ABSTRACT
Dendritic cells (DCs) expressing the chemokine receptor XCR1 are specialized in antigen cross-presentation to control infections with intracellular pathogens. XCR1-positive (XCR1+) DCs are attracted by XCL1, a γ-chemokine secreted by activated CD8+ T cells and natural killer cells. Rat cytomegalovirus (RCMV) is the only virus known to encode a viral XCL1 analog (vXCL1) that competes for XCR1 binding with the endogenous chemokine. Here we show that vXCL1 from two different RCMV strains, as well as endogenous rat XCL1 (rXCL1) bind to and induce chemotaxis exclusively in rat XCR1+ DCs. Whereas rXCL1 activates the XCR1 Gi signaling pathway in rats and humans, both of the vXCL1s function as species-specific agonists for rat XCR1. In addition, we demonstrate constitutive internalization of XCR1 in XCR1-transfected HEK293A cells and in splenic XCR1+ DCs. This internalization was independent of β-arrestin 1 and 2 and was enhanced after binding of vXCL1 and rXCL1; however, vXCL1 appeared to be a stronger agonist. These findings suggest a decreased surface expression of XCR1 during DC cultivation at 37°C, and subsequent impairment of chemotactic activity and XCR1+ DC function.
This article has an associated First Person interview with the first author of the paper.
INTRODUCTION
G protein-coupled receptors (GPCRs) represent an abundant protein family characterized by possessing seven transmembrane domains that induce an intracellular signal transduction pathway resulting in a cellular response. GPCRs are classified into six groups (A to F) based on their sequence similarity and functionality. Chemokine receptors, belonging to class A, are localized on immune cells and become active after ligand binding (Bacon et al., 2003; Jockers et al., 2012; Thiele and Rosenkilde, 2014). Conformational changes of chemokine receptors induce several signaling pathways associated with chemotaxis, intracellular Ca2+ release and other secondary messengers (Mackay, 2001; Murphy, 1994; Steen et al., 2014). Chemokines are classified into the four subfamilies CXC (α-chemokines), CC (β-chemokines), C (γ-chemokines) and CX3C (δ-chemokines) which were defined by the number and position of their N-terminal cysteine residues (Rollins, 1997). XCL1 is the only member of the C motif (γ-chemokine) subfamily (Dorner et al., 1997; Yoshida et al., 1995) that binds the GPCR XCR1. Whereas only one ligand (XCL1) is present in mice, two versions (XCL1 and XCL2) exist in the human that differ by two amino acids (Fox et al., 2015; Yoshida et al., 1996). The XCL1–XCR1 axis is involved in the cellular immune response after infection with pathogens such as viruses.
Endogenous (host) XCL1 is secreted predominantly by natural killer (NK) cells in the early innate response as well as by CD8+ T cells in the adaptive phase, and it induces chemotaxis only in XCR1-expressing dendritic cells (DCs) that are likely to represent conventional dendritic cells 1 (cDC1; Daws et al., 2019; Guilliams et al., 2014; Voisine et al., 2002). XCR1 is predominantly co-expressed on CD8a+ mouse cDC1, CD141+ human cDC1 and CD4− rat DCs, and has been suggested as a lineage marker to phenotype antigen cross-presenting DCs since XCR1+ DCs are involved in the induction of cytotoxic CD8+ T cell responses (Crozat et al., 2010; Gurka et al., 2015; Kroczek and Henn, 2012). Murine and human XCR1+ DCs were reported to cross-present antigen to CD8+ T lymphocytes (Bachem et al., 2010, 2012; Crozat et al., 2010), thereby highlighting the importance of this DC subset in infection immune control.
Cytomegalovirus (CMV), a member of the Herpesviridae, encodes various proteins with immune evasive functions to ensure its persistence in the host. Some of these viral proteins interfere with the chemokine network of the host by imitating host chemokines or chemokine receptors (Rosenkilde, 2005). The genomes of the rat CMV (RCMV) isolates England and Berlin were found to encode the only known viral γ-chemokine, viral XCL1 (vXCL1). The vXCL1 amino acid sequence is highly similar to the endogenous sequence of XCL1 present in rats, mice and humans. Rat XCL1 (rXCL1) and vXCL1 share ∼60% of their amino acid sequence, suggesting that this gene originated from a co-evolutionary adaptation process between virus and host (Geyer et al., 2015, 2014).
As XCR1+ DCs play a pivotal role in cytotoxic T cell induction, XCR1+ DCs might be targeted by CMV. It has long been known that CMV exploits the host chemokine system by means of virus-encoded chemokines that operate as agonists or antagonists. Several α- and β-chemokines have been identified in various CMV species that attract particular cell subpopulations to benefit viral dissemination (Saederup and Mocarski, 2002). Human CMV (HCMV)-encoded UL146, a vCXCL1 analog, induces Ca2+ flux and chemotaxis of neutrophils by activating chemokine receptors CXCR1 and CXCR2 (Lüttichau, 2010; Penfold et al., 1999). Mouse CMV (MCMV)-encoded m131/129 (murine chemokine 2, MCK-2) was shown to facilitate viral dissemination by recruiting myelomonocytic leukocytes and inhibiting activation and cytotoxicity of CD8+ T cells (Daley-Bauer et al., 2012; Noda et al., 2006). This viral β-chemokine analog is also conserved in the English isolate of RCMV (RCMV-E), where it was designated ECK-2 (Voigt et al., 2005).
In previous work, we demonstrated that RCMV-E-encoded vXCL1 is a post-translationally modified protein. Cleavage of the N-terminal signal peptide as well as N-and O-glycosylation were confirmed for vXCL1. CD4− XCR1+ rat DCs were selectively chemoattracted and bound by vXCL1. XCR1 was the only targeted interaction partner, as both rat XCL1 and vXCL1 competed for XCR1 on CD4− rat DCs. Hence, it was concluded that vXCL1 functionally resembles rat XCL1 (Geyer et al., 2014). In addition, the recently isolated RCMV Berlin (RCMV-B) was also shown to encode vxcl1, in analogy to the gene discovered in RCMV-E (Geyer et al., 2015).
Since both the endogenous and the viral chemokine interacted with XCR1, they might influence receptor internalization. Internalization is an important regulatory mechanism of chemokine receptor activity that can occur in both ligand-dependent and -independent manners (Neel et al., 2005). Failure to internalize upon ligand activation is linked to pathologic conditions such as the CXCR4-associated warts, hypogammaglobulinemia, immunodeficiency and myelokathexis (WHIM) syndrome caused by receptor mutations that restricts CXCR4 downregulation upon CXCL12 activation (Balabanian et al., 2005). Some viruses have been shown to exploit this pathway by encoding chemokines that modify the availability of human receptors, while the HCMV-encoded promiscuous chemokine receptor US28 has adapted this function for removal of endogenous chemokines (Bodaghi et al., 1998), which is considered to inhibit the host immune response.
In this study, we show that the vXCL1 analogs from two RCMV isolates act as agonists for XCR1 and selectively activate the XCR1 Gi signaling pathway in rats. While the human pathway was activated both by human XCL1 and rXCL1, vXCL1 exclusively activated rat XCR1 but not human XCR1, thereby defining vXCL1 as a species-specific agonist. In addition, we demonstrate constitutive internalization of XCR1 in transfected HEK293A cells that occurred independently of β-arrestin 1 and 2 (hereafter denoted β-arrestin 1/2) expression. Internalization also occurred after ligand binding in splenic XCR1+ DCs, where vXCL1 appeared to be a stronger agonist than rXCL1 given that XCR1 surface expression was further decreased in the presence of the viral γ-chemokine analog.
RESULTS
vXCL1 binds to and attracts XCR1+ CD4− DCs
The vXCL1 analogs of RCMV-B and RCMV-E share an amino acid identity of 83.3% with conserved cysteine residues at positions 30 and 67 in Fig. 1A. RCMV-B and RCMV-E vXCL1 have a total length of 114 and 115 amino acids, respectively. RCMV-B vXCL1 shares 67% and 53% identity with rXCL1 and hXCL1, respectively. RCMV-E vXCL1 is 64% identical to the endogenous XCL1 of Rattus norvegicus and 46% identical to human XCL1 (hXCL1). To examine whether these amino acid differences have an impact on XCR1 binding, recombinant rXCL1, RCMV-E vXCL1 and RCMV-B vXCL1 were overexpressed in Drosophila SL-3 cells containing a C-terminal StrepTag. Glycosylation in Drosophila cells is less complex than in higher eukaryotes; however, a direct effect of glycosylation on chemokine binding or function is unknown. Prior to assessing binding, the integrity of the recombinant chemokines was analyzed by SDS-PAGE and western blotting, which revealed a difference in size between vXCL1 and rXCL1 (Fig. 1B). Both RCMV-E and RCMV-B vXCL1 show dispersed bands at ∼22 kDa, whereas rXCL1 has a size of ∼16 kDa. RCMV-B vXCL1 western blots showed a second band at 10 kDa. Furthermore, each recombinant chemokine was verified by mass spectrometry as having a mass of ∼10 kDa, which matches the protein size without the signal peptide (data not shown).
After successful expression, DCs were incubated with the recombinant chemokines and analyzed by flow cytometry using an antibody that binds XCR1 in a non-blocking fashion. RCMV-B vXCL1 and RCMV-E vXCL1, as well as rXCL1 bound exclusively to XCR1+ CD4− DCs but not to DCs that lacked XCR1 expression (Fig. 2A). Then, to determine whether there is any difference in binding efficiency between the chemokines, XCR1+ CD4− DCs were incubated with different concentrations of both viral XCL1 analogs and rXCL1, respectively. RCMV-E vXCL1 showed higher binding efficiency compared with RCMV-B vXCL1 and rXCL1 at chemokine amounts below 1×10−7 M, but lower binding efficiency when 1×10−7 or 2×10−7 M were used (Fig. 2B). In addition, a Transwell assay was performed to test the migratory potential of XCR1+ CD4− DCs towards different concentrations of vXCL1 and rXCL1, respectively. Recombinant rXCL1 was used as a positive control. Migration of CD103-sorted rat DCs was analyzed by excluding B cells (CD45RA+) and T cells (CD3+) as those cell populations showed no migration towards rXCL1 or vXCL1 (Geyer et al., 2014). XCR1+ CD4− DCs was the only cell type that migrated towards RCMV-E vXCL1 and RCMV-B vXCL1 (Fig. 2C). The highest migration of XCR1+ DCs towards either vXCL1 analog was observed at a concentration of 1×10−8 M whereas the highest migration towards rXCL1 occurred at a concentration of 1×10−7 M. In contrast, XCR1− CD4+ DCs did not specifically migrate towards the respective chemokines. These results indicate that XCR1+CD4− rat DCs represent the equivalent DCs subset to CD8a+ XCR1+ mouse and CD141+ XCR1+ human DCs.
vXCL1 is a species-selective super-agonist for rat XCR1
GPCR ligands can exert their function either as agonists, antagonists or neutral ligands. vXCL1 has previously been demonstrated to function as an agonist on naturally occurring XCR1+ CD4− DCs (Geyer et al., 2014). To substantiate the selective interaction of vXCL1 with XCR1, a G protein signaling assay was performed in COS-7 cells transiently transfected with either rat XCR1 (rXCR1) or human XCR1 (hXCR1), where we determined the abilities of vXCL1, rXCL1 and human XCL1 to induce G protein activation (Fig. 3). As expected, endogenous rXCL1 activated rXCR1, and hXCL1 activated hXCR1 at nanomolar concentrations (white symbols). However, whereas rXCL1 also acted as an agonist on the hXCR1 receptor with a potency similar to that for rXCR1 – and thereby crossing the species barrier – this was not the case for hXCL1, which only activated hXCR1 (Fig. 3A). The opposite selectivity was observed for RCMV-E and RCMV-B vXCL1, which activated rXCR1 at ∼10-fold higher potencies than rXCL1 but were not able to activate hXCR1 notably at physiologically relevant concentrations. With increased potencies and efficacies compared to rXCL1, the viral ligands thus both acted as super-agonists on rXCR1 for G protein activation (Fig. 3B). These findings highlight a highly species-specific adaptation of RCMV-E and RCMV-B to their host.
vXCL1 binding results in diminished XCR1 surface expression
Since most GPCRs internalize after agonist binding (Violin and Lefkowitz, 2007), except for ligands acting in a biased manner (Steen et al., 2014), internalization of XCR1 was examined. For these experiments, HEK293A wild-type or Δβ-arrestin 1/2-knockout cells were transiently transfected with a SNAP-tagged version of rXCR1 to measure real-time internalization. In the absence of ligands, rXCR1 internalized rapidly at 37°C (white symbols, Fig. 4A). This rapid internalization was also observed upon pre-incubation of the cells at 4°C followed by an increase to 37°C (black symbols, Fig. 4A). Intriguingly, rXCR1 internalization was not dependent on β-arrestins as it occurred to the same extent in cells lacking β-arrestin 1/2 both at 37°C as well as at 4°C (Fig. 4A, square symbols). Addition of either rat XCL1 or the viral XCL1 analogs did not alter the rapid internalization (data not shown). Using the same method, we determined rXCR1 expression. Expression was not significantly different between the wild-type and knockout cells, indicating that expression is similar in both cell lines (Fig. 4B).
Next, we investigated XCR1 internalization in cells naturally expressing the receptor. Ligand-independent reduction of XCR1 surface expression was also observed in CD103-enriched DCs. Without addition of chemokine, the proportion of XCR1+ CD4− cells in the total DC population was reduced from 36% after 1 h to ∼14–18% after 4 h and 8 h of incubation at 37°C. However, after incubation with rXCL1 and vXCL1, a much lower degree of receptor surface expression was observed (Fig. 5A). After addition of 1×10−8 M of rXCL1 or vXCL1, XCR1 expression was significantly reduced, especially at later time points. Both vXCL1 and rXCL1 binding reduced XCR1 expression to ∼3–5% of all DCs. In a second step, we examined whether StrepTagged-ligands could be detected after binding to XCR1. DCs were stained with anti-StrepTag antibodies to determine whether the ligand–receptor complex was internalized. In the presence of chemokine, ∼55–60% of CD4− DCs (upper quadrants) expressed XCR1 after 1 h of incubation. XCR1 expression significantly decreased to 10–20% of cells after 4 h and 8 h when DCs were incubated with rXCL1 or vXCL1. Ligand and receptor could barely be detected after cell surface staining, especially at later time points, indicating that XCR1 internalization was enhanced upon binding of vXCL1 and rXCL1 (Fig. 5B).
Moreover, to determine the amount of ligand that is necessary to internalize the receptor–ligand complex, CD103-enriched DCs were incubated with the chemokines at different concentrations. vXCL1 led to increased reduction of XCR1 cell surface expression compared with rXCL1. This effect was seen at chemokine concentrations above 5×10−9 M (Fig. 6).
In order to analyze whether the decreased surface expression seen in the previously described experiments was specific for XCR1, we stained other DC surface markers (Fig. 7). After staining for 1 h, CD103+ DCs showed decreased surface expression for XCR1 but not MHCII and CD4 both in the absence as well as in the presence of recombinant chemokines (Fig. 7A,B). XCR1 surface expression continued to decrease over time and was further reduced at 4 h (Fig. 7C) and 8 h (Fig. 7D), while MHCII and CD4 expression remained stable.
Finally, since ligand–receptor engagement led to reduced XCR1 surface expression, we investigated whether the chemotactic activity of XCR1+ DCs was influenced after pre-incubation with either rXCL1 or vXCL1. After pre-incubation, DCs were transferred to the upper chamber of the Transwell insert. The lower chamber contained rXCL1, which would usually attract the cells; however, after chemokine pre-incubation, chemotaxis was compromised. XCR1 expression in the CD4− DC population was reduced to ∼5% of cells; in addition, expression was reduced to 20% of cells for the chemokine-lacking control (Fig. 8A). Ligand-independent internalization of XCR1 led to reduced migration towards rXCL1. The presence of both the vXCL1 and rXCL1 chemokines resulted in a higher reduction and in almost completely loss of chemotactic activity. Pre-incubation with rXCL1 or vXCL1 resulted in reduced migration of CD4− DCs (Fig. 8B).
DISCUSSION
XCL1 is the only endogenous ligand that targets XCR1; however, several viruses have exploited the interaction between XCL1 and XCR1. Kaposi sarcoma-associated herpesvirus (KSHV) encodes three β-chemokines vCCL-1, vCCL-2 and vCCL-3 (Nicholas, 2007) of which two interact with human XCR1. Whereas vCCL-2 has been shown to act as an antagonist, vCCL-3 has a highly selective agonistic function on XCR1 (Lüttichau, 2008; Lüttichau et al., 2007). Furthermore, the HHV-6 immunomodulatory β-chemokine receptor U51A interacts with XCL1 (Catusse et al., 2008). Hence, the XCL1–XCR1 axis appears to be an attractive target for viruses.
CMV-encoded chemokines exhibit immune evasive functions but also serve as mediators of viral dissemination by interacting with various cells that are involved in the immune response of the host (Saederup and Mocarski, 2002). The two isolates of RCMV described here encode variants of a novel γ-chemokine analog that is unknown in other viruses. The two vXCL1 proteins differ from each other and from the endogenous rat XCL1 protein in their amino acid sequences at the termini, and such changes, especially at the N-terminus, have been implicated in receptor binding and activation (Clark-Lewis et al., 1995; Gong and Clark-Lewis, 1995; Kroczek et al., 2018). For this reason, we initially analyzed the ability of the viral and endogenous chemokines to attract XCR1+ DCs and induce chemotaxis. Our results show that both vXCL1 and rXCL1 bound exclusively to XCR1+ CD4− DCs and induced migration in this DC subset. RCMV-E vXCL1 and RCMV-B vXCL1 attracted slightly higher numbers of XCR1+ CD4− DCs than host rXCL1 after incubation of DCs with lower chemokine concentrations. Furthermore, 1×10−8 M of recombinant RCMV-E vXCL1 induced the strongest migration – up to 80% of XCR1+ CD4− DCs. RCMV-B vXCL1 was less potent, attracting ∼60% of the same DC subpopulation. These differences could be due to structural changes based on altered primary sequence that might impact stability, folding and activity of the mature protein. It has been demonstrated that natural XCL1 has weak chemotactic activity unless its structure is stabilized by double disulfide bonds (Kelner et al., 1994; Matsuo et al., 2018; Tuinstra et al., 2007, 2008). The enhanced interaction between RCMV vXCL1 with XCR1+ DCs could be the result of a co-evolutionary process during which the viral analog was optimized resulting in a more potent affinity for rXCR1.
An earlier analysis revealed glycosylation sites of vXCL1 that might also have an impact on receptor binding and increased chemotaxis (Geyer et al., 2014). Glycosylation was indicated by differences in protein size with SDS-PAGE, and RCMV-E vXCL1 exhibited a higher level of glycosylation than rXCL1. RCMV-B vXCL1 had a similar size to RCMV-E vXCL1, indicating glycosylation of the protein. However, an additional band at 10 kDa was expressed, which might represent the size of the mature vXCL1 RCMV-B protein without glycosylation sites, and this could explain the observed weaker interaction with XCR1+ CD4− DCs. In addition, amino acid sequence differences potentially affecting XCL1–XCR1 binding properties might have an impact on target cells. The N-terminus of XCL1 contains a signal peptide, which is cleaved off for secretion between serine 19 and isoleucine 20, as was previously shown for RCMV-E vXCL1 (Geyer et al., 2014), and is therefore not important for receptor binding.
In previous work, we have shown that XCR1 is predominantly expressed by CD4− rat DCs, and that vXCL1 and rXCL1 selectively attract and compete for XCR1 (Geyer et al., 2014). While flow cytometry experiments revealed binding of the ligands to the receptor, it was unknown if this actually resulted in conformational changes and activation of intracellular G proteins. If activation occurred, these G proteins would dissociate into GTP-bound subunits and initiate signaling pathways that ultimately control DC migration (Liu and Shi, 2014). Control of DC migration might be important for the virus, and vXCL1 might be key to a misdirection of DCs and, ultimately, impair the immune response. Therefore, we examined whether activation of XCR1 produced a more potent signal than the signal induced by the endogenous ligand. Indeed, both vXCL1 variants activated XCR1 more than rXCL1. This has been also demonstrated for KSHV vCCL-3, a selective XCR1 agonist whose potency was 10-fold higher compared to human XCL1 with regard to chemotaxis, phosphatidylinositol turnover and Ca2+ mobilization (Lüttichau et al., 2007). In contrast, CMV vCXCL1 encoded by UL146 acts as an agonist for CXCR1 and CXCR2, but does not seem to activate these receptors more potently than the endogenous human ligands CXCL5, CXCL7 or CXCL8 (Lüttichau, 2010). In addition, we tested whether XCR1 activation occurs only in the presence of a species-specific ligand since different ligands can activate highly related GPCRs (Howard et al., 2001). We used both human and rat XCR1, and found that while vXCL1 and human XCL1 only activated their corresponding receptors, rXCL1 activated both the rat and the human XCR1 with equal potency. Binding a highly related receptor has been demonstrated with murine XCL1; in that case, the molecule exhibited a 3-fold higher potency than human XCL1 on human XCR1. However, human XCL1 and KSHV vCCL3 were unable to activate murine XCR1 (Lüttichau, 2008). Why are human, mouse or rat chemokines preferentially better suited to cross the species barrier? Viruses have adapted well to their host during evolution and it has been speculated that viral chemokines have lost the ability to do so (Lüttichau, 2008). Our finding that vXCL1 is a selective agonist for rat XCR1, as it activated its receptor with >100-fold higher potency compared to the human XCR1, supports this idea.
Upon receptor serine/threonine phosphorylation, β-arrestin binding advances G protein inactivation and ultimately leads to receptor internalization in clathrin-coated vesicles. However, β-arrestin-independent internalization has been reported for several GPCRs (Luttrell and Lefkowitz, 2002; van Koppen and Jakobs, 2004). We therefore examined receptor internalization after ligand binding by incubation of enriched DCs with recombinant RCMV-E vXCL1 and host rXCL1. XCR1 expression was significantly decreased on the surface of CD4− rat DCs after chemokine incubation compared to cells without addition of chemokine. A reduced receptor surface expression without addition of chemokine was expected since it is known that cultured DCs become activated, which results in receptor internalization. Moreover, it seemed that the receptor internalized in the presence of bound ligand and thus vXCL1 might have affected dynamics of receptor internalization. However, it is unknown how fast XCR1 is degraded after ligand binding and internalization. HCMV has been reported to induce the downregulation of cell surface expression of CCR1 and CCR5 through receptor internalization on DCs, which inhibited migration of immature DCs to sites of inflammation (Varani et al., 2005). Furthermore, typical internalization of chemokine receptors occurs upon ligand binding (Feniger-Barish et al., 1999; Zimmermann et al., 1999). Our data show that vXCL1 is necessary for chemoattraction of XCR1+ CD4− rat DCs and is responsible for receptor internalization after XCR1+ binding; however, we did not detect decreased surface expression of other DC molecules. In addition, we observed reduced surface expression over time in HEK293 cells. The disappearance of XCR1 from the DC surface after binding of vXCL1 during DC cultivation at 37°C might reflect altered receptor recycling properties or altered receptor transport to the cell surface from intracellular stores, and it is possible that DCs lose their ability to communicate with other immune cells expressing endogenous XCL1. Therefore, blocking the migratory ability of DCs would be an effective viral strategy to paralyze immune responses early after infection.
The function of various DC subsets and their role in the immune response during viral infection as well as the finely tuned balance between viral immune evasion and DC action have been intensively investigated (Alexandre et al., 2014). DCs represent an important cell population in pathogen defense, as they continuously patrol their surroundings and take up foreign antigen. Since XCR1+ DCs appear to play an important role in virus elimination, the attraction of this DC subset at an early phase of infection could lead to viral ingress and further manipulation of these cells. Since our data support the hypothesis that vXCL1 is a superior agonist to host rat XCL1, this would represent a plausible viral strategy to enhance viral dissemination or establish latency by directly infecting XCR1+ DCs. Infection with CMV triggers an extensive CD8+ T cell response based on cross-presentation of viral peptides by DCs (Jackson et al., 2011; Snyder et al., 2010). By expressing vXCL1, the virus might benefit by disturbing DC function to evade adaptive immune responses. Currently, it can only be speculated that XCR1+ rat DCs are equivalent to human and mouse cross-presenting DCs. As XCR1-expressing DCs become increasingly attractive for novel therapeutic approaches against tumors and pathogens (Brewitz et al., 2017; Fossum et al., 2015; Hartung et al., 2015; Li et al., 2017; Steinman and Banchereau, 2007), further studies on how viruses might intervene with these cross-presenting DC subsets are required.
The fact that RCMV vXCL1 activates XCR1+ DCs more effectively than the endogenous rat molecule suggests that modifications in sequence or structure, likely occurring during viral evolution, resulted in this specificity. In designing novel molecules to tackle GPCR signaling pathways, this specificity is of key to avoid cross-reactivity and potential side effects. Therefore, RCMV vXCL1 might be a useful tool to study receptor specificities and these findings could aid in targeting known and possibly also orphan GPCRs.
MATERIALS AND METHODS
Cell lines and animals
All used cell lines (Table S1) were tested for mycoplasma contamination. Spleens of at least 8-week-old male rats from the inbred Lewis strain were used for rat splenocyte preparation. Rats were purchased from Charles River (Sulzfeld, Germany) and kept under specific pathogen-free conditions in the animal facility of the Robert Koch Institute (Berlin, Germany). All animal experiments were performed according to approved guidelines.
Generation of recombinant chemokines
The DNA coding for vXCL1 and rXCL1 was chemically synthesized (Eurofins Genomics) and cloned into the expression vector pRmHa-3 through standard procedures (Wallny et al., 1995). Chemokines containing a C-terminal StrepTag (SAWSHPQFEKGGGSGGGSGGSAWSHPQFEK), separated by a glycine residue, were overexpressed in stable Drosophila SL-3 cells after transfection as described by Hartung et al. (Hartung et al., 2015; Schneider, 1972). vXCL1-StrepTag and rXCL1-StrepTag were purified from the supernatant using Heparin HP columns (GE Healthcare) followed by purification using StrepTrap HP columns (GE Healthcare). After StrepTag purification the buffer was changed to PBS using a PD-10 column (GE Healthcare). Protein concentrations were determined using a Nanodrop Spectrophotometer (Thermo Fisher Scientific). Human XCL1 was purchased from R&D Systems and reconstituted in a buffer containing 0.1% (w/v) bovine serum albumin (BSA) and 1 mM acetic acid.
Antibodies and flow cytometry
Polyclonal rabbit anti-goat (P0449, 1:2500) and goat anti-mouse (P0447, 1:1000) immunoglobulin conjugated to horseradish peroxidase (HRP) were purchased from Dako. Antibody against rXCL1 was obtained from SCBT (sc-19048, 1:1000) and the vXCL1.11 monoclonal antibody (mAb, 1:1000) was generated in the laboratory of Stipan Jonjic (University of Rijeka, Rijeka, Croatia) as described previously (Geyer et al., 2014). The following mAbs were also used: anti-CD45RA (OX-33; 202314, 1:200; 202318, 1:100), anti-CD4 (W3/25; 201516, 1:100) and anti-XCR1 (ZET; 148204, 1:800) (all from BioLegend); anti-MHCII (OX-6; BD Pharmingen; 554928, 1:100); anti-CD3 (REA223; Miltenyi Biotec; 130-102-675, 1:100); and anti-CD103 (OX-62; Cedarlane; CL083AP, 1:50). The latter was coupled to Pacific Blue-NHS (Thermo Fisher Scientific). Strep-Tactin Oyster A645 conjugate (Strep-tag® II) used for detection of recombinant chemokines was purchased from IBA (2-1557-050; 1:300). Titration of antibodies was conducted for optimal signal-to-noise-ratio. Unspecific binding to Fc receptors was blocked by pre-incubation of cells with rat γ-globulin (final concentration 45.2 µg ml−1; Dianova). Standard staining with mAbs was performed in PBS containing 2% (v/v) fetal calf serum (FCS) and 0.1% (w/v) NaN3 for 20 min on ice. For XCR1 receptor expression analysis, CD103-enriched DCs were stained with Pacific Orange (Thermo Fisher Scientific) as fixable live-dead stain and fixed with 2% (w/v) PFA. Data were acquired on a flow cytometer (LSR II and Fortessa, BD) and evaluated using FlowJo software (Tree Star).
Rat DC isolation
Rat spleens from male Lewis (Charles River) were cut and homogenized using a gentleMACS dissociator (Miltenyi Biotec). Splenocytes were digested with 500 µg ml−1 collagenase D (Roche), 20 µg ml−1 DNase I and 2% (v/v) FCS (Pan Biotech) in RPMI 1640 at 37°C for 25 min and shaking at 200 rpm. After addition of 10 mM EDTA and further incubation at 37°C for 5 min, the cell suspension was filtered through a 100 µm nylon sieve (BD Falcon™ Corning) and applied to NycoPrep density gradient (Progen; 1.073 g ml−1) centrifugation at 1700× g for 10 min at 4°C. Enriched leukocytes were recovered from the NycoPrep fraction and cells were used for chemokine binding experiments. For Transwell assays and receptor expression studies rat DCs were enriched using anti-CD103 (OX-62) MicroBeads (Miltenyi Biotec) by magnetic cell sorting according to the manufacturer's instructions.
Chemotaxis assays
A total of 1×105 to 5×105 CD103-enriched cells (purity ∼60%) stained previously with cell marker-specific fluorophore-conjugated antibodies were resuspended in 100 µl chemotaxis medium [1× RPMI 1640, 1% (w/v) BSA (low-endotoxin; Gemini Bio-products), 100 µg ml−1 penicillin-streptomycin, 50 µM β-mercaptoethanol] and applied to the upper chamber of a 24-well Transwell system (6.5-mm diameter, 5-µm pore polycarbonate membrane; Corning Costar). Recombinant chemokines were diluted in chemotaxis medium and were added to the lower chamber of the Transwell. As a control, no chemokine was added. After incubation for 150 min at 37°C and 5% CO2 the Transwell inserts were rinsed carefully to remove any attached cells and cells migrated to the lower chamber were analyzed by flow cytometry. Identification of cells was based on the expression of specific cell markers. T cells (CD3+) and B cells (CD45RA+) were excluded by gating, and XCR1+ DCs (MHCII+ CD103+ CD4−) were examined for migration. Cells were counted over 5 min at constant flow rate and a defined volume to determine absolute cell numbers of migrated and input cells. The percentage of migrated cells was calculated by dividing the number of cells in the lower chamber by the number of input cells (number of migrated cells/number of input cells×100).
XCR1 surface expression and chemokine binding to DCs
NycoPrep-enriched DCs from spleen were incubated with recombinant chemokines in MACS-PBS [1× PBS, 0.5% (w/v) BSA (low-endotoxin; Gemini Bio-products) and 2 mM EDTA] for 20 min on ice. Cells were stained with fluorophore-conjugated antibodies to analyze binding of recombinant rXCL1 and vXCL1. The anti-XCR1 mAb binds the receptor in a non-blocking fashion and therefore does not interfere with chemokine binding. To determine XCR1 expression after prolonged incubation with rXCL1 and vXCL1, CD103-enriched splenic DCs were incubated with recombinant chemokines at 37°C and 5% CO2. After incubation, cells were stained with fluorophore-conjugated antibodies directed against different cell surface markers followed by live-dead staining with PacO prior to fixation with 2% (w/v) PFA. Cells were analyzed for XCR1 expression and ligand binding by flow cytometry.
Inositol phosphate accumulation assay
COS-7 cells were cultured at 37°C and 10% CO2 in DMEM supplemented with 10% (v/v) fetal bovine serum (FBS), 2 mM L-glutamine, 180 units ml−1 penicillin and 45 µg ml−1 streptomycin. One day after seeding, the cells were transiently transfected with receptor cDNA and the chimeric G protein Gqi4myr, which converts a Gαi signal into a Gαq readout (Heydorn et al., 2004; Kostenis et al., 1998). The transfection method used was a chloroquine-based calcium phosphate precipitation method, where DNA was diluted in TE buffer, and CaCl2 was added to a concentration of 0.25 M. This mixture was then slowly mixed with an equal volume of 2× HEPES-buffered saline (HBS) and incubated for 45 min at room temperature. After incubation, the DNA mixture was dripped directly onto the cells, growth medium containing 60 µg ml−1 chloroquine was added, and cells were incubated at 37°C and 10% CO2 for 5 h. The transfection was stopped by aspirating transfection medium and adding growth medium. One day after transfection, the cells were seeded in 96-well plates (3.5×104 cells per well) and incubated with 0.5 µCi of myo-[3H]inositol in 100 µl of growth medium for 24 h. Cells were then washed with Hanks' buffered salt solution (HBSS) supplemented with CaCl2 and MgCl2 (GIBCO) and incubated at 37°C for 15 min in 100 µl of the buffer supplemented with 10 mM LiCl, followed by ligand addition and 90 min of incubation at 37°C. After incubation, the plates were put on ice, ligand medium was removed and cells were lysed with 40 µl 10 mM formic acid for 30 min on ice. 35 µl of the lysed cells was then transferred to 96-well white bottom plates. A solution of poly-L-lysine-coated YSi scintillation proximity assay (SPA) beads (Perkin Elmer) was prepared at 12.5 mg ml−1 in Milli-Q water. 80 µl of this solution was added to each well, gently mixing the solution before each addition to limit sedimentation of the beads. Plates were covered with TopSeal-A PLUS (Perkin Elmer) adhesive seals and shaken at high speed for 30 min. Afterwards, plates were centrifuged at 485 g for 5 min. The generated SPA bead-bound [3H]inositol phosphates were measured on a TopCount NXT (Packard) after 8 h incubation at room temperature. Determinations were made in duplicates.
Real-time internalization assay and receptor expression
Real-time internalization assays were performed as previously published (Foster and Bräuner-Osborne, 2017; Roed et al., 2014). Briefly, HEK293A wild-type or Δβ-arrestin1/2-knockout cells (Alvarez-Curto et al., 2016) transiently expressing SNAP-tagged rat XCR1 were seeded in white 384-well plates the day after transfection at a density of 2×104 cells/well. The following day, the medium was removed and the SNAP-tagged XCR1 was labeled with 100 nM Tag-lite SNAP-Lumi4-Tb (donor) in OptiMEM for 60 min at 4°C and 37°C. Subsequently, the cells were washed with HBBS buffer supplemented with 1 mM CaCl2, 1 mM MgCl2, 20 mM HEPES (internalization buffer, pH 7.4) and 100 μM preheated fluorescein (acceptor) was added. Constitutive receptor internalization was measured every 6 min at 37°C in PerkinElmer EnVision 2014 multilabel reader. Receptor expression was measured in the wells that were not loaded with the acceptor.
SDS-PAGE, immunoblotting and mass spectrometry analysis
Recombinant chemokines were boiled with 4×SDS loading buffer [0.3 M Tris, 12% (w/v) SDS, 40% (v/v) glycerol, 0.3 M dithiothreitol (DTT), 0.02% (w/v) Bromophenol Blue] and separated electrophoretically by size on a 15% Tris-Tricine SDS gel along with the PageRuler prestained protein ladder (Thermo Fisher Scientific). The gel was either stained with Coomassie staining solution or blotted onto a PVDF membrane (Roche) with 20 V for 1 h and stained with polyclonal anti-rat XCL1 or monoclonal vXCL1.11 antibody as given above. Detection was conducted by using ECL solution (Thermo Fisher Scientific). To verify purified chemokines by mass spectrometry, peptides were obtained from the gel by trypsin digest. Mass spectrometric analysis was performed by the Proteomics Service of the Robert Koch Institute.
Data and statistical analysis
For statistical analysis, one-way ANOVA with Dunnett's Multiple Comparison tests were performed using GraphPad 8 Software (SanDiego, CA) with P≤0.05 considered as significant. All data are shown as mean±s.e.m.
Acknowledgements
The authors would like to thank Maibritt Sigvardt Baggesen and Søren Petersen for excellent technical help.
Footnotes
Author contributions
Conceptualization: A.B., J.M., S.V.; Methodology: A.B., J.M., C.B., V.D.; Validation: A.B., J.M.; Investigation: A.B., J.M., C.B., V.D.; Data curation: S.V.; Writing - original draft: A.B., S.V.; Writing - review & editing: A.B., J.M., C.B., V.D., S.G., H.W.M., R.A.K., M.M.R., S.V.; Supervision: S.V.; Project administration: S.V.; Funding acquisition: S.V.
Funding
S.V. is supported by the Deutsche Forschungsgemeinschaft (VO 774/7-2). A.B. was supported by the Georg and Agnes Blumenthal-Stiftung. V.D. is supported by a grant from the Lundbeck Foundation (Lundbeckfonden). C.B. is partially supported by Carl og Ellen Hertz Legat, Christian Larsen and Dommer Ellen Larsens Legat, and Dagmar Marshall's Fond. M.M.R. is supported by the Novo Nordisk Foundation, The Carlsberg Foundation (Carlsbergfondet) and the Lundbeck Foundation (Lundbeckfonden).
References
Competing interests
The authors declare no competing or financial interests.