Distribution of pollen tube cell wall polymers was investigated by using mostly cytochemical reagents, enzymes and/or antibodies (Table 1). In most of the immunolocalization studies, monoclonal antibodies (mAbs) are applied on the whole pollen tube [24,25,26,27,28,29,30,31,32], allowing a cell surface labeling that may mislead in the interpretation as epitopes may have been masked by other polymers. To avoid this problem, enzyme treatments were sometimes applied on fixed tubes or pollen tubes were embedded in resin and sectioned [33]. Another possible artifact is caused by the slow penetration of the chemical fixative that arrests pollen tube growth while exocytosis is still ongoing, or the other way round, that arrests of enzymatic reactions in the wall while the tube is still growing.

Distribution of cell wall polymers in pollen tubes was investigated in plant species of the angiosperm eudicots such as A. thaliana [23,24,25,26], Nicotiana tabacum [21,22,29], Solanum chacoense [30], Camellia japonica [31], Torenia fournieri [32] and Actinidia deliciosa [33] as well as angiosperm monocots like Lilium longiflorum [48,49] and Zea mays [50]. More recent studies have focused on the gymnosperms like Pinus sylvestris [27], Picea meyeri [51] and several others [28,52].

One of the main differences between the primary cell wall of somatic and pollen tube cells is despite its presence, the low abundance of cellulose, a β-(1→4)-glucan. Instead, callose, a β-(1→3)-glucan, is the main cell wall polysaccharide. Callose is mostly localized in the inner cell wall and is absent from the tube tip of most angiosperm pollen tubes [53,54] including A. thaliana (Figure 2c) [25]. In contrast, in Cycas and several Pinus species, callose was not detected in the pollen tube wall by aniline blue staining [28]. However, in P. sylvestris, a strong fluorescence was detected with aniline blue at the tube tip [27]. In addition, callose is deposited at periodic intervals to form callose plugs that maintain the tube cell in the apical expanding region of angiosperm pollen tubes (Figure 2c) and separates the viable from the degenerating region of the tube [12]. In tobacco using CBH-I-gold and LAMP MAb [22] and in A. thaliana using CBM3a [26] and LAMP MAb [25], callose and cellulose are localized in the plugs and in the inner cell wall of the pollen tube. In contrast, in long-living and slow-growing gymnosperm pollen tubes, the callose plug deposition is apparently not a consistent feature as well as the permanent callose wall in pollen tubes [52]. This observation suggests that in the evolution of flowering plants a permanent callose wall and callose plug deposition appeared through the short-lived and fast-growing pollen tubes of angiosperms [55]. Cellulose is present in the entire cell wall of pollen tubes with a weaker staining with calcofluor white or CBH-I-gold labeling at the tip in most angiosperms [22,25,27,30]. Recently, Chebli et al. [26] using CBM3a that binds to crystalline cellulose found two populations of A. thaliana pollen tubes displaying intense or weak labeling at their tips suggesting that the labeling was related to temporal growth rate. Similarly, in T. fournieri, the distribution of cellulose changed temporally and spatially, being abundant in freshly germinated pollen tubes, and absent from the tube tip of older pollen tubes [32]. Using atomic force microscopy, Wu et al. [32] also revealed that cellulose microfibrils were more abundant in the apical region of slow-growing tubes whereas only few microfibrils were observed in the auxin-stimulated pollen tubes suggesting a correlation between cellulose abundance and pollen tube growth. In Conifers, pollen tubes have cellulose in the entire cell wall and may contain more cellulose deposited at the tube tip, which may explain the slower growth [52,56].

Pectins are complex polymers consisting of homogalacturonan (HG), which can be methyl- and acetyl-esterified, rhamnogalacturonan-I (RG-I), rhamnogalacturonan-II (RG-II), and xylogalacturonan [57]. HG is a polymer of repeated units of (1→4)-α-d-galacturonic acid. Upon block wise action of pectin methylesterases on methylesterified HG, several parallel HG chains can be cross-linked via the interaction of the negative charges of the carboxyl groups of galacturonic acid and calcium [58]. RG-II has a galacturonan backbone with four well defined oligosaccharides composed of unusual sugars such as apiose, aceric acid and 3-deoxy-d-manno-2-octulosonic acid (Kdo) of unknown function [57]. RG-II is present in the primary cell wall of all higher plants and its structure is evolutionarily conserved in the plant kingdom. In the cell wall, it exists predominantly in the form of a dimer that is cross-linked by a borate di-ester between two apiosyl residues [57]. On the other hand, RG-I consists of the repeating disaccharide unit, (1→4)-α-d-galacturonic acid-(1→2)-α-l-rhamnose. On the rhamnosyl residues, a wide variety of side chains can be detected ranging from monomers to large oligosaccharides such as (1→4)-β-d-galactan, (1→5)-α-l-arabinan, and/or type I arabinogalactan [57].

Most of the studies on the pectic wall of the pollen tube focused on the detection of HG domains using the JIM5 and JIM7 mAbs (Table 1). To date, localization data of other pectic domains such as xylogalacturonan and RG-I side chains in the pollen tube wall are scarce. Immunofluorescence labeling of weakly and highly methylesterified HG with JIM5 and JIM7 or the newly introduced LM19 and LM20 depends clearly on the species, the in vitro growth condition and the growth rate of the pollen tube [59]. It is however generally observed that in most angiosperms, highly methylesterified HG is mostly detected at the tip (Figure 2d) and weakly methylesterified HG epitopes are preferentially detected either in the entire pollen tube or in the pollen tube except in the tip region as shown in the Brassicacae (A. thaliana) [25], Solanaceae (wild potato, tobacco, petunia), Oleaceae (jasmine), Poaceae (corn) [60], Scrophulariaceae (T. fournieri) [32] and Liliaceae (lily) [48,49]. Similar observations were obtained in Picea meyeri [51] and Picea wilsonii [61] pollen tubes. In other gymnosperm species like P. macrophyllus and P. banksiana, pollen tubes were not labeled with JIM5 and JIM7 using electron microscopy observation [28].

Based on Elisa assays and dot-blots, it was clearly shown that JIM5 and JIM7 as well as LM19 and LM20 have specific but also overlapping binding capabilities to different degrees of methylesterification (DM) of HG domains after random or blockwise action of pectin methylesterases (PME) [37,62]. LM20 binds to HG with DM ranging between 85 and 16%, whereas LM19 binds to HG with DM between 66% and the demethylesterified form [37]. Using these data and assuming that PME has a blockwise action to allow calcium cross-linking in the shank of the tube, we can estimate that the DM of HG at the tip is probably over 70% and back from the tip, the DM is less than 15%. In lily, this change of methylesterification level occurs between 15 and 20 µm from the pole of the pollen tube corresponding exactly to the transition zone between the apical dome and the cylindrical shank of the cell [63]. In A. thaliana, this modification takes place closer to the tip (between 3 and 10 µm) [26]. Recently, propidium iodide (PI) was used to probe low methylesterified HG on living A. thaliana and L. longiflorum pollen tubes, which allowed following the dynamics of the pectin during pollen tube growth. The fluorescence of PI was detected at the tip and the shank of the pollen tubes with PI fluorescence oscillations preceding growth rate oscillations [45]. Finally, in certain species such as lily [49], Ornithogalum [59], tobacco [54] and others [60], periodic ring-like labeling patterns of weakly methylesterified HG were clearly visible along the pollen tube and originated from cell wall deposition during the slow growth pulses [54].

Side chains of RG-I are also present in the cell wall of pollen tubes. Using LM6 and/or LM13, epitopes associated with arabinan are evenly distributed along the entire pollen tube in A. thaliana [25], A. deliciosa [33] and P. wilsonii [61]. At the electron microscopy level, arabinan is mostly detected in the outer cell wall layer in A. thaliana [25] consistent with the location of other pectic motifs such as HG. Galactan is also detected in A. thaliana pollen tubes but the labeling intensity is weak [25] and almost absent in Picea [61].

Finally, RG-II, the major boron-binding component of the cell wall, is detected in the entire lily pollen tube cell wall [43] but is not in Picea [61].

Xyloglucan (XyG) is the major hemicellulosic polysaccharide of the primary cell wall of angiosperm eudicots and non-commelinid monocots [64,65]. Classic XyG consists of a (1→4)-β-d-glucan backbone substituted with xylose, galactose-xylose or fucose-galactose-xylose motifs [66]. In the primary cell wall, XyG interacts with cellulose via hydrogen bonds and participates in the control of cell expansion [67]. The presence of fucosylated and galactosylated XyG was assessed only recently in A. thaliana pollen tubes using CCRC-M1 [25,68] and LM15 [25]. Labeling with both mAbs is detected in the entire pollen tube wall and at the electron microscopy level; it was shown that XyG is present in the outer and inner layers of the pollen tube cell wall [25] suggesting a possible interaction with cellulose microfibrils.

All these data reveal a common organization of the pollen tube cell wall in the species from the angiosperm eudicots and the non-commelinid monocots. In contrast, with the slow-growing and short-traveling gymnosperm pollen tubes, it appears that the distribution of cell wall polymers, especially the callose, is altered [69]. However, more studies using the probes listed in Table 1 are required to draw a general conclusion. To date, the investigation on the distribution of RG-II in pollen tubes is very limited but with the development of new specific probes for each individual side chains, the gap might be soon overcome. Another attracting method that may develop in the near future in the study of cell wall dynamics during pollen tube growth is the use of sugar analogs compatible to click chemistry associated with in vivo cell imaging [70]. In this method, sugar analogs are metabolically incorporated into cell wall polymers and subsequently labeled with covalent linkages to fluorescent probes [71] as shown with alkynylated fucose analog incorporated in the pectin of A. thaliana root cell wall [72]. This method will allow for insight into the dynamics and recycling of cell wall polymers on living pollen tubes.

Only very scarce studies have focused on the biochemical characterization of the pollen tube cell wall. The most striking point observed in the few biochemical characterizations of the cell wall of tobacco, C. japonica, lily, tulipa and A. thaliana pollen tubes is consistently the high level of arabinosyl residues [25,73,74]. In A. thaliana, Ara represents 43%, Glc 20%, GalA 11%, Gal 8% and Rha 5% of the total sugars [25]. Linkage analyses show that most of the arabinosyl residues are T-Ara,5-linked and to a lesser extent 2,5-linked. These data indicate that the pectin of A. thaliana pollen tube cell wall is composed of 6% HG, 5% RG-I backbone harboring abundant chains of (1→5)-α-l-arabinan. This common feature suggests that arabinan may have an important role during pollen tube growth. Finally, linkage analyses also indicate (1) that most of the Glc is 3-linked indicating the abundance of callose [25,73] and (2) confirm that XyG is also present based on the detection of 4-Glc, 2-Xyl and T-Xyl, T-Gal and Fuc residues [25]. According to the one letter code proposed by Fry et al. [75], the unsubtituted glucosyl residue of the backbone is represented by the letter G. X, L and F represent the substitution by xylose, galactose-xylose and fucose-galactose-xylose, respectively. XyG can also be O-acetylated generally on the galactose, indicated by an underline letter, but the biological significance of O-acetylation is not known yet [76]. Generally, every four or five glucoses of the glucan backbone are not substituted allowing the cleavage of the polymer by endo-glucanase. The small oligosaccharides can then be analyzed by MALDI-TOF mass spectrometry, known as OLIMP method (OLIgosaccharide Mass Profiling) [77]. Using this technique, the fine structure of A. thaliana pollen tube XyG was recently determined. It revealed significant differences with the XyG from vegetative organs such as leaves (Figure 3). Whereas XXXG is the main fragment of leaf XyG [25,78], the fucosylated and O-acetylated fragments (XXFG and XLFG) are the major motifs in A. thaliana pollen tube XyG [25] suggesting an important role of these two features in pollen tube growth and/or in the interaction with the female tissues.

Used in the mid-90s to study the cell wall composition of algae [80] and later widely adopted to screen flax and A. thaliana cell wall mutants [81,82], Fourier transform infrared (FT-IR) microspectrometry associated with principal component analysis was recently used to study the effect of hormones, minerals or drugs on the cell wall composition of elongating pollen tubes [51,83]. It shows that treatments of P. meyeri pollen grains with brefeldin A, a drug able to inhibit the secretory pathway, reduce the pollen germination and tube growth by disrupting the secretory vesicles at the tip and significantly decrease the content of pectin in the apical region [84]. Treatments with nitric oxide donor or NO synthase inhibitor induce the accumulation of acidic HG and callose in the tip region [85]. Interestingly, blocking the release of intracellular calcium with drugs significantly alters the cell wall structure in Picea pollen tubes with the accumulation of callose and the disappearance of methylesterified HG at the tube tip [61]. Finally, treatment of T. fournieri pollen tubes with auxin promotes the growth by stimulating the synthesis of pectin, reducing the cellulose density and modifying the orientation of cellulose microfibrils as observed by atomic force microscopy [32]. However, whereas the vibrations associated with cellulose, phenolics, the methylester and the carboxyl groups of GalA are well assigned by FT-IR, the technique does not allow discriminating clearly other pectin motifs such as arabinan and/or galactan side chains of RG-I in the cell wall network. It is, however, a non destructive method that may be convenient to use in the future for determining the overall composition of pollen tube cell wall mutants compared to wild type. It may also be useful to follow the evolution of the cell wall composition in different portions of the pollen tube during growth.

Due to the technical challenges involved such as the labor intensive work to collect enough material, the number of studies focused on the biochemistry of pollen tube cell wall is limited. As a consequence, it does not permit to assess clearly the differences in the cell wall composition between angiosperm (monocot and eudicot) and gymnosperm. However, with the development of more sensitive equipments such as mass spectrometry, more information may be soon collected.

Pectin represents one of the main carbohydrate polymers found in the cell wall of pollen tubes. Given the complexity of its structure, it is predicted that 67 different glycosyltransferases, methyltransferases and acetyltransferases are required for its biosynthesis [86]. The synthesis of the different pectin motifs is carried out in the Golgi apparatus and they are secreted via Golgi-derived vesicles [87].

Based on the structure of XyG, its biosynthesis requires a combination of (1→4)-β-glucan synthases, (1→6)-α-xylosyltransferases, (1→2)-β-galactosyltransferases, (1→2)-α-fucosyltransferases and O-acetyltransferases [111]. In A. thaliana, CSL4 codes for a (1→4)-β-glucan synthase involved in the synthesis of the XyG backbone but transcriptomic data indicate that this gene is not expressed in pollen [112]. At least five XYLOGLUCAN XYLOSYLTRANSFERASEs (XXTs) have been identified in A. thaliana [76,113,114] but their implications during pollen tube growth have not been assessed yet. Two XYLOGLUCAN GALACTOSYLTRANSFERASEs were also characterized: MUR3 in A. thaliana [115] and more recently XLT2 in Tropaeolum majus [116]. However, none of these genes is expressed in pollen grains or pollen tubes [93]. MUR1 codes for an enzyme that catalyzes the first step of the biosynthesis of GDP-fucose and the mur1 mutants show a lack of labeling of fucosylated XyG epitopes with CCRC-M1 in hypocotyls, shoots or leaves [68]. However, the mutation does not affect the distribution of the fucosylated XyG epitopes in the pollen tube suggesting that this gene is not implicated in the XyG biosynthesis in pollen tubes. Mutation in the FUCOSYLTRANSFERASE gene (FUT1, FT1 or MUR2) eliminates the fucosyl residues on the XyG side chains in all major plant organs indicating that FUT1 is involved in most of the XyG fucosyltransferase activity in A. thaliana [117,118]. MUR1 and FUT1 are not expressed in pollen grains or pollen tubes but are strongly expressed in the stigma and in ovarian tissues in A. thaliana [93] suggesting that XyG from the female tissues may be highly fucosylated but more direct experimental evidence is required.

It is generally assumed that cellulose microfibrils are assembled by cellulose synthases (CESs) located at the plasma membrane in a form of rosette terminal complex. However, the possibility that the first step of cellulose biosynthesis begins in the Golgi apparatus in higher plants cannot be ruled out [119]. Different multigenic families are implicated in cellulose biosynthesis. Ten CESA (CELLULOSE SYNTHASE A) genes were found in the A. thaliana genome and it was shown by mutant analyses that they play distinct roles in the cellulose synthesis process [120,121]. Another multigenic family, CSLD (CELLULOSE SYNTHASE LIKE-D), related to CESA, seems also to be involved in cellulose biosynthesis and contains six members in the A. thaliana genome [122,123]. In rice, the genome contains 7 CESA and 5 CSLD genes [124]. In tobacco, NaCSLD1 (Nicotiana alata CELLULOSE SYNTHASE-LIKE D1) was found to be only expressed in the anther and in vitro-grown pollen tubes and was predicted to code for a cellulose synthase in pollen [125]. In N. tabacum pollen tubes, CESA and CSLD were detected along the entire length of the pollen tubes with a higher concentration at the apex [126]. Similarly, in A. thaliana, CESA6, CSLD1 and CSLD4 were found at the plasma membrane of the pollen tube both at the tip and in the shank [26,127]. Moreover, crystalline cellulose was also found inside cytoplasmic vesicles and the trans Golgi network in A. thaliana pollen tubes leading to the hypothesis that the synthesis of short cellulose microfibrils may initiate in this compartment giving the pollen tubes the head start in assembling the cell wall necessary to promote rapid elongation [26]. Alternatively, the vesicles containing the cellulose may originate from endocytosis suggesting that cellulose in the shank of the tube may be degraded and recycled [26].

The CESA1 and CESA3 mutations are both gametophytic lethal. 50% of pollen grains from the heterozygous cesa1+/− or cesa3+/− plants are significantly deformed and do not produce pollen tube [128]. The cesa6 null mutants display subtle growth defect and the pollen is not deformed [128]. On the other hand, the triple cesa2/6/9 mutants are sterile and pollen grains are strongly deformed [128]. The data indicates that CESA2 is functionally redundant with CESA6. As CESA9 is strongly expressed in pollen grains, these studies suggest that CESA1, 3, 6 and 9 play a critical role during pollen grain formation. Recently, it was shown using pCESA6::GFP-CESA6 construction that CESA6 is expressed in pollen tubes indicating that in addition to CESA1, 3 and 9, CESA6 may also play a role during pollen tube growth [26].

Three genes from the CSLD family are also implicated in pollen tube growth. Csld1 and csld4 homozygous mutants are sterile. Mutant pollen grains show abnormal germination and reduce pollen tube growth in vitro and in vivo. In addition, high level of pollen tubes is bursting due to a reduced cellulose content, abnormal callose deposition and thickened and highly irregular cell wall [123,127]. Finally, mutation in CSLC6 results in a strong reduction of pollen tube growth [129]. All these data support the fact that despite the low abundance of cellulose in the cell wall of pollen tubes, it has an important function in maintaining the integrity of the tube cell.

Callose is also synthesized at the plasma membrane by callose synthases (CALSs) located at the apex and the distal regions of tobacco pollen tubes. In longer pollen tubes, CALS accumulates also close to the callose plug [126]. Twelve putative CALS genes (CALS1-12) have been identified in the A. thaliana genome [130]. Among these 12 genes, CALS5 is implicated in the normal deposition and patterning of the exine pollen grain [131,132] but opposite data are obtained concerning the pollen viability and the ability of the pollen tube to grow normally. NaGSL1 (Nicotiana alata GLUCAN SYNTHASE-LIKE1) was shown to be a callose synthase [133] involved in the control of callose synthesis during pollen tube growth [134]. In 2011, a study brought evidence that CALS5 orthologues are expressed in pollen grains of many different angiosperms and gymnosperms but CALS5 was only expressed in fast-growing pollen tubes (i.e., angiosperms), suggesting that CALS5 plays different but crucial roles during pollen formation and/or germination and pollen tube growth [55]. This study also supports the reports showing that the callose wall in slow-growing pollen tubes is generally transiently detected and the callose plug has so far not been observed in gymnosperm pollen tubes. Based on such variability and transient expression of callose in the pollen of Pinus, Pacini et al. [135] suggested that the callose may not have a structural function but rather serves as a reserve polysaccharide. In contrast, in fast growing pollen tubes like in S. chacoense and Lilium orientalis, callose was shown to have an important mechanical function [53].

HG is synthesized in the Golgi apparatus and deposited in the expanding tube tip in a highly methylesterified form [159]. The methylesterified HG in the apical zone is thought to provide sufficient plasticity for sustaining the pollen tube growth [15]. Upon block-wise action of PMEs, de-methylesterified HG polymers can form multimers via ionic bonds between the negatively charged carboxyl groups of several HG and Ca²⁺ ions forming a pectate gel that may provide rigidification of the pollen tube cell wall [12,58]. Therefore, the control of the cell wall plasticity by PMEs is critical to ensure a proper fertilization. Conversely, upon random action of PMEs, the partial removal of methylester groups may allow the pectin-degrading enzymes, polygalacturonases (PGs) or pectate lyases (PLs) (see lower section) to cleave the HG backbone thus affecting the rigidity of the cell wall [58,160]. This sequential action of PME and PG was demonstrated in the quartet mutants that release pollen grains as tetrads due to the persistence of pectic polysaccharides [161]. The QUARTET1 gene encodes a PME that is required in association with the QUARTET3 encoding a PG for the degradation of the HG of the tetrad resulting in the normal pollen separation during microsporogenesis [162,163].

The A. thaliana genome contains 66 putative PMEs. Most of the genes encoding PMEs display a tissue-specific expression pattern, especially for 14 of them that are specifically expressed in pollen grains or pollen tubes (Table 3) [93,164]. Moreover, proteomics studies of germinated pollen grains from maize, wheat, rice and A. thaliana have identified two putative PMEs in germinated maize pollen [165], one in rice [143] and at least one in A. thaliana pollen tubes [166]. In tobacco pollen tubes, biochemical analyses have shown the presence of seven isoforms with a wide isoelectric point range [167]. Two PME isoforms were co-localized with methylesterified HG in the cell wall and Golgi-derived vesicles suggesting that PMEs are transported to the tip under an inactive precursor form [167].

The first functional characterization of PME genes and their critical roles in pollen tube growth were obtained from two knock-out mutants vanguard1(vgd1) [168] and atppme1 [169]. The disruption of VGD1 resulted in the burst of pollen tubes in vitro and marked retardation of the vgd1 pollen tube elongation in the pistil resulting in a strong reduction of male fertility and seed set [168]. The lack of AtPPME1 activity in knock-out mutants affects the shape and the growth rates of the pollen tubes, indicating that AtPPME1 is required for the integrity of the cell wall and for the tip-polarized growth [169]. Moreover, when PME from orange peel is exogenously applied to pollen tubes of Lilium formosanum and tobacco or when PME is overexpressed, the pollen tube growth is inhibited [170].

The main regulators of PMEs are PME inhibitors (PMEIs). Sequence analyses indicate that several PMEs contain, in addition to the catalytic domain, an N-terminal extension domain (the PRO region) showing similarity with PMEI domains [58]. Based on this, PMEs are classified in two distinct groups depending on the presence or absence of this putative PMEI domain. Group 1 PMEs do not have the PRO region, whereas PMEs from group 2 can contain from 1 to 3 PMEI domains [169]. It is hypothesized that the PRO region is cleaved during the maturation of PME, as so far, only PMEs lacking this domain were found in the cell wall [58].

Recently, Wolf et al. [164] indicated that the PRO region of the group 2 PMEs might regulate the release of the mature PMEs from the Golgi apparatus. It was also suggested that the PRO region might play an auto-inhibitory role during maturation [170]. PMEIs are not only found in the PRO region of PMEs but also exist as individual proteins sharing strong sequence similarities with invertase inhibitors (Inv inhibitors). However, there is no clear evidence of a similar function and mode of action between PMEIs/Inv inhibitors and PMEI domains of PMEs. The A. thaliana genome contains 76 genes coding highly putative PMEIs/Inv inhibitors. The expression patterns of PMEIs/Inv inhibitors are all regulated in a tissue-specific manner [171,172] and transcriptomic data reveal that PMEIs/Inv inhibitors are highly expressed in the A. thaliana pollen compared to other tissues [171,173]. As 9 out of the 14 PMEs specifically expressed in pollens are from the group 2 (with PMEI domains), it may suggest that PMEIs play a crucial role in regulating the activity of PMEs during pollen germination and pollen tube growth. Suppression of the expression of At1g10770 (coding for a putative PMEI) leads to a partial male sterility reducing the seed set by inhibition of pollen tube growth [174]. Moreover, treatments of pollen tubes with exogeneous PMEI also result in an abnormal germination and burst of pollen tubes in A. thaliana [175]. Finally, it was shown that AtPPME1 and AtPMEI2, both pollen specific, physically interact and in vitro assays revealed that AtPMEI2 was able to inactivate AtPPME1 [176]. The transient expression of AtPPME1 and AtPMEI2 in tobacco pollen tubes demonstrated that AtPPME1 accumulates uniformly from the tip to the shank of the pollen tube. In contrast, the localization of AtPMEI2 is restricted to the tip and in endosomal vesicles suggesting its internalization and recycling [176]. PMEs are thought to play an important role during the tip-polarized growth of the pollen tube by controlling the mechanistic properties of the cell wall. However, the function of PMEs is probably not restricted to the change of the mechanical properties of the cell wall. Weakly methylesterified HG have been implicated in other important physiological processes such as cell attachment in vegetative cells or organs [92,177,178] and in lily, associated with a stigma/style cysteine-rich adhesin (SCA), a secreted plant lipid transfer protein (LTP), in pollen tube adhesion [179,180]. Several LTPs are expressed in the transmitting tract of the A. thaliana pistil along the pollen tube path [181] that may be involved in this process. In tobacco, it was shown that TobLTP2 purified from the stigma exudates accumulate in the pistil and are able, as expansins do, to promote cell wall loosening. However, the pollen tube growth in LTP-silenced plants is similar to wild type plants, suggesting either that pollen tubes do not require this loosening protein or that other loosening agents may have act redundantly [182]. Recently, LTP5, produced in A. thaliana pollen tubes and in the transmitting tract, was proposed to play important roles in maintaining the cell polarity at the tube tip and adhesion-mediated guidance perhaps by interactions with pectins [183]. Direct evidence is lacking, but it appears highly probable that PMEs, by modifying the degree of methylesterification of HG, play a direct role in adhesion and guidance of the pollen tube.

Finally, one putative pectin acetylesterase was detected in rice pollen [143] and a recent report has shown that the overexpression of the Populus PECTIN-ACETYLESTERASE1 in tobacco affected the shape of the pollen grains and their abilities to germinate [184].

PL (PECTIN LYASE) and PLL (PECTIN LYASE-LIKE) genes are also abundantly expressed in tomato, tobacco, alfalfa and arabidopsis pollen [185,186,187] and several pollen allergens have PLL activities [188]. Out of the 26 PLL genes, 14 are expressed in pollen [189]. None of these 14 genes is specifically dedicated to the male gametophyte but 4 of them are strongly expressed in pollen grains (Table 4). To our knowledge, there is no experimental evidence of the implication of PLL in the remodeling of the pollen tube wall. Many promoter activities of the PLLs genes are similar to those exhibited by many POLYGALACTURONASEs (PGs) [190]. This may imply a close functional association between PLLs and PGs, particularly in the digestion of the pollen grain cell wall prior to germination and during pollen tube growth [190,191].

Glycoside hydrolases (GHs) are enzymes that catalyze the hydrolysis of glycosidic linkages between two or more glycosides or between a carbohydrate and a non-carbohydrate moiety. In 1991, GHs were classified into 35 families [192] and to date they are divided into over 100 families [193]. GHs involved in the reassembly of the pollen tube cell wall belong to the GH3, 9, 10, 17, 28, 35, 43 and 51 families. Their activities as well as their possible functions during pollen tube growth are presented in Table 5.

Endo-(1→4)-β-glucanases and endo-(1→3)-β-glucosidases are members of the GH9 and GH17 families, respectively (Table 5, Table 6). In A. thaliana, 25 genes are encoding GH9 proteins and 49 the GH17 proteins. Among the 25 putative CELLULASES, two are strongly expressed in pollen grains and three in in vitro-grown pollen tubes (Table 6). Four members of the GH17 family are expressed in in vitro-grown pollen tubes (Table 6). Three and five genes encoding the GH9 and GH17 are also found in the pistil tissues, respectively. Twenty years ago, these two enzyme activities were assayed in pistils and anthers of bean and they were linked to the cell wall disruption occurring during the release of the pollen grain from the anther and during the penetration of the pollen tube through the stigma [194]. Recently, and despite its low abundance in the pollen tube cell wall, it was shown that cellulose plays a crucial role by influencing the diameter of in vitro-grown pollen tubes [195]. As described previously, the cell wall of fast-growing pollen tubes is enriched in callose and several studies have shown that moderate β-(1→3)-glucanase treatments are able to stimulate the germination of pollen grains [53,196]. Moreover, it was reported that exo-β-glucanases might play an important role in the regulation of pollen tube elongation in lily [197]. On the contrary, treatments with inhibitors of glucosidases severely inhibited the growth of pollen tubes [198]. It was hypothesized that the mechanical properties of callose and more precisely the resistance to lateral deformation of the tube was regulated by glucanases [53] (Table 5).

Polygalacturonases (PGs) are involved in the HG degradation. They were characterized in the pollen from corn and other Poales [200] and have since been identified in various plants including rice, pea, tomato and arabidopsis [143,190,201,202]. PGs belong to the GH28 family composed of 69 genes in A. thaliana [203,204]. In 2000, Torki et al. [205] showed that 7 PG genes were strongly expressed in A. thaliana flowers. Publicly available microarray data are consistent with this result as 6 PG genes are strongly expressed in pollen tubes and one is expressed in the pistil (Table 6). Together with the strong expression of more than 10 PMEs and 4 PLLs in pollen tubes, these data suggest an important remodeling of the HG during pollen tube growth. PGs are involved, during pollen maturation of Brassica campestris, in the intine and/or exine formations [206,207] and in Turnera subulata, PGs are implicated in self-incompatibility [208]. PGs are released upon rehydration of triticale pollen grains [146] and are also detected in the tip region of pollen tubes in Brassica napus during papillar cell penetration [209] suggesting that PGs are probably involved in the loosening of the stigma cell wall during pollination. As PGs degrade weakly methylesterified HG and that the demethylesterification of methylesterified HG by PMEs at the tip of the tube is accompanied by the release of protons, it was suggested that this local change of pH in the pollen tube cell wall may promote PG activity [58,210]. This activation of PGs might then control the loosening of the pollen tube cell wall back from the tip and then promote the proper pulse growth of the tube (Table 5). Alternatively, pollen tube PGs may also affect the loosening of the stigma and transmitting tract cell wall to facilitate the penetration of the pollen tube (Table 5).

β-galactosidases (BGALs) belong to the GH35 family. BGALs can act on different substrates including arabinogalactan proteins, galactolipids, RG-I and RG-II side chains of pectin and XyG releasing galactose [211]. Eighteen genes are encoding BGALs in the A. thaliana genome and the functional genomics analysis reveals that BGAL expression levels are high in mature leaf, root, flower and silique but are low in young seedling [212]. Publicly available microarray data (Table 6) show that 2 BGALs are highly expressed in mature pollen grains and 2 others are highly expressed in the pistil confirming the results obtained by Ahn et al. [212]. Other BGAL genes are also strongly expressed during the microspore mitosis in developing pollen grains [199]. In rice, 35 genes are encoding BGALs; among them, two are strongly expressed in dry pollen grains [213] (Table 6). As mutants defective in BGAL have impaired fertility, it was hypothesized that BGAL may be involved in the pollen tube wall turnover in Brassica campestris by hydrolyzing arabinogalactan [214]. In 1995, it was suggested in tobacco that BGALs could be secreted from the pollen tube and released in the stylar transmitting tract to modify the branching pattern of arabinogalactan involved in the guidance of pollen tubes [215,216]. In tobacco, BGAL mRNAs are accumulated during the formation of pollen grains, presumably stored for future use during pollen germination and pollen tube growth as BGAL activities were detected in growing pollen tubes [217]. In rice, 6 isoforms of putative BGALs were detected by proteomic [143]. Finally, it was also suggested that BGALs might also be required for PME activity [218,219]. Transcriptomic data showed very strong pollen specific expression profile of one β-xylosidase gene in mature dry pollen grains [199], and in in vitro and semi-in vivo grown pollen tubes. The exact substrate of this enzyme is not known but it may degrade several cell wall polymers such as xylan, xylogalacturonan and possibly N-linked glycoproteins. All the other putative xylosidases (12 genes that belong to the families GH3 and GH31) are only expressed in the pistil (Table 6). In addition, transcriptomic data do not show a strong expression of arabinofuranosidase genes in the pollen (Table 6). This is somehow surprising considering that Ara is one of the main carbohydrates present in the cell wall of A. thaliana pollen tubes [25]. It may suggest that the Ara containing polymers are not remodeled and/or that arabinofuranosidases originate from the female tissues (Table 6). The pattern of expression of α-xylosidase genes (GH31) is close to the one found for the arabinofuranosidase genes. α-xylosidase genes are not strongly expressed in dry pollen grains or pollen tubes. However, two α-xylosidase genes are expressed in the pistil (Table 6). Analyses of the A. thaliana Atxyl1 mutants show reduced α-xylosidase activity, altered XyG composition and shorter siliques compared to the wild type. Moreover pAtxyl1::GUS expression is strong in the style [220]. It suggests that AtXYL1 may remodel the XyG of pollen tubes. However, it is difficult to directly implicate this α-xylosidase activity in pollen tube growth as seed production was apparently not affected.

By comparing the proteome of rice pollen grains and pollen tubes, Dai et al. [191] showed that several proteins involved in the cell wall biosynthesis and remodeling are either up- or down-regulated. Several RGP, PME, PG and EXPB isoforms are up-regulated upon pollen germination whereas others PGs, xylanases and EXPBs are down-regulated indicating that several proteins are dedicated to pollen germination and others to pollen tube growth. The authors suggested that the down-regulated proteins in the pollen tube may result from their release into the culture medium. In the in vivo context, the release of these proteins in the pistil tissues may facilitate the pollen tube growth. Finally, glycosylated and/or phosphorylated proteins were detected and contributed to the generation of new isoforms [191]. The activity of enzymes is commonly regulated by phosphorylation, glycosylation and/or interaction with specific inhibitors but more studies are required to assess the fine tuning and the exact function of these posttranslational modifications in pollen tube growth [191].