CN6163 PAPER CRITIQUE PLASMONIC NANOSENSORS WITH INVERSE SENSITIVITY

[AUTHOR NAME] 1

CN6163 - PAPER CRITIQUE
Laura Rodrguez-Lorenzo, Roberto de la Rica, Ramn A. lvarez-Puebla, Luis M. Liz-Marzn and Molly
M. Stevens. (July 2012) Plasmonic nanosensors with inverse sensitivity by means of enzyme-guided
crystal growth. Nature Materials, 11:7, pp. 604-607. (1)
Impact factor of Nature Materials: 35.749 in 2012 (2)
Number of times cited: 40 (3)
INTRODUCTION
This paper uses plasmonic nanosensors to create a signal-generation mechanism that inverses the
signal at ultralow concentrations (10
-20
g/ml) of the analyte the signal would be the strongest, and
as the concentration increases (up to 10
-13
g/ml) the signal would fall. In the absence of the analyte
the signal would be zero. As such the measurement would be highly sensitive to even very small
amounts of analyte, and there would be no confusion between zero and ultralow concentrations.
The plasmonic nanosensor in question, ie. gold nanostars, utilizes the localized surface plasmon
resonance (LSPR) band as a measurement signal. These nanostars are linked to glucose oxidase (GOx),
which produce hydrogen peroxide upon the addition of its substrate (glucose). Hydrogen peroxide,
being a reducing agent, will reduce surrounding silver ions into silver atoms, causing them to either
deposit on the nanostars as a coating, or to nucleate in the solution, forming nanoparticles. At low
concentrations of glucose oxidase the rate of crystallization is low, so the formation of a silver coating
on nanostars is favoured. This causes the LSPR of the coated nanostars to blueshift greatly. At higher
concentrations of glucose oxidase, the formation of silver nanoparticles is favoured, and thus the
coating on the nanostars would be thinner. As such the blueshift of the LSPR is not as much at higher
concentrations than at lower concentrations, hence the inverse sensitivity. This is summarized in the
scheme below.

Figure I. Scheme of the signal -generation mechanism by means of enzyme-guided crystal growth. (1)
The experiments done in this paper can be split into two parts. In the first part the concept of inverse
sensitivity is proven using gold nanostars covalently linked with glucose oxidase. In the second part a
CN6163 PAPER CRITIQUE PLASMONIC NANOSENSORS WITH INVERSE SENSITIVITY

[AUTHOR NAME] 2

sandwich ELISA immunoassay using these plasmonic sensors was created to test for the presence of
prostate-specific antigen (PSA). These are summarized in the figure below.
a)










b)
Figure II. Tested configurations in the two parts of experiments. a) Proof of concept of inverse sensitivity. b) Sandwich
ELISA immunoassay in phosphate buffer solution for calibration and in human serum to demonstrate robustness of
nanosensors when applied to real situations.
CRITIQUE
This paper presents a very novel idea, that of inverse sensitivity. The inverse sensitivity not only allows
for the lowering of the detection limit, but also enhances sensitivity at ultralow concentrations so that
the zero signal can be clearly differentiated from that of ultralow concentrations. As the authors
themselves have mentioned, this finds important applications not only in the biomedical field, but also
in food safety regulations and environmental policies where toxic substances at low concentrations
can be lethal. While nanosensors using LSPR shifts as a signal have already been proposed, the usage
of the LSPR shifts for inverse sensitivity represents a novel breakthrough unique to nanosensor
technology. It is thus no surprise that this article has been published in Nature Materials, which has
the highest impact factor out of other journals in the same field of materials, and the article has
received 40 citations over less than 2 years.
Generally the experiments were well designed and adequate controls were used so that the
conclusions are valid. Experimental methods were also well detailed so that people may try to
reproduce their work. In the first part, the authors tested with replacing GOx with BSA (negative
control), and also (separately) the nanostar-GOx complex but without glucose, to confirm that the
activity of the enzyme, GOx, and its product hydrogen peroxide, is absolutely critical to the formation
of the silver coating on the gold nanostars. Using these controls, the inverse sensitivity was also proven
satisfactorily to work at this stage. Subsequently, the resulting nanostructures were then studied using
XEDS spectra and maps, as well as TEM and STEM imaging, to demonstrate the theory that the
different concentrations of GOx favour either silver growth on the nanostars or nucleation into
nanoparticles.
Here the demonstration appears to be less convincing, due to a lack of evidence or methods showing
the thicker layer of silver on the nanostars at lower concentrations of GOx. Firstly there is no doubt
about the conclusion that silver nanoparticles are formed at higher concentrations but are absent or
Gold nanostar to be
coated with silver
Glucose oxidase
Gold nanostar to be
coated with silver
Polyclonal anti-PSA
Prostate-specific
antigen (PSA)
Monoclonal IgG anti-PSA
Secondary anti-IgG
Glucose oxidase
CN6163 PAPER CRITIQUE PLASMONIC NANOSENSORS WITH INVERSE SENSITIVITY

[AUTHOR NAME] 3

of very small sizes at ultralow concentrations. However in both cases a silver coating is formed on the
nanostars, according to the TEM images (figure 3a,b). The authors mention that the silver coating at
ultralow concentrations of GOx (10
-20
g/ml) has on average about 4% more silver than that at higher
concentrations (10
-14
g/ml), which leads to the larger blueshift, but do not mention anywhere in the
methods nor in the supplementary section how this number was estimated. It is thus questionable if
4% more silver could really lead to a larger blueshift of the LSPR of over 60nm (estimated from figure
2c). Could the mechanism be, instead of the differing amounts of silver deposited on the nanostars, a
question of the placement or location of the deposited silver under different regimes? Actually it is
probably possible to test the extent of the blueshift by simulating the gold nanostar with a silver
coating, either using COMSOL (finite elements), or any of the popular methods such as finite difference
time domain (FDTD) and discrete dipole approximation (DDA). This could prove that even a small
addition of silver to the nanostars could result in such a big blueshift.
Aside from the lack of evidence or methods to determine that a thicker layer of silver had indeed been
formed in the case of lower concentrations, the inverse sensitivity has been explained and
demonstrated to great effect, and the necessary controls were present.
In the second part of the experiments, the authors created a sandwich ELISA immunoassay using the
plasmonic nanosensor in order to detect ultralow concentrations of the analyte prostate-specific
antigen (PSA) in undiluted human serum. The negative controls were again present and adequate, and
the authors even took into account the high ionic strength of the medium under physiological
conditions (in the supplementary section). The authors first calibrated the immunoassay in phosphate
buffer solution, then measured the extent of the blueshifting of the LSPR band for PSA in human serum.
They concluded that human serum may have endogenous levels of PSA which corresponded to 10
-18

g/ml, by comparing the LSPR band of the blank human serum to that of the calibration curve. Also, the
different values of the slope and dynamic range between the calibration and human serum samples
were attributed to the endogenous levels of PSA and non-specific protein interactions. It is here that
a point of confusion arises.
In order to illustrate, the normalized absorbance curves for the calibration and human serum samples
are reproduced on the next page, taken from the supplementary information of the paper.
The point of confusion is in figure S9. The authors arrive at endogenous levels of PSA of 10
-18
g/ml by
comparing the peak of the black curve in figure S9 with figure S8, which does indeed give 10
-18
g/ml.
However in figure S9, after PSA was added to the human serum, the LSPR peak blueshifted. This is
contrary to the concept of inverse sensitivity, since the signal is supposed to be the strongest at the
lowest detectable concentration. As such it is illogical for the LSPR band to blueshift when adding more
PSA, if PSA was already supposed to be present in the human serum.
Figure S9 would make sense then if the blank human serum really had 0 g/ml of PSA. As such perhaps
the more logical conclusion would have been that the blank human serum showed a shifted LSPR band
compared to the calibration samples not because of endogenous levels of PSA, but because of other
CN6163 PAPER CRITIQUE PLASMONIC NANOSENSORS WITH INVERSE SENSITIVITY

[AUTHOR NAME] 4

reasons such as non-specific interactions with proteins in the serum, or a different refractive index of
the human serum with respect to PBS, or other unknown reasons. It is unsure why the authors left
such a question unaddressed. Nevertheless, this does not detract from the validity of final conclusion
that the immunoassay can indeed perform well under physiological conditions, and that the plasmonic
nanosensors are stable enough to be used in such environments while giving sensible results.


CONCLUSION
Overall the paper presents a novel idea for plasmonic nanosensors, that of inverse sensitivity. This
represents an interesting breakthrough and use for nanotechnology, and holds much promise for
lowering the detection limits, perhaps even eventually reaching the single molecule detection limit.
CN6163 PAPER CRITIQUE PLASMONIC NANOSENSORS WITH INVERSE SENSITIVITY

[AUTHOR NAME] 5

The experiments were well designed and well detailed in the paper, which makes for easy
reproduction of the results. Necessary controls were also made to ensure that their final conclusions
were valid. However two critical points can be raised:
a) The method as to how the authors determined the relative amounts of the silver coatings on
the nanostars was not mentioned, which casts some doubt on the demonstration of the
mechanism.
b) The authors had questionable results from their immunoassay test in human serum if we use
their assumptions of endogenous PSA levels.
Nevertheless, the inverse signal-generating mechanism as presented in the paper is sound and valid.
The article on the whole is well written and deserving of being published in such an impactful journal.




Figures quoted in Arabic numbers (eg. figure 2) refer to the figures found in the paper itself.





BIBLIOGRAPHY
1. Plasmonic nanosensors with inverse sensitivity by means of enzyme-guided crystal growth.
Rodriguez-Lorenzo, et al. 7, 2012, Nature Materials, Vol. 11, pp. 604-607.
2. Guide to Authors. Nature Materials. [Online] [Cited: 21 4, 2014.]
http://www.nature.com/nmat/authors/index.html#impact-factor.
3. Web of Science. [Online] [Cited: 21 4, 2014.]