Negligible Particle-Specic Antibacterial Activity of Silver

Nanoparticles
Zong-ming Xiu,

Qing-bo Zhang,

Hema L. Puppala,

Vicki L. Colvin,

and Pedro J. J. Alvarez*

,

Department of Civil and Environmental Engineering and

Department of Chemistry, Rice University, Houston, Texas 77005, United
States
*S Supporting Information
ABSTRACT: For nearly a decade, researchers have debated the
mechanisms by which AgNPs exert toxicity to bacteria and other
organisms. The most elusive question has been whether the AgNPs exert
direct particle-specic eects beyond the known antimicrobial activity of
released silver ions (Ag
+
). Here, we infer that Ag
+
is the denitive
molecular toxicant. We rule out direct particle-specic biological eects by
showing the lack of toxicity of AgNPs when synthesized and tested under
strictly anaerobic conditions that preclude Ag(0) oxidation and Ag
+
release. Furthermore, we demonstrate that the toxicity of various
AgNPs (PEG- or PVP- coated, of three dierent sizes each) accurately
follows the doseresponse pattern of E. coli exposed to Ag
+
(added as
AgNO
3
). Surprisingly, E. coli survival was stimulated by relatively low
(sublethal) concentration of all tested AgNPs and AgNO
3
(at 38 g/L
Ag
+
, or 1231% of the minimum lethal concentration (MLC)),
suggesting a hormetic response that would be counterproductive to antimicrobial applications. Overall, this work suggests
that AgNP morphological properties known to aect antimicrobial activity are indirect eectors that primarily inuence Ag
+
release. Accordingly, antibacterial activity could be controlled (and environmental impacts could be mitigated) by modulating
Ag
+
release, possibly through manipulation of oxygen availability, particle size, shape, and/or type of coating.
KEYWORDS: Silver nanoparticles, toxicity, metallic nanoparticle, silver ion, E. coli
A
s a broad-spectrum antimicrobial agent, silver nano-
particles (AgNPs) are currently the most widely
commercialized nanomaterial.
1
They are increasingly used in
medical and consumer products,
2,3
including household
antiseptic sprays and antimicrobial coatings for medical devices
that sterilize air and surfaces.
2,4,5
There is no doubt that the release of silver ions from the
crystalline core of silver nanoparticles contribute to the toxicity
of these nanomaterials. However, whether the metallic
nanoparticle itself exerts a particle-specic toxicity remains
an elusive question.
6
A specic answer would help both to
advance antimicrobial applications of AgNPs and to clarify their
potential behavior and impact in the environment. Many
toxicity studies using dierent organisms (e.g., bacteria, algae,
fungi, C. elegans, zebra sh, Lolium multif lorum, human cells)
have concluded that the toxicity of AgNPs is not solely due to
silver released from the nanoparticle (e.g., silver ions or
dissolved silver). In these cases, AgNPs often were more toxic
than equivalent concentrations of silver salts.
712
However,
these observations could be confounded by ligands in the
exposure medium that can bind to dissolved silver. These
include chloride, sulde, phosphate, or organic acids whose
presence would reduce the bioavailability (and thus the
toxicity) of released silver ions to a greater extent than that
of AgNPs.
13,14
Other studies have attributed the toxicity of
AgNPs solely to dissolved silver, possibly because of the
dierent chemical constituents of the exposure media.
15
In each
of these past examples, the studies were conducted under
aerobic conditions that promote the continued release of silver
ions (Ag
+
) from AgNPs and confound the discernment of their
relative contribution. Additionally, few studies assessed the
residual Ag
+
concentrations (dissolved or adsorbed to the
AgNPs coating), which present challenges in dierentiating true
particle-specic toxicity.
Some studies reported that AgNP size,
1619
shape,
20
surface
charge,
21
surface coating,
15
solution chemistry,
13,22
and
solubility
15,23
aect AgNPs toxicity (Table 1). However, the
extent to which these factors aected toxicity directly by
inuencing particle-specic biological eects or indirectly by
aecting silver ion release remains an open question.
Discerning the relative importance of a particle-specic eect
in the antibacterial activity of AgNPs requires careful
quantication of the silver ion concentration contributed by
the nanoparticle, as well as the role that complexing ligands
present in the exposure media could have on silver ion and
particle bioavailability.
Received: May 23, 2012
Revised: June 15, 2012
Published: July 5, 2012
Letter
pubs.acs.org/NanoLett
2012 American Chemical Society 4271 dx.doi.org/10.1021/nl301934w | Nano Lett. 2012, 12, 42714275
This study aims to resolve these outstanding questions about
AgNP toxicity and to determine whether the metallic
nanoparticle alone contributes to the antibacterial eects. The
work eliminates the possibility of silver ion release by
completing some experiments under anaerobic conditions.
Additionally, studies are conducted in a minimal medium
(NaHCO
3
buer solution, 2 mM) whose constituents have no
eect on silver bioavailability (all AgNPs/Ag
+
toxicity assays in
this work were below Ag
2
CO
3
precipitation potential (K
sp
=
0.81 10
12
, calculated by Visual MINTEC 3.0)). The eects
of nanoparticle size and surface coating on antibacterial activity
were also considered by synthesizing three types of size-
controlled AgNPs to gain insight into particle-specic eects.
Synthesis and Characterization of the PEG-AgNPs.
Particle size has frequently been reported to be a determinant
of AgNPs toxicity (Table 1). To conrm this nding, three
types of glycol-thiol-coated AgNPs (PEG-AgNPs) were
synthesized following the modied procedure as described in
Hiramatsu et al.
24
Figure 1 shows that these PEG-AgNPs have
well-controlled particles sizes (spherical shape) and narrow size
distributions with sizes ranging from 2.8 0.47, 4.7 0.20, to
10.5 0.59 nm. These three types of PEG-AgNPs are referred
to in the following text as PEG-3 nm, PEG-5 nm, and PEG-11
nm. These PEG-AgNPs are nonaggregating in both DI water
and the minimal medium. The samples remain suspended for at
least 6 months inside the anaerobic chamber.
Toxicity of AgNPs Can Be Solely Explained by the
DoseResponse of the Released Ag
+
. It is well-known that
silver nanoparticles can be oxidized in aqueous solutions
exposed to air (eq 1) resulting in the release of silver ions under
acidic conditions (eq 2)
25
+ 4Ag(0) O 2Ag O
2
2
(1)
+ +
+ +
2Ag O 4H 4Ag 2H O
2
2 (2)
Figure 2 shows that a PEGylation coating on the particles
does not protect them against these reactions; under aerobic
conditions, silver ion concentrations can be detected and they
increase over time (up to 2.1 mg/L after 5 day exposure for
PEG-5 nm at pH 4.0). This release pattern could be dierent
and highly variable in the presence of bacteria, which depending
on their metabolic state could aect the dissolved oxygen
concentration (eq 1), pH (eq 2), release, or remove polymeric
substances that coat AgNPs and increase the silver dissolution
gradient by binding the released Ag
+
. However, no silver ions
were released when these AgNPs were stored under anaerobic
conditions, suggesting a route for distinguishing the toxicity
arising from the nanoparticle with the toxicity arising from the
released silver ions. AgNPs that have limited air exposure and
whose interactions with microbes are evaluated under strict
anaerobic conditions can only impact organisms through
particle-specic eects.
To discern the toxicity contribution of the metallic
nanoparticles, PEG-AgNPs (PEG-5 and 11-nm) were synthe-
sized and assayed inside the anaerobic chamber. As noted in
Figure 2 under such conditions there is no detectable Ag
+
(<1
g/L). An E. coli strain K12 (ATCC 25404) was chosen as a
model microorganism for inactivation experiments because it is
Table 1. Proposed AgNPs Toxicity Determinants
property target organisms references
size E. coli; human alveolar macrophages 1619
shape E. coli 20
surface coating C. elegans 15
solution chemistry Nitrifying bacteria 13
surface charge B. subtilis 21
Ag
+
release C. elegans; E. coli, Nitrosomonas europaea 14, 15,
23, 39
ROS generation nitrifying bacteria; human monocytic cell;
alveolar macrophages; human hepatoma
cells; human lung broblast cells
7, 19,
4042
Figure 1. Transmission electron microscopy images and size
distribution of the lab synthesized PEG-AgNPs. (a) PEG-AgNPs-3
nm (2.8 0.47 nm); (b) PEG-AgNPs-5 nm (4.7 0.20 nm); and (c)
PEG-AgNPs-11 nm (10.5 0.59 nm). The PEG-AgNPs dispersed
homogeneously in DI water and showed narrow size distribution,
which facilitates the investigation of particles size eect on AgNPs
toxicity. The three insets are pictures of the corresponding
suspensions.
Figure 2. PEG-AgNPs (5 nm and 11 nm) dissolution (at pH 4.0)
under aerobic and anaerobic conditions. Dissolved Ag
+
concentration
increased with air exposure time for both PEG-5 nm and PEG-11 nm
nanoparticles under aerobic conditions, while no silver ions were
detected (<1 g/L) under anaerobic conditions.
Nano Letters Letter
dx.doi.org/10.1021/nl301934w | Nano Lett. 2012, 12, 42714275 4272
a facultative bacterium that exhibits equal susceptibility to silver
ions under aerobic and anaerobic conditions (Supporting
Information, Figure S2).
14
Figure 3 showed that under
anaerobic conditions the nanoparticles had no measurable
eect on E. coli up to concentrations thousands of times (6224
and 7665 times) higher than the minimum lethal concentration
(MLC) of silver ions themselves (0.025 mg/L) under similar
exposure conditions.
14
The lack of toxicity of these nano-
particles at the highest concentrations reached inside the
anaerobic chamber (i.e., 158 mg/L for 5 nm PEG-AgNPs and
195 mg/L for 11 nm PEG-AgNPs) suggest that the particles
themselves do not aect the biological activity of the microbes.
Aerobic toxicity assay using the same anaerobically synthesized
PEG-AgNPs were also completed to investigate the confound-
ing eect of the released Ag
+
. Toxicity assay (6 h exposure) of
PEG-5 nm under aerobic conditions (immediately after
transferring out of the chamber) showed enhanced toxicity,
indicating that silver ion released during the toxicity assay can
have a notable antimicrobial eect. Prolonged air exposure (48
h with magnetic stirring to increase oxygen exposure) induced
higher antibacterial toxicity of the PEG-5 nm silver nano-
particles. These results illustrate that the toxicity of AgNPs is
very sensitive to the presence of air. Oxidative dissolution of the
crystalline cores can result under aerobic conditions and
increase the concentration of soluble silver ions.
We speculated that all of the aerobic toxicity of the silver
nanoparticles could be explained by the presence of released
Ag
+
. To test this hypothesis, air-exposed PEG- and PVP-AgNPs
(commercially available, Supporting Information, Figure S1)
suspensions were tested for antimicrobial activity. These assays
were conducted inside an anaerobic chamber to stop the
increased dissolution of particles during the time scale of the
experiments. Figure 4 shows that the antimicrobial activity of all
AgNPs when expressed in terms of the measured concen-
trations of silver ion were statistically indistinguishable (p >
0.05, Supporting Information, Table S2) from the toxicity of
silver ions introduced through silver nitrate. The fact that the
doseresponse patterns of the six dierent AgNPs could be
explained by the concentration of the released Ag
+
corroborates
that the antimicrobial activity was solely due to the released Ag
+
and that no direct particle-specic eects contributed to
toxicity.
For nearly a decade, researchers have debated the
mechanisms by which AgNPs exert toxicity to bacteria and
other organisms (especially whether the AgNPs exert direct
particle-specic toxicity). These results demonstrate that the
antimicrobial activity of AgNPs is solely due to Ag
+
release and
that even relatively low (g/L) concentrations of Ag
+
(released
or adsorbed to AgNP coatings) can account for the biological
response observed in previous studies. Particle properties that
aect toxicity such as size,
1619
shape,
20
surface coating,
15
and
surface charge
21
likely aect toxicity indirectly through
mechanisms that inuence the rate, extent, location, and/or
timing of Ag
+
release. For example, AgNPs of smaller size may
exert higher toxicity due to their higher specic surface area and
associated faster Ag
+
release rate compared to larger AgNPs.
17
To fully control AgNPs toxicity will require deeper under-
standing of the release, speciation, and bioavailability of
Ag
+
.
25,26
Figure 3. Elimination of toxicity by AgNP synthesis and exposure
under anaerobic conditions that preclude oxidative Ag
+
release (less
than 1 g/L, by ICP-MS). Viability assays show no statistically
signicant toxicity with concentration up to 158 (for 5 nm AgNPs)
and 195 mg/L (for 11 nm AgNPs), which were the highest
concentration reached inside the anaerobic chamber. These nonlethal
concentrations are respectively 6224 and 7665 times higher than MLC
for Ag
+
, indicating negligible toxicity. Antibacterial assays (6 h
exposure) with the same 5-nm PEG-AgNPs under aerobic conditions
(conducted immediately after transferring the particles out of the
chamber) showed toxicity, illustrating the potential confounding eect
of Ag
+
release during exposure. Storage in an aerobic atmosphere (48
h with magnetic stirring to increase oxygen exposure) resulted in
higher Ag
+
release and higher toxicity.
Figure 4. Doseresponse of E. coli exposed to various air-exposed
AgNPs. EC
50
increases with the increasing particle size (a), suggesting
size-dependent toxicity. This is an indirect eect associated with Ag
+
release (smaller AgNPs release more Ag
+
and are more toxic).
Antibacterial activity expressed as a function of the concentration of
the released Ag
+
was statistically indistinguishable from the dose
response patterns of cells exposed to Ag
+
(added as AgNO
3
) (p >
0.05), illustrating that the released Ag
+
is the critical factor of
antibacterial activity (b).
Nano Letters Letter
dx.doi.org/10.1021/nl301934w | Nano Lett. 2012, 12, 42714275 4273
E. coli Survival Was Stimulated by Low Silver Doses.
Survival of resting E. coli cells in 2 mM NaHCO
3
buer solution
was stimulated in the presence of low Ag
+
concentrations with
13% higher bacteria viability in the AgNO
3
-treated group after
6 h exposure than in the unexposed control group (Figure 5a).
This enhanced tolerance phenomenon was also observed as a
result of exposure to lower concentrations of all tested AgNPs
with higher survival rates of 6% for PEG-AgNPs-3 nm at 2.2
mg/L, 7% for PEG-AgNPs-5 nm at 1.8 mg/L, and 13% for
PEG-AgNPs-11 nm at 2 mg/L (Figure 5b) and 11% for PVP-
AgNPs-20 nm at 16.4 mg/L, 21% for PVP-AgNPs-40 nm at 5.7
mg/L and 17% for PVP-AgNPs-80 nm at 6.7 mg/L (Figure 5c).
This apparent hormetic eect
27,28
might have been triggered by
the residual Ag
+
in air-exposed AgNPs stock suspensions
(Supporting Information, Table S1) that could not be
completely separated by ltration. The residual Ag
+
concen-
trations in AgNPs stock suspensions ranged from 0.28 mg/L
(in PEG-AgNPs-11 nm stock suspensio, 82.0 mg/L) to 7.9 mg/
L (in PVP-AgNPs-40 nm stock suspension, 5700 mg/L),
corresponding to Ag
+
concentration of 37.9 g/L (at the
tested AgNPs doses), or 1231% of the MLC for Ag
+
. This
nding suggests that sublethal concentrations of silver may
enhance bacterial tness and hinder antimicrobial applications.
A hormesis eect has also been observed in studies of the
impact of silver nanoparticles toxicity studies on human cell
lines (e.g., peripheral blood mononuclear cells, human
hepatoma derived cell line HepG2).
22,2931
Moreover, other
kinds of nanoparticles, that is, carbon nanotubes,
32
quantum
dots,
33,34
and metal nanoparticles,
35
also have shown a
stimulatory eect at sublethal exposures. Apparently, the
presence of low doses of toxicants can activate repair
mechanisms of the cells against the toxicant, and this repair
process may sometimes overcompensate for the exposure.
36
The poorly understood hormetic response of Ag
+
on E. coli
cells and other nanomaterials on all kinds of organisms
underscores the need for further study of its responsible
mechanisms.
Implications for AgNPs Antibacterial Application and
Environmental Impact. Whereas AgNPs themselves do not
signicantly exert direct particle-specic toxicity on bacteria,
AgNPs could be engineered with dierent particle formations
(e.g., surface coatings) to release Ag
+
at desired rate and
location. Furthermore, AgNPs may serve as a vehicle to deliver
Ag
+
more eectively (being less susceptible to binding and
reduced bioavailability by common natural ligands
14
) to the
bacteria cytoplasm and membrane (Figure 6), whose proton
motive force would decrease the local pH (as low as pH
3.0)
37,38
and enhance Ag
+
release (Supporting Information,
Figure S3).
Although this research was conducted with a model
bacterium (E. coli), our approach to separate the contributions
of Ag
+
from AgNPs may benet the etiology of AgNP toxicity
to higher order organisms (e.g., algae, zebra sh, C. elgans). In
these cases, however, organism-specic immune responses to
dierent nanoparticle morphologies could lead to dierent
observations, underscoring the need for caution when
extrapolating our mechanistic inferences to other biological
systems. Our approach to separate exposure to nanoparticles in
the absence of dissolved metal may also be used to advance
mechanistic understanding of the bacterial toxicity exerted by
other metal-based NPs (e.g., CuO, ZnO, QDs) that also release
toxic metal ions.
Figure 5. Survival of resting E. coli cells in 2-mM NaHCO
3
buer
solution after 6 h exposure to (a) AgNO
3
, (b) PEG-AgNPs, and (c)
PVP-AgNPs. One asterisk represents signicant decrease in viability (p
< 0.05) relative to unexposed control and corresponds to the
minimum lethal concentration (MLC). A signicant stimulatory eect
suggestive of hormesis was observed at some sublethal concentrations
of all treatments, as indicated by two asterisks.
Figure 6. Schematic of AgNPs, Ag
+
, and cell interactions. AgNPs may
serve as a vehicle to deliver Ag
+
more eectively (being less susceptible
to binding and reduced bioavailability by common natural ligands) to
the bacteria cytoplasm and membrane, whose proton motive force
would decrease the local pH (as low as pH 3.0) and enhance Ag
+
release.
Nano Letters Letter
dx.doi.org/10.1021/nl301934w | Nano Lett. 2012, 12, 42714275 4274

ASSOCIATED CONTENT
*S Supporting Information
Experimental methods for AgNPs characterization, E. coli
growth inhibition assay, anaerobic PEG-AgNPs synthesis, air-
exposed AgNPs preparation, AgNPs ltration, statistical
analysis, Figures S1,S2, Tables S1,S2, and corresponding
discussions are provided. This material is available free of
charge via the Internet at http://pubs.acs.org.

AUTHOR INFORMATION
Corresponding Author
*E-mail: alvarez@rice.edu. Phone: (713)348-5903.
Notes
The authors declare no competing nancial interest.

ACKNOWLEDGMENTS
This research was supported by a Joint US-UK Research
Program (Grant RD-834557501-0 by U.S.-EPA and U.K.-
NERC-ESPRC). We thank Xiaoyu Chai (Biostatistician of
Clinical Statistics, Fred Hutchinson Cancer Research Center)
for his assistance with the statistical analysis.

REFERENCES
(1) An Inventory of Nanotechnology-based Consumer Products
Currently on the Market. http://www.nanotechproject.org/
inventories/consumer/analysis_draft/.
(2) Quadros, M. E.; Marr, L. C. Environ. Sci. Technol. 2011, 45 (24),
1071310719.
(3) Kim, J. S.; Kuk, E.; Yu, K. N.; Kim, J. H.; Park, S. J.; Lee, H. J.;
Kim, S. H.; Park, Y. K.; Park, Y. H.; Hwang, C. Y.; Kim, Y. K.; Lee, Y.
S.; Jeong, D. H.; Cho, M. H. J. Nanomed. Nanotechnol. 2007, 3 (1),
95101.
(4) Chen, X.; Schluesener, H. J. Toxicol. Lett. 2008, 176 (1), 112.
(5) Faunce, T.; Watal, A. Nanomedicine (London, U.K.) 2010, 5 (4),
617632.
(6) Levard, C.; Hotze, E. M.; Lowry, G. V.; Brown, G. E. Environ. Sci.
Technol. 2012, 46 (13), 69006914.
(7) Choi, O.; Hu, Z. Q. Environ. Sci. Technol. 2008, 42 (12), 4583
4588.
(8) Navarro, E.; Piccapietra, F.; Wagner, B.; Marconi, F.; Kaegi, R.;
Odzak, N.; Sigg, L.; Behra, R. Environ. Sci. Technol. 2008, 42 (23),
89598964.
(9) Fabrega, J.; Fawcett, S. R.; Renshaw, J. C.; Lead, J. R. Environ. Sci.
Technol. 2009, 43 (19), 72857290.
(10) Meyer, J. N.; Lord, C. A.; Yang, X. Y. Y.; Turner, E. A.;
Badireddy, A. R.; Marinakos, S. M.; Chilkoti, A.; Wiesner, M. R.;
Auan, M. Aquat. Toxicol. 2010, 100 (2), 140150.
(11) Yin, L. Y.; Cheng, Y. W.; Espinasse, B.; Colman, B. P.; Auan,
M.; Wiesner, M.; Rose, J.; Liu, J.; Bernhardt, E. S. Environ. Sci. Technol.
2011, 45 (6), 23602367.
(12) Laban, G.; Nies, L. F.; Turco, R. F.; Bickham, J. W.; Sepulveda,
M. S. Ecotoxicology 2010, 19 (1), 185195.
(13) Choi, O.; Cleuenger, T. E.; Deng, B. L.; Surampalli, R. Y.; Ross,
L.; Hu, Z. Q. Water Res. 2009, 43 (7), 18791886.
(14) Xiu, Z. M.; Ma, J.; Alvarez, P. J. J. Environ. Sci. Technol. 2011, 45
(20), 90039008.
(15) Yang, X. Y.; Gondikas, A. P.; Marinakos, S. M.; Auan, M.; Liu,
J.; Hsu-Kim, H.; Meyer, J. N. Environ. Sci. Technol. 2012, 46 (2),
11191127.
(16) Morones, J. R.; Elechiguerra, J. L.; Camacho, A.; Holt, K.; Kouri,
J. B.; Ramirez, J. T.; Yacaman, M. J. Nanotechnology 2005, 16 (10),
23462353.
(17) Sotiriou, G. A.; Pratsinis, S. E. Environ. Sci. Technol. 2010, 44
(14), 56495654.
(18) Panacek, A.; Kvitek, L.; Prucek, R.; Kolar, M.; Vecerova, R.;
Pizurova, N.; Sharma, V. K.; Nevecna, T.; Zboril, R. J. Phys. Chem. B
2006, 110 (33), 1624853.
(19) Carlson, C.; Hussain, S. M.; Schrand, A. M.; Braydich-Stolle, L.
K.; Hess, K. L.; Jones, R. L.; Schlager, J. J. J. Phys. Chem. B 2008, 112
(43), 1360813619.
(20) Pal, S.; Tak, Y. K.; Song, J. M. Appl. Environ. Microbiol. 2007, 73
(6), 17121720.
(21) El Badawy, A. M.; Silva, R. G.; Morris, B.; Scheckel, K. G.;
Suidan, M. T.; Tolaymat, T. M. Environ. Sci. Technol. 2011, 45 (1),
2837.
(22) Kawata, K.; Osawa, M.; Okabe, S. Environ. Sci. Technol. 2009, 43
(15), 60466051.
(23) Ma, R.; Levard, C.; Marinakos, S. M.; Cheng, Y.; Liu, J.; Michel,
F. M.; Brown, G. E.; Lowry, G. V. Environ. Sci. Technol. 2012, 46 (2),
7529.
(24) Hiramatsu, H.; Osterloh, F. E. Chem. Mater. 2004, 16 (13),
25092511.
(25) Liu, J. Y.; Hurt, R. H. Environ. Sci. Technol. 2010, 44 (6), 2169
2175.
(26) Liu, J. Y.; Sonshine, D. A.; Shervani, S.; Hurt, R. H. ACS Nano
2010, 4 (11), 69036913.
(27) Kaiser, J. Science 2003, 302 (5644), 376.
(28) Iavicoli, I.; Calabrese, E. J.; Nascarella, M. A. Dose-Response
2010, 8 (4), 501517.
(29) Shin, S. H.; Ye, M. K.; Kim, H. S.; Kang, H. S. Int.
Immunopharmacol. 2007, 7 (13), 18131818.
(30) Arora, S.; Jain, J.; Rajwade, J. M.; Paknikar, K. M. Toxicol. Lett.
2008, 179 (2), 93100.
(31) Braydich-Stolle, L.; Hussain, S.; Schlager, J. J.; Hofmann, M. C.
Toxicol. Sci. 2005, 88 (2), 412419.
(32) Pulskamp, K.; Diabate, S.; Krug, H. F. Toxicol. Lett. 2007, 168
(1), 5874.
(33) Jan, E.; Byrne, S. J.; Cuddihy, M.; Davies, A. M.; Volkov, Y.;
Gunko, Y. K.; Kotov, N. A. ACS Nano 2008, 2 (5), 928938.
(34) Stern, S. T.; Zolnik, B. S.; McLeland, C. B.; Clogston, J.; Zheng,
J. W.; McNeil, S. E. Toxicol. Sci. 2008, 106 (1), 140152.
(35) Nations, S.; Wages, M.; Canas, J. E.; Maul, J.; Theodorakis, C.;
Cobb, G. P. Chemosphere 2011, 83 (8), 10531061.
(36) Calabrese, E. J. Crit. Rev. Toxicol. 2001, 31 (45), 425470.
(37) Koch, A. L. J. Theor. Biol. 1986, 120 (1), 7384.
(38) Kemper, M. A.; Urrutia, M. M.; Beveridge, T. J.; Koch, A. L.;
Doyle, R. J. J. Bacteriol. 1993, 175 (17), 56905696.
(39) Arnaout, C. L.; Gunsch, C. K. Environ. Sci. Technol. 2012, 46,
53875395.
(40) Kim, S.; Choi, J. E.; Choi, J.; Chung, K. H.; Park, K.; Yi, J.; Ryu,
D. Y. Toxicol. In Vitro 2009, 23 (6), 10761084.
(41) Foldbjerg, R.; Olesen, P.; Hougaard, M.; Dang, D. A.;
Homann, H. J.; Autrup, H. Toxicol. Lett. 2009, 190 (2), 156162.
(42) AshaRani, P. V.; Mun, G. L. K.; Hande, M. P.; Valiyaveettil, S.
ACS Nano 2009, 3 (2), 279290.
Nano Letters Letter
dx.doi.org/10.1021/nl301934w | Nano Lett. 2012, 12, 42714275 4275
1

SUPPORTING INFORMATION
Negligible Particle-Specific Antibacterial Activity
of Silver Nanoparticles

Zong-ming Xiu, Qing-bo Zhang, Hema L. Puppala, Vicki L. Colvin, Pedro J. J. Alvarez*

a
Dept. of Civil & Environmental Engineering, Rice University, Houston, TX 77005
b
Dept. of Chemical Engineering, Rice University, Houston, TX 77005

* Corresponding Author; Email: alvarez@rice.edu, PHONE: (713)348-5903

9 pages in total, with 3 figures and 2 tables
2

AgNPs and Chemicals
Glycol-thiol coated AgNPs (PEG-AgNPs: 3, 5 and 11 nm) were synthesized in the lab
1

and transferred to an anaerobic chamber for storage. Commercial polyvinylpyrrolidone-coated
AgNPs of three different sizes (PVP-AgNPs: 18, 51 and 72 nm), which have been widely used in
previous studies,
2-6
were obtained from NanoAmor (Houston, TX). The PVP-AgNPs powders
were suspended in DI water and homogenized by an ultrasonic cleaner (5510, Branson, CT). The
sizes and zeta-potential () of the six particles (Table S1) were characterized in the exposure
medium (2 mM sodium bicarbonate buffer), using JEM 2100F TEM (JEOL 2100 Field Emission
Gun Transmission Electron Microscope, Japan) and dynamic light scattering with a Malvern
Zetasizer (ZEN 3600, Malvern Instrument, UK) respectively. The morphologies of the PVP-
AgNPs are provided in Figure S1.
AgNO
3
and HNO
3
(~69.0%) was obtained from Sigma-Aldrich (St. Louis, MO); NaCl,
LB (Luria-Bertani) broth, NaHCO
3
and H
2
O
2
(30%) were all obtained from Fisher Scientific
(Fair Lawn, NJ). All chemicals used were reagent grade or better unless otherwise specified.
Table S1. Characterization of PEG- and PVP-AgNPs stock suspensions
PEG-AgNPs Sample 1 Sample 2 Sample 3
Particle size 2.8 0.5 nm 4.7 0.2 nm 10.5 0.6 nm
Zeta-potential (mV) -16.1 1.2 -13.5 0.5 -17.1 2.1
Stock concentration 44.5 mg/L 70.3 mg/L 82.0 mg/L
(Dissolved Ag
+
) (0.31 mg/L) (0.38 mg/L) (0.28 mg/L)
PVP-AgNPs Sample 1 Sample 2 Sample 3
Particle size 17.5 2.9 nm 51.4 18.7 nm 71.5 20.3 nm
Zeta-potential (mV) -27.5 1.4 -35.8 0.6 -37.1 0.6
Stock concentration 8,200 mg/L 5,700 mg/L 6,700 mg/L
(Dissolved Ag
+
) (2.93 mg/L) (7.90 mg/L) (7.15 mg/L)
3


Figure S1. TEM characterization of the commercial PVP-AgNPs (a) PVP-18nm (17.5 2.9 nm),
(b) PVP-51nm (51.4 18.7 nm) and (c) PVP-72nm (71.5 20.3 nm).

Bacteria
E. coli strain K12 (ATCC 25404) was chosen as a model microorganism for inactivation
experiments. This facilitates comparison with numerous other related antimicrobial studies, and
(because E. coli is facultative) allows for testing under both aerobic and anaerobic conditions. E.
coli exhibits equal susceptibility to silver ions under aerobic and anaerobic conditions (Figure S2)
7
. The detailed bacteria stock suspension preparation method was provided in our previous paper
7
. Sodium bicarbonate (2 mM), which is commonly used as exposure medium
8
, was chosen to
avoid ligands present in complex growth media that could bind with Ag
+
/AgNPs and affect silver
bioavailability and/or promote precipitation or other confounding effects. All AgNPs/Ag
+

toxicity assays were below the Ag
2
CO
3
precipitation potential (K
sp
=0.8110
-12
).
(a) (b) (c)
4


Figure S2. E. coli strain K12 (ATCC 25404) exhibits equal susceptibility to the silver ion (which
is the definitive molecular toxicants) under aerobic and anaerobic conditions. Reproduced with
permission from reference 7. Copyright 2011 American Chemical Society.

Preparation of air-exposed AgNP stock suspensions
To prepare air-exposed AgNPs samples, PEG-AgNPs (3, 5, 11 nm) and PVP-AgNPs (18,
51, 72 nm) stock suspensions (~2g/L) were transferred out of the chamber and exposed to air for
5 days (with pH adjusted to 4.0 to accelerate the oxidation of AgNPs and the release of Ag
+
).
The released Ag
+
concentration of each sample was monitored by ICP-OES/MS (Perkin Elmer,
Waltham, MA) to avoid complete oxidation of AgNPs. The air-exposed AgNPs suspensions
were then transferred to anaerobic chamber and resuspended in NaHCO
3
buffer solution to
achieve a constant pH (8.1) for Ag
+
dissolution equilibration (5 days). The final Ag
+

concentration for PVP-AgNPs are 9.5 mg/L, 17 mg/L. 10.5 mg/L respectively.
Anaerobic synthesis of the PEG-AgNPs
The anaerobic PEG-AgNPs were synthesized by mixing the pre-synthesized PEG-AgNPs
suspension with sodium borohydride under vigorous agitation in an anaerobic chamber. All
5

precursors (including AgNPs suspensions and NaBH
4
solution) were transferred and equilibrated
inside the anaerobic chamber for 24h to allow for oxygen depletion. Excess amount of NaBH
4

was added to ensure a complete reduction of all Ag
+
.
The AgNPs were sealed in an Amicon ultra centrifugal filter unit (molecular weight
cutoff 10,000, 2~5nm in pore size, Millipore, MA) and moved out of the chamber for
centrifugation (J2-MC Centrifuge, Beckman Coulter, CA) to get rid of the impurities (e.g.,
NaBH
4
). The centrifuged samples were transferred back to the chamber to collect the filtrate and
refilled with 2mM NaHCO
3
buffer. This washing process was repeated continuously four times
and the filtrates were monitored for dissolved silver with ICP-Ms. The final PEG-AgNPs stock
suspensions were re-suspended in 2mM NaHCO
3
buffer solution, with no detectable Ag
+
(by
ICP-Ms, detection limit: 1 g/L) in supernatants.
Preparation of filtered AgNP stock suspensions
To purify the AgNPs, PEG- and PVP-AgNPs stock suspensions were washed by DI water
and filtered through the Amicon ultra centrifugal filter units by centrifugation. AgNPs stock
suspensions were filtered continuously for ~10 times to get rid of the residual Ag
+
as much as
possible. The filtrates were analyzed for total dissolved silver with an ICP-OES/MS to obtain the
concentrations of dissolved silver.
After filtration, the retentates were resuspended and transferred to 50-ml corning tubes
(Corning, Lowell, MA). The concentrations of the AgNPs stocks were determined by nitric
acid/hydrogen peroxide digestion as described earlier
7
. These stocks were stored and tested
inside the anaerobic chamber to avoid any confounding effects caused by oxidative Ag
+
release
during the dose-response assays.
6

The minimum lethal concentration (MLC) of the six filtered AgNPs to E. coli was
compared to Ag
+
to assess their relative toxicity contribution to AgNPs. The MLC is defined as
the minimum lethal concentration of AgNPs in the exposure medium (bicarbonate buffer) that
causes statistically significant E. coli mortality (p < 0.05) relative to the control set without
AgNPs, as we reported earlier
7
.
Dose-response assay of AgNPs
Six fresh and six air-exposed AgNPs stock suspensions were tested for dose-response on
E. coli mortality inside the anaerobic chamber following the procedure as showed in Xiu et al
7
.
Specially, antimicrobial assay of the air-exposed AgNPs was conducted after 5 days of
equilibration in anaerobic chamber to ensure a complete Ag
+
dissolution. Free Ag
+

concentrations were measured in the same medium and toxicity data of each AgNPs suspension
was plotted against the free Ag
+
concentration in it. E. coli mortality in different treatments was
then determined by viable plate counts
9
and was calculated as 1-N/N
0
100%, where N and N
0

are the remaining and initial concentrations of viable bacteria (CFU/mL), respectively. The dose-
response curves of E. coli mortality versus Ag
+
concentrations was fitted using a sigmoidal
model (Eq. S1)
10
(Sigma-Plot v10.0):

(Eq. S1)
Where y is the E. coli mortality rate, y
0
is the baseline mortality rate without silver addition,
x is silver concentration (expressed as released Ag
+
), and x
0
is the Ag
+
concentration that results
in 50% mortality (EC
50
). All tests were conducted in triplicate and repeated at least three times to
ensure reproducibility.
Statistical Analyses
7

Whether differences between EC
50
values for different treatments were statistically
significant was determined using Students t-test at the 95% confidence level. T-tests were also
used to determine MLC values (i.e., the lowest concentrations that resulted in a significant (p <
0.05) decrease in cell viability, assessed by plate counts) as well as the significance of hormesis
effect relative to unexposed controls. For the dose-response assays, E. coli mortality data for
each type of air-exposed AgNPs (expressed as their corresponding released Ag
+
concentrations)
and for the AgNO
3
treatment were fitted with the Sigmoidal model (Eq. S1) using non-linear
regression. The SAS PROC NLIN procedure and R software, version 2.9.2 (R Foundation for
Statistical Computing, Vienna, Austria) was used for this purpose.
11
Whether differences in
dose-response trends were statistically significant were determined by the F-test at the 95%
significance level using the SAS/STAT version 9.2 software (SAS Institute Inc., Cary, NC).
This analysis (Table S2) inferred that the sigmoidal curves were statistically indiscernible and
thus, that the antibacterial activity of AgNPs could be fully explained by the toxicity of the
released Ag
+
.
Table S2. Dose-response assay statistical analysis results
F-test
Ag
+

control
PEG-
3 nm
PEG-
5 nm
PEG-
11 nm
PVP-
18 nm
PVP-
51 nm
PVP-
72 nm
p value 0.80 0.71 0.56 0.79 0.65 0.22

Ag
+
release at low pH found near a cell membrane under the influence of the proton motive
force
The air-exposed PVP-AgNPs (three types in total) were transferred into the anaerobic
chamber and each type of AgNPs was added to different media with different pH values. The
lowest pH used (3.0, diluted HNO
3
solution) was mimicking the local pH of a cell membrane
8

under the influence of the proton motive force. An intermediate pH (5.5, DI water) was used to
represent the inside of a lysosome in eukaryotic cells. The highest pH (8.1, 2mM NaHCO
3

solution) is representative of bicarbonate-buffered natural waters and was the same as in the
toxicity assays. The released Ag
+
concentrations in the supernatants were measured by ICP-Ms
after 72-h equilibration. As shown in Figure S2, Ag
+
release was affected by the solution pH,
with higher Ag
+
release at lower pH, suggesting increased release as AgNPs become in contact
with these organelles.
0
5
10
15
20
25
30
35


86 nm 36 nm
A
g
+

r
e
l
e
a
s
e
d

(
m
g
/
L
)
25 nm
pH 3.0
pH 5.5
pH 8.1

Figure S3. Dissolution of air-exposed PVP-AgNPs at different pH, as encountered near bacterial
cell membranes influenced by the proton motive force (pH 3), or inside lysosomes of eukaryotic
cells (pH 5.5) or in bicarbonate-buffered natural water (pH 8.1). Ag
+
release increases with
decreasing pH.

REFERENCES
(1) Hiramatsu, H.; Osterloh, F. E. Chem. Mater. 2004, 16, (13), 2509-2511.
(2) Yang, X. Environ. Sci. Technol. 2012.
9

(3) Meyer, J. N.; Lord, C. A.; Yang, X. Y. Y.; Turner, E. A.; Badireddy, A. R.; Marinakos, S.
M.; Chilkoti, A.; Wiesner, M. R.; Auffan, M. Aquat. Toxicol. 2010, 100, (2), 140-150.
(4) Laban, G.; Nies, L. F.; Turco, R. F.; Bickham, J. W.; Sepulveda, M. S. Ecotoxicology 2010,
19, (1), 185-195.
(5) Lara, H. H.; Ayala-Nunez, N. V.; Ixtepan-Turrent, L.; Rodriguez-Padilla, C. J
Nanobiotechnology 2010, 8, 1.
(6) Beer, C.; Foldbjerg, R.; Dang, D. A.; Autrup, H. Toxicol. Lett. 2011, 205, S44-S44.
(7) Xiu, Z. M.; Ma, J.; Alvarez, P. J. J. Environ. Sci. Technol. 2011, 45, (20), 9003-9008.
(8) Chen, J.; Xiu, Z.; Lowry, G. V.; Alvarez, P. J. Water Res. 2010, 45, (5), 1995-2001.
(9) Adams, L. K.; Lyon, D. Y.; Alvarez, P. J. J. Water Res. 2006, 40, (19), 3527-3532.
(10) Poch, G.; Vychodil-Kahr, S.; Petru, E. Int. J. Clin. Pharm. Th. 1999, 37, (4), 189-192.
(11) SAS/STAT(R) 9.2 User's Guide, Second Edition, the NLIN Procedure,
http://support.sas.com/documentation/cdl/en/statug/63033/HTML/default/viewer.htm#nlin_t
oc.htm.