Receptor Activation Upon Agonist Binding

This section briefly reviews the models currently applied to describe GPCRs modus operandi. Attempts are made to draw a picture of the molecular events that occur upon agonist binding that lead to G protein activation. The role of palmitoylation is discussed. Finally, modulation of the interactions with intracellular partners is envisaged in the light of the receptor susceptibility to adopt multiconformational states.

5.1. Ternary Complex Models of Receptor Activation

In the original ternary complex model (TCM) described by De Lean et al. (42), an agonist-bound activated receptor forms a complex with a G protein, resulting in its activation. This corresponds to a simple example of a receptor isomerization mechanism in which ligand-binding (A) promotes a conformation of receptor (R) that couples to and activates a G protein (G). The next level of progression toward present GPCR models involved the incorporation of different receptor conformations into the scheme. The demonstration of constitutive GPCR activity by Costa and Herz (43) indicated that receptors could couple to and activate G proteins in the absence of ligand. This required modification of the original TCM, which did not enable spontaneous formation of the R*G species; this modification resulted in the extended TCM (ETC) (ref. 44; Fig. 3A). According to the ETC, the receptor

Fig. 3. Ternary models of G protein-coupled receptor (GPCR) activation. (A) Extended ternary complex model (ETC) proposed by Samama et al. (64). According to this model, the receptor can spontaneously adopt either an inactive (R) or an active (R*) conformation. Only the activated form (R*) of the receptor can interact with the G protein (G) in the presence or the absence (constitutive activity) of an agonist (A). (B) Cubic ternary complex model (CTC) proposed by Weiss et al. (65). In this more thermodynamically complete representation of GPCR activation, both the inactive state (R) and the active state (R*) of the receptor are allowed to interact with the G protein (G).

Fig. 3. Ternary models of G protein-coupled receptor (GPCR) activation. (A) Extended ternary complex model (ETC) proposed by Samama et al. (64). According to this model, the receptor can spontaneously adopt either an inactive (R) or an active (R*) conformation. Only the activated form (R*) of the receptor can interact with the G protein (G) in the presence or the absence (constitutive activity) of an agonist (A). (B) Cubic ternary complex model (CTC) proposed by Weiss et al. (65). In this more thermodynamically complete representation of GPCR activation, both the inactive state (R) and the active state (R*) of the receptor are allowed to interact with the G protein (G).

exists in an equilibrium between an inactive conformation (R) and an active conformation (R*). In absence of agonist, the inactive form R prevails, but a certain fraction of receptors spontaneously assume the R* state because of the low-energy barrier separating the two conformations. Agonists are predicted to bind with highest affinity to R* and to shift the equilibrium to a larger proportion of receptors under the active conformation. Conversely, inverse agonists that have the ability to inhibit agonist-independent activity (also called constitutive activity) stabilize the inactive conformation R, thereby shifting the equilibrium away from R*. On the other hand, neutral antagonists do not influence the equilibrium between R and R*.

In 1996, Weiss et al. (45) proposed a more thermodynamically complete model called the cubic TCM (CTC; Fig. 3B). In this model, both the active R* and the inactive R conformations of the receptor are allowed to interact with the G protein, whereas in the ETC model only the active R* receptor state could interact with the G protein. It is presently unclear which of these models better predicts and describes experimental findings with GPCRs. On the practical side, the ETC model has fewer parameters and is simpler to use, whereas the CTC model is more comprehensive but has a greater number of nonestimatable parameters. The choice for the appropriate model may be dictated by the importance of the inactive agonist-receptor-G protein (ARG) state: GPCR systems in which the ARG state is negligible can be accurately described by the ETC model, whereas other systems in which the ARG species plays a role (e.g., cannabinoid receptors [46]) require use of the CTC model (ref. 47; Fig. 3).

Increasing evidence points to the existence of multiple conformational states for GPCRs (see Subheadings 5.3. and 5.4.). Additionally, experimental data indicate that neither the ETC nor the CTC model accurately describes the complex behavior of GPCRs. In an attempt to embrace the multiplicity of receptor conformations, multistate models in which the receptor spontaneously alternates between multiple active and inactive states have been proposed (48,49).

5.2. What Do Constitutively Active Mutants and Rhodopsin-Based Models Tell Us About Activation Mechanisms in Class A Receptors?

Some mutations appear to enhance basal activities of GPCRs and, therefore, are believed to mimic the agonist activity and to favor the active state of the receptor. This, in turn, facilitates productive interaction with intra-cellular G proteins. These mutant receptors are currently called constitu-tively active mutants (CAMs). The 5-opioid receptor was the first GPCR described as able to modulate second messengers in the absence of agonist (43). A fairly large number of CAMs were incidentally identified from mutagenesis studies on many different GPCRs. These CAMs contributed massively to the set of data that helps explain the mechanisms of receptor activation. The current hypothesis states that CAMs release the conforma-tional constraints of the GPCR inactive state. This was first postulated for the a1B-AR. Mutation of alanine 293 (A 6.34) and replacement by any of the 19 other amino acids generated a CAM, suggesting that the gain of function resulted from the loss of an intramolecular constraint (50). Indeed, the current belief is that agonist binding to a wild-type receptor introduces new molecular contacts that replace the intramolecular interactions constraining the receptor in an inactive conformation. This results in a conformational switch and subsequent receptor activation. However, many CAMs are likely activated by simple disruption of interactions that exist within the receptor inactive conformation, rather than by formation of new intramolecular bonds. Therefore, it should be remembered that the actual structure adopted by CAMs is only an approximation of the real active conformation of the receptor (for a review, see ref. 51).

The crystal structure obtained for rhodopsin corresponds to the inactive form in which 11-ds retinal is bound, and this serves as a template to postulate movement of helices III, VI, and VII upon light activation. Class A GPCRs share a good number of conserved structural determinants with rhodopsin. Therefore, the high-resolution structure of rhodopsin has been used as template for GPCR modeling of the transmembrane domains, and the helix movement model has been extended to class A receptors as a common mechanism of activation. According to this hypothesis, ligands activate GPCRs by disrupting the networks of intramolecular contacts that stabilize the ground state. This modifies the conformation of the receptor so that it optimally exposes epitopes that bind and stabilize a conformation of the G protein close to the transition state for GDP-GTP exchange and G protein activation.

Despite the availability of a high-resolution structure of rhodopsin at 2.8 A, the actual mechanism used to disrupt stabilizing intramolecular interactions remains elusive. Evidence for movements of helix VI relative to helix III have been essentially provided by several different approaches that were mostly applied to rhodopsin. Biophysical studies included Fourier transformed infrared resonance spectroscopy (FTIR), surface plasmon resonance (SPR), tryptophan ultraviolet (UV)-absorbance spectroscopy, and electron paramagnetic resonance spectroscopy (EPR) (reviewed in ref. 52). Spectral changes were also measured upon N,N'-dimethyl-.V(iodoacetyl)-.V-(7-nitrobenz-2-oxa-1,3-diazol-4-yl)ethylene-diamine (IANBD) binding to cysteine residues in the P2-AR (52,53). Additionally, several indirect strategies were used, including generation of Ws-histidine metal ion-binding sites between cytoplasmic extensions of helices III and VI in rhodopsin receptors (54), P2-ARs (55), and NK1 receptors (41). Cysteine accessibility was also determined in a P2-AR CAM (56) and random mutagenesis was performed on muscarinic m5 (57), 5-opioid (58), AT1A (59), and C5A chemo-attrac-tant (60) receptors.

In rhodopsin and biogenic amine receptors, one key event in the activation process may involve arginine (R3.50) in the highly conserved E/DRY

motif at the cytoplasmic side of helix III (Fig. 1). Protonation of this residue would disrupt the ionic interaction with a glutamic acid (E6.30) at position X1 of a basic "X1BB X2 X3B" motif (where B is a basic amino acid and X is a nonbasic amino acid) located at the junction region between intracelullar loop 3 and helix VI (Fig. 2A). Mutagenesis studies have established this mechanism for 5-HT2A receptors (61), H2 histamine receptors (62), a1B-ARs (63), and P2-ARs (64). Mutagenesis of residues clustered at the junction between helix 3 and intracellular loop 2 in the muscarinic m5 receptor suggested that some of the amino acids adjacent to the E/DRY motif are involved in maintaining the receptor in an inactive state but also alternate with residues required for G protein coupling (65). A similar role in G protein activation was postulated for the N-terminus of intracellular loop 2 in rhodopsin (66) and more recently in the V1A vasopressin receptor (67). 1H NMR analysis established a similar structure for the vasopressin and rhodop-sin intracellular 2 loops (67) but was distinct from the a2A-AR intracellular loop 2 conformation (68). This is of particular interest, because unlike the other two, the a2A-AR is not activated by mutation of the aspartate in the DRY motif and therefore diverges from the consensus model described earlier (69).

In addition to the R3.50-E6.30 salt bridge, the residue X3 (6.34) of the basic motif is hydrogen-bonded to the arginine R3.50 in rhodopsin (12). Introduction of a lysine at position X3 revealed that the residue at position 6.34 is also involved in constraining biogenic amine receptors in an inactive form in the a1B-, a2A-, Pr, and P2-ARs and in the 5-HT1B-, 5-HT2A-, and 5-HT2C-receptors (refs. 61 and 70 and references therein). However, this strategy may not generalize across all receptors. In the case of opioid receptors, the ionic interaction postulated earlier between E6.30 (X1 residue of the basic motif) and R3.50 (in the DRY motif) cannot occur, because the glutamate residue E6.30 on helix VI is replaced by a leucine. Moreover, mutation of T6.34 into a lysine does not activate the ^-opioid receptor (70). These data show that the actual interactions depend on the residues and local environments at the intracellular ends of helices III, V, and VI and that sequence differences in this region are likely to support locally different forms of activation mechanisms (71). Interestingly, in the 5-opioid receptor R258 (6.32), the second basic residue of the "X1BB X2 X3B" motif would be involved in an ionic bridge with E323 (7.43) on helix VII (58).

A group of mutations comprising tryptophan W173 (4.50) that is strictly conserved in all rhodopsin-like GPCRs (despite its location on the most variable helix IV) induced constitutive activation of the 5-opioid receptor (58). This cluster of mutations could either directly or indirectly affect the orien tation of W173, which would play a central role at the helix II-helix IV interface in controlling the orientation and outward motion of helix III during the activation process. W173 is also involved in opioid ligand binding (21) and has been located within the binding crevice in the D2 dopamine receptor (72). Because of its high conservation, W173 may represent a key switch for helix III movements in most GPCRs.

Chen et al. (73) reported that a phenylalanine F303 (6.44) on helix VI is a key residue involved in a1B-AR transmembrane movement that leads to G protein activation. This residue is highly conserved among GPCRs and is located several residues below those identified as being important for ligand interaction and receptor activation in many GPCRs. A similar role has been assigned to the equivalent phenylalanine residue in chemo-attractant C5A (60), muscarinic m5 (57), and cholecystokinin receptors (74). In the muscarinic ml receptor, the conserved F374 (6.44) in helix VI is part of a network of interactions involving a leucine residue L116 (3.43) in helix III and the asparagine N414 (7.49) of the NPXXY motif on helix VII (7, 75). Additionally, an important and specific interaction occurs in rhodopsin between the NPXXY motif and the methionine M257 (6.40) on helix VI (76). In the 8-opioid receptor, mutation of the tyrosine Y318 (7.53) of the NPXXY motif into a histidine or replacement of methionine M262 (6.36) in helix VI by a threonine led to constitutive activity (58). Interestingly, a residue equivalent to M262 is highly conserved among the peptide receptor family, and its mutation in the LH receptor is associated with precocious puberty in humans (77). These data support the view that the conserved NPXXY motif plays a central role in the conformational switch that leads to receptor activation and underscore the importance of networks of hydrophobic interactions in maintaining GPCRs in the inactive state. Following agonist binding, these networks of Van der Waals interactions may be disrupted, resulting in the removal of the hydrophobic latch between helices III, VI, and VII. This, in turn, may induce a rotation of helices VI and VII relative to helix III. From the previous examples, it can also be concluded that although activation of class A GPCRs may be associated with similar conformational changes, different receptors may employ specialized sets of intramolecular interactions to produce these changes.

A whole-receptor random mutagenesis strategy applied to the 8-opioid receptor identified 30 mutations distributed throughout the receptor sequence and allowed researchers to draw a general picture of the events leading to receptor activation (58). The N-terminus, extracellular loop 3, and upper portions of helices VI and VII constitute an outward platform that responds to extracellular ligands and initiates transmembrane signaling.

Movement of at least helices VI and VII throughout the transmembrane core then follows, in addition to local re-arrangement of the helices III, VI, and VII which are proposed for rhodopsin and several biogenic amine receptors. Again, a common structural switch might involve the cytoplasmic ends of helices III and VI identified in several class A receptors (histamine H2 receptors, ^-opioid receptors, ARs, and muscarinic receptors).

Notably, this study identified five amino acid modifications in the N-terminal domain that enhanced spontaneous activity of the 8-opioid receptor (Q12L, D21G, P28L, A30D, and R41Q) (58). Each mutation substantially modified the chemical nature of the amino acid side-chain, introducing or deleting ionic charges or modifying hydrophobicity and structural constraints. This suggests that the N-terminal portion of the receptor is folded as a domain whose structure and spatial orientation influences receptor function. This hypothesis is consistent with the rhodopsin structure, in which the N-terminal domain is folded as a P-sheet and covers the helical bundle like a lid (12). Presently, functional activity of the N-terminal region has been investigated only in glycoprotein hormone GPCRs. For example, the N-ter-minal tail of the TSH receptor has been proposed to bind spontaneously to the empty receptor and act as an inverse agonist favoring the off-state (78). The present data suggest that the short N-terminal domain of some class A GPCRs may also modulate the on-off transition.

5.3. Palmitoylation: A Modulator of Receptor Activity

Palmitoylation is a posttranslational modification that results in the attachment of a 16-carbon-long saturated acyl chain to a cysteine residue. Unlike other acyl chain additions, palmitoylation is a dynamic process. Several studies have suggested that dynamic palmitoylation could modulate receptor activity by influencing the coupling to G proteins as well as the receptor phosphorylation state.

Mutations of C-terminal cysteine residues have been reported for several GPCRs, and a variety of receptor functions were perturbed following these mutations (79-81). These cysteine residues are often believed to be palmitoylated and, therefore, are involved in the formation of a fourth intracellular loop. Dynamic modulation of the local hydrophobicity through palmitoylation may uncover or mask receptor domains that govern interactions with intracellular effectors such as heterotrimeric G proteins or receptor kinases. For example, depalmitoylation of rhodopsin increased its ability to activate Gta-light-dependent GTPase activity (82). Crystallographical data suggest that helix VIII serves in rhodopsin as a membrane-dependent conformational switch that may adopt a helical structure in the inactive state or a looplike conformation upon rhodopsin activation. Thus, one can speculate that palmitoylation of the two rhodopsin cysteine residues might modify the orientation of helix VIII on the membrane surface (83).

Presently, the impact of C-terminal cysteine mutations on the interactions between receptors and G proteins appears to be receptor type-dependent. Coupling to G proteins was either nonaffected or decreased (ref. 79 and references therein; refs. 81 and 84). The lack of palmitoylation subsequent to cysteine replacement was established only in some cases. Therefore, some of the effects observed following cysteine mutagenesis may result either from the loss of the cysteine residues themselves or from the modification they carry.

Substitution of the conserved palmitoylated cysteine residues 328/329 into serines resulted in constitutive activation of the 5-HT(4a) receptor, a receptor coupled to Gas. More recently, mutation of cysteine C328 into an arginine residue in the 5-opioid receptor as well as replacement of cysteines 348/353 by alanine residues in the ^-opioid receptor conferred agonist-independent activity to both Gai/o-coupled receptors (58,85). Although a definite link to the receptor palmitoylation state is still needed for opioid receptors, the data suggest a common role of these residues in the control of receptor activation.

Interestingly, in some cases the lack of palmitoylation appears to have differential effects on the various signaling pathways engaged by a given receptor. A palmitoyl-deficient mutant of the human endothelin (ET)A receptor was reconstituted in phospholipid vesicles together with various G proteins. This mutant was less effective in stimulating Gja and Gqa than the wild-type counterpart, but its ability to stimulate Goa was not affected (86). Similarly, Horstmeyer et al. (87) showed that the unpalmitoylated ETA receptor was still able to couple to Gas, but no longer to Gaq, in Chinese hamster ovary (CHO) cells. On the other hand, an unpalmitoylated triple cysteine mutant in positions 402, 403, and 405 of the receptor ETB was unable to couple to Ga; or Gaq proteins. However, the presence of the palmitoylated cysteine 402 was sufficient to promote coupling to Gaq but not Gaj. In the latter case, additional downstream carboxy-terminal elements appear to be required (88).

Phosphorylation by numerous kinases, including protein kinase A (PKA) and GPCR kinases (GRKs), initiates a cascade of events that leads to receptor desensitization (see Chapter 7). In addition to a role in receptor-G protein coupling, palmitoylation has been proposed as a key determinant in receptor desensitization. Increased palmitate turnover rates upon agonist stimulation has been reported for numerous receptors, including the P2-ARs

(89,90), a2A-ARs (91), muscarinic m2 (92), and 5-HT4areceptors (84). Several studies linked a lack of palmitoylation to an increased level of receptor phosphorylation. Mutation of the palmitoylated cysteine residue improved PKA phosphorylation of the P2-AR (93) and GRK phosphorylation of the adenosine A3 receptor (94). Conversely, introduction by mutagenesis of a PKA phosphorylation consensus motif at a palmitoylation site of the D1 dopamine receptor gave rise to a palmitoylation-deficient mutant that was constitutively desensitized (79). Moreover, palmitoylation at cysteine C356 and phosphorylation at tyrosine Y352 appear mutually exclusive in the bradykinin B2 receptor, maybe as a result of the close vicinity of the two residues (95). Therefore, palmitoylation could be seen as a molecular switch regulating the accessibility of phosphorylation sites involved in receptor downregulation. As mentioned previously, agonist stimulation increases palmitate turnover, thus promoting receptor depalmitoylation (89,90). This, in turn, would unmask phosphorylation sites that render them readily accessible for phosphorylation and would provide a link to receptor internaliza-tion and desensitization. Interestingly, substitution of the carboxy-terminal cysteines by glycine residues decreased both the basal- and agonist-induced level of phosphorylation of the V1a vasopressin receptor that was nonetheless internalized at a faster rate, suggesting receptor-specific effects (81).

Notably, additional palmitoylation sites have been postulated in the intracellular loops for the rat ^-opioid receptor (96) and the V1a vasopressin receptors (81) besides those identified in their C-terminal region; however, no functional role has been assigned to them yet.

5.4. Receptor Multiple Conformations and G Protein Coupling

It was long-believed that a given GPCR interacts with a particular G protein or a given family of G proteins. However, accumulating evidence has now clearly indicated that several GPCRs can simultaneously interact with G proteins that belong to different families and can activate different signaling cascades—some of which exert opposing effects (for a review, see ref. 97). The efficacy of coupling to the various G proteins may then vary according to the receptor type and the interacting G protein but also depends on the agonist.

To date, 23 Ga-subunits have been identified that are classically divided into four different families: Gai/o, Gas, Gaq/11, and Ga12. Six P- and 11 y-subunits that are differentially expressed have also been isolated (e.g., P1y2 are ubiquitous, whereas P1y1 are restricted to visual cells [98]). Consequently, the heterotrimeric combinations that can be observed are dependent on the expression pattern of each of the three components. This implies that the actual coupling of a GPCR to a given heterotrimeric G protein may vary among cells, because it is highly dependent on both the availability of the subunits in a given cell as well as their location in close vicinity to the receptor.

Availability of a,P and y-subunits is not the only factor that influences the type of coupling that will occur between a given receptor and heterotrimeric G proteins. Indeed, determinants that govern the choice of the interacting partners must be present on the receptor. A very large number of studies based on point mutations, chimeras, and synthetic mimetic peptides have pointed to intracellular loops 2 and 3 and the proximal region of the C-terminus as key regions for interaction with and activation of G proteins. Despite a plethora of data, no consensus motifs could be identified as a signature that reflects the interaction of the receptor with one of the four families of Ga proteins (97,99). Fidelity of coupling to a single G protein seems to require a combination of distinct intracellular regions on both in-tracellular loops 2 and 3 (65,100), but the exact molecular determinants that allow the receptor to distinguish among the various G protein subunits remain unclear.

Recently, Slessareva et al. (101) showed that closely related GPCRs achieve selective coupling through multiple and distinct domains located on the G protein a-subunits. This suggests that coupling selectivity ultimately involves subtle and cooperative interactions among various domains on both the G protein and the associated receptor; therefore, multiple conformational states likely exist for a given receptor. Presumably, the various conformations adopted by the receptor are also directly linked to the nature of the agonist. Multiple G protein-coupled states of the P2-AR were evidenced using various guanylyl nucleotide analogs (102). Moreover, changes in the fluorescence of a reporter group born by a purified P2-AR were monitored following binding of agonists; these revealed that the extent of changes depended on agonist efficacy (103). Therefore, a given agonist induces particular structural modifications within the receptor that will ultimately contribute to the selectivity of the coupling with the G protein. Experimental evidence was obtained when activation profiles of a Gai1- or a Gai2-subunit in fusion with the p-opioid receptor were compared upon binding of different agonists. The activation profile of the fused Gai1-subunit closely resembled that observed for the wild-type receptor, which interacts freely with the pool of G proteins present in the cell, whereas activation of the fused Gai2-subunit was only promoted by a very limited number of the agonists tested (104). Similarly, differential activation of Gao1 and Gai1 was observed with the 5-opioid agonist DADLE (105). Recently, plasmon-waveguide resonance spectros-

copy experiments using the 5-opioid receptor confirmed that the affinity of the receptor toward the G protein depends on the agonist, antagonist, or inverse agonist nature of the ligand prebound to the receptor. Moreover, the selectivity of the coupling toward a given Ga-subtype within the Gai/o-fam-ily was demonstrated to depend on the agonist DPDPE (106). Upon cat-echolamine binding, the P2-AR undergoes transitions to two kinetically distinguishable conformational states that were correlated with biological responses in cellular assays. These results support a mechanistic model for GPCR activation in which contacts between the receptor and structural determinants of the agonist stabilize a succession of conformational states with distinct cellular functions (107). The response evoked by the tachykinin NK2 receptor also differed if the receptor bound the complete form of NKA or the naturally occurring truncated NKA 4-10. NKA elevated intracellular calcium level and stimulated cyclic adenosine monophosphate (cAMP) production, whereas NKA 4-10 only affected calcium concentrations. The authors also demonstrated that PKA activation diminished cAMP production, whereas protein kinase C (PKC) activation facilitated the switch from calcium response to cAMP production. To account for these observations, multiple active and desensitized conformations with low, intermediate, or high affinities and with distinct signaling specificities were assumed for the NK2 receptor (108). Similarly, PKA-mediated phosphorylation of the P2-AR switched coupling from stimulatory Gas to inhibitory Gai/o protein (109).

Other mechanisms have been proposed to explain coupling to different pathways. Alternative splicing at the C-terminal region of the receptor may be one additional determinant of receptor-G protein selectivity. In some cases, it also modifies coupling specificity. The strict coupling to Gas observed in the case of the 5-HT4a receptor was enlarged to the Gai/o family in the 5-HT4b variant (110). Similarly, the prostaglandin receptors (111,112) distinguish themselves by their affinity for different G protein families. RNA editing in intracellular loop 3 of the 5-HT2c receptor also affected coupling selectivity and efficiency (113). Additionally, G protein coupling selectivity was reported to be modified by and 5-opioid receptors (114,115), ATI receptors and B2Rs (116), or CCR2/CCR5 (117) heterodimerization (see Part III).

5.5. Interaction With Intracellular Effectors Other Than G Proteins

Within the scope of proteomics, an increasing number of proteins were identified that interact with intracellular loops or with the carboxy-terminal tail of GPCRs (see Part II). Most of the partner candidates have been extracted from yeast two-hybrid screens, glutathione-S-transferase (GST)

pulldown assays, or gel overlays. Some possess enzymatic properties, including receptor specific (GRK) or nonspecific (PKA, PKC) kinases, nitric oxide synthase, calmodulin, or small G proteins such as Arf or RhoA. Others are scaffolding proteins that act as adaptors, including P-arrestin 1/2, MUPP-1, AKAP 79/250, NHERF 1/2; these possess several important functions. They participate in the targeting of GPCRs to specific subcellular compartments but are also responsible for the clustering of these receptors with various effectors. Finally, interacting proteins can regulate GPCR functions in an allosteric manner (for recent reviews, see refs. 10, 118, and 119). In several cases, receptor interactions with these types of molecules were shown to be ligand-dependent (118,120).

Chinese Herbs

Chinese Herbs

Worried For Your Health More Than Ever? Want To PREVENT Rather Than CURE Your Illness? Then WE are the only ones who can answer YOUR concern of time by presenting the exclusive work piece, the explicit and special eBook on CHINESE HERBS- the call of time.

Get My Free Ebook

Post a comment