Transformation-related proteins of H. saimiri
Based on the transformation potential in vivo and in vitro and on the sequence variability, H. saimiri strains were classified into the three subgroups, A, B, and C (Desrosiers and Falk, 1982; Medveczky et al., 1984, 1989). The oncogenic potential in New World primates is highest for viruses of subgroup C and lowest for viruses of subgroup B (reviewed in Broker and Fickenscher, 1999; Fickenscher and Fleckenstein, 2001; Ensser et al., 2003). In agreement with these findings, only viruses of subgroup C are capable of transforming human and rabbit T lymphocytes and of inducing lymphoproliferative disease in Old World primates (Medveczky et al., 1989; Biesinger et al., 1992; Alexander et al., 1997; Knappe et al., 2000). The major difference between the genome of various H. saimiri strains lies in the leftmost part of L-DNA, which contains a single gene in the strains of subgroups A and B (stpA and stpB, respectively), or two genes in the strains of subgroup C (stpC and tip) (Medveczky et al., 1984; Murthy et al., 1989; Biesinger et al., 1990; Albrecht et al., 1992; Ensser et al., 2003). Both tip and stpC are essential for T-cell transformation by H. saimiri of subgroup C (Duboise et al., 1998b).
The structure of StpA features an acidic domain, a hydrophobic transmembrane domain and several collagen-like repeats (GX1X2, where X1 and/or X2 is P) at an overall length of ∼170 amino acid residues, which varies in different strains (Fig. 1). The length and structure of StpB are similar to those of StpA, although sequence homology between StpA and StpB is very limited. The structure of StpC features the same components, but collagen-like repeats of StpC exist as a cluster, unlike those of StpA and StpB, which are dispersed. Sequence similarity between StpC and other Stp proteins is very low. Finally, StpC is the smallest protein of this group; its length is ∼100 amino acid residues. The size of Tip ranges from ∼210 to ∼260 amino acid residues. It has an acidic domain, one or two serine-rich regions (in some strains, Tip features duplication of a large part of its sequence, which causes the presence of two serine-rich regions and increase in the overall size), and a membrane-spanning hydrophobic domain. Tip also contains a homology with a C-terminal region of Src-family protein tyrosine kinases (PTKs) and an SH3-binding site, whose functional significance will be discussed below.
Figure 1. Structure of Herpesvirus saimiri proteins StpA, StpB, StpC, and Tip. Major structural elements and size of these proteins are shown.
Download figure to PowerPoint
Transformation potential of Stp proteins
The essential role of StpC in H. saimiri-induced T-cell transformation is consistent with the transformation potential of StpA and StpC in fibroblasts (Jung and Desrosiers, 1991; Jung et al., 1991). Noteworthy, the transformation potential of StpC in fibroblasts exceeds that of StpA (Jung et al., 1991), whereas StpB does not demonstrate any transforming ability (Choi et al., 2000), thus mirroring the relative degrees of oncogenicity for different subgroups of H. saimiri. The oncogenic potential of StpC was further demonstrated by the finding that transgenic mice expressing StpC develop multiple tumors (Murphy et al., 1994). In agreement with the finding that Tip-deficient H. saimiri of subgroup C is incapable of transforming T lymphocytes (Duboise et al., 1998b), tumors found in StpC-transgenic mice were exclusively epithelial; no lymphoid malignancies were observed in these animals (Murphy et al., 1994). Consistent with these results, lentiviral transduction of StpC cDNA to human primary T cells resulted in stable expression of StpC in these cells, but did not cause their transformation (Hasham and Tsygankov, 2004).
Effects of Stp proteins on cellular signaling
The role of StpC in H. saimiri-induced T-cell transformation and the oncogenic potential of StpC in fibroblasts are likely to be mediated by its interference with major cell signaling pathways. First, it has been shown that StpC binds to Ras and that the activities of Ras and the serine-protein kinase Erk are elevated in StpC-transformed fibroblasts (Jung and Desrosiers, 1995) (Fig. 2). Furthermore, introduction of c-ras and, especially, v-ras into the genome of stpC-deficient H. saimiri rescue the transformation potential of the virus (Guo et al., 1998). Although these findings suggest that StpC may activate Ras-mediated signaling and that activation of Ras may substitute for StpC in cell transformation, they do not provide direct evidence that the oncogenic potential of StpC is mediated by its interactions with Ras.
Figure 2. Effects of StpC and Tip on the mechanisms of cellular regulation. Functional interactions between Tip and StpC and their targets are shown by arrows. Proteins physically associated with Tip or StpC are shown as ovals. Elements of signaling pathways (circled) are indicated to the extent the effect of H. saimiri proteins on these pathways is characterized. Merging lines indicate cooperation of Tip and its target in inducing a biological effect. Activation and inhibition is shown as (+) or (−), where known. Alternative effects are shown, if supported by experimental results. Dashed lines show possible links that have not been established yet.
Download figure to PowerPoint
In contrast, most data available now argue that the transforming effect of StpC is related to its ability to activate NF-κB. First, StpA and StpC were shown to physically interact with TRAF proteins, which are known to play a crucial role in triggering signaling pathways that result in the activation of NF-κB, when StpA or StpC is co-overexpressed in epithelial cells with one of the TRAF proteins (Lee et al., 1999). Interestingly, whereas both StpA and StpC were associated with TRAF under these conditions, only StpC elevated the activity of NF-κB in these experiments (Lee et al., 1999). Furthermore, the effect of StpC on NF-κB was suppressed by a dominant-negative form of TRAF2. Finally, mutational analysis of the TRAF-binding domain of StpC conducted in this study suggested that binding of StpC to TRAF was essential for the positive effect of StpC on NF-κB activity. A related study indicated that StpB, which lacks a transformation potential, could be converted into an actively transforming protein by the introduction of the region of StpC containing 18 collagen-like repeats (Choi et al., 2000). This reconstitution of transformation potential correlated with an increase in multimerization of the StpB proteins “enriched” with collagen repeats and the activation of NF-κB in StpB-overexpressing epithelial cells (Choi et al., 2000), thus arguing that multimerization of Stp proteins is crucial for their transformation potential and their ability to activate NF-κB. A dramatic effect of StpC on the transcriptional activity and DNA binding of NF-κB was also shown in T-lymphoblastoid cells expressing StpC in a stable fashion (Merlo and Tsygankov, 2001; Sorokina et al., 2004).
StpC was shown to induce activation of I-κB kinase (IKK) and subsequent phosphorylation and degradation of I-κB in both T-lymphoid and non-lymphoid cells (Sorokina et al., 2004). Furthermore, dominant-negative forms of TRAF, NF-κB-inducing kinase (NIK) (a kinase located upstream of IKK in the pathway leading to NF-κB activation) or I-κB blocked the effect of StpC on NF-κB activity in both T-lymphoid and non-lymphoid cells (Sorokina et al., 2004). Therefore, these data indicate that the effect of StpC on NF-κB is mediated by the consensus TRAF–NIK–IKK signaling pathway (Fig. 2), which is known to induce degradation of I-κB and activation of NF-κB in multiple cell types. It is noteworthy that the effect of StpC on NF-κB is completely independent of TNF-α, a well-described stimulus of NF-κB activity. In contrast, it appears that StpC uncouples stimulation of NF-κB activity from TNF-α stimulation (Sorokina et al., 2004). Overall, the mechanism including aggregation of TRAF proteins through their physical interactions with StpC followed by constitutive activation of the NIK–IKK cascade, which leads to degradation of I-κB and subsequent activation and nuclear translocation of NF-κB, appears to be a very likely explanation for the cell type-independent effect of StpC on NF-κB.
Based on the above results, the effect of StpC on NF-κB appears to be similar to the effects of other viral proteins “usurping” the signaling pathway resulting in NF-κB activation, including Tax of HTLV-I, LMP1 of Epstein–Barr virus (EBV), and HBx of hepatitis B virus (reviewed in Li and Gaynor, 2000; Hiscott et al., 2001; Yoshida, 2001). Specifically, the molecular mechanisms, by which StpC triggers this pathway, are reminiscent of those mediating the effects of LMP1, which activates NF-κB by physically and functionally interacting with TRAF components of this pathway (Devergne et al., 1996; Kaye et al., 1996, reviewed in Hiscott et al., 2001). Considering that NF-κB activation by Tax and LMP1 is believed to be essential for cell transformation by HTLV-I and EBV, respectively (Yamaoka et al., 1996; Izumi and Kieff, 1997; Matsumoto et al., 1997; Coscoy et al., 1998; Cahir McFarland et al., 1999; Robek and Ratner, 1999; He et al., 2000), the effect of StpC on the consensus signaling leading to NF-κB activation is likely to be essential for its transformation potential. Indeed, it has been shown that mutant forms of StpC incapable of binding to TRAF and activating NF-κB are not capable of transforming fibroblasts either (Jung and Desrosiers, 1994; Lee et al., 1999). Furthermore, mutant H. saimiri encoding for a TRAF-binding-deficient form of StpC lacks the ability to transform human T cells (Lee et al., 1999). It is clear, however, that the contribution of the TRAF–NIK–IKK pathway to the transformation potential of StpC needs further elucidation, since other effects of StpC may also be involved. Indeed, the above-mentioned H. saimiri with StpC lacking the ability to interact with TRAF is still capable of inducing lymphoma in common marmosets and transforming common marmoset T cells (Lee et al., 1999).
Tip: an Lck-interacting and -activating protein
Although Tip is clearly essential for T-cell transformation caused by H. saimiri of subgroup C in vivo and in vitro (Duboise et al., 1998b), the transformation effect of Tip has not been shown in fibroblasts, unlike that of StpA or StpC (Jung et al., 1991). Therefore, studies focused on the transformation-related effects of Tip lack the clear read-out available for those focused on StpC. Another difference between Tip and StpC is that whereas studies of StpC consistently demonstrate that the NF-κB-activating signaling is the major cellular target of StpC, multiple potential targets of Tip have been identified, so our understanding of the role of Tip does not suffer from the lack of leads. It was initially shown that Tip binds to and becomes tyrosine phosphorylated by Lck, a Src-family PTK expressed primarily in T cells and playing a crucial role in T-cell signaling and biological responses (Biesinger et al., 1995; Jung et al., 1995a; Lund et al., 1996). It was later shown that Tip activates Lck in several systems in vitro and in vivo (Wiese et al., 1996; Fickenscher et al., 1997; Lund et al., 1997a) (Fig. 2). Multiple details of the mechanism by which Tip activates Lck have been revealed recently.
It was shown in the initial studies that binding of Tip to Lck requires the presence of a proline-rich SH3-binding motif and a region of homology with Src-family PTKs in Tip and involves the SH3 domain of Lck (Jung et al., 1995a) (Fig. 1). Notably, these motifs are adjacent in the sequence of Tip, and both of them plus a linker connecting them account for only 38 amino acid residues. Further analysis confirmed that the proline-rich motif (termed LBD1) mediates binding of Tip to Lck via the SH3 domain of Lck (Hartley et al., 2000). This study also revealed that an SFL motif within the Src-family PTKs of Tip (termed LBD2) mediates binding of Tip to the C-terminal fragment of Lck that contains the kinase domain, the C-terminal regulatory site and a part of the SH2 domain of this PTK. Tip containing either of its Lck-binding sites is capable of binding to Lck and stimulating its tyrosine phosphorylation, albeit at a degree dramatically lower than that characteristic for wild-type Tip, which contains two functional Lck-binding sites.
The model of Tip–Lck SH3 interactions based on the structures of free Lck SH3 and the Src SH3-APP12 complex was built in a recent NMR study (Schweimer et al., 2002). It was shown in this study that peptides containing the LBD1 with a few flanking residues are capable of binding to the Lck SH3 domain with similar affinities in the micromolar range. Interestingly, these peptides did not demonstrate any binding preference for Lck SH3 as compared to SH3 domains of other Src-family PTKs. Thus, the affinity of Lyn SH3 and Hck SH3 to the LBD1-containing peptides was several-fold higher than that of Lck SH3 (Schweimer et al., 2002). The observed lack of selectivity and the relatively low affinity for the binding of LBD1 peptides to the Src-family SH3 domains are in apparent contrast with the high specificity and affinity of Tip–Lck interactions (Biesinger et al., 1995; Wiese et al., 1996; Fickenscher et al., 1997; Hartley et al., 2000) and, thus, further argue that the interactions of Tip and Lck are not limited to the binding of LBD1 to Lck SH3. However, the interactions of LBD2 with Lck have not yet been studied in detail.
The notion that the interactions of LBD1 and LBD2 with Lck are critical to the effect of Tip on Lck has been supported not only by the mutational analysis of these sites (Hartley et al., 2000), but also by the study demonstrating that a fragment of Tip containing both LBD1 and LBD2 is sufficient for the activation of Lck (Lund et al., 1999). It appears that the activation of Lck by Tip is mediated by direct conformational changes occurring as a result of Tip–Lck interactions, since this activation does not require the presence of either Y394 or Y505 residues of Lck, which mediate physiological regulation of this PTK through its tyrosine phosphorylation (Hartley et al., 1999).
To sum up, physical interactions of Tip and Lck are mediated by the two sites on Tip, LBD1, and LBD2, which interact, respectively, with Lck SH3 and a still unidentified domain in the C-terminal part of Lck, containing the C-terminal end of its SH2, the linker between SH2 and kinase domain, the kinase domain and the C-terminal site of negative regulation (reviewed in Tsygankov, 2003). The stimulating effect of Tip on Lck kinase activity is likely to be due to the disruption of the intramolecular complex of Lck SH3 and the proline-rich motif in the linker connecting Lck SH2 and its kinase domain (reviewed in Tsygankov, 2003). The effects of this type have been shown previously in several experimental systems (Alexandropoulos and Baltimore, 1996; Briggs et al., 1997; Moarefi et al., 1997). Consistent with the finding that the SH2 domain of Lck is not essential for the Lck–Tip complex-formation (Hartley et al., 2000), the mutant form of Tip lacking tyrosine phosphorylation sites and, thus, incapable of being tyrosine-phosphorylated, retains the ability to activate Lck (Hartley and Cooper, 2000).
Lck-dependent activation of STATs by Tip
In spite of not being required for Lck activation, tyrosine phosphorylation of Tip appears to be involved in Tip-induced activation of STAT1 and STAT3 transcription factors (Lund et al., 1997b, 1999) (Fig. 2). This activation was shown to be mediated by the tyrosine phosphorylation of STAT factors by Lck. Further analysis indicated that mutant forms of Tip lacking Tyr-72 were unable to either bind to STATs or facilitate their tyrosine phosphorylation and, hence, transcriptional activity, whereas mutant Tip lacking Tyr-85, another site of tyrosine phosphorylation on Tip, behaved as wild-type Tip (Hartley and Cooper, 2000). (Note that the positions of these tyrosines are given for Tip from strain C484, which was used in these experiments. Tyr-72 and −85 of C484 correspond to Tyr-114 and −127 of C488. It should also be noted that C484 strain is designated C484m in (Ensser et al., 2003).) These results suggest a model whereby Tip binds to and activates Lck; the latter phosphorylates Tip on at least two sites, one of which recruits SH2-containing STAT1 and STAT3 transcription factors, which become tyrosine phosphorylated by Lck as a result. This model is supported by the dependence of STAT activation induced by Tip and Lck on binding of Tip to Lck (Hartley et al., 2000) and that of Tip to STAT (Hartley and Cooper, 2000). In apparent contrast to this view, it was shown that the LBD-containing fragment of Tip lacking its tyrosine phosphorylation sites, which is capable of activating Lck, could also activate STAT3 (Lund et al., 1999). It is tempting to explain this effect of the Tip LBD by a dramatic increase in Lck activity, which results in phosphorylation of STAT even in the absence of adaptor function of Tip. However, the constitutively active form of Lck expressed in the absence of Tip exhibited no such effect (Hartley and Cooper, 2000; Hartley et al., 2000), rendering this explanation unlikely. Therefore, in spite of the clearly enhancing effect of Tip on Lck-dependent activation of STATs, the exact molecular mechanism by which Tip exerts this effect remains to be elucidated.
Tip activates NF-AT
In addition, Tip was shown to activate NF-AT, another crucial transcription factor playing a major role in T-cell stimulation (Hartley et al., 2000; Merlo and Tsygankov, 2001) (Fig. 2). The mechanism of this effect of Tip is even less clear than that of the Tip-induced activation of STATs. Thus, a mutation of either Tip LBD abrogated the ability of Tip to activate NF-AT (Hartley et al., 2000). This finding, together with a well-documented involvement of Lck in the activation of NF-AT (Straus and Weiss, 1992; Baldari et al., 1993; Denny et al., 2000; Veri et al., 2001), appears to implicate Tip-induced activation of Lck in the observed stimulation of NF-AT. However, the finding that activation of NF-AT by Tip is insensitive to PTK inhibitors (Merlo and Tsygankov, 2001) argues that Tip-induced activation of Lck is not required for the effect of Tip on NF-AT.
Tip-induced downregulation of Lck in stable T-cell lines and Tip-induced apoptosis of T cells
In contrast to the results demonstrating a strong activation of Lck by Tip in several experimental systems in vivo and in vitro, downregulation of Lck in Jurkat cells expressing Tip in a stable fashion was also observed (Jung et al., 1995b; Guo et al., 1997) (Fig. 2). Lck activity and its protein level as well as the total protein tyrosine phosphorylation in these cells was decreased, and the response of these cells to T-cell-receptor activation was impaired. These reports also demonstrated that Tip partially reversed transformation of fibroblasts induced by a constitutively active form of Lck. These results supported a model according to which Tip inhibits receptor-mediated signaling of H. saimiri-transformed T cells playing a role similar to that of the EBV LMP-2A protein, which inhibits B-cell-receptor-mediated signaling and facilitates the transition to the latent form of infection. A possible explanation of these conflicting findings was the strain specificity of the effects of Tip, since downregulation of Lck was observed with Tip from H. saimiri C488 strain, whereas most, albeit not all, results demonstrating activation of Lck were obtained using Tip from H. saimiri C484. However, a direct comparison of Tip proteins from these strains of H. saimiri indicated that both Tip proteins stimulated Lck kinase activity in vivo and in vitro and that both stimulated NF-AT- and STAT3-dependent transcription in T cells (Kjellen et al., 2002).
Several authors expressed the view that these conflicting findings may correctly reflect two aspects of the effect of Tip on Lck, an initial activation and a resultant negative feedback, which leads to the downregulation of Lck in T cells expressing Tip in a stable fashion (Broker and Fickenscher, 1999; Tsygankov and Romano, 1999; Isakov and Biesinger, 2000; Fickenscher and Fleckenstein, 2001). A recent study demonstrating a strong apoptotic effect of Tip in T cells, which depends on the presence of functional Lck, supports this view (Hasham and Tsygankov, 2004). Expression of Tip at a high level, while dramatically activating Lck, invariably caused Fas-dependent apoptosis in lymphoblastoid and primary T cells transduced with a Tip-encoding lentiviral vector, so only few clones were obtained following this transduction, and those lacked Tip (Hasham and Tsygankov, 2004) (Fig. 2). In contrast, Tip did not cause apoptosis in J.CaM1.6 cells, an Lck-deficient variant of Jurkat, and various Lck-negative cells were easily transduced to express Tip in a stable fashion (Hasham and Tsygankov, 2004; Sorokina et al., 2004). Overall, these findings support the idea that the Tip-induced activation of Lck at the molecular level may cause a negative feedback at the cellular level, which results in a decrease or loss of Lck and/or Tip in the cells initially expressing both Tip and Lck. The apoptotic death of cells expressing Tip and Lck may be one of the mechanisms mediating this feedback. It is also very likely that downregulation of Lck upon its activation by Tip is mediated by a general mechanism of downregulation of activated Lck—c-Cbl-dependent ubiquitylation of Lck followed by proteosomal degradation of ubiquitylated Lck (Rao et al., 2002).
However, stable expression of Tip is not always detrimental for cell viability. H. saimiri-transformed cells express Tip, but do not die of apoptosis. Likewise, a low level of stable expression of Tip is not detrimental for cell viability for T-lymphoblastoid cell lines (Merlo et al., 1998). It is likely that the peculiar mechanism of tip gene expression employed by H. saimiri, when Tip is encoded by the second open reading frame of the bicistronic stpC/tip mRNA (Biesinger et al., 1990; Fickenscher et al., 1996), evolved to reduce the cellular level of Tip in order to prevent excessive apoptosis. It is possible that at early steps of T-cell infection by H. saimiri, Tip is expressed at a level higher than that in H. saimiri-transformed cells, and that the initial high-level expression of Tip causes death of a substantial fraction of H. saimiri-infected cells. The effect of Tip at this stage of the H. saimiri life cycle can explain the presence in the genome of H. saimiri of two viral anti-apoptotic genes, v-FLIP (Thome et al., 1997) and a Bcl-2 homologue (Nava et al., 1997; Derfuss et al., 1998), which are not expressed in H. saimiri-transformed cells (Fickenscher et al., 1996; Knappe et al., 1997). In transformed T cells the apoptotic effect of Tip may be low because of the moderate levels of Tip expression and Lck activity in these cells and, as a result, may be effectively balanced by normal cellular anti-apoptotic mechanisms and/or NF-κB, which has an anti-apoptotic effect and is activated by StpC (see above).
Binding of Tip to Tap: a possible link to the nuclear export of RNA
Another protein that may be involved in the biological effects of Tip is Tap, a protein involved in the nuclear export of mRNA (Yoon et al., 1997). Tap has been characterized as a protein interacting with the nuclear pore complex and mediating the export of cellular mRNAs and some viral RNAs that bear the constitutive transport element (Bachi et al., 2000; Paca et al., 2000; Braun et al., 2001, 2002; Wiegand et al., 2002, reviewed in Weis, 2002; Cullen, 2003). Although the putative effect of Tip on Tip-mediated RNA export has not yet been characterized, one cannot rule out that this effect may regulate the life cycle and replication of H. saimiri (Fig. 2). However, the magnitude of the observed biological effect of co-expression of Tip and Tap in Jurkat cells is very modest; only ∼10% of cells when expressing both Tip and Tap at a high level exhibited an increase in the surface density of some receptors, which correlated with an increase in cell aggregation (Yoon et al., 1997). This finding recommends caution in interpreting the role of Tip–Tap interactions in biological effects of Tip.
Binding of Tip to p80: a possible link to the lysosomal degradation
The list of Tip-associated proteins has recently been extended to include p80, a WD domain-containing protein involved in endocytosis (Park et al., 2002). Since p80 possesses binding domains for Tip and lysosomes, it can recruit Tip with the Tip-associated Lck to lysosomes. The interaction of Tip with p80 facilitated lysosomal vesicle formation and downregulation of Lck (Park et al., 2002). Furthermore, the interactions of Tip with p80 resulted in downregulation of T-cell receptor (TCR)/CD3 in a lysosome-dependent fashion (Park et al., 2002, 2003). Downregulation of CD4 was also observed, but it was independent of p80 and occurred in a lysosome-independent fashion (Park et al., 2002, 2003). Finally, downregulation of both TCR/CD3 and CD4 was shown to be dependent on the interactions between Tip and Lck. Overall, the results of these studies were interpreted as further supporting the idea that Tip acts as a latency-inducing protein by downmodulating T-cell signaling pathways (Fig. 2). However, the observed magnitude of the effects of H. saimiri-induced transformation on the expression of Lck, TCR/CD3, and CD4 in T cells is quite moderate. In fact, the lack of overt effects on a phenotype of human T cells, including their surface receptor expression, signaling, and biological responses, has always been acknowledged as an advantage of using H. saimiri for obtaining transformed T-cell lines and clones as experimental models (reviewed in Broker and Fickenscher, 1999; Tsygankov and Romano, 1999; Fickenscher and Fleckenstein, 2001). Finally, effects of Tip and p80 on Lck and surface receptors were observed only upon co-expression of Tip and p80 at high levels. It is likely, therefore, that the effects of Tip and p80 observed in this experimental system demonstrate a considerable potential of Tip–p80 interactions to downregulate Lck and surface receptors, but not the actual extent of these effects in H. saimiri-transformed T cells. The biological effect of these interactions and their relevance for the transformation-related effect of Tip require further elucidation.
Relevance of the molecular effects of Tip for H. saimiri-induced transformation
The same conclusion appears to correctly describe all findings related to the interactions of Tip with various cellular targets and to the effects of these interactions on cellular functions and, specifically, H. saimiti-induced transformation. It remains unclear, which of the reported effects of Tip, if any, is(are) linked to its role in T-cell transformation. The originally expressed views that activation of Lck by Tip is essential for the transformation-related effects of Tip was questioned, when H. saimiri encoding Tip with a mutationally disabled LBD1 was shown to retain the ability to transform T cells in culture and induce experimental lymphoma in primates (Duboise et al., 1998c). This result was initially interpreted as the demonstration of the full transformation competency of Tip incapable of interacting with Lck. However, the results obtained later argued that LBD1-deficient Tip cannot be considered fully deficient. First, it retains some ability to bind to and activate Lck (Hartley et al., 2000). Furthermore, it is quite possible that the transformation-related effect of Tip does not require a global activation of Lck, being dependent on activation of specific proteins, such as STATs. If it does not require it, LBD1-mutated Tip cannot be considered deficient, since it is capable of activating STAT3 almost to the same extent as wild-type Tip does (Hartley et al., 2000).
Molecular mechanisms of the synergism between Tip and StpC in H. saimiri-induced transformation
Overall, the two general points of view on the essential role of Tip and StpC in H. saimiri-induced T-cell transformation are as follows. One of them states that both Tip and StpC contribute to the transformation process by inducing partial effects on cellular functions, such as signal transduction, that synergize, ultimately resulting in transformation. The other view argues that StpC acts as a transforming protein, whereas Lck plays its essential role by downregulating signaling and, thus, inducing latency. The currently available data on the role of Tip in H. saimiri-induced transformation do not allow us to definitively select or reject any of the proposed mechanisms of transformation-related effects of Tip, and thus, the nature of synergism between Tip and StpC remains unclear. In addition to multiple technical problems mentioned above, the assessment of the role of Tip in H. saimiri-induced transformation is hindered by the lack of clarity with the partial transformation potential of Tip. As mentioned above, StpC is highly oncogenic in fibroblasts in cell culture and in epithelial cells in transgenic animals (Jung et al., 1991; Murphy et al., 1994). Whereas Tip clearly lacks an oncogenic potential in “classical” fibroblast systems (Jung et al., 1991), results arguing that it can induce T-cell lymphomas have been obtained (Wehner et al., 2001). These findings support the idea that Tip and StpC generate partial transforming effects that synergize facilitating T-cell transformation, although the results on the partial transforming effect of Tip need to be elucidated. First, the expression level of Tip in Tip-transgenic mice was much higher than in any other experimental system, making uncertain the degree of similarity between T cells expressing a Tip transgene and T cells infected with H. saimiri. Second, the clonality of T-cell infiltrations in Tip-transgenic mice was not shown, contrary to what is expected in a typical lymphoma, arguing that the infiltrates might be unrelated to transformation. Third, no transformed cell line was obtained from the Tip-transgenic mice.
Although the details of the molecular mechanisms by which Tip and StpC contribute to T-cell transformation remain to be elucidated, the correlation between the transformation potential of various H. saimiri strains and the organization of their transformation-related genes suggests that the “division of labor” between Tip and StpC in the strains of subgroup C enhances oncogenicity of these strains as compared to that of strains of subgroups A and B, which have a single transformation-associated protein, StpA or StpB, respectively. This “division of labor” between Tip and StpC appears to be reflected at the molecular level. Thus, the ability of StpC to interact with PTKs in general and with Src-family PTKs in particular has never been shown, whereas Tip is a strong and specific interactor of Lck, a Src-family PTK (see above). In contrast, StpA has been shown to bind to and to be phosphorylated by Src on Tyr-115 (numbering is for A11 strain) (Lee et al., 1997). Binding of StpA to Src appears to be mediated by the SH2 domain of Src. Furthermore, this phosphorylation renders StpA capable of binding to the SH2 domains of Lck and Fyn (Lee et al., 1997). In addition, StpA has been shown to bind to STAT3 and to facilitate its Src-dependent tyrosine phosphorylation and, hence, transcriptional activation (Chung et al., 2004; Park et al., 2004). These findings together with the inability of StpA to activate NF-κB under the conditions resulting in a strong StpC-induced NF-κB activation, in spite of the ability of StpA to bind to TRAF (Lee et al., 1999), suggest that the effect of StpA on T cells resembles that of Tip, while having some features of StpC-mediated effects. This conclusion based on the molecular mechanisms of the effects of StpA is consistent with the finding that StpA is oncogenic not only in fibroblasts (Jung et al., 1991), but also in transgenic mice, where StpA causes T-cell lymphomas (Kretschmer et al., 1996). Thus, StpA acts in transgenic mice similarly to Tip, which appeared to induce T-cell lymphomas (Wehner et al., 2001), and in contrast to StpC, which causes only epithelial tumors in transgenic mice (Murphy et al., 1994). Molecular effects of StpB were shown to be similar to those of StpA. Thus, StpB binds to TRAF proteins, but does not detectably increase NF-κB activity (Choi et al., 2000). Furthermore, StpB binds to and becomes tyrosine phosphorylated by Src on Tyr-118 (numbering is for B-SMHI strain) homologous to Tyr-115 of StpA in A11 strain (Choi et al., 2000; Hor et al., 2001). Binding of StpB to Src is mediated by phosphotyrosine on StpB and Src SH2 as it was previously shown for StpA (Hor et al., 2001).
The oncogenic potential of the transforming system of H. saimiri of subgroup C, which consists of two proteins, appears to be “optimized” for inducing transformation in T cells as compared to either StpA or StpB. First, the ability of Tip to bind to and to become phosphorylated by Src-family PTKs is specific for Lck (Biesinger et al., 1995; Wiese et al., 1996; Fickenscher et al., 1997), whereas StpA is phosphorylated specifically by Src (Lee et al., 1997) and only Src-mediated phosphorylation has been shown for StpB (Choi et al., 2000; Hor et al., 2001). This difference should have dramatic consequences for the ability of Tip, on the one hand, and StpA and StpB, on the other, to affect T-cell signaling, since Lck is the major Src-family PTK of T cells expressed almost exclusively in this type of cells, whereas Src is not expressed in normal human T cells (reviewed in Tsygankov, 2003). Second, Tip has been shown to facilitate Lck-dependent activation of STAT1 and STAT3 (Lund et al., 1997b; Hartley and Cooper, 2000), whereas StpA does not bind to STAT1 and activates STAT3 only (Chung et al., 2004). Third, the ability of Tip to activate NF-AT (Hartley et al., 2000; Merlo and Tsygankov, 2001) has not been shown for either StpA or StpB. Finally, StpC has apparently lost the ability of StpA and StpB to interact with Src-family PTKs, but acquired a strong NF-κB-activating potential (Lee et al., 1999; Merlo and Tsygankov, 2001; Sorokina et al., 2004). Therefore, H. saimiri strains of subgroup C are capable of strongly activating a host of transcription factors known to be crucial for T-cell activation, such as NF-AT, NF-κB, STAT1, and STAT3, with at least some of these occurring in a T cell-specific fashion. In contrast, StpA and, especially, StpB appear to activate fewer transcription factors, and this effect does not seem to be optimized for T cells.
Downregulation of Lck in the natural host as a possible defense mechanism
The extent of the observed difference in transformation potential between H. saimiri strains of subgroup C, on the one hand, and the subgroups A and B, on the other, is so profound that the squirrel monkey, a natural host of H. saimiri, appears to develop a mechanism that specifically counteracts the transformation effect of Tip. Namely, Lck is enzymatically inactive in squirrel monkey T cells, in spite of the normal expression level of Lck in these cells and the lack of mutations detrimental to Lck activity (Greve et al., 2001). The molecular basis of this inactivation is not clear, but the sensitivity of squirrel monkey Lck to pervanadate-induced stimulation suggests that Lck is downregulated in squirrel monkey T cells by the action of protein tyrosine phosphatases. Regardless of its molecular mechanism, dramatic downregulation of Lck should block the effects of Tip on squirrel monkey T cells and, thus, prevent their transformation by highly oncogenic H. saimiri viruses of subgroup C. It should also be noted that this peculiarity of squirrel monkey Lck may also play a role in protecting T cells of the natural host of H. saimiri from Tip-induced apoptosis (Hasham and Tsygankov, 2004).
Viral proteins other than Tip and StpC that might promote H. saimiri-induced transformation
In addition to stp and tip, several other genes of H. saimiri were hypothesized to be involved in viral transformation. Most of these genes are homologous to cellular proteins that may exert transformation-promoting effects. For example, open reading frame (ORF) 72 of H. saimiri encodes v-cyclin, a protein homologous to cellular cyclins of type D, which binds to Cdk6 and activates its kinase activity (Jung et al., 1994). v-Cyclin–Cdk6 complex is resistant to such Cdk6 inhibitors as p16 (Ink4a), p21 (Cip1) and p27 (Kip1). Furthermore, ectopic expression of v-cyclin prevents G1 arrest imposed by these inhibitors and stimulates cell-cycle progression in quiescent fibroblasts (Swanton et al., 1997). The structural basis of these effects was recently characterized by solving a crystal structure of the v-cyclin–Cdk6 complex (Schulze-Gahmen and Kim, 2002). Finally, v-cyclin–Cdk6 complexes were shown to directly trigger the initiation of DNA synthesis in isolated late-G1-phase nuclei (Laman et al., 2001). Although these results suggest that v-cyclin is involved in H. saimiri-induced transformation, v-cyclin is dispensable for both in vitro transformation of human and monkey T cells and induction of lymphoma in primates (Ensser et al., 2001).
Another candidate for playing a role in H. saimiri-induced transformed growth is the product of ORF13, which encodes viral IL-17 (Yao et al., 1995a,b). Cellular and viral IL-17 induces secretion of various cytokines, including IL-8, by stroma cells (Fossiez et al., 1996). Among other functions, IL-17 was shown to support proliferation of T cells (Yao et al., 1995a,b; Fossiez et al., 1996; Kennedy et al., 1996). Furthermore, the product of ORF74 is a viral homologue of the G protein-coupled IL-8 receptor (Ahuja and Murphy, 1993), which may facilitate the effect of IL-8, if it is produced by stroma cells in response to viral IL-17. However, the transformation potential of H. saimiri mutants lacking IL-17 is not impaired (Knappe et al., 1998a).
A group of viral genes consisting of two apoptosis inhibitors encoded by ORF71 (v-FLIP) (Thome et al., 1997) and ORF16 (Bcl-2 homologue) (Nava et al., 1997; Derfuss et al., 1998) was also implicated in transformation effects of H. saimiri. Indeed, anti-apoptotic proteins frequently promote cell transformation (for review, see Coultas and Strasser, 2003; Harada and Grant, 2003). However, no increase in resistance to apoptosis was detected in H. saimiri-transformed T cells (Broker et al., 1997; Kraft et al., 1998). Consistent with these results, v-FLIP was shown to be dispensable for T-cell transformation and lymphoma induction (Glykofrydes et al., 2000).
Another viral protein that might promote growth of H. saimiri-transformed T cells is the product of ORF14 designated as vSag. It is homologous to the mouse mammary tumor virus superantigen and murine M1s superntigens (Nicholas et al., 1990), is secreted from transformed cells in glycosylated form and is capable of activating human T lymphocytes (Yao et al., 1996). It was noted, however, that since no evidence for selective activation of T cells bearing specific TCR-Vβ families by vSag was presented (Yao et al., 1996; Knappe et al., 1997; Duboise et al., 1998a), this protein should be referred to as a “superantigen homologue” or “mitogen” (Broker and Fickenscher, 1999; Fickenscher and Fleckenstein, 2001). However, regardless of the mode of interaction with TCR, vSag is unlikely to be essential for T-cell transformation, since the deletion of ORF14 did not impair its ability to transform T cells in vitro and induce tumors in primates (Knappe et al., 1997, 1998b). In contrast, another study showed that a similar mutant of H. saimiri was unable to transform non-human primate T cells in vivo or in vitro (Duboise et al., 1998a). This conflict between the results obtained by two groups may be explained by either species specificity of H. saimiri transformation or differences in experimental protocols, such as comparably lower viral titers in the latter study.
Since the leftmost L-DNA region of H. saimiri, which is essential for transformation, contains not only protein-encoding ORFs, but several genes for small U RNAs as well, potential contribution of these U RNAs to the transforming effect of H. saimiri was also examined using mutant viruses. The genes encoding for U RNAs were shown to be dispensable for H. saimiri-induced transformation (Medveczky et al., 1993; Ensser et al., 1999).
Overall, little evidence has been presented for a role of any gene of H. saimiri other than stp or tip in H. saimiri-induced transformation of human T cells. The situation may be different in New World primates, in which H. saimiri establishes a semi-permissive infection, when multiple genes of H. saimiri including those homologous to cellular regulators are expressed. Although the studies described above argue that most of these proteins are not essential for transformation of New World monkey T cells, some of them may contribute to this transformation. An important additional argument against the essential role of these proteins in T-cell transformation would be successful transformation of T cells by Tip and StpC in the absence of any other viral proteins. However, attempts to transduce T cells by retro/lentiviral transduction of Tip and StpC cDNAs did not succeed, although these results can be explained by important differences between such a system and H. saimiri-induced transformation that are unrelated to the presence of other viral proteins (see Hasham and Tsygankov, 2004 as an example).
A recent report indicated that ORF5 of H. saimiri encodes another protein that interferes with T-cell signal transduction (Lee et al., 2004). ORF5-encoded protein is localized to the membrane because of myristoylation, becomes tyrosine phosphorylated by Lck and Fyn, interacts with Lck, Fyn, SLP-76, and p85 subunit of PI-3′ kinase through phosphotyrosine-SH2 interactions and facilitates TCR/CD3-mediated signaling. Finally, it is capable of partially substituting for the T-cell adaptor protein LAT in LAT-negative Jurkat cells. However, in spite of the apparently dramatic effect of the ORF5-encoded protein, it is not expressed during the period of viral latency, but only during the replication cycle (Lee et al., 2004). Therefore, an essential role for this protein in T-cell transformation is unlikely.