2.1.

Violet Diode Photoporation System

The laser source was a commercially available $405\text{-}\mathrm{nm}$ diode laser (Toptica, IBEAM-405-1V1 with an $M^2<1.2$ ) with maximum power of $40\mathrm{mW}$ . The beam was magnified by the telescope consisting of lenses L1 and L2 (Fig. 1 ). A half-wave plate and a polarizing beamsplitter in tandem were used to attenuate the beam power. A dichroic mirror at $45\mathrm{deg}$ reflected the beam to the rear entrance pupil of a high-numerical-aperture (NA) water immersion, violet corrected microscope objective (Nikon Plan Apo; $\text{magnification}=63\times$ $\mathrm{NA}=1.20$ ) with measured transmission of $\sim 84\mathrm{\%}$ at $405\mathrm{nm}$ . The laser was focused to a diffraction-limited spot approximately $0.4\mu \mathrm{m}$ in diameter. The power at the sample plane was obtained by taking the power transmission measurements through the optics. The maximum laser power dosages for each cell were first characterized by empirically observing the cell for any signs of granulation, blebbing, or necrosis. A beam shutter (Newport, United Kingdom, model 845 HP-02) placed in front of the laser was used to control the exposure times. An exposure time of $1\mathrm{s}$ was used to observe the transfection efficiency for each power level employed. Bright-field illumination in Koehler configuration was used to illuminate the sample. The image plane and laser plane were made coincident by changing the positions of lenses L1 and L2 and observing the image and laser focus. An $xyz$ stage enabled us to vary the sample position. Finally, a color CCD camera (WATEC 250D) was used to capture the videos of the process.

Fig. 1

(a) Schematic of a photoporation system using a $405\text{-}\mathrm{nm}$ diode laser: L ( $\mathrm{L}1=50\mathrm{mm}$ and $\mathrm{L}2=100\mathrm{mm}$ ), plano-convex lenses; PBS, polarizing beamsplitter; M, mirrors; DM, dichroic mirror; and LWD, long-working-distance objective. The whole system is mounted on a $60\times 90\mathrm{cm}$ optical breadboard. (b) Schematic layout of a prepared sample in a petri dish showing the laser beam focused on a cell bathed in DNA and phenol red solution.

2.2.

Cell Culture Procedure

CHO-K1 cells and HEK293 cells were cultured in T25 flasks at $37°\mathrm{C}$ and 5% $\mathrm{C}{\mathrm{O}}_2$ in modified Eagle’s medium (Sigma, United Kingdom) with 10% fetal calf serum (GlobePharm, University of Surrey, United Kingdom), $20\mu \mathrm{g}∕\mathrm{ml}$ streptomycin (Sigma, United Kingdom), and $20\mu \mathrm{g}∕\mathrm{ml}$ penicillin (Sigma, United Kingdom).

A stable tetracycline inducible system, TRex willin-GFP-HEK, was created using a TRex inducible plasmid pcDNA4/TO/myc-his (Invitrogen, United Kingdom) modified to express willin-GFP, which was then transfected into stable HEK293 cells containing a plasmid expressing a tetracycline repressor, pcDNA6/TR. TRex willin-GFP-HEK cells were cultured in T25 flasks in the presence of Dulbeco modified Eagle’s medium (Sigma, United Kingdom) with 10% fetal calf serum (GlobePharm, University of Surrey, United Kingdom), $2\mathrm{mM}$ L-glutamate, $100\text{units}∕\mathrm{ml}$ penicillin and $100\text{units}∕\mathrm{ml}$ streptomycin (Sigma, United Kingdom). Stable cells were selected by the addition with $5\mu \mathrm{g}∕\mathrm{ml}$ blasticidin and $250\mu \mathrm{g}∕\mathrm{ml}$ zeocin. Willin-GFP expression was induced with $1\mu \mathrm{g}∕\mathrm{ml}$ tetracycline (Invitrogen, United Kingdom).

Cells were routinely passaged three times a week. CHO-K1 and HEK293 cells were seeded at a density of $2.4\times {10}^4\text{cells}∕\mathrm{ml}$ onto $35\text{-}\mathrm{mm}$ glass-bottomed culture grade dishes (World Precision Instruments) to achieve 40 to 50% confluency. The cells were incubated at $37°\mathrm{C}$ for $24\mathrm{h}$ to allow cell attachment to the bottom of the glass dishes. Meanwhile, TRex willin-GFP-HEK cells were plated $48\mathrm{h}$ prior to the experiment onto $35\text{-}\mathrm{mm}$ culture dishes coated with laminin (Invitrogen, United Kingdom) to improve cells' adherence on the dishes.

2.3.

Targeted DNA Phototransfection

For each phototransfection experiment, individual CHO-K1 and HEK293 cells were exposed to up to $3.4\mathrm{mW}$ of laser power for $1\mathrm{s}$ at the focus. Before exposure, the cell monolayer was washed twice with OptiMEM (Invitrogen, United Kingdom) and then bathed with $30\mu \mathrm{l}$ of solution containing $10\mu \mathrm{g}∕\mathrm{ml}$ plasmid encoding for Mito-DsRed (Clontech) and $42.2\mu \mathrm{M}$ of phenol red (Sigma) in OptiMEM. Phenol red is a dye commonly used with cell culture media to detect changes in pH. The absorption of the solution used in the photoporation experiments was measured and a molar coefficient cross section of $\sim 1.4\times {10}^4{\mathrm{cm}}^{-1}{\mathrm{M}}^{-1}$ at $405\mathrm{nm}$ was obtained.

A type 0, $22\text{-}\mathrm{mm}$ -diam coverslip was placed on top of the cell monolayer and a region of interest was marked in the center in order to identify the region of dosed cells. On average, 50 healthy looking cells per dish were irradiated on the plasma membrane in a region $1\text{to}3\mu \mathrm{m}$ away from (and not directly above) the nucleus. After dosage, the coverslip was removed with OptiMEM and phenol red solution, and then subsequently washed twice with the same medium. The cells were then further incubated in fresh medium for up to $72\mathrm{h}$ after photoporation. Control cells were prepared in the same manner but were not exposed to the laser. The photoporated cells were observed under a fluorescent microscope for expression of the Mito-DsRed gene. For each laser power and for each irradiation time, nine experiments were performed, irradiating 50 cells per dish in the process. Therefore in total, for $n=9$ experiments, $\sim 5000$ cells were treated, enabling us to accumulate reliable statistics of the process.

2.4.

Cell Viability

HEK293 cells were plated on $35\text{-}\mathrm{mm}$ glass-bottomed dishes and were prepared similarly to transfection experiments except no DNA was added. Cells were exposed to 2.1- and $3.4\text{-}\mathrm{mW}$ laser power at exposure times from $80\mathrm{ms}\text{to}5\mathrm{s}$ . After laser exposure, cells were returned in the incubator for an hour, after which $60\mu \mathrm{l}$ of 0.3% trypan blue was added. Trypan blue is an indicator of cell membrane barrier dysfunction. In the state of necrosis, the cell’s protective membrane is often compromised, leading to intake of extracellular material. Dead cells stained blue were counted and the viability was calculated by obtaining the percentile ratio of dead cells with irradiated cells. Each data point was an average of three dishes and 50 cells per dish were exposed to the laser.

2.5.

siRNA Chemical Transfection

Willin knockdown was performed using small interfering RNA (siRNA), to a final concentration of $5\mathrm{nM}$ , specifically targeting the protein (GACAGAGCAGCAAGAUACUAUUAUU, CACAGACUAUAUGUCGGAAACCAAA, GCCUCUAUAUGAAUCUGCAGCCUGU; Invitrogen) using Gene Eraser (Startagene) according to the manufacture’s instructions. Protein expression was analyzed by Western blotting, using anti-GFP (Santa-Cruz) and anti-actin (Sigma) as a loading control.

2.6.

Fluorescence Microscopy

All fluorescence microscopy was performed on a TE2000-E, Nikon microscope. Cells expressing Mito-DsRed were imaged using TRITC HYQ, Nikon Filter cube (excitation, $530\text{to}560\mathrm{nm}$ ; emission, $590\text{to}650\mathrm{nm}$ ). Meanwhile, cells induced and expressing willin-GFP were imaged using FITC HYQ Nikon filter cube (excitation, $460\text{to}500\mathrm{nm}$ ; emission, $510\text{to}560\mathrm{nm}$ ).

3.1.

DNA Phototransfection

Phototransfection experiments on CHO-K1 and HEK293 cells were performed using a violet diode laser. For the laser parameters used, and under bright-field imaging, no visible reaction from the cell was observed. Successful transient expression was achieved with the Mito-DsRed plasmid for both cell lines, as shown in Figs. 2 and 2 . The transfection efficiency after $72\mathrm{h}$ as a function of laser power, using a $1\text{-}\mathrm{s}$ exposure time for HEK293 cells, is shown in Fig. 2. Each power level is significantly different from the control group $(p<0.05)$ with the exception of $0.8\mathrm{mW}$ , as determined by a one-way analysis of variance (ANOVA) followed by Dunnett’s statistical test. Overall, a Poisson distribution curve was obtained where the start and tail of the plot were significantly different from each other, as determined from ANOVA followed by Fischer’s pairwise test $(p<0.05)$ . As indicated in Fig. 2 an enhancement of transfection efficiency was acquired with an increase in laser power from $0.8\text{to}1.3\mathrm{mW}$ with efficiencies of $3.4\pm 3.4$ to $28.6\pm 6.3\mathrm{\%}$ , respectively, typifying a power dependence on the probability of cell poration and subsequently transfection. The optimum laser power was found to range from $1.3\text{to}2.5\mathrm{mW}$ , which yielded transfection efficiencies between $28.1\pm 5.8$ and $36.6\pm 4.8\mathrm{\%}$ . Within this range of power levels, the efficiencies were not significantly different from each other but were significantly different from 0.8 and $3.4\mathrm{mW}$ $(p<0.05)$ . Additionally, as the mitochondria were tagged, it was possible to observe their streaming, which is an indicator of the overall health of the treated cells. In comparison, the highest transfection efficiency for CHO-K1 was at $23\pm 1.5\mathrm{\%}$ . As established previously,¹ the level of spontaneous transfection was very low and negligible in comparison to the transfection efficiencies we achieved. Spontaneous transfection is defined as cells expressing the protein without exposure to the laser irradiation.

Fig. 2

Fluorescence images of transfected (a) CHO-K1 and (b) HEK293 transfected with Mito-DsRed plasmid, (c) transfection efficiency of HEK293 cells as a function of laser power at the focus using a $1\text{-}\mathrm{s}$ exposure time, and (d) transfection efficiency of HEK293 cells as a function of laser exposure time at $2.1\mathrm{mW}$ . The error bars represent standard deviation, ( $n=9$ experiments of 50 dosed cells).

The dependence of transfection efficiency on laser exposure time was also studied for a laser power of $2.1\mathrm{mW}$ for HEK293 cells. Figure 2 shows the transfection efficiency as a function of irradiation time from $80\mathrm{ms}\text{to}2\mathrm{s}$ . An ANOVA followed by Dunnett’s statistical test showed that transfection efficiencies obtained from different exposure times were significantly different compared to the control $(p<0.05)$ . Though significant transfection efficiency can already be obtained at a $80\text{-}\mathrm{ms}$ exposure, 1.0 and $1.5\mathrm{s}$ appeared to be optimal for transfection at this laser power. Increasing the exposure time from $1.5\text{to}2\mathrm{s}$ resulted in the decline of the average transfection efficiency from $36.5\pm 6.6$ to $23\pm 4.7\mathrm{\%}$ , respectively, which may be a consequence of a decrease in viability (see in the following).

3.2.

Phototoxicity of Violet Diode Phototransfection

Careful assessment of the cell viability was performed by exposing the cells to the laser and monitoring any morphological changes in the cells. Cells exposed to $2.1\mathrm{mW}$ for $6\text{to}10\mathrm{s}$ immediately showed signs of necrosis. In a further study, the viability of the cells treated with the laser at shorter exposure times were measured for two laser powers, 2.1 and $3.4\mathrm{mW}$ , as shown in Fig. 3 (laser exposure ranged from $0.08\text{to}5\mathrm{s}$ ) in the presence of trypan blue.

Fig. 3

(a) Viability of HEK293 cells exposed to laser at 2.1 and $3.4\mathrm{mW}$ at the focus at varying exposure time. Error bars $\text{represent}\pm \text{standard}$ deviation ( $n=3$ experiments of 50 dosed cells). (b) Bright-field image of laser exposed HEK293 cells. Cells pointed to by the red arrows were irradiated with the laser and have taken up the trypan blue dye, a sign of cellular necrosis. (Color online only.)

In general, viability was observed to decrease with increasing laser exposure time. For a $1\text{-}\mathrm{s}$ exposure, viability was measured from $96.4\pm 1.2$ and $95.4\pm 1.2\mathrm{\%}$ for laser powers of 2.1 and $3.4\mathrm{mW}$ , respectively. It marginally decreased from $86.7\pm 1.2$ to $65.3\pm 4.2\mathrm{\%}$ at $2\mathrm{s}$ , respectively. This small increase in laser power from $2.1\text{to}3.4\mathrm{mW}$ resulted in a sixfold decline in viability at $3\text{-}\mathrm{s}$ exposure times from $72\pm 9.2$ to $11.3\pm 12.1\mathrm{\%}$ . It was found that viability exponentially decays with exposure time and laser power. LD50, defined as dosage entailing 50% cell viability occurs at an energy density of $\sim 6.3\mathrm{MJ}∕{\mathrm{cm}}^2$ .

Notably, with laser exposures beyond the therapeutic dosage for transfection, fluorescence was observed at the targeted site, an indication of single-photon absorption, which might lead to the melting of the plasma membrane. Often when this phenomenon occurred, the cell underwent blebbing, granulation, loss of intracellular material, and necrosis. Neighboring cells, however, were not affected.

3.3.

Cell-Specific Gene Knockdown Experiments

RNA interference (RNAi) is a technology developed by Fire ¹⁴ in 1998 that enables the knockdown in expression of a specific gene. siRNA interferes with new protein expression, resulting in silencing of the gene. siRNA is made from a length of 20 to 25 nucleotides that bind specifically to the messenger RNA (mRNA) of the protein of interest. This leads to the formation of a siRNA complex, which results in mRNA cleavage and its subsequent degradation.¹⁵ The applications of this widely used gene silencing technology include studying a gene’s function, but also the potential therapeutic modification of gene expression in human diseases. Therefore, for the reasons outlined in the introduction, we explored the possible use of violet diode lasers in cell-specific gene knockdown experiments.

TRex willin-GFP-HEK cells were induced with $1\mu \mathrm{g}∕\mathrm{ml}$ tetracycline to activate willin-GFP expression. After $24\mathrm{h}$ , $5\mathrm{nM}$ siRNA manufactured to recognize specifically the willin gene, was chemically transfected into these cells. Willin-GFP protein levels were detected 24, 48, and $72\mathrm{h}$ after siRNA treatment. Western blot analysis (Fig. 4 ) shows that willin-GFP expression is significantly reduced after $48\text{to}72\mathrm{h}$ of siRNA transfection. Actin was used as a loading control and expression levels remain unchanged throughout siRNA transfection. This, therefore, indicated that the siRNA used were specific and effective in decreasing willin expression.

Fig. 4

Western blot analysis showing reduction of willin-GFP expression after $48\mathrm{h}$ of $5\mathrm{nM}$ siRNA chemical transfection. TRex willin-GFP cells were induced with $1\mu \mathrm{g}∕\mathrm{ml}$ tetracycline to express willin-GFP $24\mathrm{h}$ prior to siRNA treatment. Western blots were probed with anti-GFP and anti-actin, with the latter used as a loading control.

Therefore, for the photoporation experiments, $5\mathrm{nM}$ stock of the siRNA duplexes with the Mito-DsRed-encoding plasmid was added to the transfection medium. Initially, $10\mu \mathrm{g}∕\mathrm{ml}$ of Mito-DsRed was also added to identify cells that had been successfully transfected. Cells were targeted using a $3.4\text{-}\mathrm{mW}$ laser power at the focus and a $1\text{-}\mathrm{s}$ exposure time. Control dishes included (1) cells with Mito-DsRed plasmid and the siRNA without laser treatment; (2) cells without the Mito-DsRed plasmid but with siRNA without laser treatment, and (3) cells without both the Mito-DsRed and siRNA with laser treatment. Expression of willin-GFP was then induced by the addition of tetracycline, and cells were monitored for fluorescence over the next $2\text{days}$ . For all control dishes, spontaneous DNA transfection or knockdown was not observed.

Figure 5 shows images of photoporated TRex willin-GFP-HEK cells in the presence of the willin specific siRNA and the Mito-DsRed plasmid. Figure 5 is a successfully transfected TRex willin-GFP-HEK cell, as shown by the expression of mitochondrial targeted red fluorescent protein captured using the TRITC HYQ, Nikon Filter cube. Figure 5 is the same field of view but under bright-field imaging. Figure 5 shows the same cells, but imaged now using a FITC HYQ Nikon filter cube. As indicated by the blue and red arrows are cells that exhibit clear knocked down of willin-GFP expression, as indicated by the absence of green fluorescence. One cell, indicated by the red arrow point, was also transfected with the Mito-DsRed encoded plasmid. The laser’s specificity of action is indicated by the fact that untreated cells have not been knockdowned or transfected with Mito-DsRed, as observed in all control dishes checked under a fluorescence microscope and confirmed by Western blot analysis (data not shown).

Fig. 5

Gene knockdown using a violet diode system. (a) A TRex-willin-GFP-HEK cell fluorescing red due to the expression of the Mito-DsRed and (b) under bright-field imaging; (c) fluorescence image of the same field of view using a FITC HYQ. The Red arrow points to a cell that has been cotransfected with Mito-DsRed and willin specific siRNA. Blue arrows point to cells that have been transfected with siRNA only. (Color online only.)

3.4.

Temperature Calculations at the Focal Volume

Based on previous reports using the cw violet laser for transfection, it was conjectured that a melting of the phospholipid bilayer occurs due to the direct absorption of the phenol red in the medium, causing localized thermal effects on the irradiated site.^{8, 10} To elucidate the mechanistic process of poration, an estimate of the temperature at the beam focus was calculated based on modeling the phenol red as a sphere of radius $r$ immersed in a nonabsorbing medium. Since the absorption and the consequential increase in temperature occur only in close proximity to the focus, one can assume that the laser energy is absorbed only by the phenol red sphere. The radius $r$ of the sphere can be made equivalent to the radius of the beam spot given by $r=0.61\lambda ∕\mathrm{NA}$ , where $\lambda =405\mathrm{nm}$ and $\mathrm{NA}=1.2$ .

Since most of the incident power of the laser goes through the sphere of radius $r$ , we can assume that the substitution of focused Gaussian beam with a plane wave of incident intensity $I=P∕\left(\pi r^2\right)\sim 1\times {10}^{10}\mathrm{W}∕{\mathrm{m}}^2$ , where $P$ is the power of the laser at the focal plane going through the geometrical cross section of the sphere, will not introduce any significant error to our estimate. The Mie scattering problem can be solved for the phenol red sphere surrounded by the nonabsorbing medium. The imaginary part of the index of refraction of the phenol red sphere was determined by the equation $\widehat{n}=n[1+i(\alpha \lambda ∕4\pi n)]$ , based on the measured absorption coefficient, $\alpha =0.6{\mathrm{cm}}^{-1}$ , noting that the change in the real part of the index of refraction $n$ is negligible. This yielded an expression for the index of refraction given by $\widehat{n}=1.33+i(1.93\times {10}^{-6})$ . The solution provides us the absorption cross section $\sigma$ of the phenol red sphere, which is $\sigma =2\times {10}^{-18}{\mathrm{m}}^2$ . Hence, the heat absorbed $Q$ by the phenol red can be obtained, where $Q=I\sigma \sim 3\times {10}^{-8}\mathrm{W}$ .

Since, the cw irradiation used for the experiments lasts of the order of milliseconds to seconds and the temperature change saturates immediately after several microseconds,¹² we modeled only the steady state temperature increase, $\Delta T$ which is given by $\Delta T=(1∕4\pi )(Q∕{\kappa }_{0}r)$ ,¹⁶ where ${\kappa }_0=0.6\mathrm{W}∕\mathrm{mK}$ is the thermal conductivity of water. We assumed the same thermal conductivity for phenol red. From this, the calculated $\Delta T$ is $0.02\mathrm{K}$ .