Significant progress has been made in the study of Tap since the last committee meeting in September 1998, which has lead to one paper in press in Genes and Development, one paper submitted to RNA and another that will be submitted to Nature Cell Biology soon. We have clearly demonstrated that Tap IS the sequence-specific cellular cofactor for CTE export, with the following convincing evidence:

1. We finally have a convenient and powerful assay system to directly analyze the function of Tap in the nuclear export of CTE. The quail QCL-1 cell line is non-permissive for CTE function, but can be rendered completely permissive after transfection of a plasmid expressing human Tap protein. We also found that mouse cells are partially permissive for CTE function. Mouse Tap protein, which binds to CTE at about half the affinity of human Tap, is able to enhance CTE export, albeit less potently than human Tap. It's unclear whether the quail cell line we used lacks expression of Tap, or quail Tap is unable to bind to CTE.

2. As we discussed in the first committee meeting, randomization/selection of the CTE gave rise to wild type CTE and three CTE sequence variants that differ from the wild-type sequence at only one nucleotide. Wild type CTE is the optimal Tap binding sequence whereas CTE variants bind to Tap with lower affinity. The affinity of binding of CTE variants to Tap correlates very well with their ability to be exported, both in human cells and quail cells transfected with human Tap.

3. Using heterokaryon and microinjection assays, we have proven that Tap is a nucleocytoplasmic shuttling protein, with an NLS located at each end of the protein. The C terminal NLS is also a potent NES. NES mutants of Tap exert a dominant negative effect on the ability of wild type Tap to induce CTE export in quail cells.

4. The Rev NES can functionally replace the Tap NES and rescue Tap NES mutants in supporting CTE export in quail cells. However these two NESs apparently function through different pathways as Rev NES function is inhibited by reagents (such as LMB or DCAN) that block the export of Rev NES, while the Tap NES is not affected.

We have also shown that the N-terminal NLS of Tap is recognized and imported by Transportin (Importin b2). HnRNP M9 is also imported by Transportin (in fact, this is how transportin was identified). Interestingly while M9 is a bi-directional import/export signal, the N-terminal NLS of Tap is only an import signal, as shown by microinjection and heterokaryon assays. This is the first example of a transportin-dependent nuclear import signal. Detailed analysis of the interaction between the N-terminal NLS of Tap and Transportin will shed light on the sequence specificity and mechanism of function of Transportin.

B. Experimental design for future experiments:

My major effort now is focused on finding the nuclear export receptor/cofactor for Tap NES. Identification of such a cellular co-factor not only will complete the functional cycle of Tap in the nuclear export of CTE-containing mRNA, but will also provide us a handle to study the role Tap plays in the nuclear export of mRNA in general. Other interesting questions concerning the mechanism of function of Tap, such as why nuclear export of CTE requires two intact internal loops, even though one loop is sufficient for binding, is worth addressing when time allows.

I am currently using the yeast two hybrid system to identify the export receptor for Tap NES, using the GAL4-Tap NES fusion protein as a bait. The yeast strain pJG9-4A is used in this study. Positive clones can be identified by using either b-gal, histidine or adenine selections. I obtained ~20 positive clones in a large scale screen using adenine selection. About half of them also have positive b-gal activity. Currently I am in the process of characterizing these clones. GAL4-Tap NES-A17 mutant will be used to eliminate false positive clones. Any remaining real positives will be further analyzed in the experiments described below:

a. Specific interaction with Tap NES, in yeast and mammalian two hybrid, as well as in vitro binding by affinity column or co-immnoprecipitation.

b. Deletion mutants of such a cofactor, which retained to binding ability to Tap NES while losing other functional domains, might be an inhibitor for Tap function in human or quail cells. If Tap does play a role in the nuclear export of cellular mRNA we should also expect to see a global inhibition of cellular mRNA export (nuclear accumulation) in cell lines expressing such a dominant mutant form of the Tap NES receptor.

c. Cellular localization of the cofactor. We expect to see nuclear, or even nuclear pore staining in immunofluorescence analysis.

d. Homology to yeast protein? In yeast, Mtr2p bridges the interaction of Mex67p, the yeast homolog of Tap, with nuclear pores. If the interaction is conserved from yeast to human, we should expect to see the human Tap cofactor to be homologous to yeast Mtr2p. NCBI database search, using yeast Mtr2p sequence, did not yield any obvious human homolog though.

e. We can also expect to see that LMB and
CAN will be unable to inhibit the interaction between Tap and its export receptor. Using the mammalian RNA binding assay system we can also expect to see that Tap can interact with its export receptor when it's loaded with CTE RNA.
f. It will be interesting to see how RAN.GTP and RAN. GDP affect the interaction between Tap and its export receptor, particularly if that co-factor doesn't belong to the importin b super family.

If the two hybrid approach isn't successful, I will try to use a biochemical approach such as affinity column purification or co-immunoprecipitation to identify the cofactor. The criteria mentioned above can still be applied to analyze the candidate cofactors once they are cloned.

We just knew today that in the May 4th issue of EMBO J, Ed Hurt's group will publish a paper entitled "The Mex67p-mediated nuclear mRNA export pathway is conserved from yeast to human". Briefly they showed that the C-terminus of Tap, which is mapped as the shuttling domain in our study, interacts directly with the FG repeats of nucleoporins such as CAN and Rab in the yeast two hybrid screen and in in vitro direct binding assays. They also showed that the middle part of Tap (aa352-550) interacts with a novel human protein p15, which is homologous to yeast nuclear import factor NTF2. Co-expression of human Tap and p15 restores growth of the otherwise lethal mex67::mtr2 double knockout strain. We are in the process of sequencing positive clones identified in our yeast two-hybrid screen to see any of them is nucleoporins or p15. We will also see if Tap NES can interact with FG repeats in mammalian and yeast two hybrid, and if A17 mutation knock out this interaction. All of the experiments suggested above for a candidate Tap NES cofactor will be carried out if such a Tap NES-FG repeat interaction is true.