1.3.5 Surface enhanced Raman spectroscopy: SERS
SERS spectra were collected with a portable IM-52 Raman microscope
(Snowy Range Instruments, USA) with the 785 nm incident laser power of
100 mW and integration time of 1 second. 100 μL of each sample was
placed in the quarts cuvette Starna 18B/Q/10 and 15 spectra measurements
of each sample was averaged and used as a final result. The spectra were
baseline-corrected using the Peak software (Snowy Range
Instruments, Laramie, WY, USA).
Gold nanoparticles protein interactions
The following samples were prepared to determine GNPs-protein
interactions. To test GNPs stability in high ionic strength liquid GNPs
samples (concentration, determined with NTA, was adjusted to
1010 particles/mL for all samples) have been either
centrifuged and resuspended in 1X PBS or incubated for 30 min with 9
mg/mL bovine serum albumin (BSA, heat shock fraction, pH 7, ≥98%,
purchased from Sigma-Aldrich) and then centrifuged and resuspended in
PBS. To prepare “washed“ and “unwashed“ samples of GNPs@BSA, AuNP50,
AuNP70 and AuNS particles were incubated for 30 min with varying
concentrations BSA, including 0.1 mg/mL, 2 mg/mL, 9 mg/mL, and 15 mg/mL.
The “unwashed” samples were incubated with BSA without removal of
unbound proteins. The “washed” samples were incubated were incubated
with BSA for 30 minutes, after which they were centrifuged at 6,000 rpm
for 10 minutes and resuspended in Milli-Q water.
Further analysis of GNPs@BSA samples via extinction spectroscopy, zeta
potential and SERS measurements were performed as indicated earlier.
Finite-element method (FEM) simulation
To simulate the extinction spectra of different nanostructures and
enhanced electric field distribution around them, Wave Optics Module
from COMSOL Multiphysics software package (see www.comsol.com) was used.
Spherical nanoparticle and ellipsoidal nanoparticle were built in the
software while nanostars were built in Fusion 360 (Autodesk, USA, CA)
and then was imported to COMSOL Multiphysics. Sizes and geometries of
nanoparticles were measured from TEM images and analysed in ImageJ
software (National Institutes of Health, MD, USA). The refractive index
of bovine serum albumin was taken to be as n = 1.6021[1] and water (n = 1.33) was used as surrounding
medium. The real part and imaginary part of refractive index for gold
was taken from Johnson and Christy.[2] All
simulations were performed for the interaction of a linearly polarized
plane wave with the nanostructures. For spherical and ellipsoidal
nanoparticles, the wavelength of incident light was tuned from 300 nm to
800 nm. For nanostars, the wavelength of incident light was tuned from
400 nm to 850 nm. All enhanced electric field distributions were
calculated for 785 nm incident light, equal to the experimental laser
used for Raman measurements.
Molecular dynamics (MD) simulation
We performed MD simulations to reveal the potential link between gold
surface curvature and the adsorption of BSA protein. Four types of gold
models are constructed: plane (P), truncated cone large (TCL), truncated
cone small (TCS), and spherical tip (ST). The adsorption surface of P in
the model is the Au (111) facet, presenting an infinitely large radius
of curvature. Sizes and geometries of nanoparticles were measured from
TEM images and analysed in ImageJ software (National Institutes of
Health, MD, USA). The geometries of the truncated cones are constructed
according to the TEM sizes for the first type of spikes. The axial
direction (growing direction) of the truncated cone is aligned in the
[111] direction. The starting radii of the TCL and TCS are 7 nm and
4.5 nm, respectively. The radii narrows linearly, and the descending
ratio is around 0.11, mimicking the first type of spike. The axial
length of both models is 20 nm. The ST model was constructed by a
hemisphere with a 2.25 nm radius and a truncated cone (axial length 2.25
nm). The structure of BSA was acquired from Protein Data Bank, ID
3V03.[3] Because of the complex geometry of BSA,
the protein was rotated and allowed to approach the gold surface at
different angles. For each gold model, 10 directions were tested with an
angle interval of 90°, which are named according to the rotation angle
about the x, y, and z BSA molecule position axis (R_x_R_y_R_z).
To accelerate the simulations, a coarse-grained (CG) model of BSA was
adopted with the MARTINI potential.[4] A full
atomic model, described by a Lennard-Jones (LJ) interaction was fixed
for the simulation of the gold surfaces.[5] The
interactions between gold and protein was described by LJ
interactions.[6] Initially, the protein was placed
away from the gold surface with a minimum distance of 0.8 nm. The
protein adsorbed onto the gold surface within 1 ns of simulation. The
energy of the system was determined by averaged over 1 ns simulation
after adsorption under a canonical ensemble. The temperature was
controlled at 300 K. GROMACS was used to perform MD simulation. The
timestep were 0.02 ps. The cut-off distances for the van der Waals
interaction and electrostatic force were 1.2 nm.
Supplementary data
Table S1 Size of gold nanoparticles evaluated with NTA.