Enzymatic and chiral HPLC resolution of α-azido acids and amides

Article abstract of DOI:10.1016/S0957-4166(00)00043-4

For the first time, enzymatic resolution of α-azido acid amides has been successfully demonstrated with high yields and enantiomeric excess. In one case dynamic kinetic resolution was achieved leading to more than 50% yield of the enantiomerically pure azido acid. Chiral HPLC was also used to separate racemic α-azido acids and the separation process was automated. Two routes to enantiopure α-azido acid building blocks for solid-phase peptide synthesis have, therefore, been established. Copyright (C) 2000 Elsevier Science Ltd.

Full text of DOI:10.1016/S0957-4166(00)00043-4

Tetrahedron: Asymmetry 11 (2000) 1239±1248
Enzymatic and chiral HPLC resolution of a-azido
acids and amides
Christian W. Tornùe,a Theo Sonke,b Ilse Maes,b Hans E. Schoemakerb and
Morten Meldala,*
aCarlsberg Laboratory, Department of Chemistry, Gamle Carlsberg Vej 10, DK-2500 Valby, Denmark
bDSM Research BV, Organic Chemistry and Biotechnology Section, PO Box 18, 6160 MD Geleen, The Netherlands
Received 14 January 2000; accepted 26 January 2000
Abstract
For the ®rst time, enzymatic resolution of a-azido acid amides has been successfully demonstrated with
high yields and enantiomeric excess. In one case dynamic kinetic resolution was achieved leading to more
than 50% yield of the enantiomerically pure azido acid. Chiral HPLC was also used to separate racemic
a-azido acids and the separation process was automated. Two routes to enantiopure a-azido acid building
blocks for solid-phase peptide synthesis have, therefore, been established. # 2000 Elsevier Science Ltd. All
rights reserved.
1. Introduction
a-Azido acids are versatile precursors for natural and non-natural amino acids and can be
incorporated into peptides and subsequently reduced to amino acids.1 The azido group o€ers
complete protection of the amine so the carboxyl group can be highly activated as the acid
chloride and reacted with sterically hindered amino groups on solid-phase and in solution.
This method thereby overcomes one of the remaining obstacles in solid-phase peptide synthesis,
the problem of sterical congestion in dicult couplings.2 Azido acids can be prepared from simple
starting compounds such as a-bromo carboxylic acids, incorporated into peptides and reduced on
the solid-phase.1
The pharmaceutical industry has focused attention on peptide mimetics and non-natural amino
acids because they o€er the possibility of varying the side chains of a given peptide and restrict
the conformation, leading to putative agonists or antagonists of increased activity. The a,a-di-
substituted amino acids often stabilise or constrain peptides to b-turns,3 310-helices4,5 or extended
conformations.6,7
* Corresponding author. Tel: +45 3327 5301; fax: +45 3327 4708; e-mail: mpm@crc.dk
0957-4166/00/$ - see front matter # 2000 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(00)00043-4

1240
C. W. Tornùe et al. / Tetrahedron: Asymmetry 11 (2000) 1239±1248
Asymmetric synthesis of a-azido acids can be performed with a chiral auxiliary8 or an asym-
metric phase transfer catalyst, but these approaches are often dicult and time-consuming and
do not always give high enantiomeric excess (e.e.). Another versatile approach is enzymatic
resolution, which often discriminates very well between two enantiomers, giving good to excellent
e.e.s.9 To our knowledge, considerable e€ort has been dedicated to enzymatic syntheses of enant-
iomerically pure a-H-a-amino acids, while far less has been done on a,a-disubstituted amino
acids,10±14 and nothing on mono- and disubstituted a-azido acids. Chiral HPLC may also be used
preparatively to obtain enantiopure compounds and many di€erent chiral stationary phases
(CSPs) are available, e.g. immobilised amino acids, proteins, cyclodextrins, carbohydrates, crown
ethers and ligand-exchange columns. In addition, a new type of CSP has emerged, namely macro-
cyclic glycopeptide antibiotics immobilised on a silica gel column, that has proven ecient in
separating enantiomers.15,16 Two of these antibiotics are commercially available, namely vanco-
mycin and teicoplanin.17
Teicoplanin is produced by Actinoplanes teichomyceticus and has antimicrobial properties
against Gram-positive bacteria. It has four fused macrocyclic rings, seven aromatic rings and
three carbohydrate moieties where one has a long hydrocarbon chain attached.18 All these prop-
erties render teicoplanin diverse as a CSP and it has generally good enantioselectivity for a-sub-
stituted carboxylic acids.
Azido acids have been developed for use in SPPS.1 Because the need for enantiomerically pure
material quickly arose, an investigation into the preparation and puri®cation of enantiopure a-
azido acids was performed. The two approaches, enzymatic resolution and chiral HPLC, are both
very important and they complement each other. The enzymatic resolution can be performed on
a large scale but it takes time to develop, whereas the chiral HPLC is fast but the capacity is
limited.
2. Results and discussion
The following racemic a-azido acid amides (see Scheme 1) have been prepared from their
corresponding acids. Syntheses of the latter will be described elsewhere.19
Scheme 1. The a-azido acid amides 5±8 were prepared by activating 1±4 with SOCl2 or isobutylchloroformate and
reacting them with ammonia in CH2Cl2 or THF
The synthesis of the a-azido acid amides was relatively straightforward, with yields ranging
from 40 to 70%. Quantitative reactions were not achieved, and the prolonged reaction times
(often more than 1 day) caused some decomposition of the activated azido acid 9 before it could
react with ammonia.
In preparative chiral HPLC practice, it is standard procedure to overload the chiral column
when separating quantitative amounts of material and then collect the ®rst and the last parts of

C. W. Tornùe et al. / Tetrahedron: Asymmetry 11 (2000) 1239±1248
1241
the broad, overlapping peaks only and purify the middle fraction again. In this case, however, the
Chirobiotic T column was not overloaded and the two enantiomers could be collected directly in
high enantiomeric excess and the separation process could, therefore, be automated (see Fig. 1).
Depending on the separation of the enantiomers one can automatically inject around 75±150 mg
every 10 min on the column (250Â10 mm i.d.) resulting in separation of around 7 mg overnight,
which is enough material to do reactions on.
Figure 1. Absorption of 4 (215 nm) resolved by repeated injections on a Chirobiotic T chiral HPLC column, eluted
with 15 mM NH4Ac (pH 4):MeOH, 90:10
Data on the separated enantiomers are reported in Table 1. 2-Azidohexanoic acid 2 separated
nicely on the chiral HPLC column and gave the two enantiomers with the same speci®c rotations
but opposite signs. There are no reference values; however, experimental evidence has shown that
teicoplanin binds d-amino acids stronger than l-amino acids because the l-form is eluted ®rst
with no exceptions.15 This relation may, therefore, be used as an indication of absolute con®g-
uration in the case of enantiomeric pairs of azido acids. On this basis, the ®rst eluting enantiomer
of 2 can be assigned to (S)-2-azidohexanoic acid and the last enantiomer to the (R)-con®guration.
For 2-azidophenylacetic acid the speci®c rotations of the enantiomers corresponded well to the
previously determined values8 and the e.e.s were above 99%.
The following parameters were measured for each of the racemic a-azido acids: the retention
(k0), the enantioselectivity (ꢀ) and the enantioresolution (Rs); t1 and t2 are retention times of the
®rst and last eluting enantiomers; t0 is the dead volume; k01 and k02 are the retentions (k0) of the
®rst and second eluting enantiomers, respectively; k0 is typically low on the Chirobiotic T because
it does not retain the compounds very much; w1/2(1) and w1/2(2) are the peak widths at half height
of peaks 1 and 2, respectively; and the resolution (Rs) is a measure of the column's eciency at
separating the enantiomers. Since the enantioselectivity of 3 was low (ꢀ=1.22), separation of the
enantiomers was not attempted.

1242
C. W. Tornùe et al. / Tetrahedron: Asymmetry 11 (2000) 1239±1248
Table 1
Chiral HPLC resolution data on 2, 3 and 4. k0 ˆ ꢀt1 t0†=t0, ꢀ ˆ k0 =k0 , RS=1.18 ꢀt2 t1†=ꢀw
1
‡ w ꢀ2††,
1
2
2ꢀ1†
2
1
a
n.d.=not determined. Measured in CHCl3; baccording to chiral HPLC
In conclusion, the speci®c rotations of 2-azidophenylacetic acid 4 corresponded to the literature
values and for the other azido acid 2 no literature values have been reported but the numerical
speci®c rotations of (+)-2 and ( )-2 are similar and the e.e.s are high. Resolution of 1 on
Chirobiotic T was unsuccessful (data not shown) because the CSP could not discriminate an
a-methyl against an a-ethyl group.
Since the development of DSM's amidase process for the production of enantiomerically pure
amino acids more than a decade ago20 (see Scheme 2), over 100 a-H-a-amino acid amides have
been resolved successfully using Pseudomonas putida ATCC 12633 whole cells. Although the
aminopeptidase from P. putida can accommodate amino acid amides with an extreme diversity in
side chains, much less ¯exibility is allowed with respect to the other groups at the Ca atom: the
P. putida aminopeptidase appeared to be active for a-H-a-amino acid amides only. Because of
this, it was assumed that the activity of this aminopeptidase for a-azido acid amides, if any,
would be very low. Therefore, it was decided to use a genetically modi®ed E. coli strain containing
the P. putida l-aminopeptidase gene (pepA) for the resolution of the a-azido acid amides. Due to
the molecular biology techniques applied, the cellular aminopeptidase activity of these recombi-
nant E. coli cells is at least 25-fold higher than that of the wild type P. putida cells.21
Scheme 2. Enzymatic resolution of an a-H-a-amino acid amide by preferential hydrolysis of one enantiomer
Both 2-azidophenylacetic acid amide 8 and 2-azidohexanoic acid amide 6 were incubated with
whole cells of E. coli DH5a/pTrpLAP in a cell to substrate ratio of 1 to 10. Progress of the reac-
tions was monitored by chiral HPLC analysis of samples taken at regular time intervals. Progress
curves of these two enzymatic resolution reactions are given in Fig. 2.
As the graphs in Fig. 2 show, E. coli DH5a/pTrpLAP did have activity for both azido acid
amides tested. Because neither azido acid amide was a substrate for the negative control (E. coli
DH5a/pUC19), the azido acid amide hydrolysing activity of the pTrpLAP containing recombi-
nant E. coli strain can be solely attributed to the presence of the P. putida gene. Therefore, these

C. W. Tornùe et al. / Tetrahedron: Asymmetry 11 (2000) 1239±1248
1243
Figure 2. Development of conversion (~) and enantiomeric excess (*) of the acid formed in the resolution of 2-azi-
dohexanoic acid amide 6 (A) and 2-azidophenylacetic acid amide 8 (B) by E. coli DH5a/pTrpLAP
azido acid amides are the ®rst non-a-amino groups containing amides that can be hydrolysed
eciently by the l-aminopeptidase from P. putida.
The hydrolysis reaction of the racemic 2-azidohexanoic acid amide stopped completely upon
reaching 50% conversion after about 20 h (Fig. 2A). At that point, the enantiomeric excess of
the (S)-2-azidohexanoic acid was still >99.8%, so the P. putida l-aminopeptidase is completely
l-enantioselective for this substrate. Longer incubation times led to a slight progressive decrease
of the e.e. of the (S)-azido acid to 93% after 75 h. The exact reason for this decrease in e.e. is not
known at the moment, but the action of a racemase with activity for the 2-azidohexanoic acid
could be one of the possible explanations.
The course of the hydrolysis reaction of racemic 2-azidophenylacetic acid amide by E. coli
DH5a/pTrpLAP (Fig. 2B) dramatically di€ers from that of the aforementioned aliphatic azido
acid amide. For this substrate, 50% conversion is already reached after about 3 h, because the
speci®c activity of the P. putida l-aminopeptidase containing E. coli cells is higher for 2-azido-
phenylacetic acid amide than for 2-azidohexanoic acid amide (resp. 1.6 and 0.4 Uy/mg cellular
dry weight). More interestingly, however, this reaction does not come to a complete end upon
reaching 50% conversion, but proceeds with an about 100-fold reduced rate (0.00012 versus 0.012
{
.
moles/kg h ). Because the e.e. of the (S)-2-azidophenylacetic acid remains at over 98% through-
out the whole reaction time, this e€ect can only be explained by racemisation of the remaining
(R)-2-azidophenylacetic acid amide followed by hydrolysis of the (S)-enantiomer by the hetero-
logous l-aminopeptidase (see Scheme 3). In this process, the racemisation is the rate-limiting
step. The reaction of (RS)-2-azidophenylacetic acid amide with E. coli DH5a/pTrpLAP is an
example of a dynamic kinetic resolution, and, therefore, has a theoretical maximum yield of the
(S)-azido acid of 100%.
It was not expected that the 2-azidophenylacetic acid amide would racemise in situ, because
neither the corresponding azido acid nor the a-amino and a-hydroxy analogues, phenylglycine
amide and mandelic acid amide, respectively, racemise spontaneously under the conditions used.
Probably, the racemisation of 2-azidophenylacetic acid amide can be attributed to the presence of
three electron-withdrawing substituents rendering the a-hydrogen atom more acidic.
y
A cell preparation has a speci®c activity of 1 unit (U) per mg of dry weight, if this amount of cells hydrolyses 1 mmol
of substrate per min at standard conditions (pH 9.0 and 40ꢁC).
{
Reaction rate is given in mols per kg and hour of reaction mixture.

1244
C. W. Tornùe et al. / Tetrahedron: Asymmetry 11 (2000) 1239±1248
Scheme 3. Dynamic kinetic resolution of 2-azidophenylacetamide 8 by E. coli DH5a/pTrpLAP
In the near future e€orts will be undertaken to resolve the a-methyl substituted azido acid
amide, (‹)-2-azido-2-methylbutyric acid amide 5. Because the P. putida aminopeptidase does not
have any activity for a-alkyl disubstituted amides, the resolution of 5 will be tried with an ami-
dase from Ochrobactrum anthropi NCIMB 40321,22 which has been puri®ed and characterised.
Furthermore, the gene coding for this amidase has recently been cloned.x
3. Conclusion
Two di€erent approaches to enantiopure a-azido acids have been demonstrated successfully,
which opens up the possibility to synthesise enantiopure peptides or other kinds of pharmaco-
phores which are crucial in the pharmaceutical industry.
Teicoplanin showed very good resolution towards the tested monosubstituted a-azido acids 2±
4, and the process could easily be automated on the mg scale.
Both 2-azidohexanoic acid 2 and 2-azidophenylacetic acid 4 could be obtained in very high
enantiomeric excess from the corresponding amides by an enzymatic resolution concept using an
l-aminopeptidase from P. putida. To get acceptable reaction times, use of a genetically modi®ed
E. coli strain containing the gene for this aminopeptidase was essential. This enzymatic process
will enable the preparation of enantiomerically pure a-azido acids on a multigram scale and with
excellent enantiomerical purity.
4. Experimental
1H and 13C NMR spectra were recorded on a Bruker DRX250 (250 MHz). IR spectra were
recorded on a Perkin±Elmer 1600 FTIR instrument as neat liquids or as KBr-pellets. Optical
rotations were measured on a Perkin±Elmer 241 Polarimeter at 25ꢁC. Melting points were deter-
mined with a Buchi B-540 apparatus and were uncorrected. Separation of enantiomers was per-
formed with a Chirobiotic T 250Â10 mm (i.d.) chiral HPLC column from Astec by elution with a
solvent system of ammonium acetate (15 mM; adjusted to pH 4 with acetic acid) containing 10%
MeOH or n-heptane with 20% abs. EtOH. Detection was at 215 nm with a Waters 490E multi-
1
wavelength detector and a ¯owrate of 1 cm3 min , and this was connected to an autosampler
and an automatic fraction collector. Semipreparative reverse-phase HPLC separation was per-
1
formed on a Shodex RS Pak DS-2013 (20Â300 mm) column with a ¯ow rate of 6 cm3 min .
Detection was at 215 nm on a photodiode array detector (Waters M991). A solvent system con-
sisting of A: 0.1% TFA in water, and B: 0.1% TFA in 90% acetonitrile±10% water was used.
x
Unpublished results.

C. W. Tornùe et al. / Tetrahedron: Asymmetry 11 (2000) 1239±1248
1245
THF was distilled from sodium, CH2Cl2 was distilled from CaH2 and all solvents were stored
over molecular sieves (3 A). Water was Milli-Q (Millipore). Deuterated solvents used for NMR
were from Cambridge Isotope Laboratories and all other solvents were from Labscan.
The following commercially available chemicals were used: N,N-diisopropylethylamine
(Aldrich), isobutylchloroformate (Fluka), N-ethylmorpholine (Merck) and thionyl chloride
(Merck). The synthesis of the a-azido acids 1±4 will be described elsewhere.19
4.1. (‹)-2-Azido-2-methylbutyramide 5
(‹)-2-Azido-2-methylbutyric acid (1.70 g, 11.9 mmol) was converted to the acid chloride with
SOCl2 (4.3 mL, 59.8 mmol, 5 equiv.) in CH2Cl2 (4.3 mL) and re¯uxed under argon for 2.5 h. The
solution was concentrated and diluted with CH2Cl2 (10 mL). DIPEA (12.45 mL, 71.5 mmol, 6
equiv.) was added to a stirred suspension of NH4Cl (2.88 g, 53.9 mmol, 4.5 equiv.) in CH2Cl2
(25 mL) and cooled to 30ꢁC. The crude azido acid chloride was added dropwise to the suspen-
sion under argon, and the temperature raised to 20ꢁC over 2 h. After 72 h, H2O (30 mL) and
aq. HCl (1 M, 30 mL) were added and the phases separated. The aqueous phase was extracted
once with CH2Cl2 (30 mL) and the combined organic phases were concentrated and puri®ed by
VLC (AcOH:EtOAc:heptane, 5:15:80) a€ording 5 (1.53 g, 91%). This was crystallised from
CHCl3/heptane to obtain a very pure sample of 5. M.p. 57.4±58.3ꢁC; 250 MHz H NMR in
1
CDCl3, ꢁ ppm: 0.93 (t, 3H, J=7 Hz, H-4), 1.51 (s, 3H, H-30), 1.71 (dq, 1H, J=7 Hz, J0=14 Hz,
H-3a), 1.96 (dq, 1H, J=7 Hz, J0=14 Hz, H-3b), 6.45 (br s, 2H, NH2); 62.5 MHz 13C NMR in
0
1
CDCl3, ꢁ ppm: 9.0 (C-4), 23.2 (C-3 ), 31.7 (C-3), 68.4 (C-2), 175.2 (C-1). IR: 1666, 2107 cm .
.
Found: C, 43.10%; H, 6.92%; N 39.10%. Calculated for C5H10N4O 1/50 heptane: C, 42.84%; H,
7.22%; N, 38.86%.
4.2. (‹)-2-Azidohexanoic acid amide 6
(‹)-2-Azidohexanoic acid (1.17 g, 7.44 mmol) was dissolved in dry THF (15 mL) and cooled
to 0ꢁC under argon. DIPEA (1.30 mL, 7.46 mmol, 1 equiv.) and isobutylchloroformate (0.975
mL, 7.45 mmol, 1 equiv.) were added to the cooled solution and stirred for 5 min and then the
argon atmosphere was changed to ammonia and the reaction warmed to 20ꢁC over 2 h. H2O (15
mL) was added after 21 h and the aqueous phase was extracted with CH2Cl2 (3Â30 mL) and the
combined organic phases washed with aq. HCl (1 M, 3Â15 mL), dried, concentrated and the
residue (1ꢁ.19 g) puri®ed by VLC (AcOH:EtOAc:heptane, 5:15:80) a€ording 6 (0.48 g, 41%). M.p.
50.7±52.1 C; 250 MHz 1H NMR in CDCl3, ꢁ ppm: 0.91 (t, 3H, J=7 Hz, H-6), 1.38 (m, 4H, H-4,
H-5), 1.85 (m, 2H, H-3), 3.95 (dd, 1H, J=5 Hz, J0=7 Hz, H-2), 6.32 (s, 2H, NH2); 62.5 MHz 13
NMR in CDCl3, ꢁ ppm: 15.5 (C-6), 24.0 (C-5), 29.1 (C-4), 33.5 (C-3), 65.8 (C-2), 174.3 (C-1). IR:
C
1
.
1658, 2114 cm . Found: C, 47.30%; H, 7.77%; N, 33.59%. Calculated for C6H12N4O 1/8
EtOAc: C, 46.69%; H, 7.84%; N, 33.51%.
4.3. (‹)-2-Azidohexadecanoic acid amide 7
(‹)-2-Azidohexadecanoic acid (0.838 g, 2.82 mmol) was dissolved in dry THF (25 mL) and
cooled to 0ꢁC under argon. DIPEA (0.49 mL, 2.81 mmol, 1 equiv.) and isobutylchloroformate
(0.37 mL, 2.83 mmol, 1 equiv.) were added to the cooled solution and stirred for 10 min and then
the argon atmosphere was changed to ammonia and the reaction warmed to 20ꢁC over 2 h. H2O

1246
C. W. Tornùe et al. / Tetrahedron: Asymmetry 11 (2000) 1239±1248
(25 mL) was added after 3 days, the aqueous phase was extracted with CH2Cl2 (3Â10 mL) and
the combined organic phases washed with aq. HCl (1 M, 3Â15 mL), dried and concentrated. The
product was crystallised from DMF/H2O a€ording 7 (0.58 g, 70%). M.p. 80.3±81.5ꢁC; 250 MHz
1H NMR in d6-DMSO, ꢁ ppm: 0.81 (t, 3H, J=7 Hz, H-16), 1.20 (m, 24H, H-4-H-15), 1.62 (m,
2H, H-3), 3.62 (dd, 1H, J=6 Hz, J0=8 Hz, H-2), 7.22 (br s, 1H, NHa2), 7.53 (br s, 1H, NHb2); 62.5
MHz 13C NMR in d6-DMSO, ꢁ ppm: 16.0 (C-16), 24.2 (C-15), 27.4 (C-14), 30.9±33.4 (C-3±C-13),
1
63.4 (C-2), 173.6 (C-1). IR: 1659, 2104 cm . A small amount was further puri®ed on a Shodex
column for elemental analysis. Found: C, 64.89%; H, 11.41%; N, 18.70%. Calculated for
C16H32N4O: C, 64.82%; H, 10.88%; N, 18.90%.
4.4. (‹)-2-Azido-2-phenylacetamide 8
(‹)-2-Azido-2-phenylacetic acid (1.48 g, 8.35 mmol) was dissolved in dry THF (15 mL) and
cooled to 30ꢁC under argon. NEM (1.06 mL, 8.38 mmol, 1 equiv.) and isobutylchloroformate
(1.10 mL, 8.41 mmol, 1 equiv.) were added to the cooled solution and the resulting suspension
was ®ltered after 2 min and the precipitate was washed once with dry THF (6 mL). The ®ltrate
was again cooled to 30ꢁC and the atmosphere was saturated with ammonia (it was also bubbled
through the solution for 30 s). H2O (15 mL) was added after 0.5 h and the aqueous phase was
extracted with CH2Cl2 (3Â15 mL) and the combined organic phases were evaporated. The pro-
duct was crystallised from cold CH2Cl2 ( 78ꢁC) a€ording 8 (0.83 g, 57%). M.p. 113.1 113.7ꢁC;
250 MHz 1H NMR in CDCl3, ꢁ ppm: 5.02 (s, 1H, H-2), 6.43 (br d, 2H, NH2), 7.40 (5H, H-4, H-5,
H-6); 62.5 MHz 13C NMR in CDCl3, ꢁ ppm: 69.0 (C-2), 129.7, 131.1, 131.2 (C-4, C-5, C-6), 136.6
1
(C-3), 172.9 (C-1). IR: 1665, 2104, 3160 cm . Found: C, 54.88%; H, 4.48%; N, 31.72%. Calcu-
lated for C8H8N4O: C, 54.54%; H, 4.58%; N, 31.80%.
4.5. Resolution of (‹)-2-azidohexanoic acid 2 by chiral HPLC
A total of 18 mg of (‹)-2-azidohexanoic acid dissolved in the eluent (3.75 mL, 15 mM NH4Ac
(pH 4):MeOH, 90:10) was resolved on a Chirobiotic T column (250Â10 mm i.d.) using injections
of 150 mg every 10 min. The combined fractions where acidi®ed to pH 2 with aq. HCl (1 M) and
extracted with CH2Cl2 (3Â50 mL), dried and evaporated, giving 4.1 mg of the ®rst eluting enant-
iomer and 3.8 of the second. (S)-2: ‰ꢀŠ25= 34 (c 0.5, CHCl3); (R)-2: ‰ꢀŠ25=+36 (c 0.5, CHCl3).
D
D
4.6. Analytical resolution of (‹)-2-azidohexadecanoic acid 3 by chiral HPLC
Compound 3 (60 mg) was dissolved in DMF and eluted with abs. EtOH:n-heptane, 20:80,
giving the analytical data in Table 1. It was not resolved in quantitative amounts because the
enantioselectivity was too low (ꢀ=1.22).
4.7. Resolution of (‹)-2-azidophenylacetic acid 4 by chiral HPLC
A total of 10 mg of (‹)-2-azidophenylacetic acid dissolved in the eluent (3.5 mL, 15 mM
NH4Ac (pH 4):MeOH, 90:10) was resolved using injections of 75 mg every 10 min. The combined
fractions were acidi®ed to pH 2 with aq. HCl (1 M) and extracted with EtOAc (3Â50 mL), dried
and evaporated giving 2.6 mg of the ®rst eluting enantiomer and 3.0 of the second. (S)-4:
‰ꢀŠ25=+166 (c 0.5, CHCl3); (R)-4: ‰ꢀŠ25= 164 (c 0.25, CHCl3).
D
D

C. W. Tornùe et al. / Tetrahedron: Asymmetry 11 (2000) 1239±1248
1247
4.8. Enzymatic resolution
4.8.1. Bacterial cells
Because it was assumed that the activity of the l-aminopeptidase from P. putida for a-azido
acid amide hydrolysis would be very low, it was decided to perform the resolution experiments
with a hyperactive recombinant E. coli strain, that carried the P. putida l-aminopeptidase gene on
expression plasmid pTrpLAP. The construction of this recombinant strain, as well as its cultiva-
tion, have been described in detail elsewhere.21
4.8.2. Enzymatic resolution reaction
Per reaction, 50 mg of a-azido acid amides 6 and 8 were weighed into a reaction vial, and 2.5
mL of 400 mM TAPS±KOH bu€er, pH 9.0, 1.25 mL of 8 mM MnSO4 and 5 mL of water were
added. After pre-incubation at 40ꢁC, reactions were started by the addition of 1.25 mL of a 4 mg
(wet weight)/mL suspension of E. coli DH5a/pTrpLAP cells. For a negative control, the same
amount of E. coli DH5a/pUC19 cells was used. After certain time intervals, samples of 1 mL were
transferred into 0.5 mL of 1 M H3PO4 to stop the reaction. After passage through a 0.45 mm
®lter, these samples were analysed at 25ꢁC, on a Chirobiotic T (250Â4.6 mm I.D.) chiral HPLC
column from Astec by elution with a solvent system of 15 mM ammonium acetate (adjusted to
pH 4.0 with acetic acid) containing 10% methanol. Detection was at 215 nm and the ¯owrate was
1
0.5 cm3 min .
4.9. Abbreviations
Ahd, (‹)-2-aminohexadecanoic acid; CSP, chiral stationary phase; DIPEA, N,N-diisopropyl-
ethylamine; NEM, N-ethylmorpholine; Nle, norleucine; pepA, Pseudomonas putida l-amino-
peptidase gene; Phg, phenylglycine; SPPS, solid-phase peptide synthesis.
Acknowledgements
This work was supported by the Danish National Research Foundation.
References
1. Meldal, M.; Juliano, M. A.; Jansson, A. M. Tetrahedron Lett. 1997, 38, 2531±2534.
2. Kent, S. B. H. Ann. Rev. Biochem. 1988, 57, 957±989.
3. Crisma, M.; Valle, G.; Toniolo, C.; Prasad, S.; Rao, R. B.; Balaram, P. Biopolymers 1994, 35, 1±9.
4. Marshall, G. R.; Hodgkin, E. E.; Langs, D. A.; Smith, G. D.; Zabrocki, J.; Leplawy, M. T. Proc. Natl. Acad. Sci.
USA 1990, 87, 487±491.
5. Kessler, H.; Konat, R.; Crisma, M.; Valle, G.; Polese, A.; Formaggio, F.; Toniolo, C.; Broxterman, Q. B.;
Kamphuis, J. J. Am. Chem. Soc. 1998, 120, 4763±4770.
6. Barone, V.; Lelj, F.; Bavoso, A.; Di Blasio, B.; Grimaldi, P.; Pavone, V.; Pedone, C. Biopolymers 1985, 24, 1759±
1767.
7. Benedetti, E.; Di Blasio, B.; Pavone, V.; Pedone, C.; Bavoso, A.; Toniolo, C.; Bonora, G. M.; Leplawy, M. T.;
Hardy, P. M. J. Biosci. 1985, 8, 253±262.
8. Evans, D. A.; Britton, T. C.; Ellman, J. A.; Dorow, R. L. J. Am. Chem. Soc. 1990, 112, 4011±4030.

1248
C. W. Tornùe et al. / Tetrahedron: Asymmetry 11 (2000) 1239±1248
9. Kamphuis, J.; Meijer, E. M.; Boesten, W. H. J.; Sonke, T.; van den Tweel, W. J. J.; Schoemaker, H. E. Ann. N. Y.
Acad. Sci. 1992, 672, 510±527.
10. Kruizinga, W. H.; Bolster, J.; Kellogg, R. M.; Kamphuis, J.; Boesten, W. H. J.; Meijer, E. M.; Schoemaker, H. E.
J. Org. Chem. 1988, 53, 1826±1827.
11. Schoemaker, H. E.; Boesten, W. H. J.; Kaptein, B.; Roos, B.; Broxterman, Q. B.; van den Tweel, W. J. J.;
Kamphuis, J. Acta Chem. Scand. 1996, 50, 225±233.
12. Lalonde, J. J.; Bergbreiter, D. E.; Wong, C.-H. J. Org. Chem. 1988, 53, 2323±2327.
13. Kaptein, B.; Monaco, V.; Broxterman, Q. B.; Schoemaker, H. E.; Kamphuis, J. Recl. Trav. Chim. Pays-Bas 1995,
114, 231±238.
14. Kaptein, B.; Boesten, W. H. J.; Broxterman, Q. B.; Peters, P. J. H.; Schoemaker, H. E.; Kamphuis, J. Tetrahedron:
Asymmetry 1993, 4, 1113±1116.
15. Berthod, A.; Liu, Y.; Bagwill, C.; Armstrong, D. W. J. Chromatogr. A 1996, 731, 123±137.
16. Armstrong, D. W.; Liu, Y.; Ekborgott, K. H. Chirality 1995, 7, 474±497.
17. Armstrong, D. W.; Tang, Y.; Chen, S.; Zhou, Y.; Bagwill, C.; Chen, J.-R. Anal. Chem. 1994, 66, 1473±1484.
18. Rundlett, K. L.; Gasper, M. P.; Zhou, E. Y.; Armstrong, D. W. Chirality 1996, 8, 88±107.
19. Tornoe, C. W.; Meldal, M. J. Pep. Sci., in preparation.
20. Kamphuis, J.; Boesten, W. H. J.; Broxterman, Q. B.; Hermes, H. F. M.; van Balken, J. A. M.; Meijer, E. M.;
Schoemaker, H. E. In Advances in Biochemical Engineering/Biotechnology; Fiechter, A., Ed.; Springer-Verlag:
Berlin Heidelberg, 1990; pp. 133±186.
21. Sonke, T.; Kaptein, B.; Boesten, W. H. J.; Broxterman, Q. B.; Schoemaker, H. E.; Kamphuis, J.; Formaggio, F.;
Toniolo, C.; Rutjes, F. P. J. T. In Stereoselective Biocatalysis Handbook; Patel, R. N., Ed.; Marcel Dekker:
New York, 1999.
22. van den Tweel, W. J. J.; van Dooren, T. J. G. M.; de Jonge, P. H.; Kaptein, B.; Duchateu, A. L. L.; Kamphuis, J.
Appl. Microbiol. Biotechnol. 1993, 39, 296±300.