Vitamin D3 analogues that contain modified A- and seco-B-rings as hedgehog pathway inhibitors

Article abstract of DOI:10.1016/j.ejmech.2015.01.049

The hedgehog (Hh) signaling pathway is a developmental signaling pathway that has been implicated as a target for anti-cancer drug development in a variety of human malignancies. Several natural and synthetic vitamin D-based seco-steroids have been identified as potent inhibitors of Hh signaling with chemotherapeutic potential. These include the previously characterized analogue 4, which contains the northern CD-ring/side chain region of vitamin D3 (VD3) linked to an aromatic A-ring mimic through an ester bond. To further explore structure-activity relationships for this class of VD3-based Hh pathway inhibitors, we have designed, synthesized and evaluated several series of compounds that modify the length, composition, and stereochemical orientation of the ester linker. These studies have identified compounds 54 and 55, which contain an amine linker and an aromatic A-ring incorporating a para-phenol, as new lead compounds with enhanced potency against the Hh pathway (IC₅₀ values Combining double low line 0.40 and 0.32 μM, respectively).

Full text of DOI:10.1016/j.ejmech.2015.01.049

Accepted Manuscript
Vitamin D3 Analogues that Contain Modified A- and Seco-B-Rings as Hedgehog
Pathway Inhibitors
Albert M. DeBerardinis, Daniel S. Raccuia, Evrett N. Thompson, Chad A. Maschinot,
M. Kyle Hadden
PII:
S0223-5234(15)00069-0
10.1016/j.ejmech.2015.01.049
EJMECH 7665
DOI:
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 12 December 2014
Revised Date: 21 January 2015
Accepted Date: 24 January 2015
Please cite this article as: A.M. DeBerardinis, D.S. Raccuia, E.N. Thompson, C.A. Maschinot, M. Kyle
Hadden, Vitamin D3 Analogues that Contain Modified A- and Seco-B-Rings as Hedgehog Pathway
Inhibitors, European Journal of Medicinal Chemistry (2015), doi: 10.1016/j.ejmech.2015.01.049.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please
note that during the production process errors may be discovered which could affect the content, and all
legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT
Graphical Abstract:
I. Extended linkers
II. Epimeric linkers
III. 'Linkerless"
H
O
O
IV. Carbon-based linkers
V. Nitrogen-based linkers
H
NH
4
, IC50 = 0.74 M
HO
55, IC50 = 0.32 M
HO

ACCEPTED MANUSCRIPT
Vitamin D3 Analogues that Contain Modified A-
and Seco-B-Rings as Hedgehog Pathway
Inhibitors.
Albert M. DeBerardinis, Daniel S. Raccuia, Evrett N. Thompson, Chad A. Maschinot, M. Kyle
Hadden*
Department of Pharmaceutical Sciences, University of Connecticut, 69 N Eagleville Rd, Unit
3092, Storrs, CT 06269-3092
kyle.hadden@uconn.edu
To whom correspondence should be addressed. Phone: 1-806-486-8446. Fax: 1-860-486-6857.
Keywords: hedgehog signaling pathway, vitamin D3, Gli, basal cell carcinoma, medulloblastoma

ACCEPTED MANUSCRIPT
ABSTRACT
The hedgehog (Hh) signaling pathway is a developmental signaling pathway that has been
implicated as a target for anti-cancer drug development in a variety of human malignancies.
Several natural and synthetic vitamin D-based seco-steroids have been identified as potent
inhibitors of Hh signaling with chemotherapeutic potential. These include the previously
characterized analogue 4, which contains the northern CD-ring/side chain region of vitamin D3
(VD3) linked to an aromatic A-ring mimic through an ester bond. To further explore structure-
activity relationships for this class of VD3-based Hh pathway inhibitors, we have designed,
synthesized and evaluated several series of compounds that modify the length, composition, and
stereochemical orientation of the ester linker. These studies have identified compounds 54 and
5
5, which contain an amine linker and an aromatic A-ring incorporating a para-phenol, as new
lead compounds with enhanced potency against the Hh pathway (IC50 values = 0.40 and 0.32
µM, respectively).

ACCEPTED MANUSCRIPT
1
. Introduction
Small molecule mediated inhibition of an inappropriately activated hedgehog (Hh) signaling
pathway is being intensely pursued across academia and industry as a selective anti-cancer
strategy for a number of human cancers [1]. Mutations in the Hh pathway have been linked to the
development of basal cell carcinoma (BCC) and medulloblastoma (MB) and improper Hh
signaling may contribute to the growth and progression of other human cancers [2-4]. Hh
inhibition as a therapeutic strategy has been clinically validated with the FDA approval of
Vismodegib (GDC-0449, 1, Figure 1) for the treatment of locally advanced BCC [5]. Treatment
of an MB patient with 1 demonstrated initial positive results in clinical trials; however, relapse
occurred [6]. A point mutation in the binding site of Vismodegib on key pathway component and
signal transducer Smoothened (Smo; D473H) was identified as the cause of relapse, rendering
the patient insensitive to further treatment with 1 [7]. In addition, a recent study reported that
2
1% of patients receiving continuous treatment of 1 for BCC developed regrowth of at least one
BCC, suggesting resistance in BCC patients could also be a potential concern [8]. Finally, due to
adverse side effects, over half (54%) of basal cell nevus syndrome (BCNS) patients receiving 1
to treat BCCs withdrew from the study voluntarily [9]. These continued drawbacks of
Vismodegib underscore the need for further studies to identify novel potent and selective small
molecule antagonists of the Hh signaling pathway.
The seco-steroid vitamin D3 (VD3, cholecalciferol, 2, Figure 1), has been identified as a non-
selective inhibitor of Hh signaling [10-12]. Although it does not bind to the vitamin D receptor,
VD3 induces activation of canonical VDR signaling in Hh-dependent cell culture and murine
models of BCC [11-12]. This observation suggests that VD3 is readily converted in vitro and in
vivo to chemical structures that can bind and activate vitamin D signaling. Prolonged VDR

ACCEPTED MANUSCRIPT
activation can result in detrimental side effects; therefore, the development of VD3-based
analogues as Hh pathway inhibitors must decouple Hh antagonism and VDR activation. Our
preliminary structure-activity relationship (SAR) studies for VD3 identified Grundmann’s
alcohol, 3, as the pharmacophore for VD3-mediated inhibition of Hh signaling and also served as
a starting point for developing selective Hh pathway inhibitors based on the VD3 scaffold [13].
These studies revealed that 3 was comparable to VD3 in its ability to inhibit Hh signaling
without activating canonical VDR signaling. Analogue 3 was subsequently utilized as a handle
to develop a series of VD3 analogues in which the natural A-ring was replaced with an aromatic
mimic tethered to the CD-ring/side chain region of VD3 through an ester linker [14-15]. The
aromatic A-ring was designed to mimic the natural cyclohexanol A-ring of VD3 and analogue 4
was identified as a potent and selective inhibitor of pathway signaling in multiple Hh-dependent
cell lines (IC values = 0.74 – 5.2 µM).
5
0
Herein, we report the results of our current progress in the design, synthesis, and evaluation
of next generation VD3-based Hh pathway inhibitors utilizing analogue 4 as our lead compound.
These analogues contain modifications to the ester linkage as a means to (1) probe spatial
parameters in the seco-B and A-ring regions, (2) evaluate chemical linkages anticipated to
provide additional or enhanced binding interactions, and (3) replace the ester with tethers that are
predicted to demonstrate improved metabolic stability.
2
. Results and Discussion
2
.1. Chemistry.
Using analogue 4 as our lead scaffold, we sought to develop several series of analogues that
maintain the natural CD-ring/side chain region of VD3 and incorporate an aromatic A-ring

ACCEPTED MANUSCRIPT
mimic, while altering the ester bond (linker region or seco-B-ring). These compounds were
designed to explore optimal linker length between the CD- and A-ring regions, probe the
stereochemical aspects of the linker region, and determine to what extent the ester linkage is
responsible for potent Hh inhibition for these analogues. Finally, while our previous studies
suggested the ester of 4 remains intact in vitro and is required for optimal activity of the scaffold,
this moiety is generally considered to be a metabolic liability in vivo and we sought to explore
chemical linkers with enhanced stability.
Analogue design and synthesis first focused on generating a series of compounds that address
how augmenting the linker length and rigidity of analogue 4 affect its ability to inhibit Hh
signaling. Synthesis of this series began with conversion of the commercially available
carboxylic acids to the corresponding methyl esters using standard conditions followed by
protection of the phenol as either the benzyl or methoxymethyl ether (Scheme 1) [14]. Base-
catalyzed hydrolysis of the methyl esters afforded the carboxylic acids (6, 10-12, 16-18) for
direct coupling to 3 using standard esterification conditions (Scheme 2) [14]. Following
formation of esters 19 – 22, removal of the benzyl-protecting group was accomplished by
palladium-catalyzed reduction. This strategy was selected as a straightforward route to access
analogues 24 – 26 from commercially available coumaric acids 7 – 9, respectively, through the
simultaneous debenzylation and hydrogenation of the alkene in the α,β-unsaturated esters. This
provided ester-linked analogues with a more flexible linker region relative to analogues 30 – 32.
Removal of the methoxymethyl moiety from ester-linked analogues 27 – 29 was accomplished
using camphorsulfonic acid (±CSA) to afford the intact α,β-unsaturated esters 30 – 32.

ACCEPTED MANUSCRIPT
Lead analogue 4 was obtained from optically pure 3, which is the sole CD-ring/side chain
isomer obtained from the one-pot ozonolysis/reduction of VD3 [16]. To determine whether this
stereochemical orientation is required for activity of our scaffold, we sought to prepare the
epimer of analogue 4. Our initial synthetic plan followed a previously disclosed procedure to
access key intermediate 37 (Scheme 3A) [17]. In addition to providing 37, this strategy would
also yield several truncated CD-ring/side chain region analogues whose Hh modulation potential
would be evaluated as part of our ongoing structure-activity analysis for VD3. Formation of the
thermodynamic silyl enol ether 34 from Grundmann’s ketone (33) was followed by
phenylselenium addition and peracid (m-CPBA) treatment of the phenylselenyl-hydrindane to
afford enone 35. Luche reduction selectively afforded 36 as the alcohol with the desired C-8
inversion. Following the reported procedure of allylic alcohol 36, hydrogenation was carried out
in the presence of tris(triphenylphosphine)rhodium chloride (Wilkinson’s catalyst). While 37 was
generated through this route, the overall yield from 33 was low due to the unanticipated
formation of several byproducts in steps 2 through 5 and the incomplete reduction of 36 with up
to stoichiometric amounts of Wilkinson’s catalyst.
A recent report detailed conditions for the preparation of a secondary alcohol structurally
similar to 37 directly from a ketone related to 33 [18]. This conversion was accomplished
through the sodium metal reduction of the corresponding ketone in the presence of an acidic
resin in good yield. We anticipated that 33 would react similarly in the acid-promoted radical
reduction of the carbonyl to afford alcohol 37. Upon employment of these conditions (Scheme
3
B), 33 was readily consumed, generating multiple products (observed by TLC analysis), one of
1
which matched the polarity (R = 0.4, 6:1 Hex:EtOAc) and H NMR of 37 prepared in the initial
f
route. Intermediate 38 was isolated as the only other pure compound from this mixture (R =
f

ACCEPTED MANUSCRIPT
0
.45, 6:1 Hex:EtOAC; 37 and 38 were formed in ~3:1 ratio, respectively). The previous report
detailed the use of Amberlite IR118 as the acidic resin and 20 equivalents of sodium metal to
afford the desired alcohol as the sole product isolated in 81% yield [18]. However, since this
resin was no longer commercially available from the supplier, we replaced the IR118 for the
comparable acidic resin, Amberlite IR120. While clearly a suitable resin to generate 37 and a
superior route to that utilized in Scheme 3A, it is possible that this change could account for the
multiple products observed in the reaction. Further analysis of conditions may be required to
achieve the same degree of selectivity for this substrate with this resin. We anticipated that the
byproducts resulting from the reaction with 33 also involved epimerization at the -carbon(s) as
previously described and isolation of 38 demonstrated. Nonetheless, the one-step reaction from
3
3 employing Amberlite IR120 and sodium afforded 37 in 45% yield, an improvement over the
initial route. Analogue 40 (epi-4) was then readily prepared via esterification with 3-
benzyloxy)-benzoic acid (39) followed by palladium-catalyzed removal of the benzyl protection
(
group (Scheme 4, 40) [14]. Chemical structure configurations were verified by comparisons to
1
literature reports and assignment of chemical shifts from H NMR analysis, in particular the
characteristic methyl signal (singlet) at the C- and D-rings fusion and 1,3- interactions (C-18 and
C-8) of the hydrindane system [17-19].
With these modified esters in hand, our focus shifted towards the design and synthesis of
analogues with altered chemical compositions in the linker region between the CD- and A-ring
motifs. Direct connectivity between these regions and the incorporation of carbon- or nitrogen-
based (amine, amide) chemical linkers would address analogue generation of more drug-like and
metabolically more stable groups relative to the ester linkage. The latter subclass would also

ACCEPTED MANUSCRIPT
address exploring the effects of how varying the hydrogen bonding capacity of this region
affected Hh inhibition of the scaffold.
First, we prepared a ‘linkerless’ series—analogues incorporating the A-ring moiety of 4 as a
direct appendage to the C-ring at C-8. The initial step in synthesizing this series was the addition
of an in situ generated aryllithium species (from 3-methoxymethyl-bromophenol) to
Grundmann’s ketone 33 (Scheme 5). NMR analysis of 41 indicated a single compound was
formed and isolated from the reaction. We anticipated that 41 had the S-configuration according
to literature precedent for an analogous organolithium addition to 33 [20-21]. Removal of the
methoxymethyl protecting group afforded analogue 42. Intermediate 41 was also dehydrated
with Burgess’ reagent (methyl N-(triethylammoniosulfonyl)carbamate) to afford olefin 43
according to previous protocols [22]. A single isomer was anticipated based on a published study
on a similar scaffold, but NMR analysis indicated that two diastereomers, inseparable by flash or
thin-layer chromatography, were isolated in approximately a 3:2 ratio (major isomer shown).
Intermediate 43 was deprotected to afford 44, also isolated and evaluated for activity as a 3:2
mixture. Finally, saturation of the C-8-C-9 olefin was accomplished through palladium-catalyzed
hydrogenation to afford 45 as a mixture of inseparable diastereomers (~4:1). Mixture 45 was
deprotected to yield 46, which was also isolated as a 4:1 mixture (major isomer shown) and
inseparable by standard column chromatography. The mixtures of 44 and 46 were directly
evaluated for their anti-Hh effects without further separation.
Our next analogues concentrated on incorporating one, two, or three atom tethers between
the phenolic A-ring mimic and the CD-ring/side chain region (Schemes 6 through 8). A two-step
homologation was performed on Grundmann’s ketone (33) to yield aldehyde 48 (via enol ether
4
7) with a high degree of diastereomeric purity (Scheme 6A) [23-24]. The stereochemistry at C-

ACCEPTED MANUSCRIPT
8
was established based on the previous preparation and characterization of 48 and 48-type
compounds via a modified procedure [25]. A number of reports describe the synthesis of 48 [25-
6]; however, we found that ylide generation using sodium tert-butoxide and
methoxymethyl)triphenylphosphonium chloride (phosphonium salt) and subsequent addition to
2
(
3
3 to generate 47 was high yielding and reproducible. A portion of 48 was reduced to provide
9, a homologated version of 37, which extends the hydroxyl functionality further from the C-
4
ring [27]. Installation of the phenolic A-ring was carried out as described in Scheme 5 above.
Here, the in situ generated aryllithium species was added to 48 forming a mixture of products,
which were directly subjected to PDC-facilitated oxidation. Intermediate 50 was easily separated
as the least polar component in the mixture and subsequent removal of the methoxymethyl
protecting group afforded the one carbon extended linker analogue 51 as a single isomer.
As previously discussed, an aim of these linker-focused modification studies was to evaluate
analogues that incorporated groups with the potential to access additional strong bonding
interactions (e.g. hydrogen-bonding) with a possible target(s). Thus, a series of nitrogen-based
linker analogues was designed and synthesized (Schemes 7 and 8). Specifically, amine- and
amide-linkages were used to attach the aromatic A-ring motif to 3. First, reductive amination of
3
3 using mono-substituted aminophenols to yield amine-linked analogues following a literature
protocol was performed [28]. The stereoisomers isolated from the conditions employed were
1
assigned by comparative H NMR analysis. Interestingly, whereas a single isomer was isolated
from reactions between 2- or 3-aminophenol and 33, when 4-aminophenol was utilized, both
diastereomers were isolated. We accounted for this observation based on the reaction scale for
each respective analogue. The preparations of 52 and 53 were run on a reduced scale (one-third
to one-half) relative to synthesis of 54 (which also yielded 55). Analogues 54 and 55 combined

ACCEPTED MANUSCRIPT
to yield 50% of the amine-linked products in a 4:1 ratio. Based on this ratio, it is possible that a
small quantity of epimer was generated from the syntheses of 52 and 53, but at such a minute
quantity that it went undetected during purification.
As a direct comparison to lead compound 4, analogue 59 was designed to replace the ester
with an amide bond. Oxime 56 was formed in excellent yield from the condensation of
hydroxylamine and 33 in pyridine (Scheme 7B) [19]. Reduction of the oxime to amine 57 was
accomplished by a slightly modified literature procedure [29]. In one pot, 56 was sequentially
refluxed with lithium aluminum hydride followed by zinc in acetic acid. This route afforded
mainly amine 57, major C-8 isomer shown in Scheme 7, matching the stereochemistry of 3
(~6:1). Standard conditions were employed to couple 57 with 3-(benzyloxy)-benzoic acid to
1
afford 58, which was subsequently deprotected to yield amide 59 [30]. H NMR analysis of 59
also indicated an approximately 6:1 mixture of C-8 epimers (major isomer shown in Scheme
7
B), which was directly evaluated without further purification.
The final series of analogues prepared for analysis utilized an enamide linker to extend the
distance between the A-ring and CD-ring/side chain region while also maintaining the coplanar
nature of the natural VD3 scaffold. Triethyl phosphonoacetate was coupled to 33 via a Horner-
Wadsworth-Emmons olefination to yield α, β-unsaturated ester 60 [31]. A series of enamide-
linked analogues were directly prepared through the trimethylaluminum mediated coupling of
aniline substituents with ester 60 [32]. This series included an unsubstitued (61), ortho-, meta-,
and para-monosubstituted phenolic enamides (62, 65, and 66). In addition, allylic alcohol 67 was
generated from the lithium aluminum hydride reduction of ester 60. Fully reduced linker-region
alcohol 69 was prepared as a single isomer by Pd-catalyzed reduction of the alkene in 60
followed by lithium aluminum hydride reduction of ester 68. Together with alcohol 49,

ACCEPTED MANUSCRIPT
analogues 67 and 69 were prepared for evaluation as one- and two-carbon analogue extensions of
3
and 37, respectively.
2
.2. Biological Evaluation
Each analogue was initially assessed for its ability to inhibit Hh signaling by evaluating its
effects on endogenous Gli1 mRNA levels at a single concentration (5 µM) in C3H10T1/2 cells,
an Hh-dependent mouse embryonic fibroblast (MEF) cell line. Gli1 is a well-characterized target
gene of Hh signaling that is robustly up-regulated in this cell line following activation of Hh
signaling with exogenous Hh ligand or pathway agonist(s) [33]. Concomitant administration of a
pathway inhibitor prevents Gli1 mRNA over-expression. For these studies, we utilized a standard
oxysterol cocktail (20α-hydroxycholesterol and 22(S)-hydroxycholesterol, 5 µM each) to activate
Hh pathway signaling as previously described [11]. In addition, selectivity for Hh pathway
inhibition compared to canonical VDR activation was determined in the same cellular system by
monitoring up-regulation of Cyp24a1 mRNA, a well-characterized target gene of VDR [12,34].
SAR studies using this approach for earlier analogue generations [11,13-15] established that
structures constructed from ester-linked Grundmann’s alcohol (3) and hydrophilic aromatic A-
ring substituents could more potently and selectively inhibit Hh signaling compared to VD3 (2).
Initially, analogues with extended ester linkers relative to lead compound 4 were prepared and
screened to assess the optimal distance between the CD-ring/side chain and aromatic A-ring
regions (Table 1). Analogue 23, which incorporates an additional methylene in the linker region,
demonstrated potency and selectivity comparable to 4, suggesting this region of the binding site
is amenable to a three atom linker. Analogues 24-26 and 30-32 also contain the ester linker, but
are extended an additional methylene (from 23) to incorporate a four atom linker region between
the CD- and A-rings. In addition, evaluating the saturated (24-26) and unsaturated (30-32)

ACCEPTED MANUSCRIPT
analogues allows for a direct comparison of the flexibility/rigidity requirements in this region of
the scaffold. Each of these four atom esters demonstrated reduced inhibition of Hh signaling
compared to both 4 and 23, suggesting that more than three atoms in this region is detrimental to
the anti-Hh effects of the ester linked analogues. In addition, the saturated analogues (Gli1
expression = 15-47%), were more active than the corresponding unsaturated analogues (Gli1
expression = 57-94%). These results are most likely due to the increased flexibility of the
saturated analogues allowing the hydroxyl moiety easier access to key binding interactions. It is
interesting to note that in our previous SAR studies the 2-phenolic equivalent of 4 was inactive
[
15] whereas the corresponding four atom ester 24 regains moderate Hh inhibition. Finally, each
of these extended ester analogues demonstrated significantly reduced VDR activation when
compared to VD3.
The second subclass of linker region analogues focused on evaluation of epimeric structures
of 3 and 4, specifically at C-8 of the VD3 scaffold (Table 2). Ketone 33 and enone 35 were
equipotent at inhibiting Hh signaling (Gli1 expression = 66% and 63%, respectively) and both
were slightly less active than 3 (Gli1 expression = 46%). Epimeric alcohols 36 and 37 were
significantly less active than 3, while analogue 38, which was recovered from the radical-
promoted reduction of 33, demonstrated Hh inhibition and selectivity comparable to 3. Analogue
4
0, which was prepared specifically to evaluate whether inversion of the ester attachment at C-8
affected activity was significantly less active than lead compound 4 (Gli expression = 30% and
%, respectively). Finally, none of these isomeric analogues demonstrated significant activation
of VDR.
Our next analogue series was designed to shorten the linker region by directly appending the
-phenolic A-ring mimic to the CD-ring region or incorporating a one carbon-based linker
1
3

ACCEPTED MANUSCRIPT
between the two regions (Table 3). The “linkerless” series of analogues (42, 44, 46)
demonstrated moderate potency against the Hh signaling pathway (Gli1 expression = 39-47%)
and did not activate VDR. These data indicate that the absence of a linker region acting as
chemical spacer between the CD-ring and aromatic A-ring produces analogues with diminished
anti-Hh properties. The related one carbon linker series demonstrated a wide-range of Hh
inhibitory activity. Analogue 51, which contains a ketone-based linker and inverted
stereochemistry compared with lead compound 4, demonstrated potent down-regulation of Gli1
mRNA (11% relative to fully activated OHC controls) in an Hh-selective manner. This data
suggests that the 3-hydroxyl substituent in 51 is more suitably oriented to generate improved
binding interactions when compared with the inverted ester of analogue 40 or the directly
appended A-ring in analogue 46.
With respect to intermediate structures that did not include an aromatic A-ring mimic, 60 was
inactive, which was not unexpected based on the previous analysis of compounds lacking or
masking hydrophilic group(s) in the A- and seco-B-ring regions. Analogue 49, a one-carbon
extension of 37, was moderately active (Gli1 expression = 50%) which correlated well with the
activity demonstrated by Grundmann’s alcohol (3). An additional carbon between the CD-ring
region and the primary hydroxyl (69) provided an analogue that completely abolished Hh
signaling. Interestingly, constraining the primary hydroxyl such that it is coplanar with the CD-
ring region (67) was detrimental to its activity, suggesting that the flexibility of the saturated
linker is important for activity. Of these analogues, only 67 and 69, the latter which demonstrated
the highest level of Hh inhibition, activated VDR at a modest level (28 and 18.6-fold increase,
respectively).

ACCEPTED MANUSCRIPT
The preliminary screening results for analogues containing a nitrogen-based linker are shown
in Table 4. These compounds represent the first series to incorporate a hydrogen-bonding moiety
in the linker region. Each of the analogues that included an amine-based linker (52-55)
demonstrated potent inhibition of Hh signaling. The 3-phenol analogue (53) was the least active
of this series, highlighting the fact that modification of the linker length affects the orientation of
the aromatic A-ring in the binding pocket. Intermediate oxime 55 was also evaluated, but
demonstrated no anti-Hh activity. In addition, each of the amine-linked analogues up-regulated
Cyp24A1 at a modest level, similar to data from previously characterized analogues that
completely abrogate Gli1 expression. Analogue 59, the amide equivalent of lead compound 4,
also maintained potent Hh inhibition. Interestingly, up-regulation of Cyp24A1 by 59 was 3-fold
less than 4 and the amine-linked analogues.
The series of enamide linked analogues also demonstrated the ability to down-regulate Hh
signaling (Table 4). Analogue 61, which does not contain any functionality on the aromatic A-
ring, demonstrated modest Hh signaling inhibition. This was in direct contrast to our previous
results from the original ester-linked analogue series in which a hydrophilic moiety on the A-ring
was required for Hh inhibition [14-15]. This result suggests that extending the atoms in the linker
region can make key interactions within the analogue binding site. Each of the other enamide
analogues (62 and 65-66) were potent inhibitors of Hh signaling with modest activation of VDR.
Upon completion of the preliminary compound analysis in C3H10T1/2 cells at single
concentrations (5 µM), analogues that demonstrated potent anti-Hh signaling activity were
selected for further evaluation. In general, analogues that reduce Gli1 mRNA expression to
levels less than 20% of the OHC-activated control were defined as potent inhibitors of Hh
signaling and compounds fitting this criterion were evaluated for their ability to inhibit Hh

ACCEPTED MANUSCRIPT
signaling in a concentration-dependent fashion. These follow-up studies were performed in both
C3H10T1/2 cells and ASZ (murine BCC) cells (Table 5). Each of the compounds chosen for
further evaluation demonstrated the ability to down-regulate Gli1 mRNA expression in both cell
lines in a concentration-dependent manner with a range of analogue activity (IC values = 0.32 -
5
0
3
.9 µM in C3H10T1/2 and 0.57 – 6.7 µM in ASZ). While the majority of analogues inhibited Hh
signaling at a level comparable to lead compound 4, analogues 54 and 55 were significantly
more active. With IC values in the low to mid-nanomolar range (C3H10T1/2 = 0.40 and 0.32
5
0
µM and ASZ = 0.67 and 0.57 µM, respectively), these compounds represent the most potent
inhibitors of Hh signaling based on the VD3 scaffold identified to date.
Of additional interest, evaluation of this series revealed analogues whose ability to modulate
Hh signaling correlated well between C3H10T1/2 and ASZ cell lines. Previously, no such
activity correlation was observed for our VD3-based analogues, but was only determined for the
parent VD3 or Hh signaling inhibitors known to function through direct binding to Smo (1 and
cyclopamine) [1]. Lead compound 4 and additional VD3 analogues have demonstrated
significantly reduced activity in the ASZ cell line. None of these VD3 analogues demonstrated
the capacity to displace a tight-binding fluorescent VDR ligand from full length human VDR,
matching previously screened VD3-based synthetic analogues (up to 100 µM) [11,13-15]. These
data taken together with the comparatively modest effects on Cyp24a1 mRNA up-regulation
seen in C3H10T1/2 and ASZ cells (relative to VD3 and calcitriol) suggests the VD3 analogues
with an aromatic A-ring exhibit Hh signaling inhibition through mechanism(s) distinct from
VDR.

ACCEPTED MANUSCRIPT
3
. Conclusion
Our continuing efforts to discover potent and selective VD3-based Hh signaling inhibitors
directed the design, synthesis, and biological evaluation studies of a series of linker-modified
analogues. In particular, with 4 as a lead scaffold, analogues were rationally designed to maintain
the CD-ring/side chain region of VD3 and the aromatic A-ring of 4, while altering the tether
between these two regions. Several parameters for this region were explored in these studies,
including exploration of linker length and rigidity, orientation of the hydroxyl or phenolic
moiety, and incorporation of additional hydrogen-bonding groups within the linker region. The
spatial range available from C-8 of 3 allowed us to manipulate presentation of the active
substituents in the A and seco-B-ring regions, effectively changing the pharmacophore spatial
positioning by varying the chemical attachments between the two distinct segments (CD-rings
and A-ring mimic). This strategy allowed us to more thoroughly probe positional and chemical
modifications of the pharmacophore in this region of lead analogue 4. Analysis for this linker-
modified analogue series indicates that this strategy was successful, as demonstrated by the
identification of selective VD3-based inhibitors of Hh signaling with increased potency. Future
analogue development will explore the chemical space around the C-ring and further linker
modifications are planned, along with mechanistic and in vivo studies addressing how VD3 and
synthetic analogues exhibit Hh modulation.

ACCEPTED MANUSCRIPT
4
. Experimental Section
4
.1. General Information.
VD3 used for chemical synthesis was purchased from HBCChem, Inc. Other chemicals were
purchased from Sigma-Aldrich or Fisher Scientific. ACS or HPLC grade methanol, acetone, and
tetrahydrofuran were purchased from Fisher Scientific. Anhydrous DCM (low water, <50 ppm
water) was purchased from BrandNu Laboratories, Inc. (J.T. Baker solvent). Column
chromotagraphy was performed using silica gel purchased from Sorbtech (Sorbent
Technologies). NMR data was performed on a Bruker AVANCE 500 or 400 MHz spectrometer
and analysis with MestReNova version 8.0.0. HRMS data was analyzed at the Mass
Spectrometry Facility at the University of Connecticut, performed by Dr. You-Jun Fu. FT-IR
analysis was performed on a Bruker Alpha Platinum ATR instrument using OPUS software (v
7
.2).
4
4
.2. Synthesis of extended ester linked analogues.
.2.1 General procedure for methylation of acetic/coumaric acids.
Thionyl chloride (18.3 mmol) was added dropwise to a solution of phenylacetic or coumaric acid
5, 7 – 9; 6.1 mmol) in methanol (15 mL) at 0°C. The reaction mixture was stirred overnight and
(
allowed to warm to room temperature. After 24 h, ice was slowly added with vigorous stirring,
followed by a solution of sodium bicarbonate (80 mL; aqueous, saturated). The reaction mixture
was diluted with EtOAc (100mL) and extracted (2 X 100mL). The combined organics were
dried over Na SO , filtered, and concentrated. The crude methyl ester protected coumarates were
2
4
1
obtained in good yields (65-85%) and purity was satisfactory ( H NMR) to move directly to the
next step. Characterization of intermediates was consistent with that previously reported [35-36].

ACCEPTED MANUSCRIPT
4
.2.2. General procedure for benzyl protection of phenols.
Protection of the phenols with benzyl bromide (yields 75-90%) followed by saponification of the
methyl ester (yields 55-80%) were as detailed previously [14-15,37-39].
4
.2.3. General procedure for methoxymethyl protection of phenols.
Sodium hydride (60% mineral oil dispersion; 5.1 mmoL) was added portion-wise with vigorous
stirring to a solution of methyl ester coumarate (5.1 mmoL) in THF (30 mL) at 0°C. After
completion of the NaH addition, the mixture was stirred an additional 2 h at 0°C. Chloromethyl
methyl ether (6.63 mmoL; technical grade, 90%) was added dropwise to this mixture, which was
then allowed to warm to room temperature over 16 h. After this time, water (15 mL) was added
to the solution, followed by saturated ammonium chloride (60mL) and the aqueous layer was
washed with ethyl acetate (3 X 100 mL). The organic layers were combined, dried over MgSO4,
filtered, and concentrated. Column chromatography (SiO ) on the crude residue afforded MOM-
2
protected methyl ester coumarates in good yields (60-90%). Spectral analysis matched previous
reports [40]. Carboxylic acids were prepared by saponification of methyl esters as described
previously [15,41-43].
4
.2.4. 3-methoxymethyl propenoic acid (17).
1
H NMR (500 MHz, CDCl ) δ 11.81 (bs, 1H), 8.15 (d, J = 16.1 Hz, 1H), 7.57 (m, 1H), 7.35 (m,
3
1
3
1
H), 7.18 (m, 1H), 7.03 (d, J = 16.1 Hz, 1H). C NMR (126 MHz, CDCl ) δ 172.5, 156.1,
3
1
42.2, 131.9, 128.7, 123.7, 121.9, 117.7, 114.8, 94.5, 56.3. DART-HRMS: m/z calcd. for
+
C H O : 209.0814 [MH] . Found: 209.0815.
11
13
4
4
.2.5. General procedures for esterification and deprotection of ester-linked analogues.
Esterification of 3 and protected acids was performed as described previously (yields 60-90%)

ACCEPTED MANUSCRIPT
[
14-15]. Palladium hydroxide (10%) reduction/deprotection under an atmosphere of hydrogen in
MeOH:THF (2:1) was performed as detailed previously (yields 50-80%) [14-15]. Removal of the
methoxymethyl protecting group with ±CSA (2 equivalents) in anhydrous MeOH (2 mL) for 24-
3
6 h was done at room temperature as previously described [44]. The reaction was diluted in
ethyl acetate (50 mL) and washed with saturated NaHCO (1 X 50 mL). The organic layer was
3
collected, dried over MgSO , filtered, and concentrated. Column chromatography (SiO , various
4
2
gradients of Hex:EtOAc) afforded ester linked analogues 23-26 and 30-32 as clear oils (65-90%).
4
.2.5.1. 3-benzyloxy phenyl acetate-VD3 ester (19).
1
H NMR (500 MHz, CDCl ) δ 7.44 (m, 2H), 7.39 (m, 2H), 7.33 (m, 1H), 7.23 (m, 1H), 6.93 (m,
3
1
H), 6.88 (m, 2H), 5.16 (m, 1H), 5.07 (s, 2H), 3.59 (s, 2H), 1.97 (m, 1H), 1.80 (m, 2H), 1.60 (m,
H), 1.52 (m, 2H), 1.41 (m, 5H), 1.32 (m, 3H), 1.27 (m, 2H), 1.21 (m, 2H), 1.14 (m, 4H), 1.05
1
1
3
(
m, 1H), 0.99 (m, 1H), 0.88 (m, 10H), 0.73 (s, 3H). C NMR (126 MHz, CDCl ) δ 171.2, 158.8,
3
1
4
36.9, 135.8, 129.4, 128.5, 127.8, 127.3, 121.9, 115.8, 113.4, 77.2, 72.0, 69.8, 56.4, 51.3, 42.1,
1.8, 39.9, 39.4, 35.9, 35.3, 30.4, 29.6, 27.9, 27.0, 23.7, 22.7, 22.5, 22.5, 18.5, 17.8, 12.8.
+
DART-HRMS: m/z calcd. for C H O : 490.3447 [M] . Found: 490.3461.
3
3
46
3
4
.2.5.2. 2-benzyloxy coumarate-VD3 ester (20).
1
H NMR (500 MHz, CDCl ) δ 8.06 (d, J = 16.1 Hz, 1H), 7.56 (m, 1H), 7.45 (m, 2H), 7.39 (m,
3
2
H), 7.33 (m, 2H), 6.98 (m, 2H), 6.48 (d, J = 16.2 Hz, 1H), 5.26 (m, 1H), 5.12 (s, 2H), 1.99 (m,
1
H), 1.89 (m, 1H), 1.79 (m, 1H), 1.71 (m, 1H), 1.53 (m, 2H), 1.45 (m, 4H), 1.33 (m, 4H), 1.25
(m, 1H), 1.15 (m, 5H), 1.06 (m, 1H), 0.98 (m, 1H), 0.90 (d, J = 7.1 Hz, 3H), 0.88 (d, J = 6.5 Hz,
1
3
3
H), 0.81 (s, 3H). C NMR (126 MHz, CDCl ) δ 167.1, 157.2, 139.1, 136.4, 131.2, 128.6,
3
1
28.4, 128.0, 127.6, 123.8, 121.0, 119.4, 112.4, 71.3, 70.4, 56.5, 51.5, 41.9, 40.0, 39.5, 35.9,

ACCEPTED MANUSCRIPT
3
5.3, 30.6, 28.0, 27.1, 23.7, 22.8, 22.6, 22.5, 18.5, 18.0, 13.1. DART-HRMS: m/z calcd. for
+
C H O : 503.3525 [MH] . Found: 503.3510.
34
47
3
4
.2.5.3. 3-benzyloxy coumarate-VD3 ester (21).
1
H NMR (500 MHz, CDCl ) δ 7.65 (d, J = 16.0 Hz, 1H), 7.45 (m, 2H), 7.40 (m, 2H), 7.35 (m,
3
1
H), 7.31 (m, 1H), 7.15 (m, 2H), 7.01 (dd, J = 8.2, 2.4 Hz, 1H), 6.42 (d, J = 15.9 Hz, 1H), 5.31
(m, 1H), 5.10 (s, 2H), 2.07 (m, 1H), 1.94 (m, 2H), 1.82 (m, 2H), 1.52 (m, 5H), 1.39 (m, 4H),
1
.26 (m, 2H), 1.16 (m, 4H), 1.03 (m, 1H), 1.00 (s, 3H), 0.95 (d, J = 6.5 Hz, 3H), 0.90 (d, J = 2.1
13
Hz, 3H), 0.89 (d, J = 2.0 Hz, 3H). C NMR (126 MHz, CDCl ) δ 166.6, 159.0, 143.9, 136.6,
3
1
3
35.9, 129.8, 128.5, 128.0, 127.4, 120.9, 119.2, 116.6, 114.0, 71.6, 70.0, 56.4, 51.5, 41.9, 39.9,
9.4, 35.9, 35.3, 30.5, 27.9, 27.0, 23.7, 22.7, 22.6, 22.5, 18.5, 17.9, 13.2. DART-HRMS: m/z
+
calcd. for C H O : 502.3447 [M] . Found: 502.3468.
3
4
46
3
4
.2.5.4. 4-benzyloxy coumarate-VD3 ester (22).
1
H NMR (500 MHz, CDCl ) δ 7.63 (d, J = 16.0 Hz, 1H), 7.48 (d, J = 8.7 Hz, 1H), 7.41 (m, 4H),
3
7
.34 (1H), 6.98 (d, J = 8.6 Hz, 1H), 6.31 (d, J = 15.9 Hz, 1H), 5.29 (m, 1H), 5.10 (s, 2H), 2.05
(m, 1H), 1.92 (m, 1H), 1.81 (m, 2H), 1.50 (m, 5H), 1.38 (m, 4H), 1.23 (m, 2H), 1.13 (m, 4H),
1
.02 (m, 1H), 0.98 (s, 3H), 0.94 (d, J = 6.5 Hz, 3H), 0.89 (d, J = 2.1 Hz, 3H), 0.88 (d, J = 2.0 Hz,
1
3
3
H). C NMR (126 MHz, CDCl ) δ 167.0, 160.3, 143.7, 136.4, 129.6, 128.6, 128.0, 127.5,
3
1
27.4, 116.5, 115.1, 71.3, 70.0, 56.4, 51.5, 41.9, 40.0, 39.4, 35.9, 35.3, 30.6, 27.9, 27.1, 23.7,
+
2
2.7, 22.6, 22.5, 18.5, 18.0, 13.2. DART-HRMS: m/z calcd. for C H O : 503.3525 [MH] .
3
4
47
3
Found: 503.3510.
4
.2.5.5. 3-hydroxy phenylacetate-VD3 ester (23).

ACCEPTED MANUSCRIPT
1
H NMR (500 MHz, CDCl ) δ 7.16 (m, 1H), 6.81 (m, 1H), 6.76 (m, 1H), 6.72 (m, 1H), 5.47 (s,
3
1
H), 5.14 (m, 1H), 3.56 (s, 2H), 1.96 (m, 1H), 1.78 (m, 2H), 1.59 (m, 1H), 1.51 (m, 1H), 1.40
(m, 4H), 1.31 (m, 3H), 1.25 (m, 2H), 1.19 (m, 1H), 1.12 (m, 5H), 1.04 (m, 1H), 0.97 (m, 1H),
1
3
0
1
2
1
4
.86 (m, 10H), 0.73 (s, 3H). C NMR (126 MHz, CDCl ) δ 171.7, 155.8, 135.8, 129.6, 121.5,
3
16.2, 114.0, 77.2, 72.3, 56.4, 51.3, 41.9, 41.8, 39.9, 39.4, 35.9, 35.3, 30.4, 28.0, 27.0, 23.7, 22.8,
2.5, 22.5, 18.5, 17.8, 12.8. IR(neat) vmax 3374, 2947, 2930, 2867, 1703, 1602, 1590, 1488,
+
366, 1231, 1152, 762, 690. DART-HRMS: m/z calcd. for C H O : 400.2977 [M] . Found:
2
6
40
3
00.2976.
4
.2.5.6. 2-hydroxy hydrocoumarate-VD3 ester (24).
1
H NMR (500 MHz, CDCl ) δ 7.17 (s, 1H), 7.10 (m, 2H), 6.87 (m, 2H), 5.17 (m, 1H), 2.90 (t, J
3
=
6.3 Hz, 2H), 2.71 (t, J = 6.4 Hz, 2H), 1.98 (m, 1H), 1.78 (m, 2H), 1.55 (m, 5H), 1.40 (m, 5H),
1
3
1
.27 (m, 3H), 1.13 (m, 4H), 1.04 (m, 1H), 0.87 (m, 9H), 0.79 (s, 3H). C NMR (126 MHz,
CDCl ) δ 175.4, 154.3, 130.4, 127.9, 127.4, 120.7, 117.3, 72.8, 56.3, 51.2, 41.8, 39.8, 39.4, 35.9,
3
3
3
9
4
5.6, 35.3, 30.4, 29.7, 28.0, 27.0, 24.7, 23.7, 22.8, 22.6, 22.5, 18.5, 17.8, 12.9. IR(neat) vmax
428, 2962, 2919, 2871, 2849, 1714, 1691, 1588, 1477, 1446, 1290, 1215, 1162, 1102, 1072,
+
45, 881, 755, 681. DART-HRMS: (a) m/z calcd. for C H O : 415.3212 [MH] . Found:
2
7
43
3
+
15.3217; (b) m/z calcd. for C H O NH : 432.3478 [MNH ] . Found: 432.3489.
2
7
42
3
4
4
4
.2.5.7. 3-hydroxy hydrocoumarate-VD3 ester (25).
1
H NMR (500 MHz, CDCl ) δ 7.14 (m, 1H), 6.75 (m, 1H), 6.69 (m, 2H), 5.91 (s, 1H), 5.17 (m,
3
1
H), 2.91 (t, J = 7.8 Hz, 2H), 2.63 (t, J = 7.8 Hz, 2H), 1.99 (m, 1H), 1.79 (m, 2H), 1.62 (m, 1H),
1
.51 (m, 1H), 1.41 (m, 4H), 1.39 (m, 3H), 1.10 (m, 8H), 0.90 (d, J = 6.5 Hz, 3H), 0.88 (d, J = 2.2
13
Hz, 3H), 0.86 (d, J = 2.2 Hz, 3H), 0.84 (s, 3H). C NMR (126 MHz, CDCl ) δ 173.2, 155.9,
3
1
42.1, 129.6, 120.3, 115.2, 113.2, 71.9, 56.4, 51.2, 41.8, 39.9, 39.4, 36.2, 35.9, 35.3, 30.9, 30.4,

ACCEPTED MANUSCRIPT
2
7.9, 27.0, 23.7, 22.7, 22.5, 18.5, 17.8, 13.0. IR(neat) vmax 3320, 2947, 2847, 1730, 1701, 1589,
1
452, 1391, 1351, 1332, 1231, 1146, 1061, 1032, 997, 875, 777. DART-HRMS: (a) m/z calcd.
+
for C H O : 413.3055 [M-H] . Found: 413.3014; (b) m/z calcd. for C H O NH : 432.3478
2
7
41
3
27 42
3
4
+
[
MNH ] . Found: 432.3443.
4
4
.2.5.8. 4-hydroxy hydrocoumarate-VD3 ester (26).
1
H NMR (500 MHz, CDCl ) 7.05 (d, J = 8.1 Hz, 2H), 6.75 (d, J = 8.0 Hz, 1H), 5.78 (s, 1H), 5.15
3
(m, 1H), 2.89 (t, J = 7.8 Hz, 2H), 2.60 (t, J = 7.7 Hz, 2H), 1.99 (m, 1H), 1.78 (m, 2H), 1.63 (m,
1
H), 1.51 (m, 1H), 1.41 (m, 4H), 1.33 (m, 3H), 1.23 (m, 2H), 1.13 (m, 4H), 1.06 (m, 1H), 0.99
1
3
(
m, 1H), 0.87 (m, 12H). C NMR (126 MHz, CDCl ) δ 173.2, 154.2, 132.2, 129.2, 115.3, 71.8,
3
5
1
1
6.4, 51.2, 41.8, 39.9, 39.4, 36.6, 35.9, 35.3, 30.5, 30.2, 27.9, 27.0, 23.7, 22.7, 22.5, 18.5, 17.8,
3.0. IR(neat) vmax 3509, 2958, 2928, 2857, 1730, 1644, 1523, 1471, 1443, 13738, 1246, 1130,
110, 1079, 1033, 956, 883, 773, 666. DART-HRMS: (a) m/z calcd. for C H O : 413.3055
27
41
3
+
+
[
M-H] . Found: 413.3024; (b) m/z calcd. for C H O NH : 432.3478 [MNH ] . Found:
27 42 3 4 4
4
32.3443.
4
.2.5.9. 2-methoxymethyl coumarate-VD3 ester (27).
1
H NMR (500 MHz, CDCl ) δ 7.61 (d, J = 16.0 Hz, 1H), 7.47 (d, J = 8.7 Hz, 2H), 7.04 (d, J =
3
8
.7 Hz, 2H), 6.30 (d, J = 16.0 Hz, 1H), 5.28 (m, 1H), 5.20 (s, 2H), 3.48 (s, 3H), 2.05 (m, 1H),
.92 (m, 1H), 1.80 (m, 2H), 1.50 (m, 5H), 1.36 (m, 4H), 1.24 (m, 4H), 1.12 (m, 4H), 1.03 (m,
H), 0.97 (s, 3H), 0.92 (d, J = 6.5 Hz, 3H), 0.87 (d, J = 2.0 Hz, 3H), 0.86 (d, J = 2.0 Hz, 3H).
1
1
1
3
C NMR (126 MHz, CDCl ) δ 167.0, 158.8, 143.7, 129.5, 128.3, 116.9, 116.4, 94.2, 71.4, 56.4,
3
5
6.1, 51.5, 41.9, 40.0, 39.4, 35.9, 35.4, 30.6, 27.9, 27.1, 23.7, 22.7, 22.6, 22.5, 18.5, 18.0, 13.2.
+
DART-HRMS: m/z calcd. for C H O : 457.3317 [MH] . Found: 457.3319.
2
9
45
4

ACCEPTED MANUSCRIPT
4
.2.5.10. 3-methoxymethyl coumarate-VD3 ester (28).
1
H NMR (500 MHz, CDCl ) δ 8.07 (d, J = 16.1 Hz, 1H), 7.56 (d, J = 7.8 Hz, 1H), 7.32 (m, 1H),
3
7
.14 (d, J = 8.5 Hz, 1H), 7.01 (m, 1H), 6.47 (d, J = 16.1 Hz, 1H), 5.29 (m, 1H), 5.25 (s, 2H), 3.50
(s, 3H), 2.05 (m, 1H), 1.93 (m, 1H), 1.80 (m, 2H), 1.50 (m, 5H), 1.36 (m, 4H), 1.24 (m, 4H),
1
3
1
.13 (m, 4H), 1.02 (m, 1H), 0.99 (s, 3H), 0.93 (d, J = 6.4 Hz, 3H), 0.87 (d, J = 6.5 Hz, 6H). C
NMR (126 MHz, CDCl ) δ 167.1, 155.7, 139.2, 131.3, 128.0, 124.2, 121.8, 119.3, 114.7, 94.5,
3
7
1
1.4, 56.5, 56.2, 51.5, 41.9, 40.0, 39.4, 35.9, 35.4, 30.6, 27.9, 27.1, 23.7, 22.7, 22.6, 22.5, 18.6,
+
8.0, 13.1. DART-HRMS: m/z calcd. for C H O : 457.3317 [MH] . Found: 457.3325.
2
9
45
4
4
.2.5.11. 4-methoxymethyl coumarate-VD3 ester (29).
1
H NMR (500 MHz, CDCl ) δ 8.07 (d, J = 16.2 Hz, 1H), 7.55 (m, 1H), 7.32 (m, 1H), 7.14 (d, J =
3
8
.3 Hz, 1H), 7.02 (m, 1H), 6.46 (d, J = 16.2 Hz, 1H), 5.29 (m, 1H), 5.25 (s, 2H), 3.50 (s, 3H),
.06 (m, 1H), 1.93 (m, 1H), 1.81 (m, 2H), 1.51 (m, 5H), 1.38 (m, 4H), 1.24 (m, 2H), 1.12 (m,
H), 1.01 (m, 1H), 0.98 (s, 3H), 0.93 (d, J = 6.5 Hz, 3H), 0.87 (d, J = 1.9 Hz, 3H), 0.86 (d, J =
2
4
1
1
2
4
13
.9 Hz, 2H). C NMR (126 MHz, CDCl ) δ 167.1, 155.7, 139.2, 131.3, 128.0, 124.2, 121.8,
3
19.3, 114.7, 94.5, 71.4, 56.5, 56.2, 51.5, 41.9, 40.0, 39.5, 35.9, 35.4, 30.6, 28.0, 27.1, 23.7, 22.8,
+
2.6, 22.5, 18.6, 18.0, 13.1. DART-HRMS: m/z calcd. for C H O : 457.3317 [MH] . Found:
2
9
45
4
57.3333.
4
.2.5.12. 2-hydroxycoumarate-VD3 ester (30).
1
H NMR (500 MHz, CDCl ) δ 7.62 (d, J = 16.0 Hz, 1H), 7.26 (m, 1H), 7.10 (d, J = 7.7 Hz, 1H),
3
7
.05 (s, 1H), 6.89 (m, 1H), 6.40 (d, J = 16.0 Hz, 1H), 5.72 (bs, 1H), 5.30 (m, 1H), 2.05 (m, 1H),
1
1
.93 (m, 1H), 1.80 (m, 2H), 1.51 (m, 5H), 1.37 (m, 4H), 1.23 (m, 3H), 1.13 (m, 4H), 1.01 (m,
13
H), 0.96 (s, 3H), 0.93 (d, J = 6.5 Hz, 3H), 0.88 (d, J = 1.4 Hz, 3H), 0.87 (d, J = 1.4 Hz, 3H). C
NMR (126 MHz, CDCl ) δ 167.1, 156.2, 144.2, 135.9, 130.0, 120.6, 119.0, 117.4, 114.5, 72.0,
3

ACCEPTED MANUSCRIPT
5
6.4, 51.5, 41.9, 39.9, 39.4, 35.9, 35.4, 30.5, 27.9, 27.0, 23.7, 22.8, 22.6, 22.5, 18.5, 17.9, 13.2.
IR(neat) vmax 3435, 2928, 2867, 2850, 1694, 1681, 1636, 1597, 1448, 1219, 1188, 1155, 1131,
+
1
062, 1034, 977, 848, 782, 672. DART-HRMS: m/z calcd. for C H O : 413.3056 [MH] .
2
7
41
3
Found: 413.3071.
4
.2.5.13. 3-hydroxycoumarate-VD3 ester (31).
1
H NMR (500 MHz, CDCl ) δ 7.61 (d, J = 15.8 Hz, 1H), 7.43 (m, 1H), 6.84 (m, 1H), 6.29 (d, J =
3
1
6.0 Hz, 1H), 5.58 (s, 1H), 5.28 (m, 1H), 2.05 (m, 1H), 1.91 (m, 1H), 1.79 (m, 2H), 1.64 (m,
H), 1.50 (m, 5H), 1.37 (m, 4H), 1.23 (m, 4H), 1.12 (m, 4H), 1.01 (m, 1H), 0.96 (s, 3H), 0.92 (d,
1
13
J = 6.3 Hz, 3H), 0.87 (d, J = 2.6 Hz, 3H), 0.86 (d, J = 2.1 Hz, 3H). C NMR (126 MHz, CDCl )
3
δ 167.4, 157.7, 144.0, 129.9, 127.2, 116.2, 115.8, 71.6, 56.4, 51.5, 41.9, 40.0, 39.4, 35.9, 35.4,
3
1
4
0.6, 29.7, 28.0, 27.1, 23.7, 22.8, 22.6, 22.5, 18.6, 18.0, 13.2. IR(neat) vmax 3313, 2953, 2869,
688, 1632, 1587, 1515, 1236, 1170, 1153, 1063, 832. DART-HRMS: m/z calcd. for C H O :
2
7
41
3
+
13.3056 [MH] . Found: 413.3075.
4
.2.5.14. 4-hydroxycoumarate-VD3 ester (32).
1
H NMR (500 MHz, CDCl ) δ 8.02 (d, J = 16.2 Hz, 1H), 7.48 (m, 1H), 7.23 (m, 1H), 6.92 (m,
3
1
H), 6.85 (d, J = 8.0 Hz, 1H), 6.62 (d, J = 16.2 Hz, 1H), 5.32 (m, 1H), 2.05 (m, 1H), 1.94 (m,
1
H), 1.81 (m, 2H), 1.50 (m, 5H), 1. 38 (m, 4H), 1.24 (m, 4H), 1.12 (m, 4H), 1.01 (m, 1H), 0.98
1
3
(
s, 3H), 0.92 (d, J = 6.5 Hz, 3H), 0.87 (d, J = 2.0 Hz, 3H), 0.86 (d, J = 2.0 Hz, 3H). C NMR
(
126 MHz, CDCl ) δ 168.1, 155.2, 140.0, 131.2, 129.1, 121.8, 120.7, 119.1, 116.3, 71.8, 56.5,
3
5
1.5, 41.9, 40.0, 39.4, 35.9, 35.4, 30.6, 29.6, 28.0, 27.1, 23.7, 22.8, 22.7, 22.5, 18.5, 18.0, 13.2.
IR(neat) vmax 3320, 2946, 2928, 2865, 1674, 1602, 1365, 1304, 1274, 1217, 1153, 1108, 1092,
+
9
35, 867, 749. DART-HRMS: m/z calcd. for C H O : 413.3056 [MH] . Found: 413.3077.
2
7
41
3

ACCEPTED MANUSCRIPT
4
.3. Synthesis of inverted este-linked analogue of 4.
.3.1. Initial synthesis of epi-Grundmann’s alcohol (37).
4
The initial series of synthetic steps to prepare 37 (Scheme 3A) was as previously described [17].
The characterization of intermediates prepared via this route that were evaluated for Hh activity
was also as described.
4
.3.2. Direct reduction of Grundmann’s ketone to provide 37.
The reduction of 33 was performed according to the procedure by Blakemore et. al., with the
only modification being exchanging Amberlite IR120 (dried at 0.1 mmHg and 50ºC, 12 h) for
the Amberlite IR118 resin used in the original published protocol [18]. Following completion of
the sodium metal addition, the reaction mixture was quenched cautiously with water (dropwise
addition, 1 mL; 20 mL additional water). The mixture was then filtered and washed accordingly.
The first two chromatographic attempts yielded a mixture of alcohols (100% Hex to 10% EtOAc
in Hex) and after a third attempt (on 800 mg material; 100% Hex to 6% EtOAc in Hex), 37
(45%) and 38 (15%) were isolated as clear oils.
4
.3.3. epi-Grundmann’s alcohol (37).
1
H NMR (500 MHz, CDCl ) δ 3.57 (m, 1H), 1.98 (m, 1H), 1.86 (m, 2H), 1.79 (m, 1H), 1.60 (m,
3
2
H), 1.52 (m, 1H), 1.32 (m, 6H), 1.11 (m, 7H), 1.00 (m, 1H), 0.90 (d, J = 6.4 Hz, 3H), 0.86 (d, J
1
3
=
6.9 Hz, 6H), 0.67 (s, 3H). C NMR (126 MHz, CDCl ) δ 71.1, 57.2, 56.6, 44.6, 39.4, 39.2,
3
3
6.1, 35.9, 35.3, 27.9, 27.9, 23.8, 23.4, 22.7, 22.5, 21.8, 18.5, 11.9. IR(neat) vmax 3335, 2969,
2
932, 2882, 1466, 1407, 1378, 1340, 1305, 1159, 1127, 1107, 950, 816. DART-HRMS: a) m/z
+
calcd. for C H ONH : 284.2953 [M+NH ] . Found: 284.2959. b) m/z calcd. for C H :
1
8
34
4
4
18 33
+
2
49.2582 [M-OH] . Found: 249.2561.

ACCEPTED MANUSCRIPT
4
.3.4. α-epi-Grundmann’s alcohol (38).
1
H NMR (500 MHz, CDCl ) δ 3.89 (m, 1H), 1.97 (m, 1H), 1.84 (m, 1H), 1.58 (m, 8H), 1.44 (m,
3
13
3
H), 1.30 (m, 5H), 1.12 (m, 4H), 0.99 (s, 3H), 0.96 (m, 1H), 0.86 (m, 9H). C NMR (126 MHz,
CDCl ) δ 70.0, 55.7, 51.8, 44.8, 39.5, 36.4, 34.0, 33.4, 29.8, 27.9, 25.3, 24.5, 23.2, 22.7, 22.5,
3
2
1
2
0.7, 20.3, 19.8. IR(neat) vmax 3315, 2995, 2950, 2926, 2867, 1459, 1384, 1349, 1266, 1047,
+
009, 991, 860, 731. DART-HRMS: a) m/z calcd. for C H ONH : 284.2953 [M+NH ] . Found:
1
8
34
4
4
+
84.2957. b) m/z calcd. for C H : 249.2582 [M-OH] . Found: 249.2559.
1
8
33
4
.3.5. 3-benzyloxy benzoate-epi-VD3 ester (39).
Esterification and benzyl deprotection to afford 40 as a clear oil was as described previously
1
(
65%) [14-15]. H NMR (500 MHz, CDCl ) δ 7.64 (m, 1H), 7.45 (m, 1H), 7.39 (m, 1H), 7.33
3
(m, 1H), 7.15 (m, 1H), 5.11 (s, 2H), 5.04 (m, 1H), 2.17 (m, 1H), 1.95 (m, 1H), 1.86 (m, 1H),
1
.69 (m, 3H), 1.56 (m, 2H), 1.41 (m, 1H), 1.30 (m, 8H), 1.14 (m, 4H), 1.02 (m, 1H), 0.94 (d, J =
1
3
6
.4 Hz, 3H), 0.88 (d, J = 1.8 Hz, 3H), 0.87 (d, J = 2.1 Hz, 3H), 0.80 (s, 3H). C NMR (126
MHz, CDCl ) δ 166.0, 158.6, 136.6, 132.3, 129.2, 128.6, 128.0, 127.5, 122.1, 119.6, 115.3, 77.2,
3
7
1
4.9, 70.1, 56.5, 54.3, 45.0, 39.4, 39.1, 36.1, 35.4, 32.3, 28.0, 27.7, 23.8, 23.7, 22.7, 22.5, 21.6,
+
8.6, 11.9. DART-HRMS: a) m/z calcd. for C H O : 477.3369 [MH] . Found: 477.3384. b) m/z
3
2
45
3
+
calcd. for C H O NH : 494.3634 [M+NH ] . Found: 494.3652.
3
2
44
3
4
4
4
.3.6. 3-hydroxy benzoate-epi-VD3 ester (40, epi-4).
1
H NMR (500 MHz, CDCl ) δ 7.58 (m, 1H), 7.54 (s, 1H), 7.29 (m, 1H), 7.04 (m, 1H), 5.54 (bs,
3
1
H), 5.04 (m, 1H), 2.16 (m, 1H), 1.94 (m, 1H), 1.85 (m, 1H), 1.68 (m, 3H), 1.55 (m, 2H), 1.32
(m, 9H), 1.13 (m, 4H), 1.02 (m, 1H), 0.93 (d, J = 6.5 Hz, 3H), 0.88 (d, J = 2.2 Hz, 3H), 0.86 (d, J
1
3
=
2.2 Hz, 3H), 0.78 (s, 3H). C NMR (126 MHz, CDCl ) δ 166.2, 155.7, 132.3, 129.5, 121.8,
3
1
19.8, 116.2, 75.2, 56.5, 54.3, 45.0, 39.4, 39.1, 36.1, 35.4, 32.2, 28.0, 27.7, 23.8, 23.7, 22.8, 22.5,

ACCEPTED MANUSCRIPT
2
1.6, 18.6, 11.9. IR(neat) vmax 3406, 2950, 2925, 2867, 1716, 1687, 1603, 1588, 1454, 1287,
218, 1102, 1072, 973, 962, 755, 681. DART-HRMS: a) m/z calcd. for C H O : 387.2899
1
2
5
39
3
+
+
[
MH] . Found: 387.2894. b) m/z calcd. for C H O NH : 404.3165 [M+NH ] . Found:
25 38 3 4 4
4
04.3158.
4
4
3
.4. Synthesis of ‘linkerless’ analogues.
.4.1. 3-methoxymethyl phenyl-Grundmann’s alcohol (41).
-(MOM)-bromophenol [45] had been prepared by the procedure described above for the MOM-
protection of phenols. 3-(MOM)-bromophenol (2.16 mmoL) was dissolved in THF (20 mL) and
cooled to -78° C. A solution of n-BuLi in hexanes (2.05 mmoL) was added slowly to the aryl
bromide and stirred for 1 h at this temperature and then warmed to 0° C for 15 minutes. A
solution of 33 (1.2 mmoL) in THF (10 mL) was then added to the reaction mixture and stirred
for 12 h. The reaction was quenched by the addition of saturated ammonium chloride (50 mL)
and washed with DCM (3 X 80 mL). The organic layers were combined, dried over Na SO ,
2
4
filtered, and concentrated. The crude residue was purified using column chromatography (SiO2,
1
1
00% Hex to 5% EtOAc in Hex) and 41 was isolated as a clear oil in 89% yield. H NMR (500
MHz, CDCl ) δ 7.23 (m, 1H), 7.13 (m, 1H), 7.06 (m, 1H), 6.89 (m, 1H), 5.18 (s, 2H), 3.49 (s,
3
3
H), 2.08 (m, 1H), 1.97 (m, 1H), 1.73 (m, 3H), 1.55 (m, 3H), 1.41 (m, 1H), 1.33 (m, 2H), 1.27
(m, 1H), 1.22 (m, 2H), 1.13 (m, 5H), 1.03 (s, 3H), 0.94 (d, J = 6.5 Hz, 3H), 0.87 (d, J = 2.3 Hz,
1
3
3
1
2
3
H), 0.86 (d, J = 2.3 Hz, 3H). C NMR (126 MHz, CDCl ) δ 157.2, 150.9, 128.9, 118.1, 113.5,
3
13.1, 94.5, 76.2, 56.8, 56.1, 55.9, 42.9, 40.7, 40.2, 39.5, 35.8, 35.2, 27.9, 26.6, 23.7, 22.7, 22.5,
+
0.0, 19.4, 18.4, 13.2. DART-HRMS: a) m/z calcd. for C H O : 385.3107 [M-OH] . Found:
2
6
41
2
+
85.3078. b) m/z calcd. for C H O NH : 420.3478 [MNH ] . Found: 420.3468. c) m/z calcd.
2
6
42
3
4
4
+
for C H O : 402.3134 [M] . Found: 402.3129.
2
6
42
3

ACCEPTED MANUSCRIPT
4
.4.2. .3-phenol-Grundmann’s alcohol (42).
Removal of the methoxymethyl protecting group to afford phenol 42 as a clear oil in excellent
1
yield (85%) was as described above. H NMR (500 MHz, CDCl ) δ 7.15 (m, 1H), 6.77 (m, 1H),
3
6
2
0
1
1
1
.67 (m, 1H), 6.65 (m, 1H), 4.52 (m, 1H), 2.41 (m, 2H), 2.07 (m, 3H), 1.77 (m, 3H), 1.51 (m,
H), 1.35 (m, 4H), 1.13 (m, 5H), 0.99 (s, 3H), 0.98 (d, J = 6.6 Hz, 3H), 0.87 (d, J = 1.5 Hz, 3H),
1
3
.86 (d, J = 1.5 Hz, 3H). C NMR (126 MHz, CDCl ) δ 154.9, 145.8, 145.7, 128.8, 127.4,
3
20.5, 114.7, 112.7, 55.7, 43.7, 39.5, 37.6, 35.8, 34.8, 30.1, 28.0, 27.2, 27.0, 23.7, 22.7, 22.5,
9.8, 18.9, 18.5. IR(neat) vmax 3375, 2951, 2924, 2853, 1716, 1580, 1485, 1463, 1445, 1367,
278, 1181, 1089, 1024, 904, 866, 781, 699. DART-HRMS: a) m/z calcd. for C H O: 341.2844
2
4
37
+
[
M-OH] . Found: 341.2828.
4
.4.3. 3-methoxymethyl phenyl-(dehydro)VD3 (43).
Tertiary alcohol 41 (0.25 mmoL) was dissolved in anhydrous benzene (8 mL) and stirred at RT
for 1 h in the presence of methyl N-(triethylammoniosulfonyl)carbamate (Burgess reagent, 0.67
mmoL). The reaction mixture was heated to reflux for 2 h. After this time, the reaction was
cooled to RT, diluted with diethyl ether (75 mL) and washed sequentially with saturated sodium
bicarbonate (80 mL) and saturated sodium chloride (100 mL). The organic layer was dried over
MgSO , filtered, and concentrated. Column chromatography (SiO , 100% Hex to 4% EtOAc in
4
2
1
Hex) afforded 43 as a 3:2 mixture of isomers (43, 75%). H NMR (500 MHz, CDCl ) δ 7.19 (m,
3
1
2
H), 6.88 (m, 4H), 5.66 (m, 1H), 5.18 (s, 3H), 3.50 (s, 4H), 2.59 (m, 1H), 2.45 (m, 1H), 2.28 (m,
H), 2.09 (m, 2H), 1.94 (m, 1H), 1.80 (m, 2H), 1.53 (m, 3H), 1.37 (m, 7H), 1.13 (m, 7H), 1.00
1
3
(
m, 6H), 0.89 (m, 9H), 0.78 (s, 3H). C NMR (126 MHz, CDCl ) δ 156.8, 145.8, 145.4, 144.5,
3
1
4
40.2, 128.6, 127.5, 125.1, 121.5, 120.6, 115.8, 114.9, 113.8, 113.5, 94.5, 55.9, 55.7, 54.4, 49.9,
3.7, 42.6, 39.5, 37.6, 36.1, 36.0, 35.8, 34.8, 30.1, 28.3, 28.0, 27.2, 27.0, 25.0, 24.3, 23.8, 23.7,