p25 of the dynactin complex plays a dual role in cargo binding and dynactin regulation

Cytoplasmic dynein binds its cargoes via the dynactin complex and cargo adapters, and the dynactin pointed-end protein p25 is required for dynein–dynactin binding to the early endosomal dynein adapter HookA (Hook in the fungus Aspergillus nidulans). However, it is unclear whether the HookA–dynein–dynactin interaction requires p27, another pointed-end protein forming heterodimers with p25 within vertebrate dynactin. Here, live-cell imaging and biochemical pulldown experiments revealed that although p27 is a component of the dynactin complex in A. nidulans, it is dispensable for dynein–dynactin to interact with ΔC-HookA (cytosolic HookA lacking its early endosome–binding C terminus) and is not critical for dynein-mediated early endosome transport. Using mutagenesis, imaging, and biochemical approaches, we found that several p25 regions are required for the ΔC-HookA–dynein–dynactin interaction, with the N terminus and loop1 being the most critical regions. Interestingly, p25 was also important for the microtubule (MT) plus-end accumulation of dynactin. This p25 function in dynactin localization also involved p25's N terminus and the loop1 critical for the ΔC-HookA–dynein–dynactin interaction. Given that dynactin's MT plus-end localization does not require HookA and that the kinesin-1–dependent plus-end accumulation of dynactin is unnecessary for the ΔC-HookA–dynein–dynactin interaction, our results indicate that p25 plays a dual role in cargo binding and dynactin regulation. As cargo adapters are implicated in dynein activation via binding to dynactin's pointed end to switch the conformation of p150, a major dynactin component, our results suggest p25 as a critical pointed-end protein involved in this process.

Cytoplasmic dynein binds its cargoes via the dynactin complex and cargo adapters, and the dynactin pointed-end protein p25 is required for dynein-dynactin binding to the early endosomal dynein adapter HookA (Hook in the fungus Aspergillus nidulans). However, it is unclear whether the HookA-dyneindynactin interaction requires p27, another pointed-end protein forming heterodimers with p25 within vertebrate dynactin. Here, live-cell imaging and biochemical pulldown experiments revealed that although p27 is a component of the dynactin complex in A. nidulans, it is dispensable for dynein-dynactin to interact with ⌬C-HookA (cytosolic HookA lacking its early endosome-binding C terminus) and is not critical for dyneinmediated early endosome transport. Using mutagenesis, imaging, and biochemical approaches, we found that several p25 regions are required for the ⌬C-HookAdynein-dynactin interaction, with the N terminus and loop1 being the most critical regions. Interestingly, p25 was also important for the microtubule (MT) plus-end accumulation of dynactin. This p25 function in dynactin localization also involved p25's N terminus and the loop1 critical for the ⌬C-HookAdynein-dynactin interaction. Given that dynactin's MT plus-end localization does not require HookA and that the kinesin-1-dependent plus-end accumulation of dynactin is unnecessary for the ⌬C-HookAdynein-dynactin interaction, our results indicate that p25 plays a dual role in cargo binding and dynactin regulation. As cargo adapters are implicated in dynein activation via binding to dynactin's pointed end to switch the conformation of p150, a major dynactin component, our results suggest p25 as a critical pointed-end protein involved in this process.
Cytoplasmic dynein is a multisubunit minus-end-directed microtubule (MT) 2 motor that transports a variety of cargoes (1), and it is critical for the function of all eukaryotic cells, including neurons that are particularly sensitive to dynein defects (2). Dynein-powered transport requires the dynactin complex, which plays an important role in dynein-cargo interactions and dynein activation (3,4). A dynein-interacting subunit of the dynactin complex is p150 Glued (called p150 for simplicity), which contains an N-terminal MT-binding domain followed by coiled-coil domains CC1 and CC2 (5)(6)(7)(8)(9). The backbone of the dynactin complex is a mini-filament formed by multiple actin-related protein 1 (Arp1) subunits (10 -12). One end of the Arp1 filament is capped by the barbed-end capping proteins (10,13), and the other end binds to the pointed-end complex containing Arp11, p62, p25, and p27 (14). Although Arp11 and p62 are essential for Arp1 filament integrity and dynein function, the heterodimers p25 and p27 are unnecessary for Arp1 filament assembly and are only required for a subset of dynein functions (15)(16)(17)(18)(19). In filamentous fungi, p25 is critical for dynein-mediated transport of early endosomes and post-Golgi vesicles but not for dynein-mediated nuclear distribution (15,17,20). In Aspergillus nidulans, p25 is required for the physical interaction between dynein and early endosome (17), but the function of its binding partner p27 is unclear.
The dynein-early endosome interaction in filamentous fungi requires not only the dynactin complex but also the Hook complex (21)(22)(23). The C terminus of Hook and the two Hookinteracting proteins, FTS and FHIP, link Hook to early endosomes via the FHIP-Rab5 interaction (21)(22)(23)(24)(25). HookA (Hook in A. nidulans) interacts with dynein-dynactin, as revealed by biochemical pulldown assays (21), and this tripartite interaction requires p25 (21). Recently, mammalian Hook proteins and other dynein cargo adapters, including BICD2, have been shown to not only bind dynein-dynactin but also activate dynein's processive motility in vitro (26 -29). Because dynein cargo adapters often contain similar domain arrangements, their mechanisms of dynein-dynactin activation may be similar although not identical (1,26,28,30). In the EM structure of the BICD2-dynein-dynactin supercomplex, BICD2 directly contacts the Arp1 filament as well as its pointed end where p150's CC1A (the first part of CC1) domain also seems to contact when p150 is in its folded or "locked" conformation (3,12). These observations led to an interesting hypothesis that cargo adapters may compete with p150's CC1A domain for binding to the pointed end, thereby causing p150 to assume its open or "unlocked" conformation, which in turn activates dynein (3,29). However, the identity of pointed-end proteins involved in this process remains unknown. Although p25 is involved in the HookA-dynein-dynactin interaction, we have never tested the role of p27, the binding partner of p25, in this interaction. Because p25 and p27 are required for the stability of each other in mammalian cells (18), it has not been clear whether p27 mediates the function of p25.
In this work, we obtained evidence that p27 in A. nidulans is not critical for the HookA-dynein-dynactin interaction. Thus, within the p25-p27 heterodimer, only p25 is critical for the HookA-dynactin-dynein interaction. We identified several regions of p25 important for this interaction, including the two most critical regions, the N terminus and loop1. Importantly, we found that loss of p25 significantly decreases the MT plus-end accumulation of dynactin in live cells, and the two regions of p25 critical for HookA binding, namely the N terminus and the loop1 region, are also critical for dynactin's MT plus-end accumulation. In conjunction with published structural data (12), our results suggest that p25 is a key pointed-end component involved in both binding HookA and regulating dynactin.

p27 is not critical for the HookA-dynein-dynactin interaction
Although p25 is known to be critical for early endosome transport (17), it has been unknown whether p27, the binding partner of p25 (4,14,18,19), is involved in the same process. A previous sequence analysis indicates that both p25 and p27 adopt the left-handed parallel ␤-helix (L␤H) fold (31). The structure of human p27 has more recently been solved at 2.2 Å resolution, and it indeed forms an L␤H domain, followed by a disordered C-terminal segment (19). We identified the p27 protein homolog in A. nidulans as the product of An11815, based on its sequence homology with mammalian p27 and a predicted L␤H domain (Fig. S1). To study the function of p27, we created a deletion mutant, ⌬p27 (Fig. S2). Loss of p27 does not apparently affect colony growth (Fig. 1A). To determine whether loss of p27 affects early endosome distribution, we introduced the ⌬p27 allele into the strain carrying the early endosome marker mCherry-RabA (32,33). In filamentous fungi, cytoplasmic dynein and kinesin-3 drive bidirectional transport of early endosomes along microtubules (34,35). Because microtubule plus-ends face the hyphal tip, dynein drives early endosomes away from the hyphal tip (34, 36 -43). In this study, we compared the phenotypes of ⌬p27 with that of ⌬p25 under the same conditions. Although the ⌬p25 mutant defective in dynein-mediated early endosome transport exhibits a dramatic accumulation of early endosome in 100% of the hyphal tips, only about 38% of the ⌬p27 hyphal tips (n ϭ 203) showed a very mild accumulation of early endosomes (Fig. 1B). Representative kymographs show that whereas early endosomes are mostly For each kymograph, position of the hyphal tip is on the right side and indicated by a short arrow (hyphal tip is on the last kymograph). D, frequency of dynein-mediated retrograde early endosome movements in WT, the ⌬p27 mutant, and the ⌬p25 mutant. The frequency is defined as the number of mCherry-RabA-positive particles moving away from the hyphal tip in a single hypha within a 16-s period (n ϭ 30 hyphae for each strain). The individual values are shown in a scatter plot together with the mean and S.D. values. **** indicates p Ͻ 0.0001 (Kruskal-Wallis test, one-way ANOVA, nonparametric test, unpaired, Prism 7). If p Ͼ 0.05, the difference is considered to be nonsignificant and not specifically indicated by any * (this applies to all figures in this paper). E, Western blots showing ⌬C-HookA-S and dynactin p150 pulled down with GFP-dynein HC in the WT, ⌬p25, and ⌬p27 mutants. The bottom blot shows ⌬C-HookA-S signals in total extract of each strain used for the pulldown experiment. F, quantitative analysis on the ratio of pulled-down dynactin p150 to GFP-dynein HC (dynactin p150/GFP-dynein HC, red, n ϭ 3 individual pulldown experiments for each strain) and that of pulled-down ⌬C-HookA-S to GFP-dynein HC (⌬C-HookA-S/ GFP-dynein HC, blue, n ϭ 4 individual pulldown experiments for each strain). Values of the mutants are relative to the values of the WT controls, which are set at 1. The individual values are shown in a scatter plot together with the mean and S.D. values. * indicates p Ͻ 0.05 (one-way ANOVA, unpaired).

p25 regulates dynactin activity
nonmobile and accumulated at the hyphal tip region in the ⌬p25 mutant, they move bi-directionally in the ⌬p27 mutant and WT cells (Fig. 1C). The overall frequency of dynein-mediated transport of early endosomes away from the hyphal tip is significantly decreased in the ⌬p25 mutant but not in the ⌬p27 mutant (Fig. 1D). Consistent with a very mild effect of p27 loss on dynein-mediated early endosome transport, the interaction between HookA and dynein-dynactin is not significantly affected by the loss of p27. Specifically, we introduced S-tagged ⌬C-HookA (⌬C-HookA-S) (44) and GFP-dynein heavy chain (HC) (36,45,46) into the ⌬p25 and ⌬p27 mutants by genetic crosses and tested whether ⌬C-HookA-S can be pulled down with GFP-dynein HC in the WT, ⌬p25, and ⌬p27 cell extracts (Note that the cytosolic ⌬C-HookA is better than the fulllength HookA for testing the direct HookA-dynein-dynactin interaction in pulldown assays because it does not bind early endosomes (21), thereby excluding the possibility of indirect interactions due to binding to the same early endosome.). The amount of ⌬C-HookA-S pulled down from the ⌬p25 mutant extract was significantly diminished (Fig. 1, E and F), consistent with our previous conclusion that p25 is critical for the HookA-dynein-dynactin interaction (21). By contrast, loss of p27 did not affect significantly the amount of ⌬C-HookA-S pulled down with GFP-dynein HC (Fig. 1, E and F).
To verify that A. nidulans p27 is indeed a dynactin component, we constructed a strain containing the p27-GFP fusion for GFP antibody-based pulldown experiments. In the same experiments, we also included a strain containing the p27-GFP fusion in the ⌬p25 background, a strain containing the p25-GFP fusion, and a strain containing the p25-GFP fusion in the ⌬p27 background. We found that both dynein heavy chain and dynactin p150 are pulled down with either p25-GFP or p27-GFP in WT extracts ( Fig. 2A), and thus, both p25 and p27 are associated with dynactin. Interestingly, loss of p25 (⌬p25) significantly lowered the level of p27-GFP in total extract (Fig. 2, A  and B). More significantly, we could not even detect any dynein-dynactin pulled down with p27-GFP in the ⌬p25 mutant ( Fig. 2A). In contrast, loss of p27 (⌬p27) does not significantly affect the level of p25-GFP (Fig. 2, A and B) or the amounts of dynactin-dynein pulled down with p25-GFP (Fig.  2, A and C). Together, our results suggest that in A. nidulans p25 is required for p27 to bind dynactin, but not vice versa, and thus, p27 is more peripheral than p25 in the A. nidulans dynactin complex (Fig. 2D).

KYH) mutants exhibit a defect in the HookA-dynein-dynactin interaction
To identify the p25 residues required for the HookAdynein-dynactin interaction, we performed a structurefunction study of A. nidulans p25. Based on homology modeling, A. nidulans p25 contains, besides the core L␤H fold, two loop regions, loop1 (L1) and loop2 (L2), a C-terminal ␣-helix (C), and several clusters of conserved amino acids whose side chains face outside of the core structure, including the following: 1) Lys-36 -Tyr-83-His-113; 2) Arg-82-Lys-146 -Glu-147; and 3) Arg-21-Asp-42-Glu-118-Lys-135-Asp-153 (Fig. 3, A and B). We created the p25 ⌬L1 (called ⌬L1 for simplicity) and p25 ⌬C (called ⌬C for simplicity) strains containing deletion of the L1 region or the C-terminal ␣-helix, and also the p25 R97A/R100A/V102A strain containing mutations in L2 (called RRV for simplicity) (Fig. 3, B and C). In addition, we also created the p25 K36A/Y83A/H113A (called KYH for simplicity), p25 R82A/K146A/E147A (called RKE for simplicity) and p25 R21A/D42A/E118A/K135A/D153A (called RDEKD for simplicity) strains, which contain alanine replacements of the residues (Fig. 3, B and C). Finally, we also created the Figure 3. Analysis of A. nidulans p25 structure and colony phenotypes of various p25 mutants. A, sequence alignment of p25. The first four amino acids of the ␤-strands are highlighted in gray (each ␤-strand contains six amino acids with the last two form the turn). Three turns in between strands are colored as yellow, green, and cyan, respectively. The two loops (loop1 and loop2) and the C-terminal ␣-helix (␣-helix) are underlined. The alignment was done using CLUSTALW. Residues that are identical (*), strongly similar (:) or weakly similar (.) are indicated below the sequences. B, model of A. nidulans p25 and positions of the amino acids mutated in this study. The names of the corresponding mutants are labeled in the same color below the names of the regions or the amino acid clusters (each of them is highlighted using the same color as that of the label). The ␤-strands of the L␤H as well as the C-terminal ␣-helix (labeled on top) are highlighted in gray, and the position of loop1 is indicated with an arrow. C, colony phenotypes of a WT control strain containing p25-GFP (wild type) and various p25 mutant strains. p25 regulates dynactin activity p25 K4A/K7G/E9A/Y10G/E12A (called N-GA for simplicity) strain containing glycine and alanine substitutions of five N-terminal residues (Fig. 3, B and C). All the mutant p25 proteins were fused to GFP at the C terminus right before the stop codon, and the mutated DNA fragments replaced the WT p25 gene via homologous recombination (Fig. S3, showing the p25 ⌬L1 -GFP allele as an example). They all form colonies that are healthier than the ⌬p25 mutant (Fig. 3C). Upon microscopic examination of the early endosome marker mCherry-RabA, we found that the RKE, RRV, and RDEKD mutations have no negative effect on early endosome distribution, and thus they were not further studied (Fig. 4A). However, mutants containing the ⌬C, ⌬L1, KYH, and N-GA mutations showed abnormal accumulation of early endosomes at the hyphal tip (Fig. 4, A and B). Among the four mutants, the ⌬L1 and the N-GA mutants showed the most severe phenotype, although it was still less severe than that of the ⌬p25 mutant. Specifically, although almost 100% hyphal tips of the ⌬p25 mutant showed an abnormal accumulation of early endosomes, this was observed in about 75% (n Ͼ100) of the ⌬L1 and N-GA mutant hyphal tips, 60% (n Ͼ200) of the ⌬C mutant hyphal tips, and 40% (n Ͼ200) of the KYH mutant hyphal tips. These mutants exhibited a decrease in the frequency of dynein-mediated movements of early endosomes away from the hyphal tips, which was again most obvious in the ⌬L1 and N-GA mutants (Fig. 4C).
To determine whether the ⌬C, ⌬L1, KYH, and the N-AG mutations affect HookA binding, we introduced the ⌬C-HookA-S allele into these mutants by genetic crosses and tested whether it can be pulled down with p25-GFP containing one of these mutations (Fig. 4D). In the same assays, we also tested whether the amounts of dynactin p150 are normal in the pulled-down material. Although the WT p25-GFP pulleddown ⌬C-HookA-S, the amounts of ⌬C-HookA-S pulled down with ⌬C-p25-GFP, ⌬L1-p25-GFP, KYH-p25-GFP, and N-GA-p25-GFP were significantly decreased (Fig. 4, D and E). However, the amounts of p150 pulled down were not significantly different from normal (Fig. 4, D and E). These results indicate that the defects in binding HookA do not result from the inability of the mutant p25 proteins to bind dynactin. Rather, these mutations specifically affect the ability of p25 to mediate the HookA-dynein-dynactin interaction.

p25 regulates dynactin activity
structures representing the accumulation of the dynactin complex at the MT plus-ends (49 -52). This accumulation depends on the MT-binding domain of p150 as well as kinesin-1 (KinA) (50,52). In the absence of p25, the comets were obviously less bright, although the organization of MTs labeled by CFP-TubA (53) appeared normal (Fig. 5, A-C, and Fig. S4A). Overall, loss of p25 caused the dynactin signals to appear less concentrated at the MT plus-ends but more obvious along MTs ( Fig.  5A and Fig. S4B). This change of the dynactin localization pattern was also obvious in the p25 N-GA and ⌬L1 mutants and less so in the ⌬C or KLH mutant, although dynactin comet intensity was moderately decreased in the ⌬C mutant as indicated by our quantitative analysis (Fig. 5, D-F).
This defect in dynactin localization is not caused by the defect in HookA binding because HookA is unnecessary for the MT plus-end localization of dynactin (Fig. 6A). In fact, the p150 -GFP comet intensity in the ⌬hookA mutant was even a little higher than that in WT (Fig. 6B). Moreover, because p25 is important for the MT plus-end accumulation of dynactin, we were concerned about the possibility that the requirement of p25 for the HookA-dynein-dynactin interaction could be nonspecifically due to a defect in dynactin localization. To rule out this possibility, we tested whether ⌬C-HookA is able to bind dynein-dynactin in the ⌬kinA (kinesin-1) mutant where dynactin localizes along MTs instead of the plus-ends, a pattern similar to that in the ⌬p25 mutant (50,52). We found that the physical interaction between ⌬C-HookA and dynein-dynactin is clearly not defective in the ⌬kinA mutant (Fig. 6, C and D). This result demonstrates that the requirement of p25 for the HookA-dynein-dynactin interaction is specific rather than being due to the requirement of p25 for the MT plus-end accumulation of dynactin.
We next determined whether loss of p25 causes any defect in dynactin-KinA (kinesin-1) interaction using pulldown assays. We found that the interaction is weakened in the ⌬p25 mutant as evidenced by a decrease in the amount of S-tagged KinA (KinA-S) pulled down with p150 -GFP (Fig. 7, A and B). In contrast, the amount of S-tagged KinA (KinA-S) pulled down with p150 -GFP appeared normal in the alcA-nudA HC strain grown on glucose that shuts off the expression of the dynein HC driven by the alcA promoter (Fig. 7, A and B). Thus, although dynein is not needed for the dynactin-KinA interaction, p25 affects dynactin-KinA interaction. In a separate experiment, we also tested whether dynactin is needed for the dynein-KinA interaction by using the alcA-nudK Arp1 strain grown on glucose that shuts off the expression of Arp1. We found that the amount of KinA-S pulled down with GFP-labeled dynein HC is reduced in the absence of Arp1 (Fig. 7, C and D). Together, our results suggest that p25 enhances the dynactin-KinA interaction, and dynactin enhances the dynein-KinA interaction.

Discussion
In this study, we analyzed two proteins at the pointed end of the Arp1 filament, p25 and p27, in the A. nidulans dynactin complex, and we provided new insights into the function of p25 in cargo binding and dynactin regulation. Both p25 and p27 are conserved proteins that form structures containing an L␤H domain (Fig. 3, A and B; Fig. S1) (19). However, although p25 is critical for dynein-mediated early endosome transport (17), p27 is largely dispensable for this process. We also confirmed

p25 regulates dynactin activity
that only p25 is crucial for the interaction between dyneindynactin and ⌬C-HookA (21). Although p27 appears to be more peripheral than p25 within the dynactin complex, it is not needed for this interaction. We have identified several regions of p25 important for the interaction with HookA, and among them, the N terminus and the loop1 region are the most important.
Our result that p27 is not critical for dynein-mediated early endosome movement in A. nidulans comes as a surprise in light of the results in mammalian cells that knocking down p27 with RNAi significantly lowers the level of p25 and affects early endosome distribution (18). The reason behind this difference in the results remains unclear, but we should point out that although the core structures of p25 proteins are conserved evolutionarily, regions that form interacting sites with other pro-teins may differ subtly. For example, part of the loop1 region of A. nidulans p25 rich in amino acids Gln and Ser is not present in mammalian p25 (Fig. 3A), although it is present in Ustilago maydis or Neurospora crassa p25 homolog. In addition, whereas the human p25-p27 interaction is likely mediated by the interaction between the p27 hydrophobic ridge on face B and the p25 hydrophobic ridge on face C of their respective triangular structures (19), the analogous hydrophobic ridges in A. nidulans p25 and p27 differ slightly from that in human p25 and p27. In human p25, the hydrophobic ridge contains six hydrophobic residues from position 2 to 7, with Glu at position 1. Unlike human p25, A. nidulans p25 misses part of the rung (Note that a rung contains 3 ␤ strands and 18 amino acids.) that gives rise to the analogous position 1, and its position 2 has Gln instead of a hydrophobic residue (Fig. 8A). Moreover, the hydrophobic residue at position 6 in human or mouse p27 is replaced by Lys in A. nidulans p27 (Fig. S1), and the N terminus of A. nidulans p27 contains a stretch of 38 amino acids not present in its mammalian homologs.
Despite the subtle difference between the fungal p25/p27 dimer and mammalian p25/p27 dimer, it is most likely that the requirement of p25 in the HookA-dynein-dynactin interaction is evolutionarily conserved. The Hook protein was initially discovered in Drosophila, and three mammalian homologs (Hook1, Hook2, and Hook3) were subsequently identified (54,55). The Hook-dynein interaction was first suggested by a yeast two-hybrid study showing the interaction between the N terminus of Zyg12, the Caenorhabditis elegans Hook homolog, and dynein light intermediate chain (56). The biochemical evidence for the interaction of ⌬C-HookA with both dynein and dynactin was first obtained from pulldown assays using A. nidulans proteins (21). Such an interaction is conserved in mammalian cells (26,28,29). In addition, it was first revealed in A. nidulans by using different mutants that both the dynein and dynactin complexes are required for HookA to bind either dynein or dynactin (21). Such a tripartite complex is similar to the BICD2-dynein-dynactin complex (57), and indeed, recent structural studies have revealed that Hook3 and BICD2 interact with dynein-dynactin in a similar albeit nonidentical fashion (29,30). Hook3 and BICD2 use their N-terminal domains to bind dynein light intermediate chain (29,58,59), and the coiled-coil domains of these cargo adapters locate along the

p25 regulates dynactin activity
Arp1 mini-filament (12,30,60). The N-terminal domains in both the A. nidulans HookA and mammalian Hook3 are necessary for the Hook-dynein-dynactin interaction, and an amino acid inside the Hook N termini involved in this interaction is conserved from A. nidulans to mammalian cells (21,29). Currently, there is no experimental evidence from any higher eukaryotic system to suggest that the Hook-dynein-dynactin interaction requires p25. In structural studies, although BICD2 contacts the pointed end of the Arp1 mini-filament, such a contact was not observed in the case of mammalian Hook3, which could be because the C-terminal part of Hook3 was missing in the structure (30). However, consistent with our data demonstrating the importance of p25 in mediating the HookAdynein-dynactin interaction, a recent modeling study also suggests a role of mammalian p25 in interacting with the three mammalian Hook proteins (Hook1, Hook2, and Hook3) (61). Based on these studies and our current data, we suggest that both fungal and mammalian p25 could contact Hook proteins directly, and/or it is critical for initiating and/or maintaining the Hook-dynein-dynactin interaction.
The p25 protein in A. nidulans is dispensable for fungal nuclear distribution that requires both dynactin and dynein, but it is required for dynein-mediated transport of not only early endosomes but also post-Golgi vesicles whose movement is HookA-independent (17,20). Thus, p25 may be required for dynein-dynactin to bind other cargo adapters, an idea consistent with the result obtained from the modeling study suggesting a role of p25 in interacting with not only the Hook proteins but also the BICD proteins (BICD1, BICD2, and BICDR1) (61). However, the amino acids in p25 required for these interactions may not be conserved. In the modeling study, it was predicted that three basic amino acids of mammalian p25, Lys-74, Lys-75, and Lys-78, are involved in binding to the N-terminal part of BICD2 (61). These three amino acids are located in the loop2 region, and Lys-74 corresponds to Arg-97 in loop2 of A. nidulans p25 (Fig. 3A). However, we have not detected any obvious defect in dynein-mediated early endosome distribution exhibited by the p25 R97A/R100A/V102A (RRV) mutant (Fig. 4A), suggesting that the two basic amino acids, Arg-97 and Arg-100, in loop2 of A. nidulans p25, are not important for HookA binding. Instead, the N terminus, loop1, the C-terminal ␣-helix, and the KYH cluster are important with the N terminus, with loop1 being the most critical for HookA binding.
In this study, we revealed that p25 has an important role in MT plus-end accumulation of dynactin, which is consistent with our previous finding that loss of p25 causes a noticeable reduction in the accumulation of dynein at the MT plus-end (17). The defect in dynactin's plus-end accumulation is partly explained by a decreased interaction between dynactin and KinA (kinesin-1) when p25 is absent. We suggest that p25 could either be directly involved in initiating or maintaining the KinA-dynactin interaction, or it is required for maintaining a proper structure of dynactin for its interaction with KinA. Whether p25 affects the integrity of the Arp1 mini-filament in mammalian cells has been in dispute. In one study, knocking down p25 with RNAi does not significantly disrupt the core dynactin complex as evidenced by its apparently normal sucrose gradient sedimentation pattern (18), and in addition, mammalian p25 is not required for the Arp1-dynamitin interaction (63). However, another study suggests that p25 RNAi causes a reduction of the level of Arp1 (64). Nevertheless, the data from A. nidulans using a deletion mutant of p25 strongly argues against a role of p25 in maintaining the integrity of the Arp1 mini-filament, because the amount of Arp1 pulled down with p150 is normal from lysate lacking p25 (17). More interestingly, the amount of dynein pulled down with p150 is increased upon loss of p25 (17), which also argues against the role of p25 in maintaining the length of the Arp1 mini-filament because cryo-EM structures indicate that the Arp1 mini-filament provides sites for dynein to bind (30,60). In light of these results and the structural analysis suggesting that p150's CC1A and CC1B are close to p25 when p150 is in its folded state (Fig.  8B) (12), we suggest that loss of p25 most likely affects p150 conformation directly.
Importantly, we also found that the N terminus and loop1 of p25 are the most critical for the MT plus-end accumulation of dynactin. In comparison, the C-terminal ␣-helix or the KYH cluster has a mild or unobvious effect on the plus-end accumu- In human p25, the hydrophobic ridge on face C of the triangular structure contains six hydrophobic residues from position 2 to 7 (indicated by black arrows), with Glu at position 1 (19). In comparison, p25 in A. nidulans misses part of the rung (three ␤ strands and 18 amino acids) that gives rise to the analogous position 1, and its position 2 has Gln instead of a hydrophobic residue. B, p25 and p150 proteins are close to each other when p150 is in its folded conformation (12). The published structural data (12) were used for making this figure. The positions of the C terminus (p25(C)) and the KYH (p25(KYH)) cluster (indicated by black arrows) are close to p150's CC1B, whereas that of loop1 (p25(loop1)) (indicated by a black arrow) is close to p150's CC1A. The N terminus was not visible in the structure (12). Other pointed-end proteins such as Arp11, p62, and p27 are indicated by labels using colors corresponding to those in the structure figure. p25 regulates dynactin activity lation of dynactin. Thus, although HookA is unnecessary for dynactin's MT plus-end accumulation and the accumulation is unnecessary for dynein-dynactin to bind the cytosolic ⌬C-HookA (Fig. 6), both of these events independently require the N terminus and loop1 of p25. According to the structural data (Fig. 8B) (12), the C-terminal ␣-helix and the KYH cluster are close to p150's CC1B, although loop1 is close to CC1A. The N terminus of p25 was not clearly recognized from the structure but is likely to be close to CC1A just like loop1. It is possible that the N-GA or the ⌬loop1 mutation weakens the p25-CC1A interaction to unfold p150 abnormally, whereas the ⌬C or KYH mutation affects p150 conformation less significantly with the p25-CC1A interaction keeping p150 folded. In light of the idea that a cargo adapter competes against p150's CC1A for binding to the pointed end of the Arp1 mini-filament where it switches the folded p150 to its unfolded state (3,29), our current data suggest that p25 could be the key protein in this process. It is likely that the N terminus and loop1 of p25 bind directly to HookA as well as to CC1A of p150. In comparison, the C terminus or the KYH cluster of p25 may represent an addi-tional but less critical HookA-binding site or indirectly facilitate HookA binding by subtly modulating the p150 -p25 interaction.

A. nidulans strains and media
A. nidulans strains used in this study are listed in Table 1. For biochemical experiments, YG (yeast extract plus glucose) ϩ UU (or YUU) liquid medium was used. For live cell imaging experiments, liquid minimal medium containing 1% glycerol or 0.1% glucose plus supplements was used, and cells were cultured at 32°C overnight and observed at room temperature. Colonies were grown on YAG (YG ϩ agar) plates for 2 days at 37°C.

Live cell imaging and analyses
Fluorescence microscopy of live A. nidulans hyphae was as described (21). All images were captured at room temperature using an Olympus IX70 inverted fluorescence microscope linked to a Sensicam QE cooled CCD camera (PCO/Cooke Corp.). An UplanApo ϫ100 objective lens (oil) with a 1.35  p150-GFP-AfpyrG; p25 ⌬L1 -GFP-AfpyrG; alcA-CFP-tubA-pyr4; pyrG89; possibly ⌬nkuA::argB This work p25 regulates dynactin activity numerical aperture was used. A filter wheel system with GFP/ mCherry-ET Sputtered series with high transmission (Biovision Technologies) was used. Chroma 8600 filters for CFP were used for observing CFP signals as described previously (53). The IPLab software (Biovision Technologies) was used for image acquisition and analysis, including the measurement of comet intensity as described previously (17).

Construction of the ⌬p27 and p27-GFP strains
For constructing the ⌬p27 mutant, we obtained the deletion cassette for the An11815 gene provided by the Fungal Genetics Stock Center (FGSC) (65), which contains the A. fumigatus pyrG gene (AfpyrG) in the middle of the flanking sequences of both sides. The primers used for amplifying the flanking sequences were 5f (5Ј-GTAACGCCAGGGTTTTCCCAGTC-ACGACGGGAGTCGTTGGAGATCATAG-3Ј), 5r (5Ј-ATC-CACTTAACGTTACTGAAATCATAGGGGCGGTTTAAT-GTCC-3Ј), 3f (5Ј-CTCCTTCAATATCATCTTCTGTCCTT-TCTTTCGTCCCTACCAC-3Ј), and 3r (5Ј-GCGGATAACA-ATTTCACACAGGAAACAGCCCCTCTACTCCAGCATT-TAC-3Ј). The cassette was transformed into the TNO2A3 strain carrying the ⌬nkuA marker that greatly facilitates the selection of homologous recombination events (66). A strain containing the ⌬p27 allele was selected by PCR using two primers, 27N3 (5Ј-GAGTACTAAAACTGTGCG-3Ј) and 27C (5Ј-TTACCACCTGGCACCTA-3Ј), which are outside of the cassette. Site-specific integration of the fragment that replaces the WT p27 gene should generate a PCR product different in size from that obtained from a WT control, and this was what we found (Fig. S2). We further verified the ⌬p27 allele by sequencing both PCR products from WT and from the ⌬p27 mutant, which confirmed that the AfpyrG gene indeed replaced the An11815 gene encoding p27.

Homology modeling of A. nidulans p25
When we started this work in 2011, the p27 structure was not available (19). We first used A. nidulans p25 sequence to do a protein blast search against Protein Data Bank (PDB) protein database. The left-handed ␤-helix protein PDB 1v3w, a carbonic anhydrase from the archaeon Pyrococcus horikoshii, showed the highest sequence similarity with A. nidulans p25, and thus, this protein was selected as a template for p25 structure modeling using PHYRE2 (http://www.sbg.bio.ic.ac.uk/ phyre2/html/page.cgi?idϭindex) 3 (70). This study led to the predication that A. nidulans p25 contains, besides the core structure of the left-handed ␤-helix folds, two loop regions and a C-terminal ␣-helix. We then identified several clusters of amino acids whose side chains face outside of the core structure.
For all these p25 mutants created by transformation, correct integration of the constructs into the p25 locus was verified by PCR using primers Q1 and GFP-5R (5Ј-CCAGTGAAAA-GTTCTTCTCCTTTAC-3Ј) (Note that Q1 is at the p25 locus but not within the transformed fragment.) (Fig. S3 shows one example). All of the PCR products were sequenced to further verify the mutations.

Analyses of protein-protein interactions and Western blot analysis
The MACS GFP-tagged protein isolation kit (Miltenyi Biotec) was used to determine whether GFP-tagged p25 pulls down S-tagged ⌬C-HookA. This was done as described previously (21). Strains were grown overnight in liquid rich medium YG. Hyphal mass was harvested from overnight culture by filtering using Miracloth, and liquid was further removed by using paper towels. For each sample, 0.4 g of the hyphal mass was ground and mixed with 1.2 ml of cell lysis buffer containing 50 mM Tris-HCl, pH 8.0, and 10 g/ml of a protease inhibitor mixture (Sigma). Cell extract was centrifuged at 8,000 ϫ g for 20 min and then 16,000 ϫ g for 10 min at 4°C, and 1 ml of supernatant was used for the pulldown experiment. To pull down GFP-p25 regulates dynactin activity tagged p25 or p27 protein, 25 l of anti-GFP MicroBeads was added into 1 ml of cell extract for each sample. To pull down GFP-tagged p150 or dynein HC, 40 l of anti-GFP MicroBeads was added into 1 ml of cell extract for each sample. The microbeads/cell extracts mixture was rotated at 80 rpm for 30 min at 4°C and then applied to the Column, followed by five times gentle wash with the lysis buffer. 50 l of pre-heated (95°C) SDS-PAGE sample buffer was used to elute the proteins, and 38 l of the eluate was loaded onto each well of an SDSpolyacrylamide gel for Western analyses. Antibodies against GFP and S-tag used for Western analyses were from Clontech (polyclonal) and Cell Signaling Technology, respectively. The antibodies against dynein HC and dynactin p150 were described previously (16,71). Western analyses were performed using the alkaline phosphatase system, and blots were developed using the alkaline phosphatase color development reagents from Bio-Rad. Quantitation of the protein band intensity was done using the IPLab software as described previously (52,62).

Statistical analysis
Unpaired Student's t test was used for analyzing two data sets, and one-way ANOVA was used for analyzing multiple data sets (Prism 7 for Mac OS X, version 7.0c, 2017). For analyzing the frequency of dynein-mediated early-endosome movement, the Kruskal-Wallis test (one-way ANOVA, nonparametric test, unpaired) was performed following the recommendations of Prism 7 without assuming Gaussian distribution. All the p values are two-tailed.