Journal of Experimental Botany, Vol. 53, No. 371, pp. 1163-1176,
May 2002
© 2002 Oxford University Press
Original Papers |
Phenotype/genotype associations for yield and salt tolerance in a barley mapping population segregating for two dwarfing genes
Scottish Crop Research Institute, Invergowrie, Dundee DD2 5DA, UK
Received 12 July 2001; Accepted 12 December 2001
 |
Abstract |
|---|
 Top Abstract IntroductionMaterials and methodsResultsDiscussionAppendixReferences |
|---|
Barley traits related to salt tolerance are mapped in a population segregating for a dwarfing gene associated with salt tolerance. Twelve quantitative trait loci (QTLs) were detected for seven seedling traits in doubled haploids from the spring barley cross DerkadoxB83-12/21/5 when given saline treatment in hydroponics. The location of QTLs for seedling growth stage (leaf appearance rate), stem weight prior to elongation, and tiller number are reported for the first time. In addition, four QTLs were found for the mature plant traits grain nitrogen and plot yield. In total, seven QTLs are co-located with the dwarfing genes sdw1, on chromosome 3H, and ari-e.GP, on chromosome 5H, including seedling leaf response (SGa) to gibberellic acid (GA3). QTLs controlling the growth of leaves (GS2) on chromosomes 2H and 3H and emergence of tillers (TN2) and grain yield were independent of the dwarfing genes. Field trials were grown in eastern Scotland and England to estimate yield and grain composition. A genetic map was used to compare the positions of QTLs for seedling traits with the location of QTLs for the mature plant traits. The results are discussed in relation to the study of barley physiology and the location of genes for dwarf habit and responses to GA.
Key words: Abiotic stress, barley, dwarf habit, QTL, genetic map, stable isotopes.
|
Introduction |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAppendixReferences |
|---|
In wild barley (Hordeum vulgare spp. spontaneum C. Koch), salt tolerance is associated with droughted environments (Pakniyat et al., 1997b). Genetic variation for potential water use efficiency was found in wild barley (Handley et al., 1994) and this could be utilized for the improvement of cultivars. However, to develop cultivars with drought and salt tolerance it is necessary to understand tolerance mechanisms and their direct or indirect effects upon development, yield and quality traits. This area of research has special relevance in view of the projected shortage of water for human consumption and the need to improve the efficiency of water use in agriculture (see http://www.worldbank.org/html/cgiar/press/ watatlas.html), as barley is the major cereal crop in saline and droughted environments.
The physiological basis of salt tolerance in non-halophytes (reviewed by Yeo, 1998) depends on a number of mechanisms. These include selective uptake of K+ (Dubcovsky et al., 1996), differential transport of Na+ and K+ (Garcia et al., 1997), the synthesis of osmoprotectants (Ishitani et al., 1995), cellular compartmentalization (Barkla and Pantoja, 1996), and plant vigour (Forster et al., 1990). Given this physiological complexity, it is important to distinguish between pleiotropic effects and the direct effects of genes that confer salt tolerance. Genetic mapping and the use of 
13C/15N to integrate diverse stress responses (Robinson et al., 2000) offers a possible approach to the resolution of this common dilemma.
13C values are interpretable in terms of a well-established physiological model (Farquhar et al., 1982). Discrimination of 13C/12C in C3 species is caused primarily by stomatal control of the entry of CO2 to leaves interacting with Rubisco carbon isotope discrimination. Conditions which induce stomatal closure (e.g. water stress, either direct or through salinity), restrict CO2 supply to carboxylation sites and can be detected by a reduced 13C/12C discrimination (13C). By contrast, no model yet exists to interpret plant 15N values in such simple terms (Robinson et al., 1998). Assimilatory 15N/14N fractionations occur, but how these vary with, for example, the form, concentration and 15N of external N source(s), N assimilation in roots versus shoots, and internal recycling and losses of N, is an active research area. Here, plant 15N is used as a trait which integrates plant N metabolism without precise knowledge of underlying mechanisms or function, nevertheless the genetic controls of 15N can be mapped. Generally, differences in plant 15N among plants grown on a common N source suggest differential loss of N to the environment and is therefore a potential marker for N retention (Robinson et al., 2000).
The natural abundances of carbon and nitrogen stable isotopes (13C/12C and 15N/14N) are used to study the physiology of salt tolerance in barley grown in hydroponics (Ellis et al., 1997a). Shoot 13C was significantly correlated with other measures of salt stress such as shoot Na+ concentration (Greenway and Munns, 1980) and shoot dry weight (Pakniyat et al., 1997a). Cultivated barley showed associations between ari-e mutations and salt tolerance (Pakniyat et al., 1997a, c). The ari-e mutants, which were grown in salt solution, showed significantly less reduction in shoot dry weight, a greater reduction in shoot [Na+] and more negative shoot 13C than their non-mutant, putatively isogenic parents in response to salt. By contrast, cultivars with the sdw1 dwarfing gene showed no significant differences from the genotypes from which they derive. The population of doubled haploids (DHs), from the cross DerkadoxB83-12/21/5, contains lines with both ari-e.GP and sdw1 and so provides an opportunity to study the effects and interactions of both these commercially important genes in a common background. This population was synthesized to allow studies applicable to the production of modern germplasm for Northern Britain and was designed to contrast for malting quality, yield and disease resistance characters (Thomas et al., 1998).
The advantages of using highly polymorphic DNA-based markers for genome analysis and the examination of gene associations are well established. Some examples in barley are the use of Amplified Fragment Length Polymorphisms (AFLPs) to illuminate relationships between cultivated barley and wild barleys (Ellis et al., 1997b; Pakniyat et al., 1997b) and the use of Simple Sequence Repeats (SSRs) in the study of genetic diversity (Powell et al., 1996), the analysis of the progress in integrating the mlo gene, arguably the most important single gene for spring barley, into modern cultivars (Thomas et al., 1998) and breeding progress in spring barley (Russell et al., 2000). In this paper the use of a genetic map is reported, based largely on AFLPs and SSRs, to identify quantitative trait loci (QTL) associated with developmental, physiological and salt tolerance-related traits in a population of DHs derived from the cross between Derkado and B83-12/21/5 and compare them to QTLs for grain yield.
|
Materials and methods |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAppendixReferences |
|---|
Genotyping and gene mapping
One hundred and fifty-six DHs were produced from a cross between Derkado (prostrate semi-dwarf phenotype sdw1/ Ari-e.GP) and the Scottish Crop Research Institute (SCRI) breeding line B83-12/21/5 (erect semi-dwarf phenotype Sdw1/ ari-e.GP) (Thomas et al., 1998) (sdw1 and ari-e.GP were previously known as denso and GPert, respectively). These were genotyped with a range of Amplified Fragment Length Polymorphism (AFLP), Sequence Specific Amplification Polymorphism (S-SAP) and Simple Sequence Repeat (SSR) markers. In addition, the DHs were scored phenotypically for the presence or absence of the sdw1 and ari-e.GP dwarfing genes, and the mlo powdery mildew resistance gene (loc.cit.). Since the genotyping reported previously (Thomas et al., 1998), the population was scored for Eph, a major gene locus controlling production of epi-heterodendrin (Swanston et al., 1999) and a further 41 SSRs, which detected 54 loci. In addition, a further nine AFLP polymorphisms were apparent when checking the gels, making an overall total of 241 markers covering all seven chromosomes of the barley genome.
JOINMAP 2.0 (Stam and van Ooijen, 1995) was used to fragment the marker data into genetic linkage groups and estimate the ordering of markers within groups and the distance between them. A JUMP value between 2.0 and 5.0 was used to establish whether or not to include markers in the second round of JOINMAP 2.0 and further checks were carried out to examine the consistency of the maps so formed.
MQTL (Tinker and Mather, 1995) was used to scan the linkage groups for QTLs using the map data from JOINMAP 2.0, but excluding any markers within 1 cM of another, i.e. giving a map of 1426 cM with an average distance of 11.2 cM between markers. Scans were carried out implementing the Simple Interval Mapping (SIM) and simplified Composite Interval Mapping (sCIM) options of MQTL. One thousand random permutations of the data were used to establish thresholds for 5% genome-wide error rates for QTL main effects and QTLxenvironment interactions. Primary QTLs were identified where SIM and sCIM peaks coincided and exceeded the SIM threshold. Secondary QTLs (Mather et al., 1997), were identified where either the SIM or the sCIM peak exceeded the SIM threshold for main effects (see Appendix). For secondary QTLxenvironment interactions, the additional restriction of finding a SIM or sCIM scan exceeding the SIM threshold in an individual environment at the same position was applied.
Field trials
The DHs, together with their parents, were grown in replicated and randomized yield trials at SCRI from 1994 to 1997 and Advanta Seeds UK Ltd grew similar trials at their site near Lincoln, UK in 1996 and 1997 making a total of six trials. Fungicides were used to prevent the build-up of foliar diseases. Whole plots from each trial were harvested, dried to a constant moisture and plot grain yield (PY) recorded. Samples of grain from each plot were cleaned and the fraction retained by a 2.5 mm sieve used for grain quality analysis, in which grain nitrogen concentration (G[N]) was assayed by the Dumas combustion method. Grain 15N (G15N) was determined as described earlier (Handley et al., 1993). Adverse harvest conditions at the Lincolnshire site in 1997 led to considerable preharvest sprouting so no quality analysis was carried out on the samples from that site.
Glasshouse experiments
Plant material
To study the reaction to salt, individual DHs from the population were germinated in tubes (Ellis et al., 1997a) at 20 °C for 7 d then transferred to a controlled vernalization environment maintained at 4 °C, lit by high-pressure sodium lamps at 300 µmol m-2 s-1 for 18 h d-1. After vernalization for 6 weeks, to minimize any minor variation in vernalization requirement, the tubes were put into a hydroponics in a glasshouse maintained between 16–24 °C with natural daylight supplemented with high pressure sodium lamps. External air was supplied continuously by ventilation and an extractor fan to maintain a consistent and stable 13C background.
Experimental design
The experimental design consisted of three replicates of a randomized block arranged so that each replicate, consisting of a control and a salt treatment, was contained within 12 hydroponic tanks. In each tank, 30 experimental plants were completely surrounded by 26 guard plants and the hydroponic solution was continuously and vigorously aerated. In each treatment, in addition to the 156 DHs, the parents Derkado and B83-12/21/5, Golden Promise, Maythorpe and a C4 control, Bermuda grass, were included. The remaining 19 positions in the sixth tank were use to grow ‘replacement plants’ as were six additional tanks, three for each treatment. Replacement plants were added after 1 week when plants showed poor establishment in hydroponics and less than 1% of the total were replaced with seedlings from the same batch.
Experimental system
The nutrient solution consisted of half-strength Hewitt's solution (Hewitt, 1966), with additional Na2SiO3 (Epstein, 1994), and was mixed in bulk before each weekly change. Electrical conductivity, pH and O2 concentrations of all solutions were monitored weekly. Solutions were replenished daily with distilled water to replace evapotranspiration loss. N was supplied as 12 mol m-3 N (as Ca(NO3)2 and KNO3). All plants received the same amount, concentration and isotopic composition of N.
After 14 d growth, NaCl was added to the salt treatments in equal daily increments for 5 d. The final NaCl concentration was 150 mol m-3 and this was maintained until harvest. The time at which the seedlings were harvested, 14 d from the completion of the NaCl addition, was before stem internode extension. The date of harvest was just after the control treatment started to show lodging and was timed to avoid the complication caused by extensive lodging. CaCl2 was added to maintain a Na+:Ca2+ concentration ratio of 20:1. Regular observations were made on plant establishment, the appearance of main-stem leaves (e.g. GS2 denotes the number of fully emerged main-stem leaves at the second time of observation) and tillers (e.g. TN2 was the number of tillers that emerged from the subtending leaf sheath at the second time of observation). Plant establishment (Est) was scored visually on a 1–9 scale (1 being complete, 9 dead) scored in the first week after transplantation when differences became apparent.
At harvest, plants were separated into shoots and roots, oven-dried and weights recorded (SWt and RWt, respectively). Root material was recovered but, because of extensive intermingling of roots in the hydroponic system, an unknown fraction of the root system of each plant was lost. These losses would have been randomly distributed throughout the experiment. Hence, the root weights (RWt) reported are conservative estimates, and it is not possible to calculate whole-plant-weighted averages for 13C or 15N (see below).
Shoots and roots were oven-dried, milled and analysed separately for total N and C concentration ([N] and [C]), 13C and 15N (Handley et al., 1993). Stable isotope abundances were expressed as =1000[(R/R*) -1]
, where R is the heavy:light isotope ratio of the sample and R* that of the appropriate standard.
Response to GA3
Seeds of the DHs and parents were germinated in Petri dishes and then transplanted into Sarstedt tubes filled with perlite and with holes drilled in the bottom. The tubes were suspended in hydroponic tanks and given a GA3 treatment as described earlier (Pakniyat et al., 1997a). After 10 d, shoot lengths were measured and the seedling gibberellic acid response (SGa), was estimated as the ratio of shoot length of a genotype in the GA3 treatment to that in the control.
Statistical analysis
DHs and parental means for PY, G[N] and G15N at each site were obtained after adjustment for any significant row and column effect after analysis by using the REML directive in GENSTAT. Overall population statistics for PY, G[N] and G15N were obtained after treating all the trials as randomized complete blocks as different randomizations and layouts had been used for each trial. Whilst the hydroponics experiment was designed as a randomized complete block, REML was also used to analyse it as there were missing values. Treatment means were derived by modelling entries, treatments and their interaction as fixed components. Estimates of variance components were then obtained by modelling all effects as random components to derive an approximate analysis of variance table and then re-estimating according to a fixed model. The estimate of the between DHs component 
b2 for each trait was taken as an estimate of the additive genetic variation (VA) and the heritability of each trait estimated as the proportion of the phenotypic variation accounted for by additive genetic variation. The ratio of the square of half the DHs range to the additive genetic variation was used to estimate k, the number of effective factors controlling a trait (Mather and Jinks, 1982). The overall means for each trait were used to derive correlations.
|
Results |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAppendixReferences |
|---|
DHs analysis
By the time the experiment was harvested the effect of salt treatment was visible as a slight change in leaf colour toward blue-green (glaucous), but otherwise the treated and control plants were visually similar. Metrical results are considered in relation to the statistical analysis of the results from the doubled haploid population, seedling growth habit and the linkage map extended from that of Thomas et al. (Thomas et al., 1998). This study concentrates on the primary QTLs while the secondary QTLs that were detected are given in the Appendix for completeness.
When data on all the DHs were combined over growth habits, significant genetical variation was found for 13 (Table 1
) of the 17 traits measured [the exceptions, data not given, being shoot C (S[C]), shoot N (S[N]), shoot 15N (S15N) and root C (R[C])]. In hydroponics, there was a significant difference between the control and salt treatments for nine of the 14 traits, but not in the early seedling traits; establishment (Est), growth stage count two (GS2) and tiller number count one (TN1) and in shoot composition (S[N] and S[C]). In the case of the observations on the early developmental traits it would appear that main-stem apical primordium development and growth are not sufficiently affected by the salt treatment in time to reduce number or rate of leaf and tiller appearance (Fig. 1a) (see below for an explanation of the plot). This was a result of the strategy of gradually increasing the salt stress to permit adaptation and avoid plant death, which can complicate the analysis of results by causing variation in inter-plant competition. When plants failed to establish, before the application of the salt stress, they were replaced by a plant of the same genotype from a ‘replacement nursery’ for the same reason.
|
|
The differences in means of the DHs and their parent cultivars are shown for selected traits in Fig. 1b
. For a normal distribution the mean of the DH population should be equidistant on a straight line between the maximum and minimum. In the absence of epistasis the mid-parent value is the same as the mean of the DH population. In general, the DH population shows a reasonably normal distribution, but it is noticeable that the mid-parent shows some deviation from the DH mean. This is particularly apparent for G[N] and PY (Fig. 1c), indicating the presence of epistasis where the deviation is highly significant (P<0.001). In the hydroponic experiment, significant epistasis was also detected for GS2 but it was less significant than for the field experiments. This may reflect the fact that the means for the latter are based on five or six environments compared with two for the former.
The salt treatment showed reduced shoot (SWt) and root dry weight (RWt), fewer tillers at the second count (TN2) and a greater root [N] (R[N]) (Fig. 1b). Shoot 13C (S13C) in salt treatments were less negative, while R15N was more positive. For all these traits the minimum and maximum values for the DH population were less or greater than the parental values. There was little evidence of interactions between main effects in the hydroponics experiment, being significant for only SWt, R13C and R[N] (Table 1). Significant interactions were detected in the field experiments for grain N content (G[N]) and plot grain yield (PY), but not grain 15N (G15N). These significant interactions were associated with a reduction in maximum shoot weight, a less negative maximum for root 13C but, most markedly, an increase in R[N] in B83-12/21/5 and in the DHs.
Heritability estimates (Table 1) of the traits in the hydroponics experiment were between 0.06 and 0.34 with the traits SWt and S13C being 0.20 and 0.34, respectively, indicating a useful gain in the utility of S13C over SWt for the selection purposes. From the field trials, heritabilities of G[N] and PY were reasonably high at 0.45 and 0.38, respectively, but that of G15N was extremely low at 0.02. Given the low heritabilities of some of the traits, it is not surprising that estimates of the numbers of effective factors for the traits in the hydroponics experiment were high, but it is surprising that the estimates for grain nitrogen content and plot yield were greater than 20.
One-third of the 153 possible correlations (Table 2) between all traits (only 12 traits shown) were significant at the 5% level (ABS (r) 
0.16), but few of these were high and exceeded the absolute value of 0.5. Most of the higher positive correlations (>+0.5) involved the tiller counts and the shoot and root weights. There were high negative correlations (<-0.5) between S[N] and S15N and between G[N] and PY, as expected for a single isotopic N source. This range of correlations is similar to those reported earlier (Thomas et al., 1998) for yield and its components in the same population and to those reported elsewhere, for example, for physiological traits (Hervé et al., 2001). When the traits were organized into a developmentally meaningful sequence (i.e. from Est to PY) it was obvious that the highest correlations were within groups of traits. Earlier life-cycle events may influence critically later development and growth, but may not result in high correlations because of threshold effects, for example, senescence of the apical primordial at the tip of the main-stem apex or the death of tillers after canopy closure.
|
Seedling growth habit and GA3 response
The two independent, recessive dwarfing genes produced four distinct seedling growth habits in field trials. The phenotypes semi-prostrate dwarf, erect dwarf, tall, and double dwarf, correspond to the genotypes; sdw1+Ari-e, Sdw1+ari-e, Sdw+Ari-e, and sdw1+ari-e, respectively (the ratio of these classes was 37:43:40:36, 
21:1:1:1=0.2, NS). Analysis of the differences between the means of the four classes of seedling growth habit (Table 3) showed highly significant differences for GS2, TN1, SWt, RWt, S13C, R15N, and R13C (P<0.001). Significant treatment effects were found for six traits (SWt, RWt, S13C, R[N], R15N, and R13C) with particularly significant Treatment with Growth Habit interactions for R[N] and R13C. The tall class showed the highest SWt and RWt (Fig. 2), but was not significantly greater than the semi-dwarf class. While the double dwarf class resulted in the smallest shoots the roots were no smaller than those of the erect class. Greater tiller number was not associated with higher shoot weight as the sdw1 genotype led to more tillers in combination with either Ari-e or ari-e, i.e. semi-prostrate and double dwarf, respectively. The increase in tiller number was associated with a change in nitrogen metabolism as both the sdw1+Ari-e and sdw1+ari-e types had the least negative root 15N. Shoot 13C was greatest for the tall class with little difference between the other classes or in any of the classes of root 13C. If root and shoot weight measurements had been made after stem elongation then the relationships that were found here might change.
|
|
The seedling response to GA3 treatment (SGa) showed a significant difference (P=0.05) between the semi-prostrate dwarfs (Treated/Control 1.49±0.061) and that of the other three genotypes (erect=1.27, double dwarf=1.30, tall=1.33) which were not significantly different.
Linkage map
At LOD 4.0 eight groups (Fig. 3) which contained six or more markers were formed by JOINMAP 2.0 and reference to the previously mapped SSRs in these groups showed that chromosome 3H consisted of two groups. Suspect markers were identified from high contributions to the linkage chi-square values and eliminated after the identification of unlikely double recombinants which were then verified and corrected if appropriate and map construction begun anew. During this process, chromosome 2H fragmented into two separate groups but, as the original data indicated linkage at LOD 4.0, the two groups were re-merged for mapping and the marker order was consistent with the grouping. One hundred and fifty-two markers were spread over the eight linkage groups, covering 1426 cM.
|
QTL analysis
Twenty-five markers mapping within 1 cM of another were eliminated from the linkage maps, resulting in 127 markers suitable for QTL analysis giving an average spacing of 11.2 cM between markers. A stratified sample of 39 markers were designated background markers to absorb the effects of any neighbouring QTLs. These were spaced fairly evenly along the genome at an approximate interval of 36.5 cM. Thus, when scanning for QTLs, only peaks that were more than 35 cM apart were included. The same QTL often showed both a main effect and an interaction with environment and the one that gave the highest Test Statistic was declared as the QTL. From the scans produced by MQTL, primary QTLs (Table 4) were found for all the traits that had a heritability greater than 0.10 (SWt, S13C, G[N], and PY).
|
Over all the traits, 16 primary QTLs were revealed whose location can be postulated with a high level of confidence (Fig. 3
). The highest number of primary QTLs found for a single trait was three for SWt and G[N]. The multi-locus models of the fitted QTLs generally accounted for the same amount of the phenotypic variation as the sum of the individual loci (Table 4). The models never accounted for more than 35% of the total phenotypic variation, the highest being 31% for main effects and interactions for S13C. In general, the amount of variation accounted for by the QTLs in the multi-locus models for the traits reflected the amount of genetic variation detected for them.
The largest individual QTL effects were associated with the chromosomal regions around the two dwarfing genes sdw1 (3H) and ari-e.GP (5H) that were segregating in the DHs. From Table 4, it can be seen that sdw1 was associated with QTL main effects for S13C, R15N, and G[N]. The effect of the Derkado allele for the QTL located at the sdw1 dwarfing gene locus was to make the estimates of 13C more negative with significant effects of the Derkado allele in both control and salt treatments. The QTL accounted for 14% of the total phenotypic variation in 13C. For R15N the Derkado allele at the sdw1 QTL increased isotopic discrimination, but the effect was only significant in the salt treatment. This implies that an alteration in growth pattern that delayed internode elongation was only important in the stressed treatment. This QTL accounted for 6% of the total phenotypic variation in R15N. The effect of the Derkado allele at the sdw1 QTL for G[N] was to reduce grain nitrogen levels, usually interpreted as a ‘dilution effect’ related to an increase in yield associated with the sdw1 gene. In this study, the sdw1 gene also resulted in an overall yield increase but was only detected as a secondary QTL (see Appendix), although it exceeded the SIM and sCIM thresholds in one environment and the SIM threshold (LOD=2.6) in two others. The consistency of the negative effects for all five field trials of the Derkado allele (sdw1) for G[N] offer support for the secondary yield QTL that was detected.
By contrast to the sdw1 gene, the ari-e.GP gene was associated with QTL main effects only for S13C, response to GA3 (SGa) and a QTLxenvironment interaction for PY while the adjacent locus Bmag337 showed main effects for SWt and RWt. The Derkado QTL allele at the ari-e dwarfing gene locus resulted in effects for S13C that were positive, i.e. the plants were more stressed. In this case, the ari-e Derkado allele conditions the tall or semi-prostrate dwarf phenotype depending on the allelic constitution at sdw1, so the values for the ari-e.GP phenotype would be -0.43 in the control and -0.52 in the salt treatment, indicating an improvement of salt tolerance of 48% over the Derkado phenotype determined by sdw1. The increased response to GA3 application indicated that the Derkado allele at ari-e.GP complements a non-responsive allele in the semi-dwarf genotype. The result should, however, be treated with caution until verified by more detailed studies. They are presented to illustrate a possible mechanism for the effects of ari-e.GP. A reduction in cell size, reported for barley mutants (Stoy and Hagberg, 1967), would account for this behaviour of the ari-e.GP dwarf. For PY, however, a crossover interaction was apparent as the Derkado QTL allele (tall phenotype) at the dwarfing gene locus increased yield or had no significant effect in the SCRI trials, but gave significant yield decreases in the trials grown in Lincolnshire. When the interactions are included this particular locus accounted for 7% of the phenotypic variation.
The SSR Bmag337 maps within 10 cM of ari-e.GP so the apparent separation in mapping of QTLs for plant and root weight has to be treated with caution. At first sight the separation of plant and root weight from the dwarfing gene may seem illogical, but in this study the concern is with seedling traits before stem extension has occurred. The Derkado QTL allele that is mapped to Bmag337 increases stem weight, particularly in the Control, but there is a much lower effect on root weight. The stem weight QTL allele accounts for 10% of the phenotypic variation. Stem weight is also affected by QTL alleles that map to P21M12d on 2H and P17M62f on chromosome 4H that account for 5% and 7% of the total phenotypic variation, respectively. The number of main-stem leaves at the second count (GS2) was associated with two QTLs; the first located 23 cM from Bmac144d on chromosome 2H, showed significant effects for the Derkado QTL allele in both the control and the salt treatment, indicating that leaf emergence rates were increased by salt stress. The second QTL at Bmag318 on chromosome 3H only showed a significant effect for the Derkado QTL allele in salt treatment. These QTLs accounted for 8% and 6%, respectively, of the total phenotypic variation in the rate of leaf appearance. The number of tillers at the second count (TN2) was associated with the AFLP marker P40M38b on chromosome 7H that accounted for 6% of the total phenotypic variation. As for the GS2 QTL on chromosome 3H a significant effect of the Derkado allele resulted from the salt treatment.
The ‘–E’ option of MQTL allows the scanning of the rest of the genome for any di-genic epistatic interactions involving primary loci. When this was done, three instances were found for SWt and R15N in the hydroponic experiment and for PY in the field experiments. In each case, the interaction was, however, with loci that were not detected as primary QTLs and, in the case of SWt and R15N, nor had they been detected as secondary QTLs (see Appendix). For PY, there was a highly significant interaction between the primary QTL at ari-e.GP and the secondary QTL at sdw1 (LOD=4.8). From Table 5, it can be seen that there is considerable deviance from the genotype means with epistasis, compared to those without, so that the double dwarf group (D/B) was expected to be the highest in the absence of epistasis but only the third highest in its presence. Significant epistasis was found for R15N (LOD=3.2) where the effects were also dramatic with the B/B group changing from a very low value without epistasis to the second highest with epistasis. Significant epistasis was also found for SWt (LOD=3.2), but the effects were less marked and confined to order changes among the top 3. Biometrical analysis had only detected epistasis for PY, but had not revealed any for R15N or SWt. The discrepancies may reflect some of the assumptions in biometrical analyses that were not met in practice and the fact that the search for QTL epistasis was confined only to those involving primary QTLs.
|
|
Discussion |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAppendixReferences |
|---|
A combination of biometrical and quantitative trait analyses indicates, as expected, that where the number of effective factors is low more primary QTLs can be located, for example, GS2 and S
13C versus R15N. It has been possible to separate and pinpoint over the whole barley genome loci that affect plant development and reaction to salt stress. It is reported that (1) there are contrasting effects of the ari-e.GP and sdw1 dwarfing genes for a range of traits including stable isotope discrimination for C and N; (2) there are three novel QTLs for plant development (GS2, TN2); and (3) there are two QTLs for yield (on chromosomes 4H and 6H) that were independent of the dwarfing genes Barley, of the common cereals, is relatively salt tolerant, but the varying responses of cultivars with significant effects on plant growth (Ellis et al., 1997a) indicate the possibility of an increase in tolerance by breeding. For example, Schaller et al. reported that cv. California Mariout, derived from Egyptian barley germplasm, was the most tolerant line in their screening programme (Schaller et al., 1981). More recent work in Egypt has shown that the new cultivar Giza123 is more tolerant than cv. California Mariout (I Ahmed, personal communication).
The complexity of barley responses to salt has been widely studied during germination, in seedlings and plants, but there is still scope for confusion between true tolerance and escape due to varying plant vigour and differences in flowering time and grain maturity. The improvement of genetical analysis techniques has contributed to more precise analysis of salt tolerance. Genes for abiotic stress tolerance in the Triticeae (i.e. wheat, barley, rye, and related species) were located primarily on group 4 and group 5 chromosomes by the analysis of aneuploid chromosome stocks (Forster et al., 1990; Forster, 1992). Mano and Takeda found QTLs for salt tolerance at germination on chromosomes 1H, 4H, 5H, and 6H and, at different locations, for seedling tolerance of salt on chromosome 1H, 2H, 5H, and 6H (Mano and Takeda, 1997). The genetic control of barley shoot weight in the LinaxHS92 population (Ellis et al., 1997a) showed significant effects associated with all chromosomes except 2H with the largest effect on 5H. Shoot 13C in the same population also showed significant effects on five of the seven chromosomes, but not 3H nor 1H, with the largest effects associated with chromosome 2H and 5H. In DHs from DerkadoxB83-12/21/5 the largest effect is associated with chromosome 5H close to the ari-e.GP locus which was mapped earlier (Thomas et al., 1984). The dwarfing genes were not segregating in the LinaxHS92 population and a less stringent QTL detection methodology was applied, so it is not surprising that there are some differences between those results and the current study. The current methods for QTL detection have known limitations (Kearsey and Farquhar, 1988), but even so they have contributed to advances in understanding the genetic control of a wide range of traits in barley (Thomas, 2000). Further studies are required using techniques such as the development of chromosome substitution lines, the discovery of markers in database sequences (Pillen et al., 2000; Ramsay et al., 2000) or the use of homology between barley and rice or Arabidopsis sequences. For example, a salt tolerance mechanism similar to that of Arabidopsis (Zhu, 2000) would involve genes encoding a protein kinase (Shen et al., 2001), a Na+/H+ antiporter and a specific plasmalemma receptor. The possibility of discovering such mechanisms in Hordeum is indicated by greater than 80% homology between the sequence for AtNHX1 and database sequences for a barley leaf library (D Marshall, personal communication).
The traits establishment, the second leaf count, and both tiller counts showed no effect of salt treatment, although there were significant genotypic differences between the DHs. It was concluded that these traits were determined before the treatment was effective. Establishment reflects the vigour of seedling growth during vernalization and the successful establishment of transplants which occurred before salt concentrations became critical. Leaf counts were made at weekly intervals, as the experiment was too large for a higher frequency. The lack of a salt effect indicated that leaf expansion was not inhibited by the salt treatment, as expected from recent results (Munns et al., 2000), despite leaves becoming slightly glaucous by the end of the experiment. Leaf length was not measured, but the visual appearance of the control and treated plants indicated no large effect on leaf length. Similarly, the emergence of tillers in the leaf axils was not delayed, but the experiment was not continued long enough to permit observation of the continuing metabolic effects of salt. Leaf appearance rates can be affected by temperature and day length (Kirby and Ellis, 1980) but, by comparison with field trials, the short time-scale of the experiment meant that the change in daylength was small and growth in the glasshouse removed the correlated temperature change.
Shoot 13C was more heritable than the other seedling traits examined. Given that 13C integrates a range of processes susceptible to alteration by stress, this level of heritability is remarkably high and the fact that this is similar to that found in wheat (Ehdaie and Waines, 1994) suggests a general property for all the Triticeae. The implication is that a reasonable proportion of the phenotypic variation in plant 13C can be genetically manipulated. However, the relationship between 13C and barley yield is complex and early attempts to develop 13C as a direct assay for yield in barley have not been successful (Acevedo, 1993). The reasons for this are illuminated by the results from genome mapping, as only one of the three primary QTLs for yield is associated with a primary QTL for S13C giving the potential for a wide range of recombinants from a cross between contrasting parents.
This analysis suggests that there is little potential to manipulate shoot 15N genetically, but that there is a better prospect of manipulating root 15N, although its heritability is low indicating that phenotypic selection may be ineffective and strategies similar to those developed for improving yield would have to be adopted. Studies of wild barley populations (Robinson et al., 2000) indicate that, under stress, plant 15N changed as much as 13C. Interestingly, when ranked according to their 13C and 15N values under stress, far more wild genotypes showed positive changes in rank relative to their control treatments for 15N than for 13C. On that basis, there may be considerable opportunity for finding genotypes, which show large responses in terms of 15N/14N discrimination when grown in stressful environments. Comparisons of diverse genotypes may help to reveal the physiological mechanisms involved in N isotopic discrimination.
The major dwarfing genes found in commercial cultivars of barley cause marked changes in phenotype with, apparently pleiotropic, effects on grain size, grain composition, malting quality, and yield (Thomas et al., 1991). The outstandingly successful cultivar Golden Promise (the source of ari-e.GP) possessed a unique combination of traits; rapid establishment, short straw, early maturity, compact erect ear, short lemma awn, and ready threshing, all apparently conditioned by a single mutation at the ari-e locus (Franckowiak, 1997). Since 1980 this model has been superseded by cultivars that carry the sdw1 gene that permits an increase in grain size that is impossible in the compact ear of the ari-e genotype.
The mode of action of barley dwarfing genes has been investigated in a number of studies. Blonstein and Gale showed that changes in cell size in mutants in cv. Proctor were associated with a modification of growth rate and the extent of internode extension that varied with each internode (Blonstein and Gale, 1984). As this study was carried out in a controlled environment it is not clear how far the results can be generalized. In addition, Blonstein and Gale did not map their mutants so, while they describe erect and semi-prostrate types, their conclusions cannot be directly related to sdw1 or ari-e.GP (Blonstein and Gale, 1984). Franckowiak and Pecio assessed the reaction of a collection of dwarf barley lines to the application of gibberellic acid (GA3) (Franckowiak and Pecio, 1992). Most lines with the ari-e genotype, i.e. an ari-e mutant in Bonus, cv. Clansman and cv. Golden Promise were scored as constitutive with similar elongation in control and GA3 treatment, cv. Fleet in contrast was scored as GA-sensitive. The sdw1 genotypes showed a majority of lines with GA sensitivity (68%), but 24% were classified as constitutive and 8% as insensitive. This apparent confusion may arise for any one of a number of reasons; the cultivar names attributed to the samples may have been wrong, the seedling phenotype may be wrongly classified as delayed lower stem internode elongation is also typical of a high vernalization requirement, the timing of GA3 application may be critical or the amount and type of GA may cause variation in response (Chandler and Robertson, 1999). While a large number of dwarf mutants have been described in barley and more than 20 distinct loci have been mapped (Franckowiak, 1997) only ari-e and sdw1 have achieved commercial significance.
The characterization of the mode of action of dwarfing genes in wheat has been more extensive than in barley. It has been shown that the dwarf alleles Rht-B1 and Rht-D1 in wheat were orthologues of the maize dwarf-8 and Arabidopsis Gibberellin Insensitive genes and a mechanism was proposed involving phosphotyrosine participation in gibberellin signalling (Peng et al., 1999). The dwarf mutants in barley, which are recessive indicating a loss of function, could be useful ‘probes’ of barley metabolism. This approach was used by Chandler and Robertson who described three classes of mutant in cv. Himalaya (Chandler and Robertson, 1999). Two classes, grd (GA responsive dwarf) and gse (GA sensitive) were epistatic to the mutant slender (sln), a GA constitutive response mutant, while a third class (elo) that lacked the ability to elongate, was hypostatic. Ivandic et al. mapped a gibberellin-insensitive mutant, Dwf2, in cv. Bonus proximal to the SSR XhvOle on chromosome 4H which was aligned with the Rht-B1 and Rht-D1 loci in wheat (Ivandic et al., 1999). The SSR HvOLE was not polymorphic in DerkadoxB83-12/21/5, but was mapped in LinaxH. spontaneum Canada Park to 24 cM distal from HVM3 (Ramsay et al., 2000). A simplistic interpretation would place Dwf2 just distal to P16M72f on this study's map and excludes its involvement, as the non-mutant allele, in the stem weight QTL that is located 18 cM proximal to this marker. There are four dwarfing gene loci mapped on chromosome 2H; two on the barley morphological map sld2, slender dwarf 2, and hcm1, short culm (Franckowiak, 1997) and, in addition, two loci, gai, gibberellin-insensitive, and gal, gibberellin-less were added (Börner et al., 1999). From a comparison of other maps (Börner et al., 1999; Costa et al., 2001) with this study's map, it appears that the SSR Bmac144b could map closely to gal and Bmac134 is more than 40 cM distal from the RFLP MWG557. The GS2 QTL locus the authors map to 23 cM from Bmac144 d is therefore independent of gal and the stem weight locus at P21m12 d does not overlap with gai. It would appear that there is no direct involvement of loci where changes in GA metabolism are known to occur and the QTLs reported on chromosomes 2H and 4H are novel loci. In addition, it is concluded that sdw1 is a mutant of the gse class and that ari-e.GP is a mutant of the elo type as leaf elongation is limited by a reduction in cell size. These conclusions will be confirmed by definitive studies that dissect the fine control of seedling development and growth at the molecular level.
|
Appendix |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAppendixReferences |
|---|
Location of QTL peaks and estimates of effects where secondary QTLs were detected. Trait names in italics indicate interaction effects in contrast to main effects.
|
|
|
Acknowledgements |
|---|
The Scottish Crop Research Institute is funded by grant in aid from the Scottish Executive Environment and Rural Affairs Department.
|
Footnotes |
|---|
1 To whom correspondence should be sent. Fax: +44 (0)1382 568507. E-mail: t.ellis{at}scri.sari.ac.uk
2 Present address: Department of Plant and Soil Science, University of Aberdeen, Aberdeen AB24 3UU, UK.
3 Present address: Max-Planck-Institute of Molecular Plant Physiology, 14424 Potsdam, Germany.
|
References |
|---|
|
TopAbstractIntroductionMaterials and methodsResultsDiscussionAppendixReferences |
|---|
Acevedo E. 1993. Potential of carbon isotope discrimination as a selection criterion in barley breeding. In: Ehleringer JR, Hall AE, Farquhar GD, eds. Stable isotopes and plant carbon–water relations. London: Academic Press, 399–417.
Barkla BJ, Pantoja O. 1996. Physiology of ion transport across the tonoplast of higher plants. Annual Review of Plant Physiology and Plant Molecular Biology 47, 159–184.[Web of Science]
Börner A, Korzun V, Malyshev S, Ivandic V. 1999. Molecular mapping of two dwarfing genes differing in their GA response on chromosome 2H of barley. Theoretical and Applied Genetics 99, 607–675.
Blonstein A, Gale M. 1984. In: Semi-dwarf cereal mutants and their use in cross breeding. IAEA TECDOC 307, Vienna: International Atomic Energy Authority.
Chandler P, Robertson M. 1999. Gibberellin dose-response curves and the characterization of dwarf mutants of barley. Plant Physiology 120, 623–632.
Costa JM, Corey A, Hayes PM, Jobet C, Kleinhofs A, Kopisch-Obusch A, Sato K, Szucs P, Toojinda T, Vales MI, Wolfe RI. 2001. Molecular mapping of the Oregon Wolfe barleys: a phenotypically polymorphic double-haploid population. Theoretical and Applied Genetics 103, 415–424.
Dubcovsky J, Santa Maria G, Epstein E, Luo MC, Dvorak J. 1996. Mapping of the K+/Na+ discrimination locus Kna1 in wheat. Theoretical and Applied Genetics 92, 448–454.
Ellis RP, Forster BP, Waugh R, Bonar N, Handley LL, Robinson D, Gordon D, Powell W. 1997a. Mapping physiological traits in barley. New Phytologist 137, 149–157.
Ellis RP, Powell W, McNicol JW, Baird E, Booth A, Lawrence P. 1997b. The use of AFLPs to examine genetic relatedness in barley. Molecular Breeding 3, 359–369.
Ehdaie B, Waines JG. 1994. Genetic analysis of carbon isotope discrimination and agronomic characters in a bread wheat cross. Theoretical and Applied Genetics 88, 1023–1028.
Epstein E. 1994. The anomaly of silicon in plant biology. Proceedings of the National Academy of Sciences, USA 91, 11–17.
Farquhar GD, O'Leary MH, Berry JA. 1982. On the relationship between carbon isotope discrimination and the intercellular carbon dioxide concentration in leaves. Australian Journal of Plant Physiology 9, 121–137.[Web of Science]
Forster BP. 1992. Genetic engineering for stress tolerance in the Triticeae. In: Powell W, Hilman JR, eds. Opportunities and problems in plant biotechnology. Proceedings of the Royal Society of Edinburgh Section B (Biological Sciences) 99, 89–106.
Forster BP, Phillips MS, Miller TE, Baird E, Powell W. 1990. Chromosome location of genes controlling tolerance to salt (NaCl) and vigour in Hordeum vulgare and H. chilense. Heredity 65, 99–107.
Franckowiak JD. 1997. Revised linkage maps for morphological markers in barley, Hordeum vulgare. Barley Genetics Newsletter 26, 22–23.
Franckowiak JD, Pecio A. 1992. Entries in the collection of semidwarf barley stocks held at North Dakota State University. Barley Genetics Newsletter 21, 116–126.
Garcia A, Rizzo CA, Ud-Din J, Bartos SL, Senadhira D, Flowers TJ, Yeo AR. 1997. Sodium and potassium transport to the xylem are inherited independently in rice and the mechanism of sodium:potassium differs between rice and wheat. Plant, Cell and Environment 20, 1167–1174.
Greenway H, Munns RA. 1980. Mechanisms of salt tolerance in non-halophytes. Annual Review of Plant Physiology 31, 149–190.[Web of Science]
Handley LL, Daft J, Scrimgeour CM, Ingleby K, Sattar MA. 1993. Effects of the ecto- and VA-mycorrhizal fungi Hydnagium carneum and Glomus clarum on the 15N and 13C values of Eucalyptus globulus and Ricinus communis. Plant, Cell and Environment 16, 375–382.
Handley LL, Nevo E, Raven JA, Martinez-Carrasco R, Scrimgeour CM, Pakniyat H, Forster BP. 1994. Chromosome 4 controls potential water use efficiency (13C) in barley. Journal of Experimental Botany 45, 1661–1663.
Hervé D, Fabre F, Berrios EF, Leroux N, Chaarani GA, Planchon C, Sarrafi A, Gentzbittel L. 2001. QTL analysis of photosynthesis and water status traits in sunflower (Helianthus annuus L.) under greenhouse conditions. Journal of Experimental Botany 52, 1857–1864.
Hewitt EJ. 1966. Sand and water culture methods used in the study of plant nutrition. Farnham: Commonwealth Agricultural Bureau.
Ivandic V, Mayshev S, Korzun V, Graner A, Börner A. 1999. Comparative mappingof a gibberellic acid-insensitive dwarfing gene (Dwf2) on chromosome 4HS in barley. Theoretical and Applied Genetics 98, 728–731.
Ishitani M, Nakamura T, Han SY, Takabe T. 1995. Expression of betaine aldehyde dehydrogenase gene in barley in response to osmotic stress and abscisic acid. Plant Molecular Biology 27, 307–315.[Web of Science][Medline]
Kearsey MJ, Farquhar AGL. 1988. QTL analysis in plants; where are we now? Heredity 80, 137–142.
Kirby EJM, Ellis RP. 1980. A comparison of spring barley grown in England and Scotland. 1. Shoot apex development. Journal of Agricultural Science, Cambridge 95, 101–110.
Mano Y, Takeda K. 1997. Mapping quantitative trait loci for salt tolerance at germination and at the seedling stage of barley (Hordeum vulgare L.). Euphytica 94, 263–272.
Mather K, Jinks JL. 1982. Biometrical genetics. London, New York: Chapman and Hall.
Mather DE, Tinker NA, Laberge DE, Edny M, Jones CL, Rossnagel BG, Legge WG, Briggs KG, Irvine RB, Falk DE, Kasha KJ. 1997. Regions of the genome that affect grain quality in a North-American two-row barley cross. Crop Science 37, 544–554.
Munns R, Passioura JB, JianMin G, Chazen O, Cramer GR. 2000. Water relations and leaf expansion: importance of time scale. Journal of Experimental Botany 51, 1495–1504.
Pakniyat H, Handley LL, Thomas WTB, Connolly T, Macaulay M, Caligari PDS, Forster BP. 1997a. Comparison of shoot dry weight, Na+ content and 13C values of ari-e and other semi-dwarf barley mutants under salt stress. Euphytica 94, 7–14.
Pakniyat H, Powell W, Baird E, Handley LL, Robinson D, Scrimgeour CM, Nevo E, Hackett CA, Caligari PDS, Forster BP. 1997b. AFLP variation in wild barley (Hordeum spontaneum C. Koch) with reference to salt tolerance and associated ecogeography. Genome 40, 332–341.
Pakniyat H, Thomas WTB, Caligari PDS, Forster BP. 1997c. Comparison of salt tolerance of GPert and non-GPert barleys. Plant Breeding 116, 189–191.
Peng J, Richards DE, Hartley NM, Murphy GP, Devos KM, Flintham JE, Beales J, Fish LJ, Worland AJ, Pelica F, Sudhakar D, Christou P, Snape JW, Gale MD, Harberd NP. 1999. ‘Green revolution’ genes encode mutant gibberellin response modulators. Nature 400, 256–261.[Medline]
Pillen K, Binder A, Kreuzkam B, Ramsay L, Waugh R, Förster J, Leon J. 2000. Mapping new EMBL-derived barley microsatellites and their use in differentiating German barley cultivars. Theoretical and Applied Genetics 101, 652–660.[Web of Science]
Powell W, Machray GC, Provan J. 1996. Polymorphism revealed by simple sequence repeats. Trends in Plant Science 1, 215–222.
Ramsay L, Macaulay M, Ivanissevich SD, MacLean K, Cardle L, Fuller J, Edwards KJ, Tuvesson S, Morgante M, Massari A, Maestri E, Marmiroli N, Sjakste T, Ganal M, Powell W, Waugh R. 2000. A simple sequence repeat-based linkage map of barley. Genetics 156, 1997–2005.
Robinson D, Handley LL, Scrimgeour CM. 1998. A theory for 15N/14N fractionation in nitrate-grown vascular plants. Planta 205, 397–406.
Robinson D, Handley LL, Scrimgeour CM, Gordon DC, Forster BP, Ellis RP. 2000. Using stable isotope natural abundances (15N and 13C) to integrate the stress responses of wild barley (Hordeum spontaneum C. Koch.) genotypes. Journal of Experimental Botany 51, 41–50.
Russell JR, Ellis RP, Thomas WTB, Waugh R, Provan J, Booth A, Fuller J, Lawrence P, Young G, Powell W. 2000. A retrospective analysis of spring barley germplasm development from ‘foundation genotypes’ to currently successful cultivars. Molecular Breeding 6, 553–568.
Schaller CW, Berdegue JA, Dennett CW, Richards RA. 1981. Screening world barley collection for salt tolerance. Barley genetics IV. Proceedings of the Fourth International Barley Genetics Symposium, Edinburgh, 389–393.
Shen W, Gomez-Cadenas A, Routly EL, Ho T-DH, Simmonds JA, Gulick PJ. 2001. The salt stress-inducible protein kinase gene, Esi47, from the salt-tolerant wheatgrass Lophopyrum elongatum is involved in plant hormone signalling. Plant Physiology 125, 1429–1441.
Stam P, Van Ooijen JW. 1995. JOINMAPTM version 2.0: software for the calculation of genetic linkage maps. CPRO–DLO, Wageningen.
Stoy V, Hagberg A. 1967. Effects of growth regulators on ear density mutants of barley. Hereditas 58, 359–384.
Swanston JS, Thomas WTB, Powell W, Young G, Lawrence PE, Ramsay L, Waugh R. 1999. Using molecular markers to determine barleys most suitable for malt whisky distilling. Molecular Breeding 5, 103–109.
Thomas WTB. 2000. Molecular biological tools in cereal breeding. In: Morris PC, Bryce JH, eds. Cereal biotechnology. Cambridge, UK: Woodend Publishing, 107–136.
Thomas WTB, Baird E, Fuller JD, Lawrence P, Young G, Russell JR, Ramsay L, Waugh R, Powell W. 1998. Identification of a QTL decreasing yield in barley linked to Mlo powdery mildew resistance. Molecular Breeding 4, 381–393.
Thomas WTB, Powell W, Swanston JS. 1991. The effects of major genes on quantitatively varying charaters in barley. 4. The GPert and denso loci and quality characters. Heredity 66, 381–389.
Thomas WTB, Powell W, Wood W. 1984. The chromosomal location of the dwarfing gene present in the spring barley variety Golden Promise. Heredity 53, 177–183.
Tinker NA, Mather DE. 1995. MQTL software for simplified composite interval mapping of QTL in multiple environments. Journal of Quantitative Trait Loci http//probe.nalusda.gov8000/otherdocs/jqtl/1995-2/jqtl16r2.html
Yeo A. 1998. Molecular biology of salt tolerance in the context of whole-plant physiology. Journal of Experimental Botany 49, 915–929.
Zhu JK. 2000. Genetic analysis of plant salt tolerance using Arabidopsis. Plant Physiology 124, 941–948.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||