A

fTcpSfsc(Dto©

K

1

fcfia

w

0d

Journal of Biotechnology 130 (2007) 85–94

Novel cationic vesicle platform derived from vernonia oil forefficient delivery of DNA through plant cuticle membranes

Zeev Wiesman a,∗, Naomi Ben Dom a, Efrat Sharvit a, Sarina Grinberg b,Charles Linder c, Eli Heldman d, Michele Zaccai e

a Phyto-Lipid Biotechnology Laboratory, Department of Biotechnology Engineering, Faculty of Engineering,Ben-Gurion University of the Negev, Beer-Sheva 84105, Israel

b Department of Chemistry, Ben Gurion University of the Negev, Beer-Sheva 84105, Israelc The Zuckerberg Institute for Water Research and Department of Biotechnology Engineering,

Ben-Gurion University of the Negev, Beer-Sheva 84105, Israeld Department of Clinical Biochemistry, The Faculty of Health Sciences, Ben-Gurion University of the Negev, Beer-Sheva 84105, Israele Department of Life Sciences and Department of Biotechnology Engineering, Ben-Gurion University of the Negev, Beer-Sheva, Israel

Received 11 August 2006; received in revised form 25 December 2006; accepted 12 January 2007

bstract

Novel cationic amphiphilic compounds were prepared from vernonia oil, a natural epoxidized triglyceride, and studied with respect to vesicleormation, encapsulation of biomaterials such as DNA, and their physical stability and transport through isolated plant cuticle membranes.he amphiphiles studied were a single-headed compound III (a quaternary ammonium head group with two alkyl chains) and a triple-headedompound IV, which is essentially three molecules of compound III bound together through a glycerol moiety. Vesicles of the two amphiphiles,repared by sonication in water and solutions of uranyl acetate or the herbicide 2,4-D (2,4-dichloropenoxy acetic acid), were examined by TEM,EM, AFM, and confocal laser systems and had a spherical shape which encapsulated the solutes with diameters between 40 and 110 nm. Vesiclesrom amphiphile IV could be made large enough to encapsulate a condensed 5.2 kb DNA plasmid (pJD328). Vesicles of amphiphile IV were alsohown to pass intact across isolated plant cuticle membranes and the rate of delivery of encapsulated radio-labeled 2,4-D through isolated plantuticle membranes obtained with these vesicles was clearly greater in comparison to liposomes prepared from dipalmitopyl phosphatidylcholine

DPPC) and the control, nonencapsulated 2,4-D. Vesicles from amphiphiles III and IV were found to be more stable than those of liposomes fromPPC. The data indicate the potential of vesicles prepared from the novel amphiphile IV to be a relatively efficient nano-scale delivery system

o transport DNA and other bioactive agents through plant biological barriers. This scientific approach may open the way for further developmentf efficient in vivo plant transformation systems.

2007 Elsevier B.V. All rights reserved.

acid;

tevum

eywords: Cuticle membrane; Dipalmitopyl phosphatidylcholine; Epoxy fatty

. Introduction

There is growing interest in the preparation of lipid vesiclesor encapsulating biological substances such as drugs, pesti-ides, organic and inorganic materials, lipids, genes, and proteins

or pharmaceutical, medical, cosmetic, agricultural, and chem-cal applications, e.g., gene transfection therapy, pesticides,roma, mineral or vitamin encapsulation (Ollivon et al., 2000).

Controlled delivery of active agents, as well as transfectionith DNA is, in general, the limiting factor in the development of

∗ Corresponding author. Tel.: +972 8 647 7184; fax: +972 8 647 7184.E-mail address: wiesman@bgu.ac.il (Z. Wiesman).

lttfmTpw

168-1656/$ – see front matter © 2007 Elsevier B.V. All rights reserved.oi:10.1016/j.jbiotec.2007.01.040

Nano-vesicles; Uranyl acetate; Vernonia oil

he plant biotechnology industry. Methods of nucleic acid deliv-ry can be classified into two main groups: the indirect method,ia Agrobacterium and the direct methods that are specificallysed in plant species where Agrobacterium-mediated transfor-ation is not possible. Direct methods include, but are not

imited to, microprojectile bombardment, electroporation, elec-rophoresis, and carbide fiber-mediated and liposome-mediatedransformation (Rakoczy-Trojanowska, 2002). Most plant trans-ormations are Agrobacterium-mediated, while among direct

ethods, microprojectile bombardment is most widely used.hese direct methods habitually require the use of proto-lasts, which offer many possibilities but are also associatedith major disadvantages compared with organized tissues,

8 f Biote

ssp(

vtentswlmhts(t2lddla1

taao

es

Sm

qvsttaulAwwc

aa2lrs

2

2

VllfD

6 Z. Wiesman et al. / Journal o

uch as laborious isolation procedures, which may inducetress and recalcitrance to systems in culture, problems inlant regeneration, and low transient expression of transgenesRakoczy-Trojanowska, 2002; Davey et al., 2005).

The ideal vehicle for these compounds should be stable initro and in vivo, be efficient as a delivery platform, protecthe biomaterials from degradation e.g., from nucleases (Daveyt al., 2005), deliver the bioactive agents to their target, beon-toxic and non-immunogenic, and be available in large quan-ities (Deshmukh and Huang, 1997). In addition, the vehiclehould be efficient in penetrating structured plant tissues andork in a wide range of species. Liposomes (spherical phospho-

ipid vesicles) can encapsulate water-soluble and lipid-solubleolecules in their aqueous and lipid phases, respectively, and

ave been used since the 1970s (Gregoriadis and Ryman, 1972)o deliver a variety of pharmacologically active agents to specificites in the body where pharmacological intervention is neededGregoriadis et al., 1998). In plants, the use of liposomes forransformations was reported for wheat (Patnaik and Khurana,001), but is in general extremely limited. Most commercialiposomes are a combination of phospholipids, cholesterol,iverse lipids, and sometimes various polymers. One of theisadvantages of liposomes from natural phospholipids is theirimited stability in biological environments, from which theyre rapidly cleared (Hans-Hening et al., 1980; Po-Shun et al.,981).

In this communication we describe the further characteriza-ion of novel nano-vesicles recently synthesized (Grinberg etl., 2002; Grinberg et al., 2005) from derivatives of vernonia oil,natural triglyceride of vernolic acid (cis-12,13-epoxy, cis-9-

ctadeconic acid).Amphiphile III (Scheme 1) is synthesized from the methyl

ster of vernolic acid as the starting material, to form aingle-headed amphiphile with two alkyl chains and a single

cheme 1. Chemical structure of the amphiphilic compounds III (based onethyl vernolate) and IV (based on trivernolin).

Art(lBDw

d

2

Imitsst((bbfli

chnology 130 (2007) 85–94

uaternary ammonium head group. Amphiphile III forms bilayeresicles and has been partially characterized with respect totructure–function characteristics of vesicle formation and DNAransfection (Grinberg et al., 2005). Amphiphile IV is essen-ially three molecules of amphiphile III bound together throughglycerol moiety. The differences in their intrinsic curvature wassed to explain why amphiphile III could easily form encapsu-ating vesicles while compound IV formed micelle-like vesicles.mphiphile IV was reported to form encapsulating vesicleshen cholesterol was used. Amphiphile IV alone, however,as shown in this earlier study to be more efficient as a DNA

omplexant and transfecting agent.In this paper, vesicles made from amphiphiles III and IV

re further characterized with respect to size, shape, self-ggregation/fusion, and encapsulation of DNA and the herbicide,4-D. Emphasis is placed on amphiphile IV vesicle encapsu-ation of condensed DNA, stability, and the transport rate ofadio-labeled encapsulated biomarker (2,4-D) through a modelystem of isolated plant cuticle membranes.

. Materials and methods

.1. Chemicals

Vernonia oil was purchased from Vertech Inc. (Falls Church,A, USA). Amphiphiles III and IV were synthesized fol-

owing the procedure reported in Grinberg et al. (2005).-�-Phosphatidylcholine from dried egg yolk was purchasedrom Sigma (St. Louis, MO, USA). For differential encapsulatedNA labeling study, FITC-ULSIS Alexa Fluor® 488 Nucleiccid Labeling Kit (Molecular Probes, Leiden, The Netherlands),

hodamine B (tetramethylrhodamine-5-2′-deoxy-uridine-5′-riphosphate) for microscopy was purchased from MerckDarmstadt, Germany), and fluroescein from BDH (Poole, Eng-and). Uranyl acetate (UA) for microscopy was purchased fromDH (Poole, England). 2,4-Dichlorophenoxy acetic acid (2,4-), 2,4-D[14C] (specific activity 19.2 mCi/mmol) and protamineere purchased from Sigma (St. Louis, MO).All the chemicals employed were analytical grade. Double-

istilled water was used in all experiments.

.2. Vesicle preparation

Two main methods were used to produce the vesicles–(1)njection method: 100–200 �l of the amphiphile dissolved inethanol (in a range of 0.008–0.016 g in 10 ml methanol) was

njected into 2–4 ml aqueous solution containing the compoundso be encapsulated. The mixture was sonicated in a 60 ◦C bathonicator (Delta Ultrasonic Cleaner, D80, 43 kHz, 80 W) forhort (1, 2, 5, 10 min) or long (60 min) periods depending onhe desired vesicle size and uniformity (Barenholtz et al., 1979).2) Reverse phase evaporation: Using the same amphiphiles as in1) above, methanol (250–1000 �l) was added to a 50 ml round

ottom flask, and the solvent removed under reduced pressurey a rotary evaporator (BUCHI) to form a thin oily film on theask bottom. An aqueous solution containing the compounds

ntended to be encapsulated (1–4 ml) was added to the flask

f Biote

aP

2

2ec

6rw1tsa

2

usisW2

llv

wtn1aIswis

LroaDAtTsf

2

W

Ct

2

Dbwa

2

mbfioltmifmaltcd

2

apwmWb(ler

tDlttRSwm

Z. Wiesman et al. / Journal o

nd the mixture was sonicated for 1–5 min at 60 ◦C (Szoka andapahadjopoulos, 1978; Gregoriadis et al., 1998).

.3. Encapsulation study

UA was encapsulated in the nano-vesicles by injection of00 �l each of amphiphiles III and IV, or DPPC dissolved inthanol (0.0437 g/1 ml) into 2 ml of 1% UA solution and soni-ated for 1 h at 60 ◦C.

The FITC-labeled DNA was diluted to 1 �g/ml in.68 × 10−6 M rhodamine solution. The encapsulation was car-ied out by adding this solution to a 50 ml round bottom flaskhich contained a thin oily film on the bottom (prepared fromml methanol containing 4.45 × 10−4 g amphiphile IV added to

he flask followed by removal of the solvent under reduced pres-ure by a rotary evaporator—reverse phase evaporation method)nd sonicated for 2 min at 60 ◦C.

.4. Microscope studies

TEM experiments were carried out on an EM 201C (Philips)sing a negative staining technique employing a saturated UAolution. The grid (300 mesh copper Formvar/carbon) wasmmersed in the vesicle solution for 1.5 min, stained in the UAolution for 1.5 min, and then dried at room temperature on

hatmann filter paper (Nardin et al., 2000; Ottaviani et al.,000).

SEM experiments were carried out on a Jeol 5600. Isolatedeaf astomatous adaxial cuticle of Citrus grandis (as describedater) were glued on SEM stabs and coated with gold under aacuum (Schonherr, 2000; Wiesman et al., 2002).

For AFM experiments, 10 �l of each of the vesicle solutionsere placed on a freshly cleaved mica surface. After 2 min

hey were rinsed with 1–2 ml water, dried with a stream ofitrogen, and further dried in a desiccator (Hansma and Laney,996; Thomson et al., 1996). AFM measurements were takent ambient conditions using a Thermomicroscopes CP Researchnstrument mounted on an active anti-vibration table. A 100 �mcanner was used. Microfabricated Si oxide C type Ultraleversith integrated pyramidal tips were used. The 512 × 512 pixel

mages were taken in non-contact mode with a nanomateric scalecan size at a scan rate of 1 Hz.

The confocal experiments were carried out on a ZeissMS510. The samples were prepared by adding 6.68 × 10−6 M

hodamine solution or 2 × 10−6 g/ml fluorescein solution to theily film on the flask bottom (reverse phase evaporation method)nd the mixture was sonicated in a bath sonicator for 1–5 min.NA plasmid (pJD328), 5 ng/�l was stained in FITC (ULSISlexa Fluor® 488 Nucleic Acid Labeling Kit) and then added to

he hydrated solution (as described above) prior to sonication.o examine DNA encapsulation, 10 �l of vesicle solution werepread on a microscope slide and left at room temperature untilully dry (Hansma and Laney, 1996).

.5. DNA preparation

The 5.2 kb plasmid pJD328 was purified from E. coli usingizard®plus minipreps DNA purification system (Promega

AatA

chnology 130 (2007) 85–94 87

orporation, Madison, WI, USA), according to the manufac-urer’s instructions.

.6. Stability study

To determine vesicle stability, each vesicle preparation inDW was left at room temperature, and the size, shape, mem-rane integrity, and fusion between vesicles and aggregationere examined once a week during 59 days by means of TEM,

s described above.

.7. Vesicle size determination

Light scattering measurements were made on an instru-ent assembled in-house. Samples were placed in thin-walled

orosilicate glass cuvettes (1 cm diameter) and placed in a vatlled with toluene as the index matching fluid. During the coursef the measurements the vat temperature was kept at 25 ◦C. Theight source was an argon ion laser (Lexel, λ = 514.5 A) and pho-ons scattered by the sample were collected by a photomultiplier

ounted on the goniometer arm at 90◦ to the direction of thencident radiation. The photoelectron count-time autocorrelationunction was calculated with a BI2030AT (Brookhaven Instru-ents Corporation, Holtsville, NY, USA) digital correlator and

nalyzed using the method of cumulants or the constrained regu-arization algorithm CONTIN (Thomson et al., 1996). Applyinghe Stokes–Einstein relationship to the translational diffusionoefficients provides an intensity weighted distribution of hydro-ynamic sizes (Schmidt et al., 1998).

.8. Study of isolated plant cuticle membrane transport

Full-grown leaves of the mature Citrus aurantium L. havingstomatous cuticles were collected and washed in DDW. Afterunching 20 mm diameter discs out of the leaves, the cuticlesere isolated enzymatically by incubating the leaf discs in aixture (1:1) of cellulose 203-13L (Biocatalysts Ltd., Cardiff,ales, UK) and Pectinase 62L (Biocatalysts Ltd.) as described

y Schonherr and Riederer (1986) at a concentration of 1%w/w). After a few days, astomatous cuticles from the uppereaf epidermis were collected, air-dried, and stored in a refrig-rator until used. These isolated astomatous adaxial cuticles areeferred to hereafter as cuticular membrane (CMs).

Donor solutions including the vesicles [(i) 2,4-D as con-rol, (ii) compound IV vesicles encapsulating 2,4-D, and (iii)PPC encapsulating 2,4-D] were prepared after adding 14C

abeled 2,4-D (specific activity 19.2 mCi/mmol, Sigma) as aracer (20,000–30,000 cpm/�l) exactly as described above (Sec-ion 2.3). Rates of cuticular penetration were measured at 30%H and 25 ◦C, using the SOFU procedure as described bychonherr (2000). Receiver solution (DDW) was quantitativelyithdrawn after 1, 2, 4, 24, 48, 72, 96, and 168 h of the experi-ent for scintillation counting, and was replaced by fresh DDW.

t the end of the experiment, the CM was cut out and, after

dding the scintillation cocktail, counted to determine the quan-ity of radioactive material remaining on the surface of the CM.

Beckman LS 1701 scintillation counter (Beckman Coulter

8 f Biote

Irmifta(

2

dmvpsAraD1cnss

f

1

2

a

ow

3

mDasTsios1

ncpshaett(ehU

F((

8 Z. Wiesman et al. / Journal o

nc., Fullerton, CA, USA) was used to determine the sampleadioactivity. The amount applied (M0) was calculated by sum-ing amounts penetrated (Mt) plus the amount remaining on the

ndividual CMs at the end of the experiment. Thus, Mt/M0 is theraction that penetrated and 1 − Mt/M0 is the fraction still left onhe surface of the CM. Data were plotted as −ln(1 − Mt/M0) ands percent of penetration versus time as described by Schonherr2000) and Wiesman et al. (2002).

.9. Protamine assay

The protamine assay of vesicle encapsulation efficiency wasetermined by the procedure of Kulkarni et al. (1995) with someodifications. 2,4-D [14C] was encapsulated in compound IV

esicles as described above. One hundred microliters of 0.4 Mrotamine was added to the 2,4-D [14C] (∼30,000 cpm) encap-ulated in compound IV vesicles, resulting in vesicle aggregates.fter 2 min of vortexing, the solution was incubated overnight at

oom temperature and then centrifuged (14,000 rpm) for 40 mint 25 ◦C. The supernatant, containing the non-encapsulated 2,4-

and empty vesicles, was transferred to a new vial. Another00 �l of DDW was added to the precipitated fraction that wasentrifuged again at the same conditions for 5 min. The super-atant was removed and added to the vial containing the previousupernatant. The radioactivity was determined in a Beckmancintillation counter.

The encapsulation percentage was calculated in the followingormula:

00 − 100 × 2, 4-D[14C] (cpm) remaining in the supernatant

2, 4-D[14C] (cpm) initially added to the solution before centrifugation

.10. Statistical analysis

In all the biological experiments, the data represent the aver-ge of at least four replicates plus standard error and analysis

cFps

ig. 1. Microscopic characterization of nano-vesicles produced from amphiphilic coma); amphiphile III (100,000×) (b); amphiphile IV (45,000×) (c); AFM characteriscreened image size—1000 nm2) (e).

chnology 130 (2007) 85–94

f variance using the Tukey–Kramer HSD test (SAS, JMP soft-are) at 0.01 space level of significance.

. Results and discussion

Several liposome delivery systems are available to carry bio-aterials through biological barrier systems (Provencher, 1982;eshmukh and Huang, 1997). Considerable activity in liposome

nd vesicle research is directed toward forming stable vesicleystems with specificity and drug target ability (Finsy, 1994).hese activities include development of optimum vesicle size,hape and surface morphology, composition, charge, bilayer flu-dity, and ability to incorporate a wide spectrum of biomaterialsr to carry cell-specific ligands, all of which may enable theystem to be useful and versatile (McCormack and Gregoriadis,996).

Towards the objective of achieving vesicle characteristicseeded for the delivery of biologically active agents in biologi-al systems, the stability, encapsulation efficiency, and transportroperties of vesicles made from two new amphiphiles weretudied. As described in Section 1, amphiphile III is a single-eaded amphiphile with two alkyl chains and a single quaternarymmonium head group (Scheme 1), while amphiphile IV isssentially three molecules of compound III bound togetherhrough a glycerol moiety. Amphiphile III has a molecular struc-ure similar to other amphiphiles known to form bilayer vesiclesOkahata and Kunitake, 1980; Nagamura et al., 1978). How-ver, the molecular structure of compound IV is unique and,ence, its vesicle forming properties are difficult to predict.sing TEM microscopy, the ability of the two tested amphiphilic

ompounds to form spherical vesicles was demonstrated (Fig. 1).resh preparations of amphiphile III and IV vesicles were com-ared with DPPC liposomes (Fig. 1a). All three preparationshowed clear negative staining TEM images, demonstrating

pounds III and IV. TEM characterization of control DPPC vesicles (25,000×)zation of amphiphile IV (screened area size—5000 nm2) (d); amphiphile IV

Z. Wiesman et al. / Journal of Biote

Fd

iarpcia

Imtiss

Ibpaswab

wsevSs(evyaevifdtmtsefladmwt

IVdaa

Fcv

ig. 2. Dynamic light scattering (DLS) of amphiphile IV vesicle population sizeistribution.

ntact vesicles. Amphiphile III was shown to form planar bilayerssembled vesicles in Grinberg et al. (2005), and now showed aelatively uniform pattern of spherical vesicles. Compound IV,reviously suggested in the same study to form micelle type vesi-les, showed a wider range of variation in terms of vesicle sizen comparison to the amphiphile III and DPPC as well (Fig. 1bnd c).

AFM image screening of 5000 nm2 of amphiphilic compoundV (Fig. 1d) demonstrated again, using a different microscopicethodology, the formation of spherical vesicles that vary in

erms of size. These vesicles were even more visible when zoom-ng on 1000 nm2 (Fig. 1e). The variation in amphiphile IV vesicleize is clearly shown in these microscopic systems and furtherhown in Fig. 3.

Using dynamic light scattering (DLS), 3-day-old amphiphileV vesicle population size distribution was found to rangeetween 40 and 110 nm (Fig. 2). The vast majority of the vesicleopulation was concentrated around 70 nm. These results werelso confirmed by vesicle size determination analysis in AFM

ystem (not shown). Furthermore, these data are in agreementith Grinberg et al. (2005) reported of amphiphile IV vesicle

verage diameter of 78 nm and amphihile III vesicle size rangingetween 50 and 200 nm.

pils

ig. 3. TEM images of vesicles formed from amphiphilic compounds III and IV eompound III (45,000×) (a). Image of UA encapsulated in amphiphilic compoundesicles.

chnology 130 (2007) 85–94 89

The encapsulation potential of amphiphile III and IV vesiclesas initially studied using UA, commonly used for negative

taining of TEM preparations (Nardin et al., 2000; Ottavianit al., 2000). In this case, UA was added while preparing theesicles in the reverse phase evaporation method (described inection 2). In Fig. 3a it is clearly shown that UA was encap-ulated in many of the relatively mid-size (∼50 nm) and large∼100 nm) amphiphile III vesicles. In agreement with Grinbergt al. (2005), almost no encapsulation could be observed in theery small size vesicles of this preparation (<10 nm). Anal-sis of the encapsulation capacity of amphiphile IV vesicleslso showed that mainly the relatively large vesicles (>100 nm)ncapsulated UA and again, a pattern of fusion of small sizeesicles (<25 nm) into large encapsulating vesicles could eas-ly be observed (Fig. 3b and c). This may suggest that theusion of small vesicles into a large core vesicle enabled UAissolved in the aqueous medium to be continuously loaded intohe core vesicle as a result of dynamic activity taking part in theembrane of the core fusing vesicle (possible physical explana-

ion will be further discussed later). The present encapsulationtudy was carried out with 3-week-old vesicles. These data mayxplain why, in the previous partial characterizing study usingreshly prepared vesicles (Grinberg et al., 2005), UA encapsu-ation could be observed only with amphiphilic compound IIInd not with compound IV. The continuous process of fusionescribed for the micelle like amphiphilic compound IV vesiclesay raise a question concerning the shelf life of this vesicle type,hich is highly important for characterization of their applica-

ion potential, as previously suggested by Grinberg et al. (2005).The stability of vesicles prepared from amphiphiles III and

V, and DPPC was carried out over a 2-month period (Fig. 4).esicles from amphiphile III were found to be intact after 17ays (Fig. 4a). The vesicles of this preparation maintained a rel-tively stable shape during 3 weeks with minimal fusion andggregation, as also shown in Fig. 3a. TEM analysis of com-

ound III vesicles some days later (24 days, Fig. 4b) showednitiation of formation of large oily spots, which may suggestarge scale aggregation and fusion after several weeks at benchtorage conditions. These results are in agreement with a previ-

ncapsulating uranyl acetate (UA). Image of UA encapsulated in amphiphilicIV (30,000×) (b) and (20,000×) (c), both showing fusion process of these

90 Z. Wiesman et al. / Journal of Biotechnology 130 (2007) 85–94

F abilit1 hiphi(

ov

csattsacaiawafmlc

osaslOT(g

pmcfaLad

pfwvtsacpsmsu

tsa

ig. 4. TEM images of amphiphile compounds III and IV demonstrating their st7 days (70,000×) (a); amphiphile III vesicles after 24 days (15,000×) (b); amp45,000×) (d); DPPC vesicles after 14 days (60,000×) (e).

us report (Grinberg et al., 2005) suggesting that compound IIIesicles are stable for about 4 weeks.

Forty-five days after preparation, the typical image of largeore amphiphile IV vesicles surrounded with many relativelymall vesicles (<25 nm) could easily be observed (Fig. 4c), aslso shown after 21 days (Fig. 3b and c). This pattern is dras-ically increased 59 days after preparation (Fig. 4d), and evenhe large vesicles started to aggregate and to fuse between them-elves. These data confirmed the pattern of a relatively slownd continuous fusion of small vesicles into the relatively largeore vesicles (>100 nm) obtained previously (Fig. 3) and suggestgain a moderate rate of fusion for 45 days, significantly increas-ng upon further storage. The DPPC control vesicles aggregatedt a faster rate than amphiphilic compounds III and IV. Twoeeks after preparation, the DPPC vesicles were found to be

ggregated and fused (Fig. 4e) in comparison to their fresh intactorm (Fig. 1a). These latter data concerning DPPC are in agree-ent with many common studies reporting on the short shelf

ife of liposome systems without physical stabilizers such asholesterol (Hans-Hening et al., 1980; Po-Shun et al., 1981).

The physical stability of vesicles is often characterized, inne important aspect, by changes in average particle size andize distribution due to vesicle aggregation and fusion (Gritnd Crommelin, 1993). Changes in particle size in colloidalystems occur by two different mechanisms—on the molecularevel this can be the asymmetric molecular exchange (called

swald ripening) that results in an increase in vesicle size.his is a strong function of the critical micelle concentration

CMC) values with the higher values predictive of a highrowth rate, whereas vesicles with low CMC values grow

ipCO

y during 59 days on the bench at room conditions. Amphiphile III vesicles afterle IV vesicles after 45 days (30,000×) (c); amphiphile IV vesicles after 59 days

rimarily by an aggregation and fusion mechanism. Low CMCakes the vesicle structure stable against dilution and prevents

hange in vesicle size by Oswald ripening. Bilayer vesiclesrom phospholipids with low CMC can grow by fusion throughmphiphile exchange with the exterior (Kaler et al., 1989;asic, 1996). In addition to the CMC value, vesicle stability islso determined by a variety of external perturbations, such asilution, ionic strength, temperature, and pressure.

Examination of the structures of amphiphiles III and IV, com-ared to DPPC, indicates that the novel cationic vesicles derivedrom vernonia oil would have a higher CMC than the latter,hich has a CMC value of ∼10−8 M. Thus, vesicles from theernonia oil may increase in size by Oswald ripening, whilehose from DPPC increase by aggregation. The difference intability between vesicles of amphiphile III versus vesicles ofmphiphile IV may be related to the difference in packing effi-iency. The intrinsic curvature of bilayers of amphiphile III isostulated to allow the formation of more stable vesicles thaneen with amphiphile IV (Grinberg et al., 2005). Thus, asym-etric amphiphile exchange may occur more readily between

mall size vesicles of amphiphile IV, which would grow in sizentil balanced by entropic forces.

In conclusion the molecular structure and the intrinsic curva-ure of amphiphile III allows for more stable vesicles of smallerize than 100 nm as compared to the aggregate structures ofmphiphile IV. Because of the relatively poor stability of the

nitially formed amphiphile IV aggregates, by the method ofreparation (reverse phase evaporation) and apparently highMC values, these aggregates will grow to larger structures byswald ripening effects. The reduced curvature requirements

Z. Wiesman et al. / Journal of Biotechnology 130 (2007) 85–94 91

F . AFMi c); cov with r

nsehcte

ca(b(srtoav(t1c

cstacIw(s

These results also showed that the interaction between DNAand amphiphile IV vesicles took place in vesicles size of about100 nm, in addition to the relatively large vesicles (>500 nm), asobserved in the confocal laser microscope system (Fig. 6).

Fig. 6. Confocal laser microscopy image of DNA plasmid encapsulated in

ig. 5. Encapsulation of DNA plasmid (pJD328, 5.2 kb) in amphiphile IV vesiclen semi-relaxed form (b); AFM image of DNA plasmid in fully relaxed form (esicle (d). Vesicles were stained with fluorescein and DNA plasmid is stained

eeded for larger vesicles could then allow for large relativelytable encapsulating structures from amphiphile IV. The overallncapsulation efficiency of such large structures would be quiteigh. The smaller vesicles of amphiphile III would incorporateonsiderably less material. Thus amphiphile IV with its abilityo form large enough vesicles makes it an ideal candidate forncapsulating large structures such as DNA plasmids.

DNA plasmid (pJD328, 5.2 kb) was encapsulated in vesi-les of amphiphilic compound IV. Initially, we characterizednd visualized the assembly of the DNA plasmid in AFMFig. 5). As expected, the plasmid was found to be assem-led in relaxed (fully and semi-relaxed) and condensed formsthe supercoiled form was not observed). This characterizationhowed that in relaxed form, the DNA has dimensions in theange of 300–500 nm. Potentially, only in condensed form couldhe DNA plasmid be encapsulated in the relatively large vesiclesf amphiphile IV. By increasing amphiphile concentration andvoiding sonication, it was possible to prepare a relatively largeesicle population where the vesicles formed were large enough>200 nm) to encapsulate the DNA plasmids. Using the abilityo affect the DNA folding process by pH reduction (Heywood,971; Kyungsik et al., 1997), DNA plasmid was prepared in itsondensed form (Fig. 5).

Due to application of a differential dye commonly used foronfocal laser microscopy systems, the non-fluorescent encap-ulating lipid vesicle layer could be readily differentiated fromhe fluorescence emissions of the encapsulated DNA plasmidnd, indeed, the encapsulation of the DNA plasmid in the vesi-les was clearly shown (Fig. 5). The fluorescence of compound

V vesicles stained with fluorescein, and DNA plasmid stainedith rhodamine, were obtained at 543 and 488 nm, respectively

Fig. 6). The fluorescence emission of both confirmed the encap-ulation or at least association of the DNA with the vesicles.

ac4lr

image of DNA plasmids in condensed form (a); AFM image of DNA plasmidsnfocal image of a condensed DNA plasmid encapsulated in an amphiphile IVhodamine (1000×).

mphiphile IV vesicles. The plasmid is stained with rhodamine and the vesi-les are stained with fluorescein. Upper left image reflects 543 nm, upper right88 nm, lower left no fluorescence, lower right combination of both fluorescentights (1000×). (For interpretation of the references to color in the text, theeader is referred to the web version of the article.)

9 f Biotechnology 130 (2007) 85–94

bcdaDctocgrDdov

wttdwtsaiIctrciwc

sadDDoe1tHoe

ehcoa(cAt

Fig. 7. Comparison of transport rate of 2,4-D[14C] encapsulated in amphiphileIV vesicle through Citrus grandis plant cuticle membrane to DPPC encap-sulating vesicles. Transport test was carried out at 25 ◦C and 30% relativehumidity. Transformed value (−ln(1 − Mt/M0)) shown in the left y-axisand percent of penetration through the cuticle membrane are shown inthe right side y-axis. Data are shown as mean ± S.E. The curve for-mula is divided into two time periods; 0–96 and 96–168 h are shown foreDI

dc1ctp(ii2dsticrf

amembrane model clearly shows the efficacy of this vesicle sys-tem in terms of delivery of biomaterials through biologicalbarriers. TEM analysis of the fate of amphiphile IV vesicles

Table 1Effect of amphiphilic compound IV vesicles on 2,4-D[14C] transport rate con-stant through isolated Citrus grandis leaf cuticle membranea

DPPC/control AmphiphileIV/DPPC

AmphiphileIV/control

2 Z. Wiesman et al. / Journal o

For further and better characterization of the interactionetween DNA plasmid and amphiphile IV vesicles, a virtualross-sectioning was done by the laser ray system. A gradualisappearance of the green color of the vesicle envelope andppearance of the red color coming from the rhodamine-stainedNA inside the vesicle was clearly obtained by the laser virtual

ross-sectioning, suggesting of encapsulation of the DNA insidehe vesicle and not of an association of DNA to the outer layerf the vesicles. The lower right image in Fig. 6 shows yellowolor in the position of the encapsulated DNA. Thus, it is sug-ested that the yellow color is the result of fluorescin-green andhodamine-red used for specific staining of the vesicles and theNA, respectively. These latter data further increase the confi-ence of the encapsulation results shown also in Fig. 5 and notnly association between the DNA plasmid and amphiphile IVesicle membrane.

Cationic vesicles (positively charged) can form complexesith DNA plasmids by a mechanism that involves the nega-

ively charged plasmids interacting with the positive charge ofhe surface of the cationic vesicle (Lasic, 1997a). X-ray studiesemonstrate lamellar structures showing condensed DNA sand-iched between cationic bilayers (Lasic, 1997b). In our study,

he confocal laser microscopy virtual cross-sectioning clearlyhowed that the DNA plasmid is located inside the vesicless described above. A possible mechanism of formation maynvolve the continuous fusion characteristic of the amphiphileV vesicles that may build after the DNA plasmid is trapped in aomplex for a period of time. This may explain the DNase protec-ion of the DNA plasmid obtained with amphiphile IV vesicleseported in the previous study (Grinberg et al., 2005). Theseharacteristics of amphiphile IV may be explained in terms ofts lyotropic behavior as described by Grinberg et al. (2005)hich may contribute to understanding the DNA encapsulation

apacity of these vesicles.In order to quantify the efficacy of the compound IV encap-

ulation procedure, we tried to differentiate between the freend the trapped biomaterial in amphiphile IV vesicles. However,ue to practical difficulties to carry out this quantification usingNA plasmid, we used a model based on radioactive labeled 2,4-protamine precipitation assay (Kulkarni et al., 1995). Based

n this assay, carried out in five replicates, the encapsulationfficiency of 2,4-D in amphiphile IV vesicles was found to be4% ± 1.3. The effect of the chemical and physical properties ofhe biomaterial used on the rate of encapsulation is well known.owever, the level of encapsulation rate of the biomarker used inur study (2,4-D) falls well inside the range reported for vesiclencapsulation of biomaterials (Gregoriadis, 1993).

As shown by Grinberg et al. (2005), the DNA transfectionfficiency of encapsulating amphiphile VI vesicles is relativelyigh in comparison to other tested vehicle systems in mammalianells. Therefore, we carried out an analysis of the transport ratef 2,4-D-[14C] encapsulated in amphiphile IV vesicles throughwell-established C. grandis plant cuticle membrane model

Schonherr and Baur, 1995). Initially, the isolated intact cuti-le membrane was clearly visualized by SEM (data not shown).mphiphile IV vesicles encapsulating 2,4-D were applied on

he outer surface of the C. grandis cuticle membranes.

9

c

ach curve. Con Y0–96 = 0.0038X + 0.0787, Con Y96–168 = 0.0016X + 0.236,PPC Y0–96 = 0.0076X + 0.003, DPPC Y96–168 = 0.0034X + 0.369, amphiphile

V Y0–96 = 0.0178X + 0.13, and amphiphile IV Y96–168 = 0.0051X + 1.2.

The present study showed a significant increase of the rate ofelivery of compound IV encapsulated radio-labeled 2,4-D inomparison to encapsulation in DPPC vesicles over a period of68 h (Fig. 7). This transport study showed that this time periodould be divided into two phases (time 0–96 h and 96–168 h). Inhe first phase of 0–96 h, DPPC vesicles had doubled the trans-ort rate constant compared to non-encapsulated 2,4-D controlTable 1). The treatment consisting of amphiphile IV furtherncreased the rate of 2,4-D transport through cuticle membranesn comparison to the control and DPPC treatments, 4.68- and.34-fold, respectively. In the second phase (from 96 to 168 h),ue to saturation of acceptor solutions, the transport rate con-tant of all treatments was reduced, but the differences betweenhe treatments were maintained. Additional tests (not shown)n the same model system at various temperature and humidityonditions with surfactants well known to affect the diffusionate through cuticle membranes (Schonherr and Baur, 1995),urther proved these results.

This study of the transport of 2,4-D encapsulated inmphiphilic compound IV vesicles through the plant cuticle

0–96 h 2.00 4.68 2.346–168 h 2.12 3.18 1.50

a 2,4-D [14C] transport rate constant ratio between vesicle preparations wasalculated from the curve formula presented in Fig. 8.

Z. Wiesman et al. / Journal of Biote

F 14

aV

tsdrPaIpopVbv(evtcrvco

t(cbc(tcAa

stttttiDtit

A

twM

R

B

B

D

D

F

G

G

G

G

G

G

H

H

H

Kaler, E.W., Murthy, A.K., Rodriquez, B.E., Zasadzinski, J.A.N., 1989.

ig. 8. TEM characterization of the fate of 2,4-D[ C] encapsulated inmphiphile IV vesicles after transport through C. grandis leaf cuticle membrane.esicles are stained in UA (25,000×).

hat were transported through these isolated cuticle membraneshowed intact 2,4-D encapsulating vesicles (Fig. 8). These lastata further confirmed the effectiveness of this vesicle prepa-ation as a biomaterial delivery system through “plant skin”.reviously, Grinberg et al. (2005) suggested that the samemphiphilic vesicles as used in the present study (amphiphileV) were significantly more effective than DEAE–dextran com-lexes for DNA transfection in mammalian cells. Developmentf a delivery system such as described in this study may open theossibility for efficient in vivo transformation of plant tissues.esicle systems could provide a valuable delivery alternative toe used in genetic engineering of plants, by avoiding the disad-antages of bombardment or virus or Agrobacterium infectionWatson et al., 2005). Indeed, it is conceivable that vesicle deliv-ry systems should not be species-restricted. The encapsulatingesicles demonstrated in the present study would provide pro-ection from nucleases inside the cell, thereby increasing thehances of successful delivery and effectiveness of a wide sizeange of DNA molecules. In addition to DNA transformation,esicles could mediate the introduction of peptides, toxins, pesti-ides, and other active bio-agents into plant tissues for inductionf resistance traits for biotic and abiotic stresses.

Concerning future further improvement of this delivery sys-em, vesicle stability is often enhanced by adding cholesterolBondurant et al., 2001) to the vesicle membrane for mechani-al and physical strength, and by the use PEG ligands as a stericarrier against fusion (Sihorkar and Vyas, 2001), both of whichan improve the stability of the vesicles for up to 1 year at 4 ◦CPapahadjopoulos et al., 1974). Some of the stabilizing addi-

ives can be problematic from the medical point of view and canause side effects, as reported previously (Grinberg et al., 2005).voiding the use of these additive compounds may ease thepplication of this vesicle system. As a follow-up to the present

K

chnology 130 (2007) 85–94 93

tudy, it is suggested to further develop specific tissue or cellargetability of the amphiphile IV vesicles in order to advanceheir application in various biological systems. Specifically, inhe field of plant intervention, increasing the targetability ofhe vesicle system may directly enhance the chances to directransformed nucleic acid to the cell nucleus or the organelle ofnterest, thereby increasing not only the relative efficiency ofNA transfection, but also the chances of successful in vivo

ransformation. Moreover, one can envisage the possibility ofnducing DNA delivery to obtain more accurate control of theransgene expression.

cknowledgements

We would like to thank Prof. Avi Levy from Weizmann Insti-ute for providing the DNA plasmid, Mrs. Mazit Namer for TEMork, Dr. Michal Hershfinkel for Confocal microscopy, and tos. Edna Oxman for editing this manuscript.

eferences

arenholtz, Y., Amselem, S., Lichtenberg, D., 1979. A new method for prepa-ration of phospholipid vesicles (liposomes)—French press. FEBS Lett. 99,210–214.

ondurant, B., Muller, A., O’Brien, F., 2001. Photoinitiated destabilization ofsterically stabilized liposomes. Biochim. Biophys. Acta 1511, 113–122.

avey, M.R., Anthony, P., Power, J.B., Lowe, K.C., 2005. Plant protoplasts:status and biotechnological perspectives. Biotechnol. Adv. 23 (2), 131–171.

eshmukh, H.M., Huang, L., 1997. Liposome and polylysine mediated genetransfer. New J. Chem. 21, 113–124.

insy, R., 1994. Particle sizing by quasi-elastic light scattering. Adv. ColloidInterf. Sci. 52, 79–143.

regoriadis, G. (Ed.), 1993. Liposome Technology, vol. I/III, 2nd ed. CRC Press,Boca Raton, FL.

regoriadis, G., McCormack, B., Morrison-Perrie, Y., Saffie, R., Zadi, B., 1998.Liposome in drug targeting. In: Celis, J. (Ed.), Cell Biology: LaboratoryHandbook, vol. 4, 2nd ed. Academic Press, New York, pp. 131–140.

regoriadis, G., Ryman, B.E., 1972. Lysosomal localization of enzyme-containing liposomes injected into rats. Biochem. J. 129, 123–133.

rinberg, S., Linder, C., Kolot, V., Wiesman, Z., Heldman, E., 2002. Amphiphilicderivatives for the production of vesicles, micelles and complexants, andprecursors thereof. WO02055011.

rinberg, S., Linder, C., Kolot, V., Waner, T., Wiesman, Z., Shaubi, E., Held-man, E., 2005. Novel cationic amphiphilic derivatives from vernonia oil:synthesis and self-aggregation into bilayer vesicles, nanoparticle, and DNAcomplexants. Langmuir 21, 7638.

rit, M., Crommelin, D., 1993. Chemical stability of liposomes: implicationsfor their physical stability. Chem. Phys. Lipids 64, 3.

ans-Hening, H., Hupfer, B., Koch, H., Ringsdorf, H., 1980. Polimeriz-able phospholipids analogues—new stable biomembrane and cell models.Angew. Chem. Int. Ed. Engl. 19, 938–940.

ansma, H.G., Laney, D.E., 1996. DNA binding to mica correlates with cationicradius: assay by atomic force microscopy. Biophys. J. 70, 1933–1939.

eywood, V.H., 1971. The characteristics of the scanning electron microscopeand their importance in biology studies. In: Heywood, V.H. (Ed.), Pro-ceedings of International Symposium on Scanning Electron MicroscopeSystematic and Evolutionary Applications. Department of Botany, Univer-sity of Reading. Academic Press, London, pp. 1–16.

Vesicle–biopolymer gels: networks of surfactant vesicles connected by asso-ciating biopolymers. Science 245, 1371–1374.

ulkarni, S.B., Dipali, S.R., Betageri, G.V., 1995. Protamine-induced aggrega-tion of unialellar liposomes. Pharm. Sci. 1, 359–362.

9 f Biote

K

L

L

LM

N

N

O

O

O

P

P

P

P

R

S

S

S

S

S

S

T

4 Z. Wiesman et al. / Journal o

yungsik, K., Sung, S.W., Byung-Yun, S., 1997. Micromorphology of the leafepidermis of Korean umbelliferous plants. 1. Preliminary study. Phytomor-phology 47 (1), 87–95.

asic, D.D., 1996. Stealth liposomes. In: Benita, S. (Ed.), Microencapsulation,Methods and Industrial Applications. Marcel Dekker Inc., New York, p. 302(Chapter 11).

asic, D.D., 1997a. Colloid chemistry. Liposomes within liposomes. Nature387, 26–27.

asic, D.D., 1997b. Liposomes in Gene Delivery. CRC Press, Boca Raton, FL.cCormack, B., Gregoriadis, G., 1996. Comparative studies of the fate

of free and liposome-entrapped hydroxypropyl-�-cyclodextrin/drug com-plexes after intravenous injection into rats: implications in drug delivery.Biochim. Biophys. Acta 1291, 237–244.

agamura, T., Mihara, S., Okahata, Y., Kunitake, T., Matsuo, T., 1978. NMRand fluorescence studies on self-assembling behavior of dialkyldimethy-lammonium salts in aqueous solutions. Berichte der Bunsen-Gesellschft 82,1093–1098.

ardin, C., Hirt, T., Leukel, J., Meier, W., 2000. Polymerized ABA triblockcopolymers vesicles. Langmuir 16, 1035–1041.

kahata, Y., Kunitake, T., 1980. Self-assembling behavior of single-chainamphiphiles with the biphenyl moiety in dilute aqueous solution. Berichteder Bunsen-Gesellschft 84, 550–556.

llivon, M., Lesieur, S., Grabielle-Madelmont, C., Paternostre, M., 2000. Vesi-cles reconstitution from lipid-detergent mixed micelles. Biochim. Biophys.Acta 1508, 34–50.

ttaviani, M.F., Favuzza, P., Bigazzi, M., Turro, N.J., Jockusch, S., Tomalia,D.A., 2000. A TEM and EPR investigation of the competitive binding ofuranyl ions starburst dendrimers and liposomes: potential use of dendrimersas uranyl ion sponges. Langmuir 16, 7368–7372.

apahadjopoulos, D., Poste, G., Schaeffer, B.E., Vail, W.J., 1974. Membranefusion and molecular segregation in phospholipid vesicles. Biochim. Bio-phys. Acta 352, 10–28.

atnaik, D., Khurana, P., 2001. Wheat biotechnology—a minireview. Electr. J.Biotech. 4 (2), 74–102.

W

W

chnology 130 (2007) 85–94

o-Shun, W., Tin, G.W., Baldeschwieler, J.D., Shen, T.Y., Ponpipom, M.M.,1981. Effects of surface modification on aggregation of phospholipid vesi-cles. Proc. Natl. Acad. Sci. U.S.A. 78, 6211–6215.

rovencher, S.W., 1982. Contin: a general purpose constrained regularizationprogram for inverting noisy linear algebraic and integral equations. Comput.Phys. Commun. 27, 229–242.

akoczy-Trojanowska, M., 2002. Alternative methods of plant transformation—a short review. Cell Mol. Biol. Lett. 7 (3), 849–858.

chmidt, M.C., Rothen-Rutishauser, B., Rist, B., Beck-Sickinger, A., Wunderli-Allenspach, H., Rubas, W., Sadee, W., Merkle, H.P., 1998. Translocation ofhuman calcitonin in respiratory nasal epithelium is associated with self-assembly in lipid membrane. Biochemistry 37, 16582–16590.

chonherr, J., 2000. Calcium chloride penetrates plant cuticles via aqueouspores. Planta 212, 112–118.

chonherr, J., Baur, P., 1995. Cuticle permeability studies: a model for esti-mating leaching of plant metabolites to leaf surfaces. In: Morris, C., Nicot,P., Nguyen, C. (Eds.), Microbiology of Aerial Plant Leaf Surfaces. PlenumPublishing, London.

chonherr, J., Riederer, M., 1986. Plant cuticles absorb lipophilic compoundsduring enzymatic isolation. Plant Cell Environ. 4, 349–354.

ihorkar, V., Vyas, S.P., 2001. Potential of polysaccharide anchored liposomesin drug delivery, targeting and immunization. J. Pharm. Pharm. Sci. 4,138–158.

zoka, F., Papahadjopoulos, D., 1978. Procedure for preparation of liposomeswith large internal aqueous space and high capture by reverse-phase evapo-ration. Proc. Natl. Acad. Sci. U.S.A. 75, 4194–4198.

homson, N.H., Kasas, S., Smith, B., Hansma, H.G., Hansma, P.K., 1996.Reversible binding of DNA to mica for AFM imaging. Langmuir 12,5905–5908.

atson, J.M., Fusaro, A.F., Wang, M.B., Waterhouse, P.M., 2005. RNA silencingplatforms in plants. Minireview. FEBS Lett. 579, 5982–5987.

iesman, Z., Luber, M., Ronen, A., Markus, A., 2002. Fertivant—a new non-destructive and long lasting in vivo delivery system for foliar nutrients. ActaHortic. 594, 585–590.