- Email: [email protected]
Antibodies introduced into living cells with liposomes localize specifically and inhibit specific intracellular processes
1988, Gene Anal Techn 5:73-79
Antibodies Introduced into Living Cells with Liposomes Localize Specifically and Inhibit Specific Intracellular Processes W. S C O T T T H O M P S O N and R O B E R T H. G R O S S
We have developed a system for efficiently packaging antibodies and other macromolecules into lipsomes and then deliverin~ the encapsulated molecules into living cells through liposome-cell fusion. Fusion is very efficient, and all cells can be demonstrated to contain liposome-delivered antibodies by staining with a fluorescent second antibody. Using lupus antibodies directed against small nuclear ribonucleoprotein components of the cell, we were able to demonstrate strong nuclear localization, while control antibodies showed a general diffuse distribution throughout the cell. Lupus antibodies directed against ribosomes, on the other hand, strongly localized in the nucleolus and the cytoplasm with very little nucleoplasmic localization. Antitubulin antibodies predominantly localized in the cytoplasm. These results show that antibodies can survive liposome packaging and can retain their ability to recognize and bind to their specific antigens in the living cell. It also indicates that the nuclear envelope does not present a barrier to the liposome-introduced antibodies in Drosophila tissue culture cells. To determine if the antibodies were capable of interfering with cellular processes in vivo, we measured the effects of liposome-introduced antiribosome antibodies on translation and antitubulin antibodies on mitosis. In both cases, there was a significant inhibition suggesting that the antibodies can be used to interfere with specific functions at specific times in vivo.
R N A processing consists of all the steps involved in converting the initial primary RNA transcript into a final product. For mRNAs, this typically consists of capping, base modification, polyadenylation, packaging with proteins, splicing, an transport to the cytoplasm [1]. It is our goal to From Dartmouth College, Hanover, New Hampshire. Address reprint request to: Robert H. Gross, Department of Biological Sciences and Molecular Genetics Center, Dartmouth College, Hanover, NH 03755. Received October 1, 1987.
study the interdependencies of the various steps involved in RNA processing by transiently interfering with one or more of the steps and observing effects on the other steps. The basic approach is to introduce antibodies and other macromolecules into cells using liposome-cell fusion and measure changes in precursor-product relationships or specific mRNAs. Liposomes are superior to other delivery methods because they protect their contents from the environment during fusion, they are extremely efficient at delivering material into all the cells (eliminating the need for postfusion selection of cells), the fusion does not harm the cells (see preceding paper in this issue), and fusions can be done with large numbers of cells, making biochemical preparations practical [2]. For this approach to be feasible, several facts must first be established: 1) that antibodies can be packaged into liposomes and subsequently delivered into cells; 2) that delivered antibodies must retain their ability to interact specifically with their target antigens; 3) that delivered antibodies must be capable of crossing the nuclear envelope, because the nucleus is the sight of most of the R N A processing events we are interested in studying; and 4) that liposome-delivered antibodies must be capable of interfering with biological processes in vivo. We have used antibodies purified from lupus patients to address the above questions. Anti-Sm and anti-(U1)-RNP antibodies recognize a set of small n u c l e a r r i b o n u c l e o p r o t e i n p a r t i c l e s (snRNPs) containing U I , U2, U4, U5, and U6 s n R N A s , or j u s t the U1 s n R N A containing snRNP, respectively. As controls for these antibodies, we have used normal human serum antibodies and lupus serum-derived antiribosome antibodies. For a cytoplasmic antigen, we have used rabbit antitubulin antibodies (directed against chicken tubulin but cross-reacting with Drosophila tubulin). In the experiments described here, we have used these antibodies to demonstrate the four points listed above. The anti-Sm and anti-(U1)RNA antibodies localize in the nucleus, the antitubulin antibodies localize in the cytoplasm, and the antiribosome antibodies localize in both the nucleolus and the cytoplasm, but not the nucleoplasm. All control antibodies distributed uniformly throughout the cell. In addition, using a fluorescence-activated cell sorter, we were able to demonstrate that the antitubulin antibodies interfere with cell division and, in other experiments,
© 1988 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017 0735-0651/88/$03.50
f J -..--~. ./
- j
f
74
1988, Gene Anal Techn 5:73-79
f_--____.-r~_
that the antiribosome antibodies significantly decrease translation. Materials and Methods
Cell Culture Schneider Line 2 Drosophila melanogaster tissue c u l t u r e c e l l s ( S L 2 c e l l s ) w e r e g r o w n in Schneider's Drosophila medium (GIBCO, Grand Island, NY) supplemented with 0.1 volumes of heat-inactivated fetal bovine serum (GIBCO). Cells were grown in either 80-cm 2 plastic flasks or in 175-cm 2 tissue-culture plates (NUNC). /
Liposome Packaging Liposomes were prepared as reverse-phase evaporation vesicles according to a modification (Marc Ostro, personal communication) of the origianl procedure of Szoka and Papahadjopoulos [3] as optimized for the Drosophila system. A mixture of 8.8 ~xmole of phosphatidyl serine (Sigma) and 4.4 txmole of phosphatidyl choline (GIBCO, ultrapure, special order) was dried down in a 5-ml pear-shaped flask in a rotary evaporator under 700 mm of vacuum at 40°C. The dried phospholipids were resuspended in 660 ~1 of ethyl ether and mixed with 200 ~1 of 0.1 x Drosophila phosphatebuffered saline (1 x DPBS is 0.14 M NaC1, 2.7 mM KC1, 8.1 mM Na2HPO 4, 1.5 mM KH2PO 4) at pH 8.0 containing the antibodies to be packaged. The biphasic mixture was then sonicated in a bath-type sonicator for 30-60 seconds (Laboratory Supplies, Ultrason-X) until a homogeneous milky emulsion formed. This emulsion was then rotary evaporated at 300 mm vacuum until a viscous gel was obtained. The sample was vortexed and rotary evaporated again. It is critical not to evaporate too much at this stage. An additional 460 p~l of 0.1 x DPBS, pH 8.0, was added and the gel vortexed to get a homogeneous beige-colored emulsion having a slightly opalescent appearance. The remaining ether was then evaporated by rotary evaporation at 500 mm vacuum for 15-20 minutes.
Liposome Cell Fusion L i p o s o m e - c e l l fusion was routinely done in Schneider's Drosophila Medium (SDM without fetal b o v i n e s e r u m ) . Cells were h a r v e s t e d , washed once in SDM (no fetal bovine serum), and
fused with liposomes at 25°C by tumbling gently for 30 minutes in SDM (no fetal bovine serum) at a ratio of 6 x 108 cells/13.2 p,moles phospholipid. Under these conditions, approximately 25% of the liposome-entrapped material is delivered into the cells. After fusion, cells were pelleted and resuspended in SDM (containing 0.1 volumes of fetal bovine serum) and replated. For experiments involving f l u o r e s c e n c e microscopy, cells were "plated" directly onto microscope slides. Unless stated otherwise, cells were allowed to recover for 1-2 hours before being analyzed.
Fluorescence Microscopy For microscopy, cells adhering to microscope slides were rinsed in a gentle stream of ice-cold DPBS and then fixed for 30 minutes at I°C in 3.7% formaldehyde in DPBS (v/v), pH 7.35. The fixed ceils were washed in a gentle stream of cold DPBS and then permeabilized in -20°C acetone for 10 minutes. After air drying, the fixed and permeabilized cells were blocked with 5% (v/v) goat serum, 0.75% bovine serum albumin, 20 mM Tris, pH 8.3, for 45 minutes at 2°C, and then stained with a I:100 dilution of flurescein-conjugated rabbit antihuman (for all lupus first antibodies) or goat antirabbit (for tubulin antibodies and rabbit control) antibodies purchased from Antibodies Incorporated (Davis, CA).
Cell Sorter For cell sorting, the cells were harvested and washed once in DPBS. The cell pellet was resuspended in a staining solution of 50 ~g/ml propidium iodide, 0.1% trisodium citrate, 0.01 M NaCI, and 0. I% NP40 (4). The propidium iodide stains the DNA, which is quantitated as the cell passes through the beam of the cell sorter. Samples were a n a l y z e d in an Ortho C y t o f l u r o g r a f , model FC-200. Cells were sampled at a rate of 500 per second and the data were analyzed by computer.
Protein Synthesis To monitor protein synthesis, cells were fused and replated in medium containing a mixture of 3Hamino acids (Amersham) at 10 txCi/ml in addition to the normal amino acids of the medium. Aliquots were harvested at various times up to 2 hours and lysed by addition of SDS to 1%. Tri-
© 1988 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
75 1988, Gene Anal Techn 5:73-79
•:.. ..--
-z--
chloroacetic acid precipitable radioactivity was determined for each aliquot.
Results In order for the liposome-introduced antibodies to be useful as an experimental tool, they must be capable of surviving the packaging and fusion procedures and maintaining their ability to recognize specific antigens in the living cell. To test this, we packaged into liposomes antibodies that would be expected to localize in specific subcellular compartments and then fused these liposomes to cells. The cells were subsequently fixed and the distribution of the introduced antibodies determined by staining with a second fluorescent antibody. Thus, no matter where the liposome-introduced antibodies were located in the cell, the second fluorescent antibodies would detect them. Figure 1 shows the distribution of various antibodies after liposome-mediated introduction into Drosophila line 2 cells. The patterns for control, anti-Sm, and anti-(U1)-RNP antibodies are shown in parts a/b, c/d, and e/f, respectively. The antiSm and anti-(U1)-RNP antibodis were human antibodies from lupus patients, and the control antibodies were from a normal human donor. The control antibodies were distributed randomly throughout the cell. In contrast, the distribution of the two lupus antibodies is strongly nuclear, suggesting that the anti-Sm and anti-(U1)-RNP antibodies are binding to components in the nucleus, presumably snRNPs. In control experiments (data not shown), fusion with empty liposomes and subsequent staining with the second fluorescent antibody did not produce any fluorescence, nor was any fluorescence detected in untreated cells (i.e., there is no autofluorescence). Fusion with the first antibody and no subsequent treatment with second fluorescent antibody also did not generate any fluorescence. To illustrate that the nuclear fluorescence was not due to nonspecific binding of antibodies to a component in Drosophila nuclei, we treated cells Figure 1. Control human antibodies were introduced into cells usm--~ffff~posome-cell fusion. After 1-3 hours, cultures were fixed, permeabilized, and treated with a fluorescent second a n t i b o d y (rabbit antihuman) as d e s c r i b e d in Materials and Methods. In each photographic pair, the left micrograph is phase contrast and the right is the corresponding fluorescence micrograph. The antibodies introduced were as follows: a/b, control human; c/d, anti-Sin; e/f, anti-(Ul)-RNP; g/h, antitubulin; i/j, antiribosome.
© 1988 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
76 1988, Gene Anal Techn 5:73-79 ~j C
J
with liposome-delivered antitubulin and antiribosome antibodies. Because the tubulin antibodies were from rabbit, we also did control fusions using nonimmune rabbit antibodies. The antiribosome antibodies were from a different lupus patient. Figure 2. Histograms from fluorescence-activated cell sorter (F'A-UST.Antitubulin or control antibodies were delivered into cells with liposomes. Two hours after fusion, the cells were fixed and stained with propidium iodide as described in Materials and Methods. The stained cells were then analyzed on the FACS to determine the distribution of DNA content per cell. The ordinate represents the number of ceils counted that have the DNA content shown along the abscissa. The peak at 20-30 units represents the normal heteroploid value for these cells. The peak in the 40-50 range is twice the normal value and represents cells that have replicated their DNA but not yet divided. A, untreated cells; B, cells treated with colchicine as a positive control; C, cells fused with liposomes containing control antibodies; D, cells fused with liposomes containing antitubulin antibodies.
Figure lg/h shows the distribution of the antitubulin antibodies. Although there is some nuclear fluorescence, most of the fluorescence is in the cytoplasm, where the microtubules are located. The slight fluorescence seen over the nucleus may in fact be due to the cytoplasm that lies either above or below the nucleus as viewed in the microscope. The distribution of the antiribosome antibodies is shown in figure li/j. In this case it is very clear that the antibodies are localized in the nucleolus and the c y t o p l a s m - - t h o s e compartments of the cell that contain ribosomes. From the above results it seemed clear that antibodies could survive packaging in liposomes, could enter the cells, and find their appropriate antigens. The results with the various lupus antibodies also demonstrated that the nuclear envelope did not provide a barrier to the liposomeintroduced antibodies in these tissue-culture 188e
B
1"
t I
311 11 '1
-II
It
.-
48 G1
8e RE~
FL,
12e
IGe
2Be
1
4e
RRER
62
1 eee
8# RED
128
FL,
16e
2ee
AREA
l m
D
1
4e
G3
8e
RED
FL,
12Q
16e
2ee
1
ARER
© 1988 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
4e G1
8e RED
FL,
12e AREA
168
2OO
1988, Gene Anal Techn 5:73-79
cells. To test whether the liposome-introduced antibodies were functional and could interfere with biological processes in vivo, we measured the effects of antitubulin antibodies on cell division and of antiribosome antibodies on protein synthesis. The effect of liposome-delivered antitubulin antibodies on cell division was measured by determining the amount of DNA per cell using a fluorescence-activated cell sorter (FACS). Figure 2 shows the results of such an analysis. The peak around number 25 represents cells having normal DNA content, while the peak around 50 represents ceils having twice the normal DNA content. As a control, some cells were also treated with colchicine, which would allow DNA replication but block cell division (the microtubules can not depolymerize). A comparison of figures 2C and 2D shows that the anti-tubulin antibodies caused a shift of cells from normal to twicenormal DNA content. As shown in Figure 3, a quantitation of the areas under each of the peaks demonstrates that the antitubulin antibodies produce about a 25% decrease in the number of cells in G O + G 1. Control antibodies never demonstrated such a decrease. These experiments were done uner conditions that delivered about 50,000100,000 antibodies per cell. The results suggest that the antitubulin antibodies were interfering
with mitosis, presumably by interfering with microtubule polymerization or depolymerization. Figure 4 demonstrates the effect of liposomeintroduced antiribosome antibodies on proteitn synthesis. As in the antitubulin experiments, 50,000-100,000 antibodies were introduced into each cell. It is clear that the antiribosome antibodies cause about a 50% decrease in the rate of protein synthesis.
Discussion We have demonstrated that antibodies can be packaged into liposomes and delivered into viable cells for study (Figure 1). Control antibodies did not recognize specific cellular components and therefore did not localize to any specific subcellular compartment. The anti-Sm and anti-(U1)RNP antibodies each localized in the nucleus, which is where the majority of the snRNPs are located [5]. On the other hand, antitubulin antibodies preferentially localized in the cytoplasm, and antiribosome antibodies localized in the nuFigure 3. Effect of antitubulin AGs on the cell cycle. The data generated in Figure 3 were analyzed by the computer programs supplied with the cell sorter to determine the fraction of cells in Go + G~. This refers to the cells under the first peak (20-30). The plot shown represents the data from three separate experiments.
120
100
+
80
~
60
~
40
20
Control
+Colchicine
Control Abs
Antitubulin Abs
Treatment © 1988 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
f -~.___--
C-~-'3
-j
78 1988, Gene Anal Techn 5:73-79
f-
4000 m
control
\
3000
J
, m
~" O
2000
-•.
/
m
t~
o
_=
f
J
1000
r
E
m
anti-ribosome o
J
0
B
~
J
I
B
t
l
25
R
I
I
50
•
i
75
•
,
•
•
100
•
•
,
125
time (min)
Figure 4. Effect of antiribosome AGs on protein synthesis. CElls"~-~re fused either with antiribosome or control antibodies and then placed in medium containing 3H-amino acids. At various times, aliquots were removed and trichloroacetic acid precipitable radioactivity determined. The data shown represent the average of four separate experiments.
cleolus and the cytoplasm but not in the nucleoplasm. Because the delivered antibodies distributed themselves in distinct, different, and appropriate patterns, we conclude that liposome-delivered antibodies maintain their ability to recognize and interact with their specific antigens within living cells. Furthermore, because the introduced antisnRNP antibodies and some of the antiribosome antibodies interact specifically with components in the nucleus, the nuclear envelope does not seem to provide a barrier to movement of the antibodies in these cells. The nuclear translocation we observed is in contrast to the results of Einck and Bustin [6], who found that whole antibody molecules cannot traverse the envelope. They injected either whole IgG or F(ab)2 fragments of antihistone antibodies into human fibroblasts and found that only the F(ab) 2 fragment was capable of traversing the nuclear envelope (they did not state the organism f r o m w h i c h the a n t i b o d i e s were d e r i v e d ) . McGarry et al. [7] also studied the distribution of microinjected antibodies in human cells. They in-
jected either intact IgG, Fab, or Fc fragments of rabbit antibodies into HeLa cells. The IgG was excluded from the nucleus (after cytoplasmic injection) but the Fab and Fc fragments were not. It is not clear why liposome-introduced antibodies can enter Drosophila cell nuclei but microinjected antibodies cannot enter human nuclei. It is not likely that the method of introduction is responsible because no physical changes are being made to the antibody molecules themselves, and in both cases the antibodies are being delivered into the cytoplasm. In our cells, we have introduced human antibodies (anti-snRNPs, antiribosome, and control) as well as rabbit antibodies (antitubulin and control), and both sources have been capable of penetrating the nuclear envelope. The source of the antibody is therefore not likely to be the cause of the difference. The most likely explanation for the different nuclear permeabilities is the difference in organisms from which the cell lines have been derived, suggesting that the Drosophila nuclear envelope is more permeable to intact antibody molecules than the human nuclear envelope. Whether this indicates a different pore size for human versus Drosophila nuclei or a difference in some other parameter is not known. In the experiments described here, we added 50,000-100,000 antibody molecules per cell. Liposome-delivered antiribosome antibodies re-
© 1988 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
79'>), F
1988, Gene Anal Techn 5:73-79
E--_jJ
-
duced the rate of protein synthesis by about 50% (Figure 4), whereas antitubulin antibodies caused a 25% decrease in the rate of cytokinesis (Figures 2 and 3). The liposome delivery technique as applied here in the Drosophila system allows the antibodies to interact with specific nuclear or cytoplasmic intracellular components and potentially inhibit their function(s) in vivo. The ability of the liposome-introduced antibodies to interact with target antigens in vivo opens up several new avenues of experimental study in vivo. Immunocytochemical techniques previously applicable only in vitro can now be carried out in vivo. This approach should allow for the study of cellular mechanisms and their interrelationships on a molecular level in living cells.
References I. Lewin, B. (1985) Genes II, John Wiley & Son, Inc., New York. 2. Ostro, M. J., Giacomoni, D., and Dray, S. (1980) in Intro-
duction of Macromolecules Into Viable Mammalian Cells, Baserga, R., Croce, C., and Rovera, G., Eds., pp. 239-259 Alan R. Liss, New York. 3. Szoka, F., and Papahadjopoulos, D. (1978) Proc. Nat. Acad. Sci. 75:4194-4198. 4.
Attallah, A. M., T. J. Yeatman, P. D. Noguchi, and J. B. Johnson (1980) Science 209:404.
5. Lerner, M. R., and Steitz, J. A. (1979) Proc. Nat. Acad. Sci. 76:5495-5499. 6. Einck, L., and M. Bustin (1984) J. Cell Biol. 98:205-213. 7. McGarry, T., R. Hough, S. Rogers, and M. Rechsteiner (1983) J. Cell Biol. 96:338-246.
© 1988 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
j
Antibodies Introduced into Living Cells with Liposomes Localize Specifically and Inhibit Specific Intracellular Processes W. S C O T T T H O M P S O N and R O B E R T H. G R O S S
We have developed a system for efficiently packaging antibodies and other macromolecules into lipsomes and then deliverin~ the encapsulated molecules into living cells through liposome-cell fusion. Fusion is very efficient, and all cells can be demonstrated to contain liposome-delivered antibodies by staining with a fluorescent second antibody. Using lupus antibodies directed against small nuclear ribonucleoprotein components of the cell, we were able to demonstrate strong nuclear localization, while control antibodies showed a general diffuse distribution throughout the cell. Lupus antibodies directed against ribosomes, on the other hand, strongly localized in the nucleolus and the cytoplasm with very little nucleoplasmic localization. Antitubulin antibodies predominantly localized in the cytoplasm. These results show that antibodies can survive liposome packaging and can retain their ability to recognize and bind to their specific antigens in the living cell. It also indicates that the nuclear envelope does not present a barrier to the liposome-introduced antibodies in Drosophila tissue culture cells. To determine if the antibodies were capable of interfering with cellular processes in vivo, we measured the effects of liposome-introduced antiribosome antibodies on translation and antitubulin antibodies on mitosis. In both cases, there was a significant inhibition suggesting that the antibodies can be used to interfere with specific functions at specific times in vivo.
R N A processing consists of all the steps involved in converting the initial primary RNA transcript into a final product. For mRNAs, this typically consists of capping, base modification, polyadenylation, packaging with proteins, splicing, an transport to the cytoplasm [1]. It is our goal to From Dartmouth College, Hanover, New Hampshire. Address reprint request to: Robert H. Gross, Department of Biological Sciences and Molecular Genetics Center, Dartmouth College, Hanover, NH 03755. Received October 1, 1987.
study the interdependencies of the various steps involved in RNA processing by transiently interfering with one or more of the steps and observing effects on the other steps. The basic approach is to introduce antibodies and other macromolecules into cells using liposome-cell fusion and measure changes in precursor-product relationships or specific mRNAs. Liposomes are superior to other delivery methods because they protect their contents from the environment during fusion, they are extremely efficient at delivering material into all the cells (eliminating the need for postfusion selection of cells), the fusion does not harm the cells (see preceding paper in this issue), and fusions can be done with large numbers of cells, making biochemical preparations practical [2]. For this approach to be feasible, several facts must first be established: 1) that antibodies can be packaged into liposomes and subsequently delivered into cells; 2) that delivered antibodies must retain their ability to interact specifically with their target antigens; 3) that delivered antibodies must be capable of crossing the nuclear envelope, because the nucleus is the sight of most of the R N A processing events we are interested in studying; and 4) that liposome-delivered antibodies must be capable of interfering with biological processes in vivo. We have used antibodies purified from lupus patients to address the above questions. Anti-Sm and anti-(U1)-RNP antibodies recognize a set of small n u c l e a r r i b o n u c l e o p r o t e i n p a r t i c l e s (snRNPs) containing U I , U2, U4, U5, and U6 s n R N A s , or j u s t the U1 s n R N A containing snRNP, respectively. As controls for these antibodies, we have used normal human serum antibodies and lupus serum-derived antiribosome antibodies. For a cytoplasmic antigen, we have used rabbit antitubulin antibodies (directed against chicken tubulin but cross-reacting with Drosophila tubulin). In the experiments described here, we have used these antibodies to demonstrate the four points listed above. The anti-Sm and anti-(U1)RNA antibodies localize in the nucleus, the antitubulin antibodies localize in the cytoplasm, and the antiribosome antibodies localize in both the nucleolus and the cytoplasm, but not the nucleoplasm. All control antibodies distributed uniformly throughout the cell. In addition, using a fluorescence-activated cell sorter, we were able to demonstrate that the antitubulin antibodies interfere with cell division and, in other experiments,
© 1988 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017 0735-0651/88/$03.50
f J -..--~. ./
- j
f
74
1988, Gene Anal Techn 5:73-79
f_--____.-r~_
that the antiribosome antibodies significantly decrease translation. Materials and Methods
Cell Culture Schneider Line 2 Drosophila melanogaster tissue c u l t u r e c e l l s ( S L 2 c e l l s ) w e r e g r o w n in Schneider's Drosophila medium (GIBCO, Grand Island, NY) supplemented with 0.1 volumes of heat-inactivated fetal bovine serum (GIBCO). Cells were grown in either 80-cm 2 plastic flasks or in 175-cm 2 tissue-culture plates (NUNC). /
Liposome Packaging Liposomes were prepared as reverse-phase evaporation vesicles according to a modification (Marc Ostro, personal communication) of the origianl procedure of Szoka and Papahadjopoulos [3] as optimized for the Drosophila system. A mixture of 8.8 ~xmole of phosphatidyl serine (Sigma) and 4.4 txmole of phosphatidyl choline (GIBCO, ultrapure, special order) was dried down in a 5-ml pear-shaped flask in a rotary evaporator under 700 mm of vacuum at 40°C. The dried phospholipids were resuspended in 660 ~1 of ethyl ether and mixed with 200 ~1 of 0.1 x Drosophila phosphatebuffered saline (1 x DPBS is 0.14 M NaC1, 2.7 mM KC1, 8.1 mM Na2HPO 4, 1.5 mM KH2PO 4) at pH 8.0 containing the antibodies to be packaged. The biphasic mixture was then sonicated in a bath-type sonicator for 30-60 seconds (Laboratory Supplies, Ultrason-X) until a homogeneous milky emulsion formed. This emulsion was then rotary evaporated at 300 mm vacuum until a viscous gel was obtained. The sample was vortexed and rotary evaporated again. It is critical not to evaporate too much at this stage. An additional 460 p~l of 0.1 x DPBS, pH 8.0, was added and the gel vortexed to get a homogeneous beige-colored emulsion having a slightly opalescent appearance. The remaining ether was then evaporated by rotary evaporation at 500 mm vacuum for 15-20 minutes.
Liposome Cell Fusion L i p o s o m e - c e l l fusion was routinely done in Schneider's Drosophila Medium (SDM without fetal b o v i n e s e r u m ) . Cells were h a r v e s t e d , washed once in SDM (no fetal bovine serum), and
fused with liposomes at 25°C by tumbling gently for 30 minutes in SDM (no fetal bovine serum) at a ratio of 6 x 108 cells/13.2 p,moles phospholipid. Under these conditions, approximately 25% of the liposome-entrapped material is delivered into the cells. After fusion, cells were pelleted and resuspended in SDM (containing 0.1 volumes of fetal bovine serum) and replated. For experiments involving f l u o r e s c e n c e microscopy, cells were "plated" directly onto microscope slides. Unless stated otherwise, cells were allowed to recover for 1-2 hours before being analyzed.
Fluorescence Microscopy For microscopy, cells adhering to microscope slides were rinsed in a gentle stream of ice-cold DPBS and then fixed for 30 minutes at I°C in 3.7% formaldehyde in DPBS (v/v), pH 7.35. The fixed ceils were washed in a gentle stream of cold DPBS and then permeabilized in -20°C acetone for 10 minutes. After air drying, the fixed and permeabilized cells were blocked with 5% (v/v) goat serum, 0.75% bovine serum albumin, 20 mM Tris, pH 8.3, for 45 minutes at 2°C, and then stained with a I:100 dilution of flurescein-conjugated rabbit antihuman (for all lupus first antibodies) or goat antirabbit (for tubulin antibodies and rabbit control) antibodies purchased from Antibodies Incorporated (Davis, CA).
Cell Sorter For cell sorting, the cells were harvested and washed once in DPBS. The cell pellet was resuspended in a staining solution of 50 ~g/ml propidium iodide, 0.1% trisodium citrate, 0.01 M NaCI, and 0. I% NP40 (4). The propidium iodide stains the DNA, which is quantitated as the cell passes through the beam of the cell sorter. Samples were a n a l y z e d in an Ortho C y t o f l u r o g r a f , model FC-200. Cells were sampled at a rate of 500 per second and the data were analyzed by computer.
Protein Synthesis To monitor protein synthesis, cells were fused and replated in medium containing a mixture of 3Hamino acids (Amersham) at 10 txCi/ml in addition to the normal amino acids of the medium. Aliquots were harvested at various times up to 2 hours and lysed by addition of SDS to 1%. Tri-
© 1988 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
75 1988, Gene Anal Techn 5:73-79
•:.. ..--
-z--
chloroacetic acid precipitable radioactivity was determined for each aliquot.
Results In order for the liposome-introduced antibodies to be useful as an experimental tool, they must be capable of surviving the packaging and fusion procedures and maintaining their ability to recognize specific antigens in the living cell. To test this, we packaged into liposomes antibodies that would be expected to localize in specific subcellular compartments and then fused these liposomes to cells. The cells were subsequently fixed and the distribution of the introduced antibodies determined by staining with a second fluorescent antibody. Thus, no matter where the liposome-introduced antibodies were located in the cell, the second fluorescent antibodies would detect them. Figure 1 shows the distribution of various antibodies after liposome-mediated introduction into Drosophila line 2 cells. The patterns for control, anti-Sm, and anti-(U1)-RNP antibodies are shown in parts a/b, c/d, and e/f, respectively. The antiSm and anti-(U1)-RNP antibodis were human antibodies from lupus patients, and the control antibodies were from a normal human donor. The control antibodies were distributed randomly throughout the cell. In contrast, the distribution of the two lupus antibodies is strongly nuclear, suggesting that the anti-Sm and anti-(U1)-RNP antibodies are binding to components in the nucleus, presumably snRNPs. In control experiments (data not shown), fusion with empty liposomes and subsequent staining with the second fluorescent antibody did not produce any fluorescence, nor was any fluorescence detected in untreated cells (i.e., there is no autofluorescence). Fusion with the first antibody and no subsequent treatment with second fluorescent antibody also did not generate any fluorescence. To illustrate that the nuclear fluorescence was not due to nonspecific binding of antibodies to a component in Drosophila nuclei, we treated cells Figure 1. Control human antibodies were introduced into cells usm--~ffff~posome-cell fusion. After 1-3 hours, cultures were fixed, permeabilized, and treated with a fluorescent second a n t i b o d y (rabbit antihuman) as d e s c r i b e d in Materials and Methods. In each photographic pair, the left micrograph is phase contrast and the right is the corresponding fluorescence micrograph. The antibodies introduced were as follows: a/b, control human; c/d, anti-Sin; e/f, anti-(Ul)-RNP; g/h, antitubulin; i/j, antiribosome.
© 1988 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
76 1988, Gene Anal Techn 5:73-79 ~j C
J
with liposome-delivered antitubulin and antiribosome antibodies. Because the tubulin antibodies were from rabbit, we also did control fusions using nonimmune rabbit antibodies. The antiribosome antibodies were from a different lupus patient. Figure 2. Histograms from fluorescence-activated cell sorter (F'A-UST.Antitubulin or control antibodies were delivered into cells with liposomes. Two hours after fusion, the cells were fixed and stained with propidium iodide as described in Materials and Methods. The stained cells were then analyzed on the FACS to determine the distribution of DNA content per cell. The ordinate represents the number of ceils counted that have the DNA content shown along the abscissa. The peak at 20-30 units represents the normal heteroploid value for these cells. The peak in the 40-50 range is twice the normal value and represents cells that have replicated their DNA but not yet divided. A, untreated cells; B, cells treated with colchicine as a positive control; C, cells fused with liposomes containing control antibodies; D, cells fused with liposomes containing antitubulin antibodies.
Figure lg/h shows the distribution of the antitubulin antibodies. Although there is some nuclear fluorescence, most of the fluorescence is in the cytoplasm, where the microtubules are located. The slight fluorescence seen over the nucleus may in fact be due to the cytoplasm that lies either above or below the nucleus as viewed in the microscope. The distribution of the antiribosome antibodies is shown in figure li/j. In this case it is very clear that the antibodies are localized in the nucleolus and the c y t o p l a s m - - t h o s e compartments of the cell that contain ribosomes. From the above results it seemed clear that antibodies could survive packaging in liposomes, could enter the cells, and find their appropriate antigens. The results with the various lupus antibodies also demonstrated that the nuclear envelope did not provide a barrier to the liposomeintroduced antibodies in these tissue-culture 188e
B
1"
t I
311 11 '1
-II
It
.-
48 G1
8e RE~
FL,
12e
IGe
2Be
1
4e
RRER
62
1 eee
8# RED
128
FL,
16e
2ee
AREA
l m
D
1
4e
G3
8e
RED
FL,
12Q
16e
2ee
1
ARER
© 1988 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
4e G1
8e RED
FL,
12e AREA
168
2OO
1988, Gene Anal Techn 5:73-79
cells. To test whether the liposome-introduced antibodies were functional and could interfere with biological processes in vivo, we measured the effects of antitubulin antibodies on cell division and of antiribosome antibodies on protein synthesis. The effect of liposome-delivered antitubulin antibodies on cell division was measured by determining the amount of DNA per cell using a fluorescence-activated cell sorter (FACS). Figure 2 shows the results of such an analysis. The peak around number 25 represents cells having normal DNA content, while the peak around 50 represents ceils having twice the normal DNA content. As a control, some cells were also treated with colchicine, which would allow DNA replication but block cell division (the microtubules can not depolymerize). A comparison of figures 2C and 2D shows that the anti-tubulin antibodies caused a shift of cells from normal to twicenormal DNA content. As shown in Figure 3, a quantitation of the areas under each of the peaks demonstrates that the antitubulin antibodies produce about a 25% decrease in the number of cells in G O + G 1. Control antibodies never demonstrated such a decrease. These experiments were done uner conditions that delivered about 50,000100,000 antibodies per cell. The results suggest that the antitubulin antibodies were interfering
with mitosis, presumably by interfering with microtubule polymerization or depolymerization. Figure 4 demonstrates the effect of liposomeintroduced antiribosome antibodies on proteitn synthesis. As in the antitubulin experiments, 50,000-100,000 antibodies were introduced into each cell. It is clear that the antiribosome antibodies cause about a 50% decrease in the rate of protein synthesis.
Discussion We have demonstrated that antibodies can be packaged into liposomes and delivered into viable cells for study (Figure 1). Control antibodies did not recognize specific cellular components and therefore did not localize to any specific subcellular compartment. The anti-Sm and anti-(U1)RNP antibodies each localized in the nucleus, which is where the majority of the snRNPs are located [5]. On the other hand, antitubulin antibodies preferentially localized in the cytoplasm, and antiribosome antibodies localized in the nuFigure 3. Effect of antitubulin AGs on the cell cycle. The data generated in Figure 3 were analyzed by the computer programs supplied with the cell sorter to determine the fraction of cells in Go + G~. This refers to the cells under the first peak (20-30). The plot shown represents the data from three separate experiments.
120
100
+
80
~
60
~
40
20
Control
+Colchicine
Control Abs
Antitubulin Abs
Treatment © 1988 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
f -~.___--
C-~-'3
-j
78 1988, Gene Anal Techn 5:73-79
f-
4000 m
control
\
3000
J
, m
~" O
2000
-•.
/
m
t~
o
_=
f
J
1000
r
E
m
anti-ribosome o
J
0
B
~
J
I
B
t
l
25
R
I
I
50
•
i
75
•
,
•
•
100
•
•
,
125
time (min)
Figure 4. Effect of antiribosome AGs on protein synthesis. CElls"~-~re fused either with antiribosome or control antibodies and then placed in medium containing 3H-amino acids. At various times, aliquots were removed and trichloroacetic acid precipitable radioactivity determined. The data shown represent the average of four separate experiments.
cleolus and the cytoplasm but not in the nucleoplasm. Because the delivered antibodies distributed themselves in distinct, different, and appropriate patterns, we conclude that liposome-delivered antibodies maintain their ability to recognize and interact with their specific antigens within living cells. Furthermore, because the introduced antisnRNP antibodies and some of the antiribosome antibodies interact specifically with components in the nucleus, the nuclear envelope does not seem to provide a barrier to movement of the antibodies in these cells. The nuclear translocation we observed is in contrast to the results of Einck and Bustin [6], who found that whole antibody molecules cannot traverse the envelope. They injected either whole IgG or F(ab)2 fragments of antihistone antibodies into human fibroblasts and found that only the F(ab) 2 fragment was capable of traversing the nuclear envelope (they did not state the organism f r o m w h i c h the a n t i b o d i e s were d e r i v e d ) . McGarry et al. [7] also studied the distribution of microinjected antibodies in human cells. They in-
jected either intact IgG, Fab, or Fc fragments of rabbit antibodies into HeLa cells. The IgG was excluded from the nucleus (after cytoplasmic injection) but the Fab and Fc fragments were not. It is not clear why liposome-introduced antibodies can enter Drosophila cell nuclei but microinjected antibodies cannot enter human nuclei. It is not likely that the method of introduction is responsible because no physical changes are being made to the antibody molecules themselves, and in both cases the antibodies are being delivered into the cytoplasm. In our cells, we have introduced human antibodies (anti-snRNPs, antiribosome, and control) as well as rabbit antibodies (antitubulin and control), and both sources have been capable of penetrating the nuclear envelope. The source of the antibody is therefore not likely to be the cause of the difference. The most likely explanation for the different nuclear permeabilities is the difference in organisms from which the cell lines have been derived, suggesting that the Drosophila nuclear envelope is more permeable to intact antibody molecules than the human nuclear envelope. Whether this indicates a different pore size for human versus Drosophila nuclei or a difference in some other parameter is not known. In the experiments described here, we added 50,000-100,000 antibody molecules per cell. Liposome-delivered antiribosome antibodies re-
© 1988 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
79'>), F
1988, Gene Anal Techn 5:73-79
E--_jJ
-
duced the rate of protein synthesis by about 50% (Figure 4), whereas antitubulin antibodies caused a 25% decrease in the rate of cytokinesis (Figures 2 and 3). The liposome delivery technique as applied here in the Drosophila system allows the antibodies to interact with specific nuclear or cytoplasmic intracellular components and potentially inhibit their function(s) in vivo. The ability of the liposome-introduced antibodies to interact with target antigens in vivo opens up several new avenues of experimental study in vivo. Immunocytochemical techniques previously applicable only in vitro can now be carried out in vivo. This approach should allow for the study of cellular mechanisms and their interrelationships on a molecular level in living cells.
References I. Lewin, B. (1985) Genes II, John Wiley & Son, Inc., New York. 2. Ostro, M. J., Giacomoni, D., and Dray, S. (1980) in Intro-
duction of Macromolecules Into Viable Mammalian Cells, Baserga, R., Croce, C., and Rovera, G., Eds., pp. 239-259 Alan R. Liss, New York. 3. Szoka, F., and Papahadjopoulos, D. (1978) Proc. Nat. Acad. Sci. 75:4194-4198. 4.
Attallah, A. M., T. J. Yeatman, P. D. Noguchi, and J. B. Johnson (1980) Science 209:404.
5. Lerner, M. R., and Steitz, J. A. (1979) Proc. Nat. Acad. Sci. 76:5495-5499. 6. Einck, L., and M. Bustin (1984) J. Cell Biol. 98:205-213. 7. McGarry, T., R. Hough, S. Rogers, and M. Rechsteiner (1983) J. Cell Biol. 96:338-246.
© 1988 Elsevier Science Publishing Co., Inc., 52 Vanderbilt Ave., New York, NY 10017
j