Introduction
Starch is the major reserve of plants and the major source of food calories for humans, as well as an important raw material for the food and processing industries (James et al., Reference James, Denyerz and Myers2003). The major component of starch granules is amylopectin, which forms partially crystalline structures, while amylose constitutes the amorphous portion of the granule (Hizukuri, Reference Hizukuri and Eliasson1996; Lemke et al., Reference Lemke, Burghammer, Flot, Rössle and Riekel2004). The different molecular features of starch polymers (i.e. chain length, frequency of branching, abundance of amylose, etc.) influence both the morphology of the granule and the technological properties of starch as a raw material or foodstuff (Jobling, Reference Jobling2004). The European starch market is substantial and the interest of both the scientific community and industry in starch biosynthesis and technology is strong. Although the genetic and physiological bases of starch biosynthesis in plants are well known, the regulatory machinery controlling the formation of the complex and ordered structure of the starch granule is still not fully understood. Mechanisms of post-translational regulation are likely to play a major role in starch metabolism (Michalska et al., Reference Michalska, Zauber, Buchanan, Cejudo and Geigenberger2009; Valerio et al., Reference Valerio, Costa, Marri, Issakidis-Bourguet, Pupillo, Trost and Sparla2011; Zeeman et al., Reference Zeeman, Kossmann and Smith2010). Mutants for biosynthetic or regulatory genes of starch metabolism often produce starch granules with abnormal morphological and molecular features (Sehnke et al., Reference Sehnke, Chung, Wu and Ferl2001; Asano et al., Reference Asano, Kunieda, Omura, Ibe, Kawasaki, Takano, Sato, Furuhashi, Mujin, Takaiwa, Wu, Tada, Satozawa, Sakamoto and Shimada2002).
We describe the utilization of TILLMore (http://www.distagenomics.unibo.it/TILLMore/), a barley targeting-induced local lesions in genomes (TILLING) resource (Talamè et al., Reference Talamè, Bovina, Sanguineti, Tuberosa, Lundqvist and Salvi2008) to identify new alleles involved in starch biosynthesis and degradation in seeds. TILLING has already been successfully applied to identify starch mutants in wheat (Slade et al., Reference Slade, Fuerstenberg, Loeffler, Steine and Facciotti2005). The long-term goal of this research is the identification of barley mutants with starch granules of peculiar morphological, structural and molecular features eventually leading to novel technological properties.
Materials and methods
For the TILLING screening, five genes involved in starch metabolism were selected based also on the genomic sequence availability in barley cv. ‘Morex’, provided by Dr. Edward Schiefelbein (University of Minnesota, St. Paul, MN, USA). The genes chosen for the analysis are: β-amylase 1 (HvBMY1 accession no. EF175470), granule-bound starch synthase I (HvGBSSI accession no. AB089162), limit dextrinase 1 (HvLDA1 accession no. AF122050), starch synthase I (HvSSI accession no. AF234163) and starch synthase II (HvSSII accession no. AY133250). Primers were designed with Codons Optimized to Discover Deleterious LEsions (CODDLE; http://www.proweb.org/coddle), a tool facilitating the selection of gene regions for TILLING purposes, and Primer3 (http://frodo.wi.mit.edu/cgi-bin/primer3/primer3_www.cgi). The CODDLE program is able to identify regions where point mutations are most likely to result in deleterious effects on the gene's function (Till et al., Reference Till, Reynolds, Greene, Codomo, Enns, Johnson, Burtner, Odden, Young, Taylor, Henikoff, Comai and Henikoff2003). The following primer sequences were used for PCR and sequencing of the population and the putative mutants:
HvBMY1_For, TTTGCCTTCCGGGAGACCATGT;
HvBMY1_Rev, CGCGTTTTCGGATGCCACATTT;
HvGBSSI_For, GAGCACCCAGCCACCCACACA;
HvGBSSI_Rev, CTGCAGCATACGCCCAGACCA;
HvLDA1_For, CTCGTGTGCAGCTGACGGGAAA;
HvLDA1_Rev, GTGCCATCGTGGGCGCTGTAAT;
HvSSI_For, TGTCGCGTTCCCCATTCTGATA;
HvSSI_Rev, TGGCATGGCTACAGTTCACCAAGC;
HvSSII_For, CCGATTCGATGTATGCCGGCAAT;
HvSSII_Rev, CCAGATCGGAATCAGCGTCTCA.
The TILLING analyses were implemented using the procedure described by McCallum et al. (Reference McCallum, Comai, Greene and Henikoff2000); DNA samples of eight M3 lines were pooled and subjected to gene-specific PCR amplification using properly designed labelled-primers (MWG-Biotech). The PCR reaction and cycling were performed as described in Colbert et al. (Reference Colbert, Till, Tompa, Reynolds, Steine, Yeung, McCallum, Comai and Henikoff2001). The PCR products were then digested with a commercial endonuclease, the Surveyor® Mutation Detection Kit (Transgenomics, Omaha, NE, USA), according to the manufacturer's directions. The digested PCR products were analyzed using a detection method based on denaturing electrophoretic gels (LI-COR-4200; LI-COR Biosciences; Lincoln, NE, USA). The final validation of the results was performed by sequencing using an Applied Biosystems' 377 DNA Analyzer (Applied Biosystems, Forster City, USA). Finally, the sequences were analyzed with the PARSESNP (Taylor and Greene, Reference Taylor and Greene2003) and sorting intolerant from tolerant (SIFT) (Ng and Henikoff, Reference Ng and Henikoff2003) programs.
Results and discussion
The molecular screening was achieved on five genes involved in starch metabolism, using a cell-based heteroduplex assay, coupled with gel electrophoresis on DNA sequencers. A total number of 4906 DNA samples from individual M3 plants were screened. The analyses identified an allelic series for each of the genes examined with a total number of 29 mutations and an average of c. five mutations/gene (Table 1). The estimated mutation density was of one mutation/520 kb screened, which compares well with what was previously reported by Talamè et al. (Reference Talamè, Bovina, Sanguineti, Tuberosa, Lundqvist and Salvi2008) on the same collection. The value of the mutation density was computed by dividing the total number of identified mutations by the number of base pairs screened and corrected, considering the effective screened window. In fact, a limitation of the TILLING procedure is that mutations can escape identification when present in the terminal 80 bp of both ends of the amplicon as a result of PCR priming and electrophoresis artifacts. In our case, a correction on the effective screening window was applied by subtracting 160 bp from the length of each amplicon (Greene et al., Reference Greene, Codomo, Taylor, Henikoff, Till, Reynolds, Enns, Burtner, Johnson, Odden, Comai and Henikoff2003).
Details of the mutants identified in the five genes analyzed
jun, junction.
a,b PSSM and SIFT are values used to rank the missense mutations on the basis of their probability of affecting protein function. PSSM and SIFT can be calculated only if the missense mutation occurs in a conserved domain of the protein. Mutations are considered to be deleterious for PSSM values above 10 and SIFT values below 0.10.
c Stop codon.
Almost all the mutations detected were G/C to A/T transitions. Since a previous study proposed that NaN3 causes mutations of transition type (Olsen et al., Reference Olsen, Wang and von Wettstein1993) and because almost all of our mutations were G/C to A/T transitions, the possibility that the polymorphisms identified in TILLMore are naturally occurring as a result of seed contamination of our starting ‘Morex’ seed stock can be ruled out.
Among the 29 alleles, 13 silent mutations occurred in non-coding regions or affected the third base of a codon which does not change the aminoacid encoded by that codon; 12 mutations were classified as missense alleles, causing changes in one of the aminoacids of the protein. In four cases, non-sense alleles (two truncation mutations and two splice junction mutations) were identified. All the non-sense mutations occurred in starch synthase I and II, two genes with a crucial role in the elongation of the amylopectin chains. Severe mutations in these genes are expected to drastically reduce the content of amylopectin, hence conferring a clear phenotype (Umemoto et al., Reference Umemoto, Yano, Satoh, Shomura and Nakamura2002; Fujita et al., Reference Fujita, Yoshida, Asakura, Ohdan, Miyao, Hirochika and Nakamura2006; Sestili et al., Reference Sestili, Botticella, Bedo, Phillips and Lafiandra2009). As to the missense mutations, identified for all the genes analyzed, bioinformatic tools were applied to estimate the impact of mutations on protein function. In particular, PARSESNP and SIFT programs were utilized to identify the mutations that more likely will have a deleterious effect on protein function. In our case, the mutations GBSSI 1090 and BMY1 2253 showed position specific scoring matrix (PSSM) values of 20.5 and 10.2, respectively (mutations are considered to be deleterious for PSSM values above 10). The application of the SIFT algorithm predicted a possible deleterious effect for the mutations GBSSI 1090 (SIFT 0.0) and LDA1 1020 (SIFT 0.02). Mutations are predicted to be deleterious for SIFT values below 0.05 or even below 0.10. Since all other missense mutations were predicted to be located outside conserved domains, PSSM and SIFT values could not be calculated.
In conclusion, we detected at least one interesting allele for all the five genes analyzed in our study. These findings provide valuable genetic materials for studies on the regulation of starch biosynthesis in barley and for applications in mutation breeding.
Acknowledgements
The financial support of the University of Bologna (Strategic project ‘Starchitecture’) is gratefully acknowledged.
