Harnessing functional plasticity of enzymes: A fluorogenic probe for imaging 17β-HSD10 dehydrogenase, an enzyme involved in Alzheimer's and Parkinson's diseases

Article abstract of DOI:10.1021/ja072601x

In this paper, we describe the development of a fluorogenic substrate for 17β-hydroxysteroid-dehydrogenase type 10 (17β-HSD10), which is a multifunctional metabolic enzyme fulfilling several metabolic roles (β-oxidation of fatty acids, catabolism of isoleucine, and metabolism of steroids). In recent years, it has emerged as an important stress and pathological marker in neurons and glial cells (expression down-regulation in Parkinson's disease, up-regulation and association with β-amyloid peptide in Alzheimer's disease). Through the iterative molecular design and chemical synthesis described herein, compound 1 was developed, which possesses all required properties for a selective optical reporter substrate: alcohol-ketone optical switching, the ability to function as a good enzyme substrate (expressed in kinetic parameters), cell permeability, and cell retention. Probe 1 provides a blue-to-green/yellow bright switch and enables non-invasive, real-time imaging of 17β-HSD10 in live human cells. The selectivity of reporter 1 was established by the quantitative correlation of metabolic activity to protein expression in human kidney cell line HEK-293T.

Full text of DOI:10.1021/ja072601x

Published on Web 10/25/2007
Harnessing Functional Plasticity of Enzymes: A Fluorogenic
Probe for Imaging 17â-HSD10 Dehydrogenase, an Enzyme
Involved in Alzheimer’s and Parkinson’s Diseases
Mary K. Froemming and Dalibor Sames*
Contribution from the Department of Chemistry, Columbia UniVersity, 3000 Broadway,
New York, New York 10027
Received April 19, 2007; Revised Manuscript Received July 29, 2007; E-mail: sames@chem.columbia.edu
Abstract: In this paper, we describe the development of a fluorogenic substrate for 17â-hydroxysteroid-
dehydrogenase type 10 (17â-HSD10), which is a multifunctional metabolic enzyme fulfilling several metabolic
roles (â-oxidation of fatty acids, catabolism of isoleucine, and metabolism of steroids). In recent years, it
has emerged as an important stress and pathological marker in neurons and glial cells (expression down-
regulation in Parkinson’s disease, up-regulation and association with â-amyloid peptide in Alzheimer’s
disease). Through the iterative molecular design and chemical synthesis described herein, compound 1
was developed, which possesses all required properties for a selective optical reporter substrate: alcohol-
ketone optical switching, the ability to function as a good enzyme substrate (expressed in kinetic parameters),
cell permeability, and cell retention. Probe 1 provides a blue-to-green/yellow bright switch and enables
non-invasive, real-time imaging of 17â-HSD10 in live human cells. The selectivity of reporter 1 was
established by the quantitative correlation of metabolic activity to protein expression in human kidney cell
line HEK-293T.
short-chain substrates.3 17â-HSD10 is also required for catabo-
Introduction
lism of isoleucine, as it accepts branched hydroxyacyl-CoA
Optical cell imaging has become an indispensable tool for
the life sciences. However, few fluorescent probes are available
for direct read-out of enzyme activity in cells.1 As part of a
broad research program, we are developing optical reporter
substrates for important metabolic and signaling enzymes.2 In
this paper, we report a new fluorogenic probe that enables the
direct activity measurement of 17â-hydroxysteroid dehydroge-
nase type 10 (17â-HSD10) in live human cells, through
fluorescent microscopy. Other names for 17â-HSD10 have also
been used in the literature including endoplasmic reticulum-
associated amyloid â-binding protein (ERAB), amyloid â-pep-
tide-binding alcohol dehydrogenase (ABAD), short chain L-3-
hydroxyacyl-CoA dehydrogenase (SCHAD), hydroxyacyl-CoA
dehydrogenase (HAD), and 2-methyl-3-hydroxybutyryl-CoA
dehydrogenase (MHBD).
substrates.4 Moreover, it has been proposed that 17â-HSD10 is
involved in steroid metabolism (Figure 1).5 Several cases of
inherited 17â-HSD10 deficiency have been documented, and
this condition results in a loss of mental and motor skills.4
Reduced expression of this protein was found in brains of
Parkinson’s patients, while increased expression levels showed
protective effects in mouse models of Parkinson’s disease6 and
brain ischemia.7 Further, 17â-HSD10 binds â-amyloid peptide,
the pathological marker for Alzheimer’s disease, and it was
proposed that the resulting complex potentiates the oxidative
stress and loss of neuronal function indicative of this disease.8
Despite the importance of this emerging physiological and
pathological marker, there are no agents for direct imaging of
17â-HSD10 in live cells and tissues. The functional plasticity
(3) Yang, S.-Y.; He, X.-Y.; Schulz, H. FEBS J. 2005, 272, 4874-83.
(4) Ofman, R.; Ruiter, J. P.; Feenstra, M.; Duran, M.; Poll-The, B. T.; Zschocke,
J.; Ensenauer, R.; Lehnert, W.; Sass, J. O.; Sperl, W.; Wanders, R. J. Am.
J. Hum. Genet. 2003, 72, 1300-7.
(5) (a) Yan, S. D.; Shi, Y.; Zhu, A.; Fu, J.; Zhu, H.; Zhu, Y.; Gibson, L.;
Stern, E.; Collison, K.; Al-Mohanna, F.; Ogawa, S.; Roher, A.; Clarke, S.
G.; Stern, D. M. J. Biol. Chem. 1999, 274, 2145-56. (b) He, X.-Y.; Merz,
G.; Mehta, P.; Schulz, H.; Yang, S. Y. J. Biol. Chem. 1999, 274, 15014-
9. (c) Shafqat, N.; Marschall, H. U.; Filling, C.; Nordling, E.; Wu, X. Q.;
Bjo¨rk, L.; Thyberg, J.; Mårtensson, E.; Salim, S.; Jo¨rnvall, H.; Oppermann,
U. Biochem. J. 2003, 376, 49-60.
(6) Tieu, K.; Perier, C.; Vila, M.; Caspersen, C.; Zhang, H. P.; Teismann, P.;
Jackson-Lewis, V.; Stern, D. M.; Yan, S. D.; Przedborski, S. Ann. Neurol.
2004, 56, 51-60.
(7) Yan, S. D.; Zhu, Y.; Stern, E. D.; Hwang, Y. C.; Hori, O.; Ogawa, S.;
Frosch, M. P.; Connolly, E. S., Jr.; McTaggert, R.; Pinsky, D. J.; Clarke,
S.; Stern, D. M.; Ramasamy, R. J. Biol. Chem. 2000, 275, 27100-9.
(8) Lustbader, J. W.; et al. Science 2004, 304, 448-52.
17â-HSD10 Is an Important Physiological and Pathologi-
cal Marker. 17â-HSD10 is a multifunctional enzyme, fulfilling
several metabolic roles. It is the third enzyme in the â-oxidation
cycle, oxidizing 3-hydroxyacyl-CoAs with a preference for
(1) (a) Lawrence, D. S. Acc. Chem. Res. 2003, 36, 401-9. (b) Baruch, A.;
Jeffery, D. A.; Bogyo, M. Trends Cell Biol. 2004, 14, 29-35. (c) Zhang,
J.; Campbell, R. E.; Ting, A. Y.; Tsien, R. Y. Nat. ReV. Mol. Cell Biol.
2002, 3, 906-18.
(2) (a) Chen, G.; Yee, D. J.; Gubernator, N. G.; Sames, D. J. Am. Chem. Soc.
2005, 127, 4544-5. (b) Froemming, M. K.; Sames, D. Angew. Chem., Int.
Ed. 2006, 45, 637-42. (c) Yee, D. J.; Balsanek, V.; Bauman, D. R.;
Penning, T. M.; Sames, D. Proc. Natl. Acad. Sci. U.S.A. 2006, 103, 13304-
9. (d) Halim, M.; Tremblay, M. S.; Jockusch, S.; Turro, N. J.; Sames, D.
J. Am. Chem. Soc. 2007, 129, 7704-5.
9
14518
J. AM. CHEM. SOC. 2007, 129, 14518-14522
10.1021/ja072601x CCC: $37.00 © 2007 American Chemical Society

Harnessing Functional Plasticity of Enzymes
A R T I C L E S
Figure 1. 17â-HSD10 dehydrognease metabolizes fatty acids, amino acids, and steroids. The enzyme’s inherent substrate plasticity enabled the development
of reporter probe 1 for direct imaging of 17â-HSD10 activity in intact human cells. 17â-HSD10 is defined as 17â-hydroxysteroid dehydrogenase type 10.
alcohol probe needed to be designed that, upon oxidation to
the corresponding ketone, would afford a fluorogenic (increase
in emission intensity) or fluoromorphic (shift in emission
maximum) response (i.e., the off/on or on/on alcohol f ketone
switch). This represented the first key challenge to be addressed,
as ketones usually quench fluorescence due to facile intersystem
crossing to a non-radiative triplet excitation state.9 Our design
was founded on the idea that the quenching effect may be
diminished by engaging the electron-withdrawing ketone in a
strong electronic coupling with a well-placed electron-donating
group, affording a “push-pull” system,9,10 which is the defin-
ing element of many bright organic fluorophores. Specifically,
an electron-donating methoxy group was placed in the 6-position
to enable the electronic resonance with the ketone in the
2-position of naphthalene (Figure 2B). Initially, we synthesized
methyl alcohol 3 and the corresponding ketone 4. The latter
showed a major red-shift in fluorescence; however, signifi-
cant quenching still occurred, resulting in a decrease in
emission intensity (Figure 3A). Despite the low quantum yield
of ketone 4 (0.01, Table 1), its fluorescence is 2 orders of
magnitude greater than that of alcohol 3 at 510 nm (fluoromor-
phic switch), which allows for easy fluorimetric monitoring of
the enzyme-catalyzed oxidation. To our satisfaction, an in Vitro
kinetic assay with the purified enzyme demonstrated that probe
3 was a substrate for 17â-HSD10 (see the next section).
In order to improve the photophysical properties as well as
the kinetic parameters, derivatives 5-10 were synthesized. The
isopropyl alcohol-ketone pair 5/6 has photophysical properties
similar to those of the parent couple 3/4. In contrast, the
fluorescence of cyclic ketone 8 is not quenched; in fact, ketone
8 is much more fluorescent than alcohol 7, with a quantum yield
of 0.71 in aqueous buffer, affording an excellent fluorogenic
switch (Figure 3E). This fortunate outcome may be rationalized
by increased electronic communication between the conforma-
tionally rigid cyclic ketone and the electron-rich naphthalene.
Similarly, cyclohexenone 10 is not quenched in comparison to
cyclohexenol 9, constituting a fluoromorphic couple where the
alcohol and ketone show comparable fluorescence intensities,
with the maxima separated by more than 100 nm (Figure 3G).
Figure 2. (A) Probe design: the 2,6-disubstituted naphthalene core mimics
the size and shape of 17â-estradiol, one of the natural substrates of 17â-
HSD10. (B) Reporter substrate/product pairs for 17â-HSD10 that were
synthesized and examined.
of this enzyme, required for its multiple metabolic functions,
suggested the possibility of designing a fluorogenic reporter
substrate.
Results
De NoWo Design of Optical Probes for Alcohol-to-Ketone
Transformations. Examining the enzyme’s natural substrates,
the preferred strategy was to design a fluorogenic steroid mimic.
A 2,6-disubstituted naphthalene core was chosen to approximate
the size and shape of 17â-estradiol (Figure 2A). As 17â-HSD10
is a dehydrogenase, operating in the oxidation direction, an
(9) Yee, D. J.; Balsanek, V.; Sames, D. J. Am. Chem. Soc. 2004, 126, 2282-3
and references therein.
(10) List, B.; Barbas, C. F., III; Lerner, R. A. Proc. Natl. Acad. Sci. U.S.A.
1998, 95, 15351-15355.
9
J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007 14519

A R T I C L E S
Froemming and Sames
transfer excited state.11 Indeed, charge-transfer fluorophores
display polarity-sensitive emission which is brightest in non-
polar solvents.12
Enzyme Kinetic Parameters for Reporter Probes Ob-
tained with Purified 17â-HSD10. The first reporters we
prepared, namely acyclic alcohols 3 and 5, proved to be good
substrates, albeit with relatively high Km values (3, Km ) 100
µM; 5, Km ) 70 µM; Table 2). We hypothesized that cyclic
alcohols 7 and 9 would be better mimics of 17â-estradiol.
This indeed is the case: the Km value for substrates 7 and 9 is
8-fold lower than that for compound 3. In terms of Km, probes
7 and 9 match the physiological substrate 17â-estradiol (Km )
15 µM); however, the reporter substrates out-perform 17â-
estradiol in terms of turnover rates and catalytic efficiencies
(Table 2). Similarly, probe 1, containing the dimethylamino
group in the 6-position, exhibits excellent kinetic properties.
The systematic exercise synchronizing the desired photo-
physical properties and the substrate kinetic profiles generated
three excellent reporter candidates for the cell imaging
studies.
Direct Imaging of 17â-HSD10 Activity in Cells. The next
key goal was to examine the metabolism of the probes in intact
human cells. To this end, human kidney cells (HEK-293T) were
transiently transfected with a 17â-HSD10 plasmid and incubated
with probes 3, 5, 7, and 9. Probe conversion was examined by
fluorimetry (fluorescence of the media) and fluorescent micros-
copy. Conversion of the probe to product was easily monitored
by fluorimetry for probes 3, 7, and 9. Importantly, the probes
clearly showed faster oxidation with the transfected cells in
comparison to the null-transfected control. However, micro-
scopic imaging was precluded by the low cellular retention of
the probes and the high background fluorescence.
To overcome this problem, the 6-methoxy group in substrate
9 was replaced with a more basic and more polar dimethylamino
group, yielding probe 1. While this alteration had only a minor
effect on the in Vitro kinetic parameters (see above), it led to a
dramatic increase in cell retention, enabling the imaging of probe
1 metabolism via fluorescent microscopy. Furthermore, ketone
2 is not fluorescent in potassium phosphate buffer, whereas it
is bright green-yellow in chloroform (a mimic of cell mem-
branes) and in cells. This optical property proved advantageous
for cell microscopy, eliminating the background fluorescence
of the medium. Transfected HEK-293T cells indeed showed a
faster rate of probe oxidation in comparison to null-transfected
cells (Figure 4). Upon addition of probe 1 to the cellular media,
the oxidation to ketone 2 could be monitored in the 585 nm
filter cube, and the increase in fluorescence was linear for the
first five hours of incubation. The increase in cell fluorescence
corresponds to probe metabolism, which was rigorously con-
firmed by HPLC analysis of medium extracts (see Supporting
Information). Furthermore, the quantitative comparison between
the metabolic rates (formation of product 2) and varying
amounts of 17â-HSD10 DNA transfected into cells revealed a
linear correlation, thus demonstrating that probe 1 reports on
17â-HSD10 activity in human cells (Figure 4).
Figure 3. Fluorescent spectra for probes 1-10 in aqueous buffer and
chloroform. Probes are at a concentration of 50 µM in potassium phosphate
buffer (0.1 M, pH 7.0, 1% DMSO; A, C, E, G, I) or chloroform (B, D, F,
H, J), excited at the excitation maximum for each compound (Table 1). All
spectra are normalized to alcohol 9 in chloroform. NFU is defined as
“normalized fluorescence units”.
It is noteworthy that the emission of the cyclic ketones as well
as the acyclic ketones is completely quenched in chloroform
(Figure 3).
Last, we prepared the pair 1/2, where the 6-position bears a
strongly electron-donating dimethylamino group. While ketone
2 is not fluorescent in potassium phosphate buffer, it gives bright
green-yellow emission in chloroform and in cells, affording an
excellent fluorogenic switch (Figure 3I,J). This photophysical
profile proved highly advantageous for cell microscopy imaging,
as is discussed below. We propose that fluorescence in
chloroform (or cell membranes) originates from the charge-
(11) Bunker, C. E.; Bowen, T. L.; Sun, Y.-P. Photochem. Photobiol. 1993, 58,
499-505.
(12) (a) Stryer, L. J. Mol. Biol. 1965, 13, 482-495. (b) Cory, R. P.; Becker, R.
R.; Rosenbluth, R.; Isenberg, I. J. Am. Chem. Soc. 1968, 90, 1643-
1647.
9
14520 J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007

Harnessing Functional Plasticity of Enzymes
A R T I C L E S
Table 1. Photophysical Properties of 17â-HSD10 Probes
alcohol
ketone
λabs
(nm)
ꢀ
λmax,ex
(nm)
λmax,em
(nm)
λabs
(nm)
ꢀ
λmax,ex
(nm)
λmax,em
(nm)
(M-1 cm-
)
φa
(M-1 cm-
)
φb
1
1
In Bufferc
1
3
5
7
9
305
298
298
298
298
8 900
313
297
298
300
280
444
370
372
370
370
0.44
0.19
0.11
0.15
0.11
2
4
6
8
10
360
334
334
334
334
5 000
350
350
350
350
340
500
510
511
473
500
0.01
0.01
0.01
0.71
0.04
12 000
12 000
7 700
18 000
12 000
23 000
16 000
5 700
In Chloroform
1
3
5
7
9
320
300
300
300
300
18 000
16 000
15 000
16 000
13 000
330
303
303
304
300
370
374
374
370
370
0.01
2
4
6
8
10
390
334
334
334
334
15 000
22 000
17 000
26 000
20.000
385
510
0.70
0.0
0.0
0.0
0.0
0.06
0.05
0.05
0.04
a Relative to 2-aminopyridine in 0.1 N H2SO4 (φ ) 0.6) as a standard (excited at 300 nm). b Relative to quinine bisulfate in 0.1 N H2SO4 (φ ) 0.46) as
a standard (excited at 350 nm). c Potassium phosphate buffer (0.1 M, pH 7.0). Extinction coefficients and quantum yields have an error of (10%.
Figure 4. Metabolism of alcohol 1 in 17â-HSD10 transfected HEK-293T cells monitored by fluorescent microscopy (A-F). Images were taken after
incubating cells with probe 1 for 0 (A,D), 2 (B,E), and 4.5 h (C,F). (A-C) Null-transfected and (D-F) 17â-HSD10 transfected. Dark circles are unstained
nuclei. (G) Correlation of the metabolic rate of probe 1 to the amount of 17â-HSD10 DNA transiently transfected into HEK-293T cells. [Probe 1] ) 20 µM;
rate of probe metabolism was measured as the mean fluorescence intensity of fluorescent images over time and normalized to null-transfected cells. Data
shown are the average ( SD of three independent experiments, as described in the Supporting Information. (H) Western Blot of HEK-293T cells transfected
with 0-0.9µg 17â-HSD10. Expression of 17â-HSD10 and R-tubulin was detected in cell lysates, as described in the Supporting Information.
Conclusions
not suited for fluorescent microscopy studies, they can be used
for fluorimetric cellular assays. Substitution of the 6-methoxy
group for a dimethylamino group afforded alcohol-ketone pair
Using rational design principles, we developed probes 1, 3,
5, 7, and 9 as fluorescent reporter substrates for 17â-HSD10.
Cyclic derivatives 7 and 9 offered key advantages over the
acyclic compounds 3 and 5 in terms of improvements in both
kinetic and fluorescent properties. Although probes 7 and 9 are
(13) Powell, A. J.; Read, J. A.; Banfield, M. J.; Gunn-Moore, F.; Yan, S. D.;
Lustbader, J.; Stern, A. R.; Stern, D. M.; Brady, R. L. J. Mol. Biol. 2000,
303, 311-27.
9
J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007 14521

A R T I C L E S
Froemming and Sames
Table 2. Kinetic Data for 17â-HSD10 Probes
genic probe 1 enables non-invasive and continuous measurement
of 17â-HSD10 activity in cells, through fluorescent microscopy.
This new imaging agent will be used to elucidate the biological
functions of this important physiological marker.
a
Km
SA
kcat
1
1
substrate
(
µM)
(
µ
mol min-1 mg-
)
(min-
)
1
3
5
7
9
15 ( 3
100 ( 11
70 ( 10
13 ( 2b
12 ( 3
0.65 ( 0.07
0.41 ( 0.02
0.41 ( 0.03
0.15 ( 0.01
1.7 ( 0.1
22 ( 3
13.7 ( 0.7
12 ( 1
4.8 ( 0.4
55 ( 5
0.53 ( 0.07
0.53 ( 0.07
Acknowledgment. We thank Dr. Udo Oppermann (Oxford
University, Oxford, UK) and Dr. Shi Du Yan (Columbia
University, New York, NY) for providing purified 17â-HSD10
protein and Dr. Trevor M. Penning (University of Pennsylvania,
Philadelphia, PA) for the pcHSD10 plasmid. M.K.F. is the
recipient of a NSF predoctoral fellowship. This work was
supported by The G. Harold & Leila Y. Mathers Charitable
Foundation.
17â-estradiol
15 ( 7c
15 ( 3d
0.073 ( 0.08e
a Assay was run in 0.1 M potassium phosphate buffer (pH 9), NAD+
(500 µM), substrate, and 17â-HSD10 (0.1-0.8 ng); change in fluorescence
or absorbance was monitored as described in the Supporting Information.
b Substrate inhibition occurs. c Value taken from ref 13. d Determined by
competitive two-substrate assay using probe 3 (see ref 2c). e Determined
by following the formation of NADH at 340 nm.
Supporting Information Available: Synthetic procedures for
compounds 1-10, enzymology assay, cell imaging protocols,
and complete ref 8. This material is available free of charge
via the Internet at http://pubs.acs.org.
1/2, which exhibits excellent properties for fluorescent micros-
copy studies. Ketone 2 is non-fluorescent in water but is
fluorescent in chloroform and cell membranes, which eliminates
background fluorescence for microscopy. As a result, fluoro-
JA072601X
9
14522 J. AM. CHEM. SOC. VOL. 129, NO. 46, 2007