1 Introduction and Definition

Plant roots invest a lot of energy in the formation and release into the rhizosphere of a population of living cells at their tips. These cells are programmed to detach from the cap during root growth either as border cells or border-like cells and represent a vital biotic boundary between the root and the rhizosphere. Border cells are experimentally defined as cells that disperse into suspension individually within seconds when the root tip is put into water (Hawes et al. 2000). The number of border cells released can vary considerably from a few hundreds to several thousands and it is generally conserved for plant species within a given family. The organization of the root apical meristem (RAM) is highly correlated with the production of root border cells. In eudicotyledonous angiosperm plants, RAM can be classified into three different organizations: closed RAM (composed of highly organized tiers of cells), basic-open RAM (with cells not clearly organized into distinguishable tiers), and an intermediate-open RAM (Chapman et al. 2003; Groot et al. 2004). Hamamoto et al. (2006) demonstrated that roots with open apical organization released high numbers of individual border cells: 4,500 and 10,000 cells have been reported to be released respectively by pea and cotton roots.

Unlike border cells, border-like cells do not detach as isolated cells and are therefore defined as cells that do remain attached to each other into small groups or as sheets after their release from the root tip (Vicré et al. 2005; Driouich et al. 2007). Such cells are very clearly observed in Arabidopsis thaliana (Fig. 1a), but they have also been found in other Brassicaceae members including canola, radish, and cauliflower (Driouich et al. 2007; see also Fig. 1c, d). In terms of number, after a week of growth under laboratory conditions, radish root produces more border-like cells (907 ± 75) than does Arabidopsis (116 ± 10) or canola (375 ± 137) roots. So far border-like cells have not been observed in other families such as the Leguminoseae, the Solanaceae, or cereal species (see Hawes et al. 2003).

Fig. 1
figure 1

Border cell and border-like cell morphology. Light and electron microscopy images of border-like cells of Arabidopsis (Arabidopsis thaliana) (a, b) or radish (Raphanus sativus) (c, d, e) and border cells released from pea (Pisum sativum) root tips (f, g, h). a, c, d, f, and g: light microscopy images of border-like cells from Arabidopsis (a) and radish (c, d) and border cells of pea (f, g). Unlike border cells (f, g), border-like cells (a, c, and d) remain attached to each other into small groups or as sheets after their release from the root tip. In d border-like cells released showing distinct morphologies: elongated versus ball-shaped (*) cells. Calcofluor staining of pea border cells (f). b, e, and h: electron micrographs of border-like cells from Arabidopsis (b) or radish (e) and border cells from pea (h). BLC as well as BC contain a large number of mitochondria and Golgi stack units indicative of a high secretory activity. BC border cells, BLC border-like cells, CW cell wall, ER endoplasmic reticulum, G Golgi stacks, m mitochondria, p plastids, PM plasma membrane, RT root tip, V vacuole. Bars: 50 μm (a, c, d, f, and g); 1 μm (b, e, and h)

Full size image

Border cells and border-like cells (Fig. 1) originate from the root-cap meristem whose cells undergo a series of divisions and differentiation giving rise successively to gravity sensing columella cells and root peripheral cells. How the release of such cells is regulated is not very well understood, but it seems to rely on developmental and environmental signals. Border cell formation and release can be switched off and on independently of root development and removal of the cells was reported to stimulate mitosis in the root-cap meristem within a few minutes (Brigham et al. 1998). In relation with this observation, it has been reported that a soluble factor (named factor B), secreted into the external medium by pea border cells, is able to control cell division in the root-cap meristem, thereby influencing border cell production and release (Brigham et al. 1998; Hawes et al. 2000). The factor is sensitive to protease hydrolysis suggesting its protein nature. Also, analysis of the Arabidopsis mutant fez has demonstrated that the release of border-like cells can be genetically controlled by an NAC-domain transcription factor (named FEZ) that is active in root-cap initials. The activity of FEZ in the epidermal/lateral root-cap cell initials has been shown to promote the formation of root-cap cells including the production of border-like cells (Willemsen et al. 2008). Whereas FEZ and its own negative regulator SOMBRERO are required for root-cap cell division (Willemsen et al. 2008), root-cap differentiation seems to be under the control of four transcription factors: PLETHORA 1–3 and BABYBOOM (Aida et al. 2004; Boutilier et al. 2002). These transcription factors are under the control of the hormone auxin via the action of auxin responsive factors.

As for environmental factors, temperature, carbon dioxide, soil-based mechanical stress, aluminum, and invasion by microorganisms have been shown to influence border cell formation, morphology, and number (Miyasaka and Hawes 2001; Curlango-Rivera et al. 2010).

It should be noted that border-like cells and border cells are not dead, unwanted “garbage” cells that are thrown away by the root into the surrounding environment. They are released from the root tip by a controlled process as viable and metabolically active cells that survive for weeks within the soil environment and even for months under laboratory conditions (Vermeer and Mc Cully 1982; Hawes et al. 2000; Vicré et al. 2005). It is, therefore, totally wrong (as it is indicated in certain recent publications) to consider A. thaliana as a species that does not produce viable border-like cells.

2 Attachment and Organization Patterns of Border-Like Cells

The production of border-like cells from the root tip of A. thaliana is dependent on the stage of root development. We have observed that the release of border-like cells does occur only when seedlings are 4–5 days old and that the number of cell layers increases over time (Vicré et al. 2005). Two types of cells were observed in terms of morphology, namely ball-shaped cells, which are located at the very tip of the root and elongated cells found on the margins of the cluster (see Fig. 1c, d). Shape and size of border cells and border-like cells seem to vary depending on the species as well as on RAM organization (Table 1 and Fig. 2). Ultrastructural observations using transmission electron microscopy of high-pressure frozen/freeze-substituted cells have revealed that they notably contain a large number of mitochondria and Golgi-derived vesicles indicative of a high secretory activity at the time of their release. Similar observations were also made for radish border-like cells as well as for border cells of pea (Fig. 1b, e, h). Indeed, using immunofluorescence microscopy, it has been possible to show that border-like cells of Arabidopsis are active in secreting high amounts of polysaccharides and proteoglycans into their cell surface. Carbohydrate epitopes associated with pectins, xyloglucan (XyG), and arabinogalactan-proteins were abundant at the surface of border-like cells of Arabidopsis and radish (Figs. 3 and 4). One interesting observation made on Arabidopsis border-like cells is that XyG epitopes (recognized by CCRCM1 antibody) were not only associated with the cell wall but can also be seen detaching from the cell wall along the entire cell, forming “hairy branches” at the cell surface (Fig. 4e). Also, XyG branches can be seen bridging two neighboring cells. Such XyG bridges suggest the involvement of this polysaccharide in connecting cells together, thereby contributing to their attachment to each other (make cells adhere to each other). However, in a study aimed at investigating the role of cell wall polymers in border-like cells attachment, it has been shown that mutants with altered XyG biosynthesis (murus 2 and murus 3) (Vanzin et al. 2002; Madson et al. 2003) have normally attached border-like cells, suggesting that alteration of XyG structure does not cause any modification of border-like cell organization and morphology (Durand et al. 2009). However, the XyG in both mutants is altered in the side chain structure not the backbone, and therefore it would be of interest to examine the organization and morphology of border-like cells in a mutant with an altered XyG backbone structure or a mutant that has no XyG such as the xxt1xxt2 double mutant (Cavalier et al. 2008).

Table 1 Variation in shape and size of border cells and border-like depending on the species as well as on RAM organization. Root tips (7–12 days old) are put into a drop of water for observation of cell release. Low adhesion means that cells are loosely attached to the root tip and to each other. Strong adhesion means cells are tightly attached to root tip and to each other. Impatiens glandulifera does not release border cells once the root is put into water. Under our conditions, only one layer of tightly adhered cells to root tip is seen – see also Fig. 2u
Full size table
Fig. 2
figure 2

Morphological phenotypes of root tips showing border cells or border-like cells of species representative of the families presented in Table 1. (a) Humulus japonicus; (b) Spinacia oleracea; (c) Penstemon royalis; (d) Acacia pycnantha; (e) Papaver rhoeas; (f) Pelargonium zonale; (g) Cactus sp.; (h) Ipomoea violacea; (i) Linum usitatisimum; (j) Amaranthus caudatus; (k) Millepertuis calycinum; (l) Myosotis sp.; (m) Primula vulgaris; (n) Solanum melongena; (o) Petroselinum crispum; (p) Lactuca sativum; (q) Fragaria sp.; (r) Thymus vulgaris; (s) Zea mays; (t) Carex sp.; (u) Impatiens glandulifera. Bars = 100 μm (a and d) or 50 μm ((b–g, i–t) or 20 μm (h and u)

Full size image
Fig. 3
figure 3

Immunofluorescence labeling of border-like cells and border cells with anti-extensin (LM1), anti-arabinogalactan-protein (JIM14), and anti-pectin (LM8 and JIM5) antibodies. (a, b, c, and d) Radish border-like cells stained with the monoclonal antibodies LM1 (a) and JIM11 (b) specific extensin epitopes, LM8 specific for xylogalacturonan epitopes (c), or JIM5 specific for homogalacturonan epitopes (d). (e, f, and g) Pea border cells labeled with the monoclonal antibodies JIM14 (e) and JIM5 (f and g). RT root tip, BLC Border-like cells, BC Border cells. Bars: 25 μm (d); 50 μm (a, b, c, e, f, and g)

Full size image
Fig. 4
figure 4

Morphological phenotypes of border-like cells and polysaccharide immunostaining of cell wall epitopes in the wild-type (WT) and quasimodo mutants of Arabidopsis. Wild-type Columbia (a) and quasimodo 1–1, qua1-1 (b) are shown. Note that in qua 1–1, border-like cells are converted into border cells (not attached to each other anymore) that are embedded in a thick mucilage (M). Wild-type Columbia (c, e) and qua 2–1 mutant (d, f) border-like cells are immunostained with the monoclonal antibodies, JIM5 (c, d) or CCRCM1 (antixyloglucan antibody) (e, f). In e, the white arrowhead indicates xyloglucan “fibers” at the cell surface of border-like cells. BLC border-like cells, qua quasimodo, M mucilage, RT root tip, WT wild type. Bar: 50 μm

Full size image

A similar investigation using the qua1 mutant has clearly demonstrated that the pectic polysaccharide homogalacturonan is responsible for border-like cell attachment. In qua1 mutant border-like cells are converted into border cells and this conversion is accompanied by secretion of abundant mucilage released by the cells themselves while the root is growing (Fig. 4a, b and also Driouich et al. 2010). This self-produced mucilage embeds border cells allowing them to remain close to each other, a mucilage that controls unity. The composition of the mucilage is not fully known, but studies using different antibodies have shown that it is enriched in xylogalacturonan and arabinogalactan-protein epitopes (Durand et al. 2009). We have termed such an association of cells and mucilage in qua1 mutant “a border cell biofilm” (Driouich et al. 2010) by comparison with microbial biofilms that form in response to various stress factors and that are usually composed of polysaccharides, proteins, and extracellular DNA (exDNA) (Davey and O’Toole 2000). We hypothesize that such a switch to the biofilm mode in qua1 mutant is linked to a specialized metabolic function of border cells required for the protection of the root meristem against biotic and abiotic stress. As for bacteria, border cell mucilage might not only hold the cells together but could also protect them and facilitate cell-to-cell communication particularly during pathogen invasion of the root tip. Such a thick mucilage, which can be considered as a mucilage of unity and protection, may also have the capacity to attract and agglutinate microorganisms. It is therefore challenging to investigate whether border cells exchange specific signaling molecules and whether the mucilage has antimicrobial properties (including secreted defense proteins or peptides) as found in bacterial biofilms. Border cells and exudates from qua1 root can be easily collected or microdissected and analyzed using proteomics, transcriptomics, and glycomics. Root mucilage isolated from many plant species has been described to be composed of up to 95% sugars and little amino acids (5%) (Bacic et al. 1986; Chaboud and Rougier 1984). Also, purified pea mucilage has been analyzed and shown to contain material similar to arabinogalactan-proteins (Knee et al. 2001). More recently, it has been shown that pea root mucilage has exDNA and antidefense proteins (Wen et al. 2007, 2009; see below). Also, unlike pea border cells and unlike border cells of qua1 mutant, border-like cells of wild-type Arabidopsis do not secrete as much mucilage as it is secreted by pea border cells.

Extending on cell wall structure of border-like cells, monosaccharide composition of the cell wall of such cells in radish and Brassica napus revealed the presence of a significant content in arabinose and galactose (Fig. 5; Cannesan et al. Submitted); two sugar residues that are mainly found in rhamnogalacturonan-I, arabinogalactan-proteins, and extensin. Further analysis of radish border-like cells using immunofluorescence labeling showed that extensin epitopes – recognized by LM1 or JIM11 antibodies – are strongly expressed at the cell surface of radish border-like cells (Fig. 3a, b). Thus, these cells seem to constitutively synthesize and secrete high levels of extensin that is likely to play a role as a molecular network barrier against pathogen penetration. Extensin may even be upregulated in these cells upon pathogen attack. It is well established that extensin accumulates in the cell wall of plant cells as a response to pathogen invasion (Esquerré-Tugayé and Lamport 1979; Merkouropoulos and Shirsat 2003). Also, it has been reported that both elicitor treatment and wounding lead to a rapid in muro insolubilization of extensin by oxidative cross-linking via isodityrosine motifs (Bradley et al. 1992; Brady and Fry 1997). We speculate that this might happen in the cell wall of radish border-like cells, thus creating a cross-linked protective network by the action of peroxidase and reactive oxygen species (ROS) as occurs in the cell wall of other tissues. Indeed, oxidative burst involving ROS production, established as one of the earliest response of plant cells to pathogen invasion, does occur in border-like cells of Arabidopsis treated with elicitors (Plancot et al. unpublished).

Fig. 5
figure 5

Monosaccharide composition of cell walls extracted either from isolated border-like cells (BLC) or the root apex free of border cells (Apex) from radish. Note the high content of arabinosyl and galactosyl residues. Ara arabinose, Fuc Fucose, Gal galactose, GalUA galacturonic acid, Glc glucose, GlcUA glucuronic acid, Man mannose, Rha rhamnose, Xyl xylose

Full size image

3 Border Cells Are Involved in the Protection of the Root Meristem

Border cells are fundamental to plant–microbe interactions. Within the rhizosphere, border cells are not only important in assisting the growing root to penetrate the soil, but they also provide a protecting cover surrounding the root tip against pathogens) (Gunawardena and Hawes 2002; Wen et al. 2009). Various studies have provided compelling evidence that border cells contribute significantly to the protection of the root and consequently of the entire plants. First, the number of border cell increases in response to pathogens (Cannesan et al. 2011). Second, border cells are capable of attracting, avoiding, or repelling pathogenic microorganisms. It has been shown that exposure of pea root to the pathogen Nectria haematococca results in the formation of a sort of mantle that covers the root tip (Gunawardena and Hawes 2002; Gunawardena et al. 2005). Under the microscope the mantle was shown to consist of a mixture of both border cells and fungal hyphae, and once the mantle is removed, the root tip remains free of infection indistinguishable from nontreated roots. Here, the root development and growth were unchanged from nontreated root, indicating that the apical meristems were still active and functional even with the presence of the pathogen in the surroundings. In such a case, border cells seem to fool the pathogen by acting as a decoy allowing the protection of the root tip and the apical meristems against infection. Similar observations have also been reported showing attraction and infection of cotton border cells by the fungus Pythium dissotocum (Goldberg et al. 1989). Border cells can also repel pathogenic bacteria by means of their secreted mucilage.

Third border cells are also capable of producing antidefense molecules. Border cells of legumes and cereals have been shown to secrete a large number of antimicrobial proteins including chitinases, peptidases, and glucanases, whose profile is modified in response to pathogenic bacteria (Wen et al. 2007; De-la-Peña et al. 2008). In addition to defensive proteins and enzymes, it has recently been shown that secretions of pea root border cells contain exDNA (Wen et al. 2009). This component is likely to exert a protective function within the secretome of border cells (Wen et al. 2009). The authors have clearly demonstrated that degradation of exDNA via nucleases (DNAse I or BAL31) resulted in an increased infection of root tips by N. haematococca supporting a role of this component in the immune response of root cells. In this regard, exDNA has long been known in many biological secretions including bacterial biofilms, snail mucigels, and white human blood cell matrices where it localizes with a number of antimicrobial peptides and proteins (Fahy et al. 1993; Allesen-Holm et al. 2006). In the case of human neutrophils, such a complex called also the NET “neutrophil extracellular trap” is capable of protecting the cells against pathogens at the sites of infection (Wartha et al. 2007; Guimarães-Costa et al. 2009). The discovery of exDNA in root tip secretions is highly exciting, but it is not easy to explain how exDNA exactly inhibits/reduces pathogen infection. This novel component of plant root secretions deserves further attention and careful investigations to unravel its mechanism of action on pathogenic microorganisms.

Another interesting role of border cells is that they have the capacity to attract and immobilize parasitic nematodes, thus limiting infection of the root. The chemical signal responsible for such an attraction is not known, but it is a heat-stable, polar fraction found in border cell secretions (Hawes et al. 2000). Also root border cells could contribute to inhibit penetration and invasion of cyst nematodes, thus limiting considerable loss of food crops. In this context, transgenic plants (including rice and potato) expressing protease inhibitors were shown to reduce the invasion and migration inside the root of several nematodes species (Fuller et al. 2008). For instance, transgenic potato plants that secrete a specific peptide interfering with nematode chemoreception under the control of CaMV35S promoter were able to significantly reduce the invasion by the cyst nematode Globodera pallida (Liu et al. 2005). Whether this peptide is expressed and secreted by root border cells in potato was not investigated. Targeting of such peptides specifically to border cells was recently reported by Lilley et al. (2010). The authors have identified a gene from Arabidopsis (MDK4-20; At5g54370) that possesses sequence homology to several genes expressed in maize root-cap cells. The promoter of this gene was then successfully used to direct expression of a nematode repellent peptide specifically to root caps, root border cells, and border-like cells of potato and Arabidopsis (Lilley et al. 2010). The MDK4-20 promoter was shown to remain active for a longer period of time than the constitutive CaMV35S promoter in detached border cells and border-like cells of the transgenic plants studied. Furthermore, the study has shown clearly that the targeted peptide reduced the establishment of the nematode G. pallida on root tissues. This is also an interesting discovery which paves the way to further investigations aimed at enhancing defense capabilities of border cells/border-like cells. Thus, using the same promoter and approach, it is possible to express various antimicrobial peptides or proteins specifically in border cells and border-like cells to target specific soilborne pathogens.

4 Conclusions and Future Prospects

Two centuries ago Charles Darwin (1880) highlighted in his book the importance of the root apex (including the root cap) for plant life when he wrote “… the tip of the radicle acts like a brain”. Indeed, the root cap plays a vital role in plant root growth and health. It protects the root meristems. Both apical and root-cap meristems (stem-like cells of the root system) are fundamental in providing new cells that form new tissues for plant growth and survival. Also, it is worth noting that the root cap is the only plant organ that can be regenerated after its removal.

It is well known that the root apex absorbs all the resources (water and nutrients) required for the plant to grow from the soil. It is also well established that root tips have the capacity to develop local resistance to infection when other parts of the root are invaded by a given pathogen. Nowadays, root diseases caused by soilborne pathogens are of increasing concern as chemical pesticides are withdrawn from use in agriculture due to their high toxic effects on the environment and human health. Alternative strategies are thus required for crop protection. One of such strategies is to make use of natural molecular mechanisms of root resistance to develop novel protective compounds. Studies on border cells of the legume species pea and others have clearly shown that these are able (1) to secrete various defense molecules and (2) to confer a protective effect against pathogens. There are at least two strategies that can be used to enhance the protective capacity of root-cap cells including border cells and border-like cells. One strategy is transgenic expression of defense proteins or peptides specifically in root-cap cells. For instance, this can be done by introducing a transgene in a background of partial natural resistance to provide added protection to the plant. Studies have shown that it is possible to target genes against nematode parasitism to root tissues (Lilley et al. 2010). However, the production of transgenic crop plants is not very welcomed, especially in Europe. The other strategy is to stimulate the immune system of root cells by treatment with natural elicitors (e.g., chitin and peptidoglycans). This sort of plant “immunization” by natural stimulating agents is already used for grapevine and winter wheat protection using Iodus 40, a natural oligosaccharidic-elicitor isolated from Laminaria digitata, or Stifenia, an active extract from fenugreek seeds. This is probably the most preferred strategy as it is environmentally friendly and more cost-effective. Thus, crop protection along with environment protection in the context of an increasing world population is an exciting challenge for root biologists among other plant scientists.