Sources of carbon supporting the fast growth of developing immature moso bamboo (Phyllostachys edulis) culms: inference from carbon isotopes and anatomy

Abstract Phyllostachys edulis is a spectacularly fast-growing species that completes its height growth within 2 months after the shoot emerges without producing leaves (fast-growing period, FGP). This phase was considered heterotrophic, with the carbon necessary for the growth being transferred from the mature culms via the rhizomes, although previous studies observed key enzymes and anatomical features related to C4-carbon fixation in developing culms. We tested whether C4-photosynthesis or dark-CO2 fixation through anaplerotic reactions significantly contributes to the FGP, resulting in differences in the natural abundance of δ13C in bulk organic matter and organic compounds. Further, pulse-13CO2-labelling was performed on developing culms, either from the surface or from the internal hollow, to ascertain whether significant CO2 fixation occurs in developing culms. δ13C of young shoots and developing culms were higher (−26.3 to −26.9 ‰) compared to all organs of mature bamboos (−28.4 to −30.1 ‰). Developing culms contained chlorophylls, most observed in the skin tissues. After pulse-13CO2-labelling, the polar fraction extracted from the skin tissues was slightly enriched in 13C, and only a weak 13C enrichment was observed in inner tissues. Main carbon source sustaining the FGP was not assimilated by the developing culm, while a limited anaplerotic fixation of respired CO2 cannot be excluded and is more likely than C4-photosynthetic carbon fixation.


Introduction
Phyllostachys edulis is a large monocotyledon (Poaceae, subfamily Bambusoideae) commonly distributed in East Asia's temperate or subtropical mountains. The plant grows incredibly fast from young shoots emerging from the ground to mature bamboos (10-15 m high) within 60 days without producing leaves, the photosynthesising organs (Yen and Lee 2011). Young bamboo shoots are shoots covered by sheaths before 25 to 40 internodes about 7.5 cm in diameter simultaneously undergoing cell division (0.06 cells cell −1 h −1 ) and elongation (0.49 μm μm −1 h −1 ). The cell division zone at the base of the internode is up to 2 cm, while the elongation zone, the upper part of the internodes, is up to 12 cm (Chen et al. 2022). The elongation is completed within 21-25 days, resulting in the rapid growth of the young developing culms, during which time a large amount of carbon is required (Chen et al. 2022), after which the senescence of the sheath occurs . The young culms accumulate three-fourths of their final biomass during the initial 40 days of this fast-growing period (FGP) (Yen 2016). The carbohydrates required to support this fast growth are thought to be transported from mature bamboos (autotrophic stage) to young leafless bamboos (heterotrophic stage) via the leptomorph rhizome systems (Song et al. 2016;Wang et al. 2020;Zhai et al. 2022).
Phyllostachys edulis is a C 3 plant, based on plastome phylogenetic trees, like rice (subfamily Oryzoideae), its close relative (Bianconi et al. 2020). This is confirmed by the leaf isotope composition (δ 13 C) of P. edulis (Hanba et al. 2010). Roots or stems are, however, expected to be slightly enriched compared to leaves because of post-photosynthetic carbon fractionations (Badeck et al. 2005). Early wood in the tree rings of deciduous species is often enriched in 13 C because the carbon used in early spring is remobilised from carbon reserve (e.g. enriched starch), while the production of late wood is more directly related to the current photosynthetic carbon assimilation (Damesin and Lelarge 2003;Helle and Schleser 2004;Gessler et al. 2014).
Although not yet studied, there are good reasons to think that developing culms of bamboos should also be enriched compared to the leaves of mature bamboos if the photosynthetic products of mature bamboos are used to build storage compounds that are then transferred to developing immature culms.
Different organs of the same species also possibly exhibit different patterns of photosynthesis even though a plant is commonly recognised as a single photosynthetic type (e.g. C 3 or C 4 ). For instance, leaves show the C 3 type and stems show C 4 carbon fixation in celery and tobacco (Hibberd and Quick 2002). Although previous studies have shown that C 4 enzymes are upregulated during the heterotrophic stage of developing culms, little is known about their contribution to the carbon gain of new-born bamboo culms during the FGP. Wang et al. (2012) observed that the parenchyma cells around the vascular bundles in the developing immature bamboo culms contained chloroplasts. Wang et al. (2021) then revealed that the genes of the important enzymes related to the C 4 carbon fixation, such as the phosphoenolpyruvate carboxylase (PEPC), the nicotinamide adenine dinucleotide phosphate-malate dehydrogenase (NADP-MDH) and the NADP-malic enzyme (NADP-ME), were upregulated during the FGP. Furthermore, they revealed the activity of these key enzymes in developing culms during the FGP, suggesting that the 2-mm-thick skin of culms operates the NADP-ME subtype of C 4 carbon fixation. Mature P. edulis is therefore a typical C 3 plant but C 4 -like patterns may be present in developing culms. Carbon fixation may therefore not only occur in leaves, the autotrophic organ of mature bamboo, but also in developing culms during the FGP, which would result in a strong 13 C enrichment of developing culm organic matter compared to the leaves of mature bamboos.
It is well known that heterotrophic plant organs, such as tree stems, can assimilate carbon and that the source of CO 2 is partially from respiration (Cernusak and Marshall 2000;Wittmann et al. 2005;Berveiller et al. 2007). Forty per cent of carbon loss by respiration can be reassimilated in the current-year stems of European beech for example (Damesin 2003). The 13 C signature of the respired CO 2 is determined by the relative contributions of the decarboxylation related to the pyruvate dehydrogenase activity and those occurring during the Krebs cycle (isocitrate dehydrogenase and alphaketoglutarate dehydrogenase activities), leading to an apparent fractionation of up to 6 ‰ Tcherkez et al. 2003). Carbon fixation in tree stems can occur in darkness (Höll 1974;Langenfeld-Heyser 1989), which suggests the recycling of respired CO 2 by the anaplerotic pathway, involving the PEPC and generating 13 C-enriched organic acids due to the fractionation in favour of 13 C, with discrimination relative to gaseous CO 2 of about −5.7 ‰, including therefore those of the carbonic anhydrase (CA) involved in the hydration of CO 2 (respectively 2.2 ‰ discrimination for the PEPC and −9 ‰ for the CA, Badeck et al. 2005;Ghashghaie and Badeck 2014).
Our objective was to clarify which carbon sources contribute to the fast-growing stage of immature culms of P. edulis, from the developing culm itself or exogenously from mature bamboos. We postulated that photosynthates produced by the mature bamboos are the main carbon source for the fast growth of developing bamboo culm and that, therefore, the δ 13 C of developing bamboo does not strongly differ from the δ 13 C of all organs of mature P. edulis. To test this hypothesis, we collected samples of various organs at both the autotrophic and heterotrophic stages and analysed their natural 13 C abundance in the bulk organic matter and the fractions of different organic compounds. We further postulate that developing culms can nevertheless fix a limited amount of CO 2 during the FGP. To test this hypothesis, we operated a 13 C pulse labelling at the surface and in the hollow of the developing culm and traced the labelled carbon in the soluble polar fraction (sugars, amino acids and organic acids) in the tissues of the developing culms.

Experimental site
This study was conducted in a moso bamboo stand at the Katsura Campus, Kyoto University (Kyoto Prefecture, Japan; 34°59ʹ06.2ʹʹN, 135°40ʹ48.4ʹʹE, 110 m alt.). Mean annual temperature and mean annual precipitation in 2021 were 16.9 °C and 1552 mm. The stand management included rough selective cutting and bamboo shoot harvesting without fertilisation. The mean height of the culms of mature bamboos was 15.3 m and the average diameter at breast height (DBH) was 9.9 cm.

Sample collection of isotope analysis (natural abundance)
Between April and May 2021, three replicates of each organ were collected from three different individuals, including culms, branches, leaves, rhizomes and roots for mature bamboos; and bamboo shoots and developing culms for developing bamboos (Fig. 1). Samples were frozen in liquid nitrogen and first stored at −15 °C in a portable freezer in the field before being transferred to −20 °C in the laboratory.

CO 2 labelling experiment
The 13 CO 2 pulse labelling experiment was conducted on 11 May 2021. Six immature, leafless but vigorous culms of similar size were selected (mean DBH = 11.7 cm, mean height = 6 m). Three labelling chambers were installed on three developing culms around the 7th internode from the ground without sheaths (chamber size, Φ = 20 cm, height = 15 cm). The chambers were made of a transparent thin polycarbonate sheet wrapped around the culm on polypropylene half-circular plates at both the bottom and the top of the chamber, with semi-circular holes at their centre to accommodate the culm. They were affixed to the surface of the culm with neutral seal putty and a silicon sealant [see Supporting Information- Fig. S1]. At 13:20 h, 75 mL of 99 % 13 CO 2 was injected into each chamber from three directions (three times 25 mL with a needle inserted through the polycarbonate sheet). The concentration of 13 CO 2 in the chambers was 2.4 %, based on the volume of the chamber and the amount of 13 CO 2 injected. In the other three developing culms, 50 mL of 99 % 13 CO 2 was injected with a needle inserted into the hollow of the 7th internode without sheaths. The concentration of 13 CO 2 in the hollows was 3.7 %, based on the volume of the hollow and the amount of 13 CO 2 injected. The small pinhole was sealed by tape. Two hours after the injection of 13 CO 2 in either the chamber or the culm hollow, the 5th, 7th and 9th internodes of the six labelled and two unlabelled developing culms were sampled. Disks of 100-200 g were collected using a hand saw shortly after labelling to limit the proportion of labelled products potentially transported away from the assimilation site. Samples were frozen in liquid nitrogen and first stored at −15 °C in a portable freezer in the field before being transferred to −20 °C in the laboratory, as above.

Purification of polar fractions, structural compounds, starch and proteins from different organs
In the laboratory, all samples were freeze-dried for at least 72 h using a vacuum freeze dryer (FDU-1200, EYELA, Tokyo, Japan), and ground to a fine powder in a ball mill (MM-400, Retsch, Düsseldorf, Germany).
The polar fraction, starch, proteins and structural compounds were separated and purified according to the modified methods of Deleens and Garnier-Dardart (1977), Wanek et al. (2001), Richter et al. (2009) and Desalme et al. (2017). These purification procedures have been developed specifically to allow the carbon isotope analysis of organic fractions. Briefly, soluble and insoluble compounds were separated based on their polar properties in a mixture of methanol-chloroformwater (12/5/3, v/v/v). After centrifugation, the supernatant was mixed with 0.5 mL of methanol-chloroform (MC, 1/1, v/v) to separate pigments and lipids in the lower phase from the polar compounds in the upper phase. The pellet was digested with pronase in Tris buffer. Proteins were solubilised while starch remains in the insoluble fraction. The insoluble fraction was washed with ethanol and starch was gelatinised by hydrochloric acid. The starch fraction, recovered in the supernatant after centrifugation, was precipitated with absolute methanol. The structural fraction was recovered in the pellet washed with water at least 3 times. All fractions were oven-dried at 65 °C and weighed. The methodology detailed step by step is available [see Supporting Information-Method S2 and Fig. S2].
δ 13 C determination of total organic matter, polar fractions, structural compounds, starch and proteins For the determination of the isotope composition, 1 mg of dried samples or dried fractions was weighed in tin capsules and combusted in elemental analysers coupled to isotope ratio mass spectrometers (Thermo FLASH2000, DeltaV advantage with ConfloIV system, Thermo Fisher Scientific, Massachusetts, USA; for unlabelled samples, and Thermo NC2500, MAT252 with ConfloIII system, Thermo Fisher Scientific, Massachusetts, USA; for labelled samples). Several international standards [IAEA-CH-3 (−24.724 ‰, Coplen et al. 2006), USGS40 (−26.39 ‰, Qi et al. 2003) and USGS41a (36.55 ‰, Qi et al. 2016)] were used for calibrations. USGS 41a was further used as a running standard for every 11 samples. The analysis of the 11 samples would have been rerun if the running standard had been an outlier of all standards (outlier defined in Reed et al. 1971), which did not happen. Results (δ 13 C) were expressed in ‰ as the relative deviation of the isotope ratio of the sample ( 13 C/ 12 C, R sample ) compared to that of the international VPDB standard (Vienna PeeDee Belemnite).
The isotope composition of the weighted fractions (δ 13 all) was defined as: with Cf , the carbon ratio for each fraction refers from Bowling et al. (2008), and Wf , the dry weight of fraction. The excess amount of 13 C in a labelled sample was calculated as: . Young bamboo shoots (BS) emerge from the ground attached to mature bamboo rhizomes. Developing, immature culms grow rapidly within 3 months (heterotrophic stage, HES). Young bamboo shoots are shoots covered by sheaths before the internodes undergoing extensive elongation. Developing culms are the growing culms that undergo a period of rapid internode elongation, after which the senescence of the sheath occurs, and eventually mature into fully formed culms.
with x( 13 C) L and x( 13 C) UN , the atom fraction of 13 C in labelled and unlabelled samples respectively, and C, the total carbon content.
The atom fraction was calculated as: x

Culm anatomy and starch granule observation
The 7th internode of developing culms was cut into small cubes. Cross and vertical sections 10-30 μm thick were sliced off the frozen samples (MCR802A, Komatsu Electronics, Ishikawa, Japan) with a sliding microtome (TU-213, Yamato Kohki, Saitama, Japan). Each section was loaded on a glass slide and covered carefully with a coverslip to avoid air bubbles. After initial observation, sections were rinsed with ultrapure water and stained with 0.1 N iodine solution, and briefly rinsed with ultrapure water again. Microphotograph images were taken with a digital camera (EOS KISS X3, Canon, Tokyo, Japan) attached to a light microscope (13X50-32, Olympus, Tokyo, Japan).

Chlorophyll content measurement
Two mm wide vertical sections of the skin, middle and inner parts of the 7th internode in developing culms were cut for chlorophyll extraction. Chlorophyll was extracted from 500 mg of fresh samples using 4 mL of N,N-dimethylformamide as the extraction solvent. Extraction was conducted in darkness at 65 °C for 2 h followed by 1 h at room temperature (Berveiller and Damesin 2008). At the end of the extraction, the samples were colourless. The absorbance of the extract was measured at 470, 663.6 and 646.6 nm using a spectrophotometer (ASV-S3, As One Corporation, Osaka, Japan). The amounts of chlorophyll a and b were calculated using the extinction coefficients obtained from Porra et al. (1989).

Statistical analysis
Data presented in this study are the means of three replicates from three individuals. One-way ANOVA followed by the Tukey HSD post hoc test was performed to test the betweenorgan differences in δ 13 C in bulk organic matter and in different organic fractions, and the content in these different fractions. It was also used to test the differences in chlorophyll content between the skin, middle and inner tissues of the developing culms. Univariate student's t-test was used to test whether the excess 13 C in the polar fraction was significantly higher than 0 after pulse-labelling developing culms with 13 CO 2 . All statistical analyses were performed using R (R Core Team 2023).

Variations in natural 13 C abundance in bulk organic matter between organs
The isotope composition of the different autotrophic organs of mature bamboos (leaves and branches) and belowground systems (rhizomes and roots) ranged between −28.4 and −30.2 ‰ (−29.3 ‰ on average, Fig. 2). The isotope composition of culms was higher than that of mature bamboo branches and leaves (−28.5 and −29.8 ‰ on average, respectively; P < 0.05; Fig. 2). The bulk δ 13 C of bamboo shoots and developing culms (−26.8 and −26.3 ‰ on average; Fig. 2) were higher than that of all organs of mature bamboos (P < 0.001).

Variations in the content of biochemical fractions between organs
The content in the polar fraction was higher in the foliage than in other organs of mature bamboo and developing immature bamboo shoots (P < 0.001) but similar to that in young bamboo shoots (P > 0.05; Fig. 3A). The starch content was low in all organs, especially in the belowground part of the mature bamboos, except in the young bamboo shoots (Fig.  3B). More proteins were found in the aboveground organs of mature bamboos than in their belowground organs. While the protein content was not different in developing bamboo culms compared to mature bamboos (P > 0.1), the young bamboo shoots contained a much larger amount of proteins than all the other samples (0.48 g g −1 compared to 0.09 g g −1 on average in all other samples; P < 0.001; Fig. 3C). The content in structural compounds was much lower in young bamboo shoots (0.10 g g −1 ; P < 0.001; Fig. 3D) than in all the other organs. Low values were also observed in leaves (0.55 g g −1 , Fig. 3D), while it averaged 0.84 g g −1 in the belowground organs, 0.75 g g −1 in branches and mature culm and 0.70 g g −1 in developing bamboo culms (Fig. 3D).

Variations in natural 13 C abundance between biochemical fractions
The polar fraction was depleted in 13 C compared to the bulk organic matter and showed large variations among organs, from −35.1 ‰ in leaves to −27.5 ‰ in developing bamboo culms (Fig. 3E). The isotope composition of polar fractions from the rhizome (−28.8 ‰), less depleted than that of the other organs of the mature bamboos (< −30 ‰), was not different than that of the developing bamboo culms (−27.5 ‰; p > 0.1) and young bamboo shoots (−29.3 ‰; P > 0.1). In contrast to the polar fraction, the starch fraction was enriched in 13 C compared to the bulk organic matter and much less variable (Fig. 3F). The protein fraction was strongly depleted in all organs compared to the bulk organic matter (from −35.6 to −38.2 ‰, Fig. 3G) except for the young bamboo shoots, which showed a δ 13 C higher value than that of all other organs (−32.0 ‰; P < 0.001). δ 13 C of the structural compounds was slightly higher in the developing bamboo culm (−26.6 ‰, Fig. 3H) than that of the organs of mature bamboos (< −27.7 ‰). The weighted average of the isotopic composition of the different biochemical fractions (polar fractions, starch, proteins and structural compounds) by their concentration and carbon content (δ 13 all) was compared to the isotope composition of the bulk organic matter (δ 13 bulk), and the deviation can be attributed to the isotope composition of unknown compounds, especially lipids and other non-polar molecules, which have been solubilised in chloroform but not analysed isotopically due to their limited quantity. No pronounced differences between δ 13 bulk and δ 13 all were observed in mature organs (leaves, branches and culms) and belowground systems (rhizomes and roots). However, bulk organic matter in developing bamboo culms and young bamboo shoots was more enriched in 13 C than the analysed compounds (Fig. 4).

Chlorophyll content and anatomy of developing culms at the 7th internode
The chlorophyll content in the skin of developing culms, which includes the layer of cortical cells, was higher than in middle and inner tissues (26.1 μg g −1 FW compared to 4.4 and 0.5 μg g −1 FW on average, respectively; P < 0.001; Fig. 5A). However, when expressed per unit area, the chlorophyll content in the skin was similar to that in mature leaves (respectively 5.6 and 4.2 μg cm −2 on average). Chlorenchyma cells were distributed around the skin of developing culms ( Fig.  5C-E). In the middle part of the developing culm, chlorophyll was found preferentially near the metaxylem vessels and the phloem sieve tubes forming a bundle sheath around the vascular bundles ( Fig. 5F and G). Furthermore, a lot of starch granules were observed in parenchyma cells around the vascular bundles of developing culms. In contrast, almost no starch granules were observed in the parenchyma cells of the inner part and the chlorenchyma cells of the skin part of developing culms (Fig. 5H and I).
Distribution of labelled carbon after 13 C pulse labelling of the 7th internode of developing culms The excess 13 C of the polar fraction (PF) extracted from the 7th internode of developing culms was significantly higher than 0 near the skin (0.1 mg 13 C g −1 PF on average, Fig. 6A) while it was not significantly different from 0 deeper in the culms (0.0003 mg 13 C g −1 PF in the middle and 0.001 mg 13 C g −1 PF in the inner parts on average). When 13 CO 2 was injected in the hollow inside the 7th internode of the developing culms, little or no enrichment was recovered in any sample compared to unlabelled samples, but it was significantly different from 0 in the inner part, although the enrichment value was small (0.002 mg 13 C g −1 PF, Fig. 6B). No enrichment was found in the internodes below (5th) or above (9th) the internode 7th which was labelled, confirming that the labelled carbon did not migrate up or down during the short 2-h tracing (data not shown).

Discussion
Only slight 13 C enrichment of heterotrophic organs compared to leaves The isotope composition of the different autotrophic organs of mature bamboos (Fig. 2) was similar to those of mature leaves in a previous study (−28.3 ‰, Hanba et al. 2010), confirming that P. edulis is a C 3 plant at the autotrophic stage. The slight enrichment in 13 C of the different heterotrophic organs of mature bamboo (culms, rhizomes and roots) compared to autotrophic organs agrees with previous findings in several studies summarised by Badeck et al. (2005): roots, stems or culms (grass stems) are often slightly enriched in 13 C compared to leaves (1-2 ‰). The reasons for this phenomenon, which is often related to post-photosynthetic fractionations, are complex, and several hypotheses have been reviewed by Cernusak et al. (2009). Since the isotopic composition of starch was similar for all organs while differences were found for other fractions, a seasonal change in 13 C discrimination during photosynthesis (their hypothesis 2) or preferentially nocturnal sucrose translocation (their hypothesis 3) are therefore unlikely to explain the 13 C enrichment of immature bamboo and heterotrophic organs of mature bamboo compared to leaves. Heterogeneous carbon isotope distribution within sugars results from enzyme-dependent fractionations occurring during sugar interconversions. Because these sugars are precursors of all organic compounds in plant tissues, these fractionations impact the 13 C distribution in plants (Rossmann et al. 1991;Gilbert et al. 2012), given that different types of biochemical substances do not have the same isotope characteristics (Bowling et al. 2008). The differences in the isotope composition between organs are therefore thought to be related to their differences in the biochemical composition (Cernusak et al. 2009, their hypothesis 1). However, although different biosynthetic pathways cause carbon-related compounds in different organs to have different 13 C, the remaining part of the biosynthesis of a certain compound is used in the biosynthesis of other compounds, so the bulk δ 13 C of the organ should not be determined by internal compounds but by the balance between carbon input and output (Badeck et al. 2005;Gilbert et al. 2012). The fact that the assimilates transported from the leaves to the remaining part of the plant (input) are enriched in 13 C due to post-photosynthetic fractionations associated with mitochondrial respiration occurring in the leaves (Damesin and Lelarge 2003;Gessler et al. 2007;Kodama et al. 2008) can therefore explain the observed slight enrichment of these heterotrophic organs (Cernusak et al. 2009, their hypothesis 4). Also, additional fractionations related to incomplete decarboxylation during respiration (output) can release depleted CO 2 and contribute to the enrichment of the remaining organic matter (Damesin and Lelarge 2003;Ghashghaie and Badeck 2014). The fact that the polar fraction was similarly enriched in the belowground organs of mature bamboos and in the bamboo shoots and developing culms compared to the autotrophic organs of mature bamboos supports the hypothesis that photosynthates produced by the mature bamboos and stored in the rhizome are the main carbon source for the fast growth of developing bamboo culm (Fig. 3E). This is also consistent with the fact that the starch content was very low in the belowground part of the mature bamboos, but high in the young bamboo shoots, which may have temporarily stored starch just before the onset of the FGP (Fig. 3B). The fact that the isotopic composition of starch from young shoots was similar to that of starch from leaves of mature bamboos (Fig.  3F) suggests a possible common sugar source.
It has been shown in another bamboo (Phyllostachys nigra) that the chemical composition of the cell wall changed with time and that lignin accumulated at a later stage of development (Chang et al. 2013). It is therefore possible that the lower δ 13 C values in mature bamboo culms are related to a high proportion of lignin in their structural fraction compared to immature bamboos. Lignin is indeed depleted in 13 C compared to cellulose or other structural polysaccharides (Bowling et al. 2008) because lignin precursors are produced by the shikimate pathway, which is fuelled by the depleted products of the pyruvate dehydrogenase and the Krebs cycle (Gessler et al. 2014). However, the isotope composition of the structural compounds was very similar in developing culms, bamboo shoots and organs of mature bamboos (Fig. 3H). It . Deviation between the weighted average of the isotopic composition of the different biochemical fractions (δ 13 all) and the isotope composition of the bulk organic matter (δ 13 bulk). δ 13 all is the weighted average of the isotopic composition of the different biochemical fractions (polar fractions, starch, proteins and structural compounds) by their concentration and carbon content. Average (n = 3) with standard errors are shown. Different letters indicate significant differences among organs (one-way ANOVA and Tukey HSD post hoc test at P < 0.05). LF, BR, MC, RH, RO, DC and BS indicate leaves, branches, mature culms, rhizomes, roots, developing culms and bamboo shoots, respectively. The autotrophic stage (AUS) includes both aboveground (AG) and belowground (BG) organs of mature bamboos. The heterotrophic stage (HES) includes newly formed bamboo shoots and developing immature bamboo culms during their fast-growing period (FGP).
is therefore unlikely that the observed depletion of mature culms relative to the immature ones was related to lignin accumulation with ageing.

Photosynthesis in new-born bamboo culms during the FGP
Huge amounts of carbohydrates should be either produced or transported from mature bamboos to meet the high demand needed to support cell elongation in the internodes of the immature bamboo culms during the period of rapid growth. Using 13 C-enriched CO 2 labelling, we found direct evidence that the skin of developing culm slightly uptakes the CO 2 from the atmosphere and synthesises carbohydrates (Fig. 6A). However, because the concentration of CO 2 in the labelling chamber was much higher than in the atmosphere (2.4 %), our results revealed a potential rather than the actual CO 2 fixation. In addition, the estimated rate of CO 2 fixation by the immature culms based on excess 13 C in the polar fraction, which was about 0.1 µmol m −2 s −1 , was more than 100 times lower than expected leaf photosynthesis (Zachariah et al. 2016) and 20 times lower than the respiration of young culms (Uchida et al. 2022). Our results are therefore consistent with the low chlorophyll content in the skin of developing culms compared to leaves (Fig. 5A) and the observed translocation of photosynthates from mother to daughter ramets of moso bamboo quantified using 13 C-CO 2 pulse labelling (Zhai et al. 2022).
Previous studies have also pointed out that the developing culms of P. edulis activated the key enzymes of C 4 carbon fixation (such as NADP-ME, NADP-MDH, pyruvate phosphate dikinase [PPDK] and PEPC) during the FGP (Wang et al. 2012. A C 4 plant-like Kranz anatomy has also been evident in the middle of the developing bamboo culm in this study ( Fig. 5F and G), confirming previous observations (Wang et al. 2012). Another hypothesis to explain the slight enrichment of the bamboo shoots and developing culms is therefore that the PEPC is involved in a limited carbon fixation by immature culms during their initial growth, like C 3 -C 4 intermediate species (Lundgren et al. 2016). PEPC refixation may contribute with malate to the Krebs cycle, thus resulting in 13 C-enriched respired CO 2 (Cernusak et al. 2009, their hypothesis 5;Gessler et al. 2009). However, based on direct evidence from our 13 CO 2 pulse labelling, the assimilation of atmospheric CO 2 seems to be limited to the skin of the culm, as only the tissue near the skin was labelled when 13 CO 2 was provided around the culm, while the C 4 plant-like Kranz anatomy was found in the middle part of the culm. In addition, the isotope composition of the starch in developing culms and bamboos shoots was not different than that in any organ of mature bamboos, which definitely excludes significant C 4 -like photosynthetic carbon fixation in immature culms (Fig. 3F).
The presence of a limited amount of chlorophyll and a large number of starch granules in the parenchyma cells around the vascular bundle in the middle part of the culm nevertheless suggests that CO 2 assimilation can occur, the CO 2 being potentially supplied by the xylem sap rather than diffusing from the atmosphere or the internal hollow because of limited diffusion of CO 2 within the dense culm tissues. A high amount of water flows through the rhizome to supply the fast-growing shoots and culms with water and nutrients (Zhang et al. 2020). Berveiller & Damesin (2008) found that the activities of PEPC and NADP-ME were up to 13 and 30 times higher in stems than in leaves of European beech. High PEPC activity can supply malate to the NADP-ME enzyme which, in turn, can supply CO 2 for Rubisco after decarboxylation, recycling CO 2 derived internally from respiration, as it was postulated in stems of tobacco, celery and mikania that have characteristics of C 4 photosynthesis in stems and exhibited a C 4 plant-like Kranz anatomy (Hibberd and Quick 2002). However, it remains unclear whether enough light can reach the photosynthetic pigments in the bundle sheath and provide enough energy to sustain a C 4 -like carbon fixation under the deep shade conditions that prevail in the understory of a P. edulis because the C 4 photosynthesis became unfavourable at low PPFD (Bellasio and Farquhar 2019). In addition, the starch in the numerous granules should have been enriched compared with CO 2 derived from the oxidation of imported sugars if a large fraction of respired CO 2 was reassimilated by the PEPC and used in gluconeogenesis. This is because the fractionation in favour of 13 C by the CA involved in the hydration of CO 2 is higher than the small fractionation against 13 C by the PEPC (Badeck et al. 2005;Ghashghaie and Badeck 2014). Such enrichment was not observed in starch (Fig. 3F), suggesting that the abundant recycled CO 2 was not used for starch synthesis but for other purposes. This again suggests that the starch was instead likely synthesised from sucrose produced in mature bamboo, transported in the phloem through the rhizomes and transiently stored in granules near elongating cells ). On the other hand, carbon derived from the breakdown of stored starch in mature bamboo organs should be 13 C-enriched, as compared to primary assimilates, because starch breakdown produced 13 C-enriched sucrose (Gessler et al. 2008). Therefore, a higher contribution of stored carbon to boost developing culms may lead to enriched 13 C organic matter in immature culms.

Anaplerotic fixation of respired CO 2 in immature bamboo culms during the fast growth period
The inner part of the culms was slightly labelled when 13 CO 2 was provided in the culm hollow (Fig. 6B). This may reveal that carbon fixation is taking place in the inner parts. It has been shown that carbon fixation by the PEPC can occur in darkness in some tree stems (Höll 1974;Langenfeld-Heyser 1989) and can be stimulated in the presence of nitrate and Figure 6. Excess 13 C in after a pulse 13 CO 2 labelling at the 7th internode of developing culms during the fast-growing period. Average with their standard error (n = 3) of excess 13 C in the polar fraction (PF) extracted from the skin, middle and inner tissues of the developing culms 2 h after the injection of 13 CO 2 in either a chamber around the culm (left, A) or the culm hollow (right, B). '***' and '**' indicate that the average excess 13 C is greater than 0 (univariate student's t-test at P < 0.001 and P < 0.01).
ammonium (Müller et al. 1991;Vuorinen and Kaiser 1997). The CO 2 concentration inside culms of moso bamboos, which varies during the day and over the season from 40 000 to more than 100 000 ppm (Toshihiko Maitani, unpublished results), is in a similar range as for other bamboo species (Kiyama 1905;Zachariah et al. 2016). Therefore, it is also possible that PEPC and NADP-ME are involved in the anaplerotic fixation of respired CO 2 in the middle and inner culms where light cannot reach, providing the carbon skeleton (malate and oxaloacetate) necessary for amino-acid (Berveiller and Damesin 2008) and lipid syntheses (Smith et al. 1992), which are also needed during the FGP of bamboo. As explained above, because the fractionation in favour of 13 C by the CA involved in the hydration of CO 2 is higher than the small fractionation against 13 C by the PEPC, the organic acids issued from the anaplerotic pathway are expected to be enriched. The discrepancy between δ 13 bulk and δ 13 all (weighted average of the isotopic composition of the four biochemical fractions analysed) in developing bamboo culms and young bamboo shoots (Fig. 4), which can be attributed to the isotope composition of lipids and other non-polar molecules as mentioned above, suggests that 13 C-enriched malate, involved in the fatty acid synthesis, is partly derived from the anaplerotic fixation of CO 2 . The impact of the anaplerotic dark CO 2 fixation on the enrichment of plant tissues also depends on how much respired CO 2 can be reassimilated (Nalborczyk 1978). The use of 13 C-enriched organic acids issued from the anaplerotic pathway for nitrogen assimilation may explain the slight enrichment of the abundant proteins extracted from bamboo shoots, and why this enrichment starts to vanish when the new culm is developing (Fig. 3G). While most nitrogen needed to support the fast growth is potentially remobilised from the mature bamboos and transferred via the rhizome (Shi et al. 2021(Shi et al. , 2022Zhai et al. 2022), additional nitrogen assimilation can occur.

Conclusions
In spring, the bamboo shoots emerge, and the developing bamboos enter a phase of active and very rapid growth. Because the leaf of new-born bamboo had not yet developed, the main sources of carbohydrates during the FGP were supplied by mature bamboos and delivered through their belowground system to the developing culms, which is consistent with a slight 13 C enrichment of the bulk organic matter of immature bamboos. Although previous studies observed key enzymes related to C 4 carbon fixation, the results based on 13 C pulse labelling showed limited CO 2 uptake from the atmosphere and hollow during the FGP. Limited anaplerotic dark fixation of CO 2 is more likely to occur than C 4 photosynthetic C fixation.

Supporting Information
The following additional information is available in the online version of this article - Figure S1. Schematic illustration of 13 CO 2 labelling in developing culms of Phyllostachys edulis.
Methods S2. Procedure of purification for the polar fraction, structural compounds, starch and proteins from different organs in plants adjusted for P. edulis. Figure S2. Flow chart of the laboratory analytical procedure for carbon compound purification in plant samples.

Funding
This work was supported by JSPS KAKENHI Grant Numbers JP15H04513, JP19J11336 and JP20J15519. The authors declare that they have no competing financial interests or personal relationships that could have influenced the work reported in this paper.

Conflict of Interest Statement
None declared.