Immobility, inheritance and plasticity of shape of the yeast nucleus
© Hattier et al; licensee BioMed Central Ltd. 2007
Received: 05 June 2007
Accepted: 09 November 2007
Published: 09 November 2007
Skip to main content
© Hattier et al; licensee BioMed Central Ltd. 2007
Received: 05 June 2007
Accepted: 09 November 2007
Published: 09 November 2007
Since S. cerevisiae undergoes closed mitosis, the nuclear envelope of the daughter nucleus is continuous with that of the maternal nucleus at anaphase. Nevertheless, several constitutents of the maternal nucleus are not present in the daughter nucleus. The present study aims to identify proteins which impact the shape of the yeast nucleus and to learn whether modifications of shape are passed on to the next mitotic generation. The Esc1p protein of S. cerevisiae localizes to the periphery of the nucleoplasm, can anchor chromatin, and has been implicated in targeted silencing both at telomeres and at HMR.
Upon increased Esc1p expression, cell division continues and dramatic elaborations of the nuclear envelope extend into the cytoplasm. These "esc apades" include nuclear pores and associate with the nucleolus, but exclude chromatin. Escapades are not inherited by daughter nuclei. This exclusion reflects their relative immobility, which we document in studies of prezygotes. Moreover, excess Esc1p affects the levels of multiple transcripts, not all of which originate at telomere-proximal loci. Unlike Esc1p and the colocalizing protein, Mlp1p, overexpression of selected proteins of the inner nuclear membrane is toxic.
Esc1p is the first non-membrane protein of the nuclear periphery which – like proteins of the nuclear lamina of higher eukaryotes – can modify the shape of the yeast nucleus. The elaborations of the nuclear envelope ("escapades") which appear upon induction of excess Esc1p are not inherited during mitotic growth. The lack of inheritance of such components could help sustain cell growth when parental nuclei have acquired potentially deleterious characteristics.
The position, shape, size and orientation of organelles varies among differentiated cells, thereby allowing distinct cell types to be recognized. The approximately spherical nucleus itself generally is near the center of the cell and in yeasts the spindle pole body (SPB) provides a landmark at one pole of the nucleus, thereby allowing one to witness changes of nuclear orientation during the cell cycle [1, 2]. Related to these observations is the question of whether all portions of the nuclear envelope (NE) are necessarily inherited by the daughter nucleus at mitosis – an issue which bears directly on the ability of cells to cope with the impact of damage or change accrued during a single generation. The present investigation shows that a novel landmark of the perimeter of the yeast nucleus is not inherited.
The nuclear lamina in higher eukaryotic cells governs nuclear morphology and serves as a scaffold for the organization of the nucleoplasm, where it anchors heterochromatin and can affect both DNA replication and transcription. The best-characterized proteins of the lamina are the intermediate filament lamins, which self-associate via coiled-coil domains, generating a compact meshwork at the nuclear periphery .
Alterations in the structure, organization, and composition of the nuclear lamina are likely to account for aberrant shapes of the nuclei of malignant cells, as well as the distinct nuclear morphology of neutrophils. Moreover, the increased titer of mutant lamins can have major consequences for the shape and integrity of the nuclear surface [3–6]. Lamin orthologs are absent from yeast, and it is not known whether a structural or functional equivalent of the lamina exists in S. cerevisiae. When higher eukaryotic lamin B1 or its receptor is expressed in S. cerevisiae, these proteins concentrate at the periphery of the nucleoplasm .
Several non-membrane proteins that normally localize to the periphery of the nucleoplasm in yeast possess coiled-coil domains (Esc1p, Mlp1/2p, Sir4p, Smc5/6p) [8–11]. Esc1p (Establishes Silent Chromatin), contains three equally spaced coiled-coil domains, can function as an anchor for chromatin, interacts with Rap1p and Sir4p, promotes targeted silencing at telomeres and HMR, is needed for proper organization of the "nuclear baskets" of nuclear pores, and promotes nuclear retention of unspliced transcripts [8, 12, 13]. Esc1p is larger than lamins (187 vs 70 kDa) and lacks the C-terminal isoprenylated CAAX motif present in A and B-type lamins.
The present study demonstrates that accumulation of Esc1p at the nuclear periphery causes dramatic modifications of the nuclear envelope. These modifications are structurally distinct from those which have been reported in a screen of deletion strains  or upon deletion of an ER/NE membrane phosphatase [15–17], mutation of components of the nuclear pore complex , proteins which function in the early secretory path [19, 20], or Acc1p, which reduces very long-chain fatty acids . Interestingly, the titer of Esc1p affects levels of several transcripts, only some of which originate from telomere-proximal loci. The structural abnormalities of the NE which are caused by Esc1p are not passed on to daughter cells, showing that inheritance of components of the nuclear perimeter can be selective.
Images of cells which express Nup49p-GFP (Fig. 1C) show that escapades include nuclear pores (Fig. 1C). Ultrastructural examination also detects nuclear pores in the escapades and demonstrates that escapades are double-membrane sheets (rather than tubules) of constant width (Fig. 1D). Since escapades are limited by two layers of NE it is reasonable that – in cells expressing GFP-Esc or ER membrane proteins – their fluorescent intensity can exceed that of the rest of the NE (Fig. 1A/B). Time-lapse observations show that the position and contour of most escapades remains essentially constant for tens of minutes; however the fin-like structures occasionally fuse back to the nucleus and generate rings (Fig. 1E). Examination of cells which express a tagged histone (Htb2p-mRFP) as well as GFP-Ec1p shows that escapades include little or no chromatin (Fig 1F).
To learn whether excess of other proteins of the nuclear periphery causes similar changes, we have compared the impact of excess Esc1p to that of the non-membrane protein, Mlp1p [See Additional file 1] which colocalizes with Esc1p [9, 22], the tail-anchored inner membrane protein, Prm3p  [See Additional file 2], and the integral membrane proteins of the inner membrane, Heh1p and Heh2p  [See Additional files 3 and 4]. Excess Mlp1p is known to distribute throughout the nucleoplasm . As shown, Mlp1p has no obvious effect and Prm3p causes changes of the NE roughly comparable to those caused by Esc1p. By contrast, the Heh proteins are much more perturbing, with Heh1p not even leaving the quasi-spherical shape of the chromatin mass intact. Induction of either Esc1p or Mlp1p allows continued growth, while induction of Heh1p, Heh2p, or Prm3p is toxic (not shown). We therefore have not pursued these membrane proteins further.
Since centromeres are close to the SPB during most of the cell cycle , we also studied their relation to escapades, using cells which express lac operator arrays integrated near a centromere and GFP-tagged lac repressor. No obvious association is seen between this centromere and escapades (Fig. 2D and see Additional file 7). Equivalent experiments to localize a telomere again show no obvious association (Fig. 2E and see Additional file 8). In both cases, systematic counting of through-focal series of cells which exhibit escapades shows that only ~10–12% of the tagged loci contact escapades. A functional GFP-tagged form of the telomere-associated protein, Rap1p, which binds Esc1p , also is not detected outside the chromatin mass, as would be expected if it were in escapades (Fig. 2F and see Additional file 9). This is also the case for Sir4p .
Following the separation of sister chromatids at the onset of anaphase, the nucleus of the mother cell quickly extends into the bud, generating a dumbbell-shaped structure with a narrow bridge of NE connecting the two nuclei . Fission of the bridge leaves protruding membrane remnants that are normally resorbed by the two resulting nuclei. Escapades superficially resemble these remnants and therefore might be derived from them. Nevertheless, time-lapse microscopy of cells which exit mitosis shows that GFP-Esc1p-positive remnants are efficiently resorbed, as in control cells (see below). Moreover, cells treated with α-factor (3 hr, 1 μg/ml) [See Additional file 10] or hydroxyurea (3 hr, 0.1 M) [See Additional file 11] can generate escapades when Esc1p expression is induced in the absence of cell cycle progression.
The formation of escapades does not appear to depend on the integrity of the tubulin or actin cytoskeleton, judging from experiments in which they are induced in the presence of doses of nocodazole or latrunculin A which effectively depolymerize the corresponding cytoskeletal structures [See Additional files 12 and 13].
Maternal Retention of Escapades
94.3 +/- 3.7
97.1 +/- 2.0
98.7 +/- 0.4
94.3 +/- 1.8
94.6 +/- 2.2
93.6 +/- 2.0
Escapades could be actively retained by the maternal nucleus, excluded from the daughter nucleus, or could simply not be able to diffuse in the plane of the membrane. Since septin filaments at the bud neck can restrict transfer of proteins between mother and bud [32–35], we have examined inheritance in strains that carry mutations in septin subunits and disorganize the septin collar at the restrictive temperature (cdc3-3, cdc10-1). In these mutants, escapade exclusion from daughter nuclei is again seen (Table 1). Karmellae are also restricted to the mother in wt and in septin mutants (not shown). Consistent with the suggestion that there is no structural impasse at the bud neck, most escapades which remain in the mother do not accumulate at the neck. Moreover, electron microscopic examination shows that a significant space separates the outer nuclear membrane from the inner aspect of the plasma membrane at the bud neck during anaphase .
Despite dramatic changes in nuclear shape, cells that express excess Esc1p or GFP-Esc1p are not seriously growth impaired in liquid culture or on solid media. Furthermore, they show a distribution of DNA content comparable to wt, and can mate (not shown and Figure 6). The maternal restriction of escapades may direct most physiological consequences of their presence to the mother and thereby contribute to the lack of a major growth phenotype, as each cell division produces progeny which initially lack escapades.
kb from Telomere
The morphology of organelles is intimately related to their function and deviation from normal morphology can have profound physiological consequences. It is thus plausible that NE shape alterations contribute to events which cause cellular malfunction in disease.
Nevertheless, radical changes of the composition of the nuclear periphery can be compatible with cell survival, as in a variety of yeast deletion strains lacking transmembrane proteins which localize to the NE and ER (Nem1p, Spo7p, Ssh1p), nucleoplasmic proteins (Thp1p), or nucleoporins (Seh1p) [14, 17]. None of these proteins normally are restricted to the periphery of the nucleoplasm itself.
Interestingly, the shape of the nucleus is not affected by deletion of Esc1p or deletion of two other proteins which normally localize to the periphery of the nucleoplasm, Mlp1p and Mlp2p [9, 22, 40]. Moreover, the growth of nem1-Δ, seh1-Δ, spo7-Δ, ssh1-Δ and thp1-Δ  is not obviously affected by overexpression of Esc1p (not shown).
Mutations in nuclear lamina constituents, most notably lamin A and C, cause a diverse spectrum of diseases, the laminopathies. Laminopathies caused by excess pre-lamin A at the nuclear periphery are characterized by bleb-like expansions of the nuclear surface [41–44]. The dependence of the shape of the yeast nucleus on both nuclear membrane proteins and proteins that concentrate at the periphery of the nucleoplasm is reminiscent of a distinct laminopathy (Emery-Dreyfuss Muscular Dystrophy), which can result from either mutation of the inner nuclear membrane protein, emerin, or mutation of lamin A [45, 46]. Moreover, overexpression of a GFP-tagged form of the inner membrane protein, Prm3p, distorts the shape of the NE in much the same fashion as Esc1p, and excess Heh1p and Heh2p  grossly distort the NE and chromatin mass, while Mlp1p induction has no obvious impact. These differential effects may signify that the Heh proteins have a high affinity for chromatin, that Esc1p and Prm3p are more closely linked to the inner nuclear membrane per se than to chromatin, and that Mlp1p is relatively independent.
There is no reason to expect that Esc1p is fully comparable to higher eukaryotic lamins. For example, unlike lamins, tagged Esc1p (or Mlp1p) expressed from its own promoter does not completely encircle the nucleus, being absent from beneath the nucleolus [13, 40]. Moreover, the observation of rapid diffusion of GFP-Esc1p upon karyogamy shows that at least the overexpressed protein is mobile.
The transcriptional consequences of overexpressing Esc1p emphasize the importance of this protein (or escapades themselves) for gene expression. Since both negative and positive changes are seen, Esc1p appears to be a complex regulator, not only an enhancer of silencing. In this regard, it resembles many transcriptional regulators including Rap1p . It is also notable that ~70% of the changes which we detect occur at loci which are further than 40 kb from telomeres (which concentrate at the periphery). In addition to their intrinsic interest, these microarray data provide a possible prototype against which to evaluate the transcriptional characteristics of laminopathies, which presumably account for their cell type-specific effects.
Daughter cells differ from mothers in several regards [50, 51]. Escapades are not transferred to daughter cells, apparently due to their immobility. These structures – and karmellae – are thus part of a "lagging" domain of the nuclear perimeter. The nucleolus may also be part of this domain, judging from its association with escapades and karmellae, as well as the observation that it is one of the last nuclear components to reach the bud during anaphase [52, 53]. Extra rDNA circles  and ARS plasmids  are also retained.
This study demonstrates the extreme structural plasticity of the yeast nucleus and shows that the unusual "escapades" which can be generated are essentially immobile and therefore are not inherited. Their lack of inheritance provides a striking example of the exclusion of epigenetic change. Such mechanisms – coupled with equivalent normalization of phenotype at the molecular level – are likely to sustain cell identity through mitosis. Higher eukaryotic cells may cope with such issues by extensive disassembly and reassembly of the NE during mitosis, which could provide an opportunity to avoid incorporation of structurally aberrant components.
Strains Used in this Study
Description, Relevant Genotype
SIK1/SIK1-mRFP::kanMX6 SPC42-GFP/SPC42 [pGALp-HMG1-GFP]
IAY18 × SIK1-mRFP*
MATα SPC42-mRFP [pGALp-HMG1-GFP]
MATa cdc10-1 [pMET25-GFP-ESC1]
E. Bi (#741)
MATα SPC42-mRFP::kanMX6/SPC42 GALp-GFP-ESC1::his5+/ESC1
ATY2102 × SPC42-mRFP*
W303 ESC1/GALp-GFP-ESC1:: HIS3 lacO TELXIVL his3-11-15:: HISp-GFP-lacI-HIS3 his3-11-15:: HISp-GFP-lacI-HIS3
GA1985 × ATY1550
W303 ESC1/GALp-GFP-ESC1::his5+ lacO CENIV his3-11-15::HISp-GFP-lacI-HIS3
GA1321 × ATY1550
MATα GALp-GFP-ESC1::his5+ SIK1-mRFP::kanMX6
ATY2102 × SIK1-mRFP
MATa pep3Δ ::TRP1 GALp-GFP-ESC1::his5+
ATY2101 × BJ1601
MATa cdc3-3 [pGALp-HMG1-GFP]
E. Bi (#739)
MATa cdc10-1 [pGALp-HMG1-GFP]
E. Bi (#741)
MAT α HTB2-mRFP::kanMX6
MATa cdc3-3 [pMET25-GFP-ESC1]
E. Bi (#739)
MATa W303 his3
MATa mob1-77 GALp-GFP-ESC1::his5+/ESC1
FLY30/198 × ATY1550
ESC1/GALp-GFP-ESC1::his5+ SIK1/SIK1-mRFP::kanMX6 [pMET25p-VAC8-YFP]
ATY2102 × SIK1-mRFP*
MATa ura3-52::hmg1-GFP:URA3 GALp-ESC1::kanMX6
MATa RAP1-GFP HTB2-mRFP::kanMX6
MAT α GAL-ESC1 HTB2-mRFP::kanMX6 [pNUP49-GFP]
ATY2957 × ATY2289
MATa HTB2-mRFP::kanMX6 [pGAL-GFP-HEH1-YFP] HTB2-mRFP
MATa HTB2-mRFP::kanMX6 [pGAL-GFP-HEH2-YFP] HTB2-mRFP
MATa HTB2-mRFP::kanMX6 [pMET25-GFP-PRM3]
MATa HTB2-mRFP::kanMX6 [pGAL-MLP1]
MATa GALp-GFP-ESC1::his5+ SIK1-mRFP::kanMX6
MATa GALp-ESC1 RAP1-GFP HTB2-mRFP::kanMX6
MATa HTB2-mRFP::kanMX6 SPC42-mRFP::kanMX6
MATa GALp-GFP-ESC1::his5+ HTB2-mRFP::kanMX6
Plasmids Used in this Study
R. Wright, D. Meyer
P. Lusk, G. Blobel
p MKPL2, 2 μ/TRP1
P. Lusk, G. Blobel
Cells were grown in complete synthetic medium or the appropriate dropout medium supplemented with 2% D-glucose, D-raffinose or D-galactose. Generally strains were maintained in mid-log phase in raffinose medium and were induced by addition of galactose. All incubations were at room temperature except when ts conditional strains needed to be shifted to 37°C. As required, media were supplemented with: 200 μM latrunculin A (Sigma) diluted from a 20 mM stock in DMSO, 5 μg/ml α-factor (Sigma) diluted from a 5 mg/ml stock in sterile water, 0.1 M hydroxyurea (Sigma) diluted from a 1 M stock in sterile water, 15 μg/ml nocodazole (Sigma) diluted from a 3 mg/ml stock in DMSO, or 40 μM FM4-64 (Molecular Probes) diluted from a 2 mM stock in DMSO.
GAL-GFP-Esc1p strain ATY2102 and isogenic wild-type ATY2501 were grown overnight to mid-log phase in galactose-containing synthetic medium at room temperature. Total RNA was purified from duplicate 1 ml samples by mechanical disruption with 0.2 micron glass beads in an equal volume of hot acid phenol and used for biotinylated cRNA synthesis according to Affymetrix protocols at the Case Cancer Center Gene Expression Array Facility. Replicate samples were hybridized to Affymetrix GeneChip yeast genome Y98 arrays. Scanned chip images were analyzed with the Affymetrix GeneChip Operating Software (GCOS) using the MAS 5.1 algorithm for fold-change calculations. The results were tabulated with Microsoft Access and annotated using NETAFFX (Affymetrix) and Genespring (Silicon Genetics) software packages. In the tabulation in Table 2, the average values for GFP-Esc1p-overexpressing cells are divided by the averaged data for control cells.
GAL-GFP-Esc1p strain ATY2102 and isogenic wild-type strain ATY2501 were pre-induced in galactose media for 3 hrs at 23°C, washed and examined on agarose pads in glucose medium over 5 hr, during which time fluorescent escapades remained visible. DIC images captured every 10 min allowed measurement of the timing of bud initiation for each cell in four replicate experiments. Time measurements were all relative to the emergence of the first bud by each mother cell, M1. For example, the mother's bud interval, M2-M1 measures the delay prior to the appearance of the second bud on the mother. Correspondingly, D-M1 measures the interval prior to the appearance of the first bud on the daughter.
The time difference of budding between mothers and daughters is referred to as the budding differential and is calculated as: (D-M2 = bud differential) where D is the time at which a bud appears in the daughter cell and M2 is the time at which a second bud appears in the mother cell.
Epifluorescent, phase time-lapse, and DIC microscopy were performed on a Leica DMLB microscope equipped with a SPOT camera (Diagnostic Instruments Inc.) and the SPOT Advanced software package. Confocal microscopy for live cell Z-stack and time-lapse imaging were performed on either a Zeiss LSM510 or a Leica AOBS. Volocity (Improvision) and Metamorph (Molecular Devises) software packages were utilized for image analysis and generation of time-lapse movie sequences. Other z-stack images were collected with a DeltaVision microscope and deconvolved before inspection. For quantitation of association of escapades with various structures, through-focal series were examined systematically and the distribution of nucleoli, etc. was counted in greater than 100 cells.
Vacuole membrane staining with the vital dye FM4-64 was conducted according to  with the following modifications. Briefly, log phase galactose-induced cultures were concentrated by centrifugation and resuspended at OD600 = 1 in glucose-containing medium supplemented with 40 μM FM4-64. The cells were then incubated at 30°C with shaking for 30–60 minutes, followed by extensive washing to remove excess dye. Stained vacuoles were imaged at 546 nm with a Zeiss LSM510 confocal microscope.
For time-lapse microscopy, dilute cultures were briefly centrifuged and 2 μl aliquots from the pellet were applied to the middle of 1.5% Agarose pads prepared on microscope slides in glucose- or galactose-containing culture medium, as appropriate. Samples were overlayed with a coverslip and sealed with vaseline before examination.
To learn whether escapades or karmellae are free to move in the plane of the NE, we have crossed cells which express these structures with cells of the opposite mating type and then observed the structure of the nucleus during and after nuclear fusion. For this purpose, pairs of cells grown overnight in galactose-containing medium were mixed, sedimented, and applied to an agarose pad in glucose-containing complete synthetic medium (as above). After approximately 2 hrs, they were observed.
Transmission electron microscopy was performed on samples prepared from cultures grown to mid-log phase at 30°C. Samples were fixed in 2% glutaraldehyde followed by secondary fixation/staining in 4% potassium permanganate and uranyl acetate en bloc. Samples were dehydrated, embedded in Spurr resin, stained with lead citrate and uranyl acetate and examined in a JEOL 1200CX EM.
E. Bi, G. Blobel, V. Doye, J. Drazba, E. Fabre, S. Gasser, D. Goldfarb, J. Haber, M. Hitomi, W.-K. Huh, E. Jones, J. Kilmartin, M. Lam, J. Lithgow, M. Luca, P. Lusk, A.G. Matera, D. MacDonald, D. Meyer, F. Najm, D. Narendra, T. Rapoport, M. Rout, K. Runge, D. Shore, R. Sternglanz, P. Tiedrez, E. Townsley, A. Vasanji, K. Weis, H. Worman, R. Wright, D. Wu and Y. Zhang. Confocal Microscopy Core Facility in the Comprehensive Cancer Center of CWRU/UHC (P30 CA43703-12). T.H. was supported by an NIH training grant: "Normal and Abnormal Development" T32-HD7104. A.T.'s contribution was supported by institutional resources.
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.