brief communications
A Ran-independent pathway for export
of spliced mRNA
K. Nicole Clouse*†, Ming-juan Luo*†, Zhaolan Zhou* and Robin Reed*‡
*Department of Cell Biology, Harvard Medical School, 240 Longwood Avenue, Boston, Massachusetts 02115, USA
†These authors contributed equally to this work
‡e-mail: [email protected]
everal factors involved in mRNA export from the nucleus in
Saccharomyces cerevisiae have been identified, including
Mex67p10 and Yra1p11. The metazoan counterparts of these
proteins, TAP and Aly, have been implicated in, or directly shown
to have roles in, cellular mRNA export12–14. Of particular interest
has been Mex67p/TAP which crosslinks to poly(A)+ RNA, can be
detected at nuclear pores in vivo and interacts directly with nuclear
pore components10,15. These data suggest that Mex67p/TAP may
itself be an mRNA-export receptor10,15–17. Because Mex67p/TAP and
the other mRNA-export factors do not bear any similarity to
importin-β family members, which all contain a conserved
RanGTP-binding domain, it has been further suggested10,15,18–20 that
mRNA export does not follow the RanGTP–cargo–importin-β
family member paradigm. However, several studies in both yeast
and metazoans have suggested a role for RanGTP and/or an
importin-β family member in mRNA export (see, for example, refs
15, 21–23 and see refs 1–3 for reviews). In some of these cases, however, it has not been possible to determine whether RanGTP is
involved directly or indirectly. In the light of these observations, we
sought to test whether RanGTP is directly required for mRNA
export in metazoans.
mRNAs derived from genes that naturally contain introns are
transported to the cytoplasm by a splicing-dependent export pathway24. Recent studies indicate that the link between splicing and
export is due to specific recruitment of the export machinery (Aly
and TAP) to the mRNA during splicing14. To determine whether
RanGTP is required for splicing-dependent mRNA export, we used
the approach previously used to show that RanGTP is required for
the tRNA and snRNA export pathways9. Specifically, recombinant
RanGAP (Fig. 1a), which is normally cytoplasmic5, was injected
into Xenopus oocyte nuclei together with a mixture of adenovirus
major late (AdML) pre-mRNA, U1 snRNA and pre-tRNASer. U6
S
snRNA, which is retained in the nucleus, was included as a control
for accuracy of injection (Fig. 1b, input). Oocytes were incubated at
18 °C for 3 h, and total RNA from the nucleus or cytoplasm was
fractionated on a denaturing gel (Fig. 1b, lanes N and C, respectively). In both the absence and presence of RanGAP, the premRNA and pre-tRNASer were processed in vivo to generate mature
mRNA and tRNA, respectively. In the absence of RanGAP (Fig. 1b,
– GAP), export of U1 snRNA, mature tRNA and spliced mRNA was
observed whereas the intron and U6 snRNA were retained in the
nucleus. In the presence of RanGAP, both U1 snRNA and tRNA
export were severely inhibited9 whereas mRNA export was not (Fig.
1b, c). The same results were obtained using another spliced
mRNA, fushi tarazu (ftz) (data not shown, and see below). These
data suggest that export of spliced mRNA either requires a very low
level of RanGTP that cannot be eliminated by excess RanGAP, or
t
b
–
In
Ran
Mr (K) GAP
pu
a
GAP
Pre-mRNA
102
*
78
mRNA
U1 snRNA
49
Intron
Pre-tRNA
U6 snRNA
34
28
tRNA
N C
c
100
–
–
N C
–
N C
N C
GAP
80
Export (%)
All major nuclear export pathways so far examined follow
a general paradigm1–3. Specifically, a complex is formed
in the nucleus, containing the export cargo, a member of
the importin-β family of transporters and RanGTP. This
complex is translocated across the nuclear pore to the
cytoplasm, where hydrolysis of the GTP on Ran is stimulated by the GTPase-activating protein RanGAP4,5. The
activity of RanGAP is increased by RanBP1, which also
promotes disassembly of RanGTP–cargo–transporter
complexes6,7. Here we investigate the role of RanGTP in
the export of mRNAs generated by splicing. We show that
nuclear injection of a Ran mutant (RanT24N)8 or the normally cytoplasmic RanGAP potently inhibits the export of
both tRNA and U1 snRNA9, but not of spliced mRNAs.
Moreover, nuclear injection of RanGAP together with
RanBP1 blocks tRNA export but does not affect mRNA
export. These and other data indicate that export of
spliced mRNA is the first major cellular transport pathway
that is independent of the export co-factor Ran.
60
40
20
0
mRNA
U1
snRNA
tRNA
Figure 1 Export of tRNA and U1 snRNA, but not spliced mRNA, is blocked by nuclear
injection of RanGAP. a, Coomassie-stained SDS-polyacrylamide gel containing purified
recombinant GST–RanGAP. b, AdML pre-mRNA, U1 snRNA, U6 snRNA, and pre-tRNASer
were injected into oocytes and incubated at 18 °C for 3 h in the absence of RanGAP or
with increasing amounts of RanGAP (1.5, 3, and 7.5 µM). Input, nuclear (N) and cytoplasmic (C) lanes are indicated. The asterisk (*) indicates a breakdown product of the
pre-mRNA. c, Per cent export of AdML mRNA, U1 snRNA, and tRNA is shown.
NATURE CELL BIOLOGY VOL 3 JANUARY 2001 http://cellbio.nature.com
© 2000 Macmillan Magazines Ltd
97
ut
brief communications
Mr (K)
In
p
b
Ra
nQ
Ra 69L
nT
24
N
a
Control
2
3
1
+RanQ69L
1
2
3
1
+RanT24N
2
3
Time (h)
Pre-mRNA
mRNA
111
73
Intron
48
U1 snRNA
34
29
Pre-tRNA
U6 snRNA
tRNA
N C N C N C
N C N C N C
N C N C N C
c
Export (%)
80
60
Control
+ RanQ69L
+ RanT24N
40
20
0
1 2 3
mRNA
1 2 3
U1 snRNA
1
2 3 Time (h)
tRNA
b
–
–
+
–
+
+
GAP
BP1
102
78
100
Pre-mRNA
*
mRNA
49
Intron
Pre-tRNA
34
c
Export (%)
Mr (K)
Ran
BP1
Inp
a
ut
Figure 2 Kinetics of export of tRNA, U1 snRNA and spliced mRNA after nuclear
injection of Ran mutants. a, Coomassie-stained SDS-polyacrylamide gel containing
purified recombinant Ran mutants. b, ftz pre-mRNA, U1 snRNA, U6 snRNA, and pretRNASer were injected into oocytes and incubated for 1, 2 and 3 h in the presence
– ++
– –+
– + + GAP
– – + BP1
80
60
40
20
U6 snRNA
0
28
tRNA
tRNA
mRNA
N C N C N C
Figure 3 Export of spliced mRNA is not affected by the combination of RanGAP and
RanBP1. a, Coomassie-stained SDS-polyacrylamide gel containing purified recombinant GST–RanBP1. b, ftz pre-mRNA, U6 snRNA, and pre-tRNASer were injected into
oocytes in the absence or presence of RanGAP (0.6 µM) and RanBP1 (1.2 µM) and
incubated at 18 °C for 90 min. c, Per cent export of tRNA and ftz mRNA is shown.
that mRNA export can occur by a Ran-independent pathway.
To further investigate the possibility that RanGTP is required
for mRNA export, we analysed the kinetics of export in the presence of two different Ran mutants (Fig. 2a)9. RanT24N binds stably
to the Ran exchange factor (RCC1), thus blocking the normal production of RanGTP in the nucleus8. RanQ69L binds GTP but cannot hydrolyse it4. As shown in Fig. 2b and c, nuclear injection of
RanT24N severely inhibits both U1 snRNA and tRNA export at all
three time points9. Significantly, however, export of spliced ftz
98
of buffer alone, RanQ69L (20 µM) or RanT24N (20 µM). Input, nuclear (N) and cytoplasmic (C) lanes are indicated. c, Per cent export of ftz mRNA, U1 snRNA, and
tRNA is shown.
mRNA still occurs efficiently.
In contrast to RanT24N, RanQ69L has little effect on tRNA
export, which does not require hydrolysis of the GTP on Ran (Fig.
2b, c)9. Export of spliced mRNA also occurs efficiently in the presence of RanQ69L. In contrast, U1 snRNA export is less efficient in
the presence of RanQ69L, which is thought to be due to a requirement for GTP hydrolysis during U1 snRNA export9 (Fig. 2b, c). We
conclude that mRNA export occurs efficiently in the presence of
both RanQ69L and RanT24N, consistent with the possibility that
mRNA export does not require RanGTP.
As another approach for detecting a role for RanGTP in mRNA
export, we mislocalized both RanGAP and recombinant RanBP1
(Fig. 3a) to the nucleus. RanBP1 not only stimulates the activity of
RanGAP, but is also required for efficient disassembly of
RanGTP–export substrate–importin-β family member complexes25,26. Thus, the combination of RanBP1 and RanGAP should
potently block Ran-dependent export. As shown in Fig. 3b and c, a
low level of RanGAP does not significantly affect tRNA export,
whereas this level of RanGAP together with RanBP1 strongly
inhibits tRNA export. In contrast, mRNA export is not affected by
either RanGAP or the combination of RanGAP and RanBP1 (Fig.
3b, c). These results support the conclusion that RanGTP is not
directly required for export of spliced mRNA.
From the studies presented above, we cannot rule out the possibility that RanGTP binds so tightly to spliced mRNAs that it cannot be depleted by RanGAP or the other proteins used. Recently, we
have isolated the spliceosome and spliced mRNP in highly purified
form from nuclear extracts and shown that these complexes are
functional in splicing and mRNA export assays14,27. Thus, we next
asked if RanGTP could be detected in tight association with these
complexes. A splicing time course was carried out and complexes
were isolated from each time point (7–90 min). Total RNA present
NATURE CELL BIOLOGY VOL 3 JANUARY 2001 http://cellbio.nature.com
© 2000 Macmillan Magazines Ltd
brief communications
Pre-mRNA
mRNA
Exon 1
c
b
7 14 21 40 90
NE
Time
7 14 21 40 90 (min)
Lariat-exon
Intron
NE
a
Time
7 14 21 40 90 (min)
Mr (K)
97
66
45
31
Aly
Ran
22
15
Anti-Aly
Anti-Ran
Figure 4 Ran is not detected in purified spliceosomes or the spliced mRNP. a, AdML
pre-mRNA was incubated under splicing conditions for the indicated times, and complexes were purified. Total RNA from an aliquot of each complex was analysed. b, c,
Western blots containing aliquots of each complex were probed with a monoclonal
antibody to Ran29 (c) or an affinity-purified antibody to Aly30 (b). NE, nuclear extract.
in the purified complexes is shown in Fig. 4a, and western analysis
of total protein in each complex is shown in Fig. 4b and c.
Consistent with previous work14, the mRNA export factor Aly is
first detected at low levels early in spliceosome assembly and
becomes abundant as spliced mRNA accumulates (40- and 90-min
complexes; Fig. 4b). In contrast, Ran is detected at high levels in the
nuclear extract, but is not detected in the spliceosome or spliced
mRNP (Fig. 4c). Low levels of TAP are also present in the spliced
mRNP14. However, in contrast to TAP, Ran is not detected in any of
the complexes even on long exposures of the blot (data not shown).
Thus, these data indicate that Ran does not associate tightly with the
spliced mRNA. We conclude that the absence of a significant effect of
RanGAP, RanGAP plus RanBP1, or RanT24N on export of spliced
mRNA is not due to highly efficient or stable recruitment of RanGTP
to the spliced mRNA. In contrast to mRNP complexes, RanGTP is
detected in stable association with exportin-1 when incubated in
nuclear extract (data not shown). Thus, as reported previously, Randependent complexes can form efficiently in nuclear extract (see refs
1–3 for reviews), but do not form on mRNA.
We have shown here that tRNA and U1 snRNA export, but not
the export of spliced mRNAs, is inhibited by RanGAP, RanT24N or
RanGAP plus RanBP1. In addition, Ran is not associated with purified spliceosomal complexes or with the spliced mRNP. Taken
together, these data support the conclusion that export of spliced
mRNA is the first major cellular export pathway that does not follow
the RanGTP–transport cargo–importin-β family member paradigm.
Further studies are needed to address the possibility that RanGTP, in
association with nuclear pore components, has a direct or indirect
role during translocation of mRNA across the nuclear pore.
Methods
Plasmids encoding AdML (adenovirus major late) and ftz (fushi tarazu) pre-mRNAs were described24.
The plasmids encoding pre-tRNASer, U1∆Sm snRNA, U6 snRNA, RanBP1, Schizosaccharomyces pombe
RanGAP and the Ran mutants (RanQ69L and RanT24N) were gifts from S. Altman, I. Mattaj, M.
Moore, G. Murphy, M. Rush and D. Görlich, respectively. AdML, ftz, tRNA, U1, and U6 plasmids were
linearized with BamHI, XhoI, AvaI, BamHI and DraI, respectively, for transcription with T7 RNA polymerase. Glutathione-S-transferase (GST) fusions with RanGAP and RanBP1, and His-tagged RanT24N
and RanQ69L were expressed in Escherichia coli and affinity purified. Microinjection of Xenopus laevis
oocytes was performed as described28. Oocytes were incubated for the time points indicated, dissected,
and the RNA in the nucleus and cytoplasm was precipitated and analysed on denaturing polyacrylamide gels. Results were quantified using a phosphorimager (Bio-Rad). Per cent export was calculated
as the fraction of RNA in the cytoplasm compared to the total amount of RNA at each time point.
Spliceosomal complexes and the spliced mRNP were assembled by incubating AdML pre-mRNA in
HeLa nuclear extract. Complexes were then isolated by gel filtration and maltose-binding proteinaffinity chromatography under functional conditions14,27.
RECEIVED 23 JUNE 2000; REVISED 30 AUGUST 2000; ACCEPTED 20 SEPTEMBER 2000;
PUBLISHED 8 DECEMBER 2000.
1.
2.
3.
4.
Mattaj, I. W. & Englmeier, L. Annu. Rev. Biochem. 67, 265–306 (1998).
Stutz, F. & Rosbash, M. Genes Dev. 12, 3303–3319 (1998).
Görlich, D. & Kutay, U. Annu. Rev. Cell Dev. Biol. 15, 607–660 (1999).
Bischoff, F. R., Klebe, C., Kretschmer, J., Wittinghofer, A. & Ponstingl, H. Proc. Natl Acad. Sci. USA
91, 2587–2591 (1994).
5. Hopper, A. K., Traglia, H. M. & Dunst, R. W. J. Cell Biol. 111, 309–321 (1990).
6. Richards, S. A., Lounsbury, K. M., Carey, K. L. & Macara, I. G. J. Cell Biol. 134, 1157–1168 (1996).
7. Bischoff, F. R., Krebber, H., Smirnova, E., Dong, W. & Ponstingl, H. EMBO J. 14, 705–715 (1995).
8. Klebe, C., Bischoff, F. R., Ponstingl, H. and Wittinghofer, A. Biochemistry 34, 639–647 (1995).
9. Izaurralde, E., Kutay, U., von Kobbe, C., Mattaj, I. W. & Görlich, D. EMBO J. 16, 6535–6547 (1997).
10. Katahira, J. et al. EMBO J. 18, 2593–2609 (1999).
11. Strasser, K. & Hurt, E. EMBO J. 19, 410–420 (2000).
12. Stutz, F. et al. RNA 6, 638–650 (2000).
13. Gruter, P. et al. Mol. Cell 1, 649–659 (1998).
14. Zhou, Z., Luo, M. J., Straesser, K., Katahlra, J., Hurt, E. & Reed, R. Nature 407, 401–405 (2000).
15. Bachi, A. et al. RNA 6, 136–158 (2000).
16. Santos-Rosa, H. et al. Mol. Cell. Biol. 18, 6826–6838 (1998).
17. Kang, Y. & Cullen, B. R. Genes Dev. 13, 1126–1139 (1999).
18. Wente, S. R. Science 288, 1374–1377 (2000).
19. Cullen, B. R. Mol. Cell. Biol. 20, 4181–4187 (2000).
20. Cole, C. N. Nature Cell Biol. 2, E55–E58 (2000).
21. Stade, K., Ford, C. S., Guthrie, C. & Weis, K. Cell 90, 1041–1050 (1997).
22. Pasquinelli, A. E., Powers, M. A., Lund, E., Forbes, D. & Dahlberg, J. E. Proc. Natl Acad. Sci. USA 94,
14394–14399 (1997).
23. Schlenstedt, G., Saavedra, C., Loeb, J. D., Cole, C. N. & Silver, P. A. Proc. Natl Acad. Sci. USA 92,
225–229 (1995).
24. Luo, M. & Reed, R. Proc. Natl Acad. Sci. USA 96, 14937–14942 (1999).
25. Kutay, U., Bischoff, F. R., Kostka, S., Kraft, R. & Görlich, D. Cell 90, 1061–1071 (1997).
26. Kutay, U. et al. Mol. Cell 1, 359–369 (1998).
27. Das, R., Zhou, Z. & Reed, R. Mol. Cell 5, 779–787 (2000).
28. Hamm, J. & Mattaj, I. W. Cell 63, 109–118 (1990).
29. Hieda, M. et al. J. Cell Biol. 144, 645–655 (1999).
30. Wichmann, I., Garcia-Lozano, J. R., Respaldiza, N., Gonzalez-Escribano, M. F. & Nunez-Roldan, A.
Hum. Immunol. 60, 57–62 (1999).
ACKNOWLEDGEMENTS
We thank E. Hurt, H. Grosshans, K. Magni and R. Das for useful discussions and critical comments on
the manuscript. We thank E. Hurt, I. Mattaj, D. Görlich, and Y. Yoneda for reagents. This work was
supported by an NIH grant to R.R.
Correspondence and requests for materials should be addressed to R.R.
NATURE CELL BIOLOGY VOL 3 JANUARY 2001 http://cellbio.nature.com
© 2000 Macmillan Magazines Ltd
99