Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2025 Oct 7;129(41):10621–10630. doi: 10.1021/acs.jpcb.5c02361

Impact of Bile Acid Amidation on In Silico Farnesoid X Receptor Dynamics

Riley K Eisert-Sasse †, Arumay Biswas †, C Denise Okafor †,‡,*
PMCID: PMC12536397  PMID: 41054796

Abstract

While the behavior of amidated bile acids has been well studied in vitro, and there is a growing body of in silico research on bile acids, the effects of amidated bile acids on the dynamics and behavior of the farnesoid X receptor (FXR) are yet to be characterized. Although amidated bile acids are larger and more hydrophobic than nonamidated bile acids, their in vitro functions can often be remarkably similar. To investigate the impact of these changes on protein behavior, classical molecular dynamics simulations were performed on the ligand-binding domain of FXR bound to the primary human bile acids and their glycine- and taurine-conjugated amidates. We observe that amino acid conjugation leads to changes in critical local interactions, including salt bridges in the binding pocket and contacts at the FXR aromatic triad. Despite these effects, the global dynamic behavior of the receptor remains surprisingly consistent across all of the ligands studied. The absence of large perturbations in FXR dynamics appears to be consistent with the reports of similar in vitro activity of glyco- and tauroconjugates with primary bile acids. Moreover, these results suggest that global versus local differences in FXR dynamics may correlate differently with experimental measurements.

graphic file with name jp5c02361_0008.jpg

graphic file with name jp5c02361_0006.jpg

Introduction

Nuclear receptors are ligand-regulated transcription factors consisting of a disordered N-terminal domain, a DNA-binding domain (DBD), and a ligand-binding domain (LBD). They exhibit ligand-dependent activation; transcription of target genes is up- or downregulated as a result of lipophilic ligands binding to the LBD. Nuclear receptor LBDs share a conserved structure, comprised of 12 alpha helices that enclose a hydrophobic binding pocket (Figure A). The LBD contains a region called the activation function surface 2 (AF-2), composed of helices 3, 4, and 12. This surface forms a hydrophobic groove, which is activated upon ligand binding, allowing for the recruitment of a coactivator protein. , The farnesoid X receptor (FXR) is a nuclear receptor that is highly expressed in the liver and small intestine. It is responsible for regulating the metabolism of cholesterol and cholesterol metabolites, which include endogenous FXR ligands known as bile acids. FXR regulates over 40 genes involved in these metabolic pathways, making it a popular target for drug development.

1.

1

Open in a new tab

Structure of FXR and bile acids. (A) Structure of the farnesoid X receptor (FXR) ligand binding domain with helices labeled and ligand in ligand-binding pocket (LBP). Structure adopted from PDB: 6HL1. (B) Structure of bile acid scaffold with modifications defining different ligands used in this study. CA: cholic acid, CDCA: chenodeoxycholic acid, GCA: glyco-cholic acid, GCDCA: glyco-chenodeoxycholic acid, TCA: tauro-cholic acid, TCDCA: tauro-chenodeoxycholic acid (C) biosynthetic route of bile acid amidation in humans. Colored in blue is the oxygen atom (labeled O24) of the carboxylic acid tail that remains unmodified by amidation.

Bile acids are cholesterol derivatives synthesized in the liver. Bile acids are implicated in numerous physiological processes. They facilitate the absorption and digestion of lipids and fat-soluble vitamins, help to emulsify cholesterol with bile in the gallbladder to prevent gallstones, stimulate the movement of bile from the liver to the gallbladder, and maintain systemic homeostasis of cholesterol when excreted. It follows that the disruption of the metabolic cycle of bile acids is associated with numerous health problems including gallstones, nonalcoholic fatty liver disease, and reduced absorption of lipids and lipid-soluble nutrients.

The primary bile acids in humans are cholic acid (CA) and chenodeoxycholic acid (CDCA), which are direct products of cholesterol metabolism. , While CDCA is the most potent endogenous FXR ligand, CA does not activate FXR in luciferase assays in the absence of a bile acid transporter. In the lower gastrointestinal tract, a majority of primary bile acids are modified to become secondary bile acids. The most common secondary bile acids in humans are deoxycholic acid (DCA) and lithocholic acid (LCA), which are synthesized by the gut microbiome. In addition to the transformation of primary bile acids into secondary bile acids, microbes also conjugate other biological molecules or functional groups to bile acids, including, but not limited to, amidation, esterification, and sulfonation. Amidation, specifically the addition of glycine or taurine to CA or CDCA, is the most well-known conjugation to bile acids; , ninety-eight percent of primary bile acids excreted by the liver are amidated. These bile acid amidates (Figure B) are glyco-cholic acid (GCA), tauro-cholic acid (TCA), glyco-chenodeoxycholic acid (GCDCA), and tauro-chenodeoxycholic acid (TCDCA). ,

After CA and CDCA are synthesized from cholesterol in the liver, amidation begins when bile acid coenzyme A synthase (BACS) and bile acid-CoA/amino acid N-acetyltransferase (BAAT) catalyze a two-step reaction (Figure C). , While this process has been characterized when performed by human enzymes, recent studies illustrate that amidation is also performed by bacteria in the host microbiome, though the involved enzymes are unidentified. , The final products are stored in the gallbladder until food is eaten, at which time they are secreted into the duodenum. There, they act as antimicrobial agents; in addition to facilitating the absorbance of lipids, the detergent properties of bile acids are also able to disrupt the membranes of susceptible bacteria. Thus, amidation of bile acids by bacteria may be a defense mechanism as amidated bile acids are less effective detergents than nonamidated bile acids. ,

The behavior of CA, CDCA, GCA, TCA, GCDCA, and TCDCA has been characterized via in vitro assays measuring ligand-dependent recruitment of steroid receptor coactivator 1 (SRC-1) to the FXR LBD, and ligand-dependent transactivation of target genes by FXR. , In the coregulator recruitment studies, CDCA and its conjugates showed EC50s between 4.5 and 10 μM. Little to no effect was observed with CA and its conjugates. In cellular assays, CDCA conjugates did not activate FXR, unless a bile acid transporter was coexpressed to facilitate the uptake of the conjugates. Interestingly, the presence of the ileal bile acid transporter yielded stronger activation by TCDCA and GCDCA than by CDCA, and even produced moderate activation in GCA and TCA. A different study showed that the presence of the liver bile acid transporter revealed significant activation by CA and its conjugates. These findings (summarized in Table S1) suggest that primary bile acids amidated with glycine or taurine do not show significant differences in physiological properties compared to their nonamidated counterparts.

Although the interactions of CDCA and other bile acids with the FXR LBD have been simulated and analyzed in silico, − these interactions have not been characterized for complexes of FXR and amidated primary bile acids. While functionally similar, it makes sense that the larger size and increased lipophilicity of amidated bile acids would result in different interactions with the FXR LBD than are seen in nonamidated bile acids. This work aims to describe and evaluate the effect of amidation of primary bile acids on the FXR LBD dynamics. Because these diverging ligands are known to induce similar biological outcomes when bound to FXR, their study may yield insights into the FXR-ligand interactions and dynamics that play important, or unimportant, roles in modulating the allosteric behaviors of FXR.

Methods

Structure Preparation and Simulations

Initial structures were obtained from PDB: 6HL1, containing FXR LBD bound to CDCA, with the SRC-1 peptide (DHQLLRYLLDK). To create structures of FXR bound to CA, GCA, TCA, GCDCA, and TCDCA, the CDCA structure was modified manually using Maestro. Ligands were parametrized using Antechamber and docked using LEaP from AmberTools20. All ligands were created and parametrized with the terminal –COOH (or –SO3H in the taurine conjugates) deprotonated to –COO– (or –SO3 –) because pK a values indicate that these ligands would be deprotonated at physiological pH. Simulations were performed on complexes with and without the coactivator peptide. Classical molecular dynamics simulations were performed on the 6 complexes solvated with TIP3P water in a truncated octahedron with a 10 Å water buffer surrounding the protein–ligand complex. Na+ and Cl– ions were added to achieve a physiological salt concentration and neutral charge. tLEaP from AmberTools20 was used to generate all simulated systems. The FF14SB force field was used for proteins, and the Generalized Amber Force field (GAFF2) was used to parametrize ligands. Minimization and equilibration were performed as described previously using Amber20. , Then, 1 μs production simulations were performed in triplicate with no restraints in the NPT ensemble.

Processing and Analysis of Trajectories

Prior to analyses, the three replicate trajectories for each complex were concatenated for each complex using CPPTRAJ. All analyses described below were performed on the combined trajectories. Root mean square deviation (RMSD) and root-mean-square fluctuations (RMSF) were calculated for the complex using the best-fit deviation to rotate the complex, removing deviations caused by poor alignment rather than conformational change. The solvent-accessible surface area was calculated using the linear combinations of pairwise overlap (LCPO) method.

Hydrogen bonds were observed by measuring the distance between H–N or H–O donors and N or O acceptors over the trajectory using the “hbond” command of CPPTRAJ. The percentage of the trajectory for which the donor and acceptor remained within 3.0 Å of each other was recorded to evaluate the persistence or transience of specific hydrogen bonds. Backbone hydrogen bonds in alpha helices were excluded by counting only hydrogen bonds with a difference in the residue number greater than 4. If multiple hydrogen bonds existed between the same two residues in the same complex, then only the largest fraction was recorded. In radial distribution analysis to quantify the interaction between ligand tails and Arg331, the “dist” command of CPPTRAJ was used to measure the distance between the O24 atom and the center of mass of the NE, NH1, and NH2 atoms of the guanidium group on Arg331.

Principal components analysis (PCA) was used to identify large-scale motions of the backbone atoms. Vectors representing the largest motions of backbone residues are calculated, and the deviance of each residue along the vector from the initial structure is plotted.

Ligand contacts, communities, and suboptimal paths were determined using dynamic networks generated in VMD , and following the protocol described previously. Briefly, Cartesian coordinates were calculated for the complexes using Carma, followed by the generation of dynamic networks. Networks are composed of nodes and edges, where each node represents an amino acid or ligand in the complex. Edges are created between two nodes if heavy atoms of one node were within 4.5 Å of heavy atoms from another non-neighboring node for 75% of the trajectory. Edges are weighted by the calculated covariances, such that edge weights are inversely proportional to the pairwise correlation between the nodes.

While edges could be weighted by the absolute value of correlation, our analyses have used only positive correlation to weight edges for community and suboptimal paths. Communities were generated with the Girvan–Newman network algorithm. A community is a group of nodes within a network that have stronger and more numerous connections to other nodes within the community rather than outside of it. Ligand contacts were determined by identifying residues within 4.5 Å of any ligand atom in the dynamic network for at least 75% of the trajectory. Suboptimal paths describe communication between different sites in a protein. Suboptimal paths were generated from dynamic networks using the Floyd–Warshall algorithm. In this study, we analyze the shortest 1000 paths between the ligand (source) and AF-2/H12 region (sink = E467).

Results

RMSF Shows Similarities in Induced Motion but PCs Reveal Different Underlying Behaviors

Bile acid amidation introduces a bulky conjugate at the carboxylic acid tail that is capable of large disruptions within the receptor binding pocket. To study the effects of amidation on global FXR LBD motions, we analyzed triplicate of 1 μs-long trajectories of FXR in complex with the six bile acids, first combining each into single trajectories containing the three replicates. All analyses described, hereafter, are performed on the combined trajectory for each complex. We analyzed root-mean-square fluctuations deviation (RMSD) and fluctuations (RMSF), finding that RMSD values remain stable for all replicates over the trajectory (Figure S1). Large-scale motions of the FXR LBD remain similar between amidated and nonamidated bile acid complexes over the course of the simulation. When compared to the RMSF of apo FXR, all ligands induce more stability in the FXR LBD, especially around helices H3, H4/5, H6, and H12. A large peak is observed at residues 338–360, starting at residue L5/6 (the loop between H5 and H6) and continuing through to residue L6/7. Smaller peaks are seen at residues 263–281 (L1/2 through L2/3), 421–427 (L9/10), and 452–458 (L11/12) (Figure A,B). Except for L9/10, these regions are part of the ligand-binding pocket and are expected to move as the LBP adjusts to accommodate ligands of different shapes, sizes, and polarities. All high RMSF regions are loops in which flexibility would be expected. We calculated difference RMSFs by subtracting CA and CDCA RMSF values from those of the amidated ligands (Figure C,D). Interestingly, amidation of CA leads to more increased flexibility of FXR compared with CDCA amidation. A notable exception is at L5/6, where CDCA induces higher RMSF in FXR than its amidated counterparts (Figure D).

2.

2

Open in a new tab

Fluctuations of FXR-ligand complexes. (A) Root mean square fluctuations (RMSF) of FXR LBD in apo form, as well as complexed with CA, CDCA, GCA, GCDCA, TCA, and TCDCA. Regions with high RMSF are marked by arrows. (B) FXR LBD structure with regions of high RMSF highlighted. (C) CA ΔRMSF, calculated by subtracting CA RMSF from amidated CA complexes. ΔRMSF values range between 0 and 2 Å. (D) CDCA ΔRMSF, calculated by subtracting CDCA RMSF from amidated CDCA complexes. ΔRMSF values are largely under 1 Å, except for the L5/6 regions with a large negative value. (E) Combined fluctuations of the top two principal components (PC1 + PC2) reveal regions where amidated bile acids have higher fluctuations compared to CA and CDCA. (F) FXR LBD structure with regions of high PC fluctuations in amidated bile acid complexes are highlighted in red. Residues 338–341 (in light purple) show high fluctuations for all complexes.

To analyze the most prominent motions in FXR induced by the various ligands, we performed a principal components analysis. We examined the fluctuations of the top two principal components of each complex (Figure E), which jointly capture at least 50% of the motions in each complex (Table S2). While the PCA fluctuations largely mirror the RMSF, we noted two regions of interest: L2/3-H3 and H6–H7. Increases in fluctuations are observed in these regions in the amidated bile acid complexes (Figure E,F). The increase at H6/H7 (residues 342–366) is particularly interesting, given its proximity to L5/6 (residues 338–341), for which all complexes show a large fluctuation. This flexibility does not appear to correlate with a specific ligand property; rather, it indicates the flexibility of the loop. Thus, it is notable that amidated bile acids induce larger changes in neighboring H6, compared to CA and CDCA.

All Bile Acids Interact Similarly with FXR LBP, Amidation Results in Modifications

When FXR was newly adopted, it was predicted that amidated bile acids would bend out of the ligand-binding pocket (LBP) and directly contact the solvent when bound to FXR. However, crystal structures of FXR bound to large synthetic ligands, such as fexaramine, ivermectin, and GW4064, indicate that the ligand binding pocket can expand to accommodate larger ligands. Our analysis of the solvent-accessible surface area (SASA) supports this idea. We focused on L1/2 due to its proximity to the amino acid tails (Figures A and S2). Few to no significant differences were observed in SASA of L1/2 residues Gln263-Glu268 between the complexes, suggesting that the outer surface of the LBD is not meaningfully changed as a result of the amino acid tail (Figure S3). The exception to this observation was Arg264, for which higher SASA was observed in amidated bile acid complexes compared to nonamidated complexes (Figure B). The overall similarity in SASA values across complexes indicates that the FXR LBD largely retains the same solvent interactions, even when larger ligands are bound.

3.

3

Open in a new tab

Effects of amidation on the FXR binding pocket. (A) Amidated ligand tails are predicted to curve toward L1/2. CDCA-based ligands are used for the representation. Refer to Figure S2 for CA-based ligands. (B) Solvent-accessible surface area (SASA) data of Arg265 and Pro266 (in L1/2). An increase in SASA of Arg264 is observed for amidated bile acids. No effect of amidation is observed in SASA of Pro266. SASA data for the other L1/2 residues are shown in Figure S3. (C) RMSD of ligand calculated across simulation. Despite a higher RMSD in amidated bile acids, measurements remain stable. (D–G) Hydrogen bonding patterns of the tails in CDCA (D), TCDCA (E), and GCDCA (F,G). (H,I) Radial distribution function of the distance between the O24 atom and center of mass of Arg331-guanidium nitrogen atoms shown for (H) CDCA-based ligands and (I) CA-based ligands.

We also calculated the root-mean-square deviation (RMSD) of the ligands over the course of the trajectory (Figure C). While the amidated bile acids have slightly higher RMSD values than the nonamidated bile acids, this is likely due to their larger size. All four amidated bile acids maintain a steady RMSD throughout the simulation, illustrating their stability in LBP (Figure C). To explicitly distinguish between conformational preferences of the different ligand tails, we studied hydrogen bonding patterns of the amidated tails. Distinct patterns in hydrogen bonding were observed between Arg331 (H5) and nonconjugated, glycine-conjugated, or taurine-conjugated bile acids (Figure D–G). The nonconjugated bile acids both have the highest occupancy hydrogen bonds with Arg331, approximately 65–92% (Table S3). In addition, these ligands form bidentate hydrogen bonds using both oxygen atoms in the carboxyl group to interact with the NE and NH2 hydrogen atoms of Arg331 (Figure D). We performed a radial distribution analysis of this interaction, focusing on the distance between the O24 atom and the guanidium of Arg331, and observed that the unconjugated tails remain oriented toward Arg331 most of the simulation (Figure H,I).

Taurine-conjugated bile acids form a single hydrogen bond with Arg331 via the O24 atom and do not appear to interact with the sulfonate group (Figure E). The radial distribution shows that the taurine-conjugated tails orient toward Arg331 with slight propensity to drift, likely driven by weak interactions between the sulfonate and the pocket (Figure H,I). Conversely, glycine-conjugated bile acids form only one hydrogen bond with Arg331, between the terminal carboxyl group of the ligand and one NH2 group in Arg331 (Figure F). However, this interaction sometimes shifts to occur with the O24 atom (Figure G), which explains the diffuse nature of the radial distribution (Figure H,I). The distance between O24 and the terminal carboxyl is too long to allow bidentate hydrogen bonding with Arg331, resulting in constant fluctuations of the tail.

FXR Aromatic Triad Does Not Persist in Bile Acid Complexes

A contact analysis was performed to characterize changes induced by amidation in interactions between the ligands and FXR residues. Our criteria define two residues are “in contact” if any pair of atoms from both are within 4.5 Å for at least 75% of the trajectory. All ligands contact residues Met265, Leu287, Met290, Ala291, His294, Met328, Arg331, Ile335, Ile352, Met365, Tyr369, and Trp454 (Figure A,B). Residues that contact only some of the ligands include Phe329, Leu348, Ile357, His447, Tyr361, and Met450 (Figure A). Residues in direct contact with bile acid ligands strongly overlapped with previously identified residues of interest for both agonist and partial agonist binding.

4.

4

Open in a new tab

Effect of amidation on FXR contacts. (A) Table showing residues that interact with bile acids. (B) Ligand binding pocket showing the different residues that are in contact or form hydrogen bonds with ligands. CDCA-based ligands are used for visualization. Note that all contacting residues are depicted in the figures to illustrate their position relative to the ligand, even if they do not contact the particular ligand shown; refer to panels A and C for the specifics of interactions. (C) Prevalence of hydrogen bonds (fraction of simulation time) of different residues with the ligands. (D) Aromatic triad consisting of ligand, Tyr61, His447, and Trp469. The ligand forms hydrogen bonds between Tyr61 and His447. His447 forms cation–pi interaction with Trp469. (E) Radial distribution of His447-Trp469 cation–pi interaction distance shown for all ligands. (F) FXR LBD model with the SRC-1 coactivator peptide fragment docked at the AF-2 surface. (G) Radial distribution of His447-Trp469 cation–pi interaction distance in the presence of SRC-1 coactivator. (H) Prevalence of hydrogen bonds (percentage of simulation time) in the aromatic triad shown for before and after addition of SRC-1 coactivator.

In addition to Arg361 (Figure D–G), ligands form hydrogen bonds with Ser332, Ser342, Tyr361, Tyr369, and His447 (Figure C). Interestingly, we noticed that Tyr361 (H7) and His447 (H10) engage only with CDCA and GCDCA consistently via hydrogen bonding through their C3-hydroxyl group (Figure C). This was surprising, as these two residues, along with Trp469 (H12), constitute the “aromatic triad,” which has been established as a ligand-triggered activation switch that stabilizes AF-2 for FXR activity (Figure D). , Three components of the aromatic triad were previously identified: (i) hydrogen bonding between C3-hydroxyl and Tyr361, (ii) hydrogen bonding between C3-hydroxyl and His447, and (iii) a cation–pi interaction between His447 and Trp469 (Figure D). We observe that the hydrogen bond interactions are not preserved in some of the ligands.

To investigate whether the cation–pi interaction remained present, we measured the distance between His447 and the center of mass of the Trp469 side chain. Interestingly, this interaction was strong in all ligand complexes, with an average of around 3.5 Å observed in radial distribution plots (Figure E). Thus, only CDCA and GCDCA appear to satisfy all three interaction criteria of the aromatic triad. Yet, published reports show that amidated ligands can recruit coactivators to FXR and upregulate FXR-mediated transcription in vitro. , Therefore, our results suggest that the aromatic triad does not constitute a mandatory requirement for FXR activation. However, binding of both an agonist and a coactivator is required to completely activate FXR. It is possible that the binding of a coactivator peptide to the AF-2 surface is necessary to achieve a conformation that allows the residues of the aromatic triad to interact.

To test whether the addition of a coactivator affects the stability of the aromatic triad, we ran MD simulations with the steroid receptor coactivator-1 (SRC-1) peptide fragment docked at the AF-2 surface (Figure F). For most of the ligands, the area under the peak at 3.5 Å increases, indicating that the coactivator stabilizes the interaction between His447 and Trp469 (Figure G). The presence of the peptide slightly alters the hydrogen bonding of the ligands with Tyr361 and His447, increasing in some cases and decreasing in others (Figure H). In CDCA and GCDCA, hydrogen bond occupancies decrease upon the addition of the coactivator. The modest and variable nature of these effects suggest that the addition of the coactivator peptide does not particularly stabilize the aromatic triad.

Differences in Indirect Communication: Nonligand Hydrogen Bonds and Community Analysis

While amidated and nonamidated bile acids bind somewhat similarly to the FXR LBP and induce similar large-scale motions, behaviors outside the LBP are notably different. An analysis of hydrogen bonding across the LBD revealed differences in hydrogen bond networks, which appear specific to the base bile acid and/or the conjugation status (Figure S4). Glu334 (H5) forms a salt bridge with Arg264 (L1/2) that is interrupted by both glycine and taurine-conjugated bile acids (Figures A–C and S5). This disruption is due in part to weak interactions between the anionic tails and Arg264. Glu334 also forms a hydrogen bond with Tyr260 (H3), which is inhibited by glycine-conjugated bile acids but not taurine-conjugated bile acids, suggesting that this difference arises from the conjugate (Figures D,E and S6). In all complexes, strong hydrogen bonds form between H6 and nearby regions (Figure S4). However, some hydrogen bonds display a propensity for either CDCA- or CA-based ligands. This includes the Thr292-Thr462 (H3-L11/12) interaction, which is stronger in CDCA-based ligands, and Glu326-Arg441 and Gly322-Asn444 (both H4–H10) interactions, which show a slight preference for CA-based ligands.

5.

5

Open in a new tab

Allosteric signaling and communities in FXR complexes. (A–C) Glu334-Arg264 salt bridge is affected by the ligand structure. The salt bridge intact in CDCA (shown) and CA complexes. Amidated tails (B,C) disrupt the salt bridge. (D,E) Tyr260-Glu334 hydrogen bond is affected by the ligand structure. The glycine tail inhibits the interaction (D) but the taurine tail does not (E). (F) Correlation between ligands and FXR residues were calculated for all ligands. Glycine-conjugates showed distinguishing features including a positive correlation at L2/3, H3 and H6, and negative correlation at H7. (G) FXR LBD structure colored by correlation with the ligand. Differences between GCDCA and CDCA are evident at L2/3, H3, H6, and H7. (H) Community analysis based on the positive correlation shows that 5 of 6 complexes share the same ligand community. (I) Suboptimal paths between the ligand binding pocket and Glu467 on H12 as the sink. The shortest 1000 suboptimal paths are shown, with thinner lines identifying residues appearing in <5% of total paths. The analysis shows that 5 of 6 ligands (4 shown) utilize H3 for communication, emphasizing that amidation does not alter allosteric communication.

To determine how amidation affects the correlation between bile acids and FXR residues, we calculated cross correlation matrices reporting on correlated or anticorrelated motions. Interestingly, clear patterns in correlation were observed between the ligands and FXR, based on the amidation state (Figures F and S7). Glycine conjugates are strongly correlated with L2/3 and the bottom of H3, as well as H6 (Figures F,G and S7). Conversely, they show a strong anticorrelation with H7, contrasting with the other ligands. Compared to the glycine conjugates, the taurine conjugates and unconjugated ligands display fewer distinguishing features (Figures F, S6 and S7). Thus, bile acid amidation affects FXR motions, particularly close to the ligand binding pocket.

To determine whether local shifts in FXR dynamics alter communication across the LBD, we performed community and suboptimal path analysis. Community analysis splits a protein complex into groups of residues with correlated motions. Due to their related movement, residues in the same community may function like a rigid body, participating in shared functions and allosteric communication. , The community analysis, based on positively correlated motions (see the Methods section), reveals that all complexes contain a community consisting of residues in H2, H3, and H6 (Figure H) ,which constitutes a large portion of the binding pocket. For 5 of 6 ligands, this community contains the ligand, suggesting that amidation does not drastically alter dynamic communication within FXR. Suboptimal paths between the ligand and H12 were generated, and except for CA, paths in all complexes primarily traverse the same H3 residues (Figures I and S8). Combined, these analyses demonstrate that local differences in correlation do not perturb global communication in the LBD.

Discussion and Conclusions

Although the majority of bile acids excreted from the liver that interact with FXR are amidated, interactions between these have not been characterized by using in silico techniques. In vitro experiments indicate that the binding and transcriptional activities of these ligands are only modestly affected, suggesting that the receptor is capable of tolerating the bulky substituents without significant structural and dynamic perturbations. In this study, we compare the behavior of the FXR LBD when bound to CA and CDCA, along with the glycine and taurine conjugates. While we have modeled the amidated ligands into the FXR pocket in the same binding mode as unmodified bile acids, it is possible that the tails force them into a unique orientation. Crystallization or extensive docking studies would be required to confirm their binding modes. Thus, the findings of this study presuppose that the default binding modes persist in amidated bile acid complexes.

Previous reports speculated that the amidated tail would protrude outside the FXR binding pocket, while also maintaining a hydrogen bond with Arg331 on H5. While we do not observe the tail escaping the pocket, our simulations confirm that the glycine conjugates interact with Arg331 via their carboxylate tail. This interaction occurs less frequently in taurine conjugates. We also measured the solvent-accessible surface area of L1/2 residues and observed that apart from Arg264, SASA remained largely consistent across all complexes. This finding suggests that the LBP expands to accommodate larger ligands without significantly altering the surface conformation of the domain. RMSF values of all complexes were comparable, yet principal components revealed differences in the underlying motions that come together to form the overall fluctuation of the FXR LBD throughout the trajectory.

While the bile acids make largely similar van der Waals interactions with the surrounding amino acids, hydrogen bonding patterns appear to be base- and tail-dependent. We performed a careful analysis of the aromatic triad (also known as the FXR activation trigger), composed of Trp469, His447, and Tyr361, which is proposed to control the functional state of the receptor. Interestingly, the majority of predicted interactions within the triad do not persist in our simulations, possibly contradicting the hypothesis that these interactions are necessary for FXR activation. When simulations are performed with a coactivator peptide docked at the AF-2 surface, the results are only moderately altered. Our findings are in line with previous simulations on obeticholic acid (OCA), showing variance in the persistence of interactions. Further experimental investigations into the persistence of the triad and its relevance for FXR activity will be important to resolve the role of the interactions. Currently, no experimental studies have used mutagenesis to test the functional implications of disrupting the aromatic triad.

We note that the largest effects of the amide tails are felt in H5. The Glu334-Arg264 salt bridge between L1/2 and H5 is interrupted by the tails. Arg331 also displays distinct hydrogen bonding patterns with ligands based on their tail. Interestingly, H5 emerged in our previous simulation studies on FXR in complex with OCA. There, we observed that H5 dynamics were selectively modulated in response to the C6-ethyl substituent on the synthetic bile acid. These studies confirm that H5 plays crucial and likely diverse roles in mediating the effects of bile acids. To the best of our knowledge, the Glu334-Arg264 salt bridge has not previously been reported in FXR, nor has its functional relevance been dissected in experiments. While we cannot state with certainty that this interaction is crucial for the FXR function, our simulations would suggest that it can be disrupted by bile acid-amidated tails but still display activity levels similar to unmodified bile acids (biological activity of bile acids summarized in Table S1).

In summary, while these ligands bind to FXR and induce specific effects on residue interactions and correlative behaviors, allosteric effects and large-scale motions in silico are observed to be similar. This is evidenced in community analysis and suboptimal paths, shown to be very similar across the ligands. The preservation of global FXR motions in amidated ligand complexes seems to be consistent with experimental observations. Because conjugation did not change the effect of bile acids on SRC-1 recruitment in experiments (Table S1), we can surmise that conjugation does not significantly alter the conformational dynamics of AF-2, where coregulators bind. However, studies of other nuclear receptors emphasize that subtle conformational changes may allosterically modulate coregulator preferences. Thus, even though no differences were seen in recruitment of SRC-1, investigating the effect of amidation on a panel of coregulators, as previously described, , would likely reveal differences in coregulator binding preferences between amidated bile acids and their unconjugated counterparts. If this is observed, the subtle localized effects observed in our simulations would be relevant for explaining these effects.

Supplementary Material

jp5c02361_si_001.pdf (769.5KB, pdf)

Acknowledgments

This work was supported by an NSF CAREER award CAREER: 2144679 (to C.D.O.), and by a National Institutes of Health award DP2-GM149753 (to C.D.O.)

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acs.jpcb.5c02361.

  • RMSD, summary of previous experimental measurements, percentage contributions of PC1 and PC2, occupancy of hydrogen bonds between bile acids and Arg331, SASA, hydrogen bond prevalence heatmap, and cross-correlation values (PDF)

The authors declare no competing financial interest.

References

  1. Porter B. A., Ortiz M. A., Bratslavsky G., Kotula L.. Structure and Function of the Nuclear Receptor Superfamily and Current Targeted Therapies of Prostate Cancer. Cancers. 2019;11(12):1852. doi: 10.3390/cancers11121852. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Weikum E. R., Liu X., Ortlund E. A.. The Nuclear Receptor Superfamily: A Structural Perspective. Protein Sci. 2018;27(11):1876–1892. doi: 10.1002/pro.3496. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Kumari A., Mittal L., Srivastava M., Pathak D. P., Asthana S.. Conformational Characterization of the Co-Activator Binding Site Revealed the Mechanism to Achieve the Bioactive State of FXR. Front. Mol. Biosci. 2021;8:658312. doi: 10.3389/fmolb.2021.658312. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Liu Y., Rong Z., Xiang D., Zhang C., Liu D.. Detection Technologies and Metabolic Profiling of Bile Acids: A Comprehensive Review. Lipids Health Dis. 2018;17(1):121. doi: 10.1186/s12944-018-0774-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. de Aguiar Vallim T. Q., Tarling E. J., Edwards P. A.. Pleiotropic Roles of Bile Acids in Metabolism. Cell Metab. 2013;17(5):657–669. doi: 10.1016/j.cmet.2013.03.013. [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Hofmann A. F.. The Continuing Importance of Bile Acids in Liver and Intestinal Disease. Arch. Intern. Med. 1999;159(22):2647. doi: 10.1001/archinte.159.22.2647. [DOI] [PubMed] [Google Scholar]
  7. Guzior D. V., Quinn R. A.. Review: Microbial Transformations of Human Bile Acids. Microbiome. 2021;9(1):140. doi: 10.1186/s40168-021-01101-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Makishima M., Okamoto A. Y., Repa J. J., Tu H., Learned R. M., Luk A., Hull M. V., Lustig K. D., Mangelsdorf D. J., Shan B.. Identification of a Nuclear Receptor for Bile Acids. Science. 1999;284:1362. doi: 10.1126/science.284.5418.1362. [DOI] [PubMed] [Google Scholar]
  9. Mohanty I., Allaband C., Mannochio-Russo H., El Abiead Y., Hagey L. R., Knight R., Dorrestein P. C.. The Changing Metabolic Landscape of Bile Acids – Keys to Metabolism and Immune Regulation. Nat. Rev. Gastroenterol. Hepatol. 2024;21(7):493–516. doi: 10.1038/s41575-024-00914-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Rimal B., Collins S. L., Tanes C. E., Rocha E. R., Granda M. A., Solanki S., Hoque N. J., Gentry E. C., Koo I., Reilly E. R., Hao F., Paudel D., Singh V., Yan T., Kim M. S., Bittinger K., Zackular J. P., Krausz K. W., Desai D., Amin S., Coleman J. P., Shah Y. M., Bisanz J. E., Gonzalez F. J., Vanden Heuvel J. P., Wu G. D., Zemel B. S., Dorrestein P. C., Weinert E. E., Patterson A. D.. Bile Salt Hydrolase Catalyses Formation of Amine-Conjugated Bile Acids. Nature. 2024;626(8000):859–863. doi: 10.1038/s41586-023-06990-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Pellicciari R., Costantino G., Camaioni E., Sadeghpour B. M., Entrena A., Willson T. M., Fiorucci S., Clerici C., Gioiello A.. Bile Acid Derivatives as Ligands of the Farnesoid X Receptor. Synthesis, Evaluation, and Structure–Activity Relationship of a Series of Body and Side Chain Modified Analogues of Chenodeoxycholic Acid. J. Med. Chem. 2004;47(18):4559–4569. doi: 10.1021/jm049904b. [DOI] [PubMed] [Google Scholar]
  12. Falany C. N., Johnson M. R., Barnes S., Diasio R. B.. Glycine and Taurine Conjugation of Bile Acids by a Single Enzyme. Molecular Cloning and Expression of Human Liver Bile Acid CoA:Amino Acid N-Acyltransferase. J. Biol. Chem. 1994;269(30):19375–19379. doi: 10.1016/S0021-9258(17)32178-6. [DOI] [PubMed] [Google Scholar]
  13. Sannasiddappa T. H., Lund P. A., Clarke S. R.. In Vitro Antibacterial Activity of Unconjugated and Conjugated Bile Salts on Staphylococcus Aureus. Front. Microbiol. 2017;8:1581. doi: 10.3389/fmicb.2017.01581. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Lew J.-L., Zhao A., Yu J., Huang L., de Pedro N., Peláez F., Wright S. D., Cui J.. The Farnesoid X Receptor Controls Gene Expression in a Ligand- and Promoter-Selective Fashion. J. Biol. Chem. 2004;279(10):8856–8861. doi: 10.1074/jbc.M306422200. [DOI] [PubMed] [Google Scholar]
  15. Wang H., Chen J., Hollister K., Sowers L. C., Forman B. M.. Endogenous Bile Acids Are Ligands for the Nuclear Receptor FXR/BAR. Mol. Cell. 1999;3(5):543–553. doi: 10.1016/S1097-2765(00)80348-2. [DOI] [PubMed] [Google Scholar]
  16. Parks D. J., Blanchard S. G., Bledsoe R. K., Chandra G., Consler T. G., Kliewer S. A., Stimmel J. B., Willson T. M., Zavacki A. M., Moore D. D., Lehmann J. M.. Bile Acids: Natural Ligands for an Orphan Nuclear Receptor. Science. 1999;284:1365. doi: 10.1126/science.284.5418.1365. [DOI] [PubMed] [Google Scholar]
  17. Díaz-Holguín A., Rashidian A., Pijnenburg D., Monteiro Ferreira G., Stefela A., Kaspar M., Kudova E., Poso A., van Beuningen R., Pavek P., Kronenberger T.. When Two Become One: Conformational Changes in FXR/RXR Heterodimers Bound to Steroidal Antagonists. ChemMedChem. 2023;18(4):e202200556. doi: 10.1002/cmdc.202200556. [DOI] [PubMed] [Google Scholar]
  18. Jiang L., Zhang H., Xiao D., Wei H., Chen Y.. Farnesoid X Receptor (FXR): Structures and Ligands. Comput. Struct. Biotechnol. J. 2021;19:2148–2159. doi: 10.1016/j.csbj.2021.04.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Hazarika S., Yu T., Biswas A., Dube N., Villalona P., Okafor C. D.. Nuclear Receptor Interdomain Communication Is Mediated by the Hinge with Ligand Specificity. J. Mol. Biol. 2024;436(22):168805. doi: 10.1016/j.jmb.2024.168805. [DOI] [PubMed] [Google Scholar]
  20. Di Leva F. S., Di Marino D., Limongelli V.. Structural Insight into the Binding Mode of FXR and GPBAR1 Modulators. Handb. Exp. Pharmacol. 2019;256:111–136. doi: 10.1007/164_2019_234. [DOI] [PubMed] [Google Scholar]
  21. Merk D., Sreeramulu S., Kudlinzki D., Saxena K., Linhard V., Gande S. L., Hiller F., Lamers C., Nilsson E., Aagaard A., Wissler L., Dekker N., Bamberg K., Schubert-Zsilavecz M., Schwalbe H.. Molecular Tuning of Farnesoid X Receptor Partial Agonism. Nat. Commun. 2019;10(1):2915. doi: 10.1038/s41467-019-10853-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Schrödinger Release 2024-3; Maestro, Schrödinger, LLC, New York, NY, 2024. [Google Scholar]
  23. Wang J., Wang W., Kollman P. A., Case D. A.. Automatic Atom Type and Bond Type Perception in Molecular Mechanical Calculations. J. Mol. Graph Model. 2006;25(2):247–260. doi: 10.1016/j.jmgm.2005.12.005. [DOI] [PubMed] [Google Scholar]
  24. Case, D. A. ; Belfon, K. ; Ben-Shalom, I. Y. ; Brozell, S. R. ; Cerutti, D. S. ; Cheatham, T. E. ; Cruzeiro, V. W. D. ; Darden, T. A. ; Duke, R. E. ; Giambasu, G. ; Gilson, M. K. ; Gohlke, H. ; Goetz, A. W. ; Harris, R. ; Izadi, S. ; Izmailov, S. A. ; Kasavajhala, K. ; Kovalenko, A. ; Krasny, R. ; Kurtzman, T. ; Lee, T. S. ; LeGrand, S. ; Li, P. ; Lin, C. ; Liu, J. ; Luchko, T. ; Luo, R. ; Man, V. ; Merz, K. M. ; Miao, Y. ; Mikhailovskii, O. ; Monard, G. ; Nguyen, H. ; Onufriev, A. ; Pan, F. ; Pantano, S. ; Qi, R. ; Roe, D. R. ; Roitberg, A. ; Sagui, C. ; Schott-Verdugo, S. ; Shen, J. ; Simmerling, C. L. ; Skrynnikov, N. R. ; Smith, J. ; Swails, J. ; Walker, R. C. ; Wang, J. ; Wilson, L. ; Wolf, R. M. ; Wu, X. ; Xiong, Y. ; Xue, Y. ; York, D. M. ; Kollman, P. A. . AMBER 2020; University of California: San Francisco, 2020. [Google Scholar]
  25. Russell D. W.. The Enzymes, Regulation, and Genetics of Bile Acid Synthesis. Annu. Rev. Biochem. 2003;72(1):137–174. doi: 10.1146/annurev.biochem.72.121801.161712. [DOI] [PubMed] [Google Scholar]
  26. Mark P., Nilsson L.. Structure and Dynamics of the TIP3P, SPC, and SPC/E Water Models at 298 K. J. Phys. Chem. A. 2001;105(43):9954–9960. doi: 10.1021/jp003020w. [DOI] [Google Scholar]
  27. Maier J. A., Martinez C., Kasavajhala K., Wickstrom L., Hauser K. E., Simmerling C.. Ff14SB: Improving the Accuracy of Protein Side Chain and Backbone Parameters from Ff99SB. J. Chem. Theory Comput. 2015;11(8):3696–3713. doi: 10.1021/acs.jctc.5b00255. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Wang J., Wolf R. M., Caldwell J. W., Kollman P. A., Case D. A.. Development and Testing of a General Amber Force Field. J. Comput. Chem. 2004;25(9):1157–1174. doi: 10.1002/jcc.20035. [DOI] [PubMed] [Google Scholar]
  29. Mays S. G., Hercules D., Ortlund E. A., Okafor C. D.. The Nuclear Receptor LRH −1 Discriminates between Ligands Using Distinct Allosteric Signaling Circuits. Protein Sci. 2023;32(10):e4754. doi: 10.1002/pro.4754. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Roe D. R., Cheatham T. E.. PTRAJ and CPPTRAJ: Software for Processing and Analysis of Molecular Dynamics Trajectory Data. J. Chem. Theory Comput. 2013;9(7):3084–3095. doi: 10.1021/ct400341p. [DOI] [PubMed] [Google Scholar]
  31. Weiser J., Shenkin P. S., Still W. C.. Approximate Atomic Surfaces from Linear Combinations of Pairwise Overlaps (LCPO) J. Comput. Chem. 1999;20(2):217–230. doi: 10.1002/(SICI)1096-987X(19990130)20:2&#x0003c;217::AID-JCC4&#x0003e;3.0.CO;2-A. [DOI] [Google Scholar]
  32. David C. C., Jacobs D. J.. Principal Component Analysis: A Method for Determining the Essential Dynamics of. Proteins. 2014;1084:193–226. doi: 10.1007/978-1-62703-658-0_11. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Eargle J., Luthey-Schulten Z.. NetworkView: 3D Display and Analysis of Protein·RNA Interaction Networks. Bioinformatics. 2012;28(22):3000–3001. doi: 10.1093/bioinformatics/bts546. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Humphrey W., Dalke A., Schulten K.. VMD: Visual Molecular Dynamics. J. Mol. Graph. 1996;14(1):33–38. doi: 10.1016/0263-7855(96)00018-5. [DOI] [PubMed] [Google Scholar]
  35. Glykos N. M.. Software News and Updates Carma: A Molecular Dynamics Analysis Program. J. Comput. Chem. 2006;27(14):1765–1768. doi: 10.1002/jcc.20482. [DOI] [PubMed] [Google Scholar]
  36. Sethi A., Eargle J., Black A. A., Luthey-Schulten Z.. Dynamical Networks in TRNA:Protein Complexes. Proc. Natl. Acad. Sci. U.S.A. 2009;106(16):6620–6625. doi: 10.1073/pnas.0810961106. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Girvan M., Newman M. E. J.. Community Structure in Social and Biological Networks. Proc. Natl. Acad. Sci. U.S.A. 2002;99(12):7821–7826. doi: 10.1073/pnas.122653799. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Floyd R. W.. Algorithm 97: Shortest Path. Commun. ACM. 1962;5(6):345. doi: 10.1145/367766.368168. [DOI] [Google Scholar]
  39. Downes M., Verdecia M. A., Roecker A. J., Hughes R., Hogenesch J. B., Kast-Woelbern H. R., Bowman M. E., Ferrer J.-L., Anisfeld A. M., Edwards P. A., Rosenfeld J. M., Alvarez J. G. A., Noel J. P., Nicolaou K. C., Evans R. M. A.. Chemical, Genetic, and Structural Analysis of the Nuclear Bile Acid Receptor FXR. Mol. Cell. 2003;11(4):1079–1092. doi: 10.1016/S1097-2765(03)00104-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Mi L.-Z., Devarakonda S., Harp J. M., Han Q., Pellicciari R., Willson T. M., Khorasanizadeh S., Rastinejad F.. Structural Basis for Bile Acid Binding and Activation of the Nuclear Receptor FXR. Mol. Cell. 2003;11(4):1093–1100. doi: 10.1016/S1097-2765(03)00112-6. [DOI] [PubMed] [Google Scholar]
  41. Jin L., Feng X., Rong H., Pan Z., Inaba Y., Qiu L., Zheng W., Lin S., Wang R., Wang Z., Wang S., Liu H., Li S., Xie W., Li Y.. The Antiparasitic Drug Ivermectin Is a Novel FXR Ligand That Regulates Metabolism. Nat. Commun. 2013;4(1):1937. doi: 10.1038/ncomms2924. [DOI] [PubMed] [Google Scholar]
  42. Akwabi-Ameyaw A., Bass J. Y., Caldwell R. D., Caravella J. A., Chen L., Creech K. L., Deaton D. N., Jones S. A., Kaldor I., Liu Y., Madauss K. P., Marr H. B., McFadyen R. B., Miller A. B., Parks D. J., Spearing P. K., Todd D., Williams S. P., Wisely G. B.. Conformationally Constrained Farnesoid X Receptor (FXR) Agonists: Naphthoic Acid-Based Analogs of GW4064. Bioorg. Med. Chem. Lett. 2008;18(15):4339–4343. doi: 10.1016/j.bmcl.2008.06.073. [DOI] [PubMed] [Google Scholar]
  43. Kumari A., Mittal L., Srivastava M., Pathak D. P., Asthana S.. Deciphering the Structural Determinants Critical in Attaining the FXR Partial Agonism. J. Phys. Chem. B. 2023;127(2):465–485. doi: 10.1021/acs.jpcb.2c06325. [DOI] [PubMed] [Google Scholar]
  44. Mishra S. K., Jernigan R. L.. Protein Dynamic Communities from Elastic Network Models Align Closely to the Communities Defined by Molecular Dynamics. PLoS One. 2018;13(6):e0199225. doi: 10.1371/journal.pone.0199225. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Lesgidou N., Vlassi M.. Community Analysis of Large-scale Molecular Dynamics Simulations Elucidated Dynamics-driven Allostery in Tyrosine Kinase 2. Proteins: Struct., Funct., Bioinf. 2024;92(4):474–498. doi: 10.1002/prot.26631. [DOI] [PubMed] [Google Scholar]
  46. Yu T., Villalona P., Khan S. H., Mikeasky N., Meinert E., Magafas J., Pulahinge T., Bader A., Okafor C. D.. Enhanced Dynamic Coupling in a Nuclear Receptor Underlies Ligand Activity. J. Biol. Chem. 2025;301(2):108081. doi: 10.1016/j.jbc.2024.108081. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Mays S. G., D’Agostino E. H., Flynn A. R., Huang X., Wang G., Liu X., Millings E. J., Okafor C. D., Patel A., Cato M. L., Cornelison J. L., Melchers D., Houtman R., Moore D. D., Calvert J. W., Jui N. T., Ortlund E. A.. A Phospholipid Mimetic Targeting LRH-1 Ameliorates Colitis. Cell Chem. Biol. 2022;29(7):1174–1186e7. doi: 10.1016/j.chembiol.2022.03.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Broekema M. F., Hollman D. A. A., Koppen A., van den Ham H.-J., Melchers D., Pijnenburg D., Ruijtenbeek R., van Mil S. W. C., Houtman R., Kalkhoven E.. Profiling of 3696 Nuclear Receptor–Coregulator Interactions: A Resource for Biological and Clinical Discovery. Endocrinology. 2018;159(6):2397–2407. doi: 10.1210/en.2018-00149. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

jp5c02361_si_001.pdf (769.5KB, pdf)

Articles from The Journal of Physical Chemistry. B are provided here courtesy of American Chemical Society

RESOURCES