Skip Navigation


JXB Advance Access originally published online on February 14, 2005
Journal of Experimental Botany 2005 56(413):945-958; doi:10.1093/jxb/eri087
This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
56/413/945    most recent
eri087v1
Right arrow E-letters: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when E-letters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (8)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Thévenot, C.
Right arrow Articles by Prioul, J.-L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Thévenot, C.
Right arrow Articles by Prioul, J.-L.
Agricola
Right arrow Articles by Thévenot, C.
Right arrow Articles by Prioul, J.-L.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?

© The Author [2005]. Published by Oxford University Press [on behalf of the Society for Experimental Biology]. All rights reserved. For Permissions, please email: journals.permissions@oupjournals.org

RESEARCH PAPER

QTLs for enzyme activities and soluble carbohydrates involved in starch accumulation during grain filling in maize

Claudine Thévenot1,*, Elizabeth Simond-Côte1, Agnès Reyss1, Domenica Manicacci2, Jacques Trouverie1, Martine Le Guilloux2, Valérie Ginhoux2, Françoise Sidicina2 and Jean-Louis Prioul1

1Laboratoire Structure et Métabolisme des Plantes, Institut de Biotechnologie des Plantes, Bâtiment 630 (UMR 8618/CNRS, UPS), Université Paris-Sud, F-91405 Orsay Cedex, France
2Station de Génétique Végétale, (UMR 8120/INRA, INA P-G, CNRS, UPS), Ferme du Moulon, F-91190 Gif-sur-Yvette, France

* To whom correspondence should be addressed. Fax: +33 1 6915 6424. E-mail: thevenot{at}ibp.u-psud.fr

Received 11 May 2004; Accepted 17 November 2004


   Abstract
Top
 Abstract
Introduction
 Materials and methods
 Results
 Discussion
 References
 
ADPglucose, the essential substrate for starch synthesis, is synthesized in maize by a pathway involving at least invertases, sucrose synthase, and ADPglucose pyrophosphorylase, as shown by the starch-deficient mutants, mn1, sh1, and bt2 or sh2, respectively. To improve understanding of the relationship between early grain-filling traits and carbohydrate composition in mature grain, QTLs linked to soluble invertase, sucrose synthase, and ADPglucose pyrophosphorylase activities and to starch, sucrose, fructose, and glucose concentrations were investigated. In order to take into account the specific time-course of each enzyme activity during grain filling, sampling was carried out at three periods (15, 25, and 35 d after pollination) on 100 lines from a recombinant inbred family, grown in the field. The MQTL method associated with QTL interaction analysis revealed numerous QTLs for all traits, but only one QTL was consistently observed at the three sampling periods. Some chromosome zones were heavily labelled, forming clusters of QTLs. Numerous possible candidate genes of the starch synthetic pathway co-located with QTLs. Four QTLs were found close to the locus Sh1 (bin 9.01) coding for the sucrose synthase. In order to confirm the importance of this locus, the CAPS polymorphism of the Sh1 gene was analysed in 45 genetically unrelated maize lines from various geographical origins. The DNA polymorphism was significantly associated with phenotypic traits related to grain filling (starch and amylose content, grain matter, and ADPglucose pyrophosphorylase activity at 35 DAP). Thus, the Sh1 locus could provide a physiologically pertinent marker for maize selection.

Key words: AGPase, DNA polymorphism, maize, QTL, Sh1, soluble invertases, soluble sugars, starch, sucrose synthase


   Introduction
Top
 Abstract
 Introduction
Materials and methods
 Results
 Discussion
 References
 
The composition and the technological properties of mature grains result from a series of synthesis and maturation processes which take place several weeks before harvest. In maize, this is especially true for starch accumulation since the enzymes involved are mostly expressed during the first 35 d after pollination (DAP). The invertase activities are maximum at 3–4 DAP and decline continuously (Ou-Lee and Setter, 1985Go). ADPglucose pyrophosphorylase (AGPase) activity becomes detectable in the endosperm at 12–15 DAP, peaks at 25–35 DAP depending on the genotype, and then declines to a very low value (Prioul et al., 1990Go, 1994Go). Sucrose synthase activity (SUSY) follows the same pattern as AGPase, except that the decline is less pronounced (Prioul et al., 1990Go). In all cases these activities may be considered as nil at grain harvest (70–80 DAP). Mutations affecting genes encoding the three above-mentioned enzymes severely reduce starch accumulation, namely, miniature (mn1) for cell wall invertase, shrunken 1 (sh1) for SUSY, and shrunken 2 (sh2) or brittle 2 (bt2) for AGPase. Such mutations demonstrate that these enzymes are mandatory steps in the synthesis of ADPglucose and that ADPglucose is the essential precursor for starch-synthesizing enzymes. However, the relative contribution of each step in the metabolic pathway is not easy to establish. Using genetic variability in enzyme activity and carbohydrate concentration is a way to determine the main contributing loci, i.e. QTLs (Quantitative Trait Loci) (see review by Prioul et al., 1997Go). Such an approach has been used successfully for carbon metabolism in leaves (Causse et al., 1995Go; Pelleschi et al., 1999Go) showing the co-location of QTLs corresponding to the structural genes for some of the measured enzyme activities, and thus supporting their role as possible candidate genes (Prioul et al., 1999Go). As any quantitative trait may be used for QTL detection, many traits related to grain yield and composition have been examined in numerous situations. Concerning mature maize grains, QTLs were detected for starch, lipid, and protein contents (Goldman et al., 1993Go; Séne et al., 2000Go) and for technological properties like hardness, and dry milling traits (Sourdille et al., 1996Go; Séne et al., 2001Go).

The present study aimed at detecting QTLs for enzyme activities and carbohydrate concentrations, at three developmental stages, in order to try to relate the QTLs to those already described for mature grain composition and technological properties of the same recombinant inbred line population (Séne et al., 2000Go, 2001Go). The co-location of the observed QTLs with the markers of known gene function mapped by Causse et al. (1996)Go was analysed in order to find candidate genes. After which, the DNA polymorphism of one candidate gene was analysed by CAPS in a series of genetically unrelated lines and this was compared with phenotypic traits.


   Materials and methods
Top
 Abstract
 Introduction
 Materials and methods
Results
 Discussion
 References
 
Plant material
The same population of maize recombinant inbred lines (RILs) was used as in previous studies (Séne et al., 2000Go, 2001Go). It was derived from a cross between two contrasting genotypes, a late dent line from the American Iodent group MBS847 and a French early flint line, encoded F-2. A total of 145 RILs were used for mapping 152 RFLP probes, including a majority of expressed sequences (Causse et al., 1996Go). One-hundred RILs from the 6th selfed generation were field-grown at Gif-sur-Yvette (20 km SW Paris, France) during the summer of 1994, in a randomized complete block design with two replications. Ten plants from each line were planted in a row. Ears from three plants, in each block, were collected at each phenological stages, 15, 25, or 35 d after pollination (DAP); pollination occurred from 23 July to 6 August depending on the lines, and six kernels per ear were immediately frozen and stored in liquid nitrogen for biochemical analyses. These sampling dates were chosen because previous studies (Prioul et al., 1994Go) have shown that starch synthesis in the endosperm begins from 12–15 DAP and is maximum between 25–35 DAP. In addition, this maximum may be genotype-dependent. At 15 DAP, nine traits were measured, whereas at 25 and 35 DAP, because of the lack of variability between the parents for some enzyme activities, only six traits were chosen for measurement. The values obtained for all the traits measured, at the three stages, in the two parent lines and in seven RILs chosen at random and harvested in each block, showed no block effect; it was decided to use only kernels harvested in block 1. Composite samples of 18 frozen kernels for 15 DAP and of nine frozen kernels for 25 DAP and 35 DAP, from three plants of the same RIL, harvested at the same stage, in block 1, were made, by pooling six or three kernels per ear. The accuracy of biochemical measurements was evaluated by comparing variability in the triplicates to total trait variability using the heritability formula (broad sense). H2 values from 0.91 to 0.99 were found whatever the trait.

In addition, a panel of 44 genetically unrelated inbred lines from different origins and representative of maize genetic diversity was used, including (a) 23 dent lines F113, F252, F271, F277, F284, F292, F584, F604, F608, F618, F752, LH146, LH52, LH74, LH82, MBS847, MBS979, A188, A654, Co158, W117, W401, and CM174, (b) seven flint lines F2, F268, F283, F7, Lo32, Du101, and F591, (c) six dent/flint lines Co255, F1852, F476, F544, F670, and F7001, (d) four tropical lines CML239, CML243, CML245, and CML246, (e) one South-American Plata line Argl256, (f) two popcorn lines C6 and Ia5pop, and (g) one floury line Coest6. These inbred lines were grown in the same fields as the recombinant inbred lines (experiment 1) and were harvested at maturity. Two additional sets of these 45 inbred lines were grown in outdoor fields in France, near Toulouse (experiment 2a) and at Saint Martin de Hinx (experiment 2b), in order to repeat the evaluation of starch, amylose, and protein kernel contents. Finally, a set of these lines were grown in a glasshouse at Gif-sur-Yvette (France, experiment 3); kernels were collected at 35 DAP and stored in liquid nitrogen to measure AGPase activity.

Crude extract preparation
Three independent extracts from each composite sample were obtained by grinding decoated grains (six grains for 15 DAP or three grains for 25 and 35 DAP) in a frozen mortar containing fine sand (<10 µm). Extraction buffer (50 mM HEPES-NaOH pH 7, 10 mM MgCl2, 1 mM Na2-EDTA, 2.6 µM DTT, 0.02% Triton X100, 1% BSA) was added to a final volume of four times the grain fresh matter. After centrifugation (5 min, 15 000 g, 4 °C), the supernatant (crude extract) was immediately used for measurements of enzyme activities and soluble carbohydrate concentrations. The pellet was stored for measurement of starch concentration.

ADPglucose pyrophosphorylase activity
A 10 µl volume of crude extract was added to a final 1 ml reaction mixture containing 50 mM HEPES-NaOH pH 7, 2 mM MgCl2, 1 mM Na2-EDTA, 1.15 mM ADP glucose, 0.34 mM NAD+, 2 units phosphoglucomutase, and 2 units glucose 6-phosphate dehydrogenase and placed in a spectrophotometer cuvette at 30 °C. After stabilization for approximately 3 min (stable baseline), 1 mM PPI was added. The time-course of NAD+ reduction was measured at 340 nm over a 3 min period, in the linear part of the time-dependent absorbance increase (Bergmeyer and Bernt, 1974Go). Results were expressed in nkat g–1 fresh matter (kat=mol s–1).

Sucrose synthase and soluble invertase activities
Sucrose synthase and soluble invertase activities were measured using an aliquot of the crude extract which was desalted by passage through a G25 Sephadex column equilibrated with 50 mM HEPES-NaOH pH 7, 2 mM MgCl2, 1 mM Na2-EDTA, 2.6 µM DTT, 1% BSA. The two invertase activities were measured using their different optimum pH. Their specificity was checked, over a 2–9 pH range using the specific inhibition of alkaline invertase in TRIS-buffer. The time-course of neutral invertase activity (50 µl desalted extract) was measured at 30 °C in a 860 ml final volume reaction buffer containing 50 mM HEPES-NaOH pH 7, 2 mM MgCl2, 1 mM Na2-EDTA, 100 mM sucrose, 2.6 µM DTT, 1 mM ATP, 0.44 mM NAD+, and a desalted coupling enzyme mixture (3.5 units phospho-glucose isomerase, 2 units glucose 6-phosphate dehydrogenase and 4.2 units hexokinase). The time-course of NAD+ reduction was recorded at 340 nm over a 3 min period in the linear part of the time-dependent increase of absorbance. Sucrose synthase activity was then measured by adding 1 mM UDP to the same cuvette and measuring further NAD+ reduction; the obtained value corresponded to the sum of neutral invertase and sucrose synthase activities. The acid invertase activity of the same desalted extract (25 µl aliquot) was determined by lowering the pH with 0.2 mM Na acetate buffer pH 4.8, before adding 10 µl of 600 mM sucrose at 30 °C for 15 min. The pH was then increased by adding 50 µl of 500 mM NaH2PO4 buffer, pH 7, and samples were immediately boiled for 3 min in order to denature enzymes. Glucose and fructose derived from sucrose hydrolysis were transformed by adding 750 µl reaction buffer (50 mM HEPES-NaOH pH 7, 2 mM MgCl2, and 1 mM Na2-EDTA) containing 1 mM ATP, 0.44 mM NAD+ and the desalted coupling enzyme mixture (see above). After incubation at 30 °C for 30 min and centrifugation at 12 000 g for 3 min, the NADH formed was measured at 340 nm (Bergmeyer and Bernt, 1974Go). Results were expressed in nkat g–1 fresh matter.

Carbohydrate contents
Carbohydrate contents were determined by enzymatic assays, as above. The supernatant (crude extract) was used for measurements of soluble carbohydrate contents, and the pellet for starch content. The supernatant was boiled for 3 min, in order to denature proteins, and centrifuged at 12 000 g for 5 min. The total soluble sugar content was determined from a 25 µl aliquot to which 15 units of ß-fructofuranosidase, in 320 mM sodium citrate buffer (pH 4.6) was added directly to the spectrophotometer cuvette. Sucrose hydrolysis was carried out during 30 min at 30 °C. The resulting hexose content was determined after addition of a 750 µl reaction mixture containing 750 mM triethanolamine-NaOH buffer pH 7.6, 3 mM MgSO4, 0.5 mM NAD+, 2.7 mM ATP, and 10 µl of desalted coupling enzyme mixture (3.5 units phosphoglucose isomerase, 2 units glucose 6-phosphate dehydrogenase, and 4.2 units hexokinase) at 30 °C for 40 min. NAD+ reduction was measured at 340 nm. The free hexose content in the extract was measured in another 25 µl aliquot as described above, except that the ß-fructofuranosidase treatment was omitted. The sucrose content was calculated from the difference between the total soluble sugar and hexose contents. The glucose content in the extract was measured in another 25 µl aliquot as described above, except that the ß-fructofuranosidase treatment and the phosphoglucose isomerase were omitted. Fructose was calculated by the difference.

Starch in pellets was suspended in water and gelatinized, after addition of 5 N NaOH, for 60 min at 30 °C, under permanent shaking. After the addition of 15 units of amyloglucosidase dissolved in 320 mM sodium citrate buffer (pH 4.6), starch hydrolysis was performed for 60 min at 50 °C. The extract was cleared by centrifugation, 10 min at 12 000 g, and glucose was determined spectrophotometrically, as for the previous measurements. Results were expressed in mg g–1 fresh matter.

Measurements of kernel content by Near Infrared Reflectance Spectroscopy (NIRS)
Kernel starch, amylose, and protein contents were estimated through NIRS on 50 g of entire mature kernels from each of the three replicates for each inbred lines grown at Toulouse and Saint Martin de Hinx. Kernel water content was homogenized by drying down to 14% humidity, before NIRS determinations. Near infrared absorbance spectra (log 1/R) were collected by averaging 32 scans per sample, using a NIRS system 6500 spectrophotometer scanning every 2 nm from 400 to 2500 nm. Data acquisition, calibration, and analysis were performed using NIRS 2 (4.0) IntraSoft international (1992) software. Samples with a Mahalanobis distance from mean spectrum greater than 3 were eliminated. Kernel content was first predicted using previously determined calibration equations (C Mestres, personal communication), using the whole spectrum for starch, and only 1100 to 2500 nm for amylose and proteins. Equations were then improved by adding 30 inbred lines from our samples that were phenotypically assessed using the same methods as for the initial equations (Mestres et al., 1996Go). Coefficient of determination (cd) between measured and predicted values and standard error (se) of calibration were cd=0.899 and se=0.849 for starch, cd=0.803 and se=0.880 for amylose, and cd=0.956 and se=0.321 for protein content.

CAPS polymorphism and linkage desequilibrium
Total genomic DNA was extracted from adult leaves of the 45 genetically unrelated inbred lines, as in Tai and Tanksley (1990)Go. The Sh1 gene (X02382 [GenBank] ) was PCR-amplified in four parts (c. 2 kb) using primers designed from the published W22 line sequence (Werr et al., 1985Go). The following primer pairs were used: fragment 0, primers AACCATGTACTATCGCCCCC with TCAACAGGGAGCAAGCAGTC, or AAGCGTTTGGGATCTCTTCA with ATGGTCAACAGGGAGCAAG; fragment I, primers GACCCCTACCATCTGCACC with CCAGCTGTTCCTTGAATGCC or CATGGAGCCTAGGAGCAGC with TCCTCCACAGCCAGCTCAC; fragment II, primers TGGGATTACATTCGGGTGAA with TGGAAGGCAGTGAGTCTCTT or GTTGCACTTGCTATCAGGCC with TCGACGTCGCTGTAGATGAG; fragment III, primers CACACCGATTTCATCATCACCA with CGCTGCCCAAAAACGCTC or CTTATTGCCATGAACCACACC with GGAAGACAGGTGAACGAGCA.

PCR mixes (100 µl) included 1.25 mM MgCl2, 0.25 mM of each dNTP, 0.5 µM of each primer, and Taq DNA polymerase (Qiagen). In each PCR reaction, 100 ng of genomic DNA was added. The cycling consisted of 30 s at 94 °C, 1 min at a temperature between 58 °C and 62 °C depending on the primer pair, and 2 min at 72 °C, for 35 cycles. The PCR products were examined by electrophoresis on an 1% agarose gel (Agarose Ultra Pure, GIBCO BRL) in TAE (TRIS-acetate 0.04 M, EDTA 0.001 M) containing 6 µg ethidium bromide/100 ml buffer. The bands were extracted and purified with the QIAEX II (Qiagen) gel extraction kit. The amplified DNA was checked by sequencing at 3'- and 5'-ends and by comparison with the known W22 sequence. An aliquot (150 ng) of DNA from each inbred line was used for digestion with one of the following enzymes: NdeII or HphI on fragment 0, CfoI or RsaI on fragment I, NdeII or RsaI on fragment II, and AluI or MspI on fragment III. Restriction products were analysed on a 5% agarose electrophoresis gel (high resolution agarose, QUANTUM Appligene) in TAE stained with a highly sensitive fluorescent stain for detecting nucleic acids, GelStar (5 µl stain stock solution/50 ml buffer, TEBU). The size of the different restriction fragments was evaluated against a 20 bp ladder (Superladder-low, Eurogentec).

The molecular polymorphism was characterized by the presence/absence of each restriction fragment. Each fragment was denominated by capital letters corresponding to the first initial of the restriction enzyme (1st letter) and the fragment (2nd letter), and a number referring to the PCR fragment. For example, in NEII, ‘N’ means NdeII, ‘E’ is the size class of the restriction fragment (on an A to P scale, from smaller to larger), and ‘II’ indicates the PCR-amplified fragment where the site is located.

QTL detection
QTLs were detected by composite interval mapping (MQTL-CIM as in Tinker and Mather, 1995Go) from mean trait values. The proportion of trait variation explained by each marker was estimated by R2, using a SAS program (SAS, 1999Go; A Leonardi, unpublished data). As discussed in Séne et al. (2000)Go, the choice of cofactors in MQTL-CIM has considerable influence on the number of detected QTLs. In order to reduce the number of false positive QTLs, 5 to 9 cofactors were chosen, the cut-off being placed just before a drop in the partial R2 values arising from stepwise regression. The main effect QTLs were calculated at 5% risk, after 1000 permutations. The peak value gave the likely position between boarding markers.

An in-house program was used for estimating interaction between QTL pairs. It was based on ANOVA (A Leonardi, unpublished data), fitting a model including main-effect QTLs and the interaction between all pairs of markers. Every interaction was tested at a stringent level of significance (P <0.0005). Then, in a second step the combinations of main effect factors and interactions between markers producing the highest R2 value were chosen for the final result. Due to the method used (ANOVA), each interaction QTL was placed at a marker locus and not in-between as is the case when using interval mapping as in the main effect QTLs.

Flowering date being correlated with several trait values it could interfere with QTL detection. This point was tested by comparing QTL obtained from experimental values and from the residual of the regression between each trait value and the flowering date of the corresponding genotype.

Statistical analysis of CAPS polymorphisms
CAPS data were scored as 0/1 following absence/presence of each fragment. Indeed, the published Sh1 genomic sequence (Werr et al., 1985Go) did not allow map reconstruction since CAPS could not frequently be interpreted as presence/absence of a particular restriction site nor a precise position of an insertion/deletion. The significance of pairwise linkage disequilibrium among CAPS was tested using Fisher's exact test excluding non-informative sites (singletons), and corrected for multiple analyses using the Bonferroni procedure (Sokal and Rohlf, 1981Go). Before performing association studies, the extent of large-scale linkage disequilibrium within the 45 line panel was assessed, since it may lead to spurious interpretation of statistical associations between traits and genetic polymorphisms (Pritchard et al., 2000Go). Characterization of the 45 inbred lines for 20 unlinked RFLP markers, distributed as one single locus marker per chromosome arm, showed no linkage disequilibrium (data not shown), suggesting that association studies would not be biaised and may be interpreted as functional association between traits and polymorphisms inside or close to the gene under study. In order to determine the statistical associations between CAPS and phenotypic values, only CAPS with at least five genotypes of each allelic form were used for the peak of robustness. Non-parametric tests based on empirical reconstruction of test criteria distributions using permutations (Churchill and Doerge, 1994Go) and parametric stepwise linear multiple regression (REG procedure; SAS Institute, 1999Go) with a 0.01 threshold for entering and removing CAPS from the regression were both performed. Since results from both methods yielded consistent results, only parametric analyses are presented.


   Results
Top
 Abstract
 Introduction
 Materials and methods
 Results
Discussion
 References
 
Time-course of enzyme activities and carbohydrate contents
The two parental lines MBS and F-2 were used to search for possible differences between parents during early grain filling. Both soluble invertases, neutral-cytosolic (INCY) and acid-vacuolar (INVA), expressed on a fresh matter basis, showed maximum activities at 15 DAP in both genotypes, F-2 having the higher activity at this developmental stage (Fig. 1). A very rapid decline was observed for INVA at 25 DAP, reaching a very low value at 35 DAP. Although INCY was 2–20 times less active than the vacuolar form, the maximum activity was still seen at 15 DAP and this decreased with increasing DAP, especially in F-2; the amplitude of this variation was reduced with respect to INVA. ADPglucose pyrophosphorylase (AGPase) activity showed a different pattern with respect to invertase activities, since it was lowest at 15 DAP, increased at 25 DAP, and stayed relatively stable at 35 DAP for MBS. In F-2, AGPase activity was much higher at 15 DAP than in MBS, it increased at 25 DAP before decreasing to the same level as in MBS at 35 DAP. Sucrose synthase (SUSY) activity was lower than AGPase activity and it behaved rather similarly in the two genotypes: an increase from 15–25 DAP and a decline from 25–35 DAP. Again F-2 showed a higher activity at 15 DAP, but the difference was abolished at 25 and 35 DAP.



View larger version (41K):
[in this window]
[in a new window]
  		Fig. 1. Time-course (15, 25, and 35 DAP) for the measured traits (fresh matter: FW; activities of vacuolar (INVA) and cytosolic (INCY) invertases; sucrose synthase activity: SUSY; ADPglucose pyrophosphorylase activity: AGPase; concentrations in glucose: GLU, fructose: FRU, sucrose: SUC, and starch: STAR), in maize kernels from parental lines MBS (closed triangles) and F-2 (closed circles). Mean ±standard deviation, n=3. Mean value (open squares) and standard deviation of the same traits from 100 recombinant inbred lines (RILs), were added at each stage.

 
The soluble carbohydrate concentration varied in parallel to invertase activities, it was maximum at 15 DAP, sucrose being generally the most abundant compared to glucose and fructose (Fig. 1). Sucrose, glucose, and fructose, fell to low values at 35 DAP. Starch accumulation was four times higher in F-2 than in MBS, at 15 DAP while at 35 DAP, the F-2 value was only 2-fold higher than that of MBS (Fig. 1). Compared with MBS, the general picture for F-2 is a higher fresh matter, a lower soluble carbohydrate concentration (except for sucrose at 15 DAP), but a higher enzyme activity on a fresh matter basis, especially at the earlier sampling times.

Correlation between traits, transgression
Comparison of the parental trait values with the individual RIL values showed a larger variability in RILs (Fig. 1). The variation range in RILs is indicative of a transgressive segregation. In other words, the line with the lower trait value was much lower than in the lowest parent (MBS) and similarly the higher trait line value was much higher than in the highest parent (F-2). An interesting situation is provided by AGPase and starch where both the mean value and the maximum–minimum range were very high, especially at 25 DAP. It could be interpreted by a large genetic variability in the dates of the maximum activity of AGPase. This variability is independent of the flowering date (Table 1). Another attracting hypothesis, may be proposed when recalling that AGPase is a hetero-tetrameric enzyme composed of two types of subunits (BT2 and SH2), each being encoded by a different gene. Thus, an epistatic effect may occur between both subunit genes. The enhanced activity in the RILs at 25 DAP could mean that the reassociation of the subunit genes in the RIL is generally more favourable for activity than in the parents. The associated effect on starch is consistent with the fact that AGPase activity is frequently assumed to be a limiting step in starch synthesis.


View this table:
[in this window]
[in a new window]
  		Table 1. Linear correlation (Pearson's) between the measured traits (flowering: FLO; fresh matter: FW ; activities of vacuolar: INVA and cytosolic invertases: INCY; sucrose synthase activity: SUSY; ADPglucose pyrophosphorylase activity: AGPase; concentrations in glucose: GLU, fructose: FRU, sucrose: SUC and starch: STAR) in kernels of 100 recombinant inbred lines of maize, harvested at 15, 25, and 35 d after pollination (DAP)

 
A trait variability in RILs is a prerequisite for QTL detection. The correlations between variables were examined at the three sampling times (Table 1). As the parental lines differed in precocity a possible side-effect of the RIL flowering date could be considered, although all samplings of each plant in each genotype were made at the same time after the pollination (15, 25, 35 DAP). Among the 16 significant correlations over 21, three traits FW, glucose and fructose were consistently correlated in the same direction at the three stages: FW being higher in the earlier flowering genotypes and glucose or fructose being lower. However, the calculated mean percentage of these trait variances explained by flowering is 18%; it was 5% for the other traits.

The most common correlations were observed between soluble carbohydrates (glucose, fructose, and sucrose) whereas starch was frequently negatively correlated to soluble carbohydrates. At 15 DAP, grain fresh matter was positively correlated to AGPase and SUSY activities and negatively correlated to glucose and fructose levels. These correlations disappeared at later times, although a FW-sucrose correlation was still noticeable at 25 DAP. An AGPase-sucrose correlation was found at 35 DAP. The negative correlation between FW and glucose or fructose mainly reflects the classical decrease in both carbohydrates during development.

QTL detection
Main effect QTLs and interaction QTLs were detected at all developmental stages for most of the measured traits (Table 2). At 15 DAP, nine traits were measured (FW, INVA, INCY, SUSY, AGPase, GLU, FRU, SUC, and STAR) and 23 main effect QTLs and 10 pairs of interaction QTLs were obtained. The total variability explained by these QTLs ranged from 13% (for starch) to 55.7% (for sucrose). For 25 and 35 DAP, only six traits were measured (FW, AGPase, GLU, FRU, SUC, and STAR). Even when taking this into account, the QTL number was lower than at 15 DAP: 5 and 9 main effect, and 9 and 6 interaction pairs for 25 and 35 DAP, respectively. The total variability explained by these QTLs ranged from 20% (for glucose) to 34% (for fructose) at 25 DAP and from 6.5% (for starch) to 39.1% (for fresh matter) at 35 DAP.


View this table:
[in this window]
[in a new window]
  		Table 2. QTLs of main effect (M) and of interaction (Inter.), for the measured traits: fresh matter (FW), enzyme activities of ADPglucose pyrophosphorylase (AGPase), of vacuolar (INVA) and cytosolic (INCY) invertases, and of sucrose synthase (SUSY), concentrations in sucrose (SUC), glucose (GLU), fructose (FRU), and starch (STAR)

 
The possible effect of the RIL flowering date on the detected QTLs was examined because of the observed correlation with the other traits could introduce a bias (Table 1). The QTLs detection was also performed on the residual trait values obtained from the regression against the flowering date. Only five main effect QTLs (AGPase, GLU, FRU, FRU, SUC) and one pair of interaction QTL (STAR) were affected (Table 2, italics in brackets). Consequently, these QTLs have been removed in Fig. 2 and Table 3.




View larger version (59K):
[in this window]
[in a new window]
  		Fig. 2. Mapping of main effect and of interaction QTLs for measured traits (fresh matter, FW; activities of vacuolar (INVA) and cytosolic (INCY) invertases; activities of sucrose synthase (SUSY) and ADPglucose pyrophosphorylase (AGPase); concentration of glucose (GLU), fructose (FRU), sucrose (SUC), and starch (STAR), in maize kernels harvested at three stages, 15 (in red), 25 (in blue) and 35 (in green) DAP. All measured traits are expressed as % kernel fresh matter. The recombinant inbred lines originated from the MBS and F-2 parent lines (6th self generation). Whatever the QTL type, the bar length, in between two markers (main-effect QTLs) or in front of a marker (interaction QTLs), is proportional to the R2 value. Interaction QTLs are represented, at each marker, by a bar and, after a double arrow, the position of the pair member is given in italics by chromosome, and the distance from the telomere top is given in cM. The full description of marker names placed along each chromosome is reported in Causse et al. (1996)Go. Genetic distances are given in centiMorgan (cM) from the top of each chromosome. Bin positions have been placed using anchor markers from the University of Missouri Core Map.

 

View this table:
[in this window]
[in a new window]
  		Table 3. Search for candidate genes for the measured traits (fresh matter: FW; activities of vacuolar: INVA and cytosolic invertases: INCY; sucrose synthase activity: SUSY; ADPglucose pyrophosphorylase activity: AGPase; concentrations in glucose: GLU, fructose: FRU, sucrose: SUC and starch: STAR) in kernels of 100 recombinant inbred lines of maize, harvested at 15, 25 and 35 days after pollination (DAP)

 
The best parental allele for main effect QTLs was depending on the trait. The F-2 allele was most often favourable for QTLs related to sucrose and sucrose cleavage enzyme activities (INVA, INCY, SUSY), MBS allele was the most often favourable for AGPase activity. For the other traits, the best allele changed from one QTL to the other (FW, GLU, FRU) or QTL number was too low (STAR) to draw a conclusion. It may be noted that the allele effect is not related to the fact that a trait is increasing or decreasing with time: SUSY and AGPase varied in the same way, but the favourable allele was either F-2 or MBS. This means that the genetic effect is unlikely to be attributed to the development rate, but rather to the difference in allelic composition. Comparison of QTL location for the same trait at each stage provided few co-locations (Table 2, inset). Two QTLs were found to be in common between 15 and 25 DAP, and two others between 25 and 35 DAP but only one common QTL was found at the same locus for the three different stages: a fresh matter QTL, at bin 7.04. Although surprising at first sight, it could simply reflect that the three stages represent three development phases: onset of starch synthesis, linear filling, and end of filling, where different sets of regulatory genes are encountered.

QTL location
Depending on the developmental stage, the QTLs were more or less spread over all of the 10 chromosomes (Fig. 2). At 15 DAP, QTLs were found on all of the chromosomes, but a higher density was detected on chromosomes 5 and 9. At 25 and 35 DAP, some chromosomes were not labelled whereas others were found to be particularly rich in QTLs: chromosomes 4 and 8, and chromosomes 1 and 2, respectively.

When QTLs were abundant on one chromosome, they were often grouped in particular regions forming ‘clusters’. Some of these clusters contained only QTLs obtained at 15 DAP; for example, on chromosome 5, five QTLs (AGPase, SUC, INCY, INVA, SUC) clustered at bin 5.02–5.03 and 5 QTLs (SUC, SUSY, GLU, FRU, SUC) at bin 5.04–5.06. Other clustered QTLs were obtained when two stages were considered together: for example, seven QTLs (SUC, SUSY, GLU, GLU, FRU, SUC, FRU) at bin 5.04–5.06 for 15 and 25 DAP and six QTLs (STAR, FW, FW, FRU, FRU, FRU) at bin 2.05–2.06 for 25 and 35 DAP.

The only QTL common to the three stages (Table 2), was a fresh matter QTL located on chromosome 7 (bin 7.04), at the bnl 1407 marker (Fig. 2). In 1998, a similar location for the same trait had been found using the same RILs, at 12 DAP, in another field experiment performed in SE France (Etoile-sur-Rhône, Drôme), but under drier and more sunny conditions (C Thévenot, unpublished data).

Identification of candidate genes
Numerous genes coding for enzymes of the starch biosynthetic pathway are located on the chromosome map used in this work, thereby facilitating the search for candidate genes involved in the metabolic regulation of this pathway. Vacuolar invertases are coded either by Ivr1 or Ivr2 (Table 3). At 15 DAP, a QTL was found for the trait INVA, at the Ivr1 locus on chromosome 2 (bin 2.04). At 25 and 35 DAP, an AGPase QTL was located on chromosome 8 (bin 8.04), and on chromosome 4 (bin 4.05), respectively, close to two Bt2 loci; Bt2 genes code for the small catalytic subunit of ADPglucose pyrophosphorylase.

It is also possible to find candidate genes with less direct effect by looking for relationships between genes coding for all of the presently considered enzymes of carbohydrate metabolism (Ivr1, Ivr2, Sh2, Bt2, Sus1, and Sh1) and all of the related QTLs for the measured traits (Table 3; Fig. 2). Near the Ivr2 gene, coding for a vacuolar invertase, QTLs were found at 15 and 35 DAP (bin 5.03). Similarly, co-locations near the three Bt2 loci were found at 15 DAP (bins 1.07, 4.05, and 8.04). Thus, in the Bt2 multigenic family, three genes may be involved in controlling some stages of the starch biosynthetic pathway. However, no QTL co-located with the multigenic family Sh2 (bins 1.03 and 3.09) which codes for the large subunit of ADPglucose pyrophosphorylase. At the 15 DAP stage, a co-location of a QTL was found close to the gene Sus1 (bin 9.04) coding for one subunit of the sucrose synthase. Moreover, co-locations of QTLs at 15, 25, and 35 DAP, were found near to the gene Sh1 (bin 9.01) coding for another subunit of the sucrose synthase. Thus, a relationship may exist between Sh1 gene expression (bin 9.01), and either ADPglucose pyrophosphorylase activity or soluble sugar levels, at each physiological stage. The correlation between the three enzyme activities (Table 1) is consistent with the partial co-location of their QTLs.

Co-locations between the observed QTLs and genes coding for other enzymes of the starch biosynthetic pathway such as cell wall invertase, involved in sucrose unloading in the basal endosperm transfer cells (miniature1 mutation), sucrose-phosphate-synthase, probably involved in sucrose resynthesis after invertase cleavage, and starch-debranching enzyme were considered. In addition, two respiratory enzymes, phospho-glyceratemutase and 6-phosphogluconate dehydrogenase were also considered, since some mutations affecting respiration are likely to affect starch metabolism, as could be the case in sugary2 mutant (Pan, 2003Go) (Table 3; Fig. 2). Three genes coding for cell wall invertases (Incw1, 2 and 3) co-located with QTLs at the three considered developmental stages. At 15 DAP, two QTLs were located close to Incw1 (bin 5.04) and one QTL was near Incw2 (bin 2.05); at 25 DAP, one QTL was located close to Incw3 (bin 10.04), while at 35 DAP, it was close to Incw2 (bin 2.05), and Incw3 loci (bin 10.04). Among the three loci for sucrose-phosphate-synthase (bins 3.05, 6.01 and 8.06), the bin 8.06 co-located with AGPase QTLs at 15 DAP and 25 DAP. The starch-debranching enzyme locus (Su1, bin 4.05) co-located with one QTL at 15 and one at 25 DAP. Phospho-glyceratemutase is associated with two loci (bins 3.06 and 6.05), at 15 or 25 DAP. The 6-phosphogluconate dehydrogenase locus (bin 6.01) showed a co-location with one QTL, at 35 DAP.

Finally, it is noteworthy that seven AGPase QTLs co-located with various candidate genes including Ivr2, Bt2, Sh1, Sps, and Pmg.

Association of Sh1 molecular polymorphisms with phenotypic traits
As the Sh1 locus frequently co-located with QTLs for enzyme activity or substrate content in the F-2-MBS847 RIL family, the polymorphism of this gene was characterized within a collection of unrelated inbred lines and its statistical association with phenotypic traits assessed. Four Sh1 segments of 2 kb each were PCR-amplified and named 0 (segment in the promoter), I (exons 1–3), II (exons 4–12) and III (exons 12–16) (Fig. 3). Significant associations between CAPS and traits were detected in parts 0, II, and III. Starch content in total kernel and in endosperm was associated with two CAPS (HB0 and RPII) separated by more than 3.5 kb and showing no linkage disequilibrium between each other within this sample (Table 4). Amylose % kernel matter and quantity, either in kernel or endosperm, were significantly associated to several CAPS in segments II and III (i.e. NEII, NKII, and AKIII). These CAPS were independent from each other in these samples. Finally, AGPase activity at 35 DAP was significantly associated to the ND0 promoter fragment. As the flowering date had some influence on QTL detection, its effect was checked in the association tests. Very poor correlations were found between kernel phenotypes and AGPase activity, except for amylose quantity in experiment 2 (r= –0.425, P=0.004). Accordingly, the inclusion of the flowering date as a covariate in multiple regression analyses of the associations between CAPS and traits did not significantly change associations. The only modification was for the AKIII–amylose association which passed over the 1% threshold (P=1.9%); this association was deleted from Fig. 3 and Table 4.



View larger version (12K):
[in this window]
[in a new window]
  		Fig. 3. The Sh1 gene consists of 16 exons and 15 introns. The nucleotide numbering is from the W22 sequence (X02382 [GenBank] ). The gene was PCR-amplified in four fragments of c. 2 kb, named 0, I, II, and III. The name and position of the polymorphous restriction fragments are placed above the sequence. The first letter of the name stands for the restriction enzyme used: A, H, N, and R for AluI, HphI, NdeII, and RsaI, respectively; the second letter corresponds to the size class of the restriction fragment (on an A to P scale, from smaller to larger) and the Roman number indicates the PCR-amplified fragment where the site is located. An asterisk indicates that the exact position of the polymorphous site is known. In the other cases, a line delineates the zone where polymorphism is present.

 

View this table:
[in this window]
[in a new window]
  		Table 4. Significant associations between Sh1 CAPS fragments and phenotypic traits in a population of 45 unrelated maize inbred lines. n0/n1: sample size for absence/presence of the CAPS fragment; R2: determination coefficient; P: significance level of the multiple regression test (one factor only was significant for each experiment, in the stepwise procedure)

 

   Discussion
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
References
 
As in previous studies dealing with starch-related traits in mature grains (Goldman et al., 1993Go; Séne et al., 2000Go, 2001Go) several QTLs (explaining up to 56% of the variability) were detected for almost all of the measured traits. These QTLs were not randomly distributed, they were frequently grouped into clusters associating either traits for carbohydrate concentration (bins 2.05–2.06 and 5.05–5.06) or traits for both enzyme activity and carbohydrate concentration (bins 4.05, 5.03, 9.01, 9.07). The clustering may indicate either that several genes controlling the starch pathway are grouped in some chromosome regions or that a common regulatory gene is located at the cluster locus. Examination of the position of the structural genes of the pathway enzymes (see below) suggests that both are possible. Another property of these clusters is that they tended to be located on different chromosomes depending on the developmental stage. For example, there was a higher QTL number on chromosomes 5 and 9 at 15 DAP, on chromosomes 4 and 8 at 25 DAP and chromosomes 1 and 2 at 35 DAP, respectively. This is not surprising since 15, 25 and 35 DAP correspond to rather different physiological conditions: preparation for grain filling, active starch accumulation, and progressive reduction of starch synthesis. Thus, the set of controlling genes is likely to be different. An example of such a shift is provided by AGPase QTLs: at 35 DAP, an interaction-QTL co-located on chromosome 4 (bin 4.05) with the Bt2 gene coding the small AGPase subunit, whereas at 15 DAP none of the five main-effect AGPase QTLs (2.02, 5.03, 8.06, 9.01, 9.04) co-located with an AGPase structural gene. These five main-effect QTLs may be linked to regulatory genes. It could be noted that among the five QTL loci, two of them are close to structural genes of other enzymes of the starch pathway, Sps (8.06) and Sh1 (9.01).This raises the question of the co-regulation of enzymes of the same pathway by common loci, some of them being linked to structural genes. The co-location of main effect and interaction QTLs at bins 8.06 and 9.01 could provide further support to loci interaction. It is noteworthy that, at the Sh1 locus, the interaction involved activity and substrate QTLs.

A comparison of individual or classes of QTLs with published results on mature grain composition shows some similarities. A common locus (bin 3.06) can be found for starch component QTLs in mature grains of the same RIL family (Séne et al., 2000Go) or in different families (Goldman et al., 1993Go). Similarly, at bin 8.06, the Sps locus shows repeated co-locations either in grains with AGPase QTLs in this experiment, with starch component QTLs (Séne et al., 2000Go), with sucrose and glucose QTLs (field experiment from Etoile-sur-Rhône, Drôme, 1998; C Thévenot, unpublished data), or in adult leaves with QTLs for SPS activity, hexose and sucrose (Causse et al., 1995Go). Other common locations between this work and those found in 1998 at Etoile-sur-Rhône (C Thévenot, unpublished data) are: (i) at bin 9.08 for vacuolar invertase activity at 15 and 12 DAP, respectively, and (ii) at bin 7.04 for FW QTLs at 15, 25 and 35 DAP (Gif-sur-Yvette), and 12 DAP (Etoile).

Co-location of QTLs with genes encoding for related functions, suggest these to be candidate genes. In the present case, the co-location of a given enzyme activity and a structural gene locus possibly encoding the enzyme was only observed at one developmental stage for a vacuolar invertase QTL and the Ivr1 locus at bin 2.04, and for AGPase activity QTLs and the Bt2 gene at loci 4.05 and 8.04. However, Ivr1 transcripts as they were not detected in the endosperm (Qin et al., 2004Go) should be discarded. The endosperm-specific Bt2 gene is a better candidate since it is the functional gene expressed in the endosperm, as proved by mapping of the bt2 mutation, producing a large starch deficiency (bin 4.05). The co-location is consistent with the idea that mutation represents nil allele of a gene whereas QTL pinpoints a quantitative allelic effect. The analysis of co-locations may be broadened by considering, for an enzyme locus, not only the corresponding enzyme activity QTL but also QTLs for related enzymes and carbohydrate traits since, as already discussed, it may provide information about the functional relationships between gene expression and some QTLs of the starch biosynthesis pathway. Such co-locations were noted for the three enzyme activities or the carbohydrate concentrations and one of the structural genes (bins 1.07, 4.05, 5.03, 8.04, 9.01). This may be interpreted by regulatory relationships between enzymes of the same pathway or by the existence of common transcription factors acting at several levels along the metabolic pathway. In the case of carbohydrate concentration–enzyme activity QTL co-locations, either their concentration is dependent on enzyme activity and their position in the pathway (substrate versus product) or they act as signal molecules controlling the transcription level of the enzyme. This latter possibility gained experimental support for both vacuolar invertase and sucrose synthase (Koch, 1996Go). Whatever the interpretation, the loci where such multi-trait co-locations were observed was Sh1 (bin 9.01) since at 15, 25 or 35 DAP it was associated with AGPase, FRU, GLU, and SUC. The same locus was also pinpointed in the ‘Etoile’ field experiment at 12 DAP (SUC QTL) and by Causse et al. (1995)Go in adult leaf QTLs (AGPase, SUSY, and hexose QTLs).

Other, multi-trait-locus associations have been noted already by other authors (Séne et al., 2000Go; Goldman et al., 1993Go). At bin 3.06 a possible candidate gene, could be phosphoglucomutase since it was shown to affect starch synthesis in transgenic potato tubers (Fernie et al., 2002Go). At bin 8.06, Causse et al. (1995)Go observed, in leaves, a co-location between the SPS activity and Sps locus which is consistent with the role of this enzyme in sucrose synthesis. However, its role in the starch synthesis pathway in grain remains to be established. The co-location of the starch debranching enzyme locus (Su1, bin 4.05) with FW and starch QTLs may be of significance since the su1 mutation (sugary 1) profoundly alters starch synthesis and the sucrose/starch balance. Other co-locations with possible candidates are listed in Table 3, but they are less physiologically supported. For example, the Ivr1 and Incw1 invertases are not significantly expressed in developing grains (Qin et al., 2004Go). Co-locations with respiratory enzymes as 6-P-gluconate dehydrogenase involved in the hexose monophosphate pathway (bin 6.01) may also be significant since alteration of another respiratory enzyme (glycolytic P-fructokinase), is probably producing starch deficiency in the sugary2 mutant (Pan, 2003Go).

Amongst all suggested candidate genes, Sh1 is one of the most prominent, since it co-locates with QTLs at several developmental stages. The next step after candidate identification is the validation, as discussed in Prioul et al. (1999)Go. Several methods are possible, one of them being to check if the results obtained from the two alleles in the RIL population are coincidental or may be confirmed using a large collection of gene alleles originating from diverse genotypes. For this purpose, unrelated lines having a minimum linkage disequilibrium have been used, since they are expected to show a maximum allele diversity and minimum false positive statistical associations between molecular polymorphisms and traits. The multiple associations between CAPS in Sh1 gene and several grain-filling-related traits are in favour of the importance of Sh1 in the control of kernel starch accumulation. Since the results are based on CAPS from a small sample, they need to be confirmed on a larger genotype panel. Moreover, the exact nature of the polymorphism (position of SNPs or insertion/deletion) could be determined. Further support is currently being investigated in a larger population of unrelated lines screened for gene polymorphism through allele sequencing and SSR (single sequence repeats) genotyping. This will allow statistical associations to be detected between traits and well identified SNPs (single change polymorphism), independently of the genetic structure of the population (characterized through SSR polymorphisms), as proposed by Pritchard et al. (2000)Go and Thornsberry et al. (2001)Go.


   Acknowledgements
 
The authors are very grateful to V Capelle for efficient assistance in QTL detection, Dr C Mestres for NIRS determinations and to Dr M Hodges for careful reading of the manuscript. This work was partly funded by the French Génoplante program.


   Footnotes
 
Abbreviations: DAP, days after pollination; mn1, miniature; sh1, shrunken1; sh2, shrunken 2; bt2, brittle2; CAPS, cleaved amplified polymorphism sequence; NIRS, near infrared reflectance spectroscopy; FLO, flowering; FW, fresh matter; INVA, acid-vacuolar invertase; INCY, neutral-cytosolic invertase; SUSY, sucrose synthase; AGPase, ADPglucose pyrophosphorylase; GLU, glucose; FRU, fructose; SUC, sucrose; STAR, starch.


   References
Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 References
 
Bergmeyer HU, Bernt E. 1974. Sucrose. In: Bergmeyer HU, ed. Methods of enzymatic analysis, Vol. 3. New York: Academic Press, 1176–1179.

Churchill GA, Doerge RW. 1994. Empirical threshold values for quantitative trait mapping. Genetics 138, 963–971.[Abstract]

Causse M, Rocher JP, Henry AM, Charcosset A, Prioul JL, de Vienne D. 1995. Genetic dissection of the relationship between carbon metabolism and early growth in maize, with emphasis on key-enzyme loci. Molecular Breeding 1, 259–272.[CrossRef]

Causse M, Santoni S, Damerval C, Maurice A, Charcosset A, Deatrick J, de Vienne D. 1996. A composite map of expressed sequences in maize. Genome 39, 418–432.[Web of Science][Medline]

Fernie AR, Swiedrych A, Tauberger E, Lytovchenko A, Trethewey RN, Willmitzer L. 2002. Potato plants exhibiting combined antisense repression of cytosolic and plastid isoforms of phosphoglucomutase surprisingly approximate wild type with respect to the rate of starch synthesis. Plant Physiology and Biochemistry 40, 921–927.[CrossRef][Web of Science]

Goldman IL, Rocheford TR, Dudley JW. 1993. Quantitative trait loci influencing protein and starch concentration in the Illinois long term selection maize strains. Theoretical and Applied Genetics 87, 217–224.

Koch KE. 1996. Carbohydrate-modulated gene expression in plants. Annual Review of Plant Physiology and Plant Molecular Biology 47, 509–540.[CrossRef][Web of Science]

Mestres C, Matencio F, Pons B, Yajid M, Fliedel G. 1996. A rapid method for the determination of amylose content by using differential-scanning calorimetry. Starch-Stärke 48, 2–7.[CrossRef]

Ou-Lee TM, Setter TL. 1985. Enzyme activities of starch and sucrose pathways and growth of apical and basal maize kernels. Plant Physiology 79, 848–851.[Abstract/Free Full Text]

Pan D. 2003. Defect of phosphofructokinase activity in a sugary 2 mutant. Maize Genetics Co-operation Newsletter 77, 51–52.

Pelleschi S, Guy S, Kim JY, Pointe C, Mahé A, Barthes L, Leonardi A, Prioul JL. 1999. Ivr2, a candidate gene for a QTL of vacuolar invertase activity in maize leaves. Gene specific expression under water stress. Plant Molecular Biology 39, 373–380.[CrossRef][Web of Science][Medline]

Prioul JL, Jeannette E, Reyss A, Grégory N, Giroux M, Hannah LC, Causse M. 1994. Expression of ADP-glucose pyrophosphorylase in maize (Zea mays L.) grain and source leaf during grain filling. Plant Physiology 104, 179–187.[Abstract]

Prioul JL, Pelleschi S, Séne M, Thévenot C, Causse M, de Vienne D, Leonardi A. 1999. From QTLs for enzyme activity to candidate genes in maize. Journal of Experimental Botany 50, 1281–1288.[Abstract/Free Full Text]

Prioul JL, Quarrie S, Causse M, de Vienne D. 1997. Dissecting complex physiological functions through the use of molecular quantitative genetics. Journal of Experimental Botany 48, 1151–1163.[Abstract/Free Full Text]

Prioul JL, Reyss A, Schwebel-Dugué N. 1990. Relationships between carbohydrate metabolism in ear and in the adjacent leaf during grain filling in maize genotypes. Plant Physiology and Biochemistry 28, 489–493.

Pritchard JK, Stephens M, Rosenberg NA, Donnelly P. 2000. Association mapping in structured populations. American Journal Human Genetics 67, 170–181.

Qin L, Trouverie J, Chateau-Joubert S, Simond-Côte E, Thévenot C, Prioul JL. 2004. Involvement of the Ivr2-invertase in the perianth during maize kernel development under water stress. Plant Science 166, 371–379.[CrossRef]

SAS Institute. 1999. The SAS system for Windows, version 8.00 (SAS Inst., Cary, NC).

Séne M, Causse M, Damerval C, Thévenot C, Prioul JL. 2000. Quantitative trait loci affecting amylose, amylopectin, and starch content in maize recombinant inbred lines. Plant Physiology and Biochemistry 38, 459–472.[CrossRef]

Séne M, Thévenot C, Hoffman D, Bénétrix F, Causse M, Prioul JL. 2001. QTLs for grain milling properties, composition and vitreousness in maize recombinant inbred lines. Theoretical and Applied Genetics 102, 591–599.[CrossRef]

Sokal RR, Rohlf FJ. 1981. Biometry. Freeman WH, ed. San Francisco.

Sourdille P, Perretant MR, Charmet G, Leroy P, Gautier MF, Joudrier P, Nelson JC, Sorrells ME, Bernard M. 1996. Linkage between RFLP markers and genes affecting kernel hardness in wheat. Theoretical and Applied Genetics 93, 580–586.[CrossRef]

Tai TH, Tanskley SD. 1990. A rapid and inexpensive method for isolation of total DNA from dehydrated plant tissue. Plant Molecular Biology Reporter 8, 297–303.

Thornsberry JM, Goodman MM, Doebley J, Kresovich S, Nielsen D, Buckler ESIV. 2001. Dwarf 8 polymorphisms associate with variation in flowering time. Nature Genetics 28, 286–289.[CrossRef][Web of Science][Medline]

Tinker NA, Mather DE. 1995. MQTL: software for simplified composite interval mapping of QTL in multiple environments. JQTL 1, 2 (Electronic Journal, access at :http://probe.nalusda.gov:8000/otherdocs/jqtl/).

Werr A, Frommer WB, Maas C, Starlinger P. 1985. Structure of the sucrose synthase gene on chromosome 9 of Zea mays. EMBO Journal 4, 1373–1380.[Web of Science][Medline]


Add to CiteULike CiteULike   Add to Connotea Connotea   Add to Del.icio.us Del.icio.us    What's this?


This article has been cited by other articles:


Home page
 Plant Physiol.Home page
D. Manicacci, L. Camus-Kulandaivelu, M. Fourmann, C. Arar, S. Barrault, A. Rousselet, N. Feminias, L. Consoli, L. Frances, V. Mechin, et al.
Epistatic Interactions between Opaque2 Transcriptional Activator and Its Target Gene CyPPDK1 Control Kernel Trait Variation in Maize
Plant Physiology, May 1, 2009; 150(1): 506 - 520.
[Abstract] [Full Text] [PDF]

Home page
 GeneticsHome page
D.-L. Yang, R.-L. Jing, X.-P. Chang, and W. Li
Identification of Quantitative Trait loci and Environmental Interactions for Accumulation and Remobilization of Water-Soluble Carbohydrates in Wheat (Triticum aestivum L.) Stems
Genetics, May 1, 2007; 176(1): 571 - 584.
[Abstract] [Full Text] [PDF]

Home page
 J Exp BotHome page
R. Zhou, L. Cheng, and A. M. Dandekar
Down-regulation of sorbitol dehydrogenase and up-regulation of sucrose synthase in shoot tips of the transgenic apple trees with decreased sorbitol synthesis
J. Exp. Bot., November 1, 2006; 57(14): 3647 - 3657.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow FREE Full Text (PDF) Freely available
Right arrow All Versions of this Article:
56/413/945    most recent
eri087v1
Right arrow E-letters: Submit a response
Right arrow Alert me when this article is cited
Right arrow Alert me when E-letters are posted
Right arrow Alert me if a correction is posted
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Add to My Personal Archive
Right arrow Download to citation manager
Right arrow Search for citing articles in:
ISI Web of Science (8)
Right arrowRequest Permissions
Right arrow Disclaimer
Google Scholar
Right arrow Articles by Thévenot, C.
Right arrow Articles by Prioul, J.-L.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Thévenot, C.
Right arrow Articles by Prioul, J.-L.
Agricola
Right arrow Articles by Thévenot, C.
Right arrow Articles by Prioul, J.-L.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us  
What's this?