Arabidopsis LSH10 transcription factor and OTLD1 histone deubiquitinase interact and transcriptionally regulate the same target genes


LSH10 is a nuclear protein that interacts with OTLD1

To gain better insight into possible mechanisms by which plant histone deubiquitinases reach their target chromatin, we searched for proteins that interact with OTLD1. A truncated OTLD1 was used as bait to screen the Arabidopsis yeast two-hybrid protein interaction library29, and two independent clones encoding the LSH10 protein were identified as a putative interaction partner of OTLD1 (Supplementary Fig. 1). We then examined the subcellular localization of LSH10, which was tagged with CFP and transiently expressed in N. benthamiana leaf tissues together with a free YFP reporter that partitions between the cell cytoplasm and the nucleus, conveniently visualizing and identifying both subcellular compartments. Figure 1a shows that LSH10-CFP accumulated in the cell nucleus of the cells whereas, as expected, the free YFP fluorescence was nucleocytoplasmic. This nuclear localization of LSH10 is consistent with its interaction with OTLD1, a histone deubiquitinase that functions in the cell nucleus27.

Fig. 1: LSH10 subcellular localization, and specific interaction of LSH10 with OTLD1 in planta detected by BiFC.
figure 1

a Nuclear localization of LSH10. The LSH10 protein tagged with CFP (LSH10-CFP) and free YFP were transiently co-expressed in N. benthamiana leaves. b The BiFC assay of the interaction between LSH10 and OTLD1. The indicated combinations of proteins tagged with nYFP and cYFP were transiently co-expressed in N. benthamiana leaves. CFP signal is in cyan; YFP signal is in yellow; merged CFP and YFP signals are in cyan-yellow, and chlorophyll autofluorescence is in red. Scale bars = 20 µm. All images are single confocal sections.

Next, the physical interaction of LSH10 with OTLD1 was studied in planta using two independent approaches, bimolecular fluorescence complementation (BiFC) and fluorescence resonance energy transfer (FRET). BiFC experiments shown in Fig. 1b detected a strong fluorescent signal of the reconstituted YFP in the cells co-expressing both LSH10 and OTLD1, indicating protein interaction. This interaction was specific as no BiFC signal was observed in the cells co-expressing either OTLD1 and LSH4, a homolog of LSH10, or LSH10 and an unrelated plant viral protein MT (Fig. 1b). Furthermore, the interacting proteins colocalized with the nuclear portion of the coexpressed free YFP reporter, indicating that the OTLD1-LSH10 complexes were located in the cell nucleus, the expected subcellular site of their function (Fig. 1b). Our FRET experiments—using LSH10 tagged with GFP as donor fluorophore and OTLD1 tagged with RFP as acceptor fluorophore—confirmed and extended the BiFC findings. We used two variations of the FRET method, sensitized emission (SE-FRET) and acceptor bleaching (AB-FRET)30. In SE-FRET, protein interaction results in the transfer of the excited state energy from the GFP donor to the RFP acceptor without emitting a photon, producing the fluorescent signal with an emission spectrum similar to that of the acceptor. AB-FRET, on the other hand, detects and quantifies protein interaction from increased emission of the GFP donor when the RFP acceptor is irreversibly inactivated by photobleaching. Figure 2a summarizes the results of the SE-FRET experiments, in which the cell nuclei were simultaneously recorded in all three, i.e., donor GFP, acceptor RFP, and SE-FRET, channels and used to generate images of SE-FRET efficiency illustrated in a rainbow pseudo-color. This color scale, i.e., transition from blue to red, indicates an increase in FRET efficiency from 0 to 100%, which corresponds to the degree of protein-protein proximity during the interaction. The SE-FRET signal observed in the cell nuclei following the coexpression of LSH10 and OTLD1 was comparable to that generated in positive control experiments which expressed the translational acceptor-donor RFP-GFP fusion. Negative controls, i.e., coexpression of OTLD1-RFP with LSH4-GFP or free RFP with LSH10-GFP, produced no SE-FRET signal (Fig. 2a). The FRET data were quantified using AB-FRET (Fig. 2b, c) by recording the cell nuclei in the donor GFP channel before and after RFP photobleaching and displayed in pseudo-color to visualize the change in GFP fluorescence. Figure 2b shows that photobleaching of the RFP acceptor completely blocked its fluorescence in all protein coexpression combinations tested. Following this photobleaching, two protein combinations showed an increase in the GFP donor signal, i.e., LSH10-GFP coexpressed with OTLD1-RFP and the RFP-GFP fusion positive control. In contrast, the negative controls, i.e., LSH4-GFP coexpressed with OTLD1-RFP and LSH10-GFP coexpressed with free RFP, elicited no increase in the GFP fluorescence (Fig. 2b). Quantification of these data demonstrated that the increase in the donor fluorescence (%AB-FRET) of 13% observed following LSH10-OTLD1 coexpression was statistically significant and overall comparable to the maximal %AB-FRET of 30% achieved with RFP-GFP. Both negative controls displayed no increase in donor fluorescence (Fig. 2c). Collectively, the data in Fig. 2 indicate that LSH10 interacts with OTLD1 within living plant cells, that the interacting proteins accumulate in the cell nucleus, and that, in the LSH10-OTLD1 complex, the proteins are within <10 nm from each other, the effective range of protein interactions detected by FRET31.

Fig. 2: Specific interaction of LSH10 with OTLD1 in planta detected by FRET.
figure 2

The indicated combinations of proteins tagged with GFP (energy donor) and RFP (energy acceptor) were transiently co-expressed in N. benthamiana leaves. a SE-FRET. Images were collected from the three detection channels, i.e., donor (GFP), acceptor (RFP), and raw SE-FRET. The SE-FRET efficiency images represent the calculated corrected image after subtraction of spectral bleed-through and are presented in a rainbow pseudo-color scale, in which red denotes the highest SE-FRET signal and blue denotes the lowest signal. Scale bars = 4 µm. b AB-FRET. Images were collected from the two detection channels, i.e., donor (GFP) and acceptor (RFP), before and after photobleaching. The circle delineates a region that was photobleached and its fluorescence measured. The AB-FRET is presented in a rainbow pseudo-color scale, in which red denotes the highest GFP signal and blue denotes the lowest signal. Scale bars = 4 µm. c Quantification of AB-FRET. Data measured in the experiments described in panel b were used to calculate %AB-FRET. Error bars represent SEM of n = 13 independent cells for each measurement. The individual data points are indicated, and their numerical values are listed in Supplementary Data 2. Differences between mean values assessed by the two-tailed t-test are statistically significant for the p-values *p < 0.05, **p < 0.01, and ***p < 0.001; P ≥ 0.05 are not statistically significant (ns).

For the LSH10-OTLD1 interaction to be biologically meaningful, both genes should be expressed at least in some of the same plant tissues at the same time. Thus, we examined the expression pattern of the endogenous LSH10 and OTLD1 genes in different organs systematically throughout the plant, i.e., rosette and cauline leaves, stems, flowers, and roots of wild-type plants using reverse transcription-quantitative RT-PCR (RT-qPCR) analysis. Figure 3 shows that both genes were expressed in all tested tissues, suggesting the availability of their protein products for functional interaction during transcriptional regulation of their target genes. Obviously, the expression levels in different tissues varied. For example, LSH10 was expressed most prominently in roots, less in stems and rosette leaves, and at the lowest relative levels in flowers and cauline leaves (Fig. 3). As in earlier observations19, OTLD1 was expressed at the highest level in flowers and lower levels in all other tested tissues (Fig. 3). Overall, regardless of the exact degree of expression, LSH10 and OTLD1 transcripts were detected all tested tissues, suggesting that their protein products are available for functional interaction during transcriptional regulation of their target genes and that the interplay between the relative expression levels of LSH10 and OTLD1 may contribute to the overall regulation pattern of these target genes.

Fig. 3: Coexpression of the endogenous LSH10 and OTLD1 genes in different organs of the wild-type Arabidopsis plants.
figure 3

The indicated tissues were harvested from 36 days-old plants. Expression of LSH10 and OTLD1 was analyzed by RT-qPCR with primers listed in Supplementary Data 1. Gene expression is shown on a log scale. The individual data points are indicated, and their numerical values are listed in Supplementary Data 3. Error bars represent SEM of n = 7 or n = 8 independent biological replicates.

LSH10 has the structural features of a transcription factor

LSH10 is a 177-amino acid residue protein (Supplementary Fig. 2) encoded by the Arabidopsis At2G42610 gene. It belongs to a 10-member family of ALOG (Arabidopsis LSH1 and Oryza G1) protein family in Arabidopsis (Supplementary Fig. 3a) as well as in numerous other dicotyledonous plant species (Supplementary Fig. 3b). The members of this family, many of which remain uncharacterized, carry a highly conserved ALOG domain(also known as DOMAIN OF UNKNOWN FUNCTION 640 / DUF640) located in the center of the protein molecule. This domain is composed of 4 all-α helices, a zinc ribbon insert structure, and a nuclear localization signal (NLS) (Supplementary Fig. 2). ALOG is predicted to act as a DNA binding domain and belongs to the tyrosine recombinase/phage integrase N-terminal DBD superfamily32, in which the ALOG domain members, unlike the tyrosine recombinase members, contain a conserved zinc ribbon insert located between helices 2 and 3 with highly conserved positively charged residues at its N-terminus and the “HxxxC” and “CxC” motifs. This region can provide additional molecular contacts unique to the ALOG domain to participate in binding to DNA. The conserved ALOG sequences (Supplementary Fig. 3) and our prediction of the DNA-binding amino acid residues using three methods, DRNApred33, DP-Bind34, and DISPLAR35, indicate that LSH10 may associate with DNA via hydrogen bonding and ionic interactions with multiple conserved solvent-accessible basic residues36 in helix-1, helix-3, helix-4, and in the zinc ribbon or with a conserved acidic residue36 in helix-1, whereas the conserved hydrophobic residues in all four helices likely stabilize the core tetra-helical fold37 in the target DNA molecule (Supplementary Fig. 4). Thus, the sequence analysis of LSH10 indicates that this protein binds DNA, consistent with its proposed activity as a transcription factor.

LSH10 is a transcriptional repressor of the OTLD1 target genes

The interaction of LSH10 with OTLD1 and its potential DNA binding ability suggest that LSH10 may function as a transcription factor that directs the OTLD1 co-repressor to its target genes. In this scenario, LSH10 should function in complex with OTLD1 and, thus, repress at least a subset of the target genes repressed by OTLD1. Our previous study indicated that OTLD1 is involved in the transcriptional repression of five genes OSR2, WUS, ABI5, ARL, and GA20OX via deubiquitylation of H2B in their promoter chromatin19. Thus, we examined the effect of LSH10 loss-of-function mutations on the transcription of these genes. To this end, two Arabidopsis lsh10 T-DNA insertion lines (SALK_006965 and SK14678) were obtained from ABRC (www.arabidopsis.org/abrc/) and the homozygous mutant lines, designated lsh10-1 and lsh10-2, were generated. The lsh10-1 and lsh10-2 mutants contained a single T-DNA insertion in the exon and 5’UTR of the LSH10 gene, respectively (Fig. 4a). The RT-qPCR analysis showed that, in both lsh10-1 and lsh10-2 plants, the transcription of the LSH10 gene was virtually abolished (Fig. 4b), confirming the loss of function of this gene in both mutant lines.

Fig. 4: LSH10 is a transcriptional repressor of the OTLD1 target genes OSR2, WUS, ABI5, and ARL.
figure 4

a A schematic structure of the LSH10 gene and its loss-of-function alleles lsh10-1 and lsh10-2. The locations of the mutagenic T-DNA inserts in each of the alleles are indicated by arrowheads. b Loss of the LSH10 expression in the lsh10-1 and lsh10-2 plants. Wild-type plants, black bars; lsh10-1, gray bars; lsh10-2, white bars. c Increase in expression of the indicated target genes in the lsh10-1, and lsh10-2 mutants. Wild-type plants, black bars; lsh10-1, gray bars; lsh10-2, white bars. d Transcriptional repression of the indicated target genes in genetically complemented lsh10-1/LSH10-His6 plants. Wild-type plants, gray bars; lsh10-1/LSH10-His6, white bars. The increase in gene expression and transcriptional repression were analyzed by RT-qPCR with primers listed in Supplementary Data 1. Error bars represent SEM of n = 8 biological replicates. The individual data points are indicated, and their numerical values are listed in Supplementary Data 4 and Supplementary Data 5. Differences between mean values assessed by the two-tailed t-test are statistically significant for the p-values *p < 0.05, **p < 0.01, and ***p < 0.001; P ≥ 0.05 are not statistically significant (ns).

Next, the amounts of transcripts of each of the OSR2, WUS, ABI5, ARL, and GA20OX genes were analyzed by RT-qPCR in the lsh0-1 and lsh10-2 plants and compared to the wild-type plants. Each of the OSR2, WUS, ABI5, and ARL genes displayed a substantial and statistically significant increase in expression in both loss-of-function lines (Fig. 4c). Specifically, the amounts of the OSR2, WUS, ABI5, and ARL transcripts were elevated ca. 24.42 to 25.43-fold, 26.99 to 24.05-fold, 20.42 to 8.45-fold and 48.23 to 66.22-fold in lsh10-1 and lsh10-2, respectively. The expression of the internal reference gene EF1a was not significantly altered in any of the plant lines (Fig. 4c).

Besides testing two different alleles of the lsh10 loss-of-function mutant, we confirmed that derepression of the OTLD1 target genes resulted from the decrease in lsh10 transcription by genetic complementation of one of the alleles, lsh10-1, with the wild-type LSH10 coding sequence. We generated a transgenic lsh10-1 line, lsh10-1/LSH10-His6, that expresses wild-type LSH10 protein tagged with hexahistidine. The resulting lsh10-1/LSH10-His6 plants expressed the tagged LSH10 at higher levels than the parental lsh10-1 plants (compare Fig. 4d to Fig. 4b); these levels were comparable and even slightly, ca. 1.5-fold, higher than the levels of the endogenous LSH10 transcript in the wild-type plants (Fig. 4d). In these genetically complemented plants, we observed clear repression of all four target genes, i.e., OSR2, WUS, ABI5, and ARL, relative to the loss-of-function lsh10-1 parental plants (compare Fig. 4d to Fig. 4c) whereas no such repression was detected with the negative control EF1a gene (Fig. 4d). This analysis confirms that LSH10-His6 can functionally complement the lsh10-1 mutation. Collectively, the data in Fig. 4 indicate that LSH10 acts as a transcriptional repressor of most of the known OTLD1 target genes.

LSH10 binds to the promoter DNA sequences and associates with the chromatin of the OTLD1/LSH10 target genes to deubiquitylate H2B

The proposed function of LSH10 as a transcriptional repressor that facilitates the function of the OTLD1 co-repressor at the target chromatin implies that LSH10 binds directly to the regulatory sequences of the gene regulated both by OTLD1 and LSH10. Thus, we examined whether LSH10 can bind the promoters of the OSR2, WUS, ABI5, and ARL genes directly, using the electrophoretic mobility shift assay (EMSA). We selected 2–4 conserved motifs of the intergenic regions of each of these genes (Fig. 5a) and used them as EMSA probes for interaction with a purified recombinant LSH10 tagged with GST (glutathione-S-transferase). Figure 5b shows that each of these probes was recognized by LSH10 as detected by substantially reduced electrophoretic mobility of the GST-LSH10-probe complexes as compared to the free probe (lanes 3, 7, 11, 15, 19, 23, 27, 31, 35, 39 and lanes 1, 5, 9, 13, 17, 21, 25, 29, 33, 37 respectively). This binding was specific because it was not observed with GST alone (Fig. 5b, lanes 2, 6, 10, 14, 18, 22, 26, 30, 34, 38) and substantially reduced in the presence of competing amounts of unlabeled DNA, corresponding to each probe (Fig. 5b, lanes 4, 8, 12, 16, 20, 24, 28, 32, 36). Consistent with this binding specificity, not all selected motifs were recognized by LSH10 (e.g., Fig. 5b, lanes 40, 41, 42). Taken together, the EMSA experiments lend support to the idea that LSH10 functions as a DNA-binding protein that recognizes sequence elements within the target gene promoters. That several diverse promoters are recognized by LSH10 suggests a fuzzy type of recognition that allows a single transcription factor to bind variable consensus DNA sequences38.

Fig. 5: Binding of LSH10 to the promoter DNA sequences of the OSR2, WUS, ABI5, and ARL target genes.
figure 5

a A schematic of the OSR2, WUS, ABI5, and ARL gene promoters showing the locations of probe sequences for EMSA (identified by black rectangles and letters “a”-“i” and detailed in Supplementary Data 1) and probes for qCHIP (denoted by double-headed gray or white arrows for probes overlapping or outside of the EMSA probe locations, respectively, and detailed in Supplementary Data 1) relative to the translation initiation sites (bent arrows). b EMSA of LSH10 binding to promoter sequences of the indicated target genes. The composition of the binding mixtures is detailed above each lane of the gel image, with (+) or (−) signifying the presence or absence, respectively, of the indicated components. The electrophoretic mobilities of the LSH10-bound and free probes are indicated by arrows.

Next, we investigated the potential association of LSH10 with its target gene chromatin within plant cells, taking advantage of the fact that, in the lsh10-1/LSH10-His6 line, LSH10 is tagged with an His6 epitope, allowing us to utilize quantitative chromatin immunoprecipitation (qChIP) to detect its presence. To correlate the physical association of LSH10 with the target chromatin and the binding of LSH10 to the target promoter sequence, the qChIP primers were designed to overlap several EMSA probes (Fig. 5a). Figure 6a shows that, indeed, LSH10 was associated with the regions of the OSR2, WUS, ABI5, and ARL chromatin that contained the DNA sequences to which LSH10 was able to bind (Fig. 5). The amounts of immunoprecipitated LSH10-His6, relatively to the background signal obtained with the wild-type plants that do not express the His6 tag, were statistically significant yet varied between different target genes, potentially reflecting the different amounts of the LSH10/OTLD1-containing repressor complexes involved in the repression of each of these genes. For negative control experiments, we decided to use the ABI5 gene, but with qChIP primers outside and upstream of the EMSA probe locations (Fig. 5a), expecting a reduced or no LSH10 association with this chromatin region. Indeed, only background-level qChIP signal was observed with these primers (Fig. 6a). Overall, these data demonstrate the ability of LSH10 to bind the conserved DNA motifs in the target gene promoters and to associate with the promoter chromatin of these genes.

Fig. 6: Association of LSH10 with and effects on ubiquitylation of the chromatin of the OSR2, WUS, ABI5, and ARL genes.
figure 6

a Association of LSH10-His6 with the chromatin of the indicated target genes in the lsh10-1/LSH10-His6 plants. lsh10-1/LSH10-His6, white bars; non-specific background signal obtained in the wild-type plants that do not express the His6 epitope, gray bars. The outside probe is the qChIP probe, depicted in Fig. 5a, and located outside the EMSA probe locations. b Increase in H2B monoubiquitylation of the OSR2, WUS, ABI5, and ARL promoter chromatin in the lsh10-1 mutant plants. Wild-type plants, dark gray bars; lsh10-1, light gray bars; lsh10-2, white bars. The chromatin association of LSH10-His6 and the degree of H2B monoubiquitylation were analyzed by qChIP with probes described in Fig. 5a and listed in Supplementary Data 1. Error bars represent the SEM of n = 5 or n = 7 biological replicates with 3 technical repeats for each. The individual data points are indicated, and their numerical values are listed in Supplementary Data 6 and Supplementary Data 7. Differences between mean values assessed by the two-tailed t-test are statistically significant for the p-values *p < 0.05 and **p < 0.01; P ≥ 0.05 are not statistically significant (ns).

Finally, we examined the notion that, if LSH10 cooperates with OTLD1 to deubiquitylate the target chromatin, such deubiquitylation will be reduced in the absence of LSH10. Thus, we used qChIP to analyze the promoter chromatin of the OSR2, WUS, ABI5, and ARL genes in the loss-of-function lsh10-1 and lsh10-2 mutants in and in the wild-type plants for the presence of monoubiquitylated H2B, known to be deubiquitylated by OTLD1 in these chromatin regions27. Figure 6b shows a statistically significant degree of hyperubiquitylation of H2B in the OSR2, WUS, ABI5, and ARL chromatin in both mutants as compared to the wild-type plants. Specifically, the OSR2 chromatin of lsh10-1 and lsh10-2 plants was monoubiquitylated on average 4.24-2.46-fold more than the wild-type OSR2 chromatin, and monoubiquitylation of WUS, ABI5, and ARL was increased 2.62-1.86-fold, 4.59-4.25-fold, and 3.02-2.71-fold, respectively (Fig. 6b).



Source link

Leave a Comment