Planta (1992)187:359-366

P l ~ I l t ~ 9 Springer-Verlag 1992

Quantification of symplastic continuity as visualised by plasmodesmograms: diagnostic value for phloem-loading pathways C.E.J. Botha 1 and A . J . E . van Bel 2

1 Department of Botany, Sch6nland Laboratories, Rhodes University, P.O. Box 94, Qrahamstown 6140, South Africa 2 Department of Plant Ecology and Evolutionary Biology, University of Utrecht, P.O. Box 800.84, 3508 TB Utrecht, The Netherlands Received 24 April; accepted 7 November 1991

Abstract. The use of plasmodesmatal frequency to cor-

relate cell-cell symplastic transport capacity remains a contentious problem, as variation in cell shape, accurate determination of interface contact area between cell types, distribution (i.e. whether random or aggregated) and shape (i.e. whether single or branched), and state of permeability may confuse the issue. Additionally, variation in the methods used to determine the frequencies compounds the problem further. Data presented in this paper show that plasmodesmograms offer a means to visualise the potential transport pathway from mesophyll cells to sieve tubes. Furthermore, the results allow an instant appreciation of symplastic continuity or discontinuity and, accordingly, the potential symplastic and-or apoplastic stages involved in the overall loading process. Key words: C3, C4 plants - Phloem loading (apoplastic,

symplastic) - Plasmodesmatal modesmogram

frequencies -

Plas-

Introduction

Understanding of how assimilates produced in the mesophyll of angiospermous leaves are loaded into the phloem requires detailed studies of plasmodesmatal frequencies in the loading zone (see van Bel et al. 1988; Botha 1990; Fisher 1990a and references therein). It is assumed that the analogy here is that the greater the number (frequency) of plasmodesmata at a given interface, the greater is the potential for symplastic transport through that interface. It is therefore crucial that plasmodesmata at the various interfaces along the assumed pathway for assimilate loading are counted before the relative role(s) of symplastic and-or apoplastic transport can be meaningfully addressed. The result should be the production of a set of data which gives an accurate indication of both the number and frequency of plasmodesmata at the relevant interfaces. Unfortunately, as stated by Fisher (1990a, b) several differing methods have

been used by workers active in this field to determine actual frequencies. Whilst some degree of overlap between the various methods exists, no single system or method is universally accepted or used in frequency studies. Frequency studies are usually presented in tabular form. Tabulated data require careful study and descriptive analysis. In many instances frequency data do not give a clear indication of the relative importance of the interfaces along the route from mesophyll to functional sieve tubes. Attempts have recently been made to address this singular problem, and to present the results, graphically. Russin and Evert (1985b), using cell-cell frequencies in diagrammatic form, showed the assimilate-loading area in Populus deltoides. Subsequently, pictorial classification (pictograms) of minor veins (Gamalei 1985), and more recently, 'plasmodesmograms' (van Bel et al. 1988) have been used. Each of the above has an advantage in that the cell-cell symplastic and-or apoplastic route and phloem-loading pathway can be instantly appreciated. The structural parameter necessary to facilitate answers (at least in part) to some of these questions clearly remains the determination of plasmodesmatal frequency at the diverse cell interfaces along the route from the mesophyll cells to the phloem tissue. Plasmodesmograms are thus pertinent to studies on the role of plasmodesmata in phloem loading and intercellular transport. Notwithstanding the inherent problems associated with frequency expressed as either plasmodesmata/~tm cell-wall interface, or plasmodesmata/lxm vein, comparative plasmodesmograms constructed from data in which plasmodesmatal frequency is calculated on the basis of plasmodesmataAtm cell-wall interface or plasmodesmata/txm vein, each expressed as a percentage of the total number of plasmodesmata, may be the best form of visualization and may serve as the basis for useful comparisons. This paper serves to illustrate several points. First, that it is essential that workers agree and "standardise" cell and related cell-interface terminology. Second, where interface plasmodesmatal frequencies are

360

C.E.J. Botha and A.J.E. van Bel: Plasmodesmograms: diagnostic value for phloem-loading pathways

expressed as a percentage of the total number o f plasm o d e s m a t a , that the resultant p l a s m o d e s m o g r a m s differ little for the m o s t part. Third, that plasmodesmograms permit easy visualization of the loading pathway from mesophyll to sieve tube. Finally, that plasmodesm o g r a m s m a y be useful adjuncts to experiments aimed at determining the phloem-loading pathway.

Results and discussion

Materials and methods Plasmodesmograms for Erayrostis plana Nees, Panicum maximum Jacq., Themeda triandra var. imberbis (Retz.) A. Camus and Bromus unioloides H.B.K. were constructed based on either plasmodesmata/ lam cell-wall interface, or on plasmodesmata/lam vein. The latter method uses as the basis for calculations the formula in the paper by Robards (1976) (see also the discussion by Gunning in Robards 1976; and Fisher 1990a for further discussion), which considers the plasmodesmatal frequency (expressed as plasmodesmata/lam cellwall interface) the section thickness (in nm) the average plasmodesmatal radius (in nm) and the total interface length (in ~tm). The plasmodesmogram for Commelina benyhalensis L. was redrawn from van Bel et al. (1988); those for Amaranthus retroflexus L., Cananga odorata (Lam.) Hook. f. et Thorns., Coleus blumei Benth., Commelina benyhalensis and Sonchus oleraceus L. were drawn from data presented by Fisher (1990a, b); and that for Spinaeia oleraeea L. cv. Bloomsdale Dark Green from data by Warmbrodt and van der Woude (1990). Plasmodesmograms, comparing plasmodesmatal distribution as percent plasmodesmataAtm cell-wall interface (which reflects the number of plasmodesmata, the number of like interfaces, the number of sections, as well as the actual interface contact area) and as percent plasmodesmata/lam vein were drawn for Eraorostis plana, Panieum maximum, Themeda triandra and Bromus unioloides from data previously calculated by Botha (1990). Solid lines in the accompanying plasmodesmograms represent a distribution of at least 1%, and dashed lines < 1% of the total plasmodesmata. To avoid the confusion of massed lines in confined spaces, numbers in circles depict the actual plasmodesmatal percentage, at the interfaces of interest. f

f cl

Q

Mes

Comparison o f calculated plasmodesmatal frequencies and resultant plasmodesmoorams. Figures 1 and 2 show the difference between the plasmodesmograms based on computations and interpretation of the plasmodesmatal frequencies in Commelina benohalensis (Fig. 1: van Bel et al. 1988; Fig. 2: Fisher 1990a); labelling is as per the original data. Although the authors have used different terminologies to describe the gross cell-cell pathway from the mesophyll to the functional phloem sieve tubes, they all recognise and differentiate between connections at the vascular p a r e n c h y m a - p r o t o p h l o e m and vascular parenchyma-metaphloem-sieve tube interfaces. H o w ever, Fig. 1 (van Bel et al. 1988) shows similar numbers of plasmodesmata (and implied from this, similar frequencies) between vascular parenchyma cells (VP) and metaphloem sieve tubes (MST), and between companion cells (CC) and protophloem sieve tubes (PST). Both sets of data, however, show near-total symplastic isolation of the C C - M S T complex. The p l a s m o d e s m o g r a m derived from van Bel et al. (1988) shows sheath-cell sheath-cell (SC-SC) and V P - V P connections, whilst these lateral connections are not evident in the Fisher (1990a) data. The plasm o d e s m o g r a m drawn from Fisher (1990a) emphasises the importance of the V P ~ m e t a p h l o e m route in assimilate transport, in that 12% of the total number of plasmodesmata (based on percent plasmodesmataAtm vein) occur at this interface. The dashed lines in the Fisher-based p l a s m o d e s m o g r a m indicate a very low level of symplastic continuity at the bundle-sheath-companion-cell (BS-CC) and V P - C C interfaces, where less than 1% of the total plasmodesmata (as plasmodesmata/ ~tm vein) occur at the interfaces in question. Both plas-

1

11111111811111111

CC

Commelina benghalensis Van Bel (1988)

J

\

Commelina benghalensis

Figs. 1, 2. Plasmodesmograms showing the comparative distribution of plasmodesmata in Commelina benghalensis. Fig. 1. Distribution according to van Bel et al. (1988). Fig. 2. Plasmodesmatal frequency, expressed as plasmodesmata/~tm vein, as determined by Fisher (1990a). Numbers in circles indicate percent plasmodesmatal frequency, solid lines show plasmodesmatal connections of > 1%, whilst dotted lines show plasmodesmatal connections of < 1% plasmodesmata/lam vein for that interface. BS or SC= bundle sheath; CC=companion cell; Mes or MC= mesophyll cell; M or MST= metaphloem; P or PST= protophloem; VP = vascular parenchyma

C.E.J. Botha and A.J.E. van Bel: Plasmodesmograms: diagnostic value for phloem-loading pathways

361

3"

I

KMS

INg BS ~ BS 11811

. . . . .

A

B

J

4"

A

B

J

modesmograms (Figs. 1, 2) indicate that the metaphloem (MST, Fig. 1; M, Fig. 2) could be loaded almost exclusively by the vascular parenchyma cells, with only a small portion o f the assimilate being loaded via the C C - M interface. The Fisher-based plasmodesmogram shows that less than 1% of the total plasmodesmata occur at this interface. The van Bel-based plasmodesmogram (Fig. 1) shows many lateral plasmodesmata at the SC-SC (analogous to BS-BS) interface whereas that based on Fisher (1990a), (Fig. 2) does not.

Figs. 3, 4. Plasmodesmograms showing frequencies at all pertinent interfaces from mesophyll to phloem sieve tubes in the Ca and C4 grasses Erayrostis plana (3A, B) and Panieum maximum (4A, B), expressed as percent plasmodesmata/lam cell-wall interface (A) and percent plasmodesmata/gm vein (B). Open circles = thin-walled sieve tubes; closed circles= thick-walled sieve tubes. KMS = Kranz mesophyll cell; MS = mestome sheath; PS = parenchymatous bundle sheath. Other symbols as in Figs. 1, 2

Both show relatively high frequencies at the BS-VP interface (not quantified in Fig. 1, 37% Fig. 2) and direct connections with companion cells at the parenchymatous bundle sheath. The Fisher (1990a) data do not show the indirect VP-VP connections in the overall pathway from mesophyll to sieve tube. Clearly, there is a need not only to standardise the description o f cell types, but also to include all cell types in the calculations. This should assist in avoiding any difficulties when examining the same species!

362

C.E.J. Botha and A.J.E. van Bel: Plasmodesmograms:diagnostic value for phloem-loadingpathways

T. triandra

KMS

KMS

BS

Bs

j~

BS

VP i .7

o

A

3i

B

PS

A

B

J

Plasmodesmatal frequencies in grasses-comparative plasmodesmograms based on plasmodesmata/izm cell wall, and plasmodesmata/#m vein interface. The principal results obtained and used in this study are shown in Table 1 and Figs. 3-6. Consideration of the formula used to derive plasmodesmata/tam vein (see Robards 1976) shows that the calculation of plasmodesmatal frequency using this parameter is dependent on the accurate determination of: (i) section thickness, (ii) plasmodesmatal diameter, and (iii) total interface length. Any error in determining these parameters must influence the computed plas-

Figs. 5, 6. Plasmodesmograms(for details, see Figs. 3, 4) of Themeda triandra (SA, B) and Bromus unioloides (6A, B)

modesmatal frequencies to some extent. Careful control of section thickness and accurate measurement of plasmodesmatal diameters are essential if a degree of accuracy is expected from this method. Figures 3-6 show plasmodesmograms constructed using percent plasmodesmata/gm cell wall (Figs. 3A, 4A, 5A, 6A) and percent plasmodesmata/gm vein (Figs. 3B, 4B, 5B, 6B), respectively. In the C4 grasses (E. plana, P. maximum and T. triandra), high plasmodesmatal frequencies exist at the Kranz mesophyll-bundle sheath (KMS-BS) interface, with 49-77% of the total number

C.E.J. Botha and A.J.E. van Bel: Plasmodesmograms: diagnostic value for phloem-loading pathways Table 1. Comparative total plasmodesmata in small and minor

veins expressed as plasmodesmata/txm vein for the species used in construction of plasmodesmograms Species

Plasmodesmata

Grasses" Eragrostis plana Panicum maximum Themeda triandra Bromus unioloides

3566 1514 576 455

Dicotyledons: Amaranthus retroflexus Coleus blumei Cananga odorata Commelina benghalensis Spinacia oleracea

2511 1081 569 113 504

ofplasmodesmata associated with this interface. Relative frequencies vary, depending on the method used to determine the total number (and hence, the percentage) of plasmodesmata at a particular interface. In P. maximum, comparison of the two plasmodesmograms shows that the KMS-BS interface frequency is a substantial 21% higher when expressed as percent plasmodesmata/lam cell wall (77 and 56%, respectively). However, Fig. 4B shows that using plasmodesmata/lam vein as the basis calculation results in a remarkable 24% difference in the calculated frequency at the BS-BS interface (4 and 28%, respectively; Fig. 4A, B). Nevertheless, both methods yield very similar frequencies at the BS-VP interface (2% lower for plasmodesmataAtm vein). In Themeda, the difference at the KMS-BS interface (Fig. 5A, B) is 8% higher in the plasmodesmogram based on plasmodesmataAtm cell wall. The BS-VP interface shows a 10% frequency increase, in the plasmodesmogram based on plasmodesmata/lxm vein. Eragrostis plana, on the other hand, shows very little variation in frequencies at the KMS-BS-VP interfaces, with only a 3-5% change in the plasmodesmograms based on plasmodesmata/lam cell wall, or plasmodesmataAtm vein (Fig. 3A, B). The differences between the plasmodesmograms in which frequency is calculated as percent plasmodesmata/ ~tm cell wall and those in which it is calculated as percent plasmodesmata/~tm vein arise as a result of the measured decreased mean plasmodesmatal diameter at the BS-VP interface, from 55 to 38 nm for interfaces endarch to the BS-VP interface, which will directly affect the calculated number (and hence, percentage distribution) of plasmodesmata in the grasses discussed here. The plasmodesmograms of the C 3 grass B. unioloides (Fig. 6A, B) again demonstrate variation between the two methods. At the parenchymatous bundle sheathmestome sheath (PS-MS) interface, the plasmodesmatal frequency increases from 32 to 48 %, whilst the MS-VP interface shows a decrease from 25 to 18%, when comparing plasmodesmograms A and B. However, the 7% decrease at the MS-VP, and 15% decrease at the VP-VP interfaces do not lessen the relevance of the symplastic route, PS--*MS~VP~thin-walled sieve tube.

363

Clearly, the changes in Figs. 3B, 4B, 5B and 6B are attributable to the components of the formula given in Robard's (1976) paper. Total counts, plasmodesmatal frequency and the pitfalls in representing frequency as the percentage of plasmodes~ mata appearing at each interface. Currently, plasmodesmata/txm cell wall interface, and plasmodesmataAtm vein are the preferred methods of expressing frequency. Whilst Fisher (1990a) quite rightly points out that percent plasmodesmata is useful only for comparing the relative numbers of plasmodesmata within a species, expressing the plasmodesmatal frequency as a percentage of the total plasmodesmataAtm cell wall or total plasmodesmataAtm vein seems to be the most logical way of representing interface data in plasmodesmograms. The two methods yield different frequency data, and this may simply reflect the inaccuracy of either or both methods. The actual number of plasmodesmata involved in a frequency study thus assumes importance when dealing with total distribution when individual cell-cell frequencies are expressed as a percentage of the total plasmodesmata. Table 1 serves to illustrate this point: the range of values for total plasmodesmata (plasmodesmataAtm vein), for the species reffered to in this paper, contains high (3566, E. plana) and low (113, C. benghalensis) frequencies. Two examples which demonstrate the effect that total number of plasmodesmata counted (frequency) has on the final expression as percent plasmodesmata can be cited. In the first, P. maximum, the calculated total plasmodesmata/Ixm vein was 1514 and the calculated frequency at the VP-ST interface was 7.6 plasmodesmata/lxm vein, or 0.5% of the total plasmodesmata. In the second example, of the 113 plasmodesmataAtm vein calculated by Fisher (1990a, b) in C. benghalensis, 13.7 plasmodesmata/lxm vein, or 12.1% of the total occur at the VP-ST interface. Comparison of the plasmodesmograms indicates a substantial symplastic route in C. benghalensis via the vascular parenchyma, whilst the low (0.5%) frequency between vascular parenchyma cells and sieve tubes in P. maximum does not. However, P. maximum has 13.4 times more plasmodesmata (on a plasmodesmata/ixm vein basis) in total than C. benghalensis which must have direct bearing on plasmodesmatal frequency, when expressed as a percentage of the total. Caution must therefore be exercised when frequencies are determined from data with low overall plasmodesmatal counts in order to avoid an overinflated "high" (symplastic) and-or "low" (apoplastic) impression. Parallels between plasmodesmata/izm cell wall interface and plasmodesmata/Izm vein. The comparative plasmodesmograms of grass leaf-blade bundles show that the methods used to calculate frequency, i.e. plasmodesmata/lam cell wall interface and plasmodesmataAtm vein, in general, yield slightly different results at the interfaces centripetal to the vascular parenchyma, when compared on the basis of percentage of total plasmodesmata. These differences lead us to suggest that frequencies should be calculated using both methods for the following reasons:

364

C.E.J. Botha and A.J.E. van Bel: Plasmodesmograms: diagnostic value for phloem-loading pathways

f

f Mes1

9~

I Mes1

IIN Ill

ill

II

B$

. IJ

CC

Amaranthus

ll

CO

VP

C a n a n g a odorata

retroflexus

IIIIINItFHr

P

Sonchus oleraceus

J

lo~ "

I Mes1 i11111t

C o l e u s blumei J

I1 11 cc

Spinacia oleracea

1) Plasmodesmata/gm cell wall interface reflects the actual number o f plasmodesmata counted in relation to the actual contact area o f the cell wall interface. 2) In contrast, frequencies calculated and expressed as plasmodesmata/gm vein, although the preferred method if solute fluxes are to be estimated, is, as previously mentioned, controlled by three important variables, which are all components o f the formula given in the paper by Robards (1976): (i) section thickness (ii) plasmodesmatal diameter (they all differ slightly), and (iii)

Figs. 7-11. Plasmodesmograms showing the distribution of plasmodesmata in Amaranthus retroflexus, (Fig. 7), Cananga odorata (Fig. 8), Coleus blumei (Fig. 9), Sonchus oteraceus (Fig. 10), and Spinacia oleracea (Fig. 11). Numbers in circles shows percent plasmodesmatal frequency; solid lines indicate plasmodesmatal frequencies of > 1%, whilst dotted lines show plasmodesmatal connections with a frequency of < 1% plasmodesmata/gm vein for that interface. Ptasmodesmograms were computed from data in Fisher (1990; Figs. 7-10) and Warmbrodt and van der Woude (1990; Fig. 11). BS= bundle sheath; CC= companion cell; I= intermediary cell; Mes = mesophyll cell; PP = phloem parenchyma; S T ( M ) = sieve tube (member); VP = vascular parenchyma

total interface contact area in relation to the number o f ceils forming the particular type o f interface. Impreciseness will inevitably result if this method is relied on exclusively. 3) A factor which has been overlooked in all the most recent reviews and papers is that plasmodesmata are assumed (in calculations where frequencies have been determined using plasmodesmata/gm vein for example) to be totally randomly distributed. Clearly they are not; and more specifically, not at the BS-VP interface in the

C.E.J. Botha and A.J.E. van Bel: Plasmodesmograms: diagnostic value for phloem-loading pathways Ca or C4 grasses discussed in this paper. Large plasmodesmatal aggregates are the rule and not the exception at this interface. In terms of mathematical and statistical probability, the non-randomized plasmodesmatal distribution at this interface is cause for some uncertainty about the validity of plasmodesmata/~m vein, or any other method for that matter, as claimant to be the "ultimate" expression of frequency.

The symplastic pathway to thin- (ST) and thick~walled (TWST, late metaphloem) sieve tubes of grasses. Irrespective of the method used to calculate plasmodesmatal frequencies, a strongly developed symplastic pathway from PS--*MS~VP and K M S ~ B S ~ V P in C3 and C4 grasses is evident in Figs. 3-6. In addition, all four grasses demonstrate a convincing lateral pathway between bundle sheath cells (in the C4 grasses) and between ruestome sheath cells in the C3 B. unioloides. Symplastic cell--cell continuity seems to be virtually absent at the VP-TWST and VP-CC/ST interfaces. All plasmodesmograms show that STs are symplastically isolated from TWSTs. Whilst the reason for this is not clear, it is important to note that the CC-ST complex and VP-TWST association are not entirely isolated from other vascular parenchyma cells. Near-total symplastic isolation, as indicated by the plasmodesmograms, must imply an apoplastic step in the loading process. Comparative plasmodesmatal frequencies in minor veins of dicotyledonous leaves. Figures 7-11 are plasmodesmograms (expressed as percent plasmodesmata/ ~tm vein) for minor veins from leaves of five dicotyledonous species. Direct comparisons among these species are difficult in that only the centripetal pathways are considered here. As in the grasses, plasmodesmatal frequency between mesophyll and parenchymatous bundle sheath cells varies-from 24% (Coleus blumei, Fig. 9) to 84% (Amaranthus retroflexus, Fig. 7). Unlike the C4 species A. retroflexus, the C3 species Cananga odorata, (Fig. 8), Sonchus oleraceus (Fig. 10) and Spinacia oleracea (Fig. 11) exhibit symplastic continuity along the mesophyll-cell ( M e s ) ~ B S ~ V P ~ C C ~ S T pathway. A relatively high plasmodesmatal frequency is evident at the CC-ST interface in C. odorata and Sonchus oleraceus (15 and 13% of total plasmodesmata were recorded respectively at this interface), indicating a symplastic step in the phloem-loading process at this interface. Thus assimilates could follow the M e s ~ B S ~ V P ~ C C ~ S T route. Few connections exist between the phloem parenchyma (PP, Fig. 11) and sieve tube members (STM) in S. oleracea, and we therefore suggest that the phloem parenchyma is not in this species, involved in symplastic phloem loading. It is important to note that Coleus blumei contains specialised vascular parenchyma cells, termed intermediary cells, (I, Fig. 9) that are in close spatial association with sieve tubes. Although plasmodesmatal frequency is low at the I-ST cell interface in C. blumei (7%) when compared with the CC-ST interface in Can~ anga odorata or Sonchus oleraceus, (15 and 13 %, respectively) the high frequency (56%) at the BS-I interface indicates a direct symplastic cell-cell loading course

365

via intermediary cells, to sieve tubes along the M e s ~ B S ~ I ~ C C ~ S T route. The total absence of plasmodesmatal connections at the V P ~ S T interface in C. blumei, Cananga odorata and Sonchus oleraceus is pivotal, indicating that the vascular parenchyma cells may be involved only in apoplastic loading of sieve tubes in minor veins of these species, and that symplastic loading must follow the B S ~ V P ~ C C ~ S T route discussed above. Mixed (apoplastic and symplastic) loading can, of course, not be excluded here.

The phloem-loading pathway. It is not possible to make an absolute case for either symplastic or apoplastic transport, based on plasmodesmatal frequency alone (Warmbrodt and van der Woude 1990) as the minimal frequency required for a major contribution of assimilate loading from the mesophyU to the sieve tubes is, as yet, unknown (Fisher 1986). Recent work (Erwee et al. 1985; Madore et al. 1986; Fisher 1988; van Kesteren et al. 1988; Turgeon 1989) indicates that low-molecularweight dyes such as Lucifer Yellow are able to move over short distances symplastically from mesophyll cells to minor veins. These authors could not, however, determine if the dye entered the ST-CC complexes. Regardless of the relative roles played by the symplast or apoplast in sieve tube loading, the assumption is that companion cells are involved in, and specialised for the uptake of sugar and its subsequent transfer to sieve tubes (Esau 1977; Giaquinta 1983). However, plasmodesmatal frequencies are relatively low in a number of species at the CC-ST interface (see Evert and Mierzwa 1986; Warmbrodt and van der Woude 1990 for further discussion). Generally, lower solute concentrations have been reported for companion cells of the minor veins, compared with their associated sieve tube members (Fellows and Geiger 1974; Russin and Evert 1985a, b). The reported higher solute concentrations of the ST-CC complexes compared with other leaf cells (see discussion in Fisher 1990a) tempts a conclusion about the loading process. However, solute concentrations in the CC ST complex indicate only the total osmotically active compounds and not specifically, the osmolarity of the photosynthate in these cells. Frequency data, whether presented as tables or plasmodesmograms, give an indication of the maximum potential pathway of symplastic transport. Frequencies merely represent the number of plasmodesmatal connections determined at a particular time, for a particular interface and at a particular stage of development of the cells making up the interface. Plasmodesmatal frequency per se does not take the transport capacity of the plasmodesmata into account. Whilst the plasmodesmograms presented here do not offer conclusive evidence for the loading pathway, using either percentage plasmodesmata/~tm cell wall, or percentage plasmodesmata/~m vein as determinants of plasmodesmatal frequency, nonetheless, interpretation of the plasmodesmatal frequency at the interfaces between mesophyll and sieve tubes is greatly facilitated by their use. Despite the drawbacks of plasmodesmograms

366

C.E.J. Botha and A.J.E. van Bel: Plasmodesmograms: diagnostic value for phloem-loading pathways

(Fisher 1990a, b; van Bel a n d G a m a l e i 1991), such visual presentations o f intercellular connections m a y assist researchers to obtain u n e q u i v o c a l answers to a n u m b e r o f i m p o r t a n t questions a n d perhaps, to formulate new experiments a n d to interpret the results soundly. F o r example, is p l a s m o d e s m a t a l frequency a valid p a r a m t e r for the determination o f cell-to-cell transfer? D o sugars travel symplastically f r o m cell-to-cell via p l a s m o d e s m a t a ? Are pictograms o f p l a s m o d e s m a t a l frequencies relevant in the interpretation o f the p h l o e m - l o a d i n g p a t h w a y ? Is the p h l o e m - l o a d i n g p a t h w a y exclusively symplastic or apoplastic or a c o m b i n a t i o n o f the two? Despite the statement by R o b a r d s a n d Lucas (1990) that plasm o d e s m o g r a m s m a y be 'mere snapshots o f cytoplasmic continuity', p l a s m o d e s m o g r a m s indisputably give a clear indication o f the relative cytoplasmic continuity or apoplastic isolation existing between cells in the phloemloading p a t h w a y . Clearly, the absence of, or low frequency o f plasm o d e s m a t a at a particular interface provides conclusive evidence for apoplastic transfer between these cells. Recent experimental evidence (van Bel and Gamalei 1991) has p r o v i d e d evidence in s u p p o r t o f the coincidence o f p l a s m o d e s m o g r a m s , and the loading p a t h w a y , in families with 'symplastic' and 'apoplastic' m i n o r vein configurations. These observations have been extended to include 14CO2-1oading experiments in a n u m b e r o f symplastic and apoplastic m i n o r vein-type plants as defined by G a m a l e i (1985). It is n o t e w o r t h y that loading is prevented by p-chloromercuribenzenesulfonic acid in several apoplastic species, thus supporting the evidence gained f r o m the earlier p l a s m o d e s m o g r a m s indicating an apoplastic-loading, m i n o r vein configuration (van Bel et al. 1992, in press). Concludin9 remarks. T h e p l a s m o d e s m o g r a m s presented here are useful a n d d e m o n s t r a t e unequivocally the symplastic isolation o f the C C - S T complex in grasses (Botha 1990), which is indicative o f their d r y l a n d g r o w t h habitat, and the variability o f the C C ~ S T p a t h w a y in the d i c o t y l e d o n o u s species examined here. The symplastic isolation o f the C C - S T complex in grasses m a y represent an a d a p t a t i o n o f these grasses to load photoassimilate apoplastically u n d e r u n f a v o u r a b l e conditions. The presence o f p l a s m o d e s m a t a does not, therefore, seem to be an essential feature o f p h l o e m loading, at least in the grasses discussed in the present work, where low p l a s m o d e s m a t a l frequencies quite likely indicate an apoplastic loading step.

The Foundation for Research Development (FRD, Pretoria, South Africa), and the Rhodes University Council are gratefully acknowledged for their generous financial support. The first author wishes to acknowledge Mr. Barry Hartley for his valued assistance in the preparation of this paper.

References Botha, C.E.J. (1990) Plasmodesmatal structure and frequency in relation to assimilation in C 3 and C4 grasses in southern Africa. In: Proceedings of the International Conference on Phloem Transport and Assimilate Compartmentation, Cognac (France), p. 9, Bonnemain, J.L., ed. Univ. de Poitiers, France

Erwee, M.G., Goodwin, P.B., van Bel, A.J.E. (1985) Cell-to-cell communication in leaves of Commelina cyanea and other plants. Plant Cell Environ. 8, 173 178 Esau, K. (1977) Anatomy of seed plants, 2nd edn., John Wiley and Sons, New York Evert, R.F., Mierzwa, R.J. (1986) Pathway(s) of assimilate movement from mesophyll cells to sieve tubes in the Beta vulgaris leaf. In: Phloem transport, pp. 419-432, Cronshaw, J., Lucas, W.J., Giaquinta R.T., eds. Alan R. Liss, New York Fellows, R.J., Geiger, D.R. (1974) Structural and physiological changes in sugar beet leaves during sink to source conversion. Plant Physiol. 54, 877-885 Fisher, D.G. (1986) Ultrastructure, plasmodesmatal frequency and solute concentrations in green areas of variegated Coleus blumei Benth. leaves. Planta 169, 141-152 Fisher, D.G. (1988) Movement of Lucifer Yellow in leaves of Coleus blumei Benth. Plant Cell Environ. 11, 639-644 Fisher, D.G. (1990a) Distribution of plasmodesmata in leaves. A comparison of Cananga odorata with other species using different measures of plasmodesmatal frequency. In: Parallels in cell to cell junctions in plants and animals (Proc. NATO Adv. Res. Workshop: NATO ASI Series. Series H, Cell Biology, vol. 46), pp. 199-221, Robards, A.W., Lucas, W., Pitts, J.D., Jongsma, H.J., Spray, D.C., eds. Springer-Verlag, Berlin Fisher, D.G. (1990b) Symplastic continuity in the leaf of Sonchus oleraeeus (Asteraceae), The significance of plasmodesmata in minor veins containing transfer cells. In: Proceedings of the International Conference on Phloem Transport and Assimilate Compartmentation, Cognac (France), p. 11. Bonnemain, J.L., ed. Univ. de Poitiers, France Gamalei, Y.V. (1985) Characteristics of phloem loading in woody and herbaceous plants. Sov. Plant Physiol. 32, 656-665 Giaquinta, R.T. (1983) Phloem loading of sucrose. Annu. Rev. Plant Physiol. 34, 347-383 Madore, M.A., Oross, J.W., Lueas, W.J. (1986) Symplastic transport in Ipomea tricolor source leaves, Demonstration of functional symplastic connections from mesophyll to minor veins by a novel dye-tracer method. Plant Physiol. 82, 432-442 Robards, A.W. (1976) Plasmodesmata in higher plants. In: Intercellular communication in plants: Studies on plasmodesmata, pp. 15-57, Gunning, B.E.S., Robards, A.W., eds., SpringerVerlag Heidelberg Robards, A.W. Lucas, W.J. (1990) Plasmodesmata. Annu. Rev. Plant Physiol. Plant Mol. Biol. 41, 369-419 Russin, W.A., Evert, R.F. (1985a) Studies on the leaf of Populus deltoides (Salicaceae): morphology and anatomy. Am. J. Bot. 71, 1398-1414 Russin, W.A., Evert, R.F. (1985b) Studies on the leaf of Populus deltoides (Salicaceae): ultrastructure, plasmodesmatal frequency, and solute concentrations. Am. J. Bot. 72, 1232-1247 Turgeon R. (1989) The sink-source transition in leaves. Annu. Rev. Plant Physiol. Plant Mol. Biol. 40, 119-138 van Bel, A.J.E., van Kesteren, W.J.P., Papenhuizen, C. (1988) Ultrastructural indications for coexistence of symplastic and apoplastic phloem loading in Commelina benghalensis leaves. Planta 176, 159-172 van Bel, A.J.E., Gamalei, Y.V. (1991) Multiprogrammed phloem loading. In : Recent advances in phloem transport and assimilate compartmentation, pp. 128-139, Bonnemain, JL., Delrot, S., Lucas, W.J., Dainty, J., eds. Quest Editions Presses Acad6miques, Nantes, France van Bel, A.J.E., Gamalei, Y.V., Ammerlaan, A., Bik, L.P.M. (1992) Dissimilar phloem loading in leaves with symplastic and apoplastic minor vein configurations. Planta 186, 518 525 van Kesteren, W.J.P., van der Schoot, C., van Bel, A.J.E. (1988) Symplastic transfer of fluorescent dyes from mesophyll to sieve tube in stripped leaf tissue and partly isolated minor veins of Commelina benghalensis, Plant Physiol. 88, 667-670 Warmbrodt, R.D., Van der Woude, W.J. (1990) Leaf of Spinacia oleracea (spinach): Ultrastructure, and plasmodesmatal distribution and frequency, in relation to sieve tube loading. Am. J. Bot. 77, 1361 1377