Seedballs from two sugarbeet lines F1004 (PI 590763) and F1015 (PI 605413) were used in this research. F1004 was released in 1984 and developed from mass selection for resistance to storage rot (Campbell and Bugbee, 1984). F1015 was released in 1996 as the first publicly available sugarbeet germplasm line with sugarbeet root maggot (Tetanops myopaeformis von Röder) resistance (Campbell et al., 2000). Both lines produce multigerm seedballs. Seedballs of F1004 and F1015 were stored in the USDA-ARS sugarbeet genetics laboratory in Fargo, ND, United States at 4°C and 40% relative humidity for 24 and 7 years, respectively. Seedballs of F1004 were polished by manually rubbing seedballs to remove the out loose layer of the pericarp, while seedballs of F1015 were raw fruits without any treatment. Preliminary tests indicated that germination of both lines was very low when seeds were germinated by incubating in distilled water or wetted soil. In addition, a set of 50 additional sugarbeet lines that was cold stored in the same lab for 10–37 years was used for validating the protocol of using H₂O₂ to enhance germination.

Germination experiments were designed to determine the optimum length of time to incubate seeds in a H₂O₂ solution and evaluate the effects of environmental conditions such as seed hydration, temperature, and light on H₂O₂-stimulated germination. Also, germination of seeds that were manually isolated from seedballs was performed to identify the nature of seed dormancy for the two sugarbeet lines.

For determining the optimum length of time for incubating seeds in a H₂O₂ solution, five H₂O₂ treatments along with a water control were used. For each treatment, seedballs were incubated in a 1% H₂O₂ solution for 1, 2, 3, 4, or 7 days, then incubated in distilled water for the remainder of the 8-days period. Germination was evaluated daily. After 8 days, ungerminated seeds from each treatment were incubated for an additional 2 days in distilled water, and seeds were considered as having no germinability if no germination occurred at this time. Each treatment included three replications for each line with each replication comprised of 100 seedballs in a 9-cm diameter petri plate. Therefore, a total of 300 seedballs from each line was tested in each treatment. All petri plates were incubated at room temperature under fluorescent lights with light intensity around 11.9 μmol m^–2 s^–1 for approximately 9 h per day. On the first day, seedballs in each plate were incubated in 15 mL 1% H₂O₂ or distilled water, depending on the treatment. On each of the following days, incubating solution was decanted off and seedballs were washed three times with distilled water. To maintain treatment conditions but leave seedballs partially uncovered by liquid, 5 mL fresh 1% H₂O₂ or distilled water was added to each plate according to the treatment. Seedballs were considered germinated when radicles emerged and percentages of germinated seedballs were recorded and compared among treatments.

To determine the effect of seed hydration on H₂O₂ enhancement of germination, five treatments were used with seedballs incubated in distilled water for 0, 1, 2, 3, or 4 days prior to treatment with a 1% H₂O₂ solution. The effect of light on germination was tested using seedballs incubated either in water or H₂O₂ solution for 8 days at room temperature in the dark or under the natural light obtained through a window, with an approximate light intensity of 26.8 μmol m^–2 s^–1 for about 12 h per day. To test the influence of temperature on germination in H₂O₂ solution, two treatments were used. For the first treatment, seedballs were incubated in a 1% H₂O₂ solution at 4°C for 3 days and then incubated in H₂O₂ for an additional 4 days at room temperature; for the second treatment seedballs were incubated in 1% H₂O₂ at room temperature throughout the experiment. All experiments utilized three replicate petri plates per treatment for each of the two lines with each replicate comprised of 100 seedballs. Seedballs were rinsed with water each day and fresh 1% H₂O₂ solution or distilled water were added as described above. For validating the H₂O₂ germination protocol, two plates each containing 100 seedballs, were used for H₂O₂ treatment and water control, respectively, for each of 50 sugarbeet lines described above. Seeds were rinsed daily as described for germination experiments, and germination percentage was recorded on the 8th day. Analysis of variance (ANOVA) was conducted using the aov function in R¹ to test the significance of each factor affecting germination, and the LSD.test function in R² was used to calculate LSDs (least significant differences) at P < 0.05 for each treatment in each line to compare the significance of treatments on germination. Analyses of correlation, t-tests, and standard deviation calculations were conducted using functions of Microsoft Excel.

To determine if endogenous physiological or exogenous physical dormancy was responsible for the low germination of the two lines when seedballs were germinated in water, individual seeds were carefully isolated from seedballs. A hammer was used to slightly crack the pericarp, and seeds were carefully excised from broken seedballs. The isolated seeds were visually inspected and any seeds with visible damage were excluded. A total of 50 isolated seeds from each line were placed in a 9-cm diameter petri plate and incubated in 5 mL of distilled water. Seeds were washed daily by decanting the old rinse solution and washing seeds with water. Germination was checked daily.

RNA-sequencing (RNA-seq) was used to analyze gene transcripts during seed imbibition and up to germination. From germination tests using H₂O₂ solution it was determined that seeds very likely had lost viability if they showed no germination after 8 days in H₂O₂ solution. Samples for RNA extraction thus were prepared using an 8-days period with seedballs treated with 1% H₂O₂ solution or distilled water using two replications for each line. A total of 64 plates (two lines × two treatments × two replications × 8 days) were used with each plate comprised of 100 seedballs. Eight plates (two lines × two treatments × two replications) were sampled each day. Since the purpose of this experiment was to detect genes expressed during seed imbibition and metabolic reawakening under H₂O₂ or water treatment, germinated seedballs with roots that extended out of the seed coat were excluded to minimize mRNAs that were transcribed in the root tissue that were irrelevant to germination. Therefore, seedballs with emerged radicles were counted each day and removed from each plate before collecting samples for RNA extraction. Collected samples were immediately frozen in liquid nitrogen and stored at –80°C until processing.

Total RNA was extracted and purified using a Qiagen-RNeasy Micro Kit according to the manufacturer’s instructions. Based on a preliminary test that almost no RNA was extracted from samples using seed pericarp only (data not shown), the whole seedballs thus were used for total RNA extraction. An Agilent 2100 Bioanalyzer (Agilent RNA 6000 Nano Kit) was used to determine RNA concentration, RIN value, the 28S/18S ratio, and fragment length distribution, and a NanoDrop 2000 spectrophotometer (Thermo Scientific, Inc., Waltham, MA, United States) was used to assess the purity of the RNA samples. To prepare sequencing libraries, mRNAs were isolated using poly-T oligo-conjugated magnetic beads, and cDNA was synthesized using reverse transcriptase Super Script II (Invitrogen, Waltham, MA, United States) at 42°C and random primers after RNA fragmentation. Second strand cDNA was synthesized using DNA Polymerase I and RNase H. The cDNA fragments were modified by the addition of A-tails and ligation of sequencing adapters and purified and enriched through PCR amplification. After a quality check, samples were pooled, and single strand DNA circles (ssDNA circles) were made. DNA nanoballs (DNBs) were generated by rolling circle replication (RCR) of ssDNA circles and the DNBs were loaded into patterned nanoarrays. Sequence data comprised of 100-bp pair-end reads were collected on a DNBseq platform using the combinatorial probe-anchor synthesis (cPAS) method (Fehlmann et al., 2016).

Raw reads were filtered by removing adapters and trimming low quality bases (Phred score < 20) at the end of reads using the program SOAPnuke³. Filtered reads were aligned to the sugarbeet genome sequence, RefBeet v1.2.2 (Dohm et al., 2014; available from NCBI⁴) using the tool HISAT2 (Hierarchical Indexing for Spliced Alignment of Transcripts) (Kim et al., 2015). Transcript assemblies were generated using the computer package StringTie (Martin and Wang, 2011; Pertea et al., 2015) followed by use of Cuffcompare (Trapnell et al., 2012) to compare reconstructed transcripts to the reference annotation. The computer package CPC (Kong et al., 2007) was used to predict novel transcripts and merge novel transcripts with the reference genome to get a complete transcript reference. Transcript abundance was normalized to FPKM (fragment per kilobase of transcript per million reads mapped) values which were calculated using the computer package RSEM (Li and Dewey, 2011). Differentially expressed genes (DEGs) were detected using PoissonDis that is based on the Poisson distribution (Audic and Claverie, 1997) and limited to DEGS with a fold change > 2 and a false detection rate (FDR) at P < 0.001. Comparisons of gene expression were performed between treatments (H₂O₂ treatment vs. water control) to identify transcriptional changes in seed due to H₂O₂ and between the two sugarbeet lines (F1004 vs. F1015) to reveal transcriptional differences corresponding to different types of seed dormancy.

Gene ontology (GO) analysis was done according to Young et al. (2010) and GO functional enrichment determined using phyper, a function of R⁵. Gene functional annotation used the gene IDs and annotations in RefBeet (Dohm et al., 2014). For novel genes that did not align with any annotated genes in the sugarbeet genome, DNA sequences were compared with those stored in the NCBI GenBank database⁶ to infer gene function based on sequence similarity with a critical expectation value (E value) threshold set at 1.0 × 10^–5 (Altschul et al., 1997). Pathway analysis of DEGs was conducted through KEGG (Aoki-Kinoshita and Kanehisa, 2007). Online protein databases (such as UniProt, SwissProt, and TrEMBL) were also searched using amino acid sequences or protein names as queries to identify possible biological processes or pathways of the proteins encoded by DEGs.

In the water control, none of F1015 but 7.3% of F1004 seedballs germinated (Table 1). The 1-day treatment in 1% H₂O₂ solution had almost no effect on enhancing germination when compared with the water control (Table 1). This suggests that during the first day of incubating seedballs in H₂O₂ solution or water, only the pericarp rather than the enclosed seeds, were taking up water, and thus seed imbibition was unlikely to start on the first day. As the days of H₂O₂-treatment increased, germination of both lines increased by about 10% in the 2-day H₂O₂ treatment and by another 30% in each of the 3- and 4-days H₂O₂ treatments (Table 1). The highest germination percentage occurred in the 7-days H₂O₂ treatment in both lines with germination reaching 80.0% in F1004 and 86.7% in F1015 (Table 1), an increase of about 10% in F1004 and 20% in F1015 compared to the percentage observed in seedballs of the 4-days H₂O₂ treatment. This suggests that few seedballs require five or more days incubating in H₂O₂ solution to promote germination. The ungerminated seeds from the 7-days H₂O₂ treatment were likely unviable since they were mostly unable to germinate even after extending the H₂O₂ treatment time. Therefore, incubating seeds in H₂O₂ solution enhanced sugarbeet germination with the highest germination percentage achieved using a 7-days treatment.

For evaluating the effect of seed hydration on germination in H₂O₂ solution, seedballs were preincubated in water for 1–4 days to achieve different levels of hydration. Overall, all water-preincubated seedballs germinated more poorly in both lines, with the reduction in germination more severe as seedballs were incubated for longer durations in water (Table 2). Germination of seedballs preincubated in water for 1 day followed by H₂O₂ treatment was reduced by approximately 30% compared to seedballs that receive no water pretreatment. A further 20–30% reduction in germination was observed with 2 and 3 days of water preincubation. Seedballs preincubated in water for 4 days followed by H₂O₂ treatment showed a similar germination percentage to the water control (Table 2), indicating that H₂O₂ treatment did not enhance germination if seeds were preincubated in water for 4 or more days. Therefore, increasing hydration in seeds did not improve germination under H₂O₂ treatment.

Dark conditions increased germination by about 10% in water-incubated F1004 seedballs but had no effect on water-incubated F1015 seedballs. When using H₂O₂ treatment to induce germination, dark conditions increased germination of F1004 seedballs by 6.4% but had no effect on germination of F1015 seedballs (Table 2). Therefore, dark conditions are limitedly effective in increasing germination when utilizing a H₂O₂ treatment, with its effectiveness dependent on genotype.

Three days of cold treatment greatly decreased germination of both lines when germinated in H₂O₂ solution or water (Table 2). Only 4.3% of cold-treated F1004 seedballs germinated in the water control, which was significantly reduced from the 19.3% germination observed in the water control at room temperature. With the H₂O₂ treatment, germination of seedballs that received the 3-days cold treatment was 12.0% in F1004 and 47.7% in F1015. These germination percentages were much lower than those (80.7 and 82.0% in F1004 and F1015, respectively) obtained with seedballs that were not cold treated (Table 2). Therefore, no cold treatment was needed when using H₂O₂ to induce germination. Overall, these results suggest that using H₂O₂ solution to continuously treat seeds for 7 days in the dark at room temperature provided the highest germination percentage.

The ability of H₂O₂ to promote germination was tested using seeds from 50 randomly selected sugarbeet lines that had been cold stored for 10–37 years. Average germination percentage of these seeds using the H₂O₂ treatment reached 73.7% with a range of 38.0–99.0%. This was 49.7% greater than germination observed in water controls (average at 24.0% with a range of 1.0–70.0%). Analysis by t-tests indicated that mean percentages of the two treatment were significantly different at probability level of P < 0.01 (Table 3), which proved that the H₂O₂ protocol significantly enhanced germination of sugarbeet. Correlation analysis found that germination percentages of the 50 lines under the two treatments were highly correlated (r = 0.715, P < 0.01, Table 3), which means the H₂O₂ protocol will have a better chance to improve germination if seeds were able to germinate in water incubation. Particularly, the protocol showed a promising ability of rescuing germplasms such as F1011, F1012 and F1013 that had very low germination percentages using a water incubation (Table 3).

When germinating seeds isolated from seedballs, only 4% of F1004 seeds germinated in water incubation for 4 days (Table 4), a germination percentage that was similar to that achieved when incubating F1004 seedballs in water. However, 60% of isolated seeds of F1015 germinated when incubated in water (Table 4), which was much higher than the 0% germination of seedballs incubated in water but close to the germination percentage achieved under H₂O₂ treatment. Therefore, difficulties in germinating F1015 seedballs when incubated in water were mainly caused by the physical barrier of the seed coat. In contrast, the consistently low germination rate in F1004 seedballs and isolated seeds indicated that endogenous physiological dormancy was likely to be responsible for inhibiting germination.

Of the seedballs used for RNA-seq analysis, germination after 8 days in a H₂O₂ solution reached 82.0% in F1015 and 66.4% in F1004, which far exceeded the 0% germination in F1015 and the 7.6% germination in F1004 from water-incubated controls (Figure 1). Germination of seedballs in H₂O₂ treatment in both lines occurred predominantly on the 3rd, 4th, 5th, and 6th days, with maximum germination on the 4th day (Figure 1). This agreed with observations made in previous germination tests that seed imbibition started on the second day, germination peaked at the 3rd and 4th days and finished on the 7th and 8th days of the treatment. Therefore, seed samples on the first day of H₂O₂ and water treatments were collected as controls for gene expression before germination started, and samples collected on the second day of the treatment represented gene expression after imbibition began in a few seeds under each treatment. Seed samples from the 3rd and 4th days of the treatments, when germination was greatest, were pooled and presumed to have the greatest expression of genes related to germination. Pooled samples from the 5th and 6th days, and the 7th and 8th days of the treatments were collected to assist in identifying genes that diminish in expression during germination.

RNA extracted from the F1004 seed sample on the second day in the water control was of low quality and was removed from RNA-seq analysis. RNA extracted from the remaining 19 samples that were collected from two sugarbeet lines (F1004 and F1015) under two treatments (H₂O₂ treatment vs. water control) on different days (1st, 2nd, pools of 3rd and 4th, 5th and 6th, and 7th and 8th), however, were used to produce RNA sequencing libraries. A total of 86.8 Gb bases of RNA sequence was generated with an average of 4.6 Gb bases per sample. After quality trimming and filtering, 868 million clean reads remained, yielding an average of 45.7 million reads per sample. About 76.7% of the sequences obtained mapped to the sugarbeet RefBeet reference genome with 59.0% of them mapped uniquely to single location (Supplementary Table 1). Overall, a total of 26,668 genes were identified with 22,287 annotated and 4,401 that were novel genes with unknown functions.

Comparisons of transcriptome profiles from seed samples collected on different days all showed significant gene expressions difference between H₂O₂ treatment and water control in both F1004 and F1015 (Supplementary Figure 1). Combining all expression differences in different samples and genotypes identified a common set of 173 DEGs, of which 90 genes were expressed at higher levels in seeds treated by H₂O₂ and 83 genes were expressed at higher levels in seed samples from the water control (Table 5 and Supplementary Table 2). The 90 upregulated DEGs from the H₂O₂ treatment mostly were expressed at low levels in samples collected on the first day of the treatments but increased in samples collected on the second day and the pooled samples from the 3rd and 4th days, then declined in pooled samples for the 5th and 6th, and 7th and 8th days of the treatment (Figure 2), which mirrored germination during the 8 days of the experiment (Figure 1). An analysis of gene functions found 62 (68.9%) genes were related to growth, with 23 and 11 genes involved in the cell cycle and DNA replication, respectively, suggesting that H₂O₂ treatment promoted gene activities related to cell cycle and growth. Of the remaining H₂O₂-promoted DEGs, five were of unknown function, and 12 (13.3%) and 11 (12.2%) were related to the regulation of gene expression and stress responses, respectively. Of the DEGs related to gene expression regulation, nine were transcription regulators, two were involved in signal transduction, and one was involved in a phytohormone-signaling pathway. From the 11 DEGs related to stress responses, seven were involved in responses to oxidative stress, two were involved in defense responses, and two were involved in response to cold and osmotic stress (Table 5 and Supplementary Table 2).

Of the 83 DEGs with greater expression in the water controls, 36 (43.4%) were related to growth, 29 (34.9%) were involved in stress responses, and seven (8.4%) were regulators of gene expression. Expression of these DEGs decreased as seeds were incubated for longer durations in either H₂O₂ solution or distilled water, but expression declined more rapidly in samples treated with H₂O₂ (Figure 3). Of the DEGs with biological functions related to stress responses, eight were involved in response to oxidative stress, four were related to response to cold, 11 were involved in response to osmotic or water deprivation stress, and three corresponded to cell protection. All of these functions suggest that transcripts that were more highly expressed in water-incubated seeds were unlikely to be genes expressed during imbibition since seedballs were not subjected to cold or water deprivation during this time. Rather, these DEGs were presumably stored mRNAs that were produced while seeds were stored under dry and cold conditions or were transcribed when water content declined during seed maturation. Expression of these DEGs progressively decreased during the duration of the experiment in seedballs under both treatments, which is also consistent with a degradation of stored mRNAs during imbibition. Therefore, the faster decline in mRNAs that were likely to be present in stored seedballs under H₂O₂ treatment suggests that H₂O₂ accelerated the transition of seeds from dormancy to germination.

Germination tests using isolated seeds indicated that germination was restricted by physiological dormancy in F1004 and physical dormancy in F1015. H₂O₂ treatment increased germination for both lines, presumably by improving water and oxygen penetration through the pericarp in F1015 and by triggering gene expression changes to overcome physiological dormancy in F1004. Therefore, comparison of gene expression between these two lines when their germination was promoted by H₂O₂ could reveal changes in gene expression that are related to the alleviation of physiological dormancy.

Since previous experiments found that incubating seeds in H₂O₂ solution for 1 day had no effect on germination and seeds that remained ungerminated after 7 days of H₂O₂ treatment were likely to be unviable, the transcription data from seedballs treated with H₂O₂ for 2–6 days in the two lines were used to identify DEGs between F1004 and F1015. In total, 370 DEGs were detected between the two lines with 101 and 269 DEGs more highly expressed in F1015 and F1004, respectively (Table 6 and Supplementary Table 3). The significantly higher number of DEGs in F1004 suggests that overcoming physiological dormancy required alteration of a greater number of gene activities than were needed to overcome the physical dormancy of F1015.

Analysis of DEGs by biological function indicated a large difference between the two lines in genes involved in gene expression regulation. For example, F1004 had 28 upregulated DEGs involved in expression regulation of which 15 were transcription regulators, seven were involved in the ABA-mediated signaling pathway, one was involved in the GA-mediated signaling pathway, and five were involved in other signal transduction processes. In contrast, only seven DEGs that were involved in gene expression regulation were expressed at a higher level in F1015, including one transcription regulator, one involved in ABA-mediated signaling, three that were involved in auxin-activated signaling pathways, and two involved in other signal transduction processes (Table 6 and Supplementary Table 3). Overall, this suggests that ABA-mediated signaling pathways, along with a significant number of transcription regulators, played important roles in cell regulation toward overcoming physiological dormancy.

Another significant difference in gene expression between the two lines were the number of DEGs involved in responses to oxidative stress with 14 expressed at higher levels in F1004 but only three that were more highly expressed in F1015. The H₂O₂ treatment created an oxidative environment for both lines, but the greater number of genes in line F1004 involved in response to oxidative stress indicates that some oxidative stress response genes may participate in regulating germination rather than merely providing cell protection against oxidative stress. In addition, 38 and 172 DEGs were non-annotated with unknown functions in F1015 and F1004, respectively (Table 6 and Supplementary Table 3). The larger number of unknown DEGs in F1004 provides further evidence that release of physiological dormancy requires more complex changes in gene transcription than release of physical dormancy.