eEF1A Controls Ascospore Differentiation Through Elevated Accuracy, but Controls Longevity and Fruiting Body Formation Through Another Mechanism in Podospora anserina

Corresponding author: Philippe Silar, Institut de Génétique et Microbiologie de l'Université de Paris Sud, C.N.R.S. UMR 8621, Bât. 400, 91405 Orsay Cedex, France., silar{at}igmors.u-psud.fr (E-mail)

Communicating editor: S. SANDMEYER

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Antisuppressor mutations in the eEF1A gene of Podospora anserina were previously shown to impair ascospore formation, to drastically increase life span, and to permit the development of the Crippled Growth degenerative process. Here, we show that eEF1A controls ascospore formation through accuracy level maintenance. Examination of antisuppressor mutant perithecia reveals two main cytological defects, mislocalization of spindle and nuclei and nuclear death. Antisuppression levels are shown to be highly dependent upon both the mutation site and the suppressor used, precluding any correlation between antisuppression efficiency and severity of the sporulation impairment. Nevertheless, severity of ascospore differentiation defect is correlated with resistance to paromomycin. We also show that eEF1A controls fruiting body formation and longevity through a mechanism(s) different from accuracy control. In vivo, GFP tagging of the protein in a way that partly retains its function confirmed earlier cytological observation; i.e., this factor is mainly diffuse within the cytosol, but may transiently accumulate within nuclei or in defined regions of the cytoplasm. These data emphasize the fact that the translation apparatus exerts a global regulatory control over cell physiology and that eEF1A is one of the key factors involved in this monitoring.

ALTHOUGH intensively studied, the roles of the eukaryotic cytosolic translation elongation factor eEF1A (formerly EF-1) in cell physiology remain incompletely understood (see NEGRUTSKII and EL'SKAYA 1998 for a review). Like its bacterial counterpart EF1A (also called EF-Tu), this factor was first described as a G-protein that binds aminoacyl-tRNA to form an eEF1A:GTP:aa-tRNA ternary complex, which delivers the charged tRNA to the A site of the ribosome. As this task must be performed efficiently, but also accurately, it is not surprising that eEF1A controls the accuracy of the decoding process (SANDBAKEN and CULBERSTON 1988 ). Several data suggest that the bacterial and the eukaryotic proteins perform their functions differently, but the precise biochemical events during the elongation step catalyzed by eEF1A remain unknown (NEGRUTSKII and EL'SKAYA 1998 ). Moreover, the eukaryotic protein exhibits many additional properties and/or functions not displayed by the bacterial protein. For example, among others, eEF1A interacts with actin and tubulins (DURSO and CYR 1994 for a review), activates degradation of some proteins (GONEN et al. 1994 ), and is probably involved in signal transduction (YANG et al. 1993 ; KIM et al. 1999 ; CHEN et al. 2000 ) and cell cycle regulation (GANGWANI et al. 1998 ).

eEF1A is essential for cell viability (COTTRELLE et al. 1985 ; SILAR et al. 2000A ). Some organisms contain several genes, which may encode slightly different isoforms, but there is little information concerning the in vivo significance for the presence of these different isoforms. For example, in mammals, two isoforms are present: eEF1A1 is expressed in all tissues and eEF1A2 is expressed in muscle and the immune and nervous systems. Despite the presence of eEF1A1 in all tissues, eEF1A2 is essential because a mutation in the mouse eEF1A2 isoform entails the "wasted" phenotype that consists of a set of alterations affecting the nervous and immune systems (CHAMBERS et al. 1998 ).

Interestingly, eEF1A is also involved in cell differentiation, apoptosis, and aging. Indeed, overexpression of this factor may promote susceptibility to oncogenic transformation (TATSUKA et al. 1992 ) and a mutated form of the human eEF1A gene is oncogenic (GOPALKRISHNAN et al. 1999 ). However, the exact mechanism(s) involved in these phenotypes is (are) not yet established. eEF1A is rapidly upregulated when cells are induced to enter apoptosis with H₂O₂ and transfection with eEF1A antisense RNA protects against H₂O₂-mediated cytotoxicity (CHEN et al. 2000 and references therein). During aging, eEF1A activity diminishes in numerous cellular types, correlating with a decrease in translation efficiency (WEBSTER 1985 ). Therefore, SHEPHERD et al. 1989 overexpressed eEF1A in Drosophila melanogaster. They observed an increased longevity in adult males. However, subsequent studies led to contradictory conclusions (STEARNS and KAISER 1993 ; SHIKIMA et al. 1994 ; SHIKIMA and BRACK 1996 ). Nevertheless, decreased eEF1A activity with aging was confirmed recently in Drosophila (SHIKIMA and BRACK 1996 ). Once again, the mechanism by which eEF1A may increase life span remains unclear.

We have studied the roles of eEF1A in the filamentous fungus Podospora anserina through a thorough genetic analysis. This filamentous fungus is well suited for such analysis. Like yeast, it is easy to handle for molecular genetic studies. Interestingly, it presents a more complex life cycle than yeast.

As in many filamentous ascomycetes, sexual reproduction in P. anserina is accompanied by elaborate cell differentiation processes that make possible the evaluation of eEF1A roles during differentiation (ZICKLER et al. 1995 ). The vegetative haploid mycelium is able to differentiate specialized male (conidia) and female (ascogonium) gametes. When the medium is exhausted, sexual reproduction takes place. Cell fusion of gametes during fertilization is not followed by nuclear fusion, but by nuclear proliferation to yield a multinucleated cell. Two compatible nuclei then migrate in a specialized cell, the ascogenous hyphae. This cell gives rise to a succession of dikaryotic cells by a complex process of mitosis and particular septum formation. Some of the produced cells undergo karyogamy. Meiosis and a postmeiotic mitosis immediately follow. Finally, four binucleated ascospores differentiate in an ascus. All these steps take place in a specialized fruiting structure, the perithecium. When mature, a perithecium contains several hundred asci that are ejected by turgor pressure.

Moreover, the presence of various cell degenerative processes allows assessment of the influence of eEF1A during cell aging in P. anserina. Two such processes have been described in P. anserina, Senescence and Crippled Growth (see SILAR et al. 2001 for a review). Senescence, which is present in all wild-type strains, results in cell death (this permits definition of longevity as the time elapsed between germination and death) and is correlated with drastic modification of the mitochondrial DNA (mtDNA). More precisely, amplification of three particular regions of the mtDNA (called senDNA, senDNAß, and senDNA) correlates with Senescence, whereas the "young" mtDNA molecules disappear. The exact cause of Senescence, especially the relationship between mtDNA modification and Senescence at the present time is not clear (JAMET-VIERNY et al. 1999 ). Contrary to Senescence, Crippled Growth is present in a restricted set of strains, does not lead to cell death but to a severely impaired growth, and is not accompanied by mtDNA modification. It is associated with the presence of a nonconventional infectious element, called C, that is likely of an epigenetic nature (SILAR et al. 1999 ).

These differentiation and degenerative processes are strongly influenced by mutations located in genes coding for the cytosolic translation apparatus in a way that remains unclear. These mutations were selected as modifiers of the error level during the decoding process (see COPPIN-RAYNAL et al. 1988 for a review) either by increasing (suppressor mutations) or by decreasing this level (antisuppressor mutations). Several of these genes have been characterized at the molecular level and were proved to encode ribosomal proteins (DEQUARD-CHABLAT and SELLEM 1994 ; SILAR et al. 1997 ), termination factors (GAGNY and SILAR 1998 ), tRNAs (DEBUCHY and BRYGOO 1985 ), and eEF1A (SILAR and PICARD 1994 ). Evidence was provided that, in the case of the antisuppressor mutations located in the AS7 gene (a gene whose product is yet unknown), the ascospore production defect is likely due to the reduced error level (DEQUARD-CHABLAT and COPPIN-RAYNAL 1984 ). However, as other antisuppressor mutants are fertile, the relationship between errors and fertility is not yet fully understood. Strikingly, there is no correlation between longevity and translation error level (BELCOUR et al. 1991 ; SILAR et al. 1997 ; SILAR et al. 2001 for review). However, data suggest that elongation, not termination, of translation is a major step that regulates longevity (SILAR and PICARD 1994 ; GAGNY and SILAR 1998 ). Interestingly, some mutations in the translational genes also modify the mtDNA alteration pattern observed during Senescence through a mechanism that likely depends upon cytosolic translation (DEQUARD-CHABLAT and SELLEM 1994 ; SILAR et al. 1997 ). In contrast to Senescence, Crippled Growth is clearly promoted by decreased error level, but the involved mechanism is unknown (SILAR et al. 1999 ).

In P. anserina, eEF1A is encoded by a unique and essential gene called AS4 (SILAR and PICARD 1994 ). Several types of mutations were selected in AS4. (1) Deletion of the gene confirmed that the protein is essential for cell viability (SILAR et al. 2000A ). However, eEF1A was shown to be dispensable for male gamete production and fertilization. (2) Increase of AS4 gene dosage did not result in a large increase of cellular eEF1A and in any physiological modification (SILAR et al. 2000A ). (3) An in vitro mutagenesis followed by reintroduction of the mutant alleles in P. anserina allowed the recovery of informational suppressor alleles that increase readthrough (SILAR et al. 2000B ). Four among the six strains bearing suppressor mutations display wild-type vegetative and sexual characteristics. However, two mutations display very peculiar properties. One, AS4-55, is lethal but dominantly increases longevity and the other, AS4-56, exhibits a very complex set of phenotypes including a new growth arrest syndrome. (4) A classical genetic screen allowed recovery of mutations that decrease the readthrough associated with tRNA suppressors. These antisuppressor mutants were shown to display a global increase in translation accuracy (PICARD-BENNOUN 1976 ; see Table 1). Most of these mutations are lethal but a few are not. Mutants with the latter class of mutations exhibit an extended life span and impairment in ascospore production (SILAR and PICARD 1994 ). Like all antisuppressor mutants, they also propagate the nonconventional infectious element that is responsible for the Crippled Growth (SILAR et al. 1999 ). All three phenotypes are recessive (SILAR and PICARD 1994 ; SILAR et al. 1999 ).

Clearly, suppressor and antisuppressor mutations in eEF1A entail a very complex set of phenotypes (pleiotropy) related to differentiation and aging in Podospora. Our aim is to uncover how mutations in eEF1A can trigger such a wide range of perturbations. Since the AS4 antisuppressor mutations were recovered in a screen for mutants displaying increased translation accuracy, we have tried to determine the role of translation accuracy in ascospore formation and life span extension. The data reported in this article confirm that the defect of ascospore formation is due to elevated accuracy, but that longevity of extension is due to another process(es). However, analysis of additional mutants show that AS4 also controls fertility through the production of perithecia by a pathway, which is independent from error level control.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Strains, media, and genetic analysis:
The P. anserina strains used in this study were all derived from the S strain, ensuring a homogenous genetic background (RIZET 1952 ). Standard culture conditions, media, and genetic methods for this fungus have been described (ESSER 1974 ). Longevity was measured as described in SILAR and PICARD 1994 in at least nine independent cultures.

Selection of the various AS4 mutations used in this study was described by PICARD-BENNOUN (1976). Partial description of the physiological defects caused by the six nonlethal mutations, AS4-11, AS4-27, AS4-29, AS4-30, AS4-43, and AS4-44, was previously given by SILAR and PICARD 1994 . Their sequences, as well as those of the three lethal mutations described here, were reported by SILAR et al. 2000B and are reported in Table 1. Because AS4 is closely linked to the mating-type locus, AS4 mutations are available in only one mating type (either mat+ or mat-, depending on the mutation) except for AS4-43, which has been recombined with the mating-type locus and is thus available with both haplotypes. Strains carrying a deletion of AS4 and an ectopic copy of AS4-44, allowing the recovery of this allele associated with both mating types, were previously described (SILAR et al. 2000A ). These strains are designated AS4 {AS4-44}.

193 and leu1-1 are UGA nonsense mutations that affect spore pigmentation and leucine biosynthesis, respectively. They are both efficiently suppressed and can be used to measure in vivo suppression levels. Both su4-1 and su8-1 suppressors are isoacceptor serine-inserting tRNAs that suppress the opal UGA termination codon (DEBUCHY and BRYGOO 1985 ). The su1-1 mutation is an omnipotent suppressor mutation affecting the eRF3 release factor (GAGNY and SILAR 1998 ).

Paromomycin resistance:
The paromomycin (Pm) resistance level is the ratio obtained by dividing the diameter of the thalli after 3 (or 4) days of growth on M2 medium supplemented with 750 µg/ml of paromomycin by the diameter of the thalli after 3 (or 4) days of growth on nonsupplemented M2 medium.

Antisuppression analysis:
The antisuppression efficiency of the AS4 mutations was measured according to the following strategies: (1) the leu1-1 mutation is an UGA nonsense mutation preventing growth on M2 minimal medium and is suppressed by informational suppressors. The leu1-1 su strains grow on minimal medium at a suppressor-specific speed (COPPIN-RAYNAL 1981 ). The ratio of growth speed on M2 medium vs. growth speed on M2 supplemented with 50 µg/ml leucine can be used to quantify the suppression level at the leu1-1 mutation site. Diminution of this ratio in strains carrying an additional mutation is indicative of the efficiency of antisuppression of this mutation. (2) The 193 mutation is an UGA nonsense ascospore color mutation that is also suppressible (PICARD 1973 ). Wild-type ascospores are black and 193 ascospores are white. The suppression level at the 193 mutation site can be estimated by the intensity of the green pigmentation in 193 su ascospores. The reduced coloration of the ascospores observed, if a supplementary mutation is present, gives an estimate of the antisuppression efficiency of this additional mutation.

DNA analysis:
All DNA manipulation procedures were performed according to standard methods (AUSUBEL et al. 1987 ). mtDNA was extracted from senescent cultures of wild type and the six AS4 mutant strains by the rapid method of LECELLIER and SILAR 1994 . Modifications of the mitochondrial genome in at least nine independent senescent strains per genotype were analyzed. mtDNA was digested with HaeIII restriction enzyme, run on a 1% agarose gel, and blotted onto nylon membrane. Specific probes for the senDNA, senDNAß, and senDNA regions were used to analyze the various amplified regions (JAMET-VIERNY et al. 1997A ).

Construction of AS4-GFP fusion and isolation of strains expressing the chimeric protein:
A fragment consisting of the whole AS4 coding sequence and the AS4 promoter was amplified by PCR from pEF-1D (SILAR and PICARD 1994 ) using primers 4503 (5'-CTAAAGGGAACAAAAGCTG-3') and AS4-3' (5'-CGGGATCCCGTTTCTTGCCAGCCTTGGCAGC-3'). The first primer anneals upstream of the multiple cloning site, in which the AS4 gene is cloned. The second primer anneals at the end of the AS4 coding sequence and introduces a BamHI site in the place of the stop codon; two nucleotides were added in order for the GFP (green fluorescent protein) coding sequence to be in the same phase as AS4. The AS4 gene was cloned in the pEGFP-1 (CLONTECH, Palo Alto, CA) through several steps. First, the PCR product was cloned in pBC-SK (Stratagene, La Jolla, CA) at the EcoRV site. Second, this construct was digested by XhoI and BamHI and ligated with the pEFGP-1 that was linearized by the same enzymes, giving rise to plasmid pAS4-GFP. The AS4 coding sequence present in this plasmid was completely sequenced to ensure the absence of mutation.

The AS4-44 mutant strain was cotransformed with pAS4-GFP and pBC-hygro that carries a hygromycin resistance gene (SILAR 1995 ) as described in BRYGOO and DEBUCHY 1985 . Several hundreds of hygromycin-resistant transformants were recovered and checked for fluorescence. Eighty transformants that exhibited fluorescence were crossed with the AS4-43 mutant strain to look for the rescue of the ascospore formation defect. In most transformants, restoration of ascospore formation was observed, although not to the same level as in a cross between wild type and AS4-43. This suggested that in these transformants AS4 activity was partially restored and argued for a functional expression of the AS4-GFP chimeric protein. Two such transformants, F4 and F7, were subjected to genetic analysis by crosses with wild type. Data showed a complete cosegregation of hygromycin resistance, restoration of ascospore formation, and fluorescence. The two integrations F4 and F7 were then recombined with AS4⁺ to obtain the F4 AS4⁺ and F7 AS4⁺ strains associated with the two mating types. The F4 AS4⁺ mat- and F7 AS4⁺ mat- strains were then crossed with the leu1-1 AS4 mat+/leu1-1 AS4⁺ mat- heterokaryotic strain (SILAR et al. 2000A ). In the progeny, F4 AS4 mat+ and F7 AS4 mat+ strains were recovered, showing that F4 and F7 are able to complement the lethal AS4 deletion. Additional crosses between F4 AS4⁺ and F7 AS4⁺, on one hand, and leu1-1 AS4 mat-/ leu1-1 AS4⁺ mat+, on the other hand, obtained F4 AS4 mat- and F7 AS4 mat- strains.

Cytological analyses:
For immunofluorescence, asci were fixed in 7.4% paraformaldehyde at room temperature and crushed with a blunted hypodermic needle between a siliconized slide and a polylysine-coated coverslip, as described by THOMPSON-COFFE and ZICKLER (1994). Asci were incubated in anti-ß-tubulin (1:1200; Amersham France) 12 hr at room temperature and in secondary antibody (FITC anti-mouse, Caltag, San Francisco) at 37° for 45 min. Chromatin was visualized by adding 0.5 µg/ml 4',6-diamidino-2-phenylindole (DAPI) dihydrochloride to the final rinse. Coverslips were mounted on 90% gycerol, 10% 100 mM phosphate, pH 8.7, with 10% w/v 1,4-diazobicyclo(2,2,2)octane (Sigma, St. Louis). Controls included the use of primary or secondary antibodies alone. AS4-GFP localization was analyzed using the FITC Zeiss filter set. Both sets of cells were observed on a Zeiss Axioplan microscope and images were captured by a CCD Princeton camera.

For light microscopy, specimens were fixed in fresh Lu's fixative (butanol, proprionic acid, and 10% aqueous chromic acid, 9:6:2 v/v). After 10 min of hydrolysis at 70°, asci were stained in 2 drops of 2% hematoxylin mixed on the slide with 1 drop of ferric acetate solution.

UV mutagenesis:
To select for secondary mutations that rescue the AS4 ascospore formation defect, crosses between AS4-44 mat+ and AS4-29 mat- were set up on 60 petri plates. Just before spermatization, the plates were irradiated with 200 J/m² of UV 254 nm. After fertilization, 1200 perithecia/plate were obtained. Among these, 4 were able to discharge ascopores within 10 days of incubation.

Three-dimensional model:
The model for P. anserina eEF1A was made by using the SWISS-MODEL server accessible at http://www.expasy.ch/swissmod/SWISS-MODEL.html. The resulting model was analyzed with the Weblab ViewerLite from Molecular Simulations.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The mechanisms causing life span extension and ascospore formation defects were investigated in the six viable eEF1A mutants: AS4-11, AS4-27, AS4-29, AS4-30, AS4-43, and AS4-44.

Longevity extension and sporulation defects:
In P. anserina, senescence is accompanied by accumulation of specific mtDNA rearrangements (see DUJON and BELCOUR 1989 for a review). Some mutations affecting the translational apparatus modify the spectrum of the amplified rearranged mtDNA molecules (BELCOUR et al. 1991 ; SILAR et al. 1997 ). Therefore, we analyzed the status of the mitochondrial genome upon appearance of senescence in the six mutant strains. Fig 1 shows a typical result for the wild-type and AS4-30 mutant strains. Similar data were observed for the five other mutant strains. In all studied senescent cultures, no particular mtDNA rearrangements were detected. Thus, the AS4 mutants display a delayed accumulation of mitochondrial DNA alterations typical of the classical senescent state.

View larger version (140K): In this window In a new window Download PPT slide   Figure 1. mtDNA modification during senescence. A typical result for the analysis of senDNA and senDNAß in wild-type and AS4-30 senescent cultures is presented. mtDNA was extracted from three independent wild-type senescent cultures (left) or three independent AS4-30 senescent cultures (right). DNA was analyzed by Southern blot, as indicated in MATERIALS AND METHODS. Filters were hybridized with a probe made with DNA fragments originating from the senDNA and senDNAß regions. Both types of culture exhibit the same pattern of mtDNA modification during Senescence. The arrow points toward the 2.5-kb band that is characteristic for senDNA (the same band hybridized when a probe corresponding to the senDNA region is used). This senDNA is always present and exhibits the same structure in all wild-type senescent cultures. It is also always present in the AS4-30 senescent cultures. The other bands reveal the variable senDNAß (JAMET-VIERNY et al. 1997A ). This senDNA was found as frequently in wild-type and AS4-30 senescent cultures.

To learn when and how the AS4 mutants were impaired in the formation of ascospores, we crossed these strains with every compatible combination (see MATERIALS AND METHODS). Perithecia from crosses involving AS4-11 and especially AS4-30 were not completely barren and did yield a small progeny. No progeny were obtained from the crosses involving the four other mutants. Cytological analyses of perithecia from AS4-43 x AS4-43, AS4 {AS4-44} x AS4 {AS4-44}, and AS4-27 x AS4-43 crosses showed that sporulation was abnormal in several respects. (1) Sporulation in filamentous ascomycetes follows a postmeiotic mitosis (PMM) in which proper spindle alignment is crucial for correct ascospore formation (THOMPSON-COFFE and ZICKLER 1994 ). In four-spored species like P. anserina, formation of binucleate ascospores also requires extensive nuclear migration: First, meiotic spindles are regularly spaced along the ascus and the four PMM spindles are grouped in widely separated pairs across the long axis (Fig 2A). Second, paired nuclei issued from the PMM must turn and migrate along the ascus and then to the cell membrane in order to be isolated in the forming ascospore wall. In the AS4-43 x AS4-43, AS4 {AS4-44} x AS4 {AS4-44}, and AS4-27 x AS4-43 crosses, non-four-spored asci were the rule mainly due to faulty meiotic and especially PMM spindle alignment and to impaired nuclear migrations (Fig 2B). The issuing ascospores were either anucleate or polynucleate or badly shaped or aborted, and most asci contained unenclosed nuclei that remained free in the cytoplasm (Fig 2C). (2) After spore closure, wild-type nuclei divide and the rounded resting nuclei show characteristic decondensed chromatin and prominent nucleolus (Fig 2D). Mutant nuclei had the same appearance just after spore closure, but then often became highly condensed in the same time that the ascospores degenerated (Fig 2E). Sometimes all eight nuclei of an ascus condensed, but sometimes only part of them did so (Fig 2F). However, the proportion of normal vs. abnormal or degenerated asci was different from one mutant to the other. In AS4-43 x AS4-43, 10% of the asci showed a wild-type sporogenesis and, although not efficiently ejected from the perithecium, the corresponding ascospores were viable; 30% displayed abnormal spores with mostly abnormal numbers of nuclei (one or three or more, instead of two). The remainder 60% of the asci were blocked either during meiosis or during PMM with four or eight condensed nuclei. A similar but more severe phenotype was observed in AS4 {AS4-44} x AS4 {AS4-44}, and defects were even stronger in AS4-27 x AS4-43, with only rare ripe spores formed. Crosses involving AS4-29 showed phenotypes similar to those involving AS4-44. On the basis of the severity of the ascospore formation defect, the following classification of the AS4 mutant can be proposed: (1) AS4-30 and AS4-11, which display a moderate defect and are able to eject some mature ascospores; (2) AS4-43, which is able to produce mature ascospores but not able to eject them out of the perithecium; (3) AS4-44 and AS4-29, which do produce a few ripe but abnormal ascospores; and (4) AS4-27, which do not produce any mature ascospores.

View larger version (94K): In this window In a new window Download PPT slide   Figure 2. Defects during ascospore formation in AS4 mutants. (A) Wild-type PMM with correct position of the four mitotic spindles (arrows). (B) AS4 {AS4-44} x AS4 {AS4-44} PMM with abnormal positions of the mitotic spindles (arrows). (C) Ascus with abnormal ascospores due to abnormal spindle positioning in a AS4 {AS4-44} x AS4 {AS4-44} cross; the arrows point toward ascospores without nuclei. (D) Wild-type binucleate ascospores before maturation showing normal nuclear morphology. Vertical arrow indicates the nucleolus; slanting arrow indicates the chromatine. (E) Ascospores from a AS4 {AS4-44} x AS4 {AS4-44} cross before maturation showing condensed nuclei (arrow). All eight nuclei are highly condensed. (F) Same cross as E but, in this ascus, one ascospore carries two normal nuclei (arrowhead; compare with D), whereas the three others carry condensed nuclei (vertical arrow). Bar, 5 µm.

To study the involvement of the cytoskeleton in the defects observed in the three crosses, we analyzed their microtubule arrays during sporulation by antitubulin immunofluorescence. When compared to wild type, both cytoplasmic and spindle microtubules were normal (Fig 3A and Fig B). However, unlike the regularly spaced wild-type spindles, in the three mutants, the two meiotic and/or the four PMM spindles were mostly randomly oriented across the ascus (Fig 3C and Fig D).

View larger version (57K): In this window In a new window Download PPT slide   Figure 3. Microtubule org- anization in AS4 mutant crosses. (A) A typical first meiosis division metaphase of a AS4 {AS4-44} x AS4 {AS4-44} cross shows that cortical microtubules (slanting arrow) are normal as well as the spindle microtubule (vertical thick arrow indicates the position of the chromosomes). (B) DAPI staining corresponding to A. (C) AS4-43 x AS4-27 PMM anaphase showing abnormal orientations of the four spindles (indicated by arrowheads). Vertical arrow indicates the position of a spindle pole body with astral microtubules to the left and intranuclear microtubules to the right. (D) Corresponding DAPI staining of eight nuclei. Bar, 5 µm.

The sporulation phenotypes seen for the AS4 mutants resemble the AS7 mutant phenotypes, for which ascospore impairment was shown to be due to increased accuracy (DEQUARD-CHABLAT and COPPIN-RAYNAL 1984 ). However, as these authors reported only spindle defects and no nuclear condensation, we decided to perform a more detailed cytological analysis of an AS7-2 x AS7-2 cross. Like the AS4 mutants, AS7-2 showed both abnormal ascospore formation and condensed nuclei. In addition, even when normally delimited, ascospore maturation was stopped. Like AS4 mutants, AS7 mutants also exhibit an increased life span (BELCOUR et al. 1991 ; P. SILAR, unpublished results), but it was not determined whether accuracy was involved in life span extension. In the view of the similar defect in AS4 and AS7 mutants, we decided to determine whether increased accuracy was involved in ascospore formation but also whether it controls life span extension in AS4 mutants.

Ascospore formation, but not wild-type longevity, is restored in the presence of paromomycin:
Paromomycin is an aminoglycoside antibiotic known to increase the translation misreading frequency in vitro (PALMER and WILHEM 1978 ; SINGH et al. 1979 ). Most P. anserina strains carrying a suppressor mutation have an increased sensitivity whereas most P. anserina strains carrying an antisuppressor mutation have a decreased sensitivity to the antibiotic (COPPIN-RAYNAL 1981 ). The resistance to 750 µg/ml of paromomycin triggered by the six viable AS4 mutations was measured and compared with that of the wild type (Table 1). Three mutants are clearly more resistant (AS4-27, AS4-29, and AS4-44) and two are slightly more resistant (AS4-11 and AS4-43), whereas one was more sensitive (AS4-30) than wild type. This latter phenotype was also reported previously (COPPIN-RAYNAL 1981 ).

If a phenotype is caused by increased accuracy, it should be phenotypically suppressed by paromomycin. Therefore, we analyzed ascospore formation and longevity of AS4 strains grown on medium to which 500 µg/ml of paromomycin was added (with this concentration of paromomycin, all strains grew with a significant speed). All AS4 crosses performed on such medium were as fertile as wild-type crosses on unsupplemented medium. This is in sharp contrast to the wild-type crosses, which were sterile on the medium supplemented with paromomycin. This result strongly suggests that the ascospore formation defect seen in AS4 mutants is due to reduced error level. Strikingly, the longevity of the AS4 cultures is not restored to the wild-type level on medium with or without the antibiotics (Table 2), suggesting that elevated accuracy is not responsible for increased life span.

An omnipotent informational suppressor, but not UGA tRNA suppressors, seems to display the same effect as paromomycin on the AS4 mutations:
To confirm the phenotypic suppression of the AS4 ascospore formation defect by paromomycin, we introduced, by crosses, various suppressor mutations in the AS4 strains. As for paromomycin, if translation accuracy is involved in the determination of the phenotype, this phenotype should be suppressed. We first tested the su1-1 omnipotent suppressor that is located in the eRF3 release factor subunit (GAGNY and SILAR 1998 ). Double mutant strains containing su1-1 and either one of the AS4 mutations were constructed. All the resulting strains had a slightly improved fertility when compared with the corresponding AS4 mutant strains. All had a slightly reduced longevity (Table 2). However, since su1-1 seems to decrease longevity by itself (Table 2; BELCOUR et al. 1991 ), it was not possible to know whether the su1-1 effect was brought about through decreased accuracy or through an indirect effect.

We similarly tested the UGA tRNA suppressors su8-1 and su4-1. Clearly, these tRNA are not able to restore ascospore formation and have no effect on life span, except when su4-1 is associated with AS4-44 (Table 2).

Suppressors of the sporulation defect of the AS4 mutations confirm that translation accuracy causes ascospore formation defects but has no effect on life span extension:
Because the effect with the informational suppressors was not clear-cut, we decided to search for suppressors that would restore ascospore formation in the AS4 mutants. If ascospore formation impairment is indeed due to elevated accuracy, a subset of these suppressors should be informational suppressors.

One suppressor was recovered spontaneously in one of our AS4-44 stocks and four others were recovered after UV mutagenesis of AS4-44 x AS4-29 crosses (see MATERIALS AND METHODS). The five suppressors were crossed with the 193 tester strain and, for two of them, green spores were recovered in the progeny. The fact that an informational suppressor was present in two of these strains indicates that accuracy indeed is involved in the control of ascospore formation. Mapping of these two informational suppressors showed that one was a new allele of the su3 gene (now called su3-8). The other could not be mapped to any of the known suppressor loci and may thus define a new gene (for the present article the mutant allele is called 1976). Both su3-8 and 1976 were associated with the various AS4 mutations and both fertility and longevity were measured. Although fertility was found increased in all strains, none restored a wild-type longevity level (Table 2). Several kinds of interactions were observed depending on the mutants' combination, suppressive as seen in AS4-27 su3-8 or additive as seen in AS4-11 su3-8. Interestingly, in one of the double mutants (AS4-30 su3-8) a synthetic effect was observed and life span was so tremendously increased that the strain became almost immortal since it lived more than 20 times the life span of wild type before the experiment was stopped! This confirmed the data obtained with the paromomycin; i.e., increased accuracy is not involved in life span extension.

Among the three suppressors that did not yield green ascospores when crossed with 193, two (N1 and N3) were unlinked to AS4. Both exhibited a 90% second division segregation frequency but segregated independently. The N1 suppressor was not very efficient in restoring fertility and did not act in crosses involving AS4-27, whereas the N3 suppressor was very efficient and restored ascospore formation in all AS4 mutant strains. The two mutations did not modify translation accuracy because we did not detect any modification of the levels of suppression or the resistance to paromomycin in strains carrying these mutations. Confirmation was given by the fact that AS4-44 N1 and AS4-44 N3 did present a Crippled Growth phenotype while AS4⁺ N1 and AS4⁺ N3 did not. The N1 suppressor had a reduced fertility and strains carrying N1 lived slightly longer (14.5 ± 1.9 for AS4⁺ N1 compared to 9.5 ± 1.0 for AS4⁺ and 31.0 ± 4.5 for AS4-44 N1 compared to 21.5 ± 3.5 for AS4-44), whereas the N3 suppressor did not entail any obvious phenotype.

The fifth suppressor (N6) was so closely linked to AS4-44 that we could not recombine AS4⁺ with it. It is not an intragenic revertant because no additional mutation was found when the AS4 allele was sequenced. The informational status of this extragenic suppressor (called N6) is not clear. First, it did not modify the antisuppression efficiency of the AS4-44 mutation at the 193 mutation site in the presence of the su8-1 suppressor. Second, it did not modify the paromomycin resistance of the AS4-44 mutation. However, the AS4-44 N6 strain did not present Crippled Growth anymore, as expected for an informational suppressor. It is also noteworthy that the AS4-44 N6 strain had a slightly diminished longevity (17.0 ± 1.5 cm) when compared to the AS4-44 strain.

AS4 mutations reveal complex effects on in vivo misreading:
To more precisely correlate accuracy increase and ascospore formation defect, we analyzed the antisuppression efficiency of the six AS4 mutants. To do this, we measured the in vivo readthough at either the leu1-1 or 193 sites promoted by two different types of suppressors, tRNA suppressors (su4-1 or su8-1) and an omnipotent suppressor (su1-1) resulting from a mutation in the eRF3 release factor subunit. Although this does not measure all kinds of translation error, it should give a good indication of the AS4 mutations' behavior. The data obtained are reported in Table 1 and can be summarized as follows: First, all AS4 mutations acted on both types of suppressors. Second, the efficiency of a given AS4 allele varied with both the mutation site and the type of suppressor in such a way that the four combinations (leu1-1 su1-1, leu1-1 su4-1, 193 su1-1, and 193 su8-1) resulted in different orders in the strength of antisuppression (see DISCUSSION).

As P. anserina crosses yield dikaryotic ascospores, we were able to observe the recessivity/dominance of the AS4 translational effect by the color of the heterokaryotic 193 su AS4⁺/193 su AS4^- ascospores. Two situations were found. First, for all six mutations, the 193 su1-1 AS4⁺/193 su1-1 AS4^- ascospores had the same color as the 193 su1-1 AS4⁺ ascospores, demonstrating that the AS4 mutations were recessive when they antagonized the su1-1 omnipotent suppressor. Second, for all six mutations, the 193 su8-1 AS4⁺/193 su8-1 AS4^- ascospores exhibited a color that was intermediate between the color of the 193 su8-1 AS4⁺ and the color of the 193 su8-1 AS4^- ascospores, indicating that the mutations were semidominant when they antagonized the su8-1 tRNA suppressor.

These results show that the various AS4 alleles could not be ordered on the criterion of antisuppression at a particular locus. Similarly, no obvious correlation between paromomycin resistance and antisuppression could be detected.

In vivo localization of eEF1A by GFP tagging:
Localization of eEF1A was determined using a reporter-tagged version of the protein carrying a carboxy-terminal addition of GFP (see MATERIALS AND METHODS for details). The AS4-44 mutant strain was transformed by a plasmid carrying the eEF1A-GFP fusion gene. Among the transformants, which expressed GFP and restored the AS4-44 sporulation defect, two (F4 and F7) were selected for further analysis. Each carried a transgene located at a different position in the genome. Although these transgenes restored viability of a strain bearing a deletion of AS4 (see MATERIALS AND METHODS for experimental design), and thus showed that they expressed a fusion protein able to perform eEF1A functions, restoration of AS4-44 fertility was only partial. This result suggested that the GFP-fusion protein was less active than the wild-type eEF1A. Therefore, one could expect the fusion protein to be endowed with an antisuppressor phenotype. We could not test this directly. Nevertheless, this idea was supported by the fact that the AS4 F4 and AS4 F7 strains exhibited the following phenotypes: (1) ascospore formation defects; (2) increased longevity (24.4 ± 4.9 cm and 23.1 ± 3.5 cm, respectively; note that this life span extension is recessive); (3) Crippled Growth; (4) resistance to paromomycin similar to AS4-44; and (5) a very low production of perithecium. All these phenotypes are characteristics of AS4 antisuppressor mutants, suggesting that the chimeric gene was indeed an antisuppressor allele.

eEF1A-GFP localization was analyzed in both F4 AS4 and F7 AS4 strains. Hyphae taken from the growing edge were always brightly fluorescent (Fig 4A). The staining heterogeneity visible in most cells was mainly due to the fact that organelles and vacuoles remained unstained. Some cells also showed a few brighter foci. These most often localized in the nuclei (Fig 4D and Fig E), while others, sometimes in the same hyphae, did not co-localize with nuclei (Fig 4, B–E). This focal localization was not observed in the GFP control (i.e., ectopic integration of GFP alone expressed from the constitutive GPD promoter of Aspergillus nidulans). Hyphae taken from stationary phase cultures exhibited an even more variable pattern of eEF1A-GFP fluorescence (Fig 4, F–H). While a few cells were uniformly stained like those of hyphae from the growing edge (Fig 4F), others showed highly heterogeneous staining (Fig 4G). Surprisingly, two neighboring cells could present a completely different pattern of staining (Fig 4H), and the same was true for the GFP-control hyphae. This suggests that the staining is not representative of a particular localization for eEF1A-GFP in stationary phase mycelia. This could be due either to the fact that those hyphae are in various physiological conditions or that they correspond to eEF1A-GFP dislocalization.

View larger version (76K): In this window In a new window Download PPT slide   Figure 4. In vivo localization of eEF1A. Fluorescence emitted by the eEF1A-GFP fusion was observed as described in MATERIALS AND METHODS. (A) Hyphae from the growing edge. (B and D) Hyphae showing the accumulation of eEF1A in some regions of the cell. Slanting arrows point toward eEF1A accumulation outside nuclei. Vertical arrows point toward eEF1A accumulation inside nuclei. (C and E) Corresponding DAPI staining. (F–H) Hyphae taken 5 cm away from the growing edge (i.e., cells have stopped growing for 1 wk). (I) Nonmature ascospore.

Because AS4 mutant strains fail to produce ascospores and/or female structures, we observed young ascospores issued from crosses of F4 AS4 and F7 AS4 by wild type (Fig 4I). These ascospores displayed a uniform distribution of eEF1A within their cytosol.

Since AS4 mutants promote a modification of life span, we checked whether the localization of eEF1A was changed in senescent cells. Hyphae taken from the growing edge of senescent F4 AS4 and F7 AS4 cultures also displayed a highly heterogeneous fluorescence pattern. In this case, such a pattern may be related to the fact that senescent cells show a drastically rearranged cytoplasm (DELAY 1963 ).

Three additional AS4 antisuppressor alleles:
The AS4-4, AS4-24, and AS4-33 alleles were recovered in the same screen as the AS4-11 to AS4-44 alleles and were categorized as lethal alleles (PICARD-BENNOUN 1976 ). We analyzed these three alleles again 25 years later. AS4-24 was indeed lethal. However, we were able to recover homokaryotic strains carrying AS4-4 or AS4-33. Both were very slow growing and unable to differentiate perithecium (although they produced ascogonia). Crosses involving these strains as the male parent and the other AS4 mutant strain as the female parent were all barren, as observed in crosses involving the other AS4 antisuppressor strains. Note that the growth of both strains resembles Crippled Growth too much, so as to prevent us from checking if these strains are able, like the other AS4 antisuppressor strains, to develop the Crippled Growth alteration. Strains carrying AS4-4 could not be analyzed for longevity because reversion of their phenotype toward a wild-type phenotype occurs rapidly. However, longevity was measured for the strains carrying AS4-33. Their life span is extended as observed for the other antisuppressor strains (16.5 ± 3.5 cm; wild-type longevity is 9.5 ± 1.0 cm). Like the other AS4 antisuppressor alleles, these three alleles are recessive since strains carrying a transgenic copy of AS4⁺ have a wild-type fertility and longevity. However, the AS4-4 strain carrying the ectopic copy of AS4⁺ has a slightly increased longevity (12.4 ± 1.5 cm) as described for the AS4-44 strain carrying the same transgene.

We next analyzed whether the inability to differentiate perithecium and the slow growth of the AS4-33 strain was an extreme effect due to increased accuracy. Clearly, neither impairment in perithecium production nor normal growth was restored on paromomycin nor when either one of the su1-1, su3-8, or 1976 informational suppressors were associated with the mutation. Hence, sterility of this strain is not due to increased accuracy. Interestingly, this strain is resistant to paromomycin. A similar analysis with respect to longevity confirmed that longevity was not controlled by accuracy. For example, longevity of this strain on medium containing 500 µg/ml paromomycin is 27.5 ± 5.7 cm compared to 9.5 ± 1.0 for wild type on nonsupplemented medium.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Considering the numerous eEF1A activities, it is not surprising that mutations in eEF1A are pleiotropic (SILAR and PICARD 1994 ; SILAR et al. 2000B ). What is not clear is the exact mechanism(s) involved in each of the phenotypes. In this study, we try to clarify this pleiotropy with respect to some antisuppressor mutations, which present three main phenotypes: increased life span, fertility defect, and Crippled Growth. Because these mutations were selected with a screen connected with translation accuracy, we could expect that part of their effects were related to translation accuracy and/or other translation parameters. Another argument pointing toward this inference is the fact that the molecular characterization of the mutations (SILAR et al. 2000B ) showed them to be scattered all over the primary sequence of the protein (all regions of the protein are necessary for translation). Indeed, if one of the other known eEF1A functions was more specifically involved in the mutant phenotype, it would be expected that the mutations cluster in this region. To confirm that the amino acid changes map to a different part of the protein, we have modeled the 3D structure of P. anserina eEF1A on the basis of the recently determined structure of the yeast eEF1A in complex with eEF1B (ANDERSEN et al. 2000 ). As seen on Fig 5, the six viable amino acid changes map in a different part of the protein, either in domain I or in domain III. AS4-29 and AS4-44 map in the same loop of domain I and are characterized by the same amino acid change (glycine to aspartic acid). Interestingly, they both display phenotypes with similar strength (similar yield of ascospores, similar longevity, similar level of paromomycin resistance, and an effect on readthrough associated with all suppressors; see below), suggesting that they may act through modification of the same mechanism. On the contrary, the four other mutations are localized in different regions of eEF1A, resulting in different quantitative effects. Additionally, the eEF1A-GFP fusion gene presents the same three characteristic phenotypes associated with the other viable antisuppressor mutants. This is anticipated if one assumes that the fusion protein is less efficient during translation than the wild-type one.

View larger version (51K): In this window In a new window Download PPT slide   Figure 5. Ribbon structure for P. anserina eEF1A. Domains I, II, and III are labeled. Position of the amino acid changes of the indicated alleles are highlighted by giving the depiction of their side chain.

We already proposed a model for the control of Crippled Growth, in which a translation error was involved (SILAR et al. 1999 ). Indeed, we observed a clear relationship between accuracy and propagation of the C element that is involved in the development of this cell degeneration; all antisuppressor mutants display the degenerative process whereas wild type and suppressors do not (SILAR et al. 1999 ). We have recently discovered one exception in one AS4 suppressor mutant, but this can be related to the special characteristic of this mutant (SILAR et al. 2000B ). It is likely that mutations in AS4, like the other antisuppressor mutations, promote Crippled Growth through their effect on translation fidelity.

The relation between fertility and translation accuracy is less clear. On one hand, many high fidelity mutants have a defect in ascospore formation, but on the other hand, some others do not. Here, we show that the defect presented by the AS4 mutations is similar to the defect presented in the AS7 mutants (DEQUARD-CHABLAT and COPPIN-RAYNAL 1984 ), arguing for a similar deficiency in all the antisuppressor mutants. This defect, as for AS7 mutants, is clearly due to increased accuracy because the phenotype can be reverted by paromomycin and by suppressor mutations since it was shown that both approaches decrease error level in P. anserina wild-type and antisuppressor strains in vivo (COPPIN-RAYNAL 1981 ) and in vitro (PICARD-BENNOUN 1981 ; COPPIN-RAYNAL 1982 ). A possible explanation to account for the discrepancies between the various mutants is that there might be a threshold level in the alteration of translation before seeing any ascospore formation defect. Some mutants may not reach this threshold and be fertile.

Several models can be proposed to account for an effect of translation error in ascospore formation. First, as we proposed for Crippled Growth (SILAR et al. 1999 ), expression of a gene required during the sexual cycle may necessitate a translation error. Second, because accuracy is maintained through activated consumption of energy (KURLAND 1992 ) and because sexual reproduction takes place in P. anserina when nutrients are exhausted, it is possible that energy is lacking to complete this last step in the development of the sexual cycle. Finally, as changes in accuracy are often associated with changes in ribosome processivity (KURLAND 1992 ), it is possible that large proteins required for ascospore formation are not produced in sufficient amount in the mutants.

To differentiate between these models, we first attempted to correlate accuracy and ascospore formation defect. This was possible with the AS4 mutants because they display a wide range of alterations both at the translational and fertility levels. Strikingly, antisuppression efficiency, as measured with readthrough promoted by two types of suppressors, is very variable depending on the system used to test it. Even though some trends can be discerned (AS4-29 acts efficiently whereas AS4-44 acts moderately on all types of suppressors; AS4-30 and AS4-11 efficiently antagonize the effects of tRNA suppressors, but not that of the omnipotent suppressor; and AS4-27 and AS4-43 have the opposite outcome), different orders between the various AS4 antisuppressors are obtained even when measuring the same kind of effect (e.g., either antagonism of a deficient release factor or insertion of a tRNA suppressor). Compare, for example, the effect of AS4-29 and AS4-43 on readthrough associated with su1-1 (Table 1). This suggests that, even for the same mechanism, the AS4 mutations have a codon-specific effect. Accordingly, analysis of two EF1A antisuppressor mutants of Escherichia coli yielded the same results (TAPIO and ISAKSSON 1990 ). This prevents us from ascertaining a classification based on this criterion in order to relate it to the severity of the ascospore formation defect. However, at least in the AS4 mutants, we noted a relationship between resistance to paromomycin and ascospore impairment (the more resistant, the more affected in ascospore formation; see Table 1 and first section of RESULTS). Moreover, AS7 mutants that display a strong ascospore formation defect are also highly resistant to paromomycin (COPPIN-RAYNAL and LE COZE 1982 ). Because the paromomycin resistance level is probably an indicator of global accuracy level, this result points toward one of the global hypotheses.

We also tried to identify the cellular mechanisms impaired during ascospore formation. Indeed, the first hypothesis suggests a specific effect during ascus formation whereas the second and the third ones should promote a more global disturbance. At least two distinct phenomena impair the ascospore formation in the AS4 mutants: defects in spindle plus nuclear localization and nuclear death after high condensation of the chromatin. The observed mislocalization of both spindles and nuclei could be related to the described interaction of eEF1A with the cytoskeleton (DURSO and CYR 1994 ). However, a recent article shows that this interaction is conditional and domain dependent, thus questioning the relevance of the in vitro data (MOORE and CYR 2000 ). Interestingly, the spindle localization defect associated with the AS4 mutations is not due to faulty interactions with microtubules per se, but is due to an indirect effect through translation modification since normal ascospore formation is restored when translation accuracy is decreased. These data would therefore again point toward the global hypotheses. Confirming this is the fact that in the ascospores, no special localization of the eEF1A-GFP fusion was detected. During our search of suppressors of the fertility defect we isolated two unlinked mutations that do not seem to be involved in accuracy control. Isolation of the corresponding genes will definitely shed some light on the involved processes.

Unexpectedly, some antisuppressor mutations previously described as lethal (PICARD-BENNOUN 1976 ) are in fact viable but present a very altered growth pattern associated with a lack of perithecium production. Analysis of AS4-33 showed that, at least in this mutant, fertility is not dependent on error level. Therefore, eEF1A exerts another control on sexual reproduction independently from accuracy monitoring.

Unlike ascospore formation, longevity is clearly not controlled by accuracy in the AS4 mutants. This confirms previous data that suggested the same conclusion (BELCOUR et al. 1991 ; SILAR et al. 1997 , SILAR et al. 2000B ). How can AS4 mutations affect life span? The answer to this question is not straightforward, especially because many mechanisms contribute to life span definition in P. anserina (ROSSIGNOL and SILAR 1996 ). Strikingly, the increased life span of the AS4 antisuppressors is a recessive phenotype whereas the increased life span of AS4 suppressor mutations is dominant (SILAR et al. 2000B ). It is therefore possible that eEF1A controls longevity through several mechanisms. We are currently testing this hypothesis by associating various eEF1A alleles.

The GFP tagging of eEF1A confirms the previously reported cellular localization (GANGWANI et al. 1998 and references therein). eEF1A is mostly located in the cytoplasm, but can transiently accumulate within the nucleus. In yeast and mammals (GANGWANI et al. 1998 ), the nuclear localization is correlated with cellular proliferation. In these organisms, it was shown that eEF1A binds the ZPR1 protein that relocalizes to the nucleus when growth is stimulated. Disruption of the interactions between the two proteins results in the accumulation of cells in the G2/M phase of the cell cycle (GANGWANI et al. 1998 ). Although we previously were able to observe such a correlation between cell cycle and nuclear localization of a GFP-tagged ribosomal protein (LALUCQUE and SILAR 2000 ), we could not detect any correlation between growth status and nuclear localization of the eEF1A-GFP protein. In addition to a nuclear localization, we have also observed eEF1A foci in the cytoplasm. Similar cytoplasmic accumulation has been seen during Xenopus oogenesis (VIEL et al. 1990 ). We have no indication of the possible role of this particular localization. Note that, like MOORE and CYR 2000 , we did not observe any obvious association with the cytoskeleton, questioning the in vivo relevance of an interaction between eEF1A and the cytoskeleton in normal conditions.

Finally, there is no evident correlation between fertility and longevity in P. anserina. The case of the double mutant strain AS4-30 su3-8 is quite striking because it is almost immortal, but has a near wild-type fertility and reproduces as fast as wild type. Apart from the Rgs31 mutant strain (CONTAMINE and PICARD 1998 ), all immortal mutants previously analyzed are female sterile (TUDZYNSKI and ESSER 1979 ; DUFOUR et al. 2000 ) or show a delayed reproduction due in part to slow growth (case of PaTOM70-1, JAMET-VIERNY et al. 1997B ; V. CONTAMINE, personal communication).

	ACKNOWLEDGMENTS

We thank Corinne Vierny and all members of the RSTD laboratory for useful discussions, Sébastien Kicka and the other summer students who temporarily worked on the AS4 mutants, Latha Prabha Ganesan for correcting spelling and grammatical errors, and Françoise James for her technical assistance. We also thank Corinne Clavé for the gift of the GPD-GFP control. This work was supported by grant no. 5388 from Association pour la Recherche sur le Cancer (ARC). The work was done in compliance with the current laws governing genetic experimentation in France. P. Silar is professor at the University of Paris VII, Denis Diderot. H. Lalucque is a recipient of a fellowship from the Ministère de la Recherche.

Manuscript received January 5, 2001; Accepted for publication May 11, 2001.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ANDERSEN, G. R., L. PEDERSEN, L. VALENTE, I. CHATTERJEE, and T. G. KINZY et al., 2000  Structural basis for nucleotide exchange and competition with tRNA in the yeast elongation factor complex eEF1A:eEF1Bß. Mol. Cell 6:1261-1266[Medline].

AUSUBEL, F. M., R. BRENT, R. E. KINGSTON, D. D. MOORE, J. G. SEIDMAN et al., (Editors), 1987 Current Protocols in Molecular Biology. Wiley Interscience, New York.

BELCOUR, L., O. BEGEL, and M. PICARD, 1991  A site-specific deletion in the mitochondrial DNA of Podospora is under the control of nuclear genes. Proc. Natl. Acad. Sci. USA 88:3579-3583[Abstract/Free Full Text].

BRYGOO, Y. and R. DEBUCHY, 1985  Transformation by integration in Podospora anserina. I. Methodology and phenomenology. Mol. Gen. Genet. 200:128-131.

CHAMBERS, D. M., J. PETERS, and C. M. ABBOTT, 1998  The lethal mutation of the mouse wasted (wst) is a deletion that abolishes expression of a tissue-specific isoform of translation elongation factor 1, encoded by the Eef1a2 gene. Proc. Natl. Acad. Sci. USA 95:4463-4468[Abstract/Free Full Text].

CHEN, E., G. PROESTOU, D. BOURBEAU, and E. WANG, 2000  Rapid up-regulation of peptide elongation factor EF-1 protein levels is an immediate early event during apoptosis. Exp. Cell Res. 259:140-148[Medline].

CONTAMINE, V. and M. PICARD, 1998  Escape from premature death due to nuclear mutations in Podospora anserina: repeal versus respite. Fungal Genet. Biol. 23:223-236[Medline].

COPPIN-RAYNAL, E., 1981  Ribosomal suppressors and antisuppressors in Podospora anserina: altered susceptibility to paromomycin and relationships between genetic and phenotypic suppression. Biochem. Genet. 19:729-740[Medline].

COPPIN-RAYNAL, E., 1982  Ribosomal control of translational fidelity in Podospora anserina: a suppressor and an antisuppressor affecting the paromomycin-induced misreading in vitro. Curr. Genet. 5:57-63.

COPPIN-RAYNAL, E. and D. LE COZE, 1982  Mutations relieving hypersensitivity to paromomycin caused by ribosomal suppressors in Podospora anserina. Genet. Res. 40:149-164[Medline].

COPPIN-RAYNAL, E., M. DEQUARD-CHABLAT and M. PICARD, 1988 Genetics of ribosomes and translational accuracy in Podospora anserina, pp. 431–442 in Genetics of Translation: New Approaches, edited by M. F. TUITE, M. PICARD and M. BOLOTIN-FUKUHARA. Springer Verlag, Heidelberg, Germany.

COTTRELLE, P., M. COOL, P. THURIAUX, V. L. PRICE, and D. THIELE et al., 1985  Either one of the two yeast EF-1 genes is required for cell viability. Curr. Genet. 9:693-697[Medline].

DEBUCHY, R. and Y. BRYGOO, 1985  Cloning of opal suppressor tRNA genes of a filamentous fungus reveals two tRNA Ser/UGA genes with unexpected structural differences. EMBO J. 4:3553-3556[Medline].

DELAY, C., 1963  Observations inframicroscopiques sur le mycelium "sénescent" du Podospora anserina. C. R. Acad. Sci. Paris 256:4721-4724.

DEQUARD-CHABLAT, M. and E. COPPIN-RAYNAL, 1984  Increase of translational fidelity blocks sporulation in the fungus Podospora anserina. Mol. Gen. Genet. 195:294-299.

DEQUARD-CHABLAT, M. and C. H. SELLEM, 1994  The S12 ribosomal protein of Podospora anserina belongs to the S19 bacterial family and controls the mitochondrial genome integrity through cytoplasmic translation. J. Biol. Chem. 269:14951-14956[Abstract/Free Full Text].

DUFOUR, E., J. BOULAY, V. RINCHEVAL, and A. SAINSARD-CHANET, 2000  A causal link between respiration and senescence in Podospora anserina. Proc. Natl. Acad. Sci. USA 97:4138-4143[Abstract/Free Full Text].

DUJON, B., and L. BELCOUR, 1989 Mitochondrial DNA instabilities and rearrangements in yeasts and fungi, pp. 861–878 in Mobile DNA, edited by D. E. BERG and M. M. HOWE. American Society for Microbiology, Washington, DC.

DURSO, N. A. and R. J. CYR, 1994  Beyond translation: elongation factor-1 and the cytoskeleton. Protoplasma 180:99-105.

ESSER, K., 1974 Podospora anserina, pp. 531–551 in Handbook of Genetics, edited by R. C. KING. Plenum, New York.