Genetic Variation Controlling Wrinkled Seed Phenotypes in Pisum: How Lucky Was Mendel?

Abstract

One of the traits studied by Mendel in pea (Pisum sativum L.) was the wrinkled-seeded phenotype, and the molecular basis for a mutation underlying this phenotype was discovered in the 1990s. Although the starch-branching enzyme gene mutation identified at the genetic locus r is most likely to be that in seeds available to Mendel in the mid-1800s, it has remained an open question as to whether or not additional natural mutations in this gene exist within Pisum germplasm collections. Here, we explore this question and show that all but two wrinkled-seeded variants in one such collection correspond to either the mutant allele described previously for the r locus or a mutation at a second genetic locus, rb, affecting the gene encoding the large subunit of Adenosine diphosphoglucose (ADP-glucose) pyrophosphorylase; the molecular basis for the rb mutation is described here. The genetic basis for the phenotype of one (JI 2110) of the two lines which are neither r nor rb has been studied in crosses with a round-seeded variant (JI 281); for which extensive genetic marker data were expected. In marked contrast to the trait studied by Mendel and the rb phenotype; the data suggest that the wrinkled-seeded phenotype in JI 2110 is maternally determined, controlled by two genetic loci, and the extent to which it is manifested is very sensitive to the environment. Metabolite analysis of the cotyledons of JI 2110 revealed a profile for sucrose and sucrose-derived compounds that was more similar to that of wild-type round-seeded, than that of wrinkled-seeded r, pea lines. However, the metabolite profile of the seed coat (testa) of JI 2110 was distinct from that of other round-seeded genotypes tested which, together with analysis of recombinant inbred progeny lines, suggests an explanation for the seed phenotype.

Keywords: genetic markers; myoinositol; pea germplasm; r and rb mutations; seed coat (testa) metabolites; seed phenotype; wrinkled seeds.

Figure 1

(A) Manual assembly of contigs obtained by sequencing cDNA corresponding to sbeI in the wrinkled-seeded pea accession, JI 1194, allows the insertion sequence to be partly determined. The colour codes for the highlighted regions are given in the key. For designing primers specific to the insertion mutation in r mutants, the sequence highlighted in green was used. The positions of primer sequences are indicated in red/bold font, with reverse primer sequences underlined. The positions of primers used to provide a facile diagnostic test for RR vs. rr genotypes are shown in bold font. Diagnostic primers used in the high-throughput genetic screen of germplasm to distinguish RR from rr genotypes are shown in red font; KEY: GTCAGATG = wild-type sequence (round-seeded lines); AGTAGAAT = 8bp flanking insertion; GTAAAATA = r insertion sequence confirmed; AAAAATTC = good quality r insertion sequence, unconfirmed; ATTAGATAATTAGATA = poor quality r insertion sequence, unconfirmed; CAAAAATACAAAAATA = poor quality r insertion sequence unconfirmed; ---- = earlier and later sequence not forming a contig; ATG = initiator methionine; TAG/TAA = stop codon in wild type, deduced in mutant; Primer sequences are indicated in red/bold font, with reverse primer sequences underlined; (B) The SBEI/sbeI protein sequences deduced for a sbe1 mutant variant (JI 1194) in comparison with a wild type (JI 3316). The protein predicted for the mutant diverges from that of the wild type towards the carboxy-terminal region. The variation was validated by partial sequences determined for a second mutant line, JI 430, in comparison with a further wild type line, JI 1387. Amino acids are colour-coded: negatively charged, blue; positively charged, magenta; C, G, H, N, Q, S, T, Y, green; all other amino acids, red. Identities and similarities are indicated for the divergent mutant and wild-type carboxy-terminal sequences (*, identical; :, very similar; ., similar, amino acids).

Figure 1

(A) Manual assembly of contigs obtained by sequencing cDNA corresponding to sbeI in the wrinkled-seeded pea accession, JI 1194, allows the insertion sequence to be partly determined. The colour codes for the highlighted regions are given in the key. For designing primers specific to the insertion mutation in r mutants, the sequence highlighted in green was used. The positions of primer sequences are indicated in red/bold font, with reverse primer sequences underlined. The positions of primers used to provide a facile diagnostic test for RR vs. rr genotypes are shown in bold font. Diagnostic primers used in the high-throughput genetic screen of germplasm to distinguish RR from rr genotypes are shown in red font; KEY: GTCAGATG = wild-type sequence (round-seeded lines); AGTAGAAT = 8bp flanking insertion; GTAAAATA = r insertion sequence confirmed; AAAAATTC = good quality r insertion sequence, unconfirmed; ATTAGATAATTAGATA = poor quality r insertion sequence, unconfirmed; CAAAAATACAAAAATA = poor quality r insertion sequence unconfirmed; ---- = earlier and later sequence not forming a contig; ATG = initiator methionine; TAG/TAA = stop codon in wild type, deduced in mutant; Primer sequences are indicated in red/bold font, with reverse primer sequences underlined; (B) The SBEI/sbeI protein sequences deduced for a sbe1 mutant variant (JI 1194) in comparison with a wild type (JI 3316). The protein predicted for the mutant diverges from that of the wild type towards the carboxy-terminal region. The variation was validated by partial sequences determined for a second mutant line, JI 430, in comparison with a further wild type line, JI 1387. Amino acids are colour-coded: negatively charged, blue; positively charged, magenta; C, G, H, N, Q, S, T, Y, green; all other amino acids, red. Identities and similarities are indicated for the divergent mutant and wild-type carboxy-terminal sequences (*, identical; :, very similar; ., similar, amino acids).

Figure 2

Diagnostic PCR tests for variant wild-type (R) and mutant sbeI (r) alleles. (A) Simplex PCR assays showing a 340 bp fragment, derived from the wild-type gene, and a ~1250 bp fragment, representing the mutant allele. DNA was assayed from (a) JI 2202, RR; (b) JI 2822, RR; (c) JI 15, RR; (d) JI 399, RR; (e) JI 281, RR; (f) JI 1194, rr; (g) JI 1201, rr; (h) cv. Princess, marrowfat; (i) control PCR (lacking DNA). The right hand panel shows assays of recombinant inbred lines (RILs) derived from a cross between JI 15 and JI 1194; (B) Triplex PCR assays showing the same wild-type amplicon as shown in (A), with a smaller (298 bp) amplicon derived from the mutant (insertion to 3′ non-coding) allele, enabling heterozygotes to be scored readily. DNA was assayed from (j,k) JI 1417; (l,m) JI 1194; (n) JI 1417 × JI 1194 F1; (o) JI 1194 × JI 1417 F1; (p) control PCR (lacking DNA); (q) JI 15; (r) JI 1194; (s) JI 15; (t) JI 1194; (u,v) JI 15 and JI 1194 DNA mix; (w) control PCR (lacking DNA). Tracks labelled M show DNA markers (100 bp ladder, New England BioLabs^®, Ipswich, MA, USA).

Figure 3

(A) Comparison of the mutant protein predicted by the agpl gene sequence in JI 399, a rb mutant lacking three amino acids (ATP) compared with the wild-type protein predicted for the cultivars Cameor and Sugar snap, the last being the source of the database NCBI submission X96766; asterisks indicate amino acid identity; (B) Deletion of the region in the agpl gene sequence of six germplasm accessions, all of which have a wrinkled-seeded phenotype, in comparison to the wild-type gene in two accessions. The JI accession numbers are indicated on the left of the corresponding sequences. Nucleotides are colour-coded: C, blue; A, green; G, black; T, red. (Note that intron 2 begins after the deletion, with the nucleotides GT).

Figure 4

(A) Allelism tests of JI 2110 crossed to the wild-type round-seeded control line (cv. Cameor, top two panels) and to the wrinkled-seeded lines with mutations at rug3, rug 4 (lower four panels). The female parent is listed first for every cross. Individual or groups of F1 seeds shown underneath the corresponding cross; (B) Images of starch grains from r (JI 1194, compound) and rb (JI 399, simple) control lines, JI 2110 (simple starch grains) and a further germplasm accession (sbeI -ins, r) with compound starch grains. The lower panel shows starch grains from the four lines stained with iodine. Various magnifications were used to emphasise the differences in shape of starch grains; bar scales not shown.

Figure 5

(A) Genetic linkage map developed for JI 281 × JI 2110 RILs, based on SSAP, gene-specific and morphological markers; (B) The major genetic positions determined for the seed phenotype variation (highlighted) in JI 281 × JI 2110, based on individual generation scores. The major map positions determined for seed shape phenotype over three generations (F9, F10 and F11, round, R, wrinkled, W) are highlighted in yellow.

Figure 6

(A) Relative concentrations (% mg dry weight) of total soluble sugars, starch, and protein in cotyledons of JI 2110 in comparison with near-isogenic wild-type (WT, RR) and wrinkled-seeded sbeI (rr) mutant lines. Sugars were measured relative to an internal standard, and protein relative to a standard. Bars indicate standard deviation; three replicate seeds per line. * p < 0.05, determined by t-test; (B) Relative concentrations (% mg dry weight) of myoinositol, sucrose, raffinose, stachyose, and verbascose in cotyledons of JI 2110 in comparison with near-isogenic wild-type (WT, RR) and wrinkled-seeded sbeI (rr) mutant lines. Bars indicate standard deviation; three replicate seeds per line. * p < 0.05, ** p < 0.01, determined by t-test; (C) Concentrations of myoinositol and sucrose (μg/mg dry weight) in testas of JI 2110 in comparison with testas of the wild-type (round-seeded) genotypes, JI 281 and JI 3253. Bars indicate standard deviation; three replicate batches of testas per line. Standards of each compound enabled detection and quantification; metabolites not shown were below the lowest calibration points of the standards. ** p < 0.01, determined by t-test. ((A–C) *, ** denote comparisons of JI 2110 or sbeI with wild-type lines); (D) Concentrations of myoinositol and sucrose (μg/mg dry weight, determined using standards of each compound) in testas of JI 2110, JI 281, JI 3253 (1, 2, two replicate seed batches in every case) and a set of white-flowered JI 281 × JI 2110 RILs (four round-seeded, R; four wrinkled-seeded, W). Mean values for the R and W RILs are shown with standard error bars. ** p = 0.01, determined by t-test of myoinositol concentration in R and W RILs.