CELL BIOLOGY

Cytoplasmic streaming

In growing hairs, the majority of the cytoplasm is between the nucleus and tip. There are usually several large vacuoles in root hairs, the largest being proximal to the nucleus. The formation of cytoplasmic strands is most commonly seen around small vacuoles in the region between nucleus and tip.

Streaming is usually seen indirectly as the movement of particles in the cytoplasm, and is presumed to be based on a motor system similar to actin-myosin-based motility (see Kuroda, 1990, for review of cytoplasmic streaming in plant cells). Streaming is not easy to observe near the tip of the hair, most likely because of the high concentration of vesicles there (see later section on vesicles). From personal observations of streaming in clover and Vicia, root hairs exhibit fountain streaming, as described for pollen tubes (eg Lancelle and Hepler, 1992), and particles can move from one part of the hair to another by trans-vacuolar strands.

Bi-directionality of streaming has been reported by Emons (1987) for Equisetum hyemale root hairs for small particles about 2.5 µm in diameter, in the subapical and basal regions of the hair. Seagull and Heath (1979) have also reported bidirectional streaming, inferred from electron microscope observations of tannic acid-treated radish root hairs, but it is difficult to distinguish polarity of the microfilaments in their micrographs.

A fundamental problem with reported measurements of cytoplasmic streaming has been the assumption that the underlying velocity is constant. Clarkson et al. (1988) reported that for tomato and oilseed rape, the velocity of cytoplasmic streaming in root hair cells was highly variable in space and time, between cells and within a single cell. The most recent comprehensive study on particle movement in root hairs (where particle movement is equated with cytoplasmic streaming) is by Ayling and Butler (1993) on tomato. In a very careful study, Ayling and Butler recognised that it is difficult to distinguish experimentally-induced changes in particle velocity from those changes due to natural variability, and they used time-series analysis combined with piece-wise linear regression to provide an objective means of assessing experimental manipulation of streaming. Using data from auxin applications, Ayling and Butler were able to report that concentrations of auxin at or weaker than 10-7 M increases the velocity of streaming, concentrations of 10-5 to 10-6 M significantly reduce streaming but that hairs can recover if plain nutrient solution is used to wash away the exogenously applied hormone. Streaming rates vary between 0.6 and 0.9 µm per second in actively growing tomato hairs measured away from the tip.

For clover, streaming during rapid early hair growth has been measured at an average of 3 µm (+/- 0.87 SD) per second (Izumo Mariko, personal communication). Emons (1987) reported speeds at 25 ûC of up to 3.5 µm per second in Equisetum hyemale root hairs at the basal regions and up to 7 µm per second in the tubes of long hairs.

In a study on alfalfa root hairs development, Wood and Newcomb (1989) showed that during root hair initial emergence, cytoplasmic streaming delivers a continuous flow of cellular constituents, and presumably organelles, to the protrusion site. As the protrusion site enlarges, the pattern of cytoplasmic streaming changes until it is eventually directed into and out of the growing hair. Cytoplasmic streaming is rapid during the initial and late phases of growth, but during the period 4-6 hours after emergence, the migratory movement of the nucleus coincides with a slowing of cytoplasmic streaming. During the final phase of growth (up to ten hours after initiation) cytoplasmic streaming slows to similar rates of nearby epidermal cells. Unfortunately Wood and Newcomb didn't report actual measurements of streaming speed. Converse to this, Emons (1987) reported a higher speed of streaming in mature hairs of Equisetum hyemale: 3 to 3.5 µm per second for young growing hairs, but 5 to 7.5 µm per second for mature hairs.

The cytoskeleton

Microtubules (see eg Goddard et al. 1994; Staiger and Lloyd, 1991) and microfilaments (Figs 1 and 3)(see eg Parthasarathy et al. 1985; Staiger and Lloyd, 1991) are present in plant cells, but intermediate filaments have not been reported (although epitopes to animal intermediate filament antibodies have been found in cultured carrot cells, Dawson et al. 1985). There are many reports on the cytoskeleton in tip growing cells, but few for root hairs. The first attempt to present some qualitative and quantitative information on microtubule and microfilament organisation in radish root hairs at the electron microscope level was by Seagull and Heath (1979, 1980), who used tannic acid as an aid to preserve and reveal the filaments. They concluded that bundles of microfilament do not form an integral system but appear fragmentary and don't enter the apical dome of the hair, and they found that microtubules occurred only close to the plasma membrane. They also showed that microtubules are often associated with a single fine filament, thought to be a single actin microfilament.

However, by modifying osmotic conditions during fixation, immunofluorescence studies showed that the cortical microtubules form an integral assembly that progresses from the basal trichoblast to the apical dome (Lloyd 1983; Lloyd and Wells, 1985; Traas et al. 1985) and that bundles of microfilaments are associated with the nucleus (Lloyd et al. 1987). It was also confirmed using both tannic acid addition to chemical fixation (Lloyd et al. 1987) and by freeze-substitution (Lloyd et al. 1987; Ridge, 1988, 1990b) that bundles of microfilaments and microtubules exist endoplasmically, that is away from the cortex of the cell.

Lloyd first suggested the presence of helical cortical arrays of microtubules in root hairs (Lloyd, 1983; Lloyd and Wells, 1985) and for cotton hairs (Lloyd and Seagull, 1985). Lloyd suggested that these arrays are dynamically interconvertible, providing a spring-like device that is sensitive to environmental cues and able to shift the axis of cell expansion through 90û. In his 1985 paper (Lloyd and Seagull, 1985) Lloyd argues strongly that the hoops of microtubules regularly seen in root cortical cells are indeed helices composed of relatively long microtubules. However, in his comprehensive review on the orientation of cortical microtubules, Williamson (1991) in a long and very careful analysis of the published work, argues strongly against this dynamic helical model.

Recent work on the cytoskeleton of root hairs has been based on freeze-substitution and dry cleaving techniques (eg Emons and Derkson, 1986; Emons, 1987; Lloyd et al. 1987; Ridge, 1988, 1990b, 1993). Emons (1987) showed for Equisetum and Limnobium that microtubules are oriented parallel to the long axis of the hair (ie in the direction of growth) but in the tip region lie in random orientations. She also observed the fine filaments associated in parallel with microtubules (also shown by Ridge, 1988), and bundles of microfilaments (using phalloidin staining) deeper in the cytoplasm that lie more-or-less parallel to the long axis of the hair. Similar observations were made by Lloyd et al. (1987) and Ridge (1988, 1990b) for Vicia hirsuta.

Using freeze-substitution, Ridge (1988) found microtubules cortically throughout the root hair cells of Vicia hirsuta in various orientations, but was not able to deduce a helix from electron microscope observations. Microtubules were found endoplasmically between the nucleus and the tip. Bundles of microfilaments up to 7 µm long were found throughout the cytoplasm, parallel to the long axis of the hair, and in cross section were shown to be hexagonally-arranged with up to 35 filaments per bundle. These bundles were shown to be closely associated with the nucleus, confirming the immunofluorescence observations of Lloyd et al. (1987). Ridge also showed clear association of microfilaments with vesicles and organelles, and proposed that microfilaments have a 'channeling' function in the transport of coated vesicles after they bud from the plasma membrane.

In the apical dome of the hair, there have been reports of the cytoskeleton entering this region as fine filaments. Emons (1987) observed fuzzy staining of microfilaments after rhodamine-phalloidin staining in Equisetum, in comparison to thick bundles in the endoplasm proximal to the tip, and using freeze-substitution reported fine filaments in the tip region. In contrast, Lloyd et al. (1987) showed strong rhodamine-phalloidin staining of filaments in the apical dome of Vicia hirsuta root hairs. However, freeze-substitution observations of the same plant tissue found no cytoskeletal elements in the tip (Ridge, 1988). Microtubules have also been shown to enter the apical dome (fluorescence microscopy, Lloyd et al. 1987) but are much more difficult to observe compared to microtubules further down the tube of the hair. Microtubules have been shown to fountain out at the tip of the hair as they meet the base of the apical dome (Lloyd et al. 1987). In comparison, pollen tubes observed at the electron microscope level after freeze-substitution treatment, fine filaments of the actin network have been found in the tip (eg Tiwari and Polito, 1988; Lancelle and Hepler, 1992) but not microtubules.

Nuclear movement and the cytoskeleton

The role of the cytoskeleton in nuclear movement has been shown for plants and other organisms (eg Physcomitrella protonema, Doonan et al. 1986; Micrasterias, Meindl 1983; Aspergillus, Oakley and Morris, 1980; pollen tubes, Heslop-Harrison and Heslop-Harrison, 1989; root hairs, Lloyd et al. 1987). The nucleus passes through cytoplasm in which there is vigorous bi-directional traffic of organelles and other components. One of the most distinctive features of tip growth in filamentous plant cells such as root hairs is that the nucleus migrates in step with the tip as it extends (first recorded by Haberlandt, 1887). The varying extent to which the migrating nucleus follows the extending tip has been recorded many times (see reviews by Schnepf, 1982, 1986). Nuclear distance from the tip varies according to plants and conditions, but as an example, Vicia hirsuta nuclear-tip distance is 45 +/- 10 µm (Lloyd et al. 1987).

The nucleus does not necessarily follow the tip in a steady-state manner. In their very careful study, Wood and Newcomb (1989) found that if single root hairs of alfalfa (Medicago sativa) were observed during the 4-6 hour time frame after emergence, the nucleus made several migrations from the tip to the base of the hair. On maturity at ten hours, the nucleus usually remained at a mid-point between tip and base. During exposure to rhizobia, when root hairs grow distorted and form branches, Wood and Newcomb observed the nucleus to migrate in and out of side branches.

From a series of experiments using drugs that act against the cytoskeleton, Lloyd et al. (1987) provided some evidence that the cytoskeleton is involved in nuclear migration. Using rhodaminyl lysine phallotoxin (which labels actin) and antibodies to tubulin, root hairs of Vicia hirsuta were shown to contain axial bundles of F-actin and a complex microtubular system. To the basal side of the nucleus the microtubules were found to be cortical and net axially arranged, but in regions between the nucleus and the tip microtubule arrangement is more complex. Using both conventional and rapid-freeze techniques for electron microscopy, Lloyd et al. showed that bundles of microtubules are present in bundles endoplasmically (that is away from the cortex) and that these bundles progress from the nuclear region towards the apical dome of the hair where they fountain out upon the cortex. To provide experimental evidence for the involvement of the cytoskeleton in nuclear movement, Lloyd et al. applied an anti-microtubule herbicide (cremart) to the hairs, and, using time-lapse video microscopy, showed that degradation of the microtubules caused the nucleus to migrate to the base. Washing out the drug restored microtubules, nuclear position, and tip growth. The return of the nucleus was inhibited by the addition of cytochalasin-D, which fragments F-actin. It was concluded by Lloyd et al. that microtubules connect the nucleus to the tip, but that F-actin is involved in basipetal migration, as is known to occur when symbiotic bacteria (rhizobia) uncouple the nucleus from the tip during infection of legume root hairs (see Ridge, 1993).

Microtubule involvement of nuclear movement in other plant cells has been well described (see review of organelle movement by Williamson, 1993), but the results of Lloyd et al. (1987) have been contested by Williamson (1993) who justifiably complains that to maintain the view that the nucleus has microtubule-based anchorage to the tip requires the unproved assumptions that the fragmentation of microfilaments by cytochalasin stops tipward nuclear movement only by stopping cell growth, whereas cremart stops tipward movements directly by depolymerisation of microtubules, independently of its action on tip growth. Despite this criticism, it is clear that both components of the cytoskeleton (microfilaments and microtubules) are associated with the nucleus, and it is equally clear that the nucleus does not move passively with cytoplasmic streaming, as many of the organelles and smaller components of the cytoplasm appear to do, but in fact moves backward and forward independently of streaming. The nucleus may well use actin-based motility for its movement, but the regulation and control of this movement is clearly under active control.

The role of the cytoskeleton in root hair nuclear movement is thus far from resolved, but as rhizobia are known to uncouple the nucleus from the root hair tip, they may be a useful tool for studying the normal control of nuclear movement in root hairs.