G-actin is a highly conserved protein that is encoded by eight functional genes in plants. They can be classified, based on their sequence and temporal and spatial expression pattern, into two major groups: reproductive (ACT1, ACT3, ACT4, ACT11, and ACT12) and vegetative (ACT2, ACT7, and ACT8). The first class is expressed in pollen, ovules, and seeds while the second class is expressed in vegetative tissues, including some of the floral organs (e.g., petals, sepals, and carpel) and germinating seeds. All vegetative actins were observed to be strongly expressed in young roots (McDowell et al., 1996; Reviewed in Meagher et al., 1999, 2000; Šlajcherová et al., 2012). Of the vegetative actins, ACT7 seems most involved in shaping primary root growth, because act7 mutants have severe root elongation phenotypes and wavier roots (Gilliland et al., 2003; Kandasamy et al., 2009). At the tissue level, act7 had a reduction in cell number in the proximal meristem zone and an increase in cell number in the transition zone (Takatsuka et al., 2018). Actin organization analysis in act7 showed a delayed onset of longitudinal actin bundling at the boundary between the elongation and transition zones as observed in wild-type plants. This points to ACT7 as the main player in the second rapid cell elongation (Table 2; Kandasamy et al., 2009; Takatsuka et al., 2018). The other two vegetative actin isoforms (ACT2 and ACT8) seem to have a more relevant role in root hair formation and only show root elongation phenotypes in dominant-negative mutants (act2-2D and frz1, respectively; Nishimura et al., 2003; Kandasamy et al., 2009; Kato et al., 2010). Only when both isoforms are knocked out, the phenotype of act7 mutants could be mimicked, suggesting a secondary role of ACT2 and ACT8 in root cell elongation (Table 2) (Takatsuka et al., 2018). Consistent with phenotypical observations, act7-4 and not act2 act8 mutants show decreased vacuolar fragmentation in late meristematic epidermal cells, consistent with actin stabilization treatment by Jasplakinolide. This failure to regulate proper vacuolar constraining probably contributes to a delay in the initiation of fast cell elongation (Scheuring et al., 2016). The wavier root phenotypes observed in act7-4 and der1-2 (carrying an ACT2 point mutation) mutants correlate with the higher presence of oblique cell walls, also pointing to a role of these two isoforms in cytokinesis (Gilliland et al., 2003; Vaškebová et al., 2018).

Auxin is the most studied and important plant hormone in relation to the regulation of the actin network in the context of main root growth. Changes in the actin network have a strong effect on polar auxin transport. Conversely, auxin has a significant effect on the actin cytoskeleton itself (Reviewed in Zhu and Geisler, 2015; Zhu et al., 2016; Sahi et al., 2018; Zou et al., 2019; Arieti and Staiger, 2020). The actin cytoskeleton response to auxin has been predominantly analyzed in the root epidermis transition and elongation zones, as this hormone is strongly involved in root cell elongation during root growth and tropic responses. Actin inhibitor studies have shown that Cytochalasin D or Latrunculin B treatment alters the plasma membrane distribution of the auxin efflux carriers PIN1 and PIN3, which are key regulators of tropic growth (Geldner et al., 2001; Friml et al., 2002). Conversely, auxin treatment is known to inhibit root growth and cell elongation (Table 1; Evans et al., 1994; Rahman et al., 2007; Fendrych et al., 2018). Initial studies with the native auxin Indole-3-Acetic-Acid (IAA) reported contradicting effects on actin filament bundling (Rahman et al., 2007; Nick et al., 2009; Lanza et al., 2012; Takahashi et al., 2017). The differences between these studies suggest that precise concentration, duration, form of auxin, and root cell developmental status appear to be an important driver of actin changes. However, detailed quantitative analysis of IAA treatment reported a significative increase in density, unbundling, and higher organization of actin filaments in elongating cells of the root, which is generally maintained even after 60 min of treatment (Figure 1; Table 1; Lanza et al., 2012; Arieti and Staiger, 2020). These changes in actin organization, induced by IAA treatment, are consistent with the general trend that less bundled actin filaments are associated with less cell elongation. The synthetic auxin 1-Naphtalene-Acetic-Acid (NAA) was suggested to inhibit root growth through the actin-mediated increase of vacuolar constriction in late meristematic cells. Consistent with this hypothesis, vacuoles of plants pre-treated with actin-disrupting drugs show less sensitivity to external auxin (Table 1; Scheuring et al., 2016). The auxin analog 2,4-Dichlorophenoxyacetic acid (2,4-D) affects root growth through the inhibition of both cell division and cell elongation (Table 1; Rahman et al., 2007). Unlike IAA, 2,4-D effects mimic those of the actin-depolymerizing drug Latrunculin B (Figure 1). 2,4-D treatment disturbs auxin signaling via a different ubiquitin ligase complex than IAA, the SMAP1-SCF^TIR1 complex, thereby affecting post-translational modifications of actin (Rodríguez-Serrano et al., 2014; Takahashi et al., 2017). Each auxin analog employed in the above-mentioned studies differentially affects auxin signaling, which allows us to pinpoint the molecular mechanism of auxin-dependent actin responses (Ma and Robert, 2014).

Auxin can also affect the actin cytoskeleton via the activation of transcription: promoter activity of the vegetative actin ACT7 and transcription of several subunits of the actin nucleator ARP2/3 complex are enhanced by auxin treatment (Kandasamy et al., 2001; García-González et al., 2020). A feedback loop exists between auxin signaling and the actin cytoskeleton: treatment with the auxin transport inhibitor naphthylphthalamic acid (NPA) inhibits root growth by reducing cell division rate and induces a reduction in actin filament density (Rahman et al., 2007). In hypocotyls, NPA treatment leads to an increase in actin filament bundling and a decrease in actin bundle density in hypocotyl cells, without disturbing long actin filaments. Plants lacking the actin isoform ACT7 (act7-4) are insensitive to NPA treatment (Zhu et al., 2016). Moreover, actin seems to be required for the correct expression, localization, and trafficking of auxin carriers. act7-4 mutants have reduced expression of ABCB transporters, altered localization of PIN1 and PIN2, and increased endosomal retention of auxin transporters (Zhu et al., 2016). The wavy phenotype of act2-5 lines is connected to the delocalization from the plasma membrane of PIN2 (Lanza et al., 2012). When treated with Latrunculin B, PIN2 accumulates into intracellular vesicles and disappears largely from the plasma membrane (Kleine-Vehn et al., 2009; Glanc et al., 2019). These results indicate the importance of the actin network on the trafficking of hormone transporters. Other actin-binding protein mutants, such as the rice class II formin RMD and VILLIN4, show reduced polar auxin transport in roots (Li et al., 2014; Zou et al., 2019). Furthermore, the expression of the above-mentioned RMD gene in rice is facilitated by the action of the TIR/AFB1 pathway-dependent OsARF23 and OsARF24 (Li et al., 2014). Further work revealed that the auxin influx carrier AUX1 was necessary to mediate the actin filament response triggered by auxin treatment in roots (Arieti and Staiger, 2020).

It appears that auxin and actin are very much interlinked via the control of polar auxin transport by actin dynamics, or via the direct or transcriptional effects of auxin on the same actin dynamics. Exactly how these two important players feedback on each other and create coordinated development and tropisms is still a topic with many questions. Especially interesting is the mechanistic connection between the intracellular transport of vesicles containing auxin transporters and the status of the actin network, or the way in which auxin sensing is quickly translated into changes in the actin network.

Besides auxin, other hormones have been connected to actin organization during primary root growth, although they have been studied with less depth. For instance, brassinosteroid treatment induces a wavy phenotype that correlates with an increase of finer, shorter and more dynamic actin filaments in epidermal cells (Lanza et al., 2012). The observed phenotypes are comparable to those resulting from ACT2 mutation. Furthermore, both brassinosteroid treatment and act2-5 mutation result in PIN2 delocalization in the root epidermis. This suggests that brassinosteroids, auxin, and actin have a shared pathway regulating root waving (Table 1; Lanza et al., 2012).

Cytokinin treatment decreases root growth via a reduction in meristem size (Ioio et al., 2007; Růzǐčka et al., 2009; Takatsuka et al., 2018). Cytokinin has a negative effect on auxin signaling through its inhibition in the transition zone (Figure 1; Reviewed in Vanstraelen and Benková, 2012; Takatsuka and Umeda, 2014; Wybouw and De Rybel, 2019). The response to cytokinin is mediated through the modulation of the actin cytoskeleton at the boundary between the transition and elongation zone, as observed by an increase in actin bundling and a decrease in actin filament density after treatment (Takatsuka et al., 2018). Also, an increase in cytokinin was shown to disrupt the longitudinal rearrangement of actin filaments in roots, in agreement with the tendency of longitudinally organized actin bundles to induce root growth (Kushwah et al., 2011).

The actin cytoskeleton seems to be a nexus in mediating the complex responses to auxin during root growth. Generally, auxin tends to induce more longitudinally organized and less bundled filaments, while cytokinin tends to induce an increase in bundling and less organized actin filaments. Current models suggest that there is a cytokinin-dependent generation of an auxin minimum starting at the transition zone (Reviewed in Rutten and Tusscher, 2019). This hormone gradient correlates with the start of cell elongation and the increase in actin organization toward a more bundled and longitudinally oriented network in the elongation zone. Failure to generate the proper hormone gradients results in growth inhibition. Exactly how hormone gradients lead to changes in actin filament organization is not very clear. In the case of auxin and cytokinin it is likely auxin, which has the terminal effect, while cytokinin controls the auxin gradient. This notion could be tested by tissue-specific inhibition of cytokinin signaling to disturb the auxin gradient. However, only the systematic and comparative analysis of actin organization in specific cell types in response to the manipulation of hormonal pathways will unveil the existing interdependence in cytoskeletal control. Furthermore, the involvement of specific actin-binding proteins downstream of hormone signaling would contribute to understanding which actin organization and dynamics modifications are more relevant. Auxin and cytokinin are also very important for the regulation of root tropisms. This is the stage for another interaction between hormones and actin, which we will discuss next.

Primary root directional growth is a multistep process that goes from gravity sensing to signal transduction and response execution. The gravity vector is sensed by sedimenting statoliths which results in the relocalization of PIN auxin transporters and a differential auxin gradient, causing differential growth of cells between top and bottom of the root tip (Reviewed in Su et al., 2017). Statoliths are amyloplasts functioning in gravity perception and sediment in the central columella cells when the root orientation changes. This sedimentation has been recognized as the primary process controlling gravity perception (Sack, 1997; Reviewed in Nakamura et al., 2019). A fine network of F-actin surrounds statoliths in central columella cells. The results from actin cytoskeleton inhibitors and knockout mutants of the actin network components show that F-actin is involved in the regulation of statolith repositioning (Figure 1; White and Sack, 1990; Baluška and Hasenstein, 1997; Volkmann et al., 1999; Voigt et al., 2005; Reviewed in Blancaflor, 2013). Actin disruption by Latrunculin B or Cytochalasin D and the inhibition of myosin activity by BDM enhance root tip bending to gravity, while the actin-stabilizers Jasplakinolide or phalloidin inhibit the gravitropic response (Hou et al., 2003; Li et al., 2005; Mancuso et al., 2006). Latrunculin B treatment results in faster statolith relocalization which indicates that the fine actin network fine-tunes statolith movement and repositioning after gravistimulus (Hou et al., 2003; Zheng et al., 2015). To date, the only actin-binding protein that has been connected to statolith sedimentation and PIN relocalization during the root gravitropic response is the ARP2/3 complex (Reboulet et al., 2010; Zheng et al., 2015; Zou et al., 2016). Knock-out of the ARP2/3 subunit ARP3 leads to a slower gravitropic response which is associated with reduced statolith sedimentation and an increase in actin bundling around statoliths (Reboulet et al., 2010; Zheng et al., 2015; Zou et al., 2016). The E3 ubiquitin ligase SGR9 mediates at least part of the interaction between the actin network and statoliths. Plants lacking SGR9 show impaired statolith sedimentation resulting in a reduced gravitropic response, accompanied by an abnormal actin filament network around statoliths. This phenotype could be rescued by the application of the actin-depolymerizing drug Latrunculin B, suggesting a role of SGR9 in releasing statoliths from the actin cytoskeleton through its interactors (Nakamura et al., 2011). These experiments studied the statoliths in the endodermis of the inflorescence stem and not in roots. However, the mechanism of gravity perception between roots and shoots is likely to be very comparable at the statolith level. Further work should elucidate whether the same mechanism exists in roots. Additionally, it would be interesting to know which actin-binding proteins interact with SGR9. Experiments with decapped roots, laser ablation of columella cells, starch-less mutants, or maintaining the root cap at a specific angle suggest an alternative gravisensing mechanism at the distal elongation zone, out of the root cap (Wolverton et al., 2002, 2011; Mancuso et al., 2006). Our combined knowledge points to a significant role of actin in gravity perception via the mediation of statolith sedimentation. However, the actin cytoskeleton has also been proposed to mediate PIN3 relocalization to the lateral plasma membrane in columella cells upon gravistimulation. PIN3 clathrin-mediated endocytosis is naturally increased when seedlings are re-oriented, a phenomenon that can be enhanced by Latrunculin B treatment (Friml et al., 2002; Kleine-Vehn et al., 2010). This represents a connection between actin and gravity signal transduction through the regulation of the asymmetric redistribution of auxin. It is likely that actin is involved in the delivery of endocytic cargoes containing PIN3 to the plasma membrane. Further downstream, actin bundles could mediate the asymmetric growth of the root meristem during the gravitropic response. Although a change toward more oblique oriented microfilaments has been observed in stele cells upon gravistimulation, no clear correlation between actin organization in epidermal cells and root asymmetric growth has been determined (Blancaflor, 2002; Pozhvanov et al., 2013). Evidence in support of a direct role of actin in the asymmetric cell growth of the root tip is the observations that vesicle trafficking is defective in root epidermal cells of various actin and actin-binding protein mutants (Li et al., 2014; Mao et al., 2016; Zou et al., 2016; Oulehlová et al., 2019). The process of root-negative phototropism, where the root tip bends away from a blue or red light source, is also partly regulated via actin, since it is regulated via the same auxin transporters as gravitropism (Wan et al., 2012; Zhang et al., 2013b). In summary, actin has a role in the sedimentation and repositioning of statoliths which determines the strength of gravity perception. It is not yet clear to what extent the F-actin network is responsible for relaying the mechanical signal of sedimenting statoliths to a differential PIN distribution and gravitropic growth.

The processes of gravitropism and root growth in general are heavily influenced by light perception (Silva-Navas et al., 2016). When a seedling is exposed to light, hypocotyl growth is reduced, cotyledons open and make chlorophyll, while root length growth strongly increases. Seedlings growing fully in darkness have a strongly reduced root growth (Sassi et al., 2012). Actin filaments in elongating root cells are normally vertically oriented and bundled. However, in darkness, actin oriented seemingly randomly, with less bundling, and less alignment (Dyachok et al., 2011). Furthermore, it has been shown that PRF1 is important for root growth and development during photomorphogenesis (McKinney et al., 2001). Even though roots normally grow covered from light in the soil, Arabidopsis roots express all types of common photoreceptors in plants, including the Far-Red light sensor PHYA (Reviewed in van Gelderen et al., 2018). Seedling roots grown in Far-Red light have a bundled actin organization, mostly resembling growth in white light, but in a phyA mutant, actin organization was more resembling that of dark-grown seedlings (Dyachok et al., 2011). This is an example of a direct link between light perception and actin organization in roots. However, most studies focus on the effects of light quality and quantity in above ground tissues, such as the hypocotyl epidermis (Henty et al., 2011) or stomatal guard cells (Iwabuchi et al., 2010). Current standard growth conditions keep roots exposed to light, but roots have not evolved to grow exposed to light and therefore efforts have been made in the past years to grow Arabidopsis roots in the dark, while keeping the shoot in light. A notable example of this is the D-root system, which is simply a dark cover around a square agar plate to shield the root from light, with an added insert at the shoot-root boundary (Silva-Navas et al., 2015). This system showed that covering the roots leads to increased root growth and less free radical production compared to the roots exposed to light (Silva-Navas et al., 2016). Using an alternative root shading method to D-root it was demonstrated that components of the SCAR/WAVE-ARP2/3 system BRK1 and SCAR1 are involved in regulating root growth in covered roots (Dyachok et al., 2011). The subcellular localization of BRK1-YFP changes from occurring at three-way junctions of cells to cytoplasmic upon darkness. SCAR1 is reduced in protein extracts from both soluble and membrane extract, and the central light regulator COP1 is involved in SCAR1 degradation (Dyachok et al., 2011). Therefore, it seems clear that actin has a significant role to play in the regulation of root growth by light. However, very little work has been done until now to further explore this connection. It is also of especial importance to do this in systems such as D-root to simulate the light conditions of the soil in order to obtain results that are relevant for plant growth in field or natural conditions.