RNA-seq of life stages of the oomycete Phytophthora infestans reveals dynamic changes in metabolic, signal transduction, and pathogenesis genes and a major role for calcium signaling in development

Abstract

Background: The oomycete Phytophthora infestans causes the devastating late blight diseases of potato and tomato. P. infestans uses spores for dissemination and infection, like many other filamentous eukaryotic plant pathogens. The expression of a subset of its genes during spore formation and germination were studied previously, but comprehensive genome-wide data have not been available.

Results: RNA-seq was used to profile hyphae, sporangia, sporangia undergoing zoosporogenesis, motile zoospores, and germinated cysts of P. infestans. Parallel studies of two isolates generated robust expression calls for 16,000 of 17,797 predicted genes, with about 250 transcribed in one isolate but not the other. The largest changes occurred in the transition from hyphae to sporangia, when >4200 genes were up-regulated. More than 1350 of these were induced >100-fold, accounting for 26% of total mRNA. Genes encoding calcium-binding proteins, cation channels, signaling proteins, and flagellar proteins were over-represented in genes up-regulated in sporangia. Proteins associated with pathogenicity were transcribed in waves with subclasses induced during zoosporogenesis, in zoospores, or in germinated cysts. Genes involved in most metabolic pathways were down-regulated upon sporulation and reactivated during cyst germination, although there were exceptions such as DNA replication, where transcripts peaked in zoospores. Inhibitor studies indicated that the transcription of two-thirds of genes induced during zoosporogenesis relied on calcium signaling. A sporulation-induced protein kinase was shown to bind a constitutive Gβ-like protein, which contributed to fitness based on knock-down analysis.

Conclusions: Spore formation and germination involves the staged expression of a large subset of the transcriptome, commensurate with the importance of spores in the life cycle. A comparison of the RNA-seq results with the older microarray data indicated that information is now available for about twice the number of genes than before. Analyses based on function revealed dynamic changes in genes involved in pathogenicity, metabolism, and signaling, with diversity in expression observed within members of multigene families and between isolates. The effects of calcium signaling, a spore-induced protein kinase, and an interacting Gβ-like protein were also demonstrated experimentally. The results reveal aspects of oomycete biology that underly their success as pathogens and potential targets for crop protection chemicals.

Keywords: Gene regulation; Oomycete; Spore development; Transcriptomics; Zoospore.

Fig. 1

Overview of mRNA during development. a heatmap of mRNA levels in isolate 1306 from non-sporulating mycelia (MY), purified sporangia (SP), sporangia chilled in water to induce the cleavage of sporangia into zoospores (CL), zoospores released from the sporangia (ZO), and germinated cysts (GC). Cartoons illustrating each stage are shown in Fig. 2. Data are TMM-normalized CPM values subjected to per-gene normalization and then hierarchical clustering using the Euclidean average linkage method. b Principal component analysis of samples from isolates 1306 (squares) and 88069 (circles), labelled as in panel a. The data in both panels were filtered to only include genes with CPM ≥ 1.0

Fig. 2

Number of genes changing at developmental transitions. a Data from isolate 1306 (black bars) and isolate 88069 (white bars) showing changes between mycelia (MY), sporangia (SP), chilled sporangia (cleaving, CL), and germinated cysts (GC). In the upper-left graph, for example, the bar in the bin labelled 5 represents the number of genes showing more than 2⁵ but less than 2⁶-fold higher mRNA levels in sporangia than mycelia. b Comparison of genes induced ≥5-fold in sporangia versus mycelia, and zoospores versus sporangia, of the two isolates

Fig. 3

Comparisons of Illumina RNA-seq and Affymetrix microarray data. a Expression level calls of mRNA from sporangia. Data are based on the analysis of identical preparations of mRNA from isolate 88069. Pearson’s correlation coefficient (R) was calculated from genes with FPKM > 1. b Fold-change values of genes in sporangia versus mycelia, based on expression calls from the two technologies

Fig. 4

Expression patterns of major classes of regulatory genes. Indicated are data from nonsporulating mycelia (MY), sporangia (SP), cleaving sporangia (CL), zoospores (ZO), and germinated cysts (GC) of isolate 1306. In most panels, the genes are labeled with their five-digit gene number (i.e. 03676) and an abbreviation for the functional subclass. Labels are not shown for protein kinases and phosphatases due to space limitations, but the data are in Additional file 3. Below most panels are small heatmaps that indicate the sum of CPM values for genes in subclasses (Σ CPM, with the number of genes in parentheses), or all genes (ALL). Gene numbers not shown are presented in Additional file 5. a cation channel. b calcium binding. c cyclic nucleotide. d transcription factor. e protein kinase. f protein phosphatase

Fig. 5

Expression patterns of selected genes involved in metabolism. The figure is formatted as in Fig. 4. Gene numbers not shown are presented in Additional file 5. a amino acid transporter. b sugar transporter. c metabolizable ion transporters. d folate transporter. e β-oxidation. f glycolysis and gluconeogenesis. g proteosome. h DNA replication. i transcription. j translation

Fig. 6

Expression patterns of additional metabolic pathways. Presented in panel a are the pentose phosphate pathway, TCA cycle, other genes involved in energy production, glycerolipid and sphingolipid metabolism, other lipid metabolism pathways, purine metabolism, glyoxylate and dicarboxylate metabolism, and amino acid metabolism. Shown in panel b are pathways of pyrimidine metabolism, terpenoid metabolism, and vitamin/coenzyme metabolism.  Panel c shows sulfate esterases, inositol metabolism enzymes, and carbonic anhydrases. Shown are the summed CPM values of genes in each category. The gene numbers are provided in Additional file 5

Fig. 7

Expression patterns of genes encoding plant cell wall degrading enzymes. a heatmaps formatted as in Fig. 4. b expression of eight selected CWDE genes in ammonium sulfate or amino acid-based minimal media (Min + NH4, Min + AA), or rye media. Enzyme functions associated with each category are described in the main text

Fig. 8

Expression patterns of genes potentially playing roles in pathogenesis. a expression of RXLR genes in isolate 1306. b comparison of RXLR expression in 1306 and 88089. Also shown are the expression patterns of c ABC transporters; d secreted protease inhibitors; e secreted proteases; and f genes involved in redox homeostasis

Fig. 9

Effects of inhibitors of calcium signaling. a effect of trifluoroperazine (TFP), verapamil, or 2-APB on zoospore release from sporangia of isolate 88069. b Heatmap showing expression of 1843 cleavage-induced genes in presence and absence of inhibitors. c Comparison of effects of the three inhibitors on the induction of the 1843 genes

Fig. 10

Studies of sporulation-induced protein kinase and and interactor Gβ-like protein. a Two-hybrid analysis showing tests of interaction between full-length kinase (PKS1), kinase catalytic domain (PKS1c; amino acids 135 to 466), Gβ-like protein (GβL), and canonical Gβ. An interaction between PKS1c and GβL is indicated by the strong growth in the center upper panel of the right culture plate. b Co-immunoprecipitation of GβL with PKS1c, showing input and co-precipitated amounts. c Expression patterns of PKS1 and GβL in mycelia grown on minimal media containing casamino acids (MY min) or rye media (MY rye), sporangia (SP), sporangia undergoing zoosporogenesis (CL), zoospores (ZO), germinated cysts (GC), and tomato leaves at 2 to 5 dpi (Tom). d Northern blot of selected transformants containing GβL silencing construct, using the gene for elongation factor 1-α (EF1) as a loading control. The ratio of GβL and EF1 relative to wild type is shown below the blots. Transformants T1, T2, and T6 show levels of GβL equivalent to wild-type, while T4 and T5 are knocked-down by four to five-fold. e RT-qPCR of wild-type strain 1306 with T4 and T5, using primers for GβL (PITG_09556) and flanking gene PITG_09555, which encodes a predicted ubiquitin-ribosomal protein fusion. This confirms the knockdown and indicates that flanking genes are not co-silenced. f Relative growth of 1306, control transformants, and T4 and T5 knockdown (KD) transformants on rye media (rich media), Plich media (semidefined), and in planta. The control transformants (CONT) are the average of non-silenced transformant T1 and two strains expressing the GUS transgene. Growth on media was determined on a dry weight basis, while growth in tomato was determined by qPCR using primers for a high-copy DNA locus [72]. Values are adjusted to 1306 equals 1.0