Content uploaded by Nikolai N. Sluchanko
Author content
All content in this area was uploaded by Nikolai N. Sluchanko on Feb 05, 2018
Content may be subject to copyright.
1
Effect of the NBD-group position on interaction of fluorescently-labeled cholesterol
analogues with human steroidogenic acute regulatory protein STARD1
Kristina V. Tugaeva1,2, Yaroslav V. Faletrov3, Elvin S. Allakhverdiev4, Vladimir M.
Shkumatov3,5, Eugene G. Maksimov6, Nikolai N. Sluchanko1,6*
1A.N. Bach Institute of Biochemistry, Federal Research Center of Biotechnology of the Russian
Academy of Sciences, 119071 Moscow, Russia
2Department of Biochemistry, School of Biology, M.V. Lomonosov Moscow State University,
119991 Moscow, Russia
3Research Institute for Physical Chemical Problems, Belarusian State University, Minsk, Belarus
4Faculty of Fundamental Medicine, M.V. Lomonosov Moscow State University, 119991
Moscow, Russia
5Department of Chemistry, Belarusian State University, Minsk, Belarus
6Department of Biophysics, School of Biology, M.V. Lomonosov Moscow State University,
119992 Moscow, Russia
*Corresponding author: Dr. Nikolai N. Sluchanko
A.N.Bach Institute of Biochemistry,
Federal Research Center of Biotechnology of RAS
Moscow 119071, Russian Federation
Tel/Fax +7-495-9521384
E-mail: nikolai.sluchanko@mail.ru
Abbreviations:
20NP – 20-((NBD)amino)-pregn-5-en-3-ol;
22NC – 22-(N-(NBD)amino)-23,24-bisnorcholesterol;
25NC – 25-[N-[(NBD)-methyl]amino]-27-norcholesterol;
3NC – 3-(NBD-amino)-cholestane;
EDTA – ethylenediaminetetraacetic acid;
NBD – 7-nitrobenz-2-oxa-1,3-diazol-4-yl.
Keywords: protein-ligand interactions; steroidogenic acute regulatory protein; solvatochromism;
fluorescence; thermal stability; cholesterol
*Manuscript
Click here to view linked References
2
Abstract
Steroidogenic acute regulatory protein (StAR, STARD1) is a key factor of intracellular
cholesterol transfer to mitochondria, necessary for adrenal and gonadal steroidogenesis, and is an
archetypal member of the START protein family. Despite the common overall structural fold,
START members differ in their binding selectivity toward various lipid ligands, but the lack of
direct structural information hinders complete understanding of the binding process and
cholesterol orientation in the STARD1 complex in particular. Cholesterol binding has been
widely studied by commercially available fluorescent steroids, but the effect of the fluorescent
group position on binding remained underexplored. Here, we dissect STARD1 interaction with
cholesterol-like steroids bearing 7-nitrobenz-2-oxa-1,3-diazol-4-yl (NBD) group in different
position, namely, with 22-NBD-cholesterol (22NC), 25-NBD-cholesterol (25NC), 20-
((NBDamino)-pregn-5-en-3-ol (20NP) and 3-(NBDamino)-cholestane (3NC). While being able
to stoichiometrically bind 22NC and 20NP with high fluorescence yield and quantitative
exhaustion of fluorescence of some protein tryptophans, STARD1 binds 25NC and 3NC with
much lower affinity and poor fluorescence yield. In contrast to 3NC, binding of 20NP leads to
STARD1 stabilization and substantially increases the NBD fluorescence lifetime. Remarkably,
in terms of fluorescence response, 20NP slightly outperforms commonly used 22NC and is thus
can be used for screening of various potential ligands by a competition mechanism in the future.
1. Introduction
Steroidogenic acute regulatory protein (StAR, STARD1) is thought to be a key factor
determining adequate rates of cholesterol transfer to the inner mitochondrial membrane, where
the adrenal and gonadal steroidogenesis starts from production of pregnenolone, a single source
for various steroid hormones [1,2,3,4]. Mutations in STARD1 impair its function and often lead
to various forms of fatal lipoid congenital adrenal hyperplasia (LCAH) [5,6]. Human STARD1 is
synthesized as a 37-kDa protein (285 residues) that contains a typical N-terminal leader sequence
targeting it to mitochondria, although the 30 kDa protein devoid of this N-terminal peptide
displays almost equal ability to promote steroidogenesis, at least in COS-1 cells [7]. The N-
terminal processing of STARD1 [8] seems to be regulated by cAMP levels [9] and STARD1
phosphorylation [10], and processing rates may differ by cell type [11], however, the exact role
of the STARD1 mitochondrial import in cholesterol transfer is not clearly elucidated. Latest
research revealed the importance of protein-protein interactions in the functioning of STARD1,
which was proposed to regulate cholesterol transfer as a component of the multiprotein complex
called transduceosome [12].
STARD1 is an archetypal member of the StAR-related lipid transfer (START) protein
family composed of 15 members sharing an α/β structured domain (~210 amino acids) that forms
a hydrophobic binding cavity with a size suitable for accommodation of steroids and other lipids
[13]. The overall fold of STAR domain is very conserved, however, START members were
reported to have distinctly different selectivity toward bile acids [14], ceramides and
phospholipids [15,16], steroid hormones [13,17], presumably, due to the differences in the lining
of the ligand-binding cavities.
The repertory of proven natural ligands of STARD1 is currently limited to cholesterol,
which is also the known ligand of STARD3, STARD4, STARD6, but not STARD5 [18]. The
crystal structure of STARD1 [19] (along with that of several other START members) and its
probable solution conformation [20] have recently been reported, however, the structural
information on cholesterol binding is limited to only several, sometimes contradictory, in silico
models that lack convincing experimental verification. Moreover, despite the presence of
cholesterol during STARD1 crystallization [19], the crystal structure (PDB 3P0L) was solved in
the apoform, leaving open the question of cholesterol binding mode and orientation. Binding of
cholesterol analogues and other ligands is not systematically studied yet, however, this may
3
inform designing of artificial compounds for targeting mitochondria of malignant steroidogenic
cells [21,22].
Since cholesterol is an optically inactive substance, researchers used its radiolabeled [23]
or fluorescently labeled [23,24] analogues. Fluorescent analogues of cholesterol including
commercially available 22- (22NC) and 25-NBD-labeled cholesterols (25NC) [25,26] have been
successfully used to study lipid trafficking and different enzymes [26,27,28]. Besides obvious
advantages, the main disadvantage is the presence of a relatively bulky fluorescent group which
can have unpredictable effects on binding and other studied properties [25]. It was shown that
22NC preserves the binding properties of [14C]cholesterol toward STARD1 and its mutants [23],
however, the effect of the NBD-group position on binding to STARD1 remains underexplored.
This study is devoted to the detailed analysis of interaction between STARD1 and a
series of cholesterol analogues with different position of the NBD group using steady-state and
time-resolved fluorescence spectroscopy. To avoid previously reported problems with STARD1
aggregation [24,29,30], we used monodisperse and monomeric recombinant STARD1 [20]
obtained using our recently reported purification scheme [30].
2. Materials and methods
2.1. Materials
22NC was from Molecular Probes (Eugene, OR), 25NC was from Avanti Polar Lipids
(Birmingham, AL). Cholesterol was from Sigma-Aldrich (St. Louis, MO). NBD-labeled steroids
were synthesized for this work as described [31], via steps of reductive amination of either
pregnenolone or cholestenone and further reaction with 7-nitrobenzoxadiazole-4-yl chloride
[32,33,34]. 20NP was obtained as racemic mixture of two isomers, 20NPα and 20NPβ, which
were separated as in [30]. In this study only 20NPβ (20S-; retention time of 30.0 min [30]) was
used because it provided the most reproducible results, lower background fluorescence, and
higher fluorescence yield upon STARD1 binding.
All NBD-labeled cholesterol analogues were dissolved in 96% ethanol to get 200-300
µM stock solutions (concentration was measured by absorbance at 470 nm using extinction
coefficient of 21,000 M-1 cm-1).
Obtaining of the S195E mutant of human StAR domain (STARD1, residues 66-285) as
the maltose-binding protein fusion cleavable by 3C protease was described earlier [20,30].
2.2. Steady-state fluorescence spectra
All steady-state fluorescence spectra were recorded on a Cary Eclipse fluorescence
spectrophotometer (Varian Inc.) equipped with a temperature controller (wavelengths are
indicated; the slits width was 5 nm and typically absorbance at the excitation wavelength was
less than 0.1 to exclude the effect of inner filter). Protein samples (0.1-2 µM) were prepared on a
buffer F1 (20 mM Tris-HCl (pH 7.5), 200 mM NaCl, 0.1 mM EDTA) or buffer with Tris
replaced by Hepes for thermal stability measurements using Trp fluorescence.
2.3. Ligand-binding
STARD1S195E samples (1 μM) in buffer F1 were pre-incubated at 37 °C, and then the
intensity of either Trp or NBD fluorescence was recorded before or after each addition of small
0.5-1 µl aliquots of a ligand stock solution in ethanol, so that the final ligand concentration
varied in a range of 0-4 μM and that of ethanol was <1 %. After each addition, samples were
mixed by microsyringe and equilibrated for 5 min at 37 °C prior to measurements (460 nm
excitation; 535 nm emission). To assess apparent binding parameters, the binding curves against
4
total added ligand concentration (x) were fitted in Origin 9.0 using a quadratic equation
tolerating the difference between the total and free ligand concentrations:
;
where Fmax is fluorescence intensity (a.u.) at saturation, N – concentration of binding sites (µM),
KD – apparent dissociation constant (µM), and F(x) – fluorescence intensity (a.u.) at x ligand
concentration.
2.4. Time-resolved fluorescence measurements
Fluorescence decay kinetics were recoded using time- and wavelength-correlated single
photon counting setup based on SPC-150 module and HMP-100-50 detector (Becker&Hickl,
Germany) with excitation by 450 nm 30 ps laser pulses as described in [35].
During the experiment temperature of the sample was controlled by a cuvette holder
Qpod 2e (Quantum Northwest, USA).
2.5. Docking simulations
Docking simulations were performed using either FlexAID [36] or Autodock 4.2 with
AutoDockTools [37] and the STARD1 structural model [20]. The first approach is based on
shape complementarity between ligands and receptors, accounts for protein side chain and ligand
flexibility, but does not take into account electrostatics. Assignment of the binding site was done
using GetCleft tool of the NRGsuite (minimum probe radius – 1.6 Å, hydrogens added, volume
of the binding cavity to accommodate a ligand anchor atom estimated as 450 Å3), and the
residues from the interior of the cholesterol binding cavity were treated as flexible. The
simulations run used the genetic algorithm (1000 chromosomes, 1000 generations, share fitness
model and the rest default parameters). For Autodock, Gasteiger partial charges [38] were
calculated and assigned to atoms. The docking space was defined as a 60×60×60 Å3 box close to
geometric center of the protein. Ligands were created and prepared using HyperChem 7.01
(Hypercube). The Lamarckian genetic algorithm with default parameters was applied for rigid
docking calculations. The binding energy values were calculated automatically by Autodock.
3. Results and discussion
3.1. Interaction of STARD1 with fluorescent cholesterol analogues: the effect of the NBD-group
position
To get insight into the mode and selectivity of the cholesterol binding, we probed
STARD1 with a series of cholesterol analogues containing the fluorescent 7-nitrobenz-2-oxa-
1,3-diazol-4-yl (NBD)-group at either 20th (20NP) [30,39], 22nd (22NC), 25th (25NC) carbon
atoms of the side chain, or at the 3-OH group (3NC) (Fig. 1A). The NBD-group is sensitive to
the polarity of the environment and, being present in some commercially available NBD-
cholesterols, e.g., 22-NBD-cholesterol (=22NC), was successfully used as a reporter in binding
assays [40,41,42] without significantly interfering with binding to STARD1 [23].
After pre-incubation of STARD1S195E with a 1.5-fold ligand excess, we observed
significant differences in the NBD fluorescence, decreasing in the row 20NP > 22NC >> 25NC
> 3NC (Fig. 1B). Surprisingly, the highest intensity of fluorescence was observed in the case of
20NP (Fig. 1B). Given the reported binding of both 20NP and 22NC to STARD1 [30,40,42],
5
this may be associated with differences in the NBD-group environment and its position relative
to tryptophan residues of STARD1. Due to spectral overlap and expected proximity of the NBD-
group and Trp241 in the complex, it is possible that the binding leads to an energy transfer via
FRET (Fig. S1 and [43]) or to a formation of non-fluorescent complexes, reflecting in Trp
quenching. Assuming that four Trp residues of STARD1 roughly equally contribute to its
intrinsic fluorescence intensity, the observed Trp fluorescence quenching ((FSTARD1-
FSTARD1/20NP)/FSTARD1) as the result of 20NP binding indicates that one Trp (¼) could be
quenched completely. Most likely, this is the most proximal Trp241 separated from the NBD-
group of bound 20NP by <12 Å (Fig. S1), confirming specific binding of 20NP in the STARD1
cavity. These changes upon 20NP addition were saturable, yielding ~20 nM apparent KD (Fig.
S1), qualitatively in line with [42], however, the emission of the UV fluorescence maximum of
STARD1 in [42] (352 nm) is ~15 nm red-shifted compared to our results (~338 nm; Fig. S1).
This points to differences in protein preparations and, given the known propensity of STARD1 to
unfolding and aggregation, especially for refolded STARD1 preparations typically used in the
past [30], emphasizes the necessity of hydrodynamic characterization of STARD1 preparation
before ligand binding assays. The specificity of binding within the cavity was additionally
confirmed in the experiment, in which STARD1 was first titrated by 20NP until saturation, and
then gradually outcompeted by cholesterol (Fig. S2).
Although the intensity of fluorescence of bound 20NP was 10 or 30 times higher than
that of 25NC or 3NC, respectively (Fig. 1B), when STARD1S195E was titrated by increasing
amounts of these ligands (Fig. 1C and D) we observed binding curves with saturation at high
concentrations in all three cases. These suggested rather specific interaction occurring at a ~1:1
stoichiometry (as in [23,24,30,41,44]) and allowed us to assess binding affinities. We found that
20NP binds to STARD1 with an apparent KD of 26 ± 8 nM (Fig. 1C), which is very similar to
the earlier reported values for cholesterol (30 nM; one binding site [44]) and 22NC (32 nM; two
binding sites [42]), and is substantially lower than the apparent KD for 3NC (302 ± 29 nM) and
25NC (130 ± 23 nM) (Fig. 1D). This suggests that the NBD-group position dramatically affects
binding to STARD1.
Together, our data indicate that i) all tested ligands are able to bind to STARD1, with
different affinities, ii) the most efficient reporter is 20NP (having affinity comparable with that
of cholesterol), followed by 22NC, iii) 3NC and 25NC are poor reporters of the
STARD1/cholesterol interaction, with very low response and affinity. This may be associated
with the larger sizes of 3NC and 25NC precluding them from being fully accommodated in the
cavity. We cannot exclude that these large ligands can bind less specifically to some areas of
STARD1 outside the cavity.
The experiments were paralleled with in silico docking of cholesterol, 20NP, 25NC, and
3NC into STARD1 structure. Earlier, the opening of the omega 1 (Ω1) loop was suggested to
serve for the lipid binding and exchange [17,20,45,46]. It was concluded that the two opposite
orientations of cholesterol inside the STARD1 cavity, denoted as “IN” or “OUT” depending on
where the 3-OH group is looking [to either R188 (“IN”) or R182 (“OUT”)], are roughly
thermodynamically equivalent [45]. It is not clear, why the “IN” binding mode has become
widely assumed to be the correct one, especially in the absence of direct structural data. Our
attempts to perform docking of the cholesterol molecule into STARD1 using either Autodock
[37] or FlexAID [36] resulted in “C3 OUT” as the more preferential binding mode, and so was
the case if 20NP (or 22NC) were docked (Fig. 2). On contrary, Autodock docking of 25NC and
3NC resulted in only rare solutions with ligands lying inside the cavity, in a random orientation
(not shown), indicative of less favorable binding. FlexAID docking supported our conclusions
for cholesterol and 20NP, however, resulted in 25NC and 3NC binding with the cholesterol core
inside and the NBD-group looking outside the cavity (Fig. 2), in line with the hypothesis that the
low fluorescence response in the case of 25NC and 3NC is associated with their lower affinity
and quenching of their NBD-groups due to the partial exposure. Importantly, different scoring
functions to assess binding poses notwithstanding [36,37], the two algorithms complemented
6
each other in predicting less preferential binding of 25NC and 3NC, but favorable and consistent
binding of cholesterol and 20NP in the “C3 OUT” mode.
Fig. 1. Probing STARD1S195E by fluorescent analogues of cholesterol with different position of
the NBD-group. A. Structural formulae of the ligands studied. B. Fluorescence response of the
ligands (1.5 µM) as a result of binding to STARD1S195E (1 µM). Three spectra for each ligand
were collected, buffer-subtracted and averaged. C, D. Titration of 1 µM STARD1S195E solution
with increasing concentrations of 20NP (C), 25NC and 3NC (D) (see Methods). Vertical dashed
line indicates the 1:1 ratio. The titration experiment was repeated at least five times for each
ligand with the most representative results shown.
It was earlier proposed that R188 forms a salt bridge with E169 and plays a key role in
coordination of the 3-OH group of cholesterol [47,48], whereas R182 located near the Ω1 loop
may take part in regulation of its mobility [45]. Intriguingly, the published data on various
STARD1 mutants with substitutions of R182 and R188 residues, often associated with
pathologies like LCAH, tell in favor of the importance of these residues, but cannot help to
unambiguously distinguish whether the “C3 IN” or “C3 OUT” cholesterol binding mode is the
natively occurring one. Both residues are highly conserved in STARD1 orthologues annotated in
UniProt (14 species), however, R182 residue is more conserved among human START
members, indirectly implying its more universal role in ligand binding than R188. In line with
this, in STARD5, the equivalent position of R188 is occupied by V120 (Fig. S3), despite
cholesterol and bile acids were still proposed to bind to STARD5 in the “C3 IN” orientation
[19,49].
7
Fig. 2. Docking of cholesterol and its fluorescent analogues using Autodock (A, B) and FlexAID
(C-F). For Autodock, only the results for cholesterol and 20NP (sticks and semitransparent
surface, with 3-OH group marked red) docking are shown, 25NC and 3NC tended to locate
outside the cavity or the results were not reproducible. For FlexAID runs, the best poses of
ligands (green sticks) in the STARD1 (blue ribbon) are shown, with the binding volume
(semitransparent red surface), loop Ω1, NBD-group (green arrow), and R182 (salmon) and R188
(yellow) residues determining the orientation “IN” (C3 carbon atom of a ligand looking at R188)
or “OUT” (C3 carbon atom of a ligand looking at R182) marked. Note partial exposure of the
NBD-group in 25NC and 3NC cases.
3.2. The effect of ligand binding on thermal stability of STARD1
Binding of cognate ligands can affect stability of the START family members [17,18,44].
The STARD1S195E sample in the absence or the presence of different ligands and then was
gradually heated up upon measuring the intensity of intrinsic tryptophan fluorescence as a
function of temperature (Fig. S4 and S5). We found that, while T0.5 of STARD1S195E was equal to
49.8 ± 0.1 °C, 20NP increased this value by more than 3.5 degrees (T0.5 = 53.4 ± 0.1 °C),
whereas 3NC caused no effect (T0.5 = 49.5 ± 0.1 °C). Morevover, when added in an excess over
20NP, 3NC significantly decreased the stabilizing effect of 20NP. This supports the idea that
3NC and 20NP compete for the STARD1 binding and suggests that 3NC binding occurs with
lower affinity and fails to stabilize the protein. The presence of cholesterol also increased the
thermal stability of STARD1, although to somewhat lesser extent (from 49.8 ± 0.1 °C to 50.7 ±
0.1 °C).
The T0.5 value for our STARD1 preparation (49.8 ± 0.1 °C in the apoform) obtained
according to [30] is significantly higher than that of STARD1 preparations obtained earlier by
using CD spectroscopy [44] (T0.5 = 42.3 ± 0.1 °C in the apoform), pointing to potential
differences in the quality of preparations.
In the absence of STARD1, fluorescence lifetimes (short τ1 and long τ2) of the solvent
exposed NBD-group in 20NP (τ1 = 225 ps (79.7%); τ2 = 3435 ps (20.3%)) and 3NC (τ1 = 295 ps
(72.2%); τ2 = 2980 ps (27.8%)) were low (Fig. 3A). Upon addition of the STARD1S195E excess,
only the lifetimes for 20NP/STARD1S195E significantly increased (τ1 = 2185 ps (55%); τ2 = 9860
ps (45%)), those for 3NC remained almost unchanged (τ1 = 235 ps (76.1%); τ2 = 4145 ps
8
(23.9%)) (Fig. 3A). This suggests that the NBD-group of 20NP becomes protected upon
interaction with STARD1, whereas 3NC remains almost equally quenched in the free and bound
states, presumably, due to its incomplete accommodation in the cavity, the NBD-group exposure
and low fluorescence response.
Fig. 3. Time-resolved fluorescence spectroscopy study of the STARD1S195E interaction with
20NP and 3NC. A. The NBD fluorescence decay kinetics at 25 °C for 20NP and 3NC ligands
(1.4 µM) in the absence and the presence of STARD1S195E (2 µM). B. Hypothetical models of the
STARD1S195E complexes with 20NP and 3NC predicted by FlexAID docking. Ligands are
shown in cyan, Trp241 of STARD1 (green) is shown in magenta, loop Ω1 is orange. C, D. Color
maps demonstrating STARD1 thermal stability in the presence of 3NC (C) or 20NP (D) as
studied by changes in the fluorescence decay kinetics of the NBD-group during the heating of
the samples (25→70 °C; 1 °C/min).
Our conclusions were further confirmed by continuous time-resolved fluorescence decay
kinetics measurements upon heating of the STARD1 mixtures with either 20NP or 3NC. This
allowed to visualize changes in the lifetimes of the NBD fluorescence and additionally analyze
the thermal stability of STARD1 via unfolding-induced dissociation of NBD-ligands (Fig. 3). In
agreement with Fig. 3A and S5, at temperatures below 50 °C, the fluorescence lifetime was
dramatically larger in the case of 20NP, and gradually decreased upon heating. The sharp
decrease in its fluorescence lifetime was observed from 50 to 60 °C, reflecting the thermal
unfolding of STARD1 and the concomitant ligand dissociation (Fig. 3D). T0.5 of ~54 °C was
consistent with the results of steady-state fluorescence (Fig. S5), whereas no significant changes
in the lifetimes of 3NC were observed (Fig. 3С). This supports the hypothesis that 3NC binds to
STARD1 with lower affinity and exposed fluorescent group, and that this binding mode could
not stabilize the protein against the heat-induced unfolding.
Possible 20NP and 3NC binding modes to native STARD1 are shown in Fig. 3B. These
models, predicted by FlexAID docking and consistent with our experimental data, suggest the
“C3 OUT” as the more preferential binding mode for these fluorescence analogues (Fig. 3B).
To sum up, probing of the STARD1 cholesterol-binding cavity using the series of NBD-
labeled cholesterol analogues allowed i) to expand our knowledge about the STARD1 ligand
specificity, ii) to find an alternative efficient NBD-steroid for screening of novel STARD1
ligands (20NP), iii) to validate its binding affinity and stoichiometry (1:1) and dissect its effect
on thermal stability of STARD1, and iv) to specify the most probable orientation of the
9
cholesterol core in the cavity. The applicability of powerful time-resolved fluorescence lifetime
spectroscopy approaches to study STARD1/ligand interactions and thermal stability of the
corresponding complexes is also demonstrated.
4. Acknowledgements
This investigation was supported by Russian Science Foundation (grant 17-74-10053 to
N.N.S.). Ya.V.F. and V.M.S. acknowledge the grant from the Belarusian State Program of
Scientific Investigations (№ 20161380).
5. References
[1] D.M. Stocco, StAR protein and the regulation of steroid hormone biosynthesis, Annu Rev
Physiol 63 (2001) 193-213.
[2] B.J. Clark, J. Wells, S.R. King, D.M. Stocco, The purification, cloning, and expression of a
novel luteinizing hormone-induced mitochondrial protein in MA-10 mouse Leydig tumor
cells. Characterization of the steroidogenic acute regulatory protein (StAR), J Biol Chem
269 (1994) 28314-28322.
[3] D. Lin, T. Sugawara, J.F. Strauss, 3rd, B.J. Clark, D.M. Stocco, P. Saenger, A. Rogol, W.L.
Miller, Role of steroidogenic acute regulatory protein in adrenal and gonadal
steroidogenesis, Science 267 (1995) 1828-1831.
[4] C.B. Kallen, J.T. Billheimer, S.A. Summers, S.E. Stayrook, M. Lewis, J.F. Strauss, 3rd,
Steroidogenic acute regulatory protein (StAR) is a sterol transfer protein, J Biol Chem
273 (1998) 26285-26288.
[5] C. Bizzarri, E. Pisaneschi, M. Mucciolo, S. Pedicelli, D. Galeazzi, A. Novelli, M. Cappa,
Lipoid congenital adrenal hyperplasia by steroidogenic acute regulatory protein (STAR)
gene mutation in an Italian infant: an uncommon cause of adrenal insufficiency, Ital J
Pediatr 43 (2017) 57.
[6] W.L. Miller, Disorders in the initial steps of steroid hormone synthesis, J Steroid Biochem
Mol Biol 165 (2017) 18-37.
[7] F. Arakane, T. Sugawara, H. Nishino, Z. Liu, J.A. Holt, D. Pain, D.M. Stocco, W.L. Miller,
J.F. Strauss, 3rd, Steroidogenic acute regulatory protein (StAR) retains activity in the
absence of its mitochondrial import sequence: implications for the mechanism of StAR
action, Proc Natl Acad Sci U S A 93 (1996) 13731-13736.
[8] T. Yamazaki, C. Matsuoka, M. Gendou, S. Izumi, D. Zhao, I. Artemenko, C.R. Jefcoate, S.
Kominami, Mitochondrial processing of bovine adrenal steroidogenic acute regulatory
protein, Biochim Biophys Acta 1764 (2006) 1561-1567.
[9] I.P. Artemenko, D. Zhao, D.B. Hales, K.H. Hales, C.R. Jefcoate, Mitochondrial processing of
newly synthesized steroidogenic acute regulatory protein (StAR), but not total StAR,
mediates cholesterol transfer to cytochrome P450 side chain cleavage enzyme in adrenal
cells, J Biol Chem 276 (2001) 46583-46596.
[10] F. Arakane, S.R. King, Y. Du, C.B. Kallen, L.P. Walsh, H. Watari, D.M. Stocco, J.F.
Strauss, 3rd, Phosphorylation of steroidogenic acute regulatory protein (StAR) modulates
its steroidogenic activity, J Biol Chem 272 (1997) 32656-32662.
[11] J. Lee, T. Tong, H. Duan, Y.H. Foong, I. Musaitif, T. Yamazaki, C. Jefcoate, Regulation of
StAR by the N-terminal Domain and Coinduction of SIK1 and TIS11b/Znf36l1 in Single
Cells, Front Endocrinol (Lausanne) 7 (2016) 107.
[12] V. Papadopoulos, Y. Aghazadeh, J. Fan, E. Campioli, B. Zirkin, A. Midzak, Translocator
protein-mediated pharmacology of cholesterol transport and steroidogenesis, Mol Cell
Endocrinol 408 (2015) 90-98.
[13] F. Alpy, C. Tomasetto, Give lipids a START: the StAR-related lipid transfer (START)
domain in mammals, J Cell Sci 118 (2005) 2791-2801.
10
[14] D. Letourneau, A. Lorin, A. Lefebvre, V. Frappier, F. Gaudreault, R. Najmanovich, P.
Lavigne, J.G. LeHoux, StAR-related lipid transfer domain protein 5 binds primary bile
acids, J Lipid Res 53 (2012) 2677-2689.
[15] S.L. Roderick, W.W. Chan, D.S. Agate, L.R. Olsen, M.W. Vetting, K.R. Rajashankar, D.E.
Cohen, Structure of human phosphatidylcholine transfer protein in complex with its
ligand, Nat Struct Biol 9 (2002) 507-511.
[16] N. Kudo, K. Kumagai, R. Matsubara, S. Kobayashi, K. Hanada, S. Wakatsuki, R. Kato,
Crystal structures of the CERT START domain with inhibitors provide insights into the
mechanism of ceramide transfer, J Mol Biol 396 (2010) 245-251.
[17] D. Letourneau, M. Bedard, J. Cabana, A. Lefebvre, J.G. LeHoux, P. Lavigne, STARD6 on
steroids: solution structure, multiple timescale backbone dynamics and ligand binding
mechanism, Sci Rep 6 (2016) 28486.
[18] D. Letourneau, A. Lefebvre, P. Lavigne, J.G. LeHoux, The binding site specificity of
STARD4 subfamily: Breaking the cholesterol paradigm, Mol Cell Endocrinol 408 (2015)
53-61.
[19] A.G. Thorsell, W.H. Lee, C. Persson, M.I. Siponen, M. Nilsson, R.D. Busam, T.
Kotenyova, H. Schuler, L. Lehtio, Comparative structural analysis of lipid binding
START domains, PLoS One 6 (2011) e19521.
[20] N.N. Sluchanko, K.V. Tugaeva, E.G. Maksimov, Solution structure of human steroidogenic
acute regulatory protein STARD1 studied by small-angle X-ray scattering, Biochem
Biophys Res Commun 489 (2017) 445-450.
[21] W. Korytowski, A. Pilat, J.C. Schmitt, A.W. Girotti, Deleterious cholesterol hydroperoxide
trafficking in steroidogenic acute regulatory (StAR) protein-expressing MA-10 Leydig
cells: implications for oxidative stress-impaired steroidogenesis, J Biol Chem 288 (2013)
11509-11519.
[22] W. Korytowski, K. Wawak, P. Pabisz, J.C. Schmitt, A.C. Chadwick, D. Sahoo, A.W.
Girotti, Impairment of Macrophage Cholesterol Efflux by Cholesterol Hydroperoxide
Trafficking: Implications for Atherogenesis Under Oxidative Stress, Arterioscler Thromb
Vasc Biol 35 (2015) 2104-2113.
[23] B.Y. Baker, R.F. Epand, R.M. Epand, W.L. Miller, Cholesterol binding does not predict
activity of the steroidogenic acute regulatory protein, StAR, J Biol Chem 282 (2007)
10223-10232.
[24] Y. Tsujishita, J.H. Hurley, Structure and lipid transport mechanism of a StAR-related
domain, Nat Struct Biol 7 (2000) 408-414.
[25] D. Wustner, Fluorescent sterols as tools in membrane biophysics and cell biology, Chem
Phys Lipids 146 (2007) 1-25.
[26] E. Sezgin, F.B. Can, F. Schneider, M.P. Clausen, S. Galiani, T.A. Stanly, D. Waithe, A.
Colaco, A. Honigmann, D. Wustner, F. Platt, C. Eggeling, A comparative study on
fluorescent cholesterol analogs as versatile cellular reporters, J Lipid Res 57 (2016) 299-
309.
[27] Y.V. Faletrov, N.S. Frolova, H.V. Hlushko, E.V. Rudaya, I.P. Edimecheva, S.
Mauersberger, V.M. Shkumatov, Evaluation of the fluorescent probes Nile Red and 25-
NBD-cholesterol as substrates for steroid-converting oxidoreductases using pure
enzymes and microorganisms, FEBS J 280 (2013) 3109-3119.
[28] Y.V. Faletrov, K.I. Bialevich, I.P. Edimecheva, D.G. Kostsin, E.V. Rudaya, E.I.
Slobozhanina, V.M. Shkumatov, 22-NBD-cholesterol as a novel fluorescent substrate for
cholesterol-converting oxidoreductases, J Steroid Biochem Mol Biol 134 (2013) 59-66.
[29] H.S. Bose, M.A. Baldwin, W.L. Miller, Incorrect folding of steroidogenic acute regulatory
protein (StAR) in congenital lipoid adrenal hyperplasia, Biochemistry 37 (1998) 9768-
9775.
[30] N.N. Sluchanko, K.V. Tugaeva, Y.V. Faletrov, D.I. Levitsky, High-yield soluble
expression, purification and characterization of human steroidogenic acute regulatory
11
protein (StAR) fused to a cleavable Maltose-Binding Protein (MBP), Protein Expr Purif
119 (2016) 27-35.
[31] Y. Faletrov, A. Brzostek, R. Plocinska, J. Dziadek, E. Rudaya, I. Edimecheva, V.
Shkumatov, Uptake and metabolism of fluorescent steroids by mycobacterial cells,
Steroids 117 (2017) 29-37.
[32] M. Adamczyk, J. Grote, A straightforward synthesis of 3 alpha- and 3 beta-
aminodigoxigenin, Steroids 61 (1996) 562-564.
[33] I.F. Craig, D.P. Via, W.W. Mantulin, H.J. Pownall, A.M. Gotto, Jr., L.C. Smith, Low
density lipoproteins reconstituted with steroids containing the nitrobenzoxadiazole
fluorophore, J Lipid Res 22 (1981) 687-696.
[34] E.O. Porta, P.B. Carvalho, M.A. Avery, B.L. Tekwani, G.R. Labadie, Click chemistry
decoration of amino sterols as promising strategy to developed new leishmanicidal drugs,
Steroids 79 (2014) 28-36.
[35] E.G. Maksimov, N.N. Sluchanko, Y.B. Slonimskiy, E.A. Slutskaya, A.V. Stepanov, A.M.
Argentova-Stevens, E.A. Shirshin, G.V. Tsoraev, K.E. Klementiev, O.V. Slatinskaya,
E.P. Lukashev, T. Friedrich, V.Z. Paschenko, A.B. Rubin, The photocycle of orange
carotenoid protein conceals distinct intermediates and asynchronous changes in the
carotenoid and protein components, Sci Rep 7 (2017) 15548.
[36] F. Gaudreault, L.P. Morency, R.J. Najmanovich, NRGsuite: a PyMOL plugin to perform
docking simulations in real time using FlexAID, Bioinformatics 31 (2015) 3856-3858.
[37] G.M. Morris, R. Huey, W. Lindstrom, M.F. Sanner, R.K. Belew, D.S. Goodsell, A.J. Olson,
AutoDock4 and AutoDockTools4: Automated docking with selective receptor flexibility,
J Comput Chem 30 (2009) 2785-2791.
[38] J. Gasteiger, M. Marsili, Iterative partial equalization of orbital electronegativity—a rapid
access to atomic charges, Tetrahedron 36 (1980) 3219-3228.
[39] Y.V. Faletrov, E.V. Rudaya, H.V. Hlushko, I.P. Edimecheva, V.M. Shkumatov, Revealing
of novel affine interactions of NBD-labeled fluorescent steroids with bacterial steroid–
converting oxidoreductases using molecular docking, The 39th FEBS Congress Paris,
France, 2014, pp. 637.
[40] J. Reitz, K. Gehrig-Burger, J.F. Strauss, 3rd, G. Gimpl, Cholesterol interaction with the
related steroidogenic acute regulatory lipid-transfer (START) domains of StAR
(STARD1) and MLN64 (STARD3), FEBS J 275 (2008) 1790-1802.
[41] B.Y. Baker, D.C. Yaworsky, W.L. Miller, A pH-dependent molten globule transition is
required for activity of the steroidogenic acute regulatory protein, StAR, J Biol Chem 280
(2005) 41753-41760.
[42] A.D. Petrescu, A.M. Gallegos, Y. Okamura, J.F. Strauss, 3rd, F. Schroeder, Steroidogenic
acute regulatory protein binds cholesterol and modulates mitochondrial membrane sterol
domain dynamics, J Biol Chem 276 (2001) 36970-36982.
[43] R.P. Rambo, J.A. Tainer, Accurate assessment of mass, models and resolution by small-
angle scattering, Nature 496 (2013) 477-481.
[44] A. Roostaee, E. Barbar, J.G. Lehoux, P. Lavigne, Cholesterol binding is a prerequisite for
the activity of the steroidogenic acute regulatory protein (StAR), Biochem J 412 (2008)
553-562.
[45] M. Murcia, J.D. Faraldo-Gomez, F.R. Maxfield, B. Roux, Modeling the structure of the
StART domains of MLN64 and StAR proteins in complex with cholesterol, J Lipid Res
47 (2006) 2614-2630.
[46] D.B. Iaea, I. Dikiy, I. Kiburu, D. Eliezer, F.R. Maxfield, STARD4 Membrane Interactions
and Sterol Binding, Biochemistry 54 (2015) 4623-4636.
[47] A. Roostaee, E. Barbar, P. Lavigne, J.G. LeHoux, The mechanism of specific binding of
free cholesterol by the steroidogenic acute regulatory protein: evidence for a role of the
C-terminal alpha-helix in the gating of the binding site, Biosci Rep 29 (2009) 89-101.
12
[48] A.P. Mathieu, A. Fleury, L. Ducharme, P. Lavigne, J.G. LeHoux, Insights into steroidogenic
acute regulatory protein (StAR)-dependent cholesterol transfer in mitochondria: evidence
from molecular modeling and structure-based thermodynamics supporting the existence
of partially unfolded states of StAR, J Mol Endocrinol 29 (2002) 327-345.
[49] D. Letourneau, A. Lefebvre, P. Lavigne, J.G. LeHoux, STARD5 specific ligand binding:
comparison with STARD1 and STARD4 subfamilies, Mol Cell Endocrinol 371 (2013)
20-25.
1
SUPPLEMENTARY MATERIAL
Effect of the NBD-group position on interaction of fluorescently-labeled cholesterol
analogues with human steroidogenic acute regulatory protein STARD1
Kristina V. Tugaeva1,2, Yaroslav V. Faletrov3, Elvin S. Allakhverdiev4, Vladimir M.
Shkumatov3,5, Eugene G. Maksimov6, Nikolai N. Sluchanko1,6*
1A.N. Bach Institute of Biochemistry, Federal Research Center of Biotechnology of the Russian
Academy of Sciences, 119071 Moscow, Russia
2Department of Biochemistry, School of Biology, M.V. Lomonosov Moscow State University,
119991 Moscow, Russia
3Research Institute for Physical Chemical Problems, Belarusian State University, Minsk, Belarus
4Faculty of Fundamental Medicine, M.V. Lomonosov Moscow State University, 119991
Moscow, Russia
5Department of Chemistry, Belarusian State University, Minsk, Belarus
6Department of Biophysics, School of Biology, M.V. Lomonosov Moscow State University,
119992 Moscow, Russia
Fig. S1. Quenching of STARD1 tryptophans upon binding of 20NP. A. Spectral changes upon
addition of 20NP (indicated by arrows). Excitation at 297 nm, slits width 5 nm, temperature 37
°C. Maximum of Trp fluorescence is indicated by dashed line. B. Titration of STARD1S195E
solution (1 µM) by increasing concentrations of 20NP followed by either quenching of
tryptophan fluorescence (black) or increase in the intensity of NBD fluorescence (red). The
vertical dashed line corresponds to the 1:1 ratio. C. Docking results in Autodock showing the
proximity of Trp241 and the NBD-group of the bound 20NP and indicating possibility of
interaction (either FRET or static quenching). D. Spectral overlap in the fluorophore system
involved.
2
Fig. S2. Competitive replacement of STARD1-bound 20NP by cholesterol. STARD1S195E (1
µM) was first titrated by 20NP until saturation (2 µM) (squares) and then by increasing amounts
of cholesterol (circles). Fluorescence was excited at 460 nm. Temperature was 37 °C. Note that
concentrations of cholesterol were significally higher than its critical micelle concentration (tens
nM), therefore approximation of the curve and assessment of binding parameters were not
possible.
Fig. S3. Alignment and superposition of STARD1 and STARD5 structures exemplifying lesser
conservativity of the R188 position. The primary structure was aligned using Clustal Omega, the
level of homology is shown as greyscale. The crystal structures of STARD1 (PDB entry 3P0L)
and STARD5 (PDB entry 2R55) were superimposed and drawn in PyMol 1.69 (sagittal plane),
with the main features highlighted.
3
Fig. S4. Thermally-induced changes in the intrinsic tryptophan fluorescence of STARD1S195E.
The sample was heated at a constant rate of 1 °C/min and changes in the intensity of
fluorescence excited at 297 nm were registered at 346 nm in a range of 10-80 °C. A. Raw data
with the three regions corresponding to the (F)olded, (T)ransition, and the (U)nfolded state,
highlighted. Arrows indicate the direction of heating. B. The same data converted to a form of
temperature dependence of the so-called completeness of unfolding allowing to determine the
corresponding half-transition temperature (T0.5).
Note: a typical thermal unfolding curve could be divided into three regions corresponding to the
folded state (F), transition (T), and unfolded state (U) (Fig. S4A). Linear approximation of the
first and the third regions enabled building dependencies of completeness of transition on
temperature (Fig. S4B), which in turn allowed estimation of half-transition temperatures (T0.5)
(Fig. S5).
4
Fig. S5. Thermal stability of STARD1S195E in the absence or the presence of ligands followed by
changes in intrinsic tryptophan fluorescence of STARD1 (297 nm excitation, 346 nm emission).
A. The temperature dependences of completeness of unfolding for the samples as indicated. The
expeiment was done four times and the most typical results are presented. B. The T0.5 values for
these samples determined from the panel A. See Fig. S4 for more details.
