Introduction

Preharvest sprouting (PHS) can be a problem in barley (Hordeum vulgare L.) production, especially in malting barley (Schwarz et al. 2004). Rain or even very high humidity from near physiological maturity (PM) onward can cause sprouting in the head of standing plants in the field. This has very serious consequences for malting grain, since rapid, uniform, and complete germination is critical in the malthouse. Seed dormancy defined as the failure of viable mature seed to germinate under favorable conditions can also be a problem. Dormancy is sometimes viewed as the opposite trait to PHS. In fact, dormancy has been defined as resistance to PHS (Barr et al. 2003). However, it is not clear whether dormancy and PHS are represented by opposite alleles for a given gene.

Dormancy/non-dormancy is generally inferred by measuring germination percentage. PHS, on the other hand, is generally measured more directly by visual sprout score (SSc) and/or alpha-amylase activity (AA) directly or indirectly (e.g., falling number score). Both dormancy and PHS may be measured at some time or times from PM for up to several weeks of after-ripening.

Dormancy quantitative trait loci (QTLs) were previously mapped in several barley crosses including the US Barley Genome Project crosses of ‘Steptoe’/’Morex’(S/M) (Ullrich et al. 1993; Han et al. 1996, 1999; Gao et al. 2003), and ‘Harrington’/TR306 (Takeda 1996; Ullrich et al. 2002), as well as, ‘Chebec’/Harrington (Li et al. 2003), ‘Triumph’/Morex (Prada et al. 2004), and ‘Mona’/ Wadi Qilt 23–39 (Hordeum vulgare spontaneum) (Zhang et al. 2005).

Dormancy QTLs have been mapped to all seven barley chromosomes. Relatively, major effect (>10% variation explained) and/or consistent QTL regions have been identified across environments and genotypes/crosses on chromosomes 1H, 4H, 5H, and 7H (Ullrich et al. 1993; Han et al. 1996; Takeda 1996; Li et al. 2003; Prada et al. 2004). Specifically in the S/M cross, one relatively consistent QTL region each was identified on chromosomes 4H, 7H, and two on 5H (Ullrich et al. 1993; Han et al. 1996).

Relatively little research has been done to characterize the genetics of PHS potential per se in barley. QTL analysis of PHS as measured by AA after significant rainfall near harvest of a field grown Chebec/Harrington mapping population revealed a major effect QTL near the telomere of chromosome 5HL and a minor effect one on chromosome 2H (Li et al. 2003, 2004). King and Wettstein-Knowles (2000) showed that morphological features of epicuticular waxes and spike angle at maturity can affect ear wetting and PHS in barley and wheat (Triticum aestivum L. and T. durum L.).

An objective of a major ongoing study in our program is to determine if the barley germplasm developed and QTLs discovered in previous research on dormancy can be applied or related to the genetic control of PHS. This paper reports results from the S/M cross, which represents dormant (Steptoe) and non-dormant (Morex) six-row spring types. Our hypothesis was that some PHS QTLs will be the same as previously determined dormancy QTLs and some PHS QTLs will be different from or more important than dormancy QTLs.

Materials and methods

Grain production

The Steptoe (dormant type)/Morex (non-dormant type) mapping population consisting of 150 F1-derived doubled haploid (DH) lines, the parents, and molecular marker map developed from this population were utilized in this study (original: Kleinhofs et al. 1993; current: http://barleygenomics.wsu.edu/). The S/M DHs and parents were grown in one greenhouse and one field environment 2003–2005. Standard barley growing conditions of 20/15°C, day/night temperatures and 16 h photoperiod were used in the greenhouse. Standard barley growing conditions and two 3 m long row plots were used in the field at Spillman Agronomy Farm, Pullman, WA. Fifteen heads were harvested of each genotype in both growing environments at PM (when all green color was lost from the head) and 14 days later. All heads were frozen at −20°C immediately after harvest until analyses were performed.

Preharvest sprouting (PHS) simulation and measurement

To determine susceptibility to PHS, a modification of SSc procedures of DePauw and McCaig (1991) were used. Replication of five and ten heads of each line and parent from greenhouse and field, respectively, harvested at 0 and 14-day post-PM were treated in a mist chamber for 5 days or until Morex, the sprout susceptible parent, showed 80% sprouted heads.

The degree of PHS was determined visually and rated on a 0–5 scale; 0 = no sprouting; 1 = ≤3 roots, no shoots; 2 = ≤5 roots, ≤3 shoots; 3 = ≤8 roots, ≤5 shoots; 4 = >25% < 50% roots and shoots; 5 = >50% roots and shoots.

Roots and shoots were counted if they were ≥2 mm long. Mean SScs of the five or ten heads rated were used for genetic analysis.

Alpha-amylase activity analysis

Alpha-amylase is synthesized de novo at the onset of germination. Therefore, a measure of alpha-amylase enzyme activity gives a measure of germination in seeds even with no visible sprouting. After PHS simulation, all heads were dried at 70°C for 48 h and threshed by hand. Grain of genotypes with mean SScs of ≤1 was milled and enzyme activity determined in duplicate to achieve 10% agreement using the Megazyme Alpha-Amylase kit (http://www.megzyme.com/booklets/KCERA.pdf). Analyses were performed on greenhouse and field grown material at 0 and 14-day post-PM, respectively.

Germination test

To determine degree of dormancy, two replications of 50 seeds from 0 day and 14 days harvests from each environment were used for germination using standard procedures in Petri dishes with moist filter paper in a germinator set at 20°C for 7 days. Means of the two replications were used for genetic analysis.

Genetic analysis

The mean phenotypic data [SSc, AA, and germination percentage (GP)] for each harvest date from each growing environment were subjected to QTL analysis using MapManager QTX (Manly et al. 2001; www.mapmanager.org/mmqtx.html). Five hundred permutation tests were performed on the data to establish Likelihood Ratio Statistic (LRS) thresholds at 1 and 5% probability levels. Regression, simple interval mapping (SIM, no background control for other QTLs) and composite interval mapping (CIM, background control for other QTLs) functions were utilized. SIM identified the most significant QTL in a given trait analysis and subsequent QTLs were usually identified by CIM. SIM and CIM QTL results significant at P ≤ 0.05 are emphasized. Significant (P ≤ 0.05) QTLs indicated by regression analysis not confirmed with a significant LRS from interval mapping are treated as indicators of the potential for the existence of minor QTLs only.

Results and discussion

Phenotypic data

Steptoe, the dormant and presumed sprout resistant parent and Morex, the non-dormant and presumed sprout susceptible parent behaved as expected for each trait measured. Steptoe had zero or low SScs (0–0.2), AA (0.012–0.034 Ug−1 flour), and germination percentages (2–21%), while Morex had relatively high levels of these traits (1.1–5 SSc and 58–100% GP), Tables 1 and 2; Fig. 1. AA was not measured for Morex samples, because SScs were above the threshold for AA measurement (≤1), and they would have been extremely high. The distribution patterns for the mapping population were continuous with population means between the two parents and ranges extending to or beyond the parents (Tables 1, 2). The frequency distributions for each trait from the field environment display and characterize the continuous distribution of the mapping population (Fig. 1). The distribution patterns from the greenhouse environment (data not shown) are nearly identical to those from the field. The parents and mapping population in this study (Fig. 1c, f) behaved similarly to that in the initial dormancy study in terms of germination percentage distributions (Ullrich et al. 1993). The phenotypes of the parents and the DHLs of the mapping population confirm extensive polymorphism for each trait studied providing excellent potential for mapping relevant QTLs for each of the traits.

Table 1 Preharvest sprouting and dormancy mean ± SE phenotypic data for parents and doubled haploid (DH) line mapping population, greenhouse environment
Full size table
Table 2 Preharvest sprouting and dormancy mean ± SE phenotypic data for parents and doubled haploid (DH) line mapping population, field environment
Full size table
Fig. 1
figure 1

Frequency distributions for the preharvest sprouting (sprout score, SSc; alpha-amylase activity, AA) and dormancy (germination percentage, GP) traits measured at 0 day (a–c) and 14 days (d–f) after physiological maturity from the field environment in the Steptoe/Morex doubled haploid mapping population and parents

Full size image

QTL analysis

A relatively high number of QTLs were indicated for the various traits spread across all seven barley chromosomes measured by regression analysis (data not shown), which is similar to the results of the initial dormancy study with the S/M population (Ullrich et al. 1993). These results are also similar to the bread wheat situation in which QTLs for PHS/dormancy have been mapped to 20 of the 21 chromosomes (Kulwal et al. 2005). A number of QTL clusters were indicated in the current study, which means that QTLs for different traits overlap one another in the same chromosome region. Within a given cluster, this may indicate that one gene affects more than one trait (pleiotropy) and/or several genes for the given traits are closely linked. Many of the putative QTLs had relatively small effects, but they explained from 3 to 39% of the variation observed for a given trait. This indicates there is considerable complexity in the inheritance of these traits and there are relationships among the traits.

Simple interval mapping and CIM procedures were applied to verify the significance of QTLs indicated by regression analysis. Only those QTL regions that were verified as significant with SIM or CIM were considered as ‘real’ and are listed in Table 3. This table includes all QTLs significant by regression, as well as, SIM and CIM within each verified chromosome region. However, those QTLs with trait effects significant by regression analysis only are considered as indications of very minor genes. In all cases Morex contributed the alleles for low dormancy (high germination percentage) and high sprouting potential (high SSc), except for the QTL region on chromosome 2H (see below).

Table 3 QTL regions affecting preharvest sprouting and dormancy in the Steptoe/Morex (S/M) cross
Full size table

Sprout score

Sprout score QTLs verified with SIM or CIM were mapped to chromosomes, 2HS, 3HS, 4HS, 5HC and L and 7HC. The chromosome 2H, 4H, and 7H QTLs had relatively minor effects explaining only 5 or 6% of the variation by SIM/CIM, which translates to 0.4 SSc change (additive value). The chromosome 3H QTL in the 12.9 cM interval between ABG399A and DD1.1B had SIM/CIM effects ranging from 5 to 15%, which affects up to 0.6 SSc (Table 3; Fig. 2a). This corresponds to a very minor dormancy QTL identified from only one of eight environments in the original S/M dormancy study (Ullrich et al. 1993). Whereas, there may be minor dormancy QTL effects in this chromosome 3H region, there seems to be greater effects on PHS. Both PHS and dormancy QTLs have been implicated on the group 3 chromosomes in common wheat, but the long arm seems to be more important than the short arm in wheat (Miura et al. 2002; Kulwal et al. 2005).

Fig. 2
figure 2

Likelihood Ratio Statistic (LRS) plots based on simple or composite interval mapping with MapManager QTX of the Steptoe/Morex doubled haploid mapping population for (a) sprout score (SSc) at 0-day post-physiological maturity (PPM) from the greenhouse environment (G) on chromosome 3H, LRS 37 at peak; (b) SSc at 0 day PPM from G on chromosome 5H, LRS 54 at peak; (c) germination percentage (GP) at 0-day PPM from the field environment (F) on chromosome 3H, LRS 19 at peak; (d) GP at 0-day PPM from G on chromosome 5H, LRS 61, and 15 at peaks

Full size image

The greatest QTL effects on SSc were found on chromosome 5H, specifically in the 9.2, 11.2, and 13.7 cM intervals between CDO348B and ABC302A, WG908 and ABG495A, and ABG463 and Tel7L (note: current map notation = Tel5H), respectively (Table 3; Fig. 2b). The chromosome 5H QTL effects on SSc ranged from 6 to 31% determined by SIM/CIM. The QTL in the chromosome 5HC CDO348B–ABC302A interval had the greatest effect explaining up to 31% of the variation with an additive affect of 0.9 sprout score (SSc 0 day greenhouse). The effects of the WG908–ABG495A interval ranged from 6 to 10%. The effects of the chromosome 5HL telomere region interval ranged from 12 to 16% (Table 3). The chromosome 5H QTL regions also had relatively strong QTLs for AA and germination percentage, as well (see below).

Alpha-amylase activity

Alpha-amylase activity QTLs, verified by SIM or CIM, occur on chromosomes 2HS (WG516–DAK 213A), 3HC (ABG399A–DD1.1B), 4HS (MWG634–JS103.3), 5HC (CDO348B–ABC302A), 5HL (WG908–ABG495A), and 5HL (ABG463–Tel7L) with effects explaining 5–27% of the variation or changing enzyme activity by 0.49–1.19 U/g flour (Table 3). The chromosome 2HS, 3HC, and 4HS QTLs are all relatively minor in effect (7–8%), which was the case for SSc and germination percentage in the chromosome 2HS and 4HS intervals, as well. The chromosome 3HC AA QTL was in the same interval as SSc and germination percentage QTLs, but with smaller effects (Table 3). All three chromosome 5H QTL intervals from the centromere to long arm telomere contain relatively major AA QTLs explaining up to 27, 19, and 12% of the variation, respectively. These QTLs had the greatest affect on AA and they also overlap with SSc and germination percentage QTLs. The chromosome 4HS PHS AA QTL coincides with a previously mapped malt AA QTL (Hayes et al. 1993; Han and Ullrich 1994). These PHS and malt AA QTLs may also coincide with group 4 chromosome dormancy QTLs in common wheat (Kato et al. 2001). All AA QTLs identified in the present study coincide with previously identified dormancy QTLs in the S/M mapping population, but the chromosome 2H, 3H, and 5HL (WG908–ABG495A) dormancy QTLs were very minor and inconsistently expressing in the original studies (Ullrich et al. 1993; Han et al. 1996). In the two-row/two-row Chebec/Harrington cross, Li et al. (2003) identified a major PHS QTL based on AA in the ABG463–Tel7L interval of chromosome 5H.

Germination percentage

Dormancy QTLs as measured by germination percentage were mapped (SIM and/or CIM) to six of the seven barley chromosomes (Table 3). Relatively minor QTLs were indicated on chromosomes 1HL, 2HS, 4HS, 5HL (the WG908–ABG495A interval), and 7HC. These minor QTLs explained from 3 to 7% variation or affected about 5 to 7% germination. The chromosome 4H and 7H QTLs were previously determined to be minor (5% variation explained), but somewhat consistently expressed across environments (Ullrich et al. 1993; Han et al. 1996). These QTLs were named SD3 and SD4 for the chromosome 7H and 4H intervals, respectively (Han et al. 1996). The chromosome 2H and 5HL (WG908–ABG495A) QTLs were also previously identified as minor, and detected in only one or two of the eight environments evaluated (Ullrich et al. 1993). The very minor QTL on chromosome 1H was not previously indicated in the S/M cross. The chromosome 3HC QTL region had a 10% effect on dormancy (Table 3; Fig. 2c) and coincided with a PHS QTL as indicated above.

The largest effect QTLs occur on the chromosome 5H intervals, CDO348B–ABC302A with 16–39% effects (which changed germination percentage by about 7–20 points) and ABG463–Tel7L with 11–18% effects (which changed germination percentage by about 11–13 points), Table 3 and Fig. 2d. Again these two chromosome 5H regions coincide with PHS QTLs. These two dormancy QTLs were previously named SD1 (5HL centromere region) and SD2 (5HL telomere region) (Han et al. 1996). These chromosome 5H results are very consistent in terms of location and effect with previous interval mapping (Ullrich et al. 1993; Oberthur et al. 1995), marker assisted selection (Han et al. 1996), and fine mapping (Han et al. 1999; Gao et al. 2003) results in the S/M cross. Furthermore, both chromosome 5H dormancy QTLs were identified in the two-row/six-row Triumph/Morex cross (Prada et al. 2004) and two-row domesticated/two-row wild barley Mona/Wadi Qilt 23–39 (H. vulgare spontaneum) cross (Zhang et al. 2005). The ABG463–Tel7L QTL was previously identified in the two-row/two-row crosses of Harrington/TR306 (Takeda 1996; Ullrich et al. 2002) and Chebec/Harrington (Li et al. 2003), as well. Fine mapping of the SD1 QTL region revealed that it potentially is a complex of at least three genes occupying ∼4.4 cM between ABC324 and ABC302A (Han et al. 1999). Fine mapping the SD2 QTL revealed its position to be in the 0.7 cM interval between MWG851D and MWG851B very near the long arm telomere (Gao et al. 2003). Romagosa et al. (1999) showed that SD1 was important in determining the time of dormancy release and acts additively with SD2 in the later stages of after-ripening. This information may be relevant in explaining the relationship between dormancy and PHS potential. The SSc QTL in the SD2 QTL region was much more significant at the 14 days after-ripening measurement than the 0 day measurement. Li et al. (2004) found good synteny between the barley chromosome 5HL telomere QTL region with the chromosome 3L telomere region of rice (Oryza sativa). This led to the identification of a gene encoding Gibberellic acid-oxidase, which they proposed as a candidate gene for the chromosome 5HL seed dormancy/PHS QTL. Gibberellic acid may be involved in the release of dormancy through promoting the synthesis of hydrolytic and proteolytic enzymes associated with germination (Adkins et al. 2002).

Summary and conclusions

Considering the previously published results and those from this study, some tentative, conclusions may be drawn about the inheritance of PHS susceptibility and resistance and relationship with seed dormancy in barley. Two regions on chromosome 5H near the centromere and long arm telomere play major roles in both PHS potential and dormancy (Table 3; Fig. 2b, d). One or both regions have been consistently implicated in dormancy and/or PHS in all published barley crosses studied (Ullrich et al. 1993, 2002; Takeda 1996; Li et al. 2003; Prada et al. 2004; Zhang et al. 2005).

The chromosome 3H ABG399A–DD1.1B and chromosome 5H WG908–ABG495A QTL regions appear to be more important for control of PHS than for dormancy (Table 3). These regions showed only minor or no dormancy QTL potential in this and previous studies on barley (Ullrich et al. 1993; Han et al. 1996).

In addition to the major QTLs summarized above for PHS and dormancy, a relatively large number of minor QTLs are indicated throughout the genome (Table 3). This appears to be evidence for considerable complexity in the inheritance of these traits. A few major genes may dominate, but are influenced by potentially many other genes. This conclusion is supported by the continuous frequency distributions for the various traits (Fig. 1), as well as similar findings in other studies of these traits in barley, wheat and rice (Paterson and Sorrells 1990; Anderson et al. 1993; Ullrich et al. 1993; Wan et al. 2005). In this study the variation explained by any one QTL was a maximum of 39% and ranged down to 3% (Table 3). Evidence for additivity and epistasis among and between large- and small-effect QTLs for dormancy has been presented in previous studies in barley and wheat (Anderson et al. 1993; Oberthur et al. 1995; Romagosa et al. 1999; Kulwal et al. 2005; Mares et al. 2005).

Preharvest sprouting and dormancy are physiologically related and appear to be genetically related, as well, based on the common map locations in most instances for the two traits (Table 3; Fig. 2). Whereas, linked gene clusters may be an explanation, pleiotropy is just as likely based on the relatedness of traits (Li et al. 2003) and precedence set by fine mapping of related malting quality traits on barley chromosomes 4H (Gao et al. 2004) and 7H (Han et al. 2004). Actually, these fine mapping studies support both the occurrence of pleiotropy and linked genes in the relationship of related malting quality traits. Some of the overlapping QTLs detected in this study were not always resolved to the exact same peak within the support interval. The effects of a given QTL interval on the various traits were not usually equal either. The flanking marker and support intervals are relatively large for most of the QTLs detected in this study as have been the case in other studies for dormancy and PHS (Ullrich et al. 1993; Li et al. 2003; Prada et al. 2004; Zhang et al. 2005). There is certainly room for multiple genes affecting the same and related traits. Fine mapping of SD1, the largest-effect dormancy QTL in the S/M cross-revealed a cluster of at least three dormancy genes resolved to a 4.4 cM interval within the CDO324–ABC302A interval on chromosome 5H (Han et al. 1999). The fine mapping of SD2, the second largest-effect dormancy QTL in the S/M cross-near the chromosome 5HL telomere revealed a single gene promoting dormancy during seed development in a 0.7 cM region between MWG851D and MWG851B and the possibility of a second gene proximal to this interval that reduces dormancy during after-ripening (Gao et al. 2003). It is likely the actual number of genes is greater than the number of QTLs identified and that pleiotropy, epistasis, and clusters of linked genes all play roles in the expression of both PHS and dormancy and the relationship between the two traits.

This study adds to the growing body of knowledge about the inheritance of PHS and dormancy and the relationship between these two somewhat opposite traits in barley and cereals in general. Besides gaining a better understanding of these traits and their relationships, the identification of markers that are linked to chromosome regions that affect these traits presents opportunities for marker-assisted selection, in further genetic study, and in breeding improved cultivars. Extremes of both PHS and dormancy can be detrimental to crop production and end use quality of barley and other cereals. A moderate level of dormancy and/or PHS resistance is desirable in a cultivar. The choice of QTLs or genes is expanding to accomplish a desirable balance between PHS and dormancy. From experience thus far, the chromosome 5HL telomere QTL with a moderate effect up to 21% on PHS and up to 18% on dormancy may be the best choice for six-row barley. This QTL has consistently expressed over genotypes and environments. However, in two-row barley germplasm with the Harrington background that carries the sprout susceptible allele, this region also affects a number of positive malting quality QTL alleles, which may make it difficult to maintain or improve malting quality with reduced PHS potential (Li et al. 2003). Other QTLs that show potential, but probably require more study are the chromosome 3HC ABG399A–DD1.1B QTL and the chromosome 5HL WG908–ABG495A QTL that showed up to 15 and 19% effects on PHS, respectively, and up to 10 and 6% effects on dormancy, respectively.