with use of a kit (Sequenase version 2.0, United States Biochemical Corp.). Constrzxtion of pT-RecA and pATR4. pT-RecA, an expression plasmid of E. coli, was constructed to modify the recA gene with the 0014~4827/92%3.00
Copyright 0 1992 by Academic Press, Inc. All rights of reproduction in any form reserved.
KID0
ET AL.
FIG. 1. Subcellular localizations of the RecA protein and the T-RecA protein microinjected into the cytoplasm of HEL cells. (A) RecA protein; (B) T-RecA protein. (a, c) Phase-contrast microscopy; (b, d) fluorescence microscopy. Cells were incubated for 30-60 min after microinjection and then fixed and treated with anti-RecA antibody followed by a FITC-conjugated second antibody.
nuclear location signal of the SV40 large T antigen (Pro-Lys-LysLys-Arg-Lys-Val-Glu-Asp-Pro) (Fig. 2). Briefly, the region of the recA gene downstream from the DdeI site near the 5’-end was replaced by synthetic nucleotides in which the carboxyl terminus of the RecA protein was fused with the nuclear location signal. For expression of the T-RecA protein in mammalian cells, pATR4 was constructed (Fig. 4). In pATR4, the T-recA gene was expressed by the chicken @-actin promoter and translation was initiated by Kozak’s sequence. The selection marker hph (hygromycin B phosphotransferase) was also introduced. pAct(X), a derivative of pact-c-myb, was used as a mammalian expression vector. For construction of pAct(X) the vector was replaced by pUC19, and a 2.0-kb NcoI-XbaI fragment (c-myb gene) was replaced by a XhoI linker. Purification of RecA and T-RecA proteins from E. coli. The E. coli C600 strain harboring pTM2 and pT-RecA (Fig. 2), respectively, was induced to overproduce RecA and T-RecA proteins by treatment with nalidixic acid (60 pglml) and the proteins were purified as described [16]. Both proteins could be purified by the same procedure. Microinjection. Samples of 3-5 mg/ml of RecA or T-RecA protein were dialyzed against modified phosphate-buffered saline (reverse PBS: 2.7 mM NaCl, 137 m&f KCl, 8.1 m&f NarHPO,, 1.5 m&f KH,PO,), filtered through a Millipore filter (filter type HV, pore size
0.45 pm), and microinjected into the cytoplasm of HEL cells with a micromanipulator (Narishige Scientific Instrument, Tokyo, Japan) as described [ 171. Transient and stable transformation of CHO cells. CHO cells were transfected with plasmid DNA (pATR4) to express the T-recA gene. For transient expression, the CaPO, method 1181was used. For stable transformation, electroporation [19] was used and then the cells were selected in selection medium containing 600-600 ag/ml hygromycin B (Boeringer-Mannheim) for 10 days, and stably transformed colonies were cloned (El-E39). The Indirect immunofluorescence stainings of RecA and T-RecA. first antibody against RecA protein was raised in rabbits by repeated intradermal injections of purified RecA protein with Freund’s adjuvant (Difco). The second antibody, fluorescein isothiocyanate (FITC)-conjugated goat anti-rabbit IgG, was purchased from Cappel Co. and filtered through a Millipore filter (type HV, 0.45 pm) before use. Cells were fixed with 3.7% formaldehyde in PBS at room temperature for 15 min, permeabilized with cooled 100% methanol for 5 min, and blocked by treatment with blocking solution (2 mg/ml ovalbumin, 50 mM glycine, and 2% goat serum in PBS) for at least 30 min. A volume of 50 ~1 of the first antibody diluted 200-fold with blocking solution was applied to each coverglass and the cells were kept at 4’C
RecA WITH
\
A NUCLEAR
LOCATION
SIGNAL
BINDS
109
CHROMOSOMES
red 3’and + nuclear transport s@nal 367
bp
NGCLNAIl glu gas ctt
gly g(rc ccg
val gta cat
ala irca Cgt
clu mm ctt
thr act tya
a80 sac ttg
olu ma ctt RecA
Val
0111 Aep
OTT CM
GAG CTC
GAT CTA
Pro
tbr b CCG taa GGC att
LOCATION
SIGNAL
asp
Ias
IYS
lays
Arc
wat eta
MO Trc
AM TIT
MA TIT
COT AM OCA TIT
I carboxyl
Lps
tarmlnus
ter taK (It.2
tcf.yatgtt
tc.¶t.scacaa actatgtgtt
Kggcgcatct CCCgCgtaCa
FIG. 2. Modification of the recA gene with the nuclear location signal of SV40 large T-antigen. (A) Construction of pT-recA; The recA gene was fused with the nuclear location signal. Twelve oligonucleotides were annealed to parts B, C, and D, and these parts were ligated to form part BCD (see below). The synthetic DNA region (BCD) was then ligated with an EcoRI-DdeI 205-bp fragment of the recA gene (part A). Finally part A-BCD was ligated with a 1.8-kb 5’-fragment of recA and vector pUC19. (B) DNA sequence of the synthetic DNA region (part BCD). The nuclear location signal of 10 amino acid residues was fused at the carboxyl terminus of the RecA protein. The 3’-flanking region for termination of mRNA was the same as in the original recA gene.
overnight. They were then washed six times with cooled PBS and incubated at room temperature for at least 1 h with 50 ~1 of the second antibody diluted 100-fold with blocking solution.. They were then washed six times with cooled PBS and mounted with Mowiol 488(Hoechst) containing 1,4-diazobicyclo-[2.2.2]-octane (DABCO) [20]. Survival curves of uv-irradiated E. coli. The bacterial strains used were K12 C600 (F-, thi-1, thr-1, leuB6, lacY1, tonA21, supE44, h-), and MV1184 [ara, A(lac-proAB), rpsL, thi, (@3OlacZAM15), A(srlrecA)306::TnlO(tet’), F’(traD36, proAB+, lacF, lacZAM15)]. Strain C600 is the recA+ wild type, while MV1184 is a recA- mutant type. Bacteria in the log phase of growth were suspended in 0.1 M MgSO, at a concentration of l-2 X 10’ cells/ml and irradiated with a 15-W Toshiba germicidal lamp at a distance of 45 cm. The dose in this condition is 1.9 J/mz/s (19 erg/mm2/s). Survival curves were obtained as described [21]. RESULTS
RecA protein with a nuclear location signal was eficiently transported into the nucleus. Chromatins are shielded by nuclear membranes in mammalian cells, so if the RecA protein has no ability to pass through these membranes it cannot interact with genes. We first examined whether the RecA protein could accumulate in the
nucleus. When the purified RecA protein was microinjetted into the cytoplasm of HEL cells, it remained in the cytoplasm and did not enter the nucleus, as shown by indirect immunofluorescence staining (Fig. 1A). In eukaryotic cells, all nuclear proteins are initially synthesized in the cytoplasm and then rapidly transported into the nucleus [22-241. The precise mechanism of nuclear transport is not fully understood, but active transport has been suggested [25]. Moreover, Kalderon et al. reported that a short amino acid sequence of the SV40 large T antigen, Pro-Lys-Lys-Lys-Arg-Lys-Val, specifies its nuclear translocation [5, 61. Conjugation with this signal peptide has also been shown to direct transport of exogenous protein into the nucleus [26]. On the basis of these reports, we added the nuclear location signal of the SV40 large T antigen to RecA protein by oligonucleotide-dependent modification of the recA gene (pT-RecA, Fig. 2). This modified RecA protein (T-RecA) was produced in E. coli and purified as described under Materials and Methods. Purified TRecA protein had a slightly larger molecular weight than the RecA protein, as judged by SDS-PAGE (data
110
&I
4 6 12 16 UV Irradiation (J/m2 ) FIG. 3. Demonstration of comparable activity of the T-RecA protein to that of the wild-type RecA protein by uv-survival curves. MV1184 is a recA- strain while C600 is a recA+ strain. MV1164 was transformed with two plasmids: one was pT-RecA (Fig. 2), expressing the T-RecA protein, and the other was pRecA, expressing the intact RecA protein. pRecA was constructed to adjust the copy number in E. coli to that for pT-RecA. Since the vector of pT-RecA was high copy pUC19, the vector of pTM2 (Fig. 2) was replaced by the same pUC19 to yield pRecA. For construction, a 3.2-kb BamHI fragment containing the wild-type recA gene from pTM2 was cloned into the BamHI site of pUC19.
not shown). When the purified T-RecA protein was microinjected into the cytoplasm of HEL cells, it accumulated in the nucleus, as shown by indirect immunofluorescence staining (Fig. 1B). The T-RecA protein was as active as the RecA protein. Modification sometimes leads to inactivation
of a protein. Therefore, we examined the activity of the T-RecA protein in E. coli in vivo. As the RecA protein plays central roles in recombination and repair, recAmutants of E. coli show high sensitivity to uv-light irradiation [3]. But transformation with a plasmid encoding the recA gene complements this high sensitivity. Therefore, we compared the survival curves of recA- E. coli transformed with pT-RecA and with pRecA to determine whether the T-RecA protein was inactivated during peptide modification. Figure 3 showed that a transformant of a recA- strain (MV1184) with pT-RecA became as uv-resistant as a transformant with pRecA. This finding indicates that the T-RecA protein was as active as the RecA protein in E. coli in vivo. T-RecA protein expressed in cells was bound to chromosomes. Injection with a microcapillary is suit-
able for determining the subcellular localization of proteins of interest, but it is not a suitable method for studying the effects of the T-RecA protein in mammalian cells because it is not possible to inject the protein into a
AL.
sufficiently large number of cells. Therefore, we cloned the T-recA gene into a mammalian expression vector. Figure 4 shows the procedure used. First, the T-recA gene was cloned into the vector pAct(X) with 5’-, 3’flanking bacterial sequences (pATRl), then these redundant sequences were deleted (pATR2, pATR3), and finally, as a selection marker, the hygromycin B phosphotransferase gene hy$ was added (pATR4). The bacterial Shine-Delgarno sequence was replaced by Kozak’s sequence during oligonucleotide-directed deletion (pATR2, Fig. 4B). As shown in Fig. 5, when the resulting plasmid pATR4 was transfected into CHO cells, the T-RecA protein was strongly expressed in the cells and efficiently transported into their nuclei. Figure 5A shows the transient expression achieved by the CaPO, method. Several percent of the total cells in samples examined at 24,48, and 96 h expressed the T-RecA protein. At 24 h after transfection, scattered cells expressed T-RecA, but by 96 h colonies of cells expressing T-RecA were observed (Fig. 5A). If the protein had harmful effects, stable transformants expressing T-RecA would not be obtained. But cells expressing T-RecA stably were obtained by hygromycin B selection. Figure 5B shows one clone (clone E13) in which 100% of the progeny cells expressed TRecA. Moreover, although expression of the T-RecA protein was high, cell growth seemed to be normal and stable. Figure 6 shows indirect immunofluorescence staining of T-RecA in stable transformants (clones El3 and E23) in the late mitotic phase. The cells were dividing into two daughter cells and condensed chromosomes were stained as a thick band in the center of the rounded cells. If the T-RecA protein acts on genes, it must interact with chromatin, so this result provided clear evidence for the interaction of the T-RecA protein with chromosomes. DISCUSSION
In this study we introduced E. coli RecA protein into mammalian cells. For its efficient transport into the nuclei, modification of its carboxyl terminus with a nuclear location signal was necessary. As this modification did not cause loss of activity of RecA, this technique should be useful in studies on the effects of other exogenous proteins in the nuclei of mammalian cells, especially proteins expected to have important activities on genes or chromatins. In the case of T4 endonuclease V, even on its introduction into the cytoplasm with HVJ (Sendai virus), it functioned in the nucleus. T4 endonuclease V is sufficiently small (about 20,000 Da) to allow its passive entry into nuclei. Although the molecular weight of the RecA protein is only 38,060 Da, the smallest active form
RecA WITH
A NUCLEAR
LOCATION
SIGNAL
BINDS
111
CHROMOSOMES
A
pAot(X)
4.lkb
T-rrcA !naoflbn pATR1
6.3kb
ongonuchot-ted
B dewoIl
d&on
Met.Ala.Ile.Asp.Glu.Asn ~crcCocAG/CCA,o~~.ATG.GCT.A~.GAC.~.MC 18
#9-aCtfro \ \
7.3kb
18
.er recA
intron
HT
coding
sequence
OligOllUCleOlide
S-flanking region
FIG, 4. Construction of plasmids for expression of T-RecA in mammalian cells. (A) The T-recA gene was cloned under the control of chicken &actin promoter. pAct(X) is an expression vector for mammalian cells (see Material8 and Methods). The T-recA gene (open box, recA, 5’ shaded, nuclear location signal) with 5’- and 3’-bacterial flanking sequences (closed boxes) was cloned into the XftoI site of pAct(X) (pATR1). Redundant bacterial sequences were then completely removed by oligonucleotide-directed deletion (pATR2, see below) and by digestions with XbaI and XhoI (pATR3). Finally a 2.0-kb NruI-Hind111 fragment of the hygromycin B phosphotransferase gene (hyga) was inserted as a selection marker (pATR4). Poly A (*) means poly(A) additional signal. (B) Details of oligonucleotide-directed deletion/insertion. For deletion of 5’-bacterial flanking sequences (0.9 kb, bold triangular line) from pATR1, the vector was replaced by pUCll8 and doublestranded DNA was converted to single-stranded DNA. The ssDNA was then hybridized with the 44-mer oligonucleotide. By this hybridization, the flanking region for deletion was looped out, and the @actin intron and recA coding sequence were hybridized adjacent to the eukaryotic initiation signal (Koxak’s sequence). Ordinary oligonucleotide-directed mutagenesis was then carried out as described [13,14]. In the 44-mer oligonucleotide the mark (/) indicates a splicing acceptor of the intron.
was reported to be a tetramer [9], and its apparent mo- plasmid that expressed the T-recA gene (Figs. 4 and 5). lecular weight estimated by gel filtration was much As shown in Fig. 1, direct microinjection was suitable for larger (M. Kido, unpublished observation). Many other injecting large amounts of protein into a few cells, cerprocaryotic proteins that interact with DNA show simi- tainly, and the erythrocyte-ghost fusion method was suitable for introducing small amounts of protein into lar aggregation. When the RecA protein was microinjetted into the cytoplasm, it remained in the cytoplasm many cells, but was not suitable for introducing proteins and did not enter the nucleus (Fig. 1A). But our results of large molecular weight. The third method was someshowed that after modification with a nuclear location what different from these two methods since it did not signal, even large molecular weight proteins such as involve direct injection of the protein. But as shown in RecA can be transported into nuclei (T-RecA, Figs. 2 Fig. 5, it was very effective for introducing large and 1B). amounts of protein into a large number of cells. RecA protein without any modification could be inReferring to data described here, one can choose an jected directly into nucleus. But nuclear microinjection appropriate method to introduce an exogenous protein usually damages cells, and the amount of material in- into living cells. If the protein is easily obtained, direct jected into the nucleus must be far less than the amount microinjection and erythrocyte-ghost-mediated injecinjected into the cytoplasm. Furthermore, it is impossi- tion are available. Direct microinjection of a protein ble to inject the protein into nuclei of a larger number of into the nucleus is not impossible, but modification of cells to study its effect. the protein by a nuclear location signal facilitates its After modification of the RecA protein with a signal nuclear transport. Moreover, if the gene has already peptide, our next problems were how to introduce it into been cloned, the third method used here is also availcells and what dose would be suitable. We tried three able. methods of injection: direct microinjection (Fig. l), inAfter efficient introduction of the T-RecA protein jection by erythrocyte-ghost fusion with HVJ (Sendai into nuclei, we demonstrated its direct association with virus) [27] (data not shown), and transfection with a chromosomes (Fig. 6). If the T-RecA protein in nuclei
KID0
ET AL.
FIG. 5. Transient and stable expressions of the T-RecA protein in CHO cells. Cells were transformed with pATR4 (Fig. 4) in which the T-recA gene was cloned into a mammalian expression vector. The CaPO, precipitation method and electroporation were used for transient and stable expression, respectively, as described under Materials and Methods. Cells were stained by the indirect immunofluorescence method with anti-RecA antibody. (A) Transient expression at 96 h, (B) stable expression (clone E13). (a, c, e) Phase-contrast microscopy; (b, d, f) fluorescence microscopy.
had been present in the nucleoplasm and had no affinity to chromatin, the figure in the mitotic phase would have been different: the condensed chromosomes in the center of mitotic cells would have shown no fluorescent
signal and the fluorescence would have been observed in the surrounding area, because there is no nuclear envelope in this phase. Therefore, Fig. 6 indicates that the T-RecA protein bound to mammalian chromosomes.
RecA WITH
FIG. 6. Indirect stable transformants
A NUCLEAR
LOCATION
SIGNAL
We thank Drs. K. Ueda, H. Harada, and K. Kohno for help in syntheses of oligonucleotides. We also thank Dr. K. Kohno for valuable suggestions. This work was supported by grants from the Ministry of Education, Science and Culture of Japan and the Toray Science Foundation.
K., Sekiguchi,
M., and Okada, Y. (1975) Proc. Natl.
Tanaka,
K., Hayakawa,
H., Sekiguchi,
9.
Ogawa, T., Wabiko, H., Tsurimoto, T., Horii, T., Masukata, H., and Ogawa, H. (1978) Cold Spring Harbor Symp. Quant. Biol. 43, 909-915.
10.
5. 6. 7.
Kalderon, D., Richardson, W. D., Markham, A. F., and Smith, A. E. (1964) Nature 311,33-38. Kalderon, D., Roberts, B. L., Richardson, W. D., and Smith, A. E. (1964) Cell 39,499~509. Kohno, K., Uchida, T., Mekada, E., and Okada, Y. (1985) Somatic Cell Mol. Genet. 11, 421-431.
Horii,
T., Ogawa, T., and Ogawa, H. (1986) Proc. Natl. Acad. Sci.
USA 77,313-317. 11. 12.
Nishina, Y., Nakagoshi, H., Imamoto, F., Gonda, T. J., and Ishii, S. (1989) Nuc.!& Acids Res. 17,107-117. Yates, J. L., Warren, N., and Sugden, B. (198.5) Nature 313, 812-815.
13.
Zoller, M. J., and Smith, M. (1983) in Methods in Enzymology (Wu, R., Grossman, L., and Moldave, K., Eds.), Vol. 106, pp. 468-500, Academic Press, San Diego.
14.
Kunkel, T. A., Roberts, J. D., and Zakour, R. A. (1967) in Methods in Enzymology (Wu, R., and Grossman, L., Eds.), Vol. 154, pp. 367-382, Academic Press, San Diego.
M., and Okada, Y. (1977)
Proc. Natl. Acad. Sci. USA 74,2958-2962. 3. Clark, A. J. (1973) Annu. Reu. Genet. ‘7,67-36. 4. Radding, C. M. (1982) Annu. Rev. Genet. 16,405-437.
F., Toba, M., and Tada, A. (1979) Arch. Virol. 60,347-
Taguchi, 351.
Acad. 5%. USA 72,4071-4075. 2.
phase of
8.
REFERENCES Tanaka,
CHROMOSOMES
immunofluorescence staining of T-RecA in the late mitotic phase. Pairs of daughter cells in the late mitotic (clones El3 and E23) are shown, (A, B) Phase-contrast microscopy; (a, b) fluorescence microscopy.
Thus our results indicate that this system should be useful for studies on the effects of the T-RecA protein on mammalian chromosomes. Moreover, the same procedure is available for introduction of other bacterial proteins into the mammalian nucleus.
1.
BINDS
15.
Sanger, F., Nicklen,
S., and Coulson,
A. R. (1977) Proc. Natl.
Acad. Sci. USA 74,5463-5467. 16.
Kuramitus,
S., Hamaguchi,
K., Ogawa, T., and Ogawa, H. (1981)
J. Biochem. 90,1033-1045. 17.
Yamaizumi,
M., Horwich,
A. L., and Ruddle, F. H. (1963) Mol.
Cell. Biol. 3, 511-522. 18. 19.
Chen, C., and Okayama, H. (1967) Mol. Cell. Btil. 7.2745-2752. Kido, M., Miwatani, H., Kohno, K., Uchida, T., and Okada, Y. Cell Struct. Funct., in press.
114
KID0
20. Harlow, E., and Lane, D. (1988) Antibodies: A Laboratory Manual, p. 418, Cold Spring Harbor Laboratory, New York. 21. Miller, J. H. (1972) Experiments in Molecular Genetics, pp. 121124, Cold Spring Harbor Laboratory, New York. 22. Yamaizumi, M., Uchida, T., Okada, Y., Furusawa, M., and Mitsui, Y. (1978) Nature 273,782-7&k 23. Sugawa, H., Imamoto, N., Wataya-Kaneda, M., and Uchida, T. (1985) Exp. Cell Res. 159,419-429. Received June lo,1991 Revised version received September 5, 1991
ET AL. 24. Tsuneoka, M., Imamoto-Sondoe, N., and Uchida, T. (1986) J. Biol. Chem.
261,1829-1834.
25. Imamoto-Sonobe, N., Yoneda, Y., Iwamoto, R., Sugawa, H., and Uchida, T. (1988) Proc. Natl. Ad. Sci. USA 85,3426-3430. 26. Yoneda, Y., Arioka, T., Imamoto-Sonobe, N., Sugawa, H., Shimonishi, Y., and Uchida, T. (1987) Exp. Cell Res. 170,439-452. 27. Furusawa, M., Yamaixumi, M., Nishimura, T., Uchida, T., and Okada, Y. (1976) Methods Cell Bid. 16,73-80.