Metabolic Control of Sugarcane Internode Elongation and Sucrose Accumulation

Abstract

The relationship between metabolic changes occurring in the developing internodes of sugarcane and the final yield and sugar characteristics is poorly understood due to the lack of integration between phenotypic and metabolic data. To address this issue, a study was conducted where sugarcane metabolism was modeled based on the measurement of cellular components in the top internodes, at two stages of crop development. The study also looked at the effects of Trinexapac-ethyl (Moddus^®) on growth inhibition. The metabolome was measured using GC-analysis, while LC-MS/MS was used to measure proteome changes in the developing internodes. These data were then integrated with the metabolic rates. Regardless of the growth rate, internode elongation was restricted to the top five internodes. In contrast, sucrose and lignin accumulation was sensitive to the growth rate. Crossover plots showed that sucrose accumulation only occurred once the cell wall synthesis had slowed down. These data suggest that sucrose accumulation controlled a reduction in sucrose breakdown for metabolic activity and a reduction in demand for carbon for cell wall polysaccharide synthesis. This study also found that nucleotide sugar metabolism appears to be a key regulator in regulating carbon flow during internode development.

1. Introduction

There is a growing interest in using C4 crops like Miscanthus, switchgrass, sorghum, and sugarcane as feedstocks for biofuel production. Thesoluble sugars and cell wall fractions can be used as a source of feed or fuel [1].

There is a strong correlation between culm length, internode length, and above-ground biomass in all C4 grasses [1,2,3,4]. Therefore, to maximize yields, it is crucial to cultivate crops with larger culm volumes that contain high sugar contents and easily hydrolysable cell wall polymers. Understanding the factors that regulate the development of stem sink tissues, including their cell wall composition and soluble sugar content, is critical for enhancing the quality and yields of forage and bioenergy crops [1]. However, there is still limiting information available on this topic. Setaria viridis has become a popular model species to further understanding of the control of sink growth and development [1,5], but validation across other grasses is still lacking.

There are three important aspects of biomass accumulation, or sink strength, during growth. First, osmolytes in the vacuole must be accumulated to facilitate water uptake. Second, the synthesis of extra cell walls, cell membranes, and proteins to maintain cell functions. Third, there is the energy (respiratory) cost of the first two elements [6]. Gene expression changes associated with carbon partitioning have been described in terms of three major carbon sinks, namely sucrose storage, cell wall synthesis, and respiratory metabolism [1].

UDP-Glucose is a central metabolite in the synthesis of both sucrose and most of the cell wall polysaccharides (including cellulose, hemicellulose, and pectic polymers) [1,7,8]. Significant changes in the expression of UDP-glucose-associated proteins occur during internode development. Carbon partitioning in sugarcane is closely related to the transcription of genes associated with UDP-glucose (UDP-glc) metabolism [1,7,8,9], sugar transporters [9,10,11,12], and various cellulose synthases.

Despite this progress, the link between metabolic changes in the developing internodes in sugarcane and final yield and sugar characteristics is not well understood. This is primarily due to two factors. Firstly, studying mRNA levels only provides a partial picture of metabolic control. Interpretation of high throughput sequencing information needs to be integrated with protein levels and metabolic flux data. Many recent studies have revealed instances where mRNA abundance failed to correlate with protein abundance. This phenomenon challenges the conventional understanding and necessitates a comprehensive exploration to understand the underlying mechanisms [13]. Secondly, most metabolic studies have not provided any insight into the phenotype of the sample tissue, especially the growth rate of the tissue at the sampling times.

Here, we report on the metabolic profile of internodes with varying growth rates and model the flux into key lignocellulosic and soluble metabolic fractions. The data confirmed that the duration of internode elongation is controlled by thermal time. This study represents the first effort to integrate metabolome and transcriptome changes with metabolic rates at different stages of internode development. Cell wall polysacccharide synthesis, respiration, and other biosynthetic processes are the major demand functions in developing internodes. Sucrose storage only occurs when cell wall polysaccharide synthesis is completed.

2. Materials and Methods

2.1. Material

Research trials were conducted in the field with sugarcane variety KQ228. The trial sites were at Sugar Research Australia’s Burdekin Station, QLD (19°34′0.80″ S, 147°19′30.7″ E). The trials were planted in a completely randomized design, including three replicate plots per treatment. Each replicate consisted of 4 × 10 m of cane. Billets obtained from disease-free stalks were used as planting material. Before stick planting in August, the soil was nutrient tested and fertilized according to Six Easy Steps nutrient recommendations [14].

The sugarcane was furrow irrigated with a 7-day flood irrigation schedule throughout the cropping cycle. The trial was a completely randomized design, including two treatments with four replicate plots. Each replicate consisted of four 10 m rows with a 1.5 m spacing between rows.

MODDUS, an emulsifiable concentrate containing the active compound Trinexapac-ethyl at a concentration of 250 g L⁻¹, was applied to the crop as a foliar spray at a dose of 50 g active ingredient ha⁻¹ (0.25 label rate) using an agricultural handheld knapsack sprayer [4]. Moddus treatment occurred at 140 and 169 days after planting (DAP). The last application was 26 days before sample collection in March (crop age six months).

2.2. Non-Destructive Measurements

Six primary shoots in each plot were tagged for easy identification. Stalk elongation and phyllochron development were non-destructively measured in-field. This was done to ensure minimal disruption to canopy development by not changing shoot and leaf numbers, leaf production rates, and numbers of senescing leaves. Measurements were performed throughout the first nine months of crop development.

Modeling Growth

The growth of the culm was modeled by applying a logistic growth function

$$\mathsf{\int }\left(growth\right)=\frac{max}{1+e^{-k\left((t_{mid}-t)\right)}}$$

(1)

where ∫(length) is the change in the phenotype parameter (length, diameter, or volume) and (t) represents time or thermal time. The parameters to be fitted were maximum size (max) and $t_{mid}$ the time when half of the maximum size (length, diameter, or volume) was reached. The steepness of the growth curve is represented by k. The rate at each time point was determined with

$$\frac{d\mathrm{growth}}{dt}=\frac{-\mathit{max}·k{e}^{-k(t_{\mathit{mid}}-t)}}{{(1+e^{-k(t_{\mathit{mid}}-t)})}^2}$$

(2)

2.3. Destructive Sampling

Six culm samples were collected from the field plots at two different time points, approximately 6 and 10 months after planting. These samples were analyzed using a modified method by Berding [15]. The culm samples were disintegrated using either a garden mulcher or Dedini laboratory disintegrator at room temperature. The mulched material was then weighed to determine the fresh weight (FW) and transferred to a paper bag to be dried at 70 °C until a constant dry weight (DW) was reached (usually 6 to 7 days).

2.4. Biomass Composition

All the analyses were conducted at Celignis Analytical (Celignis https://www.celignis.com/biomass-analysis.php) using the analytical package P19 (deluxe lignocellulose: sugars, lignin, extractives, and ash, protein-corrected lignin, water-soluble sugars, uronic acids, acetyl content, and starch).

2.4.1. Extraction of Biomass Component

All extractions were carried out with a Dionex Accelerated Solvent Extractor (ASE) 200 [16]. The extractions were carried out according to the National Renewable Energy Laboratory (NREL) standard operating procedure for determining extractives in biomass [17]. Ash content was determined using a Nabertherm L-240H1SN furnace, according to the NREL operating procedure for the determination of ash in biomass [17].

2.4.2. Cell Wall Constituents

Hydrolysis of the dry extractive-free samples was performed according to a modification of the NREL standard operating procedure for the determination of structural carbohydrates and lignin in biomass [18]. The procedure was divided into two main steps: a two-stage acid hydrolysis of the samples, and gravimetric filtration of the hydrolysate to separate it from the acid-insoluble residue (AIR) [19]. Klason lignin was calculated by determining the weight difference between the AIR and its ash content. Acid-soluble lignin was measured by determining the absorbance of an aliquot of the hydrolysate at 240 nm using an Agilent 8452 UV–vis spectrophotometer. The results were then converted to ASL based on Beer’s law [20]. The lignocellulosic sugars resulting from hydrolysis were determined by ion-chromatography techniques adapted from [18]. The method consisted of diluting the hydrolysate samples 20× with a deionized water solution containing known amounts of melibiose as an internal standard [16]. The hydrolysate samples were diluted 20 times with a deionized water solution containing known amounts of melibiose as an internal standard. After dilution, the hydrolysates were filtered using 0.2 μm Teflon syringe filters and stored in 1.5 mL vials. The chromatography system included an electrochemical detector (PAD), a gradient pump, a temperature-controlled column, and a detector compartment. Complete separation of the sugars in the solution was achieved in 35 min through a Carbo-Pac PA1 guard and analytical column, connected in series. Deionized water was used as the eluent at a flow rate of 1.1 mL/min, and the column/detector temperature was 21 °C. NaOH (300 mM) was added to the post-column flow at a rate of 0.3 mL/min using a Dionex GP40 pump to create alkaline conditions for carbohydrate detection by the PAD detector.

2.4.3. Water Solubles

The sugars in the water-soluble fraction Section 2.4.1 were analyzed using ion–chromatography [16] as described above Section 2.4.2.

2.5. Extraction of Metabolites and Proteins

In this study, we followed the numbering system introduced by van Dillewijn (1952), where internodes are labeled from the top of the culm. In this system, the most recently fully expanded leaf, determined by the visibility of the dewlap, is referred to as leaf +1, and the node to which this leaf is attached is referred to as node +1. The internode below this node is designated as internode 1 [21]. Internodes 2, 4, 6, 8, and 12 were chosen for this purpose. Internodes 2 and 4 are representative of internode elongation [4,12], internode 6 and 8 of peak sucrose accumulation [22,23], and internode 12 completion of sucrose accumulation [3,12,24].

Internodes 2, 4, and 6 were removed from the stalk, and a 30 mm long section was cut from the bottom of the internode. Approximately 8 mm diameter cylindrical cores were bored off-center (avoiding the pith) and vertically down using a 12 mm cordless drill and diamond drill bit. The cylindrical samples were placed in a labeled 2 mL screw cap tube and snap frozen in liquid nitrogen and stored at $-80$ °C. The drill bit borer was sprayed with 70% (v/v) ethanol and wiped between samples.

2.6. Metabolome

The protocol followed for extraction was described in [25]. Briefly, 30 mg of homogenized leaf was added to 500 μL 100% (v/v) methanol in a cryomill tube containing internal standards ${ }^{13}{C}_6$-Sorbitol, and ${ }^{13}{C}_5$${-}^{15}N$-ValineL), 2-aminoanthracene (0.25 mg/mL) and pentafluorobenzoic acid (0.25 mg/mL). Thirty µL of each study sample was evaporated at 30 °C to complete dryness, using a CHRIST RVC 2-33 CD plus speed vacuum. To limit the amount of moisture present in the insert, 30 μL 100% methanol (LCMS grade) was added to each insert and evaporated using a speed vacuum. Samples were derivatized online using a Shimadzu AOC6000 auto-sampler robot. Derivatization was achieved by the addition of 25 μL methoxyamine hydrochloride (30 mg/mL in pyridine, Merck) followed by shaking at 37 °C for 2 h. Samples were then derivatized with 25 μL of N,O-bis (trimethylsilyl)trifluoroacetamide with trimethylchlorosilane (BSTFA with 1% TMCS, Thermo Scientific, Waltham, MA, USA) for 1 h at 37 °C. The sample was allowed to equilibrate at room temperature for 1 h before 1 μL was injected onto the GC column using a hot needle technique. Split (1:10) injections were performed for each sample. Instrumentation parameters and sample analysis. The GC-MS system used comprised an AOC6000 auto-sampler, a 2030 Shimadzu gas chromatograph, and a TQ8050NX triple quadrupole mass spectrometer (Shimadzu, Japan)with an electron ionization source (−70 eV). The mass spectrometer was tuned according to the manufacturer’s recommendations using tris-(perfluorobutyl)-amine (CF43). GC-MS was performed on a 30 m Agilent DB-5 column with 0.25 mm internal diameter column and 1 µm film thickness. The injection temperature (inlet) was set at 280 °C, the MS transfer line at 280 °C, and the ion source adjusted to 200 °C. Helium was used as the carrier gas at a flow rate of 1 mL/min, and argon gas was used in the collision cell to generate the MRM product ion. The analysis of the derivatized samples was performed under the following oven temperature program; 100 °C start temperature, held for 4 min, followed by a 10 °C/min oven temperature ramp to 320 °C with a following final hold for 11 min. Approximately 526 targets were collected using the Shimadzu Smart Metabolite Database, where each target comprised a quantifier MRM along with a qualifier MRM, which covered approximately 356 unique endogenous metabolites and multiple stable isotopically labeled internal standards. Resultant data were processed using Shimadzu LabSolutions Insight software (Kyoto, Japan), where peak integrations were visually validated and manually corrected where required. Metabolome data were analyzed using MetaboAnalyst [26] and Mapman [27].

2.7. Proteome Data Collection and Processing

A portion of the liquid nitrogen frozen samples were lyophilized and used for protein extraction to generate proteome data [28]. Each sample (100 μg) was taken for digestion and analysis with a 1D and 2D IDA nanoLC (Ultra nanoLC system, Eksigent) system. The IDA LC–MS/MS data were searched using ProteinPilot v5 (Sciex) in thorough mode. The top 6 most intense fragments of each peptide were extracted from the SWATH datasets (75 ppm mass tolerance, 10 min retention time window). Shared and modified peptides were excluded. After data processing, peptides with confidence ≥ 99% and FDR ≤ 1% (based on chromatographic feature after fragment extraction) were used for quantitation. Protein differential abundance was determined using un-normalized protein quantification values as input for package DESeq2 v1.18.1 (Love, 2014).

The Mercator annotation tool (Mercator 4.0 http://mapman.gabipd.org/web/guest/app/mercator) [29] was used to assign the MapMan “bins” to differentially expressed proteins. The Mercator output file was then used as a mapping file to assign functional categorizations to differential expressed proteins (DEPs) on MapMan. An enrichment analysis with PageMan [27] was carried out with the same list of DEPs.

2.8. Statistical Analysis

Statistical analyses were performed in R (RStudio 2023.06.0 Build 421 for Windows) using the package Agricolae [30]. One-way ANOVA tests were used to make multiple comparisons followed by a least significant difference test (LSD) [31]. TukeyHSD post hoc tests were used to compare the group means. All graphs in the boxplot format were prepared in R using the package MultiCompview. Protein differential abundance was determined using unnormalized protein quantification values as input for package DESeq2 v1.18.1 [32].

3. Results

3.1. Crop Growth

Stalk elongation was measured throughout the first nine months of crop development for both the plant and first ratoon crop. Sugarcane height and biomass did not increase linearly. Instead, growth could be best modeled by applying Equation (1) (Figure 1A).

From the stalk length data, the growth rate was calculated by applying Equation (2). Maximum growth was achieved at 1130 $D{D}_{18}$ (Figure 1B).

An ANOVA analysis showed no significant difference between the two crop classes over the entire cropping cycle. However, there was a significant difference in growth between the two crop classes over the first 360 $D{D}_{18}$ ($p<0.001$). The ratoon crop outperformed the plant crop initially (Figure 1A), but the peak growth rate and maximum growth rates between the two crop classes were not significantly different.

3.2. Biomass Accumulation

The top internodes were harvested at peak growth (approximately 1100 $D{D}_{18}$) in the presence and absence of MODDUS and at a late stage in the cropping cycle (approximately 1900$D{D}_{18}$). For this paper, these different rates of growth were defined as “fast-growth”, “slow-growth”, and “MODDUS” phenotypes (Figure 2).

Biomass accumulation followed the same pattern between these three phenotypes (Figure 3A), and the peak accumulation rate occurred at the same stage (Figure 3B). However, there was a significant difference in the amount of biomass accumulated and the accumulation rate (Figure 3).

3.3. Biomass Composition

Table S1 shows the biomass composition as a portion of total biomass at the different stages of growth and in the presence or absence of MODDUS. For comparison, the biomass composition of internodes 4 (peak biomass accumulation), 8 (late accumulation), and 12 (completed biomass accumulation) are shown (Figure S2).

According to the data, the significant differences among the various phenotypes were associated with the water-soluble components such as sucrose and hexoses, rather than the lignocellulosic components (Figure S1). There were no differences found in the content of glucan (cellulose), pentan (xylan, arabinan), and lignin among internodes 4, 8, and 12 of the three phenotypes (Figure S1). However, the sucrose content of the slow-growth and MODDUS phenotypes was significantly different from that of the fast-growing phenotype. Even in internode 12, where there was no further gain in biomass, the sucrose content was lower than that of the slow-growing internodes.

In all three stages of internode development, the hexose content was significantly higher in the fast-growing phenotype than in the slow growing phenotypes (Figure S1).

3.4. Biomass Accumulation Rates

The total mass of each component per internode was calculated from the data in Table S1 and the total internode dry weight. The thermal time for developing a new phytomer in KQ228 is 30 DD₁₈ [4]. Therefore, the internode number was multiplied by 30 to calculate rates as mg. ${\mathit{DD}}_{18}^{-1}$. The logistic Functions (1) and (2) were used to model the accumulation patterns of the different biomass components (Figure S2).

The maximum rate of biomass accumulation was evident in internodes 3 to 4 (Figure 3A). Although the maximum rate differed significantly between the three phenotypes, there were no differences in the stage of internode development where this was achieved (

Table 1. The composition of sugarcane KQ228 internodes during the different stages of development and in the presence of MODDUS. sd = standard deviation, nd = not determined.

Component	Phenotype	Parameter 1
		Max	k	tmid 2
dry weight	fast_growth	24.58 a	0.03 a	3.42 a
dry weight	moddus	10.25 b	0.02 b	3.14 a
dry weight	slow_growth	7.89 c	0.02 b	3.39 a
cellulose	fast_growth	5.44 a	0.03 a	3.05 a
cellulose	moddus	2.71 b	0.02 b	3.21 b
cellulose	slow_growth	2.22 c	0.02 c	3.06 b
hemicellulose	fast_growth	2.36 a	0.03 a	3.03 a
hemicellulose	moddus	1.21 b	0.02 b	3.13 a
hemicellulose	slow_growth	0.91 c	0.02 b	3.32 a
lignin	fast_growth	3.12 a	0.02 a	5.51 a
lignin	moddus	1.62 b	0.01 b	3.63 b
lignin	slow_growth	1.51 b	0.02 b	3.73 b
lignocellulose	fast_growth	10.80 a	0.02 a	3.67 a
lignocellulose	moddus	5.98 b	0.03 a	5.42 a
lignocellulose	slow_growth	5.06 b	0.03 a	5.65 a
sucrose	fast_growth	12.63 a	0.02 a	6.14 a
sucrose	moddus	4.82 b	0.03 b	4.71 b
sucrose	slow_growth	5.84 c	0.04 b	4.62 b

¹ Means in a row without a common superscript letter differ (p = 0.05) as analyzed by one-way ANOVA and the TUKEYHSD test. Statistical analysis between the phenotypes for each component. ² Refers to the internode where the maximum rate of change occurred.

). Maximum rates of cellulose and hemicellulose accumulation coincided with maximum biomass accumulation (Figure 3B,C). The stage of internode development where maximum biomass, cellulose, and hemicellulose accumulation occurred was independent of the growth rate per se (Table 1 and Figure 3A–C).

The maximum rates of sucrose and lignin accumulation occurred between internodes 6 and 7 in the fast-growing cane (Figure 3D). There was a significant difference in the maximum rates of lignin and sucrose accumulation between the fast-growing and slow-growing cane (Table 1). Unlike cellulose and hemicellulose accumulation, the growth rate also influenced the stage of internode development where the maximum accumulation rates of sucrose (Figure 3D) and lignin (Figure 3E) were achieved (Table 1). Both sucrose and lignin peak accumulation occurred significantly earlier in internode development in slow-growing sugarcane.

The transition from early growth to sucrose accumulation during internode development is best illustrated with crossover plots (Figure 4). These plots highlight the point in internode development where the internode development transitioned from growth to sucrose storage.

In the fast-growing sugarcane, the portion of sucrose in the total biomass accumulation was low in the first 5 internodes (Figure 4A). In the internodes older than internode 6, sucrose was the main contributor to biomass accumulation. Interestingly, the rate of sucrose accumulation in the older internodes exceed the rate of total biomass gain (Figure 4A,C,E).

The crossover point between sucrose and biomass accumulation occurred significantly earlier in the slow-growing and MODDUS phenotypes. The same was observed in the crossover point between cell wall sugars and sucrose accumulation (Figure 4B,D,F).

3.5. Metabolome

A total of 78 metabolites were identified in the metabolome analysis. A one-way ANOVA analysis revealed that 65 metabolites differed significantly ($p<0.05$) between genotypes at different stages of development.

Various measures have been used in the past for the expression of metabolic data in sugarcane, namely dry weight, fresh weight and protein content, and on a whole-internode basis (for a review see [33], which produced conflicting results. In this study, metabolite data were normalized to reflect the total water-soluble content of the internodes (Figure S3).

The whole metabolome dataset was subjected to a principal component and hierarchical cluster analysis, to assess the general effect of the stage of internode development at early season and late season growth. The score plot of PLS-DA (Figure S4) displays a distinct separation between the internodes at peak biomass and sucrose accumulation during both fast (1100$D{D}_{18}$) and slow growth (1800 $D{D}_{18}$) stages. The first (Component 1) and second (Component 2) latent variables indicate that the model effectively identified a pattern that distinguished between the groups. Samples from the internode at low biomass but high sucrose accumulation (at both growth stages) clustered on the negative side of Component 1.

This separation of metabolic expression by the stage of internode development and time in the cropping cycle was also evident in the two-factor heatmap of normalized relative concentrations. Hierarchical clustering, on the left of the heatmap, showed that metabolites belonging to particular metabolic pathways were mainly clustered together (Figure S5).

Similarly, there was a clear separation between the internodes in both growth stages in the presence and absence of MODDUS (Figure S6). Samples from the young internodes (internode 4) in the presence and absence of MODDUS clustered on the negative side of component 1, with the older internodes on the positive side. The MODDUS treatment resulted in a metabolic pattern indicative of slow growth. This metabolic process clustering was also evident in the two-factor heatmap of normalized relative concentrations in the presence and absence of MODDUS (Figure S5).

A mapping file was constructed for all the metabolites identified in this project (Table S2) for use in Mapman [27].

The difference between internodes at peak biomass or sucrose accumulation was the downregulation of metabolite concentrations (Figure 5 and Figure S6) in the transition to sucrose accumulation.

At mid-season with high growth rates, 41 metabolites were significantly decreased in the transition from peak biomass accumulation to peak sucrose accumulation, while only 5 were upregulated (

Table 2. Changes in metabolite profiles between internodes at different stages of development during peak growth (early season) and slow growth (late season).

Phenotype 1		Change 1
Stage	Comparison	Up	Down	Insignificant
Mid season	peak biomass:peak sucrose	5	41	28
Mid season:late season	peak biomass:peak biomass	41	2	31
Mid season:late season	peak sucrose:peak sucrose	38	2	34
Late season	peak sucrose:peak sucrose	1	32	41
Peak season MODDUS	peak biomass:peak biomass	8	19	24
Peak season MODDUS	peak sucrose:peak sucrose	3	17	31

¹ A cut-off of 1.2 fold and a FDR (Benjani and Hochberg) of $p<0.05$ was applied.

and Figure 5A). A similar pattern was evident in the MODDUS treatment between these two stages of internode development (Table 2 and Figure 5C).

Evidently, the metabolite content of internodes at both stages of development were significantly higher in the mid-season when the crop growth was rapid (Table 2 and Figure 5B).

3.6. Protein Levels

Protein levels are controlled by protein expression and degradation. These two processes are interconnected and tightly regulated and play crucial roles in maintaining cellular homeostasis and responding to various internal and external stimuli. This paper only considered the cellular protein homeostasis during the different growth and development phases of the internodes.

Protein analysis was restricted to material from mid-season (fast) growth. The MODDUS-treated cane was used to represent slow growth. This was done to avoid other environmental influences that would be presented by collecting material at different times during the year. A total of 7332 protein sequences were submitted to Mercator for annotation. Proteins assigned to bin 35 (not assigned annotated) and bin 50 (Enzyme classification) were excluded from further analyses. The remainder of the dataset (5117 sequences) was analyzed for differential expression using a FDR criteria [34] of $p<0.05$. Based on this criteria, a total of 928 proteins were differentially expressed across the different treatments.

A principal component analysis and hierarchical cluster analysis showed a clear visual separation between the different stages of development (peak biomass accumulation and peak sucrose accumulation, Figure 4) and between the control (fast growth) and MODDUS treatment (slow-growth) (Figure S8).

It was found that the most abundant differentially expressed proteins (DEPs) were associated with vesicle trafficking, carbohydrate metabolism, protein homeostasis, C4 photosynthesis, and cell wall organization (Figure 6). Notably, the category of C4 photosynthesis was dominated by catalytic activities associated with PEP carboxylase and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) activities.

The expression of 301 proteins were higher (FDR $p<0.05$) in the internodes at peak biomass accumulation, and 252 were higher during peak sucrose accumulation during normal growth (Figure 7A). A similar pattern of protein expression was evident in the internodes, with a very significant reduction in growth and total biomass due to the presence of MODDUS (Figure 7B).

Only 240 proteins were deferentially expressed between the internodes at peak sucrose accumulation (Figure 7C).

3.6.1. Carbohydrate Metabolism and Glycolysis

A very similar pattern of differential gene expression between internodes at peak biomass and peak sucrose accumulation was evident in large internodes (control) versus smaller internodes (MODDUS treatment) (Figure 8 and Figure S7). Proteins involved in sucrose breakdown, oxidative pentose phosphate pathway, nucleotide sugar, and pentose metabolism were significantly higher in the internodes at peak biomass accumulation. In contrast, the proteins involved in hexose-P metabolism, oligosaccharide synthesis, and fermentation were more abundant in the internodes with peak sucrose accumulation.

Pyrophosphate-dependent phosphofructokinase, pyruvate kinase, and enolase were significantly reduced during the transition from rapid growth to sucrose accumulation. Sucrose synthase (SuSy) activity exhibited a complex pattern of change, with SuSy 2 reduced and SuSy 1 and 4 increased during the transition from growth to sucrose accumulation.

Vacuolar and cell wall invertases were more abundant during peak growth and both were absent during peak sucrose accumulation (Figure S8).

During peak sucrose accumulation in the presence and absence of MODDUS, proteins involved in cellulose, hemicellulose (Figure 8 and Figure S9), and vesicle trafficking (Figure S10) were significantly reduced.

The data clearly demonstrate that both cellulose and hemicellulose synthesis were largely under coarse control, with a significant reduction in the levels of key enzymes in the two pathways (Figure 8A,B).

3.6.2. Cell Wall Sugars

Key enzymes of cellulose synthesis such as cellulose synthase (CesA) and endo-1,4-beta-glucanase (EG), and hemicellulose synthesis such as xylan O-acetyltransferase (XOAT), xylosyltransferase (XT) and xylan alpha-1,3-arabinosyltransferase (XAT), were significantly reduced in the older internodes that were accumulating sucrose.

One of the major metabolic changes during the transition from growth to sucrose accumulation resides in the pathway of nucleotide sugar production (Figure 8A,B and Figure S10). UDP-glucose dehydrogenase (UGD), UDP-glucose decarboxylase (AXS),UDP xylose synthethase (UXS), rhamnose synthase(RhaS), xylose 4 epimerase (UXE), and arabinose mutase (AraM) levels were all significantly decreased in the transition to sucrose accumulation.

3.6.3. Lignin

Evidently, there were two types of changes in the proteins that are associated with lignin metabolism (Figure S9). During rapid internode expansion and the biomass accumulation phase, lignin peroxidases were more abundant than during the sucrose accumulation phase. In contrast, proteins associated with monolignol biosynthesis such as caffeoyl-CoA 3-O-methyltransferase (CCoA-OMT), caffeoyl-CoA 3-O-methyltransferase (CCoA-OMT), cinnamoyl-CoA reductase (CCR), and p-coumaroyl-CoA: monolignol transferase (PMT) had significantly higher levels during the sucrose and lignin accumulation phase than during rapid growth.

3.6.4. Vesicle Trafficking

More than 460 proteins identified in this study were associated with vesicle trafficking. The expression of 78 of these proteins differed significantly (p < 0.05) during internode development. ECHIDNA (ECH) and the YPT/RAB guanosine triphosphatase (GTPase) interacting proteins were some of the major proteins that decreased after maximum internode elongation. RAB GTPases and retrograde trafficking, mainly mediated by Coat Protein I (COPI), and SNARE membrane fusion complexes dominated the vesicle trafficking category.

3.6.5. Vacuolar Metabolism

There were 203 tonoplast associated proteins among the identified proteins in the sugarcane internodes. Most of these were constitutively present. Twenty three proteins were differentially expressed between maximum growth and sucrose accumulation. A monosaccharide transporter and several tonoplast sucrose transporters (SUT and TST) were more highly expressed in the sucrose accumulation phase. Anion- and cation transporters, pyrophosphate-, and ATP proton translocators were more highly expressed in the rapid growth phase (Figure S12).

4. Discussion

In the sugarcane culm, the top internodes represent a range of development stages [4,24,35]. Internode 1 is starting to elongate, internode 3 attains maxim elongation, and by internode 5 elongation is completed. As an internode develops, the size of individual cells, but not the number increases [12]. Internode elongation is completed after approximately 150 °Cd (base temperature 18 °C) [4,35,36].

Rapid cell elongation would require not only rapid cell wall synthesis but new membrane materials, proteins, and other building blocks. Rapid cell enlargement is also highly energy-dependent and more than 50% of incoming sugars might be burned for energy production to support growth [37].

The importance of these processes was also evident from the data reported here, where the significant protein changes were associated with vesicle trafficking, cell wall organization, respiration, redox homeostasis, and cytoskeleton organization.

4.1. Sink Strength

Most of the reduced carbon required to support culm development and growth in sugarcane is derived from sucrose that is imported symplastically, see for a review [38].

Sink strength, is reflected by the increase in biomass of the internodes [4,6]. It is important to note that biomass accumulation is an underestimation of carbon input into the tissue as more than 50% of incoming sugars might be burned for energy production to support growth [37].

Maximum sink strength is around internode 4 to 5, which coincides with the maximum increase in internode length. However, biomass accumulation continues up to internodes 10 to 12.

Internode growth requires two processes: cell wall synthesis and cell expansion. Cell wall synthesis demands a lot of energy and resources, while cell expansion involves the uptake of water through osmotic potential and controlled changes in the cell wall’s extensibility [4]. Initially, cell wall synthesis is the most significant demand for carbon in young internodes. As internodes mature, the biggest demand for carbon shifts to sucrose accumulation. The point at which the demand for carbon shifts from cell wall synthesis to sucrose accumulation depends on the growth rate. When the growth rate slows down towards the end of the growth season or with the application of MODDUS, the crossover happens earlier in internode development. It is important to note that the demand for carbon from the other metabolic pools diminishes rapidly as internode elongation slows down. In the late stages of development, especially in the fast-growing culm, sucrose accumulation exceeds total biomass gain. This can only imply that other cellular constituents are converted to sucrose at this stage.

There has been a significant increase in the amount of knowledge available about changes in metabolite [39,40,41] and gene expression profiles [9,40,42,43] due to modern high throughput platforms. However, the connection between metabolic changes in the developing internodes in sugarcane and the final yield and sugar characteristics is not well understood. This is mainly due to a failure to demonstrate the relationship between mRNA and protein levels, or a lack of description of the development stage and phenotype.

The metabolome and proteome data presented in this manuscript provide significant insights into some of the potential metabolic control points in culm sink strength.

4.2. Carbon Availability and Regulation of Cell Wall Synthesis

The data presented here suggest that it is unlikely that UDP-glucose availability plays a significant role in determining the rate of cell wall synthesis. Instead the data from this study indicated that cell wall synthesis in the sugarcane culm is primarily under coarse control though gene expression of key enzymes in the synthesis and transport of cell wall sugars.

Early studies suggested that SuSy plays a role in the production of UDP-Glc, which is used in cellulose biosynthesis [44]. However, recent genetic analyses in Arabidopsis and aspen plants have shown that mutations or downregulation of SuSy genes have little effect on cellulose synthesis and plant growth [45]. This suggests that SuSy is unlikely to play a significant role in supplying UDP-Glc for cellulose synthesis. In contrast, mutations of two Arabidopsis cytosolic invertase genes cause severe growth defects, indicating that the invertase-mediated pathway is the main route for supplying UDP-Glc for cellulose biosynthesis [45]. Further analysis has shown that downregulation of a wood-associated cytosolic invertase gene in aspen reduces the amount of crystalline cellulose and UDP-Glc in the plant’s wood.

Therefore, the available evidence suggests that the invertase-mediated pathway is the predominant route for supplying UDP-Glc for cellulose biosynthesis. The data presented in this study showed that neutral invertase activity is high during maximum growth and remains present even at very late stages of sucrose accumulation. Previously, we also showed that downregulation of neutral invertase in sugarcane results in significant suppression of growth [46].

The composition of the cell wall does not significantly vary during the different times of the season and growth rates. This indicates that the synthesis of various components of the cell wall is closely regulated and coordinated with the growth of cells and organs. During extension growth, it is necessary to synchronize the synthesis of all the cell wall components with changes in the extensibility of the matrix component, osmolality, and the uptake of water [47].

The most striking change observed in protein levels during the transition from rapid growth to sucrose accumulation is the widespread decrease in many proteins in the nucleotide sugar synthesis pathway. These data are consistent with the rate cell wall synthesis being controlled by prevailing enzyme levels (coarse control) rather than through the availability of UDP-glucose.

The genes responsible for the synthesis of primary and secondary cell walls are expressed in a coordinated manner [9,48,49,50]. However, the transcriptional program that regulates this process is not well understood and appears to involve developmental gradients and brassinosteroids [47]. The abundance of proteins is also regulated at the level of translation [51] and turnover [52]. In Arabidopsis, it has been found that several proteins involved in nucleotide-sugar metabolism are subject to relatively rapid degradation [53]. The regulation of cellulose synthesis is also influenced by protein phosphorylation [46].

During the transition to sucrose storage in sugarcane internodes, many components of this cytoskeleton and vesicle network decrease. The most significant change in protein levels in developing sugarcane internodes is related to vesicle transport.

Cellulose and secretion, two critical components of cell wall construction, are influenced by various factors, including endomembrane pH, ion activity, changes in osmotic pressure, and cell sugar status. The initial steps of this process seem to be controlled by a protein complex that is composed of ECHIDNA (ECH) and the YPT/RAB guanosine triphosphatase (GTPase) interacting proteins. Both of these components peak during maximum internode elongation.

4.3. Sucrose Metabolism

Because the import of sucrose into the culm parenchyma is symplastic, sucrose importation into the internodes is dependent on maintaining a low sucrose concentration in the cytosol [38,54,55]. Maintaining a low sucrose concentration in the cytosol of the parenchyma cells is dependent on sucrose breakdown to provide carbon skeletons for biosynthesis and energy provision, and loading of sucrose into the vacuole [1,4,5]. Intracellular sucrose catabolism is via SuSy or cytosolic and vacuolar invertases [47]. Susy directly produces UDP-glc, which is the substrate for other nucleotide sugars needed for cell wall biosynthesis [7]. It is therefore not surprising that much research on the role of sucrose synthase (SuSy) in the accumulation of sucrose in sugarcane has been carried out [35,56,57].

When catabolism of sucrose is through the invertase route, entry of the reducing sugars into metabolism is dependent on hexokinase and fructokinase activities. As reflected in the data from the current study, hexokinase and fructokinase activities are high in sugarcane internodes [58] and hexoses are rapidly phosphorylated in the cytosol [22,23].

The changes in the levels of SuSy proteins reflect significant complexity during internode development. Although SuSy 2 levels decrease significantly during the transition from rapid growth to sucrose accumulation, SuSy 1 and Susy 4 increase. It is known that these two forms of SuSy are specifically induced by stress and hypoxia [47,59].

Both the cell wall and vacuolar invertases decrease rapidly during internode development and are absent at the stage where the internode development switches to sucrose accumulation. This was previously shown to be the case during the culm development of sugarcane [60,61]. Cytosolic invertase also decreases during internode development but is still abundant in the mature internodes.

Modeling of internode metabolism based on metabolism of${ }^{14}C$ sugars showed that more than 90% of free glucose and fructose reside in the vacuole. Hydrolysis of sucrose in the vacuole has the advantage of lowering the osmotic potential in the vacuole and facilitating cell expansion through increased water uptake into the vacuole.

Sucrose accumulation initially occurs in the symplastic compartment, possibly by loading the sucrose into the vacuole. However, as sucrose accumulation progresses, especially after internode length expansion ceases, sucrose also leaks into the apoplastic space, which can take up as much as 20% of the total internode volume.

It is interesting to note that the expression of sucrose tonoplast transporters (TSTs) was more highly expressed in the sucrose accumulation phase in the internodes. In contrast anion- and cation transporters, and pyrophosphate- and ATP proton translocators, were more highly expressed in the rapid growth phase of the internodes.

5. Conclusions

This study confirmed that the duration of internode elongation is independent of the crop’s growth rate. Instead, it is controlled by thermal time.

The sink strength of the internode is directly related to the growth rate, and the synthesis of cell wall polysaccharides and respiratory metabolism are the main internal demands on reduced carbon. The cell wall chemical composition of sugarcane seems to be independent of the growth rate or time of the season. This suggests that harvesting time should have little impact on the quality of the lignocellulosic fraction, and this should facilitate the development of processing technologies.

Both cellulose and hemicellulose synthesis seem to be primarily under coarse control. Many of the enzymes involved in cell wall synthesis are reduced during the transition from rapid growth to sucrose accumulation. Genetic manipulation of the duration of sugarcane cell wall synthesis will, therefore, pose a significant challenge, and it might be better to focus attention on the molecular basis of thermal time control of the process.

The results of this study demonstrate that sucrose accumulation is primarily controlled by four factors: reduced demand for UDP-glc for cell wall synthesis, reduced vacuolar invertase activity, increased expression of tonoplast sucrose transporters, and the volume (size) of the internode.