Mutation breeding is a way to induce crop mutation by using physical and chemical factors artificially, and breed fine varieties directly applying to production or create new germplasm resources to improve existing varieties based on the selection and identification of mutants (Zhang et al., 1996). As the fundamental source of genetic variation, mutation is of great significance to research on genetics, breeding and evolution. Colch- icine is a good polyploid mutagen and usually used for chromosome doubling. It can also induce genetic variation of biological traits besides doubling chromosome in cells (Tiwarl and Khanorkar, 1984; Liu et al., 2006).

Azuki bean is one of cultivated species, classified to Vigna of Phaseoleae in Papilionaeeae of Leguminosae. It is also one of Chinese most ancient cultivated crops. The seeds of Azuki bean are nutritious with high protein and low fat, and also have medicinal values. Azuki bean is good for health and a bean crop of Chinese traditional foreign exchange earnings (Yu, 2005). Colchicine had been used to induce allelic mutation in gramineae and vegetables, resulting in the generation of variants early maturing, multi-flowers, big leaves, multi-tillers and multi-chloroplasts (Cai et al., 1992; Zhang et al., 1996; 2000; Caroline et al., 2003; Wang, 2008). However, few studies are reported on mutation of bean plants induced by colchicine. In this research, a number of mutants inducing by colchicine were obtained, some excellent variation plants were selected. It provided valuable basic materials for the gene mapping and cloning of azuki bean, studies on functional genomics, innovation of germplasm and mutation breeding.

1 Results
1.1 Induced-mutation effect and influences on agro- nomic traits
Colchicine-induced azuki bean M1 had rather low seed- ling rate and plant survival rate compared with the control group, and its relative values were lower than 50%. The greatest change in colchicine treatment was the plant height of M₁ significantly lower compared with the control, while colchicine-treated yield factors of M₂ and M3 generation were increased than the control. M1 generation by treatment of two different concentra- tions had obviously differences in seedling rate, plant survival rate and variation frequency. The greater the treatment concentration was, the smaller the seedling rate and plant survival rate would be, while the higher the variation frequency would be, which might be caused by the relatively great damages of high concentrations on seeds (Table 1).

The plant height of M1 generation was shown the greater the treated concentration was, the shorter the plant height would be. The 0.2% colchicine had the best treatment effect by integrative analysis on node and branch number on main stem, pod number per plant, yield per plant and 100-seed weight (Table 2).

The agronomic traits of mutants in M2 and M3 generation were investigated, and the plant heights were similar to those in the control (Table 3; Table 4), but these mutants were superior to the control in pod number per plant, the number of seeds per plant, yield per plant and 100- seed weight. The seed number per plant and 100-seed weight of colchicine-induced M₂ generation exhibit sign- ificant difference at 5% and 1% levels than the control’s, respectively (Table 3). Only difference of 100-seed weight between treatments was significant, other traits had little differences. The number of seeds per plant treated by 0.2% concentration was relatively small, but yield per plant was increased 33.98% than that with 0.4% concentration, and 100-seed weight was increased 10.75%. Therefore, the 0.2% colchicine was more appropriate for the treatment.

Colchicine-treated M₃ generation had no significant di- fferences in plant height, branch number of main stem and pod number per plant compared with the control (Table 4), while significant differences occurred at 1% level in the of number of nodes on main-stem , pod width and 100-seed weight. The values of main-stem node number, main-stem branch number, pod number per plant, seed number per plant, single-plant yield and 100-seed weight by the treatment of 0.2% concentration are higher than treated by 0.4% concentration. Meanwhile, by com- prehensive analysis on plant survival rate, seedling rate, mutant frequency and agronomic traits of two kinds of treatments, 0.2% colchicine was optimal for mutagenesis of azuki bean.
  
1.2 Variation types of M₃ generation
Variations occurred in plant traits of colchicines-induced M₂ generation. The variations of leaf traits, plant archi- tecture, pod colour and seed traits, etc. such as light green leaf, sword leaf, kidney-shaped leaf, higher plants, thick stalk, black pod, dark red seed and big seeds were obtainned by treatment with different concentra- tions of colchicine. M₃ mutant types would be different to a certain degree. 0.4% colchicines-treated variations were more abundant than those by treatment of 0.2% concentration, some unique mutant types like light green leaf, sword leaf, glave kidney leaf, higher plant, few branches, thick stalk, black pod, dark red seed and big seed appeared (Table 5).

1.2.1 Leaf variation
Jingnong 6 has ovate leaves, the mutants of sword leaf, small heart-shape leaf and kidney leaf were produced in M₃ generation. Leaf-shape mutation frequency treated with 0.4% colchicine was 12.92%., A total of 19 plants of sword leaf, 10 small heart-shape leaf and 12 kidney leaf plants were found (Figure 1), including 7 of upper glave leaf and lower kidney leaf. The mutants of 17 kidney leaf and 6 small heart-shape leaf appeared in the treatment of 0.2% colchicine, with the leaf-shape mutation frequency of 2.40%.

In M₃ population treated with the 0.4% colchicine, 9 deep-green leaf plants, 8 light green leaf, 3 new leaf yellowing and 4 yellow spot leaf plants were generated with the leaf-color mutation frequency of 2.40%. In 0.2% colchicine-treated M₃ population, two plants of deep-green leaf, two of new leave etiolated plants and one of yellow spot leaf appeared with the leaf-color mutation frequency of 0.50% (Figure 2).

1.2.2 Plant trait variation
In colchicine-treated M₃ population, sprawl variant appeared with no significant differences in yield factors compared with the control. Vertical, sturdy and compact plant types got in the 0.4% colchicine-treated M₃.

Plant height of JN6 is between 40 and 60 cm in general, with the average of 42.51 cm. However, in 0.4% colch- icine-induced M₃ populations, 10 plants with the height between 70 and 87 cm were produced (Figure 3), and the average height was 78.05 cm, increased by 83.6% than the control.

Significant variations occurred in branch number of plants such as excessive branch and few branch in M₃. JN 6 Variety generally had 3 or 4 branches, while, 3 mutants with 8 branches and 7 few-branch variants with 1 branch appeared. All few-branch mutants were obtai- ned with the treatment of 0.4% colchicine.
 

1.2.3 Pod color and seed trait mutation
JN 6 has white pods, but 15 black pod mutants appeared in 0.4% colchicine-induced M₃ (Figure 4), and the mutation rate was 1.5%.

Seed colour of JN6 is russetish, while mutations of black red and light red seeds were found in M₃. The average values of L*, b* and a* for different seed color mutants were measured with Color Difference Meter (Table 6). Additionally, the pod color and seed color were closely related. The mutants with black red seed color were ordinarily those with black pod (Figure 5), and all black red seed mutants were obtained by the treatment with 0.4% colchicine.

1.2.4 Variation of growth period
The growth period of JN6 is usually from 90 to 95 days. 9 early-maturing mutants were found in M₃, earlier from 10 to 15 days than the control. 8 late-maturing mutants were later from 7 to 10 days than the control.

1.3 Analysis on trait stability of mutants in M₄ generation
Mutant traits were investigated in M₃ and M₄. Trait stab- ility of 208 mutant lines in total involving 17 variational types of leaf shape and color, plant architecture, plant height, definite growth, maturing characteristics, pod and seed color was analyzed. 142 stable inheritance lines were obtained. The traits of sprawl, yellow-white seed, higher plant mutant could be steadily inherited. Kidney leaf , small heart-shape, sword leaf mutant traits were mostly inherited. New leaf yellowing and glave kidney leaf mutants had lower genetic stability and higher segregative lines (Table 7). Some mutants of glave kidney leaf in M₃ segregated out sword, kidney, heart-shape leaves in M₄.
 

2 Discussion
Chemical mutation has advantages of simple equipments, convenient operation and so on. According to some data analysis, chemical mutagens have effects on genes with significant locus specificity, and their mutagenic effects are closely related to mutagen types. Previously, muta- gens such as ethyl methane sulphonate (EMS), diethyl sulfate (DES) and azides were commonly used in the chemical mutation breeding. However, practices prove that these mutagens have strong mutagenic effects on the trait such as leaf color and albino seedlings, but with poor mutagenic effects and poor inheritance transmissi- bility on valuable traits such as early maturity, disease resistance and high yield, etc. Previous studies showed that the chemical mutation frequency was influenced greatly by treatment time (different crop growth stages) and method. Pu et al. (2005) discussed the suitable concentration and treatment time of colchicine-induced azuki bean, and it was found that the treatment concen- tration had more influences on the seedling rate and plant survival rate of M₁ generation than treatment time. Yang et al. (1997) treated the wild soybean seeds after 4-day germination with colchicine for 3 days, and then significant changes took place in ultrastructures of cells in the root tip meristematic zone, and these cells did have the characteristics of cell differentiation without mitosis. The valuable mutation of multiple types obtained by Cai et al. (1992) in rice, wheat and barley proved that colchicine as an effective chemical mutagen. Some of valuable mutation types were obtained in this study, colchicine can be applied in azuki bean mutation bre- eding. The 0.2% and 0.4% colchicine were used to treat dry seeds of azuki bean for 12 hours, with abundant mutation types obtained. By preliminary screening, 263 mutants were selected with absolute differences in plant height, plant architecture, leaf color, leaf shape, pod color, seed color, and maturing characteristics by colchicine- induced mutation. Similarly, Wang et al. (2009) used aqueous solution mixed by the 0.05% colchicine and 2% DMSO to intermittently treat Baofeng 2 pea seeds for 48 hours, and fertile autotetraploid was found by morphological and cytological identification and com- parison on the latter generations of treated pea. Signifi- cant differences of plant shape between autotetraploid and its diploid were shown as below: high plants , thick- ened stems and tendrils, total leaf area per plant being larger, leaf color greener, flowering stage delayed, bigger pollen grains and pollen mother cell, larger pod , self-fer- tility, and bigger seed. In this research, 71.86% mutants were obtained by treatment of 0.4% colchicine, the greater the treatment concentration was, the higher the variation rate would be, and the more abundant the variation types would be. However, the mutants of better yield factors was obtained in the treatment of 0.2% colchicine.

The method of 0.2 and 0.4% colchicine treating for 12 h is effective for inducing mutation of azuki bean, with fast homozygosity, abundant variation, but chromsome doubling was induced. These mutants aslo provided valuable materials for the breeding of azuki bean and studies on gene mapping and positional cloning.

3 Materials and methods
3.1 Materials and colchicine treatment
Jingnong 6 (JN6) variety of adzuki bean was provided by Azuki Bean Research Group of Beijing University of Agricultural. 1 000 seeds of JN6 were treated with 0.2% and 0.4% colchicines for 12 hours, respectively. The treated seeds were planted in experimental plots of Beijing University of Agricultural with 40 cm row spacing, 15 cm plant spacing and double-grain dibble seeding from 2006 to 2008. 

3.2 Trait investigation and data analysis
During the growing period, seedling stage, trefoil stage, flowering stage, pod-setting stage and maturing stage were investigated, and seedling rate and plant survival rate were counted. All plants were harvested when mat- ured and dried for plant laboratory tests with items inclu- ding plant height, start nod of branches, branch number, pod length, pod width, pod number per plant, the number of seeds per plant, 100-seed weight and yield per plant. All data were processed by SPSS statistical software.

3.3 Seed Color Mensuration
Seed color was tested with CR-200b Color Difference Meter, manufactured by Japan-based Minolta Camera Co., Ltd., which used the O/D mode of diffused illumi- nation and vertical light interception. L* indicated bri- ghtness index, and the greater the value was, the brighter it would be; a* and b* were nominated as redness and yellowness index, indicating the red and yellow degree of azuki bean seeds respectively. The greater a* value was, the redder the color would be; the greater b* value was, the fresher the color would be. 5 plump seeds were selected from each seed color mutant and control, for measuring 3 times, and counted the average.

Authors’ contributions
Baomei Wu carried out this research, analyzed data and drafted partial manuscript. Hongxia Liu and Bo Zhao, Xing Tong helped with the analyses and wrote substantials parts of the paper. Xing Tong and Qingmu Zhao worked on obtaining data. Ping Wan conceived the project, performed the experiment designs, super- vised the data analysis and revised this paper. All authors read and approved the final manuscript.

Acknowledgements
This study was jointly funded by Science and Technology Rese- arch Fund of Educational Committee of Beijing City (KM₂010- 10020004), Blossom Education Through Talent Introduction Project (PXM₂007-014207-04453) and Top-notch Talent Pro- ject (PXM₂007-014207-021717) of Beijing Municipal Educa- tion Commission. We thank Guangjun Lei (Azuki Bean Research Group, College of Plant Science and Technology, Bejing Univer- sity of Agriculture) for material planting and the experiment. This paper is for commemorating respectful professor Wenlin Jin.

References
Cai G.H., Yan W.C., and Cao X., 1992, Genetic analyrds of early ripening effects induced colchicine in rice, Acta Agriculture Zhejiangensis, 4(3): 103-107

Caroline C.M., Antonio O.C., and Femando C.F., 2003, Changes allele frequencies in colchicines-treated ryegrass populations assessed with RAPA marker, R. Bras. Agrociència, 9(2): 107-112

Liu Z., and Guan S.Q., 2006, Induction of the early-mature mutant of peanut through the treatment of colchicines, Journal of Anhui Agri. Sci., 34(21): 5449-5450

Pu S.J., Li J.Y., Zhang J.J., Zhao B., and Jin W.L., 2005, Preliminary exploration to induce variations by applying colchicine to adzuki bean, Journal of Beijing Agricultural College, 20(4): 5-8

Tiwarl S.P., and Khanorkar S.M., 1984, Colchlcine induced true breeding miniature mutant in groundnut, Current Science, 53(23): 1262-1263

Wang F.B., Fu J.F., and Dong L.F., 2009, Inducing autotetraploid pea with colchicine and DMSO, Journal of Nuclear Agricultural Sciences, 23(2): 203-208

Wang J.M., 2008, Primary report on the experimental study of inducing white radish with colchicines, Journal of Anhui Agri. Sci., 36(3): 985-987

Yang M.C., Cong B., Zhao J.H., and Zhang P.F., 1997, The ultra-structural changes induced by colchicines in root meristem-atic cells of Glycine soja, Journal of Wuhan Botanical Research, 15(1): 15-18

Yu Y.H., 2005, Studys on germplasm resources of Vigna angularis, Dissertation for Ph.D., Sichuan Agricultural University, Supervisor: Sun Y.S., pp.1-3

Zhang Q.Y., Ye D.S., Zhang S.N., Jin M.Y., and Zhang Y.C., 2000, Induction effects of colchicine on agronomic characters of two-rowed barley, Fujian Academy of Agricultural Sciences, 15(2): 8-12

Zhang Q.Y., Zhang S.N., and Zhuang B.H., 1996, Preliminary studies on breeding barley by colchicine, Fujian Academy of Agricultural Sciences, 1(12): 38-41