Pollen tube growth and guidance: Occam's razor sharpened on a molecular arabinogalactan glycoprotein Rosetta Stone

Summary

Occam's Razor suggests a new model of pollen tube tip growth based on a novel Hechtian oscillator that integrates a periplasmic arabinogalactan glycoprotein–calcium (AGP-Ca²⁺) capacitor with tip-localized AGPs as the source of tip-focussed cytosolic Ca²⁺oscillations: Hechtian adhesion between the plasma membrane and the cell wall of the growing tip acts as a piconewton force transducer that couples the internal stress of a rapidly growing wall to the plasma membrane. Such Hechtian transduction opens stretch-activated Ca²⁺ channels and activates H⁺-ATPase proton pump efflux that dissociates periplasmic AGP-Ca²⁺ resulting in a Ca²⁺ influx that activates exocytosis of wall precursors. Thus, a highly simplified pectic primary cell wall regulates its own synthesis by a Hechtian growth oscillator that regulates overall tip growth. By analogy with the three cryptic inscriptions of the classical Rosetta Stone, the Hechtian Hypothesis translates classical AGP function as a Ca²⁺ capacitor, pollen tube guide and wall plasticizer into a simple but widely applicable model of tip growth. Even wider ramifications of the Hechtian oscillator may implicate AGPs in osmosensing or gravisensing and other tropisms, leading us yet further towards the Holy Grail of plant growth.

Introduction

In 1682 Nehemiah Grew in ‘The anatomy of plants’ described stamens as male organs and their pollen as necessary for fruit production. Somewhat later Amici (1824) observed pollen germinating on the stigma and suggested that the pollen tube carried sperm cells to the ovule. Over 50 years ago (Mascarenhas & Machlis, 1962) a chemotropic dependence on calcium (Ca²⁺) for pollen tube growth and guidance became evident. Such growth is of particular interest as the pollen tube tip has the simplest primary cell wall consisting largely of highly methyl esterified pectic polymers and shows the fastest known tip growth rate that is generally not continuous but pulsatile (Pierson et al., 1995) or oscillatory with associated ion fluxes, notably H⁺ and Ca²⁺ (Feijo et al., 1995) of similar periodicity. However, these ion fluxes are not in phase with tip growth rates (Michard et al., 2009) so causal relationships are not obvious (Holdaway-Clarke et al., 1997; Messerli & Robinson, 2003). With Occam's Razor as a guide we view tip growth as a biological oscillator that depends on two novel components, namely classical arabinogalactan glycoprotein (AGPs) and Hechtian adhesion sites. Based on the pH-dependent reversible binding of Ca²⁺ by AGPs, the Hechtian oscillator accounts for pollen tube H⁺ and Ca²⁺ ion currents as follows: H⁺ dissociates periplasmic AGP-Ca²⁺ thus increasing free cytosolic Ca²⁺ that coordinates exocytosis of cell wall precursors. However, cytosolic Ca²⁺ also depends on opening stretch-sensitive Ca²⁺ channels by tension from the growing wall transmitted to the plasma membrane by Hechtian adhesion sites. Because the new model includes two components, notably classical AGPs and Hechtian adhesion sites (Fig. 1) not previously considered as essential to models of extension growth, we suggest that overall tip growth consisting of the earlier mentioned components defined in the glossary (Table 1) is a Hechtian oscillator. This model (Fig. 2) may also illuminate the vexed problem of cell extension dating back to Heyn's identification of cell wall plasticity and via its hormonal regulation, as the primary factor in controlling growth by cell extension.

Figure 1

Hechtian strands of plasmolysed onion epidermis. These plasmolysed cells depict the founder event. (a) Incipient plasmolysis of onion epidermal cells in 3% KNO₃. (b) Plasmolysis in 7% KNO₃. Reprinted from Hecht (1912).

Table 1. Glossary of terms

Acid growth hypothesis: auxin lowers apoplastic pH leading to wall loosening.

AGP: arabinogalactan glycoprotein (UPAC definition of proteoglycan does not include AGPs).

AGP-Ca^2+: classical AGPs bind Ca^2+ stoichiometrically.

AGP57C: (At3g45230) a classical AGP with 14 predicted Hyp-arabinogalactan glycosylation sites.

AG peptides: small arabinogalactan glycoproteins with only two or three Hyp-AG glycomodules.

Classical AGP: with a single major polypeptide domain that contains c. 12–24 Hyp-AG glycomodules.

Embryo sac: haploid megagametophyte of flowering plants.

Exocytosis: secretion involving fusion of exocytotic vesicles with plasma membrane.

Expansin: nonenzymic protein hypothetically involved in wall loosening.

Hechtian adhesion: sites connecting plasma membrane and cell wall.

Hechtian strands: thread-like extensions of the plasma membrane of plasmolysed cells.

Hyp-AG glycomodules: small hydroxyproline-arabinogalactans.

In muro: cell wall as a unique cell compartment.

Periplasm: compartment between plasma membrane and cell wall.

RG-I: rhamnogalacturonan I repeating disaccharide: 4-α-d-GalpA-(1,2)-α-l-Rhap-(1)

RG-II: rhamnogalacturonan II homogalacturonan backbone has four complex sidechains.

Stress: applied force.

Strain: deformation.

Tip-focussed Ca^2+: Ca^2+ influx with highest cytosolic concentration at pollen tube tip.

Wall elasticity: reversible deformation.

Wall plasticity: irreversible deformation.

Figure 2

The pollen tube as a Hechtian oscillator. (a) The Hechtian oscillator commemorates the contributions of Karl Hecht. Nominally the three simple terms R, L and C form the pollen tube tip growth oscillator that shows how the wall regulates its own growth by coupling H⁺ and Ca²⁺ ion currents: Turgor pressure charges the battery i.e. stretches the wall that transmits the resulting internal stress via Hechtian adhesion to the plasma membrane stretch-activated H⁺-ATPase and Ca²⁺-channels (resistance R). Together they work in parallel where H⁺ releases bound Ca²⁺ (capacitance C) to the Ca²⁺ channels hence a major source of cytosolic Ca²⁺ that activates exocytosis (inductance L) of wall precursors. Arguably these largely determine the frequency of tip growth oscillations; both auxin (Zerzour et al., 2009) and fusicoccin (Fricker et al., 1997) enhance oscillatory growth consistent with their activation of plasma membrane H⁺ ATPase. Hechtian strands are prominent features of many cells on plasmolysis including root hairs, pollen tubes, stomatal guard cells and green algae like Closterium (Domozych et al., 2003). This implies a global role for Hechtian adhesion as a stress–strain gauge that can also act as a sensor and regulator of turgor pressure. (b) A single cycle of the oscillator based essentially on Chebli & Geitmann (2007) and Hepler et al. (2013), arbitrarily divided into 14 stages beginning conveniently with rapid tip growth and transduction of wall stress. Although depicted as a simple cycle, most stages comprise critical control points with multiple inputs. Some stages remain to be defined such as cleavage of GPI-anchored AGPs via PLC activity (Dowd et al., 2006) that is possibly also stretch activated. Others such as plasma membrane (PM) ATPase activity (Certal et al., 2008) at the tip itself are surmised although activity is strongest just behind the tip at the pollen tube shank (Hepler et al., 2006). While ion fluxes are not precisely in phase with growth their synchronicity suggests a close relationship as inferred here: 1, wall thins; 2, tube extension rate is inversely proportional to wall thickness (Kroeger et al., 2008); 3, Hechtian transduction transmits wall stress (Pont-Lezica et al., 1993); 4, increases tension in plasma membrane tethered to cell wall (Hecht, 1912); 5, proton efflux via PM H⁺-ATPase (Certal et al., 2008); 6, decreases periplasmic pH; 7, dissociates periplasmic AGP-Ca²⁺ (Lamport & Varnai, 2013; Lamport et al., 2014); 8, stretch-activated Ca²⁺ channels open (Ding & Pickard, 1993; Feijo et al., 1995; Dutta & Robinson, 2004); 9, Ca²⁺ influx via open Ca²⁺ channels; 10, cytosolic Ca²⁺ increases (Miller et al., 1992); 11, activates exocytosis of wall precursors. (Camacho & Malho, 2003); 12, wall thickness at tip increases (Picton & Steer, 1983; McKenna et al., 2009); 13, PM tension decreases; 14, minimum rate of tip extension (McKenna et al., 2009). Return to step 1. 15, Turgor drives tip extension (Hill et al., 2012). Note: Cytosolic Ca²⁺ recycles via Ca²⁺-ATPase mediated efflux (Schiott et al., 2004; Frey et al., 2015) or sequestration by Golgi exocytotic vesicles.

AGPs, generally regarded as mysterious molecules (Pickard, 2013; Pereira et al., 2015) and ‘minor’ cell wall components, are associated with many aspects of cell signalling (Seifert & Roberts, 2007) and cell expansion (Ding & Zhu, 1997). However, classical AGPs per se have not been viewed as essential regulatory components of pollen tube growth. Here we propose that classical AGPs based on their molecular properties and periplasmic location (Lamport et al., 2006), make a three-fold contribution: first, as a primary source of cytosolic Ca²⁺ waves; second, as a pectic plasticizer; and third, as Ca²⁺ signposts to the ovule. These three inscriptions on an allegorical AGP molecular Rosetta Stone translate into general plant growth that depends on the remarkable chemical properties of AGPs summarized in Lamport et al. (2014) and briefly as follows:

(1) AGP amino acid composition, rich in hydroxyproline (Lamport, 1970); (2) AGP O-Hyp glycosylation by small acidic arabinogalactan polysaccharides (Lamport, 1977) whose heterogeneity is generally exaggerated; (3) specific binding of Ca²⁺ by these glycomodules (Lamport & Varnai, 2013) amounting to a typical AGP Ca²⁺ content of c. 1% w/w; (4) AGP location initially attached to the plasma membrane by a glycosylphosphatidylinositol (GPI) anchor (Oxley & Bacic, 1999), but cleaved to allow incorporation into the growing wall; and finally, (5) AGP molecular size of c. 100 kDa hence extrusion rather than simple diffusion of AGPs through the wall matrix after GPI-anchor cleavage (Lamport et al., 2006).

Several recent mathematical models predict oscillations in tip growth based on Ca²⁺-dependent vesicle recycling and tip plasticity (Zerzour et al., 2009; Hill et al., 2012). However, none include AGPs and Hechtian feedback as essential components. Our new model supplements these biophysical approaches with the evidence for a central role of AGPs in a biochemical model of tip growth as follows.

A Hechtian model of pollen tube tip growth

Briefly, the new model (Fig. 2) proposes that a viscoplastic pectic cell wall mechanically coupled to the plasma membrane by Hechtian adhesion (Hecht, 1912) transmits wall strain to the plasma membrane and thus regulates H⁺, Ca²⁺ and other ion fluxes that regulate the exocytosis of wall precursors. Together they constitute a Hechtian oscillator (Fig. 2) consistent with Pickard's pioneering work on mechanotransduction (Pont-Lezica et al., 1993). This explains how the rapidly growing cell wall of the pollen tube tip can act as a decision maker that regulates its own growth by Hechtian feedback. Such dramatic oscillatory growth (0.1–0.5 μm s⁻¹) (Hepler et al., 2013) associated with active ion fluxes predominantly Ca²⁺ (Miller et al., 1992) H⁺, and K⁺ reflect Peter Mitchell's chemiosmotic paradox (Mitchell, 1961): ‘Not only can metabolism be the cause of transport, but also transport can be the cause of metabolism’. While there has been much effort to relate these ion fluxes to extension growth the Hechtian oscillator resolves the chemiosmotic paradox of the pollen tube by assigning specific roles to the Ca²⁺ and H⁺ ion currents essential for oscillations in tip growth (Hepler et al., 2006) driven by turgor pressure: Thus an initial acceleration of tip extension followed by exocytosis of wall precursors leads to deceleration, summarized in a simplified model (Fig. 2) partly based on Fig. 3 of Holdaway-Clarke & Hepler (2003).

Proton flux

Ion fluxes drive all growth (Armstrong, 2015): protons lead the way as the source of the chemiosmotic proton motive force (Mitchell, 1961) that generates ATP via mitochondrial F₁F_o ATP synthase (Allegretti et al., 2015). However, in reverse the synthase pumps protons (Mazhab-Jafari et al., 2016). Thus plasma membrane H⁺ ATPase proton efflux is essential for maintaining the membrane potential and ion transport (Hepler et al., 2013). However, in growing pollen tubes unevenly distributed membrane H⁺-ATPases (Certal et al., 2008) result in a pronounced proton efflux at the tube shank with an apparent much smaller oscillatory influx at the tip (Feijo et al., 1999). A large proton efflux may explain the striking difference between the optimum extracellular pH of animal and plant cells (pH 7.4 and pH 5.5); this reflects the differing compositions of their extracellular matrix and their dynamic Ca²⁺ storage that is largely peripheral in plant cells but mainly intracellular in animals. The lower external pH of plants reflects the low pK_a of uronic acids enabling pH-dependent uptake and release of Ca²⁺ from AGPs as a result of H⁺ ATPase activity. That also accounts for the massive proton efflux at the tube shank and ensures Ca²⁺ release from the abundant AGPs of stigmatic tissues, hence a possible cooperative effect on the growth of other pollen tubes. (cf. Lord, 2003). The role of the less marked tip H⁺ influx (Feijo et al., 1999) is less clear as direct measurement cannot detect a much smaller H⁺ efflux into the nanometre dimensional domain of periplasmic AGP-Ca²⁺.

AGP-Ca²⁺ as a primary source of cytosolic Ca²⁺

According to the prevailing view that ignores the largely methyl esterified status of tip pectin ‘Wall binding of Ca²⁺ accounts for the extracellular influx’ (Hepler et al., 2013) via open Ca²⁺ channels. Here we correlate the tip-focussed cytosolic Ca²⁺ oscillations with the presence of tip-localized AGPs (Fig. 3a in Mollet et al., 2002) based on their pH-dependent dissociation (Lamport & Varnai, 2013) by plasma membrane H⁺-ATPase (Koji et al., 2012). The nanometre dimensions of periplasmic AGP-Ca²⁺ and its proximity to the proton source results in a significantly lower pH than the pH in muro due to rapid dissipation of the proton concentration by diffusion, dilution and buffering.

The advantage of AGP-Ca²⁺ as a source of cytosolic Ca²⁺ at the tip arises not only from its cell surface location, a prime area of signal perception, but also from the paired glucuronic acid sidechains of AGP glycomodules (Fig. 3); these increase the total Ca²⁺ binding capacity of AGPs compared with the lower binding capacity of highly methyl esterified pectin that has largely unpaired galacturonic acid residues. Because the AGP glucuronic acid pK_a is lower than that of pectin galacturonic acid (Lamport & Varnai, 2013) AGPs bind Ca²⁺ more strongly at low pH. Thus periplasmic AGPs can also act as a sink for less firmly bound, hence more easily released, Ca²⁺ from pectin in the tip wall. Finally, the location of abundant periplasmic AGPs confers a large kinetic advantage to an AGP-Ca²⁺ capacitor (Lamport & Varnai, 2013) that can readily supply the stretch-activated Ca²⁺ channels regulated by membrane tension (Dutta & Robinson, 2004) as follows.

Figure 3

Repetitive arabinogalactan glycoprotein (AGP) glycomodule. The 15-residue glycomodule structure binds Ca²⁺ as depicted earlier and corresponds to known highly expressed AGPs (Tan et al., 2010). (a) The linkage connectivity of sugars involved in the repetitive 15-residue consensus Hyp–arabinogalactan, a conserved structure based on a β-linked galactosyl trisaccharide with paired sidechains that bind Ca²⁺. A, 1–6 arabinose residues; G, 1–3 galactose mainchain residues; G_a and G_b, galactose sidechain residues; R1 and R2, rhamnose sidechain residues; GlcA, glucuronic acid sidechain residues; Hyp, hydroxyproline. (b) Three-dimensional molecular model simulating a Hyp–AG with bound Ca²⁺. Hyp–AG interferon Hyp-polysaccharide-1 (IFNHP1) with Ca²⁺ ions (green) bound by two glucuronic acid (GlcA) sidechains (red); the galactan backbone is in dark blue and sidechains in light blue. Reprinted from Lamport & Varnai (2013). See Supporting Information Movie S1 in Lamport & Varnai (2013) – molecular dynamics simulation of Hyp-AG IFNHP1 (Gal backbone in blue) showing the conformational change on Ca²⁺ (green) binding by two GlcAs (red) over 500 ns.

Hechtian adhesion and mechanotransduction

The significance of Hechtian adhesion evidenced by thread-like elastic extensions of the plasma membrane physically connecting the membrane to the cell wall of plasmolysed cells has remained obscure for more than 100 years (Fig. 1) (Hecht, 1912). Hechtian adhesion is particularly evident in rapidly growing cell suspension cultures (Lamport, 1963) and during tip growth of root hairs (Volgger et al., 2010) and pollen tubes (Fig. 4) (Parton et al., 2001). However absence of Hechtian adhesion from the pollen tube shank (Lord, 2003) confirms a significant role during tip growth, which is further emphasized by its presence during tip growth even in chlorophycean algae like Closterium (Domozych et al., 2003). Such evolutionary conservation also supports a fundamental biological role of Hechtian adhesion in regulating plant growth inferred here: stable Hechtian adhesion arises from strong molecular anchoring forces, most likely of AGPs and formins. Arguably AGP GPI lipids with an adhesion force of c. 350 piconewtons (Cross et al., 2005), supplemented by formin transmembrane domains, enable the growing wall to transmit its stress/strain status at very low piconewton levels (cf. Buer et al., 2000) to the protoplast via multiple Ca²⁺ channels and H⁺-ATPases of the plasma membrane. High sensitivity Hechtian stress transducers are thus consistent with much evidence of stretch-activated Ca²⁺ channels (Dutta & Robinson, 2004) and a Hechtian ‘stress focussing’ structure involving AGPs suggested earlier (Gens et al., 2000). Hypothetical wall-plasma membrane wall linkers proposed earlier (Pont-Lezica et al., 1993) involve specific candidates which we identify here as AGP57C (At3g45230) and formin1 AtFH1; their well-defined molecular domains interact specifically with both plasma membrane and cell wall: AGP57C, an arabinoxylan-pectin-AGP glycoconjugate (APAP1) (Tan et al., 2013) has a C-terminal sequence directing GPI-anchor addition hence attachment to the plasma membrane, while the terminal rhamnose of an AGP glycomodule is attached to the reducing end of rhamnogalacturonan I (RG-I) a wall pectic polysaccharide. Formin1 has an N-terminal signal peptide followed by a transmembrane domain and a large extracellular domain anchored to the wall (Martiniere et al., 2011) most likely involving AGP glycomodules encoded by the Hyp glycosylation motif SPSALSPS.

Figure 4

Plasmolysed pollen tube tip showing Hechtian strands. Plasmolysis of a Lilium longiflorum pollen tube loaded with the fluorescent dye FM4-64 5-min after transfer to medium containing 0.3 M sorbitol and 2 μM FM4-64. (a) Arrowhead indicates fluorescent Hechtian strands. (b) Brightfield image. Bar, 15 μm. With permission from Parton et al. (2001).

AGP57C and formin1 fulfil the criteria for bona fide crosslinks between a wall polysaccharide and the plasma membrane; at last providing tangible molecular evidence for Hechtian adhesion and its pivotal role in Hechtian signal transduction.

Exocytosis of wall precursors

Exocytosis, the final stage of the secretory pathway from Golgi to cell surface, involves actin guidance and transport of exocytotic Golgi vesicles, docking and fusing with the plasma membrane. Although treated here as a single ‘component’ of an oscillator, exocytosis involves many proteins and elevated levels of cytosolic Ca²⁺at sites of pronounced exocytosis in growing pollen tubes (Camacho & Malho, 2003). Alteration of the Ca²⁺ gradient alters the pattern of exocytosis (Ge et al., 2007) suggesting the role of cytosolic Ca²⁺ as a coordinator of exocytosis consistent with the numerous Ca²⁺-dependent membrane processes (Luckey, 2008) and greatly decreased exocytosis when the Ca²⁺ chelator chlortetracycline decreased cytosolic Ca²⁺ (Reiss & Herth, 1978). While Ca²⁺-regulated exocytosis is a prime candidate for the regulation of oscillatory pollen tube growth, paradoxically some have concluded from the apparent lack of correlation between Ca²⁺ and secretion that although exocytosis may regulate oscillatory pollen tube growth, intracellular Ca²⁺ does not regulate oscillatory exocytosis (McKenna et al., 2009). Exocytosis, as a component of a Hechtian oscillator, connects H⁺ and Ca²⁺ ion gradients with cell wall tip growth (metabolism) and is thus a classic example of the Mitchell Paradox. Other models of tip growth based on ROP GTPases for example (Yan et al., 2009) did not include the cell wall or AGPs. However, due to technical difficulties the biochemical properties of the tip cell wall have received relatively little attention even though it is a major component of growth and its raison d’être.

Cell wall plasticity

At the tip of the Lily pollen tube maximum thickness of the wall coincides with a significantly decreased rate of tip growth (Fig. 4c in McKenna et al., 2009) which precedes a more rapid expansion (McKenna et al., 2009). As the wall thins its plastic extensibility increases with concomitant acceleration of the tip growth rate; further exocytosis restores wall thickness and thus decreases a growth rate that is inversely proportional to wall thickness (Kroeger et al., 2008). Evidently, wall plasticity plays a crucial role (Heyn, 1940) in determining rheology of the pollen tube tip primary cell wall. An explanation for the plasticity of primary walls in general is complicated by their wide range in composition with differing proportions of major components that include, cellulose, pectin, xyloglucan and structural protein (Fig. 5). However, the pollen tube tip is an ideal model of extension growth because it represents a simplified primary cell wall (originally defined by Kerr & Bailey, 1934) stripped down to a ‘single’ major macromolecular pectin component. ‘Pectin is not just jelly’ (Roberts, 1990) but forms a highly ordered composite of three major pectic polysaccharide domains: a highly methylesterified linear homogalacturonan (HG); a highly methylesterified rhamnogalacturonan I (RG-I); and a complex substituted rhamnogalacturonan I (RG-II) with intermolecular borate crosslinked sidechains, unstable at low pH (Yapo, 2011). AGP α-l-arabinosyl sidechains may interact with pectin rhamnogalacturonan RG-II by competing with the terminal α-l-Gal of RG-II sidechain-A (Yapo, 2011) with concomitant disruption of its apiosyl borate ester intermolecular crosslink.

Figure 5

Comparison of cell wall compositions. Ternary graph showing a simplified cell wall composition based on their content of pectin, cellulose and Hyp-rich glycoprotein with extremes ranging from the pollen tube tip (c. 100% pectin), secondary cell walls (nominally 100% cellulose) to Chlamydomonas (c. 100% Hyp-rich glycoprotein) with the primary cell wall of higher plants and the Charophycean alga Coleochaete representing intermediate values.

Anton Heyn's great insight defined the essential role of cell wall plasticity in cell extension (Heyn, 1940). The novel idea of the cell wall as a true plastic reflected the plastics revolution and zeitgeist of the 1930s. Curiously the simple extrapolation from plastics to plasticizer has been ignored due to the lack of candidates and a general consensus requiring the cleavage of load-bearing bonds although these remain unidentified. Not surprisingly the molecular basis of plasticity has remained speculative with many competing hypotheses including: insertion of pectin polygalacturonate as a chelator of Ca²⁺ crosslinks (Proseus & Boyer, 2006; Hepler et al., 2013) combined with fluctuations in apical stiffness ascribed to pectin demethylesterification (Zerzour et al., 2009; Bidhendi & Geitmann, 2016). An alternative ‘acid growth hypothesis’ (Kutschera, 1994) formulated almost 50 years ago (Rayle & Cleland, 1970) involves proton secretion and concomitant cleavage of putative acid labile polysaccharide crosslinks similar to the expansin hypothesis (McQueen-Mason & Cosgrove, 1994). Auxin-induced proton secretion is a major tenet of the acid growth hypothesis but it also increases cytosolic Ca²⁺ (Vanneste & Friml, 2013); this is consistent with exocytosis of AGP wall plasticizers (Schopfer, 1990; Kutschera & Niklas, 2007) during pollen tube growth. Microscopically the pollen tube tip wall appears as a single pectin layer c. 100 nm in width equivalent to >100 monomolecular layers of highly methylesterified pectin intercalated with AGPs. The Yariv reagent shows AGPs concentrated at the pollen tube tip (Jauh & Lord, 1996; Mollet et al., 2002) but also appearing as rings along the pollen tube (Li et al., 1992). Classical AGPs are ideal pectic plasticizers. By analogy with synthetic plasticizers that disrupt the orderly alignment of linear polymers, intercalation of the small bead-like Hyp-arabinogalactan glycosubstituents (Hyp-AGs) (Lamport et al., 2014) likely disrupts linear pectin alignment. Indeed, there is direct experimental evidence: the Yariv reagent inhibits tip growth but with a concomitant rapid accumulation of periplasmic AGPs (Mollet et al., 2002). That arguably decreases the level of AGPs in muro and thus decreases wall plasticity. We also infer that small arabinogalactan peptides may increase wall plasticity based on their dramatic upregulation during auxin induced root cell elongation (Pacheco-Villalobos et al., 2016). Compared with the much larger classical AGPs (> 100 kDa), the higher diffusibility of small (c. 20 kDa) AG-peptides (Van den Bulck et al., 2005) presumably enables them to plasticize thicker walls than at the pollen tube tip.

Additional experimental evidence also showed that AGPs are essential to pollen tube tip growth (Seifert & Roberts, 2007) and also with more general root epidermal cell expansion (Ding & Zhu, 1997). Indeed, double knockouts of pollen-specific agp6 and agp11 yielded pollen grains highly defective in germination and growth (Coimbra et al., 2009). Anecdotal evidence also correlates the friability of cell suspension cultures with enhanced AGP secretion.

Interestingly, when the Yariv reagent inhibits normal growth and tip extension ceases the pollen tube does not rupture, presumably because further additions such as callose thicken the tip wall (Jauh & Lord, 1996).

Fig. 2 combines wall biomechanics and biochemistry by integrating AGPs and Hechtian adhesion sites as essential components of a Hechtian oscillator. These components physically connect the wall with stretch-sensitive components of the plasma membrane that control the Ca²⁺influx essential for coordinating the exocytosis of wall precursors. Thus, the wall regulates its own growth by Hechtian feedback.

Pollen tube guidance from stigma to embryo sac

A plethora of guidance cues that may direct pollen tube growth include: K⁺, Cl⁻, Ca²⁺ (Hepler et al., 2006), glycoproteins (Sommer-Knudsen et al., 1998), reactive oxygen species (ROS) (Foreman et al., 2003), nitric oxide (Prado et al., 2016), peptides (Qu et al., 2015) and complex signalling networks that remain to be elucidated (Leydon et al., 2015). However other guidance cues can now be considered from the perspective of the female reproductive tract where AGP-rich regions visualized by anti-AGP monoclonals (Coimbra & Salema, 1997; Coimbra & Duarte, 2003) coincide with the pathway traversed from stigma to the egg cell. Indeed, the remarkable coincidence of AGPs and Ca²⁺ throughout the female reproductive tract (Coimbra & Duarte, 2003) is hardly fortuitous. Faced with competing hypotheses Occam's Razor suggests a simple Ca²⁺ guidance cue (Fig. 6): pollen tubes grown in vitro acidify their growth medium (Feijo et al., 1995); therefore in vivo they presumably dissociate AGP-Ca²⁺ of the transmitting tissue thus enabling pollen tubes to blaze a Ca²⁺ trail to the ovule. A dual source of cytosolic Ca²⁺seems likely. First, at the tip derived from its periplasmic AGP-Ca²⁺ capacitor that involves recycling or ‘reflux’ of cytosolic Ca²⁺. Second, Ca²⁺ released from surrounding tissues by the marked proton efflux at the pollen tube shank rather than at the tip (Hepler et al., 2006). Due to the cooperative effect of multiple pollen tubes (Heslop-Harrison et al., 1985) enhanced H⁺ efflux and release of Ca²⁺ from the locally abundant AGP-Ca²⁺ could contribute to reproductive success by ensuring fertilization of multiple ovules.

Figure 6

Pollen tube pathway to the embryo sac. Redrawn with permission from Qu et al. (2015) to show pollen tubes (yellow) growing through transmitting tissue lined with AGP-Ca²⁺ (red ellipsoids with green Ca²⁺). Pollen tubes signal to the transmitting tissue to supplement their endogenous Ca²⁺ levels by analogy with marathon runners who signal for a water bottle at stations along the track.

Evidence for the co-localization of AGPs and Ca²⁺ at each stage of the pollen tube pathway follows.

Stigmatic tissue

Calcium antimonate histochemical detection of abundant Ca²⁺ in stigmatic tissue of several species (Ge et al., 2007) parallels AGPs detected by the β-d-Yariv reagent and also by monoclonal antibodies JIM8 and JIM13. Significantly, in apple blossom, stigmatic receptivity was acquired concomitantly with the secretion of AGPs (Losada & Herrero, 2012).

Stylar tissue

Classic work (Mascarenhas & Machlis, 1962) showing Ca²⁺-dependent pollen tube growth and an elevated calcium content of tissues from stigma and transmitting tract of the style to the ovule has been amply confirmed (Knox et al., 1976), including Ca²⁺ chemotropism (Malho & Trewavas, 1996). As the distribution of AGPs and Ca²⁺ in the style coincide, the role of AGPs as a primary source of cytosolic Ca²⁺ seems likely.

Ovule micropyle: synergids and filiform apparatus

On its final path towards the egg cell the pollen tube makes a sharp turn from the transmitting tissue toward the micropyle (Coimbra & Duarte, 2003) (Fig. 6) guided by signals from the synergid cells and filiform apparatus both strongly expressing AGPs (Coimbra & Salema, 1997) and Ca²⁺ (Chaubal & Reger, 1990) with highest levels of Ca²⁺ in the synergid filiform apparatus (Ge et al., 2007).

Tip growth is a fine balance between cell wall thickening at high levels of Ca²⁺ (Picton & Steer, 1983) and tip bursting that occurs in low Ca²⁺ (Hill et al., 2012) and increased auxin levels (Zerzour et al., 2009). However, tip bursting also occurs at low oxygen levels physiologically relevant to the hypoxic ovary where failure of tip wall integrity (Linskens & Schrauwen, 1966) releases gametes that fertilize the egg cell.

Significantly AGP expression is absent from the rudimentary ovules of male flowers of Actinidia deliciosa (Coimbra & Duarte, 2003) while laser ablation showed that only a single synergid cell in the isolated embryo sac of Torenia fornieri was necessary and sufficient for pollen tube attraction to the female gametophyte (Higashiyama et al., 2001). However, additional signals likely involve an AGP sidechain fragment designated AMOR or ‘Activation Molecule for Response’ by the pollen tube to LURE guidance peptides from the synergids (Okuda et al., 2009). However, AMOR was identified as methyl-O-glucuronosyl-β-d-galactose (Mizukami et al., 2016) that is a key component of the AGP Ca²⁺-binding motif and thus suggests its possible role as a Ca²⁺ carrier. This complements the invariable coincidence of AGP and Ca²⁺. Clearly Occam is a guide not a guarantor!

Future research pathways guided by Occam and Darwin

In 1799 Napoleon's troops entered the Egyptian village of Rosetta (Rashid) and discovered an ancient basalt slab with a trilingual inscription in praise of King Ptolemy V (205–180 bc) that finally enabled Champollion to decipher Egyptian hieroglyphics in 1822. Rosetta is a metaphor for AGPs whose primary function remained unknown for 50 years until their structural hieroglyphics (Fig. 3) were deciphered to reveal the molecular function as an AGP-Ca²⁺ capacitor at the cell surface. This leads to a unified role for tip-localized AGPs and tip-focussed Ca²⁺ in cell extension as proposed here. However, viewed as a true plastic the control of wall plasticity at the molecular level in particular appears to be an intractable problem with widely differing views and assumptions. Nevertheless, the properties of AGPs combined with Hechtian adhesion offer a solution based on a Hechtian oscillator that generates cytosolic Ca²⁺ oscillations. This role for classical AGPs depends on their precise cell surface location and their unique chemistry: an exquisitely designed glycomotif with paired glucuronic acid residues that bind Ca²⁺ but released on demand. Such elegance would have pleased Paley but puzzled Darwin. Many questions remain. They include evolutionary origins (Verret et al., 2010). A vital clue lies embedded in the chalk cliffs of the South Downs National Park built over 200 million years of fossilized phytoplankton; the calcified cell walls of coccolithophores, typically Emiliania huxleyi, may be the evolutionary precursor to dynamic Ca²⁺ storage at the cell surface of higher plants possibly with even wider implications of the Hechtian oscillator as an osmosensor (Haswell & Verslues, 2015) or a gravisensor (Schnabl, 2002) both in search of their identity as nanoscale molecular devices.

Finally, the reason for the wide range in classical AGP structure includes the complex glycosylation of its protein core. Clearly the tiny minority of AGPs characterized biochemically only hints at their true diversity and versatility emphasized by recent bioinformatics that show higher plants have heavily invested in AGPs (Ma et al., 2017). Thus, diverse stimuli might generate specific Ca²⁺ signatures (Rudd & Franklin-Tong, 2001) based on the distribution, size and composition of AGP glycomodules. Because AGP glycosylation is only indirectly coded by the genome, precise structural oligosaccharide details cannot be predicted by bioinformatics. Hence the technical problem of rapid polysaccharide/oligosaccharide structural analysis and the determination of Ca²⁺ binding constants. Future ab initio computer simulations (cf. Fig. 3) will enable the design of novel AGP glycomodules with properties optimized for a given environment in the perpetual quest for the Holy Grail of plant growth.

Acknowledgements

Dedicated to the memory of Inez Joann Lamport (1938–2017) whose inspiration and support over many years made this work possible. Special thanks to Dean Jonathan Bacon, former Dean Timothy Flowers and Dr Steve Pearce, University of Sussex (UK), School of Life Sciences, for past laboratory facilities. Pembroke College, University of Cambridge (UK), academic home of D.T.A.L. (1955–1961) and also William Turner (first Herbal published in English) and Nehemiah Grew (Father of plant anatomy).

Author contributions

D.T.A.L. is the corresponding author; L.T.'s structural work made this paper possible and contributed to paper preparation; M.A.H. contributed to earlier structural work and paper preparation; M.J.K. has contributed many years work to this project and has provided invaluable advice.