There must be a cell-physiological necessity why many ligands use a complex GPCR, of which the folded protein chain passes through the cell membrane seven times (7 TM receptors) for signaling. Theoretically, they could instead use a single transmembrane protein helix with a ligand binding site located on a “flag” extending into the extracellular environment and a stretch ending in the cytoplasm to which a G protein (complex) can be attached/associated.

This would be the easiest and energetically cheapest way for transducing allosteric changes in the receptor induced by ligand binding. The fact that instead a much less energy-friendly multi-helix bundle complex with (at least) 7 TMs made it in evolution, is a compelling argument that, perhaps, it is not the allosteric change in one or several of the helices itself that is important. An alternative explanation is that a bundle of transmembrane protein helices enables the formation of an intramolecular microchannel for a signaling ion, particularly for H⁺ as in rhodopsins, and for Ca²⁺ as in rhodopsin-descendants (the modern GPCRs). A local and transient change in the concentration of the ion involved then initiates the signaling cascade. This consideration/argument raises the question about the advantages of a ligand binding pocket in an intramolecular microchannel, and about the mechanism, which in unstimulated cells minimizes the untimely gating of this microchannel. We hypothesize that the prenyl-group of G proteins plays a prime role in microchannel gating. Briefly, the model we propose states that prenylation serves the function of an installed hydrophobic flexible molecular valve to restrict the untimely influx of Ca²⁺, analogous to a cork into a bottle neck. The mechanism we propose is neither in conflict with the “consensus model” nor with recent detailed molecular models for GPCR activation obtained with solid-state NMR (Kimata et al., 2015) and Cryo-EM (Liang et al., 2017, 2018; Kang et al., 2018; Safdari et al., 2018). These methods require the use of wetting agents/detergents for sample preparation, and such treatment will disturb attachment of the prenyl group, even if it withstands the treatment at all. Wigglesworth (1969) reported that the use of wetting agents abolished the activities of farnesol and its juvenile hormone (JH) esters in bioassays.

Calcium is better known for its beneficial effects, e.g., as a secondary messenger in signaling pathways, as well as from its most visible role in the manufacture of the skeleton in corals, molluscs, crustaceans, vertebrates, etc., rather than for its toxicity. Yet, Ca²⁺ is the most abundant toxin on earth, and it can act as a secondary messenger because it is toxic (De Loof, 2015, 2017). This is due to the fact that changes in Ca²⁺ concentrations have profound effects on the 3D conformation of various macromolecules, and changes their functionalities as a result. Particularly, the conformation-influencing effect of changing Ca²⁺ concentrations on the proteinaceous contractile apparatus of muscle cells, as well as Ca²⁺-induced changes in chromatin configuration (Lai et al., 2009) are clear examples.

The huge concentration gradient of Ca²⁺ over the plasma membrane, which is about 20,000-fold higher in blood (2 mmol Ca²⁺) (Figure 1) compared with the cytoplasm of unstimulated (resting) cells “drives” Ca²⁺ into cells at any time that Ca²⁺ gates open up.

The lipid bilayer of biomembranes is impermeable to Ca²⁺, but many complex proteinaceous transmembrane proteins permit the passage of Ca²⁺ when properly stimulated to form a transient intramolecular microchannel. Examples are the well-documented types of canonical Ca²⁺ channels. Excess [Ca²⁺]i, exceeding 100 nmol, that entered the cell has to be pumped out of the cytoplasm, quickly and efficiently, by Ca²⁺ pumps located in the plasma membrane, known as Plasma Membrane Ca²⁺ ATPases or PMCAs, and/or by Ca²⁺ pumps in some of the intracellular membrane systems, such as the abundantly present SERCA Ca²⁺ pump (SR Ca²⁺-ATPase) in myocytes (for figures see Orrenius et al., 2003; De Loof, 2015, 2017).

G protein coupled receptors are only one of the participants in the Ca²⁺ homeostasis system. Although the amount of Ca²⁺ that -upon their activation by, e.g., ligand binding- can pass through intramolecular channels is low, the well-documented process of “Ca²⁺-induced Ca²⁺ release” can cause substantial local shifts in [Ca²⁺]i. Pump- and channel activity thus have to be kept in balance on a continuous basis. This requires a finely tuned coordination between all elements that influence Ca²⁺-influx and efflux. Indeed, it does not make sense to activate a PMCA, and simultaneously open Ca²⁺-channels in the plasma membrane. Hence, the different elements involved in Ca²⁺-homeostasis must have been in place and started functioning very early in evolution, enabling the ancestral cells to avoid Ca²⁺-induced cell death. During the next billions of years of evolution, the Ca²⁺homeostasis system has been shaped to near perfection. Ca²⁺ channel- and pump types have been remarkably well-conserved in evolution.

The mevalonate biosynthetic pathway (Figure 2), with farnesol as a key intermediate, functions as a precursor of JH(s) in insects is also evolutionarily very ancient (De Loof and Schoofs, 2019). Interestingly, the mevalonate pathway also displays a prominent role in Ca²⁺ homeostasis (this paper).

The reason why farnesyl- as a molecular valve with a role in restricting untimely Ca²⁺-influx into the cytoplasm model has not been formulated before by other researchers, is that some of the experimental data supporting this model were published half a century ago, thus long before research on GPCRs emerged. At that time, no attention was given to the possible functional importance of farnesol’s high molecular flexibility as indicated by its Rotatable Bond Count of 7 (PubChem: trans, trans-Farnesol), in combination with its horseshoe-shape (see later and Figure 6) and hydrophobicity. In addition, it was assumed that farnesol is neither a hormone, nor an “inbrome” (De Loof et al., 2015). Instead, the general view was that farnesylpyrophosphate only serves a role as a precursor for squalene in the mevalonate pathway, and that farnesol itself, if it would occur at all, has no specific function. That farnesol can have a role in itself in Ca²⁺ homeostasis has, however, been convincingly demonstrated by the electrophysiologists Luft et al. (1999) and Roullet et al. (1999).

They showed that the endogenous sesquiterpenoid farnesol (the trans–trans isomer) is a potent inhibitor of voltage-gated Ca²⁺-channels in rodents and humans (Figure 3); other types of organisms have not yet been tested in this respect. Although these authors assumed that farnesol acts from inside the cells as an inbrome and not as a hormone, a hormonal role for farnesol has been demonstrated in insects. When tested in bioassays that monitor activity of JHs, which are esters of farnesol (for figure, see De Loof et al., 2014), it has been shown that many farnesol-like substances (FLSs), in particular the JHs, are more active than farnesol, some even several orders of magnitude (Wigglesworth, 1969). Thus a hormonal role for farnesol cannot be excluded. However, the evolutionarily ancient role of farnesol/FLS was certainly not a hormonal one, because the mevalonate pathway with farnesol as one of its intermediates was most likely already present in unicellular eukaryote ancestors of all animals, as it also occurs in the choanoflagellate Opisthokonts (Cavalier-Smith, 2017; De Loof and Schoofs, 2019). The high degree of evolutionary conservation of both the mevalonate pathway and the Ca²⁺-voltage-gated channels suggests that the mode of action of farnesol underlying its role in Ca²⁺-homeostasis, may be universal in all eukaryotes.

Through another approach, De Loof (2015) and De Loof et al. (2015) advanced the hypothesis that farnesol may be the natural cognate ligand of the SERCA-Ca²⁺ pump, which is displaced by the SERCA pump blocker thapsigargin. Thapsigargin raises the cytosolic Ca²⁺ concentration by blocking the ability of the cell to pump calcium into the lumen of the sarcoplasmic and endoplasmic reticula (SER and RER). Like farnesol, the plant toxin thapsigargin, is also a sesquiterpenoid. It induces apoptosis like absence of JH does during metamorphosis of holometabolous insects (De Loof et al., 2014).

Store depletion can secondarily activate plasma membrane Ca²⁺ channels allowing an influx of calcium into the cytosol (Rogers et al., 1995; Wikipedia, 2018g). Whether the inhibitory effect of farnesol on voltage-gated Ca²⁺ channels also applies to the Ca²⁺-channels present in the SER and RER has, to our knowledge, never been investigated. During metamorphosis of holometabolous insects the well-documented drop to zero of the titre of endogenous sesquiterpenoids, farnesol and/or JHs, causes programmed cell death/apoptosis in particular in those cell types with a very well-developed RER (De Loof, 2015, 2017). This suggests that the luminal gradient of Ca²⁺ in the SER/RER collapses due to the opening of Ca²⁺ channels. Ca²⁺-induced apoptosis will result (Orrenius et al., 2003). Other mechanisms have since been suggested, e.g., Kowluru (2017) and Brooks et al. (2018).

Farnesol/FLS with a role as a hormone like in insects starts acting at the extracellular side of cells, at the contact site between the blood and the plasma membrane. Next it may diffuse into the intracellular membrane system (De Loof et al., 2014; De Loof, 2015, 2017). Yet, there is an equally important other possible mechanism of action, namely at the border between the cytoplasm and the plasma membrane with its numerous embedded helix bundle transmembrane proteins, in particular the GPCRs and their associated G-proteins. Here prenylation is the mechanism involved. Indeed, farnesyl- that is intracellularly synthesized in the mevalonate pathway, also has non-hormonal activity, as illustrated by its role in Ca²⁺-homeostasis (§2). GPCRs are key cell-surface proteins that transduce external environmental cues into biochemical signals across the membrane (Thal et al., 2018). They are intrinsically allosteric proteins that interact via spatially distinct yet conformationally linked domains with both endogenous and exogenous proteins, nutrients, metabolites, hormones, small molecules, and biological agents (Bondke Persson, 2013). This explains why they play such an important role in cell physiology and in endocrinology. Yet, their possible link with the mevalonate pathway is seldom mentioned in the literature.

This paper advances arguments in favor of the view that such link may help to clarify how allosteric changes in a GPCR may finally result in activation of the two possible downstream pathways (see later and Figure 5). The influx of relatively larger amounts of Ca²⁺ through canonical Ca²⁺ channels is a major event, with important physiological impact. However, in addition to such Ca²⁺ channels, there are also numerous transmembrane proteins in which an intramolecular microchannel exists, that upon being stretched by, e.g., ligand binding-dependent allosteric changes, allows some Ca²⁺ or/and H⁺ to enter the cytoplasm. Thus, in order to keep [Ca²⁺]i low, such micro-channels must also be kept closed as much as possible.

With respect to the evolutionary emergence of GPCRs, cell-physiological archeology (Figure 4) may help to frame an important issue in the transition from prokaryotes to eukaryotes, about 2.7 billion years ago. GPCRs are only found in eukaryotes, including yeast, choanoflagellates and animals. To date they constitute a large protein family of receptors that detect molecules outside the cell and activate internal signal transduction pathways and, ultimately, cellular responses. According to Zhang et al. (2014) GPCRs may have evolved from the prokaryotic world. More specifically, about 80% of “modern” GPCRs are thought to have evolved out of ancient microbial rhodopsins (type-I rhodopsin). Some microbial rhodopsins function as H⁺ pumps, others as cation or anion channels, as Na⁺ or Cl^- pumps or as photosensors (Kaneko et al., 2017, from whom Figure 4 is borrowed, with copyright permission). Thus, in microbial rhodopsins intramolecular transport of inorganic ions is rather the rule than the exception. The chromophore retinal was essential for the proper functioning of microbial (prokaryotic) rhodopsins. A most instructive figure on the functional conversion of rhodopsins in the course of evolution is Figure 2 in the paper by Kaneko et al. (2017).

During the course of evolution, retinal lost its monopoly. Indeed, the number of ligands that bind and activate contemporary GPCRs include light-sensitive compounds, odors, pheromones, hormones, and neurotransmitters, that vary in size from small molecules and peptides to large proteins (Mertens et al., 2004; Caers et al., 2014). This means that the number of different GPCRs is high. Indeed, 7% of all predicted protein-coding genes in the worm C. elegans are GPCRs (Bargmann, 1998; Fredriksson and Schiöth, 2005). Most of them (∼1300) encode nematode-specific chemoreceptors (Frooninckx et al., 2012).

To date, despite the progress obtained in recent high resolution structural studies (Thal et al., 2018), it is not yet fully understood how allosteric transitions (conformational 3D changes) caused by binding of, e.g., a ligand to the binding pocket of a GPCR finally yield a physiological effect inside the cell. Particularly, the role of prenylation remains enigmatic, despite some recent progress in diabetes research (Kowluru, 2017), and in uncovering the role of farnesylation of the transducing γ subunit as a prerequisite for its ciliary targeting (i.e., to outer segments of vertebrate rod photoreceptors) (Brooks et al., 2018).

Based on numerous experimental results, the widely accepted consensus on the general mode of action of GPCRs states that when a ligand binds to a GPCR, it causes a conformational (allosteric) change in the GPCR, which allows it to act as a guanine nucleotide exchange factor (GEF). The GPCR can then activate an associated G protein by exchanging the GDP bound to the G protein for a GTP. The G protein’s _α subunit, together with the bound GTP, can then dissociate from the β and γ subunits to further affect intracellular signaling proteins or target functional proteins directly depending on the α subunit type (G_αs, G_αi/o, G_αq/11, G_α12/13; Wikipedia, 2018d).

Two major classes of G proteins are well-documented (Gilman, 1987; Bastiani and Mendel, 2006; Wikipedia, 2018c). The first class functions as monomeric small GTPases (small G proteins: Rac1, Cdc42, Arf6, Rab27A; Kowluru, 2017), while the second class functions as heterotrimeric G protein(s) (Bastiani and Mendel, 2006, and many other papers). Receptor-activated G proteins are bound to the inner surface of the cell membrane. They consist of the G_α and the tightly associated G_βγ subunits. There are many classes of G_α subunits (Syrovatkina et al., 2016): Gs_α (G stimulatory), Gi_α (G inhibitory), Go_α (G other), Gq/11_α, G_αq, and G12/13_α are some examples. They behave differently in the recognition of the effector molecule, but share a similar mechanism of activation.

The classical well-documented duality in downstream signaling through GPCRs is summarized in Figure 5. Like in most other publicly available images on this topic, G proteins are shown residing closely to the cytoplasmic side of the plasma membrane. In a study on the participation of the G_αq subunit in the 20-OH-ecdysone (20E) non-genomic pathway in larval development and metamorphosis in the insect Helicoverpa armigera, it was shown that before induction by 20E G_αq was distributed throughout the cell, but that it migrated toward the plasma membrane upon 20E induction (Ren et al., 2014). This gives the impression that such translocation after ligand binding is part of some rescue/repair action. In this Helicoverpa system, G_αq is necessary for the 20E-induced intracellular Ca²⁺ release and extracellular Ca²⁺ influx.

Prenylation, which is also called “lipidation,” is the covalent addition of hydrophobic molecules to a protein or chemical compound (Zhang and Casey, 1996) (Figure 6). The _α- and _γ but not the _β G-proteins (Figure 7) are important targets of prenylation-farnesylation. Protein prenylation involves the transfer of either a farnesyl or a geranyl-geranyl moiety to a C-terminal cysteine(s) of the target protein. Three enzymes can carry out prenylation in the cell. They recognize the CaaX box at the C-terminus of the target protein. C is the cysteine that is prenylated. Any protein ending with such a terminus can be prenylated. For figure, see Figure 2 in Shen et al. (2015). Prenylation is an important process to mediate protein–protein interactions and protein-membrane interactions. Through the attachment of a hydrophobic tail a hydrophilic protein can be attached to another protein that is more hydrophobic, or to a membrane. Prenylation is only operational in eukaryotes, not in prokaryotes (Vögler et al., 2008; Wikipedia, 2018e).

A relevant question is what function the attachment of a horseshoe-shaped very flexible hydrophobic tail to a G Protein serves? Is it simply attaching a G protein to a GPCR or to the lipid part of the plasma membrane (Figure 7)? Or, given the horseshoe structure of farnesol, is it more complex?

Our model primarily focusses on the question whether or not farnesyl- has a key function as a chemical valve in restricting the untimely influx of excess Ca²⁺ (and perhaps of H⁺ as well) in GPCRs to which a prenylated G protein attaches. It has to be seen in a much broader context. Indeed, farnesyl- and farnesol are formed in the mevalonate biosynthetic pathway of all eukaryotes, indicating that this pathway is a key player in cell physiology. In vertebrate physiology, farnesylpyrophosphate is best known as the precursor for the biosynthesis of squalene, which is also the precursor for cholesterol. This precursor function is usually considered to be the key role of farnesyl-PP. This view is questioned because some invertebrates, in particular insects, do not have the gene coding for squalene synthase. Hence they cannot synthesize cholesterol by themselves, but they nevertheless did very well in evolution. In addition, farnesol serves as direct precursor for the synthesis of JH, which are only simple esters of farnesol. Vertebrates do not have these esters, but have farnesol (De Loof et al., 2015). Hence that function too is not an essential key function of the mevalonate pathway.

Its presence in the ancient unicellular ancestors of multicellular eukaryotes (Opisthokonts), in particular of animals, suggests that the key function of the mevalonate pathway must be truly essential for cell physiology (De Loof and Schoofs, 2019). Luft et al. (1999) and Roullet et al. (1999) advanced evidence for such an indispensable function, namely in Ca²⁺-homeostasis, more specifically in keeping [Ca²⁺]i low by inhibiting some types of Ca²⁺ channels. These authors only studied voltage-gated channels present in the plasma membrane. Given the multitude of different Ca²⁺ channels, their structural similarity and their distribution in various membrane-types, e.g., the important homotetrameric Ryanodine receptor(s) (RyRs: ryanodine is a plant alkaloid) in the endoplasmic reticulum, it may be possible that at least some of them also have a binding site for endogenous farnesol-like sesquiterpenoids. No experimental data have yet been reported, but it is known that the transmembrane domain of RyRs represents a chimera of voltage-gated sodium and pH-activated ion channels (Efremov et al., 2015; Zalk et al., 2015). Not only Ca²⁺ channels may have a binding site for endogenous sesquiterpenoids, some Ca²⁺ pumps have it as well. The SERCA Ca²⁺-pump has a binding site for a potent sesquiterpenoid blocker, namely thapsigargin. De Loof (2017) suggested that thapsigargin binds with greater affinity to the binding pocket of the still not yet unequivocally characterized endogenous ligand, which, perhaps, may be the endogenous sesquiterpenoid farnesol.

Given the fact that rising [Ca²⁺]i in the cytoplasm is very toxic and thus has to be removed, the question emerges whether farnesol-like endogenous sesquiterpenoids may act as “guards” that control and limit Ca²⁺ entry and passage, not only at voltage-gated Ca²⁺-channels but in all possible routes along which Ca²⁺ can pass, including routes in both the plasma membrane and in intracellular membranes. In our opinion this might very well be the case. Our major argument is that the Ca²⁺-homeostasis system must act as an integrated system involving many contributing factors all keeping [Ca²⁺]i low in the same direction. Indeed, it would be illogical to inhibit a Ca²⁺ channel, and concurrently transport Ca²⁺ in the opposite direction. Perhaps, the overlooked key function of endogenous farnesol-like sesquiterpenoids as one of the triggers for GPCR activation is to hold together the different helices of all(?) types of Ca²⁺transporting transmembrane molecules by means of “a sticky lipidic stopper.”

With respect to future research, a few remarks. Searching the potency of farnesol and some of its esters, the JHs, it has been shown that the trans–trans form is the most active isomer. The cis–cis form is much less active. However, which isomeric form is preferentially used in prenylation and under which conditions isomerization occurs is still unknown. Hopefully, in the near future, alternatives for detergents in extraction procedures, which would leave the prenyl group intact, will be developed, so that, e.g., cryo-EM and NMR studies can be used to localize these groups in the GPCR complexes.

In addition to its mere scientific interest, the role of the mevalonate pathway with respect to Ca²⁺ homeostasis and of prenylation in particular may have as yet undervalued effects in, e.g., Alzheimer’s- and other major disorders (Jeong et al., 2018). In our opinion, if one invests efforts in studying the relationship between a given disorder in which Ca²⁺-homeostasis is known to play a role, e.g., Alzheimer (Mattson and Chan, 2001; Mattson, 2012; Jeong et al., 2018) one should give attention to possible shifts in isomers, e.g., from trans–trans farnesol which naturally occurs in brain tissue and which is a good blocker of voltage-gated Ca²⁺ channels (Roullet et al., 1999), to other less potent isomers. To date, nothing is known about the presence of the less potent isomers in the healthy- and disease-states. Perhaps, some of the prenylation-dependent disorders have a problematic prenylation flip-flopping system?

Our final conclusion is that farnesol-like endogenous sesquiterpenoids with their very ancient evolutionary origin, deserve to become “noble knowns,” particularly in basic and applied biomedical research.

All authors listed have made a substantial, direct and intellectual contribution to the work, and approved it for publication.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.