Plant materials and culture media.

Six in vitro–cultured Vacccinium corymbosum L. cultivars, each representing different degrees of tissue culture recalcitrance, were used in our study. These included ‘Rowan’ and ‘NC5288’, advanced selections from North Carolina State University known for their high-quality berries and successful in-house in vitro culture. Also included were ‘Legacy’ and ‘Emerald’, which have been reported as being responsive to shoot regeneration on different media (Liu et al. 2010; Song and Hancock 2012). In addition, our study included ‘Jewel’ and ‘Pinnacle’, two cultivars known for their recalcitrance to tissue culture propagation. Previous attempts by the North Carolina State University Micropropagation and Repository Unit and PhyllaTech, LLC, to establish these cultivars in culture were unsuccessful (Phillips W, personal communication).

In vitro cultures were initially established from greenhouse-grown meristem explants, which were cultured on Lloyd & McCown Woody Plant Basal Mixture (WPM; product ID: L154; PhytoTechnology) (McCown and Lloyd 1981) supplemented with Murashige & Skoog (MS) Vitamin Solution (1000×) (product ID: M553; PhytoTechnology) (Murashige and Skoog 1962), additional Thiamine HCL (product ID: T390; PhytoTechnology) (276.85 μM) and 0.01% myo-inositol (product ID: 1703; Phytotechnology), 0.02% Ferric Sodium EDDHA (FeNa-EDDHA) (Sequestrene 138; product ID: E349; PhytoTechnology), 0.15% 2-(N-morpholino) ethanesulfonic acid (MES; product ID: M825; PhytoTechnology), 18.2 μM ZT, 2.0% D-Sucrose (product ID: S829; PhytoTechnology), and 0.6% Agar (product ID: A175; PhytoTechnology) at a pH of 5.0. After adventitious shoot formation, shoots were moved to Magenta™ vessel GA-7 plant culture boxes (Sigma Chemical Co., St. Louis, MO, USA) to be cultured on blueberry proliferation media containing WPM supplemented with MS vitamins, additional Thiamine (276.85 μM) and 0.01% myo-inositol, 0.02% Sequestrene 138, 0.15% MES, 9.12 μM ZT, 2.0% D-Sucrose, and 0.6% Agar at a pH of 5.0. In vitro plantlets were maintained under a 16-/8-h light/dark cycle and a 40-μmol·m^–2·s^–1 light intensity or dark conditions at 25 ± 1 °C, and were subcultured monthly until well-developed leaf blades from auxiliary shoots were obtained for use as explants.

To optimize the regeneration media for a broader range of blueberry cultivars, WPM supplemented with MS vitamins, 2.0% D-Sucrose, and 0.6% Agar at a pH of 5.0 (Song and Sink 2004, 2006) was selected as the basal regeneration medium. This medium was supplemented further with treatments TN1 through PZ10 (Table 1). Treatments labeled TN contained 0.045, 2.27, or 4.45 μM TDZ and 2.69 μM NAA, whereas those labeled TNZ included an additional 18.2 μM ZT. The concentration of 2.69 μM NAA and 18.2 μM ZT was selected based on its proven ability to achieve high regeneration frequencies in previous studies (Cappelletti et al. 2016; Liu et al. 2010; Song and Sink 2004, 2006). TDZ (Ghosh et al. 2018) concentrations were chosen for their documented success in media containing 4.45 μM TDZ (Liu et al. 2010), which was identified as the only treatment capable of regenerating the cultivar Jewel after five subcultures. Lower TDZ concentrations (0.0023, 0.91, and 2.27 μM) have been reported to promote in vitro proliferation and callus formation (Cappelletti et al. 2016; Liu et al. 2010). In contrast, a greater concentration of 9.08 μM TDZ resulted in no regeneration in any tested cultivars (Liu et al. 2010). Based on these findings, we tested 4.45 μM TDZ (treatments TN1 and TNZ2) alongside two lower concentrations: 2.27 μM (treatments TN3 and TNZ4) and 0.045 μM (treatments TN5 and TNZ6).

Table 1. Plant growth regulator combinations evaluated for optimized regeneration media in six V. corymbosum L. cultivars with varying tissue culture recalcitrance.

View Fullscreen

Treatments labeled CZ and PZ combined 18.2 μM ZT with either CPPU or 2iP, respectively. Previous research demonstrated that CPPU at 8.07 and 12.1 μM was more effective than ZT alone in promoting shoot regeneration from blueberry leaf sections (Liu et al. 2010). Similarly, 2iP at 4.92 μM was shown to outperform ZT in inducing shoots, buds, and meristematic nodules in leaves of blueberry cultivars Sunrise, Duke, and Georgia Gem (Liu et al. 2010). For the cultivar Duke, the inclusion of 2iP in the media has also shown potential for addressing challenges related to shoot regeneration and elongation in more recalcitrant blueberry cultivars (Cappelletti et al. 2016). Furthermore, a combination of 9.12 μM ZT and 24.6 μM 2iP has been reported to achieve regeneration frequencies ranging from 43.3% to 100% in cultivars such as Brigitta and Elliot (Song and Sink 2004). Building on these findings, we tested ZT in combination with CPPU at 8.07 μM (treatment CZ7) and 12.1 μM (treatment CZ8), as well as 2iP at 2.46 μM (treatment PZ9) and 4.92 μM (treatment PZ10). This approach allowed us to investigate whether ZT and CPPU combinations could replicate or improve on the regeneration outcomes observed with ZT and 2iP (Song and Sink 2004).

Callus induction and adventitious shoot regeneration.

Approximately 4- to 8-mm-long leaf explants from in vitro shoots of highbush blueberry cultivars Emerald, NC5288, Rowan, Legacy, Pinnacle, and Jewel were dissected carefully from the stem. The proximal portion of the leaf blade was then positioned with the abaxial side facing downward on 30 mL of regeneration media, supplemented with individual treatments TN1 through PZ10 (Table 1) and adjusted to a pH of 5.0, with each treatment consisting of two replicates, and each replicate comprising three plates. Because of the availability of in vitro plantlets, plates for ‘Rowan’ and ‘Legacy’ contained 10 leaf explants each, whereas plates for cultivars Emerald, NC5288, Pinnacle, and Jewel contained five leaf explants each. The plates were initially kept in the dark at a temperature of 25 ± 1 °C for 2 weeks, followed by a transfer to an incubation environment with cool-white fluorescent lights, also maintained at 25 ± 1 °C under a 16-/8-h light/dark cycle and a 40-μmol·m^–2·s^–1 light intensity or dark conditions. All treatments for all cultivars were initiated concurrently, and data collection was performed concomitantly across the same set of cultures to ensure consistency in scoring.

After a 2-, 4-, and 8-week culture period, data were collected on callus induction rate (percentage of the number of explants with callus in relation to the total number of explants cultured per petri dish), adventitious shoot regeneration frequency (number of explants with at least one shoot in relation to the number of explants cultured per petri dish), mean shoot numbers per explant, and callus-to-shoot conversion efficiency (percentage of the formed callus that resulted in shoot regeneration), including plate averages. After 8 weeks of culture and data collection, regenerated explants with shoots measuring at least 0.5 inch in length were transferred to plant culture boxes (Magenta™ vessel GA-7) containing blueberry proliferation media for continued culture at 25 ± 1 °C under a 16-/8-h light/dark cycle and a 40-μmol·m^–2·s^–1 light intensity or dark conditions.

Data analysis.

The data collected from the regeneration process underwent a comprehensive analysis to assess the effectiveness of 10 treatments in promoting callus formation and supporting the subsequent growth of adventitious shoots and explants. This evaluation was conducted at 2, 4, and 8 weeks of culture. However, because of contamination, data from the cultivars Pinnacle and Jewel were excluded from analysis beyond the 4-week mark. Consequently, the 8-week dataset included only ‘Emerald’, ‘Rowan’, ‘NC5288’, and ‘Legacy’.

To facilitate accurate comparisons, the means and standard errors of the means of biological replicates were calculated using RStudio v. 4.4.2 (Posit Software, PBC, Boston, MA, USA). Data normalization was performed by averaging values per explant for each plate, accounting for differences in the number of leaf explants per plate among cultivars Legacy and Rowan (10 leaf explants per plate), and cultivars Emerald, NC5288;, Jewel, and Pinnacle (five leaf explants per plate).

The normality of data were evaluated using the Shapiro-Wilk test, which returned P values indicating deviations from normality (P < 0.05), suggesting that parametric test assumptions were violated. Attempts to normalize the data using log and square root transformations did not improve normality sufficiently. Therefore, nonparametric statistical methods were applied for subsequent analyses.

To compare treatment effects within each cultivar, and cultivar responses across treatments, the Kruskal-Wallis test was used as a nonparametric alternative to analysis of variance. This rank-based test determined whether significant differences existed in callus induction, shoot formation, shoot number per explant, and callus-to-shoot conversion efficiency among treatments. When significant differences were detected (P < 0.05), Dunn’s test with Bonferroni-adjusted P values was performed for post hoc pairwise comparisons to control for multiple testing.

All statistical analyses were conducted in RStudio v. 4.4.2 (Posit Software, PBC, Boston, MA, USA), with statistical significance set at P < 0.05.

The effects of PGRs on callus induction.

Using leaf explants from in vitro shoots, we evaluated the effects of TDZ (0.045, 2.27, and 4.45 μM) and ZT (18.2 μM) in combination with NAA (2.69 μM), ZT (18.2 μM) with CPPU (8.07 and 12.1 μM), and ZT (18.2 μM) with 2iP (2.46 and 4.92 μM) on callus induction from six highbush blueberry cultivars: Emerald, NC5288, Rowan, Legacy, Pinnacle, and Jewel. Callus induction rates varied significantly across treatments in most cultivars, ranging from 10.0% to 65.0% under TDZ-based treatments (TN1–TNZ6), 30.0% to 56.7% under ZT + CPPU (treatments CZ7 and CZ8), and 46.7% to 83.3% under ZT + 2iP (treatments PZ9 and PZ10) (Fig. 1, Table 1).

View Figure

• Download as Powerpoint
• Download Figure

Kruskal-Wallis tests confirmed significant treatment effects in ‘Emerald’ (P = 0.0006), ‘Legacy’ (P < 0.0001), ‘NC5288’ (P = 0.0408), ‘Rowan’ (P < 0.0001), and ‘Pinnacle’ (P < 0.0001), whereas ‘Jewel’ showed no significant differences (Supplemental Table 2). Dunn’s post hoc tests revealed that treatments PZ9 and PZ10 (ZT + 2iP) increased callus induction significantly compared with TDZ-based treatments (TN1–TNZ6) in ‘Emerald’, ‘Legacy’, and ‘Rowan’ (P < 0.05) (Fig. 1, Supplemental Table 3). Notably, a significant difference was observed in ‘Legacy’ between treatments TN1 and TNZ2, suggesting that TDZ enhances callus induction only when used at a relatively low concentration in combination with NAA and ZT (P < 0.05) (Fig. 1, Supplemental Table 3). In ‘Pinnacle’, both ZT + 2iP (treatments PZ9 and PZ10) and ZT + CPPU (treatments CZ7 and CZ8) induced significantly greater callus formation than TDZ-based treatments (Fig. 1, Supplemental Table 3). In addition, in ‘NC5288’, treatment PZ9 increased callus induction significantly compared with treatment TN3 (Fig. 1, Supplemental Table 3).

Genotypic responses to treatments varied more under TDZ-based treatments (TN1–TNZ6) compared with ZT + 2iP (treatments PZ9 and PZ10) and ZT + CPPU (treatments CZ7 and CZ8) (Fig. 2). Kruskal-Wallis tests confirmed significant genotype effects within treatments TN1 (P = 0.0024), TNZ2 (P = 0.0002), TN3 (P = 0.0002), TNZ4 (P = 0.0057), TN5 (P = 0.0079), TNZ6 (P = 0.0037), and PZ9 (P = 0.0071) (Supplemental Table 2). Dunn’s post hoc tests showed significant genotype differences, particularly between ‘Pinnacle’ and most other cultivars under TDZ-based treatments (TN1–TNZ6), and between ‘Emerald’ and ‘Jewel’ as well as ‘Emerald’ and ‘Pinnacle’ under treatment PZ9 (Fig. 2, Supplemental Table 3).

View Figure

• Download as Powerpoint
• Download Figure

Hierarchical clustering of callus induction rates revealed a clear separation between treatment groups (Fig. 3A). Treatments containing ZT + 2iP (treatments PZ9 and PZ10) clustered together and were strongly associated with greater callus induction rates (≥ 60.0%) across most cultivars, particularly in ‘Emerald’, ‘Pinnacle’, and ‘Rowan’. In contrast, TDZ-based treatments (TN1–TNZ6) formed a distinct cluster characterized by lower callus induction rates, with treatments TN3, TN5, and TNZ6 showing the least callus formation across all cultivars. The ZT + CPPU treatments (CZ7 and CZ8) clustered separately, displaying intermediate callus induction rates ranging from 30.0% to 56.7% (Fig. 3A).

View Figure

• Download as Powerpoint
• Download Figure

PGR impact on adventitious shoot regeneration.

Using the same in vitro leaf explants, we next assessed the effects of treatments (TN1–PZ10) on shoot regeneration rates and mean shoot numbers across the six highbush blueberry cultivars. Both shoot regeneration and mean shoot production per explant varied significantly among treatments (Fig. 1). In TDZ + NAA treatments (TN1–TNZ6), shoot regeneration ranged from 3.3% to 65.0%, with an average of 0.1 to 2.3 shoots per explant (Fig. 1, Supplemental Table 1). Treatments containing ZT + CPPU (treatments CZ7 and CZ8) resulted in 10.0% to 36.7% shoot regeneration rates, producing 0.2 to 1.4 shoots per explant. When ZT was combined with 2iP (treatments PZ9 and PZ10), shoot regeneration ranged from 30.0% to 65.0%, with mean shoot numbers between 0.7 and 2.9, showing a broader range of response across cultivars (Fig. 1, Supplemental Table 1).

Kruskal-Wallis tests confirmed significant treatment effects for shoot regeneration and mean shoot number per explant across all cultivars (P ≤ 0.0001) (Supplemental Table 2). Dunn’s post hoc tests revealed that ZT + 2iP treatments (PZ9 and PZ10) increased shoot regeneration and mean shoot number per explant significantly compared with TDZ-based treatments (TN1–TNZ6) in ‘Jewel’, ‘Pinnacle’, ‘Emerald’, ‘Rowan’, and ‘Legacy’ (P < 0.05) (Fig. 1, Supplemental Table 3).

Under TDZ-based treatments (TN1–TNZ6), ‘Jewel’ and ‘Pinnacle’ failed to regenerate shoots entirely. At the same time, ‘Emerald’ showed no regeneration under treatments TN1, TN3, and TN5, which contained 0.045, 2.27, and 4.45 μM TDZ, respectively, plus NAA, as well as treatment TNZ2 (0.045 μM TDZ + NAA + ZT) (Fig. 1, Supplemental Table 1). ‘Emerald’ only regenerated shoots under treatments TNZ4 and TNZ6 (2.27 and 4.45 μM TDZ, respectively, + NAA + ZT), suggesting that TDZ promoted shoot regeneration in this cultivar only at greater concentrations and when combined with NAA and ZT (Fig. 1, Supplemental Table 1).

Conversely, ‘Rowan’ regenerated shoots only under treatments TN1 through TNZ4, which contained 0.045 and 2.27 μM TDZ, respectively, plus NAA and/or ZT, indicating the opposite trend—that TDZ promoted shoot regeneration in ‘Rowan’ only at lower concentrations, regardless of ZT addition (Fig. 1, Supplemental Table 1). In ‘Legacy’, significant differences were observed between TDZ + NAA and TDZ + NAA + ZT treatments, particularly between TN1, TN3, and TN5 vs. TNZ2 (P < 0.05), as well as between TNZ2 and TNZ6 (P < 0.05). These results suggest that the addition of ZT to lower concentrations of TDZ + NAA influenced shoot regeneration responses (Fig. 1, Supplemental Table 3).

In ‘NC5288’, ZT + 2iP treatments (PZ9 and PZ10) increased shoot formation significantly compared with treatments CZ7 and TN3 (P < 0.05), indicating that TDZ-based treatments were more effective in this cultivar, closely resembling the performance of the best-performing ZT + 2iP treatments (PZ9 and PZ10) (Fig. 1, Supplemental Table 3). In addition, ‘NC5288’ exhibited lower shoot regeneration under the CZ7 treatment compared with treatments PZ9 and PZ10, reinforcing that cytokinin type played a critical role in shoot formation responses in this cultivar (Fig. 1, Supplemental Table 3).

Genotypic responses to shoot regeneration and mean shoot number were more variable under TDZ-based treatments (TN1–TNZ6) compared with ZT + 2iP (treatments PZ9 and PZ10) and ZT + CPPU (treatments CZ7 and CZ8) (Fig. 2). Kruskal-Wallis tests confirmed significant genotype effects for shoot regeneration within treatments TN1, TNZ2, TNZ4, TN5, TNZ6, and PZ9 (P < 0.05) and for mean shoot number per explant within treatments TN1, TNZ2, TNZ4, TN5, TNZ6, and PZ10 (P < 0.05) (Supplemental Table 2). Dunn’s tests further identified significant differences, with ‘NC5288’ consistently exhibiting greater shoot regeneration and shoot numbers than most other cultivars under TDZ-based treatments (TN1–TNZ6) (Fig. 2, Supplemental Table 3).

Interestingly, genotypic variation was particularly evident in treatment PZ9 for shoot regeneration and treatment PZ10 for mean shoot number, suggesting that cytokinin concentration may influence the pattern of shoot formation. Shoot regeneration typically follows two distinct patterns: pattern I, where shoots emerge primarily at wound sites, and pattern II, which is characterized by uniform shoot formation across the explant surface (Liu et al. 2010). Under lower 2iP concentrations (2.46 μM), ‘Emerald’ and ‘Rowan’ exhibited predominantly pattern II regeneration, whereas greater 2iP concentrations (4.92 μM) induced pattern I (Fig. 4A1, A2, B1, and B2). In contrast, ‘Pinnacle’ responded inversely, displaying pattern I at lower concentrations and pattern II at greater concentrations (Fig. 4E1 and E2). ‘NC5288’, ‘Legacy’, and ‘Jewel’ consistently followed pattern I regeneration across both 2iP concentrations (Fig. 4E1, E2, D1, D2, F1, and F2, respectively).

View Figure

• Download as Powerpoint
• Download Figure

Hierarchical clustering revealed a clear distinction between treatment groups for both adventitious shoot regeneration and mean shoot number per explant (Fig. 3B and C). For shoot regeneration, ZT + 2iP treatments (PZ9 and PZ10) clustered together consistently and were strongly associated with the greatest regeneration rates, particularly in ‘Rowan’, ‘Emerald’, and ‘Pinnacle’ (≥ 50.0%) (Fig. 3B). In most cultivars, TDZ-based treatments (TN1–TNZ6) formed a separate cluster, exhibiting consistently low regeneration percentages. However, treatment TNZ2 grouped with ZT + 2iP in ‘Legacy’ and ‘NC5288’, achieving regeneration rates of ≥ 50.0% (Fig. 3B). Meanwhile, ZT + CPPU treatments (CZ7 and CZ8) formed an intermediate cluster, with moderate shoot regeneration rates ranging from 10.0% to 40.0% (Fig. 3B).

For mean shoot number per explant, ZT + 2iP (treatments PZ9 and PZ10) again clustered separately, showing the highest shoot numbers across most cultivars, particularly in ‘Rowan’ (2.92 shoots per explant), ‘Legacy’ (2.57 shoots per explant), and ‘Emerald’ (2.00 shoots per explant) (Fig. 3C). ‘NC5288’ and ‘Pinnacle’ exhibited lower shoot numbers, averaging 1.37 and 1.10 shoots per explant, respectively (Fig. 3C).

In contrast, TDZ-based treatments (TN1–TNZ6) clustered separately with consistently low shoot numbers across all cultivars, generally < 1.00, with most cultivars showing minimal or no shoot production (Fig. 3C). ‘NC5288’ displayed slightly higher shoot numbers than other cultivars under TDZ-based treatments, though still at low levels (Fig. 3C). Meanwhile, ZT + CPPU treatments (CZ7 and CZ8) occupied an intermediate cluster, producing moderate shoot numbers (1.00–2.27 shoots per explant), particularly in ‘NC5288’ and ‘Emerald’ (Fig. 3C). Although these treatments improved shoot production compared with TDZ-based treatments, they did not reach the levels observed under ZT + 2iP.

These findings reinforce ZT + 2iP as the most effective cytokinin combination for promoting both shoot regeneration and shoot proliferation, whereas TDZ-based treatments consistently resulted in poor performance across cultivars.

Influence of PGRs on callus-to-shoot conversion efficiency.

Except for ‘Pinnacle’, which failed to induce callus under TDZ-based treatments (TN1–TNZ6), successful callus induction was observed in all other genotypes (Fig. 1). However, callus induction did not result in increased shoot production consistently (Fig. 1). Notably, shoot production across all six cultivars was only stimulated after 2 weeks of culture when ZT was combined with 2iP (Supplemental Table 1). This trend persisted over more extended culture periods, as many calli failed to regenerate adventitious shoots even after 4 or 8 weeks (Supplemental Table 1).

To assess the relationship between callus formations and shoot regeneration, we analyzed callus-to-shoot conversion efficiency, which is defined as the percentage of calli that regenerated shoots successfully. Conversion rates ranged from 11.1% to 100% under TDZ-based treatments (TN1–TNZ6), from 27.8% to 66.7% under ZT + CPPU (treatments CZ7 and CZ8), and from 65.0% to 97.6% under ZT + 2iP (treatments PZ9 and PZ10) (Fig. 5, Supplemental Table 1).

View Figure

• Download as Powerpoint
• Download Figure

Kruskal-Wallis tests confirmed significant treatment effects on callus-to-shoot conversion efficiency across all cultivars (P < 0.01, Supplemental Table 2). Dunn’s post hoc tests further revealed that ZT + 2iP treatments (PZ9 and PZ10) improved conversion efficiency significantly compared with multiple TDZ-based treatments (TN1–TNZ6) in ‘Emerald’, ‘Jewel’, and ‘Rowan’ (P < 0.05, Supplemental Table 3).

‘Pinnacle’, which failed to form callus or regenerate shoots under TDZ-based treatments (TN1–TNZ6), was excluded from further analysis, as no significant differences in conversion efficiency were observed under the remaining ZT + 2iP (treatments PZ9 and PZ10) and ZT + CPPU (treatments CZ7 and CZ8) treatments (Fig. 5, Supplemental Table 3).

In ‘Emerald’, the conversion efficiency was significantly greater under treatments PZ9 and PZ10 compared with treatments TN1, TN3, TN5, and TNZ2 (P < 0.01), all of which failed to regenerate shoots from callus (Fig. 5, Supplemental Table 3). A similar trend was observed in ‘Jewel’, where treatments PZ9 and PZ10 outperformed all TDZ-based treatments significantly (P < 0.01), which were also unable to regenerate shoots from the callus (Fig. 5, Supplemental Table 3). In ‘Legacy’, treatments PZ9 improved conversion efficiency significantly over the unresponsive treatments TN3 (P = 0.0073) and (P = 0.0437), whereas treatment CZ7 (ZT + CPPU) also performed better than treatment TN3 (P = 0.0304) (Fig. 5, Supplemental Table 3). In ‘NC5288’, treatments PZ9 and PZ10 increased conversion efficiency significantly compared with treatment TN3, which failed to regenerate shoots (P < 0.05, Fig. 5, Supplemental Table 3). Last, in ‘Rowan’, treatment PZ10 outperformed treatments TN5 and TNZ6 significantly (P < 0.05), confirming its effectiveness in promoting shoot conversion (Fig. 5, Supplemental Table 3).

Intriguingly, treatment TN3 induced callus formation in five of the six cultivars. Yet, only ‘Rowan’ regenerated shoots, and even then at a low rate of 16.7% (Fig. 5). In addition, although TDZ-based treatments (TN1–TNZ6) induced callus formation successfully in ‘Emerald’ and ‘Jewel’, none except treatment TNZ4 in ‘Emerald’ (16.7% conversion) produced any shoots (Fig. 5). These findings highlight the limited effectiveness of TDZ-based treatments for shoot regeneration in the tested cultivars.

Genotypic responses to callus-to-shoot conversion efficiency were more variable under TDZ-based treatments (TN1–TNZ6) than under ZT + 2iP (treatments PZ9 and PZ10) or ZT + CPPU (treatments CZ7 and CZ8) (Supplemental Table 2). Kruskal-Wallis tests confirmed significant genotype effects within treatments TN1 (P = 0.0047), TNZ2 (P < 0.0001), TNZ4 (P = 0.0289), TN5 (P = 0.0003), and TNZ6 (P = 0.0025), whereas no significant cultivar differences were observed in ZT + CPPU (treatments CZ7 and CZ8) or ZT + 2iP (treatments PZ9 and PZ10) (Supplemental Table 2).

Hierarchical clustering revealed that ZT + 2iP treatments (treatments PZ9 and PZ10) formed a distinct group associated with the highest callus-to-shoot conversion efficiencies across all cultivars (Fig. 3D). Within this group, ‘Jewel’, ‘Pinnacle’, ‘NC5288’, ‘Legacy’, and ‘Rowan’ exhibited the greatest conversion rates (≥ 72.2%), whereas ‘Emerald’ showed slightly lower efficiency (65.0%) (Fig. 3D).

In contrast, TDZ-based treatments (TN1–TNZ6) clustered separately and were consistently associated with low or zero conversion efficiency across all cultivars. Notably, treatment TNZ2 showed slightly greater conversion rates in ‘NC5288’ and ‘Legacy’, but remained distinct from the ZT + 2iP cluster (Fig. 3D). Meanwhile, ZT + CPPU treatments (CZ7 and CZ8) occupied an intermediate position, with conversion rates ranging from 27.8% to 87.5%, depending on the cultivar (Fig. 3D). These findings reinforce the superior effectiveness of ZT + 2iP in promoting callus-to-shoot conversion across multiple cultivars, and further emphasizes the limited regenerative potential of TDZ-based treatments.