Introduction

Heterosis refers to superior growth and fitness in the offspring relative to one or both parents. Since the discovery of the phenomenon by Charles Darwin in 18761, heterosis has been widely applied in agriculture to improve the production of important crops such as maize and sorghum and farm animals such as pigs and sheep2,3. Compared to regular hybrids, heterozygotes and heterosis are permanently fixed in the allotetraploids such that mechanisms for molecular changes can be readily studied4,5. Although the genetic, genomic, and epigenetic bases of heterosis have been extensively studied2,3,6, biochemical bases for heterologous protein-protein and protein-DNA interactions in heterosis are poorly understood.

A recent study found that the heterozygote of SFT locus between an allele of the cultivated tomato and an introgressed allele from the wild tomato is associated with fruit yield heterosis7. This is consistent with the conventional example of hybrid (heterozygote) advantage for sickle cell anemia; individuals who are homozygous for the mutant allele have the disease, whereas heterozygous individuals have some susceptibility to the disease and are more tolerant to malaria infection than individuals who are homozygous for the normal allele8,9. This suggests heterozygotic (or overdominant) effect on heterosis2,5. Biochemical bases for this effect could be associated with heterodimeric interactions, leading to change in the activities or DNA-binding affinities of heteromers. For example, E2F is a mammalian transcription factor that binds directly to the adenovirus E2 promoter for activation10. Mixing individual E2F subunits together dramatically increases (100- to 1000-fold) the specific DNA binding activity. In maize, heterodimers of alcohol dehydrogenase are more stable than either homodimers, suggesting a biochemical basis for this case of single-locus heterosis11. Protein inhibitors of heterologous α-amylases and proteinases are useful for plant protection. In wheat, different subunits of these inhibitors, which are encoded by homoeologous genes on chromosomes 4A, 4B, and 4D, respectively, exist in hexaploid (AABBDD), tetraploid (AABB), and diploid (DD) wheats. Mixing subunits of the hexaploid wheat increases inhibitory activities to 133%, compared to mixing subunits of the diploid wheat (33%)12. Thus, heterodimers of protein and transcription factors are either more stable or have stronger binding activities to the promoters of downstream genes in various biological pathways, leading to heterosis2,6,13.

The heterozygote advantage model can be applied to other protein and transcription factors such as circadian clock regulators5,6,14. Circadian clocks regulate physiology and metabolism in plants and animals15,16,17,18. In Arabidopsis, the circadian clock mediates expression of 30% or more genes in any given growth conditions19,20. The central clock regulators consist of the morning regulators, CIRCADIAN CLOCK ASSOCIATED1 (CCA1) and LATE ELONGATED HYPOCOTYL (LHY), and the evening factor, TIMING OF CAB EXPRESSION1 (TOC1)16,17,21. CCA1 HIKING EXPEDITION (CHE), a TCP transcription factor, forms a negative feedback regulatory loop in mediating CCA1 repression during the day22. TOC1 is a DNA-binding transcriptional repressor of CCA1 at the evening phase23,24.

Interestingly, the morning-phased clock regulators are down-regulated, while the evening-phased ones are activated in Arabidopsis intraspecific hybrids25,26 and allotetraploids (doubled interspecific hybrids)14, which are formed by pollinating Arabidopsis thaliana with Arabidopsis arenosa pollen27. Specifically, both A. thaliana CCA1 (AtCCA1) and A. arenosa CCA1 (AaCCA1) are down-regulated in the allotetraploids during the day, which correlates with increased levels of chlorophyll and starch, promoting growth vigor26. Surprisingly, AtCCA1 is expressed at lower levels than AaCCA1 in the allotetraploids; the underlying mechanism is largely unknown. Here we take the advantage of the circadian transcriptional feedback loop to test a biochemical model for homoeologous protein-protein interactions and protein-DNA (cis-regulatory element) recognition in altering allelic expression levels. This biochemical model could explain a molecular basis for heterosis in interspecific hybrids and allotetraploids.

Results

CHE expression is affected by hd1 mutation

Chromatin modifications including histone acetylation and deacetylation mediate expression of circadian clock genes in mammals and plants14,28,29. In A. thaliana, histone deacetylation represses expression of circadian clock genes29,30. However, the genetic interaction between histone deacetylase and clock gene expression is unknown. A. thaliana Histone Deacetyase1 (AtHD1) is a general transcription repressor or activator31,32. In hybrid rice, expression of the rice histone deacetylase (OsHDT1) showed a circadian rhythm33. Here we tested if the AtHD1 affects clock gene expression and vice versa, if the circadian clock affects AtHD1 expression. In the wild-type plants, expression of AtHD1 showed slight diurnal rhythms with relatively higher expression levels at ZT9 and ZT18 (ZT, Zeitgeber time; ZT0 = dawn) (Fig. 1a). As expected, AtHD1 expression was abolished in the athd1-t1 mutant (Fig. 1a). athd1-t1 is a T-DNA insertion line in exon 2 of AtHD1, resulting in a null mutation31. In Arabidopsis, CHE was found to bind to the CCA1 promoter and repress the promoter activities22. Using qRT-PCR, we have confirmed that CHE showed high expression peaks around ZT6 and ZT9 in plants grown under 16 h light/8 h dark cycles (Fig. 1b). This is consistent with the reported CHE expression to which it peaked at 9 hours after the subjective dark period under continuous light22. In the che mutant, this oscillated expression was lost in the che mutant. In the athd1-t1 mutant31, CHE expression levels were reduced at ZT9 and ZT12 (Fig. 1c). In contrast, expression levels of AtHD1 remained unchanged in the che mutant (Fig. 1d). These data suggest that AtHD1 positively mediates CHE expression directly or indirectly. In contrast, CHE does not affect AtHD1 expression.

Figure 1: Repression of CHE in the A. thaliana athd1 mutant.
figure 1

Quantitative real-time PCR showing temporal expression of HD1 and CHE in the A. thaliana wild-type Ws or Col-0 (square) and its corresponding athd1-t1 (in Ws, circle) or che mutant (in Col, circle). The target transcript levels were normalized with the EF1α endogenous control. ZT, Zeitgeber time; ZT0 = dawn. Open bar represents light period and fill bar represents dark period. (a) HD1 expression in Ws and the athd1 mutant. (b) CHE expression in Col and the che mutant. (c) CHE expression in Ws and the athd1-t1 mutant. (d) HD1 expression in Col and the che mutant. Error bars represent standard deviation from 3 replicates.

Full size image

Roles of HD1, CHE and TOC1 in CCA1 expression

Within the central loop of the circadian clock, CHE represses CCA1 expression during the day22. In addition, using a CCA1 promoter-driven luciferase (LUC) expression in bioluminescence analysis, Pruneda-Paz et al.22 revealed that the CCA1 promoter activity was upregulated in the che mutant and down-regulated in 35S-driven CHE overexpressors22. Since CHE expression is affected in the athd1-t1 mutant, HD1 and CHE could act cooperatively to repress CCA1 expression. We tested this using a ProCCA1:LUC reporter line34 in the che or hd1 single or the che hd1 double mutant (Fig. 2a). Consistent with the previous finding22, CHE represses CCA1, and the CCA1 promoter activity was upregulated in the che mutant (Fig. 2a). In the hd1 mutant, CCA1 is diurnally expressed35. Although oscillation of the CCA1 promoter activity was observed, its peaks were dramatically reduced in the athd1-1 mutant, indicating that HD1 acts an activator for CCA1 expression. This is probably because expressions of other clock regulators, such as CHE and TOC1, are affected in the athd1 mutant30,36. Indeed, down-regulation of the CCA1 promoter activity was largely relieved in the che hd1 double mutant, suggesting that CHE and HD1 act antagonistically in gating CCA1 expression. TOC1 is an evening-expressed protein and it has long been recognized as a positive regulator of CCA137,38,39. In the toc1 mutant, CCA1 expression peak was reduced compared to the wild type plants in qRT-PCR assays38,39. Consistent with data from these studies, expression of LUC driven by the CCA1 promoter in the toc1 mutant was down-regulated in the ProCCA1:LUC reporter line (Fig. 2b). While the activity of the CCA1 promoter was upregulated in the che mutant, the LUC expression peaks were further reduced in the che toc1 double mutant. Therefore, these data suggested a role for genetic interactions between TOC1 and CHE in regulating CCA1 expression. Recent biochemical studies revealed that TOC1 indeed functions as a repressor of CCA123,24. Further studies are needed to resolve the biochemical interactions among TOC1, CHE, and CCA1.

Figure 2: ProCCA1:LUC expression in che, hd1 and toc1 mutants.
figure 2

Bioluminescence counts (Mean ± S.E.M.; in thousands) from ProCCA1:LUC expression were measured in (a) wild type Col, the che, hd1 and che hd1 mutants; and (b) toc1 and che toc1 mutants, respectively. Measurements were taken with 8-days old seedlings over a 4-days period under 16 h light (open bars) and 8 h dark (filled bars) conditions. Lines with asterisks mark the range of time points with statistically significant differences between Col and the indicated mutant (P < 0.05, Student’s t test). For hd1 or che toc1, ProCCA1:LUC expression was significantly down-regulated (P < 0.05) when compared to that in any other genotypes at all time points. Significant reduction (P < 0.05) of ProCCA1:LUC expression was also found in che and che hd1 comparison. S.E.M. = Standard error of the mean.

Full size image

Preferential binding of A. arenosa CHE homolog to the A. thaliana CCA1 promoter

In interspecific hybrids of animals and allotetraploids of plants40,41, cis- and trans-acting effects mediate expression of parental alleles originating from different progenitors. In Arabidopsis allotetraploids that are formed between A. thaliana (At) and A. arenosa (Aa), AtCCA1 is expressed at lower levels than AaCCA114 (see Supplementary Fig. S1a). The altered CCA1 expression in the allotetraploids are consistent with the upregulation of morning-phased genes in photosynthetic and starch biosynthetic pathways41 that promote biomass heterosis14. The molecular basis for this differential CCA1 allelic accumulation is unknown. We predict that this allelic CCA1 repression results from changes in the cis-regulatory elements (promoters) of CCA1 in combination of trans-acting effects by factors such as CHE and HD1 on CCA1 expression6. In A. thaliana, CHE binds to the CHE-binding site of the CCA1 promoter22,34 and represses CCA1 expression. Although the A. arenosa CCA1 promoter contains the same CHE-binding site, the flanking sequences were different compared to the A. thaliana promoter (Fig. 3a). The cis-element variation between AtCCA1 and AaCCA1 promoters could cause different binding activities by AtCHE and/or AaCHE, leading to allelic CCA1 expression changes in allotetraploids (see Supplementary Fig. S1a). To test this, we cloned cDNA of AtCHE and AaCHE to produce recombinant proteins in bacterial cells (see Supplementary Fig. S2) and purified the proteins using a maltose binding protein (MBP) that was fused to the 5′ of the CHE protein. The quantity and quality of proteins were verified by Bradford assay and 10% SDS-PAGE, respectively (Fig. 3b). In a gel electrophoretic mobility shift assay (EMSA), the same amount of the purified recombinant proteins was used to test their binding affinities to the native AtCCA1 or AaCCA1 promoter (AtCCA1p or AaCCA1p) fragment containing the CHE binding site (Fig. 3a). The recombinant AaCHE (rAaCHE) had a 1-3-fold higher binding affinity than rAtCHE to the AtCCA1p promoter (Fig. 3c, compare lanes 5 with 9 and 6 with 10) than AaCCA1p promoter (Fig. 3d, compare lanes 5 with 9 and 6 with 10). These data suggest that rAaCHE preferentially bound to the CCA1 promoter compared to rAtCHE.

Figure 3: AaCHE preferentially bound to the CCA1 promoter (CCA1p).
figure 3

(a) Sequence alignment of the AtCCA1 promoter (AtCCA1p) (from −315 to −182 relative to the transcription start site, +1) and the AaCCA1p (from −325 to −191) fragments used in gel electrophoretic mobility shift assay (EMSA). The CHE binding site (CHE BS) is indicated. (b) Coomassie stain of purified recombinant AtCHE and AaCHE with N-terminus MBP fusion. MBP: maltose binding protein. (c,d) EMSA showing preferential binding of the recombinant AaCHE (rAaCHE) to either (c) AtCCA1p or (d) AaCCA1p. MBP alone is used as a negative control in the binding reaction. Shifted DNA bands (i and ii) caused by rAaCHE binding were quantified by imageJ densitometry and normalized against the corresponding shifted band caused by rAtCHE. The relative quantifications were indicated at the bottom.

Full size image

In a competition assay using an increased amount of cold (unlabeled) AtCCA1p or AaCCA1p probe, CHE binding had a higher specificity to the AtCCA1 promoter than to the AaCCA1 promoter (Fig. 4). The cold AaCCA1p probe competitor caused more abundant mobility shift of the labeled AtCCA1p probe than the cold AtCCA1p probe competitor (compare lanes 5 with 9). Moreover, for the same amount of cold competitors, the cold AtCCA1p showed a higher affinity to rAaCHE than the cold AaCCA1p in competition with the labeled AtCCA1p. Therefore, these in vitro data suggest preferential binding of AaCHE to the AtCCA1 promoter, which agreed with and could explain the lower expression level of AtCCA1 than that of AaCCA1 homoeolog in the allotetraploids in vivo14.

Figure 4: AaCHE preferentially bound to the AtCCA1p.
figure 4

Recombinant AaCHE (rAaCHE, 4pmole) was used for EMSA in the presence of 32P-labeled AtCCA1p (15 fmol) probe (lane 2). Competition assays were performed using 1x, 5x, 25x, and 125x of cold AtCCA1p (lanes 3–6) or AaCCA1p (lanes 7–10) probe as indicated. MBP (4 pmol) is used as a negative control (lane 1). The relative imageJ densitometry quantifications of shifted DNA bands (i and ii) in the presence of cold AaCCA1p (lanes 7–10; relative to the corresponding band in the presence of cold AtCCA1p in lanes 3–6) were indicated at the bottom.

Full size image

Interactions between homoeologous CHE and HD1 proteins

HD1 is a general transcriptional repressor36,42, which could be recruited by CHE to repress target genes such as CCA1. It is unclear if CHE interacts with HD1 biochemically. Here we tested their protein-protein interactions through two approaches. A yeast two-hybrid assay has shown positive interactions between CHE and HD1 (see Supplementary Fig. S4a). Using AtCHE and AaCHE as baits, weak interactions were detected with the respective AtHD1 and AaHD1. Interestingly, by substituting AtHD1 with AaHD1 as prey when using AtCHE as bait, there was a slight increase in the interactions (AtCHE-AaHD1). Similarly, by substituting AaHD1 with AtHD1 as prey when using AaCHE as bait, the interaction level was increased. Overall, there was a trend of enhanced interactions between the proteins that originated in different species (AtCHE-AaHD1 or AaCHE-AtHD1) than that from the same-species (AtCHE-AtHD1 or AaCHE-AaHD1).

The yeast-two hybrid data were less quantitative due to the weak interaction between CHE and HD1 (as shown with the limited growth in the triple dropout media). Therefore, we also quantified the interactions between At/AaCHE and At/AaHD1 homoeologous proteins in vivo using a split green fluorescent protein (GFP) complementation in a bacterial two-hybrid assay43. AtCHE and AaCHE cDNA were cloned and fused downstream of the N-GFP, respectively. In parallel, AtHD1 and AaHD1 cDNA were fused upstream of the C-GFP, respectively (see Supplementary Fig. S3). Expression of recombinant N-GFP proteins was driven by the T7 promoter in the presence of isopropyl β-D-1-thiogalactopyranoside (IPTG), while expression of the recombinant C-GFP proteins was under the control of the PBAD promoter, which is inducible by L-(+)-arabinose. The expression and reassembly of the GFP were visualized using GFP fluorescence microscopy (Figs 5a–f). As expected, positive GFP signals could be seen for four protein (AtCHE, AaCHE, AtHD1 and AaHD1) combinations. GFP signals in cells containing the empty vectors and the leucine zipper constructs (NZ-CZ) were used as negative and positive controls, respectively. To normalize the difference in expression constructs, approximately 10,000 cells with positive GFP signals were counted for each protein combination in flow cytometer analysis to estimate relative levels of fluorescence intensities (Fig. 5g). Between the combination of AtCHE with AtHD1 and AaHD1, the GFP intensity was significantly higher in AtCHE-AaHD1 than in AtCHE-AtHD1, suggesting that AtCHE preferentially interacts with AaHD1. Similarly, the GFP intensity was significantly higher in AaCHE-AtHD1 than in AaCHE-AaHD1 combinations, indicating stronger binding of AaCHE with AtHD1 than with AaHD1. Therefore, the combination between proteins from different species (AaCHE-AtHD1 or AtCHE-AaHD1) has stronger interactions than the combination between the proteins from the same parent or species (AtCHE-AtHD1 or AaCHE-AaHD1). The result from the split GFP complementation approach is consistent with the result from the independent yeast two-hybrid assay (see Supplementary Fig. S4). Notably, when comparing AtCHE-AtHD1 and AaCHE-AaHD1 interactions, an opposite trend was observed between results from the yeast two-hybrid and split GFP assays. This discrepancy could result from the comparisons between two protein type variables (CHE and HD1) in each system, therefore making it difficult to interpret the results. In combination with the preferential binding of AaCHE to the AtCCA1 promoter (Figs 3 and 4), this stronger protein interaction between the two proteins from different species suggests a biochemical basis for down-regulation of AtCCA1 expression levels during the day in the allotetraploids14.

Figure 5: Protein-protein interactions between HDAC and CHE.
figure 5

Split GFP constructs were expressed in E. coli for quantification of protein-protein interaction. In the split constructs, the N-GFP was fused with CHE and the C-GFP was fused with HD1. (a–f) GFP fluorescent images were taken from E. coli cells containing different combinations of split GFP constructs: (a) empty vector control (NC-control), (b) positive vector control with leucine zipper constructs (NZ-CZ), (c) AtCHE-AtHD1, (d) AaCHE-AaHD1, (e) AaCHE-AtHD1, (f) AtCHE-AaHD1. (g) Flow cytometer quantification of GFP expression levels as indicators for protein-protein interaction. Normalized GFP intensity was calculated based on the fluorescence intensity divided by the percentage of cells showing positive GFP signals. Error bars represent standard deviation from 3 replicates. Scale bars = 10 μm. Asterisks mark statistically significant differences between comparisons (P < 0.05).

Full size image

AaCHE interacts stronger than AtCHE with TOC1

In the evening loop of the circadian clock, genetic studies showed that TOC1 activates CCA1 expression through tethering CHE from the CCA1 promoter22. However, TOC1 can bind to the CCA1 promoter and repress its activities23. Since our data suggest a role for genetic interactions between TOC1 and CHE in regulating CCA1 expression (Fig. 2b), we cloned the homologous AtTOC1 and AaTOC1 from the allotetraploids (see Supplementary Fig. S5) to study their interactions with CHE at biochemical levels. To achieve this, we used the yeast two-hybrid assay to test protein interactions involving AtTOC1 or AaTOC1 and AtCHE or AaCHE (Fig. 6). As expected, strong interactions between TOC1 and CHE were detected in the quadruple synthetic dropout medium (SD-Leu/-Trp/-Ade/-His) containing X-α-gal (Fig. 6a). The strength of interactions between TOC1 and CHE was quantified by a quantitative assay for X-α-gal activity44. When AtTOC1 was used as the bait protein, AtTOC1-AaCHE interaction was significantly stronger than the AtTOC1-AtCHE interaction. In contrast, the AaTOC1-AaCHE combination showed a stronger interaction than the AaTOC1-AtCHE combination when AaTOC1 was used as the bait protein (Fig. 6b). These data suggest that AaCHE interacts strongly with TOC1 than AtCHE regardless which combination is used and that not all homoeologous protein interactions are favored in hybrid combinations. Nevertheless, AaCHE was found to preferentially bind to the CCA1 promoter compared to AtCHE (Fig. 3). In addition, comparing with AtTOC1, AaTOC1 showed a stronger interaction with AtCHE or AaCHE. This in vitro result is consistent with higher expression levels of AtTOC1 than AaTOC1 in the allotetraploids in vivo14. The data collectively suggest that the observed preferential protein-protein interactions between TOC1 and CHE could act in concert with the preferential binding of AaCHE to the AtCCA1 promoter in altering the AtCCA1 expression peaks in the allotetraploids and hybrids14,25,45.

Figure 6: Interaction of TOC1 and CHE in yeast.
figure 6

(a) Double dropout medium (DDO) was used to select yeast strain containing the bait (AtTOC1 or AaTOC1) and prey (AtCHE or AaCHE) proteins. Qualitative protein-protein interaction was visualized using X-α-gal in quadruple dropout (QDO/X) medium. Yeast vectors expressing P53 (Gal4 DNA-BD fused with murine p53), T (Gal4 AD fused with SV40 large T-antigen), and Lam (a negative control vector with Gal4 BD fused with lamin) were used as controls in various indicated combinations. (b) Quantitative α-galactosidase assay for protein-protein interaction in yeast. The relative α-galactosidase activity was measured against the positive p53-T interaction. Error bars represent standard deviation from 3 biological replicates. Asterisks indicate statistically significant differences between comparisons (P < 0.05).

Full size image

Discussion

Altered gene expression and signaling pathways have been shown to potentially contribute to heterotic phenotypes in hybrids6,7,13. Results from this study provided a model for the fine-tuning of the circadian clock in Arabidopsis allotetraploids. The heterologous protein interaction model could provide a biochemical explanation for biased expression of homoeologs in interspecific hybrids or allotetraploids (Fig. 7). Within the central oscillator of the clock, CCA1, a MYB-related transcription factor, binds to the evening element of its target genes14,39. In A. thaliana, CCA1 expression is mediated by both CHE and TOC117,21. In the allotetraploids, AaCHE preferentially binds to the AtCCA1 promoter, thereby repressing its expression during the day. Such preference of protein-DNA interactions could lead to the down-regulation of the AtCCA1 homoeolog in the allotetraploids14. In this study, both the yeast two-hybrid assay and the bacterial assay with split GFP complementation confirmed the interaction of CHE and HD1. Since HD1 is a general transcriptional regulator36,42, the repression of CCA1 alleles by CHE could be mediated through preferential interactions between AtHD1 and CHE. However, the recruitment of HD1 to the CCA1 promoter through its interaction with CHE remains to be determined. As a general transcriptional repressor, HD1 could be targeted to other clock regulators and output genes in the circadian-mediated transcriptional network, which in turn, indirectly contributes to CCA1 expression variation during the day. The preferential interaction between AaCHE and TOC1 could also contribute to the allelic expression differences of CCA1 homoeologs in the allotetraploids. As a result, CCA1 expression bias could affect other downstream targets of the clock, contributing to altered metabolic and signaling pathways in the hybrids. Although the results from in vitro studies should be further tested to confirm the interaction of CHE and HD1 homoeologs in planta using alternative approaches such as bimolecular fluorescence complementation46, the presence of both transgene and endogenous duplicate genes may complicate data interpretation. In addition, generating and characterizing transgenic allotetraploid plants is not a trivial task.

Figure 7: A simplified model for CHE-mediated allelic CCA1 expression in Arabidopsis allotetraploids.
figure 7

In the morning, CHE, acts to down-regulate CCA1 expression during the day. The down-regulation of CCA1 by CHE could be mediated through its interaction with HD1 with a possibility of recruiting HD1 to the CCA1 promoter. In Arabidopsis allotetraploids, A. thaliana (At) and A. arenosa (Aa) homoeologous proteins such as CHE and HD1 could form heterologous interactions, leading to differential binding affinity of the protein complex (e.g., AaCHE-AtHD1) to the AtCCA1 (black) or AaCCA1 (grey) promoter, causing more reduction on AtCCA1 transcripts than on AaCCA1 transcripts (light/heavy arrows indicate transcript activation levels, see Supplementary Fig. S1)14,41,48. In addition to the heterologous protein complex, cis-regulatory changes in the AtCCA1 and AaCCA1 promoters could also reduce CCA1 expression levels. At night, At/AaTOC1 interacts preferentially with AaCHE, thereby, relieving CCA1 from repression. The interactions between At/AaTOC1 and At/AaHD1 homoeologous proteins have not been tested. As a transcriptional repressor, HD1 could potentially mediate CCA1 expression independent of CHE through other clock regulators (with a question mark) in the hybrids.

Full size image

In the biological network regulated by the circadian clock, hybrids or allotetraploids induce allelic expression changes of the regulatory genes such as CCA1. This allelic expression variation are associated with the parent-of-origin effect34 and histone modifications14. At the protein level, transcription factors such as CHE from maternal (red) and paternal (blue) alleles in the hybrids or allotetraploids could form homo- or heterodimers (Fig. 7), which change biochemical functions including inhibitory effects12 or binding affinities to the promoters of downstream genes in a biological pathway, leading to heterozygote advantages in the hybrids6. Given the complexity of circadian clock feedback loops, the heterologous protein-protein and/or protein-DNA interactions could alter the overall clock-mediated transcriptome in the allotetraploids, providing a potential for acquiring novel adaptive traits in the hybrids. At the systems level, overall protein stability and metabolism are predicted to play a critical role in heterosis13. In particular, one can predict that the interactions between heterologous subunits of transcription regulators such as the circadian-clock proteins enhance the strength to regulate output traits directly relevant to the development and evolution of hybrids and polyploids in plants and animals. Therefore, heterologous protein-protein and protein-DNA interactions could be generalized to provide a biochemical basis for biased expression of circadian clock genes that are associated with biomass heterosis14.

Methods

Plant materials and growth conditions

A. thaliana T-DNA mutants, che1 (SALK_143403) and hda19 (SALK_139445) were obtained from the Arabidopsis Biological Resource Center (ABRC). Sources of other plant materials include athd1-t131, toc1-10147, and ColC (ProCCA1:LUC in Col-0 ecoptype)34. All plants were grown at 22 °C under a 16/8-h light/dark cycle.

Cloning of A. thaliana (At) and A. arenosa (Aa) clock genes from the allotetraploids

Since AtCHE (At5g08330) contains no intron, therefore, genomic DNA encoding the homoeologous AtCHE and AaCHE were amplified from genomic DNA of the resynthesized Arabidopsis allotetraploid (Allo) and cloned into pGEM-T vector (Promega, Madison, WI, USA). The resulting clones, pGEM-T/AtCHE and pGEM-T/AaCHE were verified by sequencing and used for other plasmid constructions in this study. For homoeologous HD1 and TOC1 cDNA, they were amplified from cDNA prepared from Allo and cloned into pGEM-T. The resulting clones, pGEM-T/AtHD1, pGEM-T/AaHD1, pGEM-T/AtTOC1 and pGEM-T/AaTOC1 were verified by sequencing and used for cloning of plasmids for the yeast two-hybrid assay. All cloning were done using Phusion high-fidelity DNA polymerase (New England Biolabs, Ipswich, MA, USA) following manufacturer’s instructions. Primers for cloning of the At and Aa clock genes were listed in Supplementary Table S1.

Bacterial expression and purification of recombinant CHE

CHE cDNA inserts from pGEM-T/AtCHE and pGEM-T/AaCHE, respectively, were isolated using EcoRI and SalI restriction digestion and cloned into the pMalC2 vector (New England Biolabs, Ipswich, MA, USA). The homologous CHE was fused downstream of the Maltose binding protein and a TEV protease cleave site, yielding pMalC2/AtCHE and pMalC2/AaCHE, respectively. For bacterial expression of the recombinant proteins (MBP-rAtCHE or MBP-rAaCHE), corresponding plasmid was transformed into BL21-CodonPlus(DE3)-RIL competent cells (Stratagene, Santa Clara, CA, USA). Expression of the recombinant proteins was induced by subculturing 1 mL starter culture (OD600 = 0.5) into 100 mL LB media containing 1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) and shaking at 37 °C for 2.5 hours. For protein purification, bacterial cells were harvested from 40 mL induced culture ‘resuspended in 1 mL column buffer (20 mM Tris HCl pH7.4, 200 mM NaCl, 1 mM EDTA) containing 1 mg/mL lysozyme, and incubated on ice for 30 minutes. Then the bacterial cells were lysed by sonication and supernatant from the lysed cultures were collected for subsequent column purified using amylose resin (New England Biolabs, Ipswich, MA, USA) and finally eluted with 1 mL maltose (10 mM). The eluted proteins were then quantified using Bradford assay (BioRad) and 2 pmol/μL purified MBP, MBP-rAtCHE and MBP-rAaCHE were prepared as protein working stocks for subsequence electrophoretic mobility shift assay. The quality of purified proteins was verified using 10% SDS-PAGE.

Split-GFP reassembly constructs

The split-GFP reassembly vectors (pET11a-link-NGFP and pMRBAD-link-CGFP) and the positive control plasmids (pET11a-Z-NGFP and pMRBAD-Z-CGFP) were kindly provided by Lynne Regan from Yale University. Target cDNAs (AtCHE, AaCHE, AtHD1 and AaHD1) were amplified from the corresponding cloned cDNA from the allotetraploids. Restriction enzyme sites flanking the cDNA were introduced in the primers for cloning into the split-GFP vectors (see Supplementary Table S1). AtCHE and AaCHE were cloned into pET11a-link-NGFP for bacteria expression of recombinant NGFP:AtCHE and NGFP:AaCHE, respectively. AtHD1 and AaHD1 were cloned into pMRBAD-link-CGFP for bacterial expression of AtHD1:CGFP and AaHD1:CGFP, respectively.

qRT-PCR

For gene expression analyses, mature leaves from 3–4 weeks old plants were harvested. Total leaf RNAs were prepared using plant RNA reagent (Life technologies, Carlsbad, CA, USA) according to manufacturer’s instructions. cDNA synthesis was performed using Omniscript reverse transcription kit (Qiagen, Venlo, Limburg, Netherlands). Quantitative RT-PCR analysis was performed using Faststart universal SYBR green master (Rox) (Roche, Basel, Switzerland) in an ABI7500 machine (Applied Biosystems, Waltham, MA, USA).

Bioluminescence assays and data analysis

Luciferase reporter line containing the ProCCA1:LUC construct34 was crossed with various clock mutants (che, hd1 or toc1) and homozygous transgenic lines in various single or double mutants (che/hd1 or che/toc1) were selected for luciferase expression assay using a TopCount NXT luminometer and scintillation count (Perkin-Elmer, Waltham, MA, USA) and the means ± SE luciferase activity was calculated following the procedures described in ref. 34.

Electrophoretic mobility shift assay (EMSA)

For EMSA, the homologous CCA1 promoter (CCA1p) containing the CHE binding site (CHE BS) were PCR amplified using cloned AtCCA1p and AaCCA1p fragments from the allotetraploids (see Supplementary Table S2). The resulting PCR products, AtCCA1p (133 bp) and AaCCA1p (134 bp), were then gel purified. Equal molar (3 pmol) of the purified PCR products were then end-labeled with T4 polynucleotide kinase in the presence of 32P-γATP (3000 Ci/mmol, 10 mCi/mL). End-labeled probes were then purified using Sephadex G-50 column with STE buffer (10 mM Tris HCl pH8, 100 mM NaCl and 1 mM EDTA). For EMSA, 10 nM (~15 fmol) DNA were used for binding reaction in buffer containing 25 mM HEPES pH7.5, 2.5 mM DTT, 75 mM KCl, 10% glycerol and 1.25 ng poly-dIdC. Various amount of the purified MBP, MBP-rAtCHE, and MBP-rAaCHE were then used for the binding reactions. The binding reactions were incubated on ice for 20 minutes and then revolved with 5% native polyacrylamide gel electrophoresis. After electrophoresis, the gel was dried on a 3 M filter paper and exposed using X-ray film. For competition assays, 1x, 5x, 25x, and 125x unlabeled cold probes were included in the binding reactions. Densitometric intensities of the shifted DNA probes were quantified using ImageJ software (National Institutes of Health).

Protein expression for GFP fragments reassembly

A pair of GFP reassembly constructs was co-transformed into Rossetta2 (DE3) competent cells. For a negative control, both pET11a-link-NGFP and pMRBAD-link-CGFP constructs were used. For a positive control, pET11a-Z-NGFP and pMRBAD-Z-CGFP expressing the leucine zipper reassembly constructs were used43. The interacting pairs of co-transformation were 1) NGFP:AtCHE and AtHD1:CGFP; 2) NGFP:AaCHE and AaHD1:CGFP; 3) NGFP:AtCHE and AaHD1:CGFP; and 4) NGFP:AaCHE and AtHD1:CGFP, respectively. For bacterial expression, 10 ng of each Miniprep purified DNA in 5 μL water were co-transformed into 50 μL Rossetta2 (DE3) competent cells. The double transformants were selected on LB agar plates containing 100 μg/mL ampicillin, 35 μg/mL kanamycin and 40 μg/mL chloramphenicol. A single colony was picked from the triple-selection agar plate for inoculation in 5 mL LB containing ampicillin, kanamycin and chloramphenicol cultured overnight at 37 °C with constant shaking. For recombinant protein induction and GFP fragment reassembly, fresh overnight culture was diluted to 1:1000 and 5 μL of the diluted culture were plated onto a reassembly plate containing 10 μM IPTG, 0.2% arabinose and the antibiotics, ampicillin, kanamycin and chloramphenicol. The plate was divided into sections based on number of samples and controls. Each pair of treatments was repeated from 5 independent cultures and reassembly plating. The plates were incubated at 30 °C overnight followed by additional 2 days at 20 °C.

GFP fluorescence microscopy

Bacteria containing the GFP reassembly constructs were resuspended in 0.5 mL buffer (25 mM Tris-HCl, pH8.0, 10% glycerol) and the OD600 of the resuspended cells were adjusted to 2. For fluorescence microscopy, 20 μL of each resuspended cells were transferred onto a Superfrost/Plus microscope slide (Fisher Scientific, Hampton, New Hampshire, USA) and covered with 24 × 40 mm micro cover glass (VWR, Radnor, Pennsylvania, USA). Edges of the cover glasses were sealed with nail polish. Live cell images were taken on a Zeiss Axiovert 200 M inverted microscope under brightfield transmission and FITC fluorescence filter using 20× objective. Three representative fields from each slide were taken.

GFP quantification by FACSCalibur

Fresh overnight bacterial cultures were diluted to 1:10000 and 100 μL of the diluted cultures were inoculated into 5 mL LB media supplemented with 100 μg/mL ampicillin, 35 μg/mL kanamycin and 40 μg/mL chloramphenical and incubated at 37 °C with constant shaking. When OD600 reached 0.5, IPTG (15 μM) and arabinose (0.2%) were added into each 5 mL culture for recombinant proteins induction. The cultures were incubated at 20 °C incubator for 2 days. The cells were centrifuged at 2000 g for 15 minutes at 4 °C and washed twice with 25 mM Tris-HCl, pH8.0, 10% glycerol buffer. The harvested cells were resuspended in 25 mMTris-HCl buffer with OD600 = 1. FACSCalibur samples were prepared by diluting each cell culture with 25 mM Tris-HCl (pH8.0) buffer, resulting in a concentration of ~ 2 × 106 cells/mL in a 5 mL PolyStyrene Tube w/Filter Cap (Fisher Scientific, Hampton, New Hampshire, USA). FACSCalibur flow cytometer (BD Biosciences, San Jose, CA, USA) analysis was used to evaluate relative levels of fluorescence intensity in bacterial cells co-expressing the GFP reassembly constructs pair (NGFP:CHE and HD1:CGFP). Samples were run under low pressure using the FL2 detector set at 480 nm. A total of ~10,000 events were counted for each sample, and three biological replicates were performed for each treatment. Flow-cytometry data were collected and analyzed using CellQuestPro software (BD Biosciences, San Jose, CA, USA). A scatter plot was generated from the collected data. Cells with positive GFP signals were then identified and gated. Relative proportion of GFP within these gated peak regions (FL2-A) were calculated and tested for statistical significance.

Yeast two-hybrid assays

For the bait plasmids construction, homologous CHE and TOC1 inserts were cut from the corresponding clones in the pGEM-T plasmids through EcoRI and SalI restriction sites and cloned into the pGBKT7-BD respectively. For the prey plasmids construction, homologous CHE were cloned through EcoRI and SacI restriction sites; and HD1 were cloned through XhoI and XmaI restriction sites, into the pGADT7-AD respectively. For yeast two-hybrid, a matchmaker gold yeast two-hybrid system and a yeastmaker yeast transformation system 2 (Clontech, Mountain View, CA, USA) were used. Yeast two-hybrid screens were conducted according to manufacturers’ instructions. For yeast transformation, yeast strain Y2H gold was used for bait plasmid transformation while the yeast strain Y187 was used for prey plasmid transformation. For qualitative assessment of interaction between the selected target proteins pair, transformed yeast colony (2–3 mm) containing the target baits or prey plasmids were mixed and incubated at 30 °C with shaking at 200 rpm for overnight. The successful mated yeast strains were selected on double dropout media without Leu and Trp (DDO). Mated diploid yeast strains were also selected and screen on double dropout media containing X-α-gal (DDO/X), Triple dropout media (without Leu, Trp and His) containing X-α-gal (TDO/X) or quadruple dropout media (without Leu, Trp, His and Ade) containing X-α-gal (QDO/X). The selected diploid hybrid yeast containing both the bait and prey plasmids in DDO liquid media and shook (350 rpm) overnight at 30 °C. Then, 0.6 × 107 cells (OD600 = 0.6) were harvested, resuspended in 3 mL DDO media and shook (250 rpm) at 29 °C for 16 hours. For quantitation of reporter expression (α-galactosidase) resulted from positive protein-protein interaction, a quantitative p-nitrophenyl-a-D-galactopyranoside (PNG) assay was used44. To perform the assay, the subcultured cells were pelleted at 13,000 rpm for 1 minute and 200 μL supernatant was transferred to a new tube containing 600 μL assay buffer (33 mM PNG, 0.33 M sodium acetate pH4.5). The reaction was incubated at 29 °C for 24 hours, stopped by adding 200 μL 2 M Na2CO3 and OD410 of the stopped reaction was measured. For quantification, the α-galactosidase activity was estimated as the change of OD410/minute/cell. For each interacting proteins partner, the relative α-galactosidase activity was measure against the positive P53-T positive control.

Accession numbers

Sequence data from this article can be found in the GenBank/EMBL databases under the following accession numbers: At2g46830 (AtCCA1), At5g08330 (AtCHE), At5g61380 (AtTOC1), At4g38130 (AtHD1), At1g07930 (EF1α).

Additional Information

How to cite this article: Ng, D. W-K. et al. Heterologous protein-DNA interactions lead to biased allelic expression of circadian clock genes in interspecific hybrids. Sci. Rep. 7, 45087; doi: 10.1038/srep45087 (2017).

Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.