Rational Design of Vitamin D3 Analogues Which Selectively Restore Activity to a Vitamin D Receptor Mutant Associated with Rickets

Article abstract of DOI:10.1021/ol0266931

(Matrix Presented) Vitamin D₃-resistant rickets (VDRR) is associated with mutations to the Vitamin D receptor (VDR) which effect ligand-dependent transactivation. Some VDRR associated mutants directly disrupt ligand binding. Using the reported

Full text of DOI:10.1021/ol0266931

ORGANIC
LETTERS
2002
Vol. 4, No. 22
3863-3866
Rational Design of Vitamin D
3
Analogues Which Selectively Restore
Activity to a Vitamin D Receptor Mutant
Associated with Rickets
Steve L. Swann,‡ Joel J. Bergh,‡ M. Cindy Farach-Carson, and John T. Koh*
Department of Chemistry and Biochemistry, UniVersity of Delaware,
Newark, Delaware 19716
johnkoh@udel.edu
Received August 7, 2002
ABSTRACT
Vitamin D -resistant rickets (VDRR) is associated with mutations to the Vitamin D receptor (VDR) which effect ligand-dependent transactivation.
3
Some VDRR associated mutants directly disrupt ligand binding. Using the reported VDR-1,25-dihydroxy vitamin D (1,25(OH)2D ) cocrystal
3
3
structure, three 1,25(OH)2D analogues were designed to uniquely complement the rickets associated mutant VDR(Arg274fLeu). The three
3
analogues were 17 to 286 times more potent than 1,25(OH)2D with the mutant in cell-based assays and did not substantially activate cellular
3
calcium influx.
A few recent examples have shown that small molecules
(MW < 600) can restore activity to mutationally impaired
proteins that are associated with genetic disease. For example,
compounds have been discovered which can help stabilize
mutant forms of p53 associated with cancer,1 or recover
activity to mutated forms of nuclear hormone receptors
associated with refractory forms of prostate cancer,2 resis-
tance to thyroid hormone (RTH),3 and type II rickets.4 These
examples illustrate that small molecules may be used to
restore activity to at least a subset of genetic mutations and
suggest a potentially new pharmacological approach to the
treatment of genetic disease. Thus far, mutant-complementing
molecules have almost exclusively been discovered by
screening existing compounds and compound libraries. In
this study, we evaluate the ability of structure-based design
to custom design hormone analogues for a vitamin D receptor
(VDR) mutation associated with vitamin D resistant rickets
(VDRR).
The nuclear and steroid hormone receptors (NHR’s)
comprise a superfamily of ligand-dependent transcriptional
regulators that control the expression of specific eukaryotic
genes involved in development and homeostasis.5,6 The
NHRs bind to specific DNA sequences (response elements)
‡ Department of Biological Sciences, University of Delaware, Newark,
DE 19716.
(1) Foster, B. A.; Coffey, H. A.; Mornin, M. J.; Rastinejad, F. Science
1999, 286, 2507-2510.
(2) Taplin, M.-E.; Bubley, G. J.; Ko, Y.-L.; Small, E.; Upton, M.;
Rajeshkumar, B.; Balk, S. P. Cancer Res. 1999, 59, 2511-2515.
(3) Takeda, T.; Suzuki, S.; Lui, R. T.; DeGroot, L. J. J. Clin. Endocrinol.
Metab. 1995, 80, 2033-2040.
(4) Gardezi, S. A.; Nguyen, C.; Malloy, P. J.; Posner, G. H.; Feldman,
D.; Peleg, S. J. Biol. Chem. 2001, 276, 29148-29156.
(5) Aranda, A.; Pascual, A. Phys. ReV. 2001, 81, 1269-1304.
(6) Whitfield, K.; Jurutka, P. W.; Haussler, C. A.; Haussler, M. R. J.
Cell Biochem. 1999, 32/33, 110-122.
10.1021/ol0266931 CCC: $22.00 © 2002 American Chemical Society
Published on Web 09/28/2002

that are part of the promoters of hormone responsive genes
and activate gene expression in a ligand-dependent manner.
A diverse class of human genetic diseases is associated
with mutations to NHRs, including androgen insensitivity
syndrome (androgen receptor), resistance to thyroid hormone
(RTH, thyroid hormone receptor), diabetes mellitus (peroxi-
some proliferator activated receptor), rickets (VDR), Cush-
ings disease (glucocorticoid receptor), and certain forms of
prostate and breast cancer (androgen and estrogen recep-
tors).7,8 Many of these disease-associated mutations involve
amino acids that lie in and around the hormone binding
pocket and directly affect hormone binding and hormone-
dependent transactivation function.
Recent work from our lab has demonstrated that rational
molecular design can be used to generate a potent and
selective thyroid hormone analogue, which selectively
activates an RTH-associated mutant of TRâ over its R
subtype (TRR), which is associated with cardiac function.9
Efforts to complement other RTH-associated mutations of
TR mutations are currently ongoing. In this study we explore
the generality of our structure-based design approach to
complement mutations in a related NHR family member.
In response to 1,25-dihydroxy vitamin D3 (1,25(OH)2D3),
the nuclear vitamin D receptor (VDR) regulates genes
associated with phosphate and calcium homeostasis and bone
maintenance.10,11 Several studies have also demonstrated the
presence of a second membrane-associated 1,25(OH)2D3
receptor, which is responsible for rapid hormone-dependent
responses including stimulating cellular calcium influx and
activation of PKC.12
nuclear receptor mediated 1,25(OH)2D3 responsive pathways.
We envisioned that the recently reported VDR-1,25(OH)2D3
cocrystal structure could provide a basis to rationally design
1,25(OH)2D3 analogues capable of selectively complementing
rickets-associated mutations such as VDR(R274L).14
Very recently, Gardezi et al. have shown that JK-1626-
2, a known 1,25(OH)2D3 analogue, partially restores activity
to VDR(R274L) with 5% of the potency of 1,25(OH)2D3
with VDR(wt).4 This analogue was previously identified as
a potent stimulator of PKC, a membrane receptor associated
response of 1,25(OH)2D3.15
In this study we explore the use of structure-based design
to rationally design 1,25(OH)2D3 analogues which are potent
and selective activators of the rickets-associated mutant
VDR(R274L). To provide insight into the binding mode of
1,25(OH)2D3 in the rickets-associated mutant, a computa-
tional site model representing the VDR(R274L) binding
pocket was constructed in FLO/QXP,16 based on the pub-
lished X-ray coordinates of Moras et al. (Figure 1).14 The
Human Vitamin D-Resistant Rickets (VDRR) is associated
with mutations to the nuclear VDR, which result in high
serum 1,25(OH)2D3 levels and severe bone underdevelop-
ment and alopecia. Presently, over 20 VDR mutations
associated with VDRR have been reported.13 Two of these
mutations, Arg274fLeu and His305fGln, involve residues
that directly contact 1,25(OH)2D3 in the normal ligand-
receptor complex. The Arg274fLeu mutation results in a
net loss of a hydrogen bond to the 1R-hydroxyl of 1,25-
(OH)2D3 and a greater than 1000-fold decrease in hormone-
responsive transactivation compared to “wild-type” receptor,
VDR(wt) (EC50VDR(wt) ) 2nM, EC50VDR(R274L) )
>2000 nM). Patients with weakly refractory forms of
hVDRR are sometimes treatable with supraphysiological
doses of 1,25(OH)2D3; however, very high doses of 1,25-
(OH)2D3 cannot be tolerated clinically, presumably because
they will disrupt the normal homeostatic balance of non-
Figure 1. Comparison of the modeled structure of 1,25(OH)2D3
in VDR(R274L) with the wild-type ligand-receptor complex. (A)
Structure of VDR(wt) with 1,25(OH)2D3 based on cocrystal
structure. (B) Modeled structure of VDR(R274L) with 1,25(OH)2D3.
Arg274fLeu mutation results in the loss of a critical
hydrogen bond that normally exists between the 1R-OH and
the guanidine of Arg274 and opens a substantial hydrophobic
“hole” within the ligand-receptor interface, adjacent to the
position of the 1R-hydroxyl (Figure 1B).
The VDR(R274L) model suggests that hydrophobic func-
tionality in the vicinity of the 1R-OH of 1,25(OH)2D3 may
serve to replace the lost ligand-receptor hydrogen bond with
a favorable “hydrophobic bond”. Twenty candidate com-
pounds were identified as having hydrophobic functionality
of approximately the same size as the hole created by the
ArgfLeu substitution and which could be easily derived
(7) Tenbaum, S.; Baniahmad, A. Int. J. Biochem. Cell Biol. 1997, 29,
1325-1341.
(8) Maslen, G. L.; Williams, T. D.; Lewis, H.; Schafer, A. J.; Chatterjee,
V. K.; O’Rahilly, S. Nature 1999, 402, 880-882.
(9) Ye, H. F.; O’Reilly, K. E.; Koh, J. T. J. Am. Chem. Soc. 2001, 123,
1521-1522.
(10) Spiess, Y. H.; Price, P. A.; Deftos, J. L.; Manolagas, S. C.
Endcorinology 1986, 18, 1340-1346.
(11) Chen, T. C.; Castillo, L.; Korycka-Dahl, M.; Deluca, H, F. J. Nutr.
1974, 104, 1056-1060.
(14) Rochel, N.; Wurtz, J. M.; Mitschler, A.; Klaholz, B.; Moras, D.
Mol. Cell 2000, 5, 173-179.
(15) Greising, D. M.; Schwartz, Z.; Posner, G. H.; Sylvia, V. L.; Dean,
D. D.; Boyan, B. D. J. Cell Physiol. 1996, 171, 357-367.
(16) McMartin, C.; Bohacek, R. J. J. Comput.-Aided Mol. Des. 1997,
11, 333-344.
(12) Farach-Carson, M. C.; Ridall, A. L. Am. J. Kidney Dis. 1998, 31,
729-742.
(13) Malloy, P. J.; Pike, J. W.; Feldman, D. Endocr. ReV. 1999, 20, 156-
188.
3864
Org. Lett., Vol. 4, No. 22, 2002

Table 1.
1,25D3/
WT
1,25D3/
R274L
SS-I/
SS-II/
SS-III/
R274L
SS-IV/
R274L
R274L
R274L
EC50 (nM)
Eass (kcal/mol)
max activity (RLU)
2.0 ((2.2)
-8.8
100
>2000
-7.3
N/A
N/A
-7.62
23
116 ((6.7)
-8.82
85
6.8 ((0.2)
-9.51
75
24.2 ((1.2)
-10.1
64
from commercially available reagents. The candidate com-
pounds were subjected to comprehensive Monte Carlo
docking simulations and the best three analogues, SS-II, SS-
III, and SS-IV, were selected on the basis of their calculated
apparent association energies for the VDR(R274L) site model
(Eass). It is important to note that the calculations are not
expected to represent the complex series of molecular events
that represent ligand-dependent transcription; however, prior
studies have suggested that similar hormone receptor models
can be used as a reasonable guide to select compounds having
a high propensity to be potent agonists.17 The compound SS-I
had a significantly lower (less negative) association energy
presumably because the proposed A-ring substitution was
too large for the binding pocket. This compound was also
synthesized and evaluated as a negative control.
ated actions. Because small substitutions at the 1-R position
were previously shown to effect nuclear versus membrane
selectivity,15 we reasoned that the rather gross 1-R substitu-
tions represented by SS-II, SS-III, and SS-IV might impart
Scheme 1. Synthesis of D3 Analogsa
Our synthesis of the designed analogues was based on
compound 1, which was obtained from ergocalciferol fol-
lowing the six-step synthesis described by Hesse et al.18 TMS
protection of the 1R-hydroxyl allowed for its selective
deprotection after installation of the biologically active C22-
C25 TES-protected side chain. This allowed for the selective
alkylation of 1R-hydroxyl with reagents selected by molec-
ular design (Scheme 1).
The four analogues, SS-I to SS-IV, were evaluated for their
ability to elicit a transactivation response in HEK293 cells
transiently transfected with the mutant expression plasmid
pSG5VDR(R274L) luciferase reporter (VDRE-luc), and
control (pRLbasic, Promega) (Figure 2). The three designed
analogues, SS-II, SS-III, and SS-IV, all show substantial
activity in the Arg274fLeu mutant below 1 µM and are 17
to 286 times more potent than 1,25(OH)2D3 (EC50 ) >2000
nM). The analogues were able to restore 60 to 80% of the
maximum inducible activity and 8 to 30% of the potency of
1,25(OH)2D3 with VDR(wt). The compound SS-I, designed
as a negative control, has no detectable activity in the mutant
below 2000 nM.
While the analogues SS-II, SS-III, and SS-IV are potent
agonists for the mutant VDR(R274L), their utility as rickets-
complementing analogues might be limited if they also
strongly activated the rapid nonnuclear receptor mediated
responses of 1,25(OH)2D3 and perturbed normal calcium
homeostasis. The absence of a structure for the membrane-
associated receptor precludes a structure-based approach to
design ligands to be exclusive of membrane receptor associ-
a Key: (a) TMS-imidazole, THF, rt 2 h; hν, PhMe, Et3N, rt. (b)
(TES(O)C(CH3)2(CH2)2Mg)2Cu, THF/ET2O, rt; KOH, EtOH, O °C.
(c) NaH, DMF, PH(CH2)3Br; TBAF, THF, 40 °C. (d) NaH, DMF,
BnBr; TBAF, THF, 40 °C. (e) NaH, DMF, (TMSO)CH2C6H4-o-
(CH2)Cl; TBAF, THF, 40 °C. (f) NaH, DMF, m-CN-BnBr; TBAF,
THF, 40 °C.
(17) (a) Shi, Y.; Koh, J. T. Chem. Biol. 2001, 8, 501-510. (b) Swann S.
L.; Bergh, J.; Farach-Carson, M. C.; Koh, J. T. J. Am. Chem. Soc. Accepted
for publication.
(18) Andrews, D. R.; Barton, D. H. R.; Hesse, R. H.; Pechet, M. M. J.
Org. Chem. 1986, 51, 4819-4828.
Org. Lett., Vol. 4, No. 22, 2002
3865

Figure 3. Ligand-dependent stimulation of 45Ca2+ influx of
MC3T3-E1 pre-osteoblastic cells under 132 mM K+ with 20 nM
ligand.
Figure 2. Transcription response of HEK293 cells transiently
transfected with VDR-responsive luciferase promoter. Transacti-
vation activity was determined by dual luciferase assay (Promega)
and is reported as relative light units (RLU) normalized to the
maximum inducible activity of 1,25(OH)2 D3 in VDR(wt) having
100 RLU.
This work adds to a growing number of examples of small
molecules which can restore function to mutationally im-
paired proteins associated with genetic disease. Clearly only
a subset of mutations associated with genetic disease are
amenable to small molecule molecular complementation. The
rapid identification of mutant-compensating compounds of
recoverable mutants therefore presents a new and unique
challenge to the chemical community. The high-resolution
structures of disease-associated proteins enable us to use
structure-based design to rapidly identify mutant-compensat-
ing leads from existing ligands which may ultimately provide
a new approach to the treatment of genetic disease.
substantial selectivity against membrane-associated actions
of 1,25(OH)2D3 membrane-associated receptor. The designed
analogues were assayed for their ability to stimulate Ca2+
influx via Voltage-Sensitive Calcium Channels (VSCC’s)
through the membrane-associated receptor using a radio-
calcium influx assay under depolarizing conditions (high
external K+) (Figure 3).19 All of the designed mutant-specific
1,25(OH)2D3 analogues show significantly reduced activity
toward ligand-mediated calcium influx compared to 1,25-
(OH)2D3 itself, suggesting that the designed mutant-selective
analogues are capable of restoring function to the mutant
nuclear receptor VDR(R274L) without substantially affecting
the membrane-associated actions of 1,25(OH)2D3.
Acknowledgment. pSG5hVDR was kindly provided by
M. R. Haussler. We thank the National Institutes of Health
(Grant No. RO1DK54257) for financial support.
Supporting Information Available: Synthesis and char-
acterization of intermediates and analogues. This material
is available free of charge via the Internet at http://pubs.acs.org.
(19) Meszaros, J. G.; Karin, N. J.; Farach-Carson, M. C. Connect. Tissue
Res. 1996, 35, 107-111.
OL0266931
3866
Org. Lett., Vol. 4, No. 22, 2002