当前位置：文档库 › MutMap+accelerates+breeding+of+a+salt-to...

MutMap+accelerates+breeding+of+a+salt-to...

two chromosome substitution lines (CSSLs), SL502 and SL503, harboring chromosomal segments of Nona Bokra in the genetic

background of Koshihikari, a cultivar closely related to Hitomebore 10,11. We compared the growth of hst1, SL502 and SL503 at 0.75% NaCl (Supplementary Fig. 2). The 18 and 43% improved growth of hst1 as compared to SL503 and SL502, respectively, showed that this was a good starting point for breeding a salt-tolerant rice cultivar.

For rapid identification of the mutation conferring salinity tolerance in hst1, we used MutMap, a method based on whole-genome resequencing of bulked DNA of F 2 segregants 1. The hst1 line was crossed to WT Hitomebore to generate F 1 progeny, and F 2 progeny were derived from self-pollination of the F 1 progeny. Two-week-old F 2 seedlings were treated with water containing 0.75% NaCl. The progeny segregated in a 133:54 ratio for salinity-susceptible and salinity-tolerant phenotypes, respectively, conforming to a 3:1 segregation ratio (chi-squared test: χ2 = 2.2 × 10-16, nonsignificant) and thereby indicating that the salinity tolerance of hst1 is conferred by a single recessive mutation. We combined DNA from 20 F 2 progeny that had the salinity-tolerance phenotype and applied whole-genome resequencing using an Illumina GAIIx DNA sequencer. We obtained a total of 7.34 Gbp of short (75-bp) reads (Supplementary Table 1) that were aligned to the Hitomebore reference sequence

(DDBJ Sequence Read Archive DRA000927), resulting in the identification of 1,005 SNP positions. For each SNP position, the value of SNP-index (the ratio of short reads harboring SNPs different from the reference 1) was obtained and a graph relating SNP positions and SNP-index was generated for all 12 rice chromosomes (Fig. 2a , Supplementary Fig. 3). The causative SNP should be shared by all the mutant F2 plants and therefore have a SNP-index = 1, whereas SNPs unrelated to the mutant phenotype should segregate in a 1:1 ratio among the F 2 progeny, resulting in a SNP-index of ~0.5. MutMap applied to hst1

To the Editor:

Following the 2011 earthquake and tsunami that affected Japan, >20,000 ha of rice paddy field was inundated with seawater, resulting in salt contamination of the land. As local rice landraces are not tolerant of high salt concentrations, we set out to develop a salt-tolerant rice cultivar. We screened 6,000 ethyl methanesulfonate (EMS) mutant lines of a local elite cultivar, ‘Hitomebore’, and identified a salt-tolerant mutant that we

name hitomebore salt tolerant 1 (hst1). In this Correspondence, we report how we used our MutMap method 1 to rapidly identify a loss-of-function mutation responsible for the salt tolerance of hst1 rice . The salt-tolerant hst1 mutant was used to breed a salt-tolerant variety named ‘Kaijin’, which differs from Hitomebore by only 201 single-nucleotide polymorphisms (SNPs). Field trials showed that it has the same growth and yield performance as the parental line under normal growth conditions. Notably, production of this salt-tolerant mutant line ready for delivery to farmers took only two years using our approach.

Although soluble salts, such as nitrates and potassium salts, are common components of soil and essential plant nutrients, their accumulation above specific threshold

concentrations can substantially affect plant growth. There is considerable variation among plants with respect to their tolerance of salinity, and rice is considered the

most sensitive of all the cereals 2. Yields of paddy rice start to decline at salinity levels >3 dS m –1 (measured by the electrical conductivity of the extract, EC e ), beyond which a 12% reduction in yield is expected for every 1 dS m –1 increase in EC e 3. Soil salinity affects >6% of world’s total land area, causing yield losses as a result of both osmotic and ionic stresses to crop plants 2. Soil salinization due to the flooding of agricultural lands by seawater has become an additional concern since the 2004 Indian Ocean tsunami 4. In 2011, Japan was hit by the Great Tohoku Earthquake, which triggered

a devastating tsunami, altogether claiming the lives of more than 15,000 people. The tsunami extended more than 5 km inland on the Sendai Plain of Miyagi Prefecture, one of the main rice-production regions in Japan 5. An environmental impact assessment study conducted in the same area over a period of 2–7 months after the tsunami revealed wide spatial variation in the salinity level of ponded water, with EC e ranging from 0.31 to 68.2 mS cm –1 (ref. 6). Although salt concentration gradually decreased, it was too high for rice production to resume in October, 2011.

To restore rice production in tsunami-affected areas of the Tohoku region of Japan, we set out to develop and deliver a salt-tolerant rice cultivar from a line suited to local agronomic conditions. First, we carried out a genetic screen for salt tolerance using seeds pooled from approximately 6,000 independent EMS-mutagenized lines of Hitomebore 7, (Supplementary Fig. 1). We identified a mutant that survived with 1.5% NaCl supplied to the soil with irrigation water for 7 days, which we designated

hitomebore salt tolerant 1 (hst1). Seeds from a self-pollinated hst1 plant were used to further test the performance of hst1 at different NaCl concentrations (Fig. 1a ,b ). The hst1 mutant grew better than wild-type (WT) Hitomebore plants at both 0.375% and 0.75% NaCl concentrations, as measured after 14 days of treatment. The 0.375% NaCl treatment caused reductions of 38.4% and 2.9% in the fresh weight of WT and hst1 plants,

respectively. At 0.75% NaCl, WT plants dried out, with a 61.5% reduction in fresh weight, whereas hst1 plants remained green with only a 13.2% reduction in fresh weight compared with hst1 plants that received fresh water (Fig. 1b ).

Previously, the rice SHOOT K +

CONCENTRATION 1 (SKC1) gene, encoding a Na + transporter, was identified as the main quantitative trait locus (QTL) conferring salt tolerance in the indica cultivar Nona Bokra 8,9. This QTL has been used to develop

MutMap accelerates breeding of a salt-tolerant rice cultivar

C O R R E S P O N

E N C E

n p g

? 2015 N a t u r e A m e r i c a , I n c . A l l r i g h t s r e s e r v e d .

is P < 1.2 × 10-6 (Fisher’s exact test). We also carried out an allelism test by crossing hst1 to an NE0017 line homozygous for the OsRR22 Tos17 insertion and tolerant of salt stress (Supplementary Fig. 5). F 1 progeny harboring both the hst1 point mutation and NE0017 Tos17 insertion in heterozygous states were all tolerant of salinity, further confirming that the salinity tolerance of hst1 and NE0017 is caused by the mutations in the same OsRR22 gene (Supplementary Fig. 5).OsRR22 encodes a 696-amino-acid B-type response regulator protein with a predicted N-terminal receiver domain and a C-terminal Myb-like DNA-binding domain, which probably functions as a

transcription factor 2,15. OsRR22 has recently been implicated in cytokinin signaling and metabolism 2. This study is the first, to our knowledge, to find a role for OsRR22 in responding to salt stress.

Histological observations using WT

plants transformed with the b -glucuronidase (GUS) reporter gene under the control of the OsRR22 promoter revealed that OsRR22 is mainly expressed in roots and stem and leaf-sheath base (Supplementary Fig. 6). In roots, GUS expression was found in the endodermis and pericycle (Supplementary Fig. 6g–i ), suggesting that OsRR22 might be involved in regulating genes involved in osmotic responses and/or ion transport between parenchyma cells and vascular tissue cells of roots. To address how the loss of function of OsRR22 affects downstream gene expression, we carried out RNA-seq-based transcriptome analysis using libraries constructed from WT and hst1 seedlings grown under both normal and salt-treated conditions (Supplementary Fig. 7). We identified 21 genes that

showed at least twofold differences in their expression levels between WT and hst1 in response to salt stress (Supplementary Fig. 7, Supplementary Table 3). The 16 hst1-upregulated genes included OsHKT1;1/Os04g0607500, coding for a member of the high-affinity K + transporter family shown to function as a Na + transporter 16. However, its contribution to salt tolerance has not yet been characterized. Remarkably, under normal conditions, RNA-seq revealed that only 223 genes showed more than twofold expression differences between WT and hst1, out of 23,003 genes compared

(Supplementary Table 4), suggesting that the hst1 mutation affects the expression of only a small subset of genes.

To evaluate the practical utility of hst1 for rice breeding, we carried out a field trial in salt-treated paddy fields during the 2013 growing season. Four-week-old WT

F 2 progeny revealed a cluster of SNPs with high SNP-index values, showing statistically significant deviations from SNP-index = 0.5 (P < 0.01), in the region between 2.76 Mb and 8.57 Mb on chromosome 6. Of the two SNPs with SNP index = 1 identified in the candidate region, a SNP at nucleotide position 4,138,223 corresponded to the third exon of the gene Os06g0183100, predicted to encode a B-type response regulator designated as OsRR22 (ref. 12; Fig . 2b ,c and Supplementary Table 2). The SNP represented a nonsense mutation of a tryptophan (TGG) codon to a stop codon (TAG) at residue 222 in OsRR22. Another SNP at position 8,147,386 was in a noncoding region (Supplementary Table 2), suggesting that the mutation identified in OsRR22 is responsible for the salinity-tolerance phenotype of hst1.

We searched the database of rice Tos17 transposon insertion mutant lines 13,14 and

identified two independent lines, NE0017 and NF6804, with Tos17 insertions in the third and fifth exons of OsRR22, respectively (Fig. 2b ). Twenty-seven individuals obtained from self-pollination of a heterozygous NF6804 line and subjected to salinity tests using 0.375% NaCl for two weeks segregated 20:7 susceptible:tolerant, conforming to the 3:1 ratio (χ2 = 0.012, nonsignificant;

Fig. 2d ). We confirmed by PCR amplification that all seven salinity-tolerant individuals had a Tos17 insertion of OsRR22 in the

homozygous state, whereas the 20 susceptible individuals either were heterozygous for the insertion or had no insertion in OsRR22 (Supplementary Fig. 4). This perfect co-segregation of salinity tolerance and Tos17 insertion in homozygous state indicated that the loss-of-function mutation in

OsRR22 is responsible for the hst1 salinity-tolerance phenotype. The probability of

observing such co-segregation by chance

NaCI concentration

hst1

F r e s h w e i g h t (g )

0.51.01.52.02.50.375%0.75% 1.5%

0.375%0.75% 1.5%

NaCI concentration

Figure 1 hst1 is tolerant of salt stress. (a ) Phenotypes of four-week-old WT and hst1 plants grown at different NaCl concentrations. Two-week-old plants were treated (one irrigation, from week 2 to week 4) with different concentrations of NaCl, and phenotypic evaluation was done two weeks after treatment. Ten plants were tested in each treatment. (b ) Comparison of shoot fresh weight of WT and hst1 plants shown in a . Values represent weight of ten plants in each treatment. Average values of three replicates and standard deviations are shown. Asterisks indicate significant difference (Student’s t -test, **P < 0.01; normality of variables and equality of variances were validated by Kolmogorov-Smirnov test and F -test, respectively).

C O R R ES P O N

D EN C

E n p g

? 2015 N a t u r e A m e r i c a , I n c . A l l r i g h t s r e s e r v e d .

the mutant to the widely grown elite cultivar Hitomebore. After two backcrossing events followed by two consecutive selfings (BC1F3), with confirmation of inheritance of recessive hst1 allele at each step by Sanger sequencing of the HST1 locus (Supplementary Fig. 5) we developed a line we named Kaijin (Neptune in Japanese). Kaijin had the same level of salt tolerance as hst1 (Fig. 3c and Supplementary Fig. 12), and whole-genome resequencing revealed that Kaijin differed from Hitomebore WT by only 201 SNPs, which is a significant reduction from the 1,088 homozygous SNPs in the hst1 EMS mutant (Supplementary Fig. 13 and Supplementary Table 5).

Moreover, only 3 SNPs out of the 201 SNPs found in Kaijin cause nonsynonymous

changes (Supplementary Table 6). Kaijin will be registered in 2015 before its formal release to farmers in tsunami-affected areas.It is widely recognized that excessive irrigation with water containing dissolved salts eventually results in salt accumulation and that climate change could cause the land area currently in agricultural use affected by salinity to exceed 800 million ha 2. In this Correspondence, we identified a rice mutant hst1 with enhanced tolerance of salt stress

and hst1 seedlings were transplanted and grown in adjacent plots in the experimental fields of Iwate Agricultural Research Center, Kitakami, following the local farmers’ cultivation practice. Plants in the control plots were irrigated normally with fresh groundwater, while a high-salinity plot was created by periodically irrigating with water containing NaCl (Fig. 3a ,b ). Both plots were kept at similar soil moisture contents by maintaining the water at ground level. The soil salinity level in the NaCl-treated plots was monitored by measuring the electrical conductivity (EC) of soil resuspended in water to a 1:5 (soil:water) ratio

(Supplementary Fig. 8). Before the middle of July, when the EC values were still <1 mS/cm, no phenotypic differences were observed between WT and hst1 plants. With gradual increases in EC value, WT plants started to show symptoms of leaf rolling, and the major part of shoots eventually dried out toward the end of August when the EC value exceeded 6 mS cm –1. In contrast, all hst1 plants in NaCl-treated plots survived till the end of the trial in the middle of September (Fig. 3a ). Grain yield per hst1plant in salt-treated plots was more than double that of WT plants (Fig. 3b ). Salinity tolerance of hst1 in the field was

further confirmed in a trial conducted on a test plot located close to Rikuzen Takata, a coastal city heavily hit by the 2011 tsunami (Supplementary Fig. 9).

During the field test, we observed no

substantive differences in grain yield between WT and hst1 in the control plots (Fig. 3a ,b ). We further checked whether hst1 has a yield penalty under normal salt-free rice cultivation conditions. No significant

differences were observed between WT and hst1 plants for traits including plant height, grain yield, panicle number and panicle size (Supplementary Fig. 10a ,b ). We additionally compared the quality of WT and hst1 grains. Starch amylose content is a vital quality component of cereal grains including rice 17. Although minor differences were observed in grain length and width (Supplementary Fig. 10c ,d ), there was no variation between WT and hst1 grains with regard to their starch amylose content (Supplementary Fig. 11a ). We confirmed hst1 eating quality using a blind sensory test involving 23 panelists, who reported no difference in eating quality between WT and hst1 (Supplementary Fig. 11b ,c ).

To utilize the salinity-tolerance trait of hst1 in our breeding program, we backcrossed Figure 2 Identification of the hst1 mutation by MutMap. (a ) SNP-index plot of the rice chromosome 6 generated by MutMap analysis, showing a genomic region with the highest SNP-index peak harboring the candidate mutation. DNA bulked from 20 salt-tolerant F 2 progeny obtained from the cross between WT Hitomebore and hst1 was used for sequencing and MutMap analysis. Blue dots correspond to SNPs identified between hst1 and Hitomebore WT genomes. The red line represents the sliding window average values of SNP indices of 4-Mb intervals with a 10-kb increment. The yellow line shows the 99% confidence limit of SNP index value under the null hypothesis of SNP index = 0.5.

(b ) Structure of the OsRR22/HST1 gene. Open and gray boxes represent untranslated regions (UTRs) and exons, respectively, while lines denote introns. Locations of the two Tos17

insertions found in lines NE0017 and NE6804, respectively, are indicated above the gene.

(c ) Confirmation of the hst1 mutation in OsRR22 by Sanger sequencing. A red box indicates the G-to-A transition. (d ) Co-segregation of the

salinity-tolerance phenotype and Tos17 insertion of NF6804. Twenty-seven individuals obtained from selfing of an NF6804 line and subjected to salt stress segregated 7:20 for the tolerant and susceptible phenotypes, respectively. Nipponbare was used as a control. Genotyping by PCR revealed that all seven salt-tolerant individuals were homozygous (Hom) for the Tos17 insertion, whereas the susceptible individuals were

either heterozygous for (Het; 15 individuals) or entirely lacking the insertion in OsRR22 (Null; 5

individuals). Scale bar, 5 cm.

Trp (TGG)

Genotype in fth exon of OsRR22

Chromosome position (Mb)S N P i n d e x Chromosome 6

1020300.20.40.60.81.0NE0017

NF6804

OsRR22

Progeny of NF6804 sel ng

Nipponbare Tolerant

Susceptible

70Hom

Tos17Tos17

Tos17

Het

Null

Stop (T A G)

hst1

C O R R ES P O N

D EN C E

n p g

? 2015 N a t u r e A m e r i c a , I n c . A l l r i g h t s r e s e r v e d .

Developing Kaijin by backcrossing hst1 to Hitomebore. The hst1 mutant was backcrossed twice to Hitomebore WT to generate BC1F1. From BC1F1 progeny, a line harboring hst1 in heterozygous state was identified by Sanger sequencing of the OsRR22 locus. The heterozygous line was self-pollinated to make the hst1 mutation homozygous (BC1F2). Finally, this line was self-pollinated to generate the BC1F3 line, Kaijin. Whole-genome resequencing and comparison of the genomes of the original hst1 mutant and Kaijin were done to assess the number of SNPs retained following the backcrosses. Both sequence data were aligned to the Hitomebore reference sequence using MutMap pipeline, and SNPs were scored and compared (Supplementary Fig. 13).Field trials for salinity tolerance. We compared the salinity tolerance of WT and hst1 plants in the paddy filed of

Iwate Agricultural Research Center using 2 m × 2 m plots isolated by polyvinyl chloride

and used MutMap to map the causative hst1 mutation to the OsRR22 gene. The resulting variety, Kaijin, is practically equivalent to Hitomebore except for the hst1 mutation but took only two years to breed, which is far quicker than for conventional rice breeding (~10 years).

Four years after the 2011 tsunami disaster, some paddy fields in the Sendai plain are still heavily salinized as a result of ground subsidence caused by the earthquake. We expect that the Kaijin variety will contribute significantly to the recovery efforts of Tohoku rice farmers. Application of MutMap as

exemplified here is restricted to mutant lines. We could also exploit the genetic diversity of landraces by genomic selection 18 and QTL-seq 19. This report demonstrates the power of genomics-based crop breeding methods for improving our ability to provide farmers with high-yielding cultivars in response to rapidly changing environmental conditions.METHODS

Mutant screening. The rice mutant lines used in this study were generated by EMS treatment of the japonica cultivar Hitomebore at the immature embryo stage as described by Rakshit and colleagues 7. For the genetic screening, seeds from 6,000 M4 lines were bulked, and 100 g seeds were sown in each 30 cm × 60 cm tray. Y oung seedlings were subjected to salinity treatment at the 2.5-leaf stage by covering the seedling trays with 50% GEX artificial seawater containing 1.5% NaCl (obtained from GEX, Japan, http://www.gex-fp.co.jp ).

Generation of F 2 progeny and whole-genome sequencing. To generate the F 2 progeny used for bulk sequencing, hst1 was crossed to Hitomebore WT and the resulting F 1 progeny were allowed to self-pollinate and produce F 2 seeds. For whole-genome sequencing, DNA samples were extracted from rice leaves with DNeasy Plant Mini Kit (Qiagen) as previously described 1. The bulked DNA used for MutMap analysis was prepared by mixing DNA from 20 mutant F 2 individuals in an equal ratio. The Illumina sequencing libraries were prepared with TruSeq DNA LT Sample Prep Kit (Illumina) and were used for sequencing either by Illumina GAIIx (76 cycles) or Illumina HiSeq2500 (100 cycles) DNA sequencers.

MutMap analysis. The Hitomebore reference sequence was constructed by replacing nucleotides in Nipponbare with those of Hitomebore at the 124,968

SNP positions identified between the two cultivars by alignment of 12.25 Gb of Illumina Hitomebore short reads to the Nipponbare reference genome (build 5 genome sequence, http://rapdblegacy.dna.affrc.go.jp/download/index.html ) as previously described by Takagi et al.19.Illumina short reads generated from the bulked DNA sample were filtered by phred quality score and aligned to the Hitomebore reference sequence using BWA software 20. The alignment data were converted to SAM/BAM files using SAMtools 21, and low-quality SNPs were excluded with a Coval filter 22. Additionally, SNPs that were detected by self-alignment of the parental Hitomebore short reads to the Hitomebore reference sequence were excluded from the analysis. SNP-index was calculated at all SNP positions with Coval 22. All the steps were automatically processed using the MutMap_v1.2.2 pipeline (http://genome-e.ibrc.or.jp/home/bioinformatics-team/mutmap

Control WT WT

B C Control Control

Treated with 0.375% NaCI

hst1hst1

WT W T

K a i j i n

h s t 1

W T

K a i j i n

h s t 1

hst1

Salt treated

Figure 3 Field evaluation of hst1 and the new salt-tolerant variety Kaijin. (a ) A field view showing the phenotypes of control and salt-treated WT and hst1 plants. Each plot was replicated three times. (b ) Grain yield of WT and hst1 plants. Results were shown for plants from the control plot and the three replications of NaCl-treated plots (A, B and C). Values are mean ± s.d., n = 9. Asterisks indicate significant difference (Mann-Whitney U -test, **P < 0.01). (c ) The salinity-tolerance phenotype of 4-week-old WT, hst1 and Kaijin plants under control and salt-treated conditions. Kaijin was generated by backcrossing of hst1 to Hitomebore twice followed by two selfing events. Pictures were taken two weeks after NaCl treatment.

C O R R ES P O N

D EN C E

n p g

? 2015 N a t u r e A m e r i c a , I n c . A l l r i g h t s r e s e r v e d .

651–681 (2008).

3. Allen, R.G., Pereira, L.S., Raes, D. & Smith, M. Crop

evapotranspiration: guidelines for computing crop water requirements. FAO Irrigation and Drainage paper No. 56, FAO, Rome, p. 300 (1998) (https://www.wendangku.net/doc/482137791.html,/docrep/x0490e/x0490e00.htm ).4. Szczucin ′ski, W. et al. Environ. Geol. 49, 321–331

(2005).

5. Mori, N., Takahashi, T., Yasuda, T. & Yanagisawa, H.

Geophys. Res. Lett. doi:10.1029/2011GL049210 (27 September 2011).

6. Chagué-Goff, C. et al. Sediment. Geol. 282, 175–187

(2012).

7. Rakshit, S. et al. in The Handbook of Plant Mutation

Screening: Mining of Natural and Induced Alleles . (eds. Meksem, K. & Kahl, G.) 185-197 (Wiley-VCH, 2010).

8. Lin, H.X. et al. Theor. Appl. Genet. 108, 253–260

(2004).

9. Ren, Z.H. et al. Nat. Genet. 37, 1141–1146 (2005). 10. Takai, T. et al. Breed. Sci. 57, 257–261 (2007). 11. Yamamoto, T. et al. BMC Genomics 11, 267 (2010). 12. Tsai, Y.C. et al. Plant Physiol. 158, 1666–1684

(2012).

13. Hirochika, H., Sugimoto, K., Otsuki, Y., Tsugawa, H. &

Kanda, M. Proc. Natl. Acad. Sci. USA 93, 7783–7788 (1996).

14. Miyao, A. et al. Plant Cell 15, 1771–1780 (2003). 15. Argyros, R.D. et al. Plant Cell 20, 2102–2116 (2008). 16. Garciadeblás, B., Senn, M.E ., Banuelos, M.A. &

Rodríguez-Navarro, A. Plant J. 34, 788–801 (2003). 17. Olsen, K.M. et al. Genetics 173, 975–983 (2006). 18. Heffner, E.L., Sorrells, M.E. & Jannink, J.L. Crop Sci.

49, 1–12 (2009).

19. Takagi, H. et al. Plant J. 74, 174–183 (2013).

20. Li, H. & Durbin, R. Bioinformatics 25, 1754–1760

(2009).

21. Li, H. et al. Bioinformatics 25, 2078–2079 (2009). 22. Kosugi, S. et al. PLoS ONE 8, e75402 (2013).

23. Trapnell, C., Pachter, L. & Salzberg, S.L. Bioinformatics

25, 1105–1111 (2009).

sheets (Fig. 3a ). Plants in control plots were irrigated with fresh groundwater, whereas for salt treatment, plants were periodically supplied with NaCl-containing water. The soil salinity level of the NaCl-treated plots was monitored by measuring the EC value (Supplementary Fig. 8). In both plots, rice cultivation was carried out according to the locally adapted conventional methods: N:P:K fertilizer was applied at the rate of 60:60:60 kg ha –1, four-week-old seedlings were

transplanted at the spacing of 30 cm × 15 cm and rate of one seedling per hill, and water was supplied as required.

RNA-seq analysis. RNA was extracted from shoot samples using the RNeasy Plant Mini Kit (Qiagen Sciences, Germantown, Maryland, USA). For Illumina sequencing, 4 m g of RNA was used for preparation of libraries according to the protocol for the TruSeq RNA Sample Prep Kit v2 (Illumina, San Diego, California, USA). The libraries were applied to paired-end sequencing for 100 cycles on HiSeq2500 (Illumina). The sequence reads were filtered for quality in FASTAQ format, after which total reads per sample was standardized to 7,117,923 paired-end reads for all samples (DRA002623). The reads were aligned to the Hitomebore reference sequence previously developed for MutMap analysis by TopHat 23. The average depth covering coding sequences, as per the annotation provided in the rice annotation project database

(RAP-DB, http://rapdblegacy.dna.affrc.go.jp ), was employed as a metric for comparing gene expression levels between samples. Statistical significances of the observed differences (Student’s t -test, P < 0.05) were calculated from three replicate samples.

Note: Any Supplementary Information and Source Data files are available in the online version of the paper (doi:10.1038/nbt.3188).

ACKNOWLEDGMENTS

The authors thank H. Hirochika and A. Miyao

(National Institute of Agrobiological Sciences, Tsukuba, Japan) for providing the Tos17 insertion lines NE0017 and NF6804; the Rice Genome Research Center, Japan, for CSSL lines; T. Nakagawa (Shimane University, Japan) for providing the pGWB3 vector; and T. Endo (Miyagi Prefecture, Japan) for providing advice on field evaluation of rice salinity tolerance. R.T. was funded by the Science and Technology Research Promotion Program for Agriculture, Forestry, Fisheries and Food Industry, Japan, grant-in-aid for MEXT (Scientific Research on Innovative Areas 23113009) and JSPS KAKENHI (grant no. 24248004). S.K. was supported by the Gatsby Charitable Foundation, the European Research Council (ERC) and the UK Biotechnology and Biological Sciences Research Council (BBSRC).AUTHOR CONTRIBUTIONS

H.T., A.A., K.Y ., H.K. conceived the idea and carried out the screen, genetic analyses, MutMap and breeding.

A.U., H.Y ., H.U., C.M., S.N., S.K. (Kosugi), H.M., N.U. carried out genome sequencing and bioinformatic analyses. T.O. carried out the field analysis. K.O., E.K. carried out genetic transformation. S.K. (Kamoun), M.T. and R.T. conceived the idea, supervised the work and wrote the manuscript.

COMPETING FINANCIAL INTERESTS The authors declare no competing financial interests.

Hiroki Takagi 1, Muluneh Tamiru 1,

Akira Abe 1, Kentaro Yoshida 2, Aiko Uemura 1, Hiroki Yaegashi 1, Tsutomu Obara 1,

Kaori Oikawa 1, Hiroe Utsushi 1, Eiko Kanzaki 1, Chikako Mitsuoka 1, Satoshi Natsume 1, Shunichi Kosugi 3, Hiroyuki Kanzaki 1, Hideo Matsumura 4, Naoya Urasaki 5, Sophien Kamoun 2 & Ryohei Terauchi 1

1Iwate Biotechnology Research Center,

Kitakami, Iwate, Japan. 2The Sainsbury

Laboratory, Norwich Research Park, Norwich, UK. 3Kazusa DNA Research Institute, Kisarazu, Chiba, Japan. 4Gene Research Center, Shinshu University, Ueda, Nagano, Japan. 5Okinawa Prefectural Agricultural Research Center, Itoman, Okinawa, Japan. e-mail: terauchi@ibrc.or.jp or h-takagi@ibrc.or.jp

Published online 23 March 2015; doi:10.1038/nbt.3188

1. Abe, A. et al. Nat. Biotechnol. 30, 174–178 (2012).

2. Munns, R. & Tester, M. Annu. Rev. Plant Biol. 59, To the Editor:

In October 2014, the US Food and Drug Administration (FDA) issued a draft

guidance outlining its plan to overhaul the regulation of in vitro diagnostics, which are tests conducted outside of a living body to detect or diagnose diseases, conditions and infections 1,2. Most in vitro diagnostics, specifically those developed by individual laboratories (known as l aboratory-developed tests, or LDTs) for use by clinicians, have historically been exempt from premarket FDA review. As a result, medical-center laboratories have been at the forefront of precision medicine, rapidly developing tests for rare diseases and public health threats, such as HIV . Recently, however, certain LDTs (e.g., for ovarian and cervical cancers) have been associated with serious safety issues, and increasingly complex LDTs have been marketed for use in broad patient

populations—trends that suggest that a modernized regulatory system is needed to promote innovation in high-quality diagnostics and ensure patient safety.

In the wake of public concerns about the reliability of pap smears and cytological testing, in 1988 Congress passed the Clinical Laboratory Improvement Amendments Act (CLIA), which directed the Centers for Medicare and Medicaid Services (CMS) to set lab quality standards and review the analytical validity of tests (i.e., whether an LDT can detect the analyte(s) it claims to detect).

Meanwhile, the FDA has maintained that the 1976 Medical Device Amendments grants it the authority to regulate LDTs as devices but it has used ‘enforcement discretion’ not to exercise that authority, given that LDTs have historically been low-risk tests. In contrast to LDTs, in vitro diagnostics made by device manufacturers have been reviewed by the

Precision medicine and the FDA’s draft guidance on laboratory-developed tests

C O R R ES P O N

D EN C E

n p g

? 2015 N a t u r e A m e r i c a , I n c . A l l r i g h t s r e s e r v e d .