Supplemental information, Figure S1

Figure S1 DNA methylation inhibitor 5’AZA treatment released the transcriptional

silencing of SUC2 and NPTII transgenes. Related to Figures 1 and 2. (A) DNA

methylation inhibitor 5’AZA treatment released the transcriptional silencing of

transgene NPTII. Seeds were germinated on 1% glucose and 50 ng/L kanamycin

(Kan)-containing 1/2MS media supplemented with (lower panel)/without (upper panel)

DNA methylation inhibitor 5’AZA. Photographs were taken at 12-day post

germination. Col-0 without NPTII transgene served as negative control. In

comparison to 35S::SUC2 transgene control, hdp1-1, hdp1-2 and hdp2-1 mutants

displayed a kanamycin-sensitive phenotype with yellow cotyledons, as well as Col-0.

5’AZA treatments released the silencing of NPTII transgene and all mutants displayed

kanamycin-insensitive phenotype as well as 35S::SUC2 plants, except for Col-0. (B)

Upper panel: 5’AZA treatments inhibited DNA hypermethylation caused by hdp1 and

hdp2 mutations. DNA methylation-sensitive PCR (chop-PCR) was performed to test

the DNA methylation levels in 35S::SUC2 promoter region with/without 5’AZA

treatment. Lower panel: RT-qPCR results showing the transcript levels of the

transgenes SUC2 and NPTII in 5’AZA and DMSO treatment. Relative expressions

were normalized to 35S::SUC2 in each treatment. RT-qPCR results are means ± SD of

three biological replicates. (C) ChIP assay showing altered transcriptional activity in

hdp1 and hdp2 mutants transgene promoters with (lower panel)/without (upper panel)

5’AZA treatment. The occupancies of Pol II and H3K4me3 at the 35S::SUC2

transgene promoter region are reduced in hdp1-1 and hdp2-1 mutants, and the

reduction is inhibited by 5’AZA treatment. As for repressive H3K9me2 mark, it is

increased in hdp1-1 and hdp2-1 mutants and 5’AZA treatment counteracts the

increase. The enrichments were quantified by ChIP-qPCR and normalized by

comparison to input in 35S::SUC2 plants. The positions of primer pairs used for

ChIP-qPCR were labeled in 35S::SUC2 transgene diagram (Figure 2A).

Supplemental information, Figure S2

Figure S2 Pearson’s coefficient plots for RNA-Seq of 35S::SUC2, hdp1-1 and hdp2-1.

Related to Figure 2. The asterisks indicate the significance levels of the correlations.

***P-value < 0.001. The labels of x axis and y axis are log2 transformation of the

FPKM plus one.

Supplemental information, Figure S3

Figure S3 The phylogenetic analysis of HDP2. Related to Figure 3. Black sequences

are HDP2 homologous sequences from BLAST search with NCBI nr-database.

Yellow sequences are Harbinger sequences that have a myb-like DNA binding

domain. Sequences from Arabidopsis are in red. The orthologous clade of HDP2 is

labeled in green. A clade of host proteins that shares a common branching point with

transposon sequences is labeled in purple. Two clades of transposon sequences that

are intermingled with host clades are labeled in dark green and blue.

Supplemental information, Figure S4

Figure S4 DNA-binding activity of HDP2 and its role in anti-silencing and DNA

demethylation. Related to Figure 4. (A) EMSA assay showing the DNA-binding

activity of HDP2. MBP-fused HDP2 recombinant protein was incubated with four

double-stranded DNA probes from the TIRs sequence of zebrafish Harbinger3N_DR

transposon [11]. Mobility shifted bands were observed when MBP-HDP2 was

incubated with L1 and R2 probes, suggesting that HDP2 bears DNA-binding activity.

MBP served as a negative control. Bound and free probes are labeled with black

arrows. See also Supplementary information, Table S3 for the sequence of DNA

probes. (B) HDP2 showing binding activity to both unmethylated and methylated

DNA probes. The 35S-1 region (See also Figure 2A) from 35S::SUC2 promoter

region was selected to synthesize methylated and unmethylated double-stranded DNA

probes via PCR amplification. Equal amounts of methylated and unmethylated DNA

probes were incubated with increasing amounts of MBP-HDP2 proteins. MBP-HDP2

displayed similar DNA-binding patterns to these two probes. “35S-1” and “m35S-1”

represent unmethylated and methylated DNA probes, respectively. (C) HDP2

DNA-binding mutants cannot rescue the DNA hypermethylation and reduced

transcriptional activity of 35S::SUC2 promoter caused by hdp2-1 mutation. Upper

panel: chop-PCR showing HDP2 W19R and W48R DNA-binding mutants cannot

rescue the hypermethylation of hdp2-1. Chop-PCR was performed in wild-type HDP2

and HDP2 DNA-binding mutant transgenic plants to test the effects of HDP2

DNA-binding activity on DNA demethylation. Lower panel: ChIP-qPCR showing the

effects of HDP2 DNA-binding mutants on the accumulations of histone H3K4me3

and H3K9me2 modifications at 35S::SUC2 transgene promoter region. (D) ChIP

assay showing HDP2 W19R and W48R mutations abolished its enrichment at

transgene promoter regions. Wild-type and DNA-binding mutant HDP2 transgenic

plants were subjected to ChIP assay using anti-Flag antibody to test the effects of

HDP2 DNA-binding activity on its enrichment at transgene promoter regions. No

antibody was used as negative control.

Supplemental information, Figure S5

Figure S5 The domain requirement of HDP1 and HDP2 interaction. Related to

Figure 5. (A) The N terminus of HDP1 directly interacts with the C-terminus of

HDP2 in the yeast two-hybrid assay. The full-length and truncated forms of HDP1

and HDP2 were cloned into AD and BD vectors. Yeast cells harboring different

combinations were grown on selection SD(-L/W) and SD(-L/W/H) media. (B)

Bimolecular fluorescence complementation (BiFC) assay showing that HDP1 and

HDP2 interact in the nucleus. Full-length HDP1 and HDP2 were fused with split YFP.

BiFC assay was performed in tobacco (N. benthamiana) leaves. YFP fluorescence

was examined at 2-day post infiltration. BF, bright field.

Supplemental information, Figure S6

Figure S6 The N-terminal domain of IDM1 is responsible for its interaction with

HDP1 but not HDP2. Related to Figure 6. (A) The domain requirement for IDM1

interacting with HDP1 in Y2H assay. The N-terminal 1-592 amino acids, the

C-terminal 593-1189 and 859-1189 amino acids of IDM1 were fused with BD. The

full-length HDP1, HDP2 and IDM2 were fused with AD. Yeast cells transformed with

N-terminal IDM1 and HDP1 or IDM2 can grow on SD-L/T/H/3AT medium. Empty

AD and BD vectors served as negative control. (B) Full-length HDP1 is required for

its interaction with IDM1. The domain requirement for HDP1 interacting with IDM1

was tested in Y2H assay. (C) HDP1 and HDP2 do not interact with IDM2, IDM3 and

MBD7 in Y2H assay. No combinations between HDP1, HDP2 and IDM2, IDM3,

MBD7 can grow on SD-L/T/H and SD-L/T/H/3AT media. (D) HDP1 and HDP2 do

not interact with IDM2, IDM3 and MBD7 in split luciferase assay. The split luciferase

assay was performed in tobacco leaves to verify the interaction between HDP1, HDP2

and IDM2, IDM3 and MBD7. The left panel indicates the position of different

combination on tobacco leaf. Luciferase activity was examined at 2 days after

infiltration. X represents IDM2, IDM3 and MBD7 in different combinations,

respectively.

Supplemental information, Figure S7

Figure S7 A working model for the derivation of HDP1 and HDP2 and their role in

the IDM complex. Related to Figure 7. (A) Venn diagram showing the overlap of

hyper-DMRs identified from hdp1, hdp2, mbd7 and idm1 mutants. (B) HDP1 and

HDP2 are host proteins co-domesticated from an ancestor Harbinger transposon.

Analogous to their Harbinger transposon protein ancestor, HDP1 and HDP2

physically interact in the nucleus to form a small sub-complex. HDP1-HDP2 binds to

genomic target regions via the Myb-like DNA binding domain of HDP2, and serves as

a molecular platform to recruit histone acetyltransferase IDM1 via HDP1-IDM1

interaction. Another DNA-binding protein, MBD7, binds to genomic regions with

highly methylated CpG sites and interacts with IDM2 and IDM3 to form another

small sub-complex. This sub-complex can also recruit IDM1 through IDM1-IDM2

and IDM1-IDM3 interactions. The specificity of IDM1 targeting is determined by

both HDP2 and MBD7. These six proteins form a big complex (i.e., the IDM complex)

to catalyze histone acetylation. The acetylated histone marks serve as a favorable

chromatin environment for recruiting ROS1 for active DNA demethylation. Black

filled and empty cycles indicate methylated and unmethylated cytosines, respectively.

Red and empty diamonds indicate acetylated and unacetylated histone marks.

Table S5 Fisher’s exact test on weather hper-DMRs in hdp mutant are enriched inregulatory regions (promoters).

at 1-kb promoter NOT 1-kb promoter P-value of Fisher's Exact Testhdp1_DMR 355 888hdp1_sim1 321 922 0.06843hdp1_sim2 334 909 0.1851

hdp2_DMR 870 1898hdp2_sim1 719 2049 4.12E-06 *hdp2_sim2 713 2055 1.72E-06 *

at 1-kb promoter NOT 1-kb promoter P-value of Fisher's Exact Testhdp1_DMR 543 700hdp1_sim1 259 984 < 2.2e-16 *hdp1_sim2 291 952 < 2.2e-16 *

hdp2_DMR 1368 1400hdp2_sim1 608 2160 < 2.2e-16 *hdp2_sim2 615 2153 < 2.2e-16 *

Note: stars indicate statistical significance with 0.01 P-value cut-off

Gene

TE

Supplementary information, Data S1

Materials and Methods

Mutant screen and map-based cloning

The hdp1-1, hdp1-2 and hdp2-1 mutants were obtained by EMS screen as

described in our previous report [1, 2]. To clone the HDP1 and HDP2 genes, mutants

were crossed with Landsberg erecta and hybrid F1 plants were self-pollinated to

obtain the F2 population. Seedlings with long-root phenotype on 1%

sucrose-containing MS media were selected from F2 plants for calculation of mutant

linkage. Genomic DNA from mutants was re-sequenced to determine the location of

the mutation in the mapping region.

RNA-seq analysis

For RNA-seq analysis, total RNAs were extracted from 2-week-old seedlings and

submitted to RNA sequencing. Clean reads were mapped to Arabidopsis reference

genome using TopHat. FPKM were calculated using cufflinks [3]. Pearson’s coefficient

plot was generated using chart. Correlation() function from the PerformanceAnalytics

package. For the plot, we require: 1) lengths of genes should be longer than 500 bp; 2)

FPKM of at least one sample should be higher than [3].

BiFC and subcellular localization analysis

For the BiFC assay, full-length HDP1 and HDP2 sequences were cloned into

pSITE-nEYFP-C1 and pSITE-cEYFP-C1 vectors [4] to generate HDP1 and HDP2

fused split YFP constructs. Agrobacteria bearing split YFP constructs were infiltrated

into N. benthamiana leaves. YFP fluorescence was examined at 2 day-post-infiltration.

For subcellular localization of HDP1 and HDP2 proteins, full-length coding

sequences were cloned into pEarleyGate vectors to generate HDP1-YFP, HDP1-CFP

and HDP2-YFP constructs. NbFIB2-RFP was published by Duan et al. [5].

Agrobacteria bearing these constructs were infiltrated into N. benthamiana leaves or

transfected into Arabidopsis protoplasts. Fluorescence was examined at 2

day-post-infiltration in tobacco leaves or after 24 h incubation in protoplasts. For

HDP1 and HDP2 subnuclear co-localization assay, nuclei were extracted from

seedlings of 3FH-HDP2/HDP1-4myc hybrid plants and immunostaining was

performed as reported previously [6] using anti-Flag (Sigma-Aldrich, F1804) and

anti-myc (Sigma-Aldrich, C3956) antibodies.

EMSA assay

EMSA reaction was performed as described previously [7]. Full-length wild-type

and mutated HDP2 coding sequences were cloned into pMAL C2X for expression of

MBP fusion proteins in BL21 E. coli and purification as described previously [2].

DNA oligonucleotides were annealed first to generate double-stranded DNA. Then

double-stranded DNA was labeled by [ᵞ-32]P-ATP using T4 polynucleotide kinase

(NEB) to generate probes for EMSA reaction.

ChIP assay and ChIP-seq analysis

ChIP assay was performed as described [8] using 2-week-old seedlings.

Dynabeads (Invitrogen, 10004D) were used for antibody binding. The antibodies used

include, anti-pol II (Abcam, ab817), anti-H3K9me2 (Abcam, ab1220), anti-H3K4me3

(Abcam, ab8580), anti-FLAG (Sigma-Aldrich, F1804) and anti-H3K18AC (Abcam,

ab1191).

For ChIP-seq analysis, Arabidopsis genome was divided into 2 kb bins. ChIP

signal were calculated as follows [9, 10]:

Enrichment = log2( 8 + n1 ) – log2( 8 + n2 * N1/N2 )

We used N1 and N2 to represent the number of aligned ChIP and input reads in

each windows. N1 and N2 were the total reads in ChIP and input data, respectively.

To overcome sampling noise, eight pseudo counts were added. The DMCs were

defined as cytosines whose methylation level in mutant is 10% higher than that in WT

plants.

Real-time qRT-PCR

For real-time qRT-PCR, total RNAs were extracted from 2-week-old seedlings

using the RNeasy Plant Minikit (QIAGEN). After TURBO DNase I treatment

(Ambion), 2 μg of RNA was subjected to reverse transcription reaction using the

SuperScript III First-Strand Kit according to the manufacturer’s instructions

(Invitrogen). The 1st-strand cDNAs were then amplified using IQ SYBR green

supermix (BIO-RAD) with the CFX96 real-time PCR detection system (BIO-RAD).

Supplementary references

1 Lei M, La H, Lu K et al. Arabidopsis EDM2 promotes IBM1 distal

polyadenylation and regulates genome DNA methylation patterns. Proceedings of

the National Academy of Sciences of the United States of America 2014;

111:527-532.

2 Wang X, Duan CG, Tang K et al. RNA-binding protein regulates plant DNA

methylation by controlling mRNA processing at the intronic

heterochromatin-containing gene IBM1. Proceedings of the National Academy of

Sciences of the United States of America 2013; 110:15467-15472.

3 Trapnell C, Roberts A, Goff L et al. Differential gene and transcript expression

analysis of RNA-seq experiments with TopHat and Cufflinks. Nature protocols

2012; 7:562-578.

4 Martin K, Kopperud K, Chakrabarty R, Banerjee R, Brooks R, Goodin MM.

Transient expression in Nicotiana benthamiana fluorescent marker lines provides

enhanced definition of protein localization, movement and interactions in planta.

The Plant journal: for cell and molecular biology 2009; 59:150-162.

5 Duan CG, Fang YY, Zhou BJ et al. Suppression of Arabidopsis

ARGONAUTE1-mediated slicing, transgene-induced RNA silencing, and DNA

methylation by distinct domains of the Cucumber mosaic virus 2b protein. The

Plant cell 2012; 24:259-274.

6 Duan CG, Zhang H, Tang K et al. Specific but interdependent functions for

Arabidopsis AGO4 and AGO6 in RNA-directed DNA methylation. EMBO J 2015;

34:581-592.

7 Ivics Z, Hackett PB, Plasterk RH, Izsvak Z. Molecular reconstruction of Sleeping

Beauty, a Tc1-like transposon from fish, and its transposition in human cells. Cell

1997; 91:501-510.

8 Saleh A, Alvarez-Venegas R, Avramova Z. An efficient chromatin

immunoprecipitation (ChIP) protocol for studying histone modifications in

Arabidopsis plants. Nature protocols 2008; 3:1018-1025.

9 Lang Z, Lei M, Wang X et al. The methyl-CpG-binding protein MBD7 facilitates

active DNA demethylation to limit DNA hyper-methylation and transcriptional

gene silencing. Mol Cell 2015; 57:971-983.

10 Baubec T, Ivanek R, Lienert F, Schubeler D. Methylation-dependent and

-independent genomic targeting principles of the MBD protein family. Cell 2013;

153:480-492.

11 Sinzelle L, Kapitonov VV, Grzela DP et al. Transposition of a reconstructed

Harbinger element in human cells and functional homology with two

transposon-derived cellular genes. Proceedings of the National Academy of

Sciences of the United States of America 2008; 105:4715-4720.