Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Gene Therapy (1997) 4, 586–592 1997 Stockton Press All rights reserved 0969-7128/97 $12.00 H1 and HMG17 extracted from calf thymus nuclei are efficient DNA carriers in gene transfer SV Zaitsev1,2, A Haberland1, A Otto3, VI Vorob’ev2, H Haller1 and M Böttger1,4 1 Franz Volhard Clinic at the Max Delbrück Center for Molecular Medicine, 3Max Delbrück Center for Molecular Medicine, BerlinBuch; 4Medical Clinic, Department of Cardiology, Charité, Humboldt University, Berlin, Germany; and 2Institute of Cytology of the Russian Academy of Sciences, St Petersburg, Russia In this article we describe the chromatographic separation of acid nuclear protein fractions which have previously been shown to be active in DNA transfection experiments. By combining anionic and cationic ion exchangers, we were able to separate and identify some of the active proteins. In addition to HMG1, already known for its transfection activity, we have identified histone H1 and HMG17 as further transfection-active proteins. The highest transfection activity was associated with H1 and another nonidentified protein showing a somewhat higher electrophoretic mobility than H1. We have also found that the presence of CaCl2 in a low concentration in the cell culture medium is an important requirement for transfection. Keywords: nuclear extract; histone H1; HMG1; HMG17; reporter gene; transfection Introduction Effective gene transfer is an important prerequisite for gene therapy. Because nonviral delivery systems are expected to have little or no harmful side-effects, the development of such systems with high efficiency and avoiding toxic and immunogenic problems is an urgent task.1,2 One approach in this direction is the use of protein–DNA complexes as carriers of exogeneous genes into the host cell. While polylysine-ligand conjugates for targeting surface receptors of the host cell in complex with DNA via the endocytotic pathway3 or viral capsid proteins 4 have been used successfully in gene transfer, nuclear proteins have received little attention. The nonhistone protein HMG1 is the first example of a transfectionactive nuclear protein described.5–7 In a separate article, we showed that a series of nuclear protein fractions can be isolated from acid nuclear extracts by fractionated precipitation with acetone. These fractions proved to be transfection-active in complex with DNA in the presence of a low concentration of Ca ions in the cell culture medium. We demonstrated that these fractions contain histone H1, the nonhistone proteins HMG1, HMG2, HMG17 and other not yet identified proteins in different ratios.8 The aim of this study is to resolve further the constituents of these fractions by chromatographic methods, to identify the purified proteins by sequencing and to test their activity as DNA carriers for gene transfer. We envisage that with continuous purification and functional testing for transfection efficiency of the identified proteins at each purification step, we will be able to improve our delivery system based on such Correspondence: M Böttger, Franz Volhard Clinic, Wiltberg Strasse 50, 13122 Berlin, Germany Received 19 August 1996; accepted 13 February 1997 proteins. Our final goal is to establish an efficient transfection method by using physiological agents and physiological conditions which do not perturb or damage cells and which can be used in vivo. Results In a separate article8 we demonstrated that a series of nuclear protein fractions from calf thymus termed 2V– 5.5V exhibit the ability to package DNA into transfectionactive protein–DNA complexes. These protein fractions were obtained by extraction of nuclei with 5% perchloric acid and subsequent gradual acetone precipitation of these extracts. In this article, we describe the purification of the protein fractions 2V–3.5V by chromatographic methods (FPLC) which results in a considerable increase in the transfection efficiency. Each step of the purification process was checked by transfection experiments using the protein samples as DNA packaging agents. Only positive samples were investigated further. The fractions 2V-3.5V were separated on a Mono Q column by elution with a linear salt gradient of 3–73% of buffer B. Only the breakthrough material (2VA–3.5VA), not bound to the column, proved to be transfection-active and was used for additional separations using the Mono S column. Typical electrophoretograms of the Mono Q purified 2.5V fraction are shown in Figure 1a. With the exception of the transfection-active fraction 2.5VA (lane 2), all other fractions (2.5VB etc) were rejected. A second step of purification of the 2VA–3.5VA fractions was performed by means of the Mono S column under the same conditions as with the Mono Q column but using buffer C instead of buffer B. The result of this purification step for 2.5VA is shown in Figure 1b. Figure 2 shows the elution profiles of the 3VA and 3.5VA fractions after Mono S separation as examples for the resolution of the FPLC. There are large differences in the number of bands Transfection-active nuclear proteins SV Zaitsev et al 587 Figure 2 Elution profiles after Mono S chromatography of protein fractions 3.5VA (a) and 3VA (b). The fraction numbers given in the order of elution are identical with the numbers of the lanes of the electrophoresis experiment shown in Figure 1c and provide the denotation of the protein fractions. Figure 1 Acetic acid/urea polyacrylamide gel electrophoresis of protein fractions obtained from acid nuclear extracts by acetone precipitation and ion exchange chromatography (FPLC). (a) Fraction 2.5V after chromatography on Mono Q. Lane 2 contained transfection-active material indicated by an arrow (2.5VA) and was identified as histone H1. Lanes 1, 12: protein markers; lanes 3, 4: 2.5VB; lanes 5, 6: 2.5VC; lanes 9–11: 2.5VD. (b) Fraction 2.5VA after chromatography on Mono S. Lane 1: 2.5VA; lane 2: 2.5VASB (breakthrough); lanes 3–10: 2.5VAS1, lanes 11, 12: 2.5VAS2. (c) Fractions 3V and 3.5V after chromatography on Mono S. Lanes 1, 9, 12: protein marker (histone H1); lane 2: 3.5VAS3; lane 3: 3.5VAS4; lane 4: 3.5VAS5; lane 5: 3.5VAS6; lane 6: 3.5 VAS7; lane 7: 3.5VAS8; lane 8: 3.5VAS9; lane 10: 3VAS1, lane 11: 3VAS2. Bands designated by arrows in lanes 2 and 10: HMG17; in lane 4: HMG1; in lane 5: HMG2; in lane 11: H1. between both fractions. All bands were collected and studied by electrophoresis. The results are shown in Figure 1c. Fraction 2.5VA contains a single band, 3VA two bands and 3.5VA several bands. The fractions 4.5V-5.5V containing HMG1 and 2 proteins are omitted here as the transfection activity of HMG1 was previously demonstrated (Böttger et al5,6). The transfection activity of HMG1 is smaller than that of the 2V–3.5V fractions.8 Before summarizing the results of the chromatographic purification, it is necessary to describe the transfection experiments in more detail. We used the luciferase gene (pCMV Luc) as a reporter gene. pCMV Luc DNA (2 mg) were mixed with protein at a weight input ratio r i of 10– 20 and added to 2 × 105 NIH 3T3 cells, containing 8 mm CaCl2 in the culture medium. In the range of 1 , r i , 30 no significant differences in the transfection efficiency have been observed previously, although the variation in the transfection rates of several experiments of the same fraction was high. 8 Ca ions were compulsory for high transfection activity in this system. However, as shown later, the Ca concentration could be further decreased. The transfection efficiency is given in terms of relative light units. Figure 3 shows the results of transfection experiments performed with the total nonfractionated protein extract, S, the nuclear protein fraction 3V and following protein fractions 3VA, 3VAS1 and 3VAS2 Transfection-active nuclear proteins SV Zaitsev et al 588 Figure 3 A comparison of transfection efficiencies in terms of relative light units (RLU) obtained with free DNA, different protein–DNA complexes and Lipofectamin. The protein fractions S, 3V, 3VA, 3VAS1, 3VAS2 representing different purification levels by ionic exchange chromatography are used for complex formation with pCMV Luc DNA (2 mg). Complexes of ri ratios between 10 and 20 were added to 2 × 105 NIH 3T3 cells. The values are means ± s.e. from at least six different experiments. obtained after two FPLC purification steps as described above. Negative and positive controls obtained with free DNA and Lipofectamin respectively, were added. Compared with the unpurified total protein extract, S, a strong increase in the expression of the luciferase gene in terms of RLU could be seen after acetone fractionation (3V) and Mono Q separation (3VA). The increase in RLU was more than 30-fold. The purification on Mono S resulting in 3VAS1/S2 did not result in a further increase in RLU. Free DNA and cells without added DNA as a negative control were inactive (approximately 2000 RLU both). The transfection efficiencies of the fractions 3VA and 3VAS2 were comparable with that of Lipofectamin. Analogous results were obtained with 3.5V and the samples derived from it by FPLC (Figure 4). Chromatographic purification resulted in a considerable increase in the transfection efficiency. The protein fractions 3.5VAS7 and 3.5VAS8 reached the highest RLU values. In connection with the transfection studies, the results of the purification procedure can be summarized as follows. After purifying the 2V and 2.5 V fractions in two steps according to the scheme 2V→2VA→2VAS1 and 2VAS2 (2.5V in the same way), a single protein component was found in 2VAS1/S2 and 2.5VAS1/S2 which was identified as histone H1 (see Figure 1b). However, the results of purification of the 3V and 3.5V fractions were more complicated (Figure 1c). Purification was performed as before. Histone H1 was still only present in a small concentration (3VAS2, 3.5VAS9). A small but significant increase in transfection activity with (2.6 ± 0.9) × 105 RLU was found in the fractions 3VAS1 and 3.5VAS3 which was identified as bovine HMG17 by its N-terminal sequence (not shown). Furthermore, it was found that other samples such as 3.5VAS5 and in parti- Figure 4 Transfection efficiencies in terms of RLU obtained from different protein–DNA complexes. The protein fractions 3.5V, 3.5VA and 3.5VAS3–3.5VAS9 are used, 2 mg pCMV Luc, 2 × 105 NIH 3T3 cells. The values are means ± s.e. (six experiments). cular 3.5VAS7 and 3.5VAS8 were highly active in transfection (Figure 4). These results together with electrophoresis data suggest that another transfection-active protein with a somewhat higher electrophoretic mobility in the acetic acid/urea system other than H1 (Figure 1c) could be present. According to these results, we believe that the increase in transfection activity is due to both the loss of inhibiting protein components and the presence of new, nonidentified proteins. Because the presence of Ca ions is compulsory for obtaining high transfection efficiencies, it was necessary to investigate the role of Ca ions in more detail. In particular we wanted to know if Ca ions are a necessary part of the protein–DNA complex itself, ie whether they are required for the complex formation with DNA and/or as stabilizers against dilution dissociation. This latter question arose from the need to dilute the complexes during addition to the cell culture medium. However, H1–DNA complexes should be stable (see Ref. 9). We performed transfection experiments with Ca in the complex mixture using both Ca-free medium and Ca-containing cell culture medium. Only low expression could be detected if Ca was not present in the cell culture medium, independent of the presence or absence of Ca in the complex mixture (Figure 5). Thus, the presence of Ca ions is not necessary for the complex formation, but seems to be needed as a stabilizing agent against dilution dissociation. On the other hand, as discussed later, Ca ions could also play a protective role for the complexes inside the cell. In another experiment, we optimized the Ca concentration in the medium. Since Ca is known to be toxic for cells in high concentration, we also varied the time of exposure of the transfecting complexes in the Ca-containing cell culture medium to the cells. The results of such experiments using protein fraction 2.5VAS1 (H1 histone) are shown in Figure 5. After the recorded times of cell contact, the transfection mixture was replaced with fresh culture medium. It is noteworthy that the presence of 10% calf serum in the medium had no negative effect on Transfection-active nuclear proteins SV Zaitsev et al 589 Figure 5 Dependency of the transfection efficiency of the protein fraction 2.5VAS2 in the complex with DNA on the concentration of CaCl2 . Transfection efficiency in terms of RLU (2 mg pCMV Luc, 2 × 105 NIH 3T3 cells). (a) Four hours transfection time; (b) 12 h transfection time. The values are means ± s.e. (at least three experiments). the transfection process. Maximum expression was seen at 4–8 mm CaCl2 after 4 h (Figure 5a) and at 1–2 mm CaCl2 after 12 h (Figure 5b) of incubating the cells with the transfection mixture. An inverse relationship between incubation time and Ca concentration with respect to transfection efficiency seems to exist. On the other hand, at extended transfection times the morphological appearence of the cells deteriorates. Thus, we prefer to work at a transfection time of 4 h. The question whether the transfection effects observed here could be due, at least in part, to the usual calcium phosphate-mediated DNA coprecipitation was answered by our experiments. This possibility cannot be excluded because of the presence of phosphate in RPMI. Figure 6 shows the results of the DNA-Ca phosphate coprecipitation technique after preforming 2.5VAS1–DNA complexes and performing the Ca phospate precipitation in the usual way. The conditions of DNA and protein concentration were the same as in the control experiments using the 2.5VAS1 fraction as the transfection-active protein. We observed only a small positive effect of 2.5VAS1 on the transfection efficiency of the Ca phosphate method as compared with the control. It is further noteworthy with regard to this question that the control with free plasmid DNA in RPMI and in the presence of Ca++ was negative in transfection (the RLU value corresponds to that of the untreated cells of about 2000 RLU). A further point which provokes interest with respect to the practical application of these protein fractions is their cytotoxicity. Many of the nonviral vectors such as Figure 6 A comparison of the transfection efficiencies in terms of RLU obtained from 2.5VAS1–DNA complexes (open columns) and of the Ca phosphate coprecipitation method of preformed 2.5VAS1–DNA complexes (hatched columns), pCMV Luc (2 mg), 2 × 105 NIH 3T3 cells; the values are means ± s.e. (at least 10 experiments). Free DNA in the presence of CaCl2 is negative in transfection. lipofectin and polylysine are cytotoxic. Using the XTT test, we checked the cytotoxicity of 2.5VAS1 alone, in the presence of CaCl 2 and as a protein–DNA complex with and without Ca++ (Figure 7). There was no evidence for cytotoxicity under incubation conditions as used for the transfection experiments. After 16 h of incubation, the viability slowly decreased (not shown here). Finally, in order also to demonstrate the validity of our conclusions for other cell lines including primary cells, we transfected a number of different eukaryotic cell types with 2.5VAS1–pCMV Luc complexes (Figure 8). The results confirm our findings on the efficacy in transfection of these nuclear proteins. HUVEC cells are transfectable, however, with less reproducibility. There were experiments in which no transfection was observed. The permanent endothelial cell line ECV304 was constantly able to be transfected. Discussion These results confirm and extend our observations that histone H1 and other nuclear proteins like HMG1 (HMG1 however to a lower extent) are efficient mediators of transfection in complex with DNA (Böttger et al5,8). An important requirement for this activity is the presence of Ca ions in the cell culture medium. In the absence of Ca++ the transfection activity is essentially lower (Figure 5). Recently, Fritz et al10 demonstrated an effective in vitro gene transfer using histones including H1 and a modified H1 molecule with a SV 40 nucleophilic signal as the DNA carrier. However, for obtaining high expression results, the presence of chloroquine was necessary. CaCl2 was not present in the transfection mixture. Using galactosylated histones complexed with DNA and targeting the asialo- Transfection-active nuclear proteins SV Zaitsev et al 590 Figure 7 Cell viability of NIH 3T3 cells as the percentage of controls in the presence of 2.5VAS1 alone (open columns), additionally in the presence of 4 mM CaCl2 (upward hatched columns), in the presence of the 2.5VAS1–DNA complex (downward hatched columns) and, additionally, in the presence of 4 mm CaCl2 (crossed-hatched columns) as a function of the protein concentration determined by the XTT assay. The test was performed in 96-well plates with 4 × 104 NIH 3T3 cells, protein concentrations as given and 2 mg/ml pCMV Luc DNA, mean ± s.d., three experiments. Figure 8 Comparison of 2.5VAS1-mediated transfection of different cell lines and HUVEC cells. All cell types (2 × 105 cells each) were transfected with 2 mg DNA, ri ratios between 2 and 20, means ± s.e. from at least four experiments. glycoprotein receptor, Chen et al11 were able to transfect human liver cells. The best result was obtained with galactosylated H1. Without galactosylation there was, however, no detectable transfection activity. These authors used CaCl2 in the culture medium. In contrast to the results cited, we found in the presence of Ca a high transfection activity for NIH 3T3 cells and other cell lines, including human liver cells, using H1 containing protein fractions in the complex with DNA. A commercial H1 sample was also active.8 Therefore, we need to discuss the ability of certain nuclear proteins to mediate DNA transfection and the role of Ca ions in this process. We suggest that there are common features in H1 histone and other nuclear proteins such as HMG1/2 and HMG17 which can mediate DNA transfection. We suggest that the ability to bind and condense DNA9,12 together in the presence of nuclear location sequences13 are such common features. The structures of H1–DNA and HMG1–DNA complexes under similar ionic conditions are well documented by means of physicochemical and electron microscopic methods. Whereas HMG1–DNA complexes resulted in nearly spherical condensed particles of about 40 nm diameter involving one or a few DNA molecules,5,12 H1– DNA complexes often have a cable-like condensed appearance involving many DNA molecules.14,15 Furthermore, refolded DNP double fibers and large aggregated condensates (@100 nm) were found. The small size of HMG1–DNA complexes should be well suited for in vivo gene transfer studies. It is worth mentioning in this context that core histones were also shown to be transfectionactive but with lower efficiency10 (and our unpublished data). The effect that H1 inhibits or blocks transcription in cell-free systems can obviously be overcome in intact cells.16 This is not astonishing given the large number of DNA molecules involved and of the dynamic character of the electrostatic binding of these proteins to DNA. An aspect which needs further elucidation is the pathway by which the protein–DNA complexes are taken up by the cells and reach the nucleus. Possibly, Ca ions in the culture medium play an important role in this process. They could for instance transform the cytoplasmic membrane into a state more suitable for penetration of the condensed and aggregated protein–DNA complexes (see Epstein et al17). On the other hand, endocytosis cannot be excluded. In this context it is worth mentioning that we also, similar to Fritz et al,10 were able to transfect HepG2 cells by means of the fraction 2.5VAS1 representing H1 in the presence of chloroquine (data not shown). Chloroquine neutralizes the acid pH of endocytotic vesicles, inhibiting hydrolases in endosomes and lysosomes which results in decreasing degradation of macromolecules. This results in enhanced transfection rates. Hence, chloroquine could indicate that endocytosis is involved in the H1-mediated transfection. Also, the size of the complexes far exceeding 100 nm which are taken up by the cells could indicate an endocytotic mechanism. However, in the presence of Ca ions (6 mm) chloroquine is inactive in enhancing transfection (data not shown). This could mean that Ca competes with chloroquine in the function of releasing intact protein–DNA complexes into the cytoplasm. It is noteworthy that, according to our experiments, Ca does not enhance the transfection efficiency of Lipofectamin (unpublished) which supports this hypothesis. Materials and methods Protein extraction and fractionation The proteins were extracted from calf thymus nuclei with 5% perchloric acid at 0°C as suggested by Sanders and Transfection-active nuclear proteins SV Zaitsev et al Johns18 with minor modifications. Total proteins contained in the acid extract were fractionated by gradual addition of increasing volumes of cold acetone/0.1 m HCl. Fractions termed 1V, 2V,...6.5V (V, abbreviation for volumes of acetone), were obtained in this way. Each fraction was collected by centrifugation, dried and stored at −20°C (for more details, see Böttger et al8). Chromatographic separation, purification and sample labeling Chromatography was performed using an automatic FPLC gradient system with UV monitor and fraction collector from Pharmacia Biotech, Uppsala, Sweden using both the anionic and cationic exchange columns Mono Q HR 5/5 and Mono S HR 5/5 (Pharmacia Biotech, Freiburg, supplier). The transfection efficiency of the separated protein samples was taken as a criterion for a successful purification step. Chromatographic separations were performed after loading of 2 ml of fractions 2V– 3.5V in a concentration of 3 mg/ml in buffer A on the Mono Q column and elution by a linear salt gradient of 3–73% of buffer B. Then the material not bound to the column (breakthrough), exhibiting a high transfection activity, was further purified on Mono S under the same elution conditions using buffer C. All fractions obtained were precipitated by acetone, dried and stored at −20°C. The following buffers were used: A: 0.02 m NaCl, 0.05 m Tris-HCl, pH 7.4, 0.02 m b-mercaptoethanol; B: 0.82 m NaCl, 0.05 m Tris-HCl, pH 7.4, 0.02 m b-mercaptoethanol; C: 1.82 m NaCl, 0.05 m Tris-HCl, pH 7.4, 0.02 m b-mercaptoethanol. The following system was used for labeling and characterizing the samples obtained by chromatography. 2V–5.5V: fractions obtained by PCA extraction and subsequent precipitation with 2 volumes (V) to 5.5V acetone. 2VA–3.5VA: fractions 2V–3.5V, but chromatographed on Mono Q (breakthrough fractions only). 2VAS1, 2VAS2 etc: samples obtained from fraction 2VA, but additionally separated on Mono S, where numbers 1,2 etc give the order of elution. Polyacrylamide gel electrophoresis Electrophoresis was performed according to Paniym and Chalkley19 using 15% polyacrylamide slab gels and Coomassie blue staining. Protein identification Proteins were sequenced by N-terminal amino acid sequencing using a 477A sequencer (Applied Biosystems, division of Perkin Elmer, Foster City, CA, USA) and identified after a search in the SwissProt database using the Fasta programme (HUSAR). Cells NIH 3T3 cells (mouse fibroblasts) were grown in RPMI 1640 medium with l-glutamine (Serva, Heidelberg, Boehringer Ingelheim, Heidelberg, Germany, supplier) supplemented with 10% heat-inactivated fetal calf serum (FCS). The human hepatoma cell line HepG2 20 and the human hepatocellular carcinoma line Huh721 were maintained in Dulbecco’s modified Eagle’s medium (DMEM) from Gibco BRL (supplied by Life Technologies, Eggenheim, Germany) containing 10% heat-inactivated FCS. The mouse colon tumor cell line CT2622 was main- tained in RPMI 1640 medium with l-glutamine supplemented with 10% heat-inactivated FCS. Endothelial cells (HUVEC cells) were prepared from human umbilical cord veins according to Jaffe et al23 and cultivated in a modified manner published by Griesmacher et al. 24 The permanent human endothelial cell line ECV304 originally established from the vein of an apparently normal umbilical cord25 was propagated in Medium 199 (Gibco BRL, Paisley, UK, Life Technologies, Eggenstein, Germany, supplier) with 10% heat-inactivated FCS. Transfection experiments Vector DNA and proteins as specified in the text were mixed in buffer D in the presence of CaCl2 (final concentration 2 mm) at room temperature. The protein was gradually added in two to three portions to the DNA under agitation in order to avoid turbidity. Protein concentration was determined by dissolving weighed samples of lyophilized material into known volumes of buffer and, where possible (H1 and HMG1), by spectrophotometry.26 After mixing, the samples were shaken for at least 15 min. The weight input ratio ri of protein to DNA was adjusted to 10–20. The final complex volume was 100 ml. About 2 × 105 cells (here NIH 3T3 cells) per well of a 24-well plate (Falcon, Franklin Lakes, NJ, USA) grown in RPMI 1640 medium with l-glutamine supplemented with 10% FCS, were used for transfection. The plasmid pCMV Luc was used as a reporter gene. The culture medium was removed before transfection and the cells were washed twice with RPMI. The protein–DNA complexes were added directly to 0.9 ml of culture medium containing 10% FCS and 8 mm CaCl2 . This transfection mixture was then added to the cells. After 4 h of incubation under 4.5% CO2 at 37°C, the transfection mixture was removed and replaced by 1 ml RPMI 1640 containing 10% FCS. After incubation for about 30 h the cells were harvested and the luciferase activity measured by the Promega assay system (supplied by Boehringer Ingelheim, Germany) using a Lumat LB 9501/16 luminometer (Berthold, Bad Wildbad, Germany). Transfection with Lipofectamin (Gibco BRL) was performed according to the manufacturer’s recommendations. DNA (2 mg) and Lipofectamin (3 ml) were both diluted separately into 100 ml Optimem I (Gibco BRL). After gently mixing both solutions at room temperature and incubation for 15 min, the mixture was diluted with 0.8 ml Optimem I and added to the washed cells. Transfection time was 4 h, then the transfection medium was replaced with normal growth medium. Further procedures were as given previously and Buffer D: 0.15 m NaCl, 10 mm Tris-HCl, pH 7.6. Calcium phosphate coprecipitation was performed using the Profection Kit from Promega (supplied by Boehringer Ingelheim, Germany). According to the manufacturer’s recommendations, the amount of solution was adapted to 20 mm dishes (24-well plates). DNA (2 mg) were mixed with a 2 m CaCl2 solution and water. Amounts of 2.5VAS1 (histone H1) according to the input ratio ri protein–DNA (w/w) selected were added to this mixture. Then, the same volume 2 × Hepes-buffered saline was added to this mixture with vortexing, final volume 150 ml. The transfecting mixture was added dropwise to 2 × 105 cells in 1 ml culture medium. 591 Transfection-active nuclear proteins SV Zaitsev et al 592 Viability measurements Viability measurements were performed in 96-well plates by means of the cell proliferation kit II (XTT) from Boehringer Mannheim, Germany. NIH 3T3 cells (4 × 104 per well) were washed after a 4 h exposure to the protein fraction 2.5VAS1 in three different concentrations (20 mg/ml, 60 mg/ml, 100 mg/ml) in 150 ml each, additionally in the presence of CaCl2 (4 mm) to the protein, and in the presence of the protein–DNA complex (protein concentration as above and 2 mg/ml pCMV luc DNA) and of 4 mm CaCl2. Then the XTT procedure was performed as recommended by the manufacturer. The test measures the ability of cells to reduce the water soluble tetrazolium dye, XTT (sodium 3′-[1-(phenyl-amino-carbonyl)-3,4tetrazolium]-bis (4-methoxy-6-nitro) benzene sulfonic acid hydrate), to its insoluble formazan salt. The insoluble salt is coloured and can be determined spectrophotometrically. 9 10 11 12 13 14 15 Acknowledgements 16 We thank Lieselotte Winkler for skillful technical assistance, Professor Charles Coutelle (London) for a critical reading of the manuscript, and Professor Burkhard Micheel (Berlin) for support. This work was supported by the Volkswagen foundation and the BMBF. 17 References 19 1 Ledley FD. Nonviral gene therapy: the promise of genes as pharmaceutical products. Hum Gene Ther 1995; 6: 1129–1144. 2 Schofield JP, Caskey CT. Non-viral approaches to gene therapy. Br Med Bull 1995; 51: 56–71. 3 Wagner E et al. Coupling of adenovirus to transferrin– polylysine/DNA complexes greatly enhances receptormediated gene delivery and expression of transfected genes. Proc Natl Acad Sci USA 1992; 89: 6099–6103. 4 Forstova J et al. Polyoma virus pseudocapsids as efficient carriers of heterologous DNA into mammalian cells. Hum Gene Ther 1995; 6: 297–306. 5 Böttger M et al. Condensation of vector DNA by the chromosomal protein HMG1 results in efficient transfection. Biochim Biophys Acta 1988; 950: 221–228. 6 Böttger M et al. Transfection by DNA–nuclear protein HMG1 complexes: raising of efficiency and role of DNA topology. Arch Geschw Forsch 1990; 60: 265–270. 7 Sandig V et al. Direct gene transfer of HMG1 based DNA–protein complexes. J Molec Med 1995; 73: B10 (Abstr.). 8 Böttger M, Zaitsev SV, Otto A, Vorob’ev VI. Acid nuclear 18 20 21 22 23 24 25 26 extracts as mediators of gene transfer and expression (submitted for publication). Zlatanova J, Yaneva J. Histone H1–DNA interactions and their relation to chromatin structure and function. DNA Cell Biol 1991; 10: 239–248. Fritz JD, Herweijer H, Zhang G, Wolff, JA. Gene transfer into mammalian cells using histone-condensed plasmid DNA. Hum Gene Ther 1996; 7: 1395–1404. Chen J, Stickles RJ, Daichendt KA. Galactosylated histonemediated gene transfer and expression. Hum Gene Ther 1994; 5: 429–435. Böttger M et al. Condensation of supercoiled DNA by the chromosomal protein HMG1. Stud Biophys 1990; 135: 121–129. Breeuwer M, Goldfarb DS. Facilitated nuclear transport of histone H1 and other small nucleophilic proteins. Cell 1990; 60: 999–1008. Böttger M, Von Mickwitz CU, Scherneck S, Lindigkeit R. Interaction of histone H1 with superhelical DNA. Conformational studies and influence of ionic strength. Mol Biol Rep 1984; 10: 3–8. De Bernardin W, Losa R, Koller T. Formation and characterization of soluble complexes of histone H1 with supercoiled DNA. J Mol Biol 1986; 189: 503–517. Croston EG et al. Sequence-specific antirepression of histone H1mediated inhibition of basal RNA polymerase II transcription. Science 1991; 251: 643–649. Epstein et al. Extracellular calcium mimics the action of plateletderived growth factor on mouse fibroblasts. Cell Growth Differ 1992; 3: 157–164. Sanders C, Johns EW. A method for the large-scale preparation of two chromatin proteins. Biochem Soc Trans 1974; 2: 547–550. Panyim S, Chalkley R. High resolution acrylamide gel electrophoresis of histones. Arch Biochem Biophys 1969; 130: 337–346. Knowles BB, Howe CC, Aden DP. Human hepatocellular carcinoma cell lines secrete the major plasma proteins and hepatitis B surface antigen. Science 1980; 209: 497–499. Nakabayashi H et al. Growth of human hepatoma cell lines with differentiated functions in chemically defined medium. Cancer Res 1982; 42: 3858–3863. Griswold DP, Corbett TH. A colon tumor model for anticancer agent evaluation. Cancer 1975; 36: 2441–2444. Jaffe EA, Nachman RL, Becker CG, Minick CR. Culture of human endothelial cells derived from umbilical veins. J Clin Invest 1973; 52: 2745–2756. Griesmacher A, Weigel G, Schreiner W, Müller MM. Thromboxan A 2 generation by human umbilical endothelial cells. Thromb Res 1989; 56: 611–623. Takahashi K, Sawasaki Y. Human endothelial cell line, ECV304, produces pro-urokinase. In Vitro Cell Dev Biol 1991; 27A: 766– 768. Smerdon MJ, Isenberg I. Interactions between the subfractions of calf thymus H1 and nonhistone chromosomal proteins HMG1 and HMG2. Biochemistry 1976; 15: 4242–4247.