Catalyst-free conjugation and in situ quantification of nanoparticle ligand surface density using fluorogenic Cu-free click chemistry

Article abstract of DOI:10.1002/chem.201003131

A highly efficient method for functionalizing nanoparticles and directly quantifying conjugation efficiency and ligand surface density has been developed. Attachment of 3-azido-modifed RGD-peptides to PEGylated liposomes was achieved by using Cu-free click conditions. Upon coupling a fluorophore is formed, which could be utilized to monitor conjugation efficiency and the obtained ligand surface density in situ, without prior purification. Copyright

Full text of DOI:10.1002/chem.201003131

DOI: 10.1002/chem.201003131
Catalyst-Free Conjugation and In Situ Quantification of Nanoparticle Ligand
Surface Density Using Fluorogenic Cu-Free Click Chemistry
Rasmus I. Jølck, Honghao Sun, Rolf H. Berg, and Thomas L. Andresen*[a]
Surface functionalized nanoparticles have found wide-
spread applications within the fields of medical diagnostics,[1]
drug delivery,[1,2] sensor development,[3] and vaccines.[4] At-
tachment of highly specific biomolecules such as peptides,
proteins, antibodies, or apatamers enables fine-tuning of the
nanoparticle constructs resulting in highly functionalized
nanoscale materials with high specificity towards, for exam-
ple, analytes, enzymes, or over-expressed or selectively ex-
pressed receptors on diseased cells. Controlling the biomole-
cule surface density is often a crucial parameter for obtain-
ing the desired properties of the nanomaterial, in particular
within the field of drug delivery.[5] To ensure the right com-
position of the nanoparticle construct, the number of ligands
immobilized on the surface must be analyzed. This has pre-
viously been done by measuring the fluorescence intensity
of probes covalently attached to the targeting ligands.[6]
Other methods include phosphorus and amino acid analy-
sis,[7] SDS-PAGE,[8] or protein determination assays.[9] Gen-
erally, these techniques are invasive, laborious and time con-
suming, and are often limited to a semiquantitative determi-
nation of the surface density. To rapidly expand the field of
highly functionalized nanoparticles there is a crucial need
for new methods to efficiently functionalize and analyze
nanoparticles. Surface coupling reactions should be fast, effi-
cient, reproducible, mild, and ideally include a reporter
functionality that allows direct quantification of coupling ef-
ficiency without prior purification. To fulfill these require-
ments, we have directed our attention towards the pro-fluo-
rophore 3-azidocoumarin.[10] Coumarin derivatives are often
used as biological probes, since they are both biocompatible
and easy to manipulate synthetically. Introduction of an
azido functionality at the 3-position results in efficient
quenching of the coumarin fluorescence due to electron
donation from the electron-rich a-nitrogen of the azido
group into the coumarin backbone.[10] Triazole formation,
which can be achieved by the CuI-catalyzed azide/alkyne
Huisgen 1,3-dipolar cycloaddition reaction[11] (“Click” reac-
tion) or by addition of strain-promoted[12] or electron-defi-
cient alkynes,[13] eliminates the quenching through the for-
mation of a triazole ring, thus resulting in a strong fluores-
cence signal. We envisioned that 3-azidocoumarins could be
attached to the desired targeting ligand and function as both
a highly specific (orthogonal) conjugation linker molecule
and as a quantitative reporter of coupling efficiency in situ.
To test this approach we have investigated PEGylated lipo-
somes that have exposed terminal alkynes or cyclooctynes
at the distal end of the PEG and shown that post-functional-
ization with 3-azidocoumarin-modified RGD peptides[14] is
highly efficient. Coupling was achieved in high yield by the
CuI-catalyzed approach; however, the direct quantitative de-
termination of the conjugation efficiency was impaired due
to quenching of the formed fluorophore by CuII. By adopt-
ing copper-free Click conditions, excellent conjugation effi-
ciency was achieved, which was monitored by the direct and
quantitative in-situ read-out that the new method provides.
The new approach utilizes the 3-azido-7-(carboxy me-
thoxy)-chromen-2-one (3), which was synthesized in six
steps from 2,4-dihydroxybenzaldehyde (1) and N-acetylgly-
cine (2) with slight modifications to previously described
methods (Scheme 1).[10,15] Compound 3 was attached to the
N-terminal of the RGD-peptide after insertion of two gly-
cine spacers, resulting in the peptide 4. Compound 3 was
found to be fully compatible with standard conditions used
in Fmoc solid-phase peptide synthesis (SPPS). The two
alkyne-modified lipids 6 and 8 were synthesized in a single
step from commercially available DSPE-PEG2000-NH2 by
acylation with 4-pentynoic acid (5) and 7, respectively. Com-
pound 7 was synthesized as previously described else-
where.[16] The model RGD-peptide 4 used in the surface
conjugation studies was synthesized by standard Fmoc SPPS
methodology using a TentaGel resin with the RinkAmide
linker (for a detailed description of the synthesis protocol
see the Supporting Information). Visualization of the com-
pounds containing an aromatic azide functionality by
matrix-assisted laser desorption/ionization-time of flight
mass spectrometry (MALDI-TOF MS) was not possible due
to laser-induced photodissociation of the aromatic azides.[17]
However, electrospray ionization mass spectrometry (ESI-
MS) gave the expected masses. Small unilamellar vesicles
(SUVs) with the following compositions; liposome A:
DSPC/DSPE-PEG2000/6 (95:4:1) and liposome B: DSPC/
DSPE-PEG2000/8 (95:4:1) were prepared by the method de-
scribed by Bangham et al.[18] The liposomes were character-
ized by dynamic light scattering (DLS) (liposome A: (129ꢀ
[a] R. I. Jølck, Dr. H. Sun, Prof. R. H. Berg, Dr. T. L. Andresen
Department of Micro- and Nanotechnology
Technical University of Denmark (DTU)
Frederiksborgvej 399, 4000 Roskilde (Denmark)
Fax : (+45)46-77-47-91
E-mail: thomas.andresen@nanotech.dtu.dk
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201003131.
3326
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 3326 – 3331

COMMUNICATION
Scheme 1. a) Synthesis of 3-azido-7-(carboxymethoxy)-chromen-2-one (3); b) synthesis of 3-azidocoumarin-functionalized RGD-peptide 4; c) synthesis of
the alkyne-functionalized PEGylated phospholipids 6 and 8. Fmoc=9-Fluorenylmethoxycabonyl; HATU=O-(7-azabenzotriazol-1-yl)-1,1,3,3-tetramethyl-
uronium hexafluorophosphate: TFA=trifluoroacetic acid; TIS=triisopropylsilane; EDC= 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide; DMAP=
4-Dimethylaminopyridine; DSPE-PEG2000-NH2 =1,2-distearoyl-sn-glycero-3-phosphoethanolamine-N-[amino(polyethylene glycol)-2000].
1.7) nm, liposome B: (118.9ꢀ0.7) nm) and the zeta potential
was measured (liposome A: (ꢁ14.68ꢀ0.50) mV, liposome
B: (ꢁ12.22ꢀ0.81) mV).
efficiency could be directly quantified without prior purifica-
tion. The fluorescence of 9 showed a linear correlation in
the concentration range from 0.1–4.0 mm (R2 =0.99). The ob-
tained conversion plot illustrated in Figure 2c was normal-
ized according to a 60:40 distribution ratio of the functional-
ized lipid 6 due to the curvature-induced asymmetry of lipid
bilayers in 100 nm liposomes.[19] As clearly evident from the
conversion plot illustrated in Figure 2c, triazole formation
occurs rapidly within the first hour, whereas the negative
control (no copper added) remained non-fluorescent over
the entire time-span of the experiment due to lack of tria-
zole formation. However, a quenching effect was observed
shortly after reaching full conversion. This phenomenon has
not previously been described by other authors using cou-
marin as a analytical tool in protein conjugation chemis-
try,[15] even though the nature of the problem seems general
as described below. Initially, our experiments were carried
out by measuring the fluorescence in HEPES buffer giving
the plot shown in Figure 2c (triangles). The data indicated
Initially, the CuI-catalyzed approach to functionalize lipo-
somes was investigated. Site-specific conjugation of the 3-
azidocoumarin-functionalized RGD-peptide 4 to the termi-
nal alkyne-functionalized liposome A was carried out in the
presence of 4 (0.125 mm), CuSO4·5H2O (1.25 mm), and
sodium ascorbate (25 mm) (Figure 1). The same conditions
without addition of CuSO4 were used as a negative control.
The reaction mixtures were protected from sunlight, gently
stirred, and the degree of conversion was monitored by
measuring the increase in fluorescence intensity (excitation
342 nm, emission 421 nm, Figure 2a) as a function of time.
At each time point, 50 mL of the reaction solution was
added to a cuvette containing HEPES buffer (1200 mL) and
measured in a spectrofluorometer. By correlating the ob-
served fluorescence intensity to a standard curve based on
the core probe 9 in HEPES buffer (Figure 2b), the coupling
Chem. Eur. J. 2011, 17, 3326 – 3331
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3327

T. L. Andresen et al.
Figure 2. a) Excitation and emission scan of 9. lEx. max =342 nm, lEm. max
=
421 nm. b) Standard curves based on 9 in either CHCl3/MeOH/H2O
(16:8:1) (*) or HEPES buffer (~). c) Conversion plot of the CuI-cata-
lyzed azide/alkyne Huisgen 1,3-dipolar cycloaddition reaction between
the alkyne-modified liposome (liposome A; DSPC/DSPE-PEG2000/6
(95:4:1)) and 4. Fluorescence intensity measured either in CHCl3/MeOH/
H2O (16:8:1) (*) or HEPES buffer (~) and converted to degree of con-
version by using the standard curves based on 9. Negative control per-
^
formed without addition of copper ( ).
Figure 1. Liposomes consisting of DSPC/DSPE-PEG2000/6 were formulat-
ed and incubated with 4 in the presence of CuSO4 and sodium ascorbate,
and the increase of fluorescence intensity (Ex. 342 nm, Em. 421 nm) was
monitored as a function of time.
caused the observed decrease. In addition to the tempera-
ture dependency, CuII was found to efficiently quench the
fluorescence intensity of 9 (Figure 3c) and since most of the
copper in the solution is CuI that was found not to quench
the fluorescence, there is a time delay before this effect be-
comes dominant. This phenomenon, which recently has
been described for 7-aminocoumarins[20] as well as other flu-
orophores,[21] was found to cause the observed drop in inten-
sity. The plateau reached after 3 h represents the scenario
where the formed probes at the liposomes interface are fully
complexed with CuII. To regain the maximum fluorescence
intensity, the CuII-chelator ethylenediaminetetraacetic acid
(EDTA) (log K=18.76)[22] was added to the solution, but it
did not result in complete recovery of the fluorescence in-
tensity (Figure 3c). These findings encouraged us to develop
a method in which the reaction could be carried out without
the copper catalyst.
Liposome B, which contains exposed cyclooctynes at the
outer PEG-layer, was incubated with 4 but without addition
of either CuSO4 or sodium ascorbate (Figure 4). The lipo-
somes and the peptide were simply mixed and the fluores-
cence intensity was monitored as a function of time using
the CHCl3/MeOH solution as described above. The ob-
served intensity (excitation 337 nm, emission 402 nm, Fig-
to us that self-quenching was occurring on the liposome sur-
face due to a high local concentration of the coumarin
probe. To test this, we disrupted the liposome by solubiliza-
tion of the lipids in organic solvent. This was achieved by
adding 50 mL of the reaction mixture to a solution of CHCl3/
MeOH (2:1) (1200 mL), thereby forming a homogenous so-
lution with maximum intermolecular distance between the
lipids. Disruption of the liposomes was confirmed by DLS,
which showed no presence of colloids in the solution. As
clearly illustrated in Figure 2c, a parallel shift along the y
axis was observed when measuring the fluorescence in or-
ganic solvent compared to aqueous buffer, and we observed
that the reaction proceeded to completion within one hour
(Figure 2c, circles). However, as the time-dependent de-
crease in fluorescence signal remained, self-quenching was
not the only effect. We speculated that there were a number
of conditions that could influence the experiments. The fluo-
rescence signal of 9 was found to be independent of pH
(Figure 3a), whereas an increase in temperature caused a
decrease in signal intensity (Figure 3b). However, since all
measurements were performed at 308C this could not have
3328
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 3326 – 3331

Nanoparticle Functionalization Using Fluorogenic Cu-Free Click Chemistry
COMMUNICATION
Figure 3. a) The influence of pH on the fluorescence intensity of 9. b)
The influence of temperature on the fluorescence intensity of 9. c)
Quenching of the fluorescence intensity of 9 by addition of CuSO4 fol-
lowed by recovery of the fluorescence intensity by addition of EDTA.
Figure 5. a) Excitation (full line) and emission scan (dashed) of 10.
lEx. max =337 nm, lEm. max =402 nm. b) Standard curve based on 10 solubi-
lized in CHCl3/MeOH/H2O (16:8:1). c) Conversion plot of the Cu-free
strain-promoted azide/alkyne Huisgen 1,3-dipolar cycloaddition reaction
between the alkyne-modified liposome (liposome B; DSPC/DSPE-
PEG2000/8 (95:4:1)) and 4 (*). Negative control performed without addi-
tion of liposome B (&).
ure 5a) was correlated to a standard curve based on the two
core-probe 10 (Figure 5b), which was isolated as a 1:1 mix-
ture of possible regioisomers. The two regioisomers were
found to have the same fluorescence spectral properties. As
illustrated in Figure 5a, a blue shift was observed for 10
compared to the previously used probe 9. The fluorescence
of 10 was found to have a linear correlation in the concen-
tration range from 0.1–4.0 mm (R2 =0.99). By adopting the
Cu-free approach and disruption of the liposomes prior to
the fluorescence measurements, a smooth conversion plot
was achieved. The strain-promoted Click reaction proceeded
slower than the CuI-catalyzed counterpart; however, quanti-
tative conversion was achieved after approximately 8.5 h.
The employed reaction conditions were found not to alter
the size or the surface charge of the liposomes. Furthermore,
the negative control without addition of liposome B re-
mained non-fluorescent during the entire time span of the
experiment.
The average number of peptides exposed at the outer lip-
osome membrane (surface density) can be calculated by
using Equations (1)–(3). By assuming that all liposomes are
unilamellar and using an averaged lipid surface area, the
number of lipid molecules in a single liposome can be calcu-
lated as described in Equation (1),
Figure 4. Liposomes consisting of DSPC/DSPE-PEG2000/8 were formulat-
ed and incubated with 4 and the increase of fluorescence intensity (Ex.
337 nm, Em. 402 nm) was monitored as a function of time.
Chem. Eur. J. 2011, 17, 3326 – 3331
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3329

T. L. Andresen et al.
d
d
2
2
systems that cannot be dissolved by addition of organic sol-
vents as described herein, if the fluorophore surface density
is kept below the self-quenching concentration.
½4pð2Þ þ 4pð2 ꢁ hÞ ꢂ
ð1Þ
Ntot
¼
a
where Ntot is the number of lipids in a single liposome, d is
the diameter of the liposome (118.9 nm), h is the thickness
of the bilayer (4 nm), and a is the lipid head group area (PC
lipids 0.70 nm2).[23] The number of liposome particles per
milliliter of liposome solution can be calculated as described
in Equation (2),
Experimental Section
The synthesis of the compounds 3, 4, 6–10, their characterization
(1H NMR, 13C NMR, IR, MALDI-TOF MS, ESI-MS, HPLC), liposome
preparation, and protocols for surface functionalization are available in
the Supporting Information.
Mlipid ꢃ NA
ð2Þ
Nlipo
¼
Ntot ꢃ 1000
where Nlipo is the number of liposomes per milliliter and NA
is the Avogadro number. The absolute number of formed
fluorophores can be calculated by correlating the observed
fluorescence intensity to the standard curve based on 10
(Figure 5b). By dividing the number of formed fluorophores
with the number of liposome particles the average surface
density can be calculated as described in Equation (3),
Acknowledgements
The authors acknowledge financial support provided by the Danish Stra-
tegic Research Council (NABIIT) No. 2106-070033 and the Technical
University of Denmark (DTU).
Keywords: click chemistry
·
coumarin
·
liposomes
·
nanoparticles · surface functionalization
cf ꢃ V ꢃ NA
surface density ¼
ð3Þ
Nlipo
[1] a) M. A. Phillips, M. L. Gran, N. A. Peppas, Nano Today 2010, 5,
143–159; b) S. Azarmi, W. H. Roa, R. Lçbenberg, Adv. Drug Deliv-
ery Rev. 2008, 60, 863–875.
[2] a) M. E. Davis, Z. Chen, D. M. Shin, Nat. Rev. Drug Discovery 2008,
7, 771–782; b) V. Torchilin, Expert Opin. Drug Delivery 2008, 5,
1003–1025; c) R. A. Petros, J. M. DeSimone, Nat. Rev. Drug Discov-
ery 2010, 9, 615–627.
[3] a) V. Pavlov, Y. Xiao, B. Shlyahovsky, I. Willner, J. Am. Chem. Soc.
2004, 126, 11768–11769; b) J. Liu, Y. Lu, Nat. Protoc. 2006, 1, 246–
252.
[4] A. Muhs, D. T. Hickman, M. Pihlgren, N. Chuard, V. Giriens, C.
Meerschman, I. van der Auwera, F. van Leuven, M. Sugawara, M.
Weingertner, B. Bechinger, R. Greferath, N. Kolonko, L. Nagel-
Steger, D. Riesner, R. O. Brady, A. Pfeifer, C. Nicolau, Proc. Natl.
Acad. Sci. USA 2007, 104, 9810–9815.
where cf is the concentration of the formed fluorophore and
V is the reaction volume. By using Equations (1)–(3) the
strain-promoted Click reaction (liposome B) resulted in an
average surface density of 712 peptides per liposome.
From an overall perspective, this method is superior to
other post-functionalization methods, as it proceeds without
addition of a catalyst, under mild conditions, and includes a
reporter functionality by which the conversion efficiency
can be monitored directly without prior purification. Fur-
thermore, avoiding addition of Cu is highly desirable for lip-
osomal drug-delivery purposes since Cu is considered toxic
and has been reported to promote oxidation of unsaturated
phospholipids.[24] In addition, a shorter reaction time for the
strain-promoted Click reaction may be achieved by employ-
ing more reactive cycloalkynes.[25]
[5] T. L. Andresen, S. S. Jensen, K. Jørgensen, Prog. Lipid Res. 2005, 44,
68–97.
[6] J. Lu, M. Shi, M. S. Shoichet, Bioconjugate Chem. 2009, 20, 87–94.
[7] S. Zalipsky, B. Puntambekar, P. Boulikas, C. M. Engbers, M. C.
Woodle, Bioconjugate Chem. 1995, 6, 705–708.
In conclusion, we have developed a unique method to
post-functionalize liposomes and directly quantify the ach-
ieved ligand density on the liposome surface. Our results
have shown that 3-azidocoumarins are a valuable class of
linker molecules that can be utilized to give a quantitative
in-situ read out of conjugation efficiency. The CuI-catalyzed
azide/alkyne Huisgen 1,3-dipolar cycloaddition reaction re-
sulted in efficient conjugation; however, quantitative meas-
urements of the coupling efficiency were not possible due to
quenching by CuII. This limitation was solved by adopting
Cu-free conditions by using electron-deficient strain-pro-
moted alkynes. This approach resulted in a quantitative con-
version, which could be monitored by fluorescence spectros-
copy without prior workup. This method is the first to com-
bine both functionalization and quantification of the conju-
gation efficiency on the surface of liposomes in one single
step, and it is equally suited for other nanoparticle con-
structs (e.g. polymer micelles and polymersomes), including
[8] a) S. W. A. Reulen, W. W. T. Brusselaars, S. Langereis, W. J. M.
Mulder, M. Breurken, M. Merkx, Bioconjugate Chem. 2007, 18,
590–596; b) B. Gamier, A. Bouter, C. Gounou, K. G. Petry, A. R.
Brisson, Bioconjugate Chem. 2009, 20, 2114–2122.
[9] G. Bendas, D. Rothe, G. L. Scherphof, J. A. A. M. Kamps, Biochim.
Biophys. Acta Biomembr. 2003, 1609, 63–70.
[10] K. Sivakumar, S. Xie, B. M. Cash, S. Long, H. N. Bamhill, Q. Wang,
Org. Lett. 2004, 6, 4603–4606.
[11] a) C. W. Tornøe, C. Christensen, M. Meldal, J. Org. Chem. 2002, 67,
3057–3064; b) V. V. Rostovtsev, L. G. Green, V. V. Fokin, K. B.
Sharpless, Angew. Chem. 2002, 114, 2708–2711; Angew. Chem. Int.
Ed. 2002, 41, 2596–2599; c) V. V. Rostovtsev, L. G. Green, V. V.
Fokin, K. B. Sharpless, Angew. Chem. 2002, 114, 2708–2711; Angew.
Chem. Int. Ed. 2002, 41, 2596–2599.
[12] N. J. Agard, J. A. Prescher, C. R. Bertozzi, J. Am. Chem. Soc. 2004,
126, 15046–15047.
[13] Z. Li, T. S. Seo, J. Ju, Tetrahedron Lett. 2004, 45, 3143–3146.
[14] a) G. Huang, Z. Zhou, R. Srinivasan, M. S. Penn, K. Kottke-March-
ant, R. E. Marchant, A. S. Gupta, Biomaterials 2008, 29, 1676–1685;
b) L. Auzzas, F. Zanardi, L. Battistini, P. Burreddu, P. Carta, G.
Rassu, C. Curti, G. Casiraghi, Curr. Med. Chem. 2010, 17, 1255–
3330
www.chemeurj.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 3326 – 3331

Nanoparticle Functionalization Using Fluorogenic Cu-Free Click Chemistry
COMMUNICATION
1299; c) R. I. Jølck, L. N. Feldborg, S. Andersen S. M. Moghimi,
T. L. Andresen, Adv. Biochem. Eng./Biotechnol. 2010, DOI:
10.1007/10 2010 92.
[21] E. Espada-Bellido, D. M. Galinda-Riano, M. Garcia-Vargas, R. Nar-
ayanaswamy, Appl. Spectrosc. 2010, 64, 727–932.
[22] A. M. Crouch, L. E. Khotseng, M. Polhuis, D. R. Williams, Anal.
Chim. Acta 2001, 448, 231–237.
[23] S. M. Gruner, M. W. Tate, G. L. Kirk, P. T. C. So, D. C. Turner, D. T.
Keane, C. P. S. Tilcock, P. R. Cullis, Biochemistry 1988, 27, 2853–
2866.
[15] A. J. Dirks, J. J. L. M. Cornelissen, R. J. M. Nolte, Bioconjugate
Chem. 2009, 20, 1129–1138.
[16] M. K. Schultz, S. G. Parameswarappa, F. G. Pigge, Org. Lett. 2010,
12, 2398–2401.
[17] M. F. Budyka, Russ. Chem. Rev. 2008, 77, 709–723.
[18] A. D. Bangham, R. D. Horne, J. Mol. Biol. 1964, 8, 660–668.
[19] M. Rovira-Bru, D. H. Thompson, I. Szleifer, Biophys. J. 2002, 83,
2419–2439.
[20] H. S. Jung, P. S. Kwon, J. W. Lee, J. I. Kim, C. S. Hong, J. W. Kim, S.
Yan, J. Y. Lee, J. H. Lee, T. Joo, J. S. Kim, J. Am. Chem. Soc. 2009,
131, 2008–2012.
[24] a) C. Lee, J. Barnett, P. D. Reaven, J. Lipid Res. 1998, 39, 1239–
1247; b) S. Gal, I. Pinchuk, D. Lichtenberg, Chem. Phys. Lipids
2003, 126, 95–110.
[25] J. A. Codelli, J. M. Baskin, N. J. Agard, C. R. Bertozzi, J. Am. Chem.
Soc. 2008, 130, 11486–11493.
Received: October 31, 2010
Published online: February 15, 2011
Chem. Eur. J. 2011, 17, 3326 – 3331
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3331