Manufacturing and PEGylation of a dual-acting peptide for diabetes

Article abstract of DOI:10.1002/ejoc.200800722

In the given manuscript we summarize our activities to develop a scalable synthesis for a PEGylated peptide, which undergoes development for metabolic disorders. The polypeptide contains a sequence of 31 natural amino acids and carries a site-specific PEGylation at the C-terminus. Initial synthetic studies revealed that the peptide moiety of the target molecule is not an ideal target for a linear solid-phase process, as we were confronted with difficult sequences , which led to a low-yielding process. Subsequently, we developed a mixed-phase synthesis, which included production of fragments through solid-phase peptide synthesis, followed by an assembly of these fragments in solution. We were able to show that this new process delivered the desired peptide moiety in gram-scale with high purity. The overall yield was improved from 1% for the sequential solid-phase process to 9% for the fragment synthesis. Finally, a PEGylation process was installed to deliver the drug substance for preclinical and clinical testing. Wiley-VCH Verlag GmbH & Co. KGaA, 2008.

Full text of DOI:10.1002/ejoc.200800722

FULL PAPER
DOI: 10.1002/ejoc.200800722
Manufacturing and PEGylation of a Dual-Acting Peptide for Diabetes
Jochen Krüger,*[a] Torsten Minuth,[a] Werner Schröder,[a] and Jörn Werwath[a]
Keywords: Peptides / Solid-phase synthesis / Synthesis design / Mixed-phase synthesis / Fragment synthesis
In the given manuscript we summarize our activities to de-
velop a scalable synthesis for a PEGylated peptide, which
undergoes development for metabolic disorders. The poly-
peptide contains a sequence of 31 natural amino acids and
carries a site-specific PEGylation at the C-terminus. Initial
synthetic studies revealed that the peptide moiety of the tar-
get molecule is not an ideal target for a linear solid-phase
process, as we were confronted with “difficult sequences”,
which led to a low-yielding process. Subsequently, we devel-
oped a mixed-phase synthesis, which included production of
fragments through solid-phase peptide synthesis, followed
by an assembly of these fragments in solution. We were able
to show that this new process delivered the desired peptide
moiety in gram-scale with high purity. The overall yield was
improved from 1% for the sequential solid-phase process to
9% for the fragment synthesis. Finally, a PEGylation process
was installed to deliver the drug substance for preclinical and
clinical testing.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2008)
companied by a robust and scalable production process
covering a broad scale (from g to kg) with continuous out-
put of the active ingredient in high purity. Our endeavour
towards the development of a robust and scalable synthesis
of 1b is detailed in this manuscript.
Introduction
The advent of GLP-1 (glucagon-like peptide) agonists
has provided a significant impetus for novel therapeutic
concepts in the treatment of type 2 diabetes.[1] At Bayer
Schering Pharma, we pursued a dual mechanism of action,
which combined a GLP-1 agonist with the beneficial prop-
erties of a glucagon antagonist.[2] Accordingly, the active
ingredient was designed as a chimeric peptide that com-
bined structural elements of both native ligands (cf. Fig-
ure 1). Additionally, a site-specific 43 kD PEGylation was
introduced at the C-terminus to increase the metabolic sta-
bility of the peptide and to reduce CNS-mediated nausea,
which is a common adverse event associated with GLP-1
receptor agonists.
Results and Discussion
After its introduction by Merrifield in 1963,[3] solid-
phase peptide synthesis (SPPS) has evolved as a powerful
tool to rapidly access complex peptides.[4] For the supply of
initial quantities of 1a, a sequential SPPS protocol with the
use of a standard Fmoc-tBu procedure was applied to de-
liver peptidic backbone 1a on a milligram-to-gram scale.
However, after cleavage from the resin, the crude peptide
was obtained in poor quality, and tedious chromatographic
purification steps were required to obtain reasonable purity
levels of Ն95 area-% by HPLC. The total yield even for an
optimized sequential SPPS protocol did not exceed 1%.
We speculated that the low performance of the sequential
Figure 1. Structure of the hybrid peptide including structural ele- SPPS process was due to the specific amino acid sequence
ments from both glugacon (1–19) and GLP-1 (20–30) and a site-
specific PEGylation at the C-terminus.
found in 1a. Thus, the growing peptide chain tends to form
aggregates during assembly, which leads to deteriorated
swelling properties of the resin. As a consequence, the ac-
The synthesis of target molecule 1b is divided into two
cessibility of the reagents during synthesis is reduced, which
stages, which include the peptide synthesis followed by
leads to low-yielding transformations. Indeed, in the course
PEGylation. Whereas the PEGylation more or less follows
of the linear solid-phase synthesis, 15 coupling steps had to
a standard protocol, the construction of the polypeptide
be repeated due to incomplete conversion, and the unre-
backbone remains a challenge for synthetic chemistry. In
acted terminal amino groups had to be capped by acety-
particular, a clinical development program has to be ac-
lation. This showed that 1a is not a trivial target for a stan-
dard solid-phase process and can be classified as a “difficult
[a] Bayer Schering Pharma, GDD-CMC-Global Chemical Devel-
opment
sequence”.[5]
Friedrich-Ebert-Str. 42096 Wuppertal, Germany
Fax: +49-202-36-2566
As a consequence, we embarked on a mixed-phase frag-
ment synthesis, which is a well-accepted strategy to cope
E-mail: Joachim.Krueger@bayerhealthcare.com
5936
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 5936–5945

Manufacturing and PEGylation of a Dual-Acting Peptide for Diabetes
with difficult sequences in SPPS.[6] Accordingly, the pro-
tected fragments are produced by SPPS and subsequently
assembled in solution. The high convergence of such an ap-
proach adds value to a complex overall process, as the indi-
vidual arms of the synthesis can be operated in parallel and
the consumption of raw materials and reagents can be sig-
nificantly reduced. More importantly, the risk to come ac-
ross difficult sequences during SPPS is considerably re-
duced, because the fragments are smaller in size.[7] However,
the overall risk to fail with this fragment strategy is shifted
to the late stage of the synthesis, which addresses the solu-
tion-phase peptide synthesis. In particular, issues on solubil-
ity, isolation and purification protocols for the protected
and most likely amorphous intermediates are difficult to
predict in the planning phase.[8]
Figure 2. Protected fragments 2–5 resulting from a retrosynthetic
disconnection between Gly29-Arg30, Ala19-Arg20 and Ala11-Arg12.
(Pbf: 2,2,4,6,7-pentamethyldihydrobenofuran-5-yl-sulfonyl, Fmoc:
9-fluorenylmethoxycarbonyl, Trt: trityl, tBu: tert-butyl, Boc: tert-
butoxycarbonyl).
Synthesis of the Fragments
Retrosynthesis
For the synthesis of fragment 2[10] we anticipated a short
A retrosynthetic analysis for a fragment approach was solution-phase sequence as outlined in Scheme 1. The intro-
guided by the following considerations: (1) If possible, de- duction of the terminal tBu ester towards 7 has been de-
fine fragments of similar size. (2) Avoid critical amino acids scribed for a variety of N-protected cysteine derivatives;[11]
at the disconnections, which are known to undergo racemi- however, only one report details direct esterification com-
zation or side reactions upon activation (e.g., Arg, Asp, mencing from H-Cys(Trt)-OH by using HClO4 in
Cys). (3) Avoid long sequences involving S-trityl protected tBuOAc.[12] From the scale-up perspective, we considered
Cys31, which was found to undergo side reactions during the use of HClO4 as critical due to its explosive properties.
the sequential SPPS protocol.[9] (4) Select lipophilic side- Therefore, we introduced an alternative protocol for the di-
chain protecting groups to increase chances for good solu- rect conversion of 6 into 7 by employing tert-butyl tri-
bility in the later stage of the synthesis.
chloroacetimidate and methanesulfonic acid. After aqueous
Eventually, we divided target molecule 1 into four fully workup, ester 7 was isolated as a solution in toluene, and it
protected fragments as displayed in Figure 2. Protected was directly used in the subsequent coupling step with
Arg-Cys dimer 2 was proposed in order to isolate the sensi- Fmoc-Arg(Pbf)-OH. Coupling was accomplished by a
tive cysteine, whereas the remaining three building blocks TBTU protocol, and protected dipeptide 8 was obtained
3, 4 and 5 were selected for their C-terminal glycine or ala- with a purity of 92 area-% in 58% yield for both steps. Fi-
nine units. From these amino acids we expected good coup- nally, the Fmoc group was removed with a cocktail of pi-
ling kinetics in the assembly of the subsequent fragments peridine in DMF to give fragment 2. This three-step process
and a narrow side-product profile.
provides quick access to desired building block 2, which did
Scheme 1. Solution-phase synthesis of fragment 2. [MtBE: methyl tert-butyl ether, THF: tetrahydrofuran, TBTU: 2-(1H-benzotriazol-1-
yl)-1,1,2,3-tetramethyluronium tetrafluoroborate, HOBt: 1-hydroxy-1H-benzotriazole, DIPEA: N,N-diisopropylethylamine, DMF: N,N-
dimethylformamide].
Eur. J. Org. Chem. 2008, 5936–5945
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5937

J. Krüger, T. Minuth, W. Schröder, J. Werwath
FULL PAPER
Scheme 2. Solid-phase peptide synthesis (SPPS) of fragments 3–5 (AA: side-chain-protected amino acid; PG: protecting group; PG =
Fmoc for 3 and 4, PG = Boc for 5).
not require elaborate chromatographic purification. common use of acetic acid or trifluoroacetic acid, which are
Eventually, fragment 2 was obtained on a 50-g scale with difficult to remove without chromatography and which were
HPLC purity of 93 area-% and a low level of cysteine epi- found to interfere with the coupling reactions during the
merization (0.6%).
subsequent fragment condensations in solution. After cleav-
The remaining three fragments 3–5 were ideal targets for age, the crude products were isolated by precipitation from
SPPS by using an Fmoc–tBu protocol.[13] We selected chlo- diisopropyl ether to give the amorphous products, which
rotrityl resin as the solid-phase carrier for a number of were stored at –20 °C.
reasons:[14] (1) It allows selective and mild cleavage of side-
Accordingly, the three fragments were produced on a
chain protected fragments from the resin. (2) Racemization 100–200-g scale to yield 3, 4 and 5 in good purity after
upon loading is minimized. (3) It has little tendency to form precipitation (cf. Table 1). Moreover, the rather high quality
diketopiperazines. (4) It is available in multikilogram quan- of the SPPS products enabled us to proceed with the frag-
tities.
The generic SPPS protocol, which was used for the pro-
ment condensation without additional purification steps.
Table 1. Results for SPPS for fragments 3–5.
duction of fragments 3–5, is outlined in Scheme 2. Resin
loading was achieved upon addition of the first amino acid
to the resin in the presence of diisopropylethylamine to
yield functionalized resin 9. After loading was complete, the
resin was capped with MeOH to saturate the remaining
active sites. In order to maximize space–time yield for frag-
ment production, we used rather high resin loadings of 0.4–
0.6 mmolg–1.
Resin loading
[mmolg–1]
Yield
[%]
Average yield Purity[a]
per step [%]
[%]
Fragment 3
Fragment 4
Fragment 5
0.4
0.6
0.5
67
82
52
98
99
97
93
93
93
[a] Area-% according to HPLC analysis.
Subsequently, Fmoc deprotection towards 10 was
achieved by treatment of the resin with a 20% piperidine
solution in DMF. Exhaustive rinsing with DMF to effect
complete removal of the piperidine proved essential for a
high-quality peptide.
Fragment Condensation
Condensation of the fragments in solution is displayed
in Scheme 3. The chemical transformations were ac-
Peptide coupling was accomplished by a TBTU/HOBt complished by standard procedures including TBTU-medi-
protocol by using diisopropylethylamine as the base. In our ated peptide couplings and piperidine-induced Fmoc depro-
hands, the optimum concentration of running the coupling tection to give clean conversions for all steps. The stoichi-
reactions was 10 mL of solvent per 1 g of resin. The excess ometry of the fragments was optimized for each step to
amount of reagents can be reduced to 2 equiv. of the amino avoid tedious purification protocols for removal of an ex-
acid, 2 equiv. of the coupling reagent, 0.5 equiv. HOBt cess amount of the fragments.
monohydrate and 2 equiv. of the base. We monitored the
To our delight, we observed good-to-excellent solubility
progress of the peptide coupling reactions by the on-beat properties for the fully protected intermediates throughout
“Kaiser test”[15] and found complete conversion after 1 h in the sequence. However, as a drawback to this sequence, we
all cases without the need to repeat the coupling steps.
experienced that the amorphous character of the reaction
There are a variety of cleavage conditions reported in the intermediates was a major obstacle for convenient isolation
literature.[16] In our hands, a cleavage cocktail of trifluoro- and purification procedures. In particular, reagents and side
ethanol/dichloromethane (1:4) worked reliably, and the con- products, such as the piperidine dibenzofulvene adduct aris-
ditions were mild enough to yield the fully protected frag- ing from Fmoc fragmentation, were difficult to remove dur-
ments in good yields. Additionally, this cocktail avoided the ing workup. Thus, for each step, we elaborated individual
5938
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 5936–5945

Manufacturing and PEGylation of a Dual-Acting Peptide for Diabetes
Scheme 3. Solution-phase coupling of fragments 2–5 (MeCN: acetonitrile, TFE: trifluoroethanol, DTE: dithiothreitol, TIPS: triisopro-
pylsilane).
precipitation protocols in order to isolate the amorphous
We conducted comprehensive analytical characterization
products with a high level of purity as outlined in Scheme 3 of 1a, including, HPLC, MS, N-terminal sequencing, amino
and the Experimental Section. Additionally, residual traces acid analysis, SDS electrophoresis and capillary electropho-
of piperidine were removed by azeotropic distillation. This resis. Accordingly, the identity and amino acid sequence of
procedure operated under mild conditions (40–45 °C, 1a was confirmed, the HPLC purity was determined to be
15 mbar) and no degradation of the protected peptide inter- 96% and no significant racemization had occurred during
mediates was observed.
the synthesis. The main impurity arises from cysteine oxi-
The HPLC purity of the intermediates gradually de- dation, which gives the cystine dimer upon extended storage
creased from 93 area-% for 2 to 79 area-% for fully pro- periods.
tected peptide 1a prior to global deprotection without chro-
In total, the fragment approach towards 1a is a 68-step
matographic purification throughout the entire sequence convergent synthesis with a remarkable overall yield of 9%,
(Figure 3). Eventually, TFA-mediated removal of the pro- which translates into an average yield of 96% per step. In
tecting groups led to a 20–30% loss of purity. The major particular, the mixed-phase synthesis offers a significant re-
impurity found in crude 1a was assigned to a tBu alkylation duction of raw materials and resource consumption. For
product.[17] Thus, we further optimized the cleavage condi- instance, production of 1 g of target peptide 1a required a
tions with a focus on cationic scavengers. Finally, a cocktail total of 11 g of resin and 50 g of protected amino acids,
of triisopropylsilane and dithiothreitol worked best to curb whereas the initial SPPS protocol consumed 50 g of resin
the level of alkylated side products to ca. 10%.[16] Finally, and 740 g of protected amino acids for the same result.
crude peptide 1a was isolated on gram scale with a purity
PEGylation
of 55 area-% and subsequent HPLC purification yielded
The current PEGylation process[18] is depicted in
Scheme 4. Accordingly, peptide moiety 1a was coupled with
a branched 43kD PEG molecule by Michael addition of the
cysteine sulfhydryl group to the PEG maleimide.
pure 1a in not-yet fully optimized 25% yield.
The PEGylation protocol operated in an aqueous buffer
on a 4-L scale by using 4 g of the peptide and 40 g of the
PEG reagent, which is used in slight excess. The reaction
was complete after 30 min at 23 °C, and cysteine was added
to quench the remaining PEG reagent. The crude reaction
mixture was purified by ion-exchange chromatography by
employing SP-sepharose FF as the stationary phase and an
aqueous salt gradient to effect elution of the product. The
pooled product fractions were concentrated by cross-flow
filtration by employing a 10 kD Hydrosart membrane. Fol-
lowing sterile filtration (0.2 µm), the product was stored in
flexible sterile bags as a buffered solution at –78 °C and
defrosted prior to final formulation. Eventually, 2.3 g of
PEGylated drug substance 1b (Figure 4) was obtained per
batch, which corresponds to a yield of 86%.[19]
Figure 3. RP-18-HPLC chromatogram of target peptide 1a (see Ex-
perimental Section for details).
Eur. J. Org. Chem. 2008, 5936–5945
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5939

J. Krüger, T. Minuth, W. Schröder, J. Werwath
FULL PAPER
Amino Acid Analysis: Amino acid analysis was performed by using
an Eppendorf-Biotronik amino acid analyzer LC 3000 equipped
with a UV detector and the data system Chromeleon version 6.8
from Dionex. A modified standard separation program with so-
dium acetate buffer was used. The amino acids were determined
after post-column reaction with ninhydrin on an ion-exchange col-
umn filled with the ion-exchange resin BTC 2410 (4 µ material;
8ϫ150 mm; Laborservice Onken). All chemicals used were of pro
analysis or biochemical grade. Standard amino acid analysis (exam-
ple): the peptide (100 µg) was hydrolyzed with HCl (6 , 200 µL)
in vacuo at 166 °C for 1 h. Subsequently, the HCl was evaporated,
and the sample was diluted with sample buffer (pH 2.2, 400 µL).
For amino acid analysis, 20 µL of the resulting solution were used.
Homoarginine was employed as internal standard. Determination
of cysteine: the peptide (100 µg) was dissolved in performic acid
solution (formic acid/hydrogen peroxide, 9:1; 200 µL), which had
been incubated at room temperature for 2 h prior to use. The pro-
tein was treated at 4 °C for 1 h. The performic acid solution was
evaporated. The sample was then hydrolyzed with 6  HCl and
analyzed as described above. Homoarginine was used as internal
standard. The released amino acids were identified and quantified
with an external 4 n amino acid standard.
Scheme 4. PEGylation process.
N-Terminal Sequence Analysis: N-terminal sequence analyses were
performed by using the gas–liquid–solid-phase protein sequencer
ProciseTM from Applied Biosystems. The standard sequencer
program fast normal was used. The detection of PTH-amino acids
was performed on-line by using
a
RP-18-PTH-column
(220 mmϫ2 mm, 5 µ material) from Applied Biosytems. The PTH-
amino acids were identified and quantified by a 40 p standard of
PTH-amino acids. The data were collected and integrated by using
the sequencer data system 610A from Applied Biosystems. The
peak heights of the PTH-amino acids from the first sequencer cycle
were used for determination of PTH-His and PTH-Ser. N-terminal
sequence analysis of peptide: the peptide (3 nmol) was applied onto
a Biobrene® pretreated sequencer sheet. The peptide was se-
quenced over 32 cycles.
Figure 4. RP-18-HPLC chromatogram of PEGylated peptide 1b
Molecular Weight Determination: Molecular weight determination
was performed by using a LCT system from Waters Corporation
equipped with the Masslynx processing software. Mathematical
transformation of the multiply charged protonated molecular ions
[M + nH]n+ produced the molecular mass profile. The peptide Ala-
Ser-Thr-Thr-Asn-(3,5-Di-iodo)Tyr-Thr was used for internal cali-
bration. The instrument was run in the high-resolution mode. LC–
MS was carried out by using a Finnigan LCQ Deca XP MAX in
combination with a HP 1100 chromatography unit. High-resolu-
tion mass spectrometry was conducted with a LTQ-Orbitrap from
Thermo Scientific.
(see Experimental Section for details).
Conclusions
In summary we identified a scalable synthetic route
towards target peptide 1a, which encompassed a mixed-
phase fragment based synthesis. The salient features of the
new process are: (1) It is a robust and convergent 68-step
synthesis. (2) In comparison to the linear SPPS, the new
protocol is a high-yielding process. (3) Workup and purifi-
cation protocols are adapted to technically feasible pro-
cesses. (4) The current route is the basis for the production
of the peptide on technical scale.
Capillary Zone Electrophoresis: The peptide (~4 ng) dissolved in
0.1% TFA/30% ACN was hydrodynamically applied (1 s) onto the
fused silica capillary. A 100-m potassium phosphate buffer pH 3
was used as running buffer. The applied potential was 20 kV, and
the column temperature was 30 °C. The electropherograms were
measured at 210 nm. Capillary electrophoresis analyses were per-
formed by using a system from Applied Biosystems model 270A-
HT equipped with a data system from Perkin–Elmer and a variable
UV wavelength detector. The fused-silica capillary from Applied
Biosystems had an internal diameter of 50 µm, with a total length
of 72 cm. All chemicals used were of pro analysis grade.
Moreover, a PEGylation protocol was introduced, which
delivered PEGylated peptide 1b in good yields with con-
tinuous high purity.
Experimental Section
SDS-Electrophoresis: 5, 7.5, 10 and 20 µg of the peptide were dis-
solved in 20 µL sample buffer containing a reducing agent. The
samples were incubated for 10 min at 70 °C and then applied onto
a 12% Criterion XT gel and run for about 2 h. The gel was then
General: All SPPS reactions were performed by using custom-made
devices with overhead stirring and porous filter plates at the bot-
tom to allow efficient filtration.
5940
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 5936–5945

Manufacturing and PEGylation of a Dual-Acting Peptide for Diabetes
stained by using Coomassie Blue Brilliant R-250; 0.5–2 µg of the
peptide were applied in case of silver staining. All chemicals used
were of biochemical or pro analysis grade quality. The electropho-
resis chamber, the power supply and the gels were from Bio-Rad.
SDS-electrophoresis under nonreducing conditions was run as de-
scribed above but without reducing agent.
were added at 20 °C, and the mixture was stirred for 30 min. HPLC
analysis (method A) indicated 90% conversion. Therefore, ad-
ditional Fmoc-Arg(Pbf)-OH (9.1 g, 14 mmol) was added, and the
mixture was allowed to stir at 20 °C for an additional 30 min, after
which the conversion rose to 97%. For workup, water (500 mL)
was added, the layers were separated and the organic layer was
washed with water (500 mL). Subsequently, the product solution
was added dropwise to heptane (2.5 L), and the resulting suspen-
sion was stirred for 1 h at 20 °C prior to isolation. The mixture was
filtered, washed with heptane (2ϫ100 mL) and dried in vacuo to
yield Fmoc-Arg(Pbf)-Cys(Trt)-OtBu (8; 98 g) in a quality of
85 area-% (HPLC method A). For further purification, the crude
product was redissolved in toluene (490 mL), and the resulting
solution was added dropwise to diisopropyl ether (2.45 L). The re-
sulting suspension was stirred at 20 °C for 1 h. The mixture was
then filtered and washed with diisopropyl ether (2ϫ100 mL). Fi-
nally, the product was dried in vacuo at 20 °C to yield 8 (84 g) with
a purity of 92 area-% (method A). This corresponds to a yield of
58% for two steps commencing from H-Cys(Trt)-OH. The resulting
Fmoc-Arg(Pbf)-Cys(Trt)-OtBu (8; 84 g, 80 mmol) was dissolved in
DMF/piperidine (3:1, 840 mL) at 0 °C. After 10 min the reaction
mixture was slowly added to water (4.2 L) over a period of 3.75 h,
and the mixture was stirred at 0 °C for an additional 30 min. The
mixture was then filtered and washed with water (2ϫ200 mL).
Subsequently, it was redissolved in methyl tert-butyl ether, and the
aqueous layer was separated. The organic layer was added to di-
isopropyl ether (2.1 L) over a period of 1.5 h. The mixture was
stirred for 30 min at 20 °C, filtered and washed with diisopropyl
ether (2ϫ100 mL). The residue was dried in vacuo for 18 h to yield
2 (51 g, 77%). According to HPLC analysis (method C), the prod-
uct was obtained in 93% purity containing 0.6% of the dia-
stereomer H-Arg(Pbf)--Cys(Trt)-OtBu. 1H NMR (500 MHz, [D6]-
DMSO): δ = 1.30–1.65 (m, 4 H), 1.31 (s, 9 H), 1.40 (s, 6 H), 1.85
(br. m, 2 H), 1.98 (s, 3 H), 2.35–2.50 (m, 2 H), 2.42 (s, 3 H), 2.50
(s, 3 H), 2.95 (s, 2 H), 3.03 (m, 2 H), 3.11 (m, 1 H), 4.12 (m, 1 H),
6.40 (br. s, 1 H), 6.58 (br. s, 1 H), 7.22–7.35 (m, 15 H), 8.27 (br. s,
1 H) ppm. 13C NMR (125 MHz, [D6]DMSO): δ = 12.1, 17.2, 17.5,
18.9, 27.4 (3 C), 28.2 (2 C), 33.2, 42.4, 51.4, 53.9, 54.8, 66.0, 81.1,
86.2, 116.1, 124.2 (3 C), 126.7 (6 C), 128.0 (6 C), 128.9, 131.3,
137.1, 144.0 (3 C), 155.9, 157.3, 169.1, 174.9 ppm. HRMS (ESI+):
calcd. for C45H57O6N5S2 [M + H] 827.375; found 828.3814.
NMR Spectroscopy: NMR spectra were recorded with a Bruker
AC500 instrument. All spectra are calibrated against tetramethyl-
silane as an internal standard (δ = 0 ppm).
HPLC Chromatography
Method A: Column: Phenomenex Prodigy ODS3, column sym-
metry: 50ϫ4.6 mm, 3 µm; T = 45 °C; eluent: A = 1 mL formic acid
in 1 L of water, B = 1 mL formic acid in 1 L of acetonitrile, gradi-
ent: 0–10 min from 95 to 0% A, 10–20 min 100% B; flow rate:
1.2 mLmin–1; detection at 220 nm.
Method B: Column: Chiral AD-H, column symmetry:
250ϫ4.6 mm, 5 µm; T = 45 °C; eluent: A = n-heptane + 0.2%
trifluoroacetic acid, B = ethanol + 0.2% trifluoroacetic acid, 0–
20 min isocratic at 90% A; flow rate: 1 mLmin–1; detection at
210 nm.
Method C: Column: ZORBAX 300SB-C3 (Agilent), column sym-
metry: 150ϫ3.0 mm, 3.5 µm; T = 45 °C; eluent: A = water + 0.1%
trifluoroacetic acid, B = tetrahydrofuran + 0.05% trifluoroacetic
acid, gradient: 0–45 min from 90 to 0% A, 45–50 min 100% B;
flow rate: 0.5 mLmin–1; detection at 220 nm.
Method D: Column: ZORBAX 300SB-C3 (Agilent), column sym-
metry: 150ϫ3.0 mm, 3.5 µm; T = 45 °C; eluent: A = water + 0.1%
trifluoroacetic acid, B = tetrahydrofuran + 0.05% trifluoroacetic
acid, gradient: 0–15 min from 90% A to 100% B, 15–20 min 100%
B; flow rate: 0.5 mLmin–1; detection at 220 nm.
Method E: Column: Phenomenex Prodigy ODS3, column sym-
metry: 150ϫ3.0 mm, 3 µm; T = 60 °C; eluent: A = 0.1% TFA in
water, B = 0.1% TFA in acetonitrile, gradient: 0–10 min 23% B,
10–40 min from 23 to 38% B; 40–70 min from 38 to 100% B, 70–
75 min 100% B; flow rate: 1.0 mLmin–1; detection at 220 nm.
Method F: Column: Zorbax-C18-column, column symmetry:
4.6 mmϫ125 mm; 3.5 µm material; T = 40 °C; eluent: A = 0.1%
TFA in water, B = 0.1% TFA in acetonitrile, gradient: 0–10 min
0% B, 10–70 min from 0 to 60% B; 80 min 0% B; flow rate:
0.7 mLmin–1; detection at 192–300 nm.
Synthesis of Fragment Fmoc-[20–29]-OH (3)
Resin Loading: At 20 °C, the chlorotrityl resin (100 g; loading:
1.56 mmolg–1, source: Iris Biotech) was allowed to swell for 30 min
in dichloromethane (800 mL). The mixture was filtered and fresh
dichloromethane (800 mL), Fmoc-Gly-OH (14.8 g, 49.8 mmol) and
diisopropylethylamine (DIPEA; 20.2 g 156.3 mmol) were added.
The mixture was stirred for 2 h at 20 °C, filtered and then rinsed
with dichloromethane (800 mL). Subsequently, the mixture was
stirred with a solution of 5% N,N-diisopropyl ethylamine and 10%
methanol (500 mL) in dichloromethane for 10 min. After filtration,
this process was repeated allowing a reaction time of 20 min. Sub-
sequently, the resin was washed with DMF (1ϫ500 mL) and
dichloromethane (3ϫ500 mL) and dried in vacuo at 20 °C to yield
the loaded resin (110 g). The loading was determined as follows:[21]
a sample of the resin (ca. 10–20 mg) was incubated with a solution
of 20% piperidine in DMF (50 mL) for at least 4 h. The absorbance
of the supernatant was measured at 301 nm, and the loading was
Synthesis of Fragment H-[30–31]-OtBu (2): To a solution of H-
Cys(Trt)-OH (50.0 g, 138 mmol) in THF (500 mL) was added
methylsulfonic acid (33.2 g, 345 mmol) under an inert gas atmo-
sphere within 10 min. The temperature should not exceed 25 °C
during addition. Subsequently, at 20 °C a solution of tert-butyl tri-
chloroacetimidate (151 g, 690 mmol) in methyl tert-butyl ether
(1.0 L) was added, and the mixture was stirred for 18 h at 20 °C.
For workup, water (500 mL) was added, and the pH was adjusted
to 9.0 by employing aqueous NaOH (1 , 355 mL). The phases
were separated, and the organic layer was washed with water
(500 mL), dried with magnesium sulfate and concentrated in vacuo.
The residue was suspended in toluene (500 mL), and the mixture
was stirred at 20 °C for 30 min and then filtered. To the filtrate was
added water (250 mL), and the pH was adjusted to 11 with aqueous
NaOH (1 ). The layers were separated, and the toluene layer was
washed with water (2ϫ250 mL).[20] The resulting toluene solution
of H-Cys(Trt)OtBu (7) (86 area-%, method A) was diluted with
DMF (125 mL). Subsequently, HOBt monohydrate (9.3 g,
calculated according to the Lambert–Beer law with
ε =
7800 Lmol–1 cm–1. In this example, the loading was 0.38 mmolg–1.
69 mmol), N,N-diisopropylethylamine (53.5 g, 414 mmol), Fmoc- Fmoc Deprotection: After resin loading and prior to the first depro-
Arg(Pbf)-OH (53.8 g, 83 mmol) and TBTU (66.5 g, 207 mmol) tection, the resin was allowed to swell in DMF (500 mL) for 1 h.
Eur. J. Org. Chem. 2008, 5936–5945
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5941

J. Krüger, T. Minuth, W. Schröder, J. Werwath
Synthesis of Fragment Fmoc-[12–19]-OH (4)
FULL PAPER
In general, Fmoc deprotection was accomplished upon stirring of
the resin in a 20% piperidine solution in DMF (1 L) for 10 min at
20 °C. The mixture was filtered, and the process was repeated with
a fresh portion of the piperidine solution (1 L) for another 10 min.
Finally, the mixture was filtered and rinsed with DMF (5ϫ1 L).
On a routine basis, the filtrates were tested on traces of piperidine
by using the chloranil test. Accordingly, the filtrate (3 drops) was
added to a solution of acetone (3 mL) containing three drops of
a concentrated solution of chloranil (tetrachlorobenzochinone) in
toluene. The blue colour indicated the presence of amine in the
filtrate.
Resin Loading: The loading of chlorotrityl resin (100 g) was ac-
complished as outlined for fragment 3 by using Fmoc-Ala-OH
(24.9, 80.0 mmol). The loading was determined by UV
(0.56 mmolg–1).
SPPS: Fmoc deprotection, peptide coupling and cleavage from the
resin were accomplished in analogy to the method described for 3.
The following amino acids were used: Fmoc-Arg(Pbf)-OH
(3ϫ81.7 g), Fmoc-Ala-OH (39.2 g), Fmoc-Asp(OtBu)-OH
(51.8 g), Fmoc-Leu-OH (44.5 g), Fmoc-Tyr(tBu)-OH (57.8 g). In
situ analyses (Kaiser test and HPLC) revealed that all reactions
went to completion within the provided timeframe. Thus, no re-
coupling steps or acetyl capping was necessary. After cleavage from
the resin and precipitation from diisopropyl ether, 4 (109 g,
51.6 mmol) was isolated with 93 area-% (method A) according to
HPLC analysis. This corresponds to a yield of 82%. The molecular
mass was determined by LC–MS (LC: method C, MS: ion trap
ESI+): m/z = 1056 [(M+ 2 H)/2]+. 1H NMR (500 MHz, [D6]-
DMSO): δ = 8.22 (d, J = 7.6 Hz, 1 H), 8.01–8.05 (m, 2 H), 7.94–
7.99 (m, 3 H), 7.91 (br. s., 2 H), 7.89 (s, 2 H), 7.77–7.81 (m, 1 H),
7.68–7.73 (m, 2 H), 7.45–7.49 (m, 1 H), 7.40 (s, 2 H), 7.31 (s, 2 H),
7.08 (s, 2 H), 7.01 (br. s., 1 H), 6.77 (d, J = 7.9 Hz, 2 H), 6.28–6.72
(m, 7 H), 4.48–4.63 (m, 2 H), 4.17–4.34 (m, 7 H), 4.05–4.15 (m, 1
H), 3.88–3.98 (m, 1 H), 2.91–3.08 (m, 13 H), 2.74–2.79 (m, 1 H),
2.64–2.72 (m, 1 H), 2.49–2.50 (m, 1 H), 2.48 (s, 9 H), 2.43 (s, 9 H),
2.00 (s, 9 H), 1.30–1.80 (m, 42 H), 1.25 (d, J = 7.0 Hz, 3 H), 1.17–
1.18 (m, 3 H), 1.19 (s, 9 H), 0.83 (d, J = 6.4 Hz, 3 H), 0.79 (d, 3
H) ppm. [α]2D0 = –5.8 (c = 0.40, DMSO).
Coupling Reactions: To a suspension of the resin in DMF (1 L) was
added the amino acid (83 mmol, 2 equiv.), HOBt (2.8 g, 21 mmol,
0.5 equiv.), N,N-diisopropylethylamine (10.8 g, 83 mmol, 2 equiv.)
and TBTU (26.7 g, 83 mmol, 2 equiv.), and the mixture was stirred
at 20 °C for 1 h. The reaction mixture was drained, and the resin
was rinsed with DMF (3ϫ1 L). The following amino acids were
used for the synthesis of 3: Fmoc-Arg(Pbf)-OH (2ϫ54.0 g), Fmoc-
Val-OH (28.2 g), Fmoc-Leu-OH (29.4 g), Fmoc-Trp(Boc)-OH
(43.8 g), Fmoc-Lys(Boc)-OH (39.0 g), Fmoc-Ile-OH (29.4 g),
Fmoc-Phe-OH (32.2 g), Fmoc-Glu(OtBu)-OH (35.4 g). In situ
analyses were conducted by using the Kaiser test: a small aliquot
of the resin was thoroughly rinsed with ethanol. Subsequently, 2
drops of each of the following solution were added: 80% phenol in
ethanol, 5% ninhydrin in ethanol and 2% of an aqueous 1 m
KCN solution in pyridine. The mixture was heated to 100 °C for
5 min, and the resin was washed with ethanol. A blue-coloured
resin indicated the presence of amine, which corresponds to insuf-
ficient conversion. Alternatively, in selected examples the progress
of the reactions was also monitored by HPLC. Thus, a sample of
the resin was thoroughly rinsed and incubated three times, each
with of a solution of 0.5% trifluoroacetic acid in dichloromethane
for 10 min. The combined filtrates were diluted with acetonitrile
and 3 µL were injected (HPLC method A). During the production
of fragment Fmoc-[20–29]-OH (3), all coupling reactions were
complete within 1 h, as indicated by the Kaiser test and HPLC
analysis, respectively.
Synthesis of Fragment Boc-[1–11]-OH (5)
Resin Loading: The loading of chlorotrityl resin (200 g) was ac-
complished as outlined for fragment 3 by using Fmoc-Ala-OH
(50 g, 160 mmol). The loading was determined via UV
(0.51 mmolg–1).
SPPS: Fmoc deprotection, peptide coupling and cleavage from the
resin were accomplished in analogy to the method described for 3.
The following amino acids were used: Fmoc-Tyr(tBu)-OH
(119.3 g), Fmoc-Asp(OtBu)-OH (99.4 g), Fmoc-Ser(tBu)-OH
(2ϫ92.7 g), Fmoc-Thr(tBu)-OH (2ϫ96.2 g), Fmoc-Phe-OH
(93.6 g), Fmoc-Gly-OH (71.9 g), Fmoc-Gln(Trt)-OH (147.6 g),
Boc-His(Trt)-OH (120.4 g). In situ analyses (Kaiser test and
HPLC) revealed that all reactions went to completion within the
provided time frame. Thus, no recoupling steps or acetyl capping
was necessary. After cleavage from the resin and precipitation from
diisopropyl ether, Boc-[1–11]-OH (5; 135 g 63.3 mmol) was isolated
with 93 area-% (method A) according to HPLC analysis. This cor-
responds to a yield of 52%. The molecular mass was determined
by LC–MS (LC: method C, MS ion trap APES): m/z = 2132 [M +
H]+, 1068 [(M + 2 H)/2]+, 1891 [(M – Trt) + H]+. 1H NMR
Cleavage from the Resin: The resin was rinsed with dichlorometh-
ane (1 L). Subsequently, it was stirred in a mixture of dichlorometh-
ane/trifluoroethanol (4:1, 1 L) for 1 h at 20 °C. This step was re-
peated allowing for an extended reaction time of 1.5 h. The resin
was rinsed with dichloromethane (500 mL), and the filtrates were
combined. The majority of the solvent was distilled off, leaving
285 g of a concentrated solution, which was slowly added to di-
isopropyl ether (2 L) at 20 °C. It was filtered off and dried in vacuo
at 20 °C for 16 h to yield the desired fragment 3 (64 g) with a purity
of 93 area-% (HPLC; method A). This corresponds to an overall
yield of 67%. The molecular mass was determined by LC–MS (LC:
method C, MS: ion trap ESI+): m/z = 2286 [M + H]+. H NMR (500 MHz, [D6]DMSO): δ = 8.48 (s, 1 H), 8.17 (d, J = 7.0 Hz, 2
1
(500 MHz, [D6]DMSO): δ = 8.21 (d, J = 7.0 Hz, 1 H), 7.94–8.09
(m, 5 H), 7.84–7.93 (m, 5 H), 7.63–7.77 (m, 4 H), 7.46–7.54 (m, 2
H), 8.00 (d, J = 7.9 Hz, 1 H), 7.94 (d, J = 7.6 Hz, 1 H), 7.79–7.88
(m, 3 H), 7.75 (d, J = 7.0 Hz, 1 H), 7.65–7.72 (m, 2 H), 7.33–7.40
H), 7.37–7.44 (m, 2 H), 7.25–7.36 (m, 3 H), 7.07–7.25 (m, 6 H), (m, 10 H), 7.19–7.27 (m, 11 H), 7.14–7.19 (m, 10 H), 7.12 (d, J =
6.72–6.96 (m, 3 H), 6.69 (t, J = 4.9 Hz, 1 H), 6.45 (br. s., 3 H), 8.5 Hz, 2 H), 7.03–7.08 (m, 6 H), 6.88 (d, J = 7.9 Hz, 1 H), 6.82
4.64–4.71 (m, 1 H), 4.54–4.62 (m, 1 H), 4.33–4.40 (m, 1 H), 4.10–
4.32 (m, 8 H), 3.93–4.03 (m, 1 H), 3.64–3.77 (m, 2 H), 2.88–3.11
(d, J = 8.2 Hz, 2 H), 6.67 (s, 1 H), 4.71–4.81 (m, 1 H), 4.54–4.62
(m, 1 H), 4.46–4.53 (m, 1 H), 4.11–4.41 (m, 7 H), 3.90–3.95 (m, 1
(m, 11 H), 2.80–2.87 (m, 2 H), 2.70–2.79 (m, 1 H), 2.50 (s, 3 H), H), 3.78–3.84 (m, 1 H), 3.75 (d, J = 4.6 Hz, 2 H), 3.49–3.56 (m, 2
2.48 (s, 3 H), 2.43 (s, 3 H), 2.42 (s, 3 H), 2.10–2.18 (m, 2 H), 2.00 H), 3.43 (ddd, J = 15.4, 9.5, 5.3 Hz, 2 H), 3.02–3.09 (m, 1 H), 2.99
(s, 3 H), 1.99 (s, 3 H), 1.96–1.98 (m, 1 H), 1.75–1.85 (m, 1 H), 1.63– (dd, J = 13.9, 4.1 Hz, 1 H), 2.81–2.92 (m, 2 H), 2.69–2.80 (m, 2
1.72 (m, 7 H), 1.61 (s, 9 H), 1.52–1.59 (m, 4 H), 1.42–1.50 (m, 6 H), 2.59 (dd, J = 15.7, 5.3 Hz, 1 H), 2.40 (dd, J = 15.9, 7.9 Hz, 1
H), 1.40 (s, 6 H), 1.38 (s, 6 H), 1.35 (s, 9 H), 1.33 (s, 9 H), 1.30 (br.
s., 1 H), 1.12–1.26 (m, 3 H), 0.67–0.89 (m, 18 H) ppm. [α]2D0 = –9.1
(c = 0.41, DMSO).
H), 2.32 (br. s., 2 H), 1.83–1.95 (m, 1 H), 1.65–1.78 (m, 1 H), 1.35
(s, 9 H), 1.30–1.33 (m, 9 H), 1.26–1.28 (m, 3 H), 1.25 (s, 9 H), 1.16
(s, 9 H), 1.10 (s, 9 H), 1.06 (s, 9 H), 1.03 (s, 9 H), 1.00 (d, J =
5942
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 5936–5945

Manufacturing and PEGylation of a Dual-Acting Peptide for Diabetes
6.1 Hz, 3 H), 0.95 (d, J = 6.4 Hz, 3 H) ppm. [α]2D0 = +12.0 (c =
0.43, DMSO).
at 20 °C for 40 min. The reaction mixture was poured into water
(3.73 L) and stirred for 30 min at 20 °C. The mixture was filtered,
washed with water (3ϫ200 mL) and finally dried at 30 °C in vacuo
to give Boc-[1–31]-OtBu (160.5 g, 99%). HPLC (method D):
79 area-%. MS (ion trap ESI+): m/z = 1595 [(M+ 3 H)/3]+, 1715
[(M+ 4 H)/4]+, 1372 [(M+ 5 H)/5]+.
Assembly of the Fragments in Solution
Fmoc-[20–31]-OtBu: From a solution of H-[30–31]-OtBu (2; 29.9 g,
32.8 mmol, 95 area-%) in DMF (1 L) approx. 500 mL DMF was
distilled off at 40–45 °C and 15 mbar. Fresh DMF (500 mL) was
added, and the distillation protocol was repeated twice to ensure
complete removal of trace amounts of piperidine. Subsequently,
Fmoc-[20–29]-OH (3; 75.0 g 32.9 mmol, 93 area-%), HOBt mono-
hydrate (2.22 g, 16.4 mmol), DIPEA (12.7 g, 98.4 mmol) and
TBTU (21.1 g, 65.6 mmol) were added, and the mixture was stirred
at 20 °C for 30 min. The reaction mixture was added to water
(3.75 L), and the solution was stirred for 30 min at 20 °C. The mix-
ture was filtered and rinsed with water (1 L) and finally dried at
25 °C in vacuo to give Fmoc-[20–31]-OtBu (98.4 g, 97%). HPLC
(method D): 92 area-%. The molecular mass was determined by
LC–MS (LC: method C, MS: ion trap ESI+): m/z = 1549 [(M+ 2
H)/2]+.
H-[1–31]-OH (1a): To a solution of Boc-[1–31]-OtBu (159 g,
23.2 mmol) in of a mixture of dichloromethane/trifluoroethanol
(4:1, 636 mL) was added dithiothreitol (239 g, 1549 mmol), triiso-
propylsilane (239 mL, 1166 mmol) and trifluoroacetic acid
(1.35 L), and the mixture was stirred at 20 °C for 4 h. To the reac-
tion mixture was added dichloromethane (636 mL), and the re-
sulting solution was poured into cold diisopropyl ether (5.4 L)
within 2 min. The solution was stirred at 5 °C for 1 h, filtered and
rinsed with diisopropyl ether (3ϫ500 mL). Finally, it was dried for
3 d at 20 °C in vacuo to yield crude peptide 1a (126 g) with an
HPLC purity of 55 area-% (method D). The purification protocol
was demonstrated with an aliquot of crude 1a (11.9 g). The crude
material was dissolved in
a mixture of acetonitrile/water +
H-[20–31]-OtBu (12): A solution of Fmoc-[20–31]-OtBu (98.0 g,
31.65 mmol) in DMF/piperidine (3:1, 490 mL) was stirred at 20 °C
for 1 h. To the reaction mixture was added acetonitrile (490 mL),
and the resulting solution was poured within 10 min into diisopro-
pyl ether (1.96 L). The solution was stirred at 20 °C for 1 h, filtered
and washed with diisopropyl ether (3ϫ250 mL). Finally, it was
dried at 20 °C in vacuo to yield 12 (78.2 g, 86%). HPLC (method
D): 83 area-%. LC–MS (LC: method C, MS: ion trap ESI+): m/z
= 1437 [(M+ 2 H)/2]+.
0.1%TFA (30:70, 1125 mL). The resulting solution was purified by
employing a Prodigy 10 µ ODSIII 100A (Phenomenex), particle
size 10 µm, column dimension: 250ϫ50 mm. A total of 45 injec-
tions each of 25-mL (264 mg of crude peptide) were used. The elu-
ent consisted of A: water + 0.1% TFA and B: acetonitrile and a
flow of 140 mLmin–1 was applied. The following gradient was
used: 0 min injection, 0.5–23.0 min A from 75 to 20%, 23.0 to
29.0 min A at 75%. The fractions were analyzed by using HPLC
method D. Fractions with Ͼ95 area-% were pooled, acetonitrile
was removed in vacuo and the residue was freeze dried to yield
2.05 g of the pure peptide. This corresponds to a yield of 25% for
two steps commencing from 13.
Fmoc-[12–31]-OtBu: From a solution of H-[20–31]-OtBu (12;
73.3 g, 25.5 mmol, 83 area-%) in DMF (1.1 L) approx. 550 mL
DMF was distilled off at 40–45 °C and 15 mbar. Fresh DMF
(550 mL) was added, and the distillation protocol was repeated
twice to ensure complete removal of piperidine. Subsequently,
Fmoc-[12–19]-OH (4; 56.5 g 26.8 mmol), HOBt monohydrate
(1.72 g, 12.8 mmol), DIPEA (9.9 g, 76.5 mmol) and TBTU (16.4 g,
51 mmol) were added, and the mixture was stirred at 20 °C for
30 min. The reaction mixture was poured into water (1.83 L), and
the solution was stirred for 30 min at 20 °C, filtered, washed with
Amino Acid Analysis of 1a: The composition of 1a matches the
composition of the theoretical numbers very well by taking into
account the limitations of the method (cf. Table 2).
Table 2. Amino acid analysis of 1a.
Amino acid
Theory
Found
for 1a
-Amino acid
water (2ϫ250 mL) and finally dried at 30 °C in vacuo to give
Fmoc-[12–31]-OtBu (126.0 g, 99%). HPLC (method D): 79 area-
%. LC–MS (LC: method C, MS: ion trap ESI+): m/z = 1656 [(M+
3 H)/3]+.
(%)[a]
Cys
1
2
2
–
2
2
2
3
1
1
–
2
2
2
1
1
6
1
31
0.99
2.00
1.55
–
1.21
1.89
1.95
2.92
0.96
0.89
–
1.74
1.89
2.01
1.27
1.18
5.75
1.00
29.20
5.84
80.73
1.60
0.20
Ͻ0.10
Ͻ0.10
0.87
0.13
–
Asp
Thr[b]
H-[12–31]-OtBu (13): To a suspension of Fmoc-[12–31]-OtBu
(125.7 g, 25.3 mmol, 79 area-%) in dioxane/piperidine (3:1,
629 mL) was added acetonitrile (629 mL) to give a clear solution
after 5 min stirring at 20 °C. The solution was stirred at 20 °C for
an additional 40 min. Then, the reaction mixture was poured into
diisopropyl ether (4.71 L) within 5 min, and it was stirred at 20 °C
for 30 min and at 0 °C for an additional 30 min. The residue was
isolated by filtration by using a vacuum not below 100 mbar to
avoid blockage of the frit. The product was rinsed with diisopropyl
ether (3ϫ330 mL). Finally, the product was dried at 30 °C in vacuo
to obtain 13 (120.3 g, quant. yield). HPLC (method D): 77 area-%.
The molecular mass was determined by MS (ion trap ESI+): m/z
= 2372 [(M+ 2 H)/2]+.
allo-Thr
Ser[b]
Glu
Gly
Ala
Val
Ile
allo-Ile
Leu
Tyr
Phe
His
Lys
Arg
Trp
Total
1.51
Ͻ0.10
Ͻ0.10
Ͻ0.10
Ͻ0.10
0.69
Ͻ0.10
1.13
0.13
Ͻ0.10
0.10
Boc-[1–31]-OtBu: From a solution of H-[12–31]-OtBu (13; 112.1 g,
23.6 mmol, 77 area-%) in DMF (1.68 L) approx. 560 mL DMF was
distilled off at 40–45 °C and 15 mbar. Fresh DMF (560 mL) was
added, and the distillation protocol was repeated twice. Sub-
sequently, Boc-[1–11]-OH (5; 55.4 g, 26.0 mmol, 93 area-%), HOBt
monohydrate (1.60 g, 11.8 mmol), DIPEA (9.15 g, 70.8 mmol) and
TBTU (15.2 g, 47.2 mmol) was added, and the solution was stirred
Loss by hydrolysis (%)
Amino acid content
[a] Level of racemization determined at C.A.T. Analysentechnik
Tübingen, Germany. [b] Low values found for Thr and Ser are
explained by decomposition during hydrolysis; Cysteic acid was
determined from performic acid oxidation.
Eur. J. Org. Chem. 2008, 5936–5945
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5943

J. Krüger, T. Minuth, W. Schröder, J. Werwath
FULL PAPER
Protein Sequence Analysis of 1a: Automated Edman degradation
was performed. Total sequence analysis was carried out over 32
steps to assess identity with the theoretical sequence, to detect mi-
croheterogeneities, clipping and blockage of the peptide. One main
sequence (AA 1–30) was detected which confirmed the primary
structure of 1a on the amino acid level. The PTH-Cys in position
31 gave no UV signal and could therefore not be detected. No
microheterogeneity was detected.
chloric acid (11 mL). The solution was filtered through a Sartopore
2 filter (300 5441307H5–00-B) and stored at 2–8 °C for 2 d. The
second batch of the reaction mixture was purified accordingly. Both
product fractions were combined to give a total of 16.5 L of the
purified product. The cross-flow filtration unit was equilibrated
with 0.5  sodium hydroxide and subsequently flushed with water
for injection until the pH was neutral. The pooled product fractions
from chromatography were concentrated to a volume of 1 L by
using a Slice Hydrosart 10 kD membrane. Eventually, the cross-
flow filtration unit was washed with the washout buffer (1 L) to
yield a combined product phase of 2 L. The level of endotoxins
was determined (Ͻ0.12 EEmL–1), and it was filtered through a Sar-
topore 2 filter (300 5441307H5–00-B) into sterile bottles and finally
stored at –20 °C.
Molecular Weight Determination of 1a: The molecular weight of 1a
was determined by ESI-MS measurement in the high-resolution
mode with internal calibration with 3757.1 Da. This is in good
agreement with the calculated monoisotopic mass of 3757.9 Da.
Reverse-Phase Chromatography of 1a (Method F): Compound 1a
eluted as a main peak at 46 min with a purity of 94 area-%. The
cystine dimer was present at a level of ca. 1%.
The product concentration was determined by UV absorbance (c
= 1.18 gL–1). This translates into a yield of 2.36 g (86%) based on
the peptide. HPLC (method E): 100 area-%.
Capillary Zone Electrophoresis: Compound 1a eluted as one main
peak with a purity of 93%.
N-Terminal Sequence: Compound 1b (3 nmol) was applied onto a
Biobrene® pretreated sequencer sheet. The conjugate was se-
quenced over 32 cycles. One main sequence (AA 1–30) was de-
tected, which confirmed the primary structure of 1a on the amino
acid level.
SDS-Gel Electrophoresis: Compound 1a ran under reducing condi-
tions as a single band with an apparent molecular weight of about
4 kD. Only minor amounts of a dimeric structure were present,
whereas other oligomeric forms were not found.
Synthesis of PEGylated Peptide 1b
SDS-Electrophoresis: The Coomassie blue gel was stained again by
using the silver stain method to identify the free peptide. Only trace
amounts of free peptide 1a were found after silver staining, which
indicates that the peptide was completely bound to the PEG moi-
ety. PEG conjugate 1b ran as a broad band at about 70 kD.
PEGylation Buffer: A solution of sodium dihydrogen phosphate
(7.9 g) in water (5.0 L) for injection was adjusted to pH 6.5 by using
1  sodium hydroxide (11 mL).
Chromatography Buffer A: A solution of sodium acetate trihydrate
(150 g) in water (55 L) for injection was adjusted to pH 4.5 by using
2  acetic acid (750 mL).
Acknowledgments
Chromatography Buffer B: A solution of tris(hydroxymethyl)ami-
nomethane (133 g) and sodium chloride (321 g) in water (55 L) for
injection was adjusted to pH 9.5 by using 1.2  hydrochloric acid
(25 mL).
The authors are grateful to W. Jöntgen, T. Kuhlmann and G.
Pfeiffer for analytical support. J. K. thanks S. Brückner, D. Heuser,
R. Renelt and T. Scherkenbeck for their skillful preparative work,
K. Siegel for helpful discussions on the retrosynthesis as well as
H.-C. Militzer, L. Böhmer and G. Quabeck for conducting auto-
mated SPPS with a Chemspeed workstation.
Washout Buffer: A solution of sodium acetate trihydrate (5.4 g) in
water (2 L) for injection was adjusted to pH 5.0 by using 2  acetic
acid (9 mL). For all buffers endotoxins were determined prior to
use.
[1] For recent reviews on GLP-1, see: a) A. P. Abhay, D. M. Ravic-
hand, V. Seshayamma, C. P. Satish, Int. J. Endocrinol. Metab.
2006, 4, 151–166; b) C. F. Deacon, Diabetes Obes. Metab. 2007,
9, 23–31; c) F. Giorgino, L. Laviola, A. Leonardini, A. Natalic-
chio, Diabetes Res. Clin. Pr. 2006, 74, 152–155; d) M. C. Man-
cini, A. Halpern, Expert Opin. Invest. Drugs 2006, 15, 897–915;
e) C. M. Rotella, L. Pala, E. Mannucci, J. Endocrinol. Invest.
2005, 28, 746–758.
[2] a) T. H. Claus, C. Q. Pan, J. M. Buxton, L. Yang, J. C. Reyn-
olds, N. Barucci, M. Burns, A. A. Ortiz, S. Roczniak, J. N. Liv-
ingston, K. B. Clairmont, J. P. Whelan, J. Endocrinol. 2007,
192, 371–380; b) C. Q. Pan, J. M. Buxton, S. L. Yung, I. Tom,
L. Yang, H. Chen, M. MacDougall, A. Bell, T. H. Claus, K. B.
Clairmont, J. P. Whelan, J. Biol. Chem. 2006, 281, 12506–
12515.
[3] R. B. Merriefield, J. Am. Chem. Soc. 1963, 85, 2149.
[4] a) See, for instance: N. L. Benoiton, Chemistry of Peptide Syn-
thesis Sensors Update (Eds.: H. Baltes, W. Göpel, J. Hesse),
CRC Press Taylor & Francis Group, Boca Raton, 2006 pp.
125–154; b) M. Amblard, J.-A. Fehrentz, J. Martinez, G. Subra
in Peptide Synthesis and Applications (Ed.: J. Howl), Humana
Press, Totowa, New Jersey, 2005, pp. 3–25.
[5] See, for instance: N. Sewald, H.-D. Jakubke in Peptides: Chem-
istry and Biology, Wiley-VCH, Weinheim, 2002, pp. 230 ff.
[6] a) B. Riniker, A. Flörsheimer, H. Fretz, P. Sieber, B. Kamber,
Tetrahedron 1993, 49, 9307–9320; b) K. Barlos, D. Gatos, G.
Papaphotiou, W. Schäfer, Liebigs Ann. Chem. Ann. Chem.
PEGylation: All operations were carried out in a sterile environ-
ment. A 4-L glass reaction vessel was flooded with a 0.5  aqueous
solution of sodium hydroxide. It was allowed to stand overnight,
and the vessel was drained and washed with water for injection
until the pH was neutral. To a clear solution of PEGylation reagent
(40.0 g; GL2–400 MA Sunbright from NOF, Japan) in PEGylation
buffer (2 L) was added a solution of 1a (4.0 g, 69% peptide content
according to UV-absorbance) in PEGylation buffer (2 L) at 23 °C.
The pH of the resulting solution was adjusted to 6.5 by using 1 
sodium hydroxide (4.5 mL). The mixture was stirred (250 rpm) for
30 min at 23 °C before cysteine (0.96 g) dissolved in PEGylation
buffer (40 mL) was added. The reaction mixture was filtered
through a Sartopore 2 filter (300 5441307H5–00-B), and the re-
sulting solution (4.04 L) was submitted to subsequent chromatog-
raphy in batches (2ϫ2.02 L). The column was packed with 2.9–
3.0 L of SP Sepharose Fast Flow, which was equilibrated by using
ethanol (20% in water for injection). Prior to use, the column was
flooded with 0.5  sodium hydroxide and then washed with
chromatography buffer A. The column was charged with the first
portion of the reaction mixture, and it was washed with one column
volume of chromatography buffer A and then a gradient towards
100% chromatography buffer B was initiated. The desired product
eluted at a buffer A/buffer B ratio of approx. 60:40. The pH of the
product fraction (8.0 L) was adjusted to 5.0 by using 1.2  hydro-
5944
www.eurjoc.org
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2008, 5936–5945

Manufacturing and PEGylation of a Dual-Acting Peptide for Diabetes
1993, 215–220; c) B. L. Bray, Nat. Rev. Drug Discovery 2003,
2, 587–593; d) E. Atherton, R. Berry, D. Simpkin, D. Wellings,
C. Willis, J. A. Eriksen, S. Buoen, Innovation and Perspectives
in Solid Phase Synthesis & Combinatorial Libraries: Peptides,
Proteins and Nucleic Acids-Small Molecule Organic Chemistry
Diversity, Collected Papers, International Symposium, 6th,
York, United Kingdom, Aug. 31–Sept. 4 1999, 2001; e) L. An-
dersson, L. Blomberg, M. Flegel, L. Lepsa, B. Nilsson, M. Ver-
lander, Biopolymers 2000, 55, 227–250; f) T. Bruckdorfer, O.
Marder, F. Albericio, Curr. Pharm. Biotechnol. 2004, 5, 29–43.
[7] S. B. H. Kent, Annu. Rev. Biochem. 1988, 57, 959.
[8] S. Sakakibara, Biopolymers 1999, 51, 279–296.
[9] Cys31 partially underwent side reactions during SPPS via a de-
hydroalanine intermediate, which was formed by S-trityl elimi-
nation. In particular, this undesired side reaction was found
during treatment with base, for example, during piperidine-me-
diated Fmoc deprotection.
[14] See, for instance a) S. Bräse, S. Dahmen in Handbook of Combi-
natorial Chemistry (Eds.: K. C. Nicolaou, R. Hanko, W. Hart-
wig), Wiley-VCH, Weinheim, 2002, pp. 59–169; b) F. Z.
Dörwald in Organic Synthesis on Solid Phase Wiley-VCH,
Weinheim, 2000, pp. 33–131.
[15] E. Kaiser, R. L. Colescott, C. D. Bossinger, P. I. Cook, Anal.
Biochem. 1970, 34, 595.
[16] See, for instance: K. Barlos, D. Gatos in Fmoc Solid Phase
Peptide Synthesis – A Practical Approach (Eds.: W. C. Chan,
P. D. White), Oxford University Press, New York, 2000, pp.
215–228.
[17] The alkylation site has not yet been determined. We assume
that Cys, Tyr or Trp are alkylated.
[18] I. Tom, V. Lee, M. Dumas, M. Madanat, J. Ouyang,
J. Severs, J. Andersen, J. M. Buxton, J. P. Whelan, AAPS Jour-
nal 2007, 9, 227–234. http://www.aapsj.org/articles/aapsj0902/
aapsj0902025/aapsj0902025.pdf.
[19] The yield includes the peptide content of the starting peptide
(approx. 80%), which contains stabilizing salts and residual
water. Additionally, the indicated weight is based on the pep-
tide portion disregarding the PEG portion.
[20] A portion of the toluene solution was evaporated to dryness,
and the residue was analyzed (HPLC method B). Accordingly,
the ester was formed with complete retention of the stereogenic
centre (99.8%ee).
[10] Fmoc-Cys(Trt)OtBu has been described previously: L. Chen,
N. Fotouhi, D. Y. Jackson, J. Tilley, WO 00/63234, 2000.
[11] a) K. R. West, K. D. Bake, O. Sijbren, Org. Lett. 2005, 7, 2615–
2618; b) A. E. Koziol, B. Szady, E. Masiukiewicz, B. Rzeszot-
arska, M. A. Broda, A. Malgorzata, Chem. Pharm. Bull. 2001,
49, 418–423.
[12] E. Masiukiewicz, B. Rzeszotarska, Org. Prep. Proced. Int. 1999,
31, 571–572.
[13] See, for instance: W. C. Chan, P. D. White in Fmoc Solid Phase
Peptide Synthesis – A Practical Approach (Eds.: W. C. Chan,
P. D. White), Oxford University Press, New York, 2000, pp. 41–
74.
[21] M. Mergler, J. P. Durieux, The Bachem Practice of SPPS, Ba-
chem AG, 2000.
Received: July 21, 2008
Published Online: November 5, 2008
Eur. J. Org. Chem. 2008, 5936–5945
© 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
5945