Lis1 is an initiation factor for dynein-driven organelle transport

Bidirectional transport of proteins, organelles, and mRNAs along cytosolic microtubules is essential for many aspects of eukaryotic cell biology such as cell growth, cell migration, and cell–cell communication. In humans, subtle defects in the motors responsible for these movements cause neurodevelopmental and neurodegenerative diseases, highlighting the importance of microtubule-based transport in long cells such as neurons (Hirokawa et al., 2010). Most kinesin motors move toward the microtubule plus end or toward the cell periphery (anterograde motion), whereas the dynein motor moves toward the microtubule minus end or toward the nucleus (retrograde motion). Interestingly, in humans and many other eukaryotes, there are multiple cargo-transporting kinesins (15 in humans) but only a single dynein motor gene responsible for most cytoplasmic cargo transport (Vale, 2003). Thus, a critical challenge for understanding dynein-based motility is to understand how dynein is regulated to transport diverse cargos.

The dynein motor holoenzyme is large and complex (at least eight polypeptides weighing >1.2 megadaltons), and many additional proteins and protein complexes, including the dynactin complex, Lis1, and Nudel, act as regulatory factors (Kardon and Vale, 2009; Vallee et al., 2012). Lis1 was first identified as a gene mutated in type I lissencephaly, a brain developmental disease characterized by malformations of the cerebral cortex (Reiner et al., 1993). Lis1 was later linked to dynein motor function through genetic studies in Aspergillus nidulans, which demonstrated that Lis1/nudF, like dynein/nudA, was required for normal nuclear distribution (Xiang et al., 1995).

Lis1’s role in the dynein-mediated positioning of nuclei and centrosomes is well documented (Xiang et al., 1995; Faulkner et al., 2000; Dujardin et al., 2003; Lee et al., 2003; Cockell et al., 2004; Tanaka et al., 2004; Levy and Holzbaur, 2006; Vallee and Tsai, 2006; Tsai et al., 2007; Youn et al., 2009), but its role in the transport of other dynein cargos remains less clear. Several studies have shown that reducing Lis1 expression leads to defects in the distribution or transport of endosomes, Golgi, and lysosomes (Liu et al., 2000; Smith et al., 2000; Lenz et al., 2006; Ding et al., 2009; Lam et al., 2010; Zhang et al., 2010; Yi et al., 2011). In contrast, overexpression of Lis1 in mammalian tissue culture cells does not appear to affect organelle transport while affecting other dynein-mediated processes, leading to the conclusion that Lis1 is not required for the transport of some dynein cargos (Faulkner et al., 2000; Dujardin et al., 2003; Tsai et al., 2007). The role of Lis1 in organelle transport may also depend on the size of the organelle as well as the cell type being studied (Pandey and Smith, 2011; Yi et al., 2011). Thus, the breadth of dynein activities in which Lis1 participates is controversial.

How Lis1 modulates dynein function is also unclear. Recently, it was proposed that Lis1 might be specifically required for dynein functions that require pulling against high loads (McKenney et al., 2010). The finding in cultured primary neurons that large cargos were more reliant on Lis1 than smaller cargos provided additional support for this idea (Pandey and Smith, 2011; Yi et al., 2011). Although several studies have shown that Lis1 can alter dynein’s mechanochemical functions (Mesngon et al., 2006; Yamada et al., 2008; McKenney et al., 2010; Torisawa et al., 2011), it remains unknown whether Lis1 associates with dynein on moving cargo in vivo. In fact, in Saccharomyces cerevisiae, one of the most well-studied model systems for the role of Lis1 in the dynein pathway, Lis1 does not appear to interact with cargo-bound dynein (Lee et al., 2003; Markus et al., 2011). In yeast, Lis1/Pac1 is required for targeting dynein to the microtubule plus end, where it is proposed to facilitate the off-loading of dynein onto the cortex (Lee et al., 2003; Sheeman et al., 2003). Cortically positioned dynein is then thought to pull on nucleus-attached astral microtubules, biasing the movement of the daughter nucleus through the bud neck (Moore et al., 2009). Thus, the absence of cortical Lis1, but the clear presence of dynein and dynactin from cytological studies (Markus and Lee, 2011; Markus et al., 2011), suggests that Lis1 has a role in targeting and/or initiating dynein-based transport in S. cerevisiae but may not be required for subsequent motility.

Some of the controversy over Lis1’s function comes from the variety of methods used to alter Lis1 gene expression or protein activity. Here, we have turned to the model filamentous fungus A. nidulans, which has several major advantages. First, unlike S. cerevisiae, which lacks cargo-transporting kinesins (defined here as kinesin-1, -2, and -3) and uses dynein only for nuclear positioning, A. nidulans uses microtubule-based transport to move components required for rapid growth of its long hyphae (Horio and Oakley, 2005; Horio, 2007). Microtubule-based motors in the A. nidulans genome are encoded by a single dynein gene (nudA) and 11 kinesin genes representing nine classes of kinesin motors (Rischitor et al., 2004). Three of these kinesins, KinA (kinesin-1 or conventional kinesin), UncA (a kinesin-3), and UncB (also a kinesin-3), are expected to be cytoplasmic cargo-transporting kinesins (Requena et al., 2001; Zekert and Fischer, 2009; Verhey et al., 2011). Here, we will refer to the dynein heavy chain gene, nudA, as dynein, kinA as kinesin-1, and uncA and uncB as kinesin-3s. Second, the subunit compositions of dynein and dynactin in A. nidulans are similar to their mammalian counterparts, with known homologs of almost every subunit present in the A. nidulans genome. Third, unlike mammalian tissue culture models, genes can easily be knocked out or modified at the endogenous loci in A. nidulans haploid spores (Nayak et al., 2006). Finally, given the ease of genetic manipulations in A. nidulans and recent advances in whole genome sequencing, in the future, A. nidulans will be an ideal model for performing screens to identify components required for microtubule-based transport, which has not been possible in S. cerevisiae because yeast do not use microtubules for long-distance organelle transport.

Here, we set out to determine whether Lis1 is required for the motility or distribution of multiple dynein cargos and whether Lis1 regulates moving cargo-bound dynein. We begin by confirming and building on previous work showing that A. nidulans is a robust model system for performing a comprehensive analysis of multiple microtubule-based cargos (Abenza et al., 2009; Zekert and Fischer, 2009; Zhang et al., 2010, 2011). We then show that Lis1 is required for the transport of three functionally and morphologically distinct cargos, raising the possibility that Lis1 is required for the motility of most dynein cargos. We find that although the frequency of cargo transport is decreased in the genetic absence of Lis1, the velocities of both cargo and dynein are largely unaffected, providing evidence that Lis1 is not required during motility in vivo. In further support of this, we show that Lis1 is not stably associated with moving, dynein-driven cargo, whereas both dynein and dynactin are. Based on these findings, we propose that Lis1 has a general role in the initiation of dynein-based cargo motility.

The study of microtubule-based transport has a long history in A. nidulans (Osmani and Mirabito, 2004), but recent advances now permit tagging or deleting motors and motor regulatory subunits with ease at their endogenous genomic loci (Nayak et al., 2006). Using these tools in combination with high-resolution microscopy, here, we systematically analyzed the regulation of microtubule-based bidirectional transport in A. nidulans and investigated the role of the dynein regulator Lis1 in the transport of multiple cargos.

In A. nidulans, hyphal compartments contain multiple nuclei in a linear arrangement (Fig. 1 A). To examine the polarity of microtubules in hyphae, we generated a strain in which α-tubulin was tagged with GFP (Szewczyk and Oakley, 2011) and the microtubule plus-end–tracking protein EB1/AN2862 was tagged with mCherry (Fig. 1 B). We found that 98.7 ± 0.5% (SEM) of microtubules between the hyphal tip and the most proximal nucleus were polarized with their plus ends oriented toward the growing hyphal tip (Fig. 1 C), consistent with a previous study using the plus-end marker CENP-E/KipA (Konzack et al., 2005). To determine whether microtubules remain polarized in the absence of cargo-transporting motors, we deleted dynein, kinesin-1, kinesin-3/uncA, or kinesin-3/uncB from the GFP–α-tubulin, EB1-mCherry–tagged strain. We found that the vast majority of microtubules displayed normal polarity in all motor deletion strains (Fig. 1 C). Thus, the unidirectional array of microtubules within the hyphal tip makes A. nidulans an ideal model system for studying bidirectional microtubule-based transport in a highly polarized cell type.

Next, we confirmed and extended previous studies by identifying organelles that rely on microtubule-based motors for movement and distribution in A. nidulans. Previous studies identified endosomal (Abenza et al., 2009) and nuclear (Xiang et al., 1994) distribution along hyphae as dependent on microtubule-based motors. Here, we visualized and also quantified endosome motility in the hyphal tip, using an N-terminal GFP tag on Rab5/RabA to specifically label early endosomes (Abenza et al., 2009). Endosomes moved bidirectionally along the cell length (Video 1), with comparable velocities in the anterograde and retrograde directions (2.62 ± 0.83 µm/s [mean ± SD] vs. 2.66 ± 0.77 µm/s, respectively; Fig. 2 A).

To identify additional organelles dependent on microtubules for motility, we next characterized the motile behavior of peroxisomes by labeling them with a GFP tag on the peroxisomal membrane protein Pex11/PexK (Hynes et al., 2008). We found that peroxisomes, like endosomes, displayed both anterograde and retrograde movements (Video 2). Peroxisomes were significantly larger than endosomes (Fig. 2 B) and traveled with mean anterograde and retrograde velocities of 1.66 ± 0.77 and 1.63 ± 0.87 µm/s (mean ± SD; Fig. 2 A), respectively. Thus, for a given organelle, anterograde and retrograde speeds of transport are similar, but endosomes are significantly smaller in size and move more rapidly than peroxisomes (P < 0.0001, using unpaired t-tests).

To identify the motors required for the distribution and transport of these dynamic organelles, we individually deleted the genes encoding each cargo-transporting, microtubule-based motor present in the A. nidulans genome in strains expressing labeled endosomes, peroxisomes, and nuclei. Consistent with prior studies, we found that nuclei were not distributed properly in strains lacking either dynein or kinesin-1 (Fig. 2, C and D; Xiang et al., 1994; Zhang et al., 2003). Loss of kinesin-1 likely perturbs nuclear positioning as a result of mislocalization of dynein (Zhang et al., 2003); here, we confirmed that most dynein plus-end localization is lost in the absence of kinesin-1 (Fig. S1). Unlike kinesin-1, neither kinesin-3 gene (uncA or uncB) was required for nuclear positioning or dynein plus-end localization (Figs. 2 [C and D], S1, and S2 [A and B]). Thus, we have confirmed and broadened previous studies by showing that of all cargo-transporting kinesins, only kinesin-1 is required for nuclear positioning and dynein plus-end localization.

Next, we tested which microtubule-based motors were responsible for endosome and peroxisome transport. In strains lacking the dynein heavy chain gene, both endosomes and peroxisomes accumulated in the hyphal tip and were nonmotile, suggesting that dynein is responsible for the retrograde movement of these organelles (Fig. 2 [C and D] and Video 3; Abenza et al., 2009; Zhang et al., 2010). In strains lacking kinesin-1, similar to those lacking dynein, both peroxisomes and endosomes accumulated in the hyphal tip (Fig. 2, C and D), likely reflecting kinesin-1’s role in localizing dynein to the microtubule plus end (Fig. S1; Zhang et al., 2003).

Interestingly, although not required for nuclear distribution, deletion of kinesin-3/uncA caused defects in endosome and peroxisome distribution and motility (Fig. 2, C and D). In kinesin-3/uncAΔ hyphae, endosomes accumulated in small patches near nuclei (Fig. 2 C), whereas peroxisomes accumulated aberrantly in the tip region (Fig. 2 D). As was observed in dyneinΔ hyphae, endosomes and peroxisomes were immotile in kinesin-3/uncAΔ hyphae (Video 4). The only other cargo-transporting kinesin present in the A. nidulans genome, kinesin-3/uncB, was not required for the transport of either peroxisomes or endosomes (Fig. S2, A and B). Thus, for both endosomes and peroxisomes in A. nidulans, dynein and kinesin-3/UncA are the opposite polarity motors responsible for bidirectional transport.

Our observation that deletion of either dynein or kinesin-3/uncA perturbed organelle transport in both the retrograde and anterograde directions suggests some type of interdependence between plus- and minus-end–directed motors present on a shared cargo. Interdependence of bidirectional microtubule-based transport has been observed in many eukaryotic organisms (recent examples include Gross et al. [2003], Lenz et al. [2006], Barkus et al. [2008], Ally et al. [2009], Uchida et al. [2009], and Encalada et al. [2011]). To examine this further, we C-terminally tagged kinesin-3/UncA with GFP and examined its motile behavior in wild-type and dyneinΔ hyphae (Fig. S3). In wild-type hyphae, kinesin-3/UncA-GFP particles moved bidirectionally in a manner reminiscent of endosome movements (Fig. S3 A). However, in strains lacking dynein, although we detected rare anterograde-directed movements (Fig. S3 C), kinesin-3/UncA-GFP predominantly accumulated at the hyphal tip, consistent with attachment to organellar cargo (Fig. S3, B and C). Our findings suggest that in the absence of dynein, kinesin-3/UncA–mediated anterograde transport is perturbed as a result of sequestration of kinesin-3/UncA at the microtubule plus end. We propose that kinesin-3/UncA may require dynein to recycle back to the microtubule minus end.

The lack of endosome and peroxisome motility in strains lacking dynein suggests that dynein directly transports these organelles. If this were the case, dynein should colocalize with retrograde-moving endosomes and peroxisomes. To determine the localizations of dynein and its putative cargo, we tagged the endogenous dynein heavy chain with three tandem copies of GFP (3xGFP) at its C terminus. Importantly, addition of the 3xGFP tag did not disrupt dynein function, as the strain grew similarly to wild-type strains and did not share the slow growth phenotype exhibited by nud (nuclear distribution) mutants (Fig. S4, A and B). Endosome and peroxisome cargo were visualized with mCherry-Rab5/RabA and mCherry-PTS1 (peroxisome targeting signal 1), respectively. Analysis of kymographs made from time-lapse videos of these dynein and organelle-colabeled strains demonstrated the presence of discrete dynein runs that colocalized with retrograde-moving endosomes (Fig. 3 A) or peroxisomes (Fig. 3 B). Interestingly, we also observed dynein colocalization with endosomes moving toward the microtubule plus end (Fig. 3 A), suggesting that dynein is a passenger on kinesin-3/UncA–driven, anterograde-moving endosomes. These data suggest that dynein directly drives the retrograde motility of endosomes and peroxisomes within the hyphal tip.

In summary, these experiments along with other work in the field show that nuclear distribution requires dynein and kinesin-1 (which localizes dynein) but not kinesin-3s, whereas endosome and peroxisome transport requires dynein, kinesin-1 (to localize dynein), and kinesin-3/UncA but not kinesin-3/UncB. Next, we turned our focus to the role of the dynein regulator Lis1 in the distribution and transport of nuclei, endosomes, and peroxisomes.

In Aspergillus, Lis1/nudF, referred to here as Lis1, is required for the migration of nuclei along rapidly growing hyphae (Xiang et al., 1995) and has more recently been implicated in the transport of Rab5/RabA-marked early endosomes (Zhang et al., 2010). Here, we sought to confirm and further characterize Lis1’s role in endosome transport and additionally to determine whether Lis1 was required for the dynein-mediated trafficking of peroxisomes. We deleted the sole endogenous copy of Lis1 from a strain with GFP-Rab5/RabA–marked endosomes and mCherry-PTS1–marked peroxisomes. Lis1Δ hyphae displayed severe defects in both endosome (Fig. 4 A) and peroxisome (Fig. 4 B) distributions. Both classes of organelle clustered at the hyphal tip, similar to defects observed in dyneinΔ hyphae (Fig. 2, A and B). Kymographs from time-lapse videos of labeled endosomes and peroxisomes indicated that these organelles were largely immotile (Fig. 4, A and B), with defects in both anterograde and retrograde motility. Thus, Lis1 is required for the proper distribution of three morphologically and functionally distinct dynein-dependent cargos: nuclei, endosomes, and peroxisomes.

Although the majority of endosomes were immotile in Lis1Δ hyphae (Fig. 4 A), we did observe occasional bidirectional movements of endosomes (Fig. 4 C). To further examine this, we calculated the velocity of bidirectional endosome movements in the absence of Lis1. To study the process with high resolution and detail, we chose to focus on endosomes, which are significantly more dynamic than peroxisomes. Strikingly, we found that there was very little change in the velocities of anterograde (Fig. 4 D) and retrograde (Fig. 4 E) endosome movements in the absence of Lis1. The presence of occasional organelles that move at normal speeds in the complete genetic absence of Lis1 raises the possibility that Lis1 is not essential for dynein-driven endosome movements.

Our finding that endosomal cargo in Lis1Δ hyphae can move at relatively normal speeds suggested to us that dynein motility itself may not require Lis1 in vivo. To test this directly, we monitored dynein movement in Lis1Δ hyphae. Significantly, dynein-3xGFP was detected moving both toward and away from microtubule plus ends in Lis1Δ hyphae (Fig. 5 A and Video 5). Notably, the mean dynein particle velocity in either the anterograde (Fig. 5 B) or retrograde (Fig. 5 C) direction was not affected by deleting Lis1. Thus, Lis1 is not essential for the movement of either dynein or its cargo in vivo but is still necessary for proper cargo distribution along hyphae.

In the absence of Lis1, endosomes and peroxisomes accumulate at hyphal tips, and retrograde movements are quite rare despite dynein’s relatively normal capacity for motility, raising the possibility that Lis1 is required for initiating dynein-driven cargo transport. To test this, we first calculated the frequency of endosome movements by determining the number of anterograde and retrograde events that crossed a defined region within the hyphal tip during a fixed time interval. In wild-type hyphae, the frequencies of anterograde and retrograde endosome movements were 42 ± 5 per minute (SEM) and 59 ± 7 per minute, respectively (Fig. 6 A), whereas in Lis1Δ hyphae, the frequencies decreased dramatically to 3 ± 1 per minute (anterograde) and 2 ± 1 per minute (retrograde; Fig. 6 A). In contrast, we observed no bidirectional movements in hyphae lacking either dynein (Video 3) or the dynactin subunit p150/nudM (Fig. S5 A).

We also analyzed the frequency of dynein movements by counting the number of particles that arrive at or depart from the microtubule plus end over a given time period. The frequency of dynein particle anterograde movements in Lis1Δ hyphae was not significantly different from those in the wild-type strain (Fig. 6 B). However, the frequency of retrograde dynein particle movements in Lis1Δ hyphae was significantly lower than that of the wild-type strain (Fig. 6 B). Interestingly, in contrast to Lis1, deletion of the dynactin subunit p150/nudM completely abolished all dynein movement (Fig. S5 B). Our findings that the genetic absence of Lis1 does not markedly alter dynein and endosome velocities but does reduce the frequency of their movements are consistent with the hypothesis that Lis1 is necessary for initiating dynein-driven cargo transport but not for subsequent dynein motor function to transport endosomes away from the hyphal tip. In contrast, dynactin appears to be essential for the motility of both dynein and its cargo in vivo.

If the primary role of Lis1 in dynein-dependent transport were during initiation of this process, Lis1 may not remain stably associated with cargo after transport commences. To test this hypothesis, we generated a strain expressing both Lis1-GFP and mCherry-Rab5/RabA–labeled endosomes. Importantly, the addition of GFP to the C terminus of Lis1 did not disrupt Lis1 function, as this strain grew at wild-type rates (Fig. S4, A and C). Lis1-GFP localized to the plus ends of microtubules and was largely immotile (Fig. 7 A and Video 6). Although we did observe occasional retrograde-moving Lis1-GFP particles, their movements were very short in length and only very rarely colocalized with a moving endosome. To quantify this observation, we determined the percentage of moving endosomes that colocalized with either dynein or Lis1 and found that 93.4 ± 3.0% (mean ± SEM; n = 110) of moving endosomes colocalized with a dynein-3xGFP particle, whereas only 1.8 ± 1.0% (n = 148) of endosomes colocalized with a Lis1-GFP particle (Fig. 7 B). To ensure that the absence of Lis1 on moving endosomes was not a result of imaging limitations, we tagged the dynactin subunit p25 with a single GFP at its C terminus and generated a strain expressing both p25-GFP and mCherry-Rab5/RabA–labeled endosomes (Fig. 7 C). Unlike Lis1-GFP, p25-GFP was highly motile and associated with 56.4 ± 5.9% of moving endosomes (mean ± SEM; n = 161; Fig. 7 B). In these experiments, because each dynein dimer contains six GFPs, each Lis1 dimer contains two GFPs, and each dynactin complex has only a single GFP on its p25 subunit (Eckley et al., 1999), we have likely underestimated the number of endosomes that colocalize with p25 and Lis1. In summary, these results show that the majority of retrograde endosome movements are associated with dynein and dynactin but not Lis1.

We hypothesized that Lis1 could be required during transport initiation for proper loading of dynein onto its cargo. In Lis1Δ hyphae, but not wild-type hyphae, endosomes and peroxisomes accumulate aberrantly at hyphal tips (Fig. 4, A and B). If Lis1 were dispensable for loading dynein onto cargo, we expected that dynein would accumulate on these mislocalized cargos, as previously proposed (Zhang et al., 2010). To test this idea, we visualized the extent of colocalization between dynein-3xGFP and endosomes (Fig. 8, A and B) or peroxisomes (Fig. 8, C and D) in wild-type and Lis1Δ hyphae. Although we observed increased dynein-3xGFP signal in Lis1Δ hyphal tips, this signal did not colocalize with either endosomes (Fig. 8 B) or peroxisomes (Fig. 8 D) in Lis1Δ hyphae, supporting the idea that Lis1 is required for loading dynein onto its cellular cargos at the microtubule plus end.

Here, we have undertaken a comprehensive study of the role of Lis1 in dynein-based organelle transport. We analyzed the dynamics of both dynein and its cargos in the polarized haploid hyphae of A. nidulans strains lacking the single endogenous copy of the Lis1 gene. We found that Lis1 is required for the proper distribution of three functionally and morphologically distinct dynein cargos, including both small cargo (endosomes) and larger cargo (peroxisomes and nuclei). To gain insight into which stages of the dynein pathway Lis1 acts, we analyzed the frequency and velocity of endosome and dynein motions in the absence of Lis1. Surprisingly, we found that both endosomes and dynein moved at relatively normal speeds in the complete absence of Lis1 but with reduced frequency. Moreover, Lis1, unlike dynein and dynactin, is absent from nearly all moving dynein cargos, further supporting our finding that Lis1 is not strictly required for dynein-based cargo motility once movement has commenced. Based on these observations, we propose that Lis1 has a general role in the initiation of dynein-driven motility.

Despite a large number of studies in diverse biological systems, the mechanistic role of Lis1 in the transport of dynein cargos remains controversial. To clearly interpret the function of Lis1 in the transport of dynein cargos, we genetically deleted the sole copy of the endogenous Lis1 gene in A. nidulans. Many prior studies in mammalian cells or the fungus Ustilago maydis have relied on methods that lead to incomplete depletion of either the Lis1 mRNA or protein (Faulkner et al., 2000; Tai et al., 2002; Dujardin et al., 2003; Lenz et al., 2006; Ding et al., 2009; Lam et al., 2010; Pandey and Smith, 2011; Yi et al., 2011). Our finding that Lis1 is required for the proper distribution of endosomes, peroxisomes, and nuclei leads us to propose that most, if not all, dynein cargo will require Lis1 function. Consistent with our observations, decreases in Lis1 expression have been shown to cause defects in axonal transport (Liu et al., 2000) and distribution of endosomes, lysosomes, and Golgi (Liu et al., 2000; Lenz et al., 2006; Ding et al., 2009; Lam et al., 2010; Pandey and Smith, 2011; Yi et al., 2011), as well as nuclear migration and centrosome positioning (Xiang et al., 1995; Faulkner et al., 2000; Dujardin et al., 2003; Lee et al., 2003; Sheeman et al., 2003; Cockell et al., 2004; Tanaka et al., 2004; Tsai et al., 2007; Levy and Holzbaur, 2008; Youn et al., 2009).

Only a handful of studies have reported Lis1-independent dynein-based processes. First, experiments in which Lis1 or Lis1 fragments were overexpressed in mammalian tissue culture cells did not reveal any defects in the transport of lysosomes or Golgi but did show that other dynein-dependent mitotic processes were disrupted (Faulkner et al., 2000; Dujardin et al., 2003; Tsai et al., 2007). This is in contrast to RNAi depletion of Lis1 (also in cultured mammalian cells), which perturbs the distributions of Golgi, lysosomes, and late endosomes (Lam et al., 2010). One possible explanation for these differences is that Lis1 overexpression does not fully disrupt all of Lis1’s function, whereas depletion of Lis1 does. Notably, several studies suggest that Lis1 function is highly dose dependent (Hirotsune et al., 1998; Li et al., 2005; Bi et al., 2009; Wang and Zheng, 2011; Zyłkiewicz et al., 2011). Second, in cultured neurons, injection of a Lis1 function-blocking antibody or RNAi of Lis1 inhibited the motility of large but not small lysosomes or acidic organelles, raising the possibility that Lis1 is specifically required for high-load functions (Pandey and Smith, 2011; Yi et al., 2011). Here, in A. nidulans hyphae lacking the sole genomic copy of Lis1, we find no evidence for a size-dependent requirement for Lis1 function; both small endosomes and larger peroxisomes and nuclei require Lis1 for proper distribution. In filamentous fungi, internal turgor pressure directed toward the hyphal tip participates in hyphal growth (Lew, 2011), raising the possibility that dynein encounters significant opposing force while transporting cargo in the retrograde direction. However, we still observe occasional endosome movements in the absence of Lis1, indicating that Lis1 is not required to overcome any additional load imparted by turgor pressure.

One of the most extensive models of the cell biological function of Lis1 comes from studies of its role in S. cerevisiae (Fig. 9 A; Lee et al., 2003, 2005; Sheeman et al., 2003; Markus et al., 2009, 2011; Markus and Lee, 2011). In yeast, the dynein pathway is required for mitotic spindle alignment and nuclear migration (Eshel et al., 1993; Li et al., 1993). In this system, the Lis1 ortholog (also known as Pac1) is required for the targeting of dynein to the microtubule plus end, a prerequisite for off-loading the motor to a receptor, Num1, which is present on the cell cortex (Lee et al., 2003; Sheeman et al., 2003). After off-loading, cortically anchored dynein binds to cytosolic microtubules and moves toward the microtubule minus end. Because the microtubule-organizing center is embedded in the nuclear envelope of the daughter cell’s nucleus, these pulling forces bias the movement of the daughter cell’s nucleus into the daughter cell (Stuchell-Brereton et al., 2011).

Mechanistically, these results are consistent with a requirement for Lis1 to initiate but not to sustain dynein-driven motility. Because Lis1 can enhance dynein’s microtubule affinity (Yamada et al., 2008; McKenney et al., 2010), it may assist in concentrating dynein at the microtubule plus end before cortical offloading. Additionally, Lis1 may cause a conformational change in dynein that is a prerequisite for cargo binding (Markus et al., 2009). Because Lis1 is not stably localized to the cortex in yeast (Lee et al., 2003; Markus et al., 2011), it does not appear to be required for sustained pulling forces. Current data do not rule out the possibilities that Lis1 is briefly localized to the cortex and plays a role in assisting movements generated by cortically localized dynein or that Lis1 is only transiently associated with the cortex and thus not observable by fluorescence microscopy.

We hypothesize that in A. nidulans, loading of dynein onto endosomes occurs via a similar mechanism (Fig. 9 B). In support of a role for Lis1 in initiating cargo motility, we found that deletion of Lis1 decreases the frequency of endosome and dynein movements but does not dramatically alter the speeds of these movements. Additionally, we showed that dynein and dynactin were stably present on moving endosomes, but Lis1 was not. In our model for Lis1 function in A. nidulans organellar transport, dynein first arrives at the microtubule plus end in a Lis1-independent, kinesin-1–dependent manner (Fig. 9 B, step 1; Zhang et al., 2003). Next, we propose that binding of Lis1 biases dynein to a microtubule-engaged state at the microtubule plus end, perhaps inducing a conformational change in dynein (Fig. 9 B, step 2) that promotes cargo binding (Fig. 9 B, step 3). After dynein is loaded onto cargo, motility proceeds without stably bound Lis1 (Fig. 9 B, step 4). It is possible that binding of dynactin to dynein triggers the movement of dynein away from the microtubule plus end, as a recent study found that NudE (a Lis1 binding partner) and dynactin compete for interaction with the dynein intermediate chain (McKenney et al., 2011). Here, we observed stable association of dynein and dynactin but not Lis1 with moving endosomes. As is the case in yeast, we cannot exclude the possibility that Lis1 is transiently associated with endosome-bound dynein and plays some further role in the commencement of cargo movement or during cargo motility. Overall, our results support a role for Lis1 in initiating dynein-based cargo transport.

Interestingly, we find that dynein or Lis1 deletion reduces the frequency of both anterograde and retrograde transport of endosomes. As described in the Results section, such interdependence of bidirectional transport has been observed in many eukaryotic organisms. Anterograde transport of endosomes in Aspergillus is mediated by kinesin-3/UncA. Our ability to observe only rare kinesin3/UncA movements in the absence of dynein suggests that kinesin3/UncA requires dynein for recycling back to the microtubule minus end. Based on our cargo initiation model, we interpret the loss of anterograde motility in the absence of dynein or Lis1 as being a direct consequence of loss of Lis1-dependent initiation of retrograde motility, which would be required for kinesin3/UncA recycling.

A. nidulans strains were grown in yeast extract and glucose medium (Szewczyk et al., 2006) or minimal medium (Nayak et al., 2010), supplemented with 1 mg/ml uracil, 2.4 mg/ml uridine, 2.5 µg/ml riboflavin, 1 µg/ml para-aminobenzoic acid, and 0.5 µg/ml pyridoxine when required. Expression from the alcA promoter was induced by replacing glucose with 0.1% fructose. Glufosinate was used at a final concentration of 25 µl/ml (Nayak et al., 2010).

A. nidulans strains used in this study are listed in Table S1. Strains containing deletions of kinesin-1/kinA, kinesin-3/uncA, kinesin-3/uncB, dynein/nudA, and Lis1/nudF were created by homologous recombination to replace the endogenous gene with AfpyrG (Aspergillus fumigatus pyrG), Afpyro (Aspergillus fumigatus pyro), or bar (Straubinger et al., 1992) as selectable markers (Szewczyk et al., 2006). The nudA::AfpyrG and nudF::AfpyrG deletion cassettes were provided by S. Osmani (Ohio State University, Columbus, OH), J. Dunlap (Dartmouth Medical School, Hanover, NH), and the Fungal Genetics Stock Center. Gene deletions were verified by Southern blot or PCR analyses using genomic DNA isolated, as previously described (Lee and Taylor, 1990).

Constructs to delete genes or tag endogenous genes with TagGFP2 (A. nidulans codon optimized; Subach et al., 2008), GFP, 3xGFP, or mCherry were generated using fusion PCR (Szewczyk et al., 2006) or yeast homologous recombination (Orr-Weaver et al., 1983) and integrated at the endogenous locus using homologous recombination in strains lacking ku70 (Nayak et al., 2006). Tagged strains were confirmed by PCR. Some strains were created through genetic crosses, as previously described (Todd et al., 2007).

For live-cell imaging of mature hyphae, spores were inoculated on minimal medium plates containing the appropriate auxotrophic supplements and grown at 37°C for 12–24 h. Colonies were excised from plates as agar squares, inverted onto 10 µl of liquid minimal medium in an eight-chambered chamber (Lab-Tek; Thermo Fisher Scientific) or a 35-mm FluoroDish (World Precision Instruments, Inc.) and observed at 22°C. For live-cell imaging of germlings, spores were resuspended in 0.01% Tween 80 solution. The spore suspension was diluted at 1:1,000 in liquid minimal medium containing appropriate auxotrophic supplements. The spore and media mix (3 ml) was added to a 35-mm FluoroDish and incubated at 22°C for 15–20 h before imaging.

Wide-field fluorescence images were collected at 22°C using a Plan-Apochromat 100×/1.40 oil objective (Nikon) on an epifluorescence microscope (Ti; Nikon) with the Perfect Focus System (Nikon) and a charge-coupled device camera (ORCA-ER; Hamamatsu Photonics), all controlled by Elements software (Nikon). Simultaneous multicolor wide-field images were collected at 22°C using a custom-built microscope (DeltaVision OMX; Applied Precision) equipped with a Plan-Apochromat N 60×/1.12 oil objective lens (Olympus). GFP and mCherry were simultaneously excited using 488- and 592.5-nm laser lines. A series of dichroic mirrors in the OMX light path (Standard Filter Set v2; Applied Precision) was used to split the emission light from the two fluorophores to two different cooled charge-coupled device cameras (CoolSNAP HQ2; Photometrics). A narrow-band emission filter in front of each camera was used to select appropriate wavelengths (525/20 and 615/24 for GFP and mCherry, respectively). Images were acquired using DeltaVision software (v2.23) and deconvolved using softWoRx image processing software (v4.50; Applied Precision) using the enhanced ratio method. Brightness and contrast adjustments to images were subsequently made using ImageJ (v1.43; National Institutes of Health) or Photoshop CS4 (v11.0.2; Adobe), and figures were compiled in Illustrator CS4 (v. 14.0.0; Adobe).

The percentage of unidirectional microtubules in hyphal tips was determined by observing the polarity of EB1-mCherry comets. Comets with tails pointing toward the last nucleus and tips pointing toward the hyphal tip versus comets oriented in the opposite direction were scored.

For quantitative analysis of endosome velocity, we used a strain in which endosomes were marked with A. nidulans codon–optimized TagGFP2 (Subach et al., 2008)-Rab5/RabA. TagGFP2 is a GFP variant that is significantly brighter than GFP or EGFP (Subach et al., 2008). N-terminal GFP tagging of Rab5/RabA was previously shown to be functional (Abenza et al., 2009), and we observed no growth differences between tagging the endogenous rabA locus versus the argB-targeted rabA expressed under the control of the alcA promoter (Abenza et al., 2009).

Endosome, peroxisome, and dynein velocities were measured using ImageJ (v1.43). Maximum-intensity projections were generated from time-lapse sequences to define the trajectory of particles of interest. The segmented line tool was used to trace the trajectories and map them onto the original video sequence, which was subsequently resliced to generate a kymograph. The velocities of individual particles were calculated from the inverse of the slopes of kymograph traces. Dynein flux measurements were calculated from kymographs generated from 10-s time-lapse sequences by counting the number of dynein-3xGFP particles entering and leaving the microtubule plus end. Microtubules were selected for analysis based on image quality and the clarity of discrete particle movements. Endosome flux was calculated as the total number of vesicles that crossed a line drawn perpendicular to and 10 µm away from the hyphal tip over 3 min.

For quantitative analysis of organelle size, endosomes and peroxisomes were measured along their longest axis using the line tool and measure function in ImageJ (v1.43). Our measurements for peroxisomes were in the range of those described in previous EM analysis of these organelles in A. nidulans (Valenciano et al., 1998). Our measurements for endosomes were close to the theoretical diffraction limit for our imaging conditions; thus, this number represents an upper limit for endosome size. Consistent with our analysis, EM studies of endosomes in A. nidulans and mammalian cells report sizes in the range of <100–500 nm (Murk et al., 2003; Griffith et al., 2011).

Data visualization and statistical analyses were performed using GraphPad Prism (5.0d; GraphPad Software), Excel (v12.3.2; Microsoft), MATLAB (MathWorks), and ImageJ (v1.43). For Figs. 4 (D and E) and 5 (B and C), the Lis1Δ histogram data were shifted 0.25 U to the right relative to the wild-type data to visualize both datasets simultaneously; the Gaussian fits of the same data were not shifted.

Fig. S1 shows that localization of dynein to microtubule plus ends requires kinesin-1 but not kinesin-3/uncA or kinesin-3/uncB. Fig. S2 shows that endosome and peroxisome transport is unaffected by deletion of kinesin-3/uncB. Fig. S3 shows that kinesin-3/UncA accumulates in the hyphal tip in the absence of dynein. Fig. S4 shows that tagging dynein or Lis1 with GFP does not affect function. Fig. S5 shows that dynactin is required for endosome and dynein motility. Video 1 shows GFP-Rab5/RabA–labeled endosome dynamics in a wild-type hypha. Video 2 shows Pex11/PexK-GFP–labeled peroxisome dynamics in a wild-type hypha. Video 3 shows GFP-Rab5/RabA–labeled endosome dynamics in a dyneinΔ hypha. Video 4 shows GFP-Rab5/RabA–labeled endosome dynamics in a kinesin3/uncAΔ hypha. Video 5 shows dynein-3xGFP dynamics in a wild-type hypha. Video 6 shows Lis1-GFP and mCherry-Rab5/RabA–labeled endosomes in a wild-type hypha. Table S1 lists the A. nidulans strains used in this study.

We would like to thank Michael Hynes (The University of Melbourne, Melbourne, Australia), Berl Oakley (University of Kansas, Lawrence, KS), Miguel Peñalva (Centro de Investigaciones Biologicas CSIC, Madrid, Spain), and Xin Xiang (Uniformed Services University, Bethesda, MD) for strains, reagents, and protocols; Stephen Osmani, Jay Dunlap, and the Fungal Genetics Stock Center for deletion cassettes; Michael Wu and Natalie Heer for help generating strains and reagents; Anthony Roberts, Julie Huang, Mark McClintock, and Xin Xiang for comments on the manuscript; Gaudenz Danuser and Applied Precision for use of the custom-built DeltaVision OMX microscope; and the Nikon Imaging Center at Harvard Medical School for their support.

S.L. Reck-Peterson is funded by the Rita Allen Foundation, the Armenise-Harvard Foundation, and a National Institutes of Health New Innovator award (1 DP2 OD004268-01).