Abstract

   This study investigated the effects of long-term method of
   fertilization on growth and yield formation in early-season rice. Based
   on a 42-year field experiment initiated in 1981, three treatments:
   chemical N, P, and K (NPK), a double dose of chemical NPK (HNPK), and a
   combination of chemical and organic fertilizers (NPKM) were selected
   for comparison. Measurements included rice yield, yield components,
   tiller dynamic, dry matter accumulation, chlorophyll content (SPAD
   values), and transcriptome analysis of leaves at full heading stage in
   2022. Results showed that rice yield followed the order
   NPKM > HNPK > NPK. The HNPK and NPKM treatments increased by 56.64% and
   90.33%, respectively (p < 0.05). Spikelet density and 1000-grain weight
   increased by 9.16–22.89% and 3.74–4.28%, respectively. Both HNPK and
   NPKM enhanced the tillering rate by 43.84–72.73%, leading to an
   18.95–57.54% increase in effective panicles. Dry matter accumulation
   from heading to filling stages was highest in NPKM-treated rice, with
   an increase of 3.31–4.25 g/plant in dry matter transport from leaves
   and stems to spikes during the filling to maturity stages. The SPAD
   values of NPKM-treated plants were consistently higher than those of
   HNPK and NPK treatments, with a smaller decline (6.94%) from heading to
   filling stages compared to 18.23% and 26.29% for HNPK and NPK,
   respectively. Correlation analysis indicated a positive relationship
   between yield and SPAD values or dry matter accumulation and a negative
   correlation with the decline in SPAD values from full heading to
   filling stage. Transcriptome analysis revealed significant enrichment
   of photosynthesis and plant senescence pathways among treatments.
   Overall, long-term combined application of chemical and organic
   fertilizers (NPKM) improved yield components, delayed chlorophyll
   degradation, promoted dry matter accumulation and transport, and
   regulated the expression of photosynthesis-related and
   senescence-related genes, ultimately optimizing growth and yield in
   early-season rice.

   Keywords: Long-term method of fertilization, Early-season rice, Dry
   matter accumulation, SPAD, Differentially expressed genes (DEGs)

   Subject terms: Plant sciences, Environmental sciences

Introduction

   Rice (Oryza sativa) is an essential global staple crop, and ensuring
   its yield is critical for global food security. The growth and yield of
   rice are influenced by both intrinsic genetic traits and external
   environmental factors, with fertilization playing a necessary role.
   Fertilization contributed to over 40% of the increase in grain
   production^[36]1. While chemical fertilizers provide a clear short-term
   yield boost, their prolonged use without organic supplementation has
   diminished their long-term effectiveness. For example, each kilogram of
   nitrogen fertilizer increased grain production by 9.5 kg in the 1960s
   and 1970s but only 5.5 kg from the 1970s to the 1980s, similarly, the
   phosphorus fertilizer efficiency dropped from 35.5 kg to 15.5 kg during
   the same period^[37]2,[38]3. This decline is attributed to the reduced
   application of organic fertilizers, nutrient losses, and suboptimal
   usage of chemical fertilizers, which also pose ecological risks and
   hinder the sustainable development of farmlands.

   The study of long-term fertilization’s impact on crop growth and soil
   fertility is of both theoretical and practical significance. Long-term
   experiments provide unique insights into the cumulative effects of
   fertilization, surpassing conventional experiments in scope and
   relevance^[39]4–[40]6. Numerous studies have demonstrated that the
   combined application of chemical and organic fertilizers achieves the
   highest and most consistent yield increases over time^[41]7–[42]10.
   Research has revealed that long-term fertilization enhances soil
   fertility by improving soil organic matter^[43]10–[44]13, nutrient
   content^[45]4,[46]7,[47]14,[48]15, pH balance^[49]16, aggregate
   formation^[50]17,[51]18, enzyme activity^[52]19, and microbial
   community diversity and function^[53]20,[54]21. Additionally, studies
   show that integrated fertilization promotes dry matter accumulation,
   nutrient transport to grains, enzyme activity, root growth, and
   photosynthesis^[55]22–[56]28.

   Despite these advances, most studies on long-term fertilization have
   focused on soil properties, physical structure, and microbial
   diversity, with fewer investigations into crop growth dynamics and
   molecular regulation mechanisms. Addressing this gap, this study
   examines the effects of long-term fertilization on early-season rice
   (Oryza sativa) in China, where double-cropping rice systems dominate
   the Yangtze River region. Early-season rice, characterized by early
   transplantation and maturation, often suffers from low yield and
   quality, leading to limited enthusiasm among farmers. Fertilization
   plays a crucial role in enhancing its production.

   Using data from a 42-year long-term field experiment, this research
   explores the physiological and molecular mechanisms of prolonged
   fertilization on early-season rice. Specifically, it investigates leaf
   gene expression at full heading stage, dry matter accumulation, yield
   components, and SPAD values. This study aims to elucidate the
   regulatory mechanisms underlying rice growth and provide strategies for
   high-yield, efficient fertilization in early-season rice production.

Materials and methods

Field description

   The long-term field experiment was initiated in 1981 at the Red Soil
   and Germplasm Resources Research Institute in Zhanggong Town, Jinxian
   County, Nanchang City, Jiangxi Province, China (116° 20′ 24″ N, 28° 15′
   30″ E). The site is situated at an altitude of 30 m above sea level and
   experiences a mid-subtropical monsoon climate. Key climatic features
   include an average annual temperature of 17.6 °C, an effective
   cumulative temperature is 5528 °C, annual precipitation of 1785 mm,
   approximately 280 frost-free days per year, and 1950 h of annual
   sunshine. The cropping system involves double-season rice cultivation.

   The soil is classified as paddy soil derived from quaternary red clay.
   Initial soil properties prior to the experiment were as follows: pH of
   6.9, soil organic carbon (SOC) 16.3 g kg^−1, total nitrogen (TN) of
   1.49 g kg^−1, total phosphorus (TP) 0.48 g kg^−1, total potassium (TK)
   of 10.39 g kg^−1, alkali-hydrolyzed nitrogen (AN) 150.4 mg kg^−1,
   available phosphorus (AP) 4.15 mg kg^−1, and available potassium (AK)
   80.52 mg kg^−1.

Experimental design

   The experiment was conducted using a randomized complete block design
   with three replications. Each plot measured 50 m². Three treatments
   were selected from the original nine treatments:
    1. NPK: A combination of chemical nitrogen (N), phosphorus (P), and
       potassium (K).
    2. HNPK: A double dose of chemical N, P, and K.
    3. NPKM: A combination of chemical N, P, and K fertilizer, along with
       organic fertilizer.

   The fertilizer application rates for each treatment are detailed in
   Table [57]1. Urea served as the N source, applied as a base (60%) and
   tillering fertilizer (40%). Potassium chloride was used for K
   application, applied entirely as a tillering fertilizer. Calcium
   magnesium phosphate was used for P application as a base fertilizer.
   Base fertilizers were applied 1–2 days before transplanting, and
   tillering fertilizers were applied after rice began re-green.

Table 1.

   Fertilization application rates for each treatment (kg hm^−2).
   Treatment Chemical fertilizer   Organic fertilizer
             N   P[2]O[5] K[2]O    N     P[2]O[5] K[2]O
   NPK       90     45       75    0         0    0
   HNPK      180    90      150    0         0    0
   NPKM      90     45       75    123.1     9    60.75
   [58]Open in a new tab

   In the NPKM treatment, chemical fertilizers were applied in the same
   manner and quantity as the NPK treatment. Organic fertilizers were
   applied once as a base fertilizer, with winter green manure (vetch)
   used for early-season rice at a rate of 22,500 kg hm^−2. Rice cultivars
   were rotated every five years, with Lingliangyou 739 used in 2022. The
   rice was seeded on March 25 and transplanted on April 24. Manual
   transplantation followed a line-and-row design with 20 cm x 20 cm
   spacing. Alternate wetting and drying irrigation were employed
   throughout the growing season. Other agronomic practices, including
   pest and disease management, followed local high-yield cultivation
   guidelines.

Analysis of yield and yield components

   Rice was harvested at the mature stage on July 18th 2022. The yield was
   calculated based on the harvested weight and hill number per plot using
   the following formula:
   graphic file with name M1.gif 1

   Three representative plants from each plot were analyzed for yield
   components, including panicle length, grains per panicle, setting rate,
   and 1000-grain weight.

Analysis of tiller dynamics

   Ten fixed hills were marked in each plot to monitor tiller numbers
   weekly, starting seven days after transplanting until the numbers
   stabilized. The tillering rate and productive tiller percentage were
   calculated as follows:
   graphic file with name M2.gif 2
   graphic file with name M3.gif 3

Measurement of dry matter accumulation and leaf SPAD value

   Aboveground segments of rice were collected at the tillering, full
   heading, filling, and mature stages. Samples were separated into stem,
   leaf, and spike components, then baked in an oven at 105 °C for 30 min
   and subsequently dried at 85 °C for 6 h to determine dry weight. A
   chlorophyll meter (SPAD-502, Minolta Camera Co., Osaka, Japan) was used
   to measure the SPAD value (chlorophyll content) on 20 fully expanded
   top leaves per plot at each growth stage.

Leaf sample preparation and RNA-seq

   Leaf samples were collected at the full heading and filling stages,
   flash-frozen in liquid nitrogen, and stored at −80 °C. Total RNA was
   extracted using Trizol reagent (Invitrogen, Carlsbad, CA, USA) and
   purified with the RNeasy Plant Mini Kit (Qiagen, Valencia, CA, USA).
   RNA quality was assessed using the Agilent 2100 Bioanalyzer. Sequencing
   libraries were analyzed via Illumina HiSeq2000 (Illumina Inc., San
   Diego, CA, USA).

   Raw sequencing data were filtered to remove low-quality reads,
   including reads with adapters and poly-N stretches > 10%. Clean reads
   were aligned to the rice reference genome database
   (ftp://ftp.ensemblgenomes.org/pub/release42/plants/fasta/oryza_sativa/d
   na/Oryza_sativa.IRGSP-1.0.dna.toplevel.fa.gz) using HISAT (Hierarchical
   Indexing for Spliced Alignment of Transcripts)^[59]29. Expression
   levels were quantified using FPKM (Fragments Per Kilobase of transcript
   per Million fragments mapped) calculated as:

   FPKM = (Number of cDNA fragments uniquely aligned to Gene A / Total
   number of fragments uniquely aligned to all reference genes) × 10^6/
   (Transcript Length in kilobases of Gene A’s exonic regions)^[60]30.
   Differentially expressed genes (DEGs) were identified using DEseq2 with
   thresholds set at a fold change ≥ 2 and a P-value < 0.05 for
   significance.

Statistical analysis

   One-way analysis of variance (ANOVA) was performed to assess the
   effects of long-term fertilization of yield, yield components, dry
   matter accumulation, and SPAD value, with significance set at p < 0.05.
   Fisher’s least significant difference (LSD) post hoc test was used to
   determine significant differences among treatments. Graphs were
   generated using Origin 9.0 software (Microcal Software, Northampton,
   MA).

Results

Impact of long-term method of fertilization on early rice yield

   Long-term method of fertilization significantly affected early-season
   rice yield (Fig. [61]1). Both the 42-year average yield (Fig. [62]1a)
   and the 42th year’s yield (Fig. [63]1b) followed the sequence:
   NPKM > HNPK > NPK. Compared to the NPK treatment, the yields of HNPK
   and NPKM treatments increased by 56.64% and 90.33%, respectively
   (p < 0.05) after 42 years of continuous fertilization (Fig. [64]1b).

Fig. 1.

   [65]Fig. 1
   [66]Open in a new tab

   Effect of long-term fertilization on grain yield of early-season rice.
   Note: Letters upon the same color columns indicated significant
   differences among treatments at p < 0.05 level, same as below.

   Yield components were also significantly impacted by long-term method
   of fertilization (Figs. [67]2 and [68]3a). Compared to the NPK
   treatment, the number of effective panicles increased by 18.95% and
   57.54% under HNPK and NPKM, respectively (Fig. [69]3a). Spikelet
   density increased by 9.16% and 22.89% (Fig. [70]2b), while 1000-grain
   weight increased 3.74% and 4.28%, respectively (Fig. [71]2d). The
   grains per panicle were significantly higher in the NPKM treatment than
   in the chemical fertilizer treatments (NPK and HNPK) (Fig. [72]2a),
   with increases of 20.69–28.55% (Fig. [73]2c). However, the seed-setting
   rate showed a slight decline of 3.69–4.51%.

Fig. 2.

   [74]Fig. 2
   [75]Open in a new tab

   Effect of long-term fertilization on the yield components of
   early-season rice.

Fig. 3.

   [76]Fig. 3
   [77]Open in a new tab

   Effect of long-term fertilization on tillering and panicles formation
   on early-season rice.

Tiller dynamics and productive tiller percentage

   Fertilization treatments significantly influenced the tiller dynamics
   of early-season rice (Fig. [78]3b). The NPK and HNPK treatments
   exhibited a similar tillering process, where tiller numbers increased
   between 2 and 7 weeks after transplanting and subsequently decreased.
   In contrast, the NPKM treatment prolonged the tillering process, with
   tiller number peaking in the 8th week after transplanting.

   Compared to the NPK treatment, the tillering rate increased by 43.84%
   and 72.73% under HNPK and NPKM, respectively (Fig. [79]3c). The
   productive tiller percentage was significantly higher in NPK and NPKM
   treatments than in HNPK (Fig. [80]3c), there was no significant
   difference between the NPK and NPKM treatments in their productive
   tiller percentage (Fig. [81]3c), indicating that the NPKM treatment was
   more effective than both NPK and HNPK.

Dry matter accumulation

   Dry matter accumulation in early-season rice was significantly
   influenced by long-term fertilization (Fig. [82]4). Across all growth
   stages—tillering, full heading, filling, and maturity—the total dry
   matter weight followed the trend: NPKM > HNPK > NPK. Significant
   differences were observed among the three fertilization treatments at
   full heading, filling, and maturity stages, except at the tillering
   stage, where no significant difference was found between NPK and HNPK
   treatments (Fig. [83]4d).

Fig. 4.

   [84]Fig. 4
   [85]Open in a new tab

   Dry matter accumulation characteristics of early-season rice under
   long-term fertilization. Note: TS, FHS, FS, and MS represent the
   tillering stage, full heading stage, filling stage, and maturity
   stages, respectively.

   Compared to the NPK treatment, the stem dry matter accumulation of HNPK
   and NPKM increased significantly, ranging from 36.95 to 46.33% across
   all growth stages (Fig. [86]4a). Similarly, leaf dry matter
   accumulation during the tillering stage showed the sequence
   HNPK > NPKM > NPK, with HNPK and NPKM significantly outperforming NPK
   (p < 0.05) (Fig. [87]4b). No significant differences were noted in
   grain dry matter among treatments at the full heading stage; however,
   significant increases were observed during the filling and mature
   stages, following the trend NPKM > HNPK > NPK (p < 0.05, Fig. [88]4c).

   The dry matter accumulation from the tillering to full heading stage
   was significantly greater under NPKM, with increases of 90.78–107.56%
   (Table [89]2). From the full heading to filling stage, dry matter
   accumulation under NPKM and HNPK increased by 216.09% and 141.46%,
   respectively, compared to NPK (Table [90]2). No significant differences
   were observed in dry matter accumulation from the following to maturity
   stage among treatments (Table [91]2).

Table 2.

   Changes in SPAD and dry matter accumulation between adjacent growth
   stages.
   Treatment ΔS1 ΔS2 ΔS3 ΔDM1 ΔDM2 ΔDM3
   NPK 5.24 ± 0.96a 10.17 ± 0.93a 13.95 ± 1.06a 5.70 ± 0.66b 6.68 ± 1.64c
   3.42 ± 2.52a
   HNPK 5.09 ± 0.86a 7.26 ± 1.32b 13.74 ± 1.23a 6.67 ± 2.83b 16.12 ± 1.90b
   5.95 ± 2.05a
   NPKM 6.11 ± 0.82a 3.41 ± 1.65c 15.56 ± 0.34a 12.00 ± 3.47a
   21.10 ± 1.44a 5.35 ± 0.16a
   [92]Open in a new tab

   ΔS1, ΔS2, ΔS3, represent changes in SPAD values between growth stages;
   ΔDM1, ΔDM2, and ΔDM3 represent changes in dry matter accumulation
   between stages. Letters indicate significant differences among
   treatments (p < 0.05).

   Dry matter translocation from stem and leaves to spike during the
   filling to maturity stage was the highest under NPKM (5.51 g/plant),
   followed by HNPK (2.2 g/plant) and NPK (1.26 g/plant). These results
   indicate that integrating chemical and organic fertilizers (NPKM)
   significantly enhanced dry matter accumulation and improved dry matter
   translocation to spikes during grain filling and maturity.

SPAD value dynamic of leaves

   Long-term fertilization significantly influenced the chlorophyll
   content (SPAD value) of early-season rice leaves (Fig. [93]5). Compared
   to NPK, the HNPK treatment significantly increased SPAD values during
   the filling and maturity stages by 14.39% and 29.68%, respectively
   (p < 0.05). The NPKM treatment consistently produced significantly
   higher SPAD values across all growth stages, with increases of
   22.05–25.58% at the tillering stage, 38.80–58.78% at the full heading
   stage, and 57.47–104.21% at the filling and maturity stages (p < 0.05).

Fig. 5.

   [94]Fig. 5
   [95]Open in a new tab

   Chlorophyll SPAD dynamics of early-season rice leaves under long-term
   fertilization.

   From the tillering to full heading stage, SPAD values increased 5.24
   (NPK), 5.09 (HNPK), and 6.11 (NPKM) (Table [96]2). From the full
   heading to filling stage, SPAD values declined by 6.94% (NPKM), 18.23%
   (HNPK), and 26.29% (NPK). Between the filling and maturity stages, SPAD
   values decreased by 13.96 (NPK), 13.74 (HNPK), and 15.56 (NPKM).

   These findings suggest that the NPKM treatment significantly enhanced
   SPAD value and mitigated its decline from the full heading to the
   filling stage.

Relationship between yield, SPAD, and dry matter accumulation

   Correlation analysis revealed significant relationships between yield,
   SPAD values, and dry matter accumulation (Fig. [97]6). Yield was
   positively correlated with dry matter accumulation from tillering to
   full heading stage (ΔDM1). Conversely, yield was negatively correlated
   with the decline in SPAD values from full heading to filling stage
   (ΔS2).

Fig. 6.

   [98]Fig. 6
   [99]Open in a new tab

   Correlation analysis of yield, SPAD, and dry matter accumulation in
   early-season rice under long-term fertilization. Note: Letters S1, S2,
   S3, S4, DM1, DM2, DM3 and DM4 represent the SPAD and total dry matter
   of early season rice at tillering stage, fullheading stage, filling
   stage and mature sage, respectively. Symbols *and ** in the circle
   indicate significant correlations at the p < 0.05 and p < 0.01 levels.

DEGs in rice leaves at full heading stage induced by long-term fertilization

   Long-term fertilization significantly influenced leaf gene expression
   in early-season rice leaves at the full heading stage (Fig. [100]7). A
   total of 3,585 DEGs were identified between NPKM and NPK treatments,
   with 1,478 genes upregulated and 2,107 genes downregulated.
   Comparatively, 2,051 DEGs were identified between NPKM and HNPK
   treatments, of which 1,251 were upregulated and 800 were
   down-regulated, while 3,518 DEGs were detected between HNPK and NPK
   treatments, including 1,213 upregulated and 2,305 downregulated genes.
   Additionally, 320 DEGs overlapped across the three treatment
   comparisons, with 903, 882, and 413 unique DEGs identified in NPK
   versus NPKM, NPK versus HNPK, and HNPK versus NPKM, respectively.

Fig. 7.

   [101]Fig. 7
   [102]Open in a new tab

   The DEGS in early-season rice leaves at the full heading stage under
   long-term fertilization.

Senescence- and photosynthesis-related DEGs

   Gene Ontology (GO) enrichment analysis revealed that 20 DEGs were
   associated with rice senescence (Table [103]3). Several DEGs (e.g.,
   Os03g02280, Os05g45450, Os04g43990, Os01g52730, Os02g41840, Os10g27350,
   Os10g33990, and Os04g33760) were primarily involved in encoding
   proteins containing DUF584 domains, with others (e.g., Os03g54130,
   Os04g13140, and Os01g42780) encoded senescence-specific cysteine
   protease and precursors. Additional DEGs were involved in coenzymes and
   stress response processes.

Table 3.

   DEGs associated with senescence in rice leaves at the full heading
   stage.
   Gene ID FPKM DEGS Gene Function
   NPK HNPK NPKM I II III
   Os06g50330 20.324 30.823 38.798 U N N Protein EARLY-RESPONSIVE TO
   DEHYDRATION 7, chloroplastic
   Os01g52740 3.463 1.328 1.025 D D N Response to endogenous stimulus
   Os03g02280 2.192 0.186 0.681 D N N DUF584 domain containing protein,
   putative, expressed
   Os05g45450 0.409 1.277 0.597 N U N DUF584 domain containing protein,
   putative, expressed
   Os04g43990 18.011 34.5230 30.241 N U N DUF584 domain containing
   protein, putative, expressed
   Os01g52730 5.196 2.715 5.436 N N U DUF584 domain containing protein,
   putative, expressed
   Os02g41840 2.031 2.818 10.769 U N U DUF584 domain containing protein,
   putative, expressed
   Os10g27350 3.529 0.793 1.405 D D D DUF584 domain containing protein,
   putative, expressed
   Os10g33990 7.975 7.638 24.841 U N U DUF584 domain containing protein,
   putative, expressed
   Os04g33760 3.445 4.248 6.476 U N N DUF584 domain containing protein,
   putative, expressed
   Os03g54130 10.688 2.385 1.382 D D N Senescence-specific cysteine
   protease SAG39
   Os04g13140 8.189 0.274 1.442 D D U Senescence-specific cysteine
   protease SAG39
   Os01g42780 79.563 56.836 1.776 D N D Cysteine proteinase EP-B 2
   precursor
   Os05g27580 0.427 1.725 2.070 U U N Wound-induced protein WI12
   Os07g49114 10.991 5.614 2.984 D N N Wound-induced protein WI12
   Os09g01000 221.300 817.118 314.060 N U D Putative senescence-associated
   protein, partial
   Os10g07010 1.051 0.251 0.115 D D N 3-ketoacyl-CoA synthase
   Os01g21250 302.410 25.303 34.678 D D N Late embryogenesis abundant
   protein
   Os11g25160 13.029 33.960 36.780 U U N Tropinone reductase homolog
   At5g06060 isoform X1
   Os01g59570 0.318 0.307 0.101 D N D Probable LRR receptor-like
   serine/threonine-protein kinase At1g51810
   [104]Open in a new tab

   The DEGs I, II, and III refer to comparisons NPK versus NPKM, NPK
   versus HNPK, and HNPK versus NPKM, respectively.

   D, downregulated,;U, upregulated,;N +, non-regulated.

   KEGG pathway enrichment analysis indicated that photosynthesis-related
   DEGs were primarily involved in Photosystem I, Photosystem II,
   cytochrome, ferredoxin, and other related proteins (Fig. [105]8).

Fig. 8.

   [106]Fig. 8
   [107]Open in a new tab

   KEGG pathway enrichment of photosynthesis-related DEGs in rice leaves
   at the full heading stage. Note: (a) NPK versus HNPK, (b) NPK versus
   NPKM. The red, green, blue and none indicate upregulated,
   downregulated, upregulated and downregulated, non-regulated DEGs,
   respectively.

Discussion

Impact of long-term fertilization on early rice yield

   Fertilization is a critical agronomic practice for maintaining
   productivity and ensuring stable rice yields. Shen et al.^[108]31 and
   other
   studies^[109]4,[110]7,[111]8,[112]14,[113]15,[114]22,[115]24,[116]27,[1
   17]32,[118]33 have demonstrated that combining organic and chemical
   fertilizers significantly enhances crop yields, consistent with this
   study’s findings where early-season rice yields followed the trend of
   NPKM > HNPK > NPK.

   The mechanisms underpinning these effects include the rapid nutrient
   provision from chemical fertilizers and the enhancement of soil
   physical and chemical properties by organic
   fertilizers^[119]32,[120]34,[121]35. Long-term chemical fertilization
   alone has limited impact on soil fertility improvement^[122]31, whereas
   organic fertilizers enhance soil enzyme activity, microbial diversity,
   and productivity^[123]19–[124]21. However, exclusive use of organic
   fertilizers is insufficient for meeting crops’ nutrient
   needs^[125]15,[126]36. Thus, combining chemical and organic fertilizers
   ensures immediate nutrient supply while improving soil fertility over
   time, leading to sustained high yields^[127]32,[128]37.

   This study also found that HNPK significantly increased yields compared
   to NPK, suggesting that higher chemical fertilizer rates can enhance
   nutrient availability and yield potential. This contrasts with previous
   studies indicating that standard chemical fertilization suffices for
   normal growth^[129]15,[130]38. This discrepancy may arise from
   initially low fertilization levels in this experiment, suggesting that
   modern rice varieties with higher nutrient demands require adjusted
   fertilization strategies.

   Moreover, combining chemical and organic fertilizers significantly
   improved yield components, such as effective panicles, spikelet
   density, and grain weight, Amanullah et al.^[131]39 observed similar
   findings, emphasizing that long-term integration of chemical and
   organic fertilization enhances the storage capacity of rice to support
   increased dry matter production and improved yield outcomes.

   Dry matter accumulation forms the material foundation for crop yield,
   with its accumulation, transfer, and distribution significantly
   influencing crop yields^[132]40. Most dry matter originates from
   photosynthesis, primarily occurring in chloroplasts, and is determined
   by both the accumulation rate and duration^[133]41–[134]43. Chlorophyll
   content, closely associated with photosynthetic productivity, is
   positively correlated with the leaf photosynthetic rate within a
   certain range^[135]44–[136]46. Previous research has established a
   strong positive correlation between SPAD values and chlorophyll
   content, making SPAD an effective tool for estimating relative
   chlorophyll content^[137]47,[138]48.

   This study found that the combined use of mineral and organic
   fertilizers significantly increased SPAD values during the rice growth
   period. This indicates that long-term application of chemical
   fertilizer with organic fertilizers enhanced photosynthetic intensity,
   contributing to a stronger photosynthetic “source” in rice. Correlation
   analysis demonstrated a significant positive relationship between
   chlorophyll content, dry matter accumulation, and yield in rice.
   Compared with chemical fertilizers alone, combining chemical and
   organic fertilizers (NPKM treatment) significantly increased
   chlorophyll content during the full heading, filling, and maturity
   stages, thereby enhancing photosynthesis, promoting dry matter
   accumulation, and improving yield formation.

   Additionally, the NPKM treatment significantly increased the transfer
   of dry matter from stems and leaves to panicles during the filling to
   maturity stages. This finding suggests that combined fertilization
   strategies enhance the flow of assimilates, leading to improved yield
   performance in early-season rice.

Impact of long-term fertilization on growth and gene expression in early rice

   Compared with chemical fertilization treatments (NPK and HNPK), the
   NPKM treatment prolonged the tillering process and increased the number
   of effective panicles. This can be attributed to the immediate
   availability of nutrients from chemical fertilizers during the early
   growth stages and the sustained nutrient supply from organic
   fertilizers in later stages, reducing nutrient loss and supporting
   prolonged growth^[139]49–[140]51.

   Crop growth and development are influenced by genotype and
   environmental factors, with genotype exerting a decisive role. However,
   environmental factors, such as fertilization, significantly affect gene
   expression, resulting in phenotypic variations. This study identified
   that long-term fertilization significantly influenced gene expression
   in early rice leaves at the full heading stage. Consistent with
   findings from Liu et al.^[141]52,[142]53, GO and KEEG functional
   analyses of DEGs revealed significant enrichment of
   photosynthesis-related genes, particularly those involved in
   Photosystem I, Photosystem II, cytochromes, and ferredoxin. These
   findings align with the observed differences in SPAD values and dry
   matter accumulation among treatments, confirming that long-term
   fertilization regulates photosynthetic gene expression, influencing
   chlorophyll synthesis, dry matter accumulation, and, ultimately, yield.

   The combined application of chemical and organic fertilizers also
   delayed chlorophyll degradation (as indicated by SPAD values) during
   the later growth stages. This aligns with Xu et al.^[143]54, who
   demonstrated that organic fertilizers decompose gradually, replenishing
   nutrients and enhancing post-heading nutrient assimilation. Correlation
   analysis further revealed a significant negative relationship between
   the decrease in chlorophyll content from the heading to filling stages
   and yield, indicating that combined fertilization delayed leaf
   senescence, extended photosynthetic activity, and promoted dry matter
   accumulation and yield improvement. Zhang et al.^[144]55 reported
   similar results, attributing them to increased protective enzyme
   activity and reduced secondary metabolites such as malondialdehyde
   (MDA) and free proline (Pro) under combined
   fertilization^[145]25,[146]56. Transcriptome analysis in this study
   showed that genes expressed in leaves at the full heading stage were
   significantly enriched in senescence-related pathways, involving
   proteins containing the DUF584 domain, aging-related cysteine
   proteases, and associated cofactors. Proteins with the DUF584 domain
   play essential roles in regulating rice aging by modulating the
   synthesis and degradation of abscisic acid^[147]57. Cysteine further
   contributes to glutathione (GSH) formation and scavenging superoxide
   radicals, revealing molecular mechanisms by which combined
   fertilization reduces MDA and Pro levels, delays aging, enhances dry
   matter accumulation, and supports yield formation.

Future directions

   Late-season rice typically outperforms early-season rice in yield, with
   distinct growth environments and responses to fertilization. To develop
   comprehensive fertilization strategies, future studies should
   investigate the effects of long-term fertilization on late-season rice
   growth, yield, and regulatory mechanisms.

Conclusion

   Long-term method of fertilization significantly influenced the gene
   expression of early rice leaves at the full heading stage, with DEGs
   enriched in metabolic pathways associated with photosynthesis and
   senescence. These genes were primarily involved in Photosystem I,
   Photosystem II, cytochromes, ferredoxin, DUF584 domain-containing
   proteins, cysteine proteases linked to aging, and the encoding and
   synthesis of their related precursors and cofactors.

   The application of chemical fertilizer in combination with organic
   fertilizers significantly enhanced tillering and promoted dry matter
   accumulation in early-season rice, particularly from the full heading
   to the filling stage. This strategy also strengthened the transport of
   dry matter from stems and leaves to panicles during the filling to
   maturity stages. Additionally, combined fertilization significantly
   increased chlorophyll content in sword leaves and effectively mitigated
   chlorophyll degradation from the full heading to the filling stage,
   thereby prolonging photosynthetic activity and contributing to higher
   yields.

Author contributions

   Conceptualization, Zhihua Hu; methodology, Zhihua Hu, Kailou Liu;
   writing—original draft preparation, Zhihua Hu, Kailou Liu, Xiaolin Xu;
   writing—review and editing, Dandan Hu, Huijie Song, Yan Wu, Jianfu Wu;
   funding acquisition, Kailou Liu, Jianfu Wu. All authors have read and
   agreed to the published version of the manuscript.

Funding

   This research was supported by the National Natural Science Foundation
   of China (32260808; 31660596) and the key research and development plan
   “open ranking” project of Jiangxi Province, China (20223BBF61016).

Data availability

   The original datasets generated during and/or analyzed during the
   current study are available from the corresponding author on reasonable
   request.

Declarations

Competing interests

   The authors declare no competing interests.

Footnotes

   Publisher’s note

   Springer Nature remains neutral with regard to jurisdictional claims in
   published maps and institutional affiliations.

   Zhihua Hu and Kailou Liu contributed equally to this work.

References