6-Substituted quinolines as RORγt inverse agonists

Article abstract of DOI:10.1016/j.bmcl.2017.10.027

We identified 6-substituted quinolines as modulators of the retinoic acid receptor-related orphan receptor gamma t (RORγt). The synthesis of this class of RORγt modulators is reported, and optimization of the substituents at the quinoline 6-position that

Full text of DOI:10.1016/j.bmcl.2017.10.027

Accepted Manuscript
6-Substituted Quinolines as RORγt Inverse Agonists
J. Kent Barbay, Maxwell D. Cummings, Marta Abad, Glenda Castro, Kevin D.
Kreutter, David A. Kummer, Umar Maharoof, Cynthia Milligan, Rachel
Nishimura, Joan Pierce, Celine Schalk-Hihi, John Spurlino, Virginia M. Tanis,
Maud Urbanski, Hariharan Venkatesan, Aihua Wang, Craig Woods, Ronald
Wolin, Xiaohua Xue, James P. Edwards, Anne M. Fourie, Kristi Leonard
PII:
S0960-894X(17)31012-0
https://doi.org/10.1016/j.bmcl.2017.10.027
BMCL 25356
DOI:
Reference:
Bioorganic & Medicinal Chemistry Letters
To appear in:
Received Date:
Revised Date:
Accepted Date:
12 June 2017
12 October 2017
14 October 2017
Please cite this article as: Barbay, J.K., Cummings, M.D., Abad, M., Castro, G., Kreutter, K.D., Kummer, D.A.,
Maharoof, U., Milligan, C., Nishimura, R., Pierce, J., Schalk-Hihi, C., Spurlino, J., Tanis, V.M., Urbanski, M.,
Venkatesan, H., Wang, A., Woods, C., Wolin, R., Xue, X., Edwards, J.P., Fourie, A.M., Leonard, K., 6-Substituted
Quinolines as RORγt Inverse Agonists, Bioorganic & Medicinal Chemistry Letters (2017), doi: https://doi.org/
10.1016/j.bmcl.2017.10.027
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.

6-Substituted Quinolines as RORγt Inverse Agonists
Authors: J. Kent Barbay*,c, Maxwell D. Cummings*,b, Marta Abadb, Glenda Castroa, Kevin D. Kreutterc,
David A. Kummera, Umar Maharoofc, Cynthia Milliganb, Rachel Nishimuraa, Joan Piercea, Celine
Schalk-Hihib, John Spurlinob, Virginia M. Tanisa, Maud Urbanskic, Hariharan Venkatesana, Aihua Wangc,
Craig Woodsa, Ronald Wolina, Xiaohua Xuea, James P. Edwardsa, Anne M. Fouriea, Kristi Leonardc
Corresponding Authors: kbarbay@its.jnj.com (J. K. Barbay), mcummin1@its.jnj.com (M. D.
Cummings)
Addresses:
aDiscovery Immunology, Janssen Research and Development, LLC, 3210 Merryfield Row, San Diego, CA
92121
bDiscovery Sciences, Janssen Research and Development, LLC, Welsh and McKean Roads, Spring
House, PA 19477
cDiscovery Immunology, Janssen Research and Development, LLC, Welsh and McKean Roads, Spring
House, PA 19477
Keywords. RORγt, retinoic acid-related orphan nuclear receptor gamma t, Th17, IL-17
Abstract: We identified 6-substituted quinolines as modulators of the retinoic acid receptor-
related orphan receptor gamma t (RORγt). The synthesis of this class of RORγt modulators is
reported, and optimization of the substituents at the quinoline 6-position that produced
compounds with high affinity for the receptor is detailed. This effort identified molecules that
act as potent, full inverse agonists in a RORγt-driven cell-based reporter assay. The X-ray
crystal structures of two full inverse agonists from this chemical series bound to the RORγt
ligand binding domain are disclosed, and we highlight the interaction of a hydrogen-bond

acceptor on the 6-position substituent of the inverse agonist with Glu379:NH as a conserved
binding contact.
Retinoic acid receptor-related orphan receptor gamma t (RORγt) is a nuclear hormone
receptor expressed in a variety of cell types of the innate and adaptive immune system.1 This
transcription factor drives Th17 cell differentiation and the production of cytokines including
interleukin (IL)-17A and IL-17F.2 Th17 cells and IL-17 have been shown to play an important
role in the development of autoimmune disorders including psoriasis, multiple sclerosis,
rheumatoid arthritis, Crohn’s disease, and systemic lupus erythematosus.3 RORγt-deficient mice
show significantly reduced Th17 cell populations and decreased susceptibility to experimental
autoimmune encephalomyelitis as well as intestinal and skin inflammation.1,4 Clinical efficacy
with biologics targeting IL-17A, or IL-23, the cytokine that stabilizes Th17 cells, in psoriasis and
psoriatic arthritis has served to validate the critical role of IL-17A, IL-23 and Th17 cells in these
autoimmune diseases.5 Targeting RORγt, the key regulator in the Th17/IL-17 pathway, provides
a novel opportunity to treat autoimmune diseases with a small molecule drug. RORγt contains a
ligand binding domain (LBD) capable of binding small molecules that influence interactions
with co-activator and co-repressor proteins and thereby modulate gene transcription. Many
recent reports have detailed the development of RORγt modulators.6-9
We conducted a high-throughput screen to identify RORγt modulators using
ThermoFluor®, a fluorescence-based, thermal shift assay.10 Among compounds found to bind to
the LBD of RORγt was quinoline 1 (Kd 160 nM). The HTS sample was a racemate that
displayed full inverse agonism in a RORγt driven cell-based (1-hybrid) reporter assay (IC50 260

nM, 101% inhibition at 6 µM).11 Structure-activity relationships (SAR) and variations in the
mechanism of action (agonists, inverse agonists, and neutral antagonists) established during
optimization of the substituents at the 2-, 3-, and 4-positions on the quinoline core have been
described.12 Herein, we report the synthesis of compounds with changes to the substituents on the
carbon bearing the tertiary alcohol group of 1 that led to a series of full inverse agonists with
improved potency. We also describe co-crystal structures of two ligands bound to the RORγt
ligand binding domain that provide structural insights into the SAR discussed in this paper.
Figure 1. Structure of high-throughput screening hit 1.
Initially, the two aryl groups attached to the carbon atom bearing the tertiary alcohol were
explored while retaining the other structural features of the screening hit. In our initial
exploratory hit-to-lead studies, the early compounds in this series (Table 1) were tested as
racemic mixtures, while more potent analogs prepared later were resolved into single
enantiomers as detailed below. 1,2-Dimethyl imidazole 2 retained comparable binding affinity
to that of 1, while expansion of the N-methyl substituent to N-ethyl led to decreased affinity (3,
Kd 450 nM). A series of pyridines paired with the N-methyl imidazole group was also explored.
Both 4-pyridyl (4, Kd 14 nM) and 3-pyridyl analogs (5, Kd 38 nM) displayed improved binding
affinity and activity in the 1-hybrid cell-based reporter assay (IC50’s 50 nM and 140 nM,

respectively). These compounds showed considerably stronger binding affinities than the
corresponding 2-pyridyl analog (6, 960 nM).
Consistent with a specific intermolecular interaction, binding affinity and cellular
potency were sensitive to hydrogen bond acceptor position in the aryl substituents on the carbon
bearing the tertiary alcohol. For example, 1-methyl-imidazol-5-yl analog 7 displayed 5 nM
affinity for the receptor and an IC50 of 21 nM in the 1-hybrid assay, whereas the 1-methyl-
imidazol-2-yl isomer 8, with the imidazole nitrogen shifted by one position, showed a 4,000-fold
reduction in binding affinity (Kd 20,000 nM). Similarly, the 3-pyridyl analog 9 displayed a
binding affinity of 76 nM and an IC50 of 130 nM in the cell assay, while the corresponding 4-
pyridyl analog 10 displayed 20-fold weaker affinity and a corresponding reduction in cell
potency (Table 1).
Compounds 11-15 represent a series of analogs in which one of the aryl groups is fixed as
1-methyl-imidazol-5-yl. High levels of potency were achieved by incorporation of ortho-
substituted 4- or 3-pyridyls, e.g. m-methyl-4-pyridyl inverse agonist 11, which had a Kd of 0.6
nM and 1-hybrid assay IC50 of 13 nM. The protein is also able to accommodate larger fused
aryls in this position, as indicated by quinoline 14 (Kd 0.5 nM, 1-hybrid IC50 5.5 nM), as well as
5-membered heterocycles, as exemplified by thiazole 15 (Kd 1.1 nM, 1-hybrid IC50 100 nM).
Table 1. Binding and cell-based functional data of 3-phenyl-2,4-dichloroquinolines.a

RORγt cell-based
reporter assay
IC50 (nM)
ThermoFluor®
binding to RORγt
LBD, Kd (nM)
Cmpd
#
Ar1
Ar2
(% inhibition
at 6 µM)
1
2
160
190
260 (101)
600 (106)
3
4
5
450
14
820 (101)
50 (105)
140 (103)
38
6
7
960
5.3
1000 (93)
21 (103)
8
9
20000
76
>6,000 (37)
130 (100)
10
11
1500
0.60
2100 (84)
13 (104)

12
6.4
82 (101)
13
14
31
180 (100)
5.5 (101)
0.50
N
15
1.1
100 (104)
a Compounds in Table 1 were tested as racemic mixtures.
Screening hit 1 was separated into individual enantiomers 1a and 1b by chiral
chromatography. The enantiomers exhibited a 180-fold difference in binding affinity (32 nM vs
5700 nM, respectively, Table 2). This result is consistent with receptor modulation by a specific
ligand/receptor interaction and emphasizes the importance of an optimally positioned hydrogen-
bond acceptor in at least one of the aryl groups flanking the tertiary alcohol (e.g., the 3-position
nitrogen of the imidazole in only one of the enantiomers of 1).
Table 2. Activities of single enantiomers of screening hit 1.

RORγt cell-based
reporter assay IC50 (nM)
(% inhibition
ThermoFluor®
binding to RORγt
LBD, Kd (nM)
Cmpd #
at 6 µM)
1a
160
32
260 (101)
220 (100)
∼5500 (58)
1ab
1bb
5700
a Racemate.
b Single enantiomer, absolute stereochemistry not determined.
Subsequent racemic mixtures were separated by chiral chromatography to provide single
enantiomers for additional SAR exploration in this series. This work was carried out on the 2-
methoxy-3-(1-pyrazole-benzyl)-4-chloro scaffold and is detailed in Table 3. In cases where only
one enantiomer is presented in Table 3, it is the more potent isomer with regard to binding to the
RORγt LBD as measured in the ThermoFluor® assay. The 2-methoxy group was introduced to
remove potential chemical reactivity associated with 2-chloroquinolines, and a substituted benzyl
group at the 3-position yielded potent, full inverse agonists.12 As previously noted, large
differences in the binding potencies of enantiomers were sometimes observed. For example, in
this series, compound 16a (Kd 9.2 nM) bearing the original aryl substituents (1-methylimidazol-
5-yl and p-chlorophenyl) was 57-fold more potent than the less active enantiomer 16b. Analogs
17 and 18 explored the impact of replacing the tertiary alcohol with amino and methylamino
groups, respectively. Replacement of the alcohol with an amino group resulted in an
approximately 12-fold decrease in binding affinity (Kd 110 nM), and N-methylation of the amino
group further reduced binding (Kd 280 nM). The ortho-trifluoromethyl 3-pyridyl enantiomers
19a and 19b delivered strong potency, with a more modest difference in potency seen between

enantiomers (14-fold). Returning to combinations of two 5-membered heterocycles yielded
potent enantiomers 20a and 20b, with 1,2-dimethylimidazole and N-methyltriazole groups
flanking the tertiary alcohol. These cases illustrate that systems in which both heteroaryls have a
suitably positioned hydrogen bond acceptor can yield highly potent compounds for both
enantiomers. The observation that both enantiomers 20a and 20b were active suggested to us
that exploration of symmetrical, achiral compounds bearing two identical 5-membered
heterocycles might be productive. To test this, achiral molecules bis-1-methyltriazole 21 and
bis-1,2-dimethylimidazole 22 were synthesized, and, as expected, they demonstrated high
binding affinities (Kd 5.0 nM and 0.27 nM, respectively). Interestingly, both enantiomers of
secondary alcohols 23a and 23b also bound to the receptor with approximately equal potency
(Kd’s 69-91 nM), in contrast to other examples of enantiomeric pairs containing only one H-bond
acceptor (e.g. 1a/b and 16a/b). This may reflect a somewhat different binding mode or plasticity
in the receptor pocket when a ring is removed, and indicates the potential for development of
potent analogs with reduced molecular weights.
Compound 24a incorporates an N-acetyl-piperidine group as replacement for one of the
aryl groups and retains potent binding affinity (Kd 20 nM) and cellular activity (1-hybrid IC50 46
nM). The enantiomer, 24b, retains high potency and reflects the ability of the N-acetyl group to
act as an H-bond acceptor, as subsequently confirmed by X-ray crystallography (vide supra). By
contrast, replacement of the N-acetylpiperidine ring with phenyl in compounds 25a and 25b
again leads to a large divergence in the enantiomer potency (21-fold) in line with the SAR
discussed above.
Table 3. Binding and cell-based functional data of 2-methoxy-3-(1-pyrazole-benzyl)-4-
chloroquinolines.a

RORγt cell-based
reporter assay IC50
(nM) (% inhibition
at 6 µM)
ThermoFluor®
binding to RORγt
LBD, Kd (nM)
Cmpd
#
Ar1
Ar2
X
16a
16b
17
OH
OH
9.2
520
110
280
5.0
70
73 (99)
670 (87)
210 (97)
630 (90)
89 (97)
NH2
NHMe
OH
18
19a
19b
OH
170 (99)
20a
OH
1.2
13 (100)
20b
OH
OH
OH
0.38
5.0
5.3 (100)
6.6 (91)
200 (96)
21
22
0.27

23a
23b
H
H
OH
OH
69
91
270 (98)
300 (97)
24a
24b
25a
OH
OH
OH
20
33
46 (100)
59 (107)
24 (100)
3.6
25b
OH
77
140 (103)
a
All chiral compounds are single enantiomers. Absolute stereochemistry is unknown
unless specifically noted in the text.
Unlike changes previously described to the quinoline 3-position substituent,12 compounds
described in this paper uniformly displayed essentially full inverse agonism, as measured by
percent inhibition of gene transcription in the 1-hybrid cell-based functional reporter assay at an
inhibitor concentration of 6 µM. Except in cases of weak inverse agonists displaying potency in
the µM range, ligands with varied substituents on the benzylic alcohol carbon displayed close to
100% inhibition at high ligand concentration.
Molecules in Tables 1-3 were prepared by the coupling of a diaryl ketone with a 6-
lithioquinoline reagent, the latter formed initially by a halogen-lithium exchange reaction
(Scheme 1).13 6-Bromo-2,4-dichloro-3-phenylquinoline III was prepared sequentially by
acylation of 5-bromo-methyl 2-aminobenzoate I with 2-phenylacetyl chloride, lithium
hexamethyldisilazane induced cyclization, and reaction of the resultant 4-hydroxyquinolinone

with phosphorus oxychloride. 2-Methoxy-3-(1-pyrazole-benzyl)-quinoline IV, used in the
synthesis of compounds in Table 3, was prepared as previously described.12 In the case of
secondary alcohol inverse agonists (23a/b), ketone V was replaced with an aldehyde in the final
step.14
Scheme 1. (a) PhCH2COCl, Et3N, DCM; (b) LiHMDS, THF (c) POCl3, reflux; (d) n-BuLi,
THF.
The diaryl ketones used above were prepared by the general routes shown in Scheme 2.
5-Bromo-imidazoles VII were reacted with alkyl magnesium halides to generate Grignard
reagents that were reacted with either a Weinreb amide to directly produce ketones VIII or with
an aldehyde to produce a secondary alcohol which was then oxidized to the ketone. Symmetrical
ketones were made by the reaction of heterocyclic anions with ethyl-N-(methoxy)-N-
(methyl)carbamate, as illustrated in Scheme 2 for ketone X.15

Scheme 2. (a) EtMgBr, THF; Ar2-CONMe(OMe); (b) i-PrMgCl-LiCl, THF, Ar2CHO; MnO2,
1,4-dioxane, reflux; (c) n-BuLi, THF.
Compounds 17 and 18, bearing amines at the benzylic position, were synthesized from
the corresponding tertiary alcohol by acetylation of the alcohol and treatment with methanolic
ammonia or methylamine (Scheme 3).
Scheme 3. (a) NaH, Ac2O, DMF; (b) NH3, MeOH; (c) CH3NH2, MeOH.

Figure 2. Crystal structures of RORγt ligand binding domain complexes. A-C: RORγt LBD,
grey ribbon diagram; A: 25-hydroxycholesterol, green color-by-atom sticks (PDB ID: 3L0L)16;
B: compound 19b, orange color-by-atom sticks (PDB ID 5W4V); C: compound 24b, purple
color-by atom sticks (PDB ID 5W4R). Compounds 19b and 24b occupy the central 25-
hydroxycholesterol binding pocket. Note the absence of H11’ and H12 in B and C; binding of
effectors of this and other chemotypes has been observed to lead to disorder in this region of the
LBD.12,17
The crystal structure of 19b bound to the LBD of human RORγt was determined at 3.0Å
resolution. The absolute stereochemistry of 19b was observed to be R, and, similar to our
previous report on this chemical series,12 the bound ligand occupies the buried and largely
hydrophobic sterol binding pocket (Figure 2).16,17 In our earlier report we focused on
substituents at the 3-position of the quinoline core, while the present work addresses the tertiary
alcohol and pendant rings of this quinoline chemotype.

In the observed binding mode, the 4-chloro-2-methoxyquinoline core of 19b is
positioned between the sidechains of Met365 and Phe378, forming an edge-to-face π-π
interaction with the latter. The trifluoromethylpyridine ring is sandwiched between three
aromatic sidechains, making a face-to-face contact with His323 and edge-to-face interactions
with Phe377 and Phe378. The pyridine nitrogen accepts a hydrogen bond from the backbone
amide of Glu379 (Figure 3). Modeling of a likely binding mode for 19a indicates that a similar
hydrogen bonding interaction is possible for the nitrogen of the N-methylimidazole of the other
enantiomer (not shown), consistent with the high affinity observed for both enantiomers (19a/b
in Table 3).
Figure 3. Crystal structure of compound 19b bound to the RORγt ligand binding domain.
RORγt LBD: grey ribbon and grey color-by-atom; compound 19b: orange color-by-atom; bound
water molecules (red spheres) are from the published 25-hydroxycholesterol complex structure,16
and the sidechain conformations for Gln286 and Arg367 from this sterol complex are also shown
(magenta color-by-atom) relative to the respective sidechains from the LBD-19b complex
structure; selected residues are shown in stick format; the distance for one direct intermolecular

hydrogen bond is shown, in Å; selected contacts involving the modeled bound water molecules
are indicated with dashed lines.
Comparison of the present structure with the published crystal structure of the 25-
hydroxycholesterol agonist complex18 offers some insight into the specific contribution of the
tertiary hydroxyl group. Unlike the LBD-19b complex structure, bound water molecules are
resolved in the higher resolution crystal structure of the sterol complex. Overlay of bound water
molecules from the sterol complex onto the LBD-19b complex indicates that the ligand hydroxyl
group may form part of a hydrogen bond network that involves bound water molecules as well as
RORγt (Figure 3). Similar analysis suggests that the distal heteroatom of the N-methylimidazole
may be forming a water-bridged intermolecular hydrogen bond with the guanidine of Arg367
(Figure 3). Both of these modeled bridging water contacts are consistent with the available
structural information19 as well as with the observed SAR for this chemotype (Table 3, Figure 3).

Figure 4. Crystal structure of compound 24b bound to the RORγt ligand binding domain.
RORγt LBD: grey ribbon and grey color-by-atom; compound 24b: purple color-by-atom. As
described for Figure 3, bound water molecules (red spheres) are from the published 25-
hydroxycholesterol complex structure,16 and the sidechain conformations for Gln286 and Arg367
from this sterol complex are also shown (magenta color-by-atom) relative to the respective
sidechains from the LBD-24b complex structure; selected residues are shown in stick format; the
distances for two direct intermolecular hydrogen bonds are shown, in Å; selected contacts
involving the modeled bound water molecules are indicated with dashed lines.
The X-ray crystal structure of 24b bound to the RORγt LBD was also determined to 3.0Å
resolution. As for compound 19b above, the R-enantiomer of 24b was observed bound in this
complex. The N-acetylpiperidine moiety of 24b replaces the trifluoromethylpyridine of 19b.
The carbonyl of this acetyl group effectively mimics the pyridine nitrogen of 19b, again forming
an intermolecular hydrogen bond with the backbone amide of Glu379 (Figure 4). As previously
determined for 19a and 19b, the corresponding enantiomer (24a) maintains high potency, further
supporting the idea that the N-methylimidazole is able to engage in the same hydrogen bonding
contact with the backbone of Glu379. The tertiary alcohol in this structure is rotated ~180°
relative to the complex with 19b (Figure 5), in this case forming a direct intermolecular
hydrogen bond with the carbonyl oxygen of Phe377 (Figures 4 and 5). From more than a dozen
protein-ligand complex structures that we have determined for this chemical series (not shown),
the binding orientation for the hydroxyl group as seen in the complex with 19b was observed

most frequently. Despite the reorientation of the hydroxyl group, the two rings attached at the
benzylic carbon in 19b and 24b occupy similar regions of the receptor and the overall binding
mode is conserved (Figures 3-5). We see that different functional groups from the two ligands
(19b and 24b) form an intermolecular hydrogen bond with the same Glu379:NH donor. It may
be that this specific interaction drives the alternate orientation of the tertiary alcohol observed in
the complex with 24b. These structural observations, in conjunction with the SAR data
presented above showing strong sensitivity of binding affinity to hydrogen bond acceptor atom
placement in pendant rings at this position, suggest that the contact with Glu379:NH is an
important interaction within this class of RORγt modulators.
Figure 5. Superimposed RORγt binding conformations of 19b and 24b. 19b: orange color-by-
atom; 24b: purple color-by-atom; Phe378 and Glu379 of the LBD from the corresponding crystal
structures are shown, colored as 19b and 24b; distances shown, between the indicated ligand
atoms and Glu379:N of the LBD from the relevant complex structure, are in Å. The overlay is
based on superposition of the LBD in the crystal structures of the two complexes.

In summary, a high-throughput screen identified a series of 6-substituted quinolines that
bind to RORγt. Optimization of groups flanking the tertiary alcohol at this position led to the
discovery of compounds that bind to the receptor with high affinity and are full inverse agonists
in a cell-based reporter assay. We have reported the co-crystal structures of two compounds with
the RORγt ligand binding domain and described key interactions observed between potent
inverse agonists and the receptor. Further characterization of compounds in this series will be
described in additional reports.
Acknowledgment
Dr. Ulrike Hars (Crelux GmbH) determined the crystal structure of 19b complexed to RORγt
LBD.
References and Notes
1. Jetten, A. M. Nucl. Recept. Signal. 2009, 7, e003.
2. Ivanov, I. I.; McKenzie, B. S.; Zhou, L.; Tadokoro, C. E.; Lepelley, A.; Lafaille, J. J.; Cua, D.
J.; Littman, D. R. Cell 2006, 126, 1121-1133.
3. Tabarkiewicz, J.; Pogoda, K.; Karczmarczyk, A.; Pozarowski, P.; Giannopoulos, K. Arch.
Immunol. Ther. Exp. 2015, 63, 435-449.

4. (a) Withers, D. R.; Hepworth, M. R.; Wang, X.; Mackley, E. C.; Halford, E. E.; Dutton, E. E.;
Marriott, C. L.; Brucklacher-Waldert, V.; Veldhoen, M.; Kelsen, J.; Baldassano, R. N.;
Sonnenberg, G. F. Nature Medicine 2016, 22, 319-325. (b) Xue, X.; Soroosh, P.; De Leon-
Tabaldo, A.; Luna-Roman, R.; Sablad, M.; Rozenkrants, N.; Yu, J.; Castro, G.; Banie, H.; Fung-
Leung, W.-P.; Santamaria-Babi, L.; Schlueter, T.; Albers, M.; Leonard, K.; Budelsky, A. L.;
Fourie, A. M. Sci. Rep. 2016, 6, 37977.
5. (a) Campa, M.; Mansouri, B.; Warren, R.; Menter, A. Dermatol. Ther. 2016, 6, 1-12. (b)
Fragoulis, G. E.; Siebert, S.; McInnes, I. B. Annu. Rev. Med. 2016, 67, 27.1-27.17.
6. Kamenecka, T. M.; Lyda, B.; Chang, M. R.; Griffin, P. R. Med. Chem. Commun. 2013, 4, 764-
776.
7. Solt, L. A.; Burris, T. P. Trends Endocrin. Metabol. 2012, 23, 619-627.
8. Solt, L.A.; Griffin, P. R.; Burris, T. P. Curr. Opin. Lipidology 2010, 21, 204-211.
9. Fauber, B. P.; Magnuson, S. J. Med. Chem. 2014, 57, 5871-5892.
10. (a) Pantoliano, M. W.; Petrella, E. C.; Kwasnoski, J. D.; Lobanov, V. S.; Myslik, J.; Graf,
E.; Carver, T.; Asel, E.; Springer, B. A.; Lane, P.; and Salemme, F. R. J. Biomol. Screen. 2001,
6, 429-40. (b) Matulis, D.; Kranz, J. K.; Salemme, F. R.; Todd, M. J. Biochemistry 2005, 44,
5258-5266.
11. For assay details, see: Barbay, J. K.; Edwards, J. P.; Kreutter, K. D.; Kummer, D. A.;
Maharoof, U.; Nishimura, R.; Urbanski, M.; Venkatesan, H.; Wang, A.; Wolin, R. L.; Woods, C.
R.; Fourie, A.; Xue, X.; Cummings, M. D.; Leonard, K. A. PCT Int. Appl. WO 2015057203.

12. Kummer, D. A.; Cummings, M. D.; Abad, M.; Barbay, J.; Castro, G.; Wolin, R.; Kreutter, K.
D.; Maharoof, U. M.; Milligan, C.; Nishimura, R.; Pierce, J.; Schalk-Hihi, C.; Spurlino, J.;
Urbanski, M.; Venkatesan, H.; Wang, A.; Woods, C.; Xue, X.; Edwards, J. P.; Fourie, A. M.;
Leonard, K. Bioorg. Med. Chem. Lett. 2017, 27, 2047-2057.
13. For additional synthetic details, see reference 11 and: Leonard, K. A.; Barbay, J. K.;
Edwards, J. P.; Kreutter, K. D.; Kummer, D. A.; Maharoof, U.; Nishimura, R.; Urbanski, M.;
Venkatesan, H.; Wang, A.; Wolin, R. L.; Woods, C. R.; Fourie, A.; Xue, X.; Cummings, M. D.;
Jones, W. M.; Goldberg, S. PCT Int. Appl. WO 2015057205.
14. Leonard, K. A.; Barbay, K.; Edwards, J. P.; Kreutter, K. D.; Kummer, D. A.; Maharoof, U.;
Nishimura, R.; Urbanski, M.; Venkatesan, H.; Wang, A.; Wolin, R. L.; Woods, C. R.; Fourie, A.;
Xue, X.; Cummings, M. D. PCT Int. Appl. WO2015057206.
15. Hlasta, D. J.; Court, J. J. Tet. Lett. 1989, 30, 1773-1776.
16. Jin, L., Martynowski, D., Zheng, S., Wada, T., Xie, W., Li, Y. Mol Endocrinol 2010, 24,
923-929.
17. Part of helix 10/11 and all of helices 11’ and 12 have been observed to be non-helical and/or
disordered in crystal structures of effector-LBD complexes, including complexes with
representatives from the present chemical series described in our earlier report (reference 12, and
references therein). This structural observation seems generally consistent with effector
mechanism, and has been observed with LBD protein constructs that include the complete LBD
such as the 265-507 construct of our standard conditions (e.g. 24b-RORγt LBD and structures in
reference 12). The 19b-RORγt LBD complex was generated under non-standard conditions
involving treatment with V8 protease (Staphylococcus aureus endoproteinase Glu-C) that likely

led to truncation of the LBD C-terminus. In the 19b-RORγt LBD complex structure reported
here, the absence of H11’ and H12 and the non-helicity of the observed C-terminus are likely
explained by proteolytic cleavage at Glu481, near the H10/11 midpoint, rather than by a ligand-
induced conformational change. However, the observed binding mode of 19b is consistent with
that of other members of this chemical series (e.g. 24b, structures in reference 12, additional
unpublished results), and the region of interest for the present analysis is at the opposite end of
the binding site and does not include the (possibly) affected secondary structure elements.
18. At 3.0Å resolution we are unable to observe bound waters in the present structure, but they
are observed in the higher resolution crystal structure of the agonist 25-hydroxycholesterol
complexed with RORγ (reference 16).
19. The bound water nearest to the ligand hydroxyl is also observed in the digoxin complex
structure: Fujita-Sato, S., Ito, S., Isobe, T., Ohyama, T., Wakabayashi, K., Morishita, K., Ando,
O. Isono, F. J. Biol. Chem. 2011, 286, 31409-31417. Also, nearby sidechains that interact with
two of the three bound waters displayed in Figures 3 and 4 maintain the same rotamers in the
LBD-19b and LBD-24b complex structures as those observed in the published 25-
hydroxycholesterol complex structure (reference 16).

Figure 1. Structure of high-throughput screening hit 1.
Figure 2. Crystal structures of RORγt ligand binding domain complexes. A-C: RORγt LBD,
grey ribbon diagram; A: 25-hydroxycholesterol, green color-by-atom sticks (PDB ID: 3L0L)16;
B: compound 19b, orange color-by-atom sticks (PDB ID 5W4V); C: compound 24b, purple
color-by atom sticks (PDB ID 5W4R). Compounds 19b and 24b occupy the central 25-
hydroxycholesterol binding pocket. Note the absence of H11’ and H12 in B and C; binding of
effectors of this and other chemotypes has been observed to lead to disorder in this region of the
LBD.12,17
Figure 3. Crystal structure of compound 19b bound to the RORγt ligand binding domain.
RORγt LBD: grey ribbon and grey color-by-atom; compound 19b: orange color-by-atom; bound
water molecules (red spheres) are from the published 25-hydroxycholesterol complex structure,16
and the sidechain conformations for Gln286 and Arg367 from this sterol complex are also shown
(magenta color-by-atom) relative to the respective sidechains from the LBD-19b complex
structure; selected residues are shown in stick format; the distance for one direct intermolecular
hydrogen bond is shown, in Å; selected contacts involving the modeled bound water molecules
are indicated with dashed lines.
Figure 4. Crystal structure of compound 24b bound to the RORγt ligand binding domain.
RORγt LBD: grey ribbon and grey color-by-atom; compound 24b: purple color-by-atom. As
described for Figure 3, bound water molecules (red spheres) are from the published 25-
hydroxycholesterol complex structure,16 and the sidechain conformations for Gln286 and Arg367
from this sterol complex are also shown (magenta color-by-atom) relative to the respective
sidechains from the LBD-24b complex structure; selected residues are shown in stick format; the

distances for two direct intermolecular hydrogen bonds are shown, in Å; selected contacts
involving the modeled bound water molecules are indicated with dashed lines.
Figure 5. Superimposed RORγt binding conformations of 19b and 24b. 19b: orange color-by-
atom; 24b: purple color-by-atom; Phe378 and Glu379 of the LBD from the corresponding crystal
structures are shown, colored as 19b and 24b; distances shown, between the indicated ligand
atoms and Glu379:N of the LBD from the relevant complex structure, are in Å. The overlay is
based on superposition of the LBD in the crystal structures of the two complexes.

Table 1. Binding and cell-based functional data of 3-phenyl-2,4-dichloroquinolines.a
RORγt cell-based
ThermoFluor®
reporter assay
IC50 (nM)
Cmpd
#
Ar1
Ar2
binding to RORγt
(% inhibition
at 6 µM)
LBD, Kd (nM)
1
2
160
190
260 (101)
600 (106)
3
4
5
450
14
820 (101)
50 (105)
140 (103)
38
6
7
960
5.3
1000 (93)
21 (103)
8
20000
>6,000 (37)