The Cytoplasmic C-Tail of the Mouse Cytomegalovirus 7 Transmembrane Receptor Homologue, M78, Regulates Endocytosis of the Receptor and Modulates Virus Replication in Different Cell Types

Virus homologues of seven-transmembrane receptors (7TMR) are encoded by all beta- and gammaherpesviruses, suggesting important functional roles. M78 of mouse cytomegalovirus (MCMV) is representative of a family of 7TMR conserved in all betaherpesviruses. M78 family members have been found to exhibit cell-type specific effects upon virus replication in tissue culture and to affect virus pathogenesis in vivo. We reported previously that M78, for which no ligands are known, undergoes rapid, constitutive endocytosis. In this study, we have investigated the role of the M78 cytoplasmic C-tail in mediating endocytosis and consequences of C-tail deletion upon replication and pathogenesis. Mutations of M78 (C-tail truncations or point mutations) and CCR5-M78 chimeras identified two distinct regions affecting endocytosis. The first was a classical acidic di-leucine motif (DDxxxLL), located close to the C-terminus. The second region, the activity of which was suppressed by downstream sequences, included the putative 8th helix, located close to the 7th transmembrane domain. A recombinant MCMV expressing an endocytosis-deficient M78, lacking most of the C-tail (M78_CΔ155), had a cell-type specific replication phenotype. M78_CΔ155 had restricted replication in bone marrow macrophages, indistinguishable from an M78-null recombinant. In contrast, M78_CΔ155 replicated normally or with enhanced titres to wild type virus in other tested cell-types, whereas M78-null was attenuated. Distinct phenotypes for M78_CΔ155 and M78-null suggest that the C-tail deletion resulted in M78 dysfunction, rather than complete loss of function; furthermore, they highlight a cell-type specific role of M78 during replication. Infection of mice (intranasal) demonstrated that M78_CΔ155, similar to M78-null, was cleared more rapidly from the lungs than wild type virus and was severely attenuated for replication in salivary glands. It may be speculated that attenuation of both M78_CΔ155 and M78-null for replication in macrophages may have contributed to their similar pathogenic phenotypes.

Introduction required, mediated in part by GRKs, but endocytosis was not dependent on β-arrestins [27][28][29]. US28 was found to associate with lipid rafts and the C-tail was palmitoylated, but mutation of targets of palmitoylation did not block endocytosis; however, a mutation of a C-terminus proximal acidic dileucine motif ( 301 ElhcLL 306 : LL-AA) caused a significant decrease in the endocytosis rate determined via agonist internalisation [27]. A similar motif in the US27 C-tail ( 357 EeeeLL 362 ) has been suggested to be required for efficient US27 endocytosis [15]. There is evidence that M78 may be internalised via clathrin mediated endocytosis and via lipid rafts/caveolae in transfected cells [16]. However, motifs that drive endocytosis of M78 are yet to be determined.
The hypotheses of this study were that constitutive endocytosis of M78 is mediated via the cytoplasmic C-tail and that endocytosis is essential for M78 function. We have used truncated and chimeric M78 receptor constructs to demonstrate that the M78 C-tail directs rapid, constitutive endocytosis. Two regions of the C-tail are able to induce endocytosis: a region close to the C-terminus which includes an acidic di-leucine motif ( 458 DDvsaLL 464 ) and a second region overlapping the putative 8 th helix (including aa 333-347), the activity of which was unmasked in truncated C-tail constructs. A recombinant MCMV was constructed which expressed an endocytosis deficient M78 mutant lacking most of the C-tail. This recombinant was attenuated for replication similarly to M78 null virus in primary macrophages, but replicated with normal or enhanced titres compared with wild type virus in other cell types tested. In mice infected intranasally, the C-tail deletion mutant displayed a similar phenotype to M78 null virus; whereas replication early post-infection was similar to wild type, the C-tail deletion mutant was cleared more rapidly from the lungs and was severely attenuated for replication in salivary glands.

Ethics statement
Animal experiments were approved by the University of Queensland Animal Ethics Committee, in accordance with the Australian Animal Care and Protection Act (2001) and the Australian Code for the Care and Use of Animals for Scientific Purposes.

PCR mutagenesis and plasmid construction
An Expand Long template PCR kit (Roche) was used for amplification of truncated/mutated M78 amplicons, which were then cartridge purified (High Pure PCR product purification kit, Roche) and digested. The oligonucleotides used for mutagenesis (SIGMA) and cloning strategies are detailed in S1 Table. Restriction digests used enzymes from New England Biolabs (NEB); plasmid vector DNA was treated with calf intestinal phosphatase (NEB) to prevent selfligation. Digested products were run on agarose gels and extracted (cartridge purified) prior to ligation (T4 DNA ligase, NEB) and transformation (DH5α C2987, NEB). Transformants were screened by restriction digest and confirmed by sequencing (Australian Genome Research Facility), prior to preparation of DNA from selected clones for transfection (Nucleobond Xtra Midi, Machery Nagel).

Image processing and analysis
Endocytosis was quantified using immunofluorescence images from several random fields (NIKON E600, x40 objective), adjusted to minimise background and analysed in ImageJ by selecting regions bounding individual cells to measure signal intensity of the 594 (before permeabilisation) and 488 (after permeabilisation) channels. The endocytosis index was determined as the ratio of signal detected 488/594.

Mouse infections
Female BALB/c mice (Animal Resources Centre, Western Australia) were maintained at the University of Queensland Herston Medical Research Centre and used when 6 weeks old. Anaesthesia was achieved by isoflurane inhalation. Viruses were administered intranasally (5 x 10 6 PFU in 30 μl) to anaesthetized mice. Mice were euthanased by exposure to a rising concentration of CO 2 . Organs were dissected, chilled at 4°C, homogenized, and aliquots were stored at -80°C prior to quantification of infectious virus by plaque assay on MEF.

Statistical analysis
Data were analysed using Graph Pad Prism 6. Endocytosis indices were log 10 transformed and statistical significance determined using the Kruskal-Wallis H-test with Dunn's post-test. Virus titres (log 10 ) were analysed using 2-way ANOVA with Bonferroni post-test.

Results and Discussion
Identification of regions of the C-tail that direct M78 endocytosis A plasmid for expression of wild type M78 with an N-terminal HA epitope tag, described previously [16], was used as the template for truncation and substitution mutants of the cytoplasmic C-tail, generated via PCR using the primers listed in S1 Table. Endocytosis was studied in transfected cells using antibody feeding, followed by sequential staining via immunofluorescence of cells pre-and post-permeabilisation, using secondary antibodies conjugated to either AF 594 (pre-permeabilisation-detection of surface retained M78 only) or AF 488 (post-permeabilisation-detection of internalised M78 and some surface retained M78) C-terminal truncation mutants. Preliminary studies demonstrated that the M78 C-tail was required for efficient endocytosis, since a mutant deleted of 155 amino acid residues from the C-tail (M78_CΔ155: aa 1-316) retained substantial surface staining, compared with full length M78 (wtM78: aa 1-471), where surface staining was mostly absent (Fig 1 and  S1 Fig). However, internalisation of M78_CΔ155 was not completely blocked, since dual staining demonstrated a minor proportion of M78_CΔ155 that was internalised in some cells (Fig 1).
A panel of C-tail truncation mutants was analysed to identify regions of the M78 C-tail modulating endocytosis, using transient transfection of HeLa cells. Western blotting confirmed expression of each of the constructs (Fig 2). Quantification of endocytosis was achieved by antibody feeding and sequential immunofluorescence staining to determine an endocytosis index, expressed as the ratio of signal for the two different secondary antibody fluorescence channels AF 488 /AF 594 , corresponding to detection of internalised (plus some surface)/surface M78. A ratio of approx. 1 (0 on log 10 scale) indicates cell-surface retention whereas higher ratios indicate endocytosis. Results (Fig 3A) suggested that two different regions of the C-tail were capable of inducing endocytosis. The putative regulatory regions and effects of M78 mutation upon endocytosis are depicted schematically in Fig 3C. A region close to the C-terminus (aa 446-465) was suggested as a positive regulator of endocytosis, since M78_CΔ6 (aa 1-465) behaved similarly to wtM78, whereas M78_CΔ26 (aa 1-445) had substantially reduced endocytosis (P<0.001). A second positive regulator was suggested (aa 333-347), potentially within the putative 8 th membrane-associated helix, since M78_CΔ139 (aa 1-332) was deficient for endocytosis (P<0.001) whereas M78_CΔ124 (aa 1-347) was efficiently endocytosed. These results also suggested that a region downstream of aa 385 may inhibit endocytosis induced by the upstream element. This was probed by an additional set of truncation mutants (CΔ80, CΔ70, CΔ60), with results shown in Fig 3B. A region between aa 392-425 was suggested a negative regulator, since whereas M78_CΔ80 (aa 1-391) had an endocytosis index similar to wtM78, . This suggests either that the region between aa 392-425 includes motifs that have a specific inhibitory effect upon the upstream endocytosis signal, or alternatively that there is non-specific inhibition related to the length of the C-tail.  Hela cells were transfected with plasmids expressing HA-tagged M78, either wild type (wt), C-terminally truncated (CΔ6-CΔ155) or with mutations of the acidic cluster (AC) or the acidic dileucine motif (DD-AA or LL-AA). One day post-transfection, cells were analysed for endocytosis via antibody feeding (rabbit anti-HA, 1hr), then fixed and processed for immunofluorescence either before (AF 594 conjugated anti-rabbit) or after (AF 488 conjugated anti-rabbit) permeabilisation. Panels A and B show results for two independent experiments. Coverslips were mounted and images of random fields captured using the 594 and 488 fluorescence filters. Images were analysed using ImageJ software to determine the fluorescence intensity of both the 594 and 488 channels for each transfected cell (>50 cells analysed for each group). The endocytosis index (EI) is plotted as the ratio of intensities (488/594), expressed as Mutation of potential sorting motifs. Two motifs apparent in the C-tail that may potentially influence endocytosis were mutated (results shown Fig 3A & 3C). A known endocytic signal, namely an acidic di-leucine motif ( 458 DDvsaLL 464 ), was disrupted by alanine substitution of either the acidic residues (DD-AA) or the leucines (LL-AA). Both mutations resulted in significant inhibition of endocytosis (P<0.001). Clusters of acidic residues, with accompanying serine residues, have been identified as modulators of endocytosis/traffickingin some proteins, including the protease furin, the neurotransmitter transporter VMAT-2 and the human CMV envelope glycoprotein B [33][34][35]. An acidic cluster apparent in the M78 C-tail (AC: 386 DEDDDD 391 ) was modified by multiple substitutions (VQNAAA) with no apparent effect on endocytosis efficiency. Taken together, the truncation and motif mutant results suggested that the di-leucine motif close to the C-terminus was the dominant signal for endocytosis, with a secondary 'cryptic' signal including the region between aa 333-347, which is masked in the presence of a downstream region between aa 392-425.

The M78 C-tail has motifs that can direct constitutive endocytosis of an unrelated 7TMR
To determine whether the M78 C-tail is sufficient to induce endocytosis of an unrelated 7TMR and to further delineate positive and negative regulatory regions, we constructed a series of CCR5/M78 C-tail chimeras, using an approach similar to those reported previously for human CMV US28 and US27 [15,26]. The full length CCR5 coding sequence (HA-CCR5: aa 1-352), or a truncated CCR5 lacking the cytoplasmic C-tail and with a XhoI site incorporated for cloning purposes (HA-CCR5trunc: aa 1-305), were amplified by PCR (S1 Table) and introduced into an expression vector downstream of a self-cleaving signal peptide and HA-tag. CCR5/M78 chimeras were generated by insertion of M78 C-tail regions from selected mutant constructs (digested with SalI sites present within the M78 coding sequence and downstream of the stop codon) at the XhoI site of HA-CCR5trunc. The various constructs were then analysed for endocytosis as described previously.
As expected, HA-CCR5 and HA-CCR5trunc were retained mostly at the cell surface ( Fig  4A). Addition of the full length M78 C-tail (aa 309-471: HA-CCR5/M78) resulted in efficient endocytosis, similar to HA-M78, demonstrating that the M78 C-tail is sufficient to direct constitutive endocytosis (Fig 4A). The putative regulatory regions and effects of mutation upon endocytosis are depicted schematically in Fig 4B. The full length C-tail with disruption of the di-leucine motif (HA-CCR5/M78[LL-AA]) was deficient for endocytosis, confirming this motif as the dominant signal for endocytosis ( Fig 4A). However, consistent with the data from the M78 constructs, a second region of the M78 C-tail may direct endocytosis in truncation mutants lacking the di-leucine motif. Thus, HA-CCR5/M78[CΔ124] (aa 309-347) was endocytosed with an endocytosis index > 2 (log 10 ), whereas [CΔ139] (aa 309-332) was retained at the cell surface (Fig 4A & 4B). These data support the hypothesis that the region between aa 333-347 includes a secondary 'cryptic' signal for endocytosis. It is tempting to speculate that the putative 8 th helix of M78 (aa 322-351, Fig 2) is a major component of the secondary endocytosis signal. In support for this proposal, the putative 8 th helix of D6 was implicated in regulation log 10 values. Box and whiskers plots of the endocytosis index are shown, with bars indicating the group median and whiskers the 5 th and 95 th percentiles. Asterisks indicate statistical significance (Kruskal-Wallis, with Dunn's post-test) comparing mutants with wild type M78 (***P<0.001). The various constructs and effects upon endocytosis are depicted schematically in Panel C. Solid shading indicates a median EI >2, hatched shading indicates 2>EI>1, and stippled shading EI<1.
doi:10.1371/journal.pone.0165066.g003 The M78 C-tail is not required for efficient replication in most cell types To determine whether constitutive endocytosis of M78 is functionally important for virus replication, we constructed a recombinant MCMV expressing M78_CΔ155. This was generated by first disrupting M78 by insertion of a lacZ selectable marker (recombinant M78 null), then using M78 null as the parent for generation of recombinant M78_CΔ155, whereby the M78_CΔ155 sequence replaced the M78/lacZ sequence by homologous recombination, using methods similar to those employed for M33 mutation [30] (further details of plasmid constructs for generation of recombinants is given in S1 Table  Preliminary studies of transiently transfected cells confirmed that the CΔ155 truncation inhibited endocytosis compared to wild type M78 in MEF, SVEC and NMuMG cells (S2 Fig); the effect in BMM could not be assessed due to inefficient transfection. Previous studies of an M78 null mutant (Smith strain, EGFP insertion) reported a moderate replication defect ( 10-fold) following low multiplicity infection in fibroblasts (10.1 -embryonic fibroblast derived cell line) and a more pronounced attenuation (approx. 50 fold) in macrophages (IC21 -peritoneal macrophage-derived cell line) [9]. The results for the M78 null virus of this study are similar, with marked attenuation in BMM (10-80 fold reduced titre; p<0.001 from 48-168 hours post-infection) and lesser attenuation in the other cell types (SVEC, up to 3 fold; MEF, up to 4 fold; NMuMG, up to 7 fold). The M78_CΔ155 mutation attenuated replication in BMM (5-40 fold reduced titre; p<0.001 from 48-168 hours post-infection), with replication kinetics similar to that of M78 null. In contrast, M78_CΔ155 replicated at least as well as wild type MCMV in the other cell types tested, with moderately higher titres in SVEC (3-4 fold; p<0.01 at 120-168 hours post infection) and NMuMG (3-5 fold; p<0.05 at 96, 144 and 168 hours post infection). The above results suggest that the M78_CΔ155 mutation resulted in dysregulation of M78 function, with phenotypic effects distinct from knock-out of M78.

The M78 C-tail is required for replication in salivary glands following intranasal infection
Replication of the M78 null and CΔ155 recombinants was compared with wt MCMV in vivo, following intranasal inoculation (Fig 6). In the lungs, the recombinants replicated to similar titres as wt MCMV initially (day 3 p.i.), but were cleared more rapidly (approximately 20 -to 40-fold lower titres than wt MCMV by 11 days post-infection (P<0.0001)). In salivary glands, both recombinants were severely attenuated (over 1000-fold lower titres than wt MCMV at days 11 and 18 post-infection (P<0.0001)). These results are consistent with deletion of the M78 C-tail, with concomitant disruption of M78 endocytosis, resulting in dysregulation of M78 function in cell types relevant to clearance from the lung and replication in salivary glands. Given the attenuation of the C-tail deletion mutant in macrophages, but not other cell types tested in vitro, we speculate that a role of M78 during infection of macrophages may contribute to the observed in vivo phenotypes.

Conclusion
Consistent with the hypothesis that the cytoplasmic C-tail of M78 directs endocytosis, two regions were identified that induced rapid endocytosis of M78, namely an acidic di-leucine ( 458 DDxxxLL 464 ) and another cryptic signal (aa 333-347), which lies within the putative 8 th helix of M78, but has no obvious homology to known endocytic motifs. Studies of CCR5/M78 chimeras confirmed the activity of these C-tail regions was not dependent on properties of the body of the receptor. Furthermore, the chimera studies showed activity in the absence of CCR5 ligands, supporting the hypothesis that M78 endocytosis is constitutive, rather than mediated by a ubiquitous (as yet unidentified) ligand. Contrary to the hypothesis that endocytosis is essential for M78 function, an MCMV expressing an endocytosis deficient M78 mutant (M78_CΔ155), lacking most of the C-tail, was not equivalent to M78_null. With the exception of bone marrow macrophages (BMM), M78_CΔ155 had normal or enhanced replication in a variety of cell types, whereas M78_null was attenuated. The dysfunction of M78_CΔ155 resulted in a profound attenuation in vivo following intranasal infection, similar to that of M78_null. Given the known contribution of monocyte/macrophage infection to the pathogenesis of MCMV [37], it is tempting to speculate that defective macrophage replication contributed to the phenotypes observed for M78_CΔ155 and M78_null. Further studies are warranted to define the functions of M78 and related proteins of other betaherpesviruses,to explore their potential as an anti-viral therapeutic target.