Ds INSERTION LINES VALUABLE FOR RICE BREEDING
Shu-Ye Jiang and Srinivasan Ramachandran
March, 2008

Rice is a staple food for more than half of the world’s population. With the population continuously increasing and cultivable land decreasing, the quantity of rice produced may not be sufficient. However, rice yield is currently decreasing due to abiotic stresses, including drought, salinity, and cold, which may affect plants’ growth and productivity—in fact, more than 50% of rice grain production may be lost to abiotic stress. Therefore, future rice varieties should not only produce more rice grain under normal growth conditions but also minimize yield loss under various stressed growth conditions.

Rice breeding selects ideal phenotypes from a population produced from various germplasm resources, including natural or artificial mutants or their sexual hybridization. Besides natural mutation, various artificial mutations have significantly contributed to rice breeding, including physical and chemical mutation as well as tissue culture-mediated somaclonal variations. On the other hand, with the completion of rice genome sequences and the release of rice full-length cDNA data, insertion mutagenesis with the maize transposon Dissociation (Ds) has been successfully used for rice functional genomics, contributing significantly to the collection of rice germplasm resources. In this report we briefly discuss the feasibility and potential of Ds insertion mutagenesis as a tool to produce desirable traits for rice breeding.

Large collection of Ds insertion lines developed
We have developed approximately 20,000 Ds insertion lines in which Ds elements were randomly and independently inserted into different rice genes or inter-gene regions¹. From them, more than 3,000 Ds flanking sequences have been obtained using a specific PCR method followed by DNA sequencing. Thus, various mutations may be generated by knocking out rice genes or other genome regions, thereby providing a selectable resource for various phenotypes. In addition to our Ds lines collection, many other groups have also significantly contributed to the large collection of Ds insertion lines. Currently, at least 155,200 transposon insertion lines have been generated by the international rice research community (Table 1). More than 30,000 Ds flanking sequences have been obtained from these insertion lines. Because the rice genome contains less than 50,000 genes encoding various proteins, such an insertion population may be large enough to find a knockout mutant for each predicted gene, thus providing a large collection and breeding selection resource of Ds insertion lines.

Phenotypic variations among Ds insertion lines
Our entire collection of Ds insertion lines was grown under both greenhouse and field conditions. Phenotypes were observed and comparisons made between WT and mutants based on a standard evaluation system for rice available from the International Rice Research Institute (IRRI) resource (http://www.knowledgebank.irri.org/RP/morph/morphology.htm). Evaluations of approximately 20,000 independent lines revealed various visible phenotypic differences under normal growth conditions. These variations included differences in grain yield, plant height, growth duration, tiller numbers, plant stature (bent or erect culms), fertility, and so on². These results suggest that Ds insertion lines have the potential to be used as breeding germplasm to develop new rice varieties.

Ds insertion lines as germplasm resources for developing high yield rice varieties
Among available phenotypes, we were interested in lines with variations in grain yield. We first weighed and calculated the grain yield for all Ds insertion lines growing under greenhouse conditions. The Ds insertion rice lines generated variations in grain yield, ranging from 0 g per plant (sterile) to 90.5 g per plant, with a yield curve of near normal distribution (Fig. 1A). Based on this preliminary screen, we identified approximately 650 putative lines with at least 20% greater seed yield compared to WT plants. Among them, 35 lines exhibited 50% greater grain yield. One line is shown in Fig. 1 B-F. This line exhibited strong growth, a bigger size, and a more vigorous root system. As a result, it produced more seeds in each panicle and thus more grain per plant. These data suggested that Ds insertion mutagenesis can be utilized as an efficient tool to produce variations including higher yield lines for breeding.

Figure 1. Grain yield variations in Ds insertion lines and the performance of one higher yield line. (A) A distribution curve of grain yield based on the investigation of around 20,000 Ds insertion lines. X axis indicates the grain yield (gram) per plant; Y axis indicates the percentage of Ds lines with corresponding grain yield. (B) - (F) indicates stronger growth: (B and C) stronger stems; (D) vigorous roots; (E) bigger size of leaves; and (F) more seeds of a Ds insertion line. In (B) to (E), the image on the left and right represents WT and Ds line, respectively.

On the other hand, sterile rice lines were frequently observed in our Ds insertion lines. Several such lines were investigated further. Male sterile^3,4 lines provide resources from which the cytoplasmic male sterile (CMS) or photoperiod (temperature)-sensitive male sterile rice lines are derived. These two kinds of male sterile rice lines form the basis for three-line or two-line hybrid rice combinations for commercial release. One such line—a photoperiod sensitive male sterile rice plant—was further characterized. It exhibited male sterility under short day length conditions, and the sterility was recovered under long day length conditions. Further investigation showed that photoperiod sensitive localization of a myosin protein controlled the fertility transformation³.

Developing varieties with abiotic stress tolerance
Ds insertion lines were subjected to various abiotic stresses, including drought, high salinity, and cold conditions, to obtain valuable stress-responsive lines. To screen for drought-responsive lines, approximately 16,000 three week-old seedlings (WT and Ds lines) were treated with 3% or 10% PEG (polyethylene glycol, 6000) to obtain hyper-sensitive or tolerant lines, respectively. In total, we selected 84 of the best lines, including 61 sensitive and 23 tolerant lines. Among 307 of the moderate lines, 194 lines were sensitive and 113 lines were tolerant². More than 100 candidate lines were subjected to drought stress naturally under field and greenhouse conditions. The results validated the use of PEG screens to mimic drought conditions.

To screen salinity-responsive lines, 7,000 two-week-old seedlings were subjected to 50 mM NaCl to obtain hyper-sensitive lines, and to 200 mM NaCl to obtain tolerant lines. This screening produced 54 candidates, including 40 sensitive and 14 tolerant lines². One resistant line is shown in Fig. 2A.

Screening for cold-responsive lines was performed under natural winter conditions in southern China, with temperatures ranging between 10 – 20 °C. Out of 13,000 Ds insertion lines subjected to cold screens in two winter seasons, 470 cold-sensitive or -tolerant lines were obtained². By summarizing results from drought, high salinity, and cold screenings, we found that some Ds lines exhibited resistance/sensitivity to two or three stress conditions (Fig. 2B), which may be the best candidates for developing multiple-resistant rice varieties.

Figure 2. Screening for stress-responsive Ds insertion lines. (LEFT) An example of a salinity-resistant Ds line. After germination in MS media, both wild type and Ds lines were transferred for two weeks into 200 mM NaCl-containing MS media: Left, wild type; right, Ds line. (RIGHT) A summary of screening Ds lines under drought, high salinity, and cold growth conditions. We obtained 391 drought, 54 high salinity, and 470 cold-responsive Ds lines. Among them, two lines showed responsiveness to three stresses; one line was responsive to both drought and high salinity; two lines were responsive to both high salinity and cold; and 15 lines were responsive to both drought and cold conditions.

Improved breeding efficiency using molecular marker-assisted selection
Traditional breeding programs generally require seven to nine generations of conventional backcrosses to transfer an elite agronomic trait into a desirable parent. By contrast, molecular marker-assisted selection can capture desirable characters in new rice varieties within shorter periods of time. Ds insertion lines can be used as a valuable germplasm resource for rice breeding. More than 97% of Ds lines contain only a single copy of the Ds insertion; thus, the Ds transposon can be used as a molecular marker to capture desirable agronomic characters caused by the corresponding Ds insertion into a gene.

Ds insertion lines used to develop non-transgenic or marker-free rice
In our collection the Ds element is immobile. However, the Ds transposon can be released or remobilized into other regions in the presence of a transposase source by crossing with an Ac transposase-containing transgenic plant⁵. Subsequent to the remobilization of the Ds element, a footprint will usually be left behind. If a mutant has a Ds insertion into a coding region of a gene, we can develop new varieties without transgenic sequences. The footprint may cause a frame-shift mutation, making the encoded protein non-functional. Thus, the mutant phenotype can be retained by the footprint without additional foreign DNA sequences². We have successfully used this method to generate two footprint-containing plants. These plants, with no Ds elements, retained the mutant phenotype.

If the Ds element is not transposed into the coding region of a gene, new varieties can be developed by either over- or under-expressing this gene using a marker free method. Endogenous rice promoters (such as actin) can be utilized for this purpose. We have introduced two kinds of constructs to develop such rice varieties. As a result, only minimal T-DNA or maize transposon borders around 200 base pairs were retained in the final genetically engineered rice plants². These sequences do not encode a protein, so the product should be safe for commercial release.

In summary, we have generated a Ds insertion mutant population. We subjected these lines to phenotypic and abiotic stress screens. Some interesting lines have been obtained with higher yield, male sterility, or resistance/sensitivity to various abiotic stresses. Our results suggest that rice could be improved not only by introducing foreign genes but also by knocking out its endogenous genes. These results might provide a new method for rice breeders to further improve rice varieties.

Shu-Ye Jiang and Srinivasan Ramachandran*
Rice Functional Genomics Group, Temasek Life Sciences Laboratory
1 Research Link, National University of Singapore, Singapore 117604
*sri@tll.org.sg