distribution - NCBI

12 downloads 0 Views 5MB Size Report
Oct 10, 1994 - A synthetic lethal mutant slv21 was isolated, which exhibited a ts phenotype and showed nuclear accumulation of poly(A)+ RNA at 37°C. The ...
The EMBO Journal vol.13 no.24 pp.6062-6075, 1994

A novel nuclear pore protein Nup133p with distinct roles in poly(A)+ RNA transport and nuclear pore distribution Valerie Doye, Roger Wepf and Eduard C. Hurt1 EMBL, Meyerhofstrasse 1, Postfach 1022.09, D-69117 Heidelberg, Germany ICorresponding author

Communicated by E.C.Hurt

Temperature-sensitive nucleoporin nup49-316 mutant cells accumulate poly(A)+ RNA inside the nucleus when shifted to restrictive temperature. We performed a synthetic lethal screen with this mutant allele to identify further components of the mRNA export machinery. A synthetic lethal mutant slv21 was isolated, which exhibited a ts phenotype and showed nuclear accumulation of poly(A)+ RNA at 37°C. The wild-type gene complementing slv21 was cloned and sequenced. It encodes a novel protein Nupl33p which is located at the nuclear pore complex. NUPJ33 is not an essential gene, but cells in which NUP133 is disrupted grow slowly at permissive temperatures and stop growing at 37°C. Concomitant with the growth inhibition, nup133- cells accumulate poly(A)+ RNA inside the nucleus whereas nuclear import of a karyophilic reporter protein is not altered. Strikingly, nupl33cells display extensive clustering of nuclear pore complexes at a few sites on the nuclear envelope. However, the nuclear pore clustering phenotype and intranuclear accumulation of poly(A)+ RNA are not obligatorily linked, since an amino-terminally truncated Nupl33p allows normal poly(A)+ RNA export, but does not complement the clustering phenotype of nupl33- cells. Key words: nuclear envelope/nuclear pore complex/ nucleocytoplasmic transport/RNA export/yeast

Introduction In eukaryotic cells, the nuclear envelope acts as a barrier between the nucleus and the cytoplasm. Bidirectional traffic of molecules between these two compartments can, however, occur through the nuclear pore complexes (NPCs) which are elaborate structures embedded in the nuclear envelope. While small molecules can diffuse through the nuclear pore, import of proteins and snRNPs into, and export of different classes of RNA out of, the nucleus are energy-dependent and saturable processes that require specific signals on the transported substrate (for reviews see Goldfarb and Michaud, 1991; Forbes, 1992; Bossie and Silver, 1992; Izaurralde and Mattaj, 1992; Fabre and Hurt, 1994). As revealed by numerous ultrastructural studies, the NPCs are composed of a cytoplasmic and a nuclear ring with octagonal symmetry, eight spokes and a central structure refered to as plug or transporter

(reviewed in Pante and Aebi, 1993, 1994). Filaments protrude from the rings towards the cytoplasm and the nucleoplasm. On the nucleoplasmic side, these filaments end in a distal ring to form a basket-like structure which, in amphibian oocytes, is connected to a nuclear envelope lattice (Goldberg and Allen, 1992). In recent years, an increasing number of NPC components (nucleoporins) have been identified in several organisms. However, the cloned nucleoporins only account for 10-20% of the estimated mass of the NPCs that may comprise as many as 100 different polypeptides (Reichelt et al., 1990; Rout and Blobel, 1993). The first two nucleoporins identified in Saccharomyces cerevisiae, Nuplp (Davis and Fink, 1990) and Nsplp (Nehrbass et al., 1990) share numerous degenerate 'FSFG' peptides meanwhile found in several other nucleoporins from higher eukaryotes (Starr et al., 1990; Carmo-Fonseca et al., 1991; Hallberg et al., 1993; Sukegawa and Blobel, 1993; Kraemer et al., 1994; McMorrow et al., 1994) and yeast (Loeb et al., 1993). Members of a second nucleoporin subclass exhibit the repeat motif GLFG and have been so far characterized only in yeast, although a monoclonal antibody against mammalian nuclear pore proteins was used to clone some of these yeast nucleoporins (Wente et al., 1992; Wente and Blobel, 1994). Three of these GLFG nucleoporins, Nup49p, Nup Il6p (previously designated Nsp49p and Nsp l 16p) and Nup l45p were independently identified in a genetic screen for synthetic lethal mutants of Nsplp (Wimmer et al., 1992; Fabre et al., 1994). Despite the fact that they are highly conserved among nuclear pore proteins, the repetitive domains are not essential in the protein tested (Nehrbass et al., 1990; Fabre et al., 1994). On the other hand, this type of repeat sequence is not an obligatory feature of NPC proteins since nuclear pore proteins lacking any obvious repeat motifs have been found in higher eukaryotes (Greber et al., 1990; Wozniak and Blobel, 1992; Radu et al., 1993) as well as yeast (Grandi et al., 1993; Wozniak et al., 1994). Due to the recent isolation of highly purified yeast NPCs (Rout and Blobel, 1993), the development of genetic screens, and the yeast genome project, most yeast nucleoporins will probably be cloned and sequenced within a short time. Despite this increasing number of cloned nucleoporins, their precise function remains unknown. The transmembrane proteins gp2 10, POM 121 and POM 152 are supposed to anchor the NPCs within the nuclear envelope (Greber et al., 1990; Wozniak and Blobel, 1992; Hallberg et al., 1993; Wozniak et al., 1994), whereas the zinc finger motif-containing Nupl53p (Sukegawa and Blobel, 1993), which is localized at the terminal ring of the nuclear basket (Pante and Aebi, 1993), was proposed to link the nuclear pores to chromatin. p62 and its yeast homologue, Nsplp, were proposed to play a role in nuclear pore

6 Oxford University Press 6062

Poly(A)+ RNA distribution and NPC clustering biogenesis (Dabauvalle et al., 1990; Mutvei et al., 1992). However, these nucleoporins were also shown to be involved in nuclear import of NLS-containing reporter proteins both in vivo and in vitro (Finlay et al., 1991; Nehrbass et al., 1993; Schlenstedt et al., 1993). On the other hand, nupi 16 null mutants shifted to 37°C, as well as Nupl45p-depleted cells, accumulate poly(A)+ RNA inside the nucleus (Wente and Blobel, 1993; Fabre et al., 1994), but the mRNA export defect of the nup 116 mutant was proposed to be an indirect consequence of membrane sealing over the NPCs (Wente and Blobel, 1993). Although export of different RNA species may involve a common pathway of exit through the NPC, earlier steps in this process appear to be mediated by factors specific for each class of RNA (Jarmolowski et al., 1994). Indeed, several specific factors possibly involved in the export of 5S rRNA (Guddat et al., 1990), tRNA (Singh and Green, 1993; Zasloff, 1983), or polymerase II transcripts (Izaurralde et al., 1992; Pifiol-Roma and Dreyfuss, 1992) have been characterized in higher eukaryotes. To identify components involved in mRNA export in yeast, collections of temperature-sensitive mutants were screened for retention of poly(A)+ RNA inside the nucleus at the nonpermissive temperature. These screens yielded the mtr (mRNA transport; Kadowaki et al., 1992) and the rat (ribonucleic acid trafficking; Amberg et al., 1992) mutants. The MTRI gene (identical to PRP20 and SRMI) encodes the S.cerevisiae homologue of the guanine nucleotide release protein (GNRP) that interacts with two homologues of the small GTPase Ran in S.cerevisiae, and was proposed to be directly involved in mRNA export (Amberg et al., 1993; Kadowaki et al., 1993). Interestingly, the RATIO gene encodes the nucleoporin Nup 145p (A.Goldstein, T.C.Dockendorff and C.N.Cole, personal communication) that was independently shown to affect mRNA export (Fabre et al., 1994). In this study, we analysed temperature-sensitive nup49 strains for defects in nucleocytoplasmic transport. One of these mutant strains, nup49-316, accumulates poly(A)+ RNA inside the nucleus at restrictive temperature. As an alternative approach to identify components of the RNA export machinery, we set up a synthetic lethal screen based on this nup49 mutant allele. This genetic screen allowed the cloning and characterization of a novel nucleoporin, Nupl33p. Interestingly, mutations of Nupl33p lead to nuclear accumulation of poly(A)+ RNA and/or altered nuclear pore distribution.

Results Two mutant alleles of the nucleoporin Nup49p differently affect nuclear import of proteins and poly(A)+ RNA distribution Nup49p was previously shown to be a member of a nuclear pore subcomplex that contains Nsplp, Nic96p, Nup57p and a p80 component (Grandi et al., 1993, 1995). To analyse the role of Nup49p within this nucleoporin complex, conditionally lethal nup49 alleles were isolated and tested for defects in nucleocytoplasmic transport. Several thermo-sensitive alleles of NUP49 were obtained by PCR-directed mutagenesis on the carboxy-terminal domain of Nup49p. Two of these, nup49-313 and nup49316, were further characterized: nup49-313 cells stopped

-1

NIiP49

c

3WC

37CC

~**

r*

nup49-3 13 4w

nup49.316 Fig. 1. Growth properties of ts nup49 mutant strains. Equivalent amount of cells from wild-type (NUP49), nup49-313 and nup49-316 cells were spotted onto YPD plates and incubated at 24, 30 or 37°C. nup49-313 cells are already defective in growth at 30°C and are not viable at 37°C, whereas nup49-316 cells display a weaker temperaturesensitive

phenotype.

growing when shifted to 37°C, whereas nup49-316 cells were still able to grow at this temperature, although at a slow rate (Figure 1). Both nup49 mutant alleles encode proteins which exhibit more than one amino acid substitution compared with wild-type Nup49p (see Materials and methods), but no systematic attempt was made to determine which mutations contribute to the ts phenotype. In order to analyse nuclear protein import in these two temperature-sensitive nup49 mutant strains, cells were transformed with a plasmid directing the expression of the well-defined nuclear reporter protein, Mata2-lacZ (Nehrbass et al., 1993; Hall et al., 1984). The subcellular distribution of the reporter protein was examined by indirect immunofluorescence at permissive versus nonpermissive temperature. In nup49-313 cells grown at the permissive temperature (24°C), the Matca2-lacZ fusion protein was correctly localized inside the nucleus; in contrast, in nup49-313 cells shifted for 5 h to 37°C, the reporter protein was mislocalized in the cytoplasm (Figure 2). This in vivo inhibition of nuclear uptake is in accordance with in vitro data which have shown that nuclear protein import is also defective in semi-permeabilized nup49-313 cells (Schlenstedt et al., 1993). On the other hand, the Matx2 - lacZ fusion protein still accumulated in the nucleus of nup49-316 cells shifted for 5 h to the restrictive temperature (Figure 2). To complete this study, we analysed the subcellular location of poly(A)+ RNA in the two nup49 mutant strains by in situ hybridization using fluorescein isothiocyanate (FITC)-labeled oligo (dT)50. In nup49-313 and nup49-316 cells grown at the permissive temperature, poly(A)+ RNA is exported and, accordingly, gives rise to a cytoplasmic signal. nup49-316 cells shifted for 5 h to 37°C accumulated poly(A)+ RNA inside the nucleus, suggesting that the nup49-316 allele may primarily cause a defect in mRNA export (see Discussion). On the other hand, nup49-313 cells shifted for up to 8 h to 37°C showed a less pronounced poly(A)+ RNA signal inside the nucleus than did nup49316 cells (Figure 3).

A synthetic lethal screen using the nup49-316 allele identifies Nup133p as a novel nuclear pore

protein So far, synthetic lethality has been proven to be a powerful approach to find components of the nuclear pore complex 6063

V.Doye, R.Wepf and E.C.Hurt

n up49-31 3 1)NAk

ail ti-I ic /

\Noilla rs ki

K

4

24; C_

I 3 7 Ci

kl-

I nup49-316 DNI\

an ti- L .c-CZ

aNornaKllski

4 .-

_N' 24 C i5

.s

).\ 37 C

U

I Fig. 2. Nuclear import of the Mata2-LacZ reporter protein is inhibited in the temperature-sensitive nup49-313 but not nup49-316 strain. The subcellular distribution of the Mata2-LacZ reporter was analysed by indirect immunofluorescence in nup49-313 cells (top) and in nup49-316 cells (bottom) grown at 24°C and shifted for 5 h to 37°C. Cells were also stained for DNA using Hoechst 33258 and viewed by Nomarski optics.

which functionally overlap or physically interact with Nsplp (Wimmer et al., 1992; Fabre et al., 1994; Grandi et al., 1995). To identify components possibly involved in mRNA export, we screened for mutations that are lethal in a genetic background of the nup49-316 allele. For the genetic screen, the red/white colony sectoring assay based on the ade2/ade3 mutations was adopted (see also Materials and methods). The screening strain, called VD 1 RW316, was ade2/ade3, carried a disrupted nup49::TRPI gene and was complemented by plasmid-borne wild-type NUP49 and mutant nup49-316 alleles (pCH1 122-ADE3-URA3NUP49 and pUN 100-LEU2-nup49-316; see Table I). If grown at 32°C, this screening strain displayed a distinct red/white sectoring phenotype since cells can afford to lose the pCH 1122-ADE3-URA3-NUP49 plasmid and grow with pUN100-LEU2-nup49-316. After UV mutagenesis, eight non-sectoring red mutants, which no longer could lose the plasmid carrying wild-type NUP49, were

6064

obtained in the screen. These eight strains fulfilled all criteria of being synthetic lethal mutants of NUP49 (see Materials and methods). However, none of these mutants displayed allele-specificity towards the nup49-316 mutant allele, because they were also synthetic lethal with the nup49-313 mutant allele (V.Doye, unpublished data). One of these eight mutant strains, slv21, was further characterized because it exhibited a thermo-sensitive growth defect at 37°C (Figure 4) indicating that the mutation may cause thermosensitivity even in the presence of wild-type NUP49 (see below). slv2l cells also accumulated poly(A)+ RNA inside the nucleus when shifted to 37°C (data not shown). To clone the wild-type gene which caused synthetic lethality in slv2 1, the mutant strain was transformed with a yeast genomic library inserted into a single copy ARS/CEN plasmid. Four of the 5000 transformants regained a red/white sectoring phenotype. Of the re-isolated plasmids, three contained NUP49 and

Poly(A)+ \ L'IP49

RNA distribution and NPC

clustering

n1 Li p49-3 '1 O Iii._ f Li

D\E.-N

\ ma rx

24 C

ki

24 (. I

I

-)4

ii JJ

j

.

.

4

40P:Nf#

4IJ

,L..

3:7 C

W

) -Aii.s

.,7C

n u p49- 1 {i *,

;fl.,;I

24

15

4

40 Fig. 3. Temperature-sensitive nup49 mutant cells accumulate poly(A)+ RNA inside the nucleus at the restrictive temperature. Wild-type (NUP49), nup49-313 and nup49-316 cells were grown overnight at 24°C and shifted to 37°C for 8 h (NUP49 and nup49-313) or 5 h (nup49-316). Subcellular localization of poly(A)+ RNA was analysed by in situ hybridization with a FITC-labeled oligo(dT) probe. At 37°C, nup49-316 and, to a lesser extent, nup49-313 mutants, accumulate poly(A)+ RNA inside the nucleus which was stained with Hoechst 33258.

I

j

r

q

,! .f

%

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.1

Table I. Yeast strains Strain

Genotype

RS453 CH1305 VDICH VDIR VDIRW316 NUP49 nup49-313 nup49-316 slv2l s1248 WT nupl33-

Mata/a, ade2/ade2, his3/his3, leu2/Ieu2, trpl/trpl, ura3/ura3 Mata, ade2, ade3, leu2, lvs2, ura3 Mataz, ade2, his3, leu2, trpl, ura3, nup49::TRPI (pCH1 122-ADE3-URA3-NUP49) Maita, ade2, ade3, his3, leu2, Ivs2, ura3, tzup49::TRPI (pCH1 122-ADE3-URA3-NUP49) Matca, ade2, ade3, his3, leu2, IYs2, urO3, nup49::TRPI (pCH1 122-ADE3-URA3-NUP49, pUNI00-nup49-316) Mata, ade2, ade3, his3, leu2, 1vs2, ura3, nup49::TRPI (pUN90-NUP49) Mata, ade2, ade3, his3, leu2, lYs2, ura3, nup49::TRPI (pUN90-nup49-313) Mata, ade2, ade3, his3, leu2, lvs2, ura3, nup49::TRPI (pUN90-nup49-316) Mata, ade2, ade3, his3, leu2, lYs2, urca3, nup49::TRPl nupl33-1 (pCHI 122-ADE3-URA3-NUP49) [pUN90-nup49-316] Mata, ade2, ade3, leu2, trpl, ura3, nspl::HIS3, nupI33-2 (pCHI 122-ADE3-URA3-NSPI) [pRS414-TRPl-nspl's(L640S)] Mata, ade2, his3, leu2, trpl, ura3 (haploid derivative from RS453) Mata, ade2, his3, leu2, trpl, ura3, nupl33::HIS3 Mata, ade2, his3, leu2, trpl, ura3, nupl33::HIS3 (pUNIOO-LEU2-ProtA-NUP133) Mata, ade2, his3, leu2, trpl, ura3, nupl33::HIS3 (pUNIOO-LEU2-ProtA-nupl33-AN) Matn, ade2, his3, leu2, trpl, ura3, nupl33::HlS3 (pUNIOO-LEU2-ProtA-nupl33-AM)

ProtA-NUP133 ProtA-nup I33-AN ProtA-nup 133AM

carried an unrelated 8 kb genomic DNA fragment. Subsequently, a shorter 3.7 kb BamHI-BgIII restriction fragment was shown to contain the complementing activity both for the synthetic lethal and thermo-sensitive phenotype of slv2 1 (Figure 4). While this cloned DNA fragment could not complement any of the seven remaining mutants that were synthetic lethal in combination with nup49-316, it restored red/white sectoring to a mutant previously found in a synthetic lethal screen with ts tispi (s1248; H.Tekotte, unpublished data). This shows that this novel gene functionally interacts with both NUP49 and NSPI. Sequence analysis of the complementing 3.7 kb DNA restriction fragment revealed a single open reading frame of 1157 amino acids that encodes an acidic protein of 133 one

kDa, designated Nupl33p (see below). Analysis of charge and hydrophobicity distribution within the Nup l33p sequence did not reveal any specific features except two hydrophobic stretches (residues 413-429 and 641-658) and randomly distributed acidic residues. Comparison of this ORF with sequences present in the EMBL database did not reveal a striking homology to known proteins. In particular, the amino acid sequence of Nupl33p does not contain 'FSFG' or 'GLFG' repeats. However, during the course of this work, a part of the S.cerevisiae chromosome XI DNA sequence was released which contained a DNA sequence (ORF YRK402) 100% identical to the sequence of the NUP133 gene (Garcia-Cantalejo et al., 1994). 6065

V.Doye, R.Wepf and E.C.Hurt

24

24-

377C

11L11 1033

L+

LN

p

10

i

lINl I00- NLUP133 Fig. 4. The nupJ33-1 mutant allele causes synthetic lethality with nup49-316 and exhibits a temperature-sensitive growth inhibition at 37°C. The growth properties of strain VDIRW316 (nup49::TRPJ, pUN90-nup49-3 16, pCH I 122-ADE3-URA3-NUP49) carrying the wildtype NUP133 allele (NUP133+) and its derivative strain slv2l carrying the nupJ33-1 allele transformed with plasmid pUNIOO or pUNIOO-NUP133 were compared. Equivalent amount of cells were spotted either onto FOA-containing plates or onto YPD plates and incubated for 4 days at 24°C or 37°C. Both the synthetic lethal phenotype between the nupJ33-1 and nup49-316 alleles (followed by the death on FOA-containing plates) and the thermo-sensitive phenotype conferred by the nup133-1 mutant allele (followed by the death on YPD at 37°C) are complemented upon transformation by the pUNIOO-NUP133 plasmid.

To determine whether NUP133 is essential for cell growth, the NUP133 gene was disrupted in a diploid yeast strain (see Materials and methods). After sporulation of the heterozygous nupJ33::HIS31NUP133 diploid strain and tetrad analysis, four viable spores could be recovered of which the two nupl33::HIS3 containing progeny (further designated as nupl 33-) always grew slower than the NUP133+ progeny. The doubling time of nupl33cells grown in YPD at 30°C was 4.8 h, compared with 2.0 h for the wild-type strain. If nupl33- cells were shifted to 37°C they could undergo one or two divisions before cell growth stopped completely (Figure 5A). This temperature-sensitive phenotype of the nupl33- strain is fully rescued by transformation with a plasmid containing wild-type NUP133 (data not shown). In order to localize Nupl33p in the cell, the protein was tagged at its amino-terminal domain with two IgG binding domains of Staphylococcus aureus protein A, and the fusion gene was expressed in the nupl33- strain. ProtA-Nupl33p was functional since it could complement the temperature-sensitive phenotype of the nupl33strain (Figure SA) and all of its derived phenotypic defects. Analysis of a whole cell extract prepared from the ProtA-Nupl33p expressing strain by SDS-PAGE and Western blotting revealed a major band migrating at 150 kDa plus several degradation products (Figure SA). This apparent molecular weight is in good agreement with the calculated molecular mass of 133 kDa for Nupl33p and 14 kDa for the protein A moiety. The subcellular location of tagged Nup l33p was determined by indirect immunofluorescence. A punctuate staining of the nuclear envelope that was previously shown to be typical for a nuclear pore labelling was seen for ProtA-Nupl33p (Figure 5C). When double immunofluorescence staining was performed for both ProtANupl33p and known yeast nucleoporins [using the monoclonal mAb414 antibody that recognizes yeast nucleoporins including Nsplp, Nuplp, Nup2p (Afis and Blobel,

6066

1989)], both signals overlapped (data not shown). This shows that ProtA-Nupl33p co-localizes with other nuclear pore proteins at the nuclear envelope.

Intranuclear accumulation of poly(A)+ RNA and nuclear pore clustering as a consequence of the nup133 gene disruption Mutant nup49-316, as well as the synthetic lethal strain slv21 carrying the nupJ33-1 allele, display nuclear accumulation of poly(A)+ RNA. We therefore analysed poly(A)+ distribution in the nup 133- strain. When nupi 33- cells were grown at 18°C, a temperature at which no growth defect was observed as compared to wild-type cells, poly(A)+ RNA was exported normally into the cytoplasm (Figure 6A). Surprisingly, nuclear accumulation of poly(A)+ RNA was already seen when cells were grown at 24°C (Figure 6A). On the other hand, growth is also slowed down at this permissive temperature (see also Figure SA). The nuclear poly(A)+ RNA accumulation was further enhanced when cells were shifted for 5 h to 37°C (Figure 6A). This suggests that the growth defects of nupl33- cells may be due to inhibition of mRNA export. In contrast, no cytoplasmic mislocalization of the nuclear reporter protein Mata2-lacZ was observed in the nupl33 null strain, independently of whether cells were grown at 23°C or shifted for 5 h to 37°C (Figure 6B). This shows that NUP133 gene disruption does not alter the competence of nuclear pores to take up a nuclear reporter protein. It was previously reported that depletion of yeast cells of Nsp lp (Mutvei et al., 1992) and Nup ll6p (Wente and Blobel, 1993) or mutations within the NUP145 gene (Wente and Blobel, 1994; A.Goldstein, T.C.Dockendorff and C.N.Cole, personal communication), respectively, cause a decrease in the nuclear pore number, generate nuclear pores sealed with a double membrane, or lead to nuclear pore clustering. We were therefore interested to see whether the biogenesis and/or structural organization of nuclear pore complexes were abnormal in the nupl33strain. For this purpose, the immunolocalization of other nuclear pore proteins was analysed in wild-type and nupl33- cells grown at 24°C using the monoclonal mAb414 antibody (Aris and Blobel, 1989). Whereas in wild-type cells mAb414 stains the nuclear envelope in a ring-like manner, immunostaining in nupl33- cells was no longer ring-like and the nucleoporin antigens were found clustered in one or two regions of the nuclear envelope (Figure 7A). To exclude that not only certain nuclear pore antigens, but the whole nuclear pore complexes cluster in nup 133cells, nuclear pore distribution was directly analysed by freeze-fracture electron microscopy. In wild-type yeast cells, nuclear pores are normally distributed over the entire nuclear envelope with little tendency to cluster. The average pore density of NPCs was previously determined to be 10-15 pores/gm2 nuclear surface (Moor and MUhlethaler, 1963; Mutvei et al., 1992). Both the distribution and density of NPCs were similar in the ts nup49313 mutant strain grown at 24°C (Figure 7B). In contrast, NPCs extensively clustered in nupl33- cells grown at 24°C and the density could be as high as 60 pores/gim2. As a consequence of this clustering, large areas of the nuclear envelope were devoid of NPCs (Figure 7B). Since freeze-fracture electron microscopy cannot

Poly(A)+ RNA distribution and NPC clustering

A

11*1 -L

Is ,,G - %,. ,$

I

I

:.

B

ZN 1-1

.111

zia:. !",

37 C

k'`;24'LOM. wiip7r:qw

o >.

C' Pio)t-

31331) p)

-N upl

l)N'A

Fig. 5. Expression and localization of Nupl33p tagged with proteini A (A) Complementation of the ts phenotype of nupl33- strain. Cells from wild-type (WT), nupl33-, ProtA-NUP133, ProtA-nup133-AN and ProtA-nup133-AM strains were spotted onto YPD plates and incubated for 4 days at 24°C or 37°C. To follow the expression of the fusion proteins, whole cell extracts from ProtA-Nupl33, ProtA-nupl33-AN and ProtA-nupl33-AM strains were analysed by SDS-PAGE and immunoblotting using IgG-HRP conjugate. The position of the fusion proteins is indicated by a dot. (B) Amino acid sequence of the Nupl33p protein. The accession number of the NUPJ33 DNA sequence is X80066. The aminoterminal domain (residues 44-246) and middle domain (residues 592-926), deleted in ProtA-nupl33-ANp and in ProtA-nupl33-AMp, respectively, are boxed. (C) The subcellular localization of ProtA-Nupl33p in nupI33- cells was analysed by indirect immunofluorescence using rabbit antichicken IgG followed by goat anti-rabbit IgG coupled to Texas red. A punctuate nuclear envelope staining typical for nucleoporins is observed. Cells were also stained for DNA with Hoechst 33258 and viewed by Nomarski optics.

reveal structural alterations within the nuclear envelope, nupl33- cells grown at 24°C were further analysed by thin section electron microscopy. Whereas in wild-type cells NPC were randomly distributed within the nuclear envelope with no tendency for clustering (Figure 8A), the nuclear membrane of thin sectioned nup133- cells was often devoid of NPCs (in about 50% of the examined pictures; Figure 8C, left cell) or, if nuclear pores were seen, they were always clustered at one site of the nuclear envelope (Figure 8B-F). NPC clustering was normally not associated with an altered nuclear envelope structure.

However, in a few cases (5-10% of thin sectioned nuclei), aberrant nuclear envelope structures containing NPC were seen, reminiscent of grape-like NPC clusters displayed in nupl45AN mutant cells (Wente and Blobel, 1994; Figure 8C, D and F). Finally, in even rarer cases, intranuclear paired membranes containing pores were observed; these may correspond to invaginations of the nuclear envelope or to intranuclear annulate lamellae found in other eukaryotic cells (reviewed in Kessel, 1983) as well as in NUP116disrupted yeast cells (Wente and Blobel, 1993). This ultrastructural analysis of the nuclear pore clusters in 6067

V.Doye, R.Wepf and E.C.Hurt

A

ol

''14wdl

E.)

1--.

N oinars l's ki

c.m ars I,

'IS C

24 C

I%;

iK

.. 24 C

I

*7' ;- C

i,,.

C'

f.I

el

C I

I

f

.r

.-_

f

,

Th

V4

/S

fJ

:%

Fig. 6. nupl33- cells accumulate poly(A)+ RNA inside the nucleus. (A) nupl33- cells were grown overnight at 18°C or 24°C or pre-grown overnight at 24°C and then shifted for 5 h to 37°C. Subcellular localization of poly(A)+ RNA was analysed by in situ hybridization with a FITClabeled oligo(dT) probe. Nuclear accumulation of poly(A)+ RNA is already visible at 24°C, but is enhanced at 37°C. (B) Subcellular localization of the Matax2-LacZ reporter protein was analysed by indirect immunofluorescence in nupl33- cells grown at 24°C or shifted for 5 h to 37°C. At both temperatures, the reporter protein accumulates in the nucleus.

nupl33- cells also revealed that the NPC clusters do not display any strict location in respect to the bud neck or SPB (Figure 8C and D).

Intranuclear accumulation of poly(A)+ RNA and nuclear pore clustering are not linked in nup 133 mutant strains The fact that nupl33- cells grown at 18°C showed no nuclear accumulation of poly(A)+ RNA, whereas their nuclear pores were still clustered (data not shown), already indicated that these functional and structural defects are not strictly coupled. In order to determine which part of Nup l33p is required for a normal NPC distribution, deletions were made in Nup l33p, and the mutant proteins were tested for their capability to complement the phenotypic defects of the nupl33- strain (Figure 5B). Since analysis of the Nupl33p sequence revealed no specific structural domains, deletions were performed arbitrarily, exploiting convenient restrictions sites within the NUP133 gene. In one case, residues 592-926 within the middle part of Nupl33p (nupl33-AM) were deleted, thereby removing one of the two hydrophobic stretches. The second mutation was a deletion within the amino-terminal part of Nupl33p from residue 44 to 236 (nup 133-AN). To allow their detection, the truncated proteins were tagged with protein A in a similar way to wild-type Nupl33p. The two fusion proteins migrated on SDS-polyacrylamide gels at their expected molecular size and were expressed at a level similar to wild-type ProtA-Nupl33p (Figure 5A). Unlike ProtA-Nupl33p, ProtA-nupl33-AMp did not rescue the temperature-sensitive phenotype of the nup 133 strain (Figure 5A) and was no longer targeted to the nuclear envelope (Figure 9). Accordingly, it did not complement either the mRNA export defect (data not shown) or the nuclear pore clustering (Figure 9). This 6068

shows that this large deletion in the middle domain of Nup l33p leads to a non-functional protein. Thus, essential functional elements of Nupl33p, possibly involved in targeting to the NPC, are located within its middle domain. Conversely, ProtA-nupl33-ANp was able to complement the growth defect of the nupi 33- strain both at 24°C and, to a lesser extent, at 37°C (Figure 5A). Concomitantly, the defect causing nuclear accumulation of poly(A)+ RNA export was largely rescued by this mutant allele (Figure 10). When analysed for nuclear pore distribution, the ProtA-Nupl33-AN strain still showed a pattern of clustered NPCs, which means that this structural defect was not complemented by ProtA-Nupl33-ANp. However, ProtA-Nup 1 33-ANp itself was correctly targeted because it was localized in the NPC clusters labelled by mAb4l4 (Figure 9). This shows that the amino-terminal part of Nup l33p is not required for nuclear pore targeting, but is necessary for keeping the NPCs randomly distributed within the nuclear membrane. Furthermore, this experiment suggests that intranuclear accumulation of poly(A)+ RNA in nupl33- strains is not simply due to an abnormal distribution of the nuclear pore complexes.

Discussion Characterization of two nup49 mutant alleles revealed different phenotypic defects in nucleocytoplasmic transport. The stronger nup49-313 allele leads primarily to an inhibition of nuclear protein import, whereas the weaker nup49-316 allele causes accumulation of poly(A)+ RNA inside the nucleus. While these data suggest that mRNA export is inhibited in nup49-316 mutant cells, we cannot so far exclude that poly(A)+ RNA accumulation in nup49316 mutant cells could result from an impairment in mRNA synthesis, processing or stability. The allele-

Poly(A)+ RNA distribution and NPC clustering

A mAb4l4

DNA + Nornarski

B

11)-4

11

-

1

.

tAp

.IN

.s

0

....

. .. N. -\ 3'

..

I.,

-

-4

P. t. V\

-i

*1

I..

e,SI "as,

f.

f

k.

oft.

-1 ' "p

Fig. 7. NUP133 gene disruption induces nuclear pore clustering. (A) Distribution of nuclear pore antigens was analysed by indirect immunofluorescence using the inonoclonal antibody mAb414. In wild-type cells, a typical ring-like staining of the nuclear envelope is observed, whereas in nupl33- cells grown to early logarithmic phase at 24°C, the anti-nucleoporin immunostaining is concentrated in one or two foci at the nuclear periphery. The coincident Nomarski optics view is also shown. (B) Nuclear pore distribution was analysed by freeze-fracture electron microscopy in nup49-313 (left panel) and nupl33- cells (right panels) grown to early logarithmic phase at 24°C. In nup49-313 mutant cells, nuclear pores are randomly distributed over the entire nuclear envelope. In contrast, nuclear pores are clustered in nupl33- cells and accordingly, extensive areas of the nuclear envelope are devoid of NPC.

6069

V.Doye, R.Wepf and E.C.Hurt

B

A 44.

r

Ai

p

n

n

A

C n

n

E

D I

s

n

F

Fig. 8. Thin-section electron-microscopic analysis of nupl33- cells. Wild-type (A) and nupI33- (B-F) cells were grown at 24°C to early logarithmic phase before being processed for thin section EM. Arrowheads indicate nuclear pore complexes and big arrows point to nuclear pore clusters associated with abnormal nuclear envelope structure. (n): nucleus; (s): spindle pole body. Bars: (A-E) 0.5 ,um, (F) 0.25 gm.

specific phenotypic defects observed in the nup49 mutants could suggest that Nup49p functionally overlaps or physically interacts with distinct nucleocytoplasmic transport machineries. Alternatively, since Nup49p forms a NPC subcomplex together with Nsplp, Nic96p and Nup57p (Grandi et al., 1993, 1995), the various members of the complex may be involved in different nucleocytoplasmic transport processes. Mutations in one component (e.g. Nup49p) could then inhibit the organization of the subcomplex and thus indirectly interfere with more than one 6070

transport pathway. Functional characterization of the other nucleoporins (Nsplp, Nic96p and Nup57p) that build up this complex will be required to help understand more precisely the role of Nup49p within this complex. Through a synthetic lethal screen, based on the nup49316 mutant allele, a novel nuclear pore protein, Nupl33p was identified. Unlike most of the yeast nucleoporins cloned so far, Nupl33p does not contain characteristic FSFG/GLFG repeat sequences or other known sequence motifs. Since both the nupJ33-1 allele carried by the slv2l

Poly(A)+ RNA distribution and NPC clustering

I

.. ;-

1p 33-`4 i1

a nti-ProtA

nAb4l4

DNA

+

\ ornarski N

Fig. 9. Unlike ProtA-nupl33-AMp, ProtA-nupl33-ANp is localized within nuclear pore clusters. ProtA-nupl33-AM and ProtA-nupl33-AN strains were double labelled with rabbit anti-chicken IgG to detect the protein A moiety of the fusion proteins and with anti-nucleoporin mAb4l4 antibody. ProtA-nupl33-AMp is mislocalized in the cells, whereas ProtA-nupl33-ANp co-localizes with the mAb414-reactive nucleoporins within the NPC clusters. Cells were also stained with Hoechst 33258 and viewed by Nomarski optics.

strain, and the disrupted nupI33::HIS3 allele, lead to poly(A)+ RNA accumulation inside the nucleus at the restrictive temperature, Nup 133p may be required for efficient mRNA export into the cytoplasm (see above). We do not yet know what finally causes synthetic lethality when nupJ33-1 and nup49-316 alleles are combined, but an attractive explanation could be that the synergistic inhibition of RNA export finally kills the cells. The same may be true for the nupl33-I/nup49-313 combination, since the nup49-313 strain, although primarily defective in protein import, also displays a moderate intranuclear accumulation of poly(A)+ RNA. On the other hand, synthetic lethality may be more complex than anticipated, since mutant nspl alleles which do not reveal an apparent mRNA export defect are also synthetic lethal with a nup133 mutant allele. Conversely, the ProtA-nupI33AN mutant allele that complements the poly(A)+ RNA accumulation defect of nupI33- cells still leads to, a synthetic lethal phenotype when combined with nup49 mutant alleles (data not shown). Therefore, we propose that Nup133p and the Nsplp/Nup49p-containing NPC

subcomplex (Grandi et al., 1993) perform an overlapping function as part of the whole nuclear pore complex. In addition to the accumulation of poly(A)+ RNA, nupl33- cells show a striking defect in nuclear pore distribution, i.e. the nuclear pores are no longer randomly distributed all over the nuclear envelope, but become clustered in one or a few foci. Normal nuclear pore distribution appears to require at least the amino-terminal part of Nupl33p since ProtA-nupl33-ANp is still targeted to the nuclear pores, but does not restore normal NPC distribution. The co-localization of ProtA-nup 133-ANp with other nucleoporins at the clustered nuclear pore complexes further proves that Nupl33p is physically linked to these structures. As a welcome incidental result, this phenotype of NPC clustering in the nupl33- strain can be exploited as a tool to localize any putative nuclear pore protein by indirect immunofluorescence, given that antibodies or a tagged form of the protein are available. Alteration of nuclear pore distribution was previously reported for yeast cells over-expressing HMG-CoA reductase, in which most of the nuclear envelope is 6071

V.Doye, R.Wepf and E.C,Hurt

Nomarski

Fig. 10. Expression of ProlA-nup133-4Np complements the defect cawing nuclcar accumulation o f ply(^)+ RNA in the nup133- ?train. ~ o l y ( d ) +RNA export was analysed by in siru hybridization with a FITC-lakled oligo(dT) prohe in ProtA-nupl.73-AN cells grown at 24°C or shifted for 5 h to 37OC.The coincident Hwchst 33258 staining and Nomatiki optic$ views arc also ~ h o w n .

covered by stacks of nuclear envelope designated as karmellae. In the later case, NPC were reported to be enriched in akarmellar regions, indicating that yeast cells can survive with clustered pores (Wright et al., 1988). In wild-type yeast cells, derived from starved cultures (Moor and MiihFethaler, 1963; Severs et a!., 14761, a clustering phenotype wa. also reported. The later phenotype is, however, less drastic than in nup133- cells, since nuclear pores never reach densities higher than 18 pores/pm2 for cells in stationary phase (Severs er at., 1976), compared with a density of up to 60 poreslpm2 in nup133- cells. The regulatory mechanism that causes NPC to cluster in stationary and old cells is not known. In view of our results, we could speculate that repression of specific nuclear pore components such as Nup133p could be a possible mechanism to regulate NPC distribution upon starvation. Interestingly, nuclear pore clustering was also reported for two mRNA trafficking mutants of S-cerevhiae, rat2 and rat3 (Copeland el al., 1941), and for a nup145 mutant (Wente and Blobel, 1994). allelic to rat 10 (A-Goldstein, T.C.Dockendorffand C.N.Cole, personal communication). However, the nuclear pore clustering phenotype observed in the nup145-AN mutant, in which the amino-terminal

half of ~ ~ ~was1deletedldisrupted 4 5 ~ IWente and Blobel, 1944) seems to differ from that observed in nup133cells: (i) non-clustered NPCs. still detected in nupF45-AN cells, were not observed in nup133- cells; (ii) most of the nuclear pore clusters observed in nup133- cells were not associated with altered stnrcture of the nuclear envelope; (iii) only a few nup133- cells displayed a clustering phenotype associated with interconnected herniations of the nuclear envelope, previously described as grape-like clusters in nup145-AN cells; (iv) although rarely seen in nup 133- cells, intranuclear structures, somehow reminiscent of intranuclear lameIlae observed for NUPIIBdisrupted cells (Wente and Blobel, 1993) were not reported in nup145-AN cells. Despite the fact that several mRNA export mutants also share abnormal structure or distribution of NPCs, several lines of evidence indicate: that in the nup133 mutant strains, the defects causing nuclear pore clustering and intranuclear accumulation of poly(A)+ RNA are not obligatorily linked to each other: (i) in nup133- cells. nuclear pore clustering is seen at all temperatures even at 1 S°C, whereas nuclear accumulation of poly(A)+ RNA, cannot be detected at 18°C. but becomes progressively pronounced upon growing cells at higher temperatures:

Poly(A)+ RNA distribution and NPC clustering

(ii) an amino-terminal truncation of Nupl33p (ProtAnupl33-ANp) can partially complement both the temperature-sensitive phenotype and the defect in poly(A)+ RNA localization of the nupl33- strain, but does not restore random NPC distribution. Taken together, these data show that the nuclear accumulation of poly(A)+ RNA is linked to the temperature-sensitive growth phenotype of the various nupl33 mutant alleles. Whether this defect is specific for mRNA or reflects a more general inhibition of the export of the various classes or RNA remains to be analysed. In nupl33-cells grown at 37°C, drastic structural abnormalities of the nuclear pores, like a nuclear envelope seal over the NPCs as described in NUP116-disrupted cells (Wente and Blobel, 1993), are unlikely, since protein import is not affected. However, more subtle alterations of the nuclear pore structure occurring at the restrictive temperature could be the cause for RNA export defects. For instance, the grape-like clusters of NPC or the intranuclear lamellae-like structures observed in some nupl33- cells grown at 24°C could be correlated to the intranuclear accumulation of poly(A)+ RNA observed at this temperature in a few cells (Figure 6A). On the other hand, since the nuclear baskets were proposed to be directly implicated in active transport (Pante and Aebi, 1994), one could speculate that Nupl33p is part of, or interferes with, this dynamic structure. Alternatively, Nup133p may be more directly involved in RNA export. Unlike Nupl45p that was also shown to be involved in mRNA export (Fabre et al., 1994), the primary sequence of Nupl33p does not contain putative RNA-binding motifs. However, export of various RNA species was previously shown to involve specific proteinaceous factors (Izaurralde and Mattaj, 1992) which might dock at the nuclear pore complex. RNA export could therefore be mediated by specific interactions between RNA-binding proteins and specific receptors located at the NPC. Nupl33p could be such a NPC component thereby linking the RNA transport machinery to the nuclear pore complex. Further characterization of the slv21 strain, including the identification of the specific residue(s) mutated in the nupl33I allele and isolation of other components which functionally or physically interact with Nupl33p might facilitate the understanding of the involvement of Nup 133p in nuclear retention of poly(A)+ RNA. Since nothing is known about the mechanism directing the biogenesis and the distribution of the NPCs within the nuclear envelope, one can only speculate about the possible involvement of Nupl33p in NPC distribution. As normal nuclear pore distribution persists after removal of the nuclear envelope from purified nuclei (Aaronson and Blobel, 1974), the NPC arrangement is probably maintained by attachment to underlying structures. In higher eukaryotes, the nuclear lamina, the nuclear envelope lattice connecting the nucleoplasmic baskets, or the cytoplasmic 'pore connecting' fibrils (Goldberg and Allen, 1992, 1993) are potential structures which could provide such NPC anchoring sites. Interestingly, basket-like and inter-basket filaments attached to NPCs were described in yeast (Rout and Blobel, 1993). Nupl33p could therefore be a component of these anchoring structures or affect their possibly dynamic interaction with the NPCs. If, as suggested (R.A.Milligan, personal communication), nuclear pore

biogenesis arises from division of pre-existing NPCs, their subsequent migration and spreading into the nuclear envelope may require a functional Nupl33 protein. As an alternative hypothesis, pore clustering in nupl33- mutant cells may be the consequence of insertion of newly formed NPCs at specific sites within the nuclear envelope whereas in wild-type cells this insertion occurs randomly. Timecourse analysis of a GAL::NUP133 mutant strains in which the NPC clustering phenotype would be induced upon repression of Nup l 33p expression may allow validation of one of these models.

Material and methods Yeast strains, DNA manipulations and microbiological

techniques Yeast strains used in this study are listed in Table I. DNA manipulations including restriction analysis, fill-in reactions with Klenow fragment, ligations and PCR amplifications were essentially according to Maniatis et al. (1982). Microbiological techniques including yeast growth on minimal or YPD medium, plasmid transformation, plasmid recovery, gene disruption, sporulation of diploid cells and tetrad analysis were performed as described in Wimmer et al. (1992), except that in most cases minimal SD medium was supplemented by all amino acids and nutrients except those used for selection (CSM medium, BIO 101, La Jolla, CA).

Plasmids The following plasmids were used in this study: pUNI0O and pUN90: ARSU/CEN4 plasmid with respectively the LEU2 and the HIS3 markers (Elledge and Davis, 1988); pCH1 122: YCp5O derivative (ARS1/CEN4) with the URA3 and ADE3 markers (Kranz and Holm, 1990); pCHI 122URA3-ADE3-NUP49: the complete NUP49 gene was inserted as a 2.5 kb XmnI fragment (Wimmer et al., 1992) in the SmaI site of pCH 1122; pUN 100-LEU2-NUP49 and pUN90-HIS3-NUP49: containing a 3.5 kb genomic DNA insert encoding Nup49p; pUN100-LEU2-nup49313 and pUN 100-LEU2-nup49-316: pUN 100 containing a SacI-HindIll restriction fragment encoding the mutated nup49-313 and nup49-313 alleles; pUN90-HIS3-nup49-313 and pUN90-HIS3-nup49-316: a SaclHindIIl blunt-ended restriction fragment from plasmids pUN 100-LEU2nup49-313 and pUNI00-LEU2-nup49-316 encoding the mutated nup49 alleles was inserted into pUN90 plasmid at restriction sites Sacl-SmaI; pRS414-TRPI-ts nspl (L640S): pRS414 containing a SphI-HindIII restriction fragment encoding the mutated Nsplp carboxy-terminal domain (L640S) under the control of the ADHI promoter (Wimmer et al., 1992); pUNI0O-NUP133: containing a 3.8 kb BainHI-BglII genomic DNA insert encoding Nupl33p.

Generation of thermo-sensitive nup49 mutant alleles A DNA fragment (nucleotides 1425-1992) of NUP49 which encodes the last 128 amino acids of Nup49p (Wimmer et al., 1992) was randomly mutagenized by PCR according to Leung et al. (1989), using a 1:10 ratio between the dATP and the other dNTPs. The PCR product cut at XbaI/HindIII restriction sites (the latter being generated by PCR) was re-introduced into plasmid pUNIOO previously cut with the same restriction enzymes thereby replacing the wild-type DNA sequence 1425-1992 of NUP49 by the corresponding mutagenized sequences. This library of mutagenized nup49 was thus kept under the control of the authentic promoter. The DNA sequence analysis of randomly picked clones revealed a frequency of 2% point mutations. The shuffling strain VDIR used to identify temperature-sensitive (ts) nup49 mutant alleles was constructed as follows. Strains CH 1305 (Kranz and Holm, 1990) and VDICH carrying a disrupted chromosomal TRPJ::nup49 gene and complemented by pCHI 122-ADE3-URA3NUP49 were mated on selective SD (-his -ura) plates and resulting diploids were sporulated on YPA plates. Following tetrad analysis, a haploid strain (VD1R) was selected which was ade2 ade3 leu2 ura3 his3 TRPI::nup49 and contained the pCHI 122-ADE3-URA3-NUP49 plasmid. This shuffling strain VD1R was transformed with the mutagenized nup49 plasmid library and transformants were selected on SDC (-leu) plates. A total of 1400 colonies was streaked on SDC (-leu) plates and replica plated onto FOA-containing plates. Cells which grew on 5-fluoro-orotic acid (FOA) lost the pCHI 122-URA3-ADE3-NUP49

6073

V.Doye, R.Wepf and E.C.Hurt

plasmid and thus become dependent on the mutagenized nup49 gene. After 3 days at 23°C, colonies were double replica plated on YPD plates and incubated at either 23 or 37°C. Plasmid DNA was recovered from two putative ts mutant strains and the NUP49 sequence was determined. The following amino acid substitutions were found: nup49-313: K376E, 1390V, L391P, V3981, L449P; nup49-316: Q355L, D370G, 140(1V, 1439L, F447S, stop codon473R; +VNRPESLYSAYNKYKCLVQM. The ts alleles, nup49-313 and nup49-316 were excised from the recovered pUN100 plasmids as Sac -Hindlll fragments, blunt-ended at the HindlIl site and ligated into the HIS3-containing plasmid pUN90 previously cut with Sacl and Sinai. The corresponding pUN90-nup49313 and pUN90-nup49-3 16 plasmids were used to transform the shuffling strain VD1R. After loss of the NUP49-containing plasmid on FOA, the nup49-313 and nup49-316 expressing strains displayed a similar ts phenotype to the initially isolated alleles. Isolation of synthetic lethal mutants of NUP49 The tester strain VDIRW316 used for isolation of synthetic lethal mutants of nup49 was obtained by transforming the VDIR strain with plasmid pUN100-LEU2-nup49-316. If grown at 32°C, VDIRW316 exhibits a distinct red/white sectoring phenotype on SD (-leu +ura) plates, due to occasional loss of the ADE3-containing pCH 11 22-NUP49 plasmid during mitotic growth. For mutagenesis, VDlRW316 cells were grown in liquid SD (-ura -leu) medium to OD600 of 0.8, plated on SDC (-leu) plates, UV irradiated (X = 254 nm) for 35 s and incubated at 32°C. Of 400 000 plated cells, 55 000 survived the mutagenesis, most of them forming colonies with a clear red/white sectoring phenotype. Non-sectoring red colonies were picked and streaked on YPD plates (4% glucose). Thirteen of 42 colonies whose non-sectoring phenotype was stable in the second screen did not grow on FOA-containing plates thereby confirming that they could not afford to lose the URA3-containing pCH 1122-NUP49 plasmid. After back transformations, 8 synthetic lethal mutants (slv) remained that regained the sectoring phenotype when transformed with pUN90-HIS3-NUP49, but not with pUN90-HIS3-nup49-316.

Cloning of NUP133 The slv2l strain carrying pCH1122-NUP49 and pUN90-HIS3-nup49316 plasmids was transformed with a yeast genomic DNA library inserted into the LEU2-containing plasmid pUN 1(00 (Berges et al., 1994). Transformants were selected on SDC (-his -leu) plates. Four of 5000 transformants that were individually streaked on YPD plates (4% glucose) displayed a red/white sectoring phenotype. The plasmid DNA from these transformants was recovered. Restriction maps were made and showed that three contained a genomic insert corresponding to NUP49, but one carried a different insert of 8 kb in length. A 3.7 kb BacmHI-BglII restriction fragment was shown to be the minimal insert able to complement the non-sectoring phenotype of slv21.

DNA sequencing and gene disruption The 3.7 kb genomic insert containing NUPI33 was cut with several restriction enzymes and subfragments covering the entire length of the genomic insert were subcloned in pUN100 or pBluescript KS. DNA sequence analysis of both strands was performed according to Sanger et al. (1977) using either M13 universal and reverse primer or internal oligonucleotide primers. The DNA and deduced amino acid sequence of NUP133 were analysed by the GCG programs. The molecular weight was determined by using the program PEPTIDESORT. Amino acid sequences were compared using FASTA or BTFASTA. For disruption of the NUP133 gene, the H153 gene, isolated as a blunt-ended 1.15 kb BatmHI fragment from plasmid YDpH (Berben et al., 1991), was used to replace an internal 1.9 kb Ncol fragment of NUP133 ORF from plasmid pUNIOO-NUP133. By this manipulation, amino acids 45-708 were deleted from the NUP133 ORF. A linear nupl33::HIS3 DNA fragment was excised from the recombinant plasmid and used to transform diploid strain RS453 by selection on SDC (-his) plates. HIS+ transformants were characterized for correct integration of nupl33::HIS3 at the NUP133 gene locus by Southern analysis. Isolation of total DNA and Southern analysis was essentially as described (Sherman, 1990). RS453 diploids heterozygous for NUP133 were sporulated and tetrad analysis was performed.

Construction of ProtA- NUP133 fusion gene and detection of the fusion protein in yeast To construct the ProtA-NUP133 fusion gene, a BajmHI-EcoRI DNA fragment encoding two IgG binding domains from protein A expressed under the control of the NOPI promoter (Berges et al., 1994) was fused

6074

in frame to the coding sequence of NUP133 at an EcoRI site, generated by PCR at the ATG of NUP133, thereby removing the first methionine. The ProtA-NUP133 fusion gene inserted into plasmid pUNI0O was used to transform the nupl33- (nupl33::HIS3) haploid strain, yielding ProtA-NUP133. Protein A-tagged fusion proteins bearing internal deletions within the NUP133 ORF were generated by removing, respectively, the NcoI-Ncol fragment (nucleotides 360-960) encoding residues 44-236 (ProtA-nupl33-ANp) and the BglII-BamHI fragment (nucleotides 1980-2940) encoding residues 592-926 (ProtA-nupl33AMp). Whole yeast cell extracts and subcellular fractionation was performed as previously described (Grandi et al., 1993). Aliquots of the various fractions were analysed by SDS-PAGE and immunoblotting. The fusion proteins were detected by using IgG coupled to horseradish peroxidase (Dakopatts, Denmark) as described (Grandi et al., 1993).

Immunofluorescence To immunolocalize protein A fusion proteins in vivo, yeast cells expressing the tagged protein were fixed in 3.7% formaldehyde for 1 h, converted into spheroplasts using 0.2 mg/ml zymolyase 100.000 T, processed for immunofluorescence and analysed in the conventional fluorescence microscope as described earlier (Wimmer et al., 1992). Rabbit anti-chicken IgG (Medac, Hamburg, Germany) which binds to the protein A moiety was used as first antibody (1:50 dilution) followed by goat anti-rabbit IgG coupled to Texas Red (1:50 dilution). To stain for nucleoporins, mAb4l4 antibody (Aris and Blobel, 1989) (BabCO, Richemond, USA) was used at a dilution of 1:20, followed by FITC-conjugated goat anti-mouse IgG (Medac, Hamburg, Germany), dilution 1:100.

In vivo analysis of nucleocytoplasmic transport In vivO poly(A)+ RNA export from the nucleus was analysed by in situ hybridization for retention of poly(A)+ RNA inside the nucleus as described (Amberg et al., 1992) in strains grown at 18°C, 24°C or 37°C, except that the (dT)50 probe used was directly coupled to FITC. To analyse protein import, strains transformed with the Matoc2-lacZ fusion gene (YEpl3-Matax2(2-135)-lacZ, kindly provided by M.Hall, Basel, Switzerland) were grown at 23°C in liquid SDC (-leu) medium to OD600 nm = 0.2 before being transferred for 5 h to 23 or 37°C. Indirect immunofluorescence on yeast spheroplasts was previously described (Nehrbass et al., 1993), using as first antibody monoclonal mouse anti-f-galactosidase antibody (Promega, dilution 1:100). Secondary antibody was FITC-conjugated goat anti-mouse IgG (Medac, Hamburg, Germany, dilution 1: 100).

Freeze- fracture electron microscopic analysis of yeast cells Strains ts nup49-313 and nupl33- were grown overnight at 24°C in YPD medium to OD6(0) of 0.5. For freeze-fracture EM analysis, cells were centrifuged and washed by resuspending in distilled water at least four times. Dense yeast suspensions (2 pl) were frozen between two copper platelets, with a EM grid as spacer, in a modified propane jet freezer (Bal-Tec Balzers, FL). Freeze-fracturing and coating was performed in a BAF-400 (Bal-Tec, Balzers, FL). The sandwiches were loaded onto a double-fracture table while submersed in liquid nitrogen and transferred immediately onto the precooled (- 150°C) cold stage under a flow of dry nitrogen. The frozen sandwiches were fractured and etched at a temperature of -1 10°C and a pressure of 2x 10-7 mbar for 1-2 min followed by unidirectional shadowing with 2 nm Pt/C from an elevation angle of 450 and stabilized with a 20 nm thick carbon film. The single copper platelets with the replica and the frozen-etched yeast cells were dipped into water and the replica floated off. Adhering yeast cells were dissolved away from the floating replica by washing in frequently changed fresh 70% sulfuric acid for at least 20 h. After washing with double distilled water, the replica was taken up onto naked 400 mesh copper grids. The replicas were imaged in a Philips 400T microscope (Philips, Eindhoven, The Netherlands) at a magnification of 28 OOOX onto Kodak S0163 film plates.

Thin section electron microscopic analysis of yeast cells Thin section EM was performed essentially as described by Byers and Goetsch (1991) for vegetatively grown cells with few modifications. Briefly, cells grown to early logarithmic phase were rinsed in phosphate-magnesium buffer (40 mM K2HPO4-KH2PO4, pH 6.5, 0.5 mM MgCl,) and fixed for 30 min on ice in 1 ml phosphatemagnesium buffer containing 2% paraformaldehyde/2% glutaraldehyde. The fixed cells were washed in 0.1 M phosphate-citrate buffer, pH 5.8 and resuspended in I ml phosphate-citrate buffer plus a 1/10 dilution

Poly(A)+ of glusulase (Dupont, NEN) and 0.2 mg/ml Zymolyase 20T (Seikagaku Corp. Tokyo). After 2 h at 30°C, cells were washed in 0.1 M sodium acetate, pH 6.1. After postfixation (15 min in 2% osmium tetroxide) and en bloc staining (I h in I % uranyl acetate in the dark), the samples were dehydrated and embedded in Epon according to standard procedures. Thin sections collected on nickel grids coated with formvar, stabilized with carbon, were contrasted by staining with uranyl acetate and Renold's lead. Specimens were visualized as described above.

Gene bank accession number The NUP133 sequence will appear in the EMBL, GenBank and DDBJ Nucleotide Sequence Databases under the accession number X80066.

Acknowledgements We are grateful to Elmar Schiebel, Monika Matzner (Martinsried, Munich, Germany) and Maria Ericsson (EMBL, Heidelberg) for their expert assistance with the EM experiments. We would like to thank H.Tekotte for providing s1248 mutant strain, Dr A.Tartakoff (Cleveland, OH) for initial contribution to the analysis of poly(A)+ RNA distribution in some nup49 mutants and Dr M.Hall (Basel) for providing plasmid YEpl3::Mata2-iacZ. The critical reading of the manuscript by lain Mattaj, Spyros Georgatos, Bertrand S6raphin and various colleagues is gratefully acknowledged. V.D. was a recipient of a C.E.C. Fellowship and R.W. and E.C.H. of grants from the Deutsche Forschungsgemeinschaft.

References Aaronson,R.P. and Blobel,G. (1974) J. Cell Biol., 62, 746-754. Amberg, D.C., Goldstein,A.L. and Cole,C.N. (1992) Genes Dev , 6, 1173-1189. Amberg,D.C., Fleischmann,M., Stagljar,I., Cole,C.N. and Aebi,M. (1993) EMBO J., 12, 233-241. Aris,J.P. and Blobel,G. (1989) J. Cell Biol., 108, 2059-2067. Berben,G., Dumont,J., Gilliquet,V., Bolle,P.-A. and Hilger,F. (1991) Yeast, 7, 475-477. Berges,T., Petfalski,E., Tollervey,D. and Hurt,E.C. (1994) EMBO J., 13, 3136-3148. Bossie,M.A. and Silver, P.A. (1992) Curr Opin. Geniet. Dev:, 2, 768-774. Byers,B. and Goetsch,L. (1991) Methods Enzvmol., 194, 602-607. Carmo-Fonseca,M., Kern,H. and Hurt,E.C. (1991) Eur J. Cell Biol., 55, 17-30. Copeland,C.S., Amberg,D.C., Cole,C.N. and Snyder,M. (1991) J. Cell Biol., 115, 317a. Dabauvalle,M.-C., Loos,K. and Scheer,U. (1990) Chromlosoma, 100, 56-66. Davis,L.I. and Fink,G.R. (1990) Cell, 61, 965-978. Elledge,S.J. and Davis,R.W. (1988) Gene, 70, 303-312. Fabre,E. and Hurt,E.C. (1994) Curr Opin1. Cell Biol., 6, 335-342. Fabre,E., Boelens,W.C., Wimmer,C., Mattaj,I.W. and Hurt,E.C (1994) Cell, 78, 275-289. Finlay,D.R., Meier,E., Bradley,P., Horecka,J. and Forbes,D.J. (1991) J. Cell Biol., 114, 169-183. Forbes,D.J. (1992) Annu. ReE: Cell Biol., 8, 495-527. Garcfa-Cantalejo,J. et al. (1994) Yeast, 10, 231-245. Goldberg,M.W. and Allen,T.D. (1992) J. Cell Biol., 119, 1429-1440. Goldberg,M.W. and Allen,T.D. (1993) J. Cell Sci., 106, 261-274. Goldfarb,D.S. and Michaud,N. (1991) Trends Cell Biol., 1, 20-24. Grandi,P., Doye,V. and Hurt,E.C. (1993) EMBO J., 12, 3061-3071. Grandi,P., Schlaich,N., Tekotte,H. and Hurt,E.C. (1995) EMBOJ., in press Greber,U.F., Senior,A. and Gerace,L. (1990) EMBO J., 9, 1495-1502. Guddat,U., Bakken,A.H. and Pieler,T. (1990) Cell, 60, 619-626. Hall,M., Hereford,L. and Herskowitz,I. (1984) Cell, 36, 1057-1065. Hallberg,E., Wozniak,R.W. and Blobel,G. (1993) J. Cell Biol., 122, 513-521. Izaurralde,E. and Mattaj,I.W. (1992) Semiinl. Cell Biol.. 3, 279-288. Izaurralde,E., Stepinski,J., Darzynkiewicz,E. and Mattaj,I.W. (1992) J. Cell Biol., 118, 1287-1295. Jarmolowski,A., Boelens,W.C., Izaurralde,E. and Mattaji.W. (1994) J. Cell Biol., 124, 627-635. Kadowaki,T., Goldfarb,D., Spitz,L.M.. TartakoffA.M. and Ohno,M. (1993) EMBO J., 12, 2929-2937. KadowakiT., Zhao,Y. and Tartakoff,A.M. (1992) Proc . NatI Acad. Sci. USA, 89, 2312-2316. Kessel,R.G. (1983) Int. Rev. C!tol., 82, 181-313.

RNA distribution and NPC clustering

Kraemer,D., Wozniak,R.W., Blobel,G. and Radu,A. (1994) Proc. Natl Acad. Sci. USA, 91, 1519-1523. Kranz,J.E. and Holm,C. (1990) Proc. Natl Acad. Sci. USA, 87, 66296633. Leung,D.W., Chen,E. and Goeddel,D.V. (1989) Technique, 1, 11-15. Loeb,J.D.J., Davis,L.I. and Fink,G.R. (1993) Mol. Biol. Cell, 4, 209-222. Maniatis,T., Fritsch,E.T. and Sambrook,J. (1982) Molecular Cloning: A Laboratorv Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. McMorrow,I., Bastos, R., Horton,H. and Burke,B. (1994) Biochim. Biophvs. Acta, 1217, 219-223. Moor,H. and Muhlethaler,K. (1963) J. Cell Biol., 17, 609-627. Mutvei,A., Dihlmann,S., Herth,W. and Hurt,E.C. (1992) Eur J. Cell Biol., 59, 280-295. Nehrbass,U., Fabre,E., Dihlmann,S., Herth,W. and Hurt,E.C. (1993) Eur J. Cell Biol., 62, 1-12. Nehrbass,U., Kern,H., Mutvei,A., Horstmann,H., Marshallsay,B. and Hurt,E.C. (1990) Cell, 61, 979-989. Pant6,N. and Aebi,U. (1993) J. Cell Biol., 122, 977-984. Pante,N. and Aebi,U. (1994) Ciurr Opin. Struct. Biol., 4, 187-196. Pifiol-Roma,S. and Dreyfuss,G. (1992) Nature, 355, 730-732. Radu,A., Blobel,G. and Wozniak,R.W. (1993) J. Cell Biol., 121, 1-9. Reichelt,R., Holzenburg,E.L., Buhle,E.L., Jarnik,M., Engel,A. and Aebi,U. (1990) J. Cell Biol., 110, 883-894. Rout,M.P. and Blobel, G. (1993) J. Cell Biol., 123, 771-783. Sanger,F., Nicklen,S. and Coulson,A.R. (1977) Proc. Natl Acad. Sci. USA, 74, 5466-5467. Schlenstedt,G., Hurt,E.C., Doye,V. and Silver,P. (1993) J. Cell Biol., 123, 785-798. Severs,N.J., Jordan,E.G. and Williamson,D.H. (1976) J. Ultrastruct. Res., 54, 374-387. Sherman,F. (1990) Methods Enmvinol., 194, 3-20. Singh,R. and Green,M. (1993) Science, 259, 365-368. Starr,C.M., D'Onofrio,M., Park,M.K. and Hanover,J.A. (1990) J. Cell Biol., 110, 1861-1871. Sukegawa,J. and Blobel,G. (1993) Cell, 72, 29-38. Wente,S.R. and Blobel,G. (1993) J. Cell Biol., 123, 275-284. Wente,S.R. and Blobel,G. (1994) J. Cell Biol., 125, 955-969. Wente,S.R., Rout,M.P. and Blobel,G. (1992) J. Cell Biol., 119, 705-723. Wimmer,C., Doye,V., Grandi,P., Nehrbass,U. and Hurt,E. (1992) EMBO J., 11, 5051-5061. Wozniak,R.W. and Blobel,G. (1992) J. Cell Biol., 119, 1441-1449. Wozniak,R.W., Blobel,G., and Rout,M.P. (1994) J. Cell Biol., 125, 31-42. Wright,R., Basson,M., D'AriL. and Rine,J. (1988) J. Cell Biol., 107, 101-114. Zasloff,M. (1983) Proc Natl Acad Sci USA, 80, 6436-6440.

Received

oni July 11, 1994; revised on October 10, 1994

Note added in proof During the course of this work we learned that the rat3 is allelic to NUP133 (C.N.Cole, personal communication).

6075