11828
Langmuir 2008, 24, 11828-11833
Confinement of DNA in Water-in-Oil Microemulsions Anita Swami,† Gabriel Espinosa,† Samuel Guillot,‡ Eric Raspaud,† Franc¸ois Boue´,§ and Dominique Langevin*,† Laboratoire de Physique des Solides, UMR 8502 UniVersite´ Paris-Sud, 91405 Orsay, France, CRMD, UMR 6619 UniVersite´ d’Orle´ans, 45071 Orle´ans, France, and Laboratoire Leon Brillouin, CEA Saclay, France ReceiVed April 18, 2008. ReVised Manuscript ReceiVed August 19, 2008 The study of systems that allow DNA condensation in confined environments is an important task in producing cell-mimicking microreactors capable of biochemical activities. The water droplets formed in water-in-oil emulsions are potentially good candidates for such microcompartments. The anionic surfactant AOT was used here to stabilize the droplets. We have studied in detail the DNA distribution and the structural modifications of these microemulsion drops by varying the concentration and molecular weight of DNA and using various techniques such as light, X-ray, and neutron scattering, electrical conductivity, and surface tension. DNA induces the formation of large drops into which it is internalized. The size of these drops depends on the amount of DNA dissolved in water as well as on its molecular weight. The local DNA concentration is very high (>100 mg/mL). The large drops coexist with small empty drops (not containing DNA), similar to those found in the DNA-free microemulsion.
I. Introduction Cell nuclei contain all of the genomic DNA chains and can be viewed as microreactors of intense biochemical activity. Cell-sized water droplets such as liposomes were recently used as a cellmimicking system for enzymatic reactions in confined environments.1 However, to mimic the cell nuclei, the incorporated DNA chains must be also compacted. The conversion of the long and coiled DNA conformation into a highly compact form is called DNA condensation and is also crucial to achieve in gene delivery carriers to enhance the cellular intake of DNA. Getting information on the processes allowing DNA to be condensed, particularly in confined media, is therefore an important issue. In the bulk, various condensing agents such as cationic surfactants or lipids,2 neutral or cationic polymers,3 multivalent ions,4 basic proteins,5 and alcohols6 have been used to date. The link between DNA condensation and the structure and phase transitions in the dense phases of DNA has been studied in detail by Livolant’s group.7 * Corresponding author. E-mail:
[email protected]. † UMR 8502 Universite´ Paris-Sud ‡ UMR 6619 Universite´ d’Orle´ans. § CEA Saclay. (1) Yamada, A.; Yamanaka, T.; Hamada, T.; Hase, M.; Yoshikawa, K.; Baigl, D. Langmuir 2007, 23, 384. (2) (a) Bilalov, A.; Leal, C.; Lindman, B. J. Phys. Chem. B 2004, 108, 15408. (b) Mel’nikova, Y. S.; Lindman, B. Langmuir 2000, 16, 5871. (c) Radler, J. O.; Koltover, I.; Jamieson, A.; Salditt, T.; Safinya, C. R. Langmuir 1998, 14, 4272. (d) Salditt, T.; Koltover, I.; Radler, J. O.; Safinya, C. R. Phys. ReV. E 1998, 58, 889. (e) Le Ny, A.-L., M; Lee, C. T., Jr J. Am. Chem. Soc. 2006, 128, 6400. (f) Cardenas, M.; Braem, A.; Nylander, T.; Lindman, B. Langmuir 2003, 19, 7712. (g) Marchetti, S.; Onori, G. J. Phys. Chem. B 2005, 109, 3676. (h) Radler, J. O.; Koltover, I.; Salditt, T.; Safinya, C. R. Science 1997, 275, 810. (i) Wagner, K.; Harries, D.; May, S.; Kahl, V.; Radler, J. O.; Ben-Shaul, A. Langmuir 2000, 16, 303. (3) (a) Lerman, L. S Proc. Natl. Acad. Sci. U.S.A. 1971, 68, 1886. (b) Hanlon, S.; Brudno, S.; Wu, T. T.; Wolf, B. Biochemistry 1975, 14, 1648. (4) (a) He, S. Q.; Arscott, P. G.; Bloomfield, V. A. Biopolymers 2000, 53, 329. (b) Kankia, B. I.; Buckin, V.; Bloomfield, V. A. Nucleic Acids Res. 2001, 29, 2795. (c) Gueron, M.; Demaret, J. P.; Filoche, M. Biophys. J. 2000, 78, 1070. (5) (a) Garcia-Ramirez, M.; Subirana, J. Biopolymers 1994, 34, 285. (b) Kundu, T.; Rao, M. Biochemistry 1995, 34, 5143. (c) Raspaud, E., J.; Pelta, J.; de Frutos, M.; Livolant, F. Phys. ReV. Lett. 2006, 97, 068103. (6) (a) Gray, D. M.; Edmondson, S. P.; Lang, D.; Vaughan, M. Nucleic Acids Res. 1979, 6, 2089. (b) Arscott, P. G.; Ma, C.; Wenner, J. R.; Bloomfield, V. A. Biopolymers 1995, 36, 345. (c) Maestre, M. F.; Reich, C. Biochemistry 1980, 19, 5214. (7) Livolant, F.; Leforestier, A. Prog. Polym. Sci. 1996, 21, 1115.
Water droplets formed in water-in-oil (W/O) microemulsions were also reported to be good candidates for confining DNA in microcompartments. Indeed, DNA is easily solubilized into these microemulsions: because DNA is very water-soluble, it can be dissolved in the water phase used in the second stage to form a water-in-oil microemulsions (W/O). Such systems are made of water droplets coated with a surfactant monolayer and dispersed in oil. The size of the water droplets is controlled by the relative amounts of water and surfactant and is up to about 10 nm in the case of microemulsions stabilized by Na-AOT (sodium bis-(2ethylhexyl) sulfosuccinate).8 In the presence of DNA, much larger droplets have been observed.9-12 The solubilization of DNA or other polynucleotides in W/O microemulsions has been studied with systems stabilized by ionic surfactants such as CTAB (cetyltrimethylammonium bromide),9,10 SDS (sodium dodecylsulphate),10 and Na-AOT (sodium bis-(2-ethylhexyl) sulfosuccinate)10-12 aswellasnonionicsurfactantssuchastetraethyleneglycol dodecyl ether12 and lipids.13 Many studies, such as those of Luisi and co-workers, have made use of infrared (IR) and circular dichroic (CD) spectroscopy.9-14 From IR spectroscopic measurements, it was observed that the state of water inside the microemulsion is not changed as a result of the presence of DNA inside the microemulsion, whereas CD measurements show that upon decreasing water content in the microemulsion the configuration of DNA becomes reminiscent of that of the condensed Ψ form. Though the molecular configuration of DNA inside the water pool in the microemulsion has been quite well studied with these spectroscopy methods, it is yet not clear whether and how the presence of DNA in the microemulsion affects the (8) Hou, M-J.; Shah, D. O. Langmuir 1987, 3, 1086. (9) (a) Airoldi, M.; Boicelli, C. A.; Gennaro, G.; Giomini, M.; Giuliani, A. M.; Giustini, M. Phys. Chem. Chem. Phys. 2000, 2, 4636. (b) Airoldi, M.; Boicelli, C. A.; Gennaro, G.; Giomini, M.; Giuliani, A. M.; Giustini, M.; Scibetta, L. Phys. Chem. Chem. Phys. 2002, 4, 3859. (10) Anarbaev, R. O.; Elepov, I. B.; Lavrik, O. I. Biochim. Biophys. Acta 1998, 1384, 315. (11) Pietrini, A. V.; Luisi, P. L. Biochim. Biophys. Acta 2002, 1562, 57. (12) Sarkar, R.; Pal, S. K. Biopolymers 2006, 83, 675. (13) Budker, V. G.; Slattum, P. M.; Monahan, S. D.; Wolff, J. A Biophys. J. 2002, 82, 1570. (14) (a) Osfouri, S.; Stano, P.; Luisi, P. L. J. Phys. Chem. B 2005, 109, 19929. (b) Monnard, P. A.; Oberholzer, T.; Luisi, P. L. Biochim. Biophys. Acta 1997, 1329, 39.
10.1021/la802233e CCC: $40.75 2008 American Chemical Society Published on Web 09/27/2008
DNA in Water-in-Oil Microemulsions
size and/or shape of the water cores of the droplets (water pool). Several light scattering studies revealed the presence of two populations of droplets: small drops as in the DNA-free microemulsion and larger drops.11-13,14a It was suggested that these drops could host two different types of condensed DNA.12 However, in these studies, a population of large drops was sometimes found for DNA-free microemulsions, probably as a result of artifacts in the Laplace transform of the light scattering data, thus questioning the validity of the conclusions in the case of DNA addition. Although electron microscopy studies of microemulsions are difficult to achieve (the structure changes very quickly with temperature, during most of the preparation methods), recent reports in ref 13 suggest that the large objects are toroids. The purpose of the present study is to investigate the effect of various parameters (size of the water pool, type of DNA, concentration of DNA) on the size and shape of the complexes present in the microemulsions using various scattering techniques in a more quantitative way than in previous studies.
II. Experimental Section II.1. Chemicals. A di-(2-ethyl-hexyl) sulfosuccinate sodium salt (Na-AOT) was purchased from Sigma-Aldrich (product number D1685). Calf thymus DNA was purchased from Sigma (product D-1501), λ DNA was purchased from Biolabs (product N3011S, 48 500 base pairs, molecular weigth 31.5 103 kDa), and sodium bromide (product number S4547) and isooctane were purchased from from Sigma-Aldrich (product number 360066). All chemicals were used as supplied, excepted calf thymus DNA (see below). Water was obtained from a Millipore Super-Q system. II.2. Procedure. Monodisperse DNA fragments (M-DNA, 147 base pairs, 95 kDa) were prepared by enzymatic digestion of calf thymus DNA according to the method described in ref 15. The monodispersity was checked by gel electrophoresis. The length of these fragments is the intrinsic persistence length, 50 nm, so they behave as rigid rods. Larger and less monodisperse fragments (SDNA) were prepared by sonication: Calf thymus DNA was dissolved in a 50 mM citrate buffer at pH 7.0 and sonicated at 4 °C for 10 h at low intensity, having first been degassed by a nitrogen sparge. DNA was centrifuged after sonication and phenol extraction performed to remove the residual proteins. The molecular weight was checked by electrophoresis and was 240 base pairs (156 kDa) with a polydispersity of 30 kDa (defined as the half-width of the Gaussian distribution found in the electrophoresis measurements). Both types of DNA solutions were dialyzed against 20 mM NaBr to keep the DNA helix intact. The concentration of DNA was determined by a spectroscopic method, using a molar extinction coefficient of 6600 L mol-1 cm-1 at 260 nm. An aqueous solution of DNA was injected into a Na-AOT (0.1 M) solution in isooctane in different proportions, characterized by the value of the molar ratio of water to surfactant: w ) [H2O]/ [AOT]. The mixture was stirred vigorously to form a single-phase microemulsion. In all of the samples studied by the different scattering techniques, the water volume fraction was 7.4%. Under the given experimental conditions, no phase separation was observed (excepted when noted), and single-phase microemulsions were obtained. The microemulsions studied had large water cores, and most measurements have been made with w ) 44.4. We also made measurements for w ) 22.2 and below. Note that above w > 50 in DNA-free systems the maximum radius is reached and phase separation occurs. The microemulsion droplets used here were therefore larger than in most previous work. It was observed that the solubilization of DNA in the microemulsion was facilitated when heat treated (sample heated at 40 °C for 10 min). However, for DNA concentrations in water of 1 mg/mL or higher, the droplet radius increases as a function of time for 10 days. This unusual time evolution in microemulsion systems could be due to the fact that the DNA concentration in the drops’ interiors (15) Sikorav, J. L.; Pelta, J.; Livolant, F Biophys. J. 1994, 67, 1387.
Langmuir, Vol. 24, No. 20, 2008 11829 is quite high, as we will see later; therefore, the core is quite viscous. No evolution with time was observed at lower DNA concentration or if the sample was not heat treated. In the case of microemulsions with smaller droplets (w ) 22.2), no time effect was observed irrespective of the heat treatment. II.3. Methods. II.3.1. UV-Vis Absorption. UV-vis absorption spectra were measured with a Hewlett-Packard 8453 diode array spectrophotometer using 1 cm path-length quartz cuvettes. Pure isooctane was used for baseline correction. II.3.2. Dynamic and Static Light Scattering Measurements. The light scattering experiments were performed with a home-built instrument, of classical design, operating with a He-Ne laser (Melles Griot 75 mW, wavelength λ ) 632.8 nm). The detector is a Hamamatsu H7421-40 photon-counting head connected to a universal counter (Racal-Dana 1991). The correlator is a Flex2k-12 × 2 from correlator.com. The scattering-angle range is 20° < θ < 140°, corresponding to a wave vector range of 0.45 × 107 < q < 2.5 × 107 m-1 (q ) 4π n sin(θ/2)/λ), with n being the solution refractive index). The intensity of the light scattered by solutions of particles of volume Vpart and sufficiently diluted so that interactions between them are negligible can be expressed through the Rayleigh ratio, Rsample as
Rsample(q) )
4π2n2 (npart- nsolv)2 φ P(q)Vpart λ4
(1)
P(q) denotes the particles form factor, φ represents their volume fraction, and npart and nsolv are the refractive indices of the particles and the solvent, respectively (the difference is equal to the refractive index increment dn/dφ).16 When the radius of gyration Rg of the particles is small (qRg < 1),
P(q) ≈ exp -
( ) q2R2g 3
(2)
Toluene with a Rayleigh ratio of Rtoluene ) 1.406 × 10-7 m-1 was used for calibration. Then we may write Rsample(q) ) Rtoluene(n /n 2 toluene) Isample(q)/Itoluene, where ntoluene and Itoluene correspond to the toluene sample. In dynamic or quasi-elastic light scattering experiments, the normalized electric field correlation function g(1)(q, t) is measured. In the simplest case, for monodisperse particles the correlation function is given by
g(1)(q,t) ) exp(-Dq2t)
(3)
where D is the diffusion coefficient. Using the Stokes-Einstein equation, the hydrodynamic radius can be deduced from the diffusion coefficient D(q) extrapolated to q ) 0
D0 )
kT 6πηsRh
(4)
where k is the Boltzmann constant, T is the absolute temperature, and ηs the solvent viscosity.16 If radius R of the particles is small (qR, 1), D(q) is constant and equal to D0. When the particles are polydisperse, the data cannot be fitted with a single exponential, so stretched exponentials can be used instead:
g(1)(q,t) ) exp [-Dq2t]R
(5)
In the case of spheres, a polydispersity p defined as p ) 〈(R 〈R〉)2〉1/2/〈R〉. It can be easily checked that an R index larger than 0.9 corresponds to polydispersities p of less than 30%. II.3.3. Neutron and X-ray Scattering Measurements. The SANS experiments were performed at the Laboratoire Le´on Brillouin using the PAXE spectrometer. This offers a q range of 0.03 < q < 4 nm-1. SAXS experiments were performed with a SAXSess camera (AntonPaar) at the University of Graz (Institut Fu¨r Chemie, Austria). This camera is connected to an X-ray generator (Philips, PW 1730/10) (16) Pecora, R., Ed. Dynamic Light Scattering: Applications of Photon Correlation Spectroscopy; Plenum Press: New York, 1985.
11830 Langmuir, Vol. 24, No. 20, 2008
Swami et al.
operating at 40 kV and 50 mA with a sealed-tube Cu anode. A Go¨bel mirror was used to convert the divergent polychromatic X-ray beam into a focused line-shaped beam of Cu KR radiation (λ ) 0.154 nm). The 2D scattering pattern was recorded by a PI-SCX fused fiber optic taper CCD Camera from Princeton Instruments (Roper Scientific Inc., Trenton, NJ) and integrated into the 1D scattering function I(q), with q being the scattering vector. The CCD detector that was used features a 2084 × 2084 array with 24 × 24 µm2 pixel size (chip size: 50 × 50 mm2). The CCD is operated at -30 °C with 10 °C water-assisted cooling to reduce the thermally generated charge. Cosmic ray correction and background subtraction were performed on the 2D image before further data processing. The temperature of the capillary and the metallic sample holder was controlled by a Peltier element. Intensities I were normalized to transmission and incident beams and to detector efficiency through the scattering of water, yielding
I(q) ) (Fpart-Fsolv)2φP(q)Vpart
(6)
where Fpart and Fsolv are the scattering-length densities of the particles and solvent, respectively (F is the sum, normalized by the molecular volume of the scattering entity, of the coherent scattering lengths in this entity. These lengths are given for each nucleus in the case of neutrons, and for X-rays, the total is equal to the number of electrons in the entity times the classical electron radius r0 ) e2/ (4πε0mc2).17 In the case of monodisperse spheres of radius R,
P(q) )
[
3[sin(qR) - qRcos(qR)] (qR)3
]
2
Data were evaluated using generalized indirect Fourier transformation (GIFT) with polydisperse hard spheres. II.3.4. Surface Tension. Surface tension measurements were made with the bubble shape method using a commercial instrument (Tracker, ITConcept). The surface tension was measured several minutes after bubble creation, when the tension reaches a constant value. II.3.5. Electrical ConductiVity. Electrical conductivity measurements were made with a homemade setup. The solution was placed in a quartz cell and closed with a Teflon cap to avoid evaporation; the two electrodes were immersed to the same depth in the solution through two small holes on the cap, with a separation of 1 cm between them. A three-point calibration was carried out prior to the measurements, using electroconductivity standards. All of the above measurements were made at room temperature. In the following text, we will not specify the type of DNA used when the results are the same for the different DNAs.
III. Results III.1. UV-Vis Spectroscopy. The presence of DNA in the microemulsion was checked by UV spectroscopic measurements. Curve 2 in Figure 1 shows the spectrum of S-DNA in a 20 mM NaBr solution whereas curve 3 is the spectrum of DNA (of same concentration as in the case of curve 2) solubilized in the microemulsion. Curve 1 is the control experiment with a DNAfree microemulsion. The characteristic absorption band of DNA at 260 nm, as seen in curve 2, was also observed after solubilizing DNA in a microemulsion (curve 3). No absorption was detected when only buffer (20 mM NaBr) was solubilized in the microemulsion (curve 1). Thus, the UV measurements clearly indicate the presence of DNA inside the microemulsion. Furthermore, the absence of an intensity change (adding curve 1 for DNA-free microemulsions and curve 2 for oil-free DNA solutions leads to a curve very close to curve 3) indicates that DNA remains in a double-helix (17) Lindner, P., Zemb, T., Eds. Neutrons, X-rays and Light: Scattering Methods Applied to Soft Condensed Matter; Elsevier: Amsterdam, 2002.
Figure 1. UV spectra of the DNA-free microemulsion (curve 1); S-DNA in an aqueous solution (curve 2) and solubilized in the microemulsion (w) 44.4) (curve 3). The absorbance of the aqueous solution and the microemulsion has been scaled to correspond to the same DNA concentration.
form (single-stranded DNA adsorbs significantly more at 260 nm). Because DNA is very soluble in water, we assumed that when we solubilize DNA in a microemulsion, the DNA goes into the water pool of the microemulsion. To know whether the presence of DNA modifies the shape and/or size of the water pool, further studies by light, X-ray, and neutron scattering techniques were performed. III.2. Light Scattering Measurements. Before incorporating the DNA, most microemulsions had large water cores (w ) 44.4 and 22.2). Note that the scattering intensity is almost zero around w ≈ 30 as a result of the opposite contributions of surfactant and water to dn/dφ.18 The microemulsion droplets used here were then larger than in most previous work: R ≈ 10 nm for w ) 44.4. However, R increases in the presence of DNA to values similar to those reported before R ≈ 100 nm.11-14 In the case of microemulsions with w ) 22.2, the DNA-free microemulsion water cores are smaller, R ≈ 5 nm, but the droplets containing DNA are larger (R > 100 nm) and much more polydisperse. Measurements were also made for w ) 10 and 5, where similar large polydisperse drops were obtained. When the DNA concentration in the aqueous phase is very low (less than 0.1 mg/mL), the droplets are also large and polydisperse, even for w ) 44.4. The samples for concentrations above 0.5 mg/mL λ-DNA in water were turbid, and no measurements could be performed above this concentration for this type of DNA. Table 1 shows the effect on hydrodynamic radii and radii of gyration of the variation of DNA chain length for w ) 44.4. The Table also shows the effect of concentration of DNA. Observations similar to those in previous work are made by increasing the concentrations of S-DNA and M-DNA: the size gradually increases with the concentration until reaching 0.77 mg/mL, whereas no significant effect on size was observed above this concentration. Note that the droplet polydispersity is large, especially for small DNA concentrations, and the radii given are therefore averages over rather broad distributions: their ratio becomes meaningless and is not given in Table 1 when R e 0.8. In previous work with smaller drops, the maximum size was reached at small concentrations.11,13,14 No change was observed when the water content w was varied, confirming that the final (18) Zulauf, M.; Eicke, H. F. J. Phys. Chem. 1979, 83, 480.
DNA in Water-in-Oil Microemulsions
Langmuir, Vol. 24, No. 20, 2008 11831
Table 1. Hydrodynamic Radii, Radii of Gyration, and Scattering Intensities Measured for Different Samples Heated at 40 °C for 10 min Prior to Measurementsa
sample no DNA λ DNA S DNA
M DNA
DNA concentration mg/mL
Rh (nm)
R
0 0.19 0.38 0.5 0.19 0.38 0.77 1 1.55 3.1 0.19 0.38 0.77 1 1.37
10 120 110 120 25 36 47 120 95 88 29 44 67 98 70
1 0.7 0.83 0.9 0.8 0.8 0.9 0.96 0.9 1 0.77 0.85 0.83 0.92 0.95
Rg (nm)
I(q ) 0)/ Itoluene
90 69 69 40 43 42 70 51 51 40 42 46 60 45
11 100 575 700 8 75 430 620 550 800 5 75 325 330 300
Rh/Rg
1.6 1.7 1.1 1.7 1.8 1.7 1.05 1.45 1.6 1.55
a The measurements were made at 20°C, 10 days after the sample preparation for the samples containing g1 mg/mL DNA. The exponent R of the fit with a stretched exponential is also given, as is Rh/Rg.
size is independent of the initial one. In the case of λ-DNA, the maximum size is already reached at very small DNA concentration. Note that the intensity ratio between large and small drops is much higher than in former work, where the small drops could be detected by light scattering.11-13,14a This is perhaps due to the fact that here we mostly use larger w values than in former work (closer to the intensity minimum for the small drops). At size saturation, the ratio Rh/Rg is compatible with the value for spheres, 1.29; the R index approaches values of the order of E0.9, typical of polydispersities of
