Opto-perforation is an interesting alternative to conventional techniques for gene transfer into living cells. The cell membrane is perforated by femtosecond (fs) laser pulses, in order to induce an uptake of macromolecules e.g. DNA. In this study, we successfully transfected a canine cell line (MTH53a) with GFP vector or a vector coding for a GFP-HMGB1 fusion protein. The transfected cells were observed 48 hours after treatment and they were not showing any signs of apoptosis or necrosis. Based on simultaneously measured membrane potential changes during the perforation, we were able to calculate and experimentally verify that the relative volume exchanged is 0.4 times the total cell volume. Thus, for first time a quantitative predication of the amount of uptaken molecules and therefore a quantification of the transfection is possible. Additionally, this method offers new high efficient possibilities for critical transfection approaches involving special cell types, e.g. primary and stem cells.
©2008 Optical Society of America
DNA transfection is a central tool in cellular biology to study the effect of induced gene expression or negative gene regulation. For example gene transfer coding green fluorescent protein (GFP) fusion proteins [1, 2], allows to characterize the in vivo behavior and localization of these proteins, providing a central technique to understand the respective protein biology. Several methods were developed to introduce membrane impermeable molecules into the cells. Common methods are the use of viral vectors, chemical carriers and electroporation  to permeabilize the membrane so that the molecules can diffuse into the cell. However, the critical aspect in cell transfection remains the efficiency achieved, toxicity, and reproducibility depending on the characteristics of the used cells. Accordingly to this, primary cells are difficult to transfect using chemical reagents [4, 5]. Usage of electroporation might as well be limited because high electrical fields can induce irreversible cell damage to sensitive cells as primary cells and stem cells due to their limited cell population .
A novel alternative method avoiding the described problematics is the perforation of the membrane using fs-laser pulses [7, 8]. Whereas the whole membrane is perforated by electroporation, the fs-laser pulses are focused on a small region of the membrane less than one micrometer  in diameter. Due to the shortness of the laser pulses, almost no heating of the irradiated volume occurs as the applied pulse duration is shorter than the thermal conduction time. The manipulation induced by these pulses is limited to the focal volume, because the effect is based on multiphoton absorption  and therefore relies on very high photon densities. Thus, the perforation of the cell membrane by the fs-laser pulses does not damage the whole cell, affecting only a small volume of some femtoliters . The opto-perforation technique allows a “single cell targeting” and hence provides a key advantage in respect to selectivity, when compared to standard transfection.
However, the optimum parameters in terms of viability of the cells after treatment, efficiency and reproducibility are still not well known and especially the basic mechanism of the process is not yet well understood. For example, at which time pores are created, for how long they remain open and which volume is exchanged during the perforation.
In this study we combined this method with the patch-clamp technique on GFSHR-17 granulosa cells to get more insight into the mechanisms of opto-perforation. The advantage of these cells was that we have determined the parameters, which could affect the membrane potential under whole-cell configuration, in previous experiments . Thus the measurement of changes of the membrane potential allows to estimate the volume exchanged between the extracellular and the intracellular space during perforation relative to the cell volume (dilution factor) and gives an idea of the maximal life time of the induced transient pore. Additionally, the estimation of the exchanged volume could be used to calculate the concentration of the internalized material e.g. DNA during perforation.
As proof of principle we opto-perforated canine mammary cells (MTH53a) and transfected them with a GFP vector or a vector coding for a GFP fusion protein with the architectural transcription factor HMGB1 (GFP-HMGB1). Since previous experiments with MTH53a cells have shown that these cells are suitable for transfection and expression of GFP-HMGB1 (), they represent an advantageous model to study and compare transfection by opto-perforation with classical methods. The fs-laser based transfection resulted in either completely (GFP) or nucleus specific (GFP-HMGB1) fluorescing cells demonstrating the ability of the transfected cells to synthesis and process recombinant proteins.
2. Materials and methods
2.1. Laser system and microscope
The laser system used in this study is a tunable Ti:sapphire laser (Coherent, Chameleon) which generates ultrashort pulses of 140 fs at a repetition rate of 90 MHz. The accessible wavelength range is between 715 nm to 955 nm and the maximum pulse energy at 800 nm is 14 nJ. The pulse duration at the focus of the laser beam is about 210 fs due to dispersion in the optics, especially in the objective .
The laser beam is guided via a shutter (Thorlabs, SC10) and an attenuator to the microscope (Zeiss, Axiovert200) (Fig. 1). A 0.8-NA NIR water immersion objective (Zeiss, Achroplan) focuses the beam into the sample. The beam enters the tubus directly via a home built reflector cube without passing the UV lamp pathway, so that the fluorescence equipment can be used alternatively. A 0.8-NA NIR water immersion objective (Carl Zeiss AG, Achroplan) focuses the beam into the sample with a theoretical spot size of 600 nm at a central wavelength of 800 nm. The sample is placed in a chamber with a glass bottom having a thickness of 170 µm. Successful perforation of the cell membrane is observed by fluorescence visualized by a CCD-camera, concentration measurements were performed using an EM-CCD-camera (Andor Technology, iXon).
To allow the comparison of the different laser parameters as irradiation time and pulse energy, the irradiation time was regulated by a fast shutter between 30 ms and 60 ms with an accuracy of 1 ms and the pulse energy was changed by an attenuator consisting of a half waveplate and a polarizing beamsplitter cube. All experiments were realized at a central wavelength of 800 nm.
The imaging and manipulation program includes the manipulation tool for the nitrogen laser (Carl Zeiss AG, PALM), which was used to mark the manipulation area by cutting a rectangle into the glass cover slip. The microscope includes the fluorescence equipment, the UV lamp, the filters for propidium iodide (coupled to DNA) and for GFP, and the patch-clamp setup.
2.2. Patch-clamp studies
The membrane potential of the cells was measured by the patch-clamp technique  (Fig. 2) in current clamp modus using an amplifier (Axon Instruments, Axopatch-1D) and the computer interface (Instrutech Corporation).
The whole cell configuration was established on cells using a pipette-solution consisting of in mM: 100 K+-Gluconat, 40 KCl, 10 Hepes (4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid), 5 Na2ATP, 1 Glucose, 1 MgCl2, 5 EGTA (ethylene glycol tetraacetic acid), 0.25 cAMP(3’-5’-cyclic adenosine monophosphate), 0.5 cGMP (cyclic guanosine monophosphate), pH7.4 and an osmolarity 295±5. The patch electrode with the pipette-solution has a resistance of 10 MO.
During the patch-clamp measurements the cells were maintained in NaCl-media containing in mM: 121 NaCl, 5 KCl, 0.8 MgCl2, 1.8 CaCl2, 6 NaHCO3, 5.5 glucose, 25 HEPES, pH 7.4 and an osmolarity 295±5.
2.3. Cell cultures and labeling of granulosa cells
GFSHR-17 granulosa cells of rat were cultivated on cover slips using DMEM 8900 (Dulbecco’s Modified Eagle Medium) supplemented with 5% fetal calf serum (FCS) and the antibiotics penicillin, streptomycin and partricin. For laser-manipulation, a cover slip with cells was transferred in a perfusion chamber containing 0.5 ml of NaCl-media and 100 to 1000 µM of lucifer yellow (LY) or 1.5 µM of propidium iodide (PI).
2.4. Opto-perforation and viability control
PI was dissolved in the media before manipulation. The laser was focussed onto the membrane of the cell to induce a transient pore and an uptake of the solution with the dye molecules by diffusion. After treatment the cells were observed by fluorescence microscopy to verify the uptake of the dye and then washed with PBS and incubated in PBS for 90 minutes.
Then the viability of the treated cells was controlled by relabeling the sample with propidium iodide and comparing the fluorescence intensity before and after restaining. As the fluorescence intensity of the perforated cells is very low due to the small perforated area and the short interaction time of the laser, there should be an increase of fluorescence intensity in the cells with a damaged cell membrane which indicates the cells to be in a pathologic state.
2.5. Cell cultures of canine mammary cells and transfection by opto-perforation
Canine mammary cells MTH53a were cultivated in poly-L-lysine coated glass-bottom-dishes (MatTek) using M199 media (Gibco) supplemented with 20% FCS and penicillin and streptomycin. The demonstration of the principle of transfection using opto-perforationwas performed either with 50 µg/ml non-recombinant pEGFP-C1 vector plasmid (BD Bioscience Clontech) for total cell labeling or with 50 µg/ml recombinant pEGFP-C1-HMGB1 plasmid for nucleus specific labeling. The laser was focused onto the membrane to induce perforation. After treatment, the cells were washed with NaCl-media and incubated for 48 hours in the culture media. The uptake and expression of DNA was verified by fluorescence microscopy.
3.1. Membrane potential measurement during opto-perforation by fs-laser pulses
During opto-perforation by fs-laser pulses neither shrinkage nor swelling of the cells under the present conditions were observed indicating that the cell volume was not affected. However, perforation of the membrane allows a bi-directional flux of solution into and out of the cells (Fig. 2). As long as the pore is open a volume exchange between the intracellular and extracellular space takes place. In the literature, a volume of some femtoliters is predicted . For material e.g. DNA dissolved in the extracellular solution, the estimation of exchanged volume could give an indication of the concentration which can be achieved in the intracellular space after opto-perforation. Measurement of the change of membrane potential using the whole-cell configuration of the patch-clamp technique [14, 15, 16] allows to determine the exchanged volume in relation to the cell volume during opto-perforation. The verification of the uptake of molecules in these experiments was performed with PI and LY dissolved in the extracellular media. PI is only fluorescent when bound to DNA. Thus, it is not necessary to wash the sample for the fluorescence observation which could lead to the loss of the whole-cell configuration. LY offers the advantage that it could be also added to the patch pipette solution at a determined concentration. Measurement of the fluorescence of LY introduced into the cells after establishment of whole-cell with different concentration of LY in the pipette solution offered a calibration for fluorescence of LY introduced in the cells by opto-perforation. Thus allowing a verification of the theoretical calculated relative volume exchanged.
The resting membrane potential of granulosa cells was found to be about -45 mV. Fs-laser irradiation divides two regimes of depolarization depending on the focusing of the laser relative to the membrane, characterized by the formation or absence of a visible gas bubble. This gas bubble is due to chemical and accumulative thermal effects, theoretically described as low density plasma regime at MHz repetition rate and pulse energies below the optical breakdown .
In both cases the membrane potential increases about 2 to 5 mV (Fig. 3(A)). The observed depolarization time Δt in both cases is some milliseconds longer than the laser irradiation time t.
While the cases showing no visible formation of the gas bubble were characterized by a slow potential repolarization, in the second case showing the gas bubble the depolarization was followed by a second step of a strong depolarization of 10 to 20 mV (Fig. 3(A) and 3(B)). After the manipulation, the potential repolarizes slowly or stays on the new level some mV above the initial value. The depolarization continues some milliseconds after the laser irradiation, as in the first case, Δt>t. After that, the cell repolarizes in some cases but slowly compared to the depolarization. The cases without bubble formation where only the first step of depolarization was achieved, no fluorescence was detectable whereas in the case of bubble formation an uptake of chromophores could be detected. Thus, the induced gas bubble can be used as indicator of perforation of the cell membrane and successful uptake of dye molecules.
3.2. Determination of the exchanged volume during opto-perforation
According to the Nernst and Goldman equations  (for constant dye-concentration in the media and constant cell volume) the relative exchanged cell volume can be calculated by the following equation:
where V is the cell volume, α the exchanged volume, R the gas constant, T the absolute temperature, F the Faraday constant, Um the initial membrane potential and ΔUm the change of membrane potential.
Knowing the membrane potential during laser irradiation, the relative exchanged cell volume can be calculated. With the mean value of 10 mV membrane potential depolarization, the relative exchanged volume α/V is about 0.4. For the GFSHR-17 granulosa cells with a diameter of 10 µm and a volume of 500 femtoliters (fl), the exchanged volume is 200 fl. At 600 µM dye concentration in the extracellular media, the intracellular dye concentration should be about 240 µM after opto-perforation.
These quantitative considerations were verified by measuring the fluorescence intensity of fs-laser induced uptake of LY by granulosa cells. Compared to PI, LY offers the advantage of a fluorescence independent on the binding to any other molecules and is therefore an excellent dye for concentration measurements. For calibration, fluorescence intensities were determined by introducing a defined LY concentration into the cell via the patch-clamp-pipette in wholecell configuration, equilibration between the pipette-media and the cell was achieved after 15 minutes. The fluorescence can then be linked to the concentration as reference intensity.
The concentration of the LY in the pipette-media was chosen between 2 and 2000 µM. The fluorescence intensity increased linearly with increasing dye concentration (Fig. 4). The measured values fit well to the linear fit between 200 µM and 2 mM. At lower concentrations (2 and 10 µM) the fluorescence was to low to be clearly distinguished from the background.
We performed opto-perforation in presence of different concentrations of LY in the extracellular solution ([LY]o). Since we assumed a relative volume exchange of 0.4, a dilution of LY by a factor of 2.5 in the cell was expected. The mean fluorescence intensity of the cells after laser-manipulation and after washing with NaCl-media was measured for an extracellular concentration of 1 mM, 600, 200, and 100 µM at standard laser parameters, 40 ms irradiation time and 0.9 nJ pulse energy (30 to 50 cells per concentration).
At a LY concentration of 1 mM in the extracellular media the calculation after equation (2) results in an intracellular concentration of 400 a.u. after opto-perforation. Following the reference curve for the concentration (Fig. 3), we expect a fluorescence intensity of 186 a.u. after perforation. We observed a mean intensity of 155 a.u. which leads to an intracellular concentration ([LY]i) of 332 µM and a factor of exchanged media α/V of 0.33. For 600, 200, and 100 µM the corresponding factor of relative exchanged media α/V is 0.44, 0.35, and 0.37 respectively (table 3.2). The comparison of the theoretical and the measured values show that a quantitative estimation of the amount of material e.g. DNA which is taken up during opto-perforation is possible. This represents to our knowledge the first attempt to quantify this parameter.
3.3. Energy and irradiation time dependence of opto-perforation
The energy deposited into the cells has to be as low as possible to avoid damage while the efficiency of the uptake of molecules increases at longer irradiation times and higher pulse energies. The viability of the cells strongly depends on the used pulse energy, the irradiation time of the laser beam and the position of the focal point relatively to the membrane. The visualization of the viability was performed by restaining the cells with propidium iodide. In cells whose membrane is damaged, the fluorescence intensity is higher after the restaining as the fluorophore is impermeable to the membrane and it can not enter the cells if the physiologic structure is intact after the treatment. Thus the dye highlights the cells that are in a pathologic state (Fig. 5).
The parameters for the opto-perforation were chosen as follows. The central wavelength at 800 nm, a repetition rate of 90 MHz and pulse energies between 0.7 nJ and 1.1 nJ were used for the opto-perforation of the granulosa cells. The irradiation times were chosen between 30 ms and 60 ms (Fig. 6). The efficiency increased with the pulse energy and the irradiation time, while the viability observed 90 minutes after the treatment decreased. Every parameter combination was realized at 40 to 60 cells.
In the region of short irradiation time of 30 ms the viability of the cells is at up to 90% for pulse energies up to 0.9 nJ. For longer irradiation times as 60 ms and high pulse energy of 1.1 nJ the viability of the cells decreases to 40%.
In contrast, the efficiency of dye uptake is at 40% or lower for 30 ms irradiation time and pulse energies up to 1 nJ. The efficiency increases with increasing pulse energy and increasing irradiation time to 70% at 1.1 nJ and 60 ms (Fig. 6).
However, a good balance between viability and efficiency can be found at 40 ms irradiation time and thus 3.6 million pulses with an energy per pulse of 0.9 nJ. For these parameters, the viability is 90% and the efficiency 70%.
3.4. Transfection of MTH53a cells by fs-laser pulses with pEGFP-C1 and pEGFP-C1-HMGB1 vectors
The transfection of the canine mammary cells was performed at 40 ms irradiation time and thus 3.6 million pulses with an energy per pulse of 0.9 nJ, leading to a viability of the cells after treatment of about 90% and an efficiency of about 70%. The cells were treated either in presence of 50 µg/ml non-recombinant pEGFP-C1 vector or recombinant pEGFP-C1-HMGB1 vector in the culture media. The fluorescence was observed 24 and 48 h after treatment allowing an expression and processing of the respective recombinant proteins. The cells transfected with pEGFP-C1 vector showed a labeling of the complete cell (Fig. 7) by the synthesized recombinant GFP proteins. The cells transfected with the pEGFP-C1-HMGB1 vector showed a specific labeling of the nucleus (fig. 4B). These specific labeling shows that the cells are post transfected still able to synthesis the pEGFP-C1-HMGB1 fusion protein and to transport the chromatin associated architectural transcription factor HMGB1 to its native cellular localization in the nucleus.
4. Discussion and conclusion
Transfection of macromolecules represents a key technique in molecular biology. The methods currently applied e.g. viral vectors, chemical carriers, lipofection and electroporation face several critical problems in terms of the achieved efficiency, toxicity, and reproducibility. Although for multiple applications the systems used are sufficient, several experimental approaches require more sophisticated methods. Parameters like a defined amount of induced molecule intake, reduced cell damage and high transfection efficiency represent critical parameters when working with cell type requiring extremely high complexity handling e.g. stem cells. The used fs-laser based opto-perforation of cells allows a defined control of the described critical parameters. In detail the measurements of the cell membrane potential showed two different behaviors during opto-perforation.
As described previously the membrane potential increased in both cases instantaneously about 2 to 5 mV at all parameters. If no bubble was created fluorescence analysis using PI did not show an uptake of the fluorophore in contrast to the case of bubble formation showing a characteristic high PI uptake. The critical parameters for formation or non-formation of gas bubbles are combination of irradiation time, pulse energy, and position of the laser focus relatively to the membrane. It can therefore be assumed that without induced bubble formation no pore was formed in the membrane and the observed change in membrane potential is related to change of permeability of the membrane to ions which did not involve larger molecules such as chromophores. It is also possible that an inadequate focusing of the laser beam generated a very small pore which closed too quickly to allow an uptake of the molecules dissolved in the extracellular solution.
In the case of bubble formation, the potential depolarized first as described above and, as soon as the bubble was created, the potential increased by another 10 mV or more. After the manipulation, the membrane potential repolarized again or stayed at the same level but did not rise to 0 mV in most cases. These results show that there was an exchange of intra and extracellular media, especially when a bubble was induced which can be used as an indicator for successful opto-perforation. Additionally it can be concluded that the cell is still alive and it is able to repair the laser induced damage.
The maximum depolarization time (Δt=60 ms) is longer than the irradiation time (t=30 ms). Thus the induced pore seems to be open longer than the increasing time of the potential. By a change in membrane potential of 10 mV, the relative volume exchange between the cells and the extra cellular solution during opto-perforation was calculated to be about 0.4 times the cell volume. At extracellular LY concentration of 1000, 600, 200, and 100 µM, a relative volume exchangeα/V of respectively 0.33, 0.44, 0.35, and 0.37 were found. The theoretically estimated value is in perfect agreement with the average value experimentally obtained of about 0.37 (table 3.2), indicating that a quantitative estimation of the material taken up during optoperforation is possible.
The uptake of fluorochromes by perforation of the membrane by fs laser pulses was optimized in the MHz regime using propidium iodide as fluorescent molecule. In the range between 0.7 nJ and 1.1 nJ pulse energy and irradiation times between 25 ms and 60 ms, the viability of the cells varies between 20% and 90%. In the same manner, the efficiency varies between 90% and 10%. As a very satisfactory result the viability of the cells is about 80% at a pulse energy of 0.9 nJ and an irradiation time of 40 ms yielding an efficiency of succeed membrane perforation of about 70%.
At these parameters, we successfully transfected canine mammary cells with nonrecombinant pEGFP-C1 vector or recombinant pEGFP-C1-HMGB1 vector. This specific labeling shows that the fs-laser based transfection allows even to successfully transfect cells which are still able to synthesize and process recombinant proteins. For pEGFP-C1 or recombinant pEGFP-C1-HMGB1 vector, 100 to 150 cells were targeted. 48 h after opto-perforation the expression of either gene was observed in about 30% of the targeted. The experiments with PI have shown that transfer of material and cell survival after opto-perforation is achieved in 70% of the targeted cells (Fig. 6). The observed difference may be related to different cellular behavior with respect to the introduced DNA. Further experiments should clarify this issue. The possibility of quantification could be used to calculate the number of DNA molecules taken up during opto-perforation. At an extracellular concentration of the pEGFP-HMGB1 vector of 50 µg/ml, about 10 fg DNA molecules enter into the perforated cells during laser irradiation at a cell diameter of 10 µm.
The system established allows a controlled transfection including the possibility to determine the amount of uptake of molecules and thus allowing to regulate the amount of effector molecules transfected into the cell. Consequently this allows a gentle procedure for e.g. stem cell transfection opening new possibilities for stem cell based experimental and therapeutic approaches. Finally, the presented technique could also be applied for cell specific transfection in tissue (in situ) where various types of cells are present.
We would like to acknowledge Prof. Bernhard Keller (Georg-August University, Goettingen) for placing the patch-clamp setup to our disposal and BFI optilas for providing the EM-CCD camera. Parts of this work were funded by the German Research Foundation (DFG) (within the Transregio 37 and the excellence cluster REBIRTH).
References and links
3. E. Tekle, R.D. Astumian, and P.B. Chock, “Electroporation by Using Bipolar Oscillating Electric Field: An Improved Method for DNA Transfection of NIH 3T3 Cells,” Proc. Natl. Acad. Sci. USA 88, 4230–4234 (1991) [CrossRef] [PubMed]
4. J.L. Stilwell, D.M. McCarty, A. Negishi, R. Superfine, and R.J. Samulski, “Development anf Characterization of Novel Empty Adenovirus Capsids and Teir Impact on Cellular Gene Expression,” J. Virol. 12881–12885 (2003) [CrossRef] [PubMed]
5. W.S. Wold, K. Doronin, K. Toth, M. Kuppuswamy, D.L. Lichtenstein, and A.E. Tollefson, “Immune responses to adenoviruses: viral evasion mechanisms and their implications for the clinic,” Curr. Opin. Immunol. 11, 380–386 (1999) [CrossRef] [PubMed]
6. P. Menendez, L. Wang, and M. Bhatia, “Genetic Manipulation of Human Embryonic stemm cells: A System to Study Early Human Development and Potential Therapeutic Applications,” Curr. Gene Ther. 5, 375–385 (2005) [CrossRef] [PubMed]
8. D. Stevenson, B. Agate, X. Tsampoula, P. Fischer, C.T.A. Brown, W. Sibbett, A. Riches, F. Gunn-Moore, and K. Dholakia, “Femtosecond optical transfection of cells: viability and efficiency,” Opt. Express 14, 7125–7133 (2006) [CrossRef] [PubMed]
9. V. Kohli, J.P. Acker, and A.Y. Elezzabi, “Reversible permeabilization using high-intensity femtosecond laser pulses: applications to biopreservation,” Biotechnol. Bioeng. 92, 7 (2005) [CrossRef]
10. A. Vogel, J. Noack, G. Huettman, and G. Paltauf, “Mechanisms of femtosecond laser nanosurgery of cells and tissues,” Appl. Phys. B 81, 1015–1047 (2005) [CrossRef]
11. A. Ngezahayo, B. Altmann, and H.-A. Kolb, “Regulation of Ion Fluxes, Cell Volume and Gap Junctional Coupling by cGMP in GFSHR-17 Granulosa Cells,” J. Membrane Biol. 194, 165–176 (2003) [CrossRef]
12. M. Bustin, “Regulation of DNA-Dependent Activities by the Functional Motifs of the High-Mobility-Group Chromosomal Proteins,” Mol. Cell. Biol. 19, 5237–5246 (1999) [PubMed]
13. K. König, “Multophoton Microscopy in Life Sciences,” J. Microscopy 200 , 2, 83–104 (2000) [CrossRef]
14. O.P. Hamill, A. Marty, E. Neher, B. Sakmann, and F.J. Sigworth, “Improved Patch-Clamp Techniques for High-Resolution Current Recording from Cells and Cell-Free Membrane Patches,” Pflügers Arch. 391, 85–100 (1981) [CrossRef] [PubMed]
16. M. Numberger and A. Draguhn, Patch-Clamp-Technik (Spektrum Akademischer Verlag, 1996)