Substrate specificity of tuliposide-converting enzyme, a unique non-ester-hydrolyzing carboxylesterase in tulip: Effects of the alcohol moiety of substrate on the enzyme activity

Article abstract of DOI:10.1016/j.bmcl.2018.12.010

6-Tuliposides A (PosA) and B (PosB) are glucose esters accumulated in tulip (Tulipa gesneriana) as major defensive secondary metabolites. Pos-converting enzymes (TgTCEs), which we discovered previously from tulip, catalyze the conversion reactions of PosA

Full text of DOI:10.1016/j.bmcl.2018.12.010

Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Substrate specificity of tuliposide-converting enzyme, a unique non-ester-
hydrolyzing carboxylesterase in tulip: Effects of the alcohol moiety of
substrate on the enzyme activity
Yasuo Kato⁎, Takashi Futanaga, Taiji Nomura⁎
Biotechnology Research Center and Department of Biotechnology, Toyama Prefectural University, 5180 Kurokawa, Imizu, Toyama 939-0398, Japan
A R T I C L E I N F O
A B S T R A C T
Keywords:
6-Tuliposides A (PosA) and B (PosB) are glucose esters accumulated in tulip (Tulipa gesneriana) as major de-
fensive secondary metabolites. Pos-converting enzymes (TgTCEs), which we discovered previously from tulip,
catalyze the conversion reactions of PosA and PosB to antimicrobial tulipalins A (PaA) and B (PaB), respectively.
The TgTCEs, belonging to the carboxylesterase family, specifically catalyze intramolecular transesterification,
but not hydrolysis. In this report, we synthesized analogues of Pos with various alcohol moieties, and measured
the TgTCE activity together with a determination of the kinetic parameters for these analogues with a view to
probe the substrate recognition mechanism of the unique non-ester-hydrolyzing TgTCEs. It was found that D-
glucose-like structure and number of the hydroxyl group in alcohol moiety are important for substrate re-
cognition by TgTCEs. Among the analogues examined, 1,2-dideoxy analogues of PosA and PosB were found to be
recognized by the TgTCEs more specifically than the authentic substrates by lowering Km values. The present
results will provide a basis for designing simple, stable synthetic substrate analogues for crystallographic ana-
lysis of TgTCEs.
Carboxylesterase
Structure-activity relationship
Tulipalin
Tuliposide
Tuliposide-converting enzyme
6-Tuliposides (Pos), the major secondary metabolites in tulip (Tulipa
gesneriana), are glucose esters of 4-hydroxy-2-methylenebutanoate and
(3S)-3,4-dihydroxy-2-methylenebutanoate; the former and the latter
are referred to as PosA and PosB, respectively (Fig. 1).1–3 PosA and PosB
serve as precursors of the antimicrobial α-methylene-γ-butyrolactones,
tulipalins A (PaA) and B (PaB), respectively, which are formed from the
hydroxyl acids at the C-6 position of D-glucose (Fig. 1).2,4,5 We pre-
viously discovered a unique Pos-converting enzymes (tuliposide-con-
verting enzymes, TgTCEs) that specifically catalyze the conversion re-
actions of Pos to Pa (Fig. 1).6–14 Two distinct types of TgTCE, PosA-
converting enzyme (TgTCEA) and PosB-converting enzyme (TgTCEB),
are present in tulip tissues, and several isozymes with distinct expres-
sion profiles in tulip tissues have been identified for each of TgTCEA
and TgTCEB (Fig. 1): TgTCEA from bulbs6,9 and petals,8 and TgTCEB
from pollen grains,11 roots,13 and leaves.14 Both types of TgTCE belong
to the plant class I carboxylesterase family in the α/β-hydrolase fold
superfamily.12 Canonical carboxylesterase catalyzes the hydrolysis of a
carboxylic ester to form a carboxylic acid and an alcohol,15 but TgTCEs
catalyze only intramolecular transesterification of Pos to form Pa and D-
glucose in a stoichiometric manner, and never catalyze hydrolysis of
Pos to form hydroxy acids (Fig. 1). Hence, TgTCEs were identified as
unique “non-ester-hydrolyzing carboxylesterases” and classified as
lyases (EC 4.2.99.22 for TgTCEA and EC 4.2.99.23 for TgTCEB), but not
as hydrolases.12
The enzyme reaction by TgTCEs begins with a nucleophilic attack
by the catalytic Ser, whose hydroxyl group is activated by the charge
relay of the catalytic triad, on the carbonyl carbon of Pos.12 This is
followed by the formation of a tetrahedral intermediate, which is sta-
bilized by the oxyanion hole structure formed by the two Gly residues
of the HGG motif. Then, following the elimination of glucose, the acyl-
enzyme complex is formed, and an intramolecular nucleophilic attack
by a terminal hydroxyl group of Pos, but not by water, occurs, and this
nucleophilic attack results in the formation of the five-membered ring
structure of Pa. Although we proposed this reaction mechanism based
on that of the canonical ester-hydrolyzing carboxylesterase and the site-
directed mutagenesis analysis of the enzyme, it has not yet been ver-
ified experimentally by the crystallographic analysis. Moreover, there
has so far been no information regarding the requisite substrate struc-
tures to be recognized by TgTCEs. This prompted us to examine the
structure-activity relationship using structural analogues of Pos. Con-
sidering that natural Pos are chemically labile under neutral to basic
conditions,16 such information is also useful to design more simple,
⁎ Corresponding authors.
E-mail addresses: ykato@pu-toyama.ac.jp (Y. Kato), tnomura@pu-toyama.ac.jp (T. Nomura).
https://doi.org/10.1016/j.bmcl.2018.12.010
Received 14 October 2018; Received in revised form 30 November 2018; Accepted 4 December 2018
0960-894X/©2018ElsevierLtd.Allrightsreserved.
Pleasecitethisarticleas:Kato,Y.,Bioorganic&MedicinalChemistryLetters,https://doi.org/10.1016/j.bmcl.2018.12.010

Y. Kato et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Table 1
Substrate specificities of TgTCEA and TgTCEB for Pos analogues, Acyl-R.
Entry Acyl-
R
Compound Relative activity (%)a,b
TgTCEA
TgTCEB
1
2
A-
PosA
PosB
100
6.3
0.52
100
(S)-B-
3
4
5
6
7
8
A-
A-1
0.3
0.1
1.1
0.2
12
n.d.c
0.8
rac-B-
A-
rac-B-1
A-2
0.1
rac-B-
A-
rac-B-2
A-3
0.6
n.d.
7.7
rac-B-
rac-B-3
n.d.
9
A-
A-4
n.d.
n.d.
Fig. 1. Reactions catalyzed by tuliposide-converting enzymes (TgTCEs).
TgTCEA and TgTCEB catalyze intramolecular transesterification of tuliposides
A (PosA) and B (PosB) to form tulipalins A (PaA) and B (PaB), respectively, but
neither the hydrolysis of Pos nor Pa.
10
rac-B-
rac-B-4
0.08
0.84
11
12
13
A-
A-5
130
2.5
1.9
1.5
18
32
rac-B-
(S)-B-
rac-B-5
(S)-B-5
14
15
16
A-
A-6
190
3.2
3.7
1.4
7.8
25
rac-B-
(S)-B-
rac-B-6
(S)-B-6
17
18
A-
A-7
4.7
0.04
1.7
rac-B-
rac-B-7
0.08
19
20
21
A-
A-8
105
2.8
2.1
0.50
28
rac-B-
(S)-B-
rac-B-8
(S)-B-8
32
a
Specific activities of TgTCEA and TgTCEB for their authentic substrates,
PosA and PosB (at 4 mM) were set to 100%, respectively.
b
Ring-opened PaA-acid (21) and PaB-acid (rac-22) were not detected as
products for any substrates
c
n.d., not detected.
Acyl unit (Acyl-)
Alcohol unit (R-)
hydroxy acids (PaA- and PaB-acids; prepared as described in Fig. S3),
which are the hydrolytic products of substrates, was observed in the
reaction mixture.
Fig. 2. Combinations of acyl- and alcohol units for designing Pos analogues,
Acyl-R, used in this study.
We then looked into simplifying the alcohol moiety of the substrates,
by mimicking a glucose structure. As the simplest molecules, PosA
analogues with cyclohexylmethyl and tetrahydropyranylmethyl struc-
tures in an alcohol unit, A-3 and A-4, respectively, and the synonymous
analogues for rac-PosB, rac-B-3 and rac-B-4, respectively, were syn-
thesized as described in Figs. S4–S6, and subjected to the enzyme assay.
As is the case in methyl esters and NAC thioesters, these simple ana-
logues were not the preferable substrates for either TgTCE (Table 1,
entries 7–10), suggesting that the existence of hydroxyl group(s) on the
ring moiety is requisite for substrate recognition by TgTCEs. Firstly, in
order to examine the effect of a hydroxyl group at the C-1 position of
glucose, α-1-O-Me analogues of PosA and rac-PosB, A-5 and rac-B-5,
and their β-1-O-Me analogues, A-6 and rac-B-6, were synthesized as
described in Figs. S7–S9, and subjected to the enzyme assay. As shown
in Table 1 (entries 11, 12, 14, and 15), protection of the C-1 hydroxyl
group enabled TgTCEA to better recognize A-type substrates, whereas
the opposite effect was observed for TgTCEB, indicating that TgTCEA
and TgTCEB recognize the hydroxyl group at the C-1 position of glucose
in a distinct manner. We next designed substrate analogues by se-
quentially adding hydroxyl groups to the tetrahydropyranylmethyl
moiety; starting from the analogues having the C-4 hydroxyl group, we
synthesized 1,2,3-trideoxy analogues of PosA and rac-PosB, A-7 and
rac-B-7 (Figs. S10, S11), respectively, and then synthesized 1,2-dideoxy
analogues, A-8 and rac-B-8 (Figs. S12, S13), respectively. With the re-
sults obtained by the enzyme assay (Table 1, entries 17–20), we
stable substrate mimics that can be applied to co-crystallization ex-
periments with TgTCEs, which lead to the verification of the reaction
mechanism, including the formation of tetrahedral intermediate and
the intramolecular nucleophilic attack. We hereby focused on the
structural requirements of the alcohol moiety of Pos for recognition by
TgTCEs. We synthesized many analogues of PosA and PosB by combi-
nations of acyl units (for PosA, racemic (rac)-PosB, and PosB) with
various alcohol units (Fig. 2), and assessed their effects on the enzyme
activities of TgTCEA and TgTCEB. Moreover, stabilities of the synthetic
analogues in aqueous solution were examined and compared with those
of natural substrates, PosA and PosB.
TgTCEA from petals (TgTCEA1)8 and TgTCEB from pollen grains
(TgTCEB1)11 were used throughout the study. The recombinant en-
zymes were expressed in Escherichia coli, and purified as the His-tag-free
forms as described previously.9,11 Enzyme reactions were performed at
4 mM substrates under the same conditions as described previously,8
and the enzyme activity was calculated based on the amount of PaA or
PaB formed by the reaction. We first synthesized the simplest methyl
ester analogues of PosA and rac-PosB, A-1 and rac-B-1, respectively,
and their N-acetylcysteamine (NAC) thioester analogues, A-2 and rac-
B-2, respectively, according to the methods described in Figs. S1 and
S2. As shown in Table 1 (entries 3–6), all these analogues were poor
substrates for TgTCEs, suggesting that a structure similar to glucose
moiety is crucial for substrate recognition by TgTCEs. No formation of
2

Y. Kato et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Table 2
Kinetic parameters of TgTCEA for Pos analogues, Acyl-R.
Entry
Acyl-
R
Compound
kcat
Km
kcat/Km
(s−1
)
(mM)
(s−1/mM)
1
2
A-
PosA
PosB
2,760
456
77.2
6.06
9.14
41.8
0.649
1.71
302
(S)-B-
10.9
3
4
A-
A-5
1,910
45.3
103
5.27
10.9
0.819
1.84
362
(S)-B-
(S)-B-5
3.59
4.16
5
6
A-
A-6
3,180
110
110
5.79
15.3
0.618
2.16
549
(S)-B-
(S)-B-6
7.61
7.19
7
8
A-
A-8
1,580
96.1
31.9
3.88
5.26
25.7
0.339
2.28
300
(S)-B-
(S)-B-8
3.74
Data are means
s.e. (s.e.: fitting error to the Michaelis-Menten equation as calculated by a nonlinear fitting program).
identified 1,2-dideoxy analogues as acceptable substrates for both
TgTCEs, especially for TgTCEA, while 1,2,3-trideoxy analogues were
not recognized as good substrates by either TgTCE. Of the 16 synthetic
analogues, compounds 5, 6, and 8 (A- and rac-B- for each) served as
substantial substrates. We thus synthesized chiral ((S)-form) acyl unit
corresponding to natural PosB as described in Fig. S14, and the acid was
condensed with α- or β-1-O-Me, or with 1,2-dideoxy type alcohol units,
to produce chiral PosB analogues ((S)-B-5, (S)-B-6, and (S)-B-8) as
described in Fig. S15. Using these analogues, we measured the enzyme
activity (Table 1, entries 13, 16, and 21) and found that stereochemistry
at acyl side chain does not greatly affect the enzyme activity except for
TgTCEB activity toward (S)-B-6, which was 3-fold higher than that
toward rac-B-6.
was 27.7, and the ratio increased approximately three times by chan-
ging their alcohol moiety; the values for A-5/(S)-B-5, A-6/(S)-B-6, and
A-8/(S)-B-8 were 87.0, 76.4, and 80.2, respectively. These results
suggest that the deletion or protection of the C-1 hydroxyl group of
substrate improves TgTCEA discrimination between A- and B-type
substrates. This is mainly due to the decrease in the kcat values for the B-
type analogues to approximately one-tenth to one-fourth of PosB while
those for A-type substrates were comparable. On the other hand, the
opposite phenomenon was observed for TgTCEB (Table 3). The ratio for
PosB/PosA of TgTCEB was 238, and the ratio decreased by changing
their alcohol moiety to α- or β-1-O-Me forms; the values for (S)-B-5/A-5
and (S)-B-6/A-6 were 59.8 and 62.8, respectively. These changes are
due to the remarkable decrease in the kcat values, without affecting the
Km values, for B-type analogues by the protection of the C-1 hydroxyl
group. Interestingly, the value for (S)-B-8/A-8 of TgTCEB was 186,
which was comparable with that for PosB/PosA; Km value for (S)-B-8
was approximately one-third of that for PosB, while the kcat value for
PosB was five times higher than that for (S)-B-8. Very recently, during
the course of studies on an action of TgTCEs against naturally-occurring
Pos derivatives, we found that the presence of a 1-acyl group increases
the affinity of the enzymes for substrates, rather than interferes with
We next determined the kinetic parameters of TgTCEA and TgTCEB
with A-type analogues (A-5, A-6, and A-8) and chiral B-type analogues
((S)-B-5, (S)-B-6, and (S)-B-8), and compared with those for authentic
substrates, PosA and PosB. The kinetic parameters of TgTCEA and
TgTCEB are listed in Tables 2 and 3, respectively (see Figs. S16 and S17,
respectively, for Michaelis-Menten plots). Catalytic efficiency (kcat/Km)
of TgTCEA did not differ greatly between the authentic substrate PosA
and its analogues (Table 2). The ratio of kcat/Km values for PosA/PosB
Table 3
Kinetic parameters of TgTCEB for Pos analogues, Acyl-R.
Entry
Acyl-
R
Compound
kcat
Km
kcat/Km
(s−1
)
(mM)
(s−1/mM)
1
2
A-
PosA
PosB
4.27
0.119
16.5
3.95
4.36
0.267
0.222
1.08
257
(S)-B-
1,120
3
4
A-
A-5
11.9
370
0.476
6.89
5.98
3.12
0.492
0.133
1.99
119
(S)-B-
(S)-B-5
5
6
A-
A-6
18.7
310
0.420
25.9
18.6
4.89
0.789
0.906
1.01
63.4
(S)-B-
(S)-B-6
7
8
A-
A-8
27.8
205
1.66
7.46
42.8
1.70
5.21
0.650
121
(S)-B-
(S)-B-8
0.178
Data are means
s.e. (s.e.: fitting error to the Michaelis-Menten equation as calculated by a nonlinear fitting program).
3

Y. Kato et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
substrate binding; the Km values of TgTCEA and TgTCEB for PosD and
PosF, which are naturally-occurring 1,6-diacyl type analogues of PosA
and PosB, respectively, were remarkably lower than those for PosA and
PosB, respectively.16 Due to mutarotation of the glucose moiety, 1α-
and 1β-anomers exist for PosA and PosB, and both anomers serve as
substrates.12 Therefore, preferences of TgTCEA and TgTCEB for PosA
and PosB, respectively, had been considered to be primarily dependent
on the structure of the 6-acyl groups, but not on the configuration of the
C-1 position. However, given the results obtained here, not only the
structure of the 6-acyl group, but also the structure and the config-
uration of the C-1 position of substrate appear to be involved in the
preferences of TgTCEA and TgTCEB for PosA and PosB, respectively.
Since the 1,2-dideoxy analogues, A-8 and (S)-B-8, served as sub-
strates for TgTCEA and TgTCEB, respectively, with the higher affinity
(lower Km) and lower turnover number (lower kcat) than their authentic
substrates, PosA and PosB (such trend was more evident on (S)-B-8
analogue for TgTCEB than on A-8 analog for TgTCEA), an intriguing
question was raised whether A-8 and (S)-B-8 analogues behave as ef-
ficient inhibitors of TgTCEs. We then measured the TgTCE activity (at
4 mM PosA or PosB) in the presence of dideoxy analogues (at 4 or
10 mM), but no remarkable decrease in the enzyme activity was ob-
served; the rate of Pa formation was equal to the sum of those from Pos
and the analogue (data not shown). Currently, the synthesis of potent
Pos analogue type inhibitors of TgTCEs, e.g., phosphonate esters, ke-
tomethylene or dehydromethylene isosters17 with a 1,2-dideoxy-type
alcohol moiety, which are expected to bind more efficiently to TgTCEs,
is underway in our laboratory.
TgTCEB with respect to the alcohol moiety; the changes in the kcat and
Km values of TgTCEA toward A-type analogues relative to those toward
PosA were moderate, whereas kcat values of TgTCEB toward B-type
analogues and Km value toward (S)-B-8 markedly decreased relative to
those toward PosB. To clarify whether such features of each of TgTCEA
and TgTCEB are shared between the isozymes (TgTCEA isozymes from
petals [used in this study]8 and bulbs,6,9; TgTCEB isozymes from pollen
grains [used in this study],11 roots13 and leaves14), comparative kinetic
analysis toward the various substrate analogues synthesized in this
study should be performed in due course. Such analyses of substrate
recognition will give clues not only about the functional specialization
of TgTCEA and TgTCEB, which diverged from a common ancestral
enzyme,12 but also about the differentiation of unique non-ester-hy-
drolyzing carboxylesterase from a canonical ester-hydrolyzing carbox-
ylesterase.
In the present study, we found some structural features of the al-
cohol moiety of Pos that are requisite for recognition by TgTCEs, which
also opened the possibility of making competitive inhibitor of TgTCEs
that will be used to obtain the ligand-bound form of the crystal struc-
ture of TgTCEs. However, with a view to probing the substrate re-
cognition mechanism by TgTCEs, structure-activity relationship studies
regarding the acyl unit of Pos are still missing. Thus, the synthesis of a
variety of Pos analogues having distinct acyl moiety with respect to, for
example, chain length, presence or absence/position/number of double
bond, carbonyl group and hydroxyl group, is now in progress in our
laboratory.
Stability of the synthetic Pos analogues under the enzyme reaction
conditions (in 50 mM potassium phosphate buffer, pH 6.5) was assessed
by determining their half-lives (t1/2), as shown in Fig. S18. The t1/2
values for PosA, A-5, A-6, and A-8 were 42, 66, 41, and 52 h, respec-
tively, and those for PosB, (S)-B-5, (S)-B-6, and (S)-B-8 were 6.3, 9.7,
6.1, and 7.6 h, respectively. The results demonstrate that the com-
pounds having the PosA-type 6-acyl group are more stable than those
having the PosB-type 6-acyl group, but no marked improvement of the
stability was observed by changing alcohol units. Interestingly, the
ratios of t1/2 values compared to the natural substrates were almost the
same between the A-type and B-type analogues having the same alcohol
units. That is, the t1/2 ratios for A-5/PosA, A-6/PosA, and A-8/PosA
were 1.6, 0.98, and 1.2, respectively, and those for (S)-B-5/PosB, (S)-B-
6/PosB, and (S)-B-8/PosB were 1.5, 0.97, and 1.2, respectively. These
results indicate that the stabilities of Pos and their analogues depend
not only on the structure of acyl unit, but also on the structure of al-
cohol unit.
Acknowledgements
This work was supported by JSPS KAKENHI grant nos. JP26850072
and JP18K05463 (to TN), and JP16K07697 (to YK).
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://
doi.org/10.1016/j.bmcl.2018.12.010.
References
1. Tschesche R, Kämmerer F-J, Wulff G, Schönbeck F. Tetrahedron Lett. 1968;9:701.
2. Tschesche R, Kämmerer F-J, Wulff G. Chem Ber. 1969;102:2057.
3. Christensen LP, Kristiansen K. Contact Dermat. 1999;40:300.
4. Beijersbergen JCM, Lemmers CBG. Physiol Plant Pathol. 1972;2:265.
5. van Rossum MWPC, Alberda M, van der Plas LHW. Phytochemistry. 1998;49:723.
6. Kato Y, Shoji K, Ubukata M, et al. Biosci Biotechnol Biochem. 2009;73:1895.
7. Kato Y, Yoshida H, Shoji K, Sato Y, Nakajima N, Ogita S. Tetrahedron Lett.
2009;50:4751.
TgTCE belongs to the carboxylesterase family in the α/β-hydrolase
fold superfamily, and specifically catalyzes intramolecular transester-
ification, but not hydrolysis.12 This non-ester-hydrolyzing carbox-
ylesterase is an example of an enzyme with catalytic properties that are
unpredictable from its primary structure.12 The fact that TgTCEA and
TgTCEB did not form the hydrolytic product PaA- and PaB-acids from
any Pos analogues supports the reaction mechanism we proposed pre-
viously: TgTCE has rigid mechanism for an intramolecular nucleophilic
attack by a terminal hydroxyl group of the acyl unit of Pos, but not by
water, to the acyl-enzyme complex, resulting in the formation of the
five-membered ring of Pa.12 Based on the kinetic analysis, it was sug-
gested that TgTCEA has more flexible substrate recognition than
8. Nomura T, Ogita S, Kato Y. Plant Physiol. 2012;159:565.
9. Nomura T, Tsuchigami A, Ogita S, Kato Y. Biosci Biotechnol Biochem. 2013;77:1042.
10. Nomura T, Hayashi E, Kawakami S, Ogita S, Kato Y. Biosci Biotechnol Biochem.
2015;79:25.
11. Nomura T, Murase T, Ogita S, Kato Y. Plant J. 2015;83:252.
12. Nomura T. Biosci Biotechnol Biochem. 2017;81:81.
13. Nomura T, Ueno A, Ogita S, Kato Y. Biosci Biotechnol Biochem. 2017;81:1185.
14. Nomura T, Kuchida R, Kitaoka N, Kato Y. Biosci Biotechnol Biochem. 2018;82:810.
15. Gershater MC, Edwards R. Plant Sci. 2007;173:579.
16. Nomura T, Ogita S, Kato Y. Appl Biochem Biotechnol. 2018. https://doi.org/10.1007/
s12010-018-2903-3 in press.
17. Collinsová M, Jirácek J. Curr Med Chem. 2000;7:629.
4