Long-range two-photon scattering spectroscopy ruler for screening prostate cancer cells

Development of a long-range TPS ruler for the screening of prostate cancer cells with sensitivity of 5 cells per mL level is demonstrated.


Introduction
Since the introduction of Förster resonance energy transfer (FRET) by Stryer and Haugland in 1967, optical spectroscopy rulers using FRET have become an essential tool for scientists from different disciplines to address a wide range of biological activity. [1][2][3][4] Since the existing FRET spectroscopy ruler distance window is governed by dipole-dipole coupling mechanism, scientists face problems monitoring biological processes that occur at greater than 10 nm scale distance. [1][2][3][4] To overcome the distance limitation, in past few years scientists have been trying to develop nanomaterial based optical rulers. [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19] Using this approach, the current article reports for the rst time the development of a long range two-photon scattering (TPS) spectroscopy ruler, whose distance window is more than twice that of the 10 nm distance limit, as shown in Scheme 1.
When an intense laser light of frequency u is exposed to a sample, it generates scattered radiation with the frequency 2u, which is known as two-photon scattering or second harmonic (SH) light. [20][21][22][23][24][25][26][27] It is now well documented that when two plasmonic nanostructures are in close proximity, the electric eld in the hot spot is enhanced by several orders of magnitudes. [20][21][22][23][24][25][28][29][30][31] Since the two-photon scattering is directly proportional to the fourth power of the fundamental eld amplitude, theoretically a huge two-photon scattering enhancement is expected due to the formation of a nanoantenna structure. [20][21][22][23][24][25][28][29][30][31][32][33][34] The above fact motivated us to develop a long-range TPS ruler, where we have used the advantage of the strong plasmonic electric eld generated by gold nanoparticle antennas to demonstrate that the distance Scheme 1 Schematic representation illustrating the design of twophoton scattering spectroscopy ruler using gold nanoparticle and ds-DNA.
window for a TPS ruler can be tuned up to 25 nm. Our experimental nding on the distance dependent two-photon scattering are supported by nite difference time domain (FDTD) simulation investigation of the distance-dependent two-photon scattering properties. 21,[24][25][26]30,[32][33][34] In our calculation, we have included multipolar and nite size effects and these are very important for gold nano-antenna of size close to the wavelength of light. 21,[24][25][26]30 To understand the possible application of TPS ruler for biological screening, we have demonstrated the detection of prostate-specic membrane antigen (PSMA) (+) prostate cancer cells using our developed TPS ruler in solution phase. PSMA is highly expressed on the surface of prostate cancer, which is the second leading cancer in the male population. [35][36][37] PSMA is a prostate cancer-specic biomarker which has been correlated with aggressive disease and used for monitoring of therapy. [35][36][37] Since the ability to detect very low levels of PSMA may enable doctors to diagnose men with prostate cancer recurrence, it has the capability to reduce prostate cancer-related mortality and morbidity. During the last decade several articles have shown that two-photon scattering properties can be enhanced (10 6 to 10 9 ) in magnitude by using metal nanoparticle-surfaces, which is comparable to the very large enhancement of surfaceenhanced Raman scattering. [22][23][24][25][26]28,37 Here we show that an anti-PSMA aptamer based TPS ruler can be used for ultrasensitive detection of PSMA (+) LNCaP prostate cancer cells. For this purpose, we have developed a TPS ruler using the partial sequence of A9 aptamer, 35,36 which has the capability to target clinically relevant LNCaP prostate cancer. A9 aptamer is known to recognize and bind to PSMA in intracellular or cytoplasmic epitope. 35,36 Our experimental data with PSMA negative normal skin HaCaT cells, indicate that spectroscopy ruler based assay is highly selective and enables their distinction from non-targeted cancer cell lines.

Results and discussion
A long-range two-photon spectroscopy ruler from 2 to 30 nm range was designed using SH-modied ds-DNA as a coupling agent. We have followed an established technique as reported by Alivisatos et al. 12,13 Initially, thiolated ss-DNA molecules (HS-(CH 2 ) 6 -oligo) were conjugated to gold nanoparticles, followed by a passivation layer of thiolated polyethylene glycol. Aer that, hybridization was performed using a complementary sequence (oligo-(CH 2 ) 3 -SH) attached gold nanoparticle in 10 mM PBS solution containing 0.3 M NaCl. To minimize multiple tether formation during hybridization of the gold-ssDNA conjugates, the ssDNA particle ratio was limited to 25 : 1, as reported by Alivisatos et al. 12 In the next step, dimers were separated from unreacted monomers and larger assemblies by gel-electrophoresis as reported by Alivisatos et al. 12 Finally, dimers were isolated from the gel by electroelution. Aer that, the dimer structures were conrmed using dynamic light scattering (DLS) and transmission electron microscopic (TEM) techniques. Since it is known that TEM grid preparation can increase aggregation, 33,34 to conrm that the isolated assemblies were dimers, we have used dynamic light scattering (DLS) measurement and compared the data with transmission electron microscopic technique (TEM), as shown in Fig. 1D and Table 1. As shown in Fig. 1D, our experimentally measured DLS and TEM statistics data indicate that the gel-puried 20 bp dimer sample is highly enriched in dimers ($90%). Our data show that the monomers and higher agglomerates are very few. TEM statistics showed that there are an average of 950 particles. Measurement has been performed 6 times and the average values have been reported in this manuscript.
To design TPS rulers in the 2 to 30 nm range, we have used ds-DNA spacers of 4, 10, 20, 30, 40, 60 and 80 base pairs. In the design of spectroscopy rulers in the window from 3 to 30 nm Clegg et al. 18 have reported that duplex DNA lengths of less than 100 base pairs can be assumed as linear chains by a rigid-rod approximation. As a result, in our design we have estimated the "long range spectroscopy ruler" distance by taking into account the size of the 0.32 nm for each base pair, and total 1.6 nm for Au-S-(CH 2 ) n -DNA base distance in both sides of base pair. To nd out the validity of our assumption of the rigid-rod approximation, we have employed dynamic light scattering (DLS) measurement using a Malvern Zetasizer Nano instrument.
As shown in Table 1, experimental DLS data indicate that the gold nanoparticles have an average size of $25 AE 2 nm, which also can be seen clearly from the TEM data, as shown in Fig. 1. On the other hand, in nano-antenna, when two gold nanoparticles are separated by 20 base pairs (bp) dsDNA, the diameter changes to $59 AE 2 nm, which also matches quite well with the TEM data as shown in Fig. 1B. Similarly, as shown in Table  1, the end-to-end distance measured between gold nanoparticles separated by ds-DNAs of different lengths, and the DLS measurement data are quite close to our estimated distances. Measurement has been performed 6 times and average values have been reported in this manuscript. Fig. 1E and Table 2 report how the plasmon band for the gold nanoantenna varies with the separation distance. Our reported data indicate that with decreasing interparticle gap, a strong red-shi along with the increase in extinction efficiency are observed. When the gap reaches around 22 nm, the maxima of plasmon band remain unchanged from the monomer absorption band, which is at 528 nm. Observed distance dependent plasmon shi is mainly due to the fact that the magnitude of the plasmon shi for gold nanoantenna is highly dependent on the strength of the interparticle coupling, which in turn, depends on the proximity of the individual nanoparticle.
Two-photon scattering intensity from the monomer and nano-antenna were measured in solution using the hyper-Rayleigh scattering (HRS) technique. 22,23,27,28,[30][31][32][33][34] To measure the TPS intensity, a femtosecond Ti:sapphire laser beam with excitation wavelength of 760 nm was used. Details of the HRS experimental procedure we use in our laboratory have been reported before. 22,23,30 TEM and DLS experiment performed before and aer 5 min of laser exposure indicated no photothermal damage during HRS data collection time and as a result, the number of dimers remained constant. The twophoton scattering light was separated from its single-photon scattering counterpart using a 3 nm bandwidth interference lter and a monochromator. For angle-resolved two-photon scattering intensity measurement, the fundamental input beam at 760 nm light was linearly polarized, and a rotating half-wave plate was used to vary the input angle of polarization. The twophoton scattering intensity from gold nano-antenna can be expressed as: 22,23,27,28,[30][31][32][33][34] where G is a geometric factor, N w denote the number of water molecules and N NP are the number gold nano-antenna per unit volume. On the other hand, b w and b NP are the rst hyperpolarizabilities of a single water molecule and a single gold nano-antenna. Similarly, a u and a 2u are the absorption cross section at u and 2u, respectively. l is the path length and I u is the intensity of fundamental 760 nm light. Fig. 1F shows the spectrum from gold nano-antenna recorded for an excitation laser wavelength at 760 nm in the absence of 3 nm bandwidth interference lter. Our experimental data as shown in Fig. 1F clearly show a strong TPS spectral peak at 380 nm and a much weaker and broader two-photon excited luminescence peak with l max ¼ 480 nm. The reported result indicates that we can easily eliminate contribution from the two-photon uorescence in the recorded TPS signal. As shown in the Fig. 2A, the twophoton scattering intensities from gold nano-antenna with 3 nm separation exhibit a linear relationship with the square of  the powers of the fundamental incident light, which clearly implies that the signal is indeed due to the two-photon scattering process.
Our reported data clearly shows that TPS intensity is highly sensitive to small changes in the monomer separation distance and as a result, an optical ruler can be developed. Our results indicate that aer a 30 nm separation, the TPS intensity from gold nano-antenna became the same as for a monomer gold nanoparticle. It is quite remarkable to see that TPS intensity ratio is greater than one for nano-antenna when the monomer is separated by 25 nm, a distance more than twice that of 10 nm for well documented FRET ruler.
Next, to understand the origin of high TPS intensity for nanoantenna with respect to the monomer gold nanoparticle, angleresolved two-photon scattering measurement were performed using 760 nm incident light from a femtosecond Ti:sapphire laser. For angle-resolved TPS measurement, the fundamental beam at 760 nm was linearly polarized, whereas the input angle of polarization was selected using a rotating half-wave plate. It is now well documented that due to the symmetry consideration, two-photon scattering or the second-harmonic process is forbidden for cubic metals such as gold within the dipolar approximation. [22][23][24][25][26][27][28][29][30] On the other hand, since for plasmonic gold nanoparticle systems, surfaces are known to play a major role, two-photon scattering arises from imperfect metallic nanostructure surfaces. [22][23][24][25][26][27][28][29][30][32][33][34] Also in the case of nano-antenna, symmetry is broken at the interface. The two-photon scattering properties for metal nanoparticles that are small compared to the excitation wavelength of 760 nm (d ( l/10) can be explained usually in the framework of electric-dipole approximation. [22][23][24][25][26][27][28][29][30][32][33][34] As shown in Fig. 2B and D, our angle-resolved twophoton scattering measurement clearly shows that for the 25 nm monomer gold nanoparticle or nano-antenna with 22 nm separation, the two-photon scattering response is governed by the electric-dipolar contribution which arises due to the deviation of the particle shape from that of a perfect sphere. On the other hand, in case of nano-antenna with 3 nm monomer separation, since the size of the nano-antenna (25 nm + 3 nm + 25 nm) is comparable with one tenth of the wavelength of the incident light, contribution of higher-multipoles such as electric quadrupole becomes very important to the two-photon scattering spectra. To further characterize the nature of the coupled plasmon mode, a polarization analysis is performed by rotating the linear excitation polarization using a half-wave plate. In this experiment, two-photon scattering modes were selected to be polarized perpendicular to the scattering plane using an analyzer. Fig. 2C shows the polar plots for the twophoton scattering signal as a function of the angle of polarization. The observed asymmetry in the polarization line shape for the gold nano-antenna with 3 nm monomer separation, clearly evidences of the deciency of the electric-dipole approximation for this system. The above data shows clear indication of the presence of higher multipolar interactions in these nanostructures, as a result, we have observed four lobes oriented on ) with respect to monomer. Our distance dependence reported data clearly indicate that TPS spectroscopy ruler can be used within the 25 nm distance. FDTD simulations profile of distance dependent SH intensity from nano-antenna also has been included, which indicate that the theoretical fit matches very well with the experimental observation. the 45, 135, 225 and 315 axes. Very strong two-photon scattering intensity from the gold nano-antenna reported here can be due to several factors and these are, (i) electric dipole contribution due to the imperfect structure of each nanoparticle, (ii) dipole contribution due to the broken symmetry at the interface. (iii) Electric quadrupole contribution due to the size effects, (iv) enhancement of electric dipole and electric quadrupole contribution due to the enhancement of local electric elds via strong plasmon coupling. Another important factor that may cause very strong TPS intensity from gold nanoantenna is single photon resonance. 33,34 As reported in Table 2, our experimental data show that in gold nano-antenna with 3 nm separation, l PR is red shied from 528 to 580 nm. As a result, single-photon resonance enhancement contribution in TPS intensity will be much higher for gold nano-antenna with shorter distance separation than corresponding monomer.
To gain more insight into the effect of plasmon coupling on TPS intensity and ruler length, full-wave simulations were performed based on numerical solutions of Maxwell's equations, using the nite difference time domain (FDTD) simulation package. 21,[24][25][26]30,32 In these simulations, multipolar and nite size effects were included, which is very important for gold nano-antenna whose size is close to one tenth of the wavelength of fundamental excitation light. Recent reports shows that FDTD technique is suitable for plasmonic simulations and capable of nding electric E(t) and magnetic H(t) components of the electromagnetic coupling and also simulating the second harmonic intensity. 21,[24][25][26]30,32 Fig. 3 and Table 2 shows the simulated extinction spectra for nano-antenna with different spacing. From the theoretical study we found that plasmonic resonance occurs near 528 nm in the single nanoparticle, which matches quite well with experimental observation. In case of nano-antenna, plasmonic resonance closely relies on the gap distance, which also follows the reported experimental trend nicely. As reported in Fig. 3, our FDTD simulations results show the near eld distribution of second harmonic intensity for nano-antenna with different spacing. Simulation results indicate a huge enhancement of TPS intensity for gold nanoantenna with respect to monomer gold nanoparticle, which matches quite well with experimental observation.
Both experimental and theoretical TPS intensity data, as shown in Fig. 2E also indicate huge TPS intensity enhancement for gold nano-antenna with 3 nm distance separation. As shown in Fig. 2E, our experimental nding on distance dependent twophoton scattering intensity for the "spectroscopy ruler" are well supported by FDTD simulation investigation. Both the theoretical and experimental data clearly shows that our "twophoton spectroscopy" ruler can be used within the 25 nm window.
To understand how the multipolar contribution changes with the separation distance between two gold nanoparticle monomers in gold nano-antenna, angle resolved two-photon scattering measurements were performed for gold nanoantennas with different separation distances. As shown in Fig. 2C, the polar plots clearly indicate that multipolar contribution is predominant for gold nano-antenna with 3 nm separation. Recently, Chandra et al. 34 have reported the inuence of nanoparticle distance separation on SHG polarization and their reported result showed a similar plasmon ruler type of effect for shorter distance separation. Fig. 2D clearly shows that mainly dipolar contribution is predominant for gold nano-antenna with 22 nm monomer separation. All the above angle-resolved two-photon scattering measurement data and distance dependent TPS intensity data as reported in Fig. 2E, clearly indicate that the multipolar contribution plays a very important role for enormous TPS scattering intensity from gold nano-antenna.
Next to demonstrate possible application of TPS ruler for PSMA (+) cancer cell screening, we have developed spectroscopy ruler using partial sequence of the anti-PSMA A9 aptamer (5 0 -GGG AGG ACG AUG CGG ACC GAA AAA GAC CUG ACU UCU AUA CUA AGU CUA CGU UCC CAG ACG ACU CGC CCG-(CH 2 ) 6 -3 0 -SH). The underlined sequence only has been used for hybridization purpose. For this purpose, HS-(CH 2 ) 6 -modied A9 aptamer were attached with gold nanoparticles via the -SH group. Similarly HS-(CH 2 ) 3 -modied capture oligonucleotide which was complementary to the underlined sequence of A9 aptamer was attached with different gold nanoparticles. Aer that, hybridization was performed in 10 mM PBS solution containing 0.3 M NaCl, for developing the TPS ruler, as shown Fig. 4A.
To demonstrate the PSMA (+) cancer cell screening capability, we used two different cell lines in the current experiment. The well-characterized LNCaP prostate cancer cell line, which over-expresses PSMA on the cell surface, was used. [35][36][37] To understand the selectivity of cell screening process using TPS ruler, the PSMA (À) human skin HaCaT keratinocytes, a normal skin cell line was also used. Using an enzyme-linked immunosorbent assay kit, the measured amount of PSMA in LNCaP cells was 1.2 Â 10 6 per cell. No PSMA was found in HaCaT cells. Our TPS ruler based PSMA (+) cancer cell screening is based on the fact that when A9 aptamers based rulers are exposed to PSMA, a free portion of the aptamers will fold in such a way that it will specically bind to the target PSMA on the cell surface. As a result, in the presence of PSMA (+) LNCaP cancer cells, nanoantenna are broken and will form monomers, as shown in Scheme 2 and Fig. 4B. Due to the above process, in the presence of PSMA (+) cancer cells, the spectroscopy ruler distance will change abruptly, which will decrease the two-photon scattering signal tremendously, as shown in Fig. 4C. In our study, we have monitored the normalized TPS intensity change to screen the PSAM (+) cells. Fig. 4C clearly shows that as the LnCaP cells concentration increases, the TPS intensity decreases. It is very interesting to note from Fig. 4C that the sensitivity of PSAM (+) LNCaP cell screening process is about 5 cells per mL.
Since the size of PSAM (+) LNCaP cells are around 10 mM and also the number of PSMA in a single cell is about 10 6 , one cancer cell can bind with huge amount of A9 aptamer attached gold nanoparticles, as a result, TPS intensity changes by about 9% even in the presence of 5 PSMA (+) LNCaP cells per mL. As shown in Fig. 4C, TPS intensity remains constant in the presence of PSAM (À) normal skin HaCaT cells, which clearly indicate that TPS ruler based PSAM (+) cells screening process is highly selective for cells which over-express PSMA on their cell surface.

Conclusion
In conclusion, the current article reports for the rst time, the development of rigid long range two-photon scattering nanoruler whose optical window can be tuned within a 25 nm distance. Our design is based on the formation of gold nanoantenna, where monomer gold nanoparticles are separated by ds-DNA. Our experimental data indicate that the huge enhancement of two-photon scattering intensity is due to the formation of gold nano-antenna. Our reported data show that the two-photon scattering intensity from the nanoruler is highly sensitive to small changes in the monomer separation distance up to 25 nm. Experimental data with angle-resolved two-photon scattering measurement clearly shows that the multipolar contribution plays a very important role for very high TPS scattering intensity from gold nano-antenna. Theoretical insight using FDTD simulations supported our experimental observation very nicely. Simulated electric eld |E| 2 enhancement prole shows that the electric eld enhancement decreases abruptly as the optical ruler distance increases. Though we have demonstrated that TPS ruler optical window can be tuned up to 25 nm distance, the distance limit can be varied by changing the size of spherical gold nanoparticle. Since the local electromagnetic eld strengths vary with the particle size, it is expected that the distance limit will vary with the size of the nanoparticle. We have also demonstrated that a long range TPS ruler using A9 aptamer can be used for the screening of prostate-specic membrane antigen (PSMA) (+) prostate cancer cells, whose sensitivity is 5 cells per mL. Using PSMA (À) normal skin HaCaT cell, we have shown that TPS ruler based cancer cell screening assay is highly selective and it has capability to enable distinction from non-targeted cell lines. Aer appropriate engineering construction, long-range two-photon scattering rulers have the capability to be useful for better understanding of biological process.

Materials and experiments
Sodium borohydride, sodium dodecyl sulfate (SDS), sodium chloride, buffer, trisodium citrate and hydrogen tetrachloroaurate were obtained from Sigma-Aldrich. ss-DNA or A9 aptamer strands with -SH modication were purchased from Midland Certied Reagent. PSMA (+) cancer cell lines and PSMA (À) cell lines and the growth media to grow cells, were purchased from the American Type Culture Collection (ATCC, Rockville, MD).

Synthesis of gold nanoparticles
Gold nanoparticles of around 25 nm size were synthesized by using HAuCl 4 $3H 2 O and sodium citrate, using our reported procedure. 22,23,35 A JEM-2100F transmission electron microscope (TEM), UV-visible absorption spectrum and Malvern Zetasizer Nano instrument were used for the characterization of nanoparticles.
Synthesis of single-stranded DNA (ssDNA) attached gold nanoparticle ssDNA (HS-(CH 2 ) 6 -oligo) were attached to gold nanoparticle via thiol-gold chemistry. -SH modied single strand DNA with different lengths were gradually exposed to gold nanoparticles in the presence of sodium dodecyl sulfate, sodium chloride and PBS buffer for about 12-14 h. Aer attachment, unbound DNA strands were removed by centrifugation of the solution at 6000 rpm for 20 min. Similarly, complementary sequence (oligo-(CH 2 ) 3 -SH) attached gold nanoparticle were prepared in 10 mM PBS solution containing 0.3 M sodium chloride. By using 5 0 -Rh6G modied ss-DNA strands and 10 mM potassium cyanide to oxidize the gold nanoparticles, about 80-90 DNAs per gold nanoparticle were estimated from uorescence experiments.

Design of two-photon spectroscopy rulers with different length
To design two-photon scattering spectroscopy rulers of different lengths, hybridization was performed using complementary sequence attached gold nanoparticles in 10 mM PBS solution containing 0.3 M NaCl. Excess ssDNA attached gold nanoparticles which were not involved in hybridization were removed by centrifugation from the solution. In our design, we used dsDNA spacers of 4, 10, 20, 30, 40, 60 and 80 base pairs to design a spectroscopy ruler window from 3 to 30 nm. The following capture sequences were used to design the spectroscopy rulers.

Two-photon scattering intensity measurements
The hyper-Rayleigh scattering technique was used to measure the two-photon scattering intensity as we and others have previously reported. 22,23,30 For the measurement of TPS intensity, a fundamental wavelength of 760 nm from a mode-locked Ti:sapphire laser, having 150 fs pulse duration at a repetition rate of 80 MHz, was used. Detailed experimental procedures were previously reported. To make sure that there was no photothermal damage during TPS intensity collections, we performed TEM analysis before and aer the exposure of laser, which indicates no photothermal damage from gold nano-antenna within our data collection time. To remove single photon scattering signals and signals from other nonlinear processes, we used a high-pass lter, 3 nm bandwidth interference lter, and a monochromator before the two-photon scattering light was detected by a cooled photomultiplier tube. For angle-resolved TPS measurement, the fundamental beam at 760 nm was linearly polarized, whereas the input angle of polarization was selected using a rotating half-wave plate.

Finite difference time domain (FDTD) simulation
Full-eld electromagnetic wave calculations were performed using an FDTD simulation soware package. 21,[24][25][26]30 The discrete electric and the magnetic elds were evaluated on the grid points that are interspersed in both space and time.
Frequency dependent dielectric function of gold was used for the gold nano-antenna. This method consists of discretization of electric eld and magnetic eld simultaneously in space and time. For our simulation, we used the experimental excitation wavelength for excitation of gold nano-antenna. Absorbing boundary condition was used for the entire calculations. The amplitude of the electric eld was 1 V m À1 . The courant number was taken as 0.99. The entire simulation is done with 0.001 nm mesh resolution and 4000 fs time for all cases.
Cell culture PSMA (+) cancer cells and normal cells were grown according to the ATCC procedure, as we have reported before. 30,36 LNCaP cells were grown in a 5% CO 2 incubator at 37 C using ATCC medium supplemented with 10% premium fetal bovine serum (FBS) and antibiotics (10 IU mL À1 penicillin G and streptomycin) in 75 cm 2 tissue culture asks. The HaCaT cells were also grown in Dulbecco's Modied Eagle's Medicum (DMEM), as instructed by ATCC.