Crassulacean acid metabolism-cycling in Euphorbia milii

Photosynthetic characteristics of Euphorbia milii are reported for the first time. The occurrence of CAM-cycling is shown to serve as a mechanism of water conservation. This is a detailed and novel report of such CAM mode in the genus, which is abundant in constitutive CAM and C4 species. Anatomical evidence for the possible operation of a CO2-concentrating mechanism around the vascular bundle, the C2 route, is provided. Findings are important for our understanding of evolution of CAM in the genus.


Introduction
Crassulacean acid metabolism (CAM) is of frequent occurrence among the Euphorbiaceae and has appeared polyphyletically several times in the family, particularly in the genus Euphorbia. In this genus, C 4 species seem to be rare, whereas they are abundant in the genus Chamaesyse (Webster et al. 1975). In Euphorbia, CAM has been reported in 21 species, and values of d 13 C suggest its presence in 44 species (Table 1). Several of these species belong to three different clades within the genus (cladograms in Zimmermann et al. 2010).
Twenty-four species can be considered constitutive CAM on the basis of having values of d 13 C higher than 217 ‰, a criterion established by Mooney et al. (1977). In the remaining species, values of d 13 C average 224.7 ‰. A value as low as 228.9 ‰ found in E. aphylla falls into the lower mode of the bimodal frequency distribution of d 13 C in CAM plants, designated as low-level (weak) CAM (Winter and Holtum 2002;Silvera et al. 2005).
Since values of d 13 C alone are not sufficient to distinguish between C 3 species and plants that obtain up to one-third of their carbon during the night, which include weak CAM plants (Winter and Holtum 2002), measurements of physiological and biochemical variables are necessary. In order to demonstrate the operation of CAM, routine determinations include, among others, DH + , d 13 C and nocturnal CO 2 fixation. Griffiths et al. (2007) devised an ingenious method of ascertaining the occurrence of nocturnal CO 2 fixation by examining the response of the night-time CO 2 exchange rate to intercellular CO 2 concentration (C i ).
Intermediate values of d 13 C can also suggest the occurrence of C 3 metabolism with high water-use efficiency, as the data of Farquhar and Richards (1984) on wheat indicate, or of C 2 photosynthesis, as in the case of Euphorbia acuta. In wheat and maize, C 2 photosynthesis is responsible for an increase of 8 -11 % in photosynthetic rate through re-assimilation of photorespired CO 2 (Busch et al. 2013).
Plants of Euphorbia milii subgenus Euphorbia, Section Goniostema, common name crown of thorns, originally from Madagascar, are cultivated worldwide for their ornamental value. Plants are perennial armed shrubs as tall as 1 m, with fleshy stem and branches, and partly succulent leaves. According to observations by Mooney et al. (1977), CAM is present in the weak mode in leafy species of the genus. The medicinal and molluscicidal properties of the latex in E. milii have been extensively investigated (e.g. Mwine and Van Damme 2011); in contrast, literature on the physiology of the species is practically non-existent. In spite of the succulence of its leaves and the various reports of CAM in the genus, E. milii has been reported as non-CAM (Webster et al. 1975). Nevertheless, recalculation of the data of McWilliams (1970) gives a DH + of 100 mmol (g fresh mass) 21 and a dark CO 2 fixation rate of 0.1 mmol m 22 s 21 , suggesting that CAM in E. milii operates in the cycling mode, i.e. nocturnal H + accumulation and daytime but nearly no night-time CO 2 fixation (for the definition of CAM modes, see Cushman 2001).
With the aim of contributing to our knowledge of the evolution of CAM in Euphorbia, this study re-examined the possible occurrence of CAM in E. milii through daily leaf gas exchange, including P N /C i and R/C i curves (where P N is the photosynthetic rate and R is the respiration rate), measurements of dawn and dusk H + content, and determinations of d 13 C.

Plant material and cultivation
Plants of E. milii were propagated from one plant purchased at a nursery by inserting cuttings into the soil of 2-L pots filled with silty clay loam (Viveros Exotica Raphia, S.R.L., Caracas); plants were fertilized monthly with N : P : K 15 : 15 : 15 and grown in the garden for 1 year before the beginning of experiments. Plants, 50 cm tall, were maintained in the greenhouse under natural light, fully watered every other day and fertilized weekly. Day length was 12 h (06:00-18:00 h), mean maximum daily photosynthetic photon flux density (PPFD) between 09:00 and 14:30 h 507 + 22 mmol m 22 s 21 , mean air temperature 32 + 5/18.4 + 0.5 8C (day/night) and relative humidity 60 + 10 %. Water deficit was imposed by withholding watering.

Anatomy
Free-hand cross-sections of stems (average thickness 5 mm) and leaves were observed under the microscope at ×40 (stem) and ×400 (leaf). Leaf sections were stained with toluidine blue.

Succulence
Leaf and stem water content was determined as the difference between the fresh mass (FM) and the mass after drying for 72 h at 60 8C [dry mass (DM)], divided by the area in the case of leaves (FM/A) and by DM in the case of stems. Leaf thickness was measured with precision calipers. Chlorophyll (Chl) content was determined after Bruinsma (1963) in 80 % cold acetone extracts of leaf or stem sections collected at 18:00 h. The mesophyll succulence index was calculated as S m ¼ g water (mg Chl) 21 after Kluge and Ting (1978).

Stable carbon isotope composition
The d 13 C was determined with a precision of 0.15 ‰ using a ThermoFinnigan DeltaPlusXL Isotope Ratio Mass Spectrometer (San Jose, CA, USA) and PDB as the standard.

Nocturnal H 1 accumulation
Whole leaves were weighed fresh and set to boil in 50 mL distilled water for 10 min in a microwave oven at maximum power; samples were sieved through a plastic colander, leaf segments and the colander were rinsed, and the solution was made up to 100 mL. Samples were titrated to pH 7.0 for the estimation of H + corresponding to malate according to Nobel (1988), and to pH 8.4 for citrate. Since Franco et al. (1990) noted that there was a strong linear relationship between concentrations of malate and citrate determined enzymatically and by titration, in the absence of an enzymatic method for the determination, titration is an adequate alternative. Latex was collected from cut stems and leaves, suspended in distilled water and titrated likewise. The DH + was calculated as the difference between dawn and dusk H + contents.

Leaf gas exchange
The P N , R, stomatal conductance (g s ) and transpiration rate (E) were measured in the laboratory with a CIRAS 2 IRGA connected to a PLC(B) assimilation chamber (PP Systems, Amesbury, MA, USA) at an incoming CO 2 concentration (C a ) of 380 mmol mol 21 , a chamber temperature tracking ambient (24 + 1 8C) and an incident PPFD of 200 (the first morning hours) or 1000 mmol m 22 s 21 (the rest of the daytime). Records were automatically taken every 30 min. Response curves were done in six different leaves of P N and C i between 10:00 and 11:00 h, and of nocturnal CO 2 exchange to C i between 20:00 and 05:00 h.

Statistics
Values are mean + SE (n ¼ 6). Statistical significance was assessed where indicated through one-or two-way analysis of variance (ANOVA) (P , 0.05) with the Statistica package.

Results
Leaf cross-sections showed a dorsiventral anatomy, with a compact palisade parenchyma containing many chloroplasts and a spongy parenchyma with large vacuoles and fewer chloroplasts; the spongy parenchyma constituted 40 % of the whole-leaf thickness (Fig. 1). A bundle sheath with large chloroplasts located centrifugally was observed. Cross-sections of the fleshy stems AoB PLANTS www.aobplants.oxfordjournals.org (not shown) have a thick green cortex and colourless pith; the cortex was 54 % of the stem thickness on average.
In watered plants, S m in both leaves and stem green cortex was 2.7 + 0.2 g water (mg Chl) 21 ; d 13 C was 225.2 + 0.7 ‰ in leaves and 224.7 + 0.1 ‰ in stems. In the stem green cortex of watered plants, malateand citrate-H + content was 48 + 9 and 29 + 5 mmol H + (g FM) 21 , respectively, without daily oscillation. Water suspensions of latex showed no acid content.
Leaves had significant amounts of malate-and citrate-H + at dawn and dusk, contents increasing with time under drought ( Fig. 2A and B). A significant accumulation of malate-H + took place in watered plants, which remained constant up to 16 days of drought (P , 0.05). A similar trend in dawn and dusk H + content and DH + for citrate-H + was found, except that after 16 days of drought DH + became zero. Mean DH + was 18 + 2 (malate) and 18 + 4 (citrate) mmol (g FM) 21 . Changes in either morning and evening H + contents or DH + bore no relationship to changes in FM/ A, which remained relatively constant for the duration of the experiment, as did Chl content (Fig. 2). The ratio Chl a/b remained unchanged at 3.3 + 0.2. Stem water content was 11.2 + 0.5 g water (g DM) 21 , twice as high as in leaves, and did not vary with time under drought (P ¼ 0.86).
As shown in Fig. 3A, P N of watered plants became saturated at 700 mmol m 22 s 21 PPFD; apparent quantum yield was 0.047 and light-compensation point 48 mmol m 22 s 21 . The P N /C i curves (Fig. 3B) show that P N did not become saturated by C i , increasing 47 % with an increase in C i to 800 mmol mol 21 (C a ¼ 1240 mmol mol 21 ). This lack of saturation could have been due to very low g s , which remained unchanged by C i . The CO 2 compensation concentration was 32 mmol mol 21 .
Daily courses of leaf gas exchange done in plants progressively under drought showed a decrease in P N of 85 % with drought; R became nearly zero after 12 and up to 16 days without watering (Fig. 4). Mean daytime water-use efficiency calculated from these courses of leaf gas exchange was relatively high, decreasing significantly with drought only 18 %, from 4.2 + 0.1 to 3.5 + 0.1 mmol mol 21 (P ¼ 0.00).
Stem cross-sections of 1.5 cm 2 average area from watered plants introduced in the assimilation chamber showed daytime CO 2 assimilation at rates similar to those determined in leaves on a Chl basis (Table 2). Stem sections, as opposed to leaves, showed dark CO 2 uptake.  A significant decrease of 65 % in R with C i was observed without significant changes in g s (Fig. 5).
The regression of E vs. R is shown in Fig. 6. Assuming that the accumulated acids were the products of the recycling of respiratory CO 2 , the absolute recycling, i.e. the amount of CO 2 contained in acids, was calculated. Together with the E vs. R regression, it was found that recycling recovered 10 and 37 % of nocturnal CO 2 loss in watered plants and after 12 days of drought, respectively, and helped in saving water during the night by 15 % in watered plants and 2 % in plants under drought. Daytime water saving, calculated as the ratio of the absolute amount of CO 2 in acid equivalents to   Values of DH + determined at pH 7.0 were low compared with CAM species such as Kalanchoe tubiflora and Clusia minor, comparable with the cycling species T. parviflorum and T. mengessi and not as low as in Talinum teretifolium or Zamioculcas zamiifolia (Table 3).
A value of S m higher than 1 g water (mg Chl) 21 was also suggestive of CAM, as proposed by Kluge and Ting (1978). The succulent nature of leaves was corroborated by the microscopic observation of cross-sections, in which cells with a large volume and few chloroplasts are present, as in many leaf-succulent CAM plants (Kluge and Ting 1978). In facultative CAM species, values of S m are intermediate (Table 4) but, given that   Some values were re-calculated from the data in references. ND, not determined. low as well as high values of S m can be found in constitutive CAM species (Table 4), it becomes apparent that for a leaf to perform full CAM a large proportion of vacuole volume to chloroplasts is not required. The lack of significant differences in FM/A between five strong CAM and three weak CAM species (Nelson and Sage 2008) lends support to this hypothesis. The presence in E. milii of bundle sheath cells with chloroplasts is indicative of the possible operation of C 2 photosynthesis, as reported in E. acuta. This species shares with E. milii low, C 3 -like values of d 13 C (225.5 ‰ according to Webster et al. 1975, and228.5 ‰ according to Sage et al. 2011) and an intermediate value of CO 2 compensation concentration (32 mmol mol 21 ; Sage et al. 2011). The occurrence of CAM, Kranz anatomy and C 4 photosynthesis in the same leaf has been reported in Portulacaceae (Guralnick and Jackson 2001), but to date no report on CAM together with C 2 photosynthesis has been published. Given that the C 2 route of carbon fixation has been proposed as an intermediate evolutionary step from C 3 to C 4 (Sage et al. 2011), investigating the functioning of C 2 photosynthesis in E. milii would bring interesting viewpoints on the evolution of CAM and C 4 plants, particularly in Euphorbia.
A significant oscillation in H + content corresponding to citrate was found, equivalent to 12 mmol citrate (g FM) 21 after 12 days of drought, comparable with the lower end, 22 mmol (g FM) 21 , of the range in species of Clusia, a genus abundant in CAM species performing different modes (Lü ttge 2007). The role of citrate accumulation in carbon or water balance during CAM remains unclear (Lü ttge 2007); citrate does not provide net CO 2 gain, as does malate, but should prove more effective than malate in increasing C i during the day because its breakdown produces three molecules of CO 2 as opposed to one in the case of malate (Lü ttge 2007). Increased C i during the day is a photoprotective mechanism well recognized in CAM plants, as shown in plants of the facultative CAM species Talinum triangulare under drought (Pieters et al. 2003).
Photosynthetic characteristics in E. milii were consistent with those of a sun plant: high apparent quantum yield, saturating PPFD and Chl a/b ratio (Pearcy and Franceschi 1986).
The absence of net dark CO 2 fixation was consistent with the proportion of dark CO 2 uptake calculated using the regression equations of the proportion of CO 2 fixed during the night against d 13 C found by Winter and Holtum (2002) and Pierce et al. (2002). In many facultative CAM species, d 13 C tends towards low values. In a review of 23 facultative CAM species (Herrera 2009), the mean, maximum and minimum d 13 C were 223.9, 214.0 and 230.0 ‰, respectively, indicating that the variability in d 13 C values may lead researchers to classify a species as a C 3 , facultative or constitutive CAM plant. In Euphorbia aphylla, d 13 C ranged from 227.1 ‰ for the youngest cladode in the dry season during summer to 224.5 ‰ for the oldest cladode in the rainy season during winter (Mies et al. 1996), reflecting the effects on d 13 C of day/night temperatures, water availability and developmental stage.
Values of d 13 C higher in stems than in leaves suggest the occurrence of nocturnal CO 2 fixation, although this could not be demonstrated in intact plants. The occurrence of an assimilation rate in the dark amounting to a third of PPFD-saturated leaf P N strongly suggested that stems are capable of nocturnal CO 2 fixation. Stem internal CO 2 re-fixation in young twigs and branches possessing a green cortex may compensate for 60-90 % of the potential respiratory carbon loss (Pfanz et al. 2002). If stem recycling in E. milii occurred through phosphoenolpyruvate carboxylase (PEPC) activity, that would explain the apparent 13 C enrichment. There are two alternative explanations for higher d 13 C in the stems of E. milii. One explanation is that this variable was determined in sections comprising all tissues, green as well as non-autotrophic; non-autotrophic organs of C 3 plants, such as stems, have been found to be enriched in d 13 C by 1 -3 ‰ relative to leaves (Cernusak et al. 2009). Another, simpler, explanation is that barriers to CO 2 diffusion into the stem are larger than into leaves, hindering entrance of the heavier 13 CO 2 . Actual nocturnal CO 2 fixation by stems of E. milii remains to be determined accurately by methods such as carbon labelling, involving all stem tissues.
Sedum morganianum 13.0 Constitutive Kluge and Ting (1978) AoB PLANTS www.aobplants.oxfordjournals.org The response of leaf dark respiration to C i suggests the operation of a carboxylation system, most likely PEPC, which makes recycling of respiratory CO 2 possible. In the constitutive CAM plant Kalanchoe daigremontiana, a P N /C i curve during phase I of the CAM cycle showed a pronounced increase at low C i and saturation at a C i of 250 mmol mol 21 (Griffiths et al. 2007). Our results show that the response of dark respiration to C i was not an artefact caused by changes in g s , because g s remained constant in spite of increasing C i .
Water saving through respiratory CO 2 recycling was significant, as in the case of T. paniculatum, a facultative CAM species, in which the amount of water saved was 5 -12 times that lost by transpiration (Gü erere et al. 1996). Similarly, in T. calycinum, 5-44 % of water was potentially conserved by CAM-cycling (Martin et al. 1988).
Leaf water balance in E. milii seems to rest on both recycling of respiratory CO 2 and strict stomatal control, rather than on water supply from the succulent stem, as leaf FM/A remained unchanged after 16 days of drought and stem water content did not vary significantly during this time.

Conclusions
In view of the observations presented here, E. milii can be considered as a CAM-cycling species that in watered plants shows diurnal, but not nocturnal, CO 2 uptake and low DH + ; plants under drought have very low P N , equally low DH + and no net dark CO 2 exchange. The significance of the operation of such a low CAM in E. milii resides in water conservation, rather than carbon acquisition. The occurrence of C 2 photosynthesis remains to be demonstrated.

Conflicts of Interest Statement
None declared.