Neurospora Heterokaryons with Complementary Duplications and Deficiencies in Their Constituent Nuclei Provide an Approach to Identify Nucleus-Limited Genes

Introgression is the transfer of genes or genomic regions from one species into another via hybridization and back-crosses. We have introgressed four translocations (EB4, IBj5, UK14-1, and B362i) from Neurospora crassa into N. tetrasperma. This enabled us to construct two general types of heterokaryons with mat-A and mat-a nuclei of different genotypes: one type is [T + N] (with one translocation nucleus and one normal sequence nucleus), and the other is [Dp + Df] (with one nucleus carrying a duplication of the translocation region and the other being deleted for the translocation region). Self-crossing these heterokaryons again produced [T + N] and [Dp + Df] progeny. From conidia (vegetative spores) produced by the heterokaryotic mycelia, we obtained self-fertile (heterokaryotic) and self-sterile (homokaryotic) derivative strains. [T + N] heterokaryons produced homokaryotic conidial derivatives of both mating types, but [Dp + Df] heterokaryons produced viable conidial homokaryons of only the mating type of the Dp nucleus. All four [T + N] heterokaryons and three [Dp + Df] heterokaryons produced both self-sterile and self-fertile conidial derivatives, but the [Dp(B362i) + Df(B362i)] heterokaryons produced only self-sterile ones. Conceivably, the Df(B362i) nuclei may be deleted for a nucleus-limited gene required for efficient mitosis or nuclear division, and whose deficit is not complemented by the neighboring Dp(B362i) nuclei. A cross involving Dp(EB4) showed repeat-induced point mutation (RIP). Because RIP can occur in self-crosses of [Dp + Df] but not [T + N] heterokaryons, RIP alteration of a translocated segment would depend on the relative numbers of [Dp + Df] vs. [T + N] ancestors.


described three elementary chromosome translocation (T) types in
Neurospora crassa, namely, insertional (IT), quasiterminal (QT) and reciprocal (RT). ITs transfer a segment of a donor chromosome into a recipient chromosome without any reciprocal exchange, QTs transfer a distal segment of a donor chromosome to the tip of a recipient chromosome, distal to any essential gene, and presumably the donor chromosome breakpoint is capped with the tip from the recipient chromosome; and RTs reciprocally interchange the terminal segments of two chromosomes. Other chromosome rearrangements are essentially variants of these, e.g., an intrachromosomal transposition (Tp) is an IT in which the same chromosome is both donor and recipient, an inversion (In) is a Tp in which a chromosome segment is re-inserted in opposite orientation into the site from which it was derived, and there are complex rearrangements such as linked RT and IT. Three breakpoint junctions define an IT, viz, junction A created by the deletion on the donor chromosome, and junctions B and C (proximal and distal), created by the insertion into the recipient chromosome, whereas two breakpoint junctions define a QT or RT; junction A, between the breakpoint-proximal segment on the donor chromosome and the tip from the recipient chromosome, and junction B, between the breakpoint-proximal sequence on the recipient chromosome and the donor segment grafted onto it ). In the cross of an IT or QT strain with normal sequence (IT x N or QT x N), alternate segregation produces eight parental-type ascospores (i.e., 4 T + 4 N), whereas adjacent 1 segregation generates non-parental ascospores, viz, four viable ascospores containing a duplication (Dp) of the translocated segment and four inviable ones with the complementary deficiency (Df). Viable ascospores blacken (B), whereas inviable ones remain white (W). Therefore, alternate and adjacent 1 segregation produce, respectively, 8B:0W and 4B:4W asci. Since both segregations are equally likely, IT x N and QT x N crosses are characterized by 8B:0W = 4B:4W, whereas isosequential crosses (i.e., N x N or T x T) produce mostly 8B:0W asci (Perkins 1997). Adjacent 1 segregation in RT x N, yields only inviable ascospores bearing complementary duplications and deficiencies (i.e., Dp1/Df2 and Dp2/Df1), and the asci are 0B:8W. Therefore, 8B:0W = 0B:8W signals RT x N.
Dp strains (the viable segregants from 4B:4W asci) are recognizable by the characteristic barren phenotype they impart to Dp x N crosses, wherein normal looking perithecia are made, but only a few exceptional ascospores are produced (Perkins 1997). Barrenness is caused by meiotic silencing by unpaired DNA (MSUD), an RNAi-mediated process that eliminates the transcripts of any gene not properly paired during meiosis with a homologue at an allelic position (Shiu et al. 2001). Presumably, Dp-borne genes, including those underlying ascus and ascospore development, fail to properly pair in a Dp x N cross, and their silencing by MSUD renders the cross barren. The breakpoint junctions of several ITs, QTs, and RTs were defined in our laboratory (Singh 2010;Singh et al. 2010). PCR with breakpoint junction-specific primers can now be used to distinguish the Dp progeny from their T and N siblings. IT progeny contain all three breakpoints (A, B, and C), Dp progeny contain B and C, but not A, and N progeny contain none. While Dps have been extensively studied (Perkins 1997;Kasbekar 2013), the use of Dfs was limited to flagging the Dp-bearing 4B:4W asci. We now report the generation of [Dp + Df] heterokaryons with complementing duplications and deficiencies in their constituent nuclei.
They were obtained by introgressing N. crassa ITs (and a QT) into N. tetrasperma. Introgression is the transfer of genes or genomic regions from one species into another (Rieger et al. 1991).
Eight ascospores form per ascus in N. crassa, whereas four are formed in N. tetrasperma.
In both species the parental mat A and mat a nuclei fuse in the ascogenous cell to produce a diploid zygote nucleus that immediately undergoes meiosis and a post-meiotic mitosis to generate eight haploid nuclei (4 mat A + 4 mat a). In N. crassa the nuclei are partitioned into the eight initially uninucleate ascospores (Raju 1980). In contrast, N. tetrasperma ascospores are initially binucleate, receiving a pair of non-sister nuclei (1 mat A + 1 mat a) (Raju and Perkins 1994). Thus, N. crassa ascospores produce homokaryotic mycelia that are either mat A or mat a in mating type, whereas N. tetrasperma ascospores can produce heterokaryotic mycelia with nuclei of both mating types. In N. crassa, a sexual cross perforce requires mycelia from two different ascospores, one mat A, the other mat a, thus making the lifecycle "heterothallic"; whereas a heterokaryotic N. tetrasperma mycelium from a single ascospore bearing nuclei of both mating types is competent to undergo a self-cross, making the lifecycle "pseudohomothallic". However, a subset of conidia (vegetative spores) produced by a heterokaryotic N. tetrasperma mycelium can be homokaryotic by chance, and N. tetrasperma ascogenesis occasionally produces five or more (upto eight) ascospores, by replacement of one or more dikaryotic ascospore by a pair of smaller homokaryotic ones (Raju 1992). The dominant Eight-spore (E) mutant increases the frequency of such replacement, although E-homozygous crosses are infertile (Calhoun and Howe 1968). Mycelia from homokaryotic conidia or ascospores can cross with like mycelia of the opposite mating type. Therefore, N. tetrasperma is actually a facultatively heterothallic species.
Here, we have introgressed three N. crassa ITs (EB4, IBj5, and B362i) and one QT (UK14-1) into N. tetrasperma, and shown that T x N crosses produce both [T + N] and [Dp + Df] heterokaryotic progeny. We found that unlike the other heterokaryons, the [Dp(B362i) + Df(B362i)] heterokaryons produced only self-sterile conidia, possibly because the Df(B362i) nuclei are missing a putative "nucleus-limited" gene required for efficient mitosis and nuclear division. Additionally, we show that a Dp-heterozygous cross can exhibit RIP (repeat-induced point mutation), the sexual stage-specific process that induces G:C to A:T mutations in duplicated DNA (Selker 1990).
The semi-dominant MSUD suppressor strains Sad-1 A (FGSC 8740) and Sad-1 a (FGSC 8741), were gifted by Robert L. Metzenberg and are described by Shiu et al. (2001). The sad-1 locus encodes an RNA-dependent RNA polymerase essential for MSUD, and the Sad-1 suppressor allele is presumed to prevent proper pairing of its wild-type homologue thus inducing it to autogenously silence itself (Shiu et al. 2001). The MSUD testers pan-2; his-3::his-3 + Bml r A (FGSC 8755); pan-2; his-3::his-3 + Bml r a (FGSC 8756); pan-2; his-3::his-3 + mei-3 + A (FGSC 8759); and pan-2; his-3::his-3 + mei-3 + a (FGSC 8760) (hereafter designated as :: Bml r A, :: Bml r a, ::mei-3 A, and ::mei-3 a) are described by Raju et al. (2007). Another tester, rid his-3; VIIL::r ef2 -hph A (ISU3117) (hereafter ::r ef2 ) was a gift from Dr. Tom Hammond (University of Missouri). The ::Bml r and ::mei-3 testers have an extra copy of the bml (β-tubulin) or mei-3 gene inserted ectopically in the his-3 locus in chromosome 1, while the tester strain ::r ef2 has a copy of the r (round spores) gene inserted ectopically into chromosome 7. In a cross of the tester with wild type, the ectopic copy remains unpaired in meiosis and results in elimination of all its homologous transcripts, including from the paired endogenous copies. In crosses of ::Bml r , ::mei-3 and ::r ef2 with the wild type, the bml, mei-3, and r genes, respectively, are silenced. Silencing of the bml or mei-3 gene arrests normal ascus development (Raju et al. 2007;Kasbekar et al. 2011), and silencing of r causes all eight ascospores to be round instead of the normal spindle shaped. Homozygous tester A x tester a crosses do not show MSUD, nor do crosses of the testers with the Sad-1 suppressor of MSUD, and the asci developed normally (Raju et al. 2007;Kasbekar et al. 2011 four N. crassa great-grandparents and four N. tetrasperma great-grandparents (Metzenberg and Ahlgren 1969). The N. crassa great-grandparents were of the OR background, whereas the N. tetrasperma great-grandparents were of the 343.6 A E background (Metzenberg and Ahlgren 1969).
Neurospora genetic analysis was done essentially as described by Davis and De Serres (1970). Metzenberg's (2003) alternative recipe was used for making Medium N.

Outline of the introgression crosses and characterization of the resultant strains.
Crosses between N. crassa and N. tetrasperma strains are almost completely sterile. However, both N. crassa strain OR A and N. tetrasperma strain 85 A can cross with the N. crassa / N. tetrasperma hybrid strain C4,T4 a and produce viable progeny (Perkins 1991; also see Table 3 of this paper) therefore we used the C4,T4 a strain as a bridging strain for the initial introgression crosses. The N. crassa T strains were crossed with C4T4 a and T progeny from these crosses (designated T 1xC4T4 ) were distinguished from their Dp and N siblings by PCR with breakpoint junction-specific primers. Nominally, 50% of the genome of T 1xC4T4 progeny is derived from the C4,T4 a parent. The T 1xC4T4 A strains were crossed with C4,T4 a, to obtain the T 2xC4T4 progeny in a like manner. Crosses of T 2xC4T4 with the opposite mating type derivative of strain 85 were productive, and their T progeny were designated as T 1x85 . Likewise, T 1x85 x 85 yielded T 2x85 , etc.
After two to three iterations of the crosses with 85, we recovered progeny ascospores that produced mycelium of dual mating specificity characteristic of N. tetrasperma. That is, the resulting mycelium could cross with both 85A and a, and it could also undergo a self-cross. A heterokaryotic strain containing all three breakpoints (A, B and C) is potentially of genotype The [T + N] and [Dp + Df] heterokaryons are distinguishable, since the former produces homokaryotic conidial derivatives of both mating types, whereas the latter produces viable homokaryons of only the mating type of the Dp nucleus. Conidia from self-fertile heterokaryotic strains were streaked onto Vogel's -FGS medium, and well-isolated conidial germlings were transferred to SCM to distinguish self-fertile (heterokaryotic) from self-sterile (homokaryotic) conidial derivatives. The mating type of the self-sterile conidial derivatives was determined by crossing to the single mating type derivatives 85 a and 85 A. If all the self-sterile conidial derivatives are of a single mating type, then the heterokaryon from which they were derived is The results were confirmed by PCR with primers for the breakpoint junctions and mat ideomorphs and DNA of the homokaryotic conidial derivatives.

Four homokaryotic T type conidial derivatives from the self-fertile [T + N]
heterokaryons, namely, T(EB4) Nt a from the heterokaryon 3E1 (Table 2 serial number 3), (Table 2 serial number 18), T(UK14-1) Nt a from U9 (Table 2 serial number 24), and T(B362i) Nt A from 19B7 (Table 2 serial number 35) were used in the experiments whose results are summarized in Tables 3 and 4.
The introgression of these translocations into N. tetrasperma is outlined in the "Materials and  (EB4)] also fulfill this criterion, but, they were deemed to be less likely since one or more crossover is required to generate them. Eight self-fertile progeny from the self-cross of E1 were analyzed and the results, summarized in From self-crosses of strains 6E1 and 9E1 (see above) we examined 39 and 24 progeny, and found, respectively, 12 and nine were self-fertile. From a subset of self-fertile progeny we
As a control, we performed the cross Dp(B362i) a x 85 A and 39 of 61 progeny tested were self-fertile. We examined ~10 conidial derivatives from each of 10 self-fertile progeny, and in every case at least three were self-fertile, and among the self-steriles we found both mating types. This suggested that unlike [Dp(B362i) + Df (B362i) tetrasperma strain 85 were fertile, whereas their crosses with N. crassa OR strains of opposite mating type were as infertile as the interspecies OR x 85 cross (Table 3). Control crosses of the N. crassa T strains (T Nc ) with the OR strains of the opposite mating type were productive, but the crosses of the T Nc strains with the opposite mating type strain 85 derivatives were sterile (Table   3). The C4,T4 a hybrid strain produced viable ascospores in crosses with both OR A and 85 A (Table 3).
To obtain larger numbers of eight-spored asci, we crossed the T strains with E strains of the opposite mating type. The T(B362i) Nt x E crosses were infertile, and T(IBj5) Nt x E produced mostly white inviable ascospores. However, T(EB4) Nt x E and T(UK14-1) Nt x E were productive, and the 8:0 and 4:4 ascus types were produced at comparable frequencies (Table 4), which is characteristic of IT x N and QT x N crosses in N. crassa.

RIP in a Dp-heterozygous N. tetrasperma cross. In N. crassa, crosses involving Dp
strains can generate RIP-induced mutant progeny . Therefore, we expected crosses of N. tetrasperma Dp strains also would yield RIP-induced mutants. Dp(EB4) duplicates the ad-7 (adenine requiring-7) gene (Perkins 1997). Ascospores from Dp(EB4) a x E A were germinated on adenine-supplemented Vogel's-FGS medium, 130 germlings were picked to adenine-supplemented Vogel's-glucose medium, and then their growth was tested on unsupplemented Vogel's-glucose medium. Three adenine-requiring auxotrophic strains were identified among 125 progeny examined. One was a heterokaryon, but the other two were N type homokaryons. In both the homokaryons, the ad-7 locus was altered by several RIP mutations (G:C to A:T transitions) ( Table 5). Presumably, the Esm phenotype of C4,T4 a is derived from its 343.6 A E ancestor.

MSUD is relatively weak in crosses involving C4,T4 a and 85 A/a. Ramakrishnan et al.
In N. crassa, crosses of Dp(EB4) and Dp(IBj5)strains with the OR type strains were barren, whereas their crosses with the Sad type strains were fertile (Ramakrishnan et al. 2011).
Most Esm type strains gave a fertile cross with Dp(EB4) and a barren cross with Dp(IBj5), although some Esm type gave barren crosses with both the Dps, and fewer still gave fertile crosses with them. Other results of Ramakrishnan et al. (2011) suggested that N. tetrasperma strain 85 is Esm type. We crossed the N. tetrasperma Dp(EB4) and Dp(IBj5) strains with opposite mating type derivatives of strain 85 and found the crosses were as productive as those of T(EB4) and T(IBj5) with strain 85 (data not shown). These results are consistent with the classification of strain 85 as Esm or Sad type.

Discussion.
We mutant with the hope that he could rescue the [Dp + Df] heterokaryon. He had previously used this approach to rescue Sk S nuclei from a Sk S x Sk K cross in [Sk S + Sk K ] heterokaryotic ascospores (Raju 1979), but analysis of progeny nuclei in the mixed cultures from the later experiment was not easy, and his efforts were inconclusive (personal communication from Dr. N. B. Raju to DPK). D. D. Perkins also alluded to his obtaining [Dp + Df] heterokaryons (Perkins 1997), but his results remained unpublished. Therefore, to the best of our knowledge, this is the first report of heterokaryons with complementary duplications and deficiencies in their constituent nuclei.
The introgressions were facilitated by use of the C4,T4 a bridging strain (Perkins 1991).
We found that C4,T4 a is a moderate suppressor of MSUD (Supplementary Figure 1). Shiu et al. (2001) had shown that the N. crassa MSUD suppressor Sad-1 can partially breach the N. crassa / N. tetrasperma interspecies barrier. Although C4,T4 a does not suppress MSUD as strongly as Sad-1 a, its moderate suppressor phenotype might contribute to increase ascospore production in crosses with both N. crassa and N. tetrasperma strains. The N. crassa T strains were first crossed with C4T4 a, and T A progeny from this cross were crossed with C4T4 a. T progeny from the latter crosses were used to initiate two or three additional rounds of crosses with opposite mating type derivatives of N. tetrasperma strain 85. Although this sequence of introgression crosses can be continued indefinitely, we stopped once they began yielding selffertile progeny. Nominally, one nucleus in the B7, E1, I1, I2, I3, I4, I5, and U9 self-fertile progeny of Figure 1 derives 1/8 or 1/4 of its genome from the non-85 genetic background (N. crassa T or C4T4 a), and, depending on the translocation, this fraction is enriched for sequences from the translocation chromosome (1, 4, 5, 6, or 7). Therefore, in principle, by sequencing one can identify strain 85 genome segments that are replaceable by non-85 without losing the pseudohomothallic life-cycle.

Self-crosses of the [T + N] / [Dp + Df] strains also appeared to produce [T + Df ] and [N +
Df] genotypes. These genotypes can be produced by a breakpoint-proximal crossover on the donor or recipient chromosome followed by second-division segregation of the resulting Df-or N-bearing chromosome into both mat a and mat A nuclei. Breakpoint-proximal crossover produces 6:2 ascus types in IT x N and QT x N crosses in N. crassa (Perkins 1997), and in IT x E and QT x E crosses in N. tetrasperma (Table 4). Second-division segregation can also generate [Ta + TA], [Dp a + Dp A], [N a + N A], [N + Dp], and [T + Dp] progeny types. One self-fertile heterokaryon had the genotype [T a + NA + Df A]. This triple heterokaryon could have arisen from an interstitial cross-over, that in N. crassa would have produced a 6:2 tetratype ascus with two nuclei each of the T a, N A, Dp a, and Df A genotypes (see Figure 1 in Perkins 1997).
Presumably, three nuclei (T a, N A, and Df A) instead of two were partitioned during ascus development into the ascospore from which this strain was derived. In N. crassa crosses heterozygous for the Fsp-1 (Four-spore-1) or Fsp-2 (Four-spore-2) mutant, rare two-and threespored asci are occasionally formed that contain heterokaryotic ascospores (Raju 1986;Perkins et al. 2001). Presumably, our [T a + NA + Df A] strain had a similar provenance.
A Dp-borne gene (ad-7) was shown to be alterable by RIP. Since RIP occurs in self- However, the [Dp(B362i) + Df(B362i)] heterokaryons appeared to be an exception, in that, none of the six strains tested produced any self-fertile conidial derivatives. One hypothesis to account for these results is that Df(B362i) nuclei divide much less efficiently than Dp(B362i) nuclei, resulting in a rapid dwindling of their number in the [Dp(B362i) + Df(B362i)] heterokaryon.
Consequently, they are less likely to be packaged along with the Dp(B362i) nuclei during conidiation, when typically three to ten (presumably randomly picked) nuclei are partitioned into each conidium. Even if a few heterokaryotic [Dp(B362i) + Df(B362i)] conidia form, the Df(B362i) nuclei are less likely to divide upon conidial germination, thus biasing even the heterokaryotic conidia to produce homokaryotic Dp(B362i) germlings. This hypothesis provides a plausible (and exonerating) explanation for an otherwise perturbing set of results that we had obtained in the early stages of this study. A strain was initially scored as [Dp(B362i) + Df(B362i)], because of its self-fertility and the ability of its DNA to support PCR amplification of all three B362i-specific junction fragments, but later it behaved like a Dp(B362i) homokaryon, because it was now self-sterile and its newly isolated DNA failed to amplify the A fragment (Dev Ashish Giri, unpublished results). A self-cross would selectively restore the Df(B362i) nuclear fraction back to 50%, but self-crosses are not possible once the Df(B362i) nuclei are completely lost.
In our model, the putative Df(B362i)-nuclear division defect occurs in a heterokaryotic cytoplasm, hence the gene presumed to be required for efficient division must have a nullphenotype that is non-complementable by the wild-type allele in the neighboring Dp(B362i) nuclei. A gene whose null allele (Δ) is not complemented by the wild-type allele (WT) in a [Δ + WT] heterokaryon can be considered to be nucleus-limited in function (Kasbekar, 2014).
Although no nucleus-limited genes have been reported as yet, their existence in fungi is not ruled out, especially given the putative nucleus-limited behavior of the N. crassa scon c mutant (Burton and Metzenberg 1972), the DNA damage checkpoint signal in Saccharomyces cerevisiae (Demeter et al. 2000), and the MatIS gene silencing process in Aspergillus nidulans (Czaja et al. 2013). The S. cerevisiae precedent is germane to our model because it demonstrates that of two nuclei sharing the same cytoplasm, the one with damaged DNA arrests in mitosis without impeding progression through mitosis of the other with undamaged DNA.
Introgression of additional N. crassa translocations into N. tetrasperma holds out the prospect of uncovering more genotypes with putative nucleus-limited effects. Alternatively, N.
tetrasperma can now be transformed (Kasbekar 2015), therefore a quicker way to screen for nucleus-limited genes might be via engineering of targeted integration of yeast Frt sites into the different chromosome arms, and using FLP recombinase to induce crossover and produce

(EB4) A, T(IBj5) A, T(UK14-1) A, and T(B362i) A
strains of N. crassa were crossed with the C4T4 a hybrid strain. Bent arrows represent PCR with breakpoint junction-specific primers to distinguish the translocation progeny (e.g., T 1xC4T4 ) from their Dp and N siblings. T 1xC4T4 A x C4T4 a yielded T 2xC4T4 A or T 2xC4T4 a strains, which were productive in crosses with opposite mating type homokaryotic derivatives of N. tetrasperma strain 85. T 1x85 progeny were crossed with 85 a or 85 A to obtain the self-fertile heterokaryotic strains I1-I5 (for IBj5) and U9 (for UK14-1), or the T 2x85 strains (for EB4 and B362i). Crosses of T 2x85 with 85 a or 85 A produced the heterokaryons E1 and B7. From self-cross of the heterokaryons we obtained self-fertile progeny that were genotyped as [T + N] or [Dp + Df] (see Table 2). Figure 1. C4,T4 a is a weak MSUD suppressor. Ascus development in crosses of the N. crassa strains OR a and Sad-1 a, and the N. crassa / N. tetrasperma hybrid strain C4,T4 a with the MSUD tester strains ::bml A, ::mei-3 A and ::r ef2 . Meiotic silencing of the bml (β-tubulin) and mei-3 genes in the crosses with OR a disrupts ascus development, whereas its suppression in the crosses with Sad-1 a allows normal ascus development. Silencing of r in the cross with OR a causes all eight ascospores to be round, and its suppression by Sad-1 a restores the normal spindle shape. Silencing is evident in crosses of C4,T4 a with ::bml A and ::r ef2 A but not in the cross with ::mei-3 A. Partial suppression of MSUD by C4,T4 a is characteristic of Esm type strains (Ramakrishnan et al., 2011). NA-not applicable.