HET-s is a prion protein of the filamentous fungus Podospora anserina. An orthologue of this protein, called FgHET-s has been identified in Fusarium graminearum. The region of the FgHET-s protein corresponding to the prion forming domain of HET-s, forms amyloid fibrils in vitro. These fibrils seed HET-s(218–289) fibril formation in vitro and vice versa. The amyloid fold of HET-s(218–289) and FgHET-s(218–289) are remarkably similar although they share only 38% identity. The present work corresponds to the functional characterization of the FgHET-s(218–289) region as a prion forming domain in vivo. We show that FgHET-s(218–289) is capable of prion propagation in P. anserina and is able to substitute for the HET-s PFD in the full-length HET-s protein. In accordance with the in vitro cross-seeding experiments, we detect no species barrier between P. anserina and F. graminearum PFDs. We use the yeast Saccharomyces cerevisiae as a host to compare the prion performances of the two orthologous PFDs. We find that FgHET-s(218–289) leads to higher spontaneous prion formation rates and mitotic prion stability than HET-s(218–289). Then we analysed the outcome of HET-s(218–289)/FgHET-s(218–289) coexpression. In spite of the cross-seeding ability of HET-s(218–289) and FgHET-s(218–289), in vivo, homotypic polymerization is favoured over mixed fibril formation.
Prions are infectious protein particles. Prions have initially been described as disease-causing agents in mammals but a number of fungal prions have also been identified primarily in the yeast Saccharomyces cerevisiae (for recent reviews see Halfmann et al., 2010; Wickner et al., 2010). [Het-s] is a prion of the filamentous fungus Podospora anserina (Coustou et al., 1997) (see Saupe, 2011 for a review).
[Het-s] is involved in a self/non-self-recognition process termed heterokaryon incompatibility. In many if not all filamentous fungi, when cell fusion events occur between genetically unlike individuals, the heterokaryotic fusion cell is destroyed by a cell death reaction or at least severely inhibited in its growth (Glass and Dementhon, 2006). This heterokaryon incompatibility reaction is controlled by specific loci termed het loci. Classically, it is proposed that incompatibility serves to prevent spreading of cytoplasmic mycoviruses between strains and various other forms of parasitism. Recently, an alternative view emerged in which incompatibility is considered a by-product of pathogen driven divergence in genes whose primary function resides in defence against microbial pathogens of fungi (Paoletti and Saupe, 2009). In that view, incompatibility corresponds to an autoimmune condition and should be assimilated to phenomena like graft rejection in mammals and hybrid necrosis in higher plants.
The het-s locus exists as two incompatible alternative alleles termed het-s and het-S (Rizet, 1952; Turcq et al., 1990; 1991). The protein encoded by the het-s allele is a prion (Coustou et al., 1997). Strains containing the protein in its soluble non-prion state are termed [Het-s*]. Strains containing the protein in its assembled prion form are termed [Het-s]. The protein encoded by the alternative het-S allele is incapable of adopting the prion form in vivo. The incompatibility reaction leading to cell death of the fusion cell is triggered when the [Het-s] prion interacts with the antagonistic HET-S protein. Macroscopically, the incompatibility reaction leads to the formation, between the [Het-s] and het-S strains, of an abnormal contact line termed barrage. The [Het-s] prion appears spontaneously at a low frequency and is systematically transmitted to [Het-s*] by cytoplasmic contacts following cell fusion events (Beisson-Schecroun, 1962).
HET-s is a two domain protein; it displays a C-terminal domain spanning residue 218–289, which is both necessary and sufficient for prion propagation (Balguerie et al., 2003). This domain is unstructured in the soluble form of the protein and adopts an amyloid conformation in the assembled prion form of the protein (Balguerie et al., 2003; Nazabal et al., 2003; Ritter et al., 2005; Wasmer et al., 2009a). The structure of the amyloid prion form of HET-s(218–289) has been solved by solid state NMR (Wasmer et al., 2008; Van Melckebeke et al., 2010). It corresponds to a β-solenoid structure with a twofold repetition of a unitary structural motif in each monomer. This unitary motif is composed of four short β-strands delimiting a triangular hydrophobic core (Fig. 1). The N-terminal domain of HET-s is a globular α-helical domain termed HeLo domain, so far only identified in fungal proteins (Greenwald et al., 2010). The HET-s and HET-S proteins differ by 13 residues and the HET-S protein displays the same domain organization. The HeLo domain of HET-S is responsible for the incompatibility reaction (Balguerie et al., 2003; 2004; Greenwald et al., 2010). It is proposed that when HET-S interacts with the amyloid form of HET-s, the C-terminal domain of HET-S is converted to the β-solenoid fold and that β-folding of the C-terminal domain induces a refolding and activation of the toxic HeLo domain (Greenwald et al., 2010).
A HET-s homologue was found in Fusarium graminearum and termed FgHET-s. FgHET-s displays 50% identity to HET-s (55% in the HeLo domain and 38% in the PFD region) (Wasmer et al., 2010). The biochemical and structural properties of the region corresponding to the prion forming domain of FgHET-s have been analysed in vitro. FgHET-s(218–289) forms amyloid fibrils in vitro and these fibrils seed polymerization of HET-s(218–289). Reciprocally, HET-s(218–289) fibrils seed polymerization of FgHET-s(218–289). The structure of the FgHET-s(218–289) fibrils has been analysed by quenched hydrogen/deuterium exchange and solid state NMR. A structure model for FgHET-s(218–289) has been proposed. The structure of the FgHET-s(218–289) fibrils is strongly similar to that of the HET-s(218–289) fibrils in spite of limited primary sequence identity (Fig. 1). This structural similarity readily explains the cross-seeding ability of the two fibril types. Yet, the structures differ in certain aspects, primarily the length of the first β-strands of each pseudo-repeat.
Here, we show that FgHET-s(218–289) is able to behave as a prion forming domain in vivo. We have expressed a FgHET-s(218–289)–GFP fusion protein and a chimeric protein comprising the HET-s HeLo domain and the FgHET-s PFD region in P. anserina and find that both behave as prions in this host. We show that recombinant FgHET-s fibrils display [Het-s]-prion infectivity and that there is no species barrier between [Het-s] and [FgHet-s]in vivo. Then in order to compare the prion performance of the two homologous PFDs, we expressed a FgHET-s(218–289)–GFP fusion protein in S. cerevisiae which has been previously used as a host for HET-s(218–289) prion propagation (Taneja et al., 2007; Mathur et al., 2010). We found that in yeast, FgHET-s(218–289) slightly surpasses HET-s(218–289) in spontaneous prion formation rate and mitotic prion stability. We find that in vivo FgHET-s and HET-s can form distinct aggregates. It appears that the subtle structural differences existing between HET-s and FgHET-s amyloids do not lead to a species barrier but yet are sufficient to favour homotypic polymerization in vivo.
In vitro amyloid formation rates differ for FgHET-s(218–289) and HET-s(218–289)
Recombinant C-terminal histidine tagged FgHET-s(218–289) was produced in Escherichia coli and purified. For practical reasons, previous in vitro fibril formation rates of FgHET-s(218–289) were measured at acidic pH (pH 4.5) to reduce spontaneous fibril formation rates in order to permit analysis of cross-seeding between HET-s and FgHET-s fibrils (Wasmer et al., 2010). Here, we chose to compare fibril formation rates of HET-s(218–289) and FgHET-s(218–289) at pH 7, a more physiological condition. Fibril formation rate was measured by monitoring increase in light scattering at 400 nm. We found that the aggregation rate was higher for FgHET-s(218–289) than for HET-s(218–289) (Fig. 2A). The aggregation curve of FgHET-s(218–289) was biphasic with an initial rapid increase in scattering that occurred in the absence of detectable lag phase followed by the typical fibril growth phase ending in a plateau. In contrast to HET-s(218–289), FgHET-s(218–289) fibrils induce ThT (thioflavine T) fluorescence (Wasmer et al., 2010). So we could measure the fibril formation rate in parallel by light scattering and ThT fluorescence (Fig. 2B). The fibril formation rate measured by ThT fluorescence showed the typical sigmoid shape with a lag phase followed by a rapid fibril growth phase and a plateau. This experiment indicates that the assemblies leading to the initial increase in light scattering do not induce ThT fluorescence and hence are presumably not of amyloid nature. It appears that at neutral pH, FgHET-s(218–289) fibril formation proceeds via the initial formation of non-amyloid oligomers. Whether these oligomers are on- or off-pathway for the amyloid formation process is not currently known.
FgHET-s(218–289) fibrils formed in vitro efficiently induce [Het-s] when transfected into Podospora anserina
A previous study has evidenced that the recombinant FgHET-s(218–289) protein assembles into amyloid fibrils that are structurally similar to HET-s(218–289) fibrils and able to cross-seed HET-s(218–289) amyloid formation in vitro (Wasmer et al., 2010). We wanted to determine whether such fibrils display [Het-s] prion infectivity. We thus determined whether FgHET-s(218–289) fibrils are able to induce prion formation of the wild-type full-length HET-s protein when introduced in vivo. We used recombinant FgHET-s(218–289) protein purified from E. coli assembled into amyloid fibrils to transfect a [Het-s*] strain of Podospora anserina. We found that like HET-s(218–289) fibrils (but unlike Ure2p fibrils), FgHET-s(218–289) amyloids formed in vitro efficiently induce [Het-s] formation in vivo (Table 1). We conclude that recombinant FgHET-s(218–289) fibrils display [Het-s]-prion infectivity.
Table 1. Infection of P. anserina[Het-s*] strains by FgHET-s amyloids assembled in vitro.
Infected strains over total tested
FgHET-s(218–289) can function as a PFD when appended to the GFP
We next asked whether the FgHET-s(218–289) region could function as a prion forming domain in P. anserina. We constructed a plasmid allowing expression of a FgHET-s(218–289)–GFP fusion protein under control of a strong constitutive promotor. This plasmid was used to transform a Δhet-s P. anserina strain and transformants expressing this construct were analysed by fluorescence microscopy. Two phenotypic classes of transformants were obtained. In the first group, the FgHET-s(218–289)–GFP protein formed fluorescent dots and elongated aggregates (Fig. 3). The remaining transformants initially displayed a diffuse cytoplasmic fluorescence. Upon further subculture, they also spontaneously acquired a dot-like fluorescence. When transformants showing fluorescent dots were confronted to transformants with diffuse fluorescence, the latter were converted to the fluorescent dot phenotype. Based, on these experiments, the FgHET-s(218–289)–GFP protein leads to two distinct epigenetic states and the fluorescent dot state can be transmitted through cell fusion events.
Next, we tested whether transformants expressing the dot state of FgHET-s(218–289)–GFP were capable of converting a [Het-s*] strain to the prion phenotype. Confronted to [Het-s*] strains, FgHET-s(218–289)–GFP strains with fluorescent dots, induced systematically the [Het-s] state. FgHET-s(218–289)–GFP strains with diffuse fluorescence did not. Strains with diffuse FgHET-s(218–289)–GFP fluorescence were converted to the fluorescent dot phenotype upon confrontation with [Het-s] strains. We conclude from these experiments that FgHET-s(218–289) can behave as a PFD in P. anserina when appended to the GFP and that the prion form of this protein is able to convert HET-s to the prion state and vice versa.
FgHET-s(218–289) can fully substitute for the HET-s PFD in full-length HET-s
Next, we asked whether the FgHET-s(218–289) region could substitute for the HET-s PFD in full-length HET-s. We constructed a plasmid allowing for expression of a chimeric protein containing the HeLo domain of HET-s appended to the FgHET-s prion forming domain region. This allele we termed het-sFg218–289 was inserted at the resident het-s locus and expressed from the het-s promotor. het-sFg218–289 strains were initially heterokaryon compatible with [Het-s] and [Het-S] and unable to convert a [Het-s*] to the [Het-s] phenotype. Upon confrontation with [Het-s], these strains acquired the ability to form a barrage reaction to het-S strains and to convert [Het-s*] strains to the [Het-s] phenotype (Fig. 4). Thus strains expressing the het-sFg218–289 allele exist in two phenotypic states analogous to [Het-s*] and [Het-s], which we designate [Het-s*]Fg218–289 and [Het-s]Fg218–289. We noted that the barrage reaction of [Het-s]Fg218–289 strains to HET-S developed later as in the case of wild-type [Het-s] (data not shown).
We found no evidence of a species barrier between the two homologous PFDs. [Het-s] strains systematically convert [Het-s*]Fg218–289 to the [Het-s]Fg218–289 phenotype and vice versa [Het-s]Fg218–289 strains systematically convert [Het-s*] strains to the [Het-s] phenotype (Fig. 4). We conclude from these experiments that in spite of extensive sequence divergence, the 218–289 region of FgHET-s can substitute for the HET-s PFD in prion propagation and heterokaryon incompatibility. The FgHET-s(218–289) region can behave as a PFD when appended either to GFP or to the HET-s HeLo domain. In addition, we detected no species barrier between and [Het-s] and [Het-s]Fg218–289.
[Het-s]Fg218–289 spontaneous prion formation rate is lower than for het-s and [Het-s]Fg218–289 prion propagation is PaHsp104 dependent
We compared the wild-type het-s and het-sfg218–289 allele in terms of spontaneous prion formation rate. In the control experiment, 21 out of 512 tested subcultures of wild-type [Het-s*] strains (4 ± 1.4%) acquired the [Het-s] prion phenotype while only 2 out of 512 [Het-s*]Fg218–289 strains (0.4 ± 0.2%), acquired the [Het-s]Fg218–289 prion phenotype spontaneously. Thus spontaneous prion formation rate was lower for the chimeric het-sfg218–289 allele than for wild-type het-s.
We have previously reported that PaHsp104 (the P. anserina Hsp104 homologue) favours [Het-s] prion propagation but that the [Het-s] prion can be maintained in the absence of this chaperone (Malato et al., 2007). We asked whether propagation of [Het-s]fg218–289 chimeric prion is dependent on PaHsp104. We crossed the [Het-s]fg218–289 strain with a ΔPaHsp104 strain and recovered het-sfg218–289ΔPaHsp104 progeny. The het-sfg218–289ΔPaHsp104 strains failed to produce a barrage reaction to het-S and was unable to convert [Het-s*] (or [Het-s*]fg218–289) strains to the [Het-s] phenotype. We conclude from this experiment that in contrast to [Het-s], [Het-s]Fg218–289 cannot be maintained in the absence of PaHsp104. We have also overexpressed PaHsp104 in a het-sfg218–289 strain and found that, as described for [Het-s], [Het-s]Fg218–289 prion maintenance is not affected by PaHsp104 overexpression (Malato et al., 2007).
FgHET-s(218–289)–GFP prion propagation in S. cerevisiae
Next, we determine if the FgHET-s PFD could also permit prion propagation in yeast as was previously shown for HET-s(218–289). We reasoned that expression of HET-s(218–289) and FgHET-s(218–289) in the same heterologous yeast host might permit a direct comparison of the intrinsic prion properties of both PFDs.
A FgHET-s(218–289)–GFP fusion protein was expressed in yeast under control of a galactose-inducible promotor. When expression of FgHET-s(218–289)–GFP was induced by growth on media containing 2% galactose, the FgHET-s(218–289)–GFP fusion protein formed aggregates in a fraction of the cell population. Ring-like aggregates similar to those observed with HET-s(218–289) were detected although their shape was not as regular as in the case of HET-s (data not shown). In the case of HET-s(218–289) pure cultures expressing heritable prion aggregates were obtained by plating cultures grown on 2% gal medium to 0.05% galactose solid medium and by screening for colonies that homogeneously display dot-like GFP fluorescence (Taneja et al., 2007). The same experiment was performed with FgHET-s(218–289)–GFP and colonies with dot-like fluorescence could readily be obtained, indicating that FgHET-s(218–289)–GFP can exist under two epigenetic states in vivo, a diffuse fluorescence state and a heritable aggregated state corresponding to the prion form. We designate the prion form of FgHET-s(218–289)–GFP propagating in yeast [FgHet-s]y. We have verified by Western blot that in 0.05% gal FgHET-s(218–289)–GFP is expressed at the same level as HET-s(218–289)–GFP (Fig. S1), thus allowing a direct comparison between these two prion forming domains in this system.
Yeast propagating the [Het-s]y prion invariably show two dots per cells, one being perivacuolar and the other being juxtanuclear (Mathur et al., 2010). Microscopic examination of cultures in 0.05% gal with heritable aggregates of FgHET-s(218–289)–GFP revealed that the fluorescence pattern was slightly different than that observed for HET-s(218–289)–GFP. Cells expressing HET-s(218–289) in the prion form typically displayed two bright fluorescent dots (1.9 ± 0.7, n = 150) with no background diffuse fluorescence. Cells expressing FgHET-s(218–289)–GFP in the aggregated state generally displayed higher numbers of dots per cell (7.1 ± 2.2, n = 150) and a background fluorescence (Fig. 5).
To determine the rate of prion formation, liquid cultures in either SD or 2% galactose medium were plated 0.05% galactose medium and plates were screened for formation of colonies displaying the prion state of FgHET-s(218–289) or HET-s(218–289). In these conditions, FgHET-s(218–289) cultures in 2% galactose formed colonies with heritable aggregates (dot colonies) at a rate of 4.9% compared with 2% in the case of HET-s(218–289) (Table 2). Dot colonies could also form albeit with a lower frequency when liquid cultures in SD medium were used, a situation that was not observed with HET-s(218–289). With both media, the spontaneous prion formation rate was slightly higher with FgHET-s(218–289) than with HET-s(218–289).
Table 2. Frequency of appearance of colonies with heritable HET-s(218–289) and FgHET-s(218–289)–GFP aggregates.
Medium and strain
Colonies with heritable aggregates/total colony count
Spontaneous rate of prion formation
Values for FgHET-s(218–289) are higher than for HET-s(218–289) with a P-value < 0.001.
Next, we have compared the mitotic stability of the prion form of FgHET-s(218–289)–GFP and HET-s(218–289)–GFP. Mitotic stability was assayed as previously reported for HET-s(218–289)–GFP, the rate of mitotic prion loss was determined after 13–22 generations on solid 0.05% galactose medium (Taneja et al., 2007). Table S1 gives the rate of prion loss per generation for FgHET-s(218–289) and HET-s(218–289). The mitotic loss rate per generation was found to be lower by about one order of magnitude for FgHET-s(218–289)–GFP (3.4·10−4 ± 3·10−4 per generation) than for HET-s(218–289) (4.6·10−3 ± 2·10−3 per generation). Thus, the mitotic stability of the [FgHet-s]y prion is higher than for [Het-s]y.
We conclude that FgHET-s(218–289)–GFP can propagate as a prion in yeast. [FgHet-s]y slightly differs from [Het-s]y by an increased mitotic stability, a higher spontaneous formation rate and a higher number of GFP dots per cell.
FgHET-s and HET-s form distinct polymers in vivo
We found no evidence for a species barrier between FgHET-s(218–289) and HET-s(218–289) in vivo prion transmission assays and in vitro cross-seeding experiments (Wasmer et al., 2010). Yet, structural differences have been identified between the amyloid forms of the two PFDs (Wasmer et al., 2010), raising the question of whether the two proteins are able to co-aggregate. In order to analyse if FgHET-s and HET-s co-aggregate in vivo, we coexpressed both proteins labelled with different fluorescent tags in P. anserina.
Podospora anserina strains coexpressing HET-s-RFP and FgHET-s(218–289)–GFP were obtained and analysed by fluorescence microscopy. In strains coexpressing HET-s-RFP and FgHET-s(218–289)–GFP, in large aggregates there was an apparent colocalization of HET-s-RFP and FgHET-s(218–289)–GFP, in smaller aggregates however distinct localization of HET-s-RFP and FgHET-s(218–289)–GFP was observed (Fig. 6A). In the large aggregates, upon magnification of the aggregate, it appeared that there was not an overlap of the HET-s-RFP and FgHET-s(218–289)–GFP fluorescence. HET-s-RFP and FgHET-s(218–289)–GFP aggregates formed discrete patches that further assembled into higher-order aggregates (Fig. 6B). In addition, in some cells, HET-s–GFP aggregates were found to coexist with diffuse FgHET-s(218–289)–GFP fluorescence and vice versa (Fig. 6A and C).
These experiments indicate that in vivo in P. anserina although FgHET-s and HET-s can colocalize in the same higher-order aggregates, homotypic polymerization is favoured over heterotypic polymerization.
HET-s homologues in other fungal species
Existence of other HET-s homologues was analysed in current fungal genome data sets. We restricted the analysis to proteins containing both a HeLo domain and a region homologous to the HET-s(218–289) PFD region. Evolutionary distribution of close het-s homologues was found to be discontinuous since no such sequences were identified in relatively closely related species like Neurospora crassa or other fully sequenced sordariomycetidae, in contrast het-s homologues were identified in various hypocreales species which are evolutionary more distant to P. anserina. In several species, het-s homologues existed as paralogous gene families. For instance two het-s homologues were found in Fusarium graminearum and F. oxysporum and three are found in Nectria haematococca. An alignment of the 10 strongest homologues of HET-s with a HeLo and a PFD-like domain in the available fungal genomes is given in Fig. S2. Globally, as already noted for FgHET-s, conservation is stronger in the HeLo domain than in the PFD region (Wasmer et al., 2010). The phylogenetically most distant species in which a HET-s homologue (with a HeLo domain and a PFD-like domain) is detected is the dermatophyte Arthroderma otae (a eurothiomycete while all other homologues are found in sordariomyceta).
Position 33 located in the HeLo domain is polymorphic between HET-s and HET-S and was found to be critical to define het-s or het-S-allele specificity (Deleu et al., 1993; Greenwald et al., 2010). This position is a proline in HET-s and a histidine in HET-S. Most homologues encode a histidine like HET-S at that position and thus P. anserina HET-s stands out as an exception. The other residues found at position 33 are either N or R and it is of note that it was shown that H33N and H33R het-S mutants also yield proteins of the het-S specificity (Coustou et al., 1999). Based on this observation, one can speculate that these homologues (including FgHET-s) are likely to represent het-S, rather than het-s, homologues.
In the regions corresponding to the PFD most homologues contain the twofold repeat corresponding to the 21-amino-acid elementary structural unit of the β-solenoid fold. We tried to arbitrarily fit the sequence of each repeat motif of the homologues into the β-solenoid scaffold of HET-s(218–289). In spite of extensive sequence divergence, the sequences of all homologues appear to be compatible with β-solenoid fold described for het-s, as previously suggested (Wasmer et al., 2010) (Fig. S3). In all case, in these speculative models, the interior of the triangular core contains hydrophobic residues, with occasionally a S/T residue found at position 8 or 12 of the motif as seen also in FgHET-s and HET-s. Polar residues are essentially found facing the outside of the core and these positions are generally less well conserved. As already noticed, the asparagine-ladder forming N residues at positions 1 and 18 are well conserved as well as the G residue at position 17. In a number of cases, the two speculative layers of β-strands could be stabilized by ionic interactions between charged residues. When considering the ensemble of the presumed models presented in Fig. S3, of 16 predicted charge interaction, 13 would be complementary and only 3 repulsive. Of note is the fact that in three of the homologues the first repeated motif is shorten by two residues. If these sequences are to assemble into HET-s-like amyloids, this shortening precludes a simple stacking of the homologous motifs. Finally, in the Arthroderma otae sequence (EEQ35355) which is only remotely similar to the HET-s PFD only a single copy of the motif can be identified. This observation might suggest that an ancestral form of the HET-s PFD comprised only a single repeat that was then duplicated in the Sordiaromycetidae.
Lack of species barrier in prion propagation between FgHET-s and HET-s
Often, when prions from one species are used to inoculate a different species, this cross-species transmission leads to low attack rates and prolonged incubation times (which gradually decreases after serial passage in the novel host) or even complete immunity to prions of heterologous origin (see Beringue et al., 2008 for a review). This phenomenon known as species barrier can however sometimes be overcome. The species barrier (or for that matter lack there off) is important both from an applied and from a fundamental point of view. Cross-species transmission of prions has had a high impact on human affairs for instance in the BSE crisis and might also have important implications in the context of CWD epidemic. Then, the understanding of the way primary sequence differences in amyloid proteins affects cross-conversion is of importance in the amyloid field. There has been a semantic drift in the use of the term ‘species barrier’ in yeast prion studies. In most studies, what is actually measured is the capacity of a prion protein from species A to convert a prion protein from species B expressed in the same host (almost invariably S. cerevisiae). We make the same (slightly corrupted) use of the term ‘species barrier’ in the present study. In mammals and in yeast, issues of species barrier are intimately linked to the prion strain phenomenon. As a general rule the level of species barrier is dependent on the strain type and crossing of the species barrier can be accompanied by a strain switching or diversification (Tanaka et al., 2005; Collinge and Clarke, 2007; Beringue et al., 2008; Edskes et al., 2009; Wickner et al., 2009).
Here, we found no evidence of species barrier in prion propagation between HET-s and FgHET-s(218–289). These results are in complete accordance with the previous in vitro studies on amyloid cross-seeding (Wasmer et al., 2010). We however want to emphasize that the efficient cross-species transmission that we describe here between FgHET-s and HET-s does not rule out the existence of a weak species barrier. The in vivo assay that we have employed may not be sensitive enough to detect a low level of species barrier. The same restriction applies to the in vitro cross-seeding assay. With that limitation in mind, we are left with the observation that in spite of a primary sequence identity of 38%, no species barrier was detected in any of our assays. These observations are apparently in sharp contrast with the general rule that very limited sequence divergence can lead to species barrier in particular in mammalian prion systems (Beringue et al., 2008). In yeast, species barrier was detected between Sup35 PFDs displaying 94% identity (Chen et al., 2007; 2010). When the structural information on FgHET-s(218–289) and the nature of amino acid conservation are considered, the lack of species barrier is not all that surprising. Residues critical for the β-solenoid fold are conserved between FgHET-s and HET-s and the structural organization the prion amyloids of the two PFDs appear very similar (Wasmer et al., 2010). Not surprisingly, since we are dealing with structural inheritance, what matters is structure not sequence per se. This primacy of structure over sequence is illustrated in a striking way by the fact that an alternative amyloid fold of HET-s(218–289) formed at low pH, lacks [Het-s] prion infectivity (Sabate et al., 2007) whereas FgHET-s which is only 38% identical to HET-s readily infects [Het-s*] strains. One commonly accepted view of the species barrier is that it depends on the extent of the overlap in the constellation of strains that are accessible or favoured by a given primary sequence (Collinge and Clarke, 2007). No strain variants have been described in the case of [Het-s]. In that respect, the lack of species barrier between highly divergent sequences in the case of [Het-s] might be related to the lack of strain diversity.
Comparison of the prion performances of HET-s(218–289) and FgHET-s(218–289)
Since prion forming ability is conserved in the FgHET-s PFD region it is possible to ask whether there exists a hierarchy between HET-s and FgHET-s in terms of prion behaviour and how potential differences in prion performance might relate to structural differences. The FgHET-s(218–289) and HET-s(218–289) fibrils differ in terms of stability, HET-s(218–289) fibrils are more stable than FgHET-s(218–289) fibrils in face of chemical denaturation with urea or GuHCl. In addition, in vitro fibril formation rates at neutral pH are higher for FgHET-s(218–289) than for HET-s(218–289). The structural basis for this difference in stability and fibril formation rate are currently unknown. Considering that prion performances are critically dependent on fibril fragmentability and growth rate (Tanaka et al., 2006), these physicochemical characteristics (lower stability and faster fibril formation rate) might be expected to provide FgHET-s(218–289) with better prion performances compared with HET-s(218–289). When, the prion behaviour of both PFDs is compared in the same heterologous host, evolutionarily unrelated with either Podospora or Fusarium, we find that indeed spontaneous prion formation rates and meiotic stability are slightly higher for [FgHet-s]y than for [Het-s]y. It is tempting to speculate that lower stability of FgHET-s fibrils might lead to higher number of prion particles in vivo and hence ensure a better meiotic stability. Consistent with this idea, is also the observation than the number of detectable GFP dots is higher with [FgHet-s]y. In [Het-s]y cultures two dots per cell are detected, one is juxtanuclear, then other is perivacuolar and likely correspond respectively to the cellular compartments designated IPOD and JUNK in which protein aggregates accumulate (Mathur et al., 2010). In [FgHet-s]y numerous cytoplasmic dots are detected, it is possible that the presence of these additional dots is related to a lower fibril stability. Similarly, it is tempting to attribute the higher spontaneous formation rate to the increased fibril formation rate measured in vitro. However, one should exert caution before reaching this conclusion. The higher fibril formation rate measured in vivo might be due to the formation of initial non-amyloid oligomers in the in vitro assay. Whether these structures are relevant for the in vivo prion formation process is not clear at present.
At any rate, the in vitro analyses and the in vivo studies carried out in yeast suggest the intrinsic prion properties of FgHET-s(218–289) might be slightly superior to those of HET-s(218–289). Now when the HET-sFg218–289 chimeric protein is compared with full-length wild-type HET-s in P. anserina, a different picture emerges. In that experimental setting, wild-type HET-s actually appears as a more efficient prion than HET-sFg218–289. Spontaneous prion formation rates are higher for [Het-s] than for [Het-s]Fg218–289. In addition, [Het-s] can survive in the absence of the PaHsp104 prion propagation facilitator but [Het-s]Fg218–289 cannot. Several hypotheses can be envisioned to explain these contrasting results. First, one might propose that because the [Het-s]Fg218–289 chimeric prion propagates in a heterologous cellular context, its propagation is slightly disfavoured. We consider this hypothesis unlikely because, [Het-s] prion propagation appears highly promiscuous and not dependent on species-specific host factors as [Het-s] formation readily occurs in the evolutionary distant S. cerevisiae and possibly even in E. coli (Taneja et al., 2007; Wasmer et al., 2009b; Mathur et al., 2010). An alternative explanation could be that the activity of the FgHET-s PFD domain is affected in the HET-sFg218–289 chimera by the HeLo domain of HET-s. The HET-S HeLo domain is able to completely alleviate the activity of the HET-s PFD, thus it is possible to envision that the P. anserina HeLo domain exerts a slightly deleterious effect on the prion forming activity of the FgHET-s PFD (Balguerie et al., 2003). Consistent with this hypothesis of ‘incompatibility’ between heterologous HeLo and PFD domains, is the observation another chimeric protein constructed on the same model, FgHET-spa218–289 (Fusarium Helo domain fused to the HET-S PFD) failed to display the expected activity in heterokaryon incompatibility or prion propagation when expressed in P. anserina (F. Ness, unpubl. results).
Preferential homotypic polymerization in vivo
FgHET-s(218–289) and HET-s(218–289) fibrils efficiently cross-seed and we detect no species barrier in prion propagation between the two orthologues yet in vivo when both proteins are coexpressed there is evidence that FgHET-s and HET-s preferentially self-polymerize. The current structural model of FgHET-s(218–289) could explain both the lack of species barrier and the preferential homotypic fibril growth. In spite of the extensive primary sequence divergence the two structures appear very similar and share overlapping structural elements and conservation of key residues which readily explain the heterologous prion seeding ability (Fig. 1). Yet, the structures differ in certain aspects, such as the length of the first β-strands of each pseudo-repeat, or the distribution of complementary charges allowing for the formation of salt bridges. It can thus easily be conceived that an incoming FgHET-s(218–289) monomer will outcompete a HET-s(218–289) monomer for the incorporation at a FgHET-s(218–289) fibril end and vice versa. This way, in a coexpression situation, the polymerization process is preferentially homotypic (even though the heterologous seeding readily occurs when no competition with the homotypic monomer occurs). This can lead to the paradoxical situation, when FgHET-s aggregates coexist with soluble HET-s or vice versa (Fig. 6C).
Significance of the conversation of the prion forming ability of the FgHET-s PFD
Fusarium graminearum and P. anserina are not closely related species. The two species have divergent an estimated 400 Myr ago (Taylor and Berbee, 2006). This corresponds approximately to the divergence time between human and fish (Benton and Donoghue, 2007). Yet, the prion forming ability has been conserved in the FgHET-s PFD region. This retention of the prion function occurred in spite of extensive sequence divergence but with conservation of key residues critical for the β-solenoid fold (Wasmer et al., 2010) (Fig. S3). There is thus a clear indication of the existence of a selective pressure for the maintenance of the ability to form that infectious fold. The simplest explanation regarding the nature of that selective pressure would be that het-s homologues in F. graminearum and by extension in other Fusaria also function as heterokaryon incompatibility genes. This would obviously require that the Fusarium populations are polymorphic for the het-s locus and that an alternative allele of Fghet-s (of the het-S-type) exists. There has been no survey of Fusarium populations in search of such polymorphism, so such an allele has yet to be found. So currently we lack evidence for the existence of a het-s-type allele in Fusarium which is a necessary condition for evoking that heterokaryon incompatibility function is responsible for maintenance of the prion forming ability in FgHET-s. In fact, preliminary evidence suggests that that the FgHET-s allele from the sequenced reference strain behaves an allele of the het-S type (S.J. Saupe and R. Sabaté, unpubl. results). If FgHET-s is in fact a het-S-type allele, then it will be incompetent for prion formation in vivo and its PFD region can only function as a PFD in the absence of the prion inhibitory effect of its HeLo domain (Balguerie et al., 2003; Greenwald et al., 2010). Thus, we do not claim here that the naturally occurring full-length FgHET-s protein behaves as a prion but that the 218–289 region of the protein has maintained the ability to adopt the β-solenoid fold and to behave as a prion forming domain.
Strains, plasmids and media
Podospora anserina strains used in this study were wild-type het-s, het-S, Δhet-s strains and the ΔPaHsp104 strain (Malato et al., 2007). Growth medium for barrage essays and prion transmission assays was standard corn meal agar DO medium. The Δhet-s strain was constructed by inserting the nat1 cassette from the pAPI508 plasmid (El-Khoury et al., 2008) in place of the het-s ORF. The nat1 cassette was amplified with oligonucleotides 5′-CTTCCCTTCCACTTCTTCACAC-3′ and 5′-ATCCTAGATGACTTAAGACGACAGG-3′. The sequences upstream and downstream of the het-s ORF were amplified respectively with oligonucleotide pairs 5′-AAGCTTTTCGAATTGGTCTCTCAG-3′ and 5′-GGGCAGTTTGAGGGGAAAGCGAAG-3′ and 5′-GGGACTAGTACCCTCCAGCAAGGATAGC-3′ and 5′-GCGGCCGCCATGGGCACTGCATCTGGG-3′. The fragments were ligated to create the het-s::nat-1 cassette cloned as a HindIII–NcoI fragment in a pSK plasmid. This cassette was used to transform a het-sΔPaKu70 strain (El-Khoury et al., 2008). Nourseothricine-resistant transformants compatible with a het-S-tester were selected. Inactivation of het-s was verified by PCR using oligonucleotides 5′-CGACGATCACAGCTATAGCGTGGTG-3′ and 5′-ATCCGGCTTCCCTGGACCTGCTTC-3′. The strain was then backcrossed once to a het-s wild-type strains and Δhet-s (NourR and het-S compatible) Δhet-sΔKu70 strains (NourR, PhleoR and het-S compatible) were selected in the progeny.
pGPD-FgHet-s(218–289)–GFP plasmid was constructed by amplifying the region coding for the 218–289 region of FgHET-s with oligonucleotides 5′-CTCATGAGATCTTCCCAGATGCCTCTGC-3′ and 5′-GGTCATGAAGTTGAACATGATCGAGGGG-3′. The PCR product was then digested with BspHI in inserted in the unique NcoI site of the pAN–GFP plasmid (Dementhon et al., 2003). The pGPD-HET-s-RFP plasmid was constructed as follows. In a first step, the mRFP sequence was cloned as an EcoRI/BamHI restriction fragment into the pCB1004 vector (Carroll et al., 1994). Then the sequence corresponding to the GPD promotor and the het-s ORF was amplified using oligonucleotides 5′-AACCGCGGAATTCCCTTGTAT-3′ and 5′-CGTCTAGATTATCCCAGAACCC-3′ using the pGPD-het-s plasmid as template (Balguerie et al., 2003). The PCR fragment was then inserted as a SacII/XbaI fragment into the pCB1004 vector digested with SacII and SpeI.
The het-sfg218–289 chimeric allele was constructed by amplifying the pUC-het-s plasmid (HindIII–KpnI fragment of the het-s locus insert in pUC18) with oligonucleotides 5′-GTTATGCCCCTCGATCATGTTCAACTTCTGCGCAGCCGCATCAGACATAGC-3′ and 5′-GAAGATCTCGGCAGGTAGATAGATTGTGGCC-3′ in an inverse PCR. Then, the region coding for the FgHET-s 218–289 region was amplified using oligonucleotides 5′-GCTATGTCTGATGCGGCTGCGCAGAAGTTGAACATGATCGAGGGGCATAAC-3′ and 5′-GAAGAGATCTTTAATCTTCCCAGATGCCTCTGCCC-3′. Both PCR products were then mixed and reamplified in PCR product extension reaction with oligonucleotides 5′-GAAGATCTCGGCAGGTAGATAGATTGTGGCC-3′ and 5′-GAAGAGATCTTTAATCTTCCCAGATGCCTCTGCCC-3′. The final PCR product was restricted with BglII and circularized by ligation with T4 ligase. A het-s-gene replacement cassette was constructed using the following strategy. A hph cassette was amplified using oligonucleotides 5′-GGTACCGTCGACAGAAGATGATATTG-3′ and 5′-ATCTCTAGAAAGAAGGATTACCTC-3′ using pCSN43 plasmid as template. The hph cassette was then cloned downstream of the het-sfg218–289 chimeric allele. Then, a 652 bp fragment downstream of the het-s ORF was cloned downstream of the hph gene. This gene replacement cassette was then used to transform a Δhet-sΔKu70 strain and hygR NourS strains were selected. The gene replacement event was then controlled by PCR and the strain was het-sfg218–289::het-s was backcrossed to a het-S wild-type strain to eliminate the ΔKu70 mutation.
Yeast was grown on YPD. The transformants were grown on synthetic glucose media lacking tryptophan (SD-Trp) or synthetic media-Trp containing 2% raffinose (SR-Trp). To induce the expression of HET-s constructs, yeast cells were grown in SR-Trp containing 2% galactose (SR-Trp + 2% galactose) for high expression and SR-Trp containing 0.05% galactose (SR-Trp + 0.05% galactose) for low expression. Yeast cells were grown at 30°C in all media. Experiments were carried out with strain L1749 (Taneja et al., 2007).
pHet-s(PrD)–GFP-TRP1 is a centromeric plasmid encoding a fusion of GFP at the C-terminus of the PrD (amino acid 218–289) of HET-s. pFgHET-s(218–289)–GFP plasmid was constructed as a derivative of pHet-s(PrD)–GFP-TRP1 by Gap repair. The region coding for amino acids 218–289 in the FgHET-s sequence was amplified with oligonucleotides 5′-TTCTATAGACACGCAAACACAAATACACACACTAAATTACCGGATCTATGAAGTTGAACATGATCGAG-3′ and 5′-TCAACCAAAATTGGGACAACACCAGTGAATAATTCTTCACCTTTAGACATATCTTCCCAGATGCCTCTG-3′. The L1749 yeast strain (Taneja et al., 2007) was then were then transformed with the PCR product and pHet-s(PrD)–GFP-TRP1 linearized with BamHI. The pFgHET-s(218–289)-RFP plasmid was constructed as a derivative of pFgHET-s(218–289)-RFP by Gap repair. The region coding for the DsRed was amplified with oligonucleotides 5′-CAGGTTGGGAATGTTTATGGGGGCAGAGGCATCTGGGAAGATATGGTGGCCTCCTCCGAGGAC-3′ and 5′-GAAGCACCACCACCAGTAGAGACATGGGAGATCCCCCGCCTACAGGAACAGGTGGTG-3′. The L1749 yeast strain (Taneja et al., 2007) was then were then transformed with the PCR product and pFgHET-s(218–289)–GFP plasmid linearized with EcoRI.
Mitotic stability of the [Het-s]y and [FgHET-s]y prions in yeast was assayed as previously described (Taneja et al., 2007). In brief, [Het-s]y and [FgHET-s]y liquid cultures in 0.05% gal medium were plated onto 0.05% gal plates. After several days at 30°C, whole colonies were recovered and total cell count in each colony was determined to determine the number of generations of growth. Resuspended colony were then replated on 0.05% gal plates and after 24 h of growth, microcolonies were observed directly on the growth plate and the fraction of dot and diffuse-fluorescence colonies was determined. The lost rate per generation (θ) was calculated using the following formula from Boe and Rasmussen (1996):
where ψ + (g) is the fraction of cells containing the element after g generations, which is the fraction of dot-colonies over total colony count.
Prion propagation, prion formation and incompatibility assays
Methods for determination of incompatibility phenotypes, prion formation and prion propagation were as previously described (Benkemoun et al., 2006). In brief, incompatibility phenotypes were determined by confronting strains of solid corn meal agar medium to [Het-s] and [Het-S] tester strains and visualizing the formation of barrages (abnormal contact lines forming upon confrontation of incompatible strains). [Het-s] prion propagation was assayed as the ability to transmit the [Het-s] prion from a [Het-s]-donor strain to a [Het-s*] prion-free tester strain after confrontation on solid medium. Prion formation rates of [Het-s] and [Het-s]Fg218–289 were determined by measuring the fraction of [Het-s*] or [Het-s*]Fg218–289 subcultures that spontaneously acquired the prion phenotype after 5 days of growth at 26°C.
For fluorescence microscopy, synthetic medium containing 2% (w/v) agarose was poured as two 10 ml layers of medium. P. anserina hyphae were inoculated on this medium and cultivated for 16–24 h at 26°C. The top layer of the medium was then cut out and the mycelium was examined with a Leica DMRXA microscope equipped with a Micromax CCD (Princeton Instruments) controlled by the Metamorph 5.06 software (Roper Scientific). The microscope was fitted with a Leica PL APO 100× immersion lens.
Yeast cells were analysed with the same equipment. Yeast microcolonies observed directly on solid medium overlaid with a microscope slide. Similarly cells from liquid cultures were spotted on solid medium for microscopy observation.
Protein methods and protein transfection assays
HET-s(218–289) and FgHET-s(218–289) proteins were expressed in E. coli and purified as previously described. Both proteins had a C-terminal 6 histidine tag and expressed as insoluble proteins and purified under denaturing conditions using Qiagen columns. Yields were in the range of 10 mg l−1 of culture. Proteins were eluted in 6 M GuHCl 50 mM Tris–HCl pH 8, 150 mM NaCl, 200 mM imidazole. Elution buffer was replaced by 175 mM acetic acid by passage on a 5 ml Hitrap column (Amersham). Fibril formation is then triggered by a pH shift to pH 7.
The aggregation of 10 µM PaHET-s(218–289) and FgHET-s(218–289) soluble monomers were carried out at 25°C and pH 7 (in a 1:1 mixture of 175 mM acetic acid and 1 M Tris/HCl, pH 8) under constant agitation. For determining the presence of amyloid and non-amyloid aggregates in both aggregation kinetics, 25 µM thioflavin-T (Th-T), a widely used amyloid-specific dye, was added. The kinetics were followed using a Cary Eclipse spectrofluorimeter (Varian) exciting at 445 and 400 nm and recording each 1 min at 480 and 400 nm for the detection of amyloid and non-amyloid species respectively.
Protein transfection experiments were performed using a cell disruptor (Fast-prep FP120, Bio101, Qbiogen). For each test, ∼0.5 cm3 of [Het-s*] mycelium grown on solid medium is sheared (run time 30 s, speed 6) in 500 µl of STC50 buffer (0.8 M sorbitol, 50 mM CaCl2, 100 mM Tris HCl pH 7.5) and the sonicated HET-s(218–289) amyloids assembled at pH 7 (20 µl at 1 mM) in a 2 ml screw cap tube. The sheared mycelium is then diluted with 600 µl of STC50 buffer and then plated as serial dilutions onto DO-0.8 M sorbitol medium and incubated at 26°C until being confluent (7–8 days). Then several implants (at least two per mycelium) are checked for the [Het-s] phenotype in barrage tests.
This work was supported by a grant from ANR (‘PANPRIONDRUGS’). We thank Martine Sicault for technical assistance.