2.1.

Optical Setup

The schematic diagram of our optical setup is illustrated in Fig. 1. An ultrafast laser (Tsunami) was used with a wavelength of 780 nm. To study the effect of laser polarization on TEF, a polarization converter (Arcoptix S.A) was set to switch between radial polarization (RP) and azimuthal polarization (AP). The intensity of the laser was controlled by a half-wave plate and a linear polarizer. The illumination power was measured by a power meter in front of the polarization converter. To collect the emission, a $60\times$ water-immersion condenser was used as an objective. The emitted fluorescence passed through a band-pass filter of $520\pm 10\mathrm{nm}$ before being detected by the photomultiplier. The focal spot images of polarized beam were displayed in Fig. 1, which showed that the RP beam produced a sharper focus than linear polarization (LP) and AP. A three-dimensional scan was performed before each experiment to verify the apex position. All images were at a $49\mathrm{nm}/\text{pixel}$ scale with an integration time of $0.1\mathrm{ms}/\text{pixel}$.

Fig. 1

Illustration of our two-photon microscopy. The samples were scanned with three differently polarized laser beams, respectively. The upper-left image indicates the sharper focus with radially polarized illumination than the other two polarized states. The profile of the images was also affected by the shape of apexes.

2.2.

Loop-Mediated Isothermal Amplification Hepatitis C Virus cDNA

The amplification process of LAMP consisted of two steps: first, a dumbbell-like new template was formed with the help of forward/backward inner primers (FIP/BIP) and forward/backward primers (FP/BP) (Fig. 2). Second, the loop structures in the new template could be the primer of the template itself for elongating a new strand. The newly elongated amplicon then formed a flower-like DNA by self-annealing. The added forward/backward loop primers (LF/LB) were used to accelerate the elongation process. The primer sequences for HCV LAMP were listed in Table 1.

Fig. 2

Dumbbell-like new template was formed in the first step of loop-mediated isothermal amplification (LAMP). (FIP: forward inner primer, BIP: backward inner primer, FP: forward primer, BP: backward primer, LF: forward loop primer, LB: backward loop primer).

Table 1

The primer sequences for loop-mediated isothermal amplification (LAMP) on Hepatitis C virus (HCV).

The mixture of reaction solutions was prepared for general LAMP and tip-LAMP. The premixture included 0.4 mM dNTPs, 0.4 M ${\mathrm{Mg}}_2{\mathrm{SO}}_4$, $1\times \mathrm{Bst}\text{buffer}$ (New England Biolabs Inc.), 1 M betaine (Sigma-Aldrich Co.), and the primer mixture. Mg ion was the cofactor for polymerases. The betaine could reduce the required energy for separating DNA double strands. The primer mixture mentioned here contained 10 pmol of FP/BP, 20 pmol of FIP/BIP, and 5 pmol of LF/LB. Additionally, the asymmetrical cyanine dye (SYBR® Green I) green dye was added as a reporter on newly forming double-stranded DNA. All materials above were dissolved in ultrapure water. For general LAMP, target DNA templates and 8 U of large fragment Bst polymerase (76 kDa, New England Biolabs Inc.) were added to the mixture and heated to 65°C for 1 h. The HCV cDNA templates used here were prepared with the pGEM-T easy vector system (PROMEGA). The subsequent processes of tip-LAMP are explained in Sec. 2.3.

2.3.

Preparation of Nano Tips for Loop-Mediated Isothermal Amplification (Tip-LAMP)

The apex size of Pt/Ir coated tips (NanoInk Inc.) was inspected via VHX-2000 microscopy (Keyence Co.) and scanning electron microscopy (SEM) before and after gold film deposition. The Au film was deposited onto the tips by sputtering [Fig. 3(a)-1] and verified with an ellipsometer (EP3, nanofilm) to confirm the thickness. Because the electric field distribution and excitation volume at the apex were affected by the thickness of the Au film and the sharpness of the apex,^16,17 all tips used in subsequent experiments were with a 7-nm-thick Au film, which corresponded to 55 to 57 nm in apex diameter. The gold covered tips were then immersed into 1 mM 16-mercaptohexadecanoic acid (MHA) in acetonitrile for 30 min to form a MHA self-assembled monolayer. MHA is a long carbon chain structure with sulfhydryl and carboxyl groups at its two ends, respectively. The sulfhydryl group could react with the Au film surface by thiol-metal bond (Au-S bond). After being washed with ethanol and ultrapure water, the MHA coated tips were soaked into 400 mM ethyldimethylaminopropyl carbodiimide and 100 mM N-hydroxysuccinimide (NHS) in ultrapure water for 5 min to activate the carboxyl group on MHA [Fig. 3(a)-2]. For polymerase coating, Bst polymerase was dropped onto a cellulose membrane. The chemically modified tips were dried by ${\mathrm{N}}_2$ and brought into contact with the cellulose membrane by the approaching process of dip-pen nanolithography platform (NanoInk) [Fig. 3(a)-3]. The carboxyl group of MHA would covalently bind to the amine group on polymerase. Finally, the polymerase functionalized tips were mixed with the reagent premixture mentioned in Sec. 2.2 and sealed inside a reaction chamber for LAMP reaction [Fig. 3(a)-4].

Fig. 3

Tip preparation. (a) Undamaged tips were functionalized with MHA/EDC/NHS and polymerase. Heating plates with a temperature sensor and feedback control was covered in both sides of the chamber. (b) The distance of Au-fluorophore was estimation to be 6.6 to 7.1 nm ($d=\text{thickness}$). (c) Scanning electron microscopy (SEM) image of the unmodified/modified tip.

Serious quenching occurs when the fluorophore is too close to the metal surface. Therefore, we estimated the distance between the tip and the fluorophore. By assuming that the shape of the polymerase was spherical, the diameter of Bst polymerase was calculated to be 5.6 nm.¹⁸ Since the length of the MHA linker was around 1.5 nm, the Au-fluorophore distance was estimated to be 6.6 to 7.1 nm [Fig. 3(b)], which was sufficient to seriously avoid quenching through ohmic loss. The SEM images of an unmodified and modified tip are shown in Fig. 3(c).

3.1.

Excitation Power and Excitation Polarization Effect on Tip-Enhanced Fluorescence

We first inspected the background signals from our system, tip, and fluorescence solution. The system background was measured in air without a tip. The result in Fig. 4(a) indicated that system background was not relevant to the excitation power. Meanwhile, the photon count, either of the tip apex in water or of the focal spot in FITC solution, was proportional to the excitation power [the upper-left plot in Fig. 4(a)]. Even though some of the background signals slightly increased with power, the TEF photon count in FTIC solution was much greater than the background signals. In addition, the photon count at apex was found to be higher than the photon count on a flat Au surface. In view of Fig. 4, we confirmed that a metallic tip provided the strongest enhancement, and the interference from the background signals could be neglected in our experiments.

Fig. 4

Tip strongly enhances the fluorescence. (a) Tip is with stronger emission than flat Au. Upper-left plot shows the weak background signal, which can be ignored in our experiments. (b) Effect of polarization on tip-enhanced fluorescence (TEF) signal intensity. ($n=5$). (c) The focal spot diameter is around 100 smaller with RP illumination than LP. TEF with RP illumination increased the emission intensity and made the emission spot clear to be seen.

Polarized illumination affects the intensity distribution of the electrical field at the tip. According to the computer simulation, RP illumination was thought to provide sharp longitudinal field distribution at the focal point whose size was approximately $0.4\lambda$ (full width at half maximum).¹⁹ The longitudinal electric field could excite a strong and highly concentrated optical field in tip-based near-field optics.²⁰ We experimentally inspected the TEF intensity under RP and AP illuminations. Figure 4(b) indicates that the enhancement was proportional to the excitation power. In addition, an RP excitation beam provided stronger enhancement than AP in our system. Apart from this, the enhancement with RP illumination reduced the diameter of the excitation spot and the minimum requirement of the excitation intensity. Figure 4(c) displays the diameter differences of excitation spots under RP or LP illumination. We took images of the apex in FITC solution such as the upper-left two photon images shown in Fig. 1. The RP illumination reduced around 100 nm in spot diameter compared to LP. Furthermore, Fig. 4(c) shows that RP illumination was of benefit to the apex imaging with an excitation power lower than 50 mW.

The enhancement ratio (ER) of the tip in Fig. 4(a) was calculated and plotted in Fig. 5. The ER here is defined as the value of the brightest pixel in the apex image divided by the average count of a random acquired 30-pixel-square area out of the apex in the solution [Fig. 5(a)]. The ER of the tip apex in the FITC solution with RP illumination was shown in Fig 5(b). The ER of TEF was proportional to the incident laser intensity. At an excitation power of 50 mW, the ER could be five on average; at an excitation power of 95 mW, it could be around 30 on average. This rise in ER was inferred to be a result of the increase of excitation volume,²¹ and the curve should reach a plateau at a higher incident power.¹⁰ We chose a power of 50 mW for excitation in Sec. 3.2. It should be noted that the measured power of 50 mW before the polarization converter corresponded to an after-objective power of $160\mu \mathrm{W}$ in our optical system.

Fig. 5

(a) Two-photon image of the tip apex under radially polarized illumination. The enhancement ration was calculated followed the formula with red color. B is a 30-pixel square and A is a random selected rectangular region including the apex image. (b) The enhancement ratio (ER) of tip apex in fluorescein isothiocyanate (FITC) solution with radial polarized illumination.

3.2.

Tip Enhanced Fluorescence for Monitoring Loop-Mediated Isothermal Amplification

To perform the tip-LAMP, a premixture solution, polymerase functionalized tip, and $4\mathrm{ng}/\mu \mathrm{l}$ HCV cDNA templates were sealed in the reaction chamber. We heated the reaction chamber to 65°C for LAMP reaction. The two photon images of the apex were continuously captured during the reaction. The results in Fig. 6(a) indicate that the brightness of the apex images increased with the reaction time, which corresponded to the accumulation of amplicon-SYBR® Green I complex. We created histograms of the photon counts at different time points from Fig. 6(a) and divided the photon count values into three different intensity groups. The histograms shifted to the right with time [Fig. 6(b)]: before the start of the reaction, more than 90% of the pixels had an intensity of less than 1000 counts; after 70 min reaction, around 30% of the pixels had a count more than 1000, and about 25% of the pixels had a count more than 10,000.

Fig. 6

(a) Two photon images of this tip-LAMP in selected time points. (b) Grouped pixel intensity histogram of apex fluorescence images at different times during the tip-LAMP with $4\mathrm{ng}/\mu \mathrm{l}$ HCV cDNA.

To figure out the performance of tip-LAMP in quantification, tip-LAMP reactions at three different template concentrations were carried out: 4, 0.4, and $0.004\mathrm{ng}/\mu \mathrm{l}$. The premixture solution was separated into two parts as mentioned in Sec. 2.2 for general real-time LAMP and tip-LAMP. The general real-time LAMP was performed with a BIO-RAD CFX real-time PCR detection system. The results of general LAMP were shown in Fig. 7(a). In the meantime, the tip-LAMP was performed with the protocol mentioned earlier in this section. The apex images with the size of $10\times 10\text{pixels}$ were continuously acquired within the reaction time. The photon counting values in each apex image were averaged and plotted in Fig. 7(b) for all the time points. Compared to the calibration curve of general LAMP, the increasing phases of the tip-LAMP were about 10, 20, and 30 min earlier for the reactions with 4, 0.4, and $0.004\mathrm{ng}/\mu \mathrm{l}$ templates, respectively.

Fig. 7

(a) The calibration curve of general real-time LAMP performed by BIO-RAD CFX real-time PCR detection system. (b) Results of the tip-LAMP reaction with three different template concentrations.