Light-dependent maintenance of hydraulic function in mangrove branches: do xylary chloroplasts play a role in embolism repair?


  • N. Schmitz,

    1. Laboratory for Plant Biology and Nature Management, Vrije Universiteit Brussel, 1050 Brussels, Belgium
    2. Royal Museum for Central Africa, Laboratory for Wood Biology and Xylarium, Leuvensesteenweg 13, 3080 Tervuren, Belgium
    3. Plant Science Division, Research School of Biology, The Australian National University, Canberra, ACT 0200, Australia
    Search for more papers by this author
  • J. J. G. Egerton,

    1. Plant Science Division, Research School of Biology, The Australian National University, Canberra, ACT 0200, Australia
    Search for more papers by this author
  • C. E. Lovelock,

    1. The School of Biological Science, The University of Queensland, Brisbane, Qld 4072, Australia
    Search for more papers by this author
  • M. C. Ball

    1. Plant Science Division, Research School of Biology, The Australian National University, Canberra, ACT 0200, Australia
    Search for more papers by this author

Author for correspondence:
Nele Schmitz
Tel: +32 2 629 34 14


  • To clarify the role of branch photosynthesis in tree functioning, the presence and function of chloroplasts in branch xylem tissue were studied in a diverse range of mangrove species growing in Australia.
  • The presence of xylary chloroplasts was observed via chlorophyll fluorescence of transverse sections. Paired, attached branches were selected to study the effects of covering branches with aluminium foil on the gas exchange characteristics of leaves and the hydraulic conductivity of branches.
  • Xylary chloroplasts occurred in all species, but were differently distributed among living cell types in the xylem. Covering stems altered the gas exchange characteristics of leaves, such that water-use efficiency was greater in exposed leaves of covered than of uncovered branches.
  • Leaf-specific hydraulic conductivity of stems was lower in covered than in uncovered branches, implicating stem photosynthesis in the maintenance of hydraulic function. Given their proximity to xylem vessels, we suggest that xylary chloroplasts may play a role in light-dependent repair of embolized xylem vessels.


Many woody species have photosynthetic stems, at least while the stems are relatively young. Stem photosynthesis can contribute to the growth of trunks and development of buds in young woody plants (Saveyn et al., 2010). Most studies, however, have emphasized a role of stem photosynthesis in refixation of respired CO2. This would reduce carbon costs associated with maintenance of living tissues in stems, with far-reaching implications for stem survival when foliar photosynthesis may be limited by environmental conditions such as drought (Comstock et al., 1988; Teskey et al., 2008; Wittmann & Pfanz, 2008; McGuire et al., 2009).

Most attention has been given to corticular photosynthesis, particularly as high concentrations of chloroplasts can give stems a green colour. Less obvious, however, are the chloroplasts within living tissues of the xylem (Wiebe et al., 1974; Wiebe, 1975). Using chlorophyll fluorescence, Dima et al. (2006) found xylary chloroplasts to be common in a range of woody species in a Mediterranean habitat. These observations invite questions about the potential role of xylary chloroplasts in repair of embolisms and maintenance of hydraulic function. These questions arise because recent studies have demonstrated that hydrolysis of starch in xylem parenchyma occurs coincident with embolism repair (Salleo et al., 2009; Zwieniecki & Holbrook, 2009; Nardini et al., 2011; Secchi & Zwieniecki, 2011) and that prolonged darkness inhibits embolism repair in intact rice plants (Stiller et al., 2005).

In the present study, we used chlorophyll fluorescence to determine the distribution of xylary chloroplasts in stems of a diverse range of mangrove species growing naturally in wet and arid estuarine forests in Australia. We hypothesized that xylary chloroplasts would be more common in species with white than with dark bark, and in species growing in more saline habitats and under more arid climatic conditions. We also covered branches to test two hypotheses about the functions of branch photosynthesis in mangrove species. First, if branch photosynthesis contributes to the carbon status of branches, then we predicted that covering branches would increase foliar photosynthesis, presumably in response to an increase in the demand for assimilates to sustain covered branch tissues. Secondly, if xylary photosynthesis contributes to the maintenance of hydraulic function, then we predicted that hydraulic conductivity would be lower in covered than in uncovered branches.

Materials and Methods

Study sites and sample collection

The study was conducted in mangrove forests along the estuarine flood plains of the Daintree River (March and July 2010) in the wet tropics of Far North Queensland, Australia, (Lat S 16°20′, Long E 145°30′) and Giralia Bay (August 2010) in arid Western Australia (Lat S 22°43′, Long E 114°34′) (Lovelock et al., 2011). At the Daintree site we studied 13 species, of which only three were available for study at Giralia Bay (Supporting Information, Table S1). Avicennia marina was studied at three different localities in Giralia Bay that, together with the Daintree site, can be ordered according to relative soil water salinity (Table 1).

Table 1.   Data used to calculate rates of branch photosynthesis in five mangrove species assuming that increased assimilation rates in leaves compensated for losses in assimilation rates of branches covered with aluminium foil
SpeciesnCovering period (d)Covered StemS (cm²)Covered XylemV (cm3)Covered BarkV (cm3)Branch LeafS (cm²)AL(U) (μmol CO2 s−1 m−2)dAL (%)AB (μmol CO2 s−1 m−2)
  1. StemS, total stem surface area covered with aluminium foil; XylemV, the associated xylem volume; BarkV, bark volume; Branch LeafS, total leaf area of covered branches.

  2. A L, leaf assimilation rate of the associated leaves of the covered (C) and uncovered branches (U); dAL, the change in leaf assimilation rate after branch covering; and AB, the assimilation rate of the branch, where AB = [(AL(C) − AL(U)) × Branch LeafS] × Covered StemS−1.

  3. Data for Avicennia marina were collected at four sites differing in soil water salinity (ppt). Values are means ± 1 SE. Overall means were calculated as the mean of the mean values for each of the five species and sites (for A. marina).

Rhizophora apiculata5542.6 ± 8.42.8 ± 0.94.3 ± 1.0425.7 ± 80.712.9 ± 0.99.2 ± 4.211.1 ± 5.1
Ceriops australis5647.2 ± 6.73.3 ± 0.84.2 ± 0.412.4 ± 0.74.8 ± 0.2−9.5 ± 9.6−0.1 ± 0.1
Aegiceras corniculatum4467.7 ± 4.44.6 ± 0.65.1 ± 0.693.5 ± 16.14.9 ± 1.136.2 ± 35.40.7 ± 0.9
Rhizophora stylosa44112.4 ± 22.07.8 ± 2.512.6 ± 2.8168.3 ± 61.44.8 ± 1.437.0 ± 52.7−1.9 ± 2.1
Avicennia marina
 66 ppt5650.4 ± 4.74.5 ± 0.72.9 ± 0.3139.6 ± 20.27.0 ± 0.210.9 ± 4.52.4 ± 1.2
 55 ppt5674.3 ± 17.08.4 ± 3.24.2 ± 1.0163.7 ± 17.07.2 ± 0.5−11.5 ± 5.5−2.6 ± 1.2
 42 ppt64102.7 ± 16.613.0 ± 3.76.2 ± 1.2297.2 ± 49.98.2 ± 0.95.4 ± 24.7−0.8 ± 4.3
 17 ppt53187.4 ± 33.437.9 ± 9.210.5 ± 2.51652.4 ± 396.116.4 ± 0.64.3 ± 3.74.8 ± 6.0
Average of 5 species5585.6 ± 14.110.3 ± 2.76.2 ± 1.2369.1 ± 80.38.3 ± 0.710.3 ± 17.51.7 ± 2.6

Chlorophyll fluorescence

Chlorophyll fluorescence was used to detect the presence of chloroplasts in branch xylem. The trees were growing naturally in soils with pore water salinities ranging from 1 to > 50 ppt. Branches of c. 6 mm diameter were collected, with additional sampling of smaller or larger branches (range: 2.8–14 mm) when the 6 mm branch either lacked chlorophyll fluorescence in the wood or the intensity of the fluorescence decreased from pith to bark. Branch samples of different bark thickness (range: 0–2.3 mm, Table S1), different bark colour (white, green, brown) and bark texture (smooth or rough and fissured) were also studied. Transverse sections (50–100 μm thickness) were cut from freshly sampled branches with a field portable microtome and placed in a drop of water on glass slides for observation with a Zeiss Axiostar Plus, modified to support a Walz Chlorophyll Fluorometer Imaging PAM M-Series Microscopy System IV (Effeltrich, Germany). Images were collected of the spatial distribution of the fluorescence yield in response to a saturating pulse of light and compared with images of the same xylem tissue, under near-infrared light, to identify the sources of chlorophyll fluorescence emissions. To interpret the fluorescence images, additional wood sections (20–30 μm thickness) were cut and double-stained with safranin-alcian blue to visualize the amount of ray and axial parenchyma in the xylem tissue of each species.

Chloroplast role in branch functionality

An experiment where branches were shaded from light was performed on all three species growing at Giralia Bay and on Rhizophora apiculata, Ceriops australis and A. marina growing at the Daintree site (Table 1). One pair of fully exposed branches, each c. 6 mm diameter and bearing a similar leaf area, was selected on each of five trees per species and per site. One branch of each pair was covered with aluminium foil up to the apical bud, leaving the leaves uncovered. The branches were allowed to incubate for several days before gas exchange characteristics of the leaves were measured. Finally, branches were harvested for measurement of hydraulic conductivity.

Gas exchange characteristics of fully exposed, mature leaves were measured in the morning, with a Li-Cor 6400 Portable Photosynthesis System (Li-Cor Corp, Lincoln, NE, USA). Measurements were made under ambient conditions except that the incident light intensity was maintained at 1000 μmol quanta m−2 s−1. Upon completion of all gas exchange measurements, the branches were collected and immediately transported to the laboratory for measurement of hydraulic conductivity using an optical technique as previously described (Choat et al., 2011). A pressure head of 40–70 cm, generating a pressure gradient of 4–7 kPa, was imposed on branches of 8–16 cm length. Measurements were made with a perfusion solution of 1% filtered sea water (Stuart et al., 2007) to mimic the ionic composition of xylem sap in most mangrove species (Scholander et al., 1966; Ball, 1988). The flow rate was averaged over a period of 5 min with calculations based on images collected every 30 s to track the movement of the meniscus of a coloured solution in a pipette connected to the branch outlet. Leaf-specific hydraulic conductivity was then calculated as the flow rate per pressure gradient, standardized for branch length and total leaf area of the branch (KL). Images were taken of cross-sections of the branches to measure areas occupied by bark, xylem (excluding the pith) and total stem area in ImageJ 1.43u ( to calculate the respective volumes that were shaded (Table 1).


To test if the changes in gas exchange characteristics were significantly different between covered and uncovered branches a paired t-test was performed. The potential interacting effect of species and site was tested via a one-way ANOVA. The differences in KL between branches within a branch pair were tested using a paired t-test. In cases where data were not normally distributed a log transformation was performed. The match between KL values of covered and uncovered branches was tested, after log transformation, by a simple linear regression. The effect of species and sites, across species (arid vs wet) and within A. marina (sites of different soil water salinity, see Table 1), was tested via an analysis of covariance (ANCOVA).


Distribution of chloroplasts in branch xylem

Chlorophyll fluorescence was detected in the xylem of all studied tree species, although to a lesser extent in Bruguiera sexangula and Rhizophora stylosa than in the other species. The maximum diameter of branches in which fluorescence was observed in the wood varied from 6 to 13 mm across the studied species (Table S1). Within a branch, chlorophyll fluorescence typically decreased with depth from bark to pith, except in Ceriops spp. where branches with a relatively thick bark only showed chlorophyll fluorescence near the pith. Between branches of increasing bark thickness, the distribution of chlorophyll fluorescence in the xylem generally decreased, from being present along the entire branch radius to only near the bark. However, the presence of chlorophyll fluorescence and that of bark thickness were not strictly related. Also branch diameter and colour of the bark were unrelated to the distribution of chlorophyll fluorescence. Even within a species, branches that showed chlorophyll fluorescence could be thicker and/or have a thicker bark than branches without fluorescence.

No relationship was found between the spatial distribution of axial parenchyma and chlorophyll fluorescence (Fig. 1). Eight species showed chlorophyll fluorescence only in the rays (Table S1), amongst which Sonneratia caseolaris, Xylocarpus granatum and Heritiera littoralis had moderate to abundant axial parenchyma (Fig. 1c–e). The two studied Ceriops species and A. marina showed chlorophyll fluorescence in rays and vessel-associated parenchyma cells (Table S1, Figs 1b, 3c,d). In R. apiculata, chlorophyll fluorescence occurred only in the vessel-associated cells (Fig. 1a).

Figure 1.

Chlorophyll fluorescence (column 1) demonstrates the presence of xylary chloroplasts in vessel-associated parenchyma cells and ray cells (column 2) in transverse sections of branches of five mangrove species. There is no link between the species-specific distribution of vessel-associated parenchyma cells (column 3) and chlorophyll fluorescence. Species are ordered in terms of relative amounts of vessel-associated parenchyma cells. Transverse wood sections in column 3 are of the same species but from different branches of larger diameter, except for (c) where the branch was smaller than the one used for column 2. Sections were stained with safranin-alcian blue, distinguishing parenchyma (blue) from fibres and vessels (pink). (a) Rhizophora apiculata; (b) Ceriops decandra; (c) Sonneratia caseolaris; (d) Xylocarpus granatum; (e) Heritiera littoralis. Bars, 100 μm.

Effects of covering branches on shoot functions

There were no significant effects of covering branches on either the assimilation rate or stomatal conductance. Nevertheless, covering branches significantly increased leaf-level water-use efficiency (paired = −2.18, = 39, = 0.03), as measured by the ratio of assimilation rate to stomatal conductance, and decreased the intercellular CO2 concentrations (paired = 2.10, = 39, = 0.04) (Fig. 2a). The results were variable (Fig. S1) but there were no significant effects of either species (F = 1.88, df = 4, = 0.14) or sites (F = 0.66, df = 3, = 0.58). Although not significant, the mean assimilation rates of leaves attached to covered branches were 10.3% greater than those of uncovered branches on the same trees (Table 1). This higher mean value was driven by the occurrence of higher assimilation rates in leaves of covered branches in 56% of branch pairs. Nevertheless, maintenance of a constant or greater assimilation rate with decrease in intercellular CO2 concentration and increase in water-use efficiency requires up-regulation of photosynthesis in response to the covering of branches.

Figure 2.

Effects of covering branches with aluminium foil on the gas exchange and hydraulic characteristics of associated leaves and branches, respectively, in five mangrove species. (a) Percentage change in covered vs uncovered branches (means ± 1 SE) in the ratio of CO2 assimilation rate to stomatal conductance (A/g) and the intercellular CO2 concentration (Ci). Species are ordered according to increasing stem area covered (Table 1). (b) Log-transformed leaf-specific hydraulic conductivity (KL), matched well between uncovered (U) and covered (C) branch pairs in an arid (open and grey symbols) and wet (closed symbols) coastal system. Avicennia marina occupied different local sites within the arid system categorized from low to high soil water salinity (see Table 1). (c) Averaging sites and species, the mean (± 1 SE) of the log-transformed leaf-specific hydraulic conductivity (KL) for covered branches was significantly lower than for uncovered branches. Ac, Aegiceras corniculatum; Am, Avicennia marina; Ca, Ceriops australis; Ra, Rhizophora apiculata; Rs, Rhizophora stylosa.

Covering branches also affected leaf -specific hydraulic conductivity of the stems, KL. Despite variability (Fig. S1), all species responded similarly (Fig. 2b, F = 0.38, df = 4, = 0.82) with no significant effect of climatic conditions (Fig. 2b, F = 0.057, df = 1, = 0.81) or of local variations in soil water salinity for the different sites of A. marina (Fig. 2b, F = 2.79, df = 3, = 0.06). The linear relationship between KL of covered and uncovered branches shows that the branches were well matched within each tree, despite the large range of KL values between individual trees of the same and different species (Fig. 2b, = 0.76, F = 38.04, < 0.0001). Averaged over the whole study, KL was significantly lower in covered than in uncovered branches (Fig. 2c, paired = 3.34, = 29, = 0.002). This equated to a median loss in hydraulic conductivity of c. 50% in covered branches (and a mean loss of 26% based on the log10-transformed data, Fig. 2c).


The results of the present study show that xylary chloroplasts are a common feature of mangrove wood (Table S1, Fig. 1). This is consistent with the occurrence of chloroplasts in the wood of species from other biomes (Wiebe et al., 1974; Wiebe, 1975; Wittmann et al., 2001; Pilarski & Tokarz, 2006), including 20 species with diverse phenological behaviour in Greece (Dima et al., 2006).

The spatial distribution of chloroplasts within branches is suggestive of interspecific differences in axial and radial light transmission and guidance through branch tissues. In most species, the rays were the main light guides (Fig. 1b–e, Table S1), transmitting the light from the bark inwards. Additionally, in a few species, with Ceriops spp. as the clearest example, the spatial patterns of chlorophyll fluorescence revealed the presence of abundant chloroplasts in the pith. As all studied branches were relatively close to the apex, this is suggestive of axial light transmission (Sun et al., 2003) through the pith. Similarly, Vancleve et al. (1993) found photosynthetically active chloroplasts in the stem pith of poplar, a woody species with white bark. Our findings indicate that the presence of xylary chloroplasts along the branch radius depended on the light-guiding properties of the bark, partly related to the bark thickness but independent of its colour. Nevertheless, the maximum diameter and stem depth at which xylary chloroplasts occurred could not be linked to any single factor, such as bark thickness or colour, but presumably depended on a complex combination of transmittance and light-guiding properties of wood (Sun et al., 2003) and bark tissues (Trockenbrodt, 1991, 1994).

Photosynthesis by chloroplasts within bark and stem tissues can contribute to the carbon and water balances of these tissues (Pfanz et al., 2002; Aschan & Pfanz, 2003). Stem photosynthesis is believed mainly to involve refixation of respired CO2, thereby reducing carbon losses from the stems. We expected leaf photosynthetic rates to increase if the covering of stems increased their demands for carbon assimilates needed to maintain cellular functions. Contrary to our predictions, covering branches did not significantly increase assimilation rates in associated leaves, although assimilation rates in leaves of covered branches were on average 10.3% higher than those of control, uncovered branches (Table 1). Three explanations could be given for the variable response of the branches (Fig. S1), both within and between species: the short experimental period, relative to the resilience of the photosynthetic apparatus to darkness (Parolin, 2008); the need for the trees to compensate the loss in photosynthates, which is related to the local environmental conditions; and the ability to up-regulate photosynthesis in uncovered parts of the branch instead of in leaves.

How large is the contribution of branch photosynthesis to the carbon balance of shoots? An increase of 10.3% in assimilation rates of leaves attached to covered branches might seem small. However, if we assume that the higher leaf assimilation rates fully compensated for losses in photosynthetic activity of covered branches, and that all leaves within a branch behave similarly, then the average increase in leaf assimilation rates corresponds to an average stem assimilation rate of 1.7 μmol CO2 m−2 s−1 (Table 1). This rate is consistent with a recent estimate based on different methods of 1.5 μmol CO2 m−2 s−1 for corticular photosynthesis in a eucalypt species (Cernusak & Hutley, 2011). Photosynthetic assimilation rates in stems thus could be c. 20% of the average rates in leaves (Table 1; 1.7 μmol CO2 m−2 s−1/8.3 μmol CO2 m−2 s−1 × 100) and could contribute c. 5% of the total assimilation rate of an average branch (Table 1; 1.7 μmol CO2 m−2 s−1 × 85.6 10−4 m2/8.3 μmol CO2 m−2 s−1 × 369.1 10−4 m2). However, it would be incorrect to conclude from these calculations that changes in assimilation rates were sufficient to compensate for losses in stem photosynthesis, as more data would be needed to determine the carbon balance. Nevertheless, the potential contribution of bark and wood photosynthesis to the carbon balance of a tree underscores the suggestion that the unexpected daytime decrease in CO2 efflux, observed in beech and oak, is the result of C refixation (Saveyn et al., 2008). This could be extremely important under natural conditions as bark and wood photosynthesis could support the maintenance costs of the stem during times when environmental conditions constrain photosynthetic activity in leaves (Vick & Young, 2009; Cernusak & Hutley, 2011), such as during drought.

As carbon cannot be gained without the expenditure of water, recapture of respired CO2 by stem photosynthesis effectively reduces shoot water losses, thereby enhancing water-use efficiency (Eyles et al., 2009). Such savings in both carbon and water may contribute to survival in environments where water availability may be limited by seasonal drought or high soil salinity. Indeed, leaf-level water-use efficiency significantly increased within 1 wk of covering branches (Fig. 2a), underscoring the importance of stem photosynthesis for both carbon and water balances.

Finally, the loss in KL following branch covering provides evidence that stem photosynthesis contributes to maintenance of hydraulic function. Like other plant species, mangroves typically operate near their cavitation threshold, which varies with the prevailing salinity of their habitat (Sperry et al., 1988; Melcher et al., 2001). Hence the capacity to repair embolized vessels is essential for maintenance of hydraulic function regardless of whether the plants grow in high- or low-salinity environments. Indeed, diurnal variation in hydraulic conductivity of mangrove stems has been correlated with variation in the extent to which xylem vessels were embolized in both coastal and estuarine environments (Melcher et al., 2001). Refilling of embolized vessels occurs coincident with loss of starch from the xylem parenchyma, consistent with hypothetical use of sugars to create an osmotic driving force to refill embolized vessels (Salleo, 2006; Salleo et al., 2009; Zwieniecki & Holbrook, 2009). The effects of corticular and xylary photosynthesis are confounded in the present study. Nevertheless, the distribution of xylary chloroplasts in rays and in axial parenchyma cells associated with vessels, which are more abundant in trees from the arid than the wet site (Fig. 3), is consistent with a role for xylary chloroplasts in the local provision of energy and photosynthates for embolism repair, and merits further study. Thus, while their contribution to stem photosynthesis is overshadowed by corticular activity, xylary chloroplasts may play a critical role in the carbon and water balances of plants.

Figure 3.

The relative amount of vessel-associated parenchyma cells in Avicennia marina branch wood is higher in a site of relatively high salinity (a) than in one of low (b) soil water salinity. Transverse wood sections are stained with safranin-alcian blue, distinguishing parenchyma (blue) from fibres and vessels (pink). Chlorophyll fluorescence (c) indicates the presence of chloroplasts in both rays and vessel-associated parenchyma cells (d). Bars, 100 μm.


We thank Catherine Bone and Nigel Brothers for their invaluable assistance with field work and Brendan Choat for assistance with setting up the system to measure hydraulic conductivity. N.S. was funded by a postdoctoral fellowship and a mobility grant from The Research Foundation – Flanders (FWO). This research was supported by Australian Research Council Discovery Project Grant DP1096749 to M.C.B. and C.E.L.