One of the remarkable aspects of VSV pathogenesis is the ability of M protein to induce pleiotropic effects in infected cells. M protein plays multiple roles in both virus assembly and in the inhibition of host gene expression 
. M protein inhibits transcription by all three host RNA polymerases 
, inhibits nucleo-cytoplasmic RNA transport 
, and plays a role in inhibition of translation of host mRNA 
. One of the mechanisms by which M protein may serve these diverse functions is through interaction with host proteins, such as Rae1, that may also serve multiple functions in a cell.
Previous data had shown that M protein interacts with Rae1-Nup98 complexes, but did not address the ability of M protein to interact with other forms of Rae1. It was originally thought that interaction of M protein with Rae1-Nup98 complexes was responsible for blocking nuclear-cytoplasmic transport. Therefore our hypothesis was that M protein interacts with other forms of Rae1 to inhibit other steps in host gene expression, such as transcription and translation. However, the data presented here show that Rae1-Nup98 complexes are the major form of Rae1 capable of interacting with M protein. In situations where the cellular levels of Rae1 or Nup98 are altered, either by overexpressing Rae1 or silencing expression of Nup98, the low molecular weight form of Rae1 also interacts with M protein. Furthermore, rather than affecting the accumulation of host RNA in the nucleus, the major effect of silencing Rae1 expression was to make the cells more resistant to the inhibition of transcription by VSV. These results lead to a new model for how the interaction of M protein with Rae1 inhibits host gene expression. They also support the idea that Rae1-Nup98 complexes play a previously under-appreciated role in regulation of cellular transcription.
M protein does not inhibit host gene expression simply by interfering with Rae1 function, since Rae1 is not essential for host gene expression 
. This raises the question of how interaction of M protein with a sub-population of a protein that is not essential for gene expression can have a global effect on host gene expression at multiple levels. To address this paradox, we propose a model where M protein interacts with Rae1-Nup98 complexes that serve as a platform for M protein to interact with other essential host proteins, thereby, interfering with their function. The “platform hypothesis” predicts the opposite effects of silencing Rae1 expression compared to hypotheses based on M protein inhibition of Rae1 function. The latter hypotheses predict that Rae1 siRNA cells should be more sensitive than control cells to the effects of M protein, because of the lower level of Rae1 expression. In contrast the “platform hypothesis” predicts that Rae1 siRNA cells should be less sensitive to the effects of M protein than controls, since there is less Rae1 to mediate the interaction of M protein with other targets. Our data showing that Rae1 siRNA cells are relatively resistant to the inhibitory effects of M protein (, , and ) provide support for the platform hypothesis and are largely inconsistent with hypotheses based on M protein interference with Rae1 function.
The structural features of Rae1-Nup98 complexes are well-suited to mediate the interaction of M protein with other cellular targets. Rae1 is a member of the family of WD repeat proteins 
, which are known to adopt beta propeller folds 
that have large surface areas suitable for multiple protein interactions. Human Rae1, which has four WD repeats in its sequence 
, has been shown to form seven bladed β propellers with extensive surface loops 
, which provide large surface areas that could serve as interacting regions for M protein to disrupt function of other proteins associated with Rae1. Nup98 has a small globular region near its C-terminus. However, most of the remaining sequence contains FG-repeats, which are intrinsically disordered in other FG-repeat-containing proteins 
. The FG-repeat region of Nup98 provides sites of interaction with a wide variety of other cellular proteins 
. Rae1-Nup98 complexes have a larger apparent Stokes radius in gel filtration and a smaller sedimentation velocity than would be expected of compactly folded proteins. This could be due to an elongated structure, but is most likely due to the presence of disordered regions in the protein sequence. From our gel filtration data and that of Matsuoka et al
, the Stokes radius of the Rae1-Nup98 complex is estimated to be 70–75 Å, which combined with an estimated s20,w
of 5S () gives a molecular weight of approximately 150,000. This is a reasonable result for a 1
1 complex of Rae1 and Nup98.
Our data also support the idea that the interaction of M protein with Rae1-Nup98 complexes inhibits host gene expression by inhibiting host transcription (). Previous data had suggested that Rae1 and Nup98 interact with the transcriptional machinery. Both Rae1 and Nup98 are present in the nucleoplasm, as well as the nuclear envelope and cytoplasm 
. The localization of Rae1 at the nuclear envelope is affected by inhibitors of RNA polymerase I and II activity 
. This suggests that the localization of Rae1 is dependent on ongoing transcription. Similarly, the mobility of Nup98 in the nucleus is also dependent on ongoing transcription 
. Nup98 in the nucleoplasm has been shown to interact with developmentally regulated genes in Drosophila
, and altering Nup98 expression alters the expression of these genes, implicating Nup98 in the control of transcription 
Recent evidence suggests that the steps involved in gene transcription, nascent mRNA processing, and transport are coupled 
. Rae1 can be cross-linked to poly A-containing mRNA 
, and Rae1 interacts with other mRNA binding proteins 
, suggesting that Rae1 and Nup98 may be a part of larger ribonucleoprotein complexes in the nucleus. M protein, by interacting with Rae1 and Nup98, would target these complexes to inhibit both transcription and transport of nascent mRNA. Although Nup98 is likely to be important for the M protein-mediated inhibition of nuclear-cytoplasmic RNA transport, it may not be important for the inhibition of host transcription, since silencing Nup98 expression did not affect the inhibition of host transcription by VSV (). Alternatively, the level of silencing of Nup98 may not have been sufficient to have an effect on the inhibition of host transcription by VSV.
The effect of M protein on nuclear accumulation of cellular RNA depends on the cell type and the mRNA target being analyzed. Nuclear accumulation of RNA resulting from the inhibition of transport is most obvious in cells in which M protein has little if any effect on transcription, such as Xenopus
. However, in most mammalian cells, there is relatively little net accumulation of constitutively expressed mRNAs relative to pre-existing mRNAs in the nucleus during VSV infection, because their synthesis as well as their transport is inhibited by M protein. This was originally demonstrated in the pulse-chase experiments of Weck and Wagner 
, and is confirmed by our analysis of the distribution of actin mRNA (). In contrast to constitutively expressed mRNAs, mRNAs for IL-6 and cJun, which are induced by VSV infection, accumulate in the nucleus, with very little present in the cytoplasm ().
Experiments using in situ
hybridization with oligo-dT have shown an apparent accumulation of total mRNA in the nucleus of VSV-infected cells. However, this result has not been confirmed by an independent approach and may be subject to artifacts such as masking of poly A-containing mRNAs in the cytoplasm of VSV-infected cells as a result of their accumulation in poorly translating ribonucleoprotein particles 
. Further, not all nuclear-cytoplasmic transport is inhibited by M protein. For example, export of tRNA 
and RNA bearing constitutive transport element 
is resistant to the inhibition, as is export of complexes containing hnRNP-A1 and other hnRNPs 
. Silencing Rae1 expression inhibits export of hnRNP-A1 in VSV-infected cells 
, but has little if any effect on nuclear accumulation of host mRNAs (). The level of mRNA in the nucleus reflects a balance of transcription, transport, and turnover. Thus it is possible that silencing Rae1 expression may have an effect on mRNA transport in VSV-infected cells that is balanced by changes in transcription or turnover.
The inhibition of host translation in VSV-infected cells is not due to depletion of host mRNAs from the cytoplasm as a result of the inhibition of host transcription and nuclear-cytoplasmic transport 
. Instead, the translational apparatus is altered in VSV-infected cells such that translation of pre-existing host mRNAs is inhibited, and only newly appearing mRNAs are translated 
, namely those produced by the viral RNA-dependent RNA polymerase. This alteration of the translation apparatus is correlated with the dephosphorylation of the cap-binding translation factor eIF4E 
. The results presented here indicate that silencing Rae1 expression has little if any effect on this process (), suggesting that other molecular targets are involved.
In summary, the data presented here have addressed the role of host Rae1 in the M protein-mediated inhibition of host gene expression at three different levels. The data support a model in which Rae1 serves as a platform for interaction of M protein with other molecular targets. Our findings also lead us to propose a new function for Rae1 in regulating transcription in VSV- infected cells, as well as providing new insights into the mechanism of VSV-mediated inhibition of host gene expression.