‘Reverse biomimetic’ synthesis ofl-arogenate and its stabilized analogues froml-tyrosine

Article abstract of DOI:10.1039/d1sc03554a

l-Arogenate (also known asl-pretyrosine) is a primary metabolite on a branch of the shikimate biosynthetic pathway to aromatic amino acids. It plays a key role in the synthesis of plant secondary metabolites including alkaloids and the phenylpropanoids that are the key to carbon fixation. Yet understanding the control of arogenate metabolism has been hampered by its extreme instability and the lack of a versatile synthetic route to arogenate and its analogues. We now report a practical synthesis ofl-arogenate in seven steps fromO-benzyll-tyrosine methyl ester in an overall yield of 20%. The synthetic route also delivers the fungal metabolite spiroarogenate, as well as a range of stable saturated and substituted analogues of arogenate. The key step in the synthesis is a carboxylative dearomatization by intramolecular electrophilic capture of tyrosine's phenolic ring using anN-chloroformylimidazolidinone moiety, generating a versatile, functionalizable spirodienone intermediate.

Full text of DOI:10.1039/d1sc03554a

Chemical
Science
View Article Online
View Journal
EDGE ARTICLE
‘Reverse biomimetic’ synthesis of L-arogenate and
its stabilized analogues from L-tyrosine†
Cite this: DOI: 10.1039/d1sc03554a
a
a
All publication charges for this article
have been paid for by the Royal Society
of Chemistry
Louise Eagling,
Daniel J. Leonard,
Maria Schwarz,a Inaki Urruzuno,a
˜
Grace Boden,a J. Steven Wailes,b John W. Ward
and Jonathan Clayden
a
a
*
L-Arogenate (also known as L-pretyrosine) is a primary metabolite on a branch of the shikimate biosynthetic
pathway to aromatic amino acids. It plays a key role in the synthesis of plant secondary metabolites
including alkaloids and the phenylpropanoids that are the key to carbon fixation. Yet understanding the
control of arogenate metabolism has been hampered by its extreme instability and the lack of a versatile
synthetic route to arogenate and its analogues. We now report a practical synthesis of L-arogenate in
seven steps from O-benzyl L-tyrosine methyl ester in an overall yield of 20%. The synthetic route also
delivers the fungal metabolite spiroarogenate, as well as a range of stable saturated and substituted
Received 29th June 2021
Accepted 12th July 2021
analogues of arogenate. The key step in the synthesis is
a carboxylative dearomatization by
DOI: 10.1039/d1sc03554a
intramolecular electrophilic capture of tyrosine's phenolic ring using an N-chloroformylimidazolidinone
rsc.li/chemical-science
moiety, generating a versatile, functionalizable spirodienone intermediate.
aromatized to yield tyrosine and phenylalanine.1–3 Variants of
the pathway exist in which either aromatization or reductive
Introduction
amination occurs rst. The former passes through phenyl-
pyruvate or its 4-hydroxy derivative, while the latter (discovered
by Stenmark in 1974 (ref. 3)) passes through L-arogenate, 1 (also
known as pretyrosine) as the nal non-aromatic intermediate.4
The shikimate pathway is present in plants (where it may
process >30% of photosynthetically xed carbon), bacteria,
fungi, and some other organisms, but is absent from animals,
which must obtain aromatic amino acids from their diet. The
pathway is therefore an attractive target for development of new
herbicides, antimicrobial agents, and live vaccines.5
The aromatic amino acids L-Tyr and L-Phe are biosynthesized
through the shikimate pathway (Scheme 1), in which the non-
aromatic precursor prephenate is reductively aminated and
Two routes to L-arogenate 1 have been reported previously. In
one,6 the key spirocyclic cyclohexadienol intermediate was
formed from a non-biogenetic precursors by Diels–Alder cyclo-
addition. Conversion of the adduct into L-arogenate was
conrmed by NMR, but the nal product was not isolated. In
the other,7 Birch reduction and Michael addition gave a dear-
omatised intermediate that was resolved and converted into
a sample of arogenate. We now report the rst asymmetric
synthesis of an isolated sample of L-arogenate by a direct and
practical route, using arogenate's biosynthetic product L-tyro-
sine as a starting material. We characterise this strategy as
a ‘reverse biomimetic’ synthesis. The synthesis is rapid, effi-
cient, scalable, and highly amenable to the synthesis of more
stable arogenate analogues.
Scheme 1 The biosynthesis of L-Tyr and L-Phe, and (in red) our
‘reverse biomimetic’ synthesis of L-arogenate 1, introducing the
highlighted carboxyl group by an intramolecular dearomatizing
carboxylation.
Our synthetic route employs the dearomatizing cyclization8
of tyrosine's phenolic ring as an atom-economic way to install
the spirocyclohexadienol ring of the target. Dearomatising
cyclisations – whether by nucleophilic, electrophilic, or radical
methods – allow stereochemical complexity to be built from
aSchool of Chemistry, University of Bristol, Cantock's Close, Bristol BS8 1TS, UK.
E-mail: j.clayden@bristol.ac.uk
bJealott's Hill International Research Centre, Bracknell RG42 6EY, UK
† Electronic supplementary information (ESI) available. CCDC 2070018, 2070019,
2070023 and 2070024. For ESI and crystallographic data in CIF or other electronic
format see DOI: 10.1039/d1sc03554a
© 2021 The Author(s). Published by the Royal Society of Chemistry
Chem. Sci.

View Article Online
Chemical Science
Edge Article
a simple planar precursor, and have been used in the syntheses one-carbon unit could be reintroduced exactly as required for
of a number of natural products.9 Dearomatizing spirocyclisa- a possible synthesis of arogenate.
tions of phenols have been widely exploited, yet remarkably few
The trans diastereoisomer of carbamoyl chloride 5 was syn-
of these employ tyrosine, and all of them are oxidative. These thesised selectively from O-benzyl L-tyrosine methyl ester 2 by
examples produced spirolactones,10 spirolactams11 and a spi- formation of the methyl amide 3, condensation with piv-
rodienone carboxylic acid precursor to the natural product alaldehyde, and phosgene-induced cyclization of the imine 4
discorhabdin.12
under kinetic control.14,15 Heating benzyl ether 5 induces cycliza-
tion to an isoquinolinone ring,13 but quite different reactivity is
observed with the free phenol 6 (X-ray crystal structure, conrming
relative stereochemistry, shown in Fig. 1a) that was formed cleanly
by hydrogenolysis of 5 in THF. These conditions preserved the
(surprisingly stable15,16) carbamoyl chloride group.17
We found that heating 6 to 60 ꢀC for 72 h in the presence of
Et3N and KI induced dearomatizing spirocyclisation to the
dienone 7, a compound with clear structural homology to
arogenate 1, in 82% yield. Optimization of this cyclization (see
ESI†) showed that microwave irradiation with Et3N in acetoni-
trile for 10 min at 150 ꢀC gave 7 cleanly in quantitative yield
from 6 without the need for chromatographic purication.
Spirocycle 7 was fully bench stable over periods of months, but
exposure to acid or base (including attempts to hydrolyse its
lactam rings) led to decomposition products arising principally
from rearomatization.
Results and discussion
The biosynthetic decarboxylative aromatization that converts
arogenate to tyrosine – a reaction we sought to run in reverse –
involves the formal elimination of a one-carbon unit (boxed in
blue in Scheme 1) at the formic acid oxidation level. In the
course of an exploration of a Bischler–Napieralski type cycliza-
tion of some carbamoyl chloride derivatives of aromatic amino
acids,13 which typically gives isoquinolinones without dear-
omatisation, we discovered reactivity unique to L-tyrosine. We
found that tyrosine's imidazolidinone derivative 6 (Scheme 2)
underwent a dearomatizing carbonylation that allowed this
This spirocyclization contrasts with related electrophilic
intramolecular acylations of non-phenolic compounds, which
yield isoquinolinones13 without dearomatisation. There are no
other reported acylative dearomatisations of phenols.18
To generate the dienol structure characteristic of arogenate,
the spirocyclic dienone 7 was reduced cleanly under Luche
conditions19 at ꢁ78 C to a separable 3.1 : 1 mixture of diaste-
ꢀ
reoisomeric alcohols 8a and 8b, from which 8a was isolated in
72% yield. The minor diastereoisomer 8b was crystalline, and
an X-ray crystal structure (Fig. 1b) allowed us to assign the
relative congurations of 8a and 8b. In the major product 8a,
formed by approach of the reducing agent anti to the more
hindered carbonyl and tert-butyl group, the hydroxyl group is
orientated syn to the carbonyl group, as is also the case in
arogenate and prephenate.
Scheme 2 Synthesis of L-arogenate 1 by dearomatizing cyclization of
the N-chloroformylimidazolidinone derivative 6 of L-Tyr. Reagents and
conditions: (a) MeNH2, 24 h; (b) t-BuCHO, MgSO4, 16 h; (c) COCl2,
pyridine, THF, rt, 3h; (d) H2, Pd/C, THF, rt, 4 h; (e) Et3N, MeCN, 150 ꢀC
(mw), 10 min; (f) NaBH4, CeCl3$7H2O, MeOH, ꢁ78 ꢀC, 30 min; (g)
Ba(OH)2, H2O, 30 ꢀC, 18 h; (h) Ba(OH)2, BaCO3, H2O, 80 ꢀC, 5 h; (i) Fig. 1 (a) X-ray crystal structure of 6; (b) X-ray crystal structure of
Reported conditions: NaOH, Na2CO3, EtOH, 70 ꢀC, 20 h.6
minor diastereoisomer 8b.21
Chem. Sci.
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Edge Article
Chemical Science
Methods were sought for the conversion of 8a to arogenate 1 problem, with material produced by aqueous hydrolysis of 8a at
by hydrolysis of the imidazolidinone and lactam rings. Heating 80 ꢀC with Ba(OH)2 (3 equiv.) and BaCO3 (1 equiv.)22 giving
with acid led to rearomatization to products that included a sample of the stable disodium salt of arogenate, 1-Na in 41%
dihydroisoquinolone 10 and phenylalanine. Likewise, most yield, calculated by quantitative NMR.
bases promoted rearomatization to mixtures of phenylalanine-
The extreme instability of arogenate towards aromatization
and tyrosine-derived products. Barium hydroxide gave the most prompted us to explore modications to the route that might
encouraging results: treatment with 12 equiv. aqueous Ba(OH)2 provide more stable analogues with potential future utility as
for 18 h at 30 ꢀC opened the imidazolidinone ring to yield probes or modiers of the function of arogenate-metabolising
1
a sample containing >50% by H NMR of the partially hydro- enzymes. Two alternative approaches, both starting from the
lysed lactam 9, which is the natural product spiroarogenate.20
spirodienone 7, are shown in Scheme 3. Hydrogenation of 7
The purication of 9 and its separation from inorganic salts gave the cyclohexanone 12, which was reduced with remarkably
proved problematic. The ion-exchange chromatography reported6,7 high diastereoselectivity (>95 : 5) to alcohol 13a in 93% yield
for the purication of spiroarogenate and arogenate was ineffec- using K-selectride, or less selectively to a separable 47 : 53
tive, as a consequence of the dilution required by the barium salts. mixture of 13a and its crystalline alcohol epimer 13b with
In acidic solution, 9 underwent rearomatizing fragmentation (with NaBH4. The X-ray crystal structure of 13b (Fig. 3a) conrmed the
a half-life measured in hours at pH 5: see ESI†) to phenylalanine, relative stereochemistry of this pair of diastereoisomers. Heat-
ꢀ
preventing the use of standard reversed-phase chromatography. ing 13a with 3 eq. NaOH in H2O at 90 C gave tetrahydrospir-
HPLC using hydrophilic interaction liquid chromatography oarogenate 14 in a yield of 80%. Modication of the hydrolysis
(HILIC) columns was successful at basic pH, but separations were conditions to 12 equiv. Ba(OH)2 and 1 eq. BaCO3 in H2O at 80 ꢀC
poor. Eventually, successful results were obtained using a porous led directly to tetrahydroarogenate 15, which was isolated aer
graphitic carbon stationary phase in the form of a Hypercarb™ reversed-phase column chromatography as its TFA salt in 75%
column, which successfully removed more retained aromatic yield. Importantly, no elimination or other decomposition
impurities giving a 27% isolated yield of air-stable spiroarogenate products were observed from these saturated analogues, giving
9. The 1H NMR spectrum of 9 ([a]D23 ¼ ꢁ40) matched that of spi- them potential value in future biological studies.23
roarogenate isolated from Neurospora crassa (Fig. 2).20
An alternative approach to stabilized arogenate analogues
Danishefsky et al. reported the hydrolysis of 9 to 1 using entailed the addition of an additional alkyl substituent at the 4-
NaOH and Na2CO3,6 and we were able to form traces of position of the dienol ring of 1. A phenethyl moiety was selected
arogenate from 9 by this method. However, purication of 9 with the possibility that analogues of this type may also mimic
using a gradient of 0–80% acetonitrile in 100 mM ammonium NADP in the binding pocket of arogenate dehydrogenase (see
bicarbonate buffer revealed that the conditions that formed ESI†).24 16 and 17 were made (dr > 9 : 1) by 1,2-addition of
spiroarogenate 9 from 8a also formed a trace of arogenate 1,
along with another by-product, the imidazolidinone 11. Raising
the temperature of the hydrolysis of 8a with Ba(OH)2 to 70 ꢀC for
1
5 hours increased the yield of arogenate to 55% (by H NMR).
The free acid of arogenate 1 was even more unstable than spi-
roarogenate 9, decomposing instantly by rearomatization to
phenylalanine. Use of a Na2CO3/NaHCO3 buffer (10 mM at pH
9.2) in place of the ammonium bicarbonate buffer solved this
Scheme 3 Arogenate analogues stabilized towards aromatization.
Reagents and conditions: (a) H2, Pd/C, THF, rt, 16 h; (b) K-selectride
(1.6 equiv.), THF, ꢁ78 ꢀC, 1 h; (c) NaBH4 (3 equiv.), MeOH, 0 ꢀC – rt,
Fig.
2
1H NMR spectrum of spiroarogenate 9 (a) isolated from 18 h; (d) NaOH, H2O, 90 ꢀC, 16 h; (e) Ba(OH)2, BaCO3, H2O, 80 ꢀC, 20 h;
Neurospora crassa (400 MHz in D2O);19 (b) synthesized by the method (f) RCH2CH2MgBr, THF, ꢁ78 ꢀC – rt, 2 h; (g) NaOH, H2O, EtOH, 70 ꢀC,
of Scheme 2 (400 MHz in D2O).
20 h.
© 2021 The Author(s). Published by the Royal Society of Chemistry
Chem. Sci.

View Article Online
Chemical Science
Edge Article
conditions but attempted isolation of the free acid forms of
these products led to instantaneous rearomatization to their
respective phenylalanine derivatives.
Other opportunities to synthesize ring-substituted arogenate
derivatives entailed the spirocyclization of starting materials
carrying substituted phenolic rings (Scheme 4).27 Starting from
the protected form 20 of 3-iodo-L-tyrosine, a Suzuki coupling
using methylboronic acid installed the 3-methyl substituent of
21. Conversion of this compound into the chloro-
formylimidazolidinone 22 followed straightforwardly under the
conditions used in Scheme 2. Spirocyclization of 22 was unaffected
by the substitution on the ring, and heating with Et3N for 10 min at
ꢀ
150 C in the microwave gave a 1 : 1 mixture of the two diaste-
reoisomers of 23. Reduction of these gave four diastereoisomers
24a–d which were separated using standard column chromatog-
raphy and were characterized through a combination of NOE and
X-ray crystallography (the X-ray crystal structure of 24d is shown in
Fig. 3b). Hydrolysis using Ba(OH)2 at 40 ꢀC yielded samples of the
methylated spiroarogenate and arogenate analogues 25 and 26.
The free acid form of 26 showed signicantly greater stability
towards rearomatization than arogenate itself.
Fig. 3 (a) X-ray crystal structure of 13b; (b) X-ray crystal structure of 24d.
Conclusions
In summary, we report a synthesis of the primary metabolite
arogenate 1 in seven steps and an overall yield of 20%, an order of
magnitude improvement on the maximum yield previously pub-
lished.7 We also report the synthesis of the natural product spi-
roarogenate 9. Both syntheses proceed through a synthetically
versatile dienone intermediate 7, amenable to regiodiverse func-
tionalization. Arogenate and spiroarogenate display signicant
instability towards rearomatization under acidic or basic condi-
tions, but by modifying the synthetic route to introduce saturation
or alkyl substituents we were able to make analogues that are
signicantly more stable. Additionally, the ability to convert Tyr
back to arogenate efficiently opens up the possibility of making
isotopically labelled samples of arogenate for use in biosynthetic
studies. Exploration of the potential utility of these analogues and
isotopomers as tools for the investigation and modication of
plant metabolism are under way.
Author contributions
LE, DJL, JWW, and JC conceived the project. JC directed the
work. JSW provided scientic and practical input and advice.
LE, DJL, MS, IU, GB and JWW carried out the experiments. LE,
DJL, JWW and JC wrote the paper.
Scheme 4 Synthesis of 3-methylarogenate 26 from 3-iodotyrosine.
Reagents and conditions: (a) MeB(OH)2, Cs2CO3, SPhosPdG2, dioxane,
90 ꢀC, 16 h; (b) MeNH2, 24 h; (c) t-BuCHO, MgSO4, 16 h; (d) COCl2,
pyridine, THF, RT, 3 h; (e) H2, Pd/C, THF, RT, 4 h; (f) Et3N, MeCN, 150 ꢀC
(mw), 10 min; (g) NaBH4, CeCl3$7H2O, MeOH, ꢁ78 ꢀC, 30 min; (h)
Ba(OH)2, H2O, 40 ꢀC, 16 h.
Conflicts of interest
There are no conicts to declare.
Grignard reagents to the carbonyl group of the dienone 7.25 The
major diastereomers 16a and 17a were subjected to hydrolysis
conditions. Heating 16a and 17a with 6 eq. NaOH in EtOH : H2O
2 : 1 at 70 ꢀC gave the alkylated arogenate derivatives, 18 and 19,
in 51% and 57% yield.26 19 had increased stability in basic
Acknowledgements
This work was funded by the EPSRC through the Bristol
Chemical Synthesis Centre for Doctoral Training (EP/L015366/
1), the Bristol Centre for Doctoral Training in Technology-
Chem. Sci.
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Edge Article
Chemical Science
Enhanced Chemical Synthesis (EP/S024107/1), research grant
EP/L018527, and an Impact Acceleration Award; by Syngenta
through a CASE award (to L. E)., by the ERC through Advanced
Tetrahedron Lett., 2001, 42(20), 3407–3410; (g) J. Clayden,
F. E. Knowles and I. R. Baldwin, J. Am. Chem. Soc., 2005,
127, 2412–2413.
¨
Grant ROCOCO and Proof of Concept Grant QUATERMAIN, and 10 S. Ficht, M. Mulbaier and A. Giannis, Tetrahedron, 2001,
by the Basque Government through a fellowship (to I. U.). We 57(23), 4863–4866.
thank Prof Maxwell Crossley for informative discussions. We 11 N. A. Braun, M. A. Ciufolini, K. Peters and E.-M. Peter,
acknowledge the preliminary investigations carried out by Greg
McHugh and we dedicate this paper to his memory.
Tetrahedron Lett., 1998, 39(26), 4667–4670.
12 H. Tohma, Y. Harayama, M. Hashizume, M. Iwata, M. Egi
and Y. Kita, Angew. Chem., Int. Ed., 2002, 41(2), 348.
13 M. M. Amer, A. C. Carrasco, D. J. Leonard, J. W. Ward and
J. Clayden, Org. Lett., 2018, 20(24), 7977–7981.
Notes and references
1 J. R. Coggins, Herbic. Plant Metab., 1989, 38, 97–112.
2 H. Maeda and N. Dudareva, Annu. Rev. Plant Biol., 2012, 63,
73–105.
3 (a) S. L. Stenmark, D. L. Pierson, R. A. Jensen and G. I. Glover,
Nature, 1974, 247(5439), 290–292; (b) L. O. Zamir,
R. A. Jensen, B. H. Arison, A. W. Douglas, G. Albers-
14 R. Naef and D. Seebach, Helv. Chim. Acta, 1985, 68, 135–143.
15 M. M. Amer, H. Abas, D. J. Leonard, J. W. Ward and
J. Clayden, J. Org. Chem., 2019, 84(11), 7199–7206.
16 H. Abas, M. M. Amer, O. Olaizola and J. Clayden, Org. Lett.,
2019, 21(6), 1908–1911.
17 Under these conditions no Rosenmund reduction to the
formamide was observed.
¨
Schonberg and J. R. Bowen, J. Am. Chem. Soc., 1980, 102,
4499–4504.
18 Very few examples of dearomatizing reactions of phenols
with non-oxidising electrophiles are known: for an
example, see A. Venkatram, T. Colley, J. DeRuiter and
F. Smith, J. Heterocycl. Chem., 2005, 42, 297–301.
19 A. L. Gemal and J.-L. Luche, J. Am. Chem. Soc., 1981, 103(18),
5454–5459.
4 A. M. Fazel, J. R. Bowen and R. A. Jensen, Proc. Natl. Acad. Sci.
U. S. A., 1980, 77(3), 1270–1273.
5 (a) A. D. Baylis, Pest Manage. Sci., 2000, 56(4), 299–308; (b)
P. Rippert and M. Matringe, Eur. J. Biochem., 2002, 269(19),
4753–4761.
6 S. Danishefsky, J. Morris and L. A. Clizbe, J. Am. Chem. Soc.,
1981, 103(6), 1602–1604.
20 L. O. Zamir, R. Tiberio, E. Jung and R. A. Jensen, J. Biol.
Chem., 1983, 258, 6486–6491.
7 M. J. Crossley and R. C. J. Reid, J. Chem. Soc., Chem.
Commun., 1994, 2237.
21 X-ray crystal structures have been deposited with the CCDC:
Deposition numbers 6: 2070018, 8b: 2070019, 13b: 2070023,
24d: 2070024.
22 Previous work (ref. 6) had noted that addition of carbonate
salts accelerated the hydrolysis of 9.
8 (a) J. Clayden and M. N. Kenworthy, Synthesis, 2004, 1721–
1736; (b) H. Brice, J. Clayden and S. D. Hamilton, Beilstein
J. Org. Chem., 2010, 6, 22; (c) J. Clayden, C. J. Menet and
D. J. Manseld, Org. Lett., 2000, 2(26), 4229–4232; (d)
J. Clayden and M. N. Kenworthy, Org. Lett., 2002, 4(5), 787–
790.
23 We observed slow lactonization of 15 on standing in
solution.
24 P. Legrand, R. Dumas, M. Seux, P. Rippert, R. Ravelli,
J.-L. Ferrer and M. Matringe, Structure, 2006, 14, 767–776.
25 The stereochemistry of the major diastereoisomers of 16 and
17 was assigned by analogy to the major diastereoisomer 8a.
26 About 20% of the equivalent spiroarogenate analogue was
also isolated from these hydrolyses, but no aromatized
material was evident in the crude reaction product.
27 Further exploration of possible routes to arogenate
analogues is outlined in the ESI.†
9 (a) S. P. Roche and J. A. Porco, Angew. Chem., Int. Ed., 2011,
´
50(18), 4068–4093; (b) L. Pouysegu, D. Deffieux and
S. Quideau, Tetrahedron, 2010, 66(13), 5908–5917; (c)
`
G. Lemiere, S. Sedehizadeh, J. Toueg, N. Fleary-Roberts and
J. Clayden, Chem. Commun., 2011, 47(13), 3745–3747; (d)
J. Clayden, C. J. Menet and K. Tchabanenko, Tetrahedron,
2002, 58(23), 4727–4733; (e) J. Clayden, M. N. Kenworthy
and M. Helliwell, Org. Lett., 2003, 5(6), 831–834; (f)
A. Ahmed, R. A. Bragg, J. Clayden and K. Tchabanenko,
© 2021 The Author(s). Published by the Royal Society of Chemistry
Chem. Sci.