New cup out of old coffee: contribution of parental gene expression legacy to phenotypic novelty in coffee beans of the allopolyploid Coffea arabica L

Abstract

Background and aims: Allopolyploidization is a widespread phenomenon known to generate novel phenotypes by merging evolutionarily distinct parental genomes and regulatory networks in a single nucleus. The objective of this study was to investigate the transcriptional regulation associated with phenotypic novelty in coffee beans of the allotetraploid Coffea arabica.

Methods: A genome-wide comparative transcriptomic analysis was performed in C. arabica and its two diploid progenitors, C. canephora and C. eugenioides. Gene expression patterns and homeologue expression were studied on seeds at five different maturation stages. The involvement of homeologue expression bias (HEB) in specific traits was addressed both by functional enrichment analyses and by the study of gene expression in the caffeine and chlorogenic acid biosynthesis pathways.

Key results: Expression-level dominance in C. arabica seed was observed for most of the genes differentially expressed between the species. Approximately a third of the genes analysed showed HEB. This proportion increased during seed maturation but the biases remained equally distributed between the sub-genomes. The relative expression levels of homeologues remained relatively constant during maturation and were correlated with those estimated in leaves of C. arabica and interspecific hybrids between C. canephora and C. eugenioides. Functional enrichment analyses performed on genes exhibiting HEB enabled the identification of processes potentially associated with physiological traits. The expression profiles of the genes involved in caffeine biosynthesis mirror the differences observed in the caffeine content of mature seeds of C. arabica and its parental species.

Conclusions: Neither of the two sub-genomes is globally preferentially expressed in C. arabica seeds, and homeologues appear to be co-regulated by shared trans-regulatory mechanisms. The observed HEBs are thought to be a legacy of gene expression differences inherited from diploid progenitor species. Pre-existing functional divergences between parental species appear to play an important role in controlling the phenotype of C. arabica seeds.

Keywords: Coffea arabica; Allopolyploidy; additivity; caffeine; desiccation tolerance; expression-level dominance; gene expression; homeologue expression bias; phenotype; seed.

Fig. 1.

Coffee seed development. Schematic representation of seed development stages in the upper panel (adapted from Dussert et al., 2018) and, in the lower panel, cross-sections of fruits at the different developmental stages studied.

Fig. 2.

Comparison of relative homeologous gene expression distributions for 4604 genes expressed both in C. arabica developing seeds (ST3–ST4 green, ST5 yellow, ST6–ST7 blue), and in leaves of C. arabica (light grey) and of Coffea hybrids (dark grey) (χ ² test = 802, d.f. = 36, P-value < 2.2 × 10^–16).

Fig. 3.

(A) Comparison of relative homeologous gene expression between successive stages of maturation (Pearson’s correlation, P-value < 2.2 × 10^–16, correlation coefficient of 0.70, 0.90 for 6269 genes between ST3–ST4 and ST5 and 7147 genes between ST5 and ST6–ST7, respectively). (B) For divergently expressed genes between successive maturation stages, comparison of homeologous gene expression ratios (Pearson’s correlation, P-value < 2.2 × 10^–16, V0.77, 0.93 for 1004 genes between ST3–ST4 and ST5 and 277 genes between ST5 and ST6–ST7, respectively).

Fig. 4.

Comparison of homeologous gene expression distributions of C. arabica genes classified in C and E dominant up and down categories at ST6–ST7 with homeologous gene expression distribution of 8180 genes expressed at this seed maturation stage (χ ² test = 741.72, d.f. = 9, P-value < 2.2 × 10^–16 for E dominant up category, χ ² test = 144.35, d.f. = 9, P-value < 2.2 × 10^–16 for E dominant down category, χ ² test = 584.49, d.f. = 9, P-value < 2.2 × 10^–16 for C dominant up category and χ ² test = 151.61, d.f. = 9, P-value < 2.2 × 10^–16 for C dominant down category). The patterns of gene expression for the four categories considered are shown above the corresponding distribution.

Fig. 5.

Gene expression profiles for the caffeine biosynthetic pathway. Expression profiles for caffeine biosynthetic genes in C. arabica, C. canephora, and C. eugenioides (A), as well as allele-specific expression of C^a and E^a homeologue pairs in C. arabica (B). The x-axis represents seed developmental stages in chronological order and the y-axis represents the gene expression level as normalized read counts. DXMT1 (Cc01g00720), 3,7-dimethylxanthine methyltransferase (caffeine synthase); GSDA (Cc02g31500), guanosine deaminase; MXMT1 (Cc00g24720), 7-methylxanthine methyltransferase (theobromine synthase); NMN, N-methylnucleosidase; SAH, S-adenosyl-l-homocysteine; SAM, S-adenosyl-l-methionine; XMT1 (Cc09g06970), xanthosine methyltransferase.

Fig. 6.

Gene expression profiles for phenylpropanoid genes and for the chlorogenic acid biosynthetic pathway. Expression profiles for putative biosynthetic genes in C. arabica, C. canephora and C. eugenioides (left panel), as well as allele-specific expression of C^a and E^a homeologue pairs in C. arabica (right panel). The x-axis represents seed developmental stages in chronological order and the y-axis represents the gene expression level as normalized read counts. C3H, 4-coumarate 3-hydroxylase; C3′H, p-coumaroyl CoA 3-hydroxylase; C4H, trans-cinnamate 4-hydroxylase; CCoAOMT, caffeoyl-CoA 3-O-methyltransferase; 4CL, 4-coumarate-CoA ligase; COMT, caffeic acid O-methyltransferase; CSE, caffeoyl shikimate esterase; HCT, hydroxycinnamoyl-CoA:shikimate/quinate hydroxycinnamoyl transferase; HQT, hydroxycinnamoyl-CoA quinate hydroxycinnamoyl transferase; PAL, phenylalanine ammonia lyase.