Effect of the pectin contents and nanostructure on the stem straightness of two Paeonia lactiflora cultivars

Plant materials and growth conditions

Two cultivars of P. lactiflora, ‘Dong Fang Shao Nv’ and ‘Lan Tian Piao Xian’, were used in this study. The cultivars were both eight-year-old plants obtained from the Caozhou Peony Garden, Heze, China (35.14° N, 115.26° E) with normal and identical fertilization and water. The parent materials were spaced approximately 82 cm apart and 106 cm apart in rows. The plants were fertilized once a year in October or November and once every two months with a compound fertilizer for foliar application, combined with insect removal. Daily maintenance was mainly for diligent weeding and loosening the soil for air. Six plants were randomly selected for measurement at each stage. From stem elongation to flowering, every seven days was one period and five stages were divided as described by Wan et al. (2020). The growth conditions of ‘Dong Fang Shao Nv’ and ‘Lan Tian Piao Xiang’ at five stages are shown in Fig. 1A, stage S1 to S5 corresponding to the stem elongation, bud conception, color transmission, bud softness, and flowering stages of herbaceous peony, respectively. Samples from S1 to S5 were obtained by cutting 5–10 cm stem segments under the flower. All the samples were quickly frozen in liquid nitrogen and stored in a −80 °C freezer for subsequent use. At each stage, to measure the degree of stem bending, the angle of the flower and branch was measured. First, the flower was connected with the base of the stem, and second, a plumb line was extended from the base of the stem. These two lines formed a pinch angle, which was called the angle of the flower and branch. This definition was based on a previous study (Wan et al., 2020), and the schematic diagram is shown in Fig. 1B.

Determination of pectin fractions in the cell wall

The cell wall material (CWM) was extracted first as described in a previous study (Wan et al., 2020). Three hundred milligrams of stem segments were weighed and added to 1 mL of 80% ethanol solution and homogenized rapidly at room temperature. Then, the samples were placed in a water bath at 90 °C for 20 min. After cooling to room temperature, the mixtures were centrifuged at 6000 × g for 10 min at 25 °C, and the supernatant was discarded. The precipitate was washed with 1.5 mL of 80% ethanol and propanol successively, vortexed for approximately 2 min followed by centrifugation at 6000 × g 25 °C for 10 min, and the supernatant was discarded. Subsequently, 1 mL of 90% dimethyl sulfoxide solution was added to remove the starch for 15 h, and the centrifugation step was repeated. The final precipitate was dried and weighed as CWM. The CWM was weighed to 3 mg, and the carbazole colorimetric method was used to measure the pectin content of the three solubles. The reagent kit was purchased from Suzhou Kemin Biotechnology Corporation, and the manufacturer’s instructions were followed. All measurements were performed with three biological replicates.

The different kinds of pectin fractions were extracted from stems according to the method described by Fishman et al. (1993) and Wang et al. (2021) with appropriate modifications. The CWM in the pectin fraction determination step was alcohol insoluble residue (AIR) (Renard Catherine, 2005). A 0.05 g AIR sample was mixed with distilled water, shaken for 1 h at 25 °C, repeated three times, and then centrifuged at 4 °C and 10,000 × g for 15 min. The supernatant was obtained as the WSP extract and diluted 6 times for the measurement. The residue was transferred to 15 mL of cyclohexane-trans-1,2-diamine tetraacetic acid (CDTA, 0.05 M, pH 6.5 adjusted by sodium acetate) solution. The sample was oscillated for 6 h at 25 °C, centrifuged for 15 min (10,000 g, 4 °C), filtered and diluted 6 times to use as CSP. For SSP fractions, the remaining residue was incubated with 15 mL of 0.05 M Na₂CO₃ (containing 0.002 M CDTA) for 16 h at 4 °C and stirred for 6 h at 25 °C. After centrifugation and filtration, the resulting supernatant was diluted 6 times and left to be measured.

Observation of the pectin nanostructure by atomic force microscopy (AFM)

A total of 10 mL of diluted sample solutions was dropped onto freshly cleaved mica and then left at room temperature for approximately 15 min until the sample dried. The nanostructure of the pectin fractions was scanned by an atomic force microscope (BRUKER MULTIMODE 8, USA), and the probe was for Bruker’s intelligent scanning, model SCANASYST-AIR. The scanning resolution was 512 × 512 points. At least three images were observed at the scale of 10 by 10 µm² under AFM. To show the nanoscale structure of pectin more clearly, the image size shown in this manuscript is 2.5 × 2.5 µm² or 5 × 5 µm².

AFM image analysis

The images were processed using offline software (NanoScope Analysis version 1.40) for qualitative and quantitative information about pectin nanostructures as described by Paniagua et al. (2017a) and Pan et al. (2018). The “Flatten” function in the software corrected for height deviations due to drift in the Z-voltage of the scanning tube or the tilt of the sample itself, as well as the bow of the scanning tube. Since the observed pectin contained multiple morphologies, they were characterized by measuring the agglomerate width and the chain structure length. Using the ‘Section Analysis’ function in the NanoScope Analysis offline software, we measured the width and length of the chain by drawing a cross-cutting line vertically on the image or along the direction of the chain. The length of a single chain was defined as the total length of the trunk and branches (Adams et al., 2003). The data were plotted as violin plots based on the specific values of the measured molecular lengths and widths. To ensure reliable experimental results, at least 90 chains were measured at each developmental stage.

Statistical analysis

The effects of cultivars and stages and their interaction on pectin content with two-way ANOVA and the statistical significance of the differences with Tukey’s test were analyzed in SPSS 25.0.0.0. Differences in indicators were tested by t-tests, and correlations were analyzed by Spearman correlation. Violin plots were generated in R 4.2.1 (R Core Team, 2022).

Changes in angle between flower and branch

One of the most intuitive indicators to reflect stem straightness is the angle of the flower and branch. As shown in Fig. 1C, the angle of the flower and branch of ‘Dong Fang Shao Nv’ varied from 3.38° to 6.12° at five stages, and the stem remained upright. There was no significant difference in the angle of the flower and branch from S2 to S5 except for S1, when the angle of the flower and branch was significantly smaller. The angle of the flower and branch of ‘Lan Tian Piao Xiang’ gradually increased, ranging from 6.13° to 64.05° and were significantly different during the five stages. Bent stems were observed from S4 to S5. At all the stages, the angles of ‘Lan Tian Piao Xiang’ were significantly greater than those of ‘Dong Fang Shao Nv’, and the difference reached a maximum at S5. Thus, ‘Dong Fang Shao Nv’ and ‘Lan Tian Piao Xiang’ can be used as representatives of upright and bending herbaceous peonies, respectively.

Quantitative analysis of cell wall pectin contents

The contents of WSP, CSP, and SSP in the two cultivars are shown in Fig. 2. The WSP content of ‘Lan Tian Piao Xiang’ was significantly greater than that of ‘Dong Fang Shao Nv’ from S2 to S5, and the content of ‘Lan Tian Piao Xiang’ was 2.93 times greater than that of ‘Dong Fang Shao Nv’ at S2 when the difference between the two cultivars was the largest. The WSP of the two cultivars showed different trends with development, where the content of ‘Dong Fang Shao Nv’ at S1 was significantly greater than that at the other stages, while the content of ‘Lan Tian Piao Xiang’ increased first at the early stage and peaked at S2. ‘Dong Fang Shao Nv’ had a consistently higher CSP content than ‘Lan Tian Piao Xiang’ during stem development, and the largest difference (i.e., 7.36 times) was shown at S1. ‘Dong Fang Shao Nv’ had the highest CSP content at S1 and subsequently decreased, while the CSP content of ‘Lan Tian Piao Xiang’ was always at a low level. Moreover, the SSP content of ‘Dong Fang Shao Nv’ was significantly higher than that of ‘Lan Tian Piao Xiang’ from S1 to S4, and the largest difference occurred at S1, with a difference of 2.71 times. The SSP content of ‘Dong Fang Shao Nv’ changed with development, and the SSP content of ‘Lan Tian Piao Xiang’ remained high only at S1 and S2. The two-way ANOVA test, shown in Table 1, demonstrated that both the cultivar and development stage had significant effects on the pectin content of the three solubilities.

Correlation analysis of pectin content and angle between flower and branch

Correlation analysis between the three pectin contents of the two cultivars and the angles of flower and branches at different periods was conducted to analyze how the cell wall pectin contents affected the straightness of herbaceous peony stems and the results are shown in Table 2. The content of WSP was highly significant and positively correlated with the angle of flower and branch at S1 (P < 0.01) and positively correlated at S2 (P < 0.05), indicating that cultivars with relatively higher WSP contents had a greater angle between flower and branch and were more likely to bend. Contents of CSP were highly significant and negatively correlated with the angle of the flower and branch at S1 and S2 (P < 0.01). It was also demonstrated that the CSP content was greater in the upright cultivar at S1 and S2. The SSP content was negatively correlated with the angle of the flower and branch at all stages but was not significant (P > 0.05).

AFM analysis of pectin fractions

Representative topographical AFM images of WSP from ‘Dong Fang Shao Nv’ and ‘Lan Tian Piao Xiang’ at the five stages are shown in Fig. 3A. ‘Dong Fang Shao Nv’ showed an elliptic polymer shape, which tended to aggregate into chains at S2. Thin lamellar structures appeared at S2–S3 and were distributed around the agglomerates. ‘Lan Tian Piao Xiang’ showed multiple shapes at the five stages. The linear chains appeared at S2 and S4, and the chain became longer at S4, whereas the other stages retained oval aggregates.

CSP nanostructures of the two cultivars are presented in Fig. 3B. Long and branched chains were observed at S1 in ‘Dong Fang Shao Nv’, and intertwined and tightly aggregated chain structures occurred at S2. At S3, densely arranged linear chains appeared near the lamellar structure, which disappeared and was replaced with clear branched and intertwined chains at S4 and S5. In ‘Lan Tian Piao Xiang’, an equal size dot-like agglomeration structure with homogeneous distribution was shown. Then, chains were observed at S2 and S3. The branched structure occurred at S4 and changed at S5 into an entangled complex branched-chain structure, which was similar to that of ‘Dong Fang Shao Nv’ at S4.

The nanostructure of SSP in the two cultivars is shown in Fig. 3C. Short linear main chains with some agglomerates dispersed around were observed in both cultivars. ‘Dong Fang Shao Nv’ showed branched structures at S2, S4, and S5, and the chains were stacked at S4. In ‘Lan Tian Piao Xiang’, only a side chain structure from S1 to S3 was observed. The remaining stages were dominated by short chains that were widely dispersed.

Further analysis of the width of WSP of the two cultivars was conducted by means of violin plots, and the results are shown in Fig. 4. At S1, most of the WSP widths were concentrated in the range from 12 to 56 nm in ‘Dong Fang Shao Nv’ (98%), while those of ‘Lan Tian Piao Xiang’ were concentrated in the range of 30–70 nm (83%). Therefore, the median WSP nanowidth of ‘Lan Tian Piao Xiang’ was significantly larger than that of ‘Dong Fang Shao Nv’. However, the result was different at S2–S3, when the median WSP width of ‘Dong Fang Shao Nv’ was larger than that of ‘Lan Tian Piao Xiang’. ‘Lan Tian Piao Xiang’ had more WSP agglomerates at S2, resulting in a higher number of outliers on the violin plot. The violin plot of ‘Lan Tian Piao Xiang’ at S4 in Fig. 4 was narrower and longer than that of ‘Dong Fang Shao Nv’, which suggested that ‘Lan Tian Piao Xiang’ had a broader WSP than ‘Dong Fang Shao Nv’ and had a more varied WSP width range (from 22 to 142 nm). In addition, ‘Lan Tian Piao Xiang’ showed more large outliers at S5; however, most of the WSP widths were still smaller than those of ‘Dong Fang Shao Nv’.

The violin plot of the width of the contoured backbone of two cultivars was used to characterize the individual molecules of CSP samples (Fig. 5). At the first three stages of development, the CSP width distributions of the two cultivars were distinctly different. In comparison to ‘Dong Fang Shao Nv’, at S1, the violin plot of ‘Lan Tian Piao Xiang’ was more compact, suggesting a more homogeneous CSP in width, which was reflected as the agglomeration structure in Fig. 3B. Furthermore, the CSP of ‘Lan Tian Piao Xiang’ was wider than that of ‘Dong Fang Shao Nv’ at S2–S3. The CSP width of ‘Lan Tian Piao Xiang’ was smaller than that of ‘Dong Fang Shao Nv’ at both S4 and S5 when stem bending was evident, as reflected in the reduction in the median in the violin plot.

The violin plot of the frequency distribution of the CSP length measurement is shown in Fig. 6. Because the pectin structure of ‘Dong Fang Shao Nv’ developed long and branched chains while ‘Lan Tian Piao Xiang’ still showed agglomeration, the shapes of the two cultivars of violin diagrams at S1 showed great differences. At S2–S4, although most of the CSP lengths of the two cultivars were similar, as reflected by the similar median values, the number of long chains was significantly higher in ‘Lan Tian Piao Xiang’ because of the larger outliers. There were more short chains and a wider variation of chain lengths in ‘Lan Tian Piao Xiang’ according to the violin graphs at S5.

Contour widths of isolated polymer chains of SSP are measured at S1–S5 in ‘Dong Fang Shao Nv’ and ‘Lan Tian Piao Xiang’ and presented as violin plots in Fig. 7. The chain width distribution of both cultivars did not show much difference at S1–S3, except for the outliers with large values in ‘Dong Fang Shao Nv’ at S1 and S3. The violin plot distribution of the widths of the two cultivars at S4 and S5 indicated that ‘Lan Tian Piao Xiang’ had wider CSP chains.

The violin plot in Fig. 8 shows the distribution of SSP lengths for the two cultivars. The shape of the violin plot was very similar for both cultivars at all stages. Except for S3, the median SSP chain length of ‘Lan Tian Piao Xiang’ was smaller than that of ‘Dong Fang Shao Nv’ at all other stages. Moreover, at S2–S4, ‘Lan Tian Piao Xiang’ had greater SSP long-chain outlier values.