Role of fungal dynein in hyphal growth, microtubule organization, spindle pole body motility and nuclear migration

Microtubule (MT)-associated motor proteins function in subcellular motility and are primarily of two types, dynein and kinesin (Vallee and Sheetz, 1996; Vallee and Shpetner, 1990; Walker and Sheetz, 1993). Cytoplasmic dynein is a large complex consisting of a homodimer of two heavy, several intermediate and several light chains (Gibbons, 1995; Holzbaur and Vallee, 1994; Vallee, 1991, 1993). It is involved in retrograde transport in nerve axons (Paschal and Vallee, 1987; Schnapp and Reese, 1989), transport of feeding vesicles in Paramecium (Schroeder et al., 1990), and transport and localization of membranous organelles, including the Golgi apparatus, ER and endosomes (Allan, 1995; Goodson et al., 1997). It is also implicated in mitotic movements (Barton and Goldstein, 1996; Pfarr et al., 1990; Schroer, 1994; Steuer et al., 1990). Through genetic studies, cytoplasmic dynein has been found to be involved in nuclear migration in the filamentous fungi, Aspergillus nidulans (Xiang et al., 1994, 1995) and Neurospora crassa (Bruno et al., 1996; Plamann et al., 1994), apical vesicle transport in N. crassa (Plamann et al., 1994) and spindle orientation to the daughter bud in the yeast, Saccharomyces cerevisiae (Eshel et al., 1993; Li et al., 1993). However, the roles of cytoplasmic dynein in subcellular motility have not yet been fully characterized (Holzbaur and Vallee, 1994).

Of the three filamentous fungi, A. nidulans, N. crassa and now Nectria heematococca, in which the roles of cytoplasmic dynein have been investigated, mitotic structures and processes and the active motility of interphase spindle pole bodies (SPBs) can be observed clearly in living hyphae of only N. haematococca (Aist, 1995; Aist and Bayles, 1988). Moreover, of these three fungi, details of the time-course and ultrastructure of mitosis and the location and direction of mitotic forces have been thoroughly documented only in N. haematococca (Aist and Bayles, 1988, 1991; Aist et al., 1991; Aist and Berns, 1981; Jensen et al., 1991). Thus, N. haematococca is a particularly advantageous organism with which to elucidate and interpret the various nuclear phenotypes of dynein mutation, as illustrated by the results of this study.

The SPB is a nucleus-associated microtubule-organizing center (MTOC) that is involved in most nuclear motility in fungi (Aist, 1995; Hagan and Yanagida, 1997). Laser microsurgery experiments and treatment with MT-depolymerizing drugs showed that astral MTs ofN. haematococca are involved in the pulling force transmitted to the SPB during anaphase B (Aist and Berns, 1981; Aist et al., 1991). The role of the SPB in maintaining nuclear position in this fungus during interphase was demonstrated by an optical force trap experiment (Berns et al., 1992), and a similar role for the SPB was emphasized in a recent study of the fission yeast, Schizosaccharomyces pombe (Hagan and Yanagida, 1997).

In this paper we describe cloning, sequencing and disruption of the gene, DHC1, for a cytoplasmic dynein heavy chain (CDHC) in N. haematococca, and document previously unreported roles for dynein in cytoplasmic MT organization, SPB motility and anchoring of nuclei in nonmitotic cells of a filamentous fungus. Brief summaries of these results have been published (Inoue et al., 1996, 1997).

Strains of N. haematococca Berk. and Br., mating population VI (anamorph Fusarium solani (Mart.) Sacc. f. sp. pisi) used in this study included: (1) T213, the wild-type progenitor of mutants generated for this study; (2) Cu1, a stable dhc1^- transformant carrying the gene for resistance to hygromycin B (hygB) inserted via a double crossover homologous recombination event between linearized vector pUCDHS01 (Table 1, Fig. 1) and the coding sequence of the central motor domain of DHC1; (3) Cu3, an unstable dhc1^- transformant carrying plasmid pCRS2911 integrated via a single crossover homologous recombination event between circular vector pCRS2911 (Table 1, Fig. 1) and the coding sequence of the central motor domain of DHC1; and (4) S01-2, a DHC1⁺ transformant carrying pUCDHS01 at an ectopic site, leaving DHC1 intact (Fig. 1).

Cultures were grown on V8-agar plates (200 ml V8 juice, 3 g CaCO₃, 20 g agar per liter) and/or YEG-Gelrite medium plates (10 g yeast extract, 5 g glucose, 10 g Gelrite per liter) under near-UV light at 20°C. For DNA extraction, the fungus was grown in a glucoseasparagine (GA) medium (VanEtten and Stein, 1978). For regeneration of transformants from protoplasts, top (0.1% yeast extract, 0.1% casamino acids, 1.2 M sucrose, 1% agar) and base (0.1% yeast extract, 0.1% casamino acids, 0.6 M sucrose, 1.5% agar) media were used. Transformants were initially grown on V8-agar plates with 100 µg/ml hygromycin B. Resultant conidia were spread on transformation base agar, and single colonies were isolated. Transformants were grown in GA medium containing 50 µg/ml hygromycin B prior to DNA extraction.

Crosses were performed following the method described by VanEtten (1978). Wild-type T213 is female-sterile, and consequently all the CDHC mutants created in this study were female-sterile. Tester strain 3-36, a hermaphrodite obtained from Dr Hans VanEtten, University of Arizona, Tuscon, AZ, was used as the female. Both male and female strains were grown on V8-agar slants for 2 weeks. Conidia of the male strain were transferred to the female by first adding 10 ml of dH₂O to the male, agitating the culture to suspend the conidia, and then transferring the suspension to the female strain and draining. Perithecia were formed, and mature ascospores oozed from the beaks of perithecia after 4 weeks. Random ascospores were germinated on YEG-Gelrite agar. Individual colonies were picked, subcultured and analyzed.

Standard molecular techniques were performed following Sambrook et al. (1989) or as otherwise noted. Fungal genomic DNA was extracted by the method of Miao et al. (1991) and purified on CsCl gradients. Cloned DNA fragments were sequenced using TaqCycle automated sequencing with DyeDeoxy terminators (Applied Biosystems Inc., Norwalk, CT) by DNA SERVICES at the Cornell University Sequencing Facility using oligonucleotide (18-23mers) primers, designed to correspond to previously determined sequences, synthesized by the Cornell University Biotechnology Oligonucleotide Synthesis Facility. DNA sequences were analyzed using the software programs EditSeq and MapDraw (DNAStar Inc., Madison, WI). Homology searches were performed using BLAST search software against the National Center for Biotechnology Information (Bethesda, MD) database. The sequence has been deposited (GenBank accession number U84215).

A pair of degenerate oligonucleotide primers corresponding to highly conserved regions of CDHC proteins from A. nidulans (Xiang et al., 1994), S. cerevisiae (Eshel et al., 1993), rat (Zhang et al., 1993) and Drosophila (Li et al., 1994) were designed as follows: DPI, 5‘-CC(C/I)GC(C/I)GG(C/I)AC(C/I)GG(C/I)AA(A/G)AC-3’ (sense strand, targeting amino acid sequence PAGTGKT) and DPIII, 5‘-A/G)TC(A/G)CC(A/G)TC(A/G)AA(A/G)ACGATCC-3’ (antisense strand, targeting amino acid sequence WIVFDGD), corresponding to positions 1947-1953 and 2315-2321 of the sequence. The PCR cycling regimen included 35 cycles of 1 minute denaturation at 95°C, 2 minutes annealing at 37°C and 2 minutes extension at 72°C. With N. haematococca genomic DNA as template, an 1125 bp DNA fragment was amplified. The band was cut out of the gel and treated with GeneClean (BIO 101 Inc). PCR was repeated using the GeneCleantreated DNA as template and the same primers and PCR cycling regimen, to obtain sufficient DNA for cloning into the pCR(tm)II vector (In vitrogen, Carlsbad, CA) to construct vector pCRS29 (Table 1). The identity of the 1125 bp DNA fragment was confirmed by sequencing.

The 1125 bp DHC1 insert of pCRS29 was transfered from pCRS29 to pUC18 by digesting both with EcoRI and religating to produce pUCS007 (Table 1). The gene disruption vector pUCDHS01 (Table 1; Fig. 1) was constructed by inserting the 2.3 kb SalI fragment carrying hygB from pCWhyg1 (Table 1) into the XhoI restriction enzyme site (between the two P-loop motifs) in the insert of pUCS007, thereby interrupting the DHC1 sequence. pUCDHS01 was linearized with KpnI to promote a double-crossover event between homologous sequences of DHC1 on the vector and the genomic copy of DHC1 (Fig. 1A).

Transformation of N. haematococca was performed according to Wasmann and VanEtten (1996) with some modifications. Briefly, wildtype strain T213 was cultivated for 7-10 days on V8-juice agar plates to produce conidia. Conidia were suspended in water and transferred to 300 ml flasks containing 100 ml GA medium. The flasks were shaken (180 rpm) at 25°C and the mycelium was harvested after 14-16 hours. Mycelium (1 g) was suspended in osmoticum (1.2 M MgSO₄, 10 mM sodium phosphate) with 100 mg Novozyme 234 (InterSpex, Foster City, CA) and 100 µl of sterile β-galacturonidase (Sigma, St Louis, MO), and the flask was gently agitated at 30°C for 2 hours. Protoplasts were filtered through two layers of cheesecloth, pelleted by centrifugation, washed three times in 10 ml STC (1.2 M sorbitol, 25 mM CaCl₂, 25 mM Tris-HCl, pH 7.0) and resuspended in STC to a protoplast concentration of 5×10⁷/ml. pUCDHS01 or pCRS2911 (5-10 µg; see below) was mixed with 5×10⁶ protoplasts in 100 µl, and 1.2 ml of PTC (20% PEG (Sigma No. P-3640), 25 mM CaCl₂, 25 mM Tris-HCl, pH 7.0) was added in a stepwise fashion (200 µl, 400 µl and 600 µl). Protoplasts were then mixed with 4 ml of molten (48°C) regeneration top agar and poured onto a plate containing 15 ml of regeneration base agar. Regeneration plates were incubated at 32°C for 22-23 hours and overlayed with 5 ml of 1% water agar containing hygromycin B to yield a final concentration of 100 µg/ml hygromycin B. Colonies that grew through the overlay were transferred to fresh V8 agar containing 100 µg/ml hygromycin B and incubated at 20°C. Transformants were purified by isolating and culturing single conidia.

Vector pCRS2911 (Table 1; Fig. 1) was designed to aid recovery of the entire genomic copy of DHC1. A HindIII fragment of pCWhyg1 carrying hygB was inserted into the HindIII site of the polylinker in pCRS29 adjacent to the 1125 bp DHC1 fragment to construct pCRS2911, which was transformed in circular form into wild-type strain T213. One transformant, Cu3, with a mutant phenotype was chosen for further analysis (Fig. 1B). Integration of the vector via a single homologous recombination event was confirmed by gel-blot analysis (Fig. 1C). KpnI, which cuts once in pCRS2911, was used to recover the 5‘-flanking region of DHC1 (the recombinant plasmid was called pCRNDK1) and ApaI, which cuts once in pCRS2911, was used to recover the 3‘-flanking region of DHC1 (the recombinant plasmid was called pCRNDA1) (Fig. 1B). Digested DNAs were ligated and transformed into E. coli. Ampicillin-resistant colonies were selected, and plasmids bearing chromosomal DNA were isolated and purified using a Qiagen midi column (Qiagen, Inc., Chatsworth, CA). In this way, the whole DHC1 gene was recovered on two plasmids, which were subsequently sequenced.

Conidia of isolates T213 and Cu1 were incubated in a drop (approx. 20 µl) of GA medium on a glass slide in a moist chamber at 20°C overnight in the dark. The hyphal and conidial morphology of conidial germlings was observed using phase-contrast optics of a Zeiss Universal microscope. Nuclear distribution was viewed with a Zeiss Photomicroscope II by fluorescence microscopy after fixing and staining intact hyphal cells with DAPI (1 µg/ml) as described by Taga and Murata (1994).

Cellular phenotypes were observed by video-enhanced, phase-contrast microscopy (Aist and Bayles, 1988; Aist et al., 1991). Briefly, cultures grown on glass microscope slides coated with YEG-Gelrite medium were videotaped at 10,000× magnification using phase-contrast optics and real-time image processing. Measurements of hyphal tip morphology were performed using an Argus 20 image processor (Hamamatsu Photonics K. K., Hamamatsu City, Japan). Statistical analyses of these data were performed using Data Desk^® (Data Description Inc., Ithaca, NY). The schematic diagrams of daughter nuclear behavior in the mutant Cu1 were drawn by Canvas(tm) 5 (Deneva Software, Miami, FL) based on free-hand sketches made while observing video-taped sequences recorded continuously in vivo for about 1 hour. Data for plotting migration of SPBs was acquired and calculated by a custom software program. For optical force trap experiments, cells were observed with differential interference contrast optics at 7,200× magnification, and the nucleolus of interphase nuclei was trapped and pulled by a continuous wave YAG laser (Cole et al., 1995). The power of the laser at the specimen was adjusted to about 140 mW (estimated), which did not produce visible heat damage in the cells. Visualization of MTs by immunofluorescence video-microscopy was performed according to the protocol in Aist et al. (1991).

Sequencing of the 1,125 bp insert in pCRS29 (Table 1) with M13 (-40) forward and reverse primers followed by a database search revealed that the predicted amino acid sequence has 90% identity to that of the corresponding region of N. crassa CDHC and 84% identity to that of A. nidulans CDHC, and contains part of the first (PAGTGKT) and second (GNSGSGKS) P-loop motifs.

In contrast to the existence of several isoforms of cytoplasmic dynein in mammalian cells (Vaisberg et al., 1996), gel-blot analysis of the genomic DNA of N. haematococca with a 336 bp PCR probe encoding the most highly conserved region (the first P-loop motif) of DHC1, indicated that N. haematococca has only one copy of DHC1 (data not shown). Sequencing of inserts in pCRNDK1 and pCRNDA1 (Table 1) revealed that N. haematococca DHC1 has a 13,157 bp ORF interrupted by introns of 52 bp (interrupting alanine codon 101) and 55 bp (between aspartic acid codon 4203 and tyrosine codon 4204). DHC1 is predicted to encode a polypeptide of 493,514 Da with an amino acid sequence that is 78.0% identical to N. crassa CDHC RO1, 70.2% identical to A. nidulans CDHC NUDA and 24.8% identical to S. cerevisiae CDHC DYN1. All four P-loops are identical to those of N. crassa, and the first and third P-loop motifs are 100% conserved relative to A. nidulans (Table 2). Two introns are conserved in the N. haematococca and N. crassa genes, but only the first intron is found in the A. nidulans gene. The sequence has been deposited (GenBank accession number U84215).

Most hygromycin B-resistant colonies could not be distinguished from wild type on non-selective medium, and DNA gel-blot analysis of genomic DNAs showed that the vector in each of these transformants had integrated at an ectopic site (data not shown). A few colonies, however, were easily distinguished from wild type. They were small, slightly yellowish and highly compact compared to the large, white, filamentous colonies of wild type or ectopic transformants (Fig. 2). Two transformants were chosen for further analysis, one (Cu1) from the pUCDHS01- and one (Cu3) from the pCRS2911-mediated transformation experiments

(Fig. 1). Evidence that aberrant morphology resulted from gene disruption was obtained by gel-blot analysis of genomic DNA of Cu1 and Cu3 digested with StuI and probed with the 1125 bp DHC1 fragment or with hygB (Fig. 1C). StuI digestion of Cu1 yielded a diagnostic 12.1 kb fragment and StuI digestion of Cu3 yielded a diagnostic 16.9 kb fragment. These fragments (each the sum of the fragment in wild type + the transformation vector) hybridized with both probes, and wild type yielded a 9.8 kb fragment that hybridized with only the 1125 bp DHC1 fragment (probe 1).

When Cu1 and the wild-type tester strain, 6-36, were crossed, normal perithecia with 2-celled, morphologically normal ascospores were formed. Of 288 random ascospores analyzed for segregation of mutant morphology, 224 germinated. Of these 224, 123 showed wild-type colony morphology and 101 showed Cu1 colony morphology, demonstrating 1:1 segregation (P<0.01).

To test for co-segregation of the Cu1 colony morphology with hygromycin B resistance, 26 mutant and 20 wild-type progeny were chosen randomly and transferred to YEG-Gelrite medium with or without hygromycin B. All 26 mutant progeny were hygromycin B-resistant and grew as fast on hygromycin B medium as on medium without hygromycin B, whereas the 20 wild-type progeny did not grow at all on hygromycin B medium but grew with wild-type morphology on nonselective medium. This result confirmed that the DHC1 gene was disrupted by insertion of hygB at a single site and that the gene disruption caused the aberrant morphology.

Disruption of DHC1 is not lethal, but colonies grew slowly (0.83±0.10 mm/day), at about 33% of the wild-type rate (2.49±0.16 mm/day) on YEG-Gelrite plates at 20°C under near-UV light (Fig. 2). The colonies were compact and generally similar in appearance to col (colonial) (Steele and Trinci, 1977) or ro (ropy) mutants (Plamann et al., 1994) of N. crassa and the nudA mutant of A. nidulans (Xiang et al., 1994).

Mutant Cu3 was created by single crossover homologous integration between the DHC1 gene and vector pCRS2911, resulting in duplicated regions (both mutant) of the gene separated by the vector (Fig. 1). Colony morphology and growth (radial growth rate 0.99±0.49 mm/day) of Cu3 were similar to those of Cu1 (0.74±0.08 mm/day) on selective medium. After 2 days on nonselective medium, Cu3 started to grow as fast as wild type (2.88±0.31 mm/day), and its morphology began to look like that of wild type, i.e. straight hyphae were observed at the margins of the colonies. In contrast, straight hyphae have never been observed in the colonies of Cu1. Proof that the change of morphology was correlated with genetic reversion (dhc1- to DHC1+) came from gel-blot analysis of DNA from Cu3 and wild-type T213 grown in nonselective liquid culture (Fig. 1). A diagnostic 16.9 kb Stu 1 band, which hybridized to DHC1 (lane 5 of Fig. 1C) and hygB (lane 11 of Fig. 1C), disappeared and was replaced by a 9.8 kb band corresponding to the native band that did not hybridize with hygB when DNA was examined from Cu3 grown in nonselective medium (lane 5 of Fig. 1C).

In contrast to wild-type strain T213, hyphae of the CDHC-deficient mutant, Cu1, are spiral, highly branched (Fig. 3A,C) and frequently contain large vacuoles. The distance between septa (the length of one cellular compartment) is much more variable than in wild type. At low frequency (about 4.7% among 706 conidia counted), the mutant produced abnormally shaped macroconidia: comma-, boomerang-or seagull-shaped, in contrast to the 100% crescent-shaped macroconidia of wild type (Fig. 4).

The growth rate (0.92 µm/minute) of individual hyphal tips of Cu1 observed in vivo by video microscopy was 33% of the wild-type rate (2.79 µm/minute). The mean size (median crosssectional area) of Spitzenkörper was 0.81 µm², about 46% of that of wild type (1.74 µm²). The growth rate and Spitzenkörper size of tip cells of ectopic transformant SO1-2 were similar to those of wild type.

Fig. 5 illustrates the organization and distribution of organelles in hyphal tip cells. In some cells of Cu1, a distinct Spitzenkörper could not be visualized at the growing apical dome. The Spitzenkörper in Cu1 frequently shifted its position laterally within the apical dome, whereas those in control cells typically remained centrally positioned throughout extended periods of tip growth. Mitochondria were usually observed within the apical dome, as in wild type. Clusters of phase-dense, medium-sized granules, different from the usual lipid bodies, were often observed in mutant hyphal tips. Sometimes these granules were near the apical dome area where they apparently blocked the forward migration of mitochondria (15 apical domes of 39 observed), and sometimes they were behind mitochondria that were migrating forward as the apical dome advanced (Fig. 5C). This cytoplasmic phenotype has not been reported previously in a CDHC-deficient mutant of a fungus. The average distance of the first mitochondrion from the apical dome in Cu1, when the cluster of granules apparently interrupted the migration of mitochondria to the tip, was 6.38±2.31 µm (n=15), a statistically significant difference from the corresponding distances in 33 wild-type cells (1.54±0.36 µm) and 24 Cu1 cells (2.63±1.05 µm) in which the cluster of granules was behind the mitochondria.

Immunofluorescence staining of MTs showed that they were distributed throughout the tip cells of all isolates (Fig. 5E-H).

We performed immunofluorescence staining of the MTs to confirm both the existence of aster-like MT arrays at the SPBs during the post-migration period and the spatial interaction between cytoplasmic MTs and SPBs during interphase. Well-developed aster-like MT arrays radiating from SPBs of apparently migrating nuclei were observed in the wild type, T213 and the ectopic transformant, SO1-2 (Fig. 6A,B). Cytoplasmic MTs commonly radiated from the SPBs of interphase nuclei, indicating that the SPB was a focus of cytoplasmic MTs in T213 and S01-2 (Fig. 6E,F). In contrast, neither a prominent interphase aster-like MT array nor an MT focal point at the SPB was observed in the dynein mutant, Cu1 (Fig. 6C,D,G,H). These unique results indicate that the SPB in the mutant cannot focus and/or tether MTs to maintain the astral array during the post-mitotic nuclear migration period and, subsequently, during interphase.

SPBs of wild-type nuclei are always active and vigorously moving (Fig. 7A-D), and sometimes they translocate the nucleus through the cell (cf. Wilson and Aist, 1967), elongating the nucleus in the process. In contrast, such movement of SPBs was not observed in the dynein mutant, Cu1 (Fig. 7E-H), and the nucleus was always spherical. From these unique observations, we hypothesized that cytoplasmic dynein is required for manifestation of the force that pulls on and positions the SPBs of interphase nuclei and that interphase nuclei in the mutant cells are adrift because the SPBs are inactive.

To test the above hypotheses, optical force trap (Berns et al., 1991) experiments were performed (Fig. 8). We trapped the nucleolus and tried to pull it, along with the nucleus, in the opposite direction of the SPB. In the S01-2 (control) cells, the SPB was actively moving and always resisted the pulling force applied by the laser trap. The SPB consistently moved away from the laser trap and pulled the whole nucleus away from it, causing the nucleolus to escape from the trap (Fig. 8A-C). In contrast, in cells of the CDHC-deficient mutant, Cu1, the whole nucleus along with the SPB was often pulled by the laser trap (Fig. 8D-F). This unique experiment revealed that a pulling and/or anchoring force that is normally acting on the SPBs during interphase in control cells was absent from the CDHC-deficient cells.

In wild type and a DHC1+ ectopic (control) transformant, S01-2, the daughter nucleus closest to the tip was quite motile and always teardrop-shaped, with the SPB always at the narrow, leading end of the nucleus, indicating that it pulls the nucleus as it migrates (Fig. 9B). The average rates of apical post-mitotic nuclear migration in T213 and S01-2 were 9.63±2.58 µm/minute and 6.95±1.77 µm/minute, respectively (±1 s.d.; n=3). In contrast, prominent post-mitotic nuclear migration was not observed in the CDHC mutant, Cu1 (Fig. 9A). Daughter nuclei were spherical, not teardrop-shaped. They migrated toward the tip, but the rate of migration was only 1.80±0.35 µm/minute (n=3), much slower than those of T213 and S01-2 and close to the rate of tip growth alone. Obviously, the mechanism for postmitotic nuclear positioning was not operative in the mutant.

The schematic diagrams of Cu1 in Fig. 10 illustrate, for the first time in a filamentous fungus, daughter nuclear behavior in relation to septum formation following mitosis in living cells of a CDHC-deficient mutant. The daughter nuclei were separated initially by the elongating spindle. In sequence A, we observed two of the daughter nuclei passing each other, thus switching their positions in the hyphae, possibly due to continued elongation of the spindle. After the spindle elongation period in both sequences, there was no post-mitotic nuclear migration for about 5 minutes, and subsequently all daughter nuclei moved very slowly and independently, giving the impression that they were drifting passively in or with the cytoplasm. In sequence A, a cluster of nuclei was formed after septum formation. In contrast, the nuclear cluster was formed before the septum in sequence B. Note that the absence of post-mitotic nuclear migration left the daughter nuclei in clusters, and that further clustering of the nuclei resulted from slow, apparently passive, movement of nuclei with the cytoplasm, especially in Fig. 6B. This inference of passive nuclear motility is consistent with the results of the laser trapping experiment (Fig. 8).

Compared to wild type, nuclear distribution of the CDHC-deficient mutant, Cu1, was nonuniform (Fig. 3). Some areas of Cu1 hyphae contained clustered nuclei, whereas some branches were anucleate. Almost all nuclei in most mutant hyphae were clustered near septa, and the number of nuclei in a cluster varied from 2 to more than 10.

The efficiency of transformation with a circular vector was slightly higher (1.5 times) than that with a linear vector; however, the former type of integration event is unstable. The conversion of the colony morphology from mutant to wild-type phenotype was apparent within 2 days after transfer to nonselective medium. Gel-blot analysis of genomic DNA (Fig. 1) showed that reversion occurred by elimination of the transforming DNA. Eviction of transforming DNA is commonly found in genomes of transformants carrying duplicated regions and is thought to be the cause of mitotic instability (Dhawale and Marzluf, 1985; Dunne and Oakley, 1988; Keller et al., 1991; Petes and Hill, 1988; Shortle et al., 1984; Upshall, 1986).

The stable mutant was viable and showed multiple phenotypes. It is possible that additional phenotypes exist, since DHC1 was inactivated by an insertion mutation (rather than by deletion of the entire gene), leaving open the chance of a partially active protein. Although it can be argued that the deletion of the entire CDHC gene might be lethal, as suggested by Koonce and Samsó (1996) for Dictyostelium and Gepner et al. (1996) for Drosophila, evidence has accumulated that CDHC may not be essential in filamentous fungi. In N. crassa, cytoplasmic dynein function was disrupted by a number of different mutations; all were viable and showed similar phenotypes (Bruno et al., 1996). Furthermore, in A. nidulans, conditional null nudA mutants were viable, and deletion of the central region of NUDA was not lethal (Xiang et al., 1995). Disruption of p150^Glued, the gene encoding the largest subunit of the dynactin complex, which is required for proper cytoplasmic dynein function, showed phenotypes similar to the other ro mutants in N. crassa (Tinsley et al., 1996). Furthermore, Eshel (1995) found that the phenotype of a single amino acid substitution in the first P-loop motif of S. cerevisiae CDHC was not different from that of deletion of the entire dynein motor domain. Thus, one might anticipate that the presence of a truncated, partially active CDHC in Cu1 would not produce substantially misleading phenotypes.

Ascospores produced from crosses in which the mutant functioned as the male were normal. This result indicated either that cytoplasmic dynein is not involved in meiosis or that meiosis is mediated by maternal cytoplasmic dynein.

The Spitzenkörper is an aggregate of secretory vesicles at the hyphal apex that determines the form and direction of hyphal growth (Bartnicki-Garcia et al., 1989, 1995; Grove and Bracker, 1970). A prominent Spitzenkörper was observed consistently in the apical dome of growing hyphal tip cells of wild-type strain T213 of N. haematococca. However, a well-defined Spitzenkörper could not always be observed in growing hyphae of the CDHC-deficient mutant, possibly because in the mutant the Spitzenkörper is smaller and may frequently move out of the median focal plane of the apex. It is also possible that the organization of apical vesicles in the mutant cell commonly fluctuates between aggregated (visible) and dispersed (invisible). Spiral, curly, highly branched and/or knobby hyphae of the mutant may be caused by the lack of proper positioning and organization of Spitzenkörper in the apical dome, resulting in abnormal targeting of secretory vesicles. A similar conclusion was reached by Wu et al. (1998) with respect to a kinesindeficient mutant of N. haematococca and by Seiler et al. (1997) with respect to a kinesin-deficient mutant of N. crassa.

Several lines of evidence suggest that cytoplasmic dynein may be associated with retrograde transport of precursors of hyphal tip formation (Beckwith and Morris, 1995; Plamann et al., 1994; Xiang et al., 1995). In this study, we measured the Spitzenkörper and found that the average size in the mutant was considerably smaller than that in wild type. This is unique evidence linking the presumed reduction of apical transport of secretory vesicles in the mutant to a marked effect on the size of the apical cluster of the same vesicles. Our demonstration of an approximately normal distribution of MTs in tip cells of the mutant indicates that this phenotype was not due to an indirect effect on the MTs.

Induction of branches is commonly observed in fungal hyphae treated with antimicrotubule agents (Howard and Aist, 1977; Raudaskoski et al., 1994; Rupes et al., 1995; That et al., 1988). Increased branching of the N. crassa cot-1 mutant, whose colony morphology resembles that of our CDHC mutant, was suggested to be associated with regulation of linear growth of hyphae and the transport of precursors to the tip (Steele and Trinci, 1977). Seiler et al. (1997) inferred that increased branching in a kinesin-deficient mutant of N. crassa resulted from the development of multiple growth zones due, in part, to a defect in transport of secretory vesicles to the growing apex. We suspect that similar mechanisms cause increased branching in our CDHC-deficient mutant of N. haematococca.

This brings us to the question of how hyphal tip cells of a CDHC-deficient can grow at all, if dynein is involved in the apical transport of secretory vesicles. One possibility is that N. haematocca may have a different kind of MT-associated motor protein that can participate in such transport, and a kinesin has been found recently that would seem to be a good candidate (Wu et al., 1998). Alternatively, because an actin-myosin mechanism is capable of supporting hyphal tip growth (Heath, 1994), it is possible that growth of CDHC-deficient mutants is due to an altogether different transport system based on F-actin rather than MTs.

We first thought that cytoplasmic dynein was functionally involved in the motility of the SPB through SPB-associated MTs, tethering the MTs at the cell cortex, as suggested for centrosome rotation and the determination of cell division axis in Caenorhabditis elegans (Hyman, 1989; Gönczy and Hyman, 1996), SPB positioning in S. cerevisiae (Carminati and Stearns, 1997; Shaw et al., 1997; Yeh et al., 1995), maintenance of the cell cytoskeleton in Dictyostelium (Koonce and Samso, 1996; Koonce, 1996) and the positioning of the centrosome within animal cells (Kellogg et al., 1994). Instead, we demonstrated, for the first time in a fungus, that cytoplasmic dynein plays a critical role in the formation and maintenance of the aster-like array of interphase MTs. Although this result was unanticipated, it is consistent with related observations with other organisms. Cytoplasmic dynein was found to be required for formation of mitotic asters in mammalian cells (Gaglio et al., 1996). It is involved in the proper spindle formation of Xenopus eggs, especially in the focusing of the MT array at the poles (Verde et al., 1991; Vaisberg et al., 1993; Merdes et al., 1996; Heald et al., 1996). Dynein-like motors also bring MT minus ends together and organize a radial array of MTs in cytoplasmic fragments of melanophores (Rodionov and Borisy, 1997). We speculate that cytoplasmic dynein in N. haematococca is involved in holding of free MTs and/or tethering of polar MTs (Aist and Bayles, 1991) and/or crosslinking and stabilizing of MTs (Jensen et al., 1991) to form aster-like arrays at interphase. These results indicate that an MT-based motor protein like cytoplasmic dynein may be not only a motor, but also a key structural component to focus and, perhaps to tether, MTs to MTOCs in cells, as suggeested recently by Gaglio et al. (1997) and Merdes and Cleveland (1997). Whether or not DHC1 also generates a force that is transmitted to the SPBs via the aster-like MT arrays is still unknown.

Because our CDHC-deficient mutant, Cu1, fails to focus and attach cytoplasmic MTs to the SPBs, we infer that the nuclear phenotypes we observed (absence of post-mitotic nuclear migration, SPB motility and nuclear anchoring) are all a result of this failure. Thus in wild type, nuclear migration, SPB motility and anchoring of interphase nuclei all appear to be mediated by cytoplasmic MTs attached to the SPBs. Although our results are not the first or only evidence for all of these roles of aster-like MT arrays in fungi (Aist, 1995), they do provide a unique demonstration of them. Moreover, our results provide the first experimental confirmation that interphase aster-like MT arrays attached to the SPBs are involved in SPB motility and nuclear positioning in filamentous fungi. A similar conclusion was drawn recently concerning SPBs and astral MTs of the yeast, S. cerevisiae (Carminati and Stearns, 1997; Shaw et al., 1997).

Formation of both conidia and ascospores was severely suppressed in nudA mutants of A. nidulans (Morris et al., 1995; Xiang et al., 1995), whereas conidia were formed abundantly by ro mutants of N. crassa (Plamann et al., 1994; Tinsley et al., 1996) and by the CDHC-deficient mutant, Cu1, of N. haematococca. We found, however, that the latter produced abnormally shaped conidia at low frequency. We suggest that cytoplasmic dynein is critical in the conidiation process to a different extent in different fungi. This difference may be closely related to the fact that conidiation of Aspergillus requires extensive post-mitotic nuclear migration through elaborate conidiophores, whereas in Nectria and Neurospora, conidia are formed on simple conidiophores requiring less extensive, and presumably less precisely orchestrated, nuclear migration.

Nonuniform nuclear distribution is one of the most prominent phenotypes of cytoplasmic dynein mutants of filamentous ascomycetes (Bruno et al., 1996; Morris, 1976; Plamann et al., 1994). In the N. haematococca CDHC-deficient mutant, almost all nuclei were clustered near the septum, a condition referred to in the nudA mutant of A. nidulans as ‘septal clusters of nuclei’ (Morris, 1976). Our unique in vivo observations of post-mitotic nuclei (Fig. 10) confirmed that cytoplasmic dynein is required for normal nuclear behavior, especially during the period of post-mitotic nuclear migration and initial positioning of nuclei in hyphae for interphase. We showed for the first time that the septal cluster of nuclei was caused primarily by lack of post-mitotic nuclear migration, with a minor contribution from post-mitotic mass flow of cytoplasm which appeared to move some nuclei passively into even tighter clusters.

We thank Q. Wu, Dr S. Wirsel and Dr T. Arie for their technical advice, Dr T. C. Huffaker for suggestions, Drs X. Xiang and N. R. Morris for advice and generous gifts of the CDHC gene fragments and the nudA strains of A. nidulans, and Drs C. C. Wasmann and H. D. VanEtten for plasmid pCWhyg1 and N. haematococca tester strains.

We appreciate the access provided to us of his laser microbeam facility by Dr Conly L. Rieder and the expert technical assistance of Mr Richard Cole during the laser microbean experiments. This paper is part of a PhD dissertation by S. I. This research was supported by grants from the National Science Foundation (numbers DCB-8916338 to J. R. A. and MCB-9305703 and MCB-9408249 to J. R. A., O. C. Y. and B. G. T.), and by NIH National Center for Research Resources grant P41-01219, which partly supports the Wadsworth Center’s Biological Microscopy and Image Reconstruction facility as a National Biotechnological Resource.