Solid-phase synthesis of polyfunctional polylysine dendrons using aldehyde linkers

Article abstract of DOI:10.1016/j.tetlet.2014.04.127

A straightforward method for the solid-phase synthesis of C-terminally modified polylysine dendrons has been developed by applying bisalkoxybenzaldehyde and trisalkoxybenzaldehyde linkers. The method has been used for the synthesis of polylysine dendrons with a variety of C-terminal 'tail groups' such as alkyl, propargyl, and dansyl to give dendrons in high crude purity. Furthermore, the method was successful for the synthesis of dendrons with multiple N-terminal pentapeptide groups together with C-terminal alkyl and propargyl tail groups. Finally, the method was shown to be well-suited for automated synthesis.

Full text of DOI:10.1016/j.tetlet.2014.04.127

Accepted Manuscript
Solid-phase synthesis of polyfunctional polylysine dendrons using aldehyde
linkers
Daniel K. Svenssen, Sahar Mirsharghi, Ulrik Boas
PII:
S0040-4039(14)00764-3
DOI:
http://dx.doi.org/10.1016/j.tetlet.2014.04.127
TETL 44589
Reference:
Tetrahedron Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
4 March 2014
3 April 2014
30 April 2014
Please cite this article as: Svenssen, D.K., Mirsharghi, S., Boas, U., Solid-phase synthesis of polyfunctional
polylysine dendrons using aldehyde linkers, Tetrahedron Letters (2014), doi: http://dx.doi.org/10.1016/j.tetlet.
2014.04.127
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Solid-phase synthesis of polyfunctional polylysine dendrons using aldehyde linkers
Daniel K. Svenssen, Sahar Mirsharghi, Ulrik Boas
1

Solid-phase synthesis of polyfunctional polylysine dendrons using aldehyde linkers
Daniel K. Svenssena, Sahar Mirsharghi,a Ulrik Boasa, b *
a Innate Immunology Group, National Veterinary Institute, The Technical University of Denmark.
Bülowsvej 27, DK-1870 Frederiksberg, Denmark
b Center for Nanomedicine and Theranostics, The Technical University of Denmark, Building 345E,
Ørsteds Plads, DK-2800 Kongens Lyngby, Denmark
Email: uboa@vet.dtu.dk
Abstract
A straightforward method for the solid-phase synthesis of C-terminally modified polylysine
dendrons has been developed by applying bisalkoxybenzaldehyde and trisalkoxybenzaldehyde
linkers. The method has been used for the synthesis of polylysine dendrons with a variety of C-
terminal ‘tail groups’ such as alkyl, propargyl and dansyl to give dendrons in high crude purity.
Furthermore, the method was successful for the synthesis of dendrons with multiple N-terminal
pentapeptide groups together with C-terminal alkyl and propargyl tail groups. Finally, the method
was shown to be well-suited for automated synthesis.
Dendrimers and dendrons belong to a class of molecularly defined/monodisperse hyperbranched
molecules that present a high number of functional groups at their surface. The high number of
functionalitites on the surfaces of dendrons and dendrimers enables these structures to act as high
affinity ligands for various biological receptors. Therefore, these structures have great potential in
2

the development of new biomaterials.1 Dendrimers and dendrons are categorized both by the
monomers they comprise and by the number of branching points following a chain from the core, or
in lysine dendrons, from the C-terminal to the periphery (N-terminal). The number of branching
points in the chain is denoted by the generation number (Gn) of the dendron.2 Hence the
dendrimer/dendron structure grows larger for each generation in a defined manner.
Dendrimers and dendrons based on the amino acid L-lysine were initially developed by
Denkewalter3 and their application in biological research was pioneered by Tam and coworkers.4 In
this regard, there have been several reports on the synthesis of poly-L-lysine (PLL) dendrons by
solid-phase synthesis (SPS), both by Tam’s group5,6 and subsequently by other research groups.7-10
The reported SPS of PLL dendrons applies a Rink amide or Rink acid linker to give dendrons with
a primary amide or carboxylic acid as the C-terminal functional group. However, being C-terminal
protective groups, these linkers give only limited possibility to modify the C-terminus of the
dendron with an additional functionality during solid-phase synthesis.
The ability to tailor specific polyvalent structures with several functionalities may be valuable in the
preparation of, e.g., biologically active compounds. As we are interested in preparing
multifunctional dendron ligands for a large variety of receptors, we set out to develop a
straightforward strategy for the solid-phase synthesis of a large variety of polyfunctional dendron
scaffolds. A peptide linker allowing both C-terminal and N-terminal modification would pave the
way for the facile introduction of several functional groups during the synthetic sequence.
Linkers based on bis- and trisalkoxybenzaldehydes such as the backbone amide linker (BAL, with
an ortho or para substituted spacer) and a 4-(4-formyl-3-methoxyphenoxy)butanoic acid (FMPB)
linker (Figure 1) were initially developed for the solid-phase synthesis of C-terminally modified
peptides. Subsequently, they have found wide use in the synthesis of modified and cyclic peptides
as well as small non-oligomeric molecules.11-14
3

Figure 1. Bis- and trisalkoxybenzaldehyde derivatized aminomethylated polystyrene resins.
In the solid-phase synthesis of dendrons the ‘backbone amide linkage approach’ enables the
introduction of a large variety of C-terminal ‘tail’ functionalities early in the synthetic sequence,
without the need to introduce additional orthogonal protecting groups (Scheme 1). The initial step
in the synthesis is the formation of a secondary amine by reductive amination between the linker
aldehyde group and a C-terminal functionality which contains an amino group. By subsequent
acylation of the secondary amine with lysine, a dendron can be synthesized on the resin by
divergent synthesis. Due to stabilization of the alkoxybenzyl carbonium ion, the dendron substrate
with the secondary amide anchor point can be released in a ‘traceless’ manner from the resin under
mild acidic conditions (Scheme 1).
4

Scheme 1. Synthetic sequence for the formation of poly-L-lysine dendrons with a variety of C-
terminal groups.
The bisalkoxybenzaldehyde (FMPB) and trisalkoxybenzaldehyde (ortho-/para-BAL) linkers15 were
linked to aminomethylated polystyrene resin [1% divinylbenzene (DVB) cross-linked] or high
loading rigid macroporous polystyrene (10% DVB cross-linked) by an amide linkage using
(benzotriazol-1-yloxy)tripyrrolidinophosphonium hexafluorophosphate (PyBOP) and diisopropyl
amine (DIPEA).15,16 The introduction of the linker onto the resin was monitored by infrared
spectroscopy (attenuated total reflectance). Here an aldehyde band appeared at approximately 1715
cm-1 (rigid macroporous resin) or at approximately 1670 cm-1 (1% DVB cross-linked resin) together
with an amide band (approx. 1650 cm-1 for the rigid macroporous resin and 1600 cm-1 for the 1%
DVB cross-linked resin). Hence, a red-shift was generally observed with the 1% DVB cross-linked
5

resin compared to the rigid macroporous resin.17,18 The presence of aldehyde groups on the resin
was furthermore visualized colorimetrically by either the purpald test19 (where the presence of resin
aldehyde groups gave purple beads), or by reaction with 2,4-dinitrophenylhydrazine (DNPH) in
sulfuric acid (orange-red to deeply red beads with resin aldehyde groups).20 The loading of linker
groups on the resin was reduced to avoid ‘crowding’ on the resin during synthesis of the more bulky
dendrons.21 The dendron C-terminal functionality was introduced onto the aldehyde linker resin via
reductive amination using sodium cyanoborohydride in a dipolar aprotic solvent (DMF or NMP)
under mild acidic catalysis (5% acetic acid). In this step, amines with a large variety of functional
groups could be introduced. The reductive amination reactions were generally complete within 2-3
hours according to IR and purpald or DNPH tests.
On-resin IR analysis of the rigid macroporous resin after reductive amination of the aldehyde
generally showed strong reduction of the aldehyde band (approx. 1715 cm-1) in accordance with the
negative DNPH test. In contrast, with the 1% DVB cross-linked resin, only a slight reduction of the
aldehyde band (approx. 1670 cm-1) was observed relative to the amide band, although purpald and
DNPH tests were negative. However, upon reductive amination with long chain alkyl amines, both
resins showed a significantly increased C-H stretching band (approx. 2900 cm-1) relative to the
aldehyde band, which was indicative of a successful reductive amination.17,19,20 With long chain
alkyl amines, e.g., dodecylamine, a precipitate formed in NMP, so in these cases, a less polar
solvent mixture [THF/NMP (9:1)] was applied to better dissolve the amine during the reaction.
The subsequent acylation of the secondary amine on resin with Fmoc-Lys(Fmoc)-OH was generally
performed using N,N′-diisopropylcarbodiimide (DIC) twice over two hours (which worked better
compared to a 4 hours reaction time). In this procedure, the insoluble N,N′-diisopropylurea
byproduct precipitated within 10 minutes. This urea byproduct could be removed by washing the
resin with MeOH and NMP. By employing the coupling agent fluoro-N,N,N′,N′-
6

tetramethylformamidinium hexafluorophosphate (TFFH) for the acylation of the secondary amine
and reaction times two hours (twice), this precipitation could be avoided and similar coupling yields
were obtained. Phosphonium or uronium peptide coupling agents such as PyBOP or O-
(benzotriazol-1-yl)-N,N,N′,N′-tetramethyluronium hexafluorophosphate (HBTU), did not give
complete acylation of the secondary amine under these reaction conditions. For this step the content
(loading) of the amines was determined by measuring the UV absorption of the released
22
dibenzofulvene-piperidine adduct (λ=290 nm) from the Fmoc protected amines on the resin.
Generally, in the synthesis of dendron substrates, higher loadings (loading ~ 0.3 mmol/g) were
obtained with the ortho-BAL linker compared to the para-BAL and FMPB linkers (loading ≤ 0.1
mmol/g).
The subsequent dendron synthesis was carried out by Fmoc-based SPS. In some cases, Fmoc-
Lys(Boc)-OH was applied in the coupling step to release free ε-amines in the dendron product upon
acidolytic release of the dendron from the resin. In the divergent build-up of the dendron, only two
equivalents of the activated amino acids per amine were applied to afford complete acylation of the
terminal amines. This is a lower excess than the 4-5 equivalents generally needed to complete
acylations in the synthesis of peptides on solid phase. The acylations were complete within one hour
in the formation of G1-dendrons (indicated by a negative ninhydrin test). However, upon stepwise
divergent formation of G1- and G2-dendrons carrying N-terminal peptide groups (TYDKT, which
is a partial T cell epitope23), the coupling times in the synthesis of the peptide sequence usually
increased to 2-3 hours before obtaining a negative ninhydrin test. In the synthesis of dendrons with
N-terminal TYDKT peptides, delayed coloring during the ninhydrin test was occasionally observed.
This led to false negative results after Fmoc deprotection of the amino acids with bulky side-chain
groups (e.g., aspartic acid with a tert-butyl ester side-chain group, or threonine). However,
7

inconclusive ninhydrin results could be overcome by prolonged heating during the analysis (5
minutes at 80 o C), which then gave reliable colorimetric results.
In order to use the method as a tool for C-terminal fluorescence labeling of dendrons we first
reductively aminated the aldehyde linker with 1,6-diaminohexane followed by reaction of the
primary amine with fluorescein isothiocyanate. However, the introduction of fluorescein at the C-
terminus led to complex product mixtures. This may be due to the presence of two phenolic groups
and a carboxylic acid on the fluorescein structure, which are prone to reaction during the subsequent
synthesis of the dendron. However, two routes to introduce an alternative dansyl fluorophore gave a
satisfactory outcome, albeit with somewhat lower crude purities of the products. The dansyl group
could either be introduced via reductive amination with a mono-dansyl modified hexane diamine,24
followed by acylation and Fmoc-based SPS of the dendron. Alternatively, a similar result could be
obtained by reductive amination with 1,6-diaminohexane followed by selective dansylation of the
primary amine. To increase the selectivity towards the reaction with a primary amine in the
presence of a secondary amine, we found it necessary to deactivate the dansyl chloride by reaction
with hydroxybenzotriazole (HOBt) to form an intermediate HOBt-sulfonic acid ester.25
Amino groups at the dendron surface could be released as free amines or acetylated amines. Also, a
triglycine peptide was successfully introduced by segment coupling, opening up possible segment
coupling by other peptide sequences comprising a C-terminal glycine residue. The G1- and G2
dendronized triglycine derivatives were synthesized in 37-44% isolated yields possessing either
dodecyl or 6-aminohexyl tail groups (Table 1). To further investigate the scope of the present
method, the automated synthesis of dendron structures with N-terminal peptide groups was carried
out on a peptide synthesizer. Here, we synthesized two G1-dendrons with a C-terminal dodecyl and
propargyl tail and two N-terminal peptide groups. G1- and G2-dendrons with a propargyl tail group
(8 and 18) and with two and four N-terminal TYKDT peptide groups, respectively, were
8

successfully synthesized by automated synthesis (Table 1). Here, product 18 was of high crude
purity according to HPLC-MS, whereas crude product 8 showed a minor impurity with an M + 96
mass. The M + 96 mass could result from either the TFA ester or trifluoroacetamide adduct of 8
formed during the acidolytic cleavage from the resin. The propargyl tail group was chosen, since it
allows the introduction of a large variety of functionalities by click chemistry. The G1-dendron with
a dodecyl tail group 7 was synthesized both manually and by automated synthesis, the crude
product 7 prepared by automated synthesis showed some minor impurities according to HPLC-MS
with an M + 96 (TFA ester or TFA amide) together with a minor HPLC peak with mass M – 128.
The M – 128 mass was found to be a deletion sequence missing one lysine residue.17 As deletion
byproducts were not observed in the automated synthesis of dendrons with a propargyl group at the
C-terminus, this could indicate that the hydrophobic dodecyl chain influences the acylation steps
due to steric effects, occasionally giving deletion sequences. Hence, in these cases, the automated
synthesis protocol may be further optimized to resolve this problem. However, pure product 7 was
obtained by manual solid-phase synthesis. To further determine the structure of the dendronized
TYDKT peptides, derivatives 7 and 18 were analyzed by amino acid analysis (AAA) and found to
be in accordance with the expected amino acid compositions, Lys5Thr4Asp2Tyr2 and
Lys11Thr8Asp4Tyr4, respectively.
9

Table 1
Synthesized dendrons and their isolated yields24
Compound
G1-PLL((NH2)4)-N-dodecyl
G1-PLL((NH2)4)-N-(6-aminohexyl)
G1-PLL((NH2)2 (AcNH)2)-N-dodecyl
G1-PLL((NH2)2 (AcNH)2)-N-(6-aminohexyl)
G1-PLL((NH2)2 (AcNH)2)-N-propargyl
Yield (%)
1
2
3
4
5
6
7
8
9
47
72
69
52
56
G1-PLL((NH2)2(AcNH)2)-N-(6-dansylaminohexyl) 67
G1-PLL((NH2)2(TYKDT)2)-N-dodecyl
G1-PLL((NH2)2(TYKDT)2)-N-propargyl
G1-PLL(GGG)4-N-dodecyl
42
44
44
44
31
42
55
60
35
68
37
32a
10 G1-PLL(GGG)4-N-(6-aminohexyl)
11 G2-PLL((NH2)8)-N-dodecyl
12 G2-PLL((NH2)8)-N-(6-aminohexyl)
13 G2-PLL-((NH2)4(AcNH)4)-N-dodecyl
14 G2-PLL((NH2)4(AcNH)4)-N-(6-aminohexyl)
15 G2-PLL((NH2)4(AcNH)4)-N-benzyl
16 G2-PLL((NH2)4(AcNH)4)-N-propargyl
17 G2-((NH2)4(GGG)4)-N-dodecyl
18 G2-PLL((NH2)4(TYKDT)4)-N-propargyl
a A somewhat higher loading (0.11 mmol/g) and a lower isolated yield was obtained due to the lack
of capping of the free amine groups on the resin.
In summary, a straightforward method for the preparation of polyfunctional lysine dendrons in good
yields and crude purities is presented. The method paves the way for the introduction of a large
variety of C-terminal functionalities including molecular labels or functionalities which allow
further derivatisation in solution, e.g., by click chemistry. Peptide N-terminal end-groups can be
introduced by solid-phase segment coupling or by stepwise synthesis. The method can furthermore
be implemented in automated SPS of dendrons.
10

References
1. For example, see: (a) Tomalia, D. A.; Christensen, J. B.; Boas, U. Dendrimers, Dendrons and Dendritic Polymers.
Cambridge University Press 2012, (b) Boas, U.; Christensen, J. B.; Heegaard, P. M. H. Dendrimers in Medicine
and Biotechnology. RSC Publishing 2006.
2. In this paper we apply a nomenclature which is a combination of peptide nomenclature and dendrimer
nomenclature where a generation of poly-L-lysine dendron is named as follows: Gn-PLL (N-terminal end groups)n
–N-(C-terminal end group) where n= number of N-terminal end groups, and G= generation and n= the generation
number
3. Denkewalter, R. G.; Kole, J.; Lukasavage, W. J. US patent 4,289,872 (1981)
4. Tam, J.P. Proc. Natl. Acad. Sci. 1988, 85, 5409-5413
5. Posnett, D.N.; McGrath, H.; Tam, J.P. J. Biol. Chem. 1988, 263, 1719-1725
6. Lu, Y.A.; Tam, J.P. Proc. Natl. Acad. Sci. 1989, 86, 9084-9088
7. (a) Kantchev, E.A.B.; Chang, C.-C.; Cheng, S.-F.; Roche, A.-C.; Chang, D.-K. Org. Biomol. Chem. 2008, 6, 1377-
1385 (b) Kantchev, E. A. B.; Chang, C.-C.; Chang, D.-K. Biopolymers Pept. Sci. 2006, 84, 232-240.
8. Tang, Y.H.; Huang, A.Y.T.; Chen P.Y.; Chen, H.T.; Kao, C. L. Pharm. Design 2011, 17, 2308-2330
9. Papadopoulos, A.; Shiao, T. C.; Roy, R. Mol. Pharm. 2012, 9, 394-403
10. Shiao, T. C.; Roy, R. New J. Chem. 2012, 36, 324-339
11. Jensen, K.J.; Alsina, J.; Songster, M. F.; Vágner, J.; Albericio, F.; Barany, G. J. Am. Chem. Soc. 1998, 120, 5441-
5452
12. Boas, U.; Brask, J.; Jensen K.J. Chem. Rev. 2009, 109, 2092-2118
13. Swayze, E. E. Tetrahedron Lett. 1997, 38, 8465-8468
14. Sarantakis, D.; Bicksler, J. J. Tetrahedron Lett. 1997, 38, 7325-7328
15. The ortho-BAL linker was not commercially available and was therefore synthesized according to: Boas, U.;
Brask, J.; Christensen, J. B.; Jensen, K. J. J. Comb. Chem. 2002, 4, 223-228
16. Procedure for anchoring the aldehyde linker to aminomethylated polystyrene resin: the aldehyde linker, e.g.,
FMPB or ortho-/para-BAL (1 equiv.), PyBOP (1 equiv.) and HOBt (1 equiv.) were suspended in NMP
(approximately 7.3 mL/g resin). However, it was later concluded that HOBt could be omitted from the general
procedure. DIPEA (2 equiv.) was added and the mixture was swirled for 5 min at r.t. The mixture was transferred
11

to a filter syringe/glass container with a glass filter containing the aminomethylated polystyrene resin (0.5 or 1
equiv., loading: 2 mmol/g) and shaken for 1 h at r.t., and then washed with NMP (5x10 mL) and CH2Cl2 (5x10
mL). The resin was capped with Ac2O/DIPEA/CH2Cl2 10/5/85 (7-10 mL/g resin) for 2 h, or until the ninhydrin test
showed negative. The derivatized resin was washed with NMP (5x10 mL), CH2Cl2 (5x10 mL) and MeOH (2x5
mL), and then air-dried followed by drying in vacuo. The IR spectrum was recorded showing bands at 1680 cm-1
(C=O stretch, amide), around 1700 cm-1 (C=O stretch, aldehyde); positive purpald and DNPH tests were observed.
17. Analytical data including spectra and chromatograms can be found in the Supporting Information
18. A strong IR band from residual acetic anhydride was seen at approximately 1790 cm-1 after capping of the rigid
macroporous resin and the anhydride proved difficult to remove by washing. Treatment with 20% MeOH/5%
DIPEA in CH2Cl2 overnight removed the band, but sometimes gave rise to a smaller band at approximately 1770
cm-1, presumably due to a methyl acetate (C=O stretch) formed during the MeOH treatment. However, with the 1%
DVB cross-linked resin, no band from residual acetic anhydride was observed. We conclude that the reduced
swelling ability of the rigid macroporous resin must somehow have an effect on entrapment of acetic anhydride
inside the beads.
19. The purpald aldehyde test was performed as described by Cournoyer, J.J.; Kshirsagar, T.; Fantauzzi, P.P; Figliozzi,
G.M.; Makdessian T.; Yan, B. J. Comb. Chem. 2002, 4, 120-124
20. The DNPH test on the aldehyde functionality was performed according to the procedure described by Shannon, S.
K.; Barany, G. J. Comb. Chem. 2004, 6, 165-170
21. Tam, J.P.; Lu, Y.A. J. Am. Chem. Soc. 1995, 117, 12058-12063
22. Determination of the resin loading was carried out as follows: resin (3 x 5 mg) was transferred to falcon tubes and a
solution of piperidine in NMP (1:4, 25 mL) was added and the tube shaken for 30 min along with a reference tube
containing only the piperidine solution. Each of the four samples were transferred to a cuvette and the absorbance
measured by photospectrometry at 290 nm. The absorbance was used to calculate the loading of the resin according
to the formula: loading in mmol/g = 4.76 Abs/mass of resin in mg
23. Bay, S.; Lo-Man, R.; Isinaga, E.; Nakada, H.; Leclerc, C; Cantacucene, D. J. Peptide Res., 1997, 49, 620-625
24. Detailed experimental procedures on the synthesized compounds can be found in the Supporting Information
25. Data not shown
12