Glu415 in the α-tubulins plays a key role in stabilizing the microtubule–ADP-kinesin complexes

Most of the factors required during early embryogenesis are maternally provided in the animal kingdom (DeRenzo and Seydoux, 2004; Tadros and Lipshitz, 2005). To understand some features of maternal effect, a set of dominant female sterile (Fs) mutations was induced and isolated in Drosophila melanogaster (Erdelyi and Szabad, 1989; Szabad et al., 1989). It was assumed that several of the Fs mutations identify genes that are engaged in the initiation and progression of embryogenesis. Elaborating their molecular functions may shed light on the beginning of new lives.

The Tomaj^D and the Kavar^D Fs mutations identify the αTubulin67C gene (αTub67C) (Mathe et al., 1998; Venkei et al., 2006; Venkei and Szabad, 2005), which encodes α4-tubulin, the so-called maternal α-tubulin, a rather divergent α-tubulin isoform (Kalfayan et al., 1982; Kalfayan and Wensink, 1982; Matthews et al., 1989; Matthews et al., 1993; Matthies et al., 1999; Theurkauf, 1992). (For a comprehensive list of the α4-tubulin-related references see FlyBase at http://flybase.bio.indiana.edu/.)

The αTub67C gene is expressed during oogenesis in the nurse cells. The encoded α4-tubulin molecules are dumped into the egg cytoplasm and are utilized during early embryogenesis (Matthews et al., 1989; Venkei et al., 2006). They constitute about 20% of the α-tubulin pool in the Drosophila egg cytoplasm and their function cannot be replaced by the constitutively expressed α1-tubulin and α3-tubulin isoforms, products of the αTub84B and αTub84D genes (Matthews et al., 1989). Analysis of partial gain- and loss-of-function αTub67C mutant alleles suggested the involvement of α4-tubulin in meiosis, cleavage mitoses, formation of the sperm aster and in the embryonic nervous system (Mathe et al., 1998; Matthews et al., 1993; Theurkauf, 1992; Matthies et al., 1999). Study of the mutant phenotypes associated with (1) Kavar^18c, one of the Kavar^D mutations, and (2) with the complete loss of αTub67C gene function revealed that α4-tubulin is required for the fast formation of long microtubules (MTs) during the initial cleavage divisions (Venkei et al., 2006). The α4-tubulin molecules have been shown to be preferentially incorporated into the interpolar MTs and assist them in embracing the nuclear envelope. The interpolar MTs, which grow fast, push apart the daughter centrosomes along the nuclear perimeter to opposite poles in the early cleavage Drosophila embryos (Venkei et al., 2006).

Kavar^21g, another Kavar^D mutation, the result of a single base pair transition mutation (G1276→A), which causes the replacement of Glu426 by Lys [E426K-α4-tubulin (Venkei and Szabad, 2005)]. Glu426 resides in the twelfth α-helix of α4-tubulin, which – along with the eleventh α-helix and the unstructured C-terminus of the tubulin – is involved in formation of the outer surface of the MT lattice and has been implicated to be part of the landing and binding platform for the mechanoenzymes (Hoenger et al., 1998; Kikkawa and Hirokawa, 2006; Lakamper and Meyhofer, 2005; Mizuno et al., 2004; Wang and Sheetz, 2000). Mostly the positive residues of the kinesin head establish contact with the negatively charged MT surface in the landing and binding platform (Marx et al., 2006; Woehlke et al., 1997). Some of the negatively charged amino acids in the twelfth α-helix of the β-tubulins (E410, D417, E421) have been shown to be necessary to sustain the microtubule-kinesin interaction in the ATP-bound strong binding state (Uchimura et al., 2006). Removal of the C-terminus of the tubulins (by subtilisin digestion) decreases the processivity and the transport efficiency of the associated mechanoenzymes through the regulation of the rate of ADP release and the transition to the next duty cycle of the motor head (Lakamper and Meyhofer, 2005; Wang and Sheetz, 2000).

To understand the mode of E426K-α4-tubulin action and the basis of the Kavar^21g-related sterility, we analyzed ooplasmic streaming and the kinetics of lipid-droplet transport, using confocal reflection microscopy (CRM), in developing Drosophila egg primordia (Gaspar and Szabad, 2009). We report here that replacement of Glu426 by Lys in α4-tubulin brings about (1) a complete cessation of the fast ooplasmic streaming in the developing oocyte and (2) a decrease in lipid-droplet transport efficiency. E426K-α4-tubulin reduces the average transport speed and the dwell time, and consequently results in shorter than normal runs of the cargo. The E426K-α4-tubulin-related defects stem most probably from a decreased stability of the MT-kinesin complexes and are related to the impaired stabilizing forces in the ADP-bound state of the motor head, as has also been suggested by molecule models (Kikkawa and Hirokawa, 2006; Yonekura et al., 2006). The importance of Glu426 in the MT-kinesin interaction is strongly supported by the genetic interaction between Kavar^21g and loss-of-function mutant alleles of the different kinesin family members.

Since (1) the twelfth α-helix of α4-tubulin, containing Glu426, is highly conserved within all the 42 known α-tubulin isoforms in 32 species analyzed (in a BLAST analysis), (2) Glu426 in α4-tubulin corresponds to Glu415 in the evolutionary highly conserved α-tubulins, and (3) Glu415 is 100% conserved evolutionarily in all the studied α-tubulin isoforms, we propose that Glu415 plays a key role in stabilizing the microtubule–ADP-kinesin complexes.

The approach presented in this paper may be of special importance since the replacement of Glu415 by Lys in the evolutionarily highly conserved and constitutively expressed α-tubulins is expected to bring about dominant lethality. However, the same types of mutations in the α4-tubulin encoding αTub67C gene alter fate of only the egg primordia and the descending embryos and do not affect viability of the adults. The developing egg primordia thus offer a convenient system to analyze the mutation-related defects and provide a profound understanding about the roles of the different amino acids in the MT-associated functions.

Embryogenesis does not commence or proceeds abnormally inside the otherwise normal looking eggs of the Kavar^21g/+ females: several of the cleavage nuclei fail to migrate into the egg cortex leaving large areas free of nuclei (Fig. 1) (Venkei and Szabad, 2005). The Kavar^21g-related defects presumably arise from altered abilities of the MTs to function as migration routes for the motor molecules since the Kavar^21g-encoded E426K-α4-tubulin molecules do not prevent microtubule formation (Venkei and Szabad, 2005) (see also Fig. 4). In fact, cleavage spindles do form in embryos of the Kavar^21g-carrying females (see supplementary material Movies 1 and 2), although they are frequently abnormal. Because embryogenesis in eggs laid by Kavar^21g-carrying females suffers from a wide range of defects, we decided to analyze the developing late egg primordia to elucidate the mode of E426K-α4-tubulin action. As several components of the egg cytoplasm are transported along the MTs in the growing oocyte during oogenesis, we decided to study two transport processes in the developing egg primordia, (1) ooplasmic streaming and (2) transport of the lipid droplets.

During ooplasmic streaming the entire ooplasm swirls around the longitudinal axis of the egg primordia to ensure the mixing of the yolk-free cytoplasm (dumped from nurse cells into the oocyte) with the follicle-cell-derived yolk granules. The process is propelled mostly, if not exclusively, by kinesin heavy chain (KHC) in the stage 10B-12 oocytes (Palacios and St Johnston, 2002; Serbus et al., 2005) (Fig. 2). Although the yolk granules stream steadily along a linear paths at about 350 nm/second in the control oocytes [362±5 nm/second in the wild type (+/+) and 348±5 nm/second in the heterozygous (+/–) oocytes; mean ± standard error of the mean), there is no ooplasmic streaming in presence of E426K-α4-tubulin, that is in the egg primordia of the Kavar^21g/+ and the Kavar^21g/–hemizygous females (Fig. 2; `–' stands for Df(3L)55, a short deficiency that removes the αTub67C and three of the adjacent genes) (Venkei and Szabad, 2005). The absence of ooplasmic streaming leads to a gradual accumulation of the yolk granules in the posterior oocyte cytoplasm, thus leaving the anterior cytoplasm free from the yolk granules (Fig. 3). A plausible explanation for the lack of ooplasmic streaming could be an absence of the sufficiently long MTs in the oocytes, as was reported for the egg primordia of the Kavar^18c-carrying females (Venkei et al., 2006). The Kavar^18c-encoded E82K-α4-tubulin molecules become incorporated into the MTs and prevent the formation of sufficiently long MTs, which are required for ooplasmic streaming (Venkei et al., 2006; Gaspar and Szabad, 2009). In contrast to the Kavar^18c-carrying egg primordia, there are numerous and long MTs present in oocytes containing the E426K-α4-tubulin at stages 10B-12, as revealed by the distribution of Jupiter-GFP highlighted MTs (Fig. 4). However, and unlike in the wild-type oocytes, the MTs of these oocytes do not form thick parallel bundles, most probably as a consequence of the abnormal MT-kinesin interaction.

Remarkably, there is ooplasmic streaming in egg primordia of the kavar^null/– females, i.e. in the absence of α4-tubulin, although the average streaming velocity of the yolk granules drops from 348±5 to 317±8 nm/second (P<0.001, as compared with the +/– control; Fig. 2). This observation indicates that α4-tubulin is not required for ooplasmic streaming, although its presence influences the process. The E426K-α4-tubulin-related defects originate from the direct effect of the mutant E426K-α4-tubulin molecules and not from the absence of α4-tubulin function, thus excluding the dominant negative nature of Kavar^21g.

The Kavar^21g-associated mutant phenotypes and position of the Glu426→Lys replacement suggested altered interaction between MTs and mechanoenzymes and MT-based transport in the presence of E426K-α4-tubulin. To test this possibility, we constructed Kavar^21g/+ females from which one copy of a motor-encoding gene was removed (either by a mutation or by a deficiency; Table 1). We then analyzed 3- to 4-hour-old embryos and determined how far they had progressed in development (as illustrated in Fig. 1). The results are summarized in Table 1 and clearly show a genetic interaction between Kavar^21g and the studied members of the kinesin gene family. Upon removal of one copy of the genes encoding the plus end tracking motor kinesin heavy chain khc or the bipolar kinesin klp61F from Kavar^21g/+ flies, the early cleavage defects occur in a significantly larger proportion of embryos than in embryos of the control Kavar^21g/+ females (Table 1). Upon removal of one copy of the ncd gene, which encodes the Drosophila minus-end kinesin, the embryos reach a later stage in development than the control embryos (Table 1). In contrast to the kinesins, the removal of one copy of Dhc64C (which encodes cytoplasmic dynein heavy chain) or the Glued (which encodes a major subunit of the dynein regulator dynactin) did not alter the Kavar^21g-imposed defects, implying that E426K-α4-tubulin does not perceptibly affect the dynein-based transport (Table 1). (Note that the removal of one cope of any of the studied MT-motor encoding genes does not reduce the viability of the embryos from females with two normal αTub67C genes; Table 1.) The genetic interaction between Kavar^21g and the kinesins suggests an important role of Glu426 in α4-tubulin in the MT-kinesin interaction and therefore in MT-based transport.

The complete blockage of ooplasmic streaming in the presence of E426K-α4-tubulin indicates severe difficulties in MT-motor interaction. To characterize the nature of the E426K-α4-tubulin-related defects, we studied transport of the ∼0.5 μm lipid droplets in the ooplasm of the developing egg primordia. The lipid droplets, which are dumped in high numbers into the oocyte cytoplasm from the nurse cells during oogenesis, are not only the mayor lipid storage vesicles but are sites of protein sequestration and appear to be engaged in regulated inactivation of proteins, preventing the formation of toxic protein aggregates and delivery of signalling molecules (Welte, 2007). The lipid droplets have been known to be transported along the MTs by the mechanoenzymes during oogenesis and embryogenesis (Welte et al., 1998; Gross et al., 2000; Shubeita et al., 2008; Gaspar and Szabad, 2009). The refractive nature of the lipid droplets offers a convenient way, using CRM, to study their motion in the oocyte cytoplasm of the developing egg primordia (Paddock, 2002; Vesely and Boyde, 2001; Gaspar and Szabad, 2009).

Surprisingly, although E426K-α4-tubulin completely abolishes ooplasmic streaming, it does not inhibit the motion of those lipid droplets that move independently of streaming through the MT-based active transport mechanism (Fig. 4) (Gaspar and Szabad, 2009). Although the plus- and the minus-end-directed motions cannot be distinguished in the absence of ooplasmic streaming, analysis of the moving lipid droplets may well reveal the effects of E426K-α4-tubulin on MT-based transport. With this idea in mind, we focused attention on lipid-droplet motion in the ooplasm of the early stage 12 egg primordia. Considering the possible modifying effects of the viscous load that may be exerted by the flowing ooplasm on the transported lipid droplets, we analyzed wild-type, early stage 12 oocytes that were in the process of recovering from cold treatment. Note that ooplasmic streaming, as all the MT-based transports, cease upon cold treatment as a result of depolymerization of the MTs. MTs form again during the recovery from clod treatment and although the lipid-droplet transport along the MTs resumes, ooplasmic streaming is still absent during the subsequent 12-15 minutes (Gaspar and Szabad, 2009).

Although the number of the motile lipid droplets is similar over the studied cytoplasmic region and during the recording time in the egg primordia of the females with different genotype (as listed in Table 2), the average velocity of the lipid droplets is significantly lower in the presence of E426K-α4-tubulin as compared with the control (Table 2, Fig. 5). Moreover, the average dwell time and thus the run lengths decrease in the presence of E426K-α4-tubulin (Table 2). The average run length, i.e. the effective transport range of a single run, drops to 60% of wild-type run length (Table 2). To test the significance of the decrease in dwell time and run lengths in the presence of E426K-α4-tubulin, we carried out survival analyses of these parameters, i.e. we tested the survival rate of the lipid droplet runs as a function of time and distance travelled. Apparently, the dwell time and consequently the run length are significantly shorter in the presence of E426K-α4-tubulin than in the control oocytes (Table 2, Fig. 5).

Absence of α4-tubulin also results in significant decrease in the speed of the lipid droplets, as in the presence of E426K-α4-tubulin. This finding may cast doubt on the independent action of E426K-α4-tubulin from the α4-tubulin function. However, the survival analyses of both the dwell time and the run length revealed no significant departure from the wild-type control values (Table 2, Fig. 5). The dwell time findings indicate that although the presence of α4-tubulin in the MTs may increase motor processivity through velocity, i.e. the number of steps taken by the motor during a period, it has no significant influence on the dwell time of the travelling motor by altering the MT-motor dissociation rate, or by some other way. It can thus be deduced that the absence of the α4-tubulin molecules has no significant impact on the effective range of the single runs.

Apparently, the basis of the deleterious action of E426K-α4-tubulin is a reduced effective range of transport runs. However, it is not clear how E426K-α4-tubulin disturbs the transport along the MTs. Does it increase the MT-motor dissociation rate or does it bring the motor head into a rigorous state by trapping the cargo, as was described for E164A-KHC (Klumpp et al., 2003)? Analysis of the ends of the single runs revealed three types of events: (1) The lipid droplets dissociate from the track and diffuse away in about 80% of cases. Some of the droplets regain motion after a short pause. These droplets move away either (2) by using the same track (∼15%) or (3) by changing track (∼5%, i.e. there is a significant change in the direction of the motion). The average time is about 0.5 seconds in changing track and is about 0.2 seconds for in the `track-keeping' pauses. The parameters of the ending events did not change significantly in the presence of E426K-α4-tubulin, implying that the trapping hypothesis is more than unlikely (Table 3).

A fraction of the dissociating lipid droplets could be followed for more than two frame intervals after the ending event (0.195 seconds). A mean square displacement (MSD) analysis of these lipid droplets revealed that E426K-α4-tubulin does not influence the free diffusion of the lipid droplets, which otherwise followed a more-or-less linear path after they dissociated from MTs (Fig. 6). The frequency of the trapped lipid droplets that finished their motion with a pause (stuck), remained very low irrespective of the presence or absence of E426K-α4-tubulin (Table 3). The present findings exclude the possibility of the trapping mode of E426K-α4-tubulin action.

To ship a wide variety of components, the eukaryotic cells rely largely on the MTs and the associated transport mechanisms. In Drosophila, as in many other animal species, the MT-based machinery is especially important during early embryogenesis when the cleavage cycles are repeated every 8-10 minutes and the MTs need to span over large distances (Foe et al., 1993; Tadros and Lipshitz, 2005). In fact, α4-tubulin, a special α-tubulin isoform could be said to be `designed' to meet the unusual requirements associated with the MTs during the early cleavage divisions (Venkei et al., 2006). Although the overall homology is only about 70% between α4-tubulin and the evolutionarily highly conserved members of the α-tubulin family, the C-terminal structures, especially the twelfth α-helix, show 96% homology to other members of the α-tubulin family (Fig. 7). This region of the α-tubulins participates in forming the outer surface of the MTs, where most of the MT-protein interactions take place, including landing and binding of the mechanoenzymes (Marx et al., 2006). Since the Kavar^21g mutation resulted in a Glu→Lys replacement in the position 426, which is a component of the twelfth α-helix, a disturbed MT-mechanoenzyme interaction may well account for the Kavar^21g-related defects.

Indeed, the genetic interaction between Kavar^21g and members of the kinesin family genes clearly confirmed the above assumption and showed an involvement of Glu426 in interactions with (1) KHC, the major component of the classical kinesins that is involved in several MT-dependent transport processes (Saxton et al., 1991); (2) the mitotic kinesin Klp61F, the orthologue of Eg5, a bipolar kinesin that provides the force necessary to keep the mitotic spindle bipolar and acts against the spindle contractor motor; and (3) NCD, a minus-end-directed kinesin. In wild-type embryos, the poleward and plateward forces exerted by Klp61F and NCD are in a strict spatiotemporal balance in maintaining spindle integrity during the cell cycles (Cytrynbaum et al., 2003; Sharp et al., 2000). E426K-α4-tubulin appears to bring about a disturbance in the coordination of the kinesin motors as indicated by the high divergence in the size and shape of the cleavage spindles and by the consequent loss of many of the cleavage nuclei in embryos of the Kavar^21g-carrying females (Venkei and Szabad, 2005). Although a partial removal of Klp61F brings about a shift in the already unbalanced spindles toward spindle collapses, partial removal of ncd, increases the frequency of the seemingly normal divisions. Although it is not clear how removal of one copy of the KHC encoding gene brings about earlier-than-usual arrest of embryogenesis of the Kavar^21g/+-derived embryos, the involvement of KHC in embryogenesis has already been shown (Serbus et al., 2005).

To understand the mode of E426K-α4-tubulin action and the role of Glu426 in MT-kinesin interaction, we made use of the fact that α4-tubulin is not essential for the formation of normal eggs and is used only during the initial cleavage divisions (Venkei et al., 2006) (and this paper). However, once produced in the nurse cells and dumped into the oocyte cytoplasm from stage 10 of egg primordia development, much of the α4-tubulin and the E426K-α4-tubulin molecules are incorporated into the MTs and thus provide a convenient means of analysis of the role of Glu426 in MT-kinesin interaction. The egg primordia develop normally in the Kavar^21g/+ females and the MT-based transport processes are not hampered by E426K-α4-tubulin up to stage 10, the time when the amount of α4-tubulin starts to increasing in the ooplasm. Simultaneously, the MT-based transport processes become pronounced in the oocyte. In the presence of E426K-α4-tubulin the fast ooplasmic streaming is absent as it happens following the removal of KHC from the ooplasm (Serbus et al., 2005). Since absence of the wild-type α4-tubulin has little impact on the process, it is safe to conclude that E426K-α4-tubulin is responsible for the stoppage of ooplasmic streaming. E426K-α4-tubulin most probably acts through interfering with KHC, the stream-propelling motor, and not by blocking wild-type α4-tubulin function.

As revealed by analysis of lipid-droplet transport, E426K-α4-tubulin disrupts MT-based transport through decreasing the effective transport range of the carrier-cargo complexes. The basis of E426K-α4-tubulin action is almost certainly an increase in the MT-motor dissociation rate. During the duty cycle, the two-headed kinesin molecules effectively couple the chemical energy release from ATP to mechanical force exertion (Schief and Howard, 2001). In this process not only those structures that are engaged in binding the motor head domain change over the MT surface, but also the binding force is altered along with a change in the nucleotide state of the motor (Cross, 2004). The MT-motor complex reaches the most labile state during the release of ADP from the leading head (Schief et al., 2004). The speed of ADP release is relatively slow compared with the detachment rate of the trailing head, and there is an approximately 1% chance, even under normal conditions, for dissociation of the MT-motor complex (Schief et al., 2004). An assessment of the MT-kinesin head complex revealed two positively charged Arg residues in the sixth α-helix of the ADP-bound kinesin head domain residing in close vicinity of Glu426 (Fig. 8) (Kikkawa and Hirokawa, 2006). It is probable that the ionic interaction between these three amino acids is responsible for the stabilization of the motor head on the MT during this labile state (Yun et al., 2003; Yonekura et al., 2006), although it cannot be ruled out that E426K-α4-tubulin may alter the overall structure of the twelfth α-helix, and thus the destabilization of the MT-kinesin complexes may be the consequence of several altered residue-residue interactions.

In support of the former hypothesis, an analysis of the homologous region in members of the Drosophila kinesin family revealed that the two positively charged residues, especially the second one, are well conserved in the kinesin motors (Fig. 9). The charge reversal imposed by the E426K-α4-tubulin molecules reduces or abolishes the binding force and thus increase the instability of the MT-kinesin complex, leading to frequent MT-motor dissociations. Knowing that α4-tubulin comprises about 20% of the α-tubulin pool in the egg cytoplasm (Matthews et al., 1989), E426K-α4-tubulin is present in about 10% of the tubulin heterodimers in the MTs, posing possible sites for kinesin drop offs and thus reduced transport efficiency. Since many of the kinesin motors carry on various functions during the cleavage mitoses, it is understandable how Kavar^21g can bring about a wide range of defects that lead to female sterility.

The Kavar^21g-related Glu426→Lys amino acid exchange came about in the twelfth α-helix, in a region that is highly conserved in other members of the α-tubulin family (Fig. 7). Since Glu426 corresponds to Glu415 in the constitutively expressed α-tubulin isoforms and Glu415 is evolutionary 100% conserved in every member of the α-tubulin family, it appears that Glu415 plays a key role in stabilization of the ADP-kinesin head on the MT surface and is thus a key component of the MT- and kinesin-based transport.

Owing to the highly conservative nature and crucial functions of the MTs in all the eukaryotic cell types, a mutation producing a Glu415→Lys change in the constitutively expressed Drosophila α1- or in the α3-tubulin would presumably bring about dominant lethality and make analysis of the role of the studied amino acids practically impossible. However, such mutations could be of great help in understanding the motor-interacting properties and other surface-associated MT functions. Expression of the αTub67C gene during development of the egg primordia offers a convenient analysis of the encoded mutant proteins in MT-based transport and an elucidation of the role of any particular residues of the α(4)-tubulins in the process.

Kavar^21g is a dominant female-sterile mutation in the αTub67C gene of Drosophila melanogaster and encodes E426K-α4-tubulin (Erdelyi and Szabad, 1989; Venkei and Szabad, 2005). X-ray-induced reversion of Kavar^21g resulted in the formation of kavar^rX21 (designated as kavar^null), a complete loss-of-function allele of the αTub67C gene (Venkei and Szabad, 2005).

To conduct the genetic interaction tests, we generated Kavar^21g/+ (and, as control, +/+) females from which one copy of either of the following MT-motor-encoding genes were removed by a mutation or by a short deficiency: (1) kinesin heavy chain by khc⁸, a null allele (Saxton et al., 1991), (2) klp61F by the Df(3L)ED202 deficiency, (3) non-claret-disjunctional (ncd^–) by Df(3R)ca^nd (Yamamoto et al., 1989), (4) cytoplasmic dynein heavy chain by Dhc64C^4-16, a hypomorphic mutant allele (Gepner et al., 1996) and (5) Glued, a mayor component of the dynactin complex, by the null allele Gl (Ghosh-Roy et al., 2004). For an explanation of the genetic symbols see FlyBase at http://flybase.bio.indiana.edu.

To label the MTs, a Jupiter-GFP expressing transgenic line was used (Karpova et al., 2006). Jupiter-GFP and Df(3L)55 were recombined onto the same chromosome. The Jupiter-GFP Df(3L)55/Kavar^21g females were descended from a cross between Jupiter-GFP Df(3L)55/TM3, Sb Ser females and Kavar^21g/TM3, Sb Ser males. The flies were raised on standard food and kept on 25°C.

Three- to four-hour old embryos were collected from females of different genotypes (as listed in Table 1). The embryos were fixed as described by Venkei et al. (Venkei et al., 2006), stained with 1 μg/ml Hoechst 33342 (Sigma 14533) for 15 minutes. The embryos were then scored for nuclear phenotype (Fig. 1).

Ovaries were dissected and the egg primordia were prepared for analysis as described by Gaspar and Szabad (Gaspar and Szabad, 2009). Briefly, the ovarioles were dissected in BRB80 buffer (80 mM Pipes, 2 mM MgCl₂ and 1 mM EGTA), transferred to a coverslip and covered with halocarbon oil after the removal of the excess buffer. Yolk autofluorescence was excited by a 405 nm CO₂ laser beam and detected in the 425-475 nm spectrum in an Olympus FV1000 confocal microscope.

Confocal reflection microscopy (CRM) and image analysis were carried out according to Gaspar and Szabad (Gaspar and Szabad, 2009). Briefly, for reflection imaging the 633 nm He-Ne laser line with 0.6-1.0 mW power was used in combination with an LP560 emission filter and the third-in-order photomultiplier. Such arrangement allows simultaneous combination of CRM with fluorescence imaging. There were 19.5-second long high-speed recordings (∼15 f.p.s., a total of 300 frames) conducted over 20.9×20.9 μm² cytoplasm areas, two for every analyzed early stage 12 oocyte, which were selected randomly. The recordings were analyzed for the lipid-droplet motion direction and speed, other than the ooplasmic streaming. Movement of the lipid droplets was tracked by using the SpotTracker plug-in of ImageJ (http://rsb.info.nih.gov/ij/). Analysis of the trackings was done by a series of custom-made Excel macros in order to split the trackings to individual runs and pauses.

Statistical analysis of the data was carried out using SPSS 15. One-way ANOVA tests were used for comparison of the means. Pair-wise Kaplan-Meier tests and the Cox regression analysis was used for survival analysis of lipid droplet runs. The MSD plot, as described in the text, and the curve fittings were done in Microsoft Excel.