- Email: [email protected]
Stabilization of luciferase intermediates by fatty amines, amides, and nitriles
ARCHIVES
OF BIOCHEMISTRY
Vol. 294, No. 2, May
AND
BIOPHYSICS
1, pp. 361-366,
1992
Stabilization of Luciferase Intermediates Fatty Amines, Amides, and Nitriles
by
John C. Makemson,*,l J. Woodland Hastings,? and J. Martin E. Quirke$ *Department of Biological Sciences and SDepartment Florida 33199; and TBiological Laboratories, Harvard
Received
August
23, 1991, and in revised
form
December
of Chemistry, Florida International University, University University, 16 Divinity Avenue, Cambridge, Massachusetts
2, 1991
Long-chain aliphatic amides, monoand diamines, mono- and dialcohols, and nitriles were found to inhibit the bacterial luciferase reaction by binding with an enzyme intermediate (II, the luciferase-bound 4a-flavin hydroperoxide). Inhibition was determined by measuring the decay rates of the inhibitor-intermediate II complex at different inhibitor concentrations. The data fit a model which was used to estimate the KI. At high concentrations, a plot of the decay rate (k) vs l/[I] produced a straight line; extrapolation of this to l/[I] = 0 yields an estimate of the decay rate at infinite inhibitor concentration which we defined as the inhibitor-enzyme-substrate stabilization constant, kEsI. o 1992 Academic press, IW.
Bacterial luciferase catalyzes the mixed-function oxidation of a long-chain aldehyde and reduced flavin mononucleotide (FMNH.J by molecular oxygen to produce light (1). In this reaction (Fig. l), oxygen reacts rapidly with luciferase-bound FMNHB to form the luciferase 4a-flavin hydroperoxide, intermediate II, which is relatively longlived (tens of seconds at 25°C) and can either decay through the “dark pathway” to FMN and hydrogen peroxide or react with aldehyde via the light-emitting pathway (Fig. 1, top line). For the purpose of this study it is important to underscore the fact that the in vitro luciferase reaction, when initiated with FMNHz, constitutes a single turnover assay; that is, each luciferase molecule undergoes only a single catalytic cycle. Thus the decrease in luminescence is a direct measure of the decay of the longest lived luciferase intermediate. Several different inhibitors of bacterial luciferase bind to its aldehyde site. These include aliphatic aldehyde analogs (2, 3) as well as other mixed-function oxidase inhibitors (4-6). Binding to the aldehyde site may be very 1 To whom
correspondence
should
0003.9861/92 $3.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form
be addressed.
Park, Miami, 02138
tight, enough so that these inhibitors can reversibly trap reaction intermediate II (the 4a-flavin hydroperoxide), thus extending its lifetime and thereby facilitating its purification (7,8). Indeed, one of these inhibitors has been coupled to Sepharose to construct an affinity column for the purification of luciferase (9). Sobolev and Danilov (10) have shown that decylamine is also an effective inhibitor of luciferase. We have investigated the actions of different-chain-length amines, amides, and nitriles and compared these to the alcohols (3, 11) by measuring the effects of the inhibitors on the decay of luminescence. We also present a kinetic analysis of intermediate II-inhibitor decay by two different methods that define an inhibitor-enzyme-substrate intermediate stabilization constant and an inhibitor-enzyme intermediate binding constant from dark-decay kinetics. Dark decay of the intermediate-inhibitor complex occurs through release of inhibitor-forming L-FMNHOOH (intermediate II) which then decays to luciferase, FMN, and hydrogen peroxide (dark decay, Fig. 1). MATERIALS
AND
METHODS
Luciferase was purified to near homogeneity from Vibrti harveyi strain B-392 by the methods of Hastings et al. (12) and Holzman and Baldwin (9) utilizing ammonium sulfate fractionation, DEAE-cellulose fractionation (batch), and afhnity chromatography on 2,2diphenylpropylamineSepharose. SDS-polyacrylamide gels of the purified preparation revealed the bands corresponding to 40- and 36kDa subunits of bacterial luciferase. Two minor bands (24 and
’ Abbreviations used: SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; BSA, bovine serum albumin. 361
Inc. reserved.
362
MAKEMSON,
HASTINGS,
AND
QUIRKE
RCHO L + FMNH2
+ O2 __$
“Dark
L-FMNHOOH
Dew/
Scheme
+ RCOOH + hv
1%
L + FMN + H202 FIG. 1.
L + FMN + Hz0
of luciferase
L-FMNHOOH-1 reaction
and interaction
For measurements of the decay of intermediate II, luciferase was added to 1.0 ml of 0.05 M phosphate buffer, pH 7.0, at 25’C (+ inhibitor) followed by injection of 1 ml of 0.05 mu FMNHs at time zero. At intervals thereafter, 50-~1 samples were transferred to a cuvette in the photometer and 100 pl of 1 mM decanal was injected, thus initiating the luminescent reaction. The initial light intensity (I,,) was plotted vs time of aldehyde injection to obtain the decay rate of the intermediate II-inhibitor complex or of intermediate II alone. For the decay of intermediate II alone and for experiments in which the decay was rapid, at least six samples were taken at 20-s intervals. For slower decay rates, samples were removed at six but more widely separated times such that at least 30% of the intermediate formed had decayed. The decay rates were calculated using a least-squares linear regression of the natural log of Z0 vs time. The regression coefficients (r) for all data presented were greater than 0.95. Since short-chain alcohols can affect the luciferase assay without BSA (2), the decay in the presence of the same concentrations of ethanol as used as a solvent was taken as the control rate.
with
inhibitors
binding
to the aldehyde
The fatty amides were prepared from the corresponding acid chlorides (15). All other inhibitors were purchased from either Aldrich Chemical Company or Sigma Chemical Company at the highest purity available.
RESULTS AND DISCUSSION Dodecylumine. Dodecylamine decreases the rate of dark decay of luciferase intermediate II-inhibitor (Fig. 2). When the decay rate is plotted as a function of the reciprocal of this inhibitor concentration, the function appeared hyperbolic (Fig. 3). According to the model used (16), the observed rate of decay (k,,,) as a function of inhibitor concentration [I] is given by the following equation where Ki is the inhibitor binding constant, & is the rate of dark decay of intermediate II in the absence of inhibitor, and kEI is the rate of decay of the enzyme inhibitor complex: k a,,,,= tb + ItEr([~l/~dl/(1
IJJM 0.62pt.A
site.
+ [II/&).
In this analysis, [I] is known and b and kappare calculated from the decay rates in the absence or presence of inhibitor. The data were analyzed by a nonlinear statistics program (SYSTAT 3.0) that estimates Ki and kEI. From this model, the estimated &‘s closely approximated the in-
2.2 2.0 1.8 1.6 1.4 5
1.2 2% 1.0 Li 0.8 0.6
-00.0175pM
0.4
0
10
20
30
~/[DODE~YLAMINE, I
I
2
I
I
3
4
MINUTES
FIG. 2.
Effect of dodecylamine on the rate of luciferase intermediate II-dodecylamine complex decay. The concentrations of dodecylamine are indicated in the figure; the control lacked dodecylamine and had a rate of k = 2.15 min-‘.
FIG. 3.
40
50
60
LJM]
Effect of dodecylamine concentration on the decay rate of luciferase intermediate II-dodecylamine complex. The observed decay rate, k,, shown here as k, in min-‘, plotted as a function of l/[I] in pM. The linear regression line from 2.5 to 209 PM dodecylamine (l/[I] from 0.05 to 4) departs from the hyperbolic function at low (<2.5 PM) concentrations. At 17.5 nM dodecylamine, the decay was that of the uninhibited control, k = 2.15 min-‘.
LUCIFERASE TABLE
Constants
for Luciferase-Inhibitor
Amines Carbon
363
INHIBITORS I
Binding
(IQ and Stabilization
Alcohols
KI (PM)
hw
KI
length
(min-‘)
(PM)
8 9 10 11 12 14
ND” 3.21 0.45 0.061 0.021 0.18
ND 0.276 0.011 0.009 0.005 0.010
7.09 2.82 2.33 1.67 0.024 0.026
(kEsI) Amides
Nitriles hm (mini)
K (PM)
k Es1 (mini)
KI (PM)
hzsr (mini)
3.61 0.257 0.141 0.024 0.012 0.020
ND 6.03 1.89 0.52 0.058 ND
ND 0.259 0.092 0.018 0.011 ND
ND 19.4 5.1 2.2 0.73 ND
ND 0.768 0.212 0.128 0.093 ND
’ Not determined.
hibitor concentration which resulted in a 50% decrease in k,,, (Fig. 4) and the standard error of estimation of KI was always less than 10%. This was not the case for kEI; the estimates contained significant error and the model
could not be used to estimate inhibitor-enzyme intermediate decay. To obtain a measure of enzyme-inhibitor complex decay, we used plots of ka,,pvs l/[I] (Fig. 3). At high inhibitor
100 90
80 70
80
: 60 5 50 is
," 60
E
70
= 50 2 E 40
40
1
30
0.01
0.10
1 .oo
10.00
100.00
01
1.0
10.0 [AMIDE,
[ALCOHOL,~M]
1000
1000.0
10.00
100.00
JJM]
100 90
80 70 : 60 z yj 50
; 8
60
E
E
40
40
30 20 10 0 0.01
30 20 10 0.10
1.00 [NITRILE,
FIG.
50
10.00
100.00
0.01
0.10
pi]
4. Effect of inhibitor concentration upon the decay of luciferase intermediate II-inhibitor the control rate, $. This allowed normalization (removal of the effect of ethanol in the assays).
1 .oo [AMINL~M] complex.
The
rate is plotted
as a percentage
of
364
MAKEMSON,
HASTINGS,
concentrations (from 2.5 to 200 PM), these plots produced straight lines (Fig. 3) with linear regression correlation coefficients > 0.91. This was expected because hyperbolas tend to be linear at their extremes. Thus from these plots, the linear regression was used to determine the ordinate intersect (where l/[I] = 0) which would represent the decay of the enzyme-substrate-inhibitor complex at infinite inhibitor concentration. We call this value the enzyme-substrate-inhibitor stabilization constant, lZEsI (which is analogous to kzr): the lower the value, the slower is the rate of intermediate II-inhibitor complex decay (that is, intermediate II-inhibitor complex has been stabilized). Inhibitor binding. The data obtained with the alcohols were similar to those reported by Tu (3). Table I shows the Ki’s from analysis of the fatty alcohol, amide, amine, and nitrile data using the model. Figure 4 shows the inhibition of %k as a function of inhibitor concentration; the Ki’s estimated from the hyperbolic model were near 50%k. Regardless of the functional group, there is a decrease in K1 with an increase in aliphatic chain length up to 12 carbons (Fig. 5). The rate of log Ki decrease versus carbon chain length is almost parallel for the amides, nitriles, and amines, suggesting an incremental increase in the energy for transferring an extra methylene group from aqueous solution into a hydrocarbon site (aldehyde site). The alcohols did not follow this pattern, which may be due to the increase in hydration shell (hydrogen bonding) surrounding the alcohol functional group compared with the amides and nitriles. The amines certainly have a hydrogen-bonded hydration shell around the amine group, but the effect of the hydration shell may be offset by the attraction of the amine to the sulfhydryl in the active site (17). The 14-carbon alcohol had nearly the same, or slightly less, binding affinity than the 12-carbon alcohol. The Cl4 amine, however, bound 9 times less avidly than the Cl2 amine. The binding affinities of Cl2 amines and alcohols are nearly similar and just twice greater than those of nitriles, whereas amides were the lowest: Cl2 amide bound some 30 times less avidly than did Cl2 amine and alcohol but the g-carbon amide bound only 6 times less than 9carbon amine and alcohol. This difference stresses the affinity of the inhibitor (aldehyde) binding site for 12carbon rather than longer or shorter chain lengths. Intermediate II-inhibitor stabilization. Figure 6 shows the plots of k vs l/[inhibitor] for lo-, ll-, and 12-carbon fatty amides, amines, nitriles, and alcohols. In all cases the degree to which intermediate II-inhibitor complex was stabilized increased with aliphatic chain length up to 12 carbons; 14-carbon compounds were very nearly the same as 12-carbon ones (Table I). Further, amines appeared to form the most stable inhibitor complex (smallest kEsJ, followed by the alcohols and nitriles and then the amides (Fig. 7). Stabilization of 11-carbon inhibitors
AND
QUIRKE
1 00
h.7 0.10
0.01 a
9
10 CARBON
11 CHAIN
12
13
14
LENGTH
Comparison of KI (PM) to carbon chain length of luciferase inhibitors. The regression line is plotted for amides, nitriles, and amines (r = 0.996,0.987, and 0.992, respectively). The K{s of the alcohols are simply connected.
FIG. 6.
nearly equaled that of the 12-carbon inhibitors with the same functional group, which was largely different from their binding relationship (Fig. 5). The kEsI)sof the corresponding-chain-length alcohols and nitriles were almost indistinguishable (Fig. 7). As with the 1Ccarbon inhibitors in binding, stabilization by tetradecanol or tetradecamine was less (higher kEsl) than that with dodecanol or dodecamine, respectively. This further suggeststhat the hydrocarbon aldehyde site not only binds but accommodates stabilization of 12-carbon-length inhibitors better than longer- or shorter-carbon-length inhibitors. The relatively poor stabilization by the amides may be attributable to the bulkiness of the amide and its hydration shell (and therefore poor binding, see above) compared with the alcohols or nitriles. Strong stabilization by the amine could suggestthat ionic bridging is important in the active site after binding of the inhibitor to that site. Indeed, Paguatte et al. (17) have shown that cysteine106 is crucial to aldehyde binding. Such a group could be negatively charged enough so as to attract and hold the positively charged amine in the inhibitor (aldehyde) site. Note that amines stabilize the intermediate more than twice as strong as the alcohols (11 and 12 carbons) and more than 10 times as strong compared with the lo-carbon alcohol. This suggests that stabilization was more dependent on functional group than chain length of the aliphatic portion of the molecule. Decylamine was previously shown to inhibit luciferase using the coupled enzyme assay with NADH as the source of reducing power (10). In that work, the & was calculated to be 0.5 PM, which is very close to the KI of 0.3 PM obtained here for decylamine. Note that this binding represents a five times greater affinity for the inhibitor over that of the aldehyde substrate, decanal (K, = 1.5 PM). Comparison of mono- and dialcohols and amines. Table II compares the KI’s and kEsI)s for lo- and 12-carbon mono- and dialcohols and amines. The presence of the
LUCIFERASE
0.0
'
-1
,01 I/[ALCOHOL,
0.0
0.0 FIG. 6. complex.
Effect Plotted
365
INHIBITORS
0.2
0.1
0.0
us]
0.2 I/[NITRILE,
0.4
00
UM]
of high concentrations of fatty alcohols, amides, amines, as in Fig. 3 for high concentrations (>5 pM) of inhibitor.
alcohol on the second carbon greatly decreases the binding affinity for the 12-carbon inhibitor and emphasizes the relationship between the bulkiness of the functional group and binding: the bulkier the group such as 2-01 and 1,2diol compared with the l-01, the less is the binding afhnity, by 60 to 240 times. Yet, once bound, there was not such
and nitriles
on the stabilization
0.2
I/[AMIDE,
us]
01 l/[AMINE,
uM]
0.2
of luciferase
intermediate
II-inhibitor
a dramatic difference in stabilization: the 2-01 stabilizes the intermediate just less than 3 times as poorly than the l-01 even though its binding is 60 times less avid. Moreover, the 1,2-diol stabilized the intermediate only 10 times less than the l-01, yet it bound 240 times less avidly. When alcohol functional groups were at both ends of a lo-carbon chain, binding was less adversely affected TABLE
II
Comparison of 2-ols, Diols, and Diamines as Luciferase Inhibitors K Inhibitor
CUM)
k ES, (min-‘1
Cl0 0.010
l-01 l,lO-diol
2.33 5.06
0.141
l-01 2-01 1,2-diol 1,12-diol 1 -amine 1,12diamine
0.024
0.012
1.49
0.033 0.122 0.070 0.005 0.642
0.559
Cl2 0001 a
9
10 CARBON
FIG. 7. inhibitors.
Comparison
of ksst (min-‘)
11 CHAIN
12
13
14
LENGTH
to carbon
chain
length
of luciferase
5.8 1.14 0.021
31.9
366
MAKEMSON,
HASTINGS,
than was stabilization: for the lo-carbon l,lO-diol binding is 2 times less avid than for the l-01, but stabilization of the intermediate was 3.9 times less than for the l-01. For the 12-carbon compounds, binding of the 1,12-diol was 47 times less avid but the stabilization was only 6 times less than that of the l-01. For the diamine (Cl2 1,12diamine) the presence of the second charged functional group was more dramatic than for the alcohols: binding was 1500 times less avid, whereas stabilization was only 128 times as poor. The plot and constants obtained in these experiments allowed analysis of inhibitor binding and a kinetic analysis of inhibitor enzyme intermediate stabilization for nonturnover conditions: a condition in which the apparent first-order intermediate-inhibitor decay rate constant can be obtained. This has obvious and immediate application to luciferase, which is conveniently assayed by an unconventional non-turnover enzyme assay. ACKNOWLEDGMENTS We thank Dr. Therese Wilson and David Kuhn for helpful This research was supported in part by Grant NSF-8616522 U.S. National Science Foundation (to J.W.H.).
comments. from the
REFERENCES 1. Hastings, Makemson,
J. W., Potrikus, C. J., Gupta, S. C., Kurfurst, M., J. C. (1985) Adu. Microbial. Pkysiol. 26, 235-291.
and
AND
QUIRKE
2. Hastings, 3. 4. 5.
J. W., Gibson, Q. H., in Bioluminescence in Progress Eds.), pp. 151-186, Princeton Tu, S.-C. (1979) Biochemistry Nealson, K. H., and Hastings, 888-894. Makemson, J. C., and Hastings,
Friedland, J., and Spudich, J. (1966) (Johnson, F. H., and Haneda, Y., Univ. Press, Princeton, NJ. l&5940-5945. J. W. (1972) J. Biol. Chem. 247, J. W. (1979) Arch.
Biochem.
Biophys.
196,396-402. 6. Holzman, 5528.
T. F., and Baldwin,
T. 0. (1981)
Biochemi&y
20,5524-
7. Tu, S.-C. (1982) J. Biol. Ckem. 267, 3719-3725. 8. Tu, S.-C. (1986) in Methods in Enzymology (Deluca, M. A., and McElroy, W. D., Eds.), Vol. 133, pp. 128-139, Academic Press, New York. 9. Holzman, T. F., and Baldwin, T. 0. (1982) Biochemistry 2 1,61946201. 10. Sobolev, A. Y., and Danilov, V. S. (1988) Biokhimiya (Engl. Transl.) 63, 788-793 (original article: 63, 912-917). 11. Baumstark, A., Cline, T., and Hastings, J. W. (1979) Arch. Biockem. Biophys. 193, 449-455. 12. Hastings, J. W., Baldwin, T. O., and Nicoli, M. Z. (1978) in Methods in Enzymology (DeLuca, M. A., Ed.), Vol. 57, pp. 135-152, Academic Press, New York. G., and Hastings, J. W. (1971) Anal. Biochem. 39, 24313. Mitchell,
250. 14. Hastings, J. W., and Weber, G. (1963) J. Opt. Sot. Am. 53, 14101415. 15. Vogel, A. (1978) in Textbook of Practical Organic Chemistry, 4th ed., p. 1116, Longman Scientific, London. 16. Cleland, W. W. (1979) in Methods in Enzymology (Purlich, D. L., Ed.), Vol. 63, pp. 103-138, Academic Press, New York. 17. Paguatte, O., Fried, A., and Tu, S.-C. (1988) Arch. B&kern. Biophys.
264,392-399.
OF BIOCHEMISTRY
Vol. 294, No. 2, May
AND
BIOPHYSICS
1, pp. 361-366,
1992
Stabilization of Luciferase Intermediates Fatty Amines, Amides, and Nitriles
by
John C. Makemson,*,l J. Woodland Hastings,? and J. Martin E. Quirke$ *Department of Biological Sciences and SDepartment Florida 33199; and TBiological Laboratories, Harvard
Received
August
23, 1991, and in revised
form
December
of Chemistry, Florida International University, University University, 16 Divinity Avenue, Cambridge, Massachusetts
2, 1991
Long-chain aliphatic amides, monoand diamines, mono- and dialcohols, and nitriles were found to inhibit the bacterial luciferase reaction by binding with an enzyme intermediate (II, the luciferase-bound 4a-flavin hydroperoxide). Inhibition was determined by measuring the decay rates of the inhibitor-intermediate II complex at different inhibitor concentrations. The data fit a model which was used to estimate the KI. At high concentrations, a plot of the decay rate (k) vs l/[I] produced a straight line; extrapolation of this to l/[I] = 0 yields an estimate of the decay rate at infinite inhibitor concentration which we defined as the inhibitor-enzyme-substrate stabilization constant, kEsI. o 1992 Academic press, IW.
Bacterial luciferase catalyzes the mixed-function oxidation of a long-chain aldehyde and reduced flavin mononucleotide (FMNH.J by molecular oxygen to produce light (1). In this reaction (Fig. l), oxygen reacts rapidly with luciferase-bound FMNHB to form the luciferase 4a-flavin hydroperoxide, intermediate II, which is relatively longlived (tens of seconds at 25°C) and can either decay through the “dark pathway” to FMN and hydrogen peroxide or react with aldehyde via the light-emitting pathway (Fig. 1, top line). For the purpose of this study it is important to underscore the fact that the in vitro luciferase reaction, when initiated with FMNHz, constitutes a single turnover assay; that is, each luciferase molecule undergoes only a single catalytic cycle. Thus the decrease in luminescence is a direct measure of the decay of the longest lived luciferase intermediate. Several different inhibitors of bacterial luciferase bind to its aldehyde site. These include aliphatic aldehyde analogs (2, 3) as well as other mixed-function oxidase inhibitors (4-6). Binding to the aldehyde site may be very 1 To whom
correspondence
should
0003.9861/92 $3.00 Copyright 0 1992 by Academic Press, All rights of reproduction in any form
be addressed.
Park, Miami, 02138
tight, enough so that these inhibitors can reversibly trap reaction intermediate II (the 4a-flavin hydroperoxide), thus extending its lifetime and thereby facilitating its purification (7,8). Indeed, one of these inhibitors has been coupled to Sepharose to construct an affinity column for the purification of luciferase (9). Sobolev and Danilov (10) have shown that decylamine is also an effective inhibitor of luciferase. We have investigated the actions of different-chain-length amines, amides, and nitriles and compared these to the alcohols (3, 11) by measuring the effects of the inhibitors on the decay of luminescence. We also present a kinetic analysis of intermediate II-inhibitor decay by two different methods that define an inhibitor-enzyme-substrate intermediate stabilization constant and an inhibitor-enzyme intermediate binding constant from dark-decay kinetics. Dark decay of the intermediate-inhibitor complex occurs through release of inhibitor-forming L-FMNHOOH (intermediate II) which then decays to luciferase, FMN, and hydrogen peroxide (dark decay, Fig. 1). MATERIALS
AND
METHODS
Luciferase was purified to near homogeneity from Vibrti harveyi strain B-392 by the methods of Hastings et al. (12) and Holzman and Baldwin (9) utilizing ammonium sulfate fractionation, DEAE-cellulose fractionation (batch), and afhnity chromatography on 2,2diphenylpropylamineSepharose. SDS-polyacrylamide gels of the purified preparation revealed the bands corresponding to 40- and 36kDa subunits of bacterial luciferase. Two minor bands (24 and
’ Abbreviations used: SDS-PAGE, sodium dodecyl sulfate-polyacrylamide gel electrophoresis; BSA, bovine serum albumin. 361
Inc. reserved.
362
MAKEMSON,
HASTINGS,
AND
QUIRKE
RCHO L + FMNH2
+ O2 __$
“Dark
L-FMNHOOH
Dew/
Scheme
+ RCOOH + hv
1%
L + FMN + H202 FIG. 1.
L + FMN + Hz0
of luciferase
L-FMNHOOH-1 reaction
and interaction
For measurements of the decay of intermediate II, luciferase was added to 1.0 ml of 0.05 M phosphate buffer, pH 7.0, at 25’C (+ inhibitor) followed by injection of 1 ml of 0.05 mu FMNHs at time zero. At intervals thereafter, 50-~1 samples were transferred to a cuvette in the photometer and 100 pl of 1 mM decanal was injected, thus initiating the luminescent reaction. The initial light intensity (I,,) was plotted vs time of aldehyde injection to obtain the decay rate of the intermediate II-inhibitor complex or of intermediate II alone. For the decay of intermediate II alone and for experiments in which the decay was rapid, at least six samples were taken at 20-s intervals. For slower decay rates, samples were removed at six but more widely separated times such that at least 30% of the intermediate formed had decayed. The decay rates were calculated using a least-squares linear regression of the natural log of Z0 vs time. The regression coefficients (r) for all data presented were greater than 0.95. Since short-chain alcohols can affect the luciferase assay without BSA (2), the decay in the presence of the same concentrations of ethanol as used as a solvent was taken as the control rate.
with
inhibitors
binding
to the aldehyde
The fatty amides were prepared from the corresponding acid chlorides (15). All other inhibitors were purchased from either Aldrich Chemical Company or Sigma Chemical Company at the highest purity available.
RESULTS AND DISCUSSION Dodecylumine. Dodecylamine decreases the rate of dark decay of luciferase intermediate II-inhibitor (Fig. 2). When the decay rate is plotted as a function of the reciprocal of this inhibitor concentration, the function appeared hyperbolic (Fig. 3). According to the model used (16), the observed rate of decay (k,,,) as a function of inhibitor concentration [I] is given by the following equation where Ki is the inhibitor binding constant, & is the rate of dark decay of intermediate II in the absence of inhibitor, and kEI is the rate of decay of the enzyme inhibitor complex: k a,,,,= tb + ItEr([~l/~dl/(1
IJJM 0.62pt.A
site.
+ [II/&).
In this analysis, [I] is known and b and kappare calculated from the decay rates in the absence or presence of inhibitor. The data were analyzed by a nonlinear statistics program (SYSTAT 3.0) that estimates Ki and kEI. From this model, the estimated &‘s closely approximated the in-
2.2 2.0 1.8 1.6 1.4 5
1.2 2% 1.0 Li 0.8 0.6
-00.0175pM
0.4
0
10
20
30
~/[DODE~YLAMINE, I
I
2
I
I
3
4
MINUTES
FIG. 2.
Effect of dodecylamine on the rate of luciferase intermediate II-dodecylamine complex decay. The concentrations of dodecylamine are indicated in the figure; the control lacked dodecylamine and had a rate of k = 2.15 min-‘.
FIG. 3.
40
50
60
LJM]
Effect of dodecylamine concentration on the decay rate of luciferase intermediate II-dodecylamine complex. The observed decay rate, k,, shown here as k, in min-‘, plotted as a function of l/[I] in pM. The linear regression line from 2.5 to 209 PM dodecylamine (l/[I] from 0.05 to 4) departs from the hyperbolic function at low (<2.5 PM) concentrations. At 17.5 nM dodecylamine, the decay was that of the uninhibited control, k = 2.15 min-‘.
LUCIFERASE TABLE
Constants
for Luciferase-Inhibitor
Amines Carbon
363
INHIBITORS I
Binding
(IQ and Stabilization
Alcohols
KI (PM)
hw
KI
length
(min-‘)
(PM)
8 9 10 11 12 14
ND” 3.21 0.45 0.061 0.021 0.18
ND 0.276 0.011 0.009 0.005 0.010
7.09 2.82 2.33 1.67 0.024 0.026
(kEsI) Amides
Nitriles hm (mini)
K (PM)
k Es1 (mini)
KI (PM)
hzsr (mini)
3.61 0.257 0.141 0.024 0.012 0.020
ND 6.03 1.89 0.52 0.058 ND
ND 0.259 0.092 0.018 0.011 ND
ND 19.4 5.1 2.2 0.73 ND
ND 0.768 0.212 0.128 0.093 ND
’ Not determined.
hibitor concentration which resulted in a 50% decrease in k,,, (Fig. 4) and the standard error of estimation of KI was always less than 10%. This was not the case for kEI; the estimates contained significant error and the model
could not be used to estimate inhibitor-enzyme intermediate decay. To obtain a measure of enzyme-inhibitor complex decay, we used plots of ka,,pvs l/[I] (Fig. 3). At high inhibitor
100 90
80 70
80
: 60 5 50 is
," 60
E
70
= 50 2 E 40
40
1
30
0.01
0.10
1 .oo
10.00
100.00
01
1.0
10.0 [AMIDE,
[ALCOHOL,~M]
1000
1000.0
10.00
100.00
JJM]
100 90
80 70 : 60 z yj 50
; 8
60
E
E
40
40
30 20 10 0 0.01
30 20 10 0.10
1.00 [NITRILE,
FIG.
50
10.00
100.00
0.01
0.10
pi]
4. Effect of inhibitor concentration upon the decay of luciferase intermediate II-inhibitor the control rate, $. This allowed normalization (removal of the effect of ethanol in the assays).
1 .oo [AMINL~M] complex.
The
rate is plotted
as a percentage
of
364
MAKEMSON,
HASTINGS,
concentrations (from 2.5 to 200 PM), these plots produced straight lines (Fig. 3) with linear regression correlation coefficients > 0.91. This was expected because hyperbolas tend to be linear at their extremes. Thus from these plots, the linear regression was used to determine the ordinate intersect (where l/[I] = 0) which would represent the decay of the enzyme-substrate-inhibitor complex at infinite inhibitor concentration. We call this value the enzyme-substrate-inhibitor stabilization constant, lZEsI (which is analogous to kzr): the lower the value, the slower is the rate of intermediate II-inhibitor complex decay (that is, intermediate II-inhibitor complex has been stabilized). Inhibitor binding. The data obtained with the alcohols were similar to those reported by Tu (3). Table I shows the Ki’s from analysis of the fatty alcohol, amide, amine, and nitrile data using the model. Figure 4 shows the inhibition of %k as a function of inhibitor concentration; the Ki’s estimated from the hyperbolic model were near 50%k. Regardless of the functional group, there is a decrease in K1 with an increase in aliphatic chain length up to 12 carbons (Fig. 5). The rate of log Ki decrease versus carbon chain length is almost parallel for the amides, nitriles, and amines, suggesting an incremental increase in the energy for transferring an extra methylene group from aqueous solution into a hydrocarbon site (aldehyde site). The alcohols did not follow this pattern, which may be due to the increase in hydration shell (hydrogen bonding) surrounding the alcohol functional group compared with the amides and nitriles. The amines certainly have a hydrogen-bonded hydration shell around the amine group, but the effect of the hydration shell may be offset by the attraction of the amine to the sulfhydryl in the active site (17). The 14-carbon alcohol had nearly the same, or slightly less, binding affinity than the 12-carbon alcohol. The Cl4 amine, however, bound 9 times less avidly than the Cl2 amine. The binding affinities of Cl2 amines and alcohols are nearly similar and just twice greater than those of nitriles, whereas amides were the lowest: Cl2 amide bound some 30 times less avidly than did Cl2 amine and alcohol but the g-carbon amide bound only 6 times less than 9carbon amine and alcohol. This difference stresses the affinity of the inhibitor (aldehyde) binding site for 12carbon rather than longer or shorter chain lengths. Intermediate II-inhibitor stabilization. Figure 6 shows the plots of k vs l/[inhibitor] for lo-, ll-, and 12-carbon fatty amides, amines, nitriles, and alcohols. In all cases the degree to which intermediate II-inhibitor complex was stabilized increased with aliphatic chain length up to 12 carbons; 14-carbon compounds were very nearly the same as 12-carbon ones (Table I). Further, amines appeared to form the most stable inhibitor complex (smallest kEsJ, followed by the alcohols and nitriles and then the amides (Fig. 7). Stabilization of 11-carbon inhibitors
AND
QUIRKE
1 00
h.7 0.10
0.01 a
9
10 CARBON
11 CHAIN
12
13
14
LENGTH
Comparison of KI (PM) to carbon chain length of luciferase inhibitors. The regression line is plotted for amides, nitriles, and amines (r = 0.996,0.987, and 0.992, respectively). The K{s of the alcohols are simply connected.
FIG. 6.
nearly equaled that of the 12-carbon inhibitors with the same functional group, which was largely different from their binding relationship (Fig. 5). The kEsI)sof the corresponding-chain-length alcohols and nitriles were almost indistinguishable (Fig. 7). As with the 1Ccarbon inhibitors in binding, stabilization by tetradecanol or tetradecamine was less (higher kEsl) than that with dodecanol or dodecamine, respectively. This further suggeststhat the hydrocarbon aldehyde site not only binds but accommodates stabilization of 12-carbon-length inhibitors better than longer- or shorter-carbon-length inhibitors. The relatively poor stabilization by the amides may be attributable to the bulkiness of the amide and its hydration shell (and therefore poor binding, see above) compared with the alcohols or nitriles. Strong stabilization by the amine could suggestthat ionic bridging is important in the active site after binding of the inhibitor to that site. Indeed, Paguatte et al. (17) have shown that cysteine106 is crucial to aldehyde binding. Such a group could be negatively charged enough so as to attract and hold the positively charged amine in the inhibitor (aldehyde) site. Note that amines stabilize the intermediate more than twice as strong as the alcohols (11 and 12 carbons) and more than 10 times as strong compared with the lo-carbon alcohol. This suggests that stabilization was more dependent on functional group than chain length of the aliphatic portion of the molecule. Decylamine was previously shown to inhibit luciferase using the coupled enzyme assay with NADH as the source of reducing power (10). In that work, the & was calculated to be 0.5 PM, which is very close to the KI of 0.3 PM obtained here for decylamine. Note that this binding represents a five times greater affinity for the inhibitor over that of the aldehyde substrate, decanal (K, = 1.5 PM). Comparison of mono- and dialcohols and amines. Table II compares the KI’s and kEsI)s for lo- and 12-carbon mono- and dialcohols and amines. The presence of the
LUCIFERASE
0.0
'
-1
,01 I/[ALCOHOL,
0.0
0.0 FIG. 6. complex.
Effect Plotted
365
INHIBITORS
0.2
0.1
0.0
us]
0.2 I/[NITRILE,
0.4
00
UM]
of high concentrations of fatty alcohols, amides, amines, as in Fig. 3 for high concentrations (>5 pM) of inhibitor.
alcohol on the second carbon greatly decreases the binding affinity for the 12-carbon inhibitor and emphasizes the relationship between the bulkiness of the functional group and binding: the bulkier the group such as 2-01 and 1,2diol compared with the l-01, the less is the binding afhnity, by 60 to 240 times. Yet, once bound, there was not such
and nitriles
on the stabilization
0.2
I/[AMIDE,
us]
01 l/[AMINE,
uM]
0.2
of luciferase
intermediate
II-inhibitor
a dramatic difference in stabilization: the 2-01 stabilizes the intermediate just less than 3 times as poorly than the l-01 even though its binding is 60 times less avid. Moreover, the 1,2-diol stabilized the intermediate only 10 times less than the l-01, yet it bound 240 times less avidly. When alcohol functional groups were at both ends of a lo-carbon chain, binding was less adversely affected TABLE
II
Comparison of 2-ols, Diols, and Diamines as Luciferase Inhibitors K Inhibitor
CUM)
k ES, (min-‘1
Cl0 0.010
l-01 l,lO-diol
2.33 5.06
0.141
l-01 2-01 1,2-diol 1,12-diol 1 -amine 1,12diamine
0.024
0.012
1.49
0.033 0.122 0.070 0.005 0.642
0.559
Cl2 0001 a
9
10 CARBON
FIG. 7. inhibitors.
Comparison
of ksst (min-‘)
11 CHAIN
12
13
14
LENGTH
to carbon
chain
length
of luciferase
5.8 1.14 0.021
31.9
366
MAKEMSON,
HASTINGS,
than was stabilization: for the lo-carbon l,lO-diol binding is 2 times less avid than for the l-01, but stabilization of the intermediate was 3.9 times less than for the l-01. For the 12-carbon compounds, binding of the 1,12-diol was 47 times less avid but the stabilization was only 6 times less than that of the l-01. For the diamine (Cl2 1,12diamine) the presence of the second charged functional group was more dramatic than for the alcohols: binding was 1500 times less avid, whereas stabilization was only 128 times as poor. The plot and constants obtained in these experiments allowed analysis of inhibitor binding and a kinetic analysis of inhibitor enzyme intermediate stabilization for nonturnover conditions: a condition in which the apparent first-order intermediate-inhibitor decay rate constant can be obtained. This has obvious and immediate application to luciferase, which is conveniently assayed by an unconventional non-turnover enzyme assay. ACKNOWLEDGMENTS We thank Dr. Therese Wilson and David Kuhn for helpful This research was supported in part by Grant NSF-8616522 U.S. National Science Foundation (to J.W.H.).
comments. from the
REFERENCES 1. Hastings, Makemson,
J. W., Potrikus, C. J., Gupta, S. C., Kurfurst, M., J. C. (1985) Adu. Microbial. Pkysiol. 26, 235-291.
and
AND
QUIRKE
2. Hastings, 3. 4. 5.
J. W., Gibson, Q. H., in Bioluminescence in Progress Eds.), pp. 151-186, Princeton Tu, S.-C. (1979) Biochemistry Nealson, K. H., and Hastings, 888-894. Makemson, J. C., and Hastings,
Friedland, J., and Spudich, J. (1966) (Johnson, F. H., and Haneda, Y., Univ. Press, Princeton, NJ. l&5940-5945. J. W. (1972) J. Biol. Chem. 247, J. W. (1979) Arch.
Biochem.
Biophys.
196,396-402. 6. Holzman, 5528.
T. F., and Baldwin,
T. 0. (1981)
Biochemi&y
20,5524-
7. Tu, S.-C. (1982) J. Biol. Ckem. 267, 3719-3725. 8. Tu, S.-C. (1986) in Methods in Enzymology (Deluca, M. A., and McElroy, W. D., Eds.), Vol. 133, pp. 128-139, Academic Press, New York. 9. Holzman, T. F., and Baldwin, T. 0. (1982) Biochemistry 2 1,61946201. 10. Sobolev, A. Y., and Danilov, V. S. (1988) Biokhimiya (Engl. Transl.) 63, 788-793 (original article: 63, 912-917). 11. Baumstark, A., Cline, T., and Hastings, J. W. (1979) Arch. Biockem. Biophys. 193, 449-455. 12. Hastings, J. W., Baldwin, T. O., and Nicoli, M. Z. (1978) in Methods in Enzymology (DeLuca, M. A., Ed.), Vol. 57, pp. 135-152, Academic Press, New York. G., and Hastings, J. W. (1971) Anal. Biochem. 39, 24313. Mitchell,
250. 14. Hastings, J. W., and Weber, G. (1963) J. Opt. Sot. Am. 53, 14101415. 15. Vogel, A. (1978) in Textbook of Practical Organic Chemistry, 4th ed., p. 1116, Longman Scientific, London. 16. Cleland, W. W. (1979) in Methods in Enzymology (Purlich, D. L., Ed.), Vol. 63, pp. 103-138, Academic Press, New York. 17. Paguatte, O., Fried, A., and Tu, S.-C. (1988) Arch. B&kern. Biophys.
264,392-399.