pathway, and plants with intermediate anatomy probably use either or both pathways. Other studies sup-
port this hypothesis of a dual nature of phloem loading [249,272].
Furthermore, Van Bel [273] proposed that the type 1 (symplastic loading) is evolutionarily primitive
and type 2 is evolutionarily advanced. He also pointed out that type 1 is more prevalent in tropical rain
forests, type 2a predominates in steppe and deciduous forest communities, and type 2b (containing trans-
fer cells) predominates in cold deserts and arctic-alpine communities.
For plants using the symplastic pathway, one must wonder how plasmodesmata can provide the se-
lective control of phloem loading and use metabolic energy to generated osmotic and pressure gradients
between ST-CC and other leaf cells. To be able to understand the functioning of plasmodesmata, their
structure must be understood. Plasmodesmata have been studied extensively and the literature on that sub-
ject has been reviewed periodically (see, e.g., Refs. 274–276). The earlier papers were devoted to struc-
ture and distribution of plasmodesmata. Later papers also discussed function.
Briefly, plasmodesmata are tubes of cytoplasm that connect the symplast of adjacent cells through
the cell wall with the plasmalemma surrounding these tubes. On average, the diameters of plasmodesmata
are about an order of magnitude narrower than connections between sieve tube members through sieve
plates. Each plasmodesma varies in diameter across the cell wall, with the narrowest portion being near
the ends where they connect with the main body of each cell (neck region). A desmotubule, which is an
extension of (or attached to) endoplasmic reticulum of each cell, extends through each plasmodesma. (De-
spite the name, desmotubules do not appear to be open tubes.) Between the desmotubule and plas-
malemma, within the “cytoplasmic sleeve,” is an extension of the cytosol. Surrounding the desmotubule,
and apparently adhering to it, are spherical cytoplasmic sleeve subunits that almost completely occlude
the space between the desmotubule and plasmalemma in the neck region.
Plasmodesmata vary in their degree of branching; many appear unbranched whereas others are
highly branched [277]. Plasmodesmatal connections between sieve tubes and companion cells have a sin-
gle pore into a sieve tube but are branched such that multiple pores enter a companion cell [261,278]. No
suggestion has been put forth to explain a functional basis for branching.
The most likely pathway for materials moving through plasmodesmata appears to be through the cy-
toplasmic sleeve. Electron micrographic measurements of the most restricted region, the neck, and ex-
perimental studies of movement of various-sized dye molecules support that hypothesis. Madore and
coworkers [261,279] and Van Kesteren et al. [280] demonstrated movement of fluorescent dyes from
mesophyll into vascular bundles via plasmodesmata. Madore et al. [261] also showed that continuous
plasmodesmatal connections exist from mesophyll into sieve tubes. However, Erwee and Goodwin [281]
demonstrated that Ca^2 could induce blockage such that cell-to-cell movement of dye did not occur. Ro-
bards and Lucas [276] pointed out that the exclusion limit for diffusion of dye through plasmodesmata
varies from 376 Da for roots and stems to 870 Da between mesophyll and bundle sheath cells of C plants.
However, they stated that “dye-coupling results do not establish that... phloem loading occur(s) via this
symplastic pathway. However,... modeling of the phloem system must incorporate the finding that plas-
modesmata within the vascular bundle are not vestigial.”
Another approach to studying phloem loading involved the introduction of yeast invertase genes into
tomato [126], tobacco, and Arabidopsis[127]. This enzyme was secreted into the apoplast and there hy-
drolyzed any sucrose present. In tobacco and tomato, this genetic alteration resulted in accumulation of
carbohydrate in mature leaves, major inhibition of carbohydrate translocation, stunted growth, and other
malformations that are characteristic of some plant diseases. Arabidopsiswas only slightly affected.
These data indicate that the prime (exclusive?) pathway of phloem loading for tobacco and tomato is
apoplastic, whereas Arabidopsismay have a symplastic alternative.
We are still faced with the problem of proposing a mechanism of symplastic phloem loading that
meets the criteria for phloem loading discussed earlier. However, some data suggest that not all those cri-
teria need be met. Turgeon and Medville [282] suggested that, for some plants, mesophyll cells are the
osmotic/pressure origin of the mass flow process, as originally proposed by Munch. This is based on their
observation of no osmotic gradient between mesophyll and the ST-CC in willow. In addition, Richardson
et al. [283] found that sieve tube exudate of cucurbits had a rather low concentration of osmotically ac-
tive substances. If these exudation values are close to in vivo values, one could eliminate the criteria spec-
ifying that large osmotic and pressure differences must exist between ST-CC complex and mesophyll
cells. But they [147] did question the assumption that phloem exudate represents the in vivo contents of
sieve tubes. Others [226,227], using a plasmolytic method, reported that CucurbitaandColeus,plants
PRODUCTION-RELATED ASSIMILATE TRANSPORT 437