Properties of Unpaired DNA Required For Efficient Silencing in Neurospora crassa

Corresponding author: Rodolfo Aramayo, College of Science, Texas A&M University, Room 414A, Bldg. BSBW, College Station, TX 77843-3258., raramayo{at}mail.bio.tamu.edu (E-mail)

Communicating editor: J. J. LOROS

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The presence of unpaired copies of a gene during meiosis triggers silencing of all copies of the gene in the diploid ascus cell of Neurospora. This phenomenon is called meiotic silencing and on the basis of genetic studies appears to be a post-transcriptional gene silencing (PTGS) mechanism. Previously, meiotic silencing was defined to be induced by the presence of a DNA region lacking an identical segment in the homologous chromosome. However, the determinants of unpaired DNA remained a mystery. Using the Ascospore maturation-1 (Asm-1) gene, we defined what needs to be "unpaired" to silence a gene. For efficient silencing, an unpaired region of DNA needs to be of a sufficient size and contain homology to the reporter transcript. The greater the size of the loop and the larger the homology to the reporter transcript, the better the resulting meiotic silencing is. Conversely, regions not containing homology to the transcript, e.g., intergenic regions, did not silence the reporter. Surprisingly, unpaired fragments lacking a canonical promoter silenced the reporter. Additionally, we detected the unpairing-dependent loss of a transcript during meiotic silencing. Our observations further support a PTGS mechanism for meiotic silencing and offer insight into the evolutionary consequences resulting from this novel meiotic checkpoint.

MEIOTIC cells are unique in that their development entails profound changes that activate sophisticated cellular mechanisms to ensure precise chromosome duplication, repair, and recombination. Thus, these cells must undergo internal morphological development. In Neurospora, meiosis occurs inside the zygote, which forms by fusion of two haploid nuclei of opposite mating type. The zygote undergoes karyogamy, meiosis, and postmeiotic mitosis within the perithecium—the complex multicellular sexual reproductive apparatus of Neurospora (RAJU 1992 )—resulting in an ascus that contains eight haploid spores arrayed in an order that reflects their lineage (RAJU 1980 , RAJU 1992 ). Immediately following karyogamy, meiotic chromosomes align, compact, pair, and recombine to produce progeny that carry sets of newly shuffled DNA (RAJU 1980 ; KLECKNER 1996 ; COOK 1997 ; ZICKLER and KLECKNER 1998 , ZICKLER and KLECKNER 1999 ; ROEDER and BAILIS 2000 ). Inside the perithecium, the developing asci are immersed in maternal tissue and cannot be isolated in a pure form, especially at early stages in spore formation. Thus, any study of the molecular mechanisms of meiosis in Neurospora is a challenge.

Neurospora meiotic cells have developed mechanisms that control the integrity of the genomes that participate in the process. Meiotic transvection and meiotic silencing are two such mechanisms. They are the two faces of the same coin. Meiotic transvection is initiated in the diploid zygotic cell of Neurospora immediately after karyogamy, in the narrow window of time at the onset of meiosis during which homologous chromosomes align and "sense" each other, but before the first meiotic division occurs (ARAMAYO and METZENBERG 1996B ). The zygote is the only known diploid cell of Neurospora and thus the only cell in Neurospora in which the trans-sensing first seen in Drosophila (LEWIS 1954 ) can be observed. When a gene fails to sense its partner in the homologous chromosome, the resulting unpaired DNA triggers meiotic silencing, an RNA-mediated post-transcriptional gene-silencing mechanism that, once activated, persists during the subsequent meiotic divisions (ARAMAYO and METZENBERG 1996B ; SHIU et al. 2001 ), but is reset at some point prior to germination.

Meiotic transvection was discovered through a combination of design and serendipity during our studies of the complex Ascospore maturation-1 gene (Asm-1) in Neurospora (ARAMAYO and METZENBERG 1996B ). ASM-1 (the product of asm-1⁺) is an abundant nuclear protein essential for the formation of aerial hyphae, the development of protoperithecia (the haploid female sexual structures), and the maturation of the ascospores (the haploid sexual spores; ARAMAYO and METZENBERG 1996B ; ARAMAYO et al. 1996 ). Recessive, loss-of-function mutations in Asm-1 are spore autonomous within the ascus. That is, spores carrying the mutant allele fail to develop, whereas spores with the asm-1⁺ allele mature normally.

Surprisingly, we discovered that deletion alleles of Asm-1 are ascus dominant; all spores within the ascus fail to develop, including the ones carrying the asm-1⁺ allele. Several models were proposed to explain this odd genetic behavior (ARAMAYO and METZENBERG 1996B ). The favored interpretation was that pairing over the entire length of the Asm-1 gene is an essential step for normal development. This conclusion was based on the observation that crosses between two deletion mutants, each one complemented by an identically placed ectopic copy of asm-1⁺, produced abundant, viable black spores. These results are consistent with a model for transvection in which the presence of genes at identical positions in homologous chromosomes alters the developmental outcome. Genes sense each other during chromosome pairing (i.e., undergo meiotic transvection). If a given gene fails to pair with (i.e., does not sense) its partner, the silencing of all paired and unpaired copies of the gene will follow.

SHIU et al. 2001 demonstrated that mutations in the Suppressor of ascus dominance-1 (Sad-1) gene eliminate the ascus dominance of paired and unpaired copies of Asm-1. The sad-1⁺ gene encodes an RNA-dependent RNA polymerase (RdRP; SHIU et al. 2001 ; SHIU and METZENBERG 2002 ). This finding suggested that the mechanism behind ascus dominance involves RNA silencing and that it probably operates through the production of an RNA-based diffusible signal (SHIU et al. 2001 ). It also suggested that in addition to the vegetative RNA-silencing pathway, quelling, a second, meiosis-specific RNA-silencing pathway might exist in Neurospora (GALAGAN et al. 2003 ). This prediction was confirmed by our demonstration of the involvement of two genes in meiotic silencing: one coding for an Argonaute-like protein, Suppressor of meiotic silencing-2 (Sms-2; LEE et al. 2003B ), and the other coding for a Dicer-like protein, Suppressor of meiotic silencing-3 (Sms-3; M. MCLAUGHLIN, D. W. LEE, R. PRATT and R. ARAMAYO, unpublished results).

SHIU et al. 2001 interpreted their results on meiotic silencing of Asm-1 in a general context (unpaired DNA) and predicted that the silencing mechanism works at the post-transcriptional level. However, the authors did not address whether silencing could be achieved to the same extent by unpairing different functional or nonfunctional fragments of the gene. For example, the synthesis of the first strand of RNA (the predicted template for the RdRP) could depend on a canonical promoter recognized by a conventional RNA polymerase II-directed transcription complex, as would be the case for a retrotransposon undergoing activation (SANDMEYER and MENEES 1996 ). The use of Asm-1 as a reporter gene allowed us to ask these questions because ASM-1 is essential for normal ascospore development (ARAMAYO et al. 1996 ). We therefore first mapped the promoter of the gene during both asexual and sexual development and identified an Asm-1 transcript whose loss correlates with the induction of meiotic silencing. We then tested how the unpairing of different functional and nonfunctional fragments of the gene affects meiotic silencing. Silencing was observed only when regions of the gene with homology to the reporter transcript were unpaired. The efficiency of silencing increased with the length of the unpaired DNA and with the size of the unpaired loop. Additionally, we found that canonical promoter elements did not have to be present in the unpaired loop for meiotic silencing to occur. This work thus defined the primary qualitative and quantitative properties of unpaired DNA in meiotic silencing and offers further support to meiotic silencing being a post-transcriptional gene silencing (PTGS)-like mechanism.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Procedures for DNA extraction from Neurospora crassa, Southern blot analysis, and other nucleic acid manipulations were performed as described (PRATT and ARAMAYO 2002 ). Similarly, growth conditions, conidial spheroplast preparation, fungal transformation, homokaryon purification, female fertility/sterility determinations, and genetic crosses were performed as described (PRATT and ARAMAYO 2002 ). The formulas for Vogel's medium N, Westergaard's medium, and the sugar mixture of Brockman and de Serres have been described by DAVIS and DE SERRES 1970 .

Scoring of genetic crosses:
Generally the partners were co-inoculated in a petri dish and incubated at room temperature for 6 days. The bulk of the conidia were removed, and the remaining conidia were spread around the lawns with 1 ml of water. Crosses made in this way started shooting ascospores 20 days after inoculation. The degree of silencing was determined by taking several pictures of each petri dish lid. Pictures were printed and spores counted. The percentage of black (viable) or white (inviable) ascospores shot was determined from the total number of ascospores present on each picture. The strength of the observed silencing was defined as described in the Fig 2 legend.

View larger version (48K): In this window In a new window Download PPT slide   Figure 1. Promoter scanning of Asm-1. (Top) A diagram of the 12,425-bp Asm-1 chromosomal region shows the relative positions of the -transcript (shaded area) and the coding region of the gene. (Bottom) A detailed diagram depicts the 5910-bp region (3426–9336) containing the Asm-1 haploid promoter (3426–5118); the (5370), ß (5163), and (5118) transcription-start sites; the leader region (5370–6128); an intron (5238–5345, open box); a micro-open reading frame (µORF, 5600–5821; short open arrow); the transcriptionally active element (5469–5714); and the asm-1+ open reading frame (6128–8059; ARAMAYO et al. 1996 ). The horizontal bars labeled A–G represent DNA fragments inserted at the his-3 locus (LG I) in strains carrying the Asm-1(3430–9336) deletion in LG V. The ability of each fragment to complement the aerial hyphal formation, protoperithecial development, and ascospore maturation defects associated with Asm-1(3430–9336) is indicated. The asterisk indicates that construct B can direct the formation of protoperithecia in mat a strains only. The ascospore maturation of strains carrying construct G could not be determined (n.d.). The bar labeled H shows the extent of the Asm-1(4615–5469) deletion introduced into the Asm-1 locus at its native position in LG V and the behavior of the resulting strain.

View larger version (37K): In this window In a new window Download PPT slide   Figure 2. Testing cis-silencing in linkage group I. (A) A diagram of a diploid zygote cell shows the pairing of the LG V and LG I chromosomes. For simplicity, only one of the two sister chromatids is indicated. Both LG V chromosomes carry the Asm-1(3430–9336)-deletion allele. Unless otherwise indicated, the LG I Dad chromosome contains a DNA fragment corresponding to the asm-1+ region from coordinates 3426 to 9336 inserted at the his-3 locus, whereas the Mom chromosome carries a series of Asm-1-deletion alleles inserted at the same position. (B) A diagram of the Asm-1 region from coordinates 1 to 12425 shows the -transcript and the coding region of asm-1+. In crosses 1–10 (Table 2), a series of pairing Dad (solid bar) and Mom (open bar) chromosomes are presented. The numbers to the left and right of each DNA fragment indicate the positions of the left and right borders of the fragment, respectively. Whenever present, the size of the loop represents the length of unpaired DNA. The DNA region with homology to the -transcript present in the unpaired DNA is indicated by open bars. To the right, the length of total unpaired DNA [DNA (bp)], the length of unpaired DNA corresponding to the -transcript [Transcript (bp)], and the extent of meiotic silencing (Silencing?) for each cross are presented. The strength of the observed silencing is indicated as the proportion of immature (white) ascospores: +, 0–25%; ++, 25–45%; +++, 45–65%; ++++, 65–85%; and +++++, 85–100%. The percentage of mature (black) ascospores produced by a cross showing a 4:4 segregation was multiplied by 2 to calculate the proportion of immature (white) ascospores.

Plasmid construction:
The genome sequence of the Asm-1 locus is contained on contig 3.56 (Release 3, Whitehead Institute, http://www-genome.wi.mit.edu/annotation/fungi/neurospora/). We arbitrarily defined as position 1 the HindIII site located 6128 bp upstream of the translational initiation signal (ATG) for ASM-1 (ARAMAYO et al. 1996 ). Following this convention the HindIII fragment contained in pRAUW44 (ARAMAYO et al. 1996 ) maps from coordinates 1 to 12425.

A detailed description of the plasmids used in this study is presented in supplementary material online (Methods at http://www.genetics.org/supplemental/). Oligonucleotides used in this study are described in supplementary material online (Table S1 at http://www.genetics.org/supplemental/).

Total RNA isolation:
Total RNA was extracted from 6-day-old perithecia using TRIzol reagent (GIBCO-BRL, Grand Island, NY). Perithecia were harvested by removing them from Westergaard's medium, solidified with 1.5% Bacto-Agar, with a sterile razor blade, and were next ground with a mortar and pestle under liquid nitrogen. One milliter of TRIzol reagent (GIBCO-BRL) was added to the ground perithecia, and total RNA was extracted following the manufacturer's protocol. The resulting RNA was subjected to an additional LiCl precipitation purification step.

Rapid amplification of cDNA ends PCR:
First strand was synthesized by first mixing 3 µg of total RNA with 10 pmol of each ODLAM018 (coordinates 6761–6742, Table S1) and SMART II Oligo [SMART rapid amplification of cDNA ends (RACE) cDNA amplification kit; CLONTECH, Palo Alto, CA] in a 10-µl reaction volume containing a final concentration of 50 mM Tris-HCl, pH 8.3, 10 mM dithiothreitol, 75 mM KCl, 6 mM MgCl₂, and 1 mM dNTPs. The mixture was then denatured at 70° for 5 min. Two hundred units of Superscript II (Invitrogen, Carlsbad, CA) was added and the reaction was incubated at 42° for 90 min. Reaction was stopped by adding 100 µl of tricine-EDTA buffer (10 mM tricine-KOH, pH 8.5, 1 mM EDTA). PCR was performed by mixing 2.5 µl of the first-strand synthesis reaction product (cDNA) with 10 pmol of ODLAM067 (coordinates 5952–5929, Table S1) and 1x Universal Primer Mix (UPM; CLONTECH) in a 50-µl reaction volume containing a final concentration of 40 mM tricine-KOH, pH 8.7, 15 mM potassium acetate, 3.5 mM magnesium acetate, 3.75 µg/ml BSA, 0.005% (v/v) Tween-20, 0.005% Nonidet-P40, 0.2 mM dNTPs, and 1x Advantage 2 polymerase mix (CLONTECH). Reaction was incubated as follows: 1 cycle, 94° for 2 min; followed by 5 cycles of 94° for 5 sec, 72° for 3 min; followed by 5 cycles of 94° for 5 sec, 70° for 10 sec, 72° for 3 min; followed by 25 cycles of 94° for 5 sec, 68° for 10 sec, and 72° for 3 min.

Strain description:
Strains of N. crassa are described in Table 1. The following strains have been previously described: DLNCR83A and DLNCT62A (LEE et al. 2003A ), FGSC 4564 (GRIFFITHS and DELANGE 1978 ), and RPNCR4A (PRATT and ARAMAYO 2002 ). The allele hph⁺::tk⁺ consists of the hygromycin B phosphotransferase fused in frame to the herpes simplex virus thymidine kinase gene, as described by LUPTON et al. 1991 . Escherichia coli K12 XL1-Blue MR (Stratagene, La Jolla, CA) was the host for all our bacterial manipulations. When nonmethylated DNA was needed for enzyme digestions, either GM2163—an E. coli K12 derivative containing, among other markers, a dam13::Tn9 (Cam^R) and a dcm-6 mutation (New England BioLabs, Beverly, MA)—or JM110—an E. coli K12 derivative containing, among others, dam and dcm mutations (YANISCH-PERRON et al. 1985 )—was used.

A note about strains containing the Asm-1^(3430-9336)::hph⁺::mcl-1 deletion allele of Asm-1:
In these strains, the region deleted encompassed a region predicted to direct the transcription of a gene that we call myosin chain-like-1 (mcl-1⁺), on the basis of its weak homology to myosin-chain-like genes in other organisms (data not shown). Strains containing a disruption in this predicted gene are viable and do not have any detectable developmental or metabolic phenotypes.

A note about the integration at the his-3 locus:
Due to the nature of the histidine-3 (his-3) integration vectors used in this study (ARAMAYO and METZENBERG 1996A ), the lysophospholipase (lpl) gene, located downstream of the his-3 gene, was deleted during the integration of our constructs. When we arbitrarily define the HindIII restriction site present in the coding region of his-3 as position 1, the region deleted during the gene replacement spans from position 5192 to position 6046. We include this deletion [lpl^{(5192–6046)}] as part of the genotype of all our strains containing integrations at the his-3 locus (Table 1), to distinguish them from strains containing integrations at the his-3 locus obtained with a third generation of his-3-integration vectors (MARGOLIN et al. 1997 ). The genome sequence of the his-3⁺ locus can be found on contig 3.165 (Release 3, Whitehead Institute, http://www.genome.wi.mit.edu/annotation/fungi/neurospora/).

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

The leader of Asm-1 contains a promoter element:
To determine the regions capable of driving Asm-1 expression, we performed a deletion-scanning analysis of the Asm-1 promoter (Fig 1). First, we constructed two strains, one of each mating type, that contain a deletion of the 5910-bp chromosomal region spanning the promoter, coding, and downstream regions of Asm-1 in linkage group V (LG V), which was replaced with the hygromycin phosphotransferase (hph⁺) marker, to generate the Asm-1^{(3426–9336)} deletion allele (Fig 1). The resulting mutant strains had the same complex phenotype as that of the Asm-1 mutants described before (ARAMAYO et al. 1996 ). They were female sterile (i.e., they did not form protoperithecia) but male fertile (i.e., they could fertilize strains of the opposite mating type). The aerial hyphae of the mutants were stunted, compared to those of wild type; consequently, they conidiated very close to the agar surface.

We then transformed one of these strains either with the identical 5910-bp DNA fragment that was deleted in Asm-1^(3426-9336) (Fig 1, construct A) or with one of a series of 5'-truncated Asm-1 fragments (constructs B–G). All these fragments were inserted ectopically at the his-3 locus in LG I. We then tested these transformants for their ability to undergo asexual and sexual development. Asexual complementation was scored by the ability of strains containing different Asm-1 fragments to form aerial hyphae. Sexual complementation was scored by their ability to form protoperithecia and by their ability to form viable ascospores in a cross with an Asm-1^(3426-9336) deletion mutant of the opposite mating type complemented with the 5910-bp Asm-1 fragment inserted at the his-3 locus in LG I. In the absence of meiotic silencing, successful ascospore maturation complementation should result in eight mature ascospores per ascus. Failure to complement should produce four mature ascospores and four immature ones. No meiotic silencing was observed in cross 1 and crosses 3–7 (Fig 2). The meiotic silencing observed in cross 8 precluded us from determining the ascospore maturation complementation phenotype of construct G (Fig 1 and Fig 2).

As expected, deletion mutants carrying an ectopic copy of the entire Asm-1 region complemented both the asexual and sexual developmental defects associated with Asm-1^{(3426–9336)} (Fig 1, construct A). In contrast, strains carrying truncated ectopic alleles of Asm-1 lacking the promoter or the promoter plus transcription-initiation sites formed neither normal aerial hyphae nor protoperithecia (constructs B–G), with the exception of construct B, which could form protoperithecia when present in mating type a (mat a) strains (data not shown). The B, C, D, and E constructs did, however, support normal ascospore development, whereas strains carrying segment F did not. These results indicate that the region between nucleotides 5437 and 5714 can drive the expression of Asm-1 in the absence of its canonical promoter during ascospore development but not during vegetative growth (compare construct E with constructs F and G). This "transcriptionally active element" mapped to a region that encodes the leader of the Asm-1 transcript that is expressed during asexual development (ARAMAYO et al. 1996 ).

To confirm this result, we constructed a strain containing an 854-bp deletion at the normal Asm-1 locus in LG V. This deletion removed most of the promoter and all of the known transcription start sites of Asm-1 (Fig 1, coordinates 4615–5469, construct H). The resulting strain produced stunted aerial hyphae and was unable to form protoperithecia. This phenotype was similar to those strains carrying the ectopically located constructs B–G. However, this strain supported normal ascospore development (compare constructs E and H with F). These results confirm that a DNA element contained within nucleotides 5469–5714 (Fig 1) can drive the expression of Asm-1 during ascospore development. The canonical Asm-1 open reading frame (ORF) must be transcribed and translated, because mutation of the ATG translation initiation codon (located at position 6128–6130) inactivates Asm-1 function during ascospore development (KUTIL et al. 2003 ).

The -transcript of Asm-1 is a target for meiotic silencing:
To map the 5'-end(s) of the Asm-1 transcripts during sexual development, we conducted two experiments: First, we performed 5'-RACE-PCR on mRNA extracted from tissues undergoing sexual development (6 days after fertilization) and cloned and sequenced the amplified DNAs. This experiment confirmed the presence of the previously determined transcription start site of the gene [henceforth called -start (coordinate 5370); Fig 1] (ARAMAYO et al. 1996 ) and also identified two additional transcription start sites [henceforth called ß- (coordinate 5163) and -start (coordinate 5118); Fig 1]. Second, a sexual-stage-specific cDNA library was screened with a probe spanning the entire Asm-1 region (coordinates 2992–8387). The cross-reacting cDNAs were cloned and sequenced. This analysis identified cDNAs corresponding to the -, ß-, and -messages and also revealed a 107-bp intron (coordinates 5238–5345) downstream of the ß- and -transcription start sites (Fig 1). Taken together, these experiments identified the predominant transcription start sites detected for Asm-1 in wild-type cells during sexual development.

To determine which transcripts are used during ascospore development, we performed 5'-RACE-PCR on mRNA extracted from 6-day-old perithecial tissues of wild-type crosses and crosses where meiotic silencing was induced. Given that ASM-1 is essential for ascospore development, we predicted that silencing of the gene would eliminate the transcript(s) specific to ascospore development. We crossed wild-type strains with each other (Fig 3, lanes 2 and 5) and with strains carrying the Asm-1^{(3426–9336)} deletion allele (lanes 3 and 6) or carrying the Asm-1^{(3426–9336)} deletion allele complemented with the 5910-bp DNA fragment of Asm-1 inserted at the his-3 locus in LG I (lanes 4 and 7). DNA fragments corresponding to the ß- and -transcripts were present in all crosses (lanes 2–7). In contrast, DNA fragments corresponding to the -transcript could be detected only when the Asm-1 region was paired, i.e., when silencing was not induced (lanes 2 and 5). The unpairing-dependent loss of the -message strongly suggests that this transcript is a target for meiotic silencing. The ß- and -transcripts were not sensitive to meiotic silencing. Given the strong sequence similarity between the -, ß-, and -transcripts and that meiotic silencing is ascus autonomous (ARAMAYO and METZENBERG 1996B ; SHIU and METZENBERG 2002 ), the simplest explanation for their persistence is that these transcripts reside in the surrounding maternal tissue and are, therefore, protected from degradation. However, these results do not rule out that ß- and -transcripts are present inside the developing asci or that the -transcript might be present in maternal tissue in undetectable amounts, but they do strongly suggest that the -transcript is being selectively used during ascospore development.

View larger version (95K): In this window In a new window Download PPT slide   Figure 3. The -transcript of Asm-1 is a target for meiotic silencing. Ethidium bromide agarose gel [1% (w/v)] is shown. Lane 1, marker consisting of DNA digested with EcoRI and HindIII (/EcoRI/HindIII). Lanes 2–7: Reverse-transcribed PCR products were generated from total RNA as described in MATERIALS AND METHODS. RNA was extracted from the following crosses (Table 2): lane 2, cross 11 [asm-1+ A (female) x asm-1+ a (male)] (no silencing); lane 3, cross 12 [asm-1+ A (female) x asm-1(3426–9336) a (male)] (silencing induced by unpairing one copy of Asm-1); lane 4, cross 13 [asm-1+ A (female) x his-3+::Asm-1+; asm-1(3426-9336) a (male)] (silencing induced by unpairing two copies of Asm-1); lane 5, cross 14 [asm-1+ A (male) x asm-1+ a (female)] (no silencing); lane 6, cross 15 [asm-1(3426–9336) A (male) x asm-1+ a (female)] (silencing induced by unpairing one copy of Asm-1); and lane 7, cross 16 [his-3+::Asm-1+; asm-1(3426–9336) A (male) x asm-1+ a (female)] (silencing induced by unpairing two copies of Asm-1). The gel was loaded with 10% of each PCR reaction. Note that the -transcript is present only in crosses in which meiotic silencing was not induced.

Only unpaired DNA regions with homology to the transcript can silence Asm-1:
To determine what properties unpaired DNA must have to silence Asm-1, we performed crosses of the same type that were used to test sexual complementation and determined the percentage of black ascospores shot from these crosses (Fig 2).

Crosses between Asm-1^{(3426–9336)} deletion parent strains, both complemented with the 5910-bp DNA fragment inserted at his-3, produced many mature, black, viable ascospores (Fig 2, cross 1). This result obeyed the rules for meiotic transvection and indicated that pairing of the ectopic copies of Asm-1 was both necessary and sufficient for normal sporulation.

We reasoned that, if silencing of Asm-1 requires unpairing of a specific DNA region or regions present in the unpaired DNA fragment (e.g., a promoter element), these silencing elements could be mapped by the progressive-deletion method, according to the following logic: Pairing of a "Dad" chromosome carrying the entire 5910-bp Asm-1 fragment at the his-3 locus with different "Mom" chromosomes, each carrying a different deletion at the his-3 locus, results in the unpairing of the region in the Dad chromosome that corresponds to the deletion in the Mom chromosome (Fig 2). If the unpaired DNA thus generated does not induce silencing, the deletion will be recessive. In this situation, deletion mutations that do not affect expression of Asm-1 during sexual development will produce eight viable ascospores per ascus (e.g., crosses 1–6). In contrast, any mutation that blocks expression of Asm-1 during ascospore development (i.e., that prevents ascospore maturation) will affect the development of the haploid ascospore containing it. In this case, even in the total absence of silencing, half of the spores will be black and half white (e.g., cross 7). If the unpaired DNA induces silencing, the deletion will be dominant and no viable spores will form (e.g., cross 8). Also, since Mom chromosomes carried progressively larger deletions of Asm-1, the unpaired loops in the Dad chromosomes will become progressively larger. We can thus test the effect of the size of the unpaired loop on meiotic silencing (Fig 2). Thus, the lower the percentage of black ascospores, the stronger the silencing will be (Table 2).

The results of the crosses are summarized in Fig 2. The unpaired DNA in crosses 1–7 did not cause silencing, whereas it did in crosses 8 and 9. The ability to silence thus correlated both with encroachment into regions containing homology to the transcript and with increasing size of the unpaired DNA loop. Therefore, these results could be explained in three ways. First, the extent of silencing could simply correlate with the length of unpaired DNA (i.e., larger deletions cause silencing and smaller ones do not). If this is true, unpaired DNA would have to exceed some threshold length between 2288 and 2681 bp (crosses 7 and 8) to trigger silencing. Second, silencing might occur only when the unpaired DNA contains a discrete region required for silencing (e.g., homology to the canonical Asm-1 -transcript of the gene). Third, silencing could be the combined outcome of these two effects.

To discriminate among these possibilities, we decreased the length of the unpaired DNA loop from 2681 to 1492 bp, which is below the predicted threshold, by constructing a new Dad chromosome with a truncation at the 5' end of the original 5910-bp insert. The length of the unpaired transcript was held constant (cross 10). If only the size of the loop is important, then the smaller loop should not silence. If unpairing of the transcribed region is important, then the smaller loop should silence as efficiently as the larger one. If both factors are important, then the smaller loop should silence less efficiently than the larger loop but to a greater extent than a loop of equal size that does not contain the transcribed region. The result was consistent with the third possibility (compare crosses 8 and 10), suggesting that both qualitative (compare crosses 4–7 to cross 10) and quantitative (compare crosses 8 and 10) properties determine the silencing potential of unpaired DNA.

To ensure that the his-3 locus was not introducing an unknown artifact, we next tested if unpaired Asm-1 DNA at the canonical chromosomal position in LG V behaves in the same manner. We constructed a series of strains carrying progressively larger deletions of the Asm-1 gene. The deleted fragment was replaced by the hph⁺ marker gene (Fig 4). Although the presence of the hph⁺ marker gene at the site of deletion makes it difficult to compare the results in Fig 2 and Fig 4 directly, the trends were similar. Crosses 17, 18, and 19 did not result in silencing, whereas crosses 20 to 24 did. Consistent with the results shown in Fig 2, unpaired DNA with 700 bp or more of homology to the canonical transcript was required for silencing (compare cross 7 with 8 and cross 19 with 20). To confirm that the main determinant of silencing was unpaired DNA corresponding to the canonical transcript rather than the length of the unpaired DNA, we deleted a fragment of Asm-1 that is completely within the region of homology to the -transcript (Fig 4, cross 25, coordinates 5414–6690). The silencing was not as strong as in crosses 20–24, but since the total length of the unpaired DNA was shorter than that in cross 19, the enhancement of silencing was significant (compare cross 25 with cross 19).

View larger version (36K): In this window In a new window Download PPT slide   Figure 4. Testing cis- silencing in linkage group V. The presentation is as in Fig 2, except that the diagram of the diploid zygote cell represents the pairing of a wild-type LG V chromosome (Dad) with a Mom chromosome carrying a deletion allele. The coordinates indicate the unpaired region in the Dad chromosome corresponding to the region that was deleted and replaced by the hph+ gene in the Mom chromosome.

These results suggest that unpairing the promoter of Asm-1 is not sufficient to induce silencing (Fig 2, crosses 2–7 and Fig 4, crosses 18 and 19) and thus argue against the presence of a mechanism that transcriptionally inactivates the loop of unpaired DNA during meiosis and subsequent cell divisions. In addition, they show that the signal produced by the unpaired DNA loop does not propagate in cis to other paired segments of the gene. This result is in agreement with our own previous observations (KUTIL et al. 2003 ). Finally, it seems that a DNA region with at least 700 bp of homology to the canonical transcript is required for silencing the gene, which reinforces an RNA-mediated model for meiotic silencing.

Loops of unpaired DNA need not carry known promoter elements to silence:
The loop of unpaired DNA tested in our previous experiments was always linked to its normal flanking DNA and always contained segments required for transcription of Asm-1. To dissociate the unpaired loop from its surrounding sequences in the chromosome and to test whether promoter elements are necessary for meiotic silencing we unpaired different Asm-1 segments of similar sizes at his-3 and tested their abilities to silence paired wild-type copies of the gene. We call this test trans-silencing. If promoter elements are required, segments containing them will silence in trans, whereas those lacking them will not. Similarly, any silencing-active fragment of Asm-1 DNA unpaired ectopically—where it is dissociated from its normal flanking DNA in the chromosome—should silence, unless its silencing activity depends on surrounding cis-linked sequences that are normally present at its wild-type chromosomal position.

We performed this analysis by carrying out crosses between two sets of strains. All strains contained the wild-type asm-1⁺ allele at its normal position in LG V (Fig 5). At the his-3 locus in LG I, Dad chromosomes contained no insert. Mom chromosomes each contained a different Asm-1 insert. We then examined whether these segments of Asm-1 DNA would activate meiotic trans-silencing when unpaired.

View larger version (39K): In this window In a new window Download PPT slide   Figure 5. Loops of unpaired DNA need not carry known promoter elements to silence. The presentation is as in Fig 2, with the following alterations. Both LG V chromosomes carry the asm-1+ wild-type allele. On LG I, Dad chromosomes have no Asm-1 DNA at his-3, whereas Mom chromosomes carry different Asm-1 alleles inserted at his-3 (open bars). The hatched region indicates the position of the canonical promoter and the transcriptionally active element of Asm-1. The shaded area indicates the position of the region corresponding to the -transcript of the gene.

The Asm-1 region was scanned by unpairing fragments of 1503 bp (crosses 26 and 27), 1225 bp (cross 28), 1438 bp (crosses 29 and 30), 1407 bp (crosses 31 and 32), and 949 bp (crosses 33 and 34). Together, they spanned the entire Asm-1 region of the genome (Fig 5). Only the fragment spanning coordinates 7035–8442 induced silencing (crosses 31 and 32). This result contrasted with what was observed in cis-silencing (Fig 1 Fig 2 Fig 3 Fig 4). In cis, unpaired loops of 1.4 kb carrying regions of homology to the transcript as small as 0.7 kb induced silencing (Fig 2, cross 10), whereas under the experimental conditions tested here, an unpaired region of 1.4 kb belonging entirely to the transcript could not induce silencing in trans (Fig 5, crosses 29 and 30).

It seemed conceivable that silencing was not observed in some cases because the ectopic fragments lack expression signals that are normally present. To ensure that an absence of transcription was not responsible for the failure to silence, we constructed a fusion of an internal fragment of Asm-1 (coordinates 6320–7758) to the trpC promoter of Aspergillus nidulans (MULLANEY et al. 1985 ; HAMER and TIMBERLAKE 1987 ). Strains carrying this fusion inserted at the his-3 locus in LG I were crossed to wild-type strains, leaving an unpaired loop of 1.8 kb carrying a region of homology to the transcript of 1.4 kb. Silencing was not observed (crosses 35 and 36).

Next, we repeated the scanning experiment using larger unpaired fragments. Fragments of 2397 bp (crosses 37 and 38), 2064 bp (crosses 39 and 40), 2646 bp (crosses 41 and 42), and 5910 bp (crosses 43 and 44), again spanning the entire Asm-1 genomic region, were tested. Without exception, silencing was observed.

These results demonstrate that, under the conditions used here, trans-silencing was not as efficient as cis-silencing when the length of unpaired DNA is short. They also demonstrate that promoter elements in the unpaired DNA are not required for the production of the diffusible silencing signal (e.g., small RNAs). Thus, the synthesis of the first strand of RNA for meiotic silencing (the predicted template for the RdRP) does not obey the known rules for transcription (i.e., the absolute requirement for promoter elements observed for metabolic genes in haploid tissues, for example).

Bigger is better:
We consistently found a positive correlation between the length of the unpaired DNA region and the strength of silencing. This observation suggested that larger unpaired loops of DNA silence more efficiently than smaller ones. We tested this hypothesis following the same experimental design used to test trans-silencing.

First, we selected a fragment of DNA that did not silence Asm-1 in trans (Fig 5 and Fig 6, coordinates 4615–6118, crosses 26 and 27). This fragment was selected because, when unpaired, it exposes a region of homology to the Asm-1 transcript of only 748 bp, near the minimal length required for cis-silencing (Fig 2, crosses 7 and 8). We predicted that if the size of the loop matters (i.e., bigger is better), silencing should be enhanced by the addition of "neutral" DNA. The 1503-bp fragment was therefore fused to 1411 bp of DNA, and the fusion construct was integrated at the his-3 locus. When the resulting strain was crossed to wild type, Asm-1 was silenced (Fig 6, crosses 45 and 46). Since the only difference between crosses 26, 27 and 45, 46 is the total length of unpaired DNA (i.e., the length of the region of homology to the transcript was kept constant), this result implies that efficient silencing is indeed correlated with the size of the loop. We also predicted that holding the length of unpaired DNA constant (i.e., 2.9 kb) while increasing the extent of homology to the transcript should enhance silencing. The same 1503-bp Asm-1 fragment tested before was therefore fused to a 1438-bp fragment internal to the Asm-1 coding region (coordinates 6320–7758). Crosses between wild-type strains and strains carrying this fusion construct at the his-3 locus resulted in very efficient silencing (Fig 6, compare the results from crosses 45 and 46 with crosses 47 and 48).

View larger version (33K): In this window In a new window Download PPT slide   Figure 6. Bigger is better. The presentation is as in Fig 5. In crosses 45 and 46, the unpaired Asm-1 fragment corresponding to coordinates 4615–6118 was augmented by the addition of 1411 bp of DNA from bacteriophage . In crosses 47 and 48, the same fragment was fused to a 1438-bp fragment internal to the Asm-1 coding region (coordinates 6320–7758). Similarly, the 1438-bp fragment internal to the Asm-1 coding region (coordinates 6320–7758) was fused to a 852-bp upstream fragment (coordinates 4615–5467). In crosses 51–54, the loop of unpaired DNA of the Asm-1 fragment corresponding to coordinates 6320–7758 was augmented by the addition of 4878 bp (crosses 51 and 52) and 7421 bp (crosses 53 and 54), respectively, of DNA from bacteriophage .

In a further test, the fragment that did not silence Asm-1 even when coupled to the trpC promoter (Fig 5, coordinates 6320–7758, crosses 29, 30, 35, and 36) was fused to an additional 852-bp fragment of Asm-1 DNA (Fig 6). The resulting 2290-bp construct, which has 1535 bp of homology to the transcript, was integrated at the his-3 locus. When this strain was crossed to wild type, weak silencing was observed (Fig 6, compare the results of crosses 29 and 30 with crosses 49 and 50). Finally, we tested whether increasing the length of unpaired DNA while keeping the length of homology to the transcript constant (i.e., 1438 bp) would result in stronger silencing. First, the 6320–7758 fragment of Asm-1 DNA was fused to 4878 bp of DNA. The resulting 6316-bp construct was integrated at the his-3 locus. When strains carrying this construct were crossed to wild type, the result was mild silencing (crosses 51 and 52). We next increased the length of the unpaired DNA from 6316 to 8859 bp by adding additional DNA while keeping the length of homology to the transcript constant at 1438 bp. When the resulting strain was crossed to wild type, stronger silencing was observed (Fig 6, crosses 53 and 54). We therefore conclude that a longer stretch of unpaired DNA silences more efficiently than a shorter one.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

In summary, a number of elegant molecular mechanisms to control gene expression based on DNA sequence homology have evolved in N. crassa. During haploid development (i.e., vegetative growth), the presence of repeated elements can activate a vegetative silencing mechanism called quelling (COGONI and MACINO 2000 ; COGONI 2001 , COGONI 2002 ). If cells containing repeated DNA elements overcome quelling and enter the sexual phase of their life cycle, the sequences are subjected to another unique silencing mechanism called repeat-induced point mutation (RIP; SELKER 1990 , SELKER 1997 ). In this process, a series of GC-to-AT transition mutations are introduced into the duplicated sequence. Most of the remaining nonmutated bases are methylated. Both of these mechanisms belong to the repertoire of strategies that Neurospora uses to maintain the integrity of its genome. Meiotic silencing is another such process.

This work was aimed at understanding which regions of a gene need to be unpaired during meiosis to signal silencing. Detecting gene silencing during meiosis is not trivial. It requires choosing reporter genes whose products are essential and constantly required for the completion of meiosis or ascospore development or that can be easily detected (e.g., green fluorescent protein). It also requires protein products produced before the induction of meiotic silencing to be unstable. We used Asm-1 because its gene product (ASM-1) fulfills both these requirements (ARAMAYO and METZENBERG 1996B ).

In this report we established that: (1) The size of the loop and the degree of homology to the reporter transcript (i.e., Asm-1) directly correlate with the silencing efficiency by the unpaired DNA, (2) the presence of canonical promoter elements in the unpaired loop is not required for silencing, (3) the unpairing of promoter elements does not affect their transcriptional competence later, during spore development, (4) the -trancript of Asm-1 is a target for meiotic silencing, and (5) the leader of Asm-1 contains a DNA element capable of driving its expression during sexual development.

Our results provide strong confirmatory evidence that meiotic silencing is post-transcriptional, because silencing of Asm-1 can be achieved only by unpairing regions with homology to the transcript. They also raise several important questions. How is the unpaired DNA transcribed? And why do long stretches of unpaired DNA silence more efficiently than smaller ones?

Promoter elements need not be present in the unpaired DNA to induce silencing, suggesting that unpaired DNA is the target of a modified form of a known RNA polymerase. Alternatively, meiotic chromosomes might be the target of an undescribed form of transcription that originates in paired regions and extends into unpaired DNA, whenever present. Whatever the mechanism of transcription of unpaired DNA, the postulated aberrant RNAs (aRNAs) produced likely do not need poly(A) tails to play their part in meiotic silencing, as evidenced by the strong silencing activity of truncated fragments of the gene. Moreover, these "truncated" transcripts either must be immediately converted into double-strand RNAs by the SAD-1 RdRP or somehow escape the RNA-quality-control mechanisms that operate in eukaryotic cells (MOORE 2002 ), unless, of course, those mechanisms are inactive during early meiosis.

The total length of the unpaired DNA correlates with the extent of silencing. This interpretation is consistent with two possible models (Fig 7). In model I, unpaired DNA is the target of a single "transcription complex" whose intrinsic activity increases proportionally with the size of the loop. The complex is more active in larger loops than in smaller ones. In model II, unpaired DNA is the target of multiple transcription complexes, all of them with similar transcriptional activities. Larger loops silence more efficiently because they accommodate more transcription complexes than smaller ones do. The net result is the same. The larger the loop, the higher is the concentration of postulated aRNAs and small interfering RNAs (siRNAs) that would be produced by the meiotic silencing pathway.

View larger version (50K): In this window In a new window Download PPT slide   Figure 7. Models for transcription of unpaired DNA. In model I, the transcriptional activity of the single complex present in each loop is represented by the number of concentric circles. The transcription complex is more active in larger loops than in smaller ones. In model II, each complex has an equivalent transcriptional activity, but larger loops silence more efficiently because they accommodate more transcription complexes than do smaller ones. The outcome is the same: the larger the loop, the higher the concentration of siRNAs predicted to be produced by the unpaired DNA.

Paired genes do not misbehave. Unpaired genes do, regardless of their genomic location. If pairing occurs, meiosis proceeds normally. If a gene is unpaired, then the unpaired DNA will be detected by the meiotic silencing machinery. If the unpaired region is small, it poses no danger and elicits little or no response. If it is large, it activates meiotic silencing. In this regard, the meiotic machinery involved in chromosome pairing is extremely sensitive. Even the largest loop tested in this work (i.e., 8859 bp) corresponds to <0.09% of LG I. A loop this size should have negligible effect on chromosome pairing.

Clearly, the combined activities of meiotic transvection and meiotic silencing perform a critical check of the sequence composition of the genomes participating in meiosis. For example, an infection by retrotransposable elements in one of the participating genomes would reveal itself through the presence of unpaired DNA corresponding to the invading elements. Their active presence at this particular developmental stage constitutes a grave danger to the integrity of the genomes because of the highly recombinogenic nature of meiotic cells. From an evolutionary perspective, meiotic silencing does not impose a constraint on rearrangements in intergenic regions or regions of genes whose products are not essential for meiosis or ascospore development. It does, however, limit the rate of evolution of critical components of meiosis and ascospore development. In this scenario, two rapidly evolving but related genomes could accumulate a number of noncritical silent rearrangements without losing their interbreeding ability, up to a point, where a rearrangement in one critical gene could result in their reproductive isolation. It is thus appropriate to consider meiotic silencing not only as another checkpoint for meiosis, but also as a critical mechanism of reproductive isolation.

	FOOTNOTES

¹ These authors contributed equally to this work.

	ACKNOWLEDGMENTS

We thank Michael D. Manson for critical review of the manuscript. This work was supported by U.S. Public Health Service grant GM58770 to R.A.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

ARAMAYO, R. and R. L. METZENBERG, 1996a  Gene replacements at the his-3 locus of Neurospora crassa.. Fungal Genet. Newsl. 43:9-13.

ARAMAYO, R. and R. L. METZENBERG, 1996b  Meiotic transvection in fungi. Cell 86:103-113.[CrossRef][Medline]

ARAMAYO, R., Y. PELEG, R. ADDISON, and R. METZENBERG, 1996  Asm-1⁺, a Neurospora crassa gene related to transcriptional regulators of fungal development. Genetics 144:991-1003.[Abstract]

COGONI, C., 2001  Homology-dependent gene silencing mechanisms in fungi. Annu. Rev. Microbiol. 55:381-406.[CrossRef][Medline]

COGONI, C., 2002  Unifying homology effects. Nat. Genet. 30:245-246.[CrossRef][Medline]

COGONI, C. and G. MACINO, 2000  Post-transcriptional gene silencing across kingdoms. Curr. Opin. Genet. Dev. 10:638-643.[CrossRef][Medline]

COOK, P. R., 1997  The transcriptional basis of chromosome pairing. J. Cell Sci. 110(9):1033-1040.[Abstract]

DAVIS, R. H., and F. J. DE SERRES, 1970 Genetic and microbiological research techniques for Neurospora crassa, pp. 79–143 in Metabolism of Amino Acids and Amines, edited by S. P. COLOWICK and N. O. KAPLAN. Academic Press, New York/London.

GALAGAN, J. E., S. E. CALVO, K. A. BORKOVICH, E. U. SELKER, and N. D. READ et al., 2003  The genome sequence of the filamentous fungus Neurospora crassa.. Nature 422:859-868.[CrossRef][Medline]

GRIFFITHS, A. J. F. and A. M. DELANGE, 1978  Mutations of the a mating-type gene in Neurospora crassa.. Genetics 88:239-254.[Abstract/Free Full Text]

HAMER, J. E. and W. E. TIMBERLAKE, 1987  Functional organization of the Aspergillus nidulans trpC promoter. Mol. Cell. Biol. 7:2352-2359.[Abstract/Free Full Text]

KLECKNER, N., 1996  Meiosis: How could it work? Proc. Natl. Acad. Sci. USA 93:8167-8174.[Abstract/Free Full Text]

KUTIL, B. L., K. Y. SEONG, and R. ARAMAYO, 2003  Unpaired genes do not silence their paired neighbors. Curr. Genet. 43:425-432.[CrossRef][Medline]

LEE, D. W., J. R. HAAG, and R. ARAMAYO, 2003a  Construction of strains for rapid homokaryon purification after integration of constructs at the histidine-3 (his-3) locus of Neurospora crassa.. Curr. Genet. 43:17-23.[Medline]

LEE, D. W., R. J. PRATT, M. MCLAUGHLIN, and R. ARAMAYO, 2003b  An Argonaute-like protein is required for meiotic silencing. Genetics 164:821-828.[Abstract/Free Full Text]

LEWIS, E. B., 1954  The theory and application of a new method of detecting chromosomal rearrangements in Drosophila melanogaster.. Am. Nat. 88:225-239.[CrossRef]

LUPTON, S. D., L. L. BRUNTON, V. A. KALBERG, and R. W. OVERELL, 1991  Dominant positive and negative selection using a hygromycin phosphotransferase-thymidine kinase fusion gene. Mol. Cell. Biol. 11:3374-3378.[Abstract/Free Full Text]

MARGOLIN, B. S., M. FREITAG, and E. U. SELKER, 1997  Improved plasmids for gene targeting at the his-3 locus of Neurospora crassa by electroporation. Fungal Genet. Newsl. 44:34-36.

MOORE, M. J., 2002  Nuclear RNA turnover. Cell 108:431-434.[CrossRef][Medline]

MULLANEY, E. J., J. E. HAMER, K. A. ROBERTI, M. M. YELTON, and W. E. TIMBERLAKE, 1985  Primary structure of the trpC gene from Aspergillus nidulans.. Mol. Gen. Genet. 199:37-45.[CrossRef][Medline]

PRATT, R. J. and R. ARAMAYO, 2002  Improving the efficiency of gene replacements in Neurospora crassa: a first step towards a large-scale functional genomics project. Fungal Genet. Biol. 37:56-71.[CrossRef][Medline]

RAJU, N. B., 1980  Meiosis and ascospore genesis in Neurospora. Eur. J. Cell. Biol. 23:208-223.[Medline]

RAJU, N. B., 1992  Genetic control of the sexual cycle in Neurospora. Mycol. Res. 96:241-262.

ROEDER, G. S. and J. M. BAILIS, 2000  The pachytene checkpoint. Trends Genet. 16:395-403.[CrossRef][Medline]

SANDMEYER, S. B. and T. M. MENEES, 1996  Morphogenesis at the retrotransposon-retrovirus interface: gypsy and copia families in yeast and Drosophila. Curr. Top. Microbiol. Immunol. 214:261-296.[Medline]

SELKER, E. U., 1990  Premeiotic instability of repeated sequences in Neurospora crassa.. Annu. Rev. Genet. 24:579-613.[CrossRef][Medline]

SELKER, E. U., 1997  Epigenetic phenomena in filamentous fungi: Useful paradigms or repeat-induced confusion? Trends Genet. 13:296-301.[CrossRef][Medline]

SHIU, P. K. and R. L. METZENBERG, 2002  Meiotic silencing by unpaired DNA: properties, regulation and suppression. Genetics 161:1483-1495.[Abstract/Free Full Text]

SHIU, P. K. T., B. N. RAJU, D. ZICKLER, and R. METZENBERG, 2001  Meiotic silencing by unpaired DNA. Cell 107:905-916.[CrossRef][Medline]

YANISCH-PERRON, C., J. VIEIRA, and J. MESSING, 1985  Improved M13 phage cloning vectors and host strains: nucleotide sequences of the M13mp18 and pUC19 vectors. Gene 33:103-119.[CrossRef][Medline]

ZICKLER, D. and N. KLECKNER, 1998  The leptotene-zygotene transition of meiosis. Annu. Rev. Genet. 32:619-697.[CrossRef][Medline]

ZICKLER, D. and N. KLECKNER, 1999  Meiotic chromosomes: integrating structure and function. Annu. Rev. Genet. 33:603-754.[CrossRef][Medline]

This article has been cited by other articles:

				M. Stam Paramutation: A Heritable Change in Gene Expression by Allelic Interactions In Trans Mol Plant, April 8, 2009; (2009) ssp020v1. [Abstract] [Full Text] [PDF]

				C. A. Kowalczyk, A. M. Anderson, M. Arce-Larreta, and D. L. Chalker The germ line limited M element of Tetrahymena is targeted for elimination from the somatic genome by a homology-dependent mechanism Nucleic Acids Res., November 6, 2006; 34(20): 5778 - 5789. [Abstract] [Full Text] [PDF]

				M. Vyas, C. Ravindran, and D. P. Kasbekar Chromosome Segment Duplications in Neurospora crassa and Their Effects on Repeat-Induced Point Mutation and Meiotic Silencing by Unpaired DNA Genetics, March 1, 2006; 172(3): 1511 - 1519. [Abstract] [Full Text] [PDF]

				R. J. Pratt, D. W. Lee, and R. Aramayo DNA Methylation Affects Meiotic trans-sensing, Not Meiotic Silencing, in Neurospora Genetics, December 1, 2004; 168(4): 1925 - 1935. [Abstract] [Full Text] [PDF]

				E.U. SELKER Genome Defense and DNA Methylation in Neurospora Cold Spring Harb Symp Quant Biol, January 1, 2004; 69(0): 119 - 124. [Abstract] [PDF]