High-purity discrete PEG-oligomer crystals allow structural insight

Article abstract of DOI:10.1002/anie.200804623

To great (monodisperse) lengths: An improved synthesis of purer ethylene glycol (EG) oligomers allows access to 16- and 32-mers pure enough for multiple incorporation, and also to the longest (48-mer) discrete EG oligomer yet reported. The high purity enables the first crystallizations and hence the first glimpses of secondary 3₁₀-helical PEG structures.

Full text of DOI:10.1002/anie.200804623

Communications
DOI: 10.1002/anie.200804623
PEG Synthesis
High-Purity Discrete PEG-Oligomer Crystals Allow Structural
Insight**
Alister C. French, Amber L. Thompson, and Benjamin G. Davis*
Poly(ethylene glycols) (PEGs) are a class of molecules that
show wide-ranging application in biological contexts.[1] For
example, the attachment of chains of PEG to therapeutic
proteins, peptides, and even small molecule therapeutics
(PEG-ylation)[2] is a well-established method used to influ-
ence stability, immunogenicity,[3] pharmacokinetics,[1,2,4,5] and
mode of action.[4,5] This water-soluble, nontoxic polymer has
been used to improve solubility,[6] dramatically extend serum
half life,[7] and enhance the oral bioavailability[2] of substrate
molecules. PEGs have also played key roles in lipid self-
assembly, including liposome development[8] and even the
Figure 1. Statistical modeling[17] of multi-chain PEG brush: Simulated
mass spectra showing the statistical distribution of 32 copies of PEG
oligomers, PEGn, at specific purities attached to a single substrate.
a) 96% oligomer purity[20] (taken from analysis of Quanta BioDesign
amino-dPeg24 tert-butyl ester). b) 98.8% oligomer purity (target 16-mer
purity chosen as benchmark). c) 99.8% oligomer purity (taken from
actual purity of amino-PEG16 methyl ether 19 synthesized herein.
delineation of the structural mechanisms of lipid-bilayer
formation and fusion.[9]
In the development of PEG-ylated biopharmaceuticals a
first potential source of heterogeneity is low site-selectivity
and low conversion during attachment to, for example, the
protein or substrate.[6,10,11] The second potential source of
heterogeneity is that of the attached ligand itself; in polymeric
structures, such as PEG, this arises from a distribution of
different-length oligomers (polydispersity). Well-controlled
anionic polymerization can produce monofunctional PEG
with polydispersity indices (PDIs) approaching 1.04 (see
Supporting Information).[12,13] Polydisperse oligomers give
rise to mixtures of PEG-ylated conjugates and regulatory
bodies are inclined to demand further investigations before
approving products of inexactly known composition.[1,14]
To surmount these difficulties and with the goal of greater
functional precision we aim to create purer PEG-ylated
proteins. Notably, enhancement of in vivo serum half-life is
Our syntheses of 16-, 32-, and 48-mers of ethylene glycol,
focused on minimizing contamination by other chain lengths.
Despite advances in analytical, protecting, and coupling
strategies,[21,22] Hibbert and co-workersꢀ approaches to the
first discrete EG oligomer[23] in the 1930s created PEG42,[24]
which still remains one[25] of the longest reported (semi)-
discrete oligomers. A common feature is strategic desymmet-
rization (e.g. protection) that prevents polymerization during
coupling. There are then four conceptual ways to elongate the
EG: iterative coupling of a mono-protected building block to
A) one end, to B) both ends, C) chain doubling, or D) chain
tripling (Figure 2). Mode D chain tripling allows rapid
expansion but is limited[23,24,26] by difficulties in desymmetriz-
ing PEGs of > 12 units.[27] Mode A/B strategies have pro-
duced symmetrical chains of up to 44[25] and 29[28] EG units by
bidirectional extension of PEG diols with monobenzylated
PEG monotosylates.[25,28] However, although such mode A/B
iterative strategies provide longer oligomers for the first two
filial coupling generations, the length of products from a
mode C[29,27] chain doubling strategy follows the relationship
L = 2gn. Thus, once g > 2, the exponential relationship of a
mode C doubling strategy greatly outstrips the linear growth
of mode A/B iterative addition, L = n(1+2g).[30]
[16]
most significant above 20 KDa.[15] Whilst discrete PEG512
(MW 22.6 kDa) is beyond the reach of current synthesis, a
strategy can be envisaged through which the same functional
effects can be achieved additively by attaching multiple,
shorter chains. Statistical modeling[17] of such combinations
revealed that current PEG sources[18–20] would be insuffi-
ciently pure (Figure 1).
[*] A. C. French, Dr. A. L. Thompson, Prof. B. G. Davis
Department of Chemistry, University of Oxford, Chemistry Research
Laboratory
12 Mansfield Road, Oxford (UK)
Fax: (+44)1865-275-674
E-mail: ben.davis@chem.ox.ac.uk
Homepage: users.ox.ac.uk/ndplb0149
We therefore adopted mode C multi-generation, chain
doubling; tetraethylene glycol (EG4, 1) was chosen as a cheap,
readily available starting material (100 ꢁ cheaper and avail-
able in higher purity than EG6[27] or EG5.[31]) Highly pure PEG
was accessed through an ultimately efficient and scaleable
monoprotection and electrophilic activation (sulfonation)
strategy on a 35–75 gram scale. Two key features allowed
the preparation of highly pure oligomers of PEGn. Firstly,
[**] We thank Drs. John Porter and Mike Eaton from UCB-Celltech for
invaluable collaborative support; Dr. David J. Watkin for crystallo-
graphic insight and CEM for their generosity and technical support.
UCB-Celltech and the BBSRC funded an Industrial CASE Student-
ship for A.C.F. PEG=Poly(ethylene glycol).
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.200804623.
1248
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 1248 –1252

Angewandte
Chemie
ether protection,[37] in place of tert-butyl fur-
nished substrates, such as 4, with superior chro-
matographic separations. Secondly, with an
empirical connection between presence of alk-
oxide and depolymerization established, subse-
quent coupling methods were designed to mini-
mize reacting concentrations of alkoxide through
slow addition of a solution of potassium tert-
butoxide to the coupling partners, such as alco-
hols 4 or 7 and tosylates 5 or 12 in DMF. Thus,
after coupling to form the octaethylene glycol
benzyl trityl ether 7, contaminating heptaethy-
lene glycol benzyl trityl ether 9 was readily
separated by normal-phase flash chromatogra-
phy. Chain doubling, and subsequent reverse-
phase flash chromatography to remove 15-mer
gave hexadecaethylene glycol benzyl trityl ether
14 at 99.0% oligomer purity (PDI = 1.00009).
Moreover, the strategy of minimizing alkoxide
concentration through the slow addition of
base[38] allowed the preparation of 14 in an even
better 99.8% oligomer purity (PDI = 1.0000023)
without the need for reverse-phase chromatog-
raphy.
Figure 2. (Semi)discrete PEG oligomer synthesis strategies: Each circle represents an
EG unit. L is the oligomer length after g generations of coupling given a starting
material of length n. Mode A: unidirectional iterative coupling; Mode B: bidirectional
iterative coupling; Mode C: chain doubling; Mode D: chain tripling.
avoidance of PEG-chain degrading depolymerization and,
secondly, improved chromatographic contrast. This advance
was accomplished through the survey of three contrasting
hydroxy protecting groups: benzyl, tert-butyl and trityl.
EG4 mono-tert-butyl ether 2[32] and mono-benzyl ether 3[33]
were prepared from 1 using acid resin/isobutene and alkali/
BnCl, respectively; importantly 2 and 3 purified well by
normal-phase flash chromatography to yield mono-protected
hydroxy nucleophilic coupling moieties for chain doubling.
The electrophilic coupling components were accessed
through activation as sulfonate esters; despite use in earlier
examples of PEG synthesis,[24,34] attempted halogenation
methods confirmed prior observations of chain degrada-
tion.[27] Thus, reaction of 3 with alkali/TsCl[25] afforded
tosylate 5 on a 15 g scale (Scheme 1).
Coupling of 5 with 2 highlighted critical processes that
threatened purity in PEG synthesis. Ether formation was
initially performed by first treating alcohol 2 with sodium
hydride in dry THF, then adding the tosylate 5 dropwise.
After workup this gave the desired PEG8 benzyl tert-butyl
ether 6. However the product was contaminated by 3% PEG7
benzyl tert-butyl ether 8, which resulted from partial depoly-
merization of the alcohol 2;[35,36] the level of depolymerized
contaminant increased with the prolonged existence of the
intermediate alkoxide.
Next, we incorporated functionality useful for protein
modification/conjugation. Thus, after acidolytic removal of
Trt and tosylation, PEG16 benzyl tosylate 16 was treated with
sodium methoxide in DMF to give the benzyl methyl ether
from which hydrogenolytic deprotection readily furnished
PEG16 monomethyl ether 17. Further tosylation and reaction
with sodium azide then allowed access to PEG16 azido methyl
ether 18 (suitable for attachment to proteins by, e.g., site-
selective[39] copper-catalyzed Huisgen–Dimroth cycloaddi-
tion[40,41]), and by hydrogenation the PEG16 amino compound
19.
These PEG16 derivatives are the highest purity PEGs of
comparable length characterized to date, displaying PDI
values between 1.000015 and 1.000002 and percentage
purities between 99.6 and 99.9% single oligomer.[42] Remark-
ably, these exquisite purities allowed the first crystallizations
of PEGs, which in turn allowed the formation of diffracting
single crystals. Single-crystal X-ray diffraction experiments on
these gave the first indication of the 3D structure and also a
unique insight into an extended helical secondary structure of
PEGs (Figure 3; see footnote and Supporting Information for
experimental details). As a result of extended end-to-end
packing of the PEG16 17 and a period of the PEG molecule
that is not commensurate with the period of the helices or the
lattice a continuous solution emerges. Although, given the
quality of the data it is not possible to determine the structure
fully, these results clearly indicate the presence of packed
antiparallel helical strands with opposing handedness within
which the oxygen atoms reside in the core of the helix with
surrounding hydrophobic methylene groups. This behavior is
strikingly similar to that proposed[43] in the formation of
inverse micelles where again hydrophilic groups are seques-
tered in the micelle core. These helices pack together forming
interleaved opposing-strands, such that each PEG is sur-
rounded by four of the opposite handedness (Figure 3). The
Despite considerable effort, lack of chromatographic
resolution of these tBu-protected PEGs 6 and 8 under a
number conditions meant that complete separation or
removal of truncated impurities, such as 8, proved difficult
and carriage of truncated PEGnÀ1 impurities into subsequent
chain-doubling generations to give 16-mer 13 and the
corresponding 32-mer led only to compounded issues of
contamination and a rapid collapse in purity.
Two adaptations allowed us to greatly improve the
outcome of our synthetic strategy. Firstly, use of monotrityl
Angew. Chem. Int. Ed. 2009, 48, 1248 –1252
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1249

Communications
natural polypeptides.[49] We consider this first
clear structural insight into the extended shape
of PEG to be an important step in understand-
ing the structural biology of the molecule class
that is most commonly used for therapeutic
protein modification and that is playing an
important role in the understanding of bilayer
systems. Moreover, the snapshot of lateral
helix-to-helix packing observed in this case
may also offer detailed insight into crystallinity
effects seen in PEG-terminated self-assembled
monolayers.[50]
Finally, higher generation discrete PEGs,
PEG32 (20) and PEG48 (21), were synthesized.
When selective hydrogenolytic deprotection of
the benzyl trityl ether 14 was attempted to allow
further chain doubling, fully deprotected hex-
adecaethylene glycol diol 15 was instead
obtained as the major product. To explore
couplings to give access to longer PEG oligom-
ers, 15 was used in a trial desymmetrizing chain
doubling with PEG16 benzyl tosylate 16
(Scheme 1). Excitingly, this allowed not only
formation of 3rd generation but also 4th gen-
eration PEGs, giving after chromatography, not
only PEG32 derivative 20 but also PEG48 21 in
98.9% and 98.0% purity, respectively.
In summary, we have produced alcohol,
azide, and amine derivatives of hexadecaethy-
lene glycol monomethyl ether in gram quanti-
ties at purities of > 99.8% single oligomer. This
is sufficiently pure for a 32-chain protein
conjugate to be > 90% single mass; these
syntheses have also yielded PEG32 and PEG48.
Moreover, these highly pure PEGs have given
rise to the first X-ray crystal structures that
have allowed a unique insight into PEG mor-
phology. Work is currently ongoing to allow a
more complete understanding of the complex
factors that promote depolymerization and
elimination in PEG synthesis in addition to
Scheme 1. PEG Syntheses: Chain doubling mode C synthesis of highly pure PEG
oligomers. Trt=trityl, pTs =para-toluenesulfonyl, TIS=triisopropylsilane, DCM=di-
chloromethane, TFA=trifluoroacetic acid, [P]=protecting group.
distance between the centre of two similar handed coils is
6.4 ꢂ), while opposing helices are closer (4.5 ꢂ).
the development of methods that will incorporate these
highly pure PEGs into biomolecules.
The need for exquisite purity in sample preparation was
made clear by the melting points of these samples (298C for
17, 318C for 18, 238C for BnO-PEG16-OMe) which are
strikingly close to room temperature—the slightest depres-
sion in melting point would have rendered crystals unobtain-
able under many standard growth conditions. Although
limited examples[45–47] of shorter (< 9) EG derivatives have
been crystallized these have been achieved only through
heavy-metal derivatization,[45,46] templating,[45–47] and/or
unusual end-group modifications;[46] the resulting structures
are not extended, give little insight into secondary structure,
and are reminiscent only of crown ethers.
Experimental Section
Characterization and methods for all compounds can be found in the
Supporting Information.
17 (outline of synthesis): Na metal (0.5 g) was washed in
petroleum ether and dissolved in dry MeOH (10 mL) overnight. 16
(7.0 g, 7.3 mmol) was taken up in dry DMF (90 mL) and treated with
the resulting NaOMe solution. After 4 h, the reaction mixture was
concentrated, worked up (NH4Cl (aq)), extracted (CHCl3), dried
(Na2SO4) evaporated, and exposed to a 2nd reaction cycle. After 3 h,
ESI-MS indicated > 99% conversion. Work-up gave 6.9 g of crude
product which was purified by gradient flash chromatography
[Biotage SP4 system 40 mLminÀ1; SNAP 100 g KP Silica cartridge;
Solvent A ꢀ DCM, solvent B ꢀ 30% methanol in DCM; column
equilibrated at 0% B; crude product diluted with 13 mL DCM and
injected; column run 1 CV at 0% B, 3 CV to 2% B, 3 CV to 6% B,
From these unique structures it is clear that PEG is
capable of adopting an intriguing 310-helix (using Bragg
nomenclature[48]) structure that is strikingly similar in shape
and pitch to the rare 310-alpha-helical motifs found in some
1250
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 1248 –1252

Angewandte
Chemie
[2] J. Whelan, C. Pan (Bayer Pharmaceuticals Corporation, USA),
PCT Application WO 2005-US20469, 2005.
[3] A. Abuchowski, T. van Es, N. C. Palczuk, F. F. Davis, J. Biol.
Chem. 1977, 252, 3578.
[4] S. Zalipsky, Adv. Drug Delivery Rev. 1995, 16, 157.
[5] R. B. Greenwald, Y. H. Choe, J. McGuire, C. D. Conover, Adv.
Drug Delivery Rev. 2003, 55, 217.
[6] S. Brocchini, A. Godwin, S. Balan, J.-w. Choi, M. Zloh, S.
Shaunak, Adv. Drug Delivery Rev. 2008, 60, 3.
[7] P. Caliceti, F. M. Veronese, Adv. Drug Delivery Rev. 2003, 55,
1261.
[8] A. Gabizon, Adv. Drug Delivery Rev. 1997, 24, 337.
[9] B. R. Lentz, J. Lee, Mol. Membr. Biol. 1999, 16, 279.
[10] G. G. Kochendoerfer, Curr. Opin. Chem. Biol. 2005, 9, 555.
[11] Refs. [6,10] give examples of approaches that move beyond the
statistical mixtures of products obtained by partial reaction of
exposed protein side chains during typical methods using for
example, PEG-acylating reagents, PEG-alkylating reagents, or
reductive amination.
[12] N. N. Reed, K. D. Janda, J. Org. Chem. 2000, 65, 5843.
[13] J. Zhang, Y.-J. Zhao, Z.-G. Su, G.-H. Ma, J. Appl. Polym. Sci.
2007, 105, 3782.
[14] M. J. Vicent, L. Dieudonne, R. J. Carbajo, A. Pineda-Lucena,
Expert Opin. Drug Delivery 2008, 5, 593.
[15] D. Filpula, H. Zhao, Adv. Drug Delivery Rev. 2008, 60, 29.
[16] We describe a discrete oligomer of n ether-linked ethylene units
as PEGn.
[17] See Supporting Information for modelling of the effects of even
small impurities in multivalent conjugations.
Figure 3. X-ray structure of PEG16 (17). a) the close-packed 310-helix
dominated by a gauche arrangement of individual EG units; red O,
gray C, green H; b) accessible-surface diagram showing three antipar-
allel PEGs in blue and green; c) Thermal ellipsoid plot of 17 showing
the oxygen core surrounded by the hydrophobic methylene groups;
d) Packing in 17, showing the two different helix hands in green and
blue.[44]
[18] Polypure advertise material of > 95% single oligomer, up to
PEG28 prepared by sample displacement chromatography.
However, the strategy is inefficient if a specific length is required
and scales poorly for longer PEGs, limited by the diminishing
relative differences between successive oligomers.
[19] Quanta BioDesign currently advertise discrete oligomers up to
PEG24; ESI-MS analysis of “Amino-dPEG24 tert-Butyl Ester”
(lot 314521) indicates that 96% of this material is the named
oligomer.
[20] Whilst PDI is useful for discussing mixed ensembles of
oligomers, it becomes cumbersome in the situation where one
principal species is contaminated with a few others in minor
quantities. Herein, we therefore discuss oligomer purity primar-
ily on a percentage basis.
[21] M. B. Andrus, T. M. Turner, E. P. Updegraff, Z. E. Sauna, S. V.
Ambudkar, Tetrahedron Lett. 2001, 42, 3819.
[22] A. Harada, J. Li, M. Kamachi, J. Am. Chem. Soc. 1994, 116, 3192.
[23] H. Perry, Can. J. Res. Sect. B 1936, 14, 77.
[24] R. Fordyce, E. L. Lovell, H. Hibbert, J. Am. Chem. Soc. 1939, 61,
1905.
3 CV to 14% B, and 3 CV to 30% B]. to give 5.11 g (6.18 mmol, 89%
yield) of 17, a colorless oil which crystallized upon being left to stand
at room temperature.
Single-crystal X-ray diffraction data were collected using an
Enraf–Nonius KappaCCD diffractometer (MoKa radiation; l =
0.71073 ꢂ) equipped with a Cryostream N2 open-flow cooling
device,[51] and the data were collected at 150(2) K. Data were
processed using the DENZO-SMN package[52] and the structure
was solved by direct methods.[53] A number of models were examined
using PLATON (ADSYMM)[54,55] and the final model was refined on
F2 using the CRYSTALS suite.[56] CCDC 707050 contains the
supplementary crystallographic data for this paper. These data can
be obtained free of charge from The Cambridge Crystallographic
Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
[25] S. A. Ahmed, M. Tanaka, J. Org. Chem. 2006, 71, 9884.
[26] Hibbertꢀs original synthesis was a chain tripling (where L = 3gn)
that reached PEG42 by the third filial generation (g = 3).
[27] F. A. Loiseau, K. K. M. Hii, A. M. Hill, J. Org. Chem. 2004, 69,
639.
Single Crystal X-ray Diffraction Data: “C20H80O20”, Mr = 881.06,
¯
0.20 ꢁ 0.35 ꢁ 0.40 mm, Tetragonal (I4c2), a = 9.0621(11), c =
28.680(5) ꢂ, V= 2355.2(6) ꢂ3, Z = 2, m = 0.098 mmÀ1
,
1calcd
=
1.242 MgmÀ3, T= 150(2) K, 6852 reflections collected, 583 independ-
ent [R(int) = 0.129], R1 = 0.1403, wR2 = 0.2205 [I > 2s(I)], 1min,max
À0.30, 0.42 eꢂ3.
=
[28] D. Niculescu-Duvaz, J. Getaz, C. J. Springer, Bioconjugate
Chem. 2008, 19, 973.
[29] Two-generation chain doubling synthesis has allowed the syn-
thesis of 24 unit EGs.
Received: September 20, 2008
[30] It has been suggested (Ref. [28]) that bidirectional iterative
growth is quicker and more efficient than chain doubling, yet our
statistical analysis reveals that this is only true beyond L = 16n.
It should also be noted that all the steps are not of equal strategic
importance, since it is only during ether coupling that the most
difficult to remove by-products are formed.
Published online: January 13, 2009
Keywords: helical structures · mass spectrometry ·
.
poly(ethylene glycol) · polymers · synthetic methods
[31] Sigma–Aldrich prices 2008.
[32] R. D. Lousenberg, M. S. Shoichet, J. Org. Chem. 1997, 62, 7844.
[33] T. Gibson, J. Org. Chem. 1980, 45, 1095.
[1] R. Haag, F. Kratz, Angew. Chem. 2006, 118, 1218; Angew. Chem.
Int. Ed. 2006, 45, 1198.
Angew. Chem. Int. Ed. 2009, 48, 1248 –1252
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
1251

Communications
[34] G. Coudert, M. Mpassi, G. Guillaumet, C. Selve, Synth.
[43] V. Luzzati, R. Vargas, A. Gulik, P. Mariani, J. M. Seddon, E.
Rivas, Biochemistry 1992, 31, 279.
Commun. 1986, 16, 19.
[35] N. Boden, R. J. Bushby, S. Clarkson, S. D. Evans, P. F. Knowles,
A. Marsh, Tetrahedron 1997, 53, 10939.
[36] Base-catalyzed EG depolymerization during Williamson ether
synthesis is seldom noted in the literature. See Boden et al. for a
brief but excellent discussion (Ref. [35]); > 15% of EG unit loss
was observed in some examples.
[44] Figure prepared using Pymol [W. L. DeLano. The PyMOL
Molecular Graphics System (2008). DeLano Scientific, Palo
Alto, California, USA. http://www.pymol.org] and Mercury
[C. F. Macrae, P. R. Edgington, P. McCabe, E. Pidcock, G. P.
Shields, R. Taylor, M. Towler, J. van de Streek, J. Appl.
Crystallogr. 2006, 39, 453 – 457].
[37] V. E. M. Kaats-Richters, J. W. Zwikker, E. M. D. Keegstra, L. W.
Jenneskens, Synth. Commun. 1994, 24, 2399.
[45] H. Ohmoto, Y. Kai, N. Yasuoka, N. Kasai, S. Yanagida, M.
Okahara, Bull. Chem. Soc. Jpn. 1979, 52, 1209.
[38] Interestingly, the result was much reduced levels of depolyme-
rization with higher levels of elimination of tosylate; whilst this is
wasteful of tosylate, the resulting vinyl ether is much easier to
remove than a truncated but chemically identical oligomer.
[39] S. I. van Kasteren, H. B. Kramer, H. H. Jensen, S. J. Campbell, J.
Kirkpatrick, N. J. Oldham, D. C. Anthony, B. G. Davis, Nature
2007, 446, 1105.
[40] C. W. Tornoe, C. Christensen, M. Meldal, J. Org. Chem. 2002, 67,
3057.
[41] V. V. Rostovtsev, L. G. Green, V. V. Fokin, K. B. Sharpless,
Angew. Chem. 2002, 114, 2708; Angew. Chem. Int. Ed. 2002, 41,
2596.
[46] G. Weber, W. Saenger, Acta Crystallogr. Sect. B 1979, 35, 3093.
[47] D. Britton, M. K. Chantooni, Jr., J. Chem. Crystallogr. 2001, 31,
5.
[48] L. Bragg, J. C. Kendrew, M. F. Perutz, Proc. R. Soc. London Ser.
A 1950, 203, 321.
[49] C. Tonlolo, E. Benedetti, Trends Biochem. Sci. 1991, 16, 350.
[50] P. Harder, M. Grunze, R. Dahint, G. M. Whitesides, P. E.
Laibinis, J. Phys. Chem. B 1998, 102, 426.
[51] J. Cosier, A. M. Glazer, J. Appl. Crystallogr. 1986, 19, 105.
[52] Z. Otwinowski, W. Minor, Methods Enzymol. 1997, 276.
[53] A. Altomare, G. Cascarano, C. Giacovazzo, A. Guagliardi, M. C.
Burla, G. Polidori, M. Camalli, J. Appl. Crystallogr. 1994, 27, 435.
[54] A. L. Speck, Utrecht University, 2001.
[42] We have analyzed the highest purity material that is commer-
cially available (Ref. [19]); it should be noted that a lack of mass
spectra or spectral expansions in impurity regions in previous
syntheses make purities difficult to assess.
[55] Y. Le Page, J. Appl. Crystallogr. 1988, 21, 983.
[56] P. W. Betteridge, J. R. Carruthers, R. I. Cooper, K. Prout, D. J.
Watkin, J. Appl. Crystallogr. 2003, 36, 1487.
1252
www.angewandte.org
ꢀ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2009, 48, 1248 –1252