Saccharide-modified nanodiamond conjugates for the efficient detection and removal of pathogenic bacteria

Article abstract of DOI:10.1002/chem.201104069

The detection and removal of bacteria, such as E. coli in aqueous environments by using safe and readily available means is of high importance. Here we report on the synthesis of nanodiamonds (ND) covalently modified with specific carbohydrates (glyco-ND) for the precipitation of type 1 fimbriated uropathogenic E. coli in solution by mechanically stable agglutination. The surface of the diamond nanoparticles was modified by using a Diels-Alder reaction followed by the covalent grafting of the respective glycosides. The resulting glyco-ND samples are fully dispersible in aqueous media and show a surface loading of typically 0.1 mmol g^-1. To probe the adhesive properties of various ND samples we have developed a new sandwich assay employing layers of two bacterial strains in an array format. Agglutination experiments in solution were used to distinguish unspecific interactions of glyco-ND with bacteria from specific ones. Two types of precipitates in solution were observed and characterized in detail by light and electron microscopy. Only by specific interactions mechanically stable agglutinates were formed. Bacteria could be removed from water by filtration of these stable agglutinates through 10 μm pore-size filters and the ND conjugate could eventually be recovered by addition of the appropriate carbohydrate. The application of glycosylated ND allows versatile and facile detection of bacteria and their efficient removal by using an environmentally and biomedically benign material. Copyright

Full text of DOI:10.1002/chem.201104069

FULL PAPER
DOI: 10.1002/chem.201104069
Saccharide-Modified Nanodiamond Conjugates for the Efficient Detection
and Removal of Pathogenic Bacteria
Mirja Hartmann,[a] Patrick Betz,[b] Yuchen Sun,[b] Stanislav N. Gorb,[c]
Thisbe K. Lindhorst,*[a] and Anke Krueger*[b, d]
Abstract: The detection and removal
of bacteria, such as E. coli in aqueous
environments by using safe and readily
available means is of high importance.
Here we report on the synthesis of
nanodiamonds (ND) covalently modi-
fied with specific carbohydrates (glyco–
ND) for the precipitation of type 1 fim-
briated uropathogenic E. coli in solu-
tion by mechanically stable agglutina-
tion. The surface of the diamond nano-
particles was modified by using
a Diels–Alder reaction followed by the
covalent grafting of the respective gly-
cosides. The resulting glyco–ND sam-
ples are fully dispersible in aqueous
media and show a surface loading of
typically 0.1 mmolgÀ1. To probe the ad-
hesive properties of various ND sam-
ples we have developed a new sand-
wich assay employing layers of two
bacterial strains in an array format.
Agglutination experiments in solution
were used to distinguish unspecific in-
teractions of glyco–ND with bacteria
from specific ones. Two types of precip-
itates in solution were observed and
characterized in detail by light and
electron microscopy. Only by specific
interactions mechanically stable agglu-
tinates were formed. Bacteria could be
removed from water by filtration of
these stable agglutinates through 10 mm
pore-size filters and the ND conjugate
could eventually be recovered by addi-
tion of the appropriate carbohydrate.
The application of glycosylated ND
allows versatile and facile detection of
bacteria and their efficient removal by
using an environmentally and biomedi-
cally benign material.
Keywords: bacterial recognition
·
carbohydrates · diamond · nanopar-
ticles · surface chemistry
Introduction
urgent necessity. Most of the available techniques for the
sensing of bacterial contaminations require advanced labo-
ratory equipment or skills, and often several days for the
production of reliable results, which is critical in cases of
acute infections. Standard techniques for detection of bacte-
rial contamination include cultivation on agar plates,[5] meth-
ods based on DNA amplification,[6,7] such as PCR and LCR,
mass spectrometric detection[8] or immunological meth-
ods.[9,10]
Recently, nanoparticles (NP) have been reported to be
highly efficient in applications for sensing of microbes and
to require less costly equipment and/or time. Rotello and
collaborators developed a protocol based on functionalized
gold NPs (Au-NPs) carrying quaternary ammonium salts
triggering an enzyme-enhanced pathogen detection.[11]
Other reports on Au-NPs include the detection of viruses,[12]
bacteria[13] and proteins.[14] But not only Au-NPs can be ap-
plied for sensing, also macromolecular NPs[15] or NPs from
other materials, such as silver, gadolinium sulfide and iron
oxides or carbon nanomaterials have been applied for the
detection and capture of pathogenic bacteria.[16–24]
Bacterial contamination of water for human consumption or
irrigation of crops is one of the major global problems.[1–3]
Numerous bacterial species endanger the health and well-
being of millions of people all over the world in developing
countries, involving bacteraemia, meningitis and respiratory
tract infections. The problem also occurs in Europe, such as
in the case of recent EHEC infections.[4] Therefore, the effi-
cient detection and removal of such contaminations is an
[a] Dr. M. Hartmann, Prof. Dr. T. K. Lindhorst
Otto Diels Institute for Organic Chemistry
Christiana Albertina University Kiel
Otto-Hahn-Platz 3-4, 24098 Kiel (Germany)
E-mail: tklind@oc.uni-kiel.de
[b] P. Betz, Y. Sun, Prof. Dr. A. Krueger
Institute for Organic Chemistry, Wꢀrzburg University
Am Hubland, 97074 Wꢀrzburg (Germany)
E-mail: krueger@chemie.uni-wuerzburg.de
[c] Prof. Dr. S. N. Gorb
Department of Functional Morphology and Biomechanics
Zoological Institute, Christiana Albertina University Kiel
24098 Kiel (Germany)
However, many of the existing methods rely on purely
electrostatic interactions of the NPs with the respective vi-
ruses, microbes or proteins. This interaction is usually strong
but unselective. In other words, selectivity for certain types
of bacteria cannot be achieved. To solve the selectivity prob-
lem in sensing and removing bacteria an NP-based platform
is required that is conjugated to molecules that are species
[d] Prof. Dr. A. Krueger
Wilhelm Conrad Roentgen Research Center for Complex Materials
Systems
Wꢀrzburg University, 97074 Wꢀrzburg (Germany)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201104069.
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

specific. Carbohydrates are especially useful for this pur-
pose, as a majority of microbes show a pronounced specifici-
ty of certain saccharides. Thus, we set out to design suitable
nanoparticles that are terminated with specific saccharide
motifs (see below) for selective detection and/or removal of
different bacterial strains.
fied by a linker moiety to allow covalent conjugation with
a specific saccharide motif (Figure 1b). The glycosylated
ND can then interact with bacterial cells, such as fluorescing
E. coli (Figure 1c, d), through specific protein–carbohydrate
interactions. Binding to carbohydrates is a typical feature of
bacteria, which they utilize to attach to the glycosylated sur-
face of their target cells. Bacterial adhesion to cell surfaces
can eventually lead to the formation of biofilms and infec-
tion of the host organism.[29]
To accomplish adhesion effectively, many bacteria are
equipped with hairy protein appendages, called fimbriae or
pili. These adhesive organelles expose protein domains,
which function as lectins and thus mediate the molecular in-
teraction with specific carbohydrates for which they are se-
lective.[30] The best characterized pili are type 1 fimbriae,
which recognize terminal a-d-mannoside units in high-man-
nose type glycoproteins on the host-cell surface with high
specificity.[31] The lectin domain, which is known to be re-
sponsible for mannoside binding, is located at the fimbrial
tips and called FimH.[32] Type 1 fimbriae are common
throughout the family of Enterobacteriaceae, such as E. coli,
including highly pathogenic strains of UPEC and EHEC.
We, therefore, chose mannose specificity of bacterial binding
for the investigation of glycosylated ND in sensing and re-
moval of bacteria.
In this approach, the utilized nanoscale material needs to
fulfill several requirements: broad availability, easy and
stable surface functionalization, as well as dispersibility in
aqueous solution. In addition, it should be environmentally
and medically benign. A material that complies with all
these prerequisites is nanoscale diamond (ND). It is hydro-
philic, available in large quantities; it can be functionalized
in many ways and is considered a biocompatible and safe
material.[25–28] Nanodiamond possesses several advantages
over other nanoparticles and objects (like carbon nanotubes
or polymer NP), such as being a completely inert particle,
not swelling in any solvents, inherent fluorescence from lat-
tice defects (applicable in labeling applications), thermal
(up to 4508C) and mechanical long-term stability and very
low toxicity. There are several sources for nanoscale dia-
mond particles, such as high temperature high pressure
(HTHP) synthesis, chemical vapor deposition (CVD) and
detonation synthesis. The latter method can be used for the
production of large quantities of very small (5 nm) nanodia-
mond particles, which are typically strongly agglomerated
due to the highly functionalized surface. This is caused by
the oxygenation during synthesis (e.g., by reaction with the
coolant) or the purification process by using mineral
acids.[26,27] The surface of the nanodiamond can be function-
alized in many ways, and can lead to stable colloids in aque-
ous solution and a high flexibility in functional surface ter-
mination.
Results and Discussion
For the synthesis of the ND conjugates 1–3 (Scheme 1)
a suitably functionalized ND surface is needed. We applied
an efficient two-step strategy using a Diels–Alder cycloaddi-
tion as its central reaction recently reported by some of
us.[33] Thermally annealed detonation ND (8)[34] was treated
with in situ generated ortho-quinodimethane to yield a cova-
lently arylated ND (9) with surface loading of 0.14 mmolgÀ1.
After a classical aromatic sulfonation, one third of the sul-
fonic acid groups was reduced by treatment with triphenyl-
phosphine and iodine (according to X-ray photoelectron
spectroscopy (XPS), Figure S6 in the Supporting Informa-
tion).[33] This leaves a sufficient amount of charged surface
groups for the colloidal stabilization of the NPs in aqueous
environments. The thiol functions of compound 11 were
used as anchor groups. They allow the conjugation of the
saccharide moieties on the diamond surface by “thiol–ene
reaction” to the allyl aglycon of glycosides 6 and 7.[35] The
allyl glycosides 6 and 7 and the trivalent cluster mannoside
14 were synthesized according to published protocols.[36,37]
Allyl glycosides were ligated to the thiolated ND 11 to yield
the conjugates 1 and 2 after removal of the protecting
groups (see the Supporting Information). The trivalent gly-
cocluster 14 was attached to an arylated ND carrying car-
boxylic acid moieties (15)[38] by peptide coupling employing
the amino group at the focal point of the cluster mannoside
with EEDQ as the coupling agent (see the Supporting Infor-
mation). All ND conjugates were characterized by FTIR,
TGA, and z potential measurements as well as elemental
Here we report on the synthesis and application of
a novel, saccharide-modified ND material for the detection
and agglutination of bacteria. The synthetic construct con-
sists of a nanoscale diamond core (Figure 1a) that is modi-
Figure 1. Interaction entities: a) TEM image of nanodiamond (ND);
b) schematic representation of the glyco–ND conjugate; c) TEM image
of the GFP-tagged type 1 fimbriated E. coli strain PKL1162;[39] d) sche-
matic image of the same bacteria.
&
2
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Saccharide-Modified Nanodiamond Conjugates
FULL PAPER
allow for the application of
these ND glycoconjugates in
detection and agglutination of
such pathogens.
To test the specific interac-
tion of the new ND glycoconju-
gates with bacteria, they were
initially tested as inhibitors of
type 1 fimbriae-mediated bacte-
rial adhesion.[39] However, the
known
adhesion-inhibition
assay, which is routinely per-
formed to test inhibitors of bac-
terial adhesion,[40] could not be
applied to the new ND glyco-
conjugates, due to their aggluti-
nating properties (see below;
Figure S10 in the Supporting
Information). Therefore, a new
sandwich assay was developed,
that is composed of several
layers: 1) the microtiter plate
with
the
polysaccharide
mannan immobilized on its sur-
face; 2) a “capture layer” of
type 1
fimbriated
bacteria,
which is the actual analyte;
3) the ND conjugates; and 4) a
“detection layer” of fluorescent
type 1 fimbriated bacteria. In
this setup, the ND conjugates
function as “glue” between
a
layer of non-fluorescent
E. coli and a top layer of fluo-
rescing bacteria, which can be
analyzed by fluorescence read-
out (Figure 3). In this assay ag-
glutination is prevented. Hence,
it can be applied to screen the
NPsꢂ potential to form a stable
adhesive layer between bacteri-
al cells, both in an a-d-manno-
side-specific manner, as well as
Scheme 1. Synthesis of ND conjugates. a) Synthesis of nanodiamond 1 and 2 functionalized with monosacchar-
ides: i) 1. allyl alcohol, acetyl chloride, 08C to 708C, 55%; 2. Ac2O, py, room temperature, 98%; ii) [18]crown-
6, KI, toluene, 72 h; iii) SO3HCl, 508C, 20 h; iv) PPh3, I2, benzene, reflux, 20 h. b) Synthesis of nanodiamond 3
functionalized with a trivalent cluster mannoside: i) 1. BF3·Et2O, CH2Cl2, 08C to room temperature, 38 h, 43%;
2. NaOMe, MeOH, room temperature, 23 h, 94%; 3. H2, Pd/C, MeOH, room temperature, 26 h, 73%; ii) iso-
ACHTUNGTRENNUNGamylnitrite, H2O, 80 8C, 16 h; iii) EEDQ, pyridine, reflux, 24 h.
by
unspecific
interactions.
Moreover, future applications
will include the detection of
analysis (Figures S1–S4, Table S1 in the Supporting Informa-
tion). Their full dispersion as isolated nanoparticles was
shown by dynamic light scattering (DLS) and AFM meas-
urements (Figure 2, Figures S7–S9 in the Supporting Infor-
mation).
The glycosylated NDs can be regarded as mimetics of the
host cells of bacteria, as the carbohydrate decoration of the
employed ND conjugates resembles parts of the eukaryotic
glycocalyx. Hence, it could be predicted that the interaction
of fimbriated bacteria with glycosylated ND should be com-
parable to their interaction with living cells. This should
pathogenic bacteria by this sandwich setup.
The assay was performed as follows. E. coli bacteria of
the non-fluorescent type 1 fimbriated strain HB101
pPKL4[41] (“capture layer”) were allowed to adhere a-d-
mannoside-specifically to mannan-coated 96-well plates in
an incubation step at physiological temperature (378C).
After being washed, serial dilutions of the stably dispersed
ND sample in PBS buffer were applied and incubated at
378C. Unbound ND was washed away, before the wells
were coated with a suspension of the GFP-tagged E. coli
strain PKL1162[40] that in turn adhere to bound ND. Thus,
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

A. Krueger, T. K. Lindhorst et al.
this assay. Negative controls, in which the ND adhesive
layer was omitted, only showed negligibly low background
fluorescence signal.
Overall it was shown that our sandwich assay serves as
a reliable setup to test the interaction of various functional-
ized NPs (not only ND!) with living bacteria. It turned out,
that both, the saccharide-modified NDs 1–3, as well as non-
glycosylated NDs function as robust adhesives between bac-
terial analyte and detection layer. However, mannosylated
ND 1 and 3 formed adhesive films already after a very short
incubation time, owing to specific interactions between man-
nose and the fimbrial lectin, FimH. Unspecific interactions
with non-mannosylated ND, such as with 2, 8, and 10, on
the other hand, were built up more slowly, as could be seen
in incubation-duration studies (Figure S12 in the Supporting
Information). Moreover, sandwich assays with mannoside-
modified ND 1 and 3 generated fluorescence values, which
were 30% higher than the values obtained with 2, 8, and 10.
These non-glycosylated NDs can only establish unspecific
interactions with the bacteria, for example, by hydrophilic
interactions of their surface groups with the pathogens.
Thus, specific and nonspecific interactions of E. coli with
ND material occur concomitantly on surfaces, as it was
shown in the sandwich assay. In order to investigate ND–
bacteria interactions in solution, three-dimensional aggluti-
nation experiments appeared to be perfectly suited. As
known from classical hemagglutination[42] and agglutination
studies[43] cells can be precipitated through lectin–carbohy-
drate mediated interactions. To get deeper insight into the
influence of specific versus nonspecific adhesion, the type 1
fimbriated E. coli strain PKL1162 was used to test mannose-
specific interaction with the ND conjugates, and the related
non-fimbriated strain HB101 was employed to probe unspe-
cific interactions of the bacteria with different ND samples.
A suspension of fimbriated E. coli PKL1162 and six dif-
ferent ND samples (Figure 4; Table S2 and Figures S13–S17
in the Supporting Information) were serially diluted in
bidest. water and mixed in test vials at ambient temperature.
Equally, a suspension of non-fimbriated E. coli HB101 was
mixed with serial dilutions of mannosylated ND 1. Already
after a short incubation time of 5 min, precipitation was ob-
served in all cases, correlating with the ND concentrations
applied. However, the formed precipitates differed much in
their visual characteristics as well as in their mechanical sta-
bility depending on the applied ND. Whereas mannosylated
ND 1 and 3 formed well-structured crystal-like agglutinates
with type 1 fimbriated E. coli (Figure 4b, i), less compact
flocculates were formed as a consequence of unspecific in-
teractions of non-mannosylated ND (Figure 4b, ii). More-
over, mannosylated ND samples 1 and 3 led to mechanically
stable agglutinates in clear supernatant solution, which re-
mained unaffected even after heavy manual shaking of the
test vials. Apparently, strong cross-links were formed that
withstood shear stress. In contrast, flocculates resulting from
nonspecific interaction of the non-glycosylated ND samples
8, 10, or pristine hydrophilic ND (16, see the Supporting In-
formation) with the bacteria were mechanically less stable
Figure 2. AFM image of mannosylated nanodiamond 3 dropcasted on
highly ordered pyrolytic graphite (HOPG). The overview (left) shows an
even distribution of the particles. The magnification proves the absence
of agglomerates (right).
Figure 3. Setup of the sandwich assay: type 1 fimbriated E. coli (HB101
pPKL4) are immobilized onto polysaccharide (mannan)-coated microtit-
er plates. Then, serial dilutions of ND conjugates are applied; this leads
to an adhesive layer, to which fluorescent type 1 fimbriated E. coli
(PKL1162) can adhere as the detection layer. Thus, fluorescence intensity
is correlated with the adhesive properties of the applied ND conjugates.
GFP-tagged E. coli serve as the “detection layer” (Figure 3)
and the intensity of the fluorescence signal can be correlated
with the adhesive properties of ND. This novel sandwich
assay showed that ND form adhesive films between two
bacterial layers in a concentration-dependent manner (Fig-
ure S11 in the Supporting Information).
For all ND samples under investigation the measured
fluorescence was plotted against the respective applied ND
concentration on a logarithmic scale. Sigmoidal dose-re-
sponse curves were obtained; these indicate that the adhe-
sive properties of the GFP-tagged bacteria in the detection
layer are strictly dependent on the concentration of applied
ND (Figure S11 in the Supporting Information). When the
applied ND concentration was raised over a certain limit,
fluorescence did not further increase but leveled out at a pla-
teau. This observation indicates that the bacterial detection
layer does not expand beyond a monolayer but reaches
a maximum with maximal adhesion of ND. Correspondingly,
uncontrolled agglomeration of bacteria was not observed in
&
4
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Saccharide-Modified Nanodiamond Conjugates
FULL PAPER
with bacteria was shown to parallel with known standard
models of lectin–saccharide interactions. Glyco–ND can
thus be considered a suitable model for the investigation of
carbohydrate–protein interactions.
In order to test the efficiency of glyco–ND to remove bac-
teria from polluted solutions, samples of supernatants after
flocculation/agglutination were grown on agar. For all ND
samples (1, 2, 3, 8, and 16) a reduction of the amount of
E. coli in the remaining supernatant was observed (see the
Supporting Information). By far the most potent water de-
contaminating agents were shown to be the mannosylated
NDs 1 and 3. In accordance with the multivalency effect,
which is often observed in carbohydrate recognition,[45] ND
3, functionalized with the trivalent cluster mannoside per-
forms better than ND 1, which is conjugated to monovalent
mannosides.
After the basic applicability of ND conjugates for the ag-
glutination of bacteria had been shown, we set out to go
beyond the proof-of-principle. We aimed at the use of these
agglutinants in a real-world setup for the removal of patho-
gens from water. In fact, agglutination-filtration experiments
employing commercial filters (pore size 10 mm; too big to
retain bacteria) and ND 1 or pristine hydrophilic ND 16, al-
lowed the removal of different types of E. coli efficiently
from various water samples. Retention of bacteria clearly
correlated with ND concentration and ND type (Table S3,
Figures S19–S20 in the Supporting Information). The most
suitable concentration of ND 1 for the efficient retention of
bacteria was found to lie between 0.08–0.8 mgmgÀ1 E. coli
with removal rates of 93–97% (Figure 5). This was shown
by cultivation and CFU counting of the generated filtrates
in comparison to the original untreated water sample. To ex-
Figure 4. Detection and removal of bacterial cells. a) Agglutination of
bacteria in solution by using surface-modified ND. Depending on the sur-
face functionality a reversible flocculation caused by nonspecific interac-
tions or an irreversible agglutination due to stable mannose–lectin inter-
actions is observed. The latter can be used not only for the specific detec-
tion of bacteria, but also for their removal by filtration. b) Mechanically
stable agglutination with mannosylated ND 3 (i) and unstable floccula-
tion with glucosylated ND 2 (ii). c) The carbohydrate-specific agglutina-
tion of E. coli by mannosylated ND 1 is reversible by addition of a com-
peting carbohydrate in solution, such as MeMan. Only in samples with
MeMan (+MeMan) the agglutinates can be redispersed by manual shak-
ing (control: ÀMeMan). Pictures in b) and c) were taken 5 min after
shaking the vials. Samples in c) were shaken simultaneously with the
same strength.
and could be easily disintegrated by manually shaking the
vial to form a milky dispersion. The same unstable floccula-
tion was observed with non-fimbriated HB101 mixed with
mannosylated ND 1 (Figure S17 in the Supporting Informa-
tion). These observations again indicate that for the forma-
tion of stable agglutinates specific interactions are required.
The glucosylated ND 2 generated precipitates with an in-
termediate stability: lower than for the stable agglutinates
of mannosylated samples 1 and 3 and higher than for the
flocculates of non-glycosylated ND. Accordingly, agitation
of the agglutinate formed with 2 and fimbriated E. coli led
to a cloudy dispersion with some smaller agglutinates re-
maining. This can be explained by weak interactions be-
tween glucosides and the bacterial lectin FimH, as it has
been recently reported by some of us.[44]
As expected, mannose specific agglutination of ND could
be prevented by addition of a standard inhibitor of type 1
fimbriae-mediated bacterial adhesion, such as methyl a-d-
mannopyranoside (MeMan; 100 mm). Likewise, agglutina-
tion could be reversed by the addition of MeMan (Fig-
ure 4c). Hence, the interaction of the mannosylated ND
Figure 5. The agglutination-filtration experiments with E. coli show that
the addition of just 80 mgmLÀ1 of the mannosylated ND 1 is sufficient to
efficiently remove 1000 mgmLÀ1 of fimbriated bacteria PKL1162 from
the solution by filtration through a conventional filter with 10 mm pore
size. Further results including findings with non-fimbriated bacteria and
unmodified hydrophilic ND 16 are shown in Figure S19 (in the Support-
ing Information).
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!

A. Krueger, T. K. Lindhorst et al.
clude any effect of the filter, the control was equally fil-
tered.
In multiply contaminated samples, specific interactions
between carbohydrates attached to ND 1 and the bacterial
type 1 fimbriae of E. coli enabled the preferential removal
of the fimbriated strain PKL1162 (Figure S19 in the Sup-
porting Information). This ND conjugate is hence applicable
for the selective removal of specific pathogens in a very
simple setup. After filtration from the water samples, it was
possible to recycle the glyco–ND from the agglutinates by
washing with MeMan solution. The recovered ND 1 was
reused for another efficient removal of bacteria (Figure S21
in the Supporting Information). This showed the practical
applicability of our material, for example, in filter cartridges.
Recovery of the ND material after agglutination of bacteria
is of high importance if the material is used in a continuous-
flow setup. Thus, in future, for continuous removal of patho-
gens, for example, from drinking water, sanitation cartridges
could be reused after regeneration. This method would
reduce costs and enable the long-term application of glyco–
ND in difficult environments.
To get deeper insight into the nature of the observed floc-
culates and agglutinates, they were investigated by optical
and electron microscopy. In phase contrast microscopic
images of the different types of flocculates it could be clear-
ly seen that the non-fimbriated strain HB101 was not at all
agglutinated by the mannosylated ND 1. As expected, bac-
terial cells and glyco–ND particles were found mobile and
separated from each other. The fimbriated strain PKL1162,
on the other hand, was aggregated in extended lattices. A
third type of interaction with PKL1162 was seen with pris-
tine hydrophilic ND 16. Unspecific flocculates were formed,
presumably due to electrostatic interactions with the bacte-
ria.
Two types of the observed precipitates were also investi-
gated by using scanning electron microscopy (SEM, Figure 6
and Figures S22–S25 in the Supporting Information): the ag-
glutinate that was formed through specific interaction of
glyco–ND 1 and PKL1162, and the flocculates that were
generated when the non-mannosylated ND 16 interacted
with PKL1162. From the SEM images it could be seen that
non-mannosylated, surface oxidized (and hence hydrophilic)
ND 16 accumulates on the bacterial surface in large aggre-
gates (Figure 6a), presumably through nonspecific hydro-
philic interactions. On the contrary, mannosylated ND 1 was
evenly spread around the bacteria and their fimbriae (Fig-
ure 6b). As expected, glyco–ND was seen to specifically in-
teract with the fimbrial tip (Figure 6c), where the mannose-
specific lectin domain FimH is located. Individual ND parti-
cles were also seen to adhere to the fimbrial shaft (Fig-
ure 6b). Thus, SEM clearly revealed the structural differen-
ces between unspecific formation of less stable flocculates
and carbohydrate-specific generation of stable agglutinates
when bacteria are treated with different NDs.
Figure 6. SEM images of flocculates and agglutinates formed with:
a) E. coli and pristine hydrophilic nanodiamond 16. The ND particles
form large agglomerates and settle unspecifically on the cell surface.
b) Mannosylated ND 1. The particles are well dispersed and are located
on the fimbriae. c) Mannosylated ND 1. The image shows a magnification
of ND 1 specifically bound to the bacterial lectin located on the tip of
the type 1 fimbriae (arrow).
Conclusion
In conclusion, we have synthesized a series of novel nano-
diamond conjugates that are covalently modified with sac-
charide moieties using a stable ligation by an aromatic ring
introduced by a Diels–Alder reaction. As the actual surface
decoration can be flexibly chosen according to the targeted
interaction, the synthesis strategy is suited for a wide range
of applications. Depending on the respective surface termi-
nation the ND conjugates can either interact specifically or
nonspecifically with the surface of different pathogenic bac-
terial strains. The agglutination experiments with E. coli cor-
roborate ND particles as a useful means to flocculate bacte-
ria and thereby allow their removal by filtration through
a low-cost, high-throughput filter material with large pore
size. Decorated with specific carbohydrates, it is possible to
use them for the detection and removal of the respective
bacteria in a contaminated sample by the formation of me-
chanically stable agglutinates. Once separated from the mix-
ture by filtration, glyco–ND can easily be recycled by addi-
tion of a low-cost, commercially available inhibitor of bacte-
rial adhesion such as MeMan. This makes the procedure of
bacterial decontamination in solution an economic and envi-
ronmentally friendly process. Furthermore, our methodology
allows the detection of bacterial contamination by simple
visual inspection of the test solution. A novel sandwich
assay applying functionalized ND can be used for the cap-
ture and detection of bacteria; this overcomes agglomera-
tion issues observed in conventional inhibition assays. Both
setups enable a multitude of applications, such as detection
of pathogens and/or decontamination of water. Other bio-
logical applications, such as the investigation of carbohy-
&
6
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Saccharide-Modified Nanodiamond Conjugates
FULL PAPER
tion ND 8 was applied in the following concentrations: 10, 1 mgmLÀ1
,
drate-specific interactions, in vivo and in vitro, can be envis-
aged taking advantage, for example, of the non-bleaching lu-
minescence of ND carrying N-V centers.[46,47] In the future
we will study the use of the ND conjugates employing a vari-
ety of specific glycosides both in continuous filtering car-
tridges as well as in the context of glycobiology.
100, 10, and 1 mgmLÀ1. Type 1 fimbriated E. coli (PKL1162) were sus-
pended in bidest. water at a concentration of 2 mgmLÀ1. Then, this bacte-
rial suspension (900 mL) was combined with the respective ND sample
(100 mL) and the resulting agglomeration was studied by visual inspection
(Table S2 in the Supporting Information). Details of filtration and recy-
cling experiments are described in the Supporting Information.
Abbreviations: CFU, colony forming units; DLS, dynamic light scatter-
ing; EEDQ, 2-ethoxy-1-ethoxycarbonyl-1,2-dihydroquinoline; E. coli, Es-
cherichia coli; EHEC, enterohemorrhagic E. coli; FTIR, Fourier trans-
form infrared spectroscopy; LCR ligase chain reaction; MeMan, methyl
a-d-mannoside; ND, nanodiamond; NP, nanoparticle, PCR, polymerase
chain reaction; PBS, phosphate-buffered saline; PBST, PBS with 0.05%
Tween-20; TGA, thermogravimetric analysis; TLC, thin layer chromatog-
raphy; UPEC, uropathogenic E. coli.
Experimental Section
Synthesis of mannosylated ND 1: Under nitrogen atmosphere thioaryl-
ACHTUNGTRENNUNGated nanodiamond 11 (59 mg) was suspended in anhydrous benzene in an
ultrasonic bath for 15 min. Allyl 2,3,4,6-tetra-O-acetyl-a-d-mannopyrano-
side (81 mg, 0.209 mmol) and a catalytic amount of AIBN were added to
the suspension. The reaction mixture was heated, overnight, to 788C.
After the suspension cooled to room temperature, the diamond particles
were isolated by centrifugation. The precipitate was washed repeatedly
with toluene, acetone and dichloromethane in consecutive washing/cen-
trifugation cycles. Ultrasonic treatment was used in every cycle in order
to redisperse the diamond and remove adsorbed impurities. After being
washed, the sample was dried at 708C, in vacuo. The mannosylated nano-
diamond 1 (170 mg) was suspended in a mixture of methanol (4 mL),
water (5 mL) and potassium hydroxide solution (1m, 2 mL) in an ultra-
sonic bath for 15 min. The suspension was stirred, overnight, at room
temperature and the diamond particles were isolated by centrifugation.
The precipitate was washed repeatedly with water (until the supernatant
became neutral), acetone and dichloromethane in consecutive washing/
centrifugation cycles. Ultrasonic treatment was used in every cycle in
order to redisperse the diamond and remove adsorbed impurities. After
being washed, the sample was dried at 708C, in vacuo. FT-IR (vacuum-
cell): n˜ =3417, 2934, 2881, 1718, 1583, 1373, 1195, 1120, 1040 cmÀ1; ele-
mental analysis calcd (%): C 90.38, H 1.08, N 2.70, S 0.31; surface load-
ing: (calcd from TGA): 0.11 mmolgÀ1 [Dm (137–4898C) 4.0%]; frag-
ment: C17H24O6S (356 gmolÀ1); particle size: (H2O): 10%ꢀ44, 50%ꢀ59,
90%ꢀ96 nm.
Acknowledgements
We gratefully acknowledge the funding of the Deutsche Forschungsge-
meinschaft (T.K.L., S.N.G. and A.K.) and the European Commission
(A.K., contract DINAMO). We thank D. Lang for the HRTEM images
of the NDs, S. Uemura for the AFM, and Pavo Vrdoljak for XPS meas-
urements (all Wꢀrzburg).
[1] W. A. Suk, K. M. Ruchirawat, K. Balakrishnan, M. Berger, D. Car-
penter, T. Damstra, J. Pronczuk de Garbino, D. Koh, P. J. Landrigan,
I. Makalinao, P. D. Sly, Y. Xu, B. S. Sheng, Environ. Health Perspect.
2003, 111, 1340–1347.
[2] a) P. Rochelle, J. Clancey, J. Am. Water Works Assoc. 2006, 98, 163–
166; b) P. Rochelle, J. Clancey, J. Am. Water Works Assoc. 2006, 98,
169–174; c) P. Rochelle, J. Clancey, J. Am. Water Works Assoc.
2006, 98, 177–191.
[3] Pathogenic Mycobacteria in Water (Eds.: S. Pedley, J. Bartram, G.
Rees, A. Dufour, J. A. Cotruvo), World Health Organization, IWA,
London, 2004.
[4] A. R. C. Wong, J. S. Pearson, M. D. Bright, D. Munera, K. S. Robin-
son, S. F. Lee, G. Frankel, E. L. Hartland, Mol. Microbiol. 2011, 80,
1420–1438.
[5] B. S. Stevenson, S. A. Eichorst, J. T. Wertz, T. M. Schmidt, J. A.
Breznak, Appl. Environ. Microbiol. 2004, 70, 4748–4755.
[6] S. Bae, S. Wuertz, Appl. Environ. Microbiol. 2009, 75, 2940–2944.
[7] M. R. Pingle, K. Granger, P. Feinberg, R. Shatsky, B. Sterling, M.
Rundell, E. Spitzer, D. Larone, L. Golightly, F. Barany, J. Clin. Mi-
crobiol. 2007, 45, 1927–1935.
[8] N. J. C. Strachan, F. J. Nicholson, I. D. Ogden, Anal. Chim. Acta
1995, 313, 63–67.
[9] S. C. Ricke, S. D. Pillai, Crit. Rev. Microbiol. 1999, 25, 19–38.
[10] I. Safarik, M. Safarikova, S. J. Forsythe, J. Appl. Bacteriol. 1995, 78,
575–585.
The synthesis of all other ND conjugates and the saccharide precursors is
described in the Supporting Information.
Three-layered sandwich assay by using nanodiamond as adhesive layer:
To prove the cross-linking feature of the NDs towards E. coli bacteria,
the nanoparticles were assayed in a novel sandwich setup. Black 96-well
polystyrene microtiter plates (Thermo Fisher Scientific, Nunc, Maxisorp)
were coated with mannan solution (1.2 mgmLÀ1 in carbonate buffer
pH 9.5, 120 mL per well) and dried, overnight, at 378C. The plates were
washed three times with PBST (150 mL per well). Non-fluorescent type 1
fimbriated E. coli (strain pPKL4) were suspended in PBS buffer
(2.2 mgmLÀ1) and incubated in the mannan-coated wells (100 mL per
well) for 2 h at 378C to form a first bacterial layer—the capture layer.
The plates were washed three times with PBS (150 mL per well). Glyco–
ND or ND suspensions in PBS were prepared by applying ultrasound for
2 h. These suspensions were serially diluted in PBS and transferred to the
pretreated microtiter plate (50 mL per well). The plate was incubated at
378C for 1 h to allow the formation of the adhesive layer and washed
three times with PBS (100 mL per well). GFP-tagged bacteria (strain
PKL1162) were suspended in PBS (2 mgmLÀ1), dispensed to the wells
(50 mL per well) and incubated for 1 h at 378C to build up the fluorescing
detection layer. The plate was washed three times with PBS (100 mL per
well) and each well was filled with PBS (50 mL) before bacterial adhesion
was detected by fluorescence readout (excitation wavelength 485 nm,
emission wavelength 535 nm). The dose-response curves are shown in the
Supporting Information.
[11] O. R. Miranda, X. Li, L. Garcia-Gonzalez, Z.-J. Zhu, B. Yan,
U. H. F. Bunz, V. M. Rotello, J. Am. Chem. Soc. 2011, 133, 9650–
9653.
[12] J. Griffin, A. K. Singh, D. Senapati, E. Lee, K. Gaylor, J. Jones-
Boone, P. C. Ray, Small 2009, 5, 839–845.
[13] H.-W. Cheng, S.-Y. Huan, H.-L. Wu, G.-L. Shen, R.-O. Yu, Anal.
Chem. 2009, 81, 9902–9912.
[14] C. C. You, O. R. Miranda, B. Gider, P. S. Ghosh, I.-B. Kim, B. Erdo-
gan, S. A. Krovi, U. H. F. Bunz, V. M. Rotello, Nat. Nanotechnol.
2007, 2, 318–323.
Agglutination/precipitation experiments: To evaluate the adhesive poten-
tial of mannosylated ND, ND suspensions in double distilled water were
prepared by applying ultrasound for 2 h. The NDs were diluted to give
five samples at different concentrations for each functionalized ND. The
concentrations used were the same for all hydrophilic NDs (1, 2, 3, 10,
16): 5 mgmLÀ1, 500, 50, 5, and 0.5 mgmLÀ1. Thermally annealed detona-
[15] M. U. Larsen, M. Seward, A. Tripathi, N. C. Shapley, Biotechnol.
Prog. Biotechnol. Progr. 2009, 25, 1094–1102.
[16] G. Naja, P. Bouvrette, J. Champagne, R. Brousseau, J. H. T. Luong,
Appl. Biochem. Biotechnol. 2010, 162, 460–475.
[17] V. Parashar, S. K. Pandey, A. C. Pandey, Chem. Commun. 2010, 46,
3143–3145.
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
7
&
ÞÞ
These are not the final page numbers!

A. Krueger, T. K. Lindhorst et al.
[18] A. Fornara, P. Johansson, K. Petersson, S. Gustafsson, J. Qin, E.
Olsson, D. Ilver, A. Krozer, M. Muhammed, C. Johansson, Nano
Lett. 2008, 8, 3423–3428.
[32] D. Choudhury, Science 1999, 285, 1061–1066.
[33] G. Jarre, Y. Liang, P. Betz, D. Lang, A. Krueger, Chem. Commun.
2011, 47, 544–546.
[19] H. Lee, T.-J. Yoon, R. Weissleder, Angew. Chem. 2009, 121, 5767–
5770; Angew. Chem. Int. Ed. 2009, 48, 5657–5660.
[34] Y. Liang, T. Meinhardt, G. Jarre, M. Ozawa, P. Vrdoljak, A. Schçll,
F. Reinert, A. Krueger, J. Colloid Interface Sci. 2011, 354, 23–30.
[35] A. Dondoni, A. Marra, Chem. Soc. Rev. 2012, 41, 573–586.
[36] F. M. Winnik, J. P. Carver, J. J. Krepinsky, J. Org. Chem. 1982, 47,
2701–2707.
[20] A. L. Chun, Nat. Nanotechnol. 2009, 4, 698–699.
[21] S. Y. Hong, G. Tobias, K. T. Al-Jamal, B. Ballesteros, H. Ali-Boucet-
ta, S. Lozano-Perez, P. D. Nellist, R. B. Sim, C. Finucane, S. J.
Mather, M. L. H. Green, K. Kostarelos, B. G. Davis, Nat. Mater.
2010, 9, 485–490.
[37] H. A. Shaikh, F. D. Sçnnichsen, T. K. Lindhorst, Carbohydr. Res.
2008, 343, 1665–1674.
[22] P. Wu, X. Chen, N. Hu, U. C. Tam, O. Blixt, A. Zettl, C. R. Bertozzi,
Angew. Chem. 2008, 120, 5100–5103; Angew. Chem. Int. Ed. 2008,
47, 5022–5025.
[23] E. Perevedentseva, P.-J. Cai, Y.-C. Chiu, C.-L. Cheng, Langmuir
2011, 27, 1085–1091.
[38] T. Meinhardt, D. Lang, H. Dill, A. Krueger, Adv. Funct. Mater.
2011, 21, 494–500.
[39] M. Hartmann, T. K. Lindhorst, Eur. J. Org. Chem. 2011, 3583–3609.
[40] M. Hartmann, A. K. Horst, P. Klemm, T. K. Lindhorst, Chem.
Commun. 2010, 46, 330–332.
[24] a) B. K. Gorityala, J. Ma, X. Wang, P. Chen, X.-W. Liu, Chem. Soc.
Rev. 2010, 39, 2925–2934; b) K. Kostarelos, A. Bianco, M. Prato,
Nat. Nanotechnol. 2009, 4, 627–633.
[25] a) A. Krueger, J. Mater. Chem. 2011, 21, 12571–12578; b) V. N. Mo-
chalin, O. Shenderova, D. Ho, Y. Gogotsi, Nat. Nanotechnol. 2012, 7,
11–23; c) A. Krueger, D. Lang, Adv. Funct. Mater. 2012, 22, 890–
906.
[41] M. A. Schembri, G. Christiansen, P. Klemm, Mol. Microbiol. 2001,
41, 1419–1430.
[42] H. B. Donald, A. Isaacs, J. Gen. Microbiol. 1954, 10, 457–464.
[43] N. Firon, I. Ofek, N. Sharon, Carbohydr. Res. 1983, 120, 235–249.
[44] C. Grabosch, M. Hartmann, J. Schmidt-Lassen, T. K. Lindhorst,
ChemBioChem 2011, 12, 1066–1074.
[45] J. J. Lundquist, E. J. Toone, Chem. Rev. 2002, 102, 555–578.
[46] O. Faklaris, D. Garrot, V. Joshi, F. Druon, J.-P. Boudou, T. Sauvage,
P. Georges, P. A. Curmi, F. Treussart, Small 2008, 4, 2236–2239.
[47] Y.-R. Chang, H.-Y. Lee, K. Chen, C.-C. Chang, D.-S. Tsai, D.-S. Fu,
T.-S. Lim, Y.-K. Tzeng, C.-Y. Fang, C.-C. Han, H.-C. Chang, W.
Fann, Nat. Nanotechnol. 2008, 3, 284–288.
[26] A. M. Schrand, S. A. C. Hens, O. A. Shenderova, Crit. Rev. Solid
State Mater. Sci. 2009, 34, 18–74.
[27] A. Krueger, Chem. Eur. J. 2008, 14, 1382–1390.
[28] V. Vaijayanthimala, H.-C. Chang, Nanomedicine 2009, 4, 47–55.
[29] N. Sharon, Biochim. Biophys. Acta 2006, 1760, 527–537.
[30] K. Drickamer, Nat. Struct. Biol. 1995, 2, 437–439.
[31] G. Zhou, W.-J. Mo, P. Sebbel, G. Min, T. A. Neubert, R. Glockshub-
er, X.-R. Wu, T.-T. Sun, X.-P. Kong, J. Cell Sci. 2001, 114, 4095–
4103.
Received: December 29, 2011
Published online: && &&, 0000
&
8
&
www.chemeurj.org
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Saccharide-Modified Nanodiamond Conjugates
FULL PAPER
Carbohydrates
M. Hartmann, P. Betz, Y. Sun,
S. N. Gorb, T. K. Lindhorst,*
A. Krueger* ................... &&&& — &&&&
Saccharide-Modified Nanodiamond
Conjugates for the Efficient Detection
and Removal of Pathogenic Bacteria
Sweet diamond: Novel nanodiamond–
saccharide conjugates have been syn-
thesized and applied for the detection
and removal of pathogenic bacteria. A
new sandwich assay enables this detec-
tion by specific and nonspecific inter-
actions by using agglutinating nanopar-
ticles. The specific flocculation of
nanodiamond–mannose conjugates
with bacteria was used for the efficient
and repeated removal of these germs
from water samples (see scheme).
A carbohydrate-modified nanodia-
mond…
…is a valuable new material for
the straightforward detection and
efficient removal of pathogenic
bacteria from polluted water
sources. It is nontoxic and can
readily be used in difficult environ-
ments. Furthermore, sugar-specif-
icity of bacterial surface proteins
allows for identification and
targeted sectioning of particular
virulent strains through tailored
glycosylation of the nanodiamond
surface. For more details see the
Full Paper by A. Krueger, T. K.
Lindhorst et al. on page && ff.
Chem. Eur. J. 2012, 00, 0 – 0
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
9
&
ÞÞ
These are not the final page numbers!