DLC1 has been shown to exert biological functions resembling those of classical tumor suppressor genes. The diverse tumor-suppressive effects of DLC1 are strongly dependent on the presence of a functional RhoGAP domain 
. However, several studies have shown that RhoGAP activity alone is not sufficient for its tumor-suppressive function 
. Previous structural analysis from our group provided the first evidence that these functions of DLC1 could be restored only when the region between the SAM and RhoGAP domains was included 
. The identification of tensin2 as the first binding protein that interacts with this functionally important region further hinted at the relationship between tensin2 interaction and DLC1 function 
. This notion was further supported by the discovery of cten and tensin1 as binding partners of DLC1 in subsequent studies, implying that it is biologically important for tensin family proteins to work co-operatively with DLC1 
We previously proposed that DLC1 375–509 is the minimal binding region of tensin2. This region covers the key residues Y442 and S440, which were suggested by other studies to constitute an SH2 binding site for cten and tensin1 
. However, we could not provide direct evidence that the tensin2 SH2 domain was sufficient for DLC1 binding 
. In this study and our previous report, our findings consistently showed that that tensin2 PTB, rather than SH2, domain directly interacted with DLC1.
There are a few possible explanations for the discrepant findings. First, although the SH2 and PTB domains in tensin proteins have a high sequence homology, differences in their sequences may result in different binding mechanisms for DLC1 and other tensin family members. Second, different binding assays were used by different groups to study the mechanism of binding between tensin and DLC1. For instance, yeast-based binding and GST pull-down assays were used by Liao et al. and Qian et al., respectively, while in vivo co-immunoprecipitation was employed in our present study. Third, C-terminus fragments of tensin that contained either one or both of the SH2 and PTB domains were used in mapping the DLC1 binding site in the studies by Liao et al. and Qian et al. This experimental design presumed that the N-terminus region of the tensin protein neither was involved in binding nor contributed to the normal protein folding of the C-terminus region.
In the present study, we provide the first evidence that specific removal of the SH2 and PTB domains in tensin2 affects DLC1 binding to different extents ()
. To provide evidence that the PTB domain plays a critical role in DLC1-tensin2 binding, we mapped the PTB-binding domain at the N-terminus of DLC1 ()
and confirmed its role by characterizing the DLC1ΔPTB internal deletion mutants ()
. We identified an undocumented region 375–385 of DLC1 as the PTB-domain-binding site and the DLC1-tensin2 interaction was lost when this region was removed ()
. It is well established that the PTB domain recognizes a NPXY motif, as illustrated in its binding with integrin β 
. However, this motif was not found in our mapped region in DLC1, implying that the mode of PTB-mediated interaction is a probably atypical. Although DLC1ΔPTB showed loss of tensin2 binding, it was still able to localize to focal adhesions ()
. In addition, we found that the PTB domain was necessary for the focal adhesion localization of tensin2 (data not shown). Thus, in addition to acting as the physical binding site for DLC1, the PTB domain may also be important for bringing tensin2 in close proximity to DLC1 at focal adhesions and facilitating their interaction. Collectively, these results suggest that DLC1 utilizes region 375–385 in mediating tensin2 binding, while it utilizes Y442 and S440 residues for focal adhesion targeting ()
Proposed model of the focal adhesion targeting and tensin2 interaction of DLC1.
It is currently unknown whether the proposed tensin PTB binding domain in DLC1 is involved in binding with other tensins. This remains to be investigated. There have been reports that expression of the tensin1 PTB domain is sufficient for interaction with DLC1. Also, deleting the whole tensin SH2 binding motif (440–448), instead of introducing a Y442F point mutation in DLC1, is sufficient for interaction with tensin1 
. This indicates that PTB-dependent binding may be present between DLC1 and tensin1, but plays a subtle role, unlike the DLC1 and tensin2 interaction. Based on our present findings, we conclude that the PTB binding mechanism is tensin2-specific ()
, whereas other tensins utilize an SH2 binding mechanism 
. Possible compensation by other tensins may explain why DLC1ΔPTB can be localized to focal adhesions in the absence of tensin2 interactions. In addition, we found that DLC1ΔPTB showed a partial reduction in RhoGAP activity, which resulted in a partial reduction in growth suppression ()
A limitation in our present study is the use of ectopically expressed tensin2. Detection of endogenous tensin2 was not possible, as the tensin2 antibody was not available in our laboratory. Interaction between endogenous DLC1 and tensin2 must be examined to reflect their binding in the actual biological context. On the other hand, we performed preliminary quantitative real time PCR analysis to determine the mRNA expression levels of DLC1 and tensin2 in human HCC tissues. Our unpublished data showed that underexpression of both DLC1 and tensin2 was correlated with shorter overall survival when compared with those who had normal expression of either or both genes, supporting the possible functional association of DLC1-tensin2 with hepatocarcinogenesis.
Altogether, we have identified a novel tensin2 PTB binding site in DLC1 and demonstrated its involvement in tensin2 interactions. Although the removal of the PTB binding site did not affect the focal adhesion localization, it partially reduced the RhoGAP activity of DLC1, which attenuated its growth suppressive function. It would be interesting to uncover how DLC1 may control the activity of other focal adhesion molecules. We have also provided early evidence that the DLC1 paralog, DLC2, may also interact with tensin2 and localize to focal adhesions (). The conservation of the tensin binding site in DLC2 warrants further investigation into the localization control of the other DLC family members. This will clearly help to determine if DLC members are biologically redundant or if they have different compartmentalization for performing separate biological functions.