Identification of QTLs and candidate genes for high grain Fe and Zn concentration in sorghum [Sorghum bicolor (L.)Moench]

Journal Pre-proof Identification of QTLs and candidate genes for high grain Fe and Zn concentration in sorghum [Sorghum bicolor (L.)Moench] Anuradha Kotla, Rahul Phuke, K. Hariprasanna, Shivaji P. Mehtre, Abhishek Rathore, Sunita Gorthy, Rakesh K. Srivastava, Roma Das, A. Bhanu Prakash, K. Radhika, C. Tom Hash, Belum V.S. Reddy, J.V. Patil, Farzana Jabeen, D. Shashikanth, Jayakumar Jaganathan, Anil Gaddameedi, Vangala Subhasini, Santosh P. Deshpande, A. Ashok Kumar PII:

S0733-5210(19)30534-X

DOI:

https://doi.org/10.1016/j.jcs.2019.102850

Reference:

YJCRS 102850

To appear in:

Journal of Cereal Science

Received Date: 2 July 2019 Revised Date:

26 September 2019

Accepted Date: 27 September 2019

Please cite this article as: Kotla, A., Phuke, R., Hariprasanna, K., Mehtre, S.P., Rathore, A., Gorthy, S., Srivastava, R.K., Das, R., Bhanu Prakash, A., Radhika, K., Hash, C.T., Reddy, B.V.S., Patil, J.V., Jabeen, F., Shashikanth, D., Jaganathan, J., Gaddameedi, A., Subhasini, V., Deshpande, S.P., Kumar, A.A., Identification of QTLs and candidate genes for high grain Fe and Zn concentration in sorghum [Sorghum bicolor (L.)Moench], Journal of Cereal Science (2019), doi: https://doi.org/10.1016/ j.jcs.2019.102850. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier Ltd.

P1 (296 B)

× P2 (PVK 801)

Development of

mapping population F1 F2

Phenotyping for 2 year at 3 locations

Genotyping

Sorghum 342 RIL (F6) population Sorghum genetic map for Fe and Zn concentration was constructed with 2,088 markers DArTs, DArTSeqs & SSRs

1 [6]

1 [7]

64.5 64.8 65.1 65.5 65.9 66.3 66.6 67.0

Sn2650914Ch_5 Sn2653364Ch_5 Dt2646241Ch_5 Sn2650898Ch_5 Sn2648447Ch_5 Sn2657606Ch_5 Sn2654009Ch_5 Dt3933553Ch_5

67.6 68.1

Dt1905219Ch_5 Dt2651881Ch_5 Sn1969533Ch_5

68.9 69.2 69.6

Dt1990986Ch_5 Dt3933747Ch_5 Sn2647148Ch_5

70.4 70.8

Dt1898249Ch_5 Dt2654558Ch_5

73.6 74.2 74.3 74.5 74.7 74.8 75.5

Sn1947289Ch_5 Dt1949044Ch_5 Dt2013426Ch_5 Dt1966834Ch_5 Dt2649479Ch_5 Dt1900326Ch_5 Dt2649478Ch_5 Dt2678768Ch_5 Dt2020576Ch_5

76.6 76.9 77.2

Dt2029363Ch_5 Dt1997489Ch_5 Sn2654719Ch_5

78.9 79.0 79.1

Dt2656892Ch_5 Dt2234428Ch_5 Dt1927929Ch_5

1 [8]

1 [9]

80.1

Sn1990516Ch_5

96.3

Dt2648800Ch_5

81.0 81.4

Sn1910427Ch_5 Sn1915366Ch_5

96.9 97.3 97.6

Dt2009838Ch_5 Dt2647635Ch_5 Sn2649364Ch_5

82.4

Sn2703643Ch_5

98.2 98.7

Sn2646445Ch_5 Sn1945357Ch_5

83.6

Dt2215933Ch_5

84.3

Dt2049178Ch_5

85.3

Dt1993087Ch_5

86.1

Dt2710356Ch_5

87.0 87.2 87.5 87.7 88.3

Dt2674484Ch_5 Dt2655450Ch_5 Dt1962146Ch_5 Sn2009268Ch_5 Dt1975411Ch_5 Sn1929912Ch_5

89.4

Sn2650000Ch_5

90.6 91.0 91.3 91.4 91.7 92.1 92.3 93.2 93.7 94.2 94.3 94.7 94.8 95.2 95.6

Sn2658378Ch_5 Dt2735167Ch_5 Dt2655829Ch_5 Dt2650021Ch_5 Sn2650368Ch_5 Sn2651196Ch_5 Dt2684962Ch_5 Dt2650987Ch_5 Dt2648248Ch_5 Dt1878317Ch_5 Dt2023754Ch_5 Dt2658134Ch_5 Dt3933781Ch_5 Sn1940885Ch_5 Sn1989588Ch_5

101.5 102.4 102.7 102.9 103.0 103.1 103.3 103.4 103.6 103.7 104.2 104.4 104.6 105.0 105.5 105.9 106.2 106.6 108.0 108.9 109.1 109.3 109.4 109.5 110.0 110.1 111.4

Sn1928537Ch_5 Dt2648948Ch_5 Dt2655421Ch_5 Dt3626119Ch_5 Dt3627829Ch_5 Dt1947737Ch_5 Dt1940711Ch_5 Dt2648947Ch_5 Dt2215470Ch_5 Sn1928623Ch_5 Dt2217761Ch_5 Dt1936241Ch_5 Dt2650965Ch_5 Dt1945333Ch_5 Dt2647316Ch_5 Dt2647315Ch_5 Dt2649958Ch_5 Sn2647624Ch_5 Sn2645717Ch_5 Dt1935082Ch_5 Sn2645629Ch_5 Sn2657124Ch_5 Sn2647027Ch_5 Dt1987111Ch_5 Dt3933575Ch_5 Sn2653209Ch_5 Dt1975144Ch_5 Dt2645760Ch_5 Dt2645162Ch_5 Dt2644859Ch_5 Dt2650613Ch_5 Sn1899035Ch_5

112.5 113.1 113.3 113.6 114.0 114.1 114.2 114.7 115.2 115.5 116.4 116.6 116.7 117.0 117.1 117.2 117.3 117.5 118.4 119.2

Sn2658321Ch_5 Sn2647596Ch_5 Dt1879670Ch_5 Dt2649928Ch_5 Dt2648322Ch_5 Dt2658640Ch_5 Dt1879574Ch_5 Dt2658400Ch_5 Sn2645213Ch_5 Dt1887849Ch_5 Sn2207205Ch_5 Dt2650022Ch_5 Dt2654522Ch_5 Dt2657774Ch_5 Dt1874273Ch_5 Dt1980071Ch_5 Dt2645126Ch_5 Dt2648399Ch_5 Dt2653422Ch_5 Sn2710051Ch_5 Sn2672849Ch_5

120.8

Sn1948415Ch_5

121.9

Sn1906928Ch_5

122.5

Sn2668958Ch_5

Linkage Map

The total length of genetic map is 1355.52 cM with marker interval of 0.6cM was found

Analysis of grain Fe and Zn concentration - Atomic Absorption Spectroscopy

QTL Mapping

QTLs for grain Fe and Zn concentration were identified - 11 QTLs (individual) and 3 QTLs (across) environments Putative candidate genes involved in Fe and Zn metabolism were identified on SBI-07 (QTL interval of qfe7.1, qzn7.1, and qzn7.2)

Candidate gene prediction Gene synteny analysis

2

Identification of QTLs and candidate genes for high grain Fe and Zn concentration in Sorghum [Sorghum bicolor (L.)Moench]

3 4 5 6 7

Anuradha Kotla1#, Rahul Phuke1,4,#, Hariprasanna K.2, Shivaji P. Mehtre3, Abhishek Rathore1, Sunita Gorthy1, Rakesh K. Srivastava1, Roma Das1, Bhanu Prakash, A1, K. Radhika4, C. Tom Hash1, Belum V.S. Reddy1, Patil J.V.2, Farzana Jabeen4, D. Shashikanth2, Jayakumar Jaganathan1, Anil Gaddameedi1, Vangala Subhasini1, Santosh P. Deshpande1and A. Ashok Kumar1*

8 9 10 11 12 13 14 15

1. International Crops Research Institute for the Semi-Arid Tropics (ICRISAT), Patancheru, Hyderabad 502 324, Telangana, India 2. Indian Institute of Millets Research, Rajendranagar, Hyderabad 500 030, Telangana, India 3. VasantraoNaikMarathwadaKrishiVidyapeeth, Parbhani 431 402, Maharashtra, India 4. Professor JayashankarTelangana State Agricultural University, Rajendranagar, Hyderabad 500 030, Telangana, India * Corresponding author at International Crops Research Institute for the Semi-Arid Tropics, Patancheru502324, India.Tel.: +91 9000414124; E-mail: [email protected];

16

# First two authors made equal contribution.

17

Abstract

18

Sorghum is a major food crop in the semi-arid tropics of Africa and Asia. Enhancing the grain

19

iron (Fe) and zinc (Zn) concentration in sorghum using genetic approaches would help alleviate

20

micronutrient malnutrition in millions of poor people consuming sorghum as a staple food. To

21

localize genomic regions associated with grain Fe and Zn, a sorghum F6 recombinant inbred line

22

(RIL) population (342 lines derived from cross 296B × PVK 801) was phenotyped in six

23

environments, and genotyped with simple sequence repeat (SSR), DArT (Diversity Array

24

Technology) 0and DArTSeq (Diversity Array Technology) markers. Highly significant genotype

25

× environment interactions were observed for both micronutrients. Grain Fe showed greater

26

variation than Zn. A sorghum genetic map was constructed with 2,088 markers (1148 DArTs,

27

927 DArTSeqs and 13 SSRs) covering 1355.52 cM with an average marker interval of 0.6cM.

28

Eleven QTLs (individual) and 3 QTLs (across) environments for Fe and Zn were identified. We

29

identified putative candidate genes from the QTL interval of qfe7.1, qzn7.1, and qzn7.2 (across

30

environments) located on SBI-07 involved in Fe and Zn metabolism. These were CYP71B34,

31

and ZFP 8 (ZINC FINGER PROTEIN 8). After validation, the linked markers identified in this

1

1

32

study can help in developing high grain Fe and Zn sorghum cultivars in sorghum improvement

33

programs globally.

34

Keywords: Biofortification,RIL, QTL, Iron and zinc, Candidate genes, Sorghum

35

Introduction

36

Globally, more than three billion people suffer from micronutrient deficiency, particularly iron

37

(Fe) and zinc (Zn) owing to consumption of poor quality diets that lack adequate micronutrients

38

(Jin et al., 2013). Micronutrient deficiencies can impair the mental and physical development of

39

children and adolescents and result in lower IQ. Stunting and blindness prevalence in women and

40

children is disproportionately high in developing countries. Of the four billion people affected by

41

Fe deficiency globally, more than two billion are in developing countries (WHO, 2011).

42

According to Food and Agriculture Organization (FAO, 2017) two billion people still remain at

43

the risk of Zinc deficiency, the most prone are infants, young children, pregnant and lactating

44

women due to their elevated requirements of Zinc. Moreover Fe and Zn exist predominantly as

45

phytate complexes within the aleurone layer and phytate is major inhibitor of Fe and Zn

46

absorption, hence the bioavability of iron and zinc in sorghum grain is low due to presence of

47

phytate (Zhao et al., 2019). Nutrient supplementation and fortification are the most widely used

48

methods to address the micronutrient malnutrition globally. Biofortification is the process of

49

increasing the density of vitamins and minerals in a crop, through plant breeding or agronomic

50

practices so that when consumed regularly, will generate measurable improvement in vitamin

51

and mineral nutritional status. It is considered a promising approach to overcome micronutrient

52

malnutrition. Globally, sorghum is the fifth most consumed crop, and major staple for nearly 500

53

million people in Africa and Asia. The production and consumption of sorghum is significantly

54

high in the semi-arid parts of Africa and India where in more than 70 percent of the sorghum

55

grain is utilized for human consumption (FAO, 1996). It is a heat and drought tolerant C4 plant

56

with wider adaptation to a range of growing environments. Sorghum grain is gluten-free and safe

57

for people with celiac disease and is a preferred dietary choice due to its low-glycemic index and

58

high anti-oxidant levels (Ashok Kumar et al., 2015). Sorghum is good source of energy with

59

sufficient amounts of Fe and Zn, also rich in potassium, phosphorus, calcium and sodium.

60

Considering its importance as staple and rich nutrient status, sorghum is targeted for

2

61

biofortification. Sorghumcan be further improved to make it more nutrient-dense tobenefit the

62

populations suffering from micronutrient malnutrition.

63

Genetic variation in the crop is critical for improving the trait of interest. Fe and Zn being

64

quantitatively inherited, deciphering the genetic architecture of these traits is very essential for

65

development of micronutrient-efficient genotypes. Earlier studies showed that there exists

66

considerable genetic variability and high heritability for grain Fe and Zn concentration in

67

sorghum. Largest variability for grain Fe and Zn in sorghum was found in the germplasm and it

68

is feasible to exploit this variation by understanding the genetic control of grain Fe and Zn

69

(Ashok Kumar et al., 2013). Further it was indicated that it is possible to breed sorghum with

70

enhanced levels of micronutrients (Fe and Zn) in desirable maturity backgrounds (Ashok Kumar

71

et al., 2013). An understanding of trait relationships helps to formulate an appropriate breeding

72

strategy. Strong genotype × environment interaction was reported for grain Fe and Zn

73

concentration in sorghum that can affect the heritability. However, high broad sense heritability

74

for grain Fe (95.5%) and Zn concentrations (96.5%) was observed in diverse sorghum genotypes

75

and in adapted cultivars (Hariprasanna et al., 2014). Highly significant and positive correlation

76

was observed between grain Fe and Zn concentration (r = 0.853; P <0.01). Earlier studies in

77

sorghum also indicated that both Fe and Zn were quantitatively inherited and Zn is

78

predominantly under the control of additive gene action and both additive and non-additive gene

79

actions are important in controlling grain Fe concentration (Ashok Kumar et al., 2013).

80

Therefore, both parents need to be improved for Fe and Zn for developing hybrids with high Fe

81

and Zn. Further, it’s possible to predict the hybrid performance based on the mean of parental

82

lines (Ashok Kumar et al., 2013).

83

In the sorghum grain minerals are unevenly distributed. The highest concentration of Fe and Zn

84

can be found in the aleuronelayer and the embryo (in particular the scutellum) and a low

85

concentration in starchy endosperm (Chavan and Patil., 2010). The concentration of Fe and Zn

86

differ within a grain, due to accumulation typically much higher concentrations in the embryo

87

than in aleurone layer and endosperm of a grain (Singh et al., 2018). Micronutrients accumulate

88

in grain by the uptake during seed development and requisitely pass through the plant to reach

89

the grain. Thus, understanding the route of these micronutrients, and the role of genes

90

responsible for their uptake, translocation and loading into the endosperm is crucial. The map 3

91

positions of these genes and linked markers on a chromosome need to be determined for the

92

effective genetic dissection and manipulation of complex traits in crop plants. Several QTLs and

93

candidate genes governing grain Fe and Zn concentration have been reported in cereals such as,

94

rice (Anuradha et al., 2012), wheat (Velu et al., 2018); barley (Sadeghzadeh et al., 2010) and in

95

maize (Jin et al., 2013). However, no reports are available on the QTLs and genes controlling

96

grain Fe and Zn in sorghum. Therefore, we attempted the identification of QTLs and underlying

97

candidate genes in sorghum using recombinant inbred lines (RILs) population developed from a

98

cross 296B × PVK 801.

99

Material and Methods:

100

Plant material and field evaluation:

101

The RIL population consisting of 342 lines (F6 generation) developed using two contrasting

102

parents (Grain Fe/Zn, mold reaction, yield and other agronomic traits) 296B and PVK 801, was

103

used in this study for phenotyping and genotyping. The phenotypic trials were conducted for two

104

years (during postrainy seasons 2012-13 and 2013-14) at three locations viz., 1) International

105

Crops Research Institute for the Semi-Arid Tropics (ICRISAT) (17.53 o N, 78.27o E) located at

106

an altitude of 545 m above meansealevel, 2) ICAR - Indian Institute of Millets Research (IIMR)

107

(17.19o N, 78.28o E) located at an altitude of 542.3 m above mean sea level and 3)Vasantrao

108

Naik Marathwada KrishiVidyapeeth (VNMKV), Parbhani (18.45o N, 76.13o E) located at an

109

altitude of 357 m above mean sea level, which are located in major grain sorghum growing areas

110

in India. The three locations and two years form six environments, hereafter will be referred as

111

ICRISAT 2012-13(E1), IIMR 2012-13(E2), VNMKV 2012-13(E3), ICRISAT 2013-14(E4),

112

IIMR 2013-14(E5) and VNMKV 2013-14(E6). All the field experiments were conducted in post-

113

rainy seasons to obtain the best quality grain for assessing grain Fe and Zn.

114

The soil Fe and Zn in the experimental field were analysed by DTPA extractable method at

115

Charles Renard Analytical Laboratory, ICRISAT, Patancheru, and expressed as mg kg-1 (ppm).

116

These Fe and Zn contents in the soil were in the sufficient range for normal plant requirements

117

(2.6 to 4.5 mg kg-1 for Fe; 0.6 to 1.0 mg kg-1 for Zn) (Sahrawat and Wani, 2013). All the

118

recommended agronomic practices were followed for raising a good crop. Observations were

119

recorded on the following traits: Days to 50% flowering, plant height, 100-grain weight (test

120

weight), grain yield and grain Fe and Zn concentration using standard procedures. The grains 4

121

were harvested at physiological maturity stage. During harvestthe main panicles of five random

122

plants from each plot were harvested and stored separately in a cloth bag to produce clean grain

123

samples for micronutrient analysis. The remaining panicles of the plot were harvested as a bulk.

124

These panicles were sundried for 10 to 15 days. Five separately harvested panicles were

125

manually threshed for the each replicate and approximately 20 g of grain was collected for Fe

126

and Zn analysis and for 100 grain weight. Leftover grains from these panicles were added to the

127

bulk grain produced by threshing in a multi head machine thresher. The grain yield including the

128

20 g sample taken for micronutrient analysis was recorded for each plot and extrapolated to

129

tonnes per hectare for grain yield analysis. In all the experiments, self-pollinated (SP) grain

130

samples were used to estimate grain Fe and Zn concentration expressed in mg kg-1. The

131

harvested panicles were put directly in a separate cloth bag to avoid soil contamination and dried

132

in the sun to <12% post-harvest grain moisture content. Grains were cleaned from their glumes,

133

panicle chaff, and debris and transferred to new, non-metal fold envelops and stored at cold

134

temperatures. Due care was taken at each step to avoid contamination of the grains with dust or

135

other extraneous matter.

136

The grain Fe and Zn concentration of the grains from all six environments were analysed at the

137

Charles Renard Analytical Laboratory, ICRISAT, Patancheru, India following the standard

138

method (Wheal et al., 2011). The digests were filtered and Fe and Zn in the digests were

139

analysed using Inductively Coupled Plasma Optical Emission Spectrometry (ICP-OES).

140 141

Statistical Analysis

142

Analysis of Variance

143

Combined Analysis of Variance was carried out for six environments by modelling individual

144

environment (a combination of location and year) error variances using mixed model procedure.

145

Five variance components ( σg ,

146

studied using restricted maximum likelihood estimation procedure (Paterson and Thompson,

147

1971) using GenStat Software, 17th edition (VSN International, Hemel Hempstead, UK). In these

148

analyses, environment (location and year) was fitted as a fixed effect. Genotype, blocks,

149

replicates and genotype interactions with environment, were fitted as random effects. The

2

2 2 σgy2 , σgl2 , σ gyl , σe ) were estimated for each of the six traits

5

zijklm on accession m in replicate k of block l of location j and year i was

150

phenotypic observations

151

modelled as:

152

zijklm = µ + yi + ej + yeij + rijk +bijkl + gm + ( yg)im + (eg) jm + ( yeg)ijm +εijklm

153

Where

154

µ is the grand mean; yi is the fixed effect of year i; e j is the fixed effect of location j ;

yeij is the fixed effect of interaction between year i and location j; gm is the random effect of

σg2 ) ; rijk is the random effect of replication in location j and year

155

genotype m and is ~ NID(0,

156

iand is ~ NID(0,

157

year i and is ~ NID(0,

158

and year i and is

159

accession m in location j and

160

effect of the genotype m in year i and location j and

161

residual effect and ~ NID(0,

162

Analyses of variance were also conducted using data from each environment for all six traits.

163

Heritability ( H 2 , broad sense) at individual environment was estimated from analysis of

164

variance. The formula used as -

165

σ g2 H = 2 σ g + σ ε2 / r

166

Whereas Heritability estimates across the environments were estimated by the formula as-

167

H2 =

168

Where r , y , e denotes the number of replicates, years and environments respectively. The

169

relationship between grain Fe and Zn concentration and agronomic traits like days to 50%

170

flowering, plant height, 100-seed weight and grain yield were determined using Pearson

171

correlation coefficient and genotypic correlation (Karl Pearson., 1895) using BLUPs (Best Liner

172

Unbiased Predictors) of single environment as well as across the environments.

σr2 ); bijkl is the random effect of block l nested with replication k in location j and

~

σb2 ); ( yg)im

NID(0,

is the random effect of the interaction between genotype m

σ yg2 ); (eg) jm is the random effect of the interaction between ~

NID(0,

σ eg2 ); ( yeg)ijm is the random effect of the interaction ~

NID(0,

σ yg2 ); and εijklm is the random

σε2 ).

2

σ g2 2 2 σ g2 + σ yg / y + σ eg2 / e + σ yeg / ye + σ ε2 / rye

173 6

174

Genotyping, linkage map construction and QTL mapping:

175

DNA isolation (Mace et al., 2003) of RIL population and polymorphic screening with SSR

176

markers were done at Center of Excellence in Genomics (CEG) Lab at ICRISAT, Patancheru,

177

India. The DNA of 342 RIL population along with two parents was analyzed for DArT analysis

178

at Diversity Arrays Technology Pvt. Ltd. (DArT P/L), Yarralumla, ACT 2600, Australia

179

(http://www.diversityarrays.com). The SSRs, DArT and DArTseq (SNPs) markers were used to

180

genotype the parents and their derived RILs. A genetic linkage map was constructed using both

181

the co-dominant microsatellite data and the dominant DArT and SNPs marker data using

182

JoinMapv4.0 (van Ooijen et al.,2006). Markers within each group were mapped using the

183

regression mapping algorithm with the LOD score 10 and map distance was estimated using the

184

Kosambi mapping function. Quantitative trait locus (QTL) analysis was performed using the

185

composite interval mapping (CIM) by the software Win Cartographer V2.5_011 with a one cM

186

walk speed, threshold LOD of 2.5 and 1000 permutations. QTL analysis was performed using

187

the BLUPs of across environmentsand individual environments.The original RIL population

188

contained 342 lines. However, due to more than 10% non-parental data points (RIL from two

189

parents couldn’t be discriminated) few of RILs were removed in linkage mapping. Finally

190

309 lines were used for genetic linkage analysis and QTL detection.

191 192

Bioinformatics analyses:

193

The stable QTLs across locations were analyzed insilico to identify putative candidate genes

194

involved in Fe/Zn metabolism within QTLs interval. The putative genes within the identified

195

QTL

196

(http://www.plantgdb.org/SbGDB/cgi-bin/downloadGDB.pl). Further, literature search was

197

performed to identify the putative candidate genes related to Zn, Fe uptake, homeostasis and

198

accumulation in grains, that are present within the identified QTLs. Sorghum genes underlying

199

Fe/Zn QTLs were used to analyze the gene synteny with rice and maize using Phytozome 10.3

200

(Sorghum bicolor v2.1_PhytoMine) software Gene synteny - Phytozome 10.3 >JBrowse>

201

bicolor v2.1>Gene info> Homologues.

region

across

the

regions

were

retrieved

using

Plant

GDB

202 203

Results: 7

204

Mean performance, G×E interaction, heritability and trait association

205

The RIL population (296B × PVK 801) having 342 RILs of F6 generation was

206

phenotyped in six environments i.e. three locations over two years.The trials were conducted

207

using Alpha Lattice Design with three replications for agronomic traits and grain Fe and Zn

208

concentration along with parents to obtain means and variances (Table 1a & 1b). The phenotypic

209

data collected for the population during post-rainy seasons of 2012-13 and 2013-14 in six

210

environments in total were analysed statistically to obtain variance components. The mean

211

performance for grain Fe and Zn concentration in both the parents and RILs were higher in

212

VNMKV 2012-13, whereas lowest in ICRISAT 2013-14, also both the parents exhibited wider

213

range of variation for grain Fe and Zn studied in different environments.

214

The grain micronutrient (Fe and Zn) concentration showed highly significant genotypic

215

variances in all the individual environments, werein, the grain Fe concentration showed the

216

highest genotypic variance in VNMKV 2013-14 and the grain Zn concentrations showed highest

217

variance in VNMKV 2012-13. The same trend continued across environments, both the traits

218

showed highly significant genotypic variances, whereas genotype × year (σ2gy) interaction for

219

Zn was significant, but less in magnitude compared to genotypic variance. For grain Fe, the

220

genotype × year (σ2gy) interaction was non-significant. Genotype × location (σ2gl) interactions

221

were found to be non-significant for both the micronutrients but genotype × year × location

222

(σ2gyl) interactions were highly significant for both the traits, also the magnitude of variance was

223

more than genotypic variances. The frequency distribution of Fe and Zn based on across

224

environments BLUPs showed transgression beyond the parent’s values for both the traits.

225

Considerable transgressive segregation and significant differences in traits was obtained in

226

present RIL population, indicating the occurrence of large variation among the RILs specifying

227

that the RIL population is desirable for QTL mapping.

228 229

In the present study both grain Fe and Zn were highly heritable (>0.60) in individual

230

environments except for grain zinc concentration in IIMR 2013-14 environment (Table 1b).

231

Agronomic traits like days to 50% flowering, plant height, 100-seed weight and grain yield were

232

found to be more heritable than grain micronutrients (data not shown). However, a partitioned

233

genotype × environment interaction component reduced the heritability across the environments 8

234

(pooled analysis). Broad sense heritability for all the traits was high (0.30 – 0.60) across the six

235

environments. For grain Fe and Zn concentration, the heritability was high in first year (post-

236

rainy 2012-13) compared to second year (post-rainy 2013-14), whereas environment-wise IIMR-

237

2013-14 showed lowest value for Fe and Zn heritability in tune with low genotypic variance for

238

these traits in same environment.

239

Pearson’s phenotypic and genotypic correlation among traits were computed for each individual

240

environment and across the environments based on BLUPs of each individual environment and

241

across the environments (Table 2). Significant positive and negative correlations were observed

242

between the traits. Some of the traits were highly correlated while many of them had weak

243

correlations. Based on BLUPs from the individual environments for phenotypic correlation, there

244

was highly significant and high positive association between Fe and Zn concentrations in all the

245

environments (E1= 0.80, E2= 0.69, E3= 0.70, E4=0.72, E5= 0.69 and E6= 0.65; p <0.01) and

246

this trend was consistent in pooled analysis (AE= 0.79; p <0.01). The correlation of grain Zn

247

concentration and grain yield was negative, except in one environment (IIMR 2013-14(E5), but

248

was smaller in magnitude compared to association between grain iron concentration and grain

249

yield except in one environment (VNMKV 2013-14 (E6). For 100-seed weight, out of six

250

environments, grain Fe showed a significant positive association in only three environments

251

(ICRISAT 2012-13, ICRISAT 2013-14andVNMKV 2013-14). Using across environment

252

analysis, the Fe concentration showed significant positive (AE= 0.34; p<0.01) association with

253

100-seed weight. In case of Zn concentration, four environments (ICRISAT 2012-13, IIMR

254

2012-13, ICRISAT-2013-14 and VNMKV 2013-14) showed significant positive association and

255

remaining two environments (VNMKV 2012-13 and IIMR 2013-14) did not show any

256

significant association, whereas the across environment analysis of Zn concentration showed

257

significant positive (AE= 0.36; p<0.01) association with 100-seed weight.

258

The association between grain Fe concentration and days to 50% flowering ranged from (E2 = -

259

0.23; p<0.01 to E1= 0.04; non-significant), while across the environmentsshowed significant

260

negative (AE= -0.12; p<0.05) association between Fe concentration and days to 50% flowering.

261

The association of Zn concentration with days to 50% flowering ranged from (E4= E5= -0.11;

262

p< 0.05 to E6= 0.08; non-significant), while across the environments it showed significant

263

negative (AE= -0.12; p<0.05) association between Zn concentration and days to 50% flowering.

264

The association between plant height and grain Fe concentration ranged from (E2= 0.002; non9

265

significant to E1= E3= 0.29; p<0.01), while across environments it showed significant positive

266

(AE= 0.31; p<0.01) association. The plant height association with grain Zn concentration,

267

ranged from (E5= 0.04; p
268

showed significant positive association (AE= 0.33; p<0.01). The association between 100-seed

269

weight (seed size) and grain yield ranged from (E1= -0.13, p<0.05 to E2= 0.26, p<0.01), while

270

across environments it showed significant negative association (AE= -0.12, p< 0.05).

271

The association analysis using genotypic correlation also showed similar results indicating

272

significant positive correlation between grain Fe and Zn (E1 = 0.88, E2 = 0.73, E3 = 0.76, E4 =

273

0.81, E5 = 0.83 and E6 = 0.71; p <0.01) and this trend was consistent in pooled analysis (AE =

274

0.88; p < 0.01). The correlation of other agronomic traits with grain Fe and Zn showed same

275

trend as Pearson phenotypic correlation (Table 2).

276

Construction of high density genetic map

277

A total of 271 SSR markers were used for a polymorphism survey of two parental lines. Forty-

278

five out of 271 (16.60%) SSR markers detected polymorphism between 296B and PVK 801.

279

From the Diversity Arrays Technology (DArT) analysis, a set of 6,126 (70.15%) polymorphic

280

DArT clones were identified in a total of 8,732 clones. In addition, 3,331 polymorphic DArtseq

281

(SNPs) were identified in a total of 3,640 clones (91.51%) on the array of 296B and PVK 801.

282

After removal of the markers with distorted segregation and loci that contained >10% missing

283

data, a total of 2,088 markers were retained. A linkage map was constructed using 2,088

284

polymorphic DArT (1,148), DArTseq (SNPs) (927) and SSRs (13) markers of sorghum RIL

285

population. A length of 1,355.52 cM was covered across the 10 chromosomes, with an average

286

marker interval of 0.6 cM (Figure 3). The largest linkage group contained 257 markers and

287

spanned 180.6 cM (SBI-03), while the smallest linkage group contained 79 markers and spanned

288

only 102.6 cM (SBI-10). The average linkage group length was 135.5 cM with an average of

289

208.7 loci. The average adjacent marker interval length ranged from 0.51 cM (SBI-01 and SBI-

290

05) to 1.29 cM (SBI-10) followed by 0.89 cM (SBI-08) and 0.0023% of the intervals (5 out of

291

2,088) were more than 5 cM.

292

QTL analysis ofgrain Fe and Zn concentration:

10

293

The QTL analysis for individual environments (Table 3) and across the environments (Table 4)

294

was carried out to better understand the QTLs controlling grain Fe and Zn (Figure 1).

295 296

Individual Environment (E) Analysis:

297

ICRISAT

298

A total of 6 QTLs were identified for grain Fe and Zn concentration in ICRISAT 2012-13 (E1)

299

environment. The 3 QTLs for Fe concentration out of 6, were located on chromosomes SBI-06

300

and SBI-07 and 3 QTLs for Zn were on chromosome SBI-07. The phenotypic variation (PVE)

301

for Fe explained by these QTLs ranged from 5.09% to 5.82%, and for Zn from 6.96% to 9.42%.

302

Only one Fe QTL with 6.80% PVE was identified in ICRISAT 2013-14 (E4) environment on

303

chromosome SBI-01. Two significant QTLs qfe7.1 (E1) and qzn7.1 (E1) associated with high

304

PVE (Zn 9.4% and Fe 5.8%) co-located on SBI-07 (Flanking markers Sn2644846 - Dt3625344)

305

for Fe and Zn concentration were found within the region of 57 to 60.3 Mb and for both of the

306

alleles increasingly derived from PVK 801 parent.

307

IIMR

308

IIMR- 2013 (E2) and 2014 (E5) in both of the environments did not find any significant QTLs

309

VNMKV

310

One Fe and 2 Zn QTLs were identified in VNMKV 2012- 13 (E3) environment on chromosome

311

SBI-07. The PVE by these QTLs for Fe was 5.19% and for Zn 5.83% to 6.29%. One Fe QTL

312

with 5.66% PVE was identified in VNMKV2013-14 (E6) environment on chromosome SBI-07.

313

Across Environment Analysis:

314

One Fe and 2 Zn QTLs were identified across environments (pooled data of six environments)

315

QTL analysis on chromosome SBI-07 (Table 4). The PVE explained by these were for both Fe

316

and Zn QTLs

317

QTLs out of 3 were repeated from individual environment analysis.

318

Candidate gene prediction:

319

Three QTLs were identified in this study(pooled data of six environments) which were analyzed

320

in silico for identifying putative candidate genes within the QTL marker intervals. In all,

ranged from 6.66% and 5.66 to 5.74% respectively. Two (qfe7.1 and qzn7.1)

11

321

sixgenes governing Fe and Zn concentrations were found within three QTLs (Table 3). One gene

322

belonging to Cytochrome P450 family and 5 genes related to zinc finger protein family were

323

identified on chromosome SBI-07 (Table 4). In addition, we have also identified the genes on

324

QTLs from individual locations and studied the synteny with other crops such as rice and maize

325

(Figure 2). In this, we identified 42 genes governing Fe and Zn concentrations within the

326

identified QTLs.

327 328 329

Gene synteny analysis:

330

located on the present map of sorghum with published information from other well-studied

331

cereal crop species. Identified sorghum genes within QTL intervals were used to evaluate gene

332

synteny with (Zea mays) and rice (Oryza sativa) (Figure 4). The Synteny sequence level between

333

sorghum-maize ranged from 49% to 99%, while sorghum-rice ranged from 44% to 97%.

334

Putative candidate genes underlying the regions associated with grain Fe/Zn QTLs in sorghum

335

shown gene syntenic relationship with Zea mays on chromosomes 1, 2, 3, 4, 6, 8 and 10 and

336

Oryza sativa on chromosomes 1, 2, 4, 5, 6, 8, 9, 10 and 11.

The Synteny analysis was carried out using sequence information of the genes underlying QTLs

337 338 339

Discussion:

340

The grain Fe and Zn are quantitatively inherited traits and the correlation between grain Fe and

341

Zn concentration and agronomic traits is complex. Therefore, it is essential to understand the

342

genetic architecture of these traits. Measuring the constancy of grain micronutrients for

343

biofortification is critical, as sorghum is grown in diverse soil types with varying levels of

344

fertility and nutrient management in different agro-eco zones.The current study showed that the

345

parents have substantial differences for both Fe and Zn in all the environments. The average Fe

346

and Zn concentrations were highest at VNMKV followed by IIMR and ICRISAT in first year

347

analysis (2012-13),whereas in second year (2013-14) the average Fe and Zn concentrations were

348

highest at IIMR followed by VNMKV and ICRISAT.The VNMKV soils are deep black with

349

high moisture holding capacity compared to the soils of IIMR and ICRISAT. This indicates

350

Genotype × Environment interaction possibly effecting mineral uptake, translocation and

351

distribution calling for extensive testing of genotypes for commercialization. The association 12

352

between grain Fe and Zn showed significant and high positive values and the trend was similar in

353

across environment analysis. A Significant positive association between Fe and Zn gives

354

opportunity for simultaneous selection for both traits. No or weak association of grain Fe and Zn

355

with other agronomic traits including yield enables breeders to develop high yielding genotypes

356

possessing high grain Fe and Zn in different maturity backgrounds (Ashok Kumar et al., 2015).

357

The recent (2018) release of biofortified sorghum variety ‘Parbhani Shakti’ by VNMKV-

358

Parbhani, Maharasthra, India in partnership with ICRISAT vindicated this. It showed 45 ppm Fe

359

and 32 ppm Zn compared to the base level of 30 ppm Fe and 20 ppm Zn in sorghum. The grain

360

yield (3.4 tons ha-1) of Parbhani Shakti is higher than the best varieties currently grown in

361

Maharashtra,

362

09/SF18_case_study_bioavailable_micronutrients.pdf

India.

https://www.scienceforum2018.org/sites/default/files/2018-

verified on 29 March 2019

363 364

Multi-location and multi-season evaluation is a prerequisite foridentifying stable donors for

365

micronutrient

366

commercialization. In the present study, the analysis of variances for all the traits revealed highly

367

significant genotypic variances in individual environments as well as across environments

368

indicating high degree of genotypic variance for the traits studied. For agronomic traits, the

369

genotype ×year × location (σ2gyl) interaction values were significant, but lower than genotypic

370

variance, suggesting that the traits are predominantly under genetic control and influenced by

371

environments to a limited extent, which impliesthat there is no need for G×E portioning.

372

Whereas, for both micronutrients, the genotype × year × location (σ2gyl) interactions were

373

significant and also higher than genetic variances indicating environment in different years

374

played significant role in expression of grain Fe and Zn concentrations. Therefore, Fe and Zn

375

concentration in grains show variation according to micronutrients concentration in soil and their

376

availability to plants besides genotypic variability in extraction, translocation and loading.

377

Another reason for greater G×E interaction for Fe and Zn concentrations could be their

378

quantitative inheritance as reported in maize, rice and sorghum (Gregorio, 2002; Long et al.,

379

2004; Ashok Kumar et al., 2013; Hariprasanna et al., 2014), therefore the progress in the genetic

380

selection and improvement of these traits is expected to be slower than many other traits.

381

However, in spite of these challenges, there is evidence that breeding for increased levels of

382

micronutrients is feasible (Ashok Kumar et al., 2015). While stronger G×E effects call for testing

enhancement

and

foridentifying

the

most

promising

genotypes

for

13

383

the genotypes in large number of locations for many years for assessing the stability of genotypic

384

performance, large variability, high heritability, stronger positive association between Fe and Zn

385

traits offers great breeding opportunity for improving Fe and Zn in sorghum

386

Identification of QTLs helps to have better grip over the traits by identifying genomic regions

387

responsible for phenotypic variation. In that direction, to identify QTLs for Fe and Zn in

388

sorghum, a RIL population was phenotyped in multiple environments over two years and

389

genotyped using the SSR, DArT and DArTSeq markers. A linkage map length covering 1355.52

390

cM with an average marker interval of 0.6cM indicates that good genome coverage was

391

developed in the present study. In the first linkage map of sorghum using DArT and other

392

markers developed by Mace et al., (2009),the reported map length was 1603.5 cM with average

393

marker density of 0.79 cM. Compared to previous maps in sorghum, the present linkage map

394

average adjacent-marker distance is smaller with high density of markers. Genome coverage

395

reported in this map will be good source to select markers for use in whole-genome breeding

396

strategies.Compared to SSR based genetic linkage map, a well saturated genetic linkage map

397

developed using DArt/DArtseq markers for the RIL population is more suitable for precise QTL

398

mapping.

399

The QTL analysis for grain Fe and Zn in individual environments identified different QTLs in

400

different environments, which shows the effect of the environment on expression of traits beside

401

the soil type and soil Fe and Zn concentration. It is noteworthy to highlight that several genes

402

controlling Fe and Zn homeostasis in cereal grains and also grain Fe and Zn are highly

403

quantitative in nature. The Occurrence of multiple QTLs nearly in all environments with small

404

phenotypic variance, explains the cumulative effect of different genomic regions contributing to

405

a small percentage of phenotypic variation on expression of QTLs. This suggests that the

406

transport and accumulation of minerals in seed is a complex process and highly influenced by

407

environment. However, the present QTL mapping results for grain Fe and Zn showing several

408

genes with small effects cannot be investigated individually. The detection of a few QTLs with

409

large phenotypic effect is often the result of a small population size or can be the artifact of the

410

strong directional selection often used to create phenotypically divergent lines used for mapping.

411

However,in the present study the mapping population used was oflarge size and the variation in

412

parents of RIL mapping population is nearly 4 mgkg-1 for both the traits Fe and Zn, which is

14

413

lesser than variation used to select phenotypically divergent parental lines resulting in strong

414

directional selection

415

In this study, Fe and Zn QTLs were detected using a high-density genetic map, which serves as a

416

reliable reference for mapping important micronutrient traits in sorghum. A total of 11 genomic

417

regions identified for both micronutrients. A common genomic region identified with high

418

phenotypic variance Zn 9.42% (qzn7.1 (E1)) and Fe 5.82% (qfe7.2 (E1)) on SBI-0 7 in ICRISAT

419

environment. The results suggest that identification of 9 out of 11 genomic regions and QTLs

420

with high phenotypic variance on chromosome SBI-07 might be the useful target regions for

421

further improvement of both the micronutrients in sorghum. Two QTLs qfe7.2 (E1) and qzn7.2

422

(E3) on SBI-07 also were observed in across the environment analysis.The co-localized and

423

stable (across) QTLs on chromosomes SBI-07 for Fe and Zn providing evidence for the

424

significant positive correlation between for Fe and Zn in sorghum.The co-localization of QTLs

425

for Fe and Zn was previously observed in rice on chromosome 7 and 12 (Anuradha et al., 2012)

426

and in maize on chromosomes 2, 5, 7 and 9 (Jin et al., 2013). This may be due to some QTLs

427

having common molecular mechanisms controlling the uptake and translocation of the both

428

traits. The direction and intensity of association suggests a good possibility of simultaneous

429

genetic improvement of both micronutrients by co-transferring these traits into the genetic

430

background of elite lines.The stability of grain micronutrient constituents across environments as

431

well as co-localization of QTLs plays an essential role in marker-assisted breeding.

432

The availability of the complete genome sequence enabled detailed study of the identified QTLs

433

and the genes involved in grain Fe and Zn concentration in sorghum. In all, Six genes related to

434

Fe/Zn concentrations were found to be present within the selected QTLs on chromosome SBI-

435

07(pooled data of six environments). Annotations of these genes were observed as,Heme

436

binding/iron ion binding, Zinc ion binding/metal ion binding, , bZIP transcription factor, one Fe

437

gene governing Fe metabolism out of the 6, and were found in QTL with phenotypic effect

438

(6.7%) present on chromosome SBI-07.Five genes for Zn (Zinc ion binding and zinc finger

439

protein) were found within two QTLs qzn7.4 and qzn7.1. The identified gene CYP71B34 is

440

important for membrane integrity and most importantly phytochrome activities (Table 4). Zinc

441

ion binding and zinc finger protein ZFP motifs are required for their transcriptional activation

442

activity. Upon looking into the QTLs from individual locations we found 42 genes related to 15

443

Fe/Zn metabolism on chromosome SBI-07. With the progress made in identification of Fe and

444

Zn QTLs we performed gene synteny analyses with maize and rice genomes using sorghum Fe

445

and Zn genes found within QTLs interval. Significant gene sequence syntenic regions were

446

detected, the gene synteny relations could be due to close evolutionary relationship of major

447

cereal crops. These genes play a crucial role in protein metabolism wherein the stability of zinc

448

finger protein is regulated by appropriate supply of Zn to the grain. Zinc finger proteins, are

449

required for the expression and regulation of gene expression (Marschner, 1995). These results

450

strongly suggest that simultaneous selection for grain Fe and Zn will be effective in sorghum.

451

Many of these genes have been previously reported for metal homeostasis, metal chelation,

452

biosynthesis of phytosiderophore, uptake, transport, loading and storage in cereals, such as rice,

453

maize, barley and wheat. Availability of the complete genome sequence of sorghum, maize and

454

rice enabling to study comparative genomics studies for better understanding of grain Fe/Zn

455

QTLs and genes. These results will contribute to better understanding of the genetic control of

456

grain Fe and Zn concentration in sorghum. The genomic resources generated from this study

457

through QTL mapping, identification of putative candidate genes underlying these QTLs, gene

458

synteny, would provide a basis for functional analysis, which can be potentially used after

459

validation in marker-assisted selection for grain Fe and Zn in sorghum.

460 461

Conclusions

462

The present study on sorghum RIL population showed wide variability for both grain

463

micronutrients (Fe and Zn), also the presence of G × E interaction for both the micronutrients

464

indicating strong influence of environment on the expression of these traits. It is important to test

465

the genotypes in multiple environments over years to ascertain their stability. Constant positive

466

and highly significant correlation between grain Fe and Zn concentration showed that

467

simultaneous selection for both the micronutrients will be highly effective. The recent release of

468

‘Parbhani Shakti’ the first biofortified sorghum variety in India by the same authors vindicated

469

this.This work is the first report on QTL mapping utilizing high-density of DArT and DArTseq

470

markers for identifying QTLs for grain Fe and Zn in sorghum.Overall, this study identified a

471

total of 11QTLs (individual) and 3QTLs (across) environments controlling both the traits. QTLs

472

that are co-located and stable across environments, with common markers provides opportunity 16

473

for early generation testing for enhancing grain Fe and Zn in sorghum. This study may also

474

serves as a model for further genetic investigations of the sorghum grain Fe and Zn

475

concentration, as well as in other C4 crops.

476 477 478 479 480 481 482 483 484 485 486 487 488

Author contributions Experiments were designed by AAK, SPD, KA; trials were carried out by RP, KA, AAK, BVSR, SG, HK, SPM, AG and VS; the data were analyzed and interpreted by AAK, RKS, AR, RP,RD, RK, FJ; genotyping done by SPD, KA, CTH; genetic map construction and QTL analysis by KA, SPD, RKS; Candidate gene search made by KA, SG; gene synteny analysis by KA, BP. The manuscript was developed by KA, RP, AAK, RKS, SG, SPD.

Acknowledgement

489

The authors thank the funding support by Department of Biotechnology, Government of India

490

for financing the Project“Biofortifying sorghum with high grain iron and zinc concentration for

491

combating micronutrient malnutrition” vide No BT/PR 1630/AGR/2/806/2011 Dt: 28 February

492

2012. We would also like to thank the Charles Renard Analytical Laboratory for analyzing grain

493

iron and zinc concentration of samples. Thanks are also due to Senthivel, Manish Vishwakarma,

494

Manish Pandey and Rajeev Varshney for their guidance in interpreting results.We dedicate this

495

research work to DrBelumVenkataSubba Reddy, who left for heavenly abode after devoting

496

much of his life for sorghum improvement research.

497

Conflict of interest

498

The authors declare that they have no conflict of interest

499

Compliance with Ethical standards

500

No animal or humans have been used in this study. The work was done and manuscript was

501

developed with high ethical standards

502 503

17

504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546

References 1. Anuradha, K., Agarwal, S., Rao, Y. V., Rao, K. V., Viraktamath, B. C., &Sarla, N. 2012. Mapping QTLs and candidate genes for iron and zinc concentrations in unpolished rice of Madhukar × Swarna RILs. Gene, 508(2), 233-240. 2. Ashok Kumar, A., Anuradha, K., Ramaiah, B., Grando, S., Frederick, H., Rattunde, W., & Pfeiffer, W. H. 2015. Recent advances in sorghum biofortification research. Plant Breeding Reviews: Volume 39, 89-124. 3. Ashok Kumar, A., Reddy, B. V. S., Sharma, H. C., Hash, C. T., Srinivasa Rao, P., Ashok Kumar, A., Reddy, B. V., Ramaiah, B., Sahrawat, K. L., & Pfeiffer, W. H. 2013. Gene effects and heterosis for grain iron and zinc concentration in sorghum [Sorghum bicolor (L.) Moench]. Field Crops Research, 146, 86-95. 4. Chavan, U.D., and J.V. Patil. 2010. Grain sorghum processing. Ibdc Publ., Lucknow, India, 440. 5. Chen, Y., and Jia, X. 2006. Submitted (APR-2006) to the EMBL/GenBank/DDBJ databases. 6. FAO, 1996. The World Sorghum and Millet Economies: Facts, Trends and Outlook.1996. (FAO Commodities and Trade Division and the Socioeconomics and Policy Division, International Crops Research Institute for the Semi-Arid Tropics), M-71 ISBN 92-5103861-9. 7. FAO, 2017. Food and Agriculture Organization of the United Nations, Rome, Italy. Feedipedia. 8. Gregorio, G. B. 2002. Progress in breeding for trace minerals in staple crops. The Journal of nutrition, 132(3), 500S-502S. 9. Hariprasanna, K., Agte, V., &Patil, J. V. 2012. Genotype× environment interactions for grain micronutrient contents in sorghum [Sorghum bicolor (L.) Moench]. Indian Journal of Genetics and Plant Breeding, 72(4), 429-434. 10. Hindu, V., Palacios-Rojas, N., Babu, R. 2018. Identification and validation of genomic regions influencing kernel zinc and iron in maize. TheorAppl Genet; 131: 1443. https://doi.org/10.1007/s00122-018-3089-3 11. Jin, T., Zhou, J., Chen, J., Zhu, L., Zhao, Y., & Huang, Y. 2013. The genetic architecture of zinc and iron content in maize grains as revealed by QTL mapping and meta-analysis. Breeding science, 63(3), 317-324. 12. Karl Pearson (20 June 1895) Notes on regression and inheritance in the case of two parents. Proceedings of the Royal Society of London, 58 : 240–242.

13. Long, J. K., Bänziger, M., & Smith, M. E. 2004. Diallel analysis of grain iron and zinc density in southern African-adapted maize inbreds. Crop Science, 44(6), 2019-2026. 14. Mace, E.S., Buhariwalla, K.K., Buhariwalla, H.K. and Crouch, J.H., 2003. A highthroughput DNA extraction protocol for tropical molecular breeding programs. Plant Molecular Biology Reporter, 21(4), pp.459-460. 15. Mace, E.S., Rami, J.F., Bouchet, S., Klein, P.E., Klein, R.R., Kilian, A., Wenzl, P., Xia, L., Halloran, K. and Jordan, D.R., 2009. A consensus genetic map of sorghum that 18

547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593

integrates multiple component maps and high-throughput Diversity Array Technology (DArT) markers. BMC plant biology, 9(1), p.13. 16. Marschner H (1995) Mineral nutrition of higher plants. 2th ed. Academic, London 17. Mills, R. F., Peaston, K. A., Runions, J., & Williams, L. E. 2012. HvHMA2, a P1BATPase from barley, is highly conserved among cereals and functions in Zn and Cd transport. PLoS One, 7(8), e42640. 18. Nozoye, T., Inoue, H., Takahashi, M., Ishimaru, Y., Nakanishi, H., Mori, S., &Nishizawa, N. K. 2007. The expression of iron homeostasis-related genes during rice germination. Plant molecular biology, 64(1-2), 35-47. 19. Patterson HD, Thompson R .1971. Recovery of inter-block information when block sizes are unequal. Biometrika 58:545–554 20. Ricachenevsky, F. K., Punshon, T., Lee, S., Oliveira, B. H. N., Trenz, T. S., dos Santos Maraschin, F., &Guerinot, M. L. 2018. Elemental profiling of rice FOX lines leads to characterization of a new Zn plasma membrane transporter, OsZIP7. Frontiers in plant science, 9. 21. Sadeghzadeh, B., Rengel, Z., Li, C., & Yang, H. A. 2010. Molecular marker linked to a chromosome region regulating seed Zn accumulation in barley. Molecular breeding, 25(1), 167-177. 22. Singh, B. R., Timsina, Y. N., Lind, O. C., Cagno, S., & Janssens, K. 2018. Zinc and iron concentration as affected by nitrogen fertilization and their localization in wheat grain. Frontiers in plant science, 9, 307. 23. Sahrawat, K.L and Wani, S.P. 2013. Soil testing as a tool for on-farm fertility management: Experience from the semi-arid zone of India. Communications in Soil Science and Plant Analysis. 44: 1011-1032. 24. Van Ooijen, J. W. 2006. JoinMap® 4, Software for the calculation of genetic linkage maps in experimental populations. Kyazma BV, Wageningen, 33(10.1371). 25. Velu, G., Singh,R.V,. Crespo-Herrera, L., Juliana,P., Dreisigacker,S., Valluru,R., Stangoulis, J., Sohu,V.S., Mavi, G.S., Mishra,V.K., Balasubramaniam, A., Chatrath,R., Gupta, V., Singh,G.P., and Joshi,A.K. 2018. Genetic dissection of grain zinc concentration in spring wheat for mainstreaming biofortification in CIMMYT wheat breeding. Scientific Reports, vol (8), Article number: 13526 26. Wheal, M. S., Fowles, T. O., & Palmer, L. T. 2011. A cost-effective acid digestion method using closed polypropylene tubes for inductively coupled plasma optical emission spectrometry (ICP-OES) analysis of plant essential elements. Analytical Methods, 3(12), 2854-2863. 27. WHO, 2011. Micronutrient Deficiencies, Iron Deficiency Anemia. Retrieved December 4, 2011 from: http://www.who.int/nutrition/topics/ida/en/index.html. 28. Xu, Y. G., Wang, B. S., Yu, J. J., Ao, G. M., & Zhao, Q. 2010. Cloning and characterisation of ZmZLP1, a gene encoding an endoplasmic reticulum-localised zinc transporter in Zea mays. Functional Plant Biology, 37(3), 194-205. 29. Yang, X., Xu, H., Li, W., Li, L., Sun, J., Li, Y., & Hu, Y. 2011. Screening and identification of seed-specific genes using digital differential display tools combined with microarray data from common wheat. BMC genomics, 12(1), 513. 30. Zhao, Z. Y., Che, P., Glassman, K., & Albertsen, M. 2019. Nutritionally Enhanced Sorghum for the Arid and Semiarid Tropical Areas of Africa. Sorghum, 197– 207. doi:10.1007/978-1-4939-9039-9_14

19

594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609

Table 1. Means (Table 1a) & variances (Table 1b) and heritability for grain Fe and Zn in sorghum (296B × PVK 801)-derived RIL population a) The means and standard deviation of the mean for Fe/Zn measured for parents and

RILs in individual environments Trait Fe (mg kg-1)

Zn (mg kg-1)

Environment ICRISAT 12-13 (E1) IIMR 12 -13 (E2) VNMKV 12-13 (E3) ICRISAT 13-14 (E4) IIMR 13-14 (E5) VNMKV 13-14 (E6) ICRISAT12 -13 (E1) IIMR 12-13 (E2) VNMKV 12 -13 (E3) ICRISAT 13-14 (E4) IIMR 13-14 (E5) VNMKV 13-14 (E6)

296B (P1) 28.00 28.50 46.33 26.00 30.80 27.24 21.32 21.00 26.44 14.63 21.19 19.69

PVK 801 (P2) 33.4 33.00 49.4 28.2 35.9 33.6 24.33 22.00 30.43 16.46 24.82 24.06

RILs 33.60 33.00 49.26 28.00 35.85 34.00 24.63 24.76 31.43 17.33 25.66 24.72

SD (±) 5.60 6.34 6.93 4.90 5.08 7.89 4.78 5.00 6.46 3.50 4.03 5.27

610 611 612 20

b) Variances and heritability for Fe and Zn in 296B × PVK 801 derived RIL population

613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630

Pooled (across six environments ) Trait

σ2 g

Fe

4.18**

SE (±) 0.69

Zn

4.17**

0.51

σ2 g

SE (±)

H2

Fe

ICRISAT 12-13 (E1) SE σ2 g H2 (±) 15.51** 1.68 0.78

20.79**

2.17

0.81

27.44**

2.85

0.8

Zn

10.00**

12.72**

1.33

0.8

23.41**

2.43

0.8

σ2gy

SE (±)

σ2gl

SE (±)

σ2gyl

SE

H2

-0.17

0.66

-0.7

0.80

14.32**

1.18

0.58

0.71**

0.35

-0.14

0.37

5.22**

0.54

0.69

Individual environments

1.12

0.74

ICRISAT 13 -14 (E4)

IIMR 13-14 (E5)

Fe

8.72**

Zn

5.48**

VNMKV 12-13 (E3) σ2 g

H2

SE (±)

VNMKV 13-14 (E6)

1.05

0.68

10.84**

1.3

0.68

31.36**

3.23

0.77

0.6

0.73

4.60**

0.68

0.56

12.71**

1.4

0.74

2

631 632 633

IIMR 12-13 (E2)

2

2

Genotypic variance( σ g), Genotype × Year (σ gy ) , Genotype × Location ( σ gl), Genotype × Year × Location ( σ2gyl ) interactions, standard error (SE) , heritability’s (H2, broad-sense), SD= Standard Deviation, all variances ** Significant at 1% level

634

Table 2. Pearson phenotypic and genotypic correlation for agronomic traits and grain Fe and Zn concentration in sorghum RIL population ICRISAT 12-13 (E1) Traits Zn DTF PH TW GY

Fe 0.88** (0.80**) 0.03 (0.04) 0.34** (0.29**) 0.34** (0.28**) - 0.41** (-0.32**)

IIMR 12-13 (E2)

Zn

GY

-

-

-0.08 (0.06) 0.30** (0.25**) 0.34** (0.30**) -0.34** (-0.27**)

-0.23** (-0.22**) -0.02 (0.01) -0.15** (-0.13*) -

Fe 0.73** (0.69**) -0.29** (-0.23**) 0.0008 (0.002) 0.09* (0.06) -0.33** (-0.28)

VNMKV 12-13 (E3)

Zn

GY

-

-

-0.17** (-0.1) 0.14* (0.12*) 0.35** (0.30**) -0.16** (-0.16**)

0.29** (0.23**) 0.47** (0.46**) 0.29** (0.26**) -

Fe 0.76** (0.70**) 0.04 (0.01) 0.34** (0.29**) -0.01 (0.01) -0.23** (-0.21**)

Zn

GY

-

-

-0.02 (0.02) 0.35** (0.30**) -0.04 (0.023) -0.33** (-0.29**)

-0.01 (0.01) 0.17** (0.16**) 0.05 (0.045) -

635

Traits Zn DTF

ICRISAT 13-14 (E4) Fe Zn 0.81** (0.72**) -0.16** -0.12** (-0.14*) (-0.11*)

GY 0.16** (0.13*)

IIMR 13-14 (E5) Fe Zn 0.83** (0.69**) -0.09 (-0.12* 0.07) (-

GY 0.15** (0.12*)

VNMKV 13-14 (E6) Fe Zn 0.71** (0.65**) 0.1 (0.11* 0.08) (0.08)

GY -0.04 (0.03)

Across Environment Fe Zn 0.88** (0.79**) -0.18** -0.19** (-0.12*) (-0.12*)

21

GY 0.25** (0.19**)

PH

0.12** (0.09)

0.28** (0.23**)

0.30** (0.28**)

0.08 (0.06)

0.11*) 0.03 (0.04)

0.3** (0.35**)

0.26** (0.22**)

0.37** (0.31**)

0.17** (0.14**)

0.43** (0.31**)

0.39** (0.33**)

TW

0.25** (0.20**)

0.47** (0.41**)

-0.17** (-0.12*)

0.12* (0.06)

0.03 (0.02)

-0.02 (0.01)

0.17** (0.13*)

0.21** (0.18**)

-0.06 (0.03)

0.56** (0.34**)

0.49** (0.36**)

GY

-0.41** (0.29**)

-0.31** (0.24**)

-

0.06 (0.06)

0.02 (0.06)

-

-0.22** (0.16**)

-0.22** (0.19**)

-

-0.50** (0.34**)

-0.45** (0.34**)

636 637 638 639

0.34** (0.31**) -0.21** (0.12**) -

DTF = Days to 50% flowering, PH= Plant Height (cm), TW= 100 seed weight (g), Fe= Iron (mg kg-1), Zn =Zinc (mg kg-1) and GY= Grain Yield (t/ha). Value in parenthesis is phenotypic correlation * Significant at 5% level; ** Significant at 1% level

640 641 642 643 644 645 646 647 648 649

Table 3: Description of identified QTLs for grain Fe and Zn concentration in the (296B × PVK 801) - derived RIL population from six environments (E1 to E6) and across the environments (pooled data of six environments) QTL Position (cM)

Flanking Markers

LOD

Addit ive effect

PVE (%)

SBI-06

90.4

Sn2647940 - Sn2657501

90.3 - 91.5

54954413 - 54555447

4.6

0.84

5.44

qfe7.1(E1)

SBI-07

42.9

Dt3625344 - Sn2644846

36.8 - 49.6

57701276 - 60259798

3.6

0.87

5.82

qfe7.2(E1)

SBI-07

123.5

Sn2653248 - Xtxp525

123.4 -126.7

4593373 - 2340000

4.3

0.81

5.09

qzn7.1(E1)

SBI-07

42.9

Dt3625344 - Sn2644846

36.8 - 49.4

57701276 - 60259798

5.6

0.88

9.42

qzn7.2(E1)

SBI-07

54.0

Dt2646020 - Sn1925068

53.0 - 54.7

56910000 - 57064983

6.8

0.86

8.80

qzn7.3(E1)

SBI-07

57.0

Sn1895281 - Sn2650637

57.0 - 57.9

56502177- 56741914

5.9

0.77

6.96

55123911 - 55156270

3.9

0.88

5.19

QTL name

Chr. No

qfe6.1(E1)

Genetic position of marker (cM)

Physical position marker (bp)

ICRISAT 12 -13 (E1)

VNMKV 12-13 (E3) qfe7.1(E3)

SBI-07

65.5

Dt3627294 - Dt2648007

65.4 - 65.6

qzn7.1(E3)

SBI-07

64.2

Sn1875097 - Xiabtp360

63.5 - 64.2

55324468 - 55900000

4.9

0.78

6.29

qzn7.2(E3)

SBI-07

67.2

Dt3628977 - Dt2649259

67.1 - 67.6

54659732 - 54837270

4.3

0.75

5.83

ICRISAT 13 -14 (E4)

22

qfe1.1(E4)

SBI-01

107.4

Sn1933402 - Sn2650907

107.3 - 107.6

20084094 - 20071525

5.5

0.79

6.80

55409292 - 55538220

4.4

0.99

5.66

5.6

0.6

6.7

4.5

0.53

5.7

4.5

0.54

5.7

VNMKV 13 -14 (E6) qfe7.1(E6)

SBI-07

61.0

Sn2645682 - Sn2653275

60.9 - 61.4

Across the environments (pooled data of six environments) Fe qfe7.2

SBI-07

123.5

Xtxp525 Sn2653248

123.4 - 126.5

2340000 4593373

Zn qzn7.4

SBI-07

67.2

qzn7.1

SBI-07

55.5

Dt2649259 Dt3628977 Dt2645576 Sn2033434

67.1 - 67.6 55.4 - 56.2

54837270 54659732 56915197 56741980

650 651 652 653 654

23

655 656

Table 4: Details of putative candidate genes identified for Fe and Zn concentrations (Pooled data across all locations) in identified QTL region in sorghum QTL Name

Locus ID Sb07g003000

Chr 7

From 3191492

To 3193591

Transcripts 1

Sb07g021130

7

54829074

54830555

1

Sb07g021910

7

56151914

56153758

1

PF00097 PTHR22764 AT4G30400.1 zinc finger (C3HC4-type RING finger) family protein

Sb07g022160

7

56501258

56504350

1

Sb07g022165

7

56514545

56515453

1

Sb07g022570

7

57198318

57199295

1

PF01485 PTHR11685 AT3G14250.1 protein binding / zinc ion binding PF01485 PTHR11685 AT3G14250.1 protein binding / zinc ion binding AT2G41940.1 ZFP8 ZFP8 (ZINC FINGER PROTEIN 8); nucleic acid binding / transcription factor/ zinc ion binding

qfe7.2

qzn7.4

qzn7.1

Locus Description PF00067 PTHR19383 KOG0156 AT3G26300.1 CYP71B34 CYP71B34; electron carrier/ heme binding / iron ion binding / monooxygenase/ oxygen binding PF10551 PF04434 PF00098 AT4G38180.1 FRS5 FRS5 (FAR1-related sequence 5); zinc ion binding

Reference Chen and Jia, 2006 (Wheat), Xu et al., 2010 (Maize),

Nozoye et al., 2007 (Rice), Mills et al., 2012 (Barley), Ricachenevsky et al., 2018 (Rice), Yang et al.,2011 (Wheat) Nozoye et al., 2007 (Rice), Mills et al., 2012 (Barley ), Ricachenevsky et al., 2018 (Rice), Yang et al.,2011 (Wheat) Nozoye et al., 2007 (Rice), Mills et al., 2012 (Barley ) Nozoye et al., 2007 (Rice), Mills et al., 2012 (Barley) Hindu et al.,2018 (Maize)

657

24

Highlights This is the first report on QTL mapping utilizing high-density of DArT and DArTseq markers for identifying QTLs for grain Fe and Zn in sorghum. Putative candidate genes involved in Fe/Zn metabolism were identified within QTL with highest phenotypic effect. Identified sorghum genes within QTL interval were used to evaluate gene synteny with other cereals Genomic regions and the candidate genes identified plays an essential role in markerassisted breeding.

Conflict of interest The authors declare that they have no conflict of interest