X

ZTtH pUL36

UL36x1 UL37x2

UL36x2 UL37x3 UL38 UL37x1

pUL37x1 pUL38

gpUL37 PUL37m

Fig. 1 A map of the HCMV UL36-UL38 cell death suppression locus indicating the relative positions of open reading frames (ORFs) and major transcripts of the region. Rectangles represent the ORFs and include an arrowhead to denote the direction of transcription. Arrows represent the 3' nontranslated regions. A raised line connecting ORFs indicates splicing. Splicing events producing minor transcripts of the UL37 gene (Adair et al. 2003) are not shown pUL37x1 pUL38

gpUL37 PUL37m

Fig. 1 A map of the HCMV UL36-UL38 cell death suppression locus indicating the relative positions of open reading frames (ORFs) and major transcripts of the region. Rectangles represent the ORFs and include an arrowhead to denote the direction of transcription. Arrows represent the 3' nontranslated regions. A raised line connecting ORFs indicates splicing. Splicing events producing minor transcripts of the UL37 gene (Adair et al. 2003) are not shown factors similar to Bcl-2 or Bcl-xL (Goldmacher et al. 1999). To date, vMIA is the most broadly antiapoptotic CMV protein known and analogous to the cellular Bcl-2 proteins, is highly effective against a myriad of stimuli including intrinsic stresses as well as extrinsic, immune-regulated signals (Goldmacher et al. 1999; Belzacq et al. 2001; Vieira et al. 2001; Jan et al. 2002; Roumier et al. 2002; Boya et al. 2003; Andreau et al. 2004; Arnoult et al. 2004; Boya et al. 2005; McCormick et al. 2005). However, vMIA does not encode any BH-domains that characterize the cellular proteins (Goldmacher et al. 1999).

vMIA function requires an amino terminal mitochondrial-targeting domain (aa 2-34) and a carboxyl-terminal antiapoptotic domain (AAD, aa 118-147) (Hayajneh et al. 2001) that together are sufficient for function. The mitochondrial-targeting domain includes an amino-terminal hydrophobic signal followed by highly conserved basic residues, and both are required for mitochondrial trafficking (Mavinakere and Colberg-Poley 2004). Evidence suggests a mitochondrial membrane association with the targeting domain spanning the membrane and the AAD exposed to the cytoplasm (Mavinakere et al. 2006). The carboxyl-terminal AAD includes a predicted amphipathic a-helix motif (aa 126-140) critical to function (Smith and Mocarski 2005). Point mutations predicted to disrupt an a-helical structure alter amphipathicity or place charge on the hydrophobic face of the AAD a-helix, each completely abrogate vMIA function. In contrast, the hydrophilic face of the AAD a-helix tolerates significant substitutions with as many as five or six amino acid substitutions required to disrupt function (Smith and Mocarski 2005).

The growth arrest and DNA damage 45 alpha (GADD45a) protein interacts directly with vMIA in yeast and mammalian cells, fails to bind vMIA mutant proteins, and is essential for vMIA-mediated antiapoptotic activity (Smith and Mocarski 2005). Targeted knockdown of GADD45a, GADD45ß, and GADD45y reduced vMIA activity, and each GADD45 family protein individually enhanced vMIA activity. GADD45a increased both the overall amount of vMIA and that associated with mitochondrial fractions. Thus, the DNA damage response pathway is directly linked to vMIA-mediated cell death suppression. Further, vMIA was shown to bind the antiapoptotic Bcl-2 family protein Bcl-xL in mammalian cells. Collectively, these data suggest that vMIA acts together with Bcl-xL and GADD45 to regulate the mitochondrial release of proapoptotic factors (Fig. 2).

In addition to GADD45 proteins, vMIA also binds the proapoptotic Bcl-2 family protein Bax (Arnoult et al. 2004), which has more recently been connected to mitochondrial morphogenesis during life (Karbowski et al. 2006). In most instances, Bax is distributed in the cytoplasm, but Bax oligomerization and relocalization to mitochondria mediates the release of proapoptotic factors from the organelle (Antonsson et al. 2001). In the presence of vMIA, however, oligomerized Bax at mitochondria fails to promote apoptosis, suggesting sequestration as a component of the antiapoptotic mechanism (Arnoult et al. 2004; Poncet et al. 2004). Thus, the vMIA-dependent antiapoptotic mechanism is distinct from that of cellular and viral Bcl-2 proteins that prevent Bax relo-calization and oligomerization at mitochondria. Recruitment and sequestration

Fig. 2 A representation of the apoptosis pathway and CMV-mediated alterations preventing death, as described in the text. Dashed arrows indicate events prevented by the viral proteins, vICA or vMIA, as indicated. The solid arrow indicates vMIA-dependent relocalization of Bax to mitochondria. At mitochondria, a complex(s) of proteins including Bax as an oligomer, vMIA, GADD45, and Bcl-xL prevents the release of mitochondrial protein cytochrome c. vICA binds procaspase-8 and is depicted as a complex that prevents procaspase-8 activation following addition of extrinsic death signals. The mechanisms and/or direct physical interactions that promote survival in the presence of the remaining viral proteins, pUL38, IE1, IE2579aa, M45, and m41 are unclear, and these are placed according to the anticipated site of localization within the cell. For simplicity, many important regulatory components have not been included

Fig. 2 A representation of the apoptosis pathway and CMV-mediated alterations preventing death, as described in the text. Dashed arrows indicate events prevented by the viral proteins, vICA or vMIA, as indicated. The solid arrow indicates vMIA-dependent relocalization of Bax to mitochondria. At mitochondria, a complex(s) of proteins including Bax as an oligomer, vMIA, GADD45, and Bcl-xL prevents the release of mitochondrial protein cytochrome c. vICA binds procaspase-8 and is depicted as a complex that prevents procaspase-8 activation following addition of extrinsic death signals. The mechanisms and/or direct physical interactions that promote survival in the presence of the remaining viral proteins, pUL38, IE1, IE2579aa, M45, and m41 are unclear, and these are placed according to the anticipated site of localization within the cell. For simplicity, many important regulatory components have not been included of Bax at mitochondria has also been suggested as the mechanism (Karbowski et al. 2006) for vMIA-dependent disruption of reticular mitochondrial networks (McCormick et al. 2003b); however, more recent evidence of vMIA mutants that disrupt networks but fail to bind Bax (Pauleau et al. 2007) suggests other factors may also be important.

vMIA prevents apoptosis during infection; however, vMIA is not required for replication, and replication in the absence of vMIA does not induce caspase-dependent apoptosis (McCormick et al. 2005). A vMIA deletion mutant made in the laboratory-propagated strain, TownevarATCC, produces yields nearly equivalent to parental virus. In contrast, vMIA is more critical for efficient replication of the laboratory strain AD169varATCC (Reboredo et al. 2004). These strain-dependent variations may suggest that the quantity or quality of intrinsic stresses produced by individual strains varies because evidence suggests vMIA is an important regulator of the viral response to stress (McCormick et al. 2005). The phenotype produced by disruption of vMIA in AD169varATCC is highly variable (Brune et al. 2003; Yu et al. 2003; Sharon-Friling et al. 2006), perhaps due to other factors that prevent vMIA-dependent release of calcium from the endoplasmic reticulum (Sharon-Friling et al. 2006) or increase ATP levels (Poncet et al. 2006), and further analyses are needed to resolve the role of vMIA in that strain.

Although of limited impact on replication in HFs, the TownevarATCC mutant revealed a role for vMIA in regulating caspase-independent death. Caspase-3 activation underlies caspase-dependent apoptosis; however, this protease is not required for other cell death pathways that are considered to be apoptosis-like (Leist and Jaattela 2001; Lockshin and Zakeri 2002; Jaattela 2004). Thus, UL37x1 deletion can promote CMV-induced caspase-3-dependent cell death in the case of AD169varATCC (Reboredo et al. 2004), or a caspase-3-independent cell death in the case of TownevarATCC (McCormick et al. 2005), and vMIA regulates both forms of death during infection (McCormick et al. 2005). From studies so far, the context where caspase-3-independent cell death is a significant obstacle to the virus is unknown.

Chimpanzee CMV, rhesus macaque CMV, and African green monkey CMV each retain a vMIA homolog that could be identified through computer analysis (McCormick et al. 2003a). Each of these proteins share sequence similarity with the mitochondrial-targeting and AAD domains of vMIA. Rhesus macaque CMV vMIA retains similarity only to the amino- and carboxyl-terminal domains of human CMV (HCMV) vMIA and functions as an antiapoptotic protein. It is expected that all primate CMVs encode functional homologs. In contrast, the identification of rodent CMV functional homologs encoded by ORFs, m38.5 and r38.5, required more extensive analyses due to limited sequence homology (McCormick et al. 2003a, 2005; Brocchieri et al. 2005). Initial searches for murine CMV (MCMV) mitochondrial localized proteins with vMIA function were executed in HeLa cells utilizing methods that revealed vMIA (Goldmacher et al. 1999; McCormick et al. 2003a). Increasing the repertoire of stimuli revealed that m38.5 prevents proteasome inhibitor-induced, intrinsic apoptosis but not extrinsic, Fas-mediated apoptosis in HeLa cells (McCormick et al. 2005) or a telomerase-immortalized retinal epithelial cell line of human origin (Jurak and Brune 2006). Thus, MCMV m38.5 encodes an antiapoptotic protein that localizes to mitochondria (McCormick et al. 2005). The rodent CMV ORFs map to positions on the viral genomes analogous to UL37x1 (McCormick et al. 2003a; Brocchieri et al. 2005), indicating that rodent and primate CMVs each encode vMIA and vICA homologs.

Limited sequence similarity and differences in protective function in human cells suggest the human and rodent vMIA homologs retain elements that are specific to function in the appropriate host (McCormick et al. 2005). Identification of additional MCMV proteins localized to mitochondria may also suggest the potential for synergism or even replacement of m38.5 function in specific cells (Tang et al. 2006). Interestingly, vMIA apparently protects from specific apoptotic stimuli in a species-dependent fashion as well. Thus, vMIA fails to prevent mitochondrial damage induced by staurosporine in wild type murine fibroblasts, apparently due to the role of murine Bak in that setting (Arnoult et al. 2004) but prevents staurosporine-induced death in HeLa cells (Andreau et al. 2004). Thus, properties of the antiapoptotic proteins encoded by these viruses reflect the evolutionary divergence of the host. In fact, one aspect of the species barrier that restricts CMVs is reportedly due to functions that for MCMV can be provided by vMIA (Jurak and Brune 2006). Given the genomic organization and studies thus far, it is likely that m38.5 will retain vMIA functions relevant to survival in the host and that all CMVs rely on vMIA function.

Was this article helpful?

0 0

Post a comment