Despite the structural diversity in the repertoire of the endogenous 7TM receptor agonists, the conformational changes that occur upon receptor activation are believed to be overall identical. Thus, as the last two decades of biochemical and biophysical studies indicate, TM6, and to a minor degree TM7 and TM3, undergo conformational rearrangement during receptor activation [5, 6]. Centered around the highly conserved proline in the middle of TM6 (position VI:15 or 6.50) TM6 is believed to perform movements that results in space creation thereby permitting binding of intracellular signal transduction molecules like G proteins and β-arrestins [7]. (The numbering of amino acids in the helices is provided according to two numbering systems: the generic numbering system suggested by Schwartz [8], followed by the numbering system of Ballesteros and Weinstein [9],) Several crystal structures of 7TM receptors have been presented within the last decade initiated by the structure of bovine rhodopsin [10] followed by the adrenergic receptors [11–15], the adenosine receptors [16–18], additional rhodopsin variants [19–21], muscarinic receptors, [22, 23] and several others [24–26] including the chemokine receptor CXCR4 [27]. In the recent years, crystal structures of not only inactive, but also active 7TM receptors, have been identified. Thus, in the agonist-bound β2-adrenergic receptor, a relatively large rearrangement of the lower segments of TM6 is observed, when compared to the corresponding inactive structure [13–15]. This structural feature is also observed in the crystal structure of opsin in complex with a G protein peptide fragment upon comparison with dark-state rhodopsin [21, 28]. The overall arrangement of the seven transmembrane α-helices delineate the main binding pocket, and most studies in the search of functionally important residues have focused on amino acids facing this main binding pocket (delimited by TM3, TM4, TM5, TM6, and in part TM7). This is with good reason as most small molecule ligands interact with residues in this pocket [11, 12, 29]. Furthermore, most conserved microswitches of functional importance also face the main binding pocket. This includes ArgIII:26 (3.50), which is part of the conserved DRY-motif in TM3, the rotameric toggle switch TrpVI:13 (6.48), which is part of the CWxP-motif in TMVI, and TyrVII:20 (7.53), which is part the NPxxY-motif in TMVII—all of which play crucial roles during receptor activation [30, 31]. However, also residues in the region delimited by TM1, TM2, TM3, and TM7 (the so-called minor binding pocket) function as regulatory switches or major ligand anchor points [32–34].

The DRY-motif is the most conserved motif among the microswitches mentioned above (Figures 2(a) and 2(c)) [21, 30] and has been shown to directly interact with the G protein in a recent crystal structure of the β2-adrenergic receptor in complex with the G_αs-subunit—a crystal that displayed the actual signaling complex and uncovered the importance of both the DRY-motif and the NPxxY-motif in receptor activation [14, 35]. While the overall interaction between the G protein and the receptor is mainly hydrophobic within the transmembrane core, the ArgIII:26 (3.50) is sandwiched between a Tyr in the G protein and TyrVI:20 (7.53) of the NPxxY-motif, highlighting the importance of concerted action of both motifs [14, 35].

The positively charged ArgIII:26 (3.50) has been proposed to be involved in other conformational constrains of importance for receptor activation. Thus, an inactivating salt bridge (a so-called ionic lock) has been suggested between the Arg and another conserved residue, the acidic GluVI:-05 (6.30) in intracellular loop 3 (ICL3) [36]. This ionic lock is broken during receptor activation where TyrVI:20 (7.53) rotates towards the helix bundle as seen in the active crystal structures of both rhodopsin [21] and the β2-adrenergic receptor irreversibly bound to an agonist [37] or stabilized by a nanobody [13]. However, as the GluVI:-05 (6.30) is only conserved among 25 % of all class A receptors [14, 35], and not present in any of the chemokine receptors [38], the molecular interactions involved in conformational constraining of inactive receptor states and the role of ArgIII:26 (3.50) must be different in receptors without GluVI:-05(6.30). Finally, the DRY-motif interacts with ICL2 of the receptor, thereby stabilizing a position of this loop capable of interacting with a hydrophobic pocket on the G protein and directly linking the highly conserved DRY-motif to the receptor/G protein interaction [14].

Furthermore, besides being constitutively active, the viral receptors also signal promiscuously through many pathways, as compared to the predominant G_αi coupling of endogenous chemokine receptors [41, 47]. For instance the ORF74 (open reading frame 74) 7TM receptor encoded by HHV8 (human herpesvirus 8) associates with both G_αi and G_αq [63] as well as signals through MAP kinases [64] leading to the activation of numerous transcription factors, cell proliferation and transformation, VEGF secretion and angiogenesis [64–68]. The ORF74 from herpesvirus saimiri (HVS-ECRF3) also signals through both G_αi, G_α12/13, and G_αq in a ligand-dependent manner, however the constitutive activity of this receptors is constrained to G_αi and G_α12/13, but not G_αq [69, 70]. A similar broad spectrum and promiscuous signaling is also observed for the US28 (unique short 28) and UL33 (unique long 33) 7TM receptors encoded by HCMV, which signals constitutively through both G_αi, and G_αq along with MAP kinases [71–73].

Besides being evolutionary distinct from the endogenous chemokine receptors, the herpesvirus-encoded chemokine receptors cluster in four families (Figure 1): U12/UL33 of HHV6, HHV7, and CMV; U51/UL78 of HHV6, HHV7 and CMV; US27/US28 of CMV; and ORF74 of HHV8 as well as non-human herpesviruses [50]. (UL78 from CMV is evolutionarily conserved with U51 from HHV6 and HHV7. However, as the UL78 receptors have shown no functional homology to chemokine receptor, they have been excluded from current review.) The common feature of encoding chemokine receptors throughout the pox- and herpesviruses suggests that these receptors play an important role in the viral life cycle as well as in circumvention of the host immune system. A few studies of receptor-disrupted viruses have shown diminished replication in vivo in selected tissues [74, 75].

Chemokine receptors are also found in the genomes of poxviruses [76, 77]. In contrast to the broader subfamily resemblance to CC as well as CXC chemokine receptors along with the promiscuous chemokine-binding profile of many herpesvirus-encoded receptors, the poxvirus-encoded receptors solely resemble the CCR8 chemokine receptor (as illustrated in Figure 1) and only interact with CCR8-binding ligands [78, 79]. Briefly, the poxvirus-encoded receptors are located in two areas in the viral genome: 7L and 145R. The best characterized poxvirus receptors are 7L and 145R from YLDV (Yaba-like disease virus) [76, 77].

As mentioned in the introduction, it has been suggested that, in some class A 7TM receptors (e.g., the β-adrenergic receptors and in rhodopsin) the ArgIII:26 (3.50) together with AspIII:25 (3.49) and GluVI:-05 (6.30), the latter located in ICL3, form an ionic lock holding TM3 and TM6 together in the inactive state [10, 36]. During receptor activation, protonation of AspIII:25 (3.49) leads to release of the constraining interaction, thus allowing the outward movement of TM6 [5, 83]. This is also supported by charged- neutralizing mutations of AspIII:25 (3.49) suggesting that this residue is important for receptor activation [84]. However, the negatively charged residue at position VI:-05 (6.30) is not nearly as conserved as the DRY-motif indicating other possible ways to constrain the receptor in the inactive conformation [30]. During receptor activation, the interaction between adjacent Asp/GluIII:25 (3.49) and ArgIII:26 (3.50) of the DRY-motif is disturbed and ArgIII:26 (3.50) is instead able to interact with TyrV:24 (5.58) as well as directly with the G_α protein. This direct interaction with the G protein has been confirmed by the crystal structure of Opsin in complex with a small peptide from the C-terminal of the G_αt protein [21]. Receptor activation opens a pocket at the intracellular site making the interaction with the C-terminal of the G_α protein possible. This allows for the exchange of GDP with GTP thus activating the G protein for further downstream signaling [85].

From the crystal structure of the chemokine receptor CXCR4, it is evident that the overall structure is similar to the other crystal structures of class A 7TM receptors, nevertheless with a few differences mainly in the extracellular part which constitutes the chemokine ligand-recognition domain [27]. Importantly, the chemokine receptors (though belonging to class A) contain a positively charged residue at position VI:-05 (6.30) and hence the ionic lock between ArgIII:26 (3.50) and a negative charged residue at VI:-05 (6.30) does not exist in these receptors [38]. The lack of a negative charge in this position has inspired the introduction of the ionic lock in the chemokine receptors by substitution of the positive charge in VI:-05 (6.30) with a negative charge. This introduction of the putatively correct conditions for the ionic lock resulted in a reduced basal activity of the chemokine receptor CCR5, indicating that the receptor is locked in an inactive conformation with no ligand-induced activation and a strongly impaired ability to bind chemokines. Substitution of ArgIII:26 (3.50) with Ala or Gln maintained chemokine binding though still showed reduced basal activity [38]. Thus, the presence of an ionic lock between TM3 and TM6 leaves the chemokine receptor unable to switch to an active conformation; therefore, chemokine receptors must utilize a different mechanism than the classic ionic lock described above. However, the decrease in basal activity of CCR5 upon introduction of the ionic lock is in accordance with the general interpretation of the role of this motif in class A receptor activation, where loss of the ionic lock (VI:-05 (6.30) mutation), leads to constitutive activity [36]. Disregarding the lack of an ionic lock in chemokine receptors, the DRY-motif still plays an important role in the receptor activation, as exemplified in CCR5, where mutation of ArgIII:26 (3.50) to the neutral Asn disrupted both basal activity and chemokine-induced G_αi protein coupling (through calcium mobilization and GTPγS binding assays) despite retained affinity for CCL4 [86]. Interestingly, an increased basal phosphorylation of the ArgIII:26Asn- (3.50-) mutated receptor was observed along with β-arrestin-mediated endocytosis as well as a higher rate of internalization in response to CCL4 stimulation [86]. Other studies have indicated the need for an intact DRY-motif in β-arrestin1 binding to CCR5, thus supporting the importance of this motif for receptor function [87].

Another class A receptor lacking the ionic lock residue VI:-05 (6.30) is the histamine 4 receptor (H4R), which shows high constitutive activity. Introduction of GluVI:-05 (6.30) did, however, not decrease the constitutive activity [88] as expected from other studies where disruption of an already existing ionic lock leads to increased constitutive activity [36] or like the CCR5 chemokine receptor loss of activity upon ionic lock introduction [38]. Additionally, the H3R also shows constitutive activity in spite of having the putative conditions for an ionic lock [88–90] indicating that the histamine receptors have a functional difference from the general class A 7TM receptors when it comes to activation and constitutive activity. However, the H4R did show complete loss of G protein activation upon mutation of ArgIII:26 (3.50) supporting the importance of the DRY-motif for coupling to G proteins [85, 88, 91]. A similar phenomenon was observed in the H2R where charge-neutralizing mutations of ArgIII:26 (3.50) led to severely decreased basal cAMP production in terms of efficacy indicating diminished G_αs coupling. However, the mutated receptor was able to induce a response upon agonist stimulation though still with lower efficacy compared to wild type [92]. In the same study, it was shown that the charge-neutralizing mutations of ArgIII:26 also resulted in highly structurally instable receptors, where surface expression could only be detected after stabilization with either an agonist or inverse agonist, indicating a role not only in receptor activation, but also in receptor stability for this position [92]. Furthermore, the DRY-motif has also been implicated in receptor stability of the β2-adrenergic receptor [15].

Regarding ArgIII:26 (3.50), one receptor deserves special notice, namely, the CXC chemokine receptor ORF74 from equine herpesvirus 2 (EHV2). This receptor contains a DTW-motif instead of the DRY consensus, thus missing a positive charged residue in position III:26 (3.50). In spite of this, the receptor shows constitutive activity through the G_αi pathway and ligand-mediated signaling in response to the endogenous chemokine CXCL6 [55, 99]. Interestingly, introduction of the DRY-motif in the EHV2-ORF74 led to a 4-fold decrease in constitutive activity while retaining activation by the agonist, CXCL6 [55], suggesting that this receptor has been optimized to act in the absence of a positive charge in the DRY-motif.

As is also evident from Figure 2(b), the virus-encoded chemokine receptors from different pox- and herpesviruses show a larger diversity in the whole DRY-motif; primarily with large deviations within the first residue III:25 (3.49). For instance, the CXC chemokine receptor ORF74 by HHV8 contains a VRY-motif in place of the endogenous DRY-motif. This receptor is associated with Kaposi's sarcoma and shows a high degree of constitutive signaling and stimulates proliferation [100] as well as tumor transformation in mice [64]. The closest endogenous chemokine receptor to HHV8-ORF74 is CXCR2, and studies have shown that replacement of AspIII:25 (3.49) with Val in CXCR2, thus making this receptor more ORF74-like with respect to this motif, leads to constitutive activation of CXCR2 with altered signaling properties in the direction of HHV8-ORF74 signaling [101]. In contrast, the opposite mutation in HHV8-ORF74 (ValIII:25Asp thereby reintroducing the DRY-motif) did not have major effects on either ligand binding or receptor signaling [102]. Another example is the ORF74 receptor from murine herpesvirus 68 (MHV68), which contains an HRC-motif and has been indicated to activate similar oncogenic pathways as HHV8-ORF74 [103, 104]. However, further studies are needed to determine the signaling capabilities of several ORF74 receptors as well as the influence of their altered DRY-motif on constitutive activity and regulatory circumvention.

The human herpesviruses HHV6 and HHV7 both encode two 7TM receptors, U12 and U51, which contain IRY- and ERI-motifs, respectively. HHV6-U12 has been shown to act as a chemokine receptor [105], whereas HHV6-U51 has been shown to be a constitutively active G_αq-coupled CC chemokine receptor [106]. However, neither of these two receptors have the functional impact of the altered DRY-motif been studied.

The UL33 family, consisting of 7TM receptors from murine (M33), rat (R33), and human (UL33) cytomegalovirus, is known to constitutively signal through a vast array of G proteins [72]. The rodent counterparts differ from the human by containing an NRY-motif whereas UL33 contains the conserved DRY-motif. A mutational analysis of M33 revealed that ArgIII:26 (3.50) is important for the viral constitutive signaling in NFAT, CREB, and IP-turnover assays as mutation into a neutral Gln abolished constitutive activation of the receptor. The importance of this was supported by in vivo data where a virus with a missing ArgIII:26 (3.50), NRY changed to NQY, was unable to replicate in the salivary glands [107]. Very interestingly, mutation of AsnIII:25 (3.49) into the consensus AspIII:25 (3.49), NRY to DRY, leads to an increased constitutive signaling (especially through NFAT-mediated transcription) suggesting that the endogenous DRY-motif is preferable for high activity. Having lower receptor activity could be advantageous for the virus as this might be favorable for the virus life cycle [107]. In line with the results for M33, a mutation of ArgIII:26 (3.50) rendered the R33 receptor inactive with respect to G protein coupling [108]. However, in contrast to M33, it was found that replacement of AsnIII:25 (3.49) with the endogenous AspIII:25 (3.49) in R33 (NRY to DRY) did not change the constitutive activity, and that replacement with the nonpolar AlaIII:25 (3.49) (NRY to ARY) led to a diminished PLC stimulation, but an unaltered pertussis toxin-sensitive signaling indicating impaired G_αq-signaling but maintained G_αi.

Another HCMV-encoded receptor, US28, signals constitutively through several pathways such as G_αq/phospholipase C, NFκB, CREB, and MAP kinases [73, 109, 110]. This viral chemokine receptor contains the conserved DRY-motif and a mutational analysis of ArgIII:26Ala found that disruption of the DRY-motif leads to impaired G_αq protein activation and IP-turnover in spite of wild-type levels of cell-surface expression [111]. Like HHV8-ORF74, US28 has been implicated in cancer as the constitutive signaling of the receptor can activate proliferative pathways leading to tumor formation [112–114]. Additionally, HCMV has been found in glioblastomas, which could indicate a possible role in tumorigenesis [115–118].

In general, the viral chemokine receptors, ORF74, encoded by several herpesviruses have very short C-terminal tails when compared to the endogenous receptors (Figure 3). A study of several γ-herpesviruses identified an eight-amino-acid conserved region at the membrane proximal part of the C-terminal tail suggested to play a role in G protein coupling and G_α-selectivity [146]. Especially one basic residue showed importance for G_αq coupling—a residue which is conserved among the endogenous chemokine receptors suggesting an evolutionary conserved function of this residue such as G protein signaling [146]. HHV8-ORF74 is primarily located at the cell surface and deletion of the five terminal amino acids containing 3 phosphorylation sites did not seem to affect cell surface expression, though it did impact the signaling capabilities of the receptor seen by a diminished NFκB and AP-1 signaling [147]. As signaling deficiencies are seen by the removal of just five amino acids, it is tempting to consider that the length of the C-terminal tail has been optimized to only contain necessities and thus demonstrating a minimum requirement for a functional viral receptor tail. Another study also found expression levels of 12 and 24 amino acids C-terminal tail deletions similar to wild-type albeit with reduced constitutive activity in spite of retained ligand regulation by chemokines [148]. It was suggested that the C-terminal helix 8, which is present in the terminal 24 amino acids, is involved in stabilizing the interaction between the receptor and G protein, thus playing a role in mediating signals upon chemokine binding [148]. Though the C-terminal tail of ORF74 appears to be involved in signal mediation, deletion of small parts of the tail including phosphorylation sites does not seem to largely affect cell surface expression or constitutive signaling suggesting that a short tail is enough for the receptor to function to a certain degree. Having such short tails might be a way of evading host regulatory mechanisms, such as GRKs and internalization, thereby ensuring virus-mediated constitutive signaling.