Intrusive growth – the plant analog of dendrite and axon growth in animals


Cell migration is common throughout the animal kingdom (Montell, 1999), and it is a common notion that plant cells do not migrate during development. However, there are a number of tissues and cell types in plants where cell extensions invade between other groups of cells or tissues, or even into other cells – a process similar to dendrite and axon growth in animals. In such intrusive growth, which mostly occurs during tissue differentiation, the cell extensions enable the invasion of new locations. Pre-existing parts of the cell remain in their original location while new parts change their position in relation to other cells and occupy new locations – similar to the way in which plant parts move to new territories (Hart, 1990). Here, several developmental aspects of these special modes of plant cell extension growth are described, illuminating what is a common and important, but overlooked, phenomenon.

Intrusive growth

Cases of intrusive growth in plants include: elongation of fibers; growth of arms of branched sclereids; intrusive growth of the tips of fusiform initials in the cambium; growth of laticifers; growth of vessel-member elements in the secondary xylem of monocotyledons; tylosis formation; and growth of pollen tubes in the style. Some cells may grow to be several meters long, passing between the walls of hundreds of thousands of other cells. Although this intrusive growth must disrupt huge numbers of plasmodesmata and probably damages the middle lamella, no wound responses are initiated. Furthermore, the mechanisms by which plants distinguish between the penetration of a pathogen and penetration of a self cell are not known. Modern approaches have only been used to study movement between cells for the growth of pollen tubes in the style, a relatively late phenomenon in evolution.


Fibers are elongated cells that give plants strength and elasticity. They are common in the cortex, just outside the primary phloem (primary phloem fibers), in the secondary xylem and in the secondary phloem (Fahn, 1990). Fibers usually attain their length in two stages, in the beginning by symplastic growth and later by slow intrusive growth (Esau, 1965; Ghouse & Sabir, 1974; Ghouse & Yunus, 1975; Fahn, 1990). In certain trees, fibers grow in length even in the second year after the start of their differentiation (Esau, 1969). Many fibers are multinucleate and many have septa that divide them into several chambers (Esau, 1965; Fahn, 1990). Fibers several millimeters long are common in many species (Esau, 1965, 1969; Fahn, 1990; Ilvessalo-Pfäffli, 1995). Even in a small and short-lived plant, such as Arabidopsis, fibers in the inflorescence stems attain a length of > 300 µm (Lev-Yadun, 1997). Fibers in Boehmeria nivea may reach a length of 55 cm. These long fibers start as initials of c. 20 µm, and over a number of months grow to be some 27 500-times longer (Aldaba, 1927). During this growth, the fibre tips must pass through the middle lamella between tens of thousands of cells, disrupting a very large number of plasmodesmata (Wenham & Cusick, 1975; Barnett, 1981; Larson, 1994). However, no wound response following the growth of fibers is known.

The fibers of the secondary xylem are formed by the cambium, and if they elongate they do so in a tissue that is in a process of hardening during differentiation. Because the secondary xylem becomes so hard in a short time, intrusive growth of fibers in this tissue is usually limited to the tips (Larson, 1994). Even in the secondary xylem, fibre elongation in Eucalyptus globulus can be considerable (it continues for 6 d at a rate of 99 µm d−1) (Ridout & Sands, 1994).

In a very unusual case, a protophloem fibre in Pelargonium zonale grew into a parenchyma cell, indicating a loss of the control that usually restricts intrusive growth of fibers to intercellular spaces (Pizzolato & Heimsch, 1975).


Sclereids are one of the building units of the sclerenchyma, a hard tissue that protects soft tissues from damage by mechanical or biological agents. Sclereids are classified into several types according to their shape. Here, the variously branched astrosclereids and osteosclereids are considered (Fahn, 1990) – such branched sclereids grow intrusively into the area between cells and also sometimes into air spaces (Foster, 1945; Bloch, 1946; Arzee, 1953; Gaudet, 1960; Boyd et al., 1982; Heide-Jørgensen, 1990). It is believed that new protoplasmic connections, pit fields, and pits can be established between intrusive branches and adjacent cells (Boyd et al., 1982). Gaudet (1960), on the basis of sclereid morphology, proposed that the growth of the branches of sclereids in leaves of Nymphaea odourata is restricted by neighbouring cells.

Initials in the cambium

Intrusive growth of cambial initials, along with anticlinal divisions, loss of initials and transformation of fusiform to ray initials and vice versa, determine the structure of the cambium and, following that, the structure of both secondary xylem and phloem (Zagórska-Marek, 1984; Larson, 1994). Because the tissue in the cambial zone is soft, intrusive growth there does not have much physical resistance. Secondary tissues originating from the cambium are well organized and it is possible to identify many plant species using these organized structures of the secondary xylem (Schweingruber, 1990). Therefore, there must be a mechanism that regulates the intrusive growth of cambial initials in such a manner that will always give the wood structure typical for each species.


Laticifers are specialized cells or rows of cells containing latex, and compose a defense system in angiosperms (Dussourd & Eisner, 1987; Dussourd & Denno, 1991) analogous to the resin ducts of conifers (Fahn, 1979; Metcalfe, 1983; Rudall, 1987). ‘Articulated’ laticifers are composed of many cells, whereas the ‘nonarticulated’ form is multinucleate (coenocytic). Both types may be either unbranched or branched (Fahn, 1979). Branched nonarticulated laticifers can initiate only in primary tissues, but can invade other primary and secondary tissues (i.e. grow from the stem into the leaves by intrusive growth) (Mahlberg, 1959, 1961, 1963; Mahlberg & Sabharwal, 1968; Fahn, 1979; Rachmilevitz & Fahn, 1982; Cass, 1985). In certain cases, however, secondary laticifers are formed in the cambium or cortical secondary tissues (Vertrees & Mahlberg, 1978; Rudall, 1989; Van Veenendaal & Den Outer, 1990).

Mahlberg (1961) described the complicated relationships between the growing laticifers of Nerium oleander and the cells that come into contact with them:

‘The attenuated tips of the initials can be observed to bend, or grow around adjacent cells which appear to have become obstructions in their course. Growing tips follow the path of the middle lamella; they never have been observed to penetrate or fuse with adjacent cells. During their intrusive growth, laticifers establish contacts with new cells, while neighbouring cells are forced apart from their original points of contacts. In this way, new intercellular relationships are established during the entire growth of laticifers’.

The action of pectinases enables the penetration of the ends of the laticifers into the middle lamella (Mahlberg, 1963; Fahn, 1979). The first indirect evidence of this possibility was the presence of pectinase in the latex of Asclepiassyriaca (Wilson et al., 1976). Direct evidence for the role of pectinase activity in intrusive laticifer growth was shown in Nerium. The pectolytic enzymes were identified in the central vacuoles and along the middle lamella between laticifers and adjacent cells by electron microscopy using reaction products of the enzyme (Allen & Nessler, 1984).

Tracheary elements in the secondary xylem of monocotyledons

Many monocotyledons do not have secondary growth, and, where it does occur, it is little studied. In some cases, the vessel members of the secondary xylem of monocotyledons elongate up to 40 times their original length by means of intrusive growth (Philipson et al., 1971).


In the process of tylosis, cell parts invade the air-filled lumen of dead vessel members or tracheids rather than in between live cells – this makes it distinct from the other types of intrusive growth discussed here. It is a defensive process for sealing vessel members and tracheids that have lost their water for a long time and have not been re-filled. Tyloses occur following mechanical injury or fungal, bacterial or viral infections (VanderMolen et al., 1987). They are formed by intrusion of outgrowths from adjacent parenchyma cells through pit-pairs into the vessel or tracheid lumen. The tyloses form cell walls and pit pairs between neighbouring tyloses, and some accumulate starch or have their own nucleus. When the process is completed, the vessels and tracheids are filled with cellular bubbles that plug them completely (Zimmermann, 1983). Introduction of various nonhost-specific or nonpathogenic organisms into ruptured vessels has been shown to induce tylosis, but introduction of sterile vinyl particles failed to induce the process. These experiments indicate that induction of tylosis is a nonspecific response to infection (VanderMolen et al., 1987). Although the actual signal for tylose formation in cases of pathogenesis is not known, evidence from susceptible and nonsusceptible tomato plants inoculated with Fusarium indicates that there are tylosis-inhibiting molecules. It was proposed that one of the molecules that inhibits tylosis is rishitin. This inhibitor accumulated rapidly in both resistant and susceptible tomato cultivars 1–4 d after inoculation with Fusarium, but in susceptible cultivars it continued to accumulate in the following days (Harrison & Beckman, 1987).

Pollen tubes

Except for cases of selfing, pollen tubes are a foreign entity in the plant – hence, as with tyloses, they are distinct in their form of intrusive growth. Growth of pollen tubes has also received intensive coverage in recent years – considerable progress in understanding the physiological, molecular and genetic aspects has already been achieved (Nasrallah et al., 1991; Pierson & Cresti, 1992; Clarke & Newbigin, 1993; Sims, 1993; Franklin et al., 1995; Derksen, 1996; Malho & Trewavas, 1996; Woltering et al., 1997; Lord, 2000; de Ruijter & Malhó, 2000). It is, however, worth noting that pollen tube growth is a late evolutionary development in land plants: fibers, sclereids and cambium appeared a long time before their emergence. Therefore, the types of mechanisms that enable pollen tubes to penetrate tissues with no resistance and damage should have existed in plants for performing intrusive growth millions of years before the evolutionary appearence of the pollen tubes themselves.

Self and nonself

Plants ‘know’ how to recognize self and nonself via different receptors, a recognition process studied mostly with respect to pollination (Nasrallah et al., 1991; Pierson & Cresti, 1992; Clarke & Newbigin, 1993; Sims, 1993; Franklin et al., 1995; Derksen, 1996; Malho & Trewavas, 1996; Woltering et al., 1997) and pathogenic and symbiotic interactions (Lamb et al., 1989; Ebel & Cosio, 1994; Boller, 1995). Although the present article does not cover interactions with other organisms, it is interesting to note that in mutualistic symbioses, such as in the development of the Hartig net in ectomycorrhizas and the invasion of the periplasmic space of the cell lumen in endomycorrhizas, there is no wound response as would occur with many pathogens.

Internal self wounding during differentiation

The evidence for internal self wounding by intrusive growth in the process of differentiation is meager. The clearest case of internal self wounding is the emergence of lateral roots. Roots branch internally mostly from the pericycle (Fahn, 1990), but the endodermis can also contribute cells to the epidermis and root cap (Bell & McCully, 1970). The emerging lateral roots must pass through the cortex of the mother root, which should result in wound effects. Moreover, outgrowth of lateral roots through the cortex is associated with the expression of polygalacturonase, which hydrolyses the pectin of the middle lamella (Sutcliffe & Sexton, 1968; Bell & McCully, 1970; Peretto et al., 1992). Futher evidence for digestion of cortical cell protoplasm, which indicates cortical wounding in the path of lateral roots, is the pile up of empty cortical cell walls (Bonnett, 1969; Bell & McCully, 1970). The outgrowth of lateral roots through the cortex has been proposed to wound the tissue. The evidence for wounding is the shortening of the secondary vessel members in the xylem of the mother roots of Arabidopsis formed near emerging lateral roots (Lev-Yadun, 1995). Shortening of cambial initials and vessel members following wounding is well known from a variety of plant species (Lev-Yadun, 1995). Aloni & Baum (1991) found that in Luffa cylindrica, regenerative xylem vessels differentiated from the stem parenchyma around initiations of adventitious roots. This is also good evidence of internal wounding.

Open questions

Many questions arise concerning intrusive growth during fiber, sclereid or laticifer elongation:

  •  What are the signals that initiate cell elongation and intrusive growth?
  •  How does the cell tip soften the middle lamella?
  •  How does a cell determine the direction of growth? The fact that fibers, sclereids and laticifers have regular arrangements in the plant indicates that their elongation is well controlled.
  •  What are the cellular mechanisms involved in intrusive growth?
  •  When intrusive growth occurs, the plant must sense that the penetration into the tissue is not foreign or hazardous. Does the intrusive growth induce a wound responses and to what extent? If intrusive growth does not induce a wound response, why not?
  •  How can the plant distinguish between intrusive growth and the growth of fungal hyphae?
  •  Why does the intrusive growth of fibers and sclereids stop, while that of laticifers is less limited? Is it because the cells are too old? Is it because there are problems of coordination in very long cells? Is it because of mechanical resistance of the tissue? Or is it because of the activation of pathogenesis-related defense mechanisms that inhibit the growth?


In all the cases of invasion of cell extensions described here, the old part of the cell remains in position while new parts grow to new locations. Therefore, although plant cells do not migrate as in animal cells, cell parts do reach new locations as occurs with axons and dendrites. The common view that plant cells do not migrate at all is therefore inadequate. Certain parasitic vascular plants, such as Viscum spp., Loranthus spp. and Orobanche spp., invade host plants without inducing their defense mechanisms to such an extent that they block the invasion. Similarly, mutualistic symbioses such as ectomycorrhizas and endomycorrhizas do not induce wounding responses. Could the mechanisms of invasion of such parasites and symbionts be genetic modifications of mechanisms normally controlling the fibre, sclereid and laticifer differentiation used for parasitic and symbiotic interactions?


I thank Gideon Grafi, Hillel Fromm and two anonymous reviewers for their comments.