Control 84K (CK) and downregulated C3H transgenic poplar 84K (Populus alba × Populus glandulosa, 4 years) were cultivated at the Chinese Academy of Forestry Sciences, and the detail regarding the procurement of wood was described in the ESI section. Especially, the gene-specific fragments were constructed into the RNAi plant expression vector by the double-digestion technique to obtain the RNAi expression vector of the C3H gene (Supplementary Figure S1). The Populus alba × Populus glandulosa clone 84K was used as transgenic poplar receptor material. Then, the resistant plants were obtained by the Agrobacterium tumefaciens-mediated transformation of leaf disc. That includes Agrobacterium culture, infection, coculture, selective culture, screening medium, and rooting culture (Supplementary Figure S2). The transplanted greenhouse was identified by PCR after the NPT-II gene and the target gene fragment, and the greenhouse was cut and propagated at low temperature and then the transgenic plants were obtained. The poplars were debarked and smashed into small pieces, then sieved to obtain 40–60 mesh particles. The composition of CK and C3H transgenics was determined by the National Renewable Energy Laboratory (NREL) standard analytical procedure (Sluiter et al., 2008). All the chemicals used in the experiment were of analytical grade.

The determination of the Klason lignin content of CK and C3H poplar particles was based on the NREL standard analytical procedure (Sluiter et al., 2008). In detail, 0.3 g poplar sample was added to 72% H₂SO₄ to hydrolyze at 30°C for 1 h. Then, the solution was added with 84 ml deionized water for further hydrolysis at 121°C for 1 h. After the hydrolysis, the solution was filtrated using a G3 sand core funnel. The mass change before and after filtration was the weight of the Klason lignin.

Firstly, the control and C3H-downregulated poplar were delignified with acidic sodium chlorite solution (adjusted by acetic acid, pH 3.8–4.0) at 75°C for 2 h according to an earlier described procedure (Bian et al., 2012). As shown in Supplementary Figure S3, the delignified material (holocellulose) was extracted with 10% potassium hydroxide for 10 h (1:20 g ml⁻¹) at room temperature. The liquid fractions were collected and adjusted to pH 5.5–6.0 with acetic acid. Then, the neutral solution was concentrated and precipitated in ethanol (three equivalent volumes). After filtration, the precipitates were redissolved in distilled water and dialyzed against water. After the freeze-dried process, the KOH-extracted hemicelluloses were obtained (named H_-CK and H_-C3H, respectively).

To delineate the structural characteristics of the native lignin in the raw material, double enzymatic lignins (DELs) from CK and C3H were prepared. The detailed preparation process was according to our previous publications (Chen et al., 2017a; Ma et al., 2020). As shown in Supplementary Figure S4, the ball-milled powder (5 g) was mixed with the desired amount of sodium acetate buffer (pH 4.8) with a solid-to-liquid ratio of 1:20 (g/ml) and cellulase (35 FPU/g substrate). Then, the mixture was incubated at 50°C in a rotary shaker with a rotational velocity of 150 rpm for 48 h. Next, the mixture was centrifuged and the residue was washed thoroughly with sodium acetate buffer (pH 4.8) to remove the hydrolyzed carbohydrates and then freeze-dried. Finally, the dried residual solid was repeatedly subjected to ball-milling for 2 h and enzymatic hydrolysis again, as in the abovementioned processes. After washing with acidic water (pH 2.0) and freeze-drying, DEL samples were obtained (named DEL_-CK and DEL_-C3H). To increase the solubility of lignin in tetrahydrofuran (THF) for the determination of molecular weights by GPC technique, the acetylation of lignin was performed as previously described (Wang et al., 2017). All experiments in this study were conducted in duplicate, and the results reported were the average values.

Sugar analysis (neutral sugars and uronic acids) was conducted by high-performance anion-exchange chromatography (HPAEC) in a Dionex ICS-3000 system (Dionex Corporation, Sunnyvale, CA, USA) equipped with a CarboPac PA1 (4 × 250 mm) column. The weight-average (M_w) and number-average (M_n) molecular weights of the samples were determined by gel permeation chromatography (GPC) (Agilent 1200, Agilent Technologies, Santa Clara, CA, USA) with an ultraviolet (UV) detector at 240 nm. The column used was a PL-gel 10 mm mixed-B 7.5 mm i.d. column, which was calibrated with PL polystyrene standards according to a previous report (Chen et al., 2017b). The NMR spectra of the samples were recorded at 25°C in DMSO-d ₆ (or D₂O) on a Bruker AVIII 400 MHz spectrometer according to published procedures (Wen et al., 2013b; Wen et al., 2014). In detail, about 25 mg of lignin and hemicelluloses was dissolved in 0.5 ml of DMSO-d ₆ and D₂O, respectively. For quantitative 2D-HSQC spectra, the Bruker standard pulse program hsqcetgpsi was used for HSQC experiments. The spectral widths were 5,000 Hz and 20,000 Hz for the ¹H- and ¹³C-dimensions, respectively. The number of collected complex points was 1,024 for ¹H-dimension with a recycle delay of 1.5 s. The number of transients was 64, and 256-time increments were always recorded in the ¹³C-dimension. The ¹ J _CH used was 145 Hz. Prior to Fourier transformation, the data matrixes were zero filled up to 1,024 points in the ¹³C-dimension. Data processing was performed using standard Bruker Topspin-NMR software.

The inhibitory intensity of the gene is determined by detecting the expression level of the target gene at the transgenic plant. The expression of the transgene gene seriously affects the analysis of the subsequent result. Therefore, it is very important to detect the expression level of the transgenic gene. There are many ways to identify plant transgene expression at the transcriptional level, and a method for determining its specific mRNA is usually used. As shown in Supplementary Figure S5, the transcription level of the C3H gene in transgenic poplar was decreased in comparison with wild-type poplar (0.27). The expression of transgenic poplar was significantly reduced, and the average reduction was about 80%, indicating that the expression of the C3H gene was inhibited in transgenic CK.

According to Figure 1B, the plant height of wild-type plants (CK) was between 83 and 104 cm, the plant height of transgenic plants (C3H) was between 39 and 82 cm, and the average plant height was 61.93 cm. It was found that the height of transgenic poplars was lower than that of the CK, and the stems were browned to some extent (Figures 1A–C). Simultaneously, the results showed that the transgenic poplar woods had a significant decrease at different growth stages as compared to CK, and the height of C3H poplar decreased by about 34% under the same growing period. When the lignin synthetase was inhibited, it was bound to affect the growth of plant height, because lignin played a certain role in mechanical support. Hu and coauthors found that 4CL was a key gene lying upstream of C3H to adjust the lignin content (Hu et al., 1999). The 4CL suppression resulted in as much as a 45% reduction in total cell wall lignin and reportedly no impairment in growth (Hu et al., 1999). On the contrary, reductions in C3H resulted in varying effects on growth properties (Coleman et al., 2008). In this study, it was found that C3H downregulation could lead to the impairment in growth, which resulted in the relatively short plant height of C3H down-regulated poplar as compared to that of control (CK).

As shown in Table 1, regarding the chemical composition of CK and C3H, it was found that the content of Klason lignin in CK poplar wood was 25.94%, while the content of lignin was slightly decreased to 20.52% in C3H, which was slightly inconsistent with the previous reports (Pu et al., 2009; Ralph et al., 2012; Ma et al., 2021a). The reason for that was that C3H downregulation will inhibit the synthesis of lignin (Peng et al., 2016). By contrast, the contents of hemicelluloses and cellulose in the C3H-downregulated poplar woods slightly increased as compared to those of wild poplar (CK). In short, compositional analysis indicated that downregulation of C3H resulted in the slight changes of chemical composition of poplar wood, such as the decrease in lignin content.

In fact, C3H downregulation mainly affects the biosynthesis of the lignin macromolecule. Herein, the effects of C3H downregulation on the molecular weights and structural characteristics of native lignin were investigated and discussed in detail. The M_w and M_n and polydispersity index (M_w/M_n) of DEL_-CK and DEL_-C3H are displayed in Table 3. The M_w of DEL_-CK and DEL_-C3H was 8,020 and 7,410 g/mol, respectively. The higher M_w of lignin was partly related to the relatively high β-O-4 content, as observed previously (Wen et al., 2013). DEL_-C3H had relatively lower molecular weight distributions (1.81) as compared to DEL_-CK (1.90), implying that the downregulation of C3H facilitates the formation of relatively homogeneous lignin fractions. This phenomenon was similar to that of hemicellulose, implying that downregulation of C3H led to the homogenization of biomacromolecules.

To demonstrate the structural differences of native lignin between control 84K poplar (CK) and C3H-downregulated poplar (C3H) samples, the lignin samples were analyzed by the 2D HSQC NMR technique. These differences could provide some fundamental basis for obtaining ideal lignin sources for subsequent lignin valorization (Wen et al., 2013a). Figure 4 shows the chemical composition (aromatic region) and interunit linkages (side-chain region) in the 2D-HSQC spectra of DEL_-CK and DEL_-C3H according to the previous signal assignments (Wen et al., 2015). In the side-chain regions (_δC/_δH 49–92/2.5–5.7) of the 2D-HSQC spectra, the linkages of β-O-4 aryl ethers (A), resinols (B), and phenylcoumarans (C) could be obviously observed. It was found that DEL_-C3H and DEL_-CK exhibited similar but discriminative spectral patterns. Cross-signals of β-O-4 and -OCH₃ (_δC/_δH 55.7/3.70) were the prominent signals. _δC/_δH 71.9/4.87 were cross-signals of C_α-H_α correlations in the β-O-4 linkages, while the β-O-4 linkages (C_β-H_β) linked to G and S units can be distinguished at _δC/_δH 83.5/4.34 and 85.7/4.12. _δC/_δH 59.5/3.71–3.40 was assigned C_γ-H_γ correlations in the β-O-4 substructures. Meanwhile, the content of β-O-4 linkages in DEL_-C3H was higher than that of DEL_-CK, which was consistent with the results in a previous publication (Ma et al., 2021a). In addition, _δC/_δH 62.9/4.40 was assigned C_γ-H_γ correlations in γ-acylated lignin units (A′). This indicated that those lignin samples were partially γ-carbon acylated in β-O-4 aryl ether linkages and p-hydroxybenzoates (PB). In a recent publication, whether p-hydroxybenzoates acylate solely S units in transgenics poplar has not been confirmed (Ralph et al., 2012). Resinols (β-β, substructures B) can be easily identified in the spectra in conspicuous amounts. _δC/_δH 84.8/4.67, 53.4/3.04, and 71.1/3.80–4.19 were assigned their C_α-H_α, C_β-H_β and the double C_γ-H_γ correlations, respectively. The weak signal of C_α-H_α correlations of phenylcoumarans (β-5, _δC/_δH, 86.8/5.49) suggested that the low content of β-5 linkages (DEL_-CK 1.18/100Ar, DEL_-C3H 1.14/100Ar). This phenomenon could be attributed to the reduction of G units (relative to per 100Ar) as compared to that of CK based on a publication (Wang et al., 2018), in which it was reported that phenylcoumaran (β-5) was derived from the coupling of a monolignol with G units.

In the aromatic regions (_δC/_δH 100–135/5.7–8.0) of the 2D-HSQC NMR spectra (Figure 4), the chemical composition in the lignin samples (DEL_-CK and DEL_-C3H) can be clearly observed, such as syringyl (S) and guaiacyl (G) lignin units and some other lignin substructures. The C_2,6-H_2,6 correlation at _δC/_δH 103.8/6.68 represented the prominent signal for S-type lignin units, whereas the signal at _δC/_δH 106.2/7.18 was observed for the C_α-oxidized S-units (S′). Additionally, the different correlations of C₂-H₂ (_δC/_δH 110.9/6.97), C₅-H₅ (_δC/_δH 114.8/6.77), and C₆-H₆ (_δC/_δH 118.8/6.78) belonged to G-type lignin units. Specially, H_2,6 signals were detected at _δC/_δH 127.7/7.15, which increased from 0.6 to 1.1/100Ar, suggesting a striking elevation of p-hydroxyphenyl (H) units in transgenic poplar. The relative abundances of different linkages in lignin were quantified according to the previous literatures (Ma et al., 2021a; Wen et al., 2013a). The changes of the S/G ratio can intuitively reflect the compositional change of lignin samples. As is shown in Table 4, the S/G ratio in DEL_-CK was 2.82, while the S/G ratio for DEL_-C3H was 2.48. Interestingly, the relative content of H-type lignin in DEL_-C3H (1.11/100Ar) was higher than that in DEL_-CK (0.6/100Ar). This fact suggested that H-type lignin units have been elevated in C3H-downregulated poplar. Similarly, this phenomenon had been reported in a previous study, in which the increased amount of p-hydroxyphenyl unit was observed as well as a concomitant decrease of guaiacyl and syringyl units (Pu et al., 2009). In this study, the spectra shown in Figure 4 clearly showed the enhancement of PB content in DEL_-C3H. Precisely, the integral value of PB increased from 14.17 to 16.37/100Ar (as shown in Table 4), which was in agreement with a previous study (Ralph et al., 2012). In short, the C3H downregulated could increase the content of H units and p-hydroxybenzoate (PB) units in lignin. These results were also in line with the ¹³C-NMR section of H-_C3H, in which more PB units were detected. In short, 2D-HSQC spectra of native lignin and hemicellulose samples demonstrated that C3H downregulation indeed altered the chemical and structural features of these natural macromolecules to different extents. Due to these differences in composition and structure, the processing performance of transgenic poplar wood will be affected and the corresponding investigations are being explored.

The elevated lignin in the p-hydroxyphenyl (H) unit is produced by the downregulation of C3H in poplar wood, and the lignin content is also significantly reduced. A previous study investigated the effects of C3H downregulation on the lignin in alfalfa (Ralph et al., 2006). Ralph and coauthors found that the lignins rich in p-hydroxyphenyl units were produced by C3H downregulation, but the S/G ratio changed only slightly in alfalfa. Conversely, the S/G ratio of lignin was increased in the C3H downregulation poplar. Most of the relative H unit elevation was at the expense of G units rather than S units in poplar (Ralph et al., 2012). In general, genetic modification during lignin biosynthesis led to dwarfing or developmental abnormalities of the transgenic plants (Bonawitz and Chapple, 2013). However, with the growth and development of the plant, the transgenic poplar probably restores growth if there is an active cell wall feedback signaling responsible for dwarfing existing in lignin-deficient mutants (Bonawitz and Chapple, 2013). Simultaneously, as a pendant group, the PB content in lignin was increased in C3H poplar as compared to CK wood. However, the related transferase in poplar has not been identified. The identified transferase will help to understand how the transgenes affect the pendant group, such as p-hydroxybenzoates in hardwood and p-coumarate in gramineous plants.