Effects of different explants on callus induction: The frequencies of induced-callus were investigated when explants had been cultured on the callus induction medium for four weeks (Table 2). Only a few calli were induced from immature inflorescences, while most of immature inflorescences directly initiated the roots. Thus, the callus induction frequency of immature inflorescence had not calculated in this study. In two genotypes, Chuannong 16 and 98-1266, the average callus induction frequency of immature embryos were 91.8% and 93.9%, respectively, and higher than that of mature embryos. The status of callus derived from different explants inoculated on medium MS₁ was evaluated according to the method of Ye et al. (2001). Many embryonic calli (+++ status) were induced from immature embryo derived-callus, which was yellowish and more compact. It indicated that immature embryos were more suitable for callus induction than those of immature inflorescences and mature embryos. The hydrocolloid-like calli (+ status) with ivory-white were easily turned to brown and initiated roots. Based on these results, the immature embryos were selected as explants for further studies. Effects of genotype and medium on culture responses of immature embryos: Using 8 genotypes and 6 media, the effects of genotypes and media on callus induction frequencies of immature embryos were investigated (Table 3). Significant differences of callus induction frequencies of immature embryos were detected. Among these genotypes, Chuannong 16 had the highest callus induction frequency, while Chuanmai 32 had the lowest callus induction frequency. The average callus induction frequencies of Chuannong 16 and Chuanmai 32 were 92.9% and 81.8%, respectively. There was no significant variation among the callus induction frequencies of 5 genotypes (Chuannong 16, Chuanyu 16,80-8, Yiyuan 2 and 98-1266), whndile significant difference was detected between these genotypes and the other genotypes.

The significant effects of medium on callus induction were observed. Among 6 media, the frequencies of callus induction on MS₂ and N6o were higher than those of the others. The average frequencies of callus induction on MS₂ and N6₀ were 94.0% and 93.1%, respectively. The lowest frequency (83.6%) of callus induction was observed on N6₁. When MS₂ and N6₀ were selected for callus induction media, it was observed that the frequency of initiated shoot directly from explants on N6₀(60.2%) was higher than that of MS₂. It indicated that MS₂ was superior to N6₀ for callus induction. Thus, we chose MS₂ as callus induction medium in the following procedures. The plantlet regeneration competence of callus derived from different genotypes: In order to increase the number of in vitro plantlets and to investigate the green shoot differentiation competence of callus derived from medium MS₂, two regeneration strategies have been applied in this study: (1) Medium MS₁₀ containing 2 kinds of auxins, 6-BA (2 mg/I) and IBA (0.5 mg/I), which could promote the shoot-like embryoid appearance, was used to induce green shoot at first step. When the shoot grow to 1-2 cm, the shoot with callus were transferred to MS₁₁ or MSu_11-1 for root initiation, then regenerated the young plantlets; (2) Medium MS_10-1 with 2 mg/l NAA and 0.5 mg/I KT was used for the induction of shoot-like and root-like embryoid at the same time. When the calli were transferred to MS₁₀ or MS_10-1 for two weeks, the green shoots on calli of 8 genotypes were investigated (Table 4). On MS₁₀, Chuanyu 12 had the highest frequency of green shoot differentiation (83.3%). On MS_10-1, Yiyuan 2 had the highest frequency of green shoots differentiation (87.5%). Comparison of the differentiation frequency for each genotype on two media, it indicated that MS₁₀ was superior to MS_10-1 for shoot differentiation. Only one genotype, Yiyuan 2, induced more green shoots on MS_10-1 (87.5%) than MS₁₀ (75.0%)
Transient expression of GUS in particle bombarded calli: Some calli derived from immature embryos of Chuannong 16 were randomly chosen for transformation via particle bombardment. After one week of bombardment, the β-glucuronidase (GUS) activity was detected and observed as blue spots through histochemical assays. The control explants were bombarded without plasmid DNA and blue spots were not observed under the microscope. All spots were dark blue in bombarded immature embryo (Fig. 2 g). The mean frequency of blue spots was 5.6%.