Fluoromorphic substrates for fatty acid metabolism: Highly sensitive probes for mammalian medium-chain acyl-CoA dehydrogenase

Article abstract of DOI:10.1002/anie.200502675

(Chemical Equation Presented) Not fussy: A systematic study revealed considerable substrate tolerance of medium-chain acyl-CoA dehydrogenase (MCAD), a member of the β-oxidation pathway for the degradation of fatty acids. Three fluorogenic reporter substra

Full text of DOI:10.1002/anie.200502675

Angewandte
Chemie
Enzyme Promiscuity
DOI: 10.1002/anie.200502675
Fluoromorphic Substrates for Fatty Acid
Metabolism: Highly Sensitive Probes for
Mammalian Medium-Chain Acyl-CoA
Dehydrogenase**
Mary K. Froemming and Dalibor Sames*
In contrast to the accepted view in biological sciences that
enzymes are highly selective for their physiological substrates,
substrate promiscuity is a widespread phenomenon.[1] Such
functional flexibility may often be observed with enzymes
involved in the metabolism of an array of substrates, and
examples extend beyond the well-known xenobiotic metab-
olism.
The exploration of enzyme substrate flexibility for the
purpose of the development of metabolic probes and imaging
agents represents a guiding concept for a broad program in
our laboratories.[2,3] Herein we report the examination of the
substrate fidelity of medium-chain acyl-CoA dehydrogenase
(MCAD), a key enzyme in the metabolism of fatty acids,
which led to the development of fluorogenic and fluoromor-
phic probes for this enzyme.[4] These indicators allow for
selective and sensitive detection of MCAD activity in tissue
homogenate.
b-Oxidation of fatty acids represents one of the central
metabolic pathways.[5–11] Close examination of this process
reveals considerable built-in substrate flexibility. Catabolism
of a long-chain fatty acid occurs by repetitive removal of two-
carbon units (acetyl-CoA), (Figure 1). Each catalytic turn
involves four chemical steps that are catalyzed by enzymes
capable of accommodating substrates with a variety of chain
lengths. Thus, instead of having a specific enzyme for each
intermediate (C16, C14, C12!C2), this pathway consists of a few
enzymes that show broad and overlapping chain-length
specificities.
The overall mechanism for this pathway was proposed on
the basis of classical studies conducted by F. Knoop[12] who
showed that w-phenyl fatty acids were metabolized in dogs.
[*] M. K. Froemming, Prof. D. Sames
Department of Chemistry
Columbia University
3000 Broadway, New York, NY 10027 (USA)
Fax: (+1)212-932-1289
E-mail: sames@chem.columbia.edu
[**] The authors thank Professor Horst Schulz, Wenfeng Yu, and Lina
Nie (City University of New York) for assistance with rat homoge-
nates and supplying the rMCAD enzyme; Professor Jung-Ja Kim
(Medical College of Wisconsin) for supplying the pSCAD, pMCAD,
and hLCAD enzymes; and Dominic Yee for intellectual contribu-
tions. This work was generously supported by The G. Harold & Leila
Y. Mathers Charitable Foundation. M.K.F. is the recipient of a NSF
Predoctoral Fellowship.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2006, 45, 637 –642
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
637

Communications
To develop fluorogenic MCAD enzyme reporters, two key
criteria must be satisfied: 1) dehydrogenation is coupled
either to an increase in brightness (fluorogenic) or preferably
to a significant shift in emission wavelength (fluoromorphic),
and 2) these compounds serve as substrates for MCAD.
Our design was based on a general motif that consists of
an organic fluorophore attached to the b-carbon of propionyl-
CoA (Figure 3). Dehydrogenation provides the correspond-
Figure 3. Design of a MCAD fluorescent reporter. Dehydrogenation
extends the p-conjugation of the fluorophore, which results in a
change in the emission profile. (Fl=organic fluorophore).
Figure 1. The b-oxidation cycle. The four enzyme families involved in b-oxidation
are: acyl-CoA dehydrogenase (ACAD), enoyl-CoA hydratase (ECH), l-3-hydroxy-
acyl-CoA dehydrogenase (HCAD), and thiolase.
The ingenious use of the phenyl ring as a chemical label
suggested a two-carbon degradation pathway but also, from
the perspective of this paper, suggested the permissiveness of
the entire pathway.[13]
ing unsaturated system with an extended p-conjugation that
connects the fluorophore and the thioester group. We
anticipated that such dramatic expansion of the p-conjugated
system would strongly affect the emission profile, most likely
shifting the emission to longer wavelengths.
The b-oxidation spiral begins, after the activation (acyl-
CoA synthesis) and transport of fatty acids into mitochondria,
with the a,b-dehydrogenation of acyl-CoA catalyzed by acyl
coenzyme A dehydrogenases (ACADs). This flavin-depen-
dent family consists of nine isozymes, five of which are
involved in fatty acid metabolism.[10] Medium-chain acyl-CoA
dehydrogenase (MCAD), formerly known as the ꢀgeneralꢁ
ACAD owing to its ability to act on a wide range of chain
lengths (C16–C4, Figure 2), represents an attractive target for
the development of reporter substrates.
It has been shown that Knoopꢁs substrate, 3-phenylpro-
pionyl-CoA, and related compounds are substrates for
MCAD; however, neither the substrates nor the products
are fluorescent.[21–23] Consequently, we began with naphtha-
lene-based systems, namely compound 1, the dimethylamino-
substituted analogue 2 (Figure 4),[24] and their corresponding
products 3 and 4, respectively. The photophysical measure-
ments revealed that both pairs, 1/3 and 2/4, constituted
excellent fluoromorphic switches. Between compound 1 and
3, the emission maximum is red shifted by approximately
50 nm with some overlap between the two curves. A threefold
change in the emission intensity could be measured at the
emission maximum of each compound, therefore meeting the
requirements of a ratiometric probe. A comparison of 2 and 4
revealed a 190-nm shift of the emission and a 100-fold
increase in the fluorescence intensity at the emission max-
imum of compound 4 (Figure 4).[25]
As the emission switch criterion was satisfied, we next
addressed the issue of whether compounds 1 and 2 were
substrates for MCAD. The activity was tested with purified
rat-liver MCAD (rMCAD) by monitoring the reduction of an
external electron acceptor (for assay conditions, see Support-
ing Information). We were gratified to find that compound 1
was a good substrate, however, 2 was completely inert. We
hypothesized that the high polarity of the dimethyamino
group, not necessarily its size, disfavored the interaction with
the hydrophobic pocket of the enzyme. Instructed by these
results, we synthesized compounds 5–10 to systematically
explore two questions: 1) the limits of this enzyme in terms of
the size, shape, and polarity of the substrate, and 2) the effect
of substituents and other structural variations on emission
properties.
Figure 2. Substrate flexibility of medium-chain acyl-CoA dehydrogenase
(MCAD) and long-chain acyl-CoA dehydrogenase (LCAD). These
enzymes show broad and overlapping chain-length specificities. Acyl-
CoAs with eight (C8) to 18 (C18) carbon chains are substrates for
LCAD, whereas those with four (C4) to 16 (C16) are substrates for
MCAD. SCAD shows low flexibility with butanoyl-CoA (C4) being the
main substrate.
MCAD is also an important target from the medical
perspective. MCAD deficiency has recently emerged as a
common hereditary disease with severe symptoms.[14–17]
Furthermore, growing evidence implicates impaired b-oxida-
tion as an important contributor to the development of a
cluster of diseases such as hypertension, insulin resistance,
and dyslipidemia, which is referred to as ꢀmetabolic syn-
dromeꢁ.[18–20] Consequently, there is a need for a direct,
sensitive, and continuous readout of MCAD enzyme activity
in tissue samples and cell lines. Inspired by this challenge, we
aimed to develop a fluorometric assay that is based on
fluorogenic probes.
Of the series shown in Figure 5, only probes 7 and 9 were
substrates for MCAD. The adverse effect of polar substituents
was confirmed; compounds 5 and 6 were found to be inert,
whereas the conversion of the hydroxy group in 5 into a
638
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 637 –642

Angewandte
Chemie
with structural studies that revealed a narrow active site
of MCAD.[10] Interestingly, the attachment of an addi-
tional ring to form anthracene was tolerated, demon-
strating significant recognition-site flexibility as long as
the substrate can assume a narrow shape. The pyrene
system 10 was presumably too large and as a result,
showed no conversion by the enzyme (Figure 5).
Thus, besides mapping the permissiveness of MCAD,
this exercise also identified two new substrates with
favorable emission properties (Figure 6). Dehydrogen-
ation of the weakly fluorescent probe 7 led to the highly
fluorescent product 11, which resulted in an 80-fold
increase in the emission intensity at 510 nm (fluorogenic
probe), whereas conversion of 9 into 12 gave a 150-nm
shift in the emission wavelength and afforded an 80-fold
increase in emission intensity at 567 nm (fluoromorphic/
ratiometric probe). In comparison with probe 1, better
separation of emission curves and a longer emission
wavelength was achieved with substrates 7 and 9
(Figure 6).
With three fluorogenic probes in hand, a continuous
fluorometric assay was developed and the kinetic param-
eters were determined for each substrate with rMCAD as
well as pig MCAD (pMCAD; Table 1). The fluorometric
assay was validated by comparison with the standard
spectrophotometric assay (UV/Vis absorption); an excel-
lent agreement was found between these two methods
(see Supporting Information). Interestingly, all three
probes had KaMpp on the same order of magnitude (low
micromolar) as the optimum physiological substrate,
octanoyl-CoA. Probe 7 had the highest KaMpp for both
rMCAD and pMCAD, which supports the idea that polar
groups disfavor binding to the hydrophobic pocket of the
enzyme. In contrast to KaMpp, kcat values were significantly
lower when compared with octanoyl-CoA (Table 1).
Figure 4. A) Emission spectra of 1 (c) and 3 (a) (lex =340 nm).
B) Emission spectra of 2 (c) and 4 (g) (lex =390 nm). 50 mm in
100 mm potassium phosphate buffer (pH 8).
Substrate 1 had the largest turnover frequency (kcat
=
370 ꢀ 10 and 150 ꢀ 4 minꢁ1 for rMCAD and pMCAD,
respectively); substitution or extension
of this motif led to a further decrease in
this quantity (Table 1).
We noted that the decrease in
enzyme activity (umgꢁ1) for each addi-
tional phenyl ring in the substrate bears
a striking similarity to the chain-length–
activity relationship of natural fatty
acids (Figure 7). Specifically, 3-phenyl-
propionyl-CoA had very similar spe-
cific activity to the optimal physiolog-
ical substrate, octanoyl-CoA, whereas
probe 1, which contains the naphtha-
lene system, had a similar activity to
Figure 5. Concurrent exploration of MCAD substrate flexibility and emission properties of
fluorophores. Esters 7 and 9 are substrates for MCAD.
dodecanoyl-CoA (C12, lauryl-CoA).
Moreover, there was only a fivefold
difference in activity between the an-
methoxy group, as in 7, furnished a good substrate. The shape
of these probes was also of consequence as was demonstrated
through the comparison of substrate 1 with its regioisomer 8.
The latter compound, wherein the propionyl chain is attached
to C1 instead of C2, was completely inert. This is consistent
thracene probe 9 and palmitoyl-CoA (C16). We propose a
simple rationale for these findings based on similarities in the
steric demand between the regular fatty acids and synthetic
probes. These similarities become apparent when the former
substrates are viewed in a folded conformation (Figure 7). X-
Angew. Chem. Int. Ed. 2006, 45, 637 –642
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
639

Communications
Figure 7. Specific activity (umgꢁ1) for straight-chain acyl-CoA’s and
aromatic substrates with rMCAD (see Supporting Information for
assay conditions).
Ray crystallographic analysis of the enzyme–substrate com-
plex showed that the active site is not long enough to
accommodate long substrates in an extended conformation
and, as a result, folding is required (although not to the extent
shown in Figure 7).[10,26]
The next major taskcentered on the evaluation of the
possibility of using these substrates as useful probes in tissue
samples. This included addressing the issues of selectivity,
stability, and the ability to be metabolized by subsequent
enzymes of the b-oxidation pathway.
In addition to MCAD, there are two other soluble
isozymes of the ACAD family that are involved in the
metabolism of straight-chain fatty acids, namely short-chain
acyl-CoA dehydrogenase (SCAD) and long-chain acyl-CoA
dehydrogenase (LCAD). Purified pig SCAD (pSCAD) and
human LCAD (hLCAD) were obtained and tested with
probes 1, 7, and 9. Importantly, none of these probes showed
any significant activity, which demonstrates a notable selec-
tivity for MCAD within the ACAD family. To address this
question in a relevant context, competitive substrate assays
were conducted by using rat-liver homogenate (Figure 8).
Conversion of the probe by the tissue homogenate was
monitored both in the presence and absence of a competitive
physiological substrate. Addition of butanoyl-CoA and iso-
valeryl-CoA, the optimal substrates for SCAD and isovaleryl-
CoA dehydrogenase (iVAD), respectively, had no effect,
which indicates that probes 1, 7, and 9 were not substrates for
these enzymes. In contrast, octanoyl-CoA competed with the
probes, which led to a significant inhibition of fluorescent
product formation. A fivefold excess of octanoyl-CoA
completely inhibited dehydrogenation of probe 7 as a result
of the weaker binding (higher KM value) of this reporter
substrate. Dehydrogenation of all the probes was inhibited by
an excess of palmitoyl-CoA, which was consistent with
relatively low KM as well as low kcat values of this substrate
for MCAD (Figure 8). These data, including the experiments
with isolated isozymes and tissue homogenates, provided
compelling evidence for the high selectivity of the probes 1, 7,
and 9 for MCAD.
Figure 6. A) Emission spectra of 7 (c) and 11 (a) (lex =350 nm).
B) Emission spectra of 9 (c) and 12 (g) (lex =356 nm). Com-
pound 7, 9, 11, and 12 (50 mm) in potassium phosphate buffer
solution (100 mm, pH 8). The fluorescence is not affected by the assay
medium.
Table 1: Kinetic parameters of probes (see Supporting Information).
KaMpp
[mm]
kcat
[minꢁ1
]
octanoyl-CoA
rMCAD[30]
pMCAD[31]
4.0
2.3ꢀ0.1
2142
1176
rMCAD
pMCAD
homogenate
2.5ꢀ0.2[a]
2.3ꢀ0.2
2.2ꢀ0.3
370ꢀ10
150ꢀ4
–
rMCAD
pMCAD
homogenate
6.01ꢀ0.7
14.5ꢀ1.5
3.7ꢀ0.4
16.1ꢀ0.3
12.3ꢀ0.5
–
rMCAD
pMCAD
homogenate
1.1ꢀ0.2
4.5ꢀ0.5
1.9ꢀ0.4
10.4ꢀ0.3
13.1ꢀ2
–
The stability of the fluorescent products 3, 11, and 12 was
also examined. As these compounds may be metabolized by
[a] MeanꢀS.D. (n=3).
640
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 637 –642

Angewandte
Chemie
Reyeꢁs syndrome and sudden infant death syndrome. Early
diagnosis may prevent the onset of the symptoms through a
regulated diet and avoidance of fasting and excessive physical
activity.
Generally, fluorometric methods offer superior detection
sensitivity (up to two orders of magnitude) in comparison
with UV/Vis absorption methods. In this case, the fluorogenic
probe 7 allowed for reliable detection of MCAD activity with
as little as 0.4 mg of tissue homogenate protein, which is more
sensitive than the existing UV/Vis methods.[29] Further to high
sensitivity, fluorescence-based measurements may be per-
formed in a high-throughput manner in a routine 96-well-
plate format. Based on these results, a new and practical
diagnostic test may be developed that offers an attractive
alternative to the more involved methods used today (e.g.
MS-MS).
In a long-term view, the development of these probes sets
the stage for the investigation of the b-oxidation pathway in
its intact state. The possibility of measuring and imaging the
activity of MCAD, as well as the flux through this pathway,
will be examined. This represents an exciting prospect
considering the central role of this pathway in a number of
metabolic diseases.[17–19]
Figure 8. Competitive assays between the fluorogenic substrates (1, 7,
and 9) and the physiological substrates (butanoyl-CoA, isovaleryl-CoA,
octanoyl-CoA, and palmitoyl-CoA). Assays were carried out in a 96-well
plate with physiological substrate (50 mm), fluorescent substrate
(10 mm), and rat-liver homogenate as the enzyme source. The margin
of error was <20% for all assays (n=3).
Received: July 29, 2005
Published online: December 19, 2005
the subsequent enzyme(s) of the b-oxidation pathway, that is,
enoyl-CoA hydratase, they were exposed to rat-liver homo-
genate. Although no hydration occurred, a small decrease in
fluorescence was detected with 3 and 11, which was inde-
pendent of the homogenate. Rather, products 3 and 11 were
subject to photobleaching, whereas compound 12 was stable
on the time scale of the experiment. Importantly, photo-
bleaching did not significantly affect the results of the assay.
The relationship between activity and homogenate pro-
tein concentration was established for each probe and appears
to be linear within narrow ranges of protein concentrations
(See Supporting Information). As expected, enzyme activity
was dependent on the external electron acceptor, in this case,
ferrocenium hexafluorophosphate (FcPF6).[27] When the
external oxidant was omitted, probes 1 and 9 were inert,
whereas probe 7 gave a detectable residual activity (ꢂ 2% of
the total activity). As the oxidase activity (direct transfer of
electrons to oxygen)[22,23] of the purified MCAD protein was
less than 0.03% with probe 7, the residual activity could be
ascribed to either endogenous electron acceptors in the
homogenate or peroxisomal acyl-CoA oxidase (ACO).[8]
These results indicate that the ACO does not contribute to
the observed fluorescent signal in any significant manner, but
rather provides further support for high selectivity of these
probes for MCAD.
Keywords: enzyme promiscuity · fluorescent probe · medium-
.
chain acyl-CoA dehydrogenase · metabolism · oxidation
[1] P. J. OꢁBrien, D. Herschlag, Chem. Biol. 1999, 6, R91 – R105.
[2] D. J. Yee, V. Balsanek, D. Sames, J. Am. Chem. Soc. 2004, 126,
2282 – 2283.
[3] G. Chen, D. J. Yee, N. G. Gubernator, D. Sames, J. Am. Chem.
Soc. 2005, 127, 4544 – 4545.
[4] By the term ꢀfluoromorphic probeꢁ, we designate a fluorescent
compound, which, by the chemical reaction of interest, affords a
fluorescent product with a different emission wavelength profile.
The term ꢀratiometricꢁ is also used for this type of probe.
[5] W. H. Kunau, V. Dommes, H. Schulz, Prog. Lipid Res. 1995, 34,
267 – 342.
[6] S. Eaton, K. Bartlett, M. Pourfarzam, Biochem. J. 1996, 320(Pt2),
345 – 357.
[7] S. Ghisla, Eur. J. Biochem. 2004, 271, 459 – 461.
[8] K. Bartlett, S. Eaton, Eur. J. Biochem. 2004, 271, 462 – 469.
[9] N. Gregersen, P. Bross, B. S. Andresen, Eur. J. Biochem. 2004,
271, 470 – 482.
[10] J. J. Kim, R. Miura, Eur. J. Biochem. 2004, 271, 483 – 493.
[11] S. Ghisla, C. Thorpe, Eur. J. Biochem. 2004, 271, 494 – 508.
[12] F. Knoop, Beitr. Chem. Physiol. Pathol. 1904, 6, 150 – 162.
[13] The reverse process, namely fatty acid or polyketide biosynthesis
ought to possess similar substrate tolerance. In the latter case,
this phenomenon was demonstrated and exploited for the
incorporation of non-natural building units to erythronolide
antibiotics. J. R. Jacobsen, C. R. Hutchinson, D. E. Cane, C.
Khosla, Science 1997, 277, 367.
The immediate impact of this study is the ability to
monitor MCAD activity in cell and tissue homogenates in a
direct and continuous manner, therefore laying the founda-
tion for a new, sensitive, and practical diagnostic test for
MCAD deficiency.[28] MCAD deficiency has recently been
identified as a common hereditary disease with an estimated
occurrence rate of 1:15000 newborns in the US (an incidence
rate similar to that of phenylketonuria).[16] This deficiency
often results in severe symptoms that resemble those of
[14] S. Kolvraa, N. Gregersen, E. Christensen, N. Hobolth, Clin.
Chim. Acta 1982, 126, 53 – 67.
[15] W. J. Rhead, B. A. Amendt, K. S. Fritchman, S. J. Felts, Science
1983, 221, 73 – 75.
[16] B. S. Andresen, S. F. Dobrowolski, L. OꢁReilly, J. Muenzer, S. E.
McCandless, D. M. Frazier, S. Udvari, P. Bross, I. Knudsen, R.
Angew. Chem. Int. Ed. 2006, 45, 637 –642
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
641

Communications
Banas, D. H. Chace, P. Engel, E. W. Naylor, N. Gregersen, Am. J.
Hum. Genet. 2001, 68, 1408 – 1418.
[17] R. J. Wanders, P. Vreken, M. E. den Boer, F. A. Wijburg, A. H.
van Gennip, L. IJLst, J. Inherited Metab. Dis. 1999, 22, 442 – 487.
[18] G. I. Shulman, J. Clin. Invest. 2000, 106, 171 – 176.
[19] B. B. Lowell, G. I. Shulman, Science 2005, 307, 384 – 387.
[20] J. Am. Med. Assoc. 2001, 285, 2486 – 2497.
[21] K. W. Yao, H. Schulz, Anal. Biochem. 1993, 214, 528 – 534.
[22] J. T. McFarland, M. Y. Lee, J. Reinsch, W. Raven, Biochemistry
1982, 21, 1224 – 1229.
[23] J. K. Johnson, Z. X. Wang, D. K. Srivastava, Biochemistry 1992,
31, 10564 – 10575.
[24] G. Weber, F. J. Farris, Biochemistry 1979, 18, 3075 – 3078.
[25] The emission intensity is directly proportional to the brightness.
See reference [3].
[26] H. J. Lee, M. Wang, R. Paschke, A. Nandy, S. Ghisla, J. J. Kim,
Biochemistry 1996, 35, 12412 – 12420.
[27] T. C. Lehman, D. E. Hale, A. Bhala, C. Thorpe, Anal. Biochem.
1990, 186, 280 – 284.
[28] It has been shown by others that leukocytes or skin fibroblasts
may be used to monitor MCAD activity (see reference [21]).
[29] D. M. Turnbull, K. Bartlett, N. J. Watmough, I. M. Shepherd,
H. S. Sherratt, J. Inherited Metab. Dis. 1987, 10(Suppl 1), 105 –
112.
[30] K. L. Peterson, E. E. Sergienko, Y. Wu, N. R. Kumar, A. W.
Strauss, A. E. Oleson, W. W. Muhonen, J. B. Shabb, D. K.
Srivastava, Biochemistry 1995, 34, 14942 – 14953.
[31] Y. Ikeda, K. Okamura-Ikeda, K. Tanaka, J. Biol. Chem. 1985,
260, 1311 – 1325.
642
www.angewandte.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2006, 45, 637 –642