Defining the SUMO System in Maize: SUMOylation Is Up-Regulated during Endosperm Development and Rapidly Induced by Stress

Abstract

In response to abiotic and biotic challenges, plants rapidly attach small ubiquitin-related modifier (SUMO) to a large collection of nuclear proteins, with studies in Arabidopsis (Arabidopsis thaliana) linking SUMOylation to stress tolerance via its modification of factors involved in chromatin and RNA dynamics. Despite this importance, little is known about SUMOylation in crop species. Here, we describe the plant SUMO system at the phylogenetic, biochemical, and transcriptional levels with a focus on maize (Zea mays). In addition to canonical SUMOs, land plants encode a loosely constrained noncanonical isoform and a variant containing a long extension upstream of the signature β-grasp fold, with cereals also expressing a novel diSUMO polypeptide bearing two SUMO β-grasp domains in tandem. Maize and other cereals also synthesize a unique SUMO-conjugating enzyme variant with more restricted expression patterns that is enzymatically active despite a distinct electrostatic surface. Maize SUMOylation primarily impacts nuclear substrates, is strongly induced by high temperatures, and displays a memory that suppresses subsequent conjugation. Both in-depth transcript and conjugate profiles in various maize organs point to tissue/cell-specific functions for SUMOylation, with potentially significant roles during embryo and endosperm maturation. Collectively, these studies define the organization of the maize SUMO system and imply important functions during seed development and stress defense.

Figure 1.

Description of maize genes encoding central components of the SUMOylation system. Included are genes encoding SUMO-related proteins, the E1, E2, and E3 enzymes involved in conjugation, and two families of DSPs that process/release SUMO. Colored and gray boxes depict coding regions and untranslated regions, respectively. Lines indicate introns; hatched lines denote introns of unknown length. Long introns are not drawn to scale, but their lengths are indicated. Coding regions for signature protein domains and active-site residues are shown, including the active-site Cys in SAE2 and the SCE1 isoforms and the His-Asp-Cys catalytic triads in the ESD4 and OTS SUMO proteases. Domain names in brackets indicate those that are likely but not significant in Pfam. The amino acid (aa) sequence length, chromosome (Chr) location, and maize genome GRMZM accession number of each protein/gene are included to the right.

Figure 2.

The genomes of maize and other plant species encode a family of SUMO-related proteins. A, Domain structures of the plant SUMO family that includes the canonical and noncanonical SUMO isoforms, DSUL containing two SUMO-type, β-grasp domains connected by a variable linker, and SUMO-v containing a long, conserved N-terminal extension before the β-grasp domain. After processing, canonical and noncanonical SUMOs are predicted to terminate in the diGly motif required for conjugation. SUMO-v does not have this motif, and its presence in DSULs is unclear. Amino acid (aa) lengths of the processed maize isoforms are shown to the right. Detailed sequence alignments and phylogenetic comparisons for plant SUMOs can be found in Supplemental Figures S1 to S4. B, Comparison of canonical and noncanonical SUMOs with respect to sequence identity. Shown is the percentage amino acid sequence identity of each compared with Arabidopsis SUMO1. Canonical and noncanonical forms from seed plants, and closely related SUMO relatives from the seedless plants Physcomitrella patens and Selaginella moellendorffii and the algae Chlamydomonas reinhardtii and Volvox carteri, are included. C, Distribution of SUMO family members among plant species. Total gene numbers for each of the four types are indicated for each species. nd, Not detected but likely. Note that the DSUL subtype appears to be cereal specific and that noncanonical SUMOs were not obvious outside of the flowering plants.

Figure 3.

Alignment of SUMO sequences reveals conserved and divergent residues. Included are maize (Zm) SUMO1a, SUMO1b, and SUMO-v, the N- and C-terminal β-grasp domains from DSUL, along with canonical SUMOs from Arabidopsis (At) SUMO1, human (Hs) SUMO2, and yeast (Sc) Smt3. The dashed line locates the β-grasp domain. White arrowheads identify the Lys residues in Arabidopsis SUMO1 shown to be modified by poly-SUMOylation (see Fig. 6B; Colby et al., 2006; Miller et al., 2010). The black arrowhead locates possible processing sites by DSPs that expose the C-terminal diGly motif essential for conjugation in canonical SUMOs; the released amino acids are indicated by the bracket. Residue numbers are shown for each polypeptide; the length of each is shown to the right. Gray and black boxes identify similar and conserved amino acids, respectively. Dashes denote gaps.

Figure 4.

Cereals encode a novel SCE1 subfamily. A, The maize SCE1a to SCE1d and SCE1e to SCE1g isoforms have divergent amino acid sequences. The asterisk highlights the active-site Cys. Arrowheads locate residues responsible for the distinctive negative (red) and positive (blue) electrostatic surface charges that discriminate class II (SCE1f) from class I (SCE1b) E2s (see D). Gray and black boxes identify similar and conserved amino acids, respectively. Dashes denote gaps. B, Phylogenetic analysis of SCE1 sequences reveals a cereal-specific class II subfamily. The phylogenetic relationship of all available plant SCE1 protein sequences was assessed by the neighbor-joining method using MEGA. The scale bar indicates the P distance. A larger version of this tree with more discernible bootstrap values and species abbreviations is shown in Supplemental Figure S6. Alignments of all available monocot SCE1 sequences based on identity/similarity and amino acid properties are shown in Supplemental Figure S7 and S8, respectively. C, Class I and class II SCE1 E2s are predicted to have similar three-dimensional folds. Maize SCE1b (cyan) and SCE1f (magenta) were threaded into the human HsUBC9 (Protein Data Bank [PDB] code 1U9A) template using SWISS-MODEL, and the ribbon diagrams were superimposed. The active-site Cys is colored green and shown in space filling. C, C terminus; N, N terminus. D, Electrostatic surface charges surrounding the active-site Cys (green) are distinct between SCE1b and SCE1f. Surface electrostatics shown in two orientations were calculated using the Adaptive Poisson-Bolzmann Solver plugin in PyMOL. Blue indicates positive charges and red indicates negative charges.

Figure 5.

A recombinant SUMO conjugation system generated with maize components. Various combinations of SUMO1a, the SAE1 and SAE2a polypeptides of the E1 heterodimer, and the SCE1b and SCE1f E2 enzymes bearing 6His, HA, FLAG, and Myc epitopes were expressed in E. coli. Crude extracts prepared after an 8-h induction at 30°C were probed with antibodies against AtSUMO1 (6His-SUMO1a), HA (HA-SAE1), FLAG (SAE2a-FLAG), and Myc (SCE1-Myc; right gel) or anti-6His antibodies (6His-SUMO1a; left gel). SCE1(C-S) indicates the Ser substitutions of the active-site Cys. A, The E1 and E2 enzymes are sufficient to drive SUMOylation. SUMO1a and the SAE1/2 heterodimer were coexpressed with wild-type or C-S versions of SCE1b (left gels) and an E. coli-optimized version of SCE1f (right gels). B, The truncated SAE2a splice variant T2 (Trunc) is functional but less active than full-length SAE2a (FL). The SAE2a polypeptides bearing a C-terminal FLAG tag were coexpressed individually with SUMO1a, SAE1, and SCE1b. Asterisk locates the T2 SAE2a truncation. C, Direct comparison of the conjugating activity of the class II E2 SCE1f with representative class I E2s, SCE1b, and SCE1d. D, Both class I and class II SCE1s catalyze the formation of poly-SUMO chains. Wild-type SUMO1a (WT) or the Lys-less K0 mutant blocked in forming SUMO chains were coexpressed with the SAE1/2 heterodimer and either SCE1b or SCE1f.

Figure 6.

Mass spectrometric analyses identify SUMO-SUMO attachment sites in maize SUMO1a. SUMO conjugates were generated with 6His-SUMO1a(M1-R,H89-R)-GG, using the E. coli-based system shown in Figure 5, followed by purification via nickel-nitrilotriacetic acid agarose (Ni-NTA) affinity chromatography. A, Silver-stained gels of purified ZmSUMO1a conjugates assembled by SCE1b or SCE1f. Red boxes highlight regions of the gel that were analyzed by mass spectrometry. SUMO footprints detected by MS/MS on specific SUMO1 Lys residues are shown; numbers in parentheses indicate the number of PSMs bearing these footprints. B, Positions of SUMOylated Lys residues mapped onto the amino acid sequence of 6His-SUMO1a(M1-R,H89-R)-GG. Red and black arrowheads indicate Lys residues (K) with and without detectable SUMO footprints, respectively. Line widths reflect the number of mass spectrometry-detected PSMs for each peptide: red lines denote that at least one peptide had a SUMO footprint, and black lines denote unmodified peptides. C, Predicted three-dimensional ribbon diagram of maize SUMO1a generated by threading its sequence onto the three-dimensional model of human SUMO2 (PDB code 2AWT) using SWISS-MODEL. The positions of unmodified and SUMOylated Lys residues are shown in black and red, respectively. C, C terminus. GG represents the terminal diGly motif in mature SUMO1a.

Figure 7.

Transcriptome analysis uncovers both constitutive and tissue-specific expression patterns for maize SUMO system components. RNA-seq experiments representing 80 developmentally distinct maize tissues (Stelpflug et al., 2015) were mined for reads per kilobase per million reads (RPKM) values for individual SUMO pathway genes. Note that maize Siz1b, Siz1c, and Ots1c transcripts are subdivided into 5′ and 3′ fragments as a result of prior gene annotation errors. Tissues were harvested during vegetative (V) or reproductive (R) stages of growth. Primary roots were divided into zones: Z1 (1 cm of root tip including root cap and division zone), Z2 (root elongation zone), Z3 (bottom half of the differentiation zone), and Z4 (top half of the differentiation zone). DAP, Days after pollination; DAS, days after sowing; DZ, differentiation zone; EZ, elongation zone; MZ, meristematic zone; SAM, shoot apical meristem.

Figure 8.

SUMOylation profiles vary dramatically among maize tissues. Total protein extracts derived from the indicated tissues were subjected to immunoblot analysis with anti-At SUMO1 or anti-histone H3 antibodies (control). Black arrowheads locate free SUMO. Brackets locate high-molecular-mass SUMO conjugates. White arrowheads highlight several SUMO conjugates that appear to be tissue and/or development specific. A shorter exposure of the blots around free SUMO1 was included to better show variations in abundance. A, Analysis of various tissues. Included are whole seeds collected at increasing DAP, endosperm (Endo) and embryos from 16-DAP seeds, 10-d-old shoots from the second leaf sampled before (−) or after (+) a 30-min heat shock at 42°C, 10-d-old roots, and immature tassels. Cob, husk, and silk tissues were collected from an unpollinated silking ear. B, Time course of seed development showing that the rise of SUMO conjugates follows an increase in SCE1 protein. Seeds were collected at the indicated times from self-pollinated field-grown plants. SCE1 proteins were detected using antibodies against Arabidopsis SCE1. Immunoblotting with anti-histone H3 antibodies was included to show protein loading.

Figure 9.

SUMOylation in maize is strongly induced by heat and oxidative stress. Crude extracts from 10-d-old leaves (A and C–E) or roots (B) were subjected to immunoblot analyses with anti-AtSUMO1 or anti-histone H3 antibodies (control). Free SUMO and SUMO conjugates are highlighted by the arrowheads and brackets, respectively. Asterisks locate species nonspecifically recognized by the anti-AtSUMO1 antibodies. In all but F, immunoblotting with anti-histone H3 antibodies was included to confirm nearly equal protein loading. A, SUMO conjugates increase in abundance soon after exposure to heat stress. Seedlings grown at 28°C were shifted to 42°C for 1 h and returned to 28°C for 1 h. Leaves were collected at the indicated times starting at the upshift in temperature. B, Stress-induced SUMOylation also occurs in roots and is induced by heat and oxidative stress. At left, roots from seedlings grown at 28°C were exposed to 42°C for 1 h and then returned to 28°C for 1 h before harvest. At right, roots were exposed for 30 min to varying concentrations of hydrogen peroxide (H₂O₂) before harvest. C, Heat-induced SUMOylation is affected by the duration of high temperature. Seedlings grown at 28°C were shifted to 42°C for varying times (arrows) and then returned to 28°C. Leaves were collected at the indicated times starting at the temperature upshift. D, SUMOylation is induced rapidly by high temperature. Seedlings grown at 28°C were exposed to the indicated temperatures (°C) for 30 min before harvest. E, A refractory period is required to induce a second upshift in SUMOylation in response to high temperature. Seedlings grown at 28°C were shifted to 42°C for 30 min and then returned to 28°C for varying times (h) before a second 30-min exposure to 42°C. Leaves were collected either at the beginning of the heat stress or at the end of the 30-min pulse. The time between the start of each pulse is indicated. F, SUMOylation occurs mainly in the nucleus. Leaves were collected prior to or immediately after a shift from 28°C to 42°C for 30 min. Total extracts (T) were separated into cytoplasmic (C) and nuclear (N) fractions by Percoll gradient centrifugation and subjected to immunoblot analysis with anti-AtSUMO1 antibodies. Immunoblotting with anti-PUX1 and anti-histone H3 antibodies was included to confirm enrichment of the cytoplasmic and nuclear compartments, respectively.