Transportation of Aspergillus nidulans Class III and V Chitin Synthases to the Hyphal Tips Depends on Conventional Kinesin

Cell wall formation and maintenance are crucial for hyphal morphogenesis. In many filamentous fungi, chitin is one of the main structural components of the cell wall. Aspergillus nidulans ChsB, a chitin synthase, and CsmA, a chitin synthase with a myosin motor-like domain (MMD) at its N-terminus, both localize predominantly at the hyphal tip regions and at forming septa. ChsB and CsmA play crucial roles in polarized hyphal growth in A. nidulans. In this study, we investigated the mechanism by which CsmA and ChsB accumulate at the hyphal tip in living hyphae. Deletion of kinA, a gene encoding conventional kinesin (kinesin-1), impaired the localization of GFP-CsmA and GFP-ChsB at the hyphal tips. The transport frequency of GFP-CsmA and GFP-ChsB in both anterograde and retrograde direction appeared lower in the kinA-deletion strain compared to wild type, although the velocities of the movements were comparable. Co-localization of GFP-ChsB and GFP-CsmA with mRFP1-KinArigor, a KinA mutant that binds to microtubules but does not move along them, was observed in the posterior of the hyphal tip regions. KinA co-immunoprecipitated with ChsB and CsmA. Co-localization and association of CsmA with KinA did not depend on the MMD. These findings indicate that ChsB and CsmA are transported along microtubules to the subapical region by KinA.


Introduction
Hyphae of filamentous fungi represent extremely polarized cells. Since "polarized growth" is a general growth mechanism common in the elongation of cells as distinct as neurons or pollen tubes, as well as in the differentiation of embryos [1], the mechanism of hyphal tip growth in filamentous fungi may provide a helpful basic model. This mechanism is thought to be highly elaborate and is supported by the continuous delivery of vesicles containing proteins required the transport of secretory vesicles, early endosomes, peroxisomes, and mRNPs containing mRNA [11,[43][44][45][46][47][48][49][50][51][52][53]. Zhang et al. [53] and Yao et al. [51] showed that dynein and dynactin localize along microtubules in the ΔkinA mutant, which provide good support for the idea that KinA transports dynein/dynactin along microtubules to the plus ends.
Although the localization of chitin synthases has already been investigated in some filamentous fungi, their localization mechanisms remain largely unknown. In N. crassa chitin synthases are thought to be transported on special vesicles called chitosomes [54,55]. In basidiomycete dimorphic yeast Ustilago maydis, it was reported that the proper localization of class V chitin synthase Msc1 depends on the presence of microtubules and/or actins, that its transport along microtubules does not depend on its MMD, and that the MMD is required for the tethering of Mcs1 at the hyphal tips to increase the efficiency of its exocytosis [36,50]. In this study, we report the transporting mechanisms of classes III and V chitin synthases, ChsB and CsmA, to hyphal tips in A. nidulans and the role of the MMD of CsmA in the transport process.

Constructions of plasmids and A. nidulans strains
To create the strains that produced CsmA with an EGFP tag at its N-terminus, we constructed the plasmid pMAEC as follows: The 0.8-kb KpnI-PstI fragment of pM-ALC2 [25] was cloned into the KpnI-PstI site of pUC19, yielding pUCCSM2. The 0.7-kb fragment encoding EGFP was amplified by PCR from pEGFP (Clontech) using primers 5 0 -GCGTTAACATGGTGAG CAAGGGCGAGG-3 0 and 5 0 -GCCCATGGTCTTGTACAGCTCGTCCAT-3 0 , in which the underlined letters represent HincII and NcoI recognition sites, respectively, digested with HincII and NcoI, and cloned into the NcoI site of pUCCSM2, yielding pUCCSM. The 1.5-kb KpnI-PstI fragment of pUCCSM was cloned into the KpnI-PstI sites of pM-ALC2, yielding pMAEC. The 4.1-kb EcoRI and PstI fragment of pMAEC was used for the transformation of the ABPU1 and SNR1 (ΔkinA) strains (Table 1). Two transformants obtained from each strain, in which the DNA fragment encoding EGFP was integrated into the 5 0 -region of the csmA cording region, were selected and confirmed by Southern blot analysis using the 1.0-kb SpeI-XhoI fragment of pM-ALK-CHS5 [23] as a probe, and were designated as EGFP-CsmA1 and 2, and EGFP-Cs-mAΔkinA1 and 2, respectively (Table 1).
We obtained at least two transformants in each case after the confirmation of their relevant DNA structures by Southern blot analysis. They exhibited the same phenotypes with respect to growth and hyphal morphology. We used one of them for further experiments.

Light/fluorescence microscopy
Live-cell imaging of germlings and young hyphae: cells were grown on cover slips in 0.5 ml MMGly (de-repression of the alcA promoter) were used. Although GFP-ChsB was expressed under the native promoter, to compare the data of GFP-CsmA, which was expressed under the alcA promoter, MMGly was selected. Cells were incubated at 30°C overnight or 1 day. The cover slips were mounted on a microscope slide. Tempcontrol mini (Pepcon) was used as needed to control the temperature of the specimens during microscopy. Images were captured using an Axiophot microscope using a Planapochromatic 63 times oil immersion objective lens, the Zeiss AxioCam MRM camera (Zeiss, Jena, Germany), and the HBO103 mercury arc lamp (Osram) or HXP 120 (Zeiss, Jena, Germany) possessing faster speed wavelength switching. Images were collected and analyzed using the AxioVision and Zen system (Zeiss). Kymographs were made using ImageJ software (http://rsb.info.nih.gov/ij/) and Zen system (Zeiss).

Localization of ChsB and CsmA
In previous work chitin synthase class III ChsB, expressed from the native promoter and fused to GFP, was observed at hyphal tips and forming septa in A. nidulans [24]. Here we analyzed the localization mechanism for ChsB and the class V chitin synthase CsmA (S1 Fig). To analyze the dynamic localization of CsmA in living hyphae, we replaced the csmA gene with a C-terminally GFP tagged version (CsmA-GFP). These strains, however, exhibited phenotypic abnormalities similar to those observed in csmA-deletion mutants and only weak fluorescence of CsmA-GFP (data not shown). This suggested that the GFP tag at the C-terminus interferes with the function of CsmA. Thus, we constructed a strain expressing CsmA tagged with GFP at its N-terminus (GFP-CsmA) under the control of the alcA promoter (Table 1). This strain grew well and did not exhibit phenotypic abnormalities when the alcA promoter was derepressed (medium containing glycerol as a carbon source) or when the alcA promoter was induced (medium containing threonine and fructose as carbon sources). The correct insertion of the construct in two strains expressing GFP-CsmA (EGFP-CsmA1 and 2) was confirmed by Southern blot analysis (S2 Fig). The production of the GFP-CsmA protein was confirmed by Western blot analysis using anti-GFP antibodies (S3A Fig).
Fluorescence of GFP-CsmA was observed at hyphal tips and at forming septa in addition as moving spots in the cytoplasm (Fig 1A, and data not shown). At the apical membrane, GFP-CsmA was found at the Spitzenkörper and/or as a crescent along the apical membrane ( Fig 1B, S1 and S2 Movies). Likewise, GFP-ChsB was observed at hyphal tips and forming septa [24]. However GFP-ChsB was found mainly at the Spitzenkörper at the hyphal tips under this condition (Fig 1C, S3 Movie).
Typical localization patterns of GFP-ChsB or GFP-CsmA in the ΔkinA or ΔuncA are shown in Fig 2A and 2C. In the absence of KinA, both GFP-ChsB and GFP-CsmA showed large accumulations at subapical regions (Fig 2A and 2C). The profiles of GFP signal intensities along hyphae are shown in Fig 2B and 2D. In wild type hyphae, the maximum peaks of GFP-ChsB and GFP-CsmA were restricted sharply at the tip of hyphae. In contrast, the hyphae in kinA-deletion strains displayed wider peaks of GFP-ChsB and GFP-CsmA at approximately 1 μm behind the tip of hyphae. In the uncA-deletion strains, GFP-ChsB and GFP-CsmA localized at the apex of hyphal tips as in wild type stains (Fig 2A and 2C). However GFP-ChsB and GFP-CsmA appeared to be less diffused behind the apex. In addition, multiple punctate structures were observed along the hyphae. In the profiles of GFP signal intensities, several small peaks were often observed throughout the hyphae in the uncA-deletion strains (Fig 2B and 2D).

Role of KinA and UncA in the transportation of ChsB and CsmA
We investigated the roles of kinesins, KinA and UncA, in the transportation of ChsB and CsmA. To visualize the transportation events of GFP-ChsB and GFP-CsmA, kymographs were recorded at hyphal tips in wild type, the kinA-deletion, and the uncA-deletion strains (  (S4-S9 Movies). The signals forming vertical lines in the kymographs represent immobile accumulations of GFP-ChsB ( Fig 3A) and GFP-CsmA (Fig 3B). With the used microscopy setup it is impossible to detect single GFP molecules. The transport events shown in the kymographs are thought to represent events that multiple GFP-ChsB or GFP-CsmA molecules are transported together. It is not defined here whether the spots represent secretory vesicles, chitosomes, or early endosomes.

Transport frequency of ChsB and CsmA
The transport frequency of GFP-ChsB and GFP-CsmA in both anterograde and retrograde direction appeared higher in the wild type than in the kinAor the uncA-deletion strains. To compare the transport frequencies, the transport events were quantified from the kymographs (Fig 4). In addition, the velocity of GFP spots was calculated from the kymographs and compared between wild type and the kinesin-deletion strains (Fig 5).  In wild type the frequency of anterograde and retrograde movements visualized by GFP-ChsB were 6.1 ± 1.8 and 7.5 ± 2.7 (mean ± S.D., n = 82 and 100) events per minute, respectively. In the kinA-deletion strain those (the frequencies of GFP-ChsB anterograde and retrograde movements) were reduced to 2.1 ± 3.1 and 2.3 ± 2.9 (mean ± S.D., n = 38 and 47) events per minute, respectively. In the uncA-deletion strain the frequencies were also reduced to 2.6 ± 1.4 and 3.7 ± 1.9 (mean ± S.D., n = 42 and 58) events per minute, respectively. The frequencies of transports visualized by GFP-ChsB in the kinA-deletion and the uncA-deletion strains were both significantly lower than those in wild type in anterograde (34% in ΔkinA and 46% in ΔuncA) and retrograde direction (31% in ΔkinA and 48% in ΔuncA) (Figs 3 and 4A).
In wild type the frequencies of GFP-CsmA anterograde and retrograde movements were 5.7 ± 1.3 and 5.2 ± 1.6 (mean ± S.D., n = 100 each) events per minute, respectively. In the kinA-deletion strain those (the frequencies of GFP-CsmA anterograde and retrograde movements) were significantly reduced to 0.8 ± 0.8 and 1.2 ± 1.1 (mean ± S.D., n = 40 and 45) events per minute, respectively. In the uncA-deletion strain the frequencies of GFP-CsmA anterograde and retrograde movement were reduced to 2.2 ± 1.8 and 2.1 ± 1.9 (mean ± S.D., n = 40 and 50) events per minute, respectively. Both frequencies of moving spots visualized by GFP-CsmA in the kinA-deletion and the uncA-deletion strains were significantly lower than those in wild type in both anterograde (15% in ΔkinA, 39% in ΔuncA) and retrograde (23% in ΔkinA, 40% in ΔuncA) direction (Figs 3 and 4A). Generally, the frequencies of GFP-ChsB movement were relatively higher than those of GFP-CsmA. The difference might represent different expression levels of GFP-ChsB and GFP-CsmA.
To analyze the distribution of transport events along the hyphae, the hyphae were divided into 5 μm-segments starting from hyphal tips and the frequencies of the movements of GFP-ChsB and GFP-CsmA were quantified in each segment in each strain (Fig 4B-4E). In cases where movements of GFP-ChsB and GFP-CsmA were observed in longer distances than 5 μm, they were counted in multiple segments. The frequencies of GFP-ChsB and GFP-CsmA in both anterograde and retrograde movements in the kinA-deletion strain were lower than those of the wild type strain. Especially in the region close to the tips (0-15 μm), movements of GFP-ChsB and GFP-CsmA were hardly observed in both anterograde and retrograde directions. The frequencies of GFP-ChsB and GFP-CsmA in both anterograde and retrograde direction in the uncA-deletion strain were also lower than those of the wild type strain along the hyphae, in contrast, especially behind the tip region (20-30 μm).
In the wild type strain the velocities of GFP-CsmA anterograde and retrograde movement were 1.6 ± 0.1 and 1.7 ± 0.1 (μm/sec., mean ± S.E., n = 100 each), respectively (Fig 5A). The velocities of GFP-CsmA showed no significant difference compared to the velocities of GFP-ChsB in both anterograde and retrograde movements, respectively. The velocities of anterograde and retrograde movements also showed no significant difference in both GFP-ChsB and GFP-CsmA. In the kinA-deletion strain the velocities of GFP-CsmA anterograde and retrograde movement were 1.5 ± 0.1 and 1.4 ± 0.1 (μm/sec., mean ± S.E., n = 40 and 45), respectively. In the uncA-deletion strain those were 1.6 ± 0.1 and 1.6 ± 0.1 (μm/sec., mean ± S.E., n = 40 and 50), respectively. There were no significant differences in the velocities of anterograde movements in the kinesin deletion strains, whereas the velocities of retrograde movements were slightly reduced in the kinA-deletion strain (81%) but not in the uncA-deletion strain.
The distribution of velocities in the anterograde and retrograde movements of GFP-ChsB and GFP-CsmA was analyzed in each strain (Fig 5B-5E). Velocities faster than 2 μm/sec were observed in wild type at all the anterograde and retrograde movements of GFP-ChsB and GFP-CsmA. However, the high speeds were not clearly found in the kinA-deletion strain. The average speed of early endosomes transport is slower than that of secretory vesicles (unpublished results). The higher speed peaks might indicate those signals derived from secretory vesicles.

Transport of ChsB and CsmA to the hyphal tips along microtubules by KinA
To further analyze the function of KinA in the transport of GFP-ChsB and GFP-CsmA, we constructed strains co-expressing either GFP-ChsB and mRFP-KinA rigor (EGFP-ChsB/ mRFP-KinA rigor strain), or GFP-CsmA and mRFP-KinA rigor (EGFP-CsmA/mRFP-KinA rigor strain). The KinA rigor protein, which has a point mutation in its ATPase domain, binds tightly to microtubules but does not move along them. Consequently, KinA rigor decorates microtubules [48]. GFP-ChsB and GFP-CsmA localized to large accumulations at subapical regions similar to those in kinA-deletion strains. In addition, GFP-ChsB and GFP-CsmA were observed along microtubules decorated with mRFP-KinA rigor throughout the hyphae (Fig 6A and  6B). These results suggest that ChsB and CsmA are transported to the hyphal tip regions along microtubules by KinA.
Only the chitin synthase domain of CsmA without the MMD (GFP-ΔMA) did not show a clear accumulation at the hyphal tip membrane (data not shown) [25]. When the GFP-ΔMA was co-expressed with mRFP-KinA rigor in the EGFP-ΔMA/ mRFP-KinA rigor 1 strain, the GFP-ΔMA also exhibited a similar localization pattern with GFP-CsmA co-expressed with mRFP-KinA rigor , along microtubules and at subapical large accumulations (Fig 6C), suggesting that CsmA is transported to the hyphal tip regions by KinA independently of its MMD.

Interaction of ChsB and CsmA with KinA
To further examine the interaction between ChsB and KinA rigor , or CsmA and KinA rigor in vivo, we constructed strains co-expressing FLAG-ChsB and mRFP-KinA rigor , or CsmA-HA and mRFP-KinA rigor (FLAG-ChsB/mRFP-KinA rigor strain or CsmA-HA/mRFP-KinA rigor strain, respectively). The strains were subjected to a fractionation experiment (see Materials and methods). FLAG-ChsB and CsmA-HA were detected in a 10,000 x g pellet (P10) and in a 100,000 x g pellet (P100) that were obtained by cell fractionation of the crude cell extract of A. nidulans [24,61]. In S. cerevisiae, P100 fraction contains transport vesicles, endosomal membranes, and Golgi cisternae [63]. FLAG-ChsB and CsmA-HA were co-immunoprecipitated with mRFP-KinA rigor from the P100 fractions of the FLAG-ChsB/mRFP-KinA rigor strain and CsmA-HA/mRFP-KinA rigor strain, respectively (Fig 7A and 7C). CsmA-HA was also co-immunoprecipitated with wild type KinA from the P100 fraction of the strain expressing CsmA-HA and GFP-KinA (Fig 7B). These results suggest that transport vesicles containing ChsB or CsmA are transported through the interaction with KinA. In addition, CsmA without the MMD (ΔMA) and CsmA truncated mutant protein lacking 10 amino acids from the actin binding domain of the MMD (D10M) also co-immunoprecipitated with mRFP-KinA rigor (S4 Fig)[25], suggesting that the interaction of transport vesicles containing CsmA with KinA is independent of the MMD.
We examined the interaction between ChsB and CsmA, but we did not detect co-immunoprecipitation of CsmA-HA or FLAG-ChsB from the P100 fractions of the strains that produce these tagged proteins together with or without mRFP-KinA rigor (FLAG-ChsB/CsmA-HA/ mRFP-KinA rigor strain or FLAG-ChsB/CsmA-HA strain, respectively) (data not shown). These results suggest that ChsB and CsmA are transported on different vesicles by KinA.

Discussion
In filamentous fungi, protein machineries required for polarized growth, such as formin and regulators of polarity establishment, accumulate in the Spitzenkörper at the hyphal tips [3,5,7,10,12]. Chitin synthases of the classes I, III, and VII of Neurospora crassa (Chs3, Chs1, and Chs6) also localized at the Spitzenkörper and in subapical regions as punctate structures [55,64]. In the previous studies, we investigated the localization of class III and V chitin synthases in A. nidulans, ChsB and CsmA, respectively [24,25]. In this study, we investigated the transport mechanisms of ChsB and CsmA in living hyphae, especially the roles of kinesin-1 (KinA) and kinesin-3 (UncA) in the transportation of ChsB and CsmA. Our results suggest that these chitin synthases are delivered to hyphal tips along microtubules through the interaction with conventional kinesin (kinesin-1, KinA).
The localizations of GFP-CsmA and GFP-ChsB were disturbed especially around the hyphal tip region by the kinA deletion (Figs 2-5). In addition these chitin synthases co-immunoprecipitated with KinA. These results suggest that KinA transports CsmA and ChsB. Movements of ChsB and CsmA were hardly observed in kinA-deletion mutants and the number of movements with higher velocities were reduced in the mutants. Since it is suggested that smaller vesicles tend to move faster than larger vesicles [65], it is possible that these chitin synthases are mainly transported on small vesicles, such as secretory vesicles or chitosomes, at extreme hyphal tip regions by KinA. However, KinA is required for the transport of cytoplasmic dynein to the microtubule plus ends [52] and dynein itself is involved in membrane transport such as early endosomes in several fungi [11,47,66,67]. Since ChsB and CsmA accumulated at subapical regions of the hyphae in the kinA-deletion mutant and dynein is not transported to those regions in the kinA-deletion mutants [52], it is also possible that the defects of ChsB and CsmA transport in the kinA-deletion strains include the direct and/or indirect effect of mislocalization of dynein.
In U. maydis, Schuster et al. reported that class V chitin synthase Mcs1 was also transported by kinesin-1 [50]. These results indicate that the transport mechanism of class V chitin synthases is conserved between ascomycete and basidiomycete fungi. In contrast, the transport of the class III chitin synthase in N. crassa does not depend strictly on microtubules [64], suggesting that the transport mechanism of class III chitin synthase of A. nidulans is at least partially different from that of another ascomycete N. crassa. Since CsmA-HA and FLAG-ChsB did not co-immunoprecipitate, they are likely transported in different vesicle populations. In N. crassa, it is also suggested that classes I, III, and VII chitin synthases are contained in different vesicular subpopulations [64]. Thus chitin synthases of different classes are possibly transported in different vesicle subpopulations.
Kinesin-3, UncA, is involved in the transport of early endosome and peroxisome [11,48,68]. In U. maydis, kinesin-3 is also involved in the transport of early endosomes and the secretion of acid phosphatase [46,67,69], but is not involved in the transport of Mcs1 [50]. Although the transports of GFP-ChsB and GFP-CsmA were disturbed to some extent in the uncA-deletion strain, the localization at hyphal tips appeared nearly normal in the mutants. This suggests that UncA is involved in the transport of ChsB and CsmA, but not essential especially in hyphal tip regions. In A. nidulans, actin forms a ring structure at the subapical plasma membrane [70]. Endocytosis is suggested to occur nearby this structure, and the resultant endocytic vesicles fuse with each other in the cytoplasm to form early endosomes [7,[70][71][72][73]. The early endosomes are transported along microtubules by kinesin-3 and dynein in the plus and minus end directions, respectively, in hyphae of A. nidulans and U. maydis [6,11,46,48,49] and the endosome system is thought to be involved in the recycling of membrane proteins at hyphal tips [74]. It is possible that kinesin-3, UncA, is involved in the transport of ChsB and CsmA through the recycling pathway. Alternatively the defects of early endosome transport might affect the transport of ChsB and CsmA indirectly. It is necessary to distinguish secretory vesicles and early endosomes in same hyphae in order to investigate the dependency of KinA and UncA on the transport.
It is well known that actin is also crucial for hyphal tip growth and that myosins are motor proteins that move on actin filaments. A. nidulans contains three genes, myoA, myoB, and myoE that encode myosin heavy chains belonging to the myosin families, myosin-1, myosin-2, and myosin-5, respectively [8,9,75]. Moreover, MyoE participates in vesicle transport in hyphae [8]. Mutagenesis of both, myosin-5 and kinesin-1 caused severe growth defects in A.
nidulans and U. maydis [9,69], suggesting that the function of kinesin-1 could be partly substituted by myosin-5 and vice versa. Since myosin-5 functions in the transport of Mcs1 to the cell surface in a microtubule-independent manner [50], it is possible that a myosin-5, MyoE, is also involved in the transport of CsmA and ChsB. This idea is supported by the fact that the inhibition of actin polymerization by cytochalasin A severely disturbed the localization of these chitin synthases (data not shown) [25]. The MMDs of class V chitin synthases belong to the myosin-17 family. The binding of the MMD to actin filaments is required for the proper localization and function of CsmA [25], whereas no motor activity of the MMD of Mcs1 was detected in vitro [50]. The point mutation at the P-loop or Switch I in the MMD of CsmA does not affect its function and localization [25], suggesting that there is no motor activity or at least the motor activity is not essential for the transport mechanism. The idea is supported by our results that CsmA and CsmA without the MMD is transported to the hyphal tips by KinA.
If the MMD of CsmA is not necessary for the transport to hyphal tips, the role of MMD remains unrevealed. Schuster et al. suggested that the MMD of Mcs1 functions in the tethering of vesicles containing Mcs1 near the sites of exocytosis [50]. Our results showed that the deletion of kinesin-1 or kinesin-3 caused similar effects on the transport of ChsB and CsmA. The role of MMD may be important for the localization at hyphal tips after their transportation, rather than the transport towards hyphal tips. GFP-CsmA was found at the Spitzenkörper and/or as a crescent along the apical membrane ( Fig 1B). In contrast, GFP-ChsB was found mainly at the Spitzenkörper at hyphal tips ( Fig 1C). The difference in the localization between CsmA and ChsB was also observed in a significant population of hyphal tips by indirect immunofluorescence analysis [24,25]. ChsB and CsmA are likely to be transported to the Spitzenkörper in a similar way. Then, CsmA might be localized and/ or stay at special regions along the apical membrane depending on the interaction of MMD with the actin cytoskeleton. (ΔkinA), and EGFP-ChsB/ΔuncA1 (ΔuncA) strains were detected using the same antibody. P10, 10,000 x g pellet; P100, 100,000 x g pellet; S100, 100,000 x g supernatant.