1. Computational and Systems Biology

Identification of a carbohydrate recognition motif of purinergic receptors

  1. Lifen Zhao
  2. Fangyu Wei
  3. Xinheng He
  4. Antao Dai
  5. Dehua Yang
  6. Hualiang Jiang
  7. Liuqing Wen  Is a corresponding author
  8. Xi Cheng  Is a corresponding author
  1. State Key Laboratory of Drug Research, Carbohydrate-Based Drug Research Center and National Center for Drug Screening, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, China
  2. University of Chinese Academy of Sciences, China
  3. School of Pharmaceutical Science and Technology, Hangzhou Institute of Advanced Study, China
https://doi.org/10.7554/eLife.85449
Share this article
Cite this article
  1. Lifen Zhao
  2. Fangyu Wei
  3. Xinheng He
  4. Antao Dai
  5. Dehua Yang
  6. Hualiang Jiang
  7. Liuqing Wen
  8. Xi Cheng
(2023)
Identification of a carbohydrate recognition motif of purinergic receptors
eLife 12:e85449.
https://doi.org/10.7554/eLife.85449
  2. Download BibTeX
  3. Download .RIS

Abstract

As a major class of biomolecules, carbohydrates play indispensable roles in various biological processes. However, it remains largely unknown how carbohydrates directly modulate important drug targets, such as G-protein coupled receptors (GPCRs). Here, we employed P2Y purinoceptor 14 (P2Y14), a drug target for inflammation and immune responses, to uncover the sugar nucleotide activation of GPCRs. Integrating molecular dynamics simulation with functional study, we identified the uridine diphosphate (UDP)-sugar-binding site on P2Y14, and revealed that a UDP-glucose might activate the receptor by bridging the transmembrane (TM) helices 2 and 7. Between TM2 and TM7 of P2Y14, a conserved salt bridging chain (K2.60-D2.64-K7.35-E7.36 [KDKE]) was identified to distinguish different UDP-sugars, including UDP-glucose, UDP-galactose, UDP-glucuronic acid, and UDP-N-acetylglucosamine. We identified the KDKE chain as a conserved functional motif of sugar binding for both P2Y14 and P2Y purinoceptor 12 (P2Y12), and then designed three sugar nucleotides as agonists of P2Y12. These results not only expand our understanding for activation of purinergic receptors but also provide insights for the carbohydrate drug development for GPCRs.

Editor's evaluation

This study describes a valuable model for the interaction of nucleotides, an important group of signalling molecules in health and disease, with their receptors. By combining experimental and theoretical methods, the authors provide compelling evidence for the mechanism of receptor activation. The results will be useful for drug design to target these receptors.

https://doi.org/10.7554/eLife.85449.sa0

eLife digest

Sugars and other types of carbohydrates are biomolecules which play a range of key roles in the body. In particular, they are important messengers that help to coordinate immune responses. For example, a carbohydrate known as UDP-Glucose (a kind of UDP-sugar) can activate P2Y14, a receptor studded through the surface of many cells; this event then triggers a cascade of molecular events associated with asthma, kidney injury and lung inflammation. Yet it remains unclear how exactly UDP-Glucose recognizes P2Y14 – and, more broadly, how carbohydrates interact with purinergic receptors, the class of proteins that P2Y14 belongs to.

To examine this question, Zhao et al. combined functional experiments in the laboratory with molecular dynamics simulations, a computational approach. This work revealed that UDP-Glucose may activate P2Y14 by bridging its segments anchored within the cell membrane. A component of P2Y14, known as the KDKE chain, was found to have an important role in distinguishing between highly similar types of UDP-sugars. This allowed Zhao et al. to design three sugar molecules which could activate another purinergic receptor that also contained a KDKE chain.

Purinergic receptors are promising therapeutic targets. A finer understanding of how they recognise the molecules that activate them is therefore important to be able to identify and design new drug compounds.

Introduction

As significant components of the organism, carbohydrates play indispensable roles in energy supply, cell signaling, and immune responses (Gagneux and Varki, 1999). Dysregulation of carbohydrates has been proved to be associated with the development of various diseases (Reily et al., 2019). However, it is still elusive how carbohydrates directly act on major therapeutic targets, including G-protein coupled receptors (GPCRs) (Cheng and Jiang, 2019; Hauser et al., 2017). P2Y purinoceptor 14 (P2Y14) represents an outstanding model system for understanding carbohydrate modulation of GPCRs. It belongs to P2Y purinoceptor subfamily, consisting of receptors responding to nucleotides, including adenosine diphosphate (ADP) and UDP (Ralevic and Burnstock, 1998). Distinct from the other purinoceptors, P2Y14 is potently activated by UDP and a class of carbohydrates, that is, UDP-sugars (Abbracchio et al., 2006; Jacobson et al., 2020). UDP-sugars activate P2Y14 with a relative potency order of UDP-glucose (UDP-Glc), UDP-galactose (UDP-Gal), UDP-glucuronic acid (UDP-GlcA), and UDP-N-acetylglucosamine (UDP-GlcNAc) (Chambers et al., 2000; Hamel et al., 2011; Ko et al., 2009; Ko et al., 2007). These sugar nucleotides act as important signaling molecules via P2Y14 to mediate many physiological processes (Amison et al., 2017; Breton and Brown, 2018; Ferreira et al., 2017; Lazarowski, 2010; Müller et al., 2005; Sesma et al., 2016). Particularly, UDP-Glc regulates immune responses and associate with asthma, kidney injury, and lung inflammation (Amison et al., 2017; Breton and Brown, 2018; Ferreira et al., 2017; Müller et al., 2005; Sesma et al., 2016). As an isomer of UDP-Glc, UDP-Gal is present in various cell models, including physiologically relevant primary cultures of human bronchial epithelial cells (Lazarowski, 2010). It remains unknown how these sugar nucleotides are recognized by P2Y14.

As the closest homolog to P2Y14, P2Y purinoceptor 12 (P2Y12) has not been reported to be activated by any sugar nucleotide (Jacobson et al., 2020; Ralevic and Burnstock, 1998). P2Y12 is potently activated ADP. The reported agonist-bound structures of P2Y12 provide insights to understand the nucleotide activation of P2Y purinoceptors. The crystal structures of P2Y12 show that a full agonist 2-methylthio-adenosine-5’-diphosphate (2MeSADP, a close analogue of ADP) binds to an extracellular pocket consisting of transmembrane (TM) helices and extracellular loops (Zhang et al., 2014a). Since P2Y12 is highly similar to P2Y14 with 45.67% amino acid sequence identity, it would be interesting to investigate whether this receptor is also sensible to sugar nucleotides.

Here, we combined molecular docking, molecular dynamics (MD) simulations, and functional study to reveal the molecular mechanism how P2Y14 is activated by a sugar nucleotide. The ligand-binding models of different UDP-sugars (UDP-Glc, UDP-Gal, UDP-GlcA, and UDP-GlcNAc) were quantitatively characterized to identify the sugar recognition site of P2Y14. Both P2Y14 and P2Y12 were employed to unveil a conserved sugar-binding motif. Multiple carbohydrates were designed and validated as their agonists targeting the conserved functional motif.

Results

Identification of sugar-binding site in P2Y14

Both UDP and UDP-Glc potently activate P2Y14 with EC50 values of 50.9±6.1 nM and 40.3±1.5 nM, respectively (Figure 1A–C). Compared with UDP, UDP-Glc showed an increased potency on P2Y14 at high concentration (Figure 1B–C), suggesting that the sugar moiety of UDP-Glc contributes to activating P2Y14. To investigate how UDP-Glc regulates the P2Y14 via its sugar moiety, we used molecular docking to construct UDP-Glc-bound models of P2Y14 and compared them with UDP-bound P2Y14 models (Figure 1D–G). Because the protein structure of P2Y14 is unrevealed, we employed the X-ray structures of P2Y12 (Zhang et al., 2014a) as templates to constructed homology models of human P2Y14. The molecular docking showed that both UDP and UDP-Glc bind to an extracellular pocket consisting of TM helices 2–7 and extracellular loop (ECL) 2 (Figure 1D), which is corresponding to a known agonist binding pocket of P2Y12 (Zhang et al., 2014a). The docking score of UDP-Glc is –9.3 kcal/mol, and that of UDP is –8.5 kcal/mol, indicating that both UDP and UDP-Glc bind to P2Y14.

Figure 1 with 1 supplement see all
Download asset Open asset
Identification of sugar-binding sites in P2Y purinoceptor 14 (P2Y14).

(A) Chemical structures of uridine diphosphate (UDP) and UDP-glucose (UDP-Glc). (B) Concentration-response curves of calcium mobilization for UDP or UDP-Glc in HEK293 cells transiently co-transfected with human P2Y14 and Gαqi5. Data are shown as mean ± SEM (n=3). See also Figure 1—source data 1 and Supplementary file 1. (C) Concentration-dependent comparison of calcium mobilization for UDP and UDP-Glc in HEK293 cells transiently co-transfected with human P2Y14 and Gαqi5 (n=3); **p<0.01, ***p<0.001, ****p<0.0001. (D) Ligand-bound model of P2Y14. Protein and compound are shown in cartoon and stick representations. (E, F) Docking models of UDP (E) and UDP-Glc (F) to P2Y14. Key residues are highlighted in yellow. Two sub-pockets for ligand binding are marked with circles. (G) Decomposition of ligand-binding energy for each receptor residue (n=10). (H) Calcium mobilization concentration-response curves for UDP or UDP-Glc in HEK293 expressing P2Y14 wild-type (WT) and mutants (n=3). See also Figure 1—source data 2 and Supplementary file 1. (I) Comparison of EC50s for UDP-Glc and UDP in HEK293 cells expressing P2Y14 mutants in calcium mobilization assay (n=3); **p<0.01, ***p<0.001. See also Figure 1—source data 3.

Figure 1—source data 1

Potency of uridine diphosphate (UDP) or UDP-glucose (UDP-Glc) in HEK293 cells expressing P2Y purinoceptor 14 (P2Y14).

https://cdn.elifesciences.org/articles/85449/elife-85449-fig1-data1-v1.docx
Figure 1—source data 2

Potency of uridine diphosphate (UDP) or UDP-glucose (UDP-Glc) in HEK293 cells expressing P2Y purinoceptor 14 (P2Y14) wild-type (WT) and mutants.

https://cdn.elifesciences.org/articles/85449/elife-85449-fig1-data2-v1.docx
Figure 1—source data 3

Comparison of EC50s for uridine diphosphate-glucose (UDP-Glc) and UDP in HEK293 cells expressing P2Y purinoceptor 14 (P2Y14) mutants.

https://cdn.elifesciences.org/articles/85449/elife-85449-fig1-data3-v1.docx

Compared with the UDP-bound receptor model (Figure 1E), the UDP-Glc-bound model showed extra interactions between the glucose moiety and the TM2, TM3, TM7, and ECL2 of P2Y14 (Figure 1F), enhancing the binding of UDP-Glc. Based on these molecular docking models, we further decomposed the ligand-binding energy to each receptor residue (Figure 1G). Five residues (K772.60, D812.64, N903.21, I170ECL2, and E2787.36; superscript indicates Ballesteros-Weinstein residue numbering; Ballesteros and Weinstein, 1995) were predicted to stabilize UDP-Glc binding (Figure 1F and G), while they made few energetic contributions (van der Waals energy >–0.25 Rosetta energy unit) to UDP binding (Figure 1E and G). As shown in Figure 1E and F, two sub-pockets of P2Y14 were unveiled for ligand binding. The sub-pocket 1 is formed by 16 residues of TMs 3–7 and ECL2 (Figure 1G, Figure 1—figure supplement 1A,B) and binds to the nucleotide moiety of the agonist, that is, UDP. The sub-pocket 2 is the specific sugar-binding site involving K772.60, D812.64, N903.21, I170ECL2, and E2787.36 (Figure 1F and G). These residues are primarily charged or polar amino acids, which could made hydrogen bonds with the glucose hydroxyl groups of UDP-Glc (Figure 1F). To validate the proposed sugar-binding sites, we designed single-point mutations of these five residues (K77A, D81A, N90A, I170A, and E278A). Among these mutants, D81A showed significantly reduced activities by UDP-Glc compared with the wild-type (WT) group (Figure 1H). However, substitution of D812.64 by alanine did not significantly affect the receptor activities by UDP (Figure 1H). Interestingly, K77A mutation diminished both UDP-Glc- and UDP-induced calcium mobilization (Figure 1H), but it showed greater impact on UDP-Glc-induced receptor responses than UDP-induced ones (Figure 1H, I), suggesting extra interactions between K772.60 and sugar moiety of UDP-Glc. Two mutations on TM3 and ECL2 (N90A and I170A) did not significantly affect the receptor responses by UDP or UDP-Glc (Figure 1—figure supplement 1C, D). These findings indicate that sub-pocket 2 residues of TM2 provide major contributions to stabilizing the sugar moiety of UDP-Glc.

UDP-Glc as a ‘glue’ for P2Y14 activation

The molecular docking employs rigid side chains of the receptor and does not include the influence of explicit environment on molecular interactions. To investigate how UDP-Glc interacts with P2Y14, we performed all-atom MD simulations of the P2Y14 receptor with and without UDP-Glc (Figure 2A). We used the molecular docking model of P2Y14 to construct the simulation systems. Each system was replicated to performed three independent simulations (see Materials and methods for detailed information). Apo P2Y14 and UDP-Glc-bound P2Y14 simulation models showed different conformations in TM6 and TM7 (Figure 2A and B). In UDP-Glc-bound P2Y14 simulations, the extracellular tip of TM6 shifted over 3 Å and TM7 over 4 Å toward the receptor core, compared with the apo P2Y14 simulations (Figure 2A and B). This inward shift of TM6 and TM7 allowed formation of polar and ionic interactions with the UDP-Glc (Figure 2B and C). During UDP-Glc-bound P2Y14 simulations, two charged residues K2777.35 and E2787.36 formed hydrogen bonds with the glucose 6’ hydroxyl group of UDP-Glc to keep TM7 close to the receptor core (Figure 2B and C), and an arginine residue (R2536.55) formed a salt bridge with the phosphate group of UDP-Glc to stabilize the inward shift of TM6 (Figure 2B and C). Consistently, compared with WT group (EC50 of 40.3±1.5 nM), single-point mutations (R253A and E278A) of TM6 and TM7 helices resulted to diminished UDP-Glc-induced calcium mobilization (EC50 of 808.6±43.6 nM for R253A and 60.2±3.6 nM for E278A) (Figure 1H, Figure 2D). In addition, at the extracellular side, the distance between TM5 and TM6 of UDP-Glc-bound P2Y14 was 5.9 Å shorter than that in the apo system (Figure 2—figure supplement 1A, B). Y1895.41 and T2576.59 made stable hydrophobic interactions to maintain the tight compact between TM5 and TM6 in UDP-Glc-bound-P2Y14 simulations, while TM6 did not interact with TM5 at the extracellular side in the apo simulations (Figure 2—figure supplement 1A, B). Compared with WT group (EC50 of 40.3±1.5 nM), a mutation of TM6 (T257A) showed significantly reduced UDP-Glc-induced responses (EC50 of 504.9±15.9 nM) (Figure 2—figure supplement 1C), fully agreeing with our simulation models. Collectively, these data suggest that UDP-Glc might serve as intramolecular ‘glue’ to make a tight helical bundle of P2Y14, involving TM6 and TM7.

Figure 2 with 1 supplement see all
Download asset Open asset
Comparison of the apo P2Y purinoceptor 14 (P2Y14) and uridine diphosphate-glucose (UDP-Glc)-bound P2Y14 simulation systems.

(A–B) Side (A) and top (B) views of representative models of apo P2Y14 and UDP-Glc-bound P2Y14. Receptor is shown as cartoon. Ligand and key residues are shown as sticks. Movement of the extracellular tips of transmembrane (TM)6 and TM7 toward the receptor core is shown by arrows. See Supplementary file 2 for computational characterization of conformational changes. (C) Key interactions between P2Y14 and UDP-Glc. Hydrogen bonds and salt bridges are displayed as red dashed lines. See Supplementary file 2 for pairwise interaction details. (D) Concentration-response curves of calcium mobilization for UDP-Glc in HEK293 expressing P2Y14 wild-type (WT) and mutants. Data are shown as mean ± SEM (n=3). See also Figure 2—source data 1 and Supplementary file 1.

Figure 2—source data 1

Potency of uridine diphosphate-glucose (UDP-Glc) in HEK293 expressing P2Y purinoceptor 14 (P2Y14) wild-type (WT) and mutants.

https://cdn.elifesciences.org/articles/85449/elife-85449-fig2-data1-v1.docx

Molecular recognition of P2Y14 via sugar-binding site

P2Y14 could be activated by different UDP-sugars with distinct potencies. With only one group substitution at the sugar moiety, UDP-Glc induced stronger activity on P2Y14 (EC50=40.3 ± 1.5 nM) than the other UDP-sugars, that is, UDP-Gal (EC50=78.3 ± 9.2 nM), UDP-GlcA (EC50=59.9 ± 4.8 nM), and UDP-GlcNAc (EC50=184.4 ± 11.8 nM) (Figure 3A and B). To investigate how P2Y14 recognizes different sugar moieties, we performed MD simulations of the human P2Y14 receptor complex with UDP-Gal, UDP-GlcA, and UDP-GlcNAc, respectively, and compared them with the UDP-Glc-bound P2Y14 simulations. We observed that UDP-Gal, UDP-GlcA, and UDP-GlcNAc bound to P2Y14 at the same pocket as UDP-Glc. Similar to UDP-Glc, their uridine groups occupied the sub-pocket 1 of P2Y14, while their sugar moieties bound to the sub-pocket 2 during simulations (Figure 3C–F, Figure 3—figure supplements 1 and 2). At the sub-pocket 2, a stable salt bridging chain formed by four charged residues (K772.60, D812.64, K2777.35, and E2787.36) were observed in all systems (Figure 3C–F). The negative charged glutamic acid residue E2787.36 linked TM2 and TM7 helices by forming salt bridges with K772.60 and K2777.35, while the other negative charged residue D812.64 forming a salt bridge with K772.60 to further stabilize these ionic interactions (Figure 3C–F).

Figure 3 with 4 supplements see all
Download asset Open asset
Sugar moiety recognition of P2Y purinoceptor 14 (P2Y14).

(A) Chemical structures of uridine diphosphate-glucose (UDP-Glc), UDP-glucuronic acid (GlcA), UDP-N-acetylglucosamine (GlcNAc), and UDP-galactose (Gal). (B) Concentration-response curves of calcium mobilization for different UDP-sugars in HEK293 cells transiently co-transfected with human P2Y14 and Gαqi5. Data are shown as mean ± SEM (n=3). See also Figure 3—source data 1. (C–F) Molecular recognition of P2Y14 for UDP-Glc (C), UDP-GlcA (D), UDP-GlcNAc (E), and UDP-Gal (F). Receptor, ligands, and key residues are shown in cartoon and stick representations. Hydrogen bonds and salt bridges are displayed as red dashed lines. See Supplementary file 2 for pairwise interaction details. (G–I) Concentration-response curves of calcium mobilization for UDP-GlcA (G), UDP-GlcNAc (H), and UDP-Gal (I) in HEK293 expressing P2Y14 wild-type (WT) and mutants. Data are shown as mean ± SEM (n=3). See also Figure 3—source data 2.

Figure 3—source data 1

Potency of uridine diphosphate-glucuronic acid (UDP-GlcA), UDP-N-acetylglucosamine (GlcNAc), and UDP-galactose (Gal) in HEK293 expressing P2Y purinoceptor 14 (P2Y14).

https://cdn.elifesciences.org/articles/85449/elife-85449-fig3-data1-v1.docx
Figure 3—source data 2

Potency of uridine diphosphate-glucuronic acid (UDP-GlcA), UDP-N-acetylglucosamine (GlcNAc), and UDP-galactose (Gal) in HEK293 expressing P2Y purinoceptor 14 (P2Y14) wild-type (WT) and mutants.

https://cdn.elifesciences.org/articles/85449/elife-85449-fig3-data2-v1.docx

In simulations, different sugar moieties bound to the K772.60-D812.64-K2777.35-E2787.36 salt bridging chain with distinct binding modes (Figure 3C–F, Figure 3—figure supplement 3). For UDP-Glc, both K2777.35 and E2787.36 could form hydrogen bonds with the glucose 6’ hydroxyl group to keep TM7 close to the receptor core (Figure 3C, Figure 3—figure supplement 4A). However, in UDP-GlcA-bound P2Y14 simulations, at the corresponding position, the 5’ carboxyl group of the sugar moiety failed to form hydrogen bond with the negatively charged E2787.36 (Figure 3D, Figure 3—figure supplement 4B). Compared with that of WT group (EC50=59.9 ± 4.8 nM), the single-point mutation of E278A did not decrease the UDP-GlcA-induced calcium mobilization (EC50=38.2 ± 2.2 nM) (Figure 3G), supporting with the proposed sugar-binding model (Figure 3D). Substitution of 2’ hydroxyl group by an acetamido group led to a rotation of the sugar moiety of UDP-GlcNAc in simulations (Figure 3E). Consequentially, the 6’ hydroxyl of N-acetylglucosamine group flipped to form hydrogen bonds with K772.60 and D812.64 instead of K2777.35 and E2787.36 (Figure 3E, Figure 3—figure supplement 4C). Consistently, single-point mutation of D81A made significant effect to reduce the UDP-GlcNAc-induced receptor activities (Figure 3H). Compared with the other three UDP-sugars, UDP-Gal has a different orientation of 4’ hydroxyl group. The 4’ hydroxyl group of galactose formed a stable hydrogen bond with K772.60 and disrupted the interaction between 6’ hydroxyl group with E2787.36 (Figure 3F, Figure 3—figure supplement 4D). Compared with UDP-Glc, UDP-Gal had more interactions with TM2 and less interactions with TM7 (E2787.36). Substitution of E2787.36 by alanine did not significantly affect the UDP-Gal-induced receptor response (Figure 3I), agreeing with the proposed UDP-Gal-binding model (Figure 3F). For all UDP-sugars, at least three residues of K772.60, D812.64, K2777.35, and E2787.36 participated in ligand binding (Figure 2, Figure 3). Both computational models and experimental data indicate the K772.60-D812.64-K2777.35-E2787.36 salt bridging chain as a sugar-binding site of P2Y14, which can recognize different sugar moieties. The interactions of ligands with the TM7 might determine the ligand potency on P2Y14.

Conserved sugar-binding motif for P2Y12 and P2Y14

In previous sections, we have identified K772.60-D812.64-K2777.35-E2787.36 salt bridging chain as an important functional site for sugar moiety recognition and UDP-sugar activation of P2Y14. These four residues (K2.60, D2.64, K7.35, and E7.36) are conserved between P2Y14 and its closest homolog, that is, P2Y12 (Figure 4A, Figure 4—figure supplement 1A). Therefore, we assumed that P2Y12 can also be activated by carbohydrate ligands. P2Y12 is activated by ADP (Herbert and Savi, 2003), but it has not been reported to be activated by any sugar nucleotide. To validate our assumption, we designed and synthesized three ADP-sugars, that is, ADP-glucose (ADP-Glc), ADP-glucuronic acid (ADP-GlcA), and ADP-mannose (ADP-Man), and then tested whether they can activate P2Y12 (Figure 4B and C). We docked ADP-Glc and ADP-Man to the X-ray structure of P2Y12 (Zhang et al., 2014a). The docking scores are –9.4 kcal/mol for ADP-Glc, for –10.0 kcal/mol for ADP-GlcA, and –9.3 kcal/mol for ADP-Man (Figure 4—figure supplement 1B), suggesting they stably bound to P2Y12. Consistently, in calcium mobilization assays, ADP-Glc, ADP-GlcA, and ADP-Man activated P2Y12 with EC50 values of 3.4±0.4 μM, 1.3±0.1 μM, and 12.3±0.9 μM, respectively (Figure 4C). Single-point mutations of K802.60, D842.64, K2807.35, and E2817.36 significantly diminished ADP-Glc-, ADP-GlcA-, and ADP-Man-induced responses, compared with WT P2Y12 (Figure 4C). These findings not only validate our assumption that P2Y12 can be activated by sugar nucleotides, but also indicate the conserved KDKE salt bridging chain as a functional motif for sugar binding. In addition, we found a less conserved phenylalanine residue F2777.32, which is adjacent to the KDKE motif of P2Y12 (Figure 4—figure supplement 2A). The alanine substitution of F2777.32 significantly reduced ADP-Glc-induced responses of P2Y12, compared with that of the WT group (Figure 4—figure supplement 2B, C). However, single-point mutations at the corresponding position of P2Y14 did not significantly affect the UDP-Glc-induced activation (Figure 4—figure supplement 2C–F). These findings indicate different nature of P2Y12 and P2Y14.

Figure 4 with 4 supplements see all
Download asset Open asset
Adenosine diphosphate (ADP)-sugars binding to P2Y purinoceptor 12 (P2Y12).

(A) Sequence log of the alignment between P2Y12 and P2Y14. Sequences of P2Y12 and P2Y14 involving 359 species were collected for making sequence alignments. See Supplementary file 3 and Figure 4—source data 1 for species repertoire information. The height of a letter is proportional to the relative frequency of that residue at a particular site. Four residues of K2.60-D2.64-K7.35-E7.36 (KDKE) sugar-binding motif are marked by arrows, with the corresponding residues in P2Y12 and P2Y14. (B) Chemical structure of ADP-glucose (Glc), ADP-glucuronic acid (GlcA), and ADP-mannose (Man). (C) Calcium mobilization concentration-response curves for ADP-Glc, ADP-GlcA, and ADP-Man in HEK293 expressing P2Y12 wild-type (WT) and mutants. Data are shown as mean ± SEM (n=3). See also Figure 4—source data 1 and Supplementary file 1.

Figure 4—source data 1

Potency of adenosine diphosphate-glucose (ADP-Glc), ADP-glucuronic acid (GlcA), and ADP-mannose (Man) in HEK293 expressing P2Y purinoceptor 12 (P2Y12) wild-type (WT) and mutants.

https://cdn.elifesciences.org/articles/85449/elife-85449-fig4-data1-v1.docx
Figure 4—source data 2

The neighbor joining tree of P2Y purinoceptor 12 (P2Y12) showing evolutionary range of species.

https://cdn.elifesciences.org/articles/85449/elife-85449-fig4-data2-v1.docx
Figure 4—source data 3

The neighbor joining tree of P2Y purinoceptor 13 (P2Y13) showing evolutionary range of species.

https://cdn.elifesciences.org/articles/85449/elife-85449-fig4-data3-v1.docx
Figure 4—source data 4

The neighbor joining tree of P2Y purinoceptor 14 (P2Y14) showing evolutionary range of species.

https://cdn.elifesciences.org/articles/85449/elife-85449-fig4-data4-v1.docx
Figure 4—source data 5

The neighbor joining tree of GPR87 showing evolutionary range of species.

https://cdn.elifesciences.org/articles/85449/elife-85449-fig4-data5-v1.docx

Discussion

Mediated by UDP-sugars, P2Y14 plays an important role in immune responses and inflammation (Arase et al., 2009; Barrett et al., 2013; Breton and Brown, 2018; Ferreira et al., 2017; Müller et al., 2005; Sesma et al., 2012; Sesma et al., 2016), and possibly insulin resistance (Wang et al., 2008). Breton et al. found that kidney collecting duct intercalated cells present high levels of P2Y14, which is activated by UDP-Glc to promote neutrophil infiltration and renal inflammation (Breton and Brown, 2018). Exerting excessive P2Y14-mediated inflammatory reactions, high concentration of UDP-sugars was observed in extracellular tissue surrounding airway epithelial cells and lung secretions of cystic fibrosis patients (Ferreira et al., 2017; Müller et al., 2005; Sesma et al., 2016). UDP-Glc is also released from liver cells in obese states, possibly via hepatocellular apoptosis, leading to liver inflammation and insulin resistance (Wang et al., 2008). All these results indicate importance of UDP-sugar regulation of P2Y14 in pathological progresses.

In this work, we built the molecular model of UDP-Glc-bound P2Y14 to answer the long-standing question of sugar nucleotide regulation of the purinergic receptor. Binding to an extracellular pocket involving TMs 2 and 7 (Figure 1), the UDP-Glc might serve as intramolecular ‘glue’ attaching to TM6 and TM7 to activate P2Y14 (Figure 2). The agonist-induced remarkable conformational changes of TM6 and TM7 are also reported for P2Y12 (Zhang et al., 2014a). Compared with the AZD1283-bound (antagonist-bound) P2Y12 structure (Zhang et al., 2014b), the extracellular part of TM6 in the 2MeSADP-bound (agonist-bound) P2Y12 structure shifts over 10 Å and TM7 over 5 Å toward the center of TM helix bundle (Zhang et al., 2014a). The close parallel of P2Y12 and P2Y14 in the agonist-induced conformational changes indicates a common ligand-induced activation mechanism shared by purinergic receptors. In addition, in the studies involving the other UDP-sugars, we also found that the interactions between sugar-moieties of agonists with TM7 (E2787.36) is determinant for UDP-sugars’ potencies (Figure 3).

The carbohydrate-binding site has not been fully characterized for GPCRs. Except for P2Y14, it has not been reported that the other members of P2Y12-like subfamily can be directly activated by carbohydrates. Integrated computational modeling with mutagenesis study, we identified a conserved carbohydrate-binding motif (KDKE) for both P2Y14 and P2Y12 (Figure 4). The KDKE motif not only participates in receptor activation by bridging TM2 and TM7 (K772.60, D812.64, K2777.35, and E2787.36) (Figure 2), but also recognize different sugar moieties, including glucose, galactose, glucuronic acid, and N-acetylglucosamine groups (Figure 3, Figure 4). Remarkably, this KDKE motif can distinguish isomers as UDP-Glc and UDP-Gal. Our MD simulations showed that KDKE motif attracted the 6’ hydroxyl group of glucose and interacted with the 4’ hydroxyl group of galactose (Figure 3C and F). Consistent with our observations, a previous structure-activity relationship study revealed that selective mono-fluorination of the 6’ hydroxyl group of the glucose moiety results to fourfold less potency on P2Y14 (Ko et al., 2009). As another member of P2Y12-like subfamily, P2Y13 also has the conserved KDKE site (Figure 4—figure supplement 3), suggesting it might be regulated by carbohydrates. GPR87 is a close homolog of P2Y14 with the sequence identity of 44.94%. GPR87 has K/R2.60, D2.64, K/E7.35, and E7.36 at the corresponding positions of the KDKE sugar-binding motif, indicating varied carbohydrate sensitivities of this receptor in different species (Figure 4—figure supplement 4). The species repertoires of these receptors consist of many amniotes but a few anamniotes (Supplementary file 3a). Distant relatives of the amniote receptors might show different sensitivities to carbohydrates. Similar to the KDKE motif of the receptors, the UDP-sugar-binding sites consisting of charged residues have been discovered for sugar transferases (Gerlach et al., 2018; Hao et al., 2021). In typical glycosylation transfers as TarP and SseK3, two aspartic acids and one positively charged residue (arginine or lysine) participate in recognition of 3’ or 4’ hydroxyl groups of GlcNAc or GalNAc moiety (Gerlach et al., 2018; Hao et al., 2021). However, a salt bridging chain has not been observed in these sugar-binding sites. The different arrangements of UDP-sugar-binding sites between P2Y14 and these sugar transferases might be determinant for their sugar selectivity.

In conclusion, we revealed a conserved carbohydrate-binding motif in both P2Y12 and P2Y14, extending our understanding of how carbohydrates regulate GPCRs. Our molecular models of different sugar nucleotides provide great details for carbohydrate activation and recognition of these receptors, which would inspire further carbohydrate drug development for GPCRs. Whether the other carbohydrate-binding motifs exist in GPCRs is currently unknown. Further investigations focused on carbohydrate regulation of GPCRs will continue to add both new concepts and physiological understanding to the field.

Materials and methods

Key resources table
Reagent type (species) or resourceDesignationSource or referenceIdentifiersAdditional information
Gene (Homo sapiens)P2Y14GenBankNM_001081455.2
Gene (Homo sapiens)P2Y12GenBankNM_022788.5Optimized
Strain, strain background (Escherichia coli)Trans5α Chemically Competent CellTransGen BiotechCat.No: CD201-01
Cell line (Homo sapiens)HEK293ATCCCRL-1573
AntibodyAnti-HA primary antibody
(Rabit monoclonal)		Cell Signaling
Technology		Cat.No: 3724.FCM (1:800)
AntibodyGoat anti-rabbit IgG(H+L) FITC conjugate secondary antibody
(Goat monoclonal)		TransGen BiotechCat.No: HS111FCM (1:200)
Recombinant DNA reagentpCDNA3-HA(plasmid)This paperHA version of pCDNA3
Sequence-based reagentP2Y14-K77A-FThis paperPCR primersGACTTTTCCT
TTCGCGATCC
TTGGTGAC
Sequence-based reagentP2Y14-K77A-RThis paperPCR primersGTCACCAAGG
ATCGCGAAAG
GAAAAGTC
Sequence-based reagentP2Y14-D81A-FThis paperPCR primersCAAGATCCT
TGGTGCCTC
AGGCCTTGG
Sequence-based reagentP2Y14-D81A-RThis paperPCR primersGACCAAGGCC
TGAGGCACCAA
GGATCTTG
Sequence-based reagentP2Y14-N90A-FThis paperPCR primersGTCCCTGGCA
GCTGGCCGTG
TTTGTGTGCAG
Sequence-based reagentP2Y14-N90A-RThis paperPCR primersCTGCACACAAA
CACGGCCAGCT
GCCAGGGAC
Sequence-based reagentP2Y14-I170A-FThis paperPCR primersGAGGTTACACA
AGCAAAATGTA
TAGAACTG
Sequence-based reagentP2Y14-I170A-RThis paperPCR primersGTTCTATACATT
TTGCTTGTGTAA
CCTC
Sequence-based reagentP2Y14-R274A-FThis paperPCR primersCAAAAGAAA
TCTTGGCGT
ATATGAAAG
AATTC
Sequence-based reagentP2Y14-R274A-RThis paperPCR primersGAATTCTTTC
ATATACGCCA
AGATTTCTTTT
G
Sequence-based reagentP2Y14-K277A-FThis paperPCR primersCTTGCGGTAT
ATGGCAGAAT
TCACTCTG
Sequence-based reagentP2Y14-K277A-RThis paperPCR primersCAGAGTGAAT
TCTGCCATAT
ACCGCAAG
Sequence-based reagentP2Y14-E278A-FThis paperPCR primersGCGGTATAT
GAAAGCATT
CACTCTGCT
AC
Sequence-based reagentP2Y14-E278A-RThis paperPCR primersGTAGCAGAG
TGAATGCTTT
CATATACCG
Sequence-based reagentP2Y12-K80A-FThis paperPCR primersCACATTCCC
ATTCGCGAT
CCTGTCAGA
TG
Sequence-based reagentP2Y12-K80A-RThis paperPCR primersCATCTGACAG
GATCGCGAAT
GGGAATGTG
Sequence-based reagentP2Y12-D84A-FThis paperPCR primersCAAGATCCTGT
CAGCTGCCAAG
CTCGGTAC
Sequence-based reagentP2Y12-D84A-RThis paperPCR primersGTACCGAGCTT
GGCAGCTGACA
GGATCTTG
Sequence-based reagentP2Y12-F277A-FThis paperPCR primersGAGAACACTC
TGGCCTACGT
CAAGGAATC
Sequence-based reagentP2Y12-F277A-RThis paperPCR primersGATTCCTTGAC
GTAGGCCAGAG
TGTTCTC
Sequence-based reagentP2Y12-K280A-FThis paperPCR primersCTGTTCTACGT
CGCGGAATCCA
CATTG
Sequence-based reagentP2Y12-K280A-RThis paperPCR primersCAATGTGGATT
CCGCGACGTAG
AACAG
Sequence-based reagentP2Y12-E281A-FThis paperPCR primersGTTCTACGTCA
AGGCATCCACA
TTGTGGC
Sequence-based reagentP2Y12-E281A-RThis paperPCR primersGCCACAATGTG
GATGCCTTGAC
GTAGAAC
Commercial assay or kitKOD-plus-Ver.2TOYOBOCat.No:
KOD-211
Commercial assay or kitMycoBlue Mycoplasma DetectorVazymeCat.No: D101-01
Software, algorithmGraphPad Prism 6GraphPad Prism 6

Chemicals

UDP-GlcNAc was prepared from D-GlcNAc as reported previously (Zheng et al., 2022). UDP-Glc and UDP-GlcA were prepared from Sucrose (Wang et al., 2022). UDP-Gal was prepared from D-Gal (Muthana et al., 2012). ADP-Man was synthesized by a two-step strategy. In detail, Man-1-p was first synthesized from D-Man using NahK from Bifidobacterium longum (Nishimoto and Kitaoka, 2007) and ATP as phosphorylation donor. Man-1-p was purified from the reaction mixture by the silver nitrate precipitation method (Wen et al., 2016; Wen et al., 2015). Then, ADP-Man was synthesized from Man-1-p and ATP by a GDP-mannose pyrophosphorylase from Pyrococcus furiosus, which could take ATP as substrate.

Cell lines

Request a detailed protocol

HEK293 were purchased from ATCC and the identity of the cell line was confirmed by carrying out fingerprinting (Shanghai Genening Biotechnologies Inc, Shanghai, China). In addition, HEK293 cells were tested using the MycoBlue Mycoplasma Detector Kit (Vazyme,China), indicating that the cells were not contaminated by Mycoplasma.

Cell culture and transient transfections

Request a detailed protocol

HEK293 cells were cultured in Dulbecco’s modified Eagle’s medium with 10% fetal bovine serum. All cells were maintained at 37°C in humidified incubators with 5% CO2 and 95% air. Human P2Y14 or P2Y12 receptors and G protein α-subunit (Gαqi5) were transiently co-transfected with HEK293 cells using PolyJet In Vitro DNA Transfection Reagent (SignaGen) according to the manufacturer’s instructions. Thus, a mixture of 1 μg of receptor DNA and 1 μg of Gαqi5 DNA was used to transfect with the six-well plate cells at 90% confluency. HEK293 cells transiently expressing P2Y14 or P2Y12 receptor were subsequently used for the intracellular Ca2+ assays after 48 hr post-transfection.

Cell surface expression

Request a detailed protocol

Human P2Y14 or P2Y12 was cloned into a pcDNA3 vector with HA tag for expression in HEK293 cells. Mutants of P2Y14 or P2Y12 were constructed according to Fast Mutagenesis System (TransGen). Cell surface expression of P2Y14 or P2Y12 was analyzed by flow cytometry. HEK293 cells were transfected with pCDNA3-HA-P2Y14 or P2Y12 in six-well plate overnight. After having been incubated with rabbit anti-HA primary antibody (1:800, CST) for 1 hr at 4°C, the cells were incubated with goat anti-rabbit IgG(H+L) FITC conjugate secondary antibody (1:200, TransGen) for 50 min at 4°C. Data were collected with a flow cytometer (FACS Calibur, BD) and analyzed with FlowJo software.

Intracellular Ca2+ mobilization

Request a detailed protocol

Intracellular Ca2+ assays were carried out as follows. HEK293 cells were seeded (80,000 cells/well) into Matrigel-coated 96-well plate 24 hr prior to assay. The cells were incubated with 2 μM Fluo-4 AM (Invitrogen) diluted in HBSS solution (0.4 g L−1 KCl, 0.12 g L−1 Na2HPO4·12H2O, 0.06 g L−1 KH2PO4, 0.35 g L−1 NaHCO3, 0.14 g L−1 CaCl2, 0.10 g L−1 MgCl2·6H2O, 0.05 g L−1 MgSO4, and 8.0 g L−1 NaCl) at 37°C for 50 min. After dye loading, the cells were treated with the compounds of interest. Then, calcium response (relative fluorescence unit) was measured using Flexstation 3 (Molecular Device) with fluorescence excitation made at 485 nm and emission at 525 nm.

Molecular modeling, docking, and energy decomposition

Request a detailed protocol

Using the crystal structures of agonist-bound P2Y12 (PDB codes 4PXZ, 4PY0) (Zhang et al., 2014a) as templates, we employed Modeller (Sali and Blundell, 1993) to construct the human P2Y14 models. The human P2Y12 models are also built using these P2Y12 crystal structures (PDB codes 4PXZ, 4PY0) (Zhang et al., 2014a). The models with the lowest root mean square deviations from their template structures were selected for further analysis. A ligand was docked to the receptor using Schodinger Glide software in SP mode with default parameters (Friesner et al., 2004). A pocket binding to the ligand with Glide G-scores below –6.5 kcal/mol was considered as a possible ligand-binding pocket. To involve receptor flexibility, we used RosettaLigand (Davis and Baker, 2009) to generate representative ligand-bound receptor models. After Rosetta-based docking, the top 1000 models with lowest binding energy score were selected. Then, they were further scored with the ligand-binding energy between ligand and receptor. The top 10 models with the lowest ligand-binding energy were selected for energy decomposition. The van der Waals energy of ligand binding was mapped to each receptor residue by residue_energy_breakdown utility (Davis and Baker, 2009). The model with the lowest ligand binding energy was used as the representative model.

Modeling and simulations

Request a detailed protocol

To build a simulation system, we place the molecular model into a 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphocholine lipid bilayer. The lipid-embedded complex model was solvated in periodic boundary condition box (80 Å × 80 Å × 120 Å) filled with TIP3P water molecules and 0.15 M KCl using CHARMM-GUI (Wu et al., 2014). Each system was replicated to perform three independent simulations. On the basis of the CHARMM36m all-atom force field (Guvench et al., 2011; Huang et al., 2017; MacKerell et al., 1998), MD simulations were conducted using GROMAS 5.1.4 (Hess et al., 2008; Van Der Spoel et al., 2005). After 100 ns equilibration, a 500 ns production run was carried out for each simulation. All productions were carried out in the NPT ensemble at temperature of 303.15 K and a pressure of 1 atm. Temperature and pressure were controlled using the velocity-rescale thermostat (Bussi et al., 2007) and the Parrinello-Rahman barostat with isotropic coupling (Aoki and Yonezawa, 1992), respectively. Equations of motion were integrated with a 2 fs time step; the LINCS algorithm was used to constrain bond length (Hess, 2008). Non-bonded pair lists were generated every 10 steps using distance cutoff of 1.4 nm. A cutoff of 1.2 nm was used for Lennard-Jones (excluding scales 1–4) interactions, which were smoothly switched off between 1 nm and 1.2 nm. Electrostatic interactions were computed using particle-mesh-Ewald algorithm with a real-space cutoff of 1.2 nm. The last 200 ns trajectory of each simulation was used to calculate average values.

Sequence analysis

Request a detailed protocol

To analyze the conservation of residual sites, we collected sequences of receptors from UniProt database involving 359 species. Among these, we only found five anamniotes, including Astyanax mexicanus, Xenopus tropicalis, Microcaecilia unicolor, Geotrypetes seraphini, and Xenopus laevis. See Supplementary file 3 for species repertoire information. The multiple sequence alignments were performed using Clustal Omega. Logplots generated for these alignments by WebLog. In each logplot, the height of a letter is proportional to the information content of an amino acid in bits, which was calculated by Equation 1 as follows:

(1) I=log2⁡N-∑pilog2⁡pi

where N is the number of all sequences and pi is the probability of the amino acid in all sequences. A large value of the unit bits indicates a high conservation of a particular site.

Statistics

Statistical analyses were performed using GraphPad Prism 6 (GraphPad Software). EC50 values for compounds were obtained from concentration-response curves by nonlinear regression analysis. Comparison of two compounds or two constructs was analyzed by unpaired t test to determine statistical difference. All statistical data are given as mean ± SEM of at least three independent experiments performed in duplicate or triplicate.

Data availability

All data generated or analyzed during this study are included in the manuscript and supplement files.

References

    1. Ballesteros JA
    2. Weinstein H
    (1995)
    Integrated methods for the construction of three-dimensional models and computational probing of structure-function relations in G protein-coupled receptors
    Methods Neurosci 25:366–428.
    1. Ralevic V
    2. Burnstock G
    (1998)
    Receptors for purines and pyrimidines
    Pharmacological Reviews 50:413–492.

Decision letter

  1. Dan Larhammar
    Reviewing Editor; Uppsala University, Sweden
  2. Kenton J Swartz
    Senior Editor; National Institute of Neurological Disorders and Stroke, National Institutes of Health, United States
  3. Dan Larhammar
    Reviewer; Uppsala University, Sweden
  4. Hugo Gutierrez de Teran
    Reviewer; Uppsala University, Sweden

Our editorial process produces two outputs: (i) public reviews designed to be posted alongside the preprint for the benefit of readers; (ii) feedback on the manuscript for the authors, including requests for revisions, shown below. We also include an acceptance summary that explains what the editors found interesting or important about the work.

Decision letter after peer review:

Thank you for submitting your article "Identification of a carbohydrate-recognition motif of purinergic receptors" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, one of whom is a member of our Board of Reviewing Editors, and the evaluation has been overseen by Kenton Swartz as the Senior Editor.

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

I apologize for the delay in handling your manuscript. It took some time to find the second reviewer and when reviewers 1 and 2 had submitted their reports, I was deeply involved in hosting a conference and other time-demanding matters at the end of the academic semester.

Essential Revisions (for the authors):

Reviewing Editor (Recommendations for the authors):

In addition to the comments provided by the other two reviewers, I wish to bring the authors' attention to the following matters:

1) The authors mention the sequence alignment repertoire of species on lines 360, 642-644, and in the legend to Figure 4A, but refer the reader to Supplementary file 3 for specific information apart from the fact that 379 species were included. It is far more important to describe the evolutionary range of species than the number of species, so I urge the authors to provide this information in the text. As far as I can see, the list in the supplementary file includes large numbers of mammals and birds and some turtles, lizards, and a snake, thus a panel consisting of amniotes. A quick look identified no amphibians nor any more distant relatives of the amniotes such as rayfinned fishes or cartilaginous fishes, although my tentative exploration of the members of this receptor family indicates that they existed before the origin of vertebrates. If the species repertoire is restricted to amniotes, this must be clearly stated in the text in the places discussing alignment and sequence conservation.

2) Another aspect of the alignment displayed in Figure 4 is that some positions are extremely difficult to read, for instance, positions 1, 2, 3, 4, and 266. The less frequent residues at many positions are totally impossible to see. But it may suffice to highlight the predominant residue at each position in this way if the authors do not discuss any of the secondary or tertiary amino acids that occur at a certain position.

3) Why does the alignment in Figure 4 use both capital F and small f? All other amino acid abbreviations use capital letters as far as I can see. Position 34 has both F and f – what is the difference? This may be explained somewhere in the text, but if so I must have missed it.

Reviewer #1 (Recommendations for the authors):

I will proceed with a detailed list of suggestions and criticisms following the article reading, irrespective of minor/major suggestions, the last of which will have an easy match for the authors with the weaknesses identified in the public review:

1) Line 59 – typo in the full name of 2MeSADP. Line 61 – please mention EL2 as part of the binding site for this ligand, it makes important interactions.

2) Lines 74-77 – Figure 1B and Figure 1H show a decrease in max EFFICACY for some ligands (UDP-Glc, 1H; K77A and D81A for the UDP-Glc mutations, panel of 1H). The way this is presented is confusing, I would classify UDP as a PARTIAL agonist as compared to UDP-Glc based on this graph. The implications of the effect of the mutations on UDP-Glc are however more delicate, it seems that these residues are involved in the activation of the receptor since the mutation leads to partial efficacy (reduction of the max efficacy) for this ligand. You must use the appropriate pharmacological parameters (Bmax) for presenting and discussing these results.

3) Lines 84-87 – docking scores are not sensible enough to prove this affirmation. The stability of binding has nothing to do with a score of docking, which can only be used to rank docking poses and, in the best cases, to distinguish good and bad binders (according to this data, both UDP and UDP-Glc bind equally well, which correlates with the experimental data and that is as far as you can go).

4) Lines 105 – The effect of a mutation like E278A on the potency (as measured in functional assay) of UDP-Glc is as low as 1.5 fold, which is far from being significant, despite the P value to compare with UDP, and in this sense, I don't see a proof for the specific role of this residue in the sugar binding. K77 and D81 definitely have a role which, despite EC50, modifies Bmax (see my comment below, you must discuss the last value)

5) The MD trajectories are minimally analyzed. The salt bridge analysis and inter-helical distances (Supplementary file 2-appendix table 1) are fine. But for the proposed H-bonds, only the distance between each of the four residues here postulated as forming the carbohydrate-binding site (K277, E278, R253, K77) are monitored with respect to the minimum distance to a sugar atom, (Figure 3 Figure supplement 2). It is well known that the occurrence of a H Bond depends not only on the distance but also on the geometry, so the percentage of occurrence of a H Bond should be reported. In addition, RMSF of the ligand would be informative as to the stability of the binding pose (even if it is an evolved binding pose from the initial docking, we need to verify that this is stable).

6) In this sense, it is stated (l 168 – 170) that E278 causes an electrostatic repulsion only on the UDP-GlcA derivative and this explains the increase in efficacy of this molecule upon E278A mutation. I don't see such a selective repulsion as all molecules except UDP-Glc show the same increased distance to this residue, and as happened before, the experimental data is again overinterpreted since E278A (Figure 3G, compare purple vs black curves) does not enhance the activity of UDP-GlcA in a meaningful way (60 vs 38 nM EC50, p<0.01).

7) The same applies to the discussion of the binding mode of UDP-GlcNAc (l177, Figure 3H). A selective interaction is postulated with D81 based on MD going away from the initial interaction with E278A; but indeed alanine mutation of both residues affects equally. Please modify your interpretation of the data accordingly.

8) Generally speaking, I am convinced that the carbohydrates binding site identified is consistent with the mutagenesis data, but the discussion of specific interactions for each molecule goes beyond what can be demonstrated by the data, and one should draw general lines when not certain of specific interactions, or clearly state these are hypothesis, not all of them backed up by the experimental data.

9) P2Y12 data (Figure 4) – It seems clear that the conserved positions for the four amino acids discussed affect the ADP-sugar activation on P2Y12, but at any point, the authors comment the much lower potencies of the WT for the sugar-derivatives of ADP on this receptor, as compared to the UDP-sugars on P2Y12. The nature of the binding site of the two receptors can probably explain why this difference in potency.

l250 – incomplete sentence, modify accordingly.

Reviewer #2 (Recommendations for the authors):

1. As the author noted in line 117, the molecular docking process only considers the rigidity of the receptor, thereby excluding the receptor's flexibility. To provide a more comprehensive understanding of the system's dynamic properties and validate the docking results, the use of the representative binding pose from molecular dynamic simulations is commendable. Furthermore, it is noteworthy that K77 is relatively far away from the UDP in the present predicted binding pose, despite functional data indicating its importance. MD analysis may offer some insights into this discrepancy.

2. In Figure G, the residue contribution is solely based on the docking results. However, using MM/PB(GB)SA for a similar decomposition analysis can yield more precise outcomes. As such, it would be worthwhile to perform this analysis using the MD trajectory.

3. Having only a single replica with 500 ns of simulation may not provide enough sampling to achieve meaningful results. It is recommended that the author provides the agonist RMSD vs time to assess the convergence of the simulations. Furthermore, performing at least three replicas can increase the statistical significance of the findings and further validate the results.

https://doi.org/10.7554/eLife.85449.sa1

Author response

Essential Revisions (for the authors):

Reviewing Editor (Recommendations for the authors):

In addition to the comments provided by the other two reviewers, I wish to bring the authors' attention to the following matters:

1) The authors mention the sequence alignment repertoire of species on lines 360, 642-644, and in the legend to Figure 4A, but refer the reader to Supplementary file 3 for specific information apart from the fact that 379 species were included. It is far more important to describe the evolutionary range of species than the number of species, so I urge the authors to provide this information in the text. As far as I can see, the list in the supplementary file includes large numbers of mammals and birds and some turtles, lizards, and a snake, thus a panel consisting of amniotes. A quick look identified no amphibians nor any more distant relatives of the amniotes such as rayfinned fishes or cartilaginous fishes, although my tentative exploration of the members of this receptor family indicates that they existed before the origin of vertebrates. If the species repertoire is restricted to amniotes, this must be clearly stated in the text in the places discussing alignment and sequence conservation.

We appreciate your comments and have updated the sequence alignment repertoire of species for P2Y12, P2Y13, P2Y14 and GPR87 in Supplementary file 3. The gene names of these receptors (P2YR12, P2YR13, P2YR14 and GPR87) were used to collect protein sequences in the UniProt database. If a species had multiple entries for the same protein, the entry with the longest sequence and the evidence for protein existence was selected. The updated Supplementary file 3a includes 359 species, which are common in P2Y12, P2Y13, P2Y14 and GPR87. To describe the evolutionary range of species, we generated neighbor joining trees for these 359 species of each receptor (source data for Figure 4) in the revised manuscript.

It is true that large numbers of amniotes are included in the species repertoires. In fact, among the 359 common species (Supplementary file 3a), we only found five anamniotes, including Astyanax mexicanus, Xenopus tropicalis, Microcaecilia unicolor, Geotrypetes seraphini, and Xenopus laevis. We analyzed species for each receptor. P2Y12, P2Y13, P2Y14 and GPR87 involve 420, 419, 460 and 409 species, respectively. For anamniotes, P2Y12, P2Y13, P2Y14 and GPR87 only have 14, 20, 56 and 14 protein sequences, respectively. We further analyzed the sequence alignments for anamniote receptors. As shown in Author response images 1-4, Figure 4—figure supplement 1, Figure 4—figure supplement 3 and Figure 4—figure supplement 4, the KDKE motif are conserved in the 359 common species, but less conserved in the anamniotes. Therefore, distant relatives of the amniote receptors might show different sensitivities to carbohydrates. We have included this information in the text in the place discussing alignment and sequence conservation in the revised manuscript.

Author response image 1
Download asset Open asset
Conservation of each residue on P2Y12 in 14 anamniotes.

The height of a letter is proportional to the relative frequency of that residue at a particular site. Residues of KDKE sugar-binding motif are labeled with red arrows. See Supplementary file 3 for species repertoire information.

Author response image 2
Download asset Open asset
Conservation of each residue on P2Y13 in 20 anamniotes.

The height of a letter is proportional to the relative frequency of that residue at a particular site. Residues of KDKE sugar-binding motif are labeled with red arrows. See Supplementary file 3 for species repertoire information.

Author response image 3
Download asset Open asset
Conservation of each residue on P2Y14 in 359 common species (left) and 56 anamniotes (right).

The height of a letter is proportional to the relative frequency of that residue at a particular site. Residues of KDKE sugar-binding motif are labeled with red arrows. See Supplementary file 3 for species repertoire information.

Author response image 4
Download asset Open asset
Conservation of each residue on GPR87 in 14 anamniotes.

The height of a letter is proportional to the relative frequency of that residue at a particular site. Residues of KDKE sugar-binding motif are labeled with red arrows. See Supplementary file 3 for species repertoire information.

2) Another aspect of the alignment displayed in Figure 4 is that some positions are extremely difficult to read, for instance, positions 1, 2, 3, 4, and 266. The less frequent residues at many positions are totally impossible to see. But it may suffice to highlight the predominant residue at each position in this way if the authors do not discuss any of the secondary or tertiary amino acids that occur at a certain position.

Thanks for your comments. We have remade all alignment plots by highlighting the predominant residue at each position in the revised manuscript (Figure 4—figure supplements 1, 3, 4).

3) Why does the alignment in Figure 4 use both capital F and small f? All other amino acid abbreviations use capital letters as far as I can see. Position 34 has both F and f – what is the difference? This may be explained somewhere in the text, but if so I must have missed it.

Thanks for your comments. We have remade all alignment plots by employing consistent representations for residues in the revised manuscript (Figure 4—figure supplements 1, 3, 4).

Reviewer #1 (Recommendations for the authors):

I will proceed with a detailed list of suggestions and criticisms following the article reading, irrespective of minor/major suggestions, the last of which will have an easy match for the authors with the weaknesses identified in the public review:

1) Line 59 – typo in the full name of 2MeSADP. Line 61 – please mention EL2 as part of the binding site for this ligand, it makes important interactions.

The full name of 2MeSADP is 2-methylthio-adenosine-5'-diphosphate. And we agree that the extracellular loops including ECL2 are part of the binding site for this ligand. We have revised the manuscript to correct the typo and include this information.

2) Lines 74-77 – Figure 1B and Figure 1H show a decrease in max EFFICACY for some ligands (UDP-Glc, 1H; K77A and D81A for the UDP-Glc mutations, panel of 1H). The way this is presented is confusing, I would classify UDP as a PARTIAL agonist as compared to UDP-Glc based on this graph. The implications of the effect of the mutations on UDP-Glc are however more delicate, it seems that these residues are involved in the activation of the receptor since the mutation leads to partial efficacy (reduction of the max efficacy) for this ligand. You must use the appropriate pharmacological parameters (Bmax) for presenting and discussing these results.

We appreciate the above comments. As the reviewer suggested, we performed additional radiolabeled ligand binding assay, try to measure the binding parameters of P2Y14 with UDP or UDP-Glc. Unfortunately, we could not obtain a specific binding signal using the 3H-labeled uridine diphosphate D-glucose (3H-UDP-Glc, PerkinElmer, #NET1163250UC). The reported literature also showed no reasonable specific binding window when using 3H-UDP-Glc as radioligand (J Biomol Screen 2011, 16(9):1098). Based on the current results of intracellular ca2+ mobilization, we agree with the reviewer’s comment that UDP is a partial agonist as compared to UDP-Glc which maybe attribute to the extra binding sub-pocket2 by UDP-Glc and require further validation.

3) Lines 84-87 – docking scores are not sensible enough to prove this affirmation. The stability of binding has nothing to do with a score of docking, which can only be used to rank docking poses and, in the best cases, to distinguish good and bad binders (according to this data, both UDP and UDP-Glc bind equally well, which correlates with the experimental data and that is as far as you can go).

We agree with your comments. The docking score of UDP-Glc is -9.3 kcal/mol and that of UDP is -8.5 kcal/mol, indicating that both UDP and UDP-Glc bind to P2Y14. Accordingly, we have modified the statement in the revised manuscript.

4) Lines 105 – The effect of a mutation like E278A on the potency (as measured in functional assay) of UDP-Glc is as low as 1.5 fold, which is far from being significant, despite the P value to compare with UDP, and in this sense, I don't see a proof for the specific role of this residue in the sugar binding. K77 and D81 definitely have a role which, despite EC50, modifies Bmax (see my comment below, you must discuss the last value)

We agree with your comments and removed the statement that the effect of a mutation (like E278A) on the potency of UDP-Glc is significant in the revised manuscript.

5) The MD trajectories are minimally analyzed. The salt bridge analysis and inter-helical distances (Supplementary file 2-appendix table 1) are fine. But for the proposed H-bonds, only the distance between each of the four residues here postulated as forming the carbohydrate-binding site (K277, E278, R253, K77) are monitored with respect to the minimum distance to a sugar atom, (Figure 3 Figure supplement 2). It is well known that the occurrence of a H Bond depends not only on the distance but also on the geometry, so the percentage of occurrence of a H Bond should be reported. In addition, RMSF of the ligand would be informative as to the stability of the binding pose (even if it is an evolved binding pose from the initial docking, we need to verify that this is stable).

Thanks for the valuable comments. We calculated the percentage of occurrence of hydrogen binding between the ligand and the carbohydrate-binding site (K277, E278, R253 and K77) (Supplementary file 2c) for the proposed hydrogen bonds. We also calculated the ligand RMSF to show the stability of the ligand-binding pose in MD simulations (Figure 3—figure supplement 1). These results have been added in the revised manuscript.

6) In this sense, it is stated (l 168 – 170) that E278 causes an electrostatic repulsion only on the UDP-GlcA derivative and this explains the increase in efficacy of this molecule upon E278A mutation. I don't see such a selective repulsion as all molecules except UDP-Glc show the same increased distance to this residue, and as happened before, the experimental data is again overinterpreted since E278A (Figure 3G, compare purple vs black curves) does not enhance the activity of UDP-GlcA in a meaningful way (60 vs 38 nM EC50, p<0.01).

We agree with your comments. For UDP-Glc, both K2777.35 and E2787.36 could form hydrogen bonds with the glucose 6’ hydroxyl group to keep TM7 close to the receptor core in MD simulations (Figure 3C, Figure 3—Figure supplement 4A). However, in UDP-GlcA-bound P2Y14 simulations, at the corresponding position, the 5’ carboxyl group of the sugar moiety failed to form hydrogen bond with the negative charged E2787.36 (Figure 3D, Figure 3—Figure supplement 4B). Compared with that of WT group (EC50 = 59.9 ± 4.8 nM), the single-point mutation of E278A did not decrease the UDP-GlcA-induced calcium mobilization (EC50 of 38.2 ± 2.2 nM) (Figure 3G), supporting the proposed sugar-binding model (Figure 3D). Accordingly, we have modified the statement in the revised manuscript.

7) The same applies to the discussion of the binding mode of UDP-GlcNAc (l177, Figure 3H). A selective interaction is postulated with D81 based on MD going away from the initial interaction with E278A; but indeed alanine mutation of both residues affects equally. Please modify your interpretation of the data accordingly.

We agree with your comments. Single-point alanine mutation of both D81 and E278 significantly reduce the UDP-GlcNAc-induced receptor activities (Figure 3H). Accordingly, we have modified the statement in the revised manuscript.

8) Generally speaking, I am convinced that the carbohydrates binding site identified is consistent with the mutagenesis data, but the discussion of specific interactions for each molecule goes beyond what can be demonstrated by the data, and one should draw general lines when not certain of specific interactions, or clearly state these are hypothesis, not all of them backed up by the experimental data.

Thanks for your comments. In the revised manuscript, we have removed or modified the statements that are not fully backed up by the experimental data.

9) P2Y12 data (Figure 4) – It seems clear that the conserved positions for the four amino acids discussed affect the ADP-sugar activation on P2Y12, but at any point, the authors comment the much lower potencies of the WT for the sugar-derivatives of ADP on this receptor, as compared to the UDP-sugars on P2Y12. The nature of the binding site of the two receptors can probably explain why this difference in potency.

Discovery of highly potent P2Y12 agonists requires screening of a large number of compounds. In this work, we only designed, synthesized and tested three ADP-sugars (ADP-Glc, ADP-GlcA and ADP-mannose) to validate our assumption that sugar nucleotides can activate human P2Y12. It is possible there are the other ADP-sugars, which are highly potent P2Y12 agonists. It is technically challenging to synthesize ADP-sugars. Currently, we can only obtain ADP-Glc, ADP-GlcA and ADP-Man. Once the other ADP-sugars are available for us, we will test them and try to discover highly potent agonists in the future work.

To investigate the different nature of the binding site of P2Y12 and P2Y14, we identified 22 sugar nucleotide-binding positions in two receptors, including 16 highly conserved positions and six less conserved positions (Author response image 5). Among these less conserved ones, the F7.32 (F277 at 7.32 position) of human P2Y12 is adjacent to the KDKE motif (Figure 4—figure supplement 2A). The alanine substitution of F7.32 significantly reduced ADP-Glc-induced calcium mobilizations of human P2Y12, compared with that of the WT group (Figure 4—figure supplement 2B, C). But the mutations of the corresponding residue of human P2Y14 (R274A or R274F) did not significantly affect the UDP-Glc-induced responses (Figure 4—figure supplement 2D-F). These results suggest the phenylalanine at 7.32 position of human P2Y12 might contribute to the interaction with ADP-Glc, while the residue at the corresponding position of human P2Y14 might not participate in binding of UDP-Glc. This is consistent with the assumption that the nature of the binding site of the two receptors is different. We have included this information in the revised manuscript.

Author response image 5
Download asset Open asset
Sequence log of the alignment between P2Y14 and P2Y12.

Sequences of P2Y14 and P2Y12 involving 359 species were collected for making sequence alignments. See Supplementary file 3 for species repertoire information. The height of a letter is proportional to the relative frequency of that residue at a particular site. Sugar nucleotide-binding residues are highlighted in yellow. Four residues of KDKE sugar-binding motif are marked by arrows, with the corresponding residues in P2Y14 and P2Y12.

l250 – incomplete sentence, modify accordingly.

Thank you for pointing this out. We have modified the sentence in the revised manuscript accordingly.

Reviewer #2 (Recommendations for the authors):

1. As the author noted in line 117, the molecular docking process only considers the rigidity of the receptor, thereby excluding the receptor's flexibility. To provide a more comprehensive understanding of the system's dynamic properties and validate the docking results, the use of the representative binding pose from molecular dynamic simulations is commendable. Furthermore, it is noteworthy that K77 is relatively far away from the UDP in the present predicted binding pose, despite functional data indicating its importance. MD analysis may offer some insights into this discrepancy.

We performed three independent MD simulations for the UDP-bound P2Y14 system. The system setup is same as the UDP-Glc-bound P2Y14 simulation system. The average minimal distance between the UDP and K77 was 5.9 ± 0.1 Å (Author response image 6), consisting with the docking model. This result suggests that K77 does not stably bind to UDP to affect the receptor activation.

Author response image 6
Download asset Open asset
Binding of UDP to P2Y14 in MD simulations.

(A) A representative simulation model showing UDP in the sub-pocket of P2Y14. Key residues are highlighted in yellow. Two sub-pockets for ligand binding are marked with circles. (B) Massive atom distance between UDP and K77 in three independent simulations of UDP-bound P2Y14. Three replicated simulations were indicated by different colors.

2. In Figure G, the residue contribution is solely based on the docking results. However, using MM/PB(GB)SA for a similar decomposition analysis can yield more precise outcomes. As such, it would be worthwhile to perform this analysis using the MD trajectory.

We performed MMPBSA calculation using the MD trajectory of UDP-Glc-bound P2Y14 with implicit membrane. The calculation followed the Amber 20 MMPBSA.py protocol, and as in our manuscript, we used the last 200 ns of the trajectories. The ligand-binding energy was decomposed for each receptor residue (Author response image 7). The total contribution of two negatively charged residues D81 and E278 are more than 25 kcal/mol, suggesting they made negative contributions to UDP-Glc binding. This is inconsistent with our experimental observations that D81 and E278 are key residues for UDP-Glc-induced activation of P2Y14 (Figure 1H). It's important to note that MMPBSA uses an implicit solvent model and thus may not effectively capture solvent-induced interactions between sugar-moiety and charged residues. This discrepancy might explain the deviations between the calculation outcomes and experimental results.

To further characterize the ligand-binding pose, we calculated the percentage of occurrence of hydrogen binding between the ligand and the carbohydrate-binding site (K277, E278, R253 and K77) for the proposed H-bonds (Supplementary file 2c). We also calculated the ligand RMSF to show the stability of the ligand-binding pose in MD simulations (Figure 3—figure supplement 1). These results have been added in the revised manuscript.

Author response image 7
Download asset Open asset
Decomposition of UDP-Glc-binding contribution for each receptor residue using MMPBSA.

3. Having only a single replica with 500 ns of simulation may not provide enough sampling to achieve meaningful results. It is recommended that the author provides the agonist RMSD vs time to assess the convergence of the simulations. Furthermore, performing at least three replicas can increase the statistical significance of the findings and further validate the results.

We performed three replicas for each simulation system in our work as mentioned in the method section. To clarify this, we added simulation information in the main text in the revised manuscript. We calculated the agonist RMSDs vs time to show the convergence of the simulations (Figure 3—figure supplement 2). This information is also included in the revised manuscript.

https://doi.org/10.7554/eLife.85449.sa2

Article and author information

Author details

  1. Lifen Zhao

    State Key Laboratory of Drug Research, Carbohydrate-Based Drug Research Center and National Center for Drug Screening, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai, China
    Contribution
    Validation, Investigation, Methodology, Writing - original draft
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0009-0000-4158-790X

  2. Fangyu Wei

    1. State Key Laboratory of Drug Research, Carbohydrate-Based Drug Research Center and National Center for Drug Screening, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai, China
    2. University of Chinese Academy of Sciences, Beijing, China
    Contribution
    Methodology
    Competing interests
    No competing interests declared

  3. Xinheng He

    1. State Key Laboratory of Drug Research, Carbohydrate-Based Drug Research Center and National Center for Drug Screening, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai, China
    2. University of Chinese Academy of Sciences, Beijing, China
    Contribution
    Visualization, Methodology
    Competing interests
    No competing interests declared

  4. Antao Dai

    State Key Laboratory of Drug Research, Carbohydrate-Based Drug Research Center and National Center for Drug Screening, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai, China
    Contribution
    Investigation, Methodology
    Competing interests
    No competing interests declared

  5. Dehua Yang

    1. State Key Laboratory of Drug Research, Carbohydrate-Based Drug Research Center and National Center for Drug Screening, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai, China
    2. University of Chinese Academy of Sciences, Beijing, China
    Contribution
    Supervision, Investigation, Methodology
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-3028-3243

  6. Hualiang Jiang

    1. State Key Laboratory of Drug Research, Carbohydrate-Based Drug Research Center and National Center for Drug Screening, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai, China
    2. University of Chinese Academy of Sciences, Beijing, China
    3. School of Pharmaceutical Science and Technology, Hangzhou Institute of Advanced Study, Hangzhou, China
    Contribution
    Supervision
    Competing interests
    No competing interests declared

  7. Liuqing Wen

    1. State Key Laboratory of Drug Research, Carbohydrate-Based Drug Research Center and National Center for Drug Screening, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai, China
    2. University of Chinese Academy of Sciences, Beijing, China
    Contribution
    Supervision, Funding acquisition, Methodology, Writing - review and editing
    For correspondence
    lwen@simm.ac.cn
    Competing interests
    No competing interests declared

  8. Xi Cheng

    1. State Key Laboratory of Drug Research, Carbohydrate-Based Drug Research Center and National Center for Drug Screening, Shanghai Institute of Materia Medica, Chinese Academy of Sciences, Shanghai, China
    2. University of Chinese Academy of Sciences, Beijing, China
    3. School of Pharmaceutical Science and Technology, Hangzhou Institute of Advanced Study, Hangzhou, China
    Contribution
    Conceptualization, Supervision, Funding acquisition, Investigation, Methodology, Project administration, Writing - review and editing
    For correspondence
    xicheng@simm.ac.cn
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0003-3735-645X

Funding

Shanghai Municipal Science and Technology Major Project

  • Xi Cheng

Lingang Laboratory grant (LG202102-01-01)

  • Xi Cheng

Lingang Laboratory grant (LG-QS-202206-08)

  • Liuqing Wen

Fund of Youth Innovation Promotion Association (2022077)

  • Xi Cheng

National Key Research and Development Program of China (2021YFA1301900)

  • Xi Cheng

National Natural Science Foundation of China (22007092)

  • Liuqing Wen

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

This work was partially supported by Shanghai Municipal Science and Technology Major Project (to XC and LW), Lingang Laboratory grant (LG202102-01-01 to XC, LG-QS-202206-08 to LW), Fund of Youth Innovation Promotion Association (2022077 to XC), National Key Research and Development Program of China (2021YFA1301900 to XC), and National Natural Science Foundation of China (22007092 to LW).

Senior Editor

  1. Kenton J Swartz, National Institute of Neurological Disorders and Stroke, National Institutes of Health, United States

Reviewing Editor

  1. Dan Larhammar, Uppsala University, Sweden

Reviewers

  1. Dan Larhammar, Uppsala University, Sweden
  2. Hugo Gutierrez de Teran, Uppsala University, Sweden

Version history

  1. Received: December 8, 2022
  2. Preprint posted: March 6, 2023 (view preprint)
  3. Accepted: October 8, 2023
  4. Version of Record published: November 13, 2023 (version 1)

Copyright

© 2023, Zhao et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 109
    Page views
  • 27
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Open citations (links to open the citations from this article in various online reference manager services)

Cite this article (links to download the citations from this article in formats compatible with various reference manager tools)

  1. Lifen Zhao
  2. Fangyu Wei
  3. Xinheng He
  4. Antao Dai
  5. Dehua Yang
  6. Hualiang Jiang
  7. Liuqing Wen
  8. Xi Cheng
(2023)
Identification of a carbohydrate recognition motif of purinergic receptors
eLife 12:e85449.
https://doi.org/10.7554/eLife.85449

Categories and tags

Research organism

Further reading

    1. Computational and Systems Biology
    2. Physics of Living Systems

    Beta-cell intrinsic dynamics rather than gap junction structure dictates subpopulations in the islet functional network

    Jennifer K Briggs, Anne Gresch ... Richard KP Benninger
    Research Article

    Diabetes is caused by the inability of electrically coupled, functionally heterogeneous -cells within the pancreatic islet to provide adequate insulin secretion. Functional networks have been used to represent synchronized oscillatory [Ca2+] dynamics and to study -cell subpopulations, which play an important role in driving islet function. The mechanism by which highly synchronized -cell subpopulations drive islet function is unclear. We used experimental and computational techniques to investigate the relationship between functional networks, structural (gap-junction) networks, and intrinsic -cell dynamics in slow and fast oscillating islets. Highly synchronized subpopulations in the functional network were differentiated by intrinsic dynamics, including metabolic activity and KATP channel conductance, more than structural coupling. Consistent with this, intrinsic dynamics were more predictive of high synchronization in the islet functional network as compared to high levels of structural coupling. Finally, dysfunction of gap junctions, which can occur in diabetes, caused decreases in the efficiency and clustering of the functional network. These results indicate that intrinsic dynamics rather than structure drive connections in the functional network and highly synchronized subpopulations, but gap junctions are still essential for overall network efficiency. These findings deepen our interpretation of functional networks and the formation of functional sub-populations in dynamic tissues such as the islet.

    1. Computational and Systems Biology
    2. Immunology and Inflammation

    Quantitative analyses of T cell motion in tissue reveals factors driving T cell search in tissues

    David J Torres, Paulus Mrass ... Judy L Cannon
    Research Article Updated

    T cells are required to clear infection, and T cell motion plays a role in how quickly a T cell finds its target, from initial naive T cell activation by a dendritic cell to interaction with target cells in infected tissue. To better understand how different tissue environments affect T cell motility, we compared multiple features of T cell motion including speed, persistence, turning angle, directionality, and confinement of T cells moving in multiple murine tissues using microscopy. We quantitatively analyzed naive T cell motility within the lymph node and compared motility parameters with activated CD8 T cells moving within the villi of small intestine and lung under different activation conditions. Our motility analysis found that while the speeds and the overall displacement of T cells vary within all tissues analyzed, T cells in all tissues tended to persist at the same speed. Interestingly, we found that T cells in the lung show a marked population of T cells turning at close to 180o, while T cells in lymph nodes and villi do not exhibit this “reversing” movement. T cells in the lung also showed significantly decreased meandering ratios and increased confinement compared to T cells in lymph nodes and villi. These differences in motility patterns led to a decrease in the total volume scanned by T cells in lung compared to T cells in lymph node and villi. These results suggest that the tissue environment in which T cells move can impact the type of motility and ultimately, the efficiency of T cell search for target cells within specialized tissues such as the lung.

    1. Computational and Systems Biology

    Multicellular factor analysis of single-cell data for a tissue-centric understanding of disease

    Ricardo Omar Ramirez Flores, Jan David Lanzer ... Julio Saez-Rodriguez
    Research Article

    Biomedical single-cell atlases describe disease at the cellular level. However, analysis of this data commonly focuses on cell-type centric pairwise cross-condition comparisons, disregarding the multicellular nature of disease processes. Here we propose multicellular factor analysis for the unsupervised analysis of samples from cross-condition single-cell atlases and the identification of multicellular programs associated with disease. Our strategy, which repurposes group factor analysis as implemented in multi-omics factor analysis, incorporates the variation of patient samples across cell-types or other tissue-centric features, such as cell compositions or spatial relationships, and enables the joint analysis of multiple patient cohorts, facilitating the integration of atlases. We applied our framework to a collection of acute and chronic human heart failure atlases and described multicellular processes of cardiac remodeling, independent to cellular compositions and their local organization, that were conserved in independent spatial and bulk transcriptomics datasets. In sum, our framework serves as an exploratory tool for unsupervised analysis of cross-condition single-cell atlases and allows for the integration of the measurements of patient cohorts across distinct data modalities.