Mitochondrion 5 (2005) 352–358 www.elsevier.com/locate/mito
Mitochondrial leader sequence-plasmid DNA conjugates delivered into mammalian cells by DQAsomes co-localize with mitochondria Gerard G.M. D’Souza, Sarathi V. Boddapati, Volkmar Weissig* Department of Pharmaceutical Sciences, School of Pharmacy, Bouve´ College of Health Sciences, Northeastern University, 360 Huntington Avenue, 211 Mugar Building, Boston, MA 02115, USA Received 6 April 2005; received in revised form 7 July 2005; accepted 12 July 2005
Abstract In the last decade the increase in therapeutic strategies aimed at mitochondrial targets has resulted in the need for novel delivery systems for the selective delivery of drugs and DNA into mitochondria. In this study, we have continued our efforts towards the development of the first mitochondriotropic drug and DNA delivery system (DQAsomes). Prepared from derivatives of the self-assembling mitochondriotropic bola-amphiphile dequalinium chloride, these vesicles bind and transport DNA to mitochondria in living mammalian cells where upon they have been shown to release the DNA on contact with mitochondrial membranes. We present data to demonstrate that oligonucleotides as well as plasmid DNA conjugated to a mitochondrial leader sequence (MLS) co-localize with mitochondria when delivered into mammalian cells by DQAsomes. In contrast to a commercially available DNA delivery vector, our vesicles appear to have a pronounced specificity for mitochondria. Further, the data strongly suggest that linear conjugates might be better suited to delivery into mitochondria and that in the absence of a mitochondria specific vector, the presence of a MLS-peptide conjugated to the DNA is alone not sufficient to direct the accumulation of DNA at mitochondria. q 2005 Elsevier B.V. and Mitochondria Research Society. All rights reserved. Keywords: Mitochondria; Drug delivery; Mitochondrial gene therapy; Mitochondrial drug targeting; Dequalinium; Oligonucleotides; Plasmid DNA; Mitochondrial leader sequence peptides; DNA peptide conjugates
1. Introduction Abbreviations DOPE, dioleoyl phosphatidyl ethanolamine; DOTAP, dioleoyl oxytrimethylammonio propane; DMRIE, 1-2 dimyristoyloxypropyl 3- dimethylhydroxyethyl ammonium bromide; TCEP, Tris (2 carboxy ethyl) phosphine; mOTC, mouse ornithine transcarbamylase; hMDH, human malate dehydrogenase; COX, cytochrome c oxidase. * Corresponding author. Tel.: C1 617 373 3212; fax: C1 617 373 8886. E-mail address:
[email protected] (V. Weissig).
It has become progressively more evident that mitochondrial dysfunction contributes to a variety of human disorders such as neurodegenerative and neuromuscular diseases, obesity, diabetes, ischemiareperfusion injury and cancer. Increased efforts directed towards the study of mitochondria as targets for pharmacological intervention, have made
1567-7249/$ - see front matter q 2005 Elsevier B.V. and Mitochondria Research Society. All rights reserved. doi:10.1016/j.mito.2005.07.001
G.G.M. D’Souza et al. / Mitochondrion 5 (2005) 352–358
mitochondrial research one of the fastest growing disciplines in biomedicine (Larson and Luft, 1999; Murphy and Smith, 2000; Schon and DiMauro, 2003; Singh, 2000). However, current therapeutic strategies that have focused on mitochondrial targets have frequently been limited by the unpredictability of the mitochondrial concentration of the active moiety and the often high or even toxic concentrations needed to reach a mitochondrial target (D’Souza and Weissig, 2004; Morin et al., 2001). Consequently, there is a growing need for strategies to render drugs mitochondriotropic (to facilitate their selective accumulation at or inside mitochondria) or alternatively, novel delivery systems for the selective delivery of drugs to or into mitochondria. Our efforts to develop the first mitochondriotropic delivery system are based on the use of mitochondriotropic quinolinium derivatives like dequalinium (D’Souza et al., 2003; Lasch et al., 1999; Weissig et al., 2001a,b, 1998, 2000). Dequalinium and its derivatives are self-assembling amphiphiles that resemble ‘bola’-form electrolytes, i.e. they are symmetric molecules with two charge centers separated by a hydrophobic chain. Such ‘bola’-form like amphiphiles are able to form liposome-like cationic vesicles, which we termed ‘DQAsomes’ (Weissig et al., 1998). Our strategy for the delivery of DNA into mitochondria requires that DQAsomes deliver an MLS-DNA conjugate intact to the outer membranes of mitochondria, where upon the MLS-peptide should mediate the mitochondrial uptake of the DNA. In previous studies we have shown that DQAsomes bind DNA and selectively release DNA at isolated mitochondria (D’Souza et al., 2003; Weissig et al., 2001a,b, 1998). In mammalian cell culture DQAsomes selectively release plasmid DNA at the site of mitochondria in living cells (D’Souza et al., 2003). However, all these studies have been performed with non-MLS conjugated DNA and so mitochondrial internalization of the DNA was neither expected nor observed. This paper describes our first study of the sub-cellular disposition of DNA conjugated to an MLS peptide and delivered intracellularly by our mitochondriotropic vesicles. We have employed confocal fluorescence microscopic techniques to determine if we have been successful in achieving the import of MLS conjugated oligonucleotides and plasmid DNA to mitochondria of live mammalian cells.
353
2. Materials and methods 2.1. Materials Dequalinium chloride was purchased from Sigma Corp. (St Louis, MO, USA). The cyclohexylderivative of dequalinium was synthesized in C.R Ganellin’s laboratory (Christopher Ingold Laboratories, Dept. of Chemistry, University College, London) as described earlier in (Weissig et al., 2001a,b). Mitotrackerw Red CMXRos was purchased from Molecular Probes (Eugene, OR, USA). LabelITe nucleic acid fluorescein labeling kit was purchased from Mirus Corp. (Madison, WI, USA). Plasmid DNA (pGL3) was purchased from Promega (WI, USA). Lipofectine was purchased from Invitrogen (Carlsbad, CA, USA). The MLS-oligonucleotide (mOTC-oligo) consisted of a 5 0 fluorescein labeled oligonucleotide (5 0 -CTCCCTCACCATTGG0 CAGCCTA-3 ) coupled at the 3 0 end to the N-terminal end of the mouse ornithine transcarbamylase (mOTC) mitochondrial leader sequence (H2N– FNLRILLNNAAFRNGHNFMVRNFRCGQPLQN– COOH) was purchased from Eurogentec (Belgium). 2.2. Labeling of plasmid DNA The pDNA used in these studies was pre labeled with fluorescein using the commercially available Label-ITe nucleic acid labeling kit (Mirus Corp.). This kit uses proprietary chemistry to covalently label double stranded DNA at purine residues. Briefly, 20 mg of pDNA was incubated with 20 ml of LabelITe reagent for 1 h at 37 8C. The DNA was then isolated by ethanol precipitation and spectrophotometricaly assayed for DNA content at 260 nm. Fluorescence measurements at an excitation wavelength of 492 nm and an emission wavelength of 518 nm were used to confirm and quantify labeling. 2.3. Conjugation of MLS peptide to plasmid DNA Maleimide labeled GeneGripe pDNA was purchased from Gene Therapy Systems, San Diego, CA. The leader sequence of the human mitochondrial matrix protein malate dehydrogenase (hMDH) (H2N– MLSALARPAG AALRRSFSTS AQNNAKVAVL GASC–COOH) with a C-terminal cysteine, was
354
G.G.M. D’Souza et al. / Mitochondrion 5 (2005) 352–358
synthesized by Invitrogen Corporation, Carlsbad, CA. A 1 mg amount of lyophilized peptide was dissolved in 1 ml of 0.1 mM Tris (2-carboxyethyl) phosphine hydrochloride (TCEP) in 100 mM sodium phosphate buffer, pH 7.0. A 0.1 ml aliquot was allowed to stand at room temperature for 2 h to allow reduction of the sulphydryl group. After reduction TCEP was not removed from the peptide solution. The reduced ligand was added in 20 molar excess (w150 pmol) to 20 mg (w4.5 pmol) of maleimide-labeled plasmid and incubated overnight at 4 oC to yield the conjugate (hMDH–pDNA). After the conjugation protocol, restriction enzymes BamHI and BstXI were used to excise a 472 bp region of the plasmid carrying the 100 bp PNA binding GeneGrip site and the corresponding band electrophoresed on a 1.2% pre-cast agarose E-gele (Invitrogen Corporation, Carlsbad, CA) was monitored for retardation to verify the success of the conjugation reaction.
2.6. Preparation of cationic vector DNA complexes
2.4. Restriction digest of hMDH–pDNA conjugates
2.7. Cell culture
BamHI and the restriction buffer were obtained from Roche Biochemicals, Indianapolis, IN. MLSDNA (0.5–1 mg DNA) was mixed with w0.5 ml of enzyme in appropriately diluted buffer solution and incubated in a water bath for 1 h at 37 8C.
BT20 cells were grown in eagles modified essential medium (EMEM), on 22 mm circular cover slips in six well cell culture plates till they were approximately 60–80% confluent.
DQAsome/DNA complexes (DQAplexes) used for these studies were prepared after the DNA binding capacity of each preparation was determined in order to enable us to choose the correct vector: DNA ratio, i.e. to exclude the presence of any excess of free DNA. The complexes were then generated by simply mixing DNA (oligonucleotides or pDNA) with the requisite amount of DQAsomes. Lipofectine/DNA complexes (lipoplexes) were prepared according to the manufacturer’s instructions using either oligonucleotides or pDNA. The requisite amounts of DNA and liposomal carrier were separately diluted in serum free media allowed to stand for 10 min at room temperature and then mixed together and allowed to stand at room temperature for a further 5 min before addition to the cell culture well.
2.8. Cellular uptake studies 2.5. Preparation and characterization of DQAsomes DQAsomes were prepared as previously described (Weissig et al., 1998, 2001a,b). A weighed quantity of quinolinium compound (to yield final concentration of 10 mM) was dissolved in methanol in a round-bottom flask and the methanol evaporated on a rotary evaporator. A 5-ml volume of HEPES buffer, pH 7.4, was added to the solvent-free dequalinium film and the suspension sonicated with a Model 100 sonic dismembrator (Fisher Scientific) for 45 min at 10– 12 W until a uniform opalescent solution was obtained. The preparations were then centrifuged for 10 min at 3000 rpm to remove metal particles and the particle size of the resulting preparations was then measured by light scattering analysis in a Coulter N4 particle size analyzer. (Beckmann Coulter Inc., Fullerton.CA).
All incubations were performed in serum free medium to eliminate possible precipitation events associated with the interaction of the vesicular carriers with serum proteins. BT20 cells were incubated for 10 h with 1 mg DNA complexed with DQAsomes or liposomal vector. For control, cells were exposed to naked DNA and empty vesicles. After removing the medium containing non-internalized material, cells were thoroughly washed with Cellscrube buffer (Gene Therapy Systems, San Diego, CA) to remove surface associated complex and incubated with the mitochondria specific dye Mitotrackerw Red CMXRos for 5 min to stain mitochondria. The cells were then washed in sterile phosphate buffered saline (PBS) and mounted on slides in Fluoromount G medium for analysis by confocal fluorescence microscopy.
G.G.M. D’Souza et al. / Mitochondrion 5 (2005) 352–358
2.9. Confocal fluorescence microscopy Confocal images were taken with a Zeiss Meta 510 LSM and using Zeiss AIM confocal imaging software. A pinhole size of 1 airy unit was used which resulted in optical slices of around 0.7 mm. Black levels were adjusted using both unstained control slides and slides stained only with Mitotrackerw Red CMXRos. All other slides were then imaged at the same settings. All images shown are taken from the middle of a stack that runs from the top of the cell to the bottom and are thus representative of the inside of the cell. After capture of the individual channels the green channel was analysed for co-localization with the red channel using the built in tool in the Zeiss AIM confocal imaging software. Composite images generated by this operation are presented along with the individual channels used to generate them.
3. Results and discussion 3.1. DQAsome mediated cellular uptake and sub-cellular distribution of mOTC-oligo conjugate As a first step towards evaluating the effect of DQAsomes on the sub-cellular disposition of an
355
MLS-DNA conjugate we studied the intracellular delivery of a small MLS-oligonucleotide conjugate. The conjugate was obtained commercially and was selected for our initial experiments because it had already been shown to associate with mitochondria when delivered intracellularly in what was the first of only two reports of DNA delivery to mitochondria in living mammalian cells (Geromel et al., 2001). BT20 cells were exposed for 10 h to mOTC-oligo complexed with DQAsomes followed by confocal fluorescence microscopic analysis. The characteristic red staining pattern in Fig. 1(a) demonstrates that the exposure to the complex did not adversely affect mitochondrial viability. The green intracellular fluorescence in Fig. 1(b) shows that complexation of the mOTC-oligo with DQAsomes resulted in the cellular uptake of the conjugate. Most importantly when the green and red channels were overlaid an appreciable part of the green intracellular fluorescence colocalized with stained mitochondria as indicated by the white arrows (Fig. 1(c)). Regions of co-localization were pseudo-colored in white for easy observation (Fig. 1(c)). The observed co-localization indicates that the mOTC-oligo conjugates delivered into mammalian cells by complexation with DQAsomes are in definite association with the
Fig. 1. Representative confocal fluorescence micrographs of BT20 cells stained with Mitotrackerw Red CMXRos (red) after exposure for 10 h to fluorescein labeled oligonucleotide-MLS peptide (green); (a)–(c): complexed with DQAsomes, (d)–(f): complexed with Lipofectine, (a,d): red channel, (b,e): green channel, (c,f): overlay of red and green channels with white indicating co-localization of red and green fluorescence.
356
G.G.M. D’Souza et al. / Mitochondrion 5 (2005) 352–358
mitochondria and suggests that the conjugates have been imported into mitochondria. While DQAsomes appear to have delivered the vast majority of mOTC-oligo conjugates into close proximity to mitochondria, almost all of the intracellular mOTC-oligo fluorescence in cells exposed to mOTC-oligo complexed with Lipofectine appears inside the nucleus and virtually none is found at or near mitochondria (Fig. 1(f)). Such exclusive nuclear localization of an oligonucleotide bearing a mitochondrial leader sequence peptide is remarkable even though a nuclear-targeted carrier delivered the conjugate into the cell. This finding suggests first that complexing the conjugate with Lipofectine seems to make the mitochondrial leader sequence peptide ineffective as a mitochondrial targeting moiety and second that the dissociation of the oligonucleotides from their carrier probably takes place only upon approaching the nucleus, thus enabling the small oligonucleotide to enter the nucleus. Any earlier dissociation of the DNA from its cationic carrier, i.e. during or shortly after endosomal escape of the lipoplex, should have led to at least some random interaction events of the MLS peptide with the mitochondrial protein import machinery and a subsequent accumulation of the oligonucleotide inside mitochondria (already demonstrated for peptide nucleic acids conjugated to the leader sequence peptide of the nuclear-encoded human COX subunit VIII (Chinnery et al., 1999)). Our data with Lipofectine seem to be in agreement with data reported recently by Flierl et al. (2003). All of their attempts to utilize three different commercially available cationic liposome formulations (DOTAP/DOPE; Fugenee 6 and Escorte) have failed to deliver PNA-MLS peptide/oligonucleotide (PPO) complexes to mitochondria of living myoblasts. Instead they report that the PPO complexes remained associated with the outside of the cell, indicating a complete failure of these three liposome formulations to mediate any cell internalization. From the data reported by Flierl et al. and by us, it appears that cationic liposomes might not generally be suited for the delivery of oligonucleotides or DNA to mammalian mitochondria. However, to thoroughly evaluate the potential of cationic liposomes for the transport of DNA and other molecules to mitochondria, more data still needs to be gathered.
3.2. DQAsome mediated cellular uptake and subcellular distribution of hMDH–pDNA conjugate As the final step in our preliminary demonstration that DQAsomes can mediate the delivery of DNA to mitochondria of mammalian cells in culture, we studied the uptake of hMDH–pDNA conjugates in cultured BT20 cells. To conjugate hMDH MLS to the plasmid DNA we used a novel molecular biological approach based on the use of peptide nucleic acid (PNA) linkers. This technology, known as GeneGripe, utilizes a short, linear, horseshoe-shaped peptide nucleic acid (PNA) molecule called a PNA clamp (US Patent 6,165,720). PNA clamps hybridize with duplex DNA with very high affinity in a sequence specific manner and can be conjugated to a variety of molecules for gene tracking and targeting (Zelphati et al., 1999, 2000). GeneGripe labeled plasmids carry 8–10 functional groups linked to a distinct 100 bp region of the 5 kb plasmid via hairpin PNA clamps. Maleimide was selected as the functional group in order to facilitate easy coupling with the hMDH peptide synthesized with a C-terminal cysteine. Theoretically, a conjugation carried out in such a manner would result in the N-terminal peptide attached to the DNA in the correct orientation as well as the location of the peptide being restricted to the 100 bp binding site on the plasmid. The GeneGripe method thus represented the best compromise in terms of ease of use and the required functionality. Fig. 2 shows images representative of the confocal fluorescence micrographs obtained in our study. The green and red channels used to generate the overlaid images in the far right column are shown separately in the preceding columns to facilitate a careful comparison of the observed staining patterns. The image set in the top row shows the results obtained with DQAsomes. The image from the green channel shows that cellular internalization of the hMDH–pDNA conjugate was indeed successful (Fig. 2(b)). Once again, the characteristic mitochondrial staining pattern seen with the red channel is a strong indicator of mitochondrial viability in the imaged cells. From the composite image obtained by overlaying the green and red channels it can be seen that a sizeable fraction of the intracellular green fluorescence co-localized with the red mitochondrial
G.G.M. D’Souza et al. / Mitochondrion 5 (2005) 352–358
357
Fig. 2. Representative confocal fluorescence micrographs of BT20 cells stained with Mitotrackerw Red CMXRos (red) after exposure for 10 h to fluorescein labeled MLS-pDNA conjugate (green) complexed with cyclohexyl-DQAsomes; (a–c): circular pDNA conjugate, (d–f): linearized pDNA conjugate. (a,d) green channel, (b,e): red channel, (c,f): overlay of red and green channels with white indicating co-localization of red and green fluorescence.
fluorescence (depicted as white areas in Fig. 2(c)). These observations indicate that in addition to mediating the cellular uptake of the hMDH–pDNA conjugate, the use of DQAsomes resulted in a definite association of an appreciable fraction of the internalized conjugate with mitochondria. Interestingly, we observed that the presence of the MLS peptide did not appear to influence the amount of DNA internalized by any of the preparations (i.e. with the same vector and equivalent amounts of DNA, the presence of an MLS peptide did not change the apparent level of intracellular DNA fluorescence). The presence of an MLS peptide only influenced the final co-localization of the DNA with mitochondria. These observations seem to suggest that the mitochondriotropic vector is primarily responsible for the mitochondria-specific targeting while the MLS peptide is only necessary for the final import of the conjugate into mitochondria. In an effort to evaluate the effect of conjugate architecture on mitochondrial uptake, BamHI was used to linearize the circular pDNA in the conjugate. BamHI cuts the plasmid at a single site close to the PNA binding region to generate a linear conjugate
with the MLS peptide at one end. Cells were then exposed to equal amounts of intact and linearized conjugate complexed with the vector and analyzed by confocal microscopy for observable differences in uptake and disposition of the conjugate. Fig. 2 (bottom row) shows confocal fluorescence micrographs of cells exposed to linearized hMDH–pDNA conjugate complexed with DQAsomes. In the case of the linearized conjugate higher levels of co-localization than with the same amount of non-linearized conjugate were seen. Our observations indicate that linearized hMDH–pDNA conjugate complexed with DQAsomes resulted in higher levels of mitochondrial association than circular DNA thereby suggesting the possibility that a linear conjugate might be better suited to delivery and import into the mitochondria. In conclusion, using confocal fluorescence microscopy we have shown that not only are DQAsomes capable of selectively mediating the association of an MLS-oligonucleotide with mitochondria in living cells, but have shown for the very first time that DQAsomes are capable of delivering a MLS-pDNA conjugate to the mitochondria of living cells and that the conjugate becomes associated with
358
G.G.M. D’Souza et al. / Mitochondrion 5 (2005) 352–358
mitochondria in a manner that suggests successful import into the organelle. Given the lack of an easily available and detectable mitochondrial reporter gene and the fact that the functional expression of a transgene inside the mitochondrial matrix has never been demonstrated so far, we believe that our results represent a significant step towards the delivery of therapeutic transgenes to mitochondria and collectively these data serve as the first ‘proof of principle’ in our proposed strategy for the use of DQAsomes to deliver genetic material to mitochondria. Experiments are now in progress to demonstrate the mitochondrial expression of a transgene delivered to mitochondria by these mitochondriotropic vesicles.
Acknowledgements 4.The authors extend their sincere thanks to Michelle Ocana and Mark Chafel, (Harvard Medical School, Center for Neurodegeneration and Repair, Boston MA) for assisting with the use of their confocal imaging facility, Prof. Vladimir P. Torchilin (Northeastern University, Dept. of Pharmaceutical Sciences, Boston, MA) for generous access to particle size and fluorescence measurement equipment and to Prof. Josephine Modica-Napolitano (Merrimack College, Dept. of Biology, Andover, MA) and Prof. Robert N. Lightowlers (Mitochondrial Research Group, University of Newcastle upon Tyne, UK) for helpful suggestions and discussion. This work was supported in part by a grant from the United Mitochondrial Disease Foundation, Pittsburg. PA, USA.
References Chinnery, P.F., Taylor, R.W., Diekert, K., Lill, R., Turnbull, D.M., Lightowlers, R.N., 1999. Peptide nucleic acid delivery to human mitochondria. Gene Ther. 6 (12), 1919–1928. D’Souza, G.G., Weissig, V., 2004. Approaches to mitochondrial gene therapy. Curr. Gene Ther. 4 (3), 317–328.
D’Souza, G.G., Rammohan, R., Cheng, S.M., Torchilin, V.P., Weissig, V., 2003. DQAsome-mediated delivery of plasmid DNA toward mitochondria in living cells. J. Control Release 92 (1–2), 189–197. Flierl, A., Jackson, C., Cottrell, B., Murdock, D., Seibel, P., Wallace, D.C., 2003. Targeted delivery of DNA to the mitochondrial compartment via import sequence-conjugated peptide nucleic acid. Mol. Ther. 7 (4), 550–557. Geromel, V., Cao, A., Briane, D., Vassy, J., Rotig, A., Rustin, P., Coudert, R., Rigaut, J.P., Munnich, A., Taillandier, E., 2001. Mitochondria transfection by oligonucleotides containing a signal peptide and vectorized by cationic liposomes. Antisense Nucleic Acid Drug Dev. 11 (3), 175–180. Larson, N.-G., Luft, R., 1999. Revolution in mitochondrial medicine. FEBS Lett. 455, 199–202. Lasch, J., Meye, A., Taubert, H., Koelsch, R., Mansa-ard, J., Weissig, V., 1999. Dequalinium vesicles form stable complexes with plasmid DNA which are protected from DNase attack. Biol. Chem. 380 (6), 647–652. Morin, D., Hauet, T., Spedding, M., Tillement, J., 2001. Mitochondria as target for antiischemic drugs. Adv. Drug Deliv. Rev. 49 (1–2), 151–174. Murphy, M.P., Smith, R.A., 2000. Drug delivery to mitochondria: the key to mitochondrial medicine. Adv. Drug Deliv. Rev. 41 (2), 235–250. Schon, E.A., DiMauro, S., 2003. Medicinal and genetic approaches to the treatment of mitochondrial disease. Curr. Med. Chem. 10 (23), 2523–2533. Singh, K.K., 2000. Mitochondrial me and the Mitochondrion journal. Mitochondrion 1 (1), 1–2. Weissig, V., Lasch, J., Erdos, G., Meyer, H.W., Rowe, T.C., Hughes, J., 1998. DQAsomes: a novel potential drug and gene delivery system made from Dequalinium. Pharm. Res. 15 (2), 334–337. Weissig, V., Lizano, C., Torchilin, V.P., 2000. Selective DNA release from DQAsome/DNA complexes at mitochondria-like membranes. Drug Deliv. 7 (1), 1–5. Weissig, V., D’Souza, G.G., Torchilin, V.P., 2001. DQAsome/DNA complexes release DNA upon contact with isolated mouse liver mitochondria. J. Control Release 75 (3), 401–408. Weissig, V., Lizano, C., Ganellin, C.R., Torchilin, V.P., 2001. DNA binding cationic bolasomes with delocalized charge center. A structure-activity relationship study. S.T.P. Pharma. Sci. 11 (1), 91–96. Zelphati, O., Liang, X., Hobart, P., Felgner, P.L., 1999. Gene chemistry: functionally and conformationally intact fluorescent plasmid DNA. Hum. Gene Ther. 10 (1), 15–24. Zelphati, O., Liang, X., Nguyen, C., Barlow, S., Sheng, S., Shao, Z., Felgner, P.L., 2000. PNA-dependent gene chemistry: stable coupling of peptides and oligonucleotides to plasmid DNA. Biotechniques 28 (2), 304–310 (see also 312–314, 316).
