Uses of a saponin and method for its isolation

Abstract

The present invention relates to the novel use of saponins having acetyl residues on one of their sugar residues. These saponins are able to enhance the transfection efficiency to a surprisingly much higher extent than already known saponins and even than Lipofectamin.

Claims

1. A method for an in vitro delivery of a nucleic acid, a lipid, a peptide and/or a protein to a cell, the method comprising the step of incubating the cell with the nucleic acid, the lipid, the peptide and/or the protein with a saponin according to formula (I): ##STR00008## wherein R.sup.1 is a xylose residue or an arabinose residue bonded with its C1 atom to the corresponding xylose residue of formula (I); and R.sup.2 is independent from other R.sup.2 residues in the same molecule H or an acetyl residue, with the proviso that at least two acetyl residues are present in the saponin.

2. The method according to claim 1, wherein the cell is a eukaryotic cell.

3. A transfection composition, comprising at least one transfection reagent chosen from the group consisting of liposomal-based transfection reagents and polymer-based transfection reagents as well as a saponin according to formula (I): ##STR00009## wherein R.sup.1 is a xylose residue or an arabinose residue bonded with its C1 atom to the corresponding xylose residue of formula (I); and R.sup.2 is independent from other R.sup.2 residues in the same molecule H or an acetyl residue, with the proviso that at least two acetyl residues are present in the saponin.

4. The transfection composition according to claim 3, wherein R.sup.1 is a xylose residue and/or the saponin carries exactly two acetyl groups.

5. The transfection composition according to claim 4, wherein the acetyl groups are bonded to the oxygen atoms in C3 and C4 position of the corresponding quinovose residue.

6. A method for an in vitro transfection, comprising the step of incubating a cell with a nucleic acid in the presence of a saponin according to formula (I): ##STR00010## wherein R.sup.1 is a xylose residue or an arabinose residue bonded with its C1 atom to the corresponding xylose residue of formula (I); and R.sup.2 is independent from other R.sup.2 residues in the same molecule H or an acetyl residue, with the proviso that at least two acetyl residues are present in the saponin.

7. The method according to claim 6, wherein the cell is a eukaryotic cell.

8. The method according to claim 6, wherein the nucleic acid forms part of a nanoparticle.

9. The method according to claim 6, wherein the saponin is used in a concentration lying in a range of 1 ng/mL to 25 ng/mL.

10. The method according to claim 6, wherein the saponin is used in combination with at least one transfection reagent chosen from the group consisting of liposomal-based transfection reagents and polymer-based transfection reagents.

11. A method for isolating a saponin according to formula (I): ##STR00011## wherein R.sup.1 is a xylose residue or an arabinose residue bonded with its C1 atom to the corresponding xylose residue of formula (I); and R.sup.2 is independent from other R.sup.2 residues in the same molecule H or an acetyl residue, with the proviso that at least two acetyl residues are present in the saponin from Gypsophila elegans M. Bieb, comprising the following steps: cutting roots of Gypsophila elegans M. Bieb, freeze-drying and grinding the cut roots to obtain a root powder, extracting the root powder with a high-concentrated organic solvent to obtain a root extract, removing the high-concentrated organic solvent from the root extract to obtain a dry extract, dissolving the dry extract in a low-concentrated organic solvent to obtain an extract solution, subjecting the extract solution to at least one chromatographic separation step to obtain a purified saponin solution, and removing any solvent from the purified saponin solution to obtain a purified saponin powder.

Description

(1) Further details of aspects of the present invention will be explained with respect to an exemplary embodiment and accompanying Figures. In the Figures:

(2) FIG. 1 shows an HPLC chromatogram of dry raw extract from the roots of Gypsophila elegans M. Bieb;

(3) FIG. 2 shows an LC-ESI-MS chromatogram of the substance present in peak P2 of the chromatogram of FIG. 1 after further purification by semi-preparative HPLC;

(4) FIG. 3 shows and LC-ESI-MS chromatogram of the isolated saponin GE1741;

(5) FIG. 4 shows an MS/MS spectrum of the isolated saponin GE1741;

(6) FIG. 5 shows the results of flow cytometric measurements;

(7) FIG. 6 shows the transfection efficiency of nanoplexes±triterpene saponins in Neuro2a cells;

(8) FIG. 7 shows the chemical structure of GE1741 elucidated by mass spectrometry and NMR spectroscopy;

(9) FIG. 8 shows a plot on the transfection efficiency of different transfection enhancers;

(10) FIG. 9 shows the impact on GE1741 on the transfection efficiency of different commercial transfection methods;

(11) FIG. 10 shows the results of a first experiment for examining the toxicity of GE1741 on Neuro2a cells;

(12) FIG. 11 shows results of a second experiment for examining the toxicity of GE1741 on Neuro2a cells and

(13) FIG. 12 shows the results of a targeted suicide gene therapy.

EXEMPLARY EMBODIMENT: ISOLATION OF THE SAPONIN GE1741 FROM GYPOSPHILA ELEGANS M. BIEB

(14) Seeds from Gypsophila elegans M. Bieb were seeded on a field in the state of Brandenburg, Germany. The mature plants were harvested and the roots were cut. The roots were washed, freeze-dried and ground to powder. The powder was extracted by 90% methanol in water. Following filtration, the methanol was removed by vacuum distillation. The remaining aqueous extract was finally freeze-dried (dry extract). The dry extract was dissolved in 30% methanol at a concentration of 60 mg/mL.

(15) The solution (each 0.5 mL) was subjected to semi-preparative HPLC using a Kinetex® 5 μm C18 100 Å, LC Column 250×10.0 mm. A methanol(A)/water, 0.01% TFA (B) gradient was used from 30% to 90% (A) over 20 min and then to 30% (A) over 10 min. Flow rate was 4 mL/min. Detection wavelength was at 210 and 300 nm. A corresponding chromatogram is depicted in FIG. 1, wherein the upper curve shows the detector signal at 210 nm and the lower curve the detector signal at 300 nm. The peak at a retention time (RT)≈18 to 20 (≈19) min (marked as P2) was collected. A high number of isolation cycles was repeated and the collected peaks were finally pooled. The methanol was evaporated by vacuum centrifugation and the remaining water was removed by freeze-drying.

(16) The dried material was dissolved in 50% acetonitrile at a concentration of 4 mg/mL. The solution (each 0.5 mL) was subjected to semi-preparative HPLC using a Kinetex® 5 μm C18 100 Å, LC Column 250×10.0 mm. An acetonitrile (A)/water, 0.01% TFA (B) gradient was used from 30% to 50% (A) over 24 min and then to 30% (A) over 1 min. Detection wavelength was at 210 nm. Flow rate was 4 mL/min. The peak at RT≈18 min was collected. Several cycles were repeated. The acetonitrile was evaporated by vacuum centrifugation and the remaining water was removed by freeze-drying.

(17) The samples were analyzed by using an Agilent 6200 Series Q-TOF LC-ESI-MS/MS system. For LC-ESI-MS analysis of the recovered fractions a Kinetex® C18 HPLC column (2.6 μm, 100 Å, (150×4.6 mm), and acetonitrile (A)/water (B) gradient, 0.01% formic acid was used starting with 30% to 70% (A) over 30 min using a flow rate of 0.7 mL/min. The corresponding chromatogram is depicted in FIG. 2. Two prominent peaks were detected. A mass of 1538.67 was assigned to the substance of the first peak, and a mass of 1741.73 was assigned to the substance of the second peak. The second peak was chosen for further purification.

(18) The dried material of the second peak was dissolved in 50% acetonitrile at a concentration of 4 mg/mL. The solution (each 0.5 mL) was subjected to semi-preparative HPLC using the semi-preparative Kinetex® C-18 column (see above). An acetonitrile (A)/water, 0.01% TFA (B) gradient was used from 30% to 50% (A) over 24 min and then to 30% (A) over 1 min using a flow rate of 4.0 mL/min. Detection wavelength was at 210 nm. The peak at RT≈18 min was collected in several repeating cycles. The acetonitrile was evaporated by vacuum centrifugation and the remaining water was removed by freeze-drying.

(19) The dry material was dissolved in 50% acetonitrile and subjected (0.5 mL per run) to HPLC using an Ultrasep ES PEO, LC Column, 250×4 mm, 5 μm. The column was conditioned by 5% acetonitrile. The polarity of the solvent was abruptly changed by applying an acetonitrile(A)/water, 0.01% TFA (B) gradient starting with 100% to 5% (A) from 1.25 to 21 min. Detection wavelength was at 210 nm and flow rate was 1 ml/min. The product (GE1741; also referred to as gypsophilosid A) was collected at RT 4-6 min. The acetonitrile was removed by vacuum centrifugation. After freeze drying GE1741 was obtained as white powder.

(20) The product was once again analyzed by LC-ESI-MS as outlined above. The corresponding chromatogram is depicted in FIG. 3. Only one prominent peak was detected, indicating a successful purification. A mass of 1741.73 was assigned to the substance of this peak. This mass gave the product its name, GE1741.

(21) For elucidating the chemical structure of GE1471, LC-ESI-MS/MS measurements were performed. A corresponding spectrum is shown in FIG. 4.

(22) Based on this mass spectrometric data and further data obtained by NMR spectroscopy (see below for details), it could be established that the structure of this saponin corresponds to formula (VIII), wherein R.sup.1 is a xylose residue. However, it is assumed that the biological function of GE1741 will most probably not depend on the concrete structure of residue R.sup.1. Rather, the presence of at least one acetyl residue on the quinovose residue of GE1741 is considered to be a relevant factor for the properties of GE1741 that will be explained in the following in more detail. While GE1741 comprises two acetyl residues (one in C3 position and one in C4 position), preliminary data suggests that the amount and position of the acetyl residues can be varied within the indicated limits without significantly changing the properties of the respective saponin.

(23) Testing the Properties of GE1741

(24) Cell Culture

(25) Murine neuroblastoma cells (Neuro2a cells, ATTC CCL-131™) and human colorectal carcinoma cells (HTC-116, (ATCC® CCL-247™) were cultured in Dulbecco's Modified Eagle's Medium (DMEM), containing 1 g/L D-Glucose, 10% FBS and stable glutamine. Neuro-2A-cells were incubated at 37° C. and 5% CO.sub.2. These cells were then incubated with DNA nanoparticles (YD) with or without the novel saponin GE1741.

(26) The DNA nanoparticles used for the transfection experiments contained DNA encoding for the green fluorescent protein (GFP). Therewith, the transfection efficiency could very easily be monitored by detecting the fluorescence of the cells that have been incubated with the corresponding DNA nanoparticles. The corresponding incubation period was chosen to be 48 hours in the present case.

(27) As negative control (left panel of FIG. 5), no DNA nanoparticles at all have been added to the Neuro2a cells. A fluorescence indicating an apparent transfection efficiency of 0.37% could be observed. This value represents the measuring error.

(28) If peptide minicircle DNA particles (PM) have been added to the Neuro2a cells without any transfection enhancer, a transfection efficiency of only 1.35% could be observed based on the detected fluorescence (middle panel of FIG. 5). If, however, the same peptide minicircle DNA particles have been added together with GE1741 (the novel saponin), a transfection efficiency of 98.65% could be observed based on the detected fluorescence (right panel of FIG. 5). These flow cytometric measurements detecting fluorescent cells already indicates a high transfection efficiency enhancement by GE1741.

(29) Formulation of PD/YD—Nanoplexes

(30) 20 mg of positively charged poly-lysine peptides with (Y) and without (P) an integrin receptor targeting amino sequence were purchased from Genecust. The vector of p-EGFP-N3, coding for the green fluorescing protein (GFP), was obtained and propagated with DH5α—E. Coli cells (1.645 mg/mL). StemMACS™ eGFP mRNA (20 μg) and GFP encoding minicircle DNA (Gene Bank Accession: U55761) were used as further nucleic acids. In order to conduct the transfection, nanoplexes were formulated as follows: The poly-lysine peptides (P or Y) and the p-EGFP-N3 (D), mRNA (m) or minicircle (M) vector were diluted in water (each 50 μL) and mixed thoroughly by fast pipetting in a ratio of 4:1. The nanoplexes were allowed to form in a 30-minute incubation step. Thereafter, the nanoplex suspension was diluted with OptiMEM to a total volume of 1 mL. The commercial transfection reagents TransIT-X2® Dynamic Delivery System, Xtreme GENET™ HP Transfection Reagent, Genecellin and Lipofectamin® were formulated as described by the manufacturer.

(31) Sapofection (Transfection with Triterpene Saponins)

(32) Neuro2a cells or HTC-116 cells (15,000 cells/well) were seeded in a 24-well-plate with a well volume of 400 μL culture medium and incubated for 24 h. The transfection reagents were formulated as described above and admixed with saponin solution, if required. The culture medium was replaced with the transfection medium with a final amount of 500 ng DNA/RNA. After a 48 h (DNA transfections) or 24 h (RNA transfections) incubation period, the transfection medium was removed, the cells were trypsinized and transferred in a polystyrene tube for flow cytometry (Cytoflex). For each measurement 10,000 cells were acquired. The transfection efficiency was determined by the analysis software Cyflogic (by comparison of the sample plots with the negative control in terms of fluorescence.

(33) Cell Impedance Measurements

(34) For real-time toxicity measurements the impedance measuring device iCELLigence® was deployed. 8000 Neuro2a cells per well were seeded into two 8-well E-plates L8 and incubated for 24 h in a volume of 800 μL. For transfection toxicity studies 50 μL of reagent was added after the respective volume of culture medium was removed. A non-toxic concentration of 2 μg/mL GE1741 was applied. For permeability studies increasing concentrations of GE1741 from 2.5 μg/mL to 60 μg/mL were applied to the cells. Each 10 minutes the impedance/viability was measured. The results were analyzed and displayed with the RTCA Data Analysis Software.

(35) NMR Spectroscopy

(36) The samples for NMR-spectroscopy were prepared by dissolving 2 mg of GE 1741 in 600 μl d.sub.4-methanol (99.95% deuterium content, Deutero, Kastellaun, Germany) and by dissolving 4 mg of GE 1741 in 600 μl d.sub.5-pyridine (99.9% deuterium content, Deutero, Kastellaun, Germany) to yield a concentration of 1.9 and 3.8 mM/l, respectively. The solutions were transferred to 5 mm sample tubes which were sealed to prevent evaporation of solvent or accumulation of water.

(37) NMR spectra were recorded at 300 K at 600 (.sup.1H frequency) on Bruker AV-III spectrometers (Bruker Biospin, Rheinstetten, Germany) using cryogenically cooled 5 mm TCI-triple resonance probe equipped with one-axis self-shielded gradients. The software used to control the spectrometer was topspin 3.5 pl6. Temperature had been calibrated using do-methanol and the formula of Findeisen et al. (Findeisen 2007).

(38) Experiments using the sample in methanol: One-dimensional .sup.1H- and .sup.13C-spectra were recorded at 600 MHz using 32 and 18432 scans, respectively. In case of the carbon spectrum proton broadband decoupling was applied. Homonuclear spectra: A DQF-COSY (Piantini et al., 1982), a TOCSY (Braunschweiler and Ernst, 1983; Bax and Davis, 1985a) and a NOESY (Jeener et al., 1979) were recorded using 2048×512 complex data points and acquisition times of 204 and 61 ms in F2 and F1, respectively. DQF-COSY and TOCSY (mixing time 80 ms) were recorded using 8 scans, the NOESY using 64 scans with a mixing time of 80 msec. The heteronuclear spectra were recorded as follows: A .sup.13C-HSQC (Bodenhausen and Ruben, 1980), .sup.13C-HSQC-TOCSY, .sup.13C-HSQC-CLIP-COSY and .sup.13C-HSQC-NOESY were recorded using 512×1024 complex data points, acquisition times of 51 and 45 ms in F2 and F1, respectively, and 48, 96, 96 and 288 scans, respectively. A .sup.13C DEPT-HMQC (Kessler et al., 1989), and a .sup.13C-HQQC (Kessler et al., 1991) were recorded using 512×512 complex data points, acquisition times of 51 and 23 ms in F2 and F1, respectively, and 64 and 60 scans, respectively. For the determination of the .sup.1J HC and the .sup.3J HH a .sup.13C-HSQC was recorded without refocussing and carbon decoupling using 16384×512 complex data points, acquisition times of 1638 and 73 ms in F2 and F1, respectively, and using 32 scans. All the above heteronuclear spectra were recorded using a BIRD pulse for suppression of protons bound to .sup.12C (Bax and Subramanian, 1986). A gradient-enhanced .sup.13C-HMBC (Bax and Summers, 1986; Cicero et al., 2001) was recorded using 2048×1024 complex data points, acquisition times of 204 and 33 m sin F2 and F1, respectively, and 128 scans.

(39) Experiments using the sample in pyridine: One-dimensional .sup.1H- and .sup.13C-spectra were recorded at 600 MHz using 32 and 14349 scans, respectively. In case of the carbon spectrum proton broadband decoupling was applied. A DQF-COSY (Piantini et al., 1982), was recorded using 2048×512 complex data points, 8 scans and an acquisition times of 204 and 61 ms in F2 and F1, respectively. A .sup.13C-HSQC (Bodenhausen and Ruben, 1980), was recorded using 512×1024 complex data points, 32 scans and an acquisition times of 51 and 45 ms in F2 and F1, respectively. A gradient-enhanced .sup.13C-HMBC (Bax and Summers, 1986; Cicero et al., 2001) was recorded using 2048×1024 complex data points, acquisition times of 204 and 33 ms in F2 and F1, respectively, and 64 scans.

(40) Data were processed using topspin3.2, typically a squared sine bell shifted by 90° was used in both dimensions, in case of the HMBC a sine bell was used in F2. Datasets were processed to yield a data matrix of 4096 by 2048 points.

(41) Isolation and Characterization of GE1741

(42) Initially the seed and root extract was tested regarding transfection enhancing activity. The results are shown in FIG. 6. The HPLC chromatogram of the dry root extract (cf. FIG. 1) was analyzed, the peaks P2, P3 and P4 were collected and their biological activity was tested for their biological activity. These results are also shown in FIG. 6. Subsequently, P2 was subjected to further HPLC purification, as explained above. Subsequent LC-ESI-MS guided semi-preparative purification steps (cf. FIGS. 2 and 3) led to an isolation of a peak, which activity was repeatedly tested. LC-ESI-MS/MS measurements detected in the negative ionization mode the deprotonated molecular ion signal [M-H]—at m/z 1741 (cf. FIG. 4). The MS/MS fragment ion signal at m/z 955 was assigned to the saponin after secession of the di-acetyl-pentasaccharide chain linked to C-28 of the quillaic acid, and m/z 460 resulted after the additional cleavage of the trisaccharide unit at C-3. The molecular formula was determined to be C.sub.79H.sub.122O.sub.42 (calculated: 1741.73325) from the measured mass for the negatively charged ion (HR-ESI-MS m/z 1741.7336). This corresponds to a mass difference of only 0.2 ppm.

(43) 1D- and 2D-NMR Spectroscopy of GE1741

(44) The structure of the bidesmosic quillaic acid type saponin GE1741 was fully elucidated by 1D and 2D-NMR spectroscopic experiments, including .sup.1H, .sup.13C and HSQC, HMBC, HQQC, DQF-COSY, TOCSY, HSQC-TOCSY and NOESY using d4-methanol. Amongst the signals in the .sup.13C NMR spectrum of GE174, two double intensity resonances (δC 75.33, 71.43 ppm) indicated completely overlapping of C-atoms. As partial structural units, a triterpene moiety and the existence of 8 sugars with two points of acetylation was seen. Using HSQC and HQQC experiments, six sp3-hybridised resonances (δC 10.9, 16.4, 17.7, 24.8, 27.1, 33.3 ppm) were recognized as methyl groups in the triterpene moiety, and additionally three methyl groups were assigned to deoxy-sugars (δC 17.0, 17.6, 18.3). The .sup.1H NMR gave singlet resonances (δH 0.74, 0.87, 0.94, 1.00, 1.16, 1.38 ppm) for the triterpene methyl-functions, and doublets (δH 1.16, 1.23, 1.31) for the deoxy-functions.

(45) The C30 corpus displayed two methyl groups less than expected which was explained by the occurrence of an aldehyde function (δC 211.5 ppm, δH 9.47 ppm) and an acidic function (δC 175.9 ppm). Together with resonances for a double bond at δC 123.2 (CH), 144.8 ppm (quaternary C), a broad triplet signal (δH 5.31 ppm), a substituted oxymethin function at C-3 (δC 86.4, δH dd 3.86) combined with HMBC correlation data elucidated a hydroxy-oleanan-12-ene type backbone. The corresponding results are shown in FIG. 7 and will be explained in more detail in the following

(46) .sup.2,3J CH correlations from the aldehyde proton δ 9.47 ppm to C-4 (56.3 ppm), CH3-24 (δ 10.9), and C-5. The long-range correlation of H-3 (δ 3.86) to the quaternary position C-4 and C-24 suggested the CHO-function to be at C-23. The second hydroxylation at C-16 (δC 74.5) was seen by .sup.2,3J-CH signals from protons of H-18 (δ 2.94), CH2-22 (δ 1.93, 1.16). A total number of eight monosaccharides were detected in the structure of GE1741 by .sup.1H- and .sup.13C NMR spectroscopy by observation of characteristic chemical shifts of anomer resonances. The nomenclature of GE1741 is as follows: 3-O-(ß-D-Galactopyranosyl-(1.fwdarw.2)-[ß-D-xylopyranosyl-(1.fwdarw.3)]-ß-D-glucuronopyranosyl)-28-O-(ß-D-xylopyranosyl-(1.fwdarw.3)-[ß-D-xylopyranosyl-(1.fwdarw.4)]-α-L-rhamnopyranosyl-(1.fwdarw.2)-[3,4-di-(O-acetyl)-ß-D-quinovopyranosyl]-(1.fwdarw.4)-ß-D-fucopyranosyl)-quillaic acid

(47) A trisaccharide unit is linked to C-3 and a hexasaccharide to C-28 in this bisdesmosidic quillaic acid type saponin. The sugar identities, and linkages with respective 1H/13C connectivities were unambiguously identified by a combination of HMBC (cf. FIG. 7), HSQC (high resolution), HSQC-DEPT, HQQC, HSQC-TOCSY, and DQF-COSY, .sup.1H/.sup.1H-TOCSY, .sup.1H/.sup.1H-NOESY, and non-decoupled HSQC. The .sup.1J CH-connectivities of all sugar methin atoms were clearly derived by HSQC recorded with high resolution in the indirect dimension (TD: F2 1024, F1 2048).

(48) The trisaccharide linked to C-3 consist of a glucuronic acid where the free carboxylic acid function at C-6 was not detected by .sup.13C nor by HMBC, a galactose, and a xylose with anomeric resonances at δH/δC in [ppm] and .sup.3J HH/.sup.1J CH in [Hz]: β-GlcA-1 (δ 4.42, 104.6; d, 7.4, 161), β-Gal-1 (δ 4.80, 103.7; d, 7.8, 163), β-Xyl-1 (δ 4.57, 104.9; d, 7.8, 162). The pentasaccharide unit at C-28 consist of β-Fuc-1 (δ 5.29, 94.8; d, 7.8, 166), α-Rha-1 (δ 5.36, 101.5; d, 1.4, 171), β-Xyl′-1 (δ 4.49, 107.0; d, 7.2, 159), β-Qui-1 (δ 4.61, 105.8; d, 8.2, 161), β-Xyl″-1 (δ 4.51, 105.6; d, 7.0, 162).

(49) Stereochemical properties of the glycosidic linkages were determined by characteristic values of .sup.3J HH and .sup.1J CH coupling constants J (Hz), respectively. A large trans-diaxial coupling (.sup.3J HH: 7-8 Hz) for H-1/H-2 in case of glucose and galactose type sugars define a β-configuration. Characteristic .sup.1J CH coupling values indicate the anomeric configuration of hexopyranoses with ˜170 Hz for an α, and ˜160 Hz for a β configuration. The detected values were in absolute accordance to these literature data and can also be applied for the quinovose (6-deoxy-glc), fucose (6-deoxygal) and rhamnose (6-deoxy-mannose).

(50) The .sup.1J CH coupling values J in [Hz] for the anomers were directly derived from the HMBC. In the .sup.1H NMR of GE1741, the so-called ‘bulk area’ with non-anomeric protons occurred in the range 3.15-3.95 ppm, partly four to five sugar proton signals overlapped. .sup.3J HH methin proton coupling pattern in the different sugar ring systems were detected and extracted by a non-decoupled HSQC experiment measured in the highly crowded signal range (.sup.1H: δ 3.15-3.95 ppm, and .sup.13C: δ 60-90 ppm where .sup.1J CH and almost all .sup.3J HH coupling constants were visualized and accurately determined. The non-decoupled HSQC clearly distinguished and elucidated GlcA from a possible GalA linked to C-3 seen by a large trans-diaxial coupling pattern of J 8.0 Hz (triplet) of GlcA-H-4 (δ 3.55 ppm) to H-3 and H-5. Also the TOCSY signal GlcA-H-1 (4.42 ppm) (spin-lock time: 50 ms), and the HSQC-TOCSY detected the position GlcA-5 indicating large coupling constants in this spin-system.

(51) The identified anomer 1H signals were used as ‘structural-reporter groups’ (Vliegenthart, 2006), (Bubb, 2003), (Duus et al., 2000)). All independent spin-systems with scalar couplings in the existing ‘bulk region’ (δH: 3.15-4.7 ppm) in the oligosaccharide chains of GE1741 were elucidated by a combination of DQF-COSY, .sup.1H/.sup.1H-TOCSY, and were confirmed by a HSQC-TOCSY experiment. The long-range HMBC correlation signals (cf. FIG. 7) clearly identified the interglycosidic connecting points, e.g. connection to C-28 (δ 177.1) seen by Fuc-H-1 (δ 5.29), and the sequence and linkages of the branched pentasaccharide units with the central sugar unit Fucose: Fuc-H-4 (δ 3.76) to Qui-C-1 (δ 105.8), Fuc-H-2 (δ 3.79) to Rha-C-1 (δ 101.5), Rha-H-4 (δ 3.51) to Xyl′-C-1 (δ 107.0), and Xyl′-H-4 (δ 3.52) to Xyl″-C-1 (δ 105.6).

(52) The two significant acetylation shifts observed for Qui-H-3 (δ 5.01) and Qui-H-4 (δ 4.63) clearly indicated the acetylation in the quinovose and moved the sugar methin resonances out of the ‘proton bulk region’. The two points of acetylation were clearly elucidated by .sup.2,3J CH long range cross-signals in the HMBC (cf. FIG. 7 from Qui-H-3 and Qui-H-4 to the acetate carbonyl functions (δ 172.1, 171.6). This double acetylation in the quinovose unit of GE1741 was unexpected and can be seen as a relevant structural feature and possibly as an important factor for the high activity of this saponin in the conducted assays.

(53) The connection of the branched trisaccharide unit starting at glucuronic acid is seen by the .sup.2,3J CH from GlcA-H-1 (δ 4.42) to C-3 (δ 86.4), and vice versa from H-3 (δ 3.86) to GlcA-H-1. Galactose is linked to GlcA seen by .sup.2,3J CH from GlcA-H-2 (δ 3.64) to Gal-C-1 (δ 103.7) and vice-versa by Gal-H-1 (δ 4.80) to Fuc-C-2 (δ 75.1). The xylose is bound to GlcA-C-3 seen by .sup.2,3J CH of Xyl-H-1 (δ 4.57) to GlcA-C-3 (δ 86.6), and in vice versa direction from Glc-H-3 to Xyl-C-1. The inter-glycosidic connectivities were also partly seen in the NOESY showing the through-space interaction over the glycosidic bonds. All structure relevant .sup.2,3J CH long-range signals in the sugar units are presented in FIG. 7.

(54) Transfection Efficiencies of Gypsophila elegans M. Bieb. Saponins

(55) The transfection efficiency from crude extracts, via several purification steps through to the isolated saponin GE1741, was determined. All tested saponins increased the transfection efficiency significantly compared to a solely application of PD nanoplexes. The seed and root extract yielded comparable transfection efficiencies (˜40%). However a lower concentration was used for the root extract (5 μg/mL). P2 implied a significantly higher transfection enhancing potential (72%) compared to P3 (29%) and P4 (27%). GE1741, which was isolated from P2, achieved a remarkable efficiency with 60% transfected cells.

(56) These results are shown in FIG. 6 depicting the transfection efficiency of nanoplexes±triterpene saponins in Neuro2a cells. Biological activities of different purification steps were investigated until the final isolation of GE1741. P2, which was obtained after HPLC purification of the raw extract, achieved the highest efficiency, even exceeding the isolated GE1741. An asterisk (*) marks results being significant to the transfection method without GE1741 co-application as determined by a t-test with p<0.05 and n≥3.

(57) Delivery of Nucleic Acids

(58) FIG. 8 compares the enhancement of the transfection efficiency by GE1741 (“novel saponin”) with the enhancement of the transfection efficiency by a prior art saponin (SO1861) as well as the current “gold standard” Lipofectamin.

(59) As negative control, Neuro2a cells have been incubated without an addition of particles comprising a nucleic acid.

(60) Upon adding particles comprising a peptide bound to mRNA (PmRNA), a transfection efficiency in the single-digit range could be observed. After adding a saponin as enhancer to the respective cell suspension, the transfection efficiency could be increased to approximately 15% when using SO1861, but to even 35% when using GE1741.

(61) When incubating Neuro2a cells with particles comprising a peptide bound to DNA (PD), once again a transfection efficiency in the single-digit range could be observed. When SO1861 was added to the according solution, the transfection efficiency was increased to approximately 35%. If GE1741 was used instead of SO1861, a transfection efficiency of almost 60% could be observed.

(62) When incubating Neuro2a cells with nanoparticles comprising a peptide bound to minicircle DNA (PM), once again a transfection efficiency in the single-digit area could be observed. After adding SO1861 to the according cell suspension, the transfection efficiency could be increased to more than 40%. If, on the other hand, GE1741 was added to the cell suspension, the transfection efficiency was increased to almost 80%.

(63) A transfection efficiency of almost 80% when using peptide minicircle DNA nanoparticles is even better than a transfection efficiency of approximately 70% observed for the “gold standard” Lipofectamin. Thus, GE1741 is not only better than previously isolated saponins, but shows an even higher enhancement of the transfection efficiency than the gold standard being currently on the market.

(64) The asterisks in FIG. 8 indicate a significant difference to the transfection method without saponin co-application by applying a t-test (p<0.05, n≥3).

(65) Combination of GE1741 with Commercial Transfection Methods

(66) HCT-116 cells were transfected with commercial transfection reagents in order to examine the impact of GE1741 on the transfection efficiency. The results are shown in FIG. 9. The addition of GE1741 led to an increase in transfection efficiency of each transfection method. While the transfection efficiency with TransIT was only slightly increased by 3 percentage points and that of Xtreme Gene by 8 percentage points, GE1741 boosted the transfection efficiency of Genecellin by 30 percentage points from 12% (without GE1741) to 42% (with GE1741).

(67) Cell Toxicity and Permeability of GE1741

(68) The impedance, which shows a direct correlation with cell viability, was investigated in order to determine the toxicity and the permeabilizing properties of GE1741.

(69) GE1741 does not exhibit cell-toxic effects when used in a concentration that is fully sufficient for enhancing the transfection efficiency. To test the toxic effects of GE1741, this saponin was applied in different concentrations to Neuro2a cells and incubated over a period of 50 hours. At concentrations of 2.5 μg/mL and 6.25 μg/mL, a normalized cell index is (within the error of measurement) identical to the normalized cell index observed for the negative control that has been subjected to a transfection process without an addition of GE1741. This is indicated in FIG. 10 by other three overlaid curves 1. If the GE1741 concentration is increased to 12.5 μg/mL, a slightly decreased normalized cell index could be observed (curve 2). The normalized cell index slightly further decreased upon increasing the GE1741 concentration to 25 μg/mL (curve 3).

(70) A significant effect on the cell viability could only be observed at a GE1741 concentration of 50 μg/mL (curve 4). In this case, the normalized cell index markedly decreased approximately 12 hours after transfection.

(71) Since an enhancement of the transfection efficiency could already be observed at a GE1741 concentration of 2 μg/mL (this concentration was used in the experiments the results of which are shown, e. g, in FIGS. 5 and 8), it can be concluded that GE1741 exhibits no cytotoxic effects in the relevant concentration range.

(72) In another toxicity experiment, DNA nanoparticles (curve 9), DNA nanoparticles with 2 μg/mL GE1741 (GE1741; curve 10) or 2 μg/mL GE1741 alone (curve 11) was added to Neuro2a cells. Thereby, the DNA nanoparticles comprised DNA encoding for the Green Fluorescent Protein (GFP). Buffer was used as negative control (curve 12). After adding the respective substances to the cells at the time point marked with an arrow and the word “transfection”, no significant difference in the cell viability could be observed. Thus, no toxic effects were revealed for any of the tested substances.

(73) Transfection Efficiency

(74) The transfection efficiency of GE1741 was also examined by transfecting cells with DNA encoding saporin. Saporin is a cytotoxic protein. If the transfection is successful, the transfected cells will die. The corresponding results are shown in FIG. 12. The negative control (curve 5) refers to Neuro2a cells incubated without any peptide DNA nanoparticles. Curve 6 refers to a transfection experiment of Neuro2a cells with nanoparticles comprising DNA encoding for saporin. The normalized cell index for this setup approximately corresponds to the normalized cell index of the negative control, indicating a very poor transfection efficiency. A likewise high normalized cell index could be observed if peptide DNA nanoparticles comprising DNA encoding for the green fluorescent protein were added to the cells together with 2 μg/mL GE1741. As shown in the previous Figures, such combination of peptide DNA nanoparticles and GE1741 serves for a highly efficient transfection. Since GFP is non-toxic for Neuro2a cells, the normalized cell index of this experimental setup (curve 7) roughly corresponds to the normalized cell index of the negative control (curve 5).

(75) If, however, peptide DNA nanoparticles comprising DNA encoding for saporin are added together with 2 μg/mL GE1741 to Neuro2a cells, no further cell growth could be observed at approximately 15 hours after transfection (curve 8). Rather, the normalized cell index remains on a value below 3 that is considerable lower than in case of the negative control. This clearly shows the highly efficient transfection of the Neuro2a cells with DNA encoding for the cytotoxic protein saporin.

(76) Summarizing, new transfection enhancing compounds could be identified by using a bioassay-guided isolation strategy on the mostly unexplored plant Gypsophila elegans M. Bieb. With this strategy, an alternating approach of purification and activity tests was conducted to find the highest active compound for advanced biochemical experiments. The HPLC chromatogram of the raw root extract revealed a number of peaks at a notable retention time, which were directly tested on a peptide-based transfection method. The distinct transfection enhancing effect of P2, compared to lower levels of P3 and P4, led to the assumption of a highly active compound in the specific fraction. The further purification and isolation via LC/MS and the identification of GE1741 via MS/MS provided a compound, the strong activity of which was confirmed. The lower efficiency of GE1741 towards P2 could be explained by the presence of a saponin mixture in P2 comprising, besides GE1741, a saponin the terminal xylose residue of which is replaced by arabinose.

(77) The universal applicability of GE1741 in transfection experiments was investigated. A delivery of different nucleic acids, incorporated into nanoplexes, could be considerably increased. mRNA transfections were less effective than DNA transfections due to a faster degradation by cellular enzymes. However, a missing transcription step allowed the highest efficiency even after 24 h incubation time and therefore representing a rapid transfection method. Minicircle transfection yielded the highest efficiency, presumably caused by the smaller size of minicircle-compared to plasmid DNA.

(78) The impact of GE1741 on commercial transfection methods was evaluated in order to underline the universal and simple applicability. An increase of transfection efficiency could be achieved for each method. A particular significant increase could be achieved for Genecellin transfections.

(79) A common problem and limitation for the use of transfection reagents in specific in vitro and in vivo experiments are the toxic effects, which often come along with the high efficiency. By conducting impedance measurements, representing the cell viability, GE1741-mediated transfections were analyzed in terms of toxicity. The results revealed that cells that are transfected with a combination of nanoplexes and GE1741 were as viable as the untreated negative control. The application of different GE1741 concentrations showed a toxicity starting above a concentration of 6.25 μg/mL and first lytic effects at a concentration of 50 μg/mL, indicated by a considerable impedance drop. As transfections efficiency measurements were conducted in a concentration of 2 μg/mL, ultimately, a high activity with no toxicity can be confirmed for GE1741.

(80) GE1741 is a novel and valuable compound for the process of saponin-mediated transfection, the so-called sapofection.

LIST OF PUBLICATIONS CITED IN THE PRECEDING SECTIONS

(81) Findeisen, M.; Brand, T.; Berger, S., A 1H-NMR thermometer suitable for cryoprobes. Magn Reson Chem 2007, 45 (2), 175-8. Piantini, U.; Sorensen, O. W.; Ernst, R. R. Multiple quantum filters for elucidating NMR coupling networks. Journal of the American Chemical Society 1982, 104, 6800-6801. Bax, A.; Davis, D. G. Practical aspects of two-dimensional transverse NOE spectroscopy. Journal of Magnetic Resonance (1969) 1985, 63, 207-213. Braunschweiler, L.; Ernst, R. R. Coherence transfer by isotropic mixing: Application to proton correlation spectroscopy. Journal of Magnetic Resonance (1969) 1983, 53, 521-528. Jeener, J.; Meier, B. H.; Bachmann, P.; Ernst, R. R. Investigation of exchange processes by two-dimensional NMR spectroscopy. The Journal of Chemical Physics 1979, 71, 4546-4553 Bodenhausen, G.; Ruben, D. J. Natural abundance nitrogen-15 NMR by enhanced heteronuclear spectroscopy. Chemical Physics Letters 1980, 69, 185-189. Kessler, H.; Schmieder, P.; Kock, M.; Reggelin, M. Selection of methyl resonances in proton-detected heteronuclear shift correlation, the HQQC experiment. Journal of Magnetic Resonance (1969) 1991, 91, 375-379. Kessler, H.; Schmieder, P.; Kurz, M. Implementation of the DEPT sequence in inverse shift correlation; the DEPT-HMQC. Journal of Magnetic Resonance (1969) 1989, 85, 400-405. Bax, A.; Subramanian, S. Sensitivity-enhanced two-dimensional heteronuclear shift correlation NMR spectroscopy. Journal of Magnetic Resonance (1969) 1986, 67, 565-569. Cicero, D. O.; Barbato, G.; Bazzo, R. Sensitivity enhancement of a two-dimensional experiment for the measurement of heteronuclear long-range coupling constants, by a new scheme of coherence selection by gradients. Journal of magnetic resonance (San Diego, Calif.: 1997) 2001, 148, 209-213. Bax, A.; Summers, M. F. Proton and carbon-13 assignments from sensitivity-enhanced detection of heteronuclear multiple-bond connectivity by 2D multiple quantum NMR. Journal of the American Chemical Society 1986, 108, 2093-2094.