2.1.

Cell Culturing

All cell lines were purchased from the European collection of cell cultures (ECACC, distributor Sigma-Aldrich Poole, UK). E14g2a cells were donated. Cells were cultured in a $37°\mathrm{C}$ , 5% carbon dioxide $\left(\mathrm{C}{\mathrm{O}}_2\right)$ , and 85% humid incubator. Cells were routinely grown in T25 vented-top culture flasks (Nunc^™), subcultured twice weekly at a ratio 1:4 Chinese hamster ovary (CHO-K1), human neuroblastoma (SK-N-SH) and human embryonic kidney (HEK-293) cells were grown in minimum essential medium (MEM) (Invitrogen, United Kingdom) with 1% penicillin-streptomycin (PEST) (Sigma, United Kingdom) supplemented with 10% foetal bovine serum (FBS). The mouse/rat neuroblastoma x glioma hybrid (NG108-15) cells were grown in Dulbecco’s modified eagle’s medium (DMEM) (Invitrogen, United Kingdom) with 1% PEST and 10% FBS. Mouse embryonic stem cells (E14g2a) were grown in Knockout DMEM (KDMEM) (Invitrogen, United Kingdom) modified with 1% nonessential amino acids (Invitrogen, United Kingdom), 1% L-glutamine (Invitrogen, United Kingdom), 1% sodium pyruvate (Invitrogen, United Kingdom), 0.1% $\beta$ -mercaptoethanol (Sigma, United Kingdom), supplemented with the cytokine leukemia inhibitory factor (LIF) (Millipore, United Kingdom) and 10% FBS (Biosera, United Kingdom). They were grown in 0.2% gelatin (Sigma, United Kingdom) coated on the surface of T25 culture flasks.

To prevent detachment of the monolayer when working with NG108-15 and HEK-293 cell lines, the glass surface of the petri dishes was coated with a thin layer of $2\mu \mathrm{g}∕{\mathrm{cm}}^2$ laminin (Sigma, United Kingdom).

2.2.

DNA Preparation

Preparation of all DNA plasmids and a description of the plasmid pDsRed2-Mito has been previously reported.^{5, 6} A plasmid encoding the transcription factor Gata-6 gene (pCAGSIH-Gata-6) was donated.

2.3.

Optical Setup

The optical setup is shown in Fig. 1 . Femtosecond laser beam pulses emitted by a titanium sapphire laser ( $790\mathrm{nm}$ , $80\mathrm{MHz}$ , $200\mathrm{fs}$ , average power at the beam $\text{focus}=60\mathrm{mW}$ ) were magnified by a simple two-lens telescope to match the back aperture of the $60\times$ Nikon objective lens with a numerical aperture (NA) of 0.8. This created a diffraction limited spot of $\sim 1.1\mu \mathrm{m}$ diam. A mechanical shutter (Newport, United Kingdom, model no. 845HP-02) placed in the beam path regulated the time of beam exposure at the cell sample plane. The sample was placed on an xyz stage (Newport, United Kingdom) and illuminated using a Koehler arrangement. The cell sample was imaged by a long working distance objective onto a Watec color camera (WAT-250D) placed below the sample stage.

Fig. 1

Phototransfection setup for our studies. An infrared Gaussian beam (beam $\text{diameter}=1.5\mathrm{mm}$ ) is emitted by a titanium sapphire laser bounced off mirror (M1) and expanded via lenses L1 $(f=50\mathrm{mm})$ and L2 $(f=175\mathrm{mm})$ . A half-wave plate $(\lambda ∕2)$ and polarizing beamsplitter cube (PBS) were used to attenuate the power output, while mirrors (M2 and M3) served as a periscope. Mirror M4 reflected the beam onto the back aperture of a $60\times$ objective lens $(\mathrm{NA}=0.8)$ . A light-emitting diode (LED) provided sample lighting when arranged into Koehler illumination via passing through lenses (L5, L4, and L3) and two apertures. The sample imaging system consisted of a long working distance (LWD), $f=200\mathrm{mm}$ , $50\times$ objective lens $(\mathrm{NA}=0.55)$ , a tube lens (L6), and a charge-coupled device (CCD) camera. These were connected to the output computer by a data-capture card.

2.4.

Sample Preparation and Phototransfection

For phototransfection experiments, $\sim {10}^4\text{cells}$ in $2\mathrm{mL}$ of complete medium were seeded in $35\text{-}\mathrm{mm}$ -diam type-zero glass-bottomed petri dishes ( $23\mathrm{mm}$ $\mathrm{diam}=\text{glass}$ working area (World Precision Instruments, Stevenage, United Kingdom). These were incubated to subconfluence over $24\mathrm{h}$ in optimum growth conditions. Then, the monolayer was washed twice with $2\mathrm{mL}$ of OptiMEM (Invitrogen, United Kingdom) each time to remove the serum. Thereafter, the cells were covered in $60\mu \mathrm{L}$ of serum-free medium containing $10\mu \mathrm{g}∕\mathrm{mL}$ of pDsRed2-Mito plasmid DNA (pDNA). The sample chamber was then covered with a $22\text{-}\mathrm{mm}$ -diam type-1 coverslip (BDH, Poole, United Kingdom). Targeted phototransfection of individual cells was then performed via, first, imaging the beam at the bottom coverslip. Then, by carefully adjusting the translation stage in the axial direction while both the sample and the beam were imaged through a bright field, the beam was focused onto the plasma membrane of the cell. Laser irradiation per cell was thereafter administered by three shots of ultrashort duration while avoiding visual cellular response. Because the Rayleigh range of the beam was $\sim 5\mu \mathrm{m}$ , we estimate the cell membrane to be within one Rayleigh range for the creation of the transient holes necessary for successful phototransfection.

Therefore, the top surface of the cell’s plasma membrane was directly exposed to the pDNA and fs beam focus, the area where the multiphoton effect is confined¹⁷ (see Fig. 2 ). Alignment of the beam focus to the cell’s plasma membrane therefore promotes the generation of free electrons $\left({\mathrm{e}}^{-}\right)$ , which photochemically react with the membrane resulting in the induction of transient holes¹⁶ through which pDNA diffuses into the cytosol. Following irradiation, the DNA-containing medium was aspirated, the monolayer washed once with OptiMEM, covered in $2\mathrm{mL}$ condition medium, and incubated under optimum growth conditions for $48\mathrm{h}$ before live cell fluorescence analysis.

Fig. 2

Sample chamber illustrated. Irradiation of the top surface of adherent cells facilitated by targeted delivery of the infrared fs laser beam permitted transient opening of the plasma membrane and subsequent diffusion of surrounding plasmid DNA.

All experiments were performed in triplicate, and each experiment was repeated five times. In all cases, the transfection efficiency (measured in percent) was calculated according to Tsukakoshi ¹² and Stevenson ¹³ using the expression: $N_{\mathrm{cor}}=[(E∕D).100]∕X_{\mathrm{D}}$ , where $N_{\mathrm{cor}}$ is the population corrected transfection efficiency, $E$ is number of cells transiently expressing the pDNA after a suitable amount of time has passed, $D$ number of cells dosed on a given experiment, and $X_{\mathrm{D}}$ is the ratio of proliferation that has occurred by the dosed cells between dosing and the measurement of expression.

2.5.

Stem Cell Differentiation and Selection

E14g2a cells were seeded at concentration ${10}^6\text{cells}∕\mathrm{mL}$ in gelatin-coated glass-bottomed petri dishes and incubated overnight. To stimulate cell differentiation, these were phototransfected with $15\mu \mathrm{g}∕\mathrm{mL}$ of Gata-6 pDNA at $50\text{-}\mathrm{mW}$ average power levels at the beam focus and $40\text{-}\mathrm{ms}$ pulse duration. After laser treatment, the cells were rinsed with serum-free KDMEM and incubated in optimum growth conditions in complete medium and selecting with $200\mu \mathrm{g}∕\mathrm{ml}$ of hygromycin (Invitrogen, United Kingdom). The medium was changed every two days.

2.6.

Reverse Transcriptase Polymerase Chain Reaction (rtPCR)

Total RNA was isolated from differentiated embryonic stem (ES) cells by trizol reagent (Invitrogen, United Kingdom). For rtPCR analysis, cDNA was synthesized from $5\mu \mathrm{g}$ of total RNA, with an oligo-dT primer and transcriptor reverse transcriptase (rt) (Roche, United Kingdom). One-twentieth of the single-strand cDNA products were utilized for each polymerase chain reaction (PCR) amplification experiment. PCR was performed according to the manufacturer’s instructions using the ready-mix taq PCR reaction kit (Sigma, United Kingdom). Primer (Eurofins, United Kingdom) sets (Gata-4, Oct-4, and Nanog) described by Fujikura ¹⁸ were used at $10\mathrm{pmol}$ and $40\text{cycles}$ per reaction. Amplified cDNAs were run on a 1.5% agarose gel.

2.7.

Cell Cycle Arrest by Thymidine and Colcemid

CHO-K1 and HEK-293 cells were reversibly synchronized at the M and S phases of the cell-division cycle. Cells were plated in complete medium in T25 tissue culture flasks at concentration ${10}^6\text{cells}∕\mathrm{mL}$ and incubated in optimum growth conditions over $48\mathrm{h}$ prior to arresting. Next, a nonlethal metaphase arrest was induced by treating the cells with $10\mu \mathrm{g}∕\mathrm{mL}$ of colcemid (Invitrogen, United Kingdom), followed by a further incubation of $6\mathrm{h}$ . Thereafter, the cells were harvested using 0.25% trypsin (Sigma, United Kingdom), plated for phototransfection and left to recover for one cell cycle $(\sim 18\mathrm{h})$ .^{19, 20, 21} Addition of colcemid to the cells prevents the centrioles from organizing the microtubules, which are necessary for chromosome migration to the poles during cell division.²² Cells were synchronized at the S phase via a thymidine-1 (TdR) (Sigma, United Kingdom) double block. Briefly, they were treated with $2\mathrm{mM}$ of TdR for $18\mathrm{h}$ , released into complete medium for $9\mathrm{h}$ , and then incubated in medium containing $2\mathrm{mM}$ TdR for another $18\mathrm{h}$ . Afterward, the cells were again harvested with 0.25% trypsin, plated for phototransfection, and allowed to recover for one cell cycle $(\sim 18\mathrm{h})$ . Essentially, TdR is uptaken by cells and subsequently phosphorylated to thymidine-monophosphate and thymidine-triphosphate (TIP). The accumulation of TIP results in the feedback inhibition of other nucleoside triphosphates, particularly deoxycytidine-triphosphate (dCTP), which is critical for DNA synthesis.²³

3.1.

Phototransfection and Differing Cell Types Including Stem Cells

Our previous studies have focused on the commonly used cell line, the CHO-K1.⁵ Therefore, using the same optical parameters and optical setup, we established the transient transfection efficiency (see Table 1 ) of a range of differing mammalian cell types, including pluripotent stem cells (E14g2a) and neuroblastomas (SK-N-SH and NG108-15), that are difficult to transfect by conventional transfection technologies.²⁴

Table 1

Summary of the cell lines phototransfected with a gene-encoding red fluorescent protein (pDsRed2-Mito). Experiments were performed in triplicate with a minimum of 150 individual cells treated per experiment. The transfection efficiency values reported were corrected as previously described.

All cell types were successfully transfected, but of particular note, was the possibility of phototransfecting embryonic stem cell colonies. Indeed, the effect of the exposure of stem cells to transfection reagents and/or chemicals is important because they may alter their metabolic balance and influence their biochemistry. Figure 3 depicts successful phototransfection of nondifferentiated pluripotent E14g2a stem cells and also indicates that our system allows the selection of specific cells within a mass of cells.

Fig. 3

A $50\times$ Mitutoyo long working distance objective lens $(\mathrm{NA}=0.55)$ was used to capture image A displaying a brightfield view of the E14g2a cells. Surrounded within the red circle is the sample of interest (i.e., the nondifferentiated cell colony). $10\mu \mathrm{g}∕\mathrm{mL}$ of DsRed2-Mito pDNA was phototransfected into these cells at $50\mathrm{mW}$ (at the beam focus) and $40\mathrm{ms}$ . Then, $48\mathrm{h}$ later a $10\times$ Nikon objective lens $(\mathrm{NA}=0.25)$ was employed to capture images B (brightfield image) and C (fluorescent image) in the presence of a TRITC filter cube. (Color online only.)

The successful laser transfection of stem cells has been previously reported,²⁵ but crucially we explored the possibility of taking this technology an important step further. Previously, we and others have commonly transfected a fluorescent reporter gene into mammalian cells, but within the field of stem cell biology, it would be a significant step forward to use physiologically relevant genes. Indeed, Uchugonova and Konig²⁶ reported on an urgent requirement in stem cell research on technologies for noninvasive, marker-free observation of growth, proliferation, and stability of living stem cells under physiological conditions. Therefore, here we report the induced differentiation of mouse embryonic stem cells into a defined tissue by introducing a specific marker-free plasmid DNA via femtosecond laser pulses.

Mouse stem cells (E14g2a) can be differentiated into extraembryonic endoderm by the activation of the transcriptional factor Gata-6, which is considered to be one of the master regulator genes.¹⁸ Using our established parameters, we phototransfected the endoderm-associated transcription factor Gata-6 gene into E14g2a cells. These cells were further cultured as outlined in the methods, and as a result of the Gata-6–induced differentiation events, a morphological change in E14g2a cells $96\mathrm{h}$ post-treatment was observed. Figure 4 displays images of this result, showing an altered morphology of E14g2a colonies producing identifiable individual cells with spindle- and stellate-shaped morphology, which is characteristic of extraembryonic endoderm tissue. ¹⁸

Fig. 4

(a) and (b) (negative controls) are brightfield images of the E14g2a cell colonies $48\text{-}\mathrm{h}$ postroutine subculturing, growing in the presence of LIF. Obtained $96\mathrm{h}$ after phototransfection at $50\mathrm{mW}$ (at the beam focus), $40\mathrm{ms}$ with the Gata-6 vector. (c) and (d) are the morphologically altered E14g2a cells selected in the presence of LIF. All images were captured through a $10\times$ Nikon microscope objective lens $(\mathrm{NA}=0.25)$ .

To biochemically prove this morphological change, the differentiated cells were tested by rtPCR for the expression of key transcriptional factors. In Fig. 5 , the upregulation and/or downregulation of three different transcriptional factors, including Gata-4, Oct-4, and Nanog, is shown. Our results confirm that the stem cells have been differentiated as we can see: first, the amplification of Gata-4 (lane 2) in the transfected cells as compared to the undifferentiated cells (lane 5) and, second, downregulation of both Oct-4 and Nanog (lanes 3 and 4) in the differentiated cells as compared to the undifferentiated cells (lanes 6 and 7). This is in agreement with what has been previously observed for these transcription factors.¹⁸

Fig. 5

rt-PCR analysis of germ layer markers. Lane 1: Molecular weight marker: Hyperladder 1 molecular weight marker (Bioline), Lane 2: rtPCR product for Gata-4, Lane 3: rtPCR product for Oct-4, Lane 4: rtPCR product for Nanog gene transcripts in differentiated cells (Gata-6). Lane 5: rtPCR product for Gata-4, Lane 6: rtPCR product for Oct-4, Lane 7: rtPCR product for Nanog gene transcripts in undifferentiated cells.

3.2.

Culture Passage Number and Transfection Efficiency

Successful phototransfection of stem cells into recognizable tissue is in itself significant, but next we explored methods to enhance this technology even further. Generally, in all transfection experiments, the passage number of cell lines can affect not only the transfection efficiency, but also protein expression.²⁷ We therefore phototransfected a plasmid encoding a mitochondrially targeted red fluorescent protein (pDsRed2-Mito) into CHO-K1 cells of different culture passage number to establish its effect on phototransfection. As can be seen in Fig. 6 , from passage 19 to 30 then, the efficiency is between 40 and 60%, peaking at approximately passage 28, but then drops significantly once the cells are at passage 34. This confirmed that, as with other transfection techniques, the passage number of a cell line plays a significant role in establishing the efficiency of phototransfection.²⁷

Fig. 6

CHO-K1 cells phototransfected (at an average power at the beam $\text{focus}=60\mathrm{mW}$ and $40\mathrm{ms}$ time of exposure) illustrate the influence of culture passage number on the phototransfection efficiency. Error bars represent the standard error of the mean $(n=5$ experiments of 50 treated cells). Using analysis of variance (ANOVA) followed by Dunnett’s and Fisher’s tests: The asterisk means data sets are significantly different from each other.

3.3.

Cell Arresting and Phototransfection Efficiency

We next explored the role of the cell cycle in the phototransfection of mammalian cells. In Fig. 7 , is shown the effect of chemically inducing CHO-K1 and HEK-293 cells to be arrested at two different phases (M and S) of the cell cycle, thereby synchronizing the population of the cells. The results show that arresting cells at two different stages of the cell cycle can significantly enhance phototransfection efficiency, with those in the S-phase the most efficient. The reasons for this are dependent on the processes occurring in the different phases of the cell cycle. In the M phase, the nuclear membrane disappears and, as a result, gene transfection efficiency increases as plasmid DNA can easily access the nuclear machinery.^{28, 29} While in the S phase, this is the point in the cycle where DNA is synthesized; thus, plasmid DNA will be also copied and transcribed, increasing its overall copy number and thus subsequently allowing more transcription.

Fig. 7

Arresting CHO-K1 and HEK-293 at the M phase gave a higher phototransfection compared to cells treated while at random stages of the cycle. However, arresting at the S phase gave highly enhanced phototransfection efficiency; this was observed in both cell lines. The corrected values of transfection efficiency are presented. Error bars represent the standard error of the mean ( $n=5$ experiments of 50 dosed cells). Using ANOVA followed by Dunnett’s and Fisher’s tests: The asterisk means data sets are significantly different from each other.