2033
Journal of Cell Science 112, 2033-2041 (1999)
Printed in Great Britain © The Company of Biologists Limited 1999
JCS0385
A nuclear localization signal can enhance both the nuclear transport and
expression of 1 kb DNA
James J. Ludtke, Guofeng Zhang, Magdolna G. Sebestyén and Jon A. Wolff*
Departments of Pediatrics and Medical Genetics, Waisman Center, University of Wisconsin-Madison, 1500 Highland Ave.,
Madison, WI 53705, USA
*Author for correspondence (e-mail: [email protected])
Accepted 7 April; published on WWW 26 May 1999
SUMMARY
Although the entry of DNA into the nucleus is a crucial
step of non-viral gene delivery, fundamental features of
this transport process have remained unexplored. This
study analyzed the effect of linear double stranded DNA
size on its passive diffusion, its active transport and its
NLS-assisted transport. The size limit for passive diffusion
was found to be between 200 and 310 bp. DNA of 310-1500
bp entered the nuclei of digitonin treated cells in the
absence of cytosolic extract by an active transport process.
Both the size limit and the intensity of DNA nuclear
transport could be increased by the attachment of strong
nuclear localization signals. Conjugation of a 900 bp
expression cassette to nuclear localization signals
increased both its nuclear entry and expression in
microinjected, living cells.
INTRODUCTION
Rouault et al., 1998). Our laboratory has found that while the
non-covalent association of an NLS with plasmid DNA (using
another method) can increase its efficiency of expression, the
mechanism was not by increasing DNA nuclear entry (Fritz et
al., 1996). Inferring that the NLS has to be tightly associated
with the gene, we developed a new approach for covalently
attaching NLSs to DNA (Sebestyén et al., 1998). The
covalently-attached NLS was functional in the digitonintreated cell system but was unable to enhance DNA nuclear
entry in living cells after microinjection.
Size is likely a critical factor affecting the efficiency of DNA
nuclear entry. The structure of the nuclear pore imposes two
size limits on nuclear transport (Pante and Aebi, 1996). The
peripheral channels of the nuclear pore, about 9-10 nm
diameter, allow small solutes and proteins up to 50-60 kDa to
freely diffuse in and out (except if there is some kind of active
retention on either side). Larger proteins need an NLS in order
to be actively transported through the central channel of the
pore. Colloidal gold particles less than ~25 nm in diameter
could be transported into the nuclei of HeLa cells when
coated with SV40 NLS peptide-conjugated BSA, or with
nucleoplasmin, an endogenous karyophilic protein (Dworetzky
et al., 1988; Feldherr and Akin, 1990).
Previous studies have shown that small, single-stranded
oligonucleotides can diffuse into the nucleus (Chin et al., 1990;
Hagstrom et al., 1997; Leonetti et al., 1991). We explored the
size limit of double stranded DNA diffusion into the nucleus
by using linear DNAs of different sizes. We have also
examined the effect of NLS conjugated streptavidin on the
nuclear transport of biotinylated linear DNAs (up to 3 kb in
size) in digitonin permeabilized or microinjected cells. The
The transport of DNA into the nucleus of mammalian cells is
of importance to the study of DNA viruses such as adenovirus
and Herpes virus (Sebestyén et al., 1998; Whittaker and
Helenius, 1998; Kasamatsu and Nakanishi, 1998) and the
development of non-viral vectors for gene therapy. A
cornerstone of these efforts is the tremendous advances in our
understanding of how proteins and RNA are transported in and
out of the nucleus (Mattaj and Englmeier, 1998; Pennisi, 1998).
These macromolecules contain (or bind to proteins that
contain) nuclear localization signals (NLSs) or export signals
that bind to nuclear transport proteins (e.g. karyopherins α and
β), which mediate their directional movement through the
nuclear pore. Viral proteins also contain NLSs and the exact
mechanism by which they enable DNA nuclear entry is under
active study.
A variety of non-viral methods of gene transfer are under
development for gene therapy (Wolff, 1994; Wolff and
Trubetskoy, 1998). The therapeutic gene, usually in the form
of a plasmid DNA, must be transported from the exterior of the
cell to the nucleus in order for it to be expressed and have its
healing effect. While naked DNA (linear or plasmid) can
traverse the nuclear pores of intact nuclei (Dowty et al., 1995;
Hagstrom et al., 1997), the process is inefficient as compared
to that of viral genomes and karyophilic proteins. A manifest
approach to increase the efficiency of foreign DNA nuclear
entry is to link the DNA with an NLS. Several groups have
reported that the non-covalent association of an NLS with
DNA can enhance DNA nuclear transport (Collas and
Alestrom, 1997; Dean, 1997; Kaneda et al., 1989; Langle-
Key words: Nuclear localization signal, Expression, DNA transport
2034 J. J. Ludtke and others
extended form of the SV40 T antigen NLS (NLS/NL) was used
since it was found to be more efficient for the nuclear transport
of large macromolecules such as IgM (Hubner et al., 1997;
Rihs et al., 1991; Yoneda et al., 1992). Our results show that
this NLS/NL is able to enhance the nuclear uptake of DNA
both in digitonin treated and microinjected cells in a sizedependent way. The tight association of NLS/NL with a 900
bp DNA increased both its physical entry into the nucleus and
expression of an encoded marker gene, a finding with
therapeutic implications.
MATERIALS AND METHODS
DNA production
Fluorescein-labeled oligonucleotides were purchased from and PAGE
purified by Integrated DNA Technologies (Coralville, IA). DNAs of
different sizes were produced by using the polymerase chain reaction
(PCR) to amplify a small region of lambda DNA. 78, 118, 200, 318,
518 and 997 base pair (bp) PCR products were produced by using one
common primer (5′ GAC GGA TCC AAC CTG ACA AAG GTT
AAA TTA GAA A 3′) and a primer specific for each individual size:
78 bp (5′ GAC GGA TCC TTT GTT TTT ATC TGC GAG ATT ATT
T 3′), 118 (5′ GAC GGA TCC CTG ATG AAA CAA GCA TGT CAT
CGT A 3′), 200 (5′ ATCACCGAAAATTCAGGATAATGTGCAA 3′),
318 (5′ GAC GGA TCC GTC CTT GTC GAT ATA GGG ATG AAT
C 3′), 518 (5′ GAC GGA TCC TTT CTT TGA GTA ATC ACT TCA
CTC A 3′), and 997 (5′ GAC GGA TCC CTT CCG ATA GTG CGG
GTG TTG AAT G 3′). Both primers had a BamHI restriction site at
the 5′ end. The reaction mixture consisted of 10 mM Tris-HCl (pH
8.3), 50 mM KCl, 1.5 mM MgCl2, 25 µM dNTP (Pharmacia,
Piscataway, NJ), 0.8 µM of each primer, 0.2 pM template, and 2.5
units of Ex Taq (PanVera, Madison, WI) polymerase per 100 µl of
PCR reaction. Cycling parameters were calculated based on template
length and annealing temperature of the primers. Purification of
amplified labeled-DNA consisted of washing on a Millipore UltrafreeMC filter (30 kDa cut-off) in ‘base import buffer’ (20 mM Hepes, pH
7.3, 110 mM potassium acetate, 5 mM sodium acetate, 2 mM
magnesium acetate, 0.5 mM EGTA) diluted 10-fold. Samples were
washed 5-10 times (a minimum of 2,500 fold dilution of
contaminants), and selected filtrates were checked for the presence of
unincorporated fluorescent oligonucleotides using both agarose gel
electrophoresis and spectrofluorometry.
In order to produce DNA with biotin at only one end, the common
primer above (with the BamHI site) was used with primers specific
for 518 (5′ GAC GGT ACC TTT CTT TGA GTA ATC ACT TCA
CTC A 3′), 997 (5′ GAC GGT ACC CTT CCG ATA GTG CGG GTG
TTG AAT G 3′), 1500 (5′ GAC GGT ACC ATT TTT AGC CCA AGC
CAT TAA TGG A 3′), and 2000 (5′ GAC GGT ACC TTA GGC TTA
TCT ACC AGT TTT AGA C 3′) bp (which had an Acc65I site instead
of a BamHI site) in a PCR reaction as above.
DNA labeling
DNAs that were produced with 2 BamHI sites were digested with 0.52 units BamHI (New England Biolabs, NEB) per pmol DNA at 37°C
for 4 hours. The recessed 3′ ends of the digested DNA were then filled
using 0.2-0.6 units Klenow (exo-) (NEB) per pmol DNA (according
to the manufacturer’s instructions) in the presence of 10 µM biotin
dATP, Texas Red dCTP, unlabeled dTTP and unlabeled dGTP for 3
hours at 37°C. This resulted in a biotin and Texas Red molecule on
each end of the DNA. The Klenow was inactivated by a 10 minute
incubation at 75°C. Unincorporated nucleotide was removed by
washing the DNA 5-10 times (at least 2,500-fold dilution of the
contaminant) on a Millipore Ultrafree-MC filter (30 kDa cut-off) in
‘base import buffer’. A small aliquot of DNA was assayed for biotin
incorporation by binding to 5 µg of streptavidin. 80-100% of the
biotin-labeled DNA was shifted on an ethidium bromide stained 2%
agarose gel.
DNAs with one BamHI site and one Acc65I site were digested with
BamHI as above, but filled in with 10 µM biotin dATP, dCTP, dGTP
and dTTP. This DNA was then digested with Acc65I and filled with
10 µM Texas Red dCTP, dTTP, dATP and dGTP. This produced DNA
with a biotin on one end and a Texas Red molecule on the other. 7090% of the biotin-labeled DNA was shifted when assayed for
streptavidin binding on an ethidium bromide stained 1% agarose gel.
In order to equalize the number of fluorochromes on a per weight
basis, some DNAs were labeled by incorporating fluorescent
nucleotides during the PCR reaction as previously described
(Hagstrom et al., 1997).
The 900 bp green fluorescent protein (GFP) expression vector
consisted of a minimal CMV promoter (−117 to +17 bp in relation
to the transcriptional start site of the human immediate early
cytomegalovirus promoter obtained from base pairs 619-753 of the
pCI mammalian expression vector; Promega, WI) 5′ to the complete,
717-bp S65T GFP coding sequence (Clontech), which was followed
by a minimal polyA signal sequence (AAT AAA AGA TCT TTA
TTT TCA TTA GAT CTG TGT GTT GGT TTT TTT GTG TG)
(Briggs and Proudfoot, 1989). Production and labeling of this vector
was by PCR using a biotin-labeled 5′ primer (5′ GCA CCA AAA
TCA ACG GGA CTT TCC AAA ATG T 3′) and a Cy5-labeled 3′
primer (5′ CAC ACA AAA AAA CCA ACA CAC AGA TCT AAT
GA 3′).
Streptavidin-peptide conjugation
Streptavidin was covalently conjugated as previously reported
(Adam et al., 1992) to either a 39 amino acid peptide
(H-CKKKSSSDDEATADSQHSTPPKKKRKVEDPKDFPSELLS)
containing a functional SV40 large T antigen NLS (Kalderon et al.,
1984) (streptavidin-NLS/NL), or to a mutant version of it that
has a K to T amino acid change at position 24 (corresponding to
the K 128 position of the SV40 T antigen, known to be transport
deficient; Kalderon et al., 1984). The mutant peptide also has
an additional K near the amino terminus at position 2
(H-CKKKKSSSDDEATADSQHSTPPKTKRKVEDPKDFPSELLS)
(streptavidin-NLS/MNL) to keep the overall charge of the two
peptides equal. The average number of peptide molecules per
streptavidin monomer was estimated to be 2 (and thus, 8 for the
streptavidin tetramer) by analysis of mobility shift on SDSpolyacrylamide gel electrophoresis.
Streptavidin-NLS/NL or streptavidin-NLS/MNL was bound to
fluorescein biotin at a ratio of 50:1 on ice for 30 minutes. Removal of
unbound fluorescein biotin was accomplished by gel filtration on a
NAP 5 column (Pharmacia).
Streptavidin-NLS/NL or streptavidin-NLS/MNL was added to
biotinylated DNA at a molar excess of between 2 and 10.
Transport assay
Transport assays were performed as previously described (Adam et
al., 1992; Hagstrom et al., 1997) for one hour. Omission of ‘energy’
in a transport assay consisted of replacing the ATP, GTP, creatine
phosphate, and creatine phosphokinase with buffer. Wheat germ
agglutinin (WGA) treatment involved incubating the digitonin-treated
cells in import buffer containing 0.16 mg/ml WGA for 10 minutes at
room temperature and including 1 mg/ml WGA in the transport
solution.
Microinjection
HeLa cells grown on untreated glass coverslips were microinjected in
the cytoplasm using an IM 200 microinjector (Narishige, Tokyo)
(Dowty et al., 1995). All microinjected materials were co-injected
with 1 mg/ml of either 580 kDa fluorescein-labeled dextran (Sigma)
or 500 kDa Cascade Blue-labeled dextran (synthesized by Vladimir
Budker). The microinjected cells were fixed in 2% formaldehyde and
NLS enhanced DNA transport 2035
mounted in Gel/Mount. Nail polish was used to create a raised border
on the slide to prevent shearing of the cells.
Microscopy and quantitation
Images of the samples were collected by epifluorescence microscopy
on a Nikon Optiphot with a SenSys CCD Camera (Photometrics,
Tucson, AZ) using a ×100 oil plan apo objective with NA 1.4. In our
previous publication (Hagstrom et al., 1997), confocal microscopy
was used to determine the subcellular location of fluorescent DNA.
Based on comparisons of samples analyzed by both confocal
microscopy and epifluorescent microscopy, and given the large size
of the HeLa cell nuclei, it was determined that epifluorescence
microscopy was sufficient to ascertain whether the labeled DNA was
intranuclear.
When DNA is end labeled with one or two fluorochromes, different
size DNAs will have the same number of fluorochromes per DNA
molecule but different amounts of label per µg. Thus it was
problematic to perform direct quantitative comparisons among the
different DNA sizes. So, to compare different sized DNAs, qualitative
analysis of the cytoplasmic signal relative to the nuclear signal on an
individual cell basis was used to judge the efficiency of nuclear
accumulation. Unless otherwise indicated, the photomicrographs in
the figures show representative examples.
To compare DNAs of the same size, quantitative analyses were
used to compare the ratio of DNA transport with and without
streptavidin-NLS/NL (Fig. 5O). Mean pixel values in the center of
nuclei were determined and background values were subtracted.
Background-subtracted mean pixel values were averaged within
groups and their standard deviations calculated. Two-sample t-tests,
without the assumption of equal variances, were applied to the mean
pixel values. In order to protect against potential problems with nonnormality, the two-sample Wilcoxon test was also computed. If the
Wilcoxon P-value was larger than the t-test P-value, the Wilcoxon
P-value was used. Otherwise, the t-test P-values were used. The
averages for NLS samples were divided by the averages for nonNLS samples to obtain a ‘+ to –’ ratio seen in Fig. 5O. The standard
error of the ratio was calculated using the delta method (Agresti,
1990).
RESULTS
Nuclear accumulation of DNA is size dependent
Linear DNA fragments of various sizes were labeled with
Texas Red fluorophores and assayed for their ability to
accumulate in the nuclei of digitonin-permeabilized HeLa cells
in the absence of rabbit reticulocyte lysate (RRL) (Fig. 1). Our
previous study showed that linear DNA of ~1 kb in size is
transported into the nucleus by an active process that is
inhibited by RRL (Hagstrom et al., 1997). Seventy bp DNA
accumulated strongly in nucleoli and exhibited diffuse staining
throughout the rest of the nucleus, with occasional punctate
nuclear staining (Fig. 1A). The 110 and 200 bp DNAs
distributed similarly to 70 bp DNA, but had additional small
punctate accumulation in the non-nucleolar regions of the
nucleus (Fig. 1B,C). 310 bp and 510 bp DNA showed binding
to the nuclear membrane, and exhibited large and frequent
punctate staining in the nucleus with no nucleolar staining (Fig.
1D,E). The 1 kb DNA sample also had punctate DNA staining
in the nuclei (Fig. 1F) which was decreased as compared to the
510 bp DNA specimen (Fig. 1E). The 1.5 kb DNA sample
exhibited smaller and fewer nuclear spots (Fig. 1G), while the
2 and 3 kb DNAs had no punctate staining and minimal diffuse
staining (Fig. 1H,I).
Fig. 1. Effect of DNA size on sub-cellular distribution and nuclear
accumulation pattern in digitonin-permeabilized cells at 37°C with
‘energy’ but no RRL. 70 bp (A), 110 bp (B), 200 bp (C), 310 bp (D),
510 bp (E), 1 kb (F), 1.5 kb (G), 2 kb (H), 3 kb (I). 200-300 ng DNA
labeled by PCR incorporation of Texas Red fluorophores (A, B, DF), at one end with a Cy5 fluorophore (C) or at one end with one
Texas Red fluorophore (G-I) was used per well.
The ability of the different sized DNAs to enter the nuclei
of digitonin-permeabilized cells was also analyzed in the
presence of RRL and ‘energy’ at 37°C (see Fig. 5). The
addition of RRL did not prevent the nuclear accumulation of
the 70 to 200 bp DNAs (Fig. 5A,C,E) but inhibited the punctate
nuclear staining from the DNAs greater than 310 bp (Fig.
5F,H,J,L,N) as previously noted for 1 kb DNA (Hagstrom et
al., 1997).
The DNAs of different sizes were microinjected along with
fluorescently-labeled dextran (580 kDa) into the cytoplasm of
HeLa cells and harvested after about 4 hours (Fig. 2). Only the
2036 J. J. Ludtke and others
Fig. 3. Effect of DNA size on sub-cellular distribution and nuclear
accumulation pattern in digitonin-permeabilized cells at 4°C with
neither ‘energy’ nor RRL. 70 bp (A), 110bp (B), 200 bp (C), 310 bp
(D), 510 bp (E) and 1 kb (F). 200 ng DNA labeled by PCR
incorporation of Texas Red fluorophores (A-B and D-E), or by Cy5labeled primer incorporation (C). Representative examples are
displayed.
Fig. 2. Effect of DNA size on sub-cellular distribution and nuclear
accumulation pattern in cells at 37°C about 4 hours after
cytoplasmic injection. 70 bp (A), 70 bp + streptavidin-NLS/NL (B),
110 bp (C), 110 bp + streptavidin-NLS/NL (D), 200 bp (E), 310 bp
(F), 310 bp + streptavidin-NLS/NL (G), 510 bp (H), 510 bp +
streptavidin-NLS/NL (I), 1 kb (J), 1 kb + streptavidin-NLS/NL (K).
220 ng/µl - 1.75 µg/µl DNA labeled at both ends with one Texas
Red fluorophore (A-D and F), at one end with one Cy5 fluorophore
(E), or at one end with one Texas Red fluorophore (G-K) was
microinjected. Representative examples of transporting cells are
displayed.
cells in which the dextran was excluded from the nuclei were
included in the analyses. The 70 and 110 bp DNA accumulated
as a non-nucleolar haze in the nuclei (Fig. 2A,C). The 200 bp
DNA exhibited both a punctate nuclear accumulation as well
as a non-nucleolar haze (Fig. 2E,F). The 310 bp DNA sample
exhibited less nuclear accumulation (Fig. 2F) while 510 bp and
1 kb DNA samples had very little (Fig. 2H,J). DNAs of 1.5, 2
and 3 kb size were unable to efficiently enter nuclei under these
conditions (data not shown).
DNA can passively enter intact nuclei in a sizedependent manner
The DNAs of different sizes labeled with Texas Red
fluorophores were also assayed for their ability to enter the
nuclei of digitonin-treated cells at 4°C and without ‘energy’
(Fig. 3). The 70 bp DNA accumulated evenly throughout nuclei
(Fig. 3A), while 110 bp and 200 bp DNAs displayed both hazy
and small, punctate staining (Fig. 3B,C). The 310 bp, 510 bp
and 1 kb DNAs had a very small level of nuclear haze and little
of the punctate staining (Fig. 3D-F) that was prominent at 37°C
and with ‘energy’ (Fig. 1D-F). Nuclear rimming was more
prominent in nuclei exposed to DNA greater than 310 bp (Fig.
3). These results indicate that there is a transition in the
accumulation efficiency of 200 and 310 bp double-stranded
DNA at 4°C.
NLS enhanced DNA transport 2037
The effect of DNA size on NLS-mediated transport in
digitonin-treated cells
The following experiments were conducted to determine the
effect of DNA size on the nuclear entry of DNA conjugated to
a karyophilic protein. DNAs of various sizes were labeled with
a Texas Red fluorophore and a biotin at one or each end. The
DNA was mixed with streptavidin containing various
covalently-linked NLSs. If the NLS was the canonical SV40 T
ag NLS (PKKKRKV)(Yoneda et al., 1992), the DNA
aggregated because of charge interactions between the
negatively-charged DNA and the positively-charged NLSs
(data not shown). When streptavidin containing covalentlyattached NLS/NL (streptavidin-NLS/NL) or a non-functional
NLS (streptavidin-NLS/MNL) was mixed with linear DNA
containing a single biotin, more than 90% of the DNA was
shifted from its original position when assayed by agarose gel
electrophoresis (Fig. 4; compare lane 2 to lane 3 and 4, arrow
A). Biotinylation had no effect on the migration of the DNA
(data not shown). Aggregation was not observed with
streptavidin-NLS/NL or NLS/MNL (Fig. 4) because it contains
an extended form of the SV40 T ag NLS (labeled NLS/NL)
that is less positively charged (+1.5). On the basis of size, it
appeared that the streptavidin complexes mostly contained one
or two DNA molecules (Fig. 4; lanes 3 and 4, arrows B and C,
respectively).
The complexes of streptavidin-NLS/NL and biotinylated
DNA of various sizes were applied to digitonin-permeabilized
cells in the presence of RRL and ‘energy’ at 37°C (Fig. 5).
With streptavidin-NLS/NL, biotinylated DNA (up to 3 kb)
accumulated in the nucleus. 70 bp and 110 bp DNA
accumulated both with and without streptavidin-NLS/NL
present (Fig. 5A-D). DNA of sizes 310 bp and larger all formed
punctate staining in the presence of streptavidin-NLS/NL (Fig.
5G,I,K,M,O), while virtually no nuclear accumulation
occurred without streptavidin-NLS/NL (Fig. 5F,H,J,L,N). The
ratios of mean pixel values of NLS-mediated to non-NLSmediated transport are shown in Fig. 5P. All DNA sizes tested
exhibited a significant (P< 0.0002) enhancement of transport
(at least 2.5-fold). When the 510 bp biotinylated DNA was
bound to streptavidin-NLS/MNL, it was not able to enter nuclei
(data not shown). Without DNA, streptavidin-NLS/NL entered
the nucleus while streptavidin-NLS/MNL did not (data not
shown).
SDS/PAGE analysis indicated that the modified streptavidin
used in the experiments contained on average two peptides per
streptavidin monomer (data not shown) and therefore, eight
peptides per streptavidin tetramer. If the streptavidin monomer
Fig. 4. Effect of modified streptavidin on biotinylated DNA
migration in 1% agarose gel electrophoresis and subsequent ethidium
bromide staining. 1 µg of 1 kb ladder (NEB) (lane 1), 200 ng of 900
bp biotinylated and Cy5-labeled GFP expression vector (lane 2), 200
ng of 900 bp biotinylated and Cy5 labeled GFP expression vector +
111 ng streptavidin-NLS/NL (lane 3), 200 ng of 900 bp biotinylated
and Cy5 labeled GFP expression vector + 111 ng streptavidinNLS/MNL (lane 4). Arrow A shows unbound DNA, arrow B shows
streptavidin complexes with one DNA molecule, and arrow C shows
streptavidin complexes with two DNA molecules.
contained less than two peptides, DNA nuclear import was less
efficient (data not shown).
Wheat germ agglutinin (WGA) was used to determine
whether the enhanced uptake of the complexes of streptavidinNLS/NL and DNA was mediated via the nuclear pore (Forbes,
1992; Garcia-Bustos et al., 1991). Complexes containing
DNAs ranging in size from 70 bp to 510 bp were incubated
with digitonin-permeabilized cells with and without WGA.
WGA prevented the DNA of all sizes from entering the nuclei
resulting in an increased signal at the nuclear rim (data not
shown). The intensity of nuclear fluorescence with WGA was
significantly less (P≤ 0.007) than the fluorescence without
WGA for all sizes of DNA tested (data not shown).
Nuclear transport of DNA complexes with
streptavidin-NLS/NL in living cells
The complexes of streptavidin-NLS/NL and DNA of various
sizes were microinjected into the cytoplasm of HeLa cells and
Table 1. GFP expression in microinjected cells
GFP+/DNA+
GFP+/DNA−
GFP−/DNA+
Streptavidin
Total cells
Weak
Strong
Weak
Strong
Weak
Strong
NLS/NL
(percentages)
272
11
(4.0)
11
(4.0)
8
(2.9)
0
(0)
12
(4.4)
4
(1.5)
NLS/MNL
(percentages)
197
1
(0.5)
0
(0)
2
(1.0)
0
(0)
8
(4.1)
0
(0)
One day after the Cy5-labeled, biotinylated 900 bp GFP expression cassette complexed with streptavidin NLS/NL or streptavidin NLS/MNL was microinjected
in the cytoplasm of HeLa cells, the number of cells (without nuclear dextran) that contained GFP expression and/or nuclear DNA were determined. For
GFP+/DNA+ cells, weak and strong refer to the strength of both GFP and nuclear DNA signal which was correlated in these cells. For GFP+/DNA− cells, weak
and strong refer to the strength of the GFP signal. For GFP−/DNA+ cells, weak and strong refer to the strength of the DNA nuclear signal.
2038 J. J. Ludtke and others
Fig. 5. Effect of DNA size on its nuclear
accumulation in the presence of RRL in
digitonin-permeabilized cells at 37°C
and enhancement by binding
streptavidin-NLS/NL. 70 bp (A), 70 bp +
streptavidin-NLS/NL (B), 110 bp (C),
110 bp + streptavidin-NLS/NL (D), 200
bp (E) 310 bp (F), 310 bp + streptavidinNLS/NL (G), 510 bp (H), 510 bp +
streptavidin-NLS/NL (I), 1 kb (J), 1 kb +
streptavidin-NLS/NL (K), 2 kb (L), 2 kb
+ streptavidin-NLS/NL (M), 3 kb (N)
and 3 kb + streptavidin-NLS/NL (O).
200-300ng DNA labeled with one Texas
Red or Cy5 fluorophore and one biotin at
each end (A-D) or at one end (E-O).
Representative examples are displayed.
The averages of the mean pixel values in
the center of 7-31 nuclei were measured.
The values for streptavidin-NLS/NL
samples were divided by values without
streptavidin-NLS/NL to obtain a ‘+ to –’
ratio (P).
harvested after about 4 hours (Fig. 2). Complexes containing
70 bp - 510 bp DNAs all accumulated relatively efficiently with
the majority of the microinjected nuclei containing a strong
signal (Fig. 2B,D,G,I). Complexes with ~1 kb DNA still
accumulated, but only approximately 10% of the nuclei had
substantial staining to the degree shown in Fig. 2K. (see Table
1 for quantitation of a 900 bp DNA fragment). Complexes with
2 or 3 kb DNA yielded weakly-positive signals in only a few
percentage of nuclei (data not shown). The DNAs smaller than
310 bp entered nuclei relatively efficiently without any
streptavidin-NLS/NL (compare Fig. 2A,C,E,F). A 900 bp DNA
fragment did not enter the nucleus as efficiently when
complexed with non-functional streptavidin-NLS/MNL (Fig.
6F and Table 1).
NLS/NL enhancement of GFP expression after
microinjection
To better understand the relationship between DNA nuclear
transport and gene expression, a 900 bp GFP expression
cassette, which was labeled with a biotin on one end and a Cy5
fluorophore on the other end, was complexed with streptavidinNLS/NL or streptavidin-NLS/MNL. These complexes were
microinjected into the cytoplasm of HeLa cells and harvested
after 18-24 hours. Only cells with dextran-negative nuclei were
analyzed (to exclude cells in which the injected material
entered nuclei from either microinjection or cell division).
Complexes with functional streptavidin-NLS/NL expressed
GFP more efficiently than complexes with non-functional
streptavidin-NLS/MNL (Fig. 6 and Table 1). The extent of GFP
expression appeared to correlate with the amount of nuclear
DNA when the cell expressed at all. However, there were also
non-expressing cells with DNA accumulated in the nucleus
(Table 1). None of the cells injected with the streptavidinNLS/MNL complex had strong GFP or DNA staining. The 900
bp GFP DNA without any streptavidin also expressed poorly
after cytoplasmic microinjection (data not shown). Initial
studies suggest that GFP expression was not enhanced when
streptavidin-NLS/NL was complexed with a two kb DNA
fragment that contained a GFP expression cassette (data not
shown). When the nucleus was compromised during the
microinjection experiment (when dextran fluorescence was
observed in the nucleus), the 900 bp DNA bound to either
streptavidin-NLS/NL or streptavidin-NLS/MNL expressed
similarly (18.2% for NLS/NL and 17.7% for NLS/MNL).
The decreased DNA signal in Fig. 6F as compared to 6C
may be due to increased degradation of DNA that remains in
the cytoplasm as compared to that in the nucleus. Microscopic
examination of the microinjected cells at 15 minute, 4 hours,
and overnight indicated that there was a several-fold decrease
in the cytoplasmic signal from 15 minute to 4 hours, but only
a slight decrease thereafter.
DISCUSSION
The size of the linear DNA fragment had a substantial effect
on its pattern of nuclear accumulation. DNA less than or equal
to 200 bp distributed diffusely throughout nuclei in both
digitonin-treated (Fig. 1) and living, microinjected cells (Fig.
2). The ability of DNA less than 200 bp in length to enter the
nuclei of digitonin-treated cells at 4°C and without ‘energy’
(Fig. 3A-C) suggests that DNA of this size diffuses passively
NLS enhanced DNA transport 2039
Fig. 6. An NLS can enhance the transport and expression of DNA.
660 ng/µ l – 1 µ g/µ l of a biotinylated and Cy5 labeled 900 bp GFP
expression vector that was bound to either a functional streptavidinNLS/NL or a mutant (streptavidin-NLS/MNL) was cytoplasmically
microinjected with 500 kDa Cascade Blue dextran into HeLa cells
and harvested after 18-24 hours. DNA + streptavidin-NLS/NL, DAPI
channel (A), DNA + streptavidin-NLS/NL, FITC channel (B), DNA
+ streptavidin-NLS/NL, Cy5 channel (C), DNA + streptavidinNLS/MNL, DAPI channel (D), DNA + streptavidin-NLS/MNL,
FITC channel (E), DNA + streptavidin-NLS/MNL, Cy5 channel (F).
The signal in E is the same as background in cells without GFP. The
best examples are displayed.
into the nucleus. This would put the upper limit for passive
nuclear entry of DNA between 132 kDa (200 bp) and 204 kDa
(310 bp), which is larger than the upper molecular weight limit
for the passive nuclear entry of globular proteins. This may be
because naked, linear DNA does not form a globular particle
and has a persistence length of ~50 nm (Cantor and Schimmel,
1980). 200 bp DNA, which has a length of about 70 nm and a
diameter of about 2 nm, must pass through the nuclear pore
peripheral channels that are 9-10 nm in diameter and 100 nm
in length (Bustamante et al., 1994). Perhaps if the DNA is too
much larger than its persistence length, efficient diffusion
through the long skinny channel is inhibited.
DNAs between 310 bp and 1.5 kb in size exhibited less
nuclear haze and more punctate staining in the digitonintreated cells than did DNAs less than 310 bp (Fig. 1). Minimal
punctate nuclear staining was observed with the 2 or 3 kb
DNA. Thus the upper limit for the effective transport of naked
DNA in digitonin-treated cells is approximately 1.5 kb. The
cellular basis for this upper limit is not clear given that the
mechanism by which the naked DNA enters the nucleus is not
known. Our previous study demonstrated that the nuclear entry
of ~1 kb DNA required energy and was inhibited by WGA and
N-ethylmaleimide (NEM) (Hagstrom et al., 1997). The fact
that DNA transport did not require additional cytosolic factors
led us to postulate that DNA could interact directly with the
nuclear pore. Recently, it has been shown that the nuclear
transport of fragments of karyopherin β1 and β2 (transportin)
also do not require additional cytosolic factors (Kose et al.,
1997; Nakielny and Dreyfuss, 1998). While interactions of
these proteins with the nuclear pore have been identified, the
mechanism by which these and other proteins translocate
through the pore is unknown (Mattaj and Englmeier, 1998).
Nuclear rimming was observed for DNAs greater than or equal
to 310 bp (Figs 1, 3). Perhaps the nuclear translocation of DNA
greater than ~1.5 kb is hindered. Studies are in progress to
determine whether such DNA is sequestered at the nuclear
pores or other peri-nuclear structures.
As the size of DNA increases above 200 bp, the nuclear
transport of DNA in microinjected cells diverges from that in
digitonin-treated cells (compare Fig. 1 to Fig. 2). Very little
punctate staining was observed in cells microinjected with
uncomplexed 510 bp or 1 kb DNA. The addition of RRL to the
digitonin-permeabilized cells blocked the punctate nuclear
staining generated by the 510 bp or 1 kb DNA (Fig. 5), thus
paralleling the results in microinjected cells. RRL also altered
the intranuclear localization of the 70 and 110 bp DNAs
(compare Figs 1 and 5). Studies are in progress to identify the
factors in RRL that affect the nuclear transport of DNA in
digitonin-permeabilized cells.
The coupling of the karyophilic protein, streptavidinNLS/NL, to DNAs greater than 110 bp substantially increased
their DNA transport in the presence of RRL (Fig. 5). This result
is consistent with our previous study (that used the CPI
alkylating reagent to covalently attach many NLSs to DNA)
and indicates that RRLs inhibitory effect can be overcome by
an NLS. In both methods, there is a dose-response relationship
between the number of NLSs and the efficiency of nuclear
transport. For covalent modification using CPI, the DNA
needed to contain at least 3-4 NLSs/100 bp for efficient nuclear
entry (Sebestyén et al., 1998), while streptavidin containing 8
NLSs per tetramer transported DNA more efficiently than did
streptavidin containing 6 NLSs per tetramer (data not shown).
A possible structural correlative to this threshold is the dense
docking sites on the cytoplasmic fibrils that extend out from
the nuclear pore complex (Feldherr et al., 1984; Richardson et
al., 1988).
Linkage of DNA to streptavidin-NLS/NL via biotin
increased the nuclear transport of fluorescently-labeled DNA
in living cells after microinjection (Figs 2, 6). The overall level
of uptake was more marked for DNA less than 1.5 kb in length.
Since most of the streptavidin complexes contained 2 DNA
molecules (Fig. 4), this size limit may be different for single
DNA molecules.
The NLS/NL peptide was chosen for these studies because
it was more effective than the shorter SV40 T ag NLS in
enabling nuclear transport of IgM, a large protein (Jans and
Hubner, 1996; Yoneda et al., 1992). The importance of the
phosphorylation sites within the NLS/NL peptide for DNA
nuclear transport in living cells remains to be determined. Its
reduced positive charge (as compared to the +5 charge of the
minimal SV40 T ag NLS) could also reduce masking from
electrostatic interactions with negatively-charged DNA
(Henkel et al., 1992).
Fluorescently-labeled DNA is useful for determining the
relative amounts of DNA that physically enter the nucleus.
The fluorescent tag should preferably be covalently linked to
the DNA since even strong non-covalent DNA binders such
as TOTO (Molecular Probes) can move from the exogenous
DNA to endogenous chromosomal DNA (J. Wolff et al.,
unpublished results). Nonetheless, covalently-attached
fluorescent tags can also be problematic if DNA degradation
occurs. DNA degradation is not an issue in the digitonin-
2040 J. J. Ludtke and others
permeabilized cell system with or without RRL (Hagstrom et
al., 1997) but may be problematic for microinjection. The
present study used DNA which contained a fluorescent tag
and an NLS at opposite ends. If degradation had occurred in
the living cells, then the fluorescent tag would have separated
from the NLS and the nuclear fluorescent signal would not
be dependent upon the presence of an NLS. The fact that
fluorescently-labeled DNA alone or linked to streptavidin
containing a mutant-NLS control peptide entered the nucleus
less efficiently argues against the possibility that the
enhanced nuclear transport of the DNA and streptavidinNLS/NL complexes is due to DNA degradation. The
substantially decreased nuclear signal in cells microinjected
with complexes of 2 kb DNA and streptavidin-NLS/NL
complexes also argues against DNA degradation (data not
shown). In addition, streptavidin-NLS/NL complexed with
DNA containing fluoroscein and biotin at one end and Texas
Red at the other end was also used for microinjections. The
fluoroscein and Texas Red nuclear signals were congruent
(data not shown). Even though cytoplasmic degradation of
DNA may occur, the nuclear signal observed with
streptavidin-NLS/NL bound to biotin DNA is not a result of
degradation.
The effect of NLS conjugation on DNA expression is
another measure of DNA nuclear transport. Given that the
effect of streptavidin-NLS/NL was greater for DNA of a
smaller size, a minimal GFP expression cassette was designed
to be 900 bp. If the minimal cassette alone or conjugated with
a non-functional NLS was microinjected cytoplasmically,
approximately 1% of the HeLa cells whose nuclei were intact
expressed GFP. This low expression is mainly due to the
abbreviated CMV promoter. Linkage of this biotinylated
900 bp GFP expression cassette to the functional NLS
(streptavidin-NLS/NL) resulted in GFP expression in ~10% of
the cells in whose nuclei were intact (Table 1). Strong GFP
expression (i.e. Fig. 6B) was observed in 4% of the cells
injected with complexes containing the functional NLS but
never in cells injected with complexes containing mutant NLS.
These results are consistent with the increased nuclear entry of
fluorescent DNA conjugated to the streptavidin-NLS/NL (Figs
2, 5, 6). Several cells microinjected with DNA conjugated to
the functional NLS had strong nuclear DNA fluorescence but
no GFP expression (Table 1). Given that DNA degradation is
an unlikely explanation (see above paragraph), the delivered
DNA may not be located within a nuclear compartment capable
for expression or the cell is not in a proper state for expression.
These effects may be heightened by the weakness of the
expression vector.
In conclusion, this study demonstrates that the covalent
modification of a DNA (by the attachment of the biotin ligand)
can enable its increased expression in mammalian cells. A
recent report describes the ability for an NLS to increase
expression but it did not ascertain the NLSs effect on actual
DNA transport into the nucleus by physical methods (Zanta et
al., 1999). Our report is the first to correlate the effect of an
NLS on both DNAs nuclear transport and expression in the
same cell.
We thank Vladimir Budker for discussions and for preparing the
cascade blue-dextran. This work was supported in part by NIH grant
R01-DK49117.
REFERENCES
Adam, S. A., Sterne-Marr, R. and Gerace, L. (1992). Nuclear protein import
using digitonin-permeabilized cells. Meth. Enzymol. 219, 97-110.
Agresti, A. (1990). Categorical Data Analysis. Wiley, New York.
Briggs, N. L. and Proudfoot, N. J. (1989). Definition of an efficient synthetic
poly(A) site. Genes Dev. 3, 1019-1025.
Bustamante, J. O., Liepins, A. and Hanover, J. A. (1994). Nuclear pore
complex ion channels (Review). Mol. Membr. Biol. 11, 141-150.
Cantor, C. and Schimmel, P. (1980). Biophysical Chemistry (ed. A. C.
Bartlett). WH Freeman, San Francisco.
Chin, D. J., Green, G. A., Zon, G., Szoka, F. C., Jr. and Straubinger, R. M.
(1990). Rapid nuclear accumulation of injected oligodeoxyribonucleotides.
New Biol. 2, 1091-1100.
Collas, P. and Alestrom, P. (1997). Nuclear localization signals: a driving
force for nuclear transport of plasmid DNA in zebrafish. Biochem. Cell Biol.
75, 633-640.
Dean, D. A. (1997). Import of plasmid DNA into the nucleus is sequence
specific. Exp. Cell Res. 230, 293-302.
Dowty, M. E., Williams, P., Zhang, G., Hagstrom, J. E. and Wolff, J. A.
(1995). Plasmid DNA entry into postmitotic nuclei of primary rat myotubes.
Proc. Nat. Acad. Sci. USA 92, 4572-4576.
Dworetzky, S. I., Lanford, R. E. and Feldherr, C. M. (1988). The effects of
variations in the number and sequence of targeting signals on nuclear uptake.
J. Cell Biol. 107, 1279-1287.
Feldherr, C. M., Kallenbach, E. and Schultz, N. (1984). Movement of a
karyophilic protein through the nuclear pores of oocytes. J. Cell Biol. 99,
2216-2222.
Feldherr, C. M. and Akin, D. (1990). The permeability of the nuclear
envelope in dividing and nondividing cell cultures. J. Cell Biol. 111, 1-8.
Forbes, D. J. (1992). Structure and function of the nuclear pore complex.
Annu. Rev. Cell Biol. 8, 495-527.
Fritz, F. D., Herweijer, H., Zhang, G. and Wolff, J. A. (1996). Gene transfer
into mammalian cells using histone-condensed plasmid DNA. Hum. Gene
Ther. 7, 1395-1404.
Garcia-Bustos, J., Heitman, J. and Hall, M. N. (1991). Nuclear protein
localization. Biochim. Biophys. Acta 1071, 83-101.
Hagstrom, J. E., Ludtke, J. J., Bassik, M. C., Sebestyén, M. G., Adam, S.
A. and Wolff, J. A. (1997). Nuclear import of DNA in digitoninpermeabilized cells. J. Cell Sci. 110, 2323-2331.
Henkel, T., Zabel, U., van Zee, K., Muller, J. M., Fanning, E. and Baeuerle,
P. A. (1992). Intramolecular masking of the nuclear location signal and
dimerization domain in the precursor for the p50 NF-kappa B subunit. Cell
68, 1121-1133.
Hubner, S., Xiao, C. Y. and Jans, D. A. (1997). The protein kinase CK2
site (Ser111/112) enhances recognition of the simian virus 40 large Tantigen nuclear localization sequence by importin. J. Biol. Chem. 272,
17191-17195.
Jans, D. A. and Hubner, S. (1996). Regulation of protein transport to the
nucleus: central role of phosphorylation. Physiol. Rev. 76, 651-685.
Kalderon, D., Roberts, B. L., Richardson, W. D. and Smith, A. E. (1984).
A short amino acid sequence able to specify nuclear location. Cell 39, 499509.
Kaneda, Y., Iwai, K. and Uchida, T. (1989). Increased expression of
DNA cointroduced with nuclear protein in adult rat liver. Science 243,
375-378.
Kasamatsu, H. and Nakanishi, A. (1998). How do animal DNA viruses get
to the nucleus? [Review]. Annu. Rev. Microbiol. 52, 627-686.
Kose, S., Imamoto, N., Tachibana, T., Shimamoto, T. and Yoneda, Y.
(1997). Ran-unassisted nuclear migration of a 97-kD component of nuclear
pore-targeting complex. J. Cell Biol. 139, 841-849.
Langle-Rouault, F., Patzel, V., Benavente, A., Taillez, M., Silvestre, N.,
Bompard, A., Sczakiel, G., Jacobs, E. and Rittner, K. (1998). Up to 100fold increase of apparent gene expression in the presence of Epstein-Barr
virus oriP sequences and EBNA1 - implications of the nuclear import of
plasmids. J. Virol. 72, 6181-6185.
Leonetti, J. P., Mechti, N., Degols, G., Gagnor, C. and Lebleu, B. (1991).
Intracellular distribution of microinjected antisense oligonucleotides. Proc.
Nat. Acad. Sci. USA 88, 2702-2706.
Mattaj, I. W. and Englmeier, L. (1998). Nucleocytoplasmic transport: The
soluble phase. Annu. Rev. Biochem. 67, 265-306.
Nakielny, S. and Dreyfuss, G. (1998). Import and export of the nuclear
protein import receptor transportin by a mechanism independent of GTP
hydrolysis. Cur. Biol. 8, 89-95.
NLS enhanced DNA transport 2041
Pante, N. and Aebi, U. (1996). Molecular dissection of the nuclear pore
complex. Crit. Rev. Biochem. Mol. Biol. 31, 153-199.
Pennisi, E. (1998). The nucleus’s revolving door. Science 279, 11291131.
Richardson, W. D., Mills, A. D., Dilworth, S. M., Laskey, R. A. and
Dingwall, C. (1988). Nuclear pore migration involves two steps: rapid
binding at the nuclear envelope followed by slower translocation through
the nuclear pores. Cell 52, 655-664.
Rihs, H. P., Jans, D. A., Fan, H. and Peters, R. (1991). The rate of nuclear
cytoplasmic protein transport is determined by the casein kinase II site
flanking the nuclear localization sequence of the SV40 T-antigen. EMBO J.
10, 633-639.
Sebestyén, M. G., Ludtke, J. J., Bassik, M. C., Zhang, G., Budker, V.,
Lukhtanov, E. A., Hagstrom, J. E. and Wolff, J. A. (1998). DNA vector
chemistry: the covalent attachment of signal peptides to plasmid DNA.
Nature Biotechnol. 16, 80-85.
Sebestyén, M. G. and Wolff, J. A. (1998). Nuclear transport of exogenous
DNA. In Gene Therapy (ed. T. Friedman), pp. 151-183. Cold Spring Harbor
Press, Cold Spring Harbor, NY.
Whittaker, G. R. and Helenius, A. (1998). Nuclear import and export of
viruses and virus genomes. Virology 246, 1-23.
Wolff, J. (1994). Gene Therapeutics. Birkhauser, New York.
Wolff, J. A. and Trubetskoy, S. (1998). The cambrian period of nonviral gene
delivery. Nature Biotechnol. 16, 421-422.
Yoneda, Y., Semba, T., Kaneda, Y., Noble, R. L., Matsuoka, Y., Kurihara,
T., Okada, Y. and Imamoto, N. (1992). A long synthetic peptide containing
a nuclear localization signal and its flanking sequences of SV40 T-antigen
directs the transport of IgM into the nucleus efficiently. Exp. Cell Res. 201,
313-320.
Zanta, M. A., Belguise-Valladier, P. and Behr, J.-P. (1999). Gene delivery:
a single nuclear localization signal peptide is sufficient to carry DNA to the
cell nucleus. Proc. Nat. Acad. Sci. USA 96, 91-96.