Incomplete Penetrance and Variable Expressivity

Copyright  2004 by the Genetics Society of America

Incomplete Penetrance and Variable Expressivity of a Growth Defect as a Consequence of Knocking Out Two Kⴙ Transporters in the Euascomycete Fungus Podospora anserina Herve´ Lalucque and Philippe Silar1 Institut de Geńe´tique et Microbiologie UMR 8621, Universite´ de Paris-Sud, 91405 Orsay Cedex, France Manuscript received May 23, 2003 Accepted for publication September 28, 2003 ABSTRACT We describe an example of incomplete penetrance and variable expressivity in the filamentous fungus Podospora anserina, two genetic properties classically associated with mutations in more complex organisms, such as green plants and animals. We show that the knockouts of two TRK-related K⫹ transporters of this ascomycete present variability in their phenotype that cannot be attributed to fluctuations of the genetic background or the environment. Thalli of the knockout strains derived from independent monokaryotic ascospores or from a single monokaryotic ascospore and cultivated under standard growth conditions may or may not present impaired growth. When impaired, thalli exhibit a range of phenotypes. Environmental conditions control expressivity to a large extent and penetrance to a low extent. Restoration of functional potassium transport by heterologous expression of K⫹ transporters from Neurospora crassa abolishes or strongly diminishes the growth impairment. These data show that incomplete penetrance and variable expressivity can be an intrinsic property of a single Mendelian loss-of-function mutation. They also show that such variability in the expression of a mutant phenotype can be promoted by a phenomenon not obviously related to the well-known chromatin structure modifications, i.e., potassium transport. They provide a framework to understand human channelopathies with similar properties.

I

NCOMPLETE penetrance and variable expressivity have been described in animals and plants, in which alleles in many genes may trigger phenotypic changes in only a subset of carriers or cause variable severity in the phenotype. Classically, both effects are attributed to interactions with other genes, environmental fluctuation, and epigenetic variation in gene expression (Griffiths et al. 2000; Rakyan et al. 2002). In this last case, the epigenetic variation is effected by changing the chromatin structure (Rakyan et al. 2002). Presently, there is a large scientific effort to understand the cause of this variable phenotypic expression. Two classes of epigenetic phenomena may trigger variable expression in the yeasts Saccharomyces cerevisiae and Schizosaccharomyces pombe: the silencing phenomena at the telomeres, the silent mating-type loci, and other silent regions (Ayoub et al. 1999; Maillet et al. 2001) and the yeast prions phenomena (Derkatch et al. 1996). Although these kinds of processes potentially lead to incomplete penetrance and variable expressivity in the context of a multicellular organism, these unicellular yeasts are not well suited to follow the phenotypic expression of a character after meiosis in a multicellular

Sequence data from this article have been deposited with the EMBL/GenBank Data Libraries under accession nos. AY077729 and AL627362. 1 Corresponding author: Institut de Geńe´tique et Microbiologie UMR 8621, Universite´ de Paris-Sud, 15, rue George Clemenceau, 91405 Orsay Cedex, France. E-mail: [email protected] Genetics 166: 125–133 ( January 2004)

individual. On the contrary, filamentous fungi are suitable for such studies. First, after meiosis they rapidly generate a multicellular thallus that can rapidly differentiate complex structures, usually devoted to sexual or asexual reproduction. Second, there is communication between the different parts of the mycelium, suggesting that a thallus is a true organism (Olsson 2001). Although researchers studying filamentous fungi are aware of phenotypic variations of their favorite organisms, usually appearing as invasive sectors of morphological variations (Silar and Daboussi 1999), few studies have been undertaken to understand the symptoms, etiology, and molecular basis of these phenomena. We show that the knockouts of two potassium transporters of the TRK family result in a variable-sector phenotype that can be best described as incomplete penetrance and variable expressivity. We were able to eliminate both genetic and environmental effects as the cause of the phenotypic variations. Our results suggest that stochastic events are the cause of variability in this system. Although potassium is known to regulate morphogenesis in Neurospora crassa and other eukaryotes (Shaw and Hoch 2001; Levin et al. 2002), this is the first report in which potassium transporters are shown to control the stability of a morphological pattern. MATERIALS AND METHODS Strains, media, and genetic analysis: All strains used are derived from the S strains (Rizet 1952), which have been

126

H. Lalucque and P. Silar

kept in laboratories for over 50 years, with annual self-crossing to prevent accumulation of mutations. Apart from the matingtype region, all genomic DNA is identical between the mat⫹ and mat⫺ strains, including Patrk-1 and Patrk-2. The mid26 strain was provided by A. Sainsard-Chanet (Begel et al. 1999). The AS6-5 strain carries a mutation that decreases translation error rate and is used as a tester to determine if the cytoplasmic element responsible for Crippled Growth, a cell degeneration displayed by some Podospora anserina mutant strains, is present in mycelia (Silar et al. 1999). The mutation Patrk-1452 was selected to prevent the AS4-44 strain from exhibiting the Crippled Growth cell degeneration (Silar et al. 1999). The Patrk-1452 AS4-44 double mutant was crossed with wild type. Progeny analysis verified that Patrk-1452 segregated as a single Mendelian mutation independently of AS4-44. Phenotypic analysis showed that the Patrk-1452 mutant strain grew slowly both at 37⬚ on M2 medium and on germination (G) medium at 27⬚ and did not grow at all when both conditions were combined, i.e., on G medium at 37⬚. Addition of ammonium ions in the growth medium strongly inhibited the growth of Patrk-1452, which explained the slow growth phenotype on G medium, as this medium contained a large amount of ammonium. The reduced growth on G medium could be attributed to a defect in germination resulting in microscopic Patrk-1452 thallus after 2 days of ascospore incubation (wild-type thalli had a diameter of 1–2 cm under the same conditions). The growth and germination defect could be alleviated by the addition of 100 mm KCl to the media. The mutant phenotypes cosegregated in a 2:2 Mendelian fashion when crossed with wild type, showing that all the phenotypes were due to a single mutation. The mutation had a first division segregation frequency of 27%. M2 minimal medium has the following composition: KH2PO4 0.25 g/liter, K2HPO4 0.3 g/liter, MgSO4/7H2O 0.25 g/liter, urea 0.5 g/liter, thiamine 0.05 mg/liter, biotin 0.25 ␮g/liter, citric acid 2.5 mg/liter, ZnSO4 2.5 mg/liter, CuSO4 0.5 mg/liter, MnSO4 125 ␮g/liter, boric acid 25 ␮g/liter, sodium molybdate 25 ␮g/liter, iron alum 25 ␮g/liter, dextrin 5 g/liter, and agar 12.5 g/liter. In M2 medium without potassium (M2 ⫺ K), potassium phosphate ions were replaced by their sodium counterparts. In the M2 medium with high potassium content (M2 ⫹ K) KCl was added to a final concentration of 100 mm. G medium is Bacto-peptone 15 g/liter, ammonium acetate 6 g/liter, and agar 12 g/liter. All petri plates were poured using an automatic dispenser under standardized conditions. Genetic analysis of P. anserina was previously described (Esser 1974). Ascospores were germinated on G medium supplemented with 100 mm KCl for 24 hr and then transferred on the growth medium. Sectors were checked after 10 days of growth at 27⬚ or 20 days of growth at 18⬚. Cloning and knockout of Patrk-1: The wild-type allele of Patrk-1 was cloned by complementation of the 37⬚ temperature sensitivity of Patrk-1452 on germination medium, using a cosmid bank derived from the S strain and provided by M. DequardChablat. The cosmid was then subcloned into pBC-Hygro (Silar 1995) and the smallest DNA fragment able to complement the growth defect was sequenced. We identified in the sequenced region a large coding sequence (CDS) interrupted by one intron. Intron position and expression of the gene was validated by reverse transcriptase-PCR experiments. The CDS coded for a protein homologous to plant and fungal transporters of the TRK family with 56% identity with the central region of the N. crassa TRK-1 protein (Haro et al. 1999). Complete sequence was made on two independent PCR products derived from Patrk-1452 genomic DNA, identifying two deleted nucleotides in codon 150 in the Patrk-1452 mutant allele. To knock out the Patrk-1 gene, we took advantage of the presence of an XbaI site 60 nucleotides before the start codon

and an EcoRI site at position 1742 of the coding sequence. Replacement of the DNA region within these two restriction sites by the complete pBC-Phleo vector (Silar 1995) should result in a null allele, because the beginning of the gene as well as the first 582 codons of the coding sequence is missing. Knockout vector was made by reassembling the 0.7-kb BamHIEcoRI fragment and the 4.0-kb BamHI-XbaI fragment that surround the Patrk-1 gene into the pBC-Phleo plasmid (Silar 1995) digested by EcoRI and XbaI, yielding plasmid p⌬Trk-1. Protoplasts of the S strain were transformed with p⌬Trk-1 linearized at the BamHI site and the phleomycin-resistant transformants were selected. Several were unable to grow at 37⬚ on G medium; two were further analyzed by Southern blotting. Both had the correct replacement of the Patrk-1 region by the pBC-Phleo DNA fragment (data not shown). Apart from their phleomycin resistance, these strains were indistinguishable from the original Patrk-1452 strain. One, ⌬Patrk-1, was selected for further studies. Knockout of Patrk-2 : To knock out the Patrk-2 gene, four primers were designed to amplify by PCR two 1.5-kb regions located upstream and downstream the Patrk-2 CDS. For each primer, a restriction site was added in 5⬘ to facilitate the cloning of the PCR products and several bases (boldface type) were added to allow an efficient digestion. The name of the restriction enzyme is indicated before the sequence primer and the restriction site is underlined. The first pair of primers was ⌬Trk-2-1 (SpeI 5⬘ CCACTAGTTAGTCTCGTGCTTGCCC TCC 3⬘) and ⌬Trk-2-2 (ApaI 5⬘ AAGGGCCCCGGATACAAC CAGGAACAAG 3⬘); the second pair of primers was ⌬Trk-2-3 (NotI 5⬘ ATTTGCGGCCGCCAGCTCGTCTCTGGCGAGAC 3⬘) and ⌬Trk-2-4 (SpeI 5⬘ CCACTAGTTGCCGAGGTGCGAGAG ATCC 3⬘). These two PCR products were amplified by PCR from P. anserina genomic DNA with the Pfu polymerase turbo hot start (Stratagene, La Jolla, CA). Both PCR products were isolated, digested by the ad hoc enzymes, and cloned into the vector pBC-Hygro (Silar 1995) linearized by ApaI and NotI. This constructed plasmid pBC-⌬T2. Protoplasts of the wildtype strain were transformed by pBC-⌬T2 linearized at the SpeI restriction site. A double crossing-over event allowed replacement of the entire coding sequence of Patrk-2 by the pBCHygro sequence. Because Patrk-2 is close to the centromere of chromosome V, 23 transformants were crossed with strains carrying the su8-1 mutation, which is linked to the centromere of chromosome V. For two transformants, the analysis of the progeny of these crosses revealed that the integration was located close to the centromere of chromosome V since no recombination was observed between su8-1 and the integrated hygromycin B resistance marker. The confirmation of the deletion of Patrk-2 gene (⌬Patrk-2) in those two transformants was obtained by Southern blot. Transgenic expression of the N. crassa TRK-1 and HAK-1 in P. anserina: The transgenic expressions of the trk-1 and hak-1 coding sequences have been made under the control of the Patrk-1 promoter. Plasmids PLS5 and pRH-11 carrying the coding sequences of trk-1 and hak-1 were provided by A. Rodriguez-Navarro. The Patrk-1 promoter was amplified by PCR from P. anserina genomic DNA with the Pfu polymerase turbo hot start (Stratagene). For some primers, a restriction site was added in 5⬘ to facilitate the cloning of the PCR products and several bases (boldface type) are added to allow an efficient digestion. The name of the restriction enzyme is indicated before the sequence primer and the restriction site is underlined. The primers used to amplify the Patrk-1 promoter were trk-1-2 (5⬘ AAATAGTCTGCCATCACAATC 3⬘) and 452-37 (NotI 5⬘ ATAAGAATGCGGCCGCTGGACCCACGCCTCAAA TGC 3⬘). The PCR product was isolated, digested by NotI and XbaI, and cloned into the vector pCB1530 linearized with the same enzymes. pCB1530 contains a Bialaphos resistance

K⫹ Transporter Knockouts in P. anserina marker (Sweigard et al. 1997), giving rise to plasmid pCBpromPatrk-1. The N. crassa trk-1 and hak-1 coding sequences were amplified by PCR from pLS5 and pRH1.1, respectively, and using two designed primers, for the trk-1 gene, Trk-1-1 (XbaI 5⬘ GCTCTAGATTGTCGCCCACCATGGAACG 3⬘) and Trk-1-2 (SpeI 5⬘ GGACTAGTGATGCAAATTCCGCCCTTCG 3⬘) and for the hak-1 gene, Hak-1-1 (XbaI 5⬘ GCTCTAGACCATAAA AAAAAAGATGGAC 3⬘) and Hak-1-2 (BamHI 5⬘ CGGGATCC GAATGGGAGTTGTTCAGTTG 3⬘). The trk-1 and hak-1 PCR products were isolated, digested by the ad hoc enzymes, and cloned into the plasmid pCB-promPatrk-1 that had been linearized by the same enzymes to yield plasmids pT1-trk-1 and pT1-hak-1, respectively. Strain Patrk-1452 was transformed with these two plasmids to allow selection of functional transgenic copies on the basis of restoration of growth at 37⬚ on germination medium. In both cases, numerous Bialaphos resistant transformants with such ability were obtained, showing that the transgenes were expressed. In each case, two independent transformants were selected for further studies and crossed with the ⌬Patrk-1 strain of opposite mating type. Analysis of the progeny revealed that, in each transformant, the transgenes had integrated at a single locus. In the two transformants with pT1-trk-1, integration was unlinked to Patrk-1452. In the case of pT1-hak-1, one of the two transgenes integrated close to the Patrk-1452 locus preventing the analysis of the effect of this transgene in the ⌬Patrk-1 background. For each plasmid, the two transgenes displayed the same phenotype, showing that the phenotypes were not due to the integration point but to the expression of the transgenes. RESULTS

During a search for mutations that interfere with the propagation of the cytoplasmic and infectious element responsible for the Crippled Growth cell degeneration in P. anserina, we isolated Patrk-1452, a mutation that deletes two nucleotides of codon 150 of Patrk-1, a gene encoding a 939-amino-acid protein similar to K⫹ transporters of the TRK family (see materials and methods for the properties of the mutations and the cloning of the relevant gene; GenBank accession no. AY077729). In the mutant strain, the growth is thermosensitive at 37⬚, reduced on ammonium-containing medium, and displays variable sectors of alteration (see next section). Because the Patrk-1452 mutation still permitted the production of a truncated 150-amino-acid polypeptide, a knockout of Patrk-1 (⌬Patrk-1) was constructed by replacing the first two-thirds of the gene by a phleomycin-resistant marker. The ⌬Patrk-1 strain had exactly the same phenotypes as the Patrk-1452 mutant, showing that these can be attributed to the lack of the protein PaTRK-1. Evidence that the protein encoded by Patrk-1 was involved in K⫹ homeostasis came from the fact that addition of KCl, but not of NaCl, corrected the thermal sensitivity and reduced growth on ammonium-containing media of ⌬Patrk-1. When analyzed in wild type ⫻ ⌬Patrk-1 crosses, the thermal sensitivity and reduced growth on ammonium-containing medium segregated in a Mendelian fashion, as due to a single mutation. On the contrary, sectors of growth alteration displayed a non-Mendelian behavior.

127

⌬Patrk-1 displays a growth alteration with an incomplete penetrance and a variable expressivity: In the progeny of wild type ⫻ ⌬Patrk-1 crosses, 33 of 40 ⌬Patrk-1 thalli derived from independent monokaryotic ascospores presented a shape and fertility similar to that of wild type when cultivated at 18⬚ on M2 minimal medium (Figure 1A, “normal” shape). The remaining 7 displayed altered sectors of growth (Figure 1B). Despite variability in the growth alteration (see below), all sectors lacked aerial hyphae and were female sterile but male fertile. Under the same conditions, wild type and ⌬Patrk-1 carrying a transgenic wild-type Patrk-1 never presented this phenomenon. We called this impaired growth Wavy because of the waves observed on most altered forms. Some thalli derived from independent ascospores presented altered sectors very early after germination while others displayed them later in the culture. The sectors presented variable forms that were clearly revealed when mycelium explants were inoculated onto fresh medium and allowed to develop (Figure 1, C–E). This kind of variability is reminiscent of incomplete penetrance and variable expressivity of some plant and animal mutations, as described in Griffiths et al. (2000). We could conclude that variability was not due to a hidden secondary mutation present in the stock because when several cultures issued from the same initial ascospores were set up, similar variability was encountered in 5 of 20 cultures and 9 of 20 cultures for two independent ascospores. It was also not due to a variation in the medium because some thalli displayed two distinct forms on the same culture plate (Figure 1, F–H). Stability after subculture demonstrated that the altered phenotype was faithfully transmitted during mitotic divisions. Nevertheless, altered cultures upon extensive subculture eventually ended up displaying the form of Figure 1E. This suggested that all forms reflected the same underlying physiological alteration and that the form in Figure 1E displayed the strongest physiological modification and therefore was called “strong Wavy.” The other forms were designated as “weak Wavy” (Figure 1C) and “medium Wavy” (Figure 1D). In yeast, deletions of genes encoding potassium transporters of the TRK family result in plasma membrane hyperpolarization (Madrid et al. 1998). We assayed membrane polarization by sensitivity to hygromycin B and tetramethylammonium (TMA). Cells that have a hyperpolarized membrane are sensitive to these drugs (McCusker et al. 1987; Navarre and Goffeau 2000). Sensitivity to hygromycin B and TMA was measured in ⌬Patrk-1 from both its normal and its strong Wavy forms and compared to wild type. Loss of PaTRK-1 entailed membrane hyperpolarization since it was hypersensitive to hygromycin B and TMA (Figure 2). Noticeably, the Wavy form was more resistant to hygromycin B and TMA, suggesting that it has a membrane less polarized than the normal form. The addition of carbonyl cyanide 3-chlorophen-

128


Figure 1.—Incomplete penetrance and variable expressivity of Patrk-1 knockout. (A) The normal form observed in 80% of the cultures made on M2 medium at 18⬚. (B) A typical sector of altered growth seen in the remaining cultures. (C–H) Variable forms obtained with replication of mycelium explants taken from sectors depicted in B. (I and J) Normal and most altered forms on medium without potassium at 18⬚, respectively. (K) Altered form along with a normal area of growth on medium supplemented with 100 mm KCl at 18⬚.

ylhydrazone, a compound that depolarizes the plasma membrane (Heytler 1963) had no effect on the transition between the normal and altered forms. This showed

Figure 2.—Growth rate of wild-type, normal, and altered forms of ⌬Patrk-1 on M2 medium supplemented with hygromycin B (top) or tetramethylammonium (TMA, bottom) at 18⬚. Experiments were made in duplicate and the same values were obtained (error due to measuring procedure was ⬍0.1 mm/day in all cases).

that the polarization difference was not the cause of, but was simply correlated with Wavy. Nuclear and mitochondrial mutations cannot account for variability: To ensure that the variable forms were not due to the appearance of additional nuclear mutations during growth, we crossed as the female parent a ⌬Patrk-1 normal culture with either a ⌬Patrk-1 altered culture or a ⌬Patrk-1 normal one. In the progeny, 17/ 56 and 13/50 thalli, respectively, presented sectors of alteration with the same range of phenotype as the previously analyzed Patrk-1452 or ⌬Patrk-1 cultures. When cultivated on the semidefined cornmeal medium at 27⬚, these altered strains were able to revert to normal (see below), showing that the growth alteration was reversible. Overall, these data showed that no classical Mendelian mutation could account for the growth alteration. To ascertain that Wavy was not due to mitochondrial DNA (mtDNA) modifications akin to those seen in P. anserina during senescence (Silar et al. 2001), mtDNA structure was analyzed. First, a ⌬Patrk-1 culture was crossed as a male partner to a female mid26 strain. This strain carries a mitochondrial mutation that delays the mtDNA rearrangements seen during senescence and transmits it to its progeny when used as female (Begel et al. 1999). In this genetic background, we expected the penetrance to be low if mtDNA rearrangements were involved in the thalli alteration. In the progeny, the ⌬Patrk-1 mid26 strains showed the same culture variability with the same frequencies as the ⌬Patrk-1 strains (3 of 16 tested monokaryotic ascospores). Second, no obvious mtDNA modification could be detected using restriction analysis of mtDNA in two independent altered cultures recovered in the progeny of the cross (data not shown). This indicated that the mitochondrial modifications related to those observed during P. anserina senescence were not implicated. Expressivity but not penetrance is greatly influenced by external conditions: Temperature and potassium

K⫹ Transporter Knockouts in P. anserina

129

TABLE 1

TABLE 2

Effect of temperature and potassium concentration on penetrance

Effect of the expression of TRK-1 and HAK-1 on sector frequency

M2 [K⫹] 18⬚ 27⬚

5.3 mm 6/23 (26%) 5/35 (14%)

M2 ⫹ K 105.3 mm 7/50 (14%) 25/49 (51%)

M2 ⫺ K 0 mm NT 3/50 (6%)

Frequency of sectors of altered growth presented by thalli issued from the progeny of a ⌬Patrk-1 ⫻ ⌬Patrk-1 cross, cultivated under the indicated conditions (number of sectors per number of thalli). NT, not testable due to lack of obvious differences between the forms.

concentration in the medium were varied to evaluate their influence on penetrance or expressivity in the ascospore progeny issued from a ⌬Patrk-1 ⫻ ⌬Patrk-1 cross. Frequency of sector formation was within an order of magnitude (Table 1) and thus initial sector setup always involved a minority of cells. In no instances have we found conditions that would promote a large fraction of the cells to initiate sectors. This suggested that penetrance was modestly dependent upon the environmental conditions of temperature and potassium concentration. Note that this frequency reflected not only the frequency of sector appearance but also that of their positive selection during growth. The frequency of appearance was related to cell number, a parameter difficult to evaluate on petri plate-grown thalli, especially for syncytial fungi. The selection frequency was also difficult to evaluate since it depended on the respective growth characteristics of the normal and altered forms. These were greatly influenced by environmental factors. For example, utilization of cornmeal medium resulted in no observable sectors but altered cultures grown on this medium progressively reverted to normal, suggesting that the altered form is counterselected under these conditions. In other words, expressivity, unlike penetrance, was greatly influenced by culture conditions. This was seen directly on the sectored thalli but most visible in thalli generated by inoculating mycelium explants from the normal and altered portions onto fresh medium. Increasing the temperature resulted in the destabilization of the growth alteration. The normal form of ⌬Patrk-1 looked similar at 18⬚ and 27⬚ on M2 medium. At 27⬚, all the Wavy cultures displayed weak Wavy. Explants taken from these cultures frequently generated normal thalli when inoculated at 27⬚, but when replicated at 18⬚ they regenerated the range of weak, medium, and strong forms previously described. Conversely, we observed that inoculating mycelium explants of Wavy cultures that had been obtained at 18⬚ onto M2 medium at 27⬚ resulted most often in a progressive reversion toward the normal form. This usually followed

⌬Patrk-1 7/28 (25%)

⌬Patrk-1 ⫹ TRK-1

⌬Patrk-1 ⫹ HAK-1

0/125 (0%)

2/102 (2%)

Frequency of sectors of altered growth in thalli carrying no transgenes (⌬Patrk-1), one transgene expressing the TRK-1 protein (⌬Patrk-1 ⫹ TRK-1), or the HAK-1 protein (⌬Patrk-1 ⫹ HAK-1).

a sequence of shapes starting from weak Wavy toward strong Wavy and going through medium Wavy. When explants from these cultures were inoculated onto M2 at 18⬚, they generated the same variability as the ones issued directly from ascospores. Normal M2 medium contains 5.3 mm potassium. Modifying ion concentration resulted in morphological modifications. When sodium ions were used in place of potassium in this medium, both normal or Wavy forms presented a very similar appearance, different from those observed on classical M2. This was especially visible at 18⬚ where the two kinds of cultures were essentially identical (Figure 1, I and J). However, this was only a temporary masking effect since replicating these cultures on M2 medium at 18⬚ resulted in the recovery of the original differences. Increasing the concentration of potassium by adding 100 mm KCl resulted in almost no effect on the normal phenotype, whereas the altered forms all looked the same. They grew slowly as a spindly, unpigmented mycelium with very frequent areas of reverted growth (Figure 1K). These sectors were likely obtained because they were positively selected during growth as the normal mycelium had a tremendous growth advantage over the altered one. Expression of N. crassa TRK-1 and HAK-1 transporters partially complement the growth defect: To evaluate the role of PaTRK-1 through potassium transport vs. a more specific role in the growth alteration, we decided to express the TRK-1 and HAK-1 transporters of the related fungus N. crassa in the strains mutated for Patrk-1. TRK-1 is the homolog of PaTRK-1 and should function in a similar fashion. HAK-1 belongs to another class of transporters and exhibits several differences in its transport properties when compared withTRK-1 (Haro et al. 1999). Two plasmids expressing TRK-1 and HAK-1 from the Patrk-1 promoter were constructed and introduced into the Patrk-1452 strain by transformation (see materials and methods for detailed experimental procedure). Expression of TRK-1 restored most of the phenotypic deficiencies of ⌬Patrk-1, including the poor growth at 37⬚ and on germination medium, as well as growth impairment. Indeed, no Wavy sector was observed on thalli issued from 125 ascospores (Table 2). In the same exper-

130


iment, the control ⌬Patrk-1 ascospores presented 7 sectored cultures out of 28 tested. However, upon extensive subcultures, some thalli carrying the TRK-1 expression transgenes rarely presented altered growth sectors. These were similar in appearance to weak Wavy, suggesting a mild alteration. They could be attributed to a partial complementation of the ⌬Patrk-1 by the transgenes. Expression of HAK-1 partially restored the phenotypic deficiency of ⌬Patrk-1 or Patrk-1452. Growth at 37⬚ or on ammonium-containing medium was ameliorated but not up to the level of wild type. Sector formation was inhibited but not completely eliminated. Indeed, 2 of 102 thalli carrying the HAK-1 expression transgenes presented Wavy sectors (Table 2). These were intermediate in appearance to the medium and strong Wavy. Overall, these data suggest that restoration of a functional potassium transport eliminated almost completely, in the case of the TRK-1 expression, or partially, in the case of the HAK-1 expression, all the phenotypes, including the incomplete penetrance and variable expressivity. This suggested that the latter were related to abnormal potassium transport. Knockout of Patrk-2, a second isoform, also displays a variable growth defect: The relation between variability and potassium transport was further substantiated by the study of a second isoform of TRK transporters, PaTRK-2. The Patrk-2 gene was discovered through the systematic sequencing of the pericentromeric region of chromosome V (Silar et al. 2003; GenBank accession no. AL627362). PaTRK-2 displays 27% identity (36% in the most conserved regions) and 43% similarity with PaTRK-1. Search in release 3 of the N. crassa complete genomic sequence showed that a trk-2 isoform is also present in this fungus (the sequence is available at http://www-genome.wi. mit.edu/annotation/fungi/neurospora/). To evaluate its role, we replaced the Patrk-2 gene with a hygromycin B resistance marker (see materials and methods). Unlike the ⌬Patrk-1 strains, the ⌬Patrk-2 strains did not display any obvious potassium transportrelated phenotype at all temperatures or on all tested media. We could not assay membrane polarization by measuring sensitivity to hygromycin B because the marker used to inactivate Patrk-2 was a gene for resistance to this antibiotic. We nonetheless assayed membrane polarization by resistance to TMA and found no significant difference from wild type (Figure 3C). However, we observed incomplete penetrance and variable expressivity of a growth alteration, as observed for ⌬Patrk-1 (Figure 3, A and B). This degenerative process displays many properties similar to those of Wavy. It was not related to any obvious mtDNA alteration (data not shown). Expressivity was strongly influenced by growth conditions to such a point that no alteration could be stably maintained at temperature above 20⬚. Although very similar to the ⌬Patrk-1 growth alter-

ation, the ⌬Patrk-2 growth alteration differed by three main properties. First, penetrance was low, as ⬍1% of the thalli directly issued from ascospores were altered (3 of 407 tested thalli directly presented a growth defect; however, the growth alteration frequently appeared upon subculture). Second, expressivity seemed less variable because the altered thalli were often altered completely and their morphology was less variable. Thalli were usually very spindly, allowing for sectors to resume normally, as observed for ⌬Patrk-1 cultures on medium with high K⫹ concentration. Third, the growth alteration was very unstable and upon replication often disappeared. The double mutant strain ⌬Patrk-1 ⌬Patrk-2 was recovered in the progeny of a cross between the single mutant strains. In every respect, this strain behaved like the ⌬Patrk-1 single mutant. Interestingly, this strain was able to grow normally on medium with very low potassium content, suggesting the presence of at least a third transporter. Wavy is different from Crippled Growth: Crippled Growth is a degenerative process with phenotypic properties similar to those of Wavy. It appears as sectors of slower growth, abnormal hyphae production, and female sterility in strains with fewer translation errors, resulting in a bistable growth pattern with a normal state and a crippled state (Silar et al. 1999). These sectors are caused by the de novo appearance and subsequent spreading of C, a cytoplasmic and infectious element (Silar et al. 1999). The Patrk-1452 mutation was selected as a suppressor of Crippled Growth (see materials and methods). It was thus necessary to check if Wavy was different from Crippled Growth. We searched for the presence of C in Wavy cultures. The growing margin of the AS6-5 tester strain was thus inoculated with donor explants of a Patrk-1452 culture growing in the Wavy form. No Crippled Growth sectors were observed downstream from the donor explants, indicating that the C element was not present in Wavy sectors. Under the same conditions, contamination with explants containing C yields Crippled Growth sectors in about one case out of two (Silar et al. 1999). Further evidence that Wavy was different from Crippled Growth were the observations that Wavy, unlike Crippled Growth, could not be induced in stationary phase and that it was not reverted by stresses, which reverted Crippled Growth (Silar et al. 1999). DISCUSSION

In this article, we present the results of the inactivation of two potassium transporter genes in P. anserina. One of these, Patrk-1, seems to play a major role, since two potassium transport-related phenotypes are obtained when it is inactivated, i.e., impairment of growth in the presence of ammonium and at high temperature. Both defects could be compensated by addition of exogenous potassium. As described for the other transport-

K⫹ Transporter Knockouts in P. anserina

131

Figure 3.—Growth alteration in ⌬Patrk-2. Normal (A) and altered (B) forms of ⌬Patrk-2. (C) Growth rate of wild-type, normal, and altered forms of ⌬Patrk-2 on M2 medium supplemented with tetramethylammonium (TMA) at 18⬚. Experiments were carried out as in Figure 2.

ers of the same family (Madrid et al. 1998), it also entails an increase in plasma membrane polarization. The second, Patrk-2, did not display any phenotypes related to potassium transport, suggesting a minor role in potassium import, at least in the conditions that we explored, or it may have another role unrelated to potassium transport. The fact that deletion of both transporters does not inhibit growth in medium with a low amount of potassium suggests that an additional transporter is present. It could be related to the HAK-1 transporter of N. crassa. We could not clone such a transporter by PCR amplification with degenerated oligonucleotides. A possible cause is the rapid evolution of the primary sequences of this class of transporter. The main conclusion obtained from their inactivation is that both Patrk-1 and Patrk-2 are involved in regulating the growth pattern of the mycelium. When both are present, the mycelium adopts a single morphology; i.e., the mycelium is monostable. However, whenever one of them is absent, the mycelium can adopt various growth patterns, with impact on its overall morphology; i.e., the mycelium is multistable. One growth regimen is like wild type and the resulting mycelium is healthy and able to perform sexual reproduction. In the others, the growth is altered with many different macroscopic aspects, but all forms lack aerial hyphae and are sterile as female. Likely, all the forms reflect a common physio-

logical modification with scaling of expression. Importantly, the multistability leads to an incomplete penetrance and variable expressivity of the growth alteration. This is especially visible in the knockout of Patrk-1, which seems to be the main transporter. The knockouts of Patrk-1 and Patrk-2 thus behave formally as metastable epialleles (Rakyan et al. 2002), but with the important difference that variability is not related to the expression level of Patrk-1 or Patrk-2. The situation is reminiscent of our previous observation that decreasing the translation error level through mutations leads to a bistable growth pattern in P. anserina (Silar et al. 1999). One state is not different from wild type and the other, Crippled Growth, is altered in a way similar to that presented here. We showed that the bistable pattern is caused by the appearance and spreading of C, a cytoplasmic and infectious element (Silar et al. 1999). Our data show that C does not cause the growth alteration presented by Patrk-1 and Patrk-2. It thus seems that bi- or multistable patterns of growth can be promoted by different means in P. anserina and are likely not uncommon. This is an important issue since, had we not searched for it, we would not have detected the growth impairment presented by the ⌬Patrk-2 strains because of its very low penetrance. On an evolutionary perspective, it is possible that this growth alteration, even when present in ⬍1% of the

132


germinating thalli, is sufficient to diminish fitness. It is therefore possible that some genes are selected by evolution only to ensure a correct monostable growth regimen. These would not easily be uncovered by routine phenotypic screen as presently made in some organisms like yeast. Note that we could easily detect this phenomenon in the filamentous P. anserina because the apical growth of thalli recapitulates in a spatial arrangement a temporal succession of events, which are not easily observable in yeast. Both the fact that the ⌬Patrk-1 and ⌬Patrk-2 strains display a similar growth defect and the fact that the growth defect is complemented by two N. crassa transporters with different properties suggest that alteration of potassium transport is related to the multistability of the growth pattern. It is known that ions are involved in morphogenesis in fungi (Shaw and Hoch 2001), but the precise mechanism involved in variability is quite puzzling. The only cause that we can attribute to the growth pattern of ⌬Patrk-1 and ⌬Patrk-2 strains is a stochastic activation of a physiological disorder resulting in growth alteration, but whose primary effectors are yet unknown. Because penetrance is insensitive to external conditions, we can speculate that the trigger of the mechanism is mostly genetically regulated. On the contrary, its expression depends greatly on environmental factors. This is reminiscent of the morphological variation observed in Drosophila melanogaster and Arabidopsis thaliana when the HSP90 protein is inhibited (Rutherford and Lindquist 1998; Queitsch et al. 2002). In the case of HSP90, the observed array of variation has recently been attributed to variation in the chromatin structure (Sollars et al. 2003). Although K⫹ ion can possibly affect chromatin stability, alternative possibilities exist to explain how it may work. Indeed, recently, a post-translational epigenetic mechanism affecting potassium transport has been described in paramecium (Ling et al. 2001). Since genetic analysis is easy in P. anserina, uncovering the underlying mechanism of the variable growth alteration is feasible. Because potassium transport plays a crucial role in the process, mutants defective in this process can be isolated and evaluated for their effect on the growth alteration. Cloning of the relevant genes should identify the molecular actors that participate in establishing the incomplete penetrance and should permit formulation of a testable model to explain the variability of the growth impairment. In conclusion, our data provide important information concerning variation processes during pluricellular growth and the role of a protein related to potassium transporters in their control. Our observations are evocative of the incomplete penetrance and variable expressivity seen in some human channelopathies (Vincent 1998) and other genetic disorders, for which genetic or environmental factors are actively searched for. Such searches may end up being fruitless as differences could

be due to stochastic events during development or early life. We thank F. Malagnac, A. K. Sobering, and C. Vierny for useful discussion and comments on the manuscripts and Luca Pattarello for technical help. This work was supported by grant “Aide aux jeunes e´quipes” from Centre National de la Recherche Scientifique. H. Lalucque is a recipient of a fellowship from the Ministe`re de la Recherche and P. Silar is professor at the University of Paris 7, Denis Diderot. The work was done in compliance with the current laws governing genetic experimentation in France.

LITERATURE CITED Ayoub, N., I. Goldshmidt and A. Cohen, 1999 Position effect variegation at the mating-type locus of fission yeast: a cis-acting element inhibits covariegated expression of genes in the silent and expressed domains. Genetics 152: 495–508. Begel, O., J. Boulay, B. Albert, E. Dufour and A. Sainsard-Chanet, 1999 Mitochondrial group II intron, cytochrome c oxidase, and senescence in Podospora anserina. Mol. Cell. Biol. 19: 4093– 4100. Derkatch, I. L., Y. O. Chernoff, V. V. Kushnirov, S. G. IngeVechtomov and S. W. Liebman, 1996 Genesis and variability of [PSI] prion factors in Saccharomyces cerevisiae. Genetics 144: 1375–1386. Esser, K., 1974 Podospora anserina, pp. 531–551 in Handbook of Genetics, edited by R. C. King. Plenum, New York. Griffiths, A. J. F., J. H. Miller, D. Suzuki, R. C. Lewontin and W. M. Gelbart, 2000 An Introduction to Genetic Analysis. W. H. Freeman, San Francisco. Haro, R., L. Sainz, F. Rubio and A. Rodriguez-Navarro, 1999 Cloning of two genes encoding potassium transporters in Neurospora crassa and expression of the corresponding cDNAs in Saccharomyces cerevisiae. Mol. Microbiol. 31: 511–520. Heytler, P. G., 1963 Uncoupling of oxidative phosphorylation by carbonyl cyanide phenylhydrazones. I. Some characteristics of m-Cl-CCP, action on mitochondria and chloroplasts. Biochemistry 2: 357–361. Levin, M., T. Thorlin, K. R. Robinson, T. Nogi and M. Mercola, 2002 Asymmetries in H⫹/K⫹-ATPase and cell membrane potentials comprise a very early step in left-right patterning. Cell 111: 77–89. Ling, K. Y., W. J. Haynes, L. Oesterle, C. Kung, R. R. Preston et al., 2001 K(⫹)-channel transgenes reduce K(⫹) currents in Paramecium, probably by a post-translational mechanism. Genetics 159: 987–995. Madrid, R., M. J. Gomez, J. Ramos and A. Rodriguez-Navarro, 1998 Ectopic potassium uptake in trk1 trk2 mutants of Saccharomyces cerevisiae correlates with a highly hyperpolarized membrane potential. J. Biol. Chem. 273: 14838–14844. Maillet, L., F. Gaden, V. Brevet, G. Fourel, S. G. Martin et al., 2001 Ku-deficient yeast strains exhibit alternative states of silencing competence. EMBO Rep. 2: 203–210. McCusker, J. H., D. S. Perlin and J. E. Haber, 1987 Pleiotropic plasma membrane ATPase mutations of Saccharomyces cerevisiae. Mol. Cell. Biol. 7: 4082–4088. Navarre, C., and A. Goffeau, 2000 Membrane hyperpolarization and salt sensitivity induced by deletion of PMP3, a highly conserved small protein of yeast plasma membrane. EMBO J. 19: 2515–2524. Olsson, S., 2001 Colonial growth of fungi, pp. 125–141 in Biology of the Fungal Cell, edited by R. J. Howard and N. A. R. Gow. Springer-Verlag, Berlin. Queitsch, C., T. A. Sangster and S. Lindquist, 2002 Hsp90 as a capacitator of phenotypic variation. Nature 417: 618–624. Rakyan, V. K., M. E. Blewitt, R. Druker, J. I. Preis and E. Whitelaw, 2002 Metastable epialleles in mammals. Trends Genet. 18: 348–351. Rizet, G., 1952 Les pheńome`nes de barrages chez Podospora anserina. I. Analyse geńe´tique des barrages entre souches S et s. Rev. Cytol. Biol. Veg. 13: 51–92.

K⫹ Transporter Knockouts in P. anserina Rutherford, S. L., and S. Lindquist, 1998 Hsp90 as a capacitator of morphological evolution. Nature 396: 336–342. Shaw, B. D., and H. C. Hoch, 2001 Ions as regulators of growth and development, pp. 73–89 in Biology of the Fungal Cell, edited by R. J. Howard and N. A. R. Gow. Springer-Verlag, Berlin. Silar, P., 1995 Two new easy to use vectors for transformations. Fungal Genet. Newsl. 42: 73. Silar, P., and M. J. Daboussi, 1999 Non-conventional infectious elements in filamentous fungi. Trends Genet. 15: 141–145. Silar, P., V. Haedens, M. Rossignol and H. Lalucque, 1999 Propagation of a novel cytoplasmic, infectious and deleterious determinant is controlled by translational accuracy in Podospora anserina. Genetics 151: 87–95. Silar, P., H. Lalucque and C. Vierny, 2001 Cell degeneration in the model system Podospora anserina. Biogerontology 2: 1–17.

133

Silar, P., C. Barreau, R. Debuchy, S. Kicka, B. Turcq et al., 2003 Characterization of the genomic organization of the region bordering the centromere of chromosome V of Podospora anserina by direct sequencing. Fungal Genet. Biol. 39: 250–263. Sollars, V., X. Lu, L. Xiao, X. Wang, M. D. Garfinkel et al., 2003 Evidence for an epigenetic mechanism by which Hsp90 acts as a capacitor for morphological evolution. Nat. Genet. 33: 70–74. Sweigard, J., F. Chumley, A. Carroll, L. Farrall and B. Valent, 1997 A series of vectors for fungal transformation. Fungal Genet. Newsl. 44: 52–53. Vincent, G. M., 1998 The molecular genetics of the long QT syndrome: genes causing fainting and sudden death. Annu. Rev. Med. 49: 263–274. Communicating editor: J. J. Loros