N. Winssinger et al.
step in the sequence, the deprotection of Mtt was complete
in under 5 min by using either 1% TFA in dichloromethane
or in 10 min with neat hexafluoroisopropanol (HFIP). How-
ever, neither conditions were deemed ideal; TFA is corro-
sive and is not recommended for use in automated synthe-
sizers, whereas HFIP does not swell well the resin. Diluting
the HFIP with dichloroethane gave comparably fast reaction
while providing a better resin swelling profile. Finally, to
keep the amine resulting from the deprotection protonated
as a salt, a small excess of 1-hydroxybenzotriazole (HOBt)
was included in the deprotection solution. This proved to be
beneficial for the Mtt deprotection as well, affording com-
plete deprotection in less than 5 min. We then used the best
conditions identified in the solid-phase synthesis of oligo-
higher specificity for the perfect match versus mismatched
compared with unmodified (average ratio of PM/MM for
modified PNA and unmodified PNA are 3.2 and 2.1, respec-
tively). These results are in good agreement with previous
reports on the benefit of including g-modifications originat-
ing from l-amino acids.
We next investigated the scope of the chemistry that
could be achieved with PES, focusing particularly on reac-
tions that have proven to be important for library synthesis
in providing diversity and complexity.[14] To this end, a resin
containing a first codon with the four different possible nu-
cleobases (TGCA)[15] including a serine modified derivative
(A) was used to evaluate different reactions (Figure 4). The
outcomes of all reactions were assessed by MALDI and LC/
MS analysis of the crude reaction cleavage product. Palladi-
um-catalyzed reactions certainly stand as one of the most
utilized reaction type in library synthesis. Resin 15 contain-
ing an Mtt-protected PNA tag was derivatized with 4-iodo-
benzoic acid to obtain 16, which was engaged in Suzuki–
Miyaura cross-coupling[16] with styrenyl boronic acid under
classical coupling conditions (palladium tetrakis) to afford a
complete cross-coupling reaction (17). Heck reaction[17] was
equally productive, affording 18 in excellent yield. Stille
cross-coupling[18] also proceeded as anticipated to yield the
expected biaryl 19. Palladium-catalyzed p-allylation[19] reac-
tions were also found to be very productive. Treatment of
PNA-tagged amine 20 with an excess of allyl acetate in the
presence of 10% palladium-Dppe catalyst cleanly afforded
the expected double allylation product 21. Reaction of a
PNA-tagged secondary amine (proline: 22) with a more con-
gested p-allyl system (1-phenyl allyl acetate) still afforded
the desired compound 23 in good yield (more than 85%
conversion, regioselectivity extrapolated from the same re-
action performed in solution). Ring-closing metathesis
(RCM) reactions have been used extensively in DOS, and
ruthenium-based catalysts are known to be compatible with
peptides.[20] Indeed, substrate 24 containing two terminal al-
kenes underwent smooth RCM with first generation Grubbs
catalyst to yield 25.[21] The Cu-catalyzed azide alkyne cyclo-
addition[22] has proven to be one of the most robust reac-
tions, and azide substrate 26 was converted into the triazole
27 in excellent yield. Beyond transition-metal-catalyzed re-
actions, the aldehyde function is a versatile starting point for
diversification. PNA-encoded aldehyde 28 could be smooth-
ACHTUNGTRENNUNGmers containing one modified PNA monomer per codon. As
shown in Figure 2, the crude cleavage of an 18-mer PNA
showed a single major peak, corresponding to the desired
oligomer, attesting to the efficiency of the chemistry.
We next investigated the impact of serine-modified PNA
in the context of microarray hybridization. To this end, mi-
croarrays containing combinatorial permutations of four sets
of codons representing 625 combinations of 14 mers were
used. The arrays are particularly informative with regards to
sequence selectivity because, for any given sequence on the
array, there are 19 other sequences that share three out of
the four codons. We chose a sequence that had previously
been identified as particularly promiscuous due to a long
stretch of GC nucleotides and measured the fluorescent in-
tensity for the perfect match as well as the brightest mis-
match starting at 2 nm with two-fold dilutions (i.e., 1, 0.5,
0.25 and 0.125 nm, and 62.5, 31.2, and 16 pM). As shown in
Figure 3, the modified PNA was three to five times more in-
tense at a given concentration, resulting in more sensitive
detection threshold (31.2 pM) relative to unmodified PNA
(125 pM). More importantly, the modified PNA showed a
ly engaged in reductive amination using NaBHACHTUNGTRENNUNG(OAc)3,
Horner–Wadsworth–Emmons olefination (phosphonate,
DBU, LiCl), or Knoevenagel condensation using piperidine
to obtain 29, 30 and 31, respectively, in excellent yields. In-
terestingly, performing the Knoevenagel condensation with
1,3-cyclohexanedione yielded a double addition product re-
sulting in formation of tetrahydroxantene 32. Aldehyde 28
could also be engaged in the Baylis–Hillman reaction[23]
with acrylates using quinuclidine to afford the addition
product 33 in very good yield. Substitution reactions were
also examined as an entry into privileged heterocyclic
motifs.[24] Starting with chloroacetamide resin 34, the chlor-
ide could be readily displaced by nucleophilic thiols, such as
Figure 3. Left: Plot of the fluorescent intensities measured for the perfect
matched hybridization (PM) and brightest mismatched hybridization
(MM) for serine modified PNA (black) and unmodified PNA (gray);
Right: Images of the array at 0.5 nm for the unmodified PNA (left) and
ser-modified PNA (right). Unmodified PNA sequence: K-GAA CCC
GGT GGA CG-K
ACHTUNGTRENNUNG(Cy3); K-GAA CCC GGT GGA CG-KACHTUNGTERN(NUGN Cy3). Bold
letters denote serine-modified residues.
&
4
&
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!
Chouikhi, Dalila
