(MeO–PEG–NH2), to increase retention by the membrane.3
Since PEGylated peptides have received much recent attention
due to their enhanced therapeutic and pharmacokinetic
potential,12–14 in some cases a bioactive peptide made by
MEPS might also be used directly without cleavage from PEG.
Fmoc (9-fluorenylmethoxycarbonyl) peptide synthesis
chemistry was chosen due to its widespread application and
mild deprotection conditions (piperidine/DMF). Hydroxy-
methylphenylacetic acid (HMPA) was used as linker to attach
the first amino acid in the peptide sequence to the
MeO–PEG–NH2 enabling facile cleavage of the final peptide
via acidolysis. Peptide coupling proceeds optimally in polar
aprotic solvents such as dimethylformamide (DMF) or N-methyl
pyrrolidinone (NMP), and DMF was used in both reaction
and diafiltration steps. The use of these solvents in com-
bination with OSN membranes has only recently become
possible with the development of membranes15 which possess
good stability in these aprotic solvents and make this approach
viable.
suggested that traces of impurities would still remain in the
system after each purification step even after 10 volumes of
washing solvent, but this did not affect the final peptide purity.
Apparently the level of impurities that can be tolerated in the
system without provoking side reactions is higher than antici-
pated, and so there is potential to reduce the washing volume.
Encouraged by this first success, Thymopentin (H-Arg-Lys-
Asp-Val-Tyr-OH) was synthesized as a second demonstration
of the MEPS process. Thymopentin (TP-5) is a derivative of
the naturally occurring hormone thymopoietin, with potential
for the treatment of rheumatoid arthritis, AIDS and other
primary immunodeficiencies.16 Besides being a potential active
pharmaceutical, this peptide includes a range of amino acids
from aromatic (Tyr), acidic (Asp) and basic (Lys and Arg) to
hydrophobic (Val). It also contains the largest Fmoc/Boc
protected amino acid Fmoc-Arg(Boc)2, MW 597 g molÀ1
,
and so this synthesis represents a significant further challenge
for MEPS. RP-HPLC analyses of TP-5 produced by both
MEPS and SPPS are illustrated in Fig. 2.
The most important consideration for successful realisation
of this approach was the choice of membrane. This must
possess excellent long term stability in the reaction solvent
(DMF) and high selectivity between MeO–PEG–peptide,
and side reaction products and excess reagents, including
unreacted amino acids, activators and deprotection reagents.
Membrane performance should not be affected by frequent
switching of the reaction media between DMF solution in
the coupling step and 20% piperidine/DMF solution in the
deprotection step.
The purity of the MEPS product (as a percentage of total
(TP-5 plus peptidic by-products)) was estimated at 94%.
MALDI-TOF analysis (Fig. 3) confirmed the target product
molecular weight of MH+ 680 and identified the two impurities
as formed by deletion of Asp, MH+ 564, and Lys, MH+ 550.
TP-5 produced by SPPS under the same reagent excess
(2 equivalents) was only 77% pure and the main impurity
was identified as deletion of Arg, MH+ 524. This result
demonstrates the key advantage of liquid phase synthesis
over SPPS—a higher purity was obtained using the same
After screening a range of commercial and developmental
membranes we identified a ceramic OSN membrane that met
these requirements: the Inopor ZrO2 coated membrane with
3 nm pore size and hydrophobic surface modification (Innocermic,
Germany). Based on the membrane characterisation data it was
estimated that 10–12 solvent wash volumes should be sufficient
for removal of all excess reagents from around 1.0 to less than
0.01 equivalents.
To prove the MEPS concept we started by producing
a
model peptide H-Tyr-Ala-Tyr-Ala-Tyr-OH. This was
chosen as it includes one of the largest protected amino acids
Fmoc-Tyr(tBu)-OH, and one of the smallest hydrophobic
amino acids, Fmoc-Ala-OH, thus providing information on
the performance of the MEPS process with respect to different
molecular sizes and properties of amino acids. PyBOP was
chosen for the coupling reaction as it is one of the largest of
the commercial activators available, and its successful removal
presents a challenge for the MEPS process. DIC was used
for the esterification linking the first amino acid onto
MeO-PEG-HMPA. Thus the synthesis also provided insight
into the behaviour of the post-reaction species derived from
both activators during diafiltration. Finally, this first experi-
ment also sought to establish membrane stability at high
concentrations of organic base (piperidine) during the depro-
tection step (see Table S1, ESIw).
Fig. 2 HPLC chromatograms of peptide TP-5 produced by MEPS and
SPPS processes, and TP-5 standard purchased from Sigma-Aldrich
(UK). The target TP-5 peptide was eluted at 10.3 minutes. Both
syntheses (MEPS and SPPS) were performed under the same reaction
conditions of 2 equivalents of reagents per 1 equivalent peptide, and
single reaction cycles. Peptide purity was determined as a ratio between
the target peptide TP-5 peak area and the total area of the peaks
corresponding to peptide sequences in the solution. The purity of TP-5
produced by MEPS was determined as B94% (two impurities eluted at
10.0 minutes and 10.4 minutes) while TP-5 produced by SPPS was
B77% pure (one impurity eluted at 10.5 minutes). The large peaks
eluted between 19–23 minutes were PEGylated wastes such as
MeO-PEG-HMPA and the peak eluted at 13 minutes was not of
peptide origin as confirmed by MALDI-TOF analysis and was not
taken into account for the purity calculations.
The model peptide produced contained no peptide by-products
as confirmed by RP-HPLC and MALDI-TOF analysis. This
absence of any detectable peptide impurities indicates that
membrane purification is efficient at removing un-reacted
protected amino acids throughout the synthesis. Our calculations
ꢀc
This journal is The Royal Society of Chemistry 2010
Chem. Commun., 2010, 46, 2808–2810 | 2809
So, Sheung
