Reducing nitrogen with dense planting increase sweetpotato storage roots quantity via favoring starch synthesis over lignification

TL;DR

Reducing nitrogen and using dense planting improves sweetpotato root growth by boosting starch production and reducing lignin.

Contribution

The study reveals how medium nitrogen with dense planting enhances sweetpotato root development through hormonal and genetic mechanisms.

Findings

MDMN increased root number, length, and volume compared to traditional low-density high-N methods.

MDMN upregulated genes for starch synthesis and downregulated lignin-related genes.

MDMN altered hormone levels to favor starch accumulation over lignification.

Abstract

Excessive nitrogen (N) inhibits crop development, yet the synergistic effects of reducing nitrogen with dense planting on sweetpotato storage root (SR) formation remain unclear. This study evaluated four N levels (0, 23, 46, and 69 mg N kg−1 soil) and three planting spacings (0.25, 0.20, and 0.15 m). Results indicated that medium density with medium N (MDMN) significantly outperformed the traditional low-density high-N (LDHN) treatment. From 15 to 30 days after planting (DAP), MDMN enhanced root number, length, and volume, while by 45 DAP it increased SR number, root activity, and N accumulation. MDMN upregulated IbKNOX1 and IbARF5, elevating ZR and IAA levels while reducing GA3, thereby promoting starch accumulation via increased IbSUS and IbAGP expression. Concurrently, lignin biosynthesis was suppressed through downregulation of IbPOD and lignin-related genes. Collectively, MDMN promotes SR formation by optimizing hormonal balance to favor starch synthesis over lignification.

•MDMN improved storage root number, and N uptake vs. high-N low density•MDMN increased ZR and IAA but decreased GA3 during storage root formation•MDMN upregulated starch biosynthesis genes, boosting sucrose and starch accumulation•MDMN suppressed lignin biosynthesis, lowering lignin and promoting storage root initiation

MDMN improved storage root number, and N uptake vs. high-N low density

MDMN increased ZR and IAA but decreased GA3 during storage root formation

MDMN upregulated starch biosynthesis genes, boosting sucrose and starch accumulation

MDMN suppressed lignin biosynthesis, lowering lignin and promoting storage root initiation

Agricultural science; Agronomy; Agricultural techniques; Agricultural plant products

Figures

Climate during the sweetpotato growing seasonFigure 2Layout of pot test(A) The plant spacing of plants in each pot. The units of numbers in the figure are meters. The seedlings on both sides of the pot and between treatments are not used as selection for investigation (plants marked with × at the top of the figure). The actual sampling range for each treatment is six plants (plants marked with 1–6 at the top of the figure).(B) Test layout for 36 pots.Figure 3Dynamic changes of sweetpotato root traits and biomass under different planting densities and nitrogen application treatments(A) Phenotyp

Dynamic changes of N distribution and root activity under different planting densities and N application treatments(A) N content in different parts of sweetpotato plant.(B) N accumulation in sweetpotato roots.(C) N distribution ratio of sweetpotato roots.(D) Sweetpotato root activity. Planting distances, LD (0.25 m), MD (0.20 m). N application levels, MN (46 mg N kg^−1^ soil), HN (69 mg N kg^−1^ soil). DAP, days after planting. One-factor ANOVA, LSD. Data are represented as mean ± SEM. The error line represents the standard error of the mean (three biological replicates). ∗, ∗∗, and ∗∗∗, signi

Dynamic changes of sucrose and starch contents, related enzyme activities and gene expression in sweetpotato under different planting density and N application(A) Sucrose and starch contents in sweetpotato root.(B) SUS and AGPase activities in sweetpotato root.(C) Five significantly expressed genes of *IbSUS* and *IbAGPase*.Planting distances: LD (0.25 m), MD (0.20 m). N application levels, MN (46 mg N kg^−1^ soil), HN (69 mg N kg^−1^ soil). DAP, days after planting. One-factor ANOVA, LSD. Data are represented as mean ± SEM. The error line represents the standard error of the mean (three biolo

Dynamic changes of lignin synthesis indexes in sweetpotato under different planting densities and N application treatments(A) relative expression of lignin biosynthesis genes in sweetpotato roots.(B) POD enzyme activity in sweetpotato root.(C) Lignin content in sweetpotato root.Planting distances, LD (0.25 m), MD (0.20 m). N application levels, MN (46 mg N kg^−1^ soil), HN (69 mg N kg^−1^ soil). DAP, days after planting. One-factor ANOVA, LSD. Data are represented as mean ± SEM. The error line represents the standard error of the mean (three biological replicates). ∗, ∗∗, and ∗∗∗, significant

Dynamic changes of hormone-related gene expression and hormone content in sweetpotato under different planting densities and N application treatments(A) *IbKNOX1* gene expression and ZR content in sweetpotato.(B) *IbARF5* gene expression and IAA content in sweetpotato.(C) Gene expression and GA content of *IbKO*, *IbGDI1*, and *IbGA20ox* in sweetpotato. One-factor ANOVA, LSD. Data are represented as mean ± SEM. The error line represents the standard error of the mean (three biological replicates). ∗, ∗∗, and ∗∗∗, significant differences at *p* < 0.05, *p* < 0.01 and *p* < 0.001, respectively.

Full text

Introduction

Sweetpotato ranks as the seventh most cultivated food crop globally, and it holds great potential in addressing the issue of global food shortages.1 Storage root (SR) is the main harvest organ of sweetpotato, generally differentiates from adventitious roots (ARs, sprout from trimmed stems). AR subsequently progress into fibrous root (FR) with a diameter smaller than 2 mm, followed by the development of thicker roots ranging from 2 to 5 mm, ultimately evolving into SR exceeding 5 mm in diameter.2^,^3^,^4 AR with higher values of lateral root number, length, and surface area are more likely to differentiate into SR.5

In the context of sweetpotato SR formation, endogenous phytohormones such as auxin (IAA), gibberellin (GA), and cytokinin (CTK) play pivotal roles in regulating lignin content and the activity of cambium cells.6 High IAA concentrations accelerate cell division and promote sweetpotato SR formation, while low IAA concentrations increase root lignification and inhibit SR formation.7^,^8 Auxin response factors (ARFs) are a specific class of proteins involved in auxin signaling pathways and have been implicated in these regulatory processes.9 The expression of ARF5 is localized to the cambium region, where it plays a dual role in restricting the population of undifferentiated cambium cells and promoting the differentiation of vascular cambium cells.10 Exogenous GA_3_ upregulated lignin biosynthesis genes (IbPAL, IbC4H, Ib4CL, IbCCoAOMT, and IbCAD) in young sweetpotato roots, enhancing lignin deposition and xylem development, thereby reducing SR diameter and number.11 In addition, gibberellin biosynthesis is positively regulated by ent-kaurene oxidase (KO), GA 20-oxidase (GA2Oox), and GA Insensitive Dwarf1 (GID1) genes.12^,^13 Zeatin riboside (ZR), a CTK derivative, exhibits rapid accumulation in the vascular cambium during the early stages of SR formation, facilitating cell division and the development of SRs.14 Research by Kondo et al.15 have demonstrated that the activation of CTK signaling hinders primary xylem development and the formation of protoxylem vessels in Arabidopsis roots. Knotted 1-like (KNOXI) positively regulates CTK biosynthesis16^,^17 and negatively regulates GA biosynthesis.18^,^19 The KNOX1 protein controls cell differentiation by negatively regulating lignin biosynthesis in Arabidopsis.20 In sweetpotato, KNOXI expression patterns resemble ZR distribution in the vascular cambium, with expression levels twice as high in young SRs as in FR.21

Hormones also influence formation of sweetpotato SR by regulating—sucrose to starch conversion.22^,^23 For example, IAA is involved in AR development24^,^25 and enhances ADP-glucose pyrophosphorylase (AGPase) activity, thereby promoting SR starch synthesis in sweetpotato.4^,^26^,^27 CTK induces the formation of sweetpotato SR once sucrose concentration reaches a critical level.28^,^29 In addition, CTK activates AGPase and starch synthase (SS) during potato tuber initiation, promoting starch accumulation.30^,^31^,^32 The increase in ZR content in rice and corn grains can promote starch accumulation.33^,^34 The exogenous application of GA_3_ down-regulated IbAGPase expression, thereby inhibiting starch synthesis and SR formation.11^,^35^,^36^,^37

Excessive nitrogen (N) fertilization disrupts carbon-nitrogen metabolic coordination in crops, suppressing sink organ formation while diminishing N use efficiency (NUE), thereby exacerbating economic losses despite elevated input costs.38^,^39 Reducing N levels while employing dense-planting practices can enhance N utilization efficiency, increase the number of effective panicles, and improve various morphological parameters of the root system (such as fresh weight, volume, count, length, and dry weight of the roots) in rice.40^,^41 Similarly, this N reduction combined with dense planting has been found to elevate the panicle number and grain yield per unit area in maize.8^,^42 However, the effect of reducing N with dense planting on sweetpotato N utilization, root growth and development is currently not well understood. In summary, there is limited knowledge regarding how the combination of reducing N levels with dense-planting influences sweetpotato phytohormone metabolism, carbohydrate metabolism, lignin biosynthesis, and its implications for SR formation. To address these knowledge gaps, multiple planting density and N level treatments were designed in this study to explore the effect of reduced N with dense planting on sweetpotato SR formation. The study aimed to elucidate its effects on phytohormone metabolism, carbohydrate metabolism, and lignin biosynthesis throughout the process of sweetpotato SR formation.

Results

Root morphological traits

During the sweetpotato growing season (Figure 1), multiple planting densities and nitrogen (N) levels were applied (Figure 2) to investigate the effects of reduced N input combined with dense planting on storage root (SR) formation. Based on these treatments, representative phenotypic images of LDHN and MDMN from 15 to 45 days after planting (DAP) are presented in Figure 3A. The effects of planting density, N level, and their interaction on total root length (TRL), total root surface area (TRSA), and TRSA per unit area as well as on total root volume (TRV), average root diameter (ARD), total root tip number (TRTN), and TRTN per unit area were significant at 15 days after planting (DAP). Under the same planting density, TRL, TRSA, TRSA per unit area, TRV, ARD, TRTN, and TRTN per unit area initially exhibited both increasing and decreasing trends with a decrease in N level, peaking at MN. Conversely, under the same N level, as planting density increased, TRSA per unit area increased in HN and NN. Whereas, in MN and LN, the measurements initially exhibited an increase followed by a decrease, reaching their peak at MD. TRTN per unit area followed a similar pattern, peaking at MD. Meanwhile, TRL, TRSA, TRV, ARD, and TRTN decreased. Compared to low density with medium N (LDMN), medium density with medium N (MDMN) exhibited a significant decrease in TRL, TRSA, TRV, and TRTN, but a significant increase in TRSA per unit area and TRTN per unit area (Table S2). Compared with LDHN, MDMN significantly (p < 0.05) increased TRL, TRSA, TRSA per unit area, TRV, ARD, TRTN, and TRTN per unit area (Figure 3B).Figure 1. Climate during the sweetpotato growing seasonFigure 2Layout of pot test(A) The plant spacing of plants in each pot. The units of numbers in the figure are meters. The seedlings on both sides of the pot and between treatments are not used as selection for investigation (plants marked with × at the top of the figure). The actual sampling range for each treatment is six plants (plants marked with 1–6 at the top of the figure).(B) Test layout for 36 pots.Figure 3. Dynamic changes of sweetpotato root traits and biomass under different planting densities and nitrogen application treatments(A) Phenotypic changes of sweetpotato roots at 15, 30, and 45 DAP under different treatments. The scale bars represent 5 cm.(B) Effect of treatments on root morphological traits at 15 DAP (TRL, total root length; TRSA, total root surface area; TRV, total root volume; ARD, average root diameter; TRTN, total root tip number).(C) Dynamic changes in related root biomass and development at 15, 30, and 45 DAP. Planting distances: LD (0.25 m), MD (0.20 m). N application levels, MN (46 mg N kg^−1^ soil), HN (69 mg N kg^−1^ soil). DAP, days after planting. One-factor ANOVA, LSD. Data are represented as mean ± SEM; values followed by lowercase letters within a column are significantly different among treatments (p < 0.05). The error line represents the standard error of the mean (three biological replicates). ^∗^, ^∗∗^, and ^∗∗∗^, significant differences at p < 0.05, p < 0.01 and p < 0.001, respectively. ns represents non-significance.

As shown in Table S3, planting density and N level significantly (p < 0.05) affected root number per plant and per unit area, potential storage root (PSR) length and diameter, FR weight per plant, and root weight at 15 and 30 DAP. No significant (p > 0.05) effect was observed on single PSR weight at 15 DAP, while significant (p < 0.05) effects were detected on single PSR weight and root weight per unit area at 30 DAP. N level significantly (p < 0.05) influenced root weight per unit area at 15 DAP, whereas planting density had no significant (p > 0.05) effect. Their interaction significantly (p < 0.05) affected root number and weight per plant at 15 DAP, root number per unit area and PSR length at 30 DAP, as well as PSR diameter and FR weight per plant at both 15 and 30 DAP. Under the same planting density, reductions in N level led to initial increases followed by decreases in root number per plant and per unit area, PSR length, diameter and weight, FR weight per plant, and root weight per plant and per unit area, with peaks observed at MN. Under the same N level, increases in planting density resulted in reductions in root number per plant, PSR length and diameter, single PSR weight, and FR and root weight per plant, whereas root number and weight per unit area initially increased and then decreased, also peaking at MN. Furthermore, compared with LDMN, MDMN exhibited reductions in root traits per plant but a significant (p < 0.05) increase in root number per unit area. Although no significant difference (p > 0.05) was found in root weight per unit area, an increasing trend was observed. Compared with LDHN, MDMN significantly (p < 0.05) increased root number and weight per plant and per unit area. While single PSR weight showed no difference (p > 0.05) at 15 DAP, it was significantly (p < 0.05) higher at 30 DAP (Figure 3C).

As shown in Table S4 at 45 DAP, both planting density and nitrogen level significantly (p < 0.05) affected SR number per plant and per unit area, SR diameter, single SR weight, and SR weight per plant and per unit area. Nitrogen level significantly (p < 0.05) influenced SR length, whereas planting density had no significant (p > 0.05) effect. Their interaction exhibited a significant (p < 0.05) effect only on SR diameter. Under the same planting density, reducing nitrogen level initially increased then decreased SR number per plant and per unit area, SR length and diameter, and SR weight per plant and per unit area, with peak values at the MN level. In contrast, single SR weight consistently decreased. At the same nitrogen level, increasing planting density reduced SR number per plant, SR length and diameter, single SR weight, and SR weight per plant. Conversely, SR number and weight per unit area initially increased then decreased, peaking at the MN level. Compared to LDMN, MDMN showed reduced SR traits per plant but exhibited an increasing trend in SR number and weight per unit area. Relative to LDHN, MDMN significantly (p < 0.05) increased SR number and weight per unit area, as well as singleSR weight (Figure 3C), despite no differences (p > 0.05) in per-plant metrics. Under these experimental conditions, MDMN outperformed all other planting density-nitrogen level combinations. Future studies will investigate its effects on root nitrogen uptake capacity, nitrogen distribution, starch synthesis, lignin biosynthesis, and phytohormone metabolism in comparison to LDHN.

Root uptake ability and N distribution

Compared to LDHN, MDMN showed a significant decrease (p < 0.01) in the N content of roots, leaves, and stems from 15 to 45 DAP, except for roots at 15 DAP (Figure 4A). Meanwhile, MDMN significantly (p < 0.05) increased root N accumulation, distribution, and activity at 15, 30, and 45 DAP (Figures 4B–4D).Figure 4. Dynamic changes of N distribution and root activity under different planting densities and N application treatments(A) N content in different parts of sweetpotato plant.(B) N accumulation in sweetpotato roots.(C) N distribution ratio of sweetpotato roots.(D) Sweetpotato root activity. Planting distances, LD (0.25 m), MD (0.20 m). N application levels, MN (46 mg N kg^−1^ soil), HN (69 mg N kg^−1^ soil). DAP, days after planting. One-factor ANOVA, LSD. Data are represented as mean ± SEM. The error line represents the standard error of the mean (three biological replicates). ∗, ∗∗, and ∗∗∗, significant differences at p < 0.05, p < 0.01 and p < 0.001, respectively. ns represents non-significance.

Carbohydrate metabolism

Compared to the LDHN, MDMN significantly increased the contents of and sucrose starch from 15 to 45 DAP (p < 0.01). Moreover, compared with LDHN, MDMN significantly (p < 0.05) increased SUS and AGPase activity, expression of IbSUS2, IbSUS5, IbAGPS1, IbAGPS2, and IbAGPL2 in roots from 15 to 45 DAP (Figures 5B and 5C).Figure 5. Dynamic changes of sucrose and starch contents, related enzyme activities and gene expression in sweetpotato under different planting density and N application(A) Sucrose and starch contents in sweetpotato root.(B) SUS and AGPase activities in sweetpotato root.(C) Five significantly expressed genes of IbSUS and IbAGPase.Planting distances: LD (0.25 m), MD (0.20 m). N application levels, MN (46 mg N kg^−1^ soil), HN (69 mg N kg^−1^ soil). DAP, days after planting. One-factor ANOVA, LSD. Data are represented as mean ± SEM. The error line represents the standard error of the mean (three biological replicates). ∗, ∗∗, and ∗∗∗, significant differences at p < 0.05, p < 0.01 and p < 0.001, respectively. ns represents non-significance.

Lignin biosynthesis

In Figure 6A, compared with LDHN, MDMN increased the expression of IbPAL, IbC4H, Ib4CL, IbHCT, IbCCoAMOT, and IbCAD from 15 to 45 DAP. Significant differences were observed in the expression of IbPAL and Ib4CL at 45 DAP, IbC4H from 30 to 45 DAP, and IbHCT, IbCCoAMOT, and IbCAD from 15 to 45 DAP. However, MDMN significantly (p < 0.01) elevated the expression of IbPOD at 15 DAP, but significantly (p < 0.001) decreased it from 30 to 45 DAP. Furthermore, MDMN significantly (p < 0.01) increased root peroxidase (POD) activity and lignin content at 15 DAP, but significantly (p < 0.05) decreased both at 30 and 45 DAP (Figures 6B and 6C).Figure 6. Dynamic changes of lignin synthesis indexes in sweetpotato under different planting densities and N application treatments(A) relative expression of lignin biosynthesis genes in sweetpotato roots.(B) POD enzyme activity in sweetpotato root.(C) Lignin content in sweetpotato root.Planting distances, LD (0.25 m), MD (0.20 m). N application levels, MN (46 mg N kg^−1^ soil), HN (69 mg N kg^−1^ soil). DAP, days after planting. One-factor ANOVA, LSD. Data are represented as mean ± SEM. The error line represents the standard error of the mean (three biological replicates). ∗, ∗∗, and ∗∗∗, significant differences at p < 0.05, p < 0.01 and p < 0.001, respectively. ns represents non-significance.

Phytohormone metabolism

In comparison to LDHN, MDMN significantly (p < 0.05) increased root IbKNOX1, IbARF5 expression, as well as the concentrations of ZR and IAA from 15 to 45 DAP (Figures 7A and 7B). Furthermore, MDMN significantly (p < 0.01) increased IbKO, IbGID1, and IbGA2OoX expression and GA_3_ concentration at 15 DAP. However, MDMN significantly (p < 0.05) decreased IbKO, IbGID1, and IbGA2OoX expression, and GA_3_ concentration at 30 and 45 DAP, except for IbGID1 expression at 45 DAP and GA_3_ concentration at 30 DAP (Figure 7C).Figure 7. Dynamic changes of hormone-related gene expression and hormone content in sweetpotato under different planting densities and N application treatments(A) IbKNOX1 gene expression and ZR content in sweetpotato.(B) IbARF5 gene expression and IAA content in sweetpotato.(C) Gene expression and GA content of IbKO, IbGDI1, and IbGA20ox in sweetpotato. One-factor ANOVA, LSD. Data are represented as mean ± SEM. The error line represents the standard error of the mean (three biological replicates). ∗, ∗∗, and ∗∗∗, significant differences at p < 0.05, p < 0.01 and p < 0.001, respectively. ns represents non-significance.

Discussion

Effect of reducing N with dense planting on sweetpotato root N uptake and distribution and development

Previous studies have indicated a significant interaction between planting density and N level in various crops like wheat,43 maize,44^,^45 and rice.46 In this study, this interaction also influenced sweetpotato SR diameter at 45 DAP, aligning with the conclusions drawn in earlier studies.47^,^48 However, the interaction did not exert a significant effect on sweetpotato SR length, number, or single root weight at 45 DAP, consistent with the findings of Guertal and Kemble.49 Importantly, the MDMN treatment increased the SR number per unit area compared with other treatments, highlighting the effectiveness of combining reduced N input with dense planting to enhance SR formation.

Moreover, the strategy of reducing N levels while increasing planting density has proven effective in enhancing various morphological parameters of root systems across different crops. Specifically, in rice,40^,^41 wheat,50 and soybeans,51 significant improvements were observed in root number, fresh weight, length, volume, and dry weight. Consistent with this pattern, the MDMN treatment in our study markedly increased adventitious root length, surface area, surface area per unit area, root volume, average root diameter, number of root tips, and root tips per unit area at 15 DAP, as well as the weight of PSR at 30 DAP. These improvements are not only quantitative but also qualitatively meaningful: they indicate that dense planting under reduced N conditions can stimulate root proliferation and functional activity. According to Villordon et al.,5 a well-developed root system is beneficial to SR formation. The findings of this study corroborate the hypothesis, compared with LDHN, the MDMN treatment led to an increase in the number of SRs per plant by 45 DAP. This highlights reducing N with dense planting as a promising agronomic approach for balancing nutrient efficiency with yield improvement in sweetpotato.

Additionally, MDMN significantly enhanced sweetpotato root activity by promoting better root development,52 aligning with the conclusions drawn Zhai et al.,42 who reported similar results in rice. A well-developed root with higher activity can also enhance its absorptive capacity,53 effectively improving N absorption.54 Reducing N with dense planting can promote root development and increase root activity, thereby promoting N uptake, as observed in crops like rice55 and maize.56 In this study, the MDMN treatment led to a decrease in root N content compared to LDHN, in line with the observations by Zheng et al.,43 who noted a significant reduction in N content in the wheat ear with reduced N levels and dense planting. However, the increase in root N accumulation and distribution observed with MDMN contrasts with the results of Zheng et al.43 This discrepancy could be attributed to two factors: firstly, an appropriate N supply may improve dry matter flow to the roots,57 leading to increased N accumulation and distribution in the roots; secondly, dense planting may enhance N acquisition capacity and promote N flow to productive organs, as observed in maize.58 Overall, reducing N with dense planting has a positive effect on sweetpotato root N uptake and distribution, ultimately stimulating root development. Taken together, these findings indicate that reducing N combined with dense planting does not merely conserve fertilizer inputs but also reprograms N allocation dynamics, thereby stimulating root development and improving overall nutrient-use efficiency. This integrated response provides a promising pathway for sustainable sweetpotato cultivation, balancing high productivity with reduced environmental costs.

Effect of reducing N with dense planting on carbohydrate metabolism of sweetpotato root

The main sign of SR formation is the development of vascular cambium cells into parenchyma cells and the accumulation of starch.11^,^35^,^59^,^60 Sucrose synthase (SUSy) and ADP glucose pyrophosphorylase (AGPase) are pivotal enzymes that determined the sucrose and starch content, and SR formation of sweetpotato.61^,^62^,^63 IbSUS2 and IbSUS5,64 along with IbAGPS1 and IbAGPS2,65 may play a key role in SR development. Previous studies have demonstrated that excessive N inhibits the SUS activity and starch content in maize66 and the AGPase activity as well as starch content in tartary buckwheat.67 Similarly, excessive N inhibited starch synthesis in sweetpotato roots by suppressing the expression of AGPa and AGPb.68 Furthermore, in wheat69^,^70 and rice,71 the strategy of reducing N with dense planting has been proven to significantly promotes starch synthesis and increases starch content. Our findings support this view; compared with LDHN, MDMN significantly increased the expression of IbSUS2, IbSUS5, IbAGPS1, IbAGPS2, and IbAGPL2, activities of SUS and AGPase, contents of sucrose and starch in sweetpotato roots at 15 to 45 DAP. Beyond demonstrating enzyme upregulation, these results point to a broader regulatory mechanism in which moderate N reduction coupled with dense planting alleviates the inhibitory effect of excessive N, thereby enhancing carbon partitioning into starch. This metabolic adjustment ensures sufficient sink strength for SR formation and highlights a promising strategy to synchronize nutrient supply with storage organ development.

Effect of reducing N with dense planting on sweetpotato root lignin biosynthesis

The suppression of genes associated with lignin biosynthesis leading to a reduction in lignin accumulation represents a critical process in SR formation.35 Nevertheless, lignin content alone may not fully determine the degree of lignification; for instance, in the Arabidopsis mutant stem, a substantial decrease in lignin content did not significantly alter the degree of lignification in the casparian strip.72 The findings of this study corroborate this observation. In comparison to the LDHN, MDMN significantly increased lignin content by upregulating the expression of IbPAL, IbC4H, Ib4CL, IbHCT, IbCCoAOMT, IbCAD, and IbPOD, as well as POD enzyme activity at 15 DAP. Despite these increases, anatomical analysis indicated that MDMN had a lower degree of root lignification. Interestingly, MDMN showcased a greater number of root vessels and a heightened N absorption capacity at 15 DAP, suggesting that higher root lignin content in MDMN facilitates the formation of xylem vessels, thereby enhancing the root’s water and nutrient transport capabilities.73 Fortuitously, in contrast to the LDHN, MDMN notably decreased lignin content along with IbPOD expression and POD enzyme activity between 30 and 45 DAP. This aligns with previous research indicating that the POD enzyme is crucial for determining lignin monomer polymerization and overall lignin content.74^,^75 Taken together, these results indicate that reducing N with dense planting does not simply decrease lignin, but rather modulates its temporal pattern to balance structural reinforcement with storage tissue development. This dual regulation ensures both efficient resource acquisition in the early stage and enhanced sink capacity in the later stage, thereby promoting SR formation and yield.

Effect of reducing N with dense planting on sweetpotato root endogenous phytohormone metabolism

Phytohormones act as central regulators of SR formation by coordinating cell division, vascular differentiation, and carbon allocation. KNOX1 positively regulates ZR concentration,76^,^77 while ARF5 positively responds to IAA levels.77 Additionally, KO, GA2Oox, and GID1 are involved in GA_3_ biosynthesis. Notably, higher IAA and ZR concentrations, coupled with lower GA_3_ concentrations, are advantageous for increased SR formation.78 Higher IAA and ZR concentrations have been shown to stimulate vascular cambium activity and reduce lignin content,22^,^78 while lower GA_3_ concentrations can decrease lignin content and inhibit lignification.79 An increase in ZR content in rice and corn grains can enhance starch accumulation.33^,^34 IAA enhances AGPase activity, thereby promoting SR starch synthesis sweetpotato,4^,^26^,^27 while the exogenous application of GA_3_ leads to the downregulation of IbAGPase expression, thereby inhibiting starch synthesis and SR formation.11^,^35^,^36^,^37 The results of this study indicate that compared with LDHN, MDMN significantly increased the expression of IbKNOX1 and IbARF5 in PSR, thus significantly elevating ZR and IAA levels, while simultaneously decreasing IbKO, IbGID1, and IbGA2OoX expression and GA_3_ concentrations in PSR. In line with these insights, MDMN predominantly enhances starch accumulation while suppressing lignin biosynthesis, thus facilitating more SR formation. These findings suggest that reducing N with dense planting reprograms the endogenous hormonal environment to create a developmental context that simultaneously promotes carbon sink formation and suppresses excessive structural reinforcement. Such hormonal fine-tuning may represent a key mechanism through which agronomic management practices translate into improved storage organ yield and quality. Importantly, it also highlights the potential of optimizing N input and planting density as a strategy to indirectly manipulate phytohormone signaling for sustainable sweetpotato production.

Conclusion

The cultivation model of dense planting combined with reduced N input (e.g., MDMN) represents a promising and sustainable strategy for enhancing sweetpotato productivity. Our findings suggest that this approach promotes SR formation by coordinating nutrient uptake with developmental processes: optimized N acquisition improves root system efficiency, enhances carbon partitioning toward starch deposition, and dynamically suppresses lignin synthesis, thereby favoring SR fomation. This integrated response highlights the potential of agronomic management to reprogram physiological and molecular pathways that determine yield and quality. At the same time, the current study also reveals key knowledge gaps. The precise regulatory networks linking N metabolism, carbohydrate allocation, lignin biosynthesis, and phytohormone signaling remain incompletely understood. Moreover, most evidence derives from expression profiles rather than functional validation, underscoring the need for genetic transformation and molecular biology approaches to verify the roles of pivotal genes. Future research should therefore focus on disentangling these complex regulatory interactions and establishing a comprehensive model of SR development under varying N and planting density regimes. Such work would not only advance our fundamental understanding of SR biology but also inform the design of high-efficiency, resource-saving cultivation systems for sustainable sweetpotato production.

Limitations of the study

This study elucidates the regulatory role of coordinated nitrogen reduction and optimized planting density in sweetpotato SR formation by integrating morphological, physiological, and molecular evidence. While the experiments were conducted under controlled pot conditions, this approach allowed precise manipulation of nitrogen supply and planting density, providing clear mechanistic insights; validation under field conditions will further strengthen the applicability of the findings. Moreover, the present analysis focused on a single cultivar (“Pushu32”), and future evaluation across diverse genetic backgrounds will help assess the broader relevance of the proposed regulatory framework. Finally, key candidate genes associated with hormone signaling, starch biosynthesis, and lignification were identified through integrative physiological and expression analyses, and their functional verification represents an important direction for subsequent studies.

Resource availability

Lead contact

Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Chengcheng Si (ccsi@hainanu.edu.cn).

Materials availability

This study did not generate new unique reagents.

Data and code availability

This article analyzes existing publicly available data and those datasets are listed in the key resources table. This article does not report original code. Any other additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments

We would like to express our gratitude for the financial support from the 10.13039/501100001809National Natural Science Foundation of China (32060716, 32560771), the Science and Technology Special Fund of Hainan Province (ZDYF2022XDNY264, ZDYF2025XDNY070), the Project of Sanya Yazhou Bay Science and Technology City (SCKJ-JYRC-2023-24), and the Innovation and Entrepreneurship Training Program of Hainan University Nanfan College (NFCX2024ZD-30).

Author contributions

Investigation, H.P. and X.H.; visualization, H.P. and X.H.; data curation, X.G., X.Z., R.A., H.X., and Y.L.; methodology, X.Z. and R.A.; writing – original draft preparation, H.P. and X.H.; writing – review and editing, C.S.; supervision, C.S.; funding acquisition, C.S. All authors have read and agreed to the published version of the manuscript.

Declaration of interests

The authors declare no potential conflict of interest.

STAR★Methods

Key resources table

REAGENT or RESOURCESOURCEIDENTIFIERChemicals, peptides, and recombinant proteinsUrea (46% N)Sinofert Holdings Limited, Beijing, ChinaN/APotassium sulfate (50% K_2_O)Sinofert Holdings Limited, Beijing, ChinaN/ACalcium superphosphate (16% P_2_O_5_)Sinofert Holdings Limited, Beijing, ChinaN/ACritical commercial assaysLignin Assay KitBoxbio, Beijing, ChinaCat# AKSU010UOligonucleotidesPrimers for *RT–qPCR,*see Table S1This paperN/ASoftware and algorithmsWinRHIZO 2009 root analysis softwareQuebec, CanadaN/AMicrosoft Excel 2016Microsoft,Redmond,WA,USAhttps://www.microsoft.comSPSS19.0IBM,Armonk,NY,USAhttps://www.ibm.com/products/spss-statisticsMedPeerBeijing MedPeer Technology Co., Ltd., Beijing, Chinahttps://www.medpeer.cnGraphPad Prism 8.0.2.263GraphPad Software Inc., San Diego, CA, USAhttps://www.graphpad.comCaseViewer 2.43DHISTECH CaseViewer, Budapest, Hungaryhttps://www.3dhistech.com/software/caseviewer/TBtools v1.098669Chen et al.1https://github.com/CJ-Chen/TBtoolsOtherEPSON EXPRESSION 10000XL scannerJapanN/AqTOWER3G RT-qPCR machineJena, GermanyN/ARNA extraction kitTiangen, Beijing, ChinaCat# DP437HiScript II Q RT SuperMixVazyme, Nanjing, ChinaCat# R223electronic balanceSartorius, GermanyMCA125S–2CCN–I–QP99Dumas N analyzerSKALAR, Breda, NetherlandsPrimacs SN100

Experimental model and study participant details

Sweetpotato plant materials

Cultivar selection: The sweetpotato cultivar used in this study was 'Pushu 32', selected as the experimental material to evaluate growth responses under different planting distances and nitrogen(N) application levels.

Growth environment: A pot experiment was carried out from October to November 2023 at the Sanya Nanfan Research Institute, Hainan University, located in Yazhou (18°30′ N, 110°59′ E). Detailed climate data for the study can be found in Figure 1 (available via https://www.qweather.com/).

Soil properties: Before the experiment, the soil used in the pot experiment was collected from the field. Its basic physicochemical properties are as follows: the organic matter in the top 20 cm of the soil profile was measured to be 0.82%, the soil was ascertained to contain 88.76 mg kg^-1^ of alkaline hydrolyzable N, 4.24 mg kg^-1^ of phosphorus (P), and 22.05 mg kg^-1^ of potassium (K). The pot experiment was conducted outdoors under natural environmental conditions.

Method details

Pot experiment design

The experimental design consisted of a split-plot arrangement, setting 3 plant distances (LD: 0.25 m, MD: 0.20 m, HD: 0.15 m) and 4 N levels (NN: 0 mg N kg^−1^ soil, LN: 23 N kg^−1^ soil, MN: 46 mg N kg^−1^ soil, HN: 69 mg N kg^−1^ soil). Among them, LDHN stands for traditional cultivation models.

In addition to N fertilizer, the pots were supplemented with phosphate fertilizer (40 mg P_2_O_5_ kg^−1^ soil) and potassium fertilizer (85 mg K_2_O kg^−1^ soil), and all fertilizers were applied as a base before planting. The fertilizers used in this study included urea (46% N), potassium sulfate (50% K_2_O content), and calcium superphosphate (16% P_2_O_5_), supplied by Sinofert Holdings Limited, Beijing, China.

The pot in the experiment was 5 m long, 0.3 m wide and 0.35 m high (Figure S1). 24 plants were planted per pot, planted in single rows (Figure 2A), and 9 pots were planted per treatment for sampling in 3 periods (15, 30 and 45 days after planting (DAP)), making a total of 36 pots planted (Figure 2B). The management practices in the experiment conform to standard agricultural practices. 24 plants were planted per pot in single rows, and the seedlings on both sides of the pot and between treatments were not used as selection for investigation (plants marked with × at the top of the figure). The actual sampling range for each treatment is 6 plants (plants marked with 1-6 at the top of the figure).

Investigation of agronomic traits and preparation of dry and fresh samples

Five plants were taken from each treatment (at least one plant was randomly taken from each pot) to investigate agronomic traits at 15, 30 and 45 DAP.

Root trait investigation: The number and weight of total roots, the weight of fibrous roots (FR), the length, diameter (thickest part) and weight of potential storage roots (PSR) were investigated on 15 and 30 DAP, and the six thickest roots per plant were set as PSR.22 The number and weight of stored roots (SR) per plant, the weight of Single SR, and the length and diameter of SR were investigated 45 DAP.

Root structure parameter determination: One plant was randomly taken from each pot 15 DAP, the root system of the plants was scanned using EPSON EXPRESSION 10000XL scanner (made in Japan), and the root structure parameters were obtained through WinRHIZO 2009 root analysis software (Quebec, Canada). The dry matter rate was calculated using the formula: dry weight/fresh weight × 100%.

Dry sample preparation: 6 plants were randomly collected from each treatment (2 plants were mixed evenly into 1 replicate). After the fresh weights of leaves, stems, PSR and SR were investigated, they were placed in an oven at 105°C for enzyme removal, spread out and dried, and then dried at 60°C to constant weight, weighed for dry weight, and calculated for dry matter rate (formula: dry weight/fresh weight × 100%). The dry samples were then ground to powder and stored for determination of N, sucrose, starch and lignin content.

Fresh sample preparation: PSR and SR from 6 plants (2 plants were mixed into 1 replicate) were randomly collected, cut into small pieces, placed in liquid N for quick freezing, and stored in a refrigerator at -80°C for determination of hormone concentration, enzyme activity and gene expression level.

N content and N utilization determination

10 mg of dried samples were weighed using an electronic balance (Sartorius, Germany, MCA125S–2CCN–I–QP99), transferred to a tin boat, and subsequently analyzed for N content at Dumas N analyzer (SKALAR, Breda, Netherlands, the Primacs SN100). Each treatment was repeated three times. The calculation formula are as follows:

$$[eqn]$$ $$[eqn]$$

Root activity determination

The sweetpotato root activity between 15 and 45 DAP was assessed using an enhanced version of the TTC method.80 An approximately 0–1 cm segment of the absorptive root tip was severed and washed in sterile H_2_O for 10 minutes. This root segment was subsequently placed in a test tube filled with 15 mL of TTC solution (0.08% TTC in a 0.05 M sodium phosphate buffer, pH 7.4), and incubated in darkness at 30 °C for 24 hours. After incubation, the TTC solution was discarded, and the root tip was again cleaned with sterile H_2_O. Subsequently, the cleaned root tip was submerged in 10 mL of 95% ethanol–water mixture and heated in a water bath at 80 °C for 20 minutes. Absorbance was measured at a wavelength of 485 nm to quantify the reaction. To ensure reliability, each condition was replicated three times, and the results were averaged for consistency.

Starch and sugar contents determination

The carbohydrate contents in the root of sweetpotato were determined using anthrone colorimetry according to the method of Taufik and Guntarti.81 Briefly, 5 mL of hyper pure water was added to the fine-dried potential storage root powder (0.1 g). The resulting powders were subjected to centrifugation at 3000 rpm for 5 minutes. This process was repeated four times, and the supernatants from each cycle were pooled in a 50 mL centrifuge tube for subsequent sucrose measurement. The precipitate was transferred to a 50 mL centrifuge tube with 10 mL of 3N hydrochloric acid (HCl) for starch determination. The above-mentioned sugar and starch contents were measured at 640 nm. The standard curve was established for sucrose dilution solution, and distilled water was added to a 2 mL anthrone reagent. Finally, the absorbance value was recorded at 640 nm.

Carbohydrate-metabolizing enzyme activity determination

SUS activity determination: The SUS activity in potential storage roots of sweetpotato was measured according to the method outlined by Tomlinson et al.82 Briefly, 0.2 g of potential storage root was weighed and added with 0.6 mL of enzyme extract, followed by the addition of steel beads for homogenization. The homogenate was centrifuged at 12000 rpm for 10 min after bead removal. Subsequently, 80 uL of enzyme solution (40 uL of supernatant and 40 uL of precipitation suspension) was mixed with 360 uL of enzyme analytical solution and incubated at 30°C water bath for 0.5 h. The absorbance at 340 nm was measured, and SUS activity was determined accordingly.

AGPase activity determination: The AGPase activity was measured as described by Kerr et al.83 The unit of enzyme activity was recorded as μmol min^-1^ g^-1^ FW for AGPase.

Lignin content determination

To quantify the lignin content in roots, a lignin assay kit (Boxbio, Beijing, China, AKSU010U) was utilized. For the assay, a 5 mg sample was combined with 500 μL of reagent 2 and 20 μL of perchloric acid in a glass test tube, and then thoroughly mixed. The mixture was then heated in a water bath maintained at 80 °C for 40 minutes. Following the heating process, the test tubes were cooled to room temperature naturally. Subsequently, an additional 500 μL of reagent 2 was introduced to the mixture, stirred thoroughly, and allowed to stabilize at room temperature. The clear supernatant was collected for further analysis. In a separate blank test tube, 20 μL of the supernatant was mixed with 980 μL of acetic acid (CH_3_COOH), and the absorbance of this mixture was measured at a wavelength of 280 nm.

Lignin enzymes activity determination

Total soluble proteins were extracted from the root tissues for Peroxidase (POD) activity analysis, following a modified version of the protocol by Li.84 Guaiacol was employed as the substrate, which, under the catalytic action of POD, is oxidized by H_2_O_2_ into a tea-brown coloration. The assay involved mixing 0.1 mL root extract supernatant with 1 mL of 2% H_2_O_2_, 1 mL of 0.05 M guaiacol, and 2.9 mL of 0.05 M phosphate buffer. This mixture was then incubated at 37°C for 15 minutes. A control was set by boiling the enzyme solution for 5 minutes to deactivate it. To halt the reaction, 2.0 mL of 20% trichloroacetic acid was added. The solution was centrifuged at 5,000 rpm for 10 minutes, and the absorbance of the supernatant was measured at 470 nm to determine POD activity. Each treatment was repeated three times to ensure the reliability of the results.

Phytohormones concentration determination

The content of ZR, IAA, and GA_3_ in sweetpotato roots was assessed using an enzyme-linked immunosorbent assay (ELISA), as per the method described by Yang et al.85 The antibodies utilized in the ELISA were developed by the Phytohormones Research Institute at China Agricultural University.

Each well of 96-well plates was coated with 100 μL of a buffer (1.5 g L^-1^ sodium carbonate, 2.93 g L^-1^ sodium bicarbonate, 0.02 g L^-1^ sodium azide, pH 9.6) containing 0.25 mg L^-1^ of specific antigens for these hormones. The plates designated for ZR and GA_3_ analyses were incubated at 37 °C for 4 hours, while those for IAA were incubated at 4 °C for 16 hours and subsequently allowed to stand at room temperature for 30 minutes.

Following incubation, the plates were washed four times using a solution composed of 1 L of phosphate buffer with 1 mL of Tween-20. Subsequently, 50 μL of either grain extracts or standard solutions of ZR, IAA, and GA_3_, along with 50 μL of 20 mg L^-1^ specific antibodies for each hormone, were added to each well. Following another similar incubation period, the wells were washed again four times.

A mixture of 100 μL of 10-20 mg o-phenylenediamine and 10 mL of a specific buffer solution (5.10 g citric acid, 18.43 g Na_2_HPO_4_·12H_2_O, 1 L water, 1 mL Tween-20, pH 5.0) was added to each well. Subsequently, 4 mL of 30% H_2_O_2_ was introduced and thoroughly mixed. The plates were then placed in a humidified chamber, and the reaction was stopped by adding H_2_SO_4_ to each well. The color development in each well was quantified using an ELISA reader (model EL310, Bio-TEK, Winooski, VT) at 490 nm.

Relative gene expression analysis

Total RNA was extracted from sweetpotato roots using a specific RNA extraction kit supplied by Tiangen (Beijing, China, DP437). Following the extraction, the cDNA synthesis was carried out using the HiScript II Q RT SuperMix, specifically designed for real-time quantitative PCR (RT-qPCR) applications (Vazyme, Nanjing, China, R223). This setup was intended to facilitate accurate quantification of gene transcript levels.

The RT-qPCR analyses utilized the SYBR Green I method, which utilizes a fluorescence chimeric dye for accurate detection. The fluorescent dye solution used in this process was TB Green® Premix Ex Taq™ II (RR820A, Takara Bio Inc, Beijing, China). The RT-qPCR procedures were performed with precision using gene-specific primers provided by Shanghai Shenggong Biological Company, China. These primers were utilized on a state-of-the-art qTOWER3G RT-qPCR machine (Jena, Germany).

For normalization and to ensure data accuracy, the housekeeping gene IbPLD was employed as an internal control within the experiments. Relative gene expression levels were determined by applying the 2^-ΔΔCT^ method, ensuring robust quantification across samples. The experimental protocol included three independent biological replicates to confirm the reproducibility of the results. The specific primers used in these assays are presented in Table S1, providing transparency and reproducibility for future research. The raw expression data of these genes are provided in Data S1.

Quantification and statistical analysis

Data collation and visualization: Data from this study were processed and analyzed using Microsoft Excel 2016 (Microsoft, Redmond, WA, USA) for calculating averages. A variety of software tools were employed to generate visualizations: MedPeer (http://www.medpeer.cn, accessed 07 April 2024) and GraphPad Prism 8.0.2.263 (GraphPad Software Inc., San Diego, CA, USA) for graphing data; CaseViewer 2.4 (3DHISTECH CaseViewer, Budapest, Hungary) and TBtools v1.09866986 for additional imaging and analysis; Microsoft PowerPoint 2016 (Microsoft, Redmond, WA, USA) to enhance presentation materials.

Statistical tests: including one-factor and two-way ANOVA, were performed using SPSS 19.0 (IBM, Armonk, NY, USA). Post-hoc comparisons were conducted using the least significant difference (LSD) test to ascertain statistical significance. All statistical details of experiments can be found in the figure legends and results section, including: (1) Exact value of n: n=3 for all measurements, representing biological replicates; (2) Definition of n: For dry/fresh sample preparation, gene expression analysis, enzyme activity assay, hormone concentration determination, N content measurement, and lignin content quantification, each biological replicate consists of 2 mixed plants; for root morphological trait analysis at 15 DAP, each biological replicate is 1 plant per pot (3 pots per treatment); (3) Measures of central tendency: Mean; (4) Measures of dispersion and precision: Standard error (SEM), and all quantitative data in the figures are presented as mean ± SEM. Statistical significance was denoted as follows: Values followed by lowercase letters within a column are significantly different among treatments (*P<*0.05). The error line represents the standard error of the mean (three biological replicates). ^∗^, ^∗∗^, and ^∗∗∗^, significant differences at P<0.05, P<0.01 and P<0.001, respectively. ns represents non-significance.