Plant Mol Biol (2014) 86:367–380 DOI 10.1007/s11103-014-0234-5
A novel method to identify the DNA motifs recognized by a defined transcription factor Xiaoyu Ji · Liuqiang Wang · Xianguang Nie · Lin He · Dandan Zang · Yujia Liu · Bing Zhang · Yucheng Wang
Received: 21 June 2014 / Accepted: 29 July 2014 / Published online: 10 August 2014 © Springer Science+Business Media Dordrecht 2014
Abstract The interaction between a protein and DNA is involved in almost all cellular functions, and is vitally important in cellular processes. Two complementary approaches are used to detect the interactions between a transcription factor (TF) and DNA, i.e. the TF-centered or protein–DNA approach, and the gene-centered or DNA– protein approach. The yeast one-hybrid (Y1H) is a powerful and widely used system to identify DNA–protein interactions. However, a powerful method to study protein– DNA interactions like Y1H is lacking. Here, we developed a protein–DNA method based on the Y1H system to identify the motifs recognized by a defined TF, termed TF-centered Y1H. In this system, a random short DNA sequence insertion library was generated as the prey DNA sequences to interact with a defined TF as the bait. Using this system, novel interactions were detected between DNA motifs and the AtbZIP53 protein from Arabidopsis. We identified six
motifs that were specifically bound by AtbZIP53, including five known motifs (DOF, G-box, I-box, BS1 and MY3) and a novel motif BRS1 [basic leucine zipper (bZIP) Recognized Site 1]. The different subfamily bZIP members also recognize these six motifs, further confirming the reliability of the TF-centered Y1H results. Taken together, these results demonstrated that TF-centered Y1H could identify quickly the motifs bound by a defined TF, representing a reliable and efficient approach with the advantages of Y1H. Therefore, this TF-centered Y1H may have a wide application in protein–DNA interaction studies. Keywords Cis-acting element · Gene expression regulation · Protein–DNA interaction · bZIP · Transcription factor · Yeast one-hybrid
Introduction Xiaoyu Ji and Liuqiang Wang have contributed equally to this work. Electronic supplementary material The online version of this article (doi:10.1007/s11103-014-0234-5) contains supplementary material, which is available to authorized users. X. Ji · Y. Wang (*) Key Laboratory of Biogeography and Bioresource in Arid Land, Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Üru¯mqi 830011, Xinjiang, China e-mail: [email protected] L. Wang · X. Nie · L. He · D. Zang · B. Zhang State Key Laboratory of Tree Genetics and Breeding (Northeast Forestry University), 26 Hexing Road, Harbin 50040, China Y. Liu College of Food Engineering, Harbin University of Commerce, 1 Xuehai Street, Harbin 150028, China
The interaction between a protein and DNA is vitally important in cellular processes, and is involved in almost all cellular functions, such as transcriptional regulation, chromosome maintenance, DNA replication and repair, and chromosome segregation at mitosis (Dey et al. 2012). The interaction between DNA and protein is also the fundamental regulatory mechanism of gene expression regulation, forming the backbone of gene regulatory networks. A study of gene regulatory networks can provide important information about the gene control principles and, thereby, reveal the mechanisms of organismal cellular processes, such as growth, development, differentiation, and the response to environmental change. Therefore, it is very important to identify the regulatory elements in promoters recognized by specific transcription factors (TFs). To detect the interactions between TFs and DNA, two complementary
13
368
approaches have been developed, i.e. the TF-centered or protein-to-DNA approach, and gene-centered or DNA-toprotein approach (Reece-Hoyes et al. 2011; Arda and Walhout 2010). In the TF-centered approach, a TF or set of TFs of interest were studied to determine the DNA sequences recognized by these TF(s), such as by systematic evolution of ligands by exponential enrichment (SELEX), chromatin immunoprecipitation (ChIP) and DNA adenine methyltransferase identification (DamID) (Kim and Ren 2007). On the other hand, in the gene-centered approach, one or more regulatory DNA elements were investigated to determine which TFs could bind to these DNA sequences (Walhout 2006), such as by the yeast one-hybrid (Y1H), and electrophoretic mobility shift assay (EMSA) methods. TFs play a vital role in various biological processes, either repressing or activating transcription by binding to cis-acting genomic regions to control expression of their target genes. Many TFs can both repress and activate transcription, depending on the cellular context, which provides the first level of gene control. Therefore, determination of the cis-acting elements recognized by a certain TF is important to understand the function of TFs and revealing the regulatory networks in which they are involved. Y1H is a gene-centered (DNA-to-protein) approach, which has the advantages of being simple and efficient and can detect protein–DNA interactions in vivo. These advantages of Y1H have made it one of the most widely used genecentered technologies, and has great potential to reveal TFs that interact with specific cis-acting elements and to delineate gene regulatory networks involving different genes (Mitsuda et al. 2010). However, this method can only determine the TFs that bind to a defined DNA motif, it cannot determine the type of DNA motifs that bind to a defined TF. Therefore, development of a TF-centered Y1H that retains the advantages of Y1H may have great potential to reveal the cis-acting elements recognized by a defined TF or set of TFs, and will have wide application in the study of the interactions between DNAs and proteins. Basic leucine zippers (bZIPs) TFs are present throughout the plant kingdom, and contain two structural features: a basic region that is responsible for DNA binding, and an adjacent heptad leucine repeat domain responsible for dimerization. These two structural features constitute the conserved basic domain in a bZIP protein (Vinson et al. 1989). There are other conserved domains present in bZIP proteins, such as proline-rich, glutamine-rich and acidic domains, which play roles in transcriptional activation (Lee et al. 2006). In plants, bZIPs are important regulators involved in many physiological processes, including morphogenesis, seed formation, plant senescence, photomorphogenesis, light signaling, and abiotic and biotic stress responses (Lee et al. 2006; Mallappa et al. 2006; Alves et al. 2013; Zou et al. 2008). To date, only a few
13
Plant Mol Biol (2014) 86:367–380
bZIPs have been investigated. In addition, plant bZIP proteins preferentially bind to ACGT core sequence such as G-boxes, C-boxes and A-boxes, and also bind certain nonpalindromic binding sites (Jakoby et al. 2002). However, the bZIP proteins might bind to other, as yet unknown, motifs, and knowledge of these DNA motifs bound by TFs will be important to reveal their functions. Therefore, study of the cis-acting elements recognized by bZIP will help to fully understand their functions. In the present study, we developed a method to determine the DNA sequences recognized by a defined TF, based on the Y1H system. Using this method, we identified six motifs that interact with bZIP proteins in Arabidopsis; among these motifs, five motifs were identified that were not known to interact with bZIP proteins. These results showed that this system is a simple, reliable and efficient method to identify the DNA sequence bound by TFs of interest, and may be widely applicable to study the function of TFs and reveal novel DNA motifs. In addition, the identified motifs recognized by bZIPs will aid our understanding of the function of bZIPs.
Materials and methods Construction of a random short DNA sequence insertion library Y1H assay was performed using Matchmaker One-Hybrid System of Clontech (Clontech™, Palo Alta, CA, USA). For insertion of a random DNA fragments in pHIS2, three single-stranded DNA sequences were synthesized, named as Y1, Y2 and Y3. Y1: “CTCACTATAGGGCGAATTCCCANNNNNNCGGGGAGCTCACGCGTTCGCGA”; Y2: “CTCACTATAGGGCGAATTCCCC(T)NNNNNNC GGGGAGCTCACGCGTTCGCGA”; and Y3: “CGCGA ACGCGTGAGCTC”. The underlined ‘Ns’ are random DNA sequences that were used to determine the cis-acting element motifs recognized by a certain TF. The flanking sequences of the underlined DNA sequences are same as the two flanking sequences of the Sma I site in pHIS2. PCR was performed using Y1 and Y2 as templates, and Y3 as the primer. The PCR reaction system included: 2 μL of Y1 or Y2 (10 μM), 0.5 μL of dNTP (10 mM each), 3 μL of Y3 (10 μM), 1 μL PCR buffer, and 0.5 U ExTaq (Takara, Dalian, China) with the volume of 10 μL. The PCR reaction conditions were as follows: 94 °C for 90 s, 55 °C for 15 min, 50 °C for 30 min (one cycle). The pHIS2 vector was Sma I (Promega) digested and the digested vector was purified by agarose gel electrophoresis (0.8 % agarose). For addition of a single ‘T’ base at the linear pHIS2 terminals, the following reagents were included in the PCR: 1 μg of purified pHIS2, 0.5 μL of
Plant Mol Biol (2014) 86:367–380
dTTP (10 mM), 1 μL 10× PCR buffer, and 0.5 U ExTaq (Takara), in 10 μL. The PCR was carried out at 74 °C for 30 min, and was purified by agarose gel electrophoresis, to obtain the T-vector of pHIS2. The T-A ligation system includes 0.2 μg of T-vector of pHIS2, 0.5 μL of PCR production (PCR products of Y1 + Y3 or Y2 + Y3), 1 μL of 10× ligation buffer, 3 U of T4 ligase (Promega), PEG6000 at final concentration of 10 % (w/v). The ligation conditions were 12 °C for 20 h, then 10 U Sma I (Promega) was added, and incubated at 25 °C for 4 h to linearize the empty pHIS2. The two ligations (Y1 + Y3 and Y2 + Y3) were then mixed and the mixture was transformed into DH5α Escherichia coli competent cells using heat shock. After incubation at 42 °C for 90 s, 1 mL of LB medium was added to the transformation mixture, which was incubated at 37 °C for 1 h, then 10, 50 and 100 μL of transformation mixture were plated to evaluate the transformation efficiency. LB medium (4 mL) was added to the remaining reactions. After incubation at 37 °C for 14 h, the culture was used for plasmid isolation. The isolated plasmids formed the random DNA insertion library, which was used for screening the DNA motifs recognized by a definite TF. The ORF of AtbZIP53 was cloned into pGADT7-Rec2 (designated as: pGADT7-AtbZIP53) at the site between the SMART III Sequence and the CDS III Sequence using the infusion method, following the procedures in the user manual (In-Fusion® HD Cloning Kit). Screening the random DNA insertion library The following were combined into a sterile 15-mL tube: 2 μg pGADT7-AtbZIP53, 1.5 μg of the random DNA insertion library, 10 μL of Herring Testes Carrier DNA. The mixture was added to 600 μL of competent Y187 yeast cells. The transformation and selection methods were performed according to the manufacturer’s protocol (BD Matchmaker™ Library Construction & Screening Kits User Manual). The positive clones were selected on the SD/-His/-Leu/-Trp (TDO) medium supplied with 30 mM 3-AT (3-Amino-1, 2, 4-triazole). The positive clones were further selected on the high stringency selection mediums (supplied with 80 mM 3-AT) to select the clones having high binding affinities to AtbZIP53. For all the Y1H assay in this study, the primary culture of yeast transformants were transferred into fresh medium for further culture, and were grown to about 0.6 at OD600 before spotting. The densities of yeast cultures were measured at OD600, and each transformants was adjusted to equal density for spotting. The positive transformants were further confirmed by spotting yeast cells with serial dilutions (1/1, 1/10, 1/100, 1/1,000) onto TDO medium supplied with 50 mM 3-AT.
369
Fig. 1 Serial deletion of the inserted sequences to determine the novel motif recognized by AtbZIP53. The insertion sequences that contain the sequences recognized by AtbZIP53, named BRS1, was studied. pGADT7-AtbZIP53/pHIS2-BRS1D1-8: the serially deleted insertion sequences (D1–D8) interacted with AtbZIP53. D1–D8: the sequences of serially deleted insertion D1–D8 (right panel). Positive control: pGADT7-p53 interacting with p53HIS2 (p53HIS2/pGADT7P53); Negative control: pGADT7-AtbZIP53 interacting with p53HIS2 (p53HIS2/pGADT7-AtbZIP53)
The transformants grown at SD/-Leu/-Trp (DDO) were used as growth positive controls. Analysis of the insertion sequences of positive clones The pHIS2 plasmids were rescued from the positive clones identified by TF-centered Y1H analysis, and sequenced. The insertion sequences were analyzed using PLACE (http:// www.dna.affrc.go.jp/PLACE/) and PlantCARE (http:// bioinformatics.psb.ugent.be/webtools/plantcare/html/) to identify whether they were known motifs. Determination of the novel motif recognized by AtbZIP53 The sequence “CAGTGCGC”, containing a novel DNA motif, was serially deleted to determine the exact DNA sequence recognized by AtbZIP53. The two flanking sequences of the insertion may also be part of a novel DNA motif; therefore, three bases of each flanking sequence, together with the insertion, i.e. “CCCCAGTGCGCGGG”, was used in serial deletion analysis. Three tandem copies of these deletions were cloned into pHIS2, and their interactions with AtbZIP53 assayed using Y1H analysis. The sequences of the serial deletions of the novel motif are shown in Fig. 1. The novel motif was designed as BRS1 (NRS: bZIP Recognized Sequence). To study the presents of BRS1 in the genome of Arabidopsis, the BRS1 sequence “GTGCG” was search on the promoter regions of all the Arabidopsis genes using Patmatch program with the patter search of “TAIR10 Loci Upstream Sequences—1,000 bp (DNA)” in the Tair database (http://www.arabidopsis.org/cgi-bin/patmatch/nph-patmatch.pl).
13
370
Y1H assay of the binding of AtbZIP53 to different motifs and the truncated promoters Three tandem copies of the BRS1 (“GTGCG”), Dof (“AAAG”), I-box (“GATAA”), G-box (“CACGTG”), BS1 “AGCGGG”, and MY3 motif (“CGACG”) together with their respective mutants “ATGCG”, “GCGCG”, “GTACG”, “GTGAG”, “GTGCA”, “GCAAG”, “ACAAA” (mutants of BRS1, named BRSM1, BRSM2, BRSM3, BRSM4, BRSM5, BRSM6 and BRSM7, respectively); “CAAG”, “ACAG”, “AACG”, “AAAA”, “CAAA”, “ACCG”, and “CCCA” (mutants of Dof, named DM1, DM2, DM3, DM4, DM5, DM6 and DM7, respectively); “AATAA”, “GCTAA”, “GACAA”, “GATCC”, “ACCAA”, “GCCCC” and “ACCCC” (mutants of I-box, named IM1, IM2, IM3, IM4, IM5, IM6 and IM7, respectively); “CCCGTG”, “CAAGTG”, “CACATG”, “CACGCG”, “CCCGCG”, “CCAACG”, and “ACAAAA” (mutants of G-box, named GM1, GM2, GM3, GM4, GM5, GM6 and GM7, respectively); “CGCGGG”, “AACGGG”, “AGAGGG”, “AGCAAA”, “CACGGG”, “CAAGGG”, and “CAAAAA” (mutants of BS1, named BM1, BM2, BM3, BM4, BM5, BM6 and BM7, respectively); “AGTCG”, “CATCG”, “CGCCG”, “CGTAG”, “CGTCA”, “CACAG”, and “AACAA” (mutants of MY3, named MM1, MM2, MM3, MM4, MM5, MM6 and MM7, respectively were cloned into pHIS2 (see Supporting Information Table S2 for primers), respectively. The above motifs were mutated following this principle, i.e. “A/T” was mutated to “C” and “C/G” was mutated to “A”. Y1H screening analysis was performed to determine whether AtbZIP53 was able to activate gene expression by interacting with the promoters containing these identified motifs. The sequences of BRS1, Dof, I-box, G-box, BS1, and MY3 were searched against the promoter regions of genes in Arabidopsis in the Tair database (http://www.arabidopsis.org/) to find the genes that were putatively regulated by AtbZIP53. The criteria for promoter selection were the promoter regions (0 to −1,000 bp) of genes contain fewer kinds of the motifs of BRS1, Dof, I-box, G-box, BS1, or MY3, but have more copies of the contained motif. The genes of AT1G51090, AT5G64860, AT1G51090, AT3G10410, AT2G22450, and AT2G43550, which contain BRS1, Dof, I-box, G-box, BS1, and MY3 motifs, respectively, in their promoter regions were selected for further study. Vector pHIS2 harboring the truncated promoter of AT1G51090 (without Dof, I-box, G-box, BS1, and MY3) containing BRS1 “GTGCG” (pHIS2-BRS1p+) or with the deleted “GTGCG” (pHIS2-BRS1p−) (as negative controls) driving HIS3 gene, were generated as reporter vectors (see Supporting Information Table S2 for primers). Similarly, the pHIS2 harboring the truncated promoters of AT5G64860, AT1G51090, AT3G10410, AT2G22450, and AT2G43550,
13
Plant Mol Biol (2014) 86:367–380
which contained the motif of Dof, I-box, G-box, BS1, and MY3, respectively, but lacked the other five corresponding motifs (named as pHIS2-Dofp+, pHIS2-I-boxp+, pHIS2G-boxp+ pHIS2-BS1p+ and pHIS2-MY3p+), were generated as reporter vectors (see Supporting Information Table S2 for primers). Meanwhile, pHIS2 harboring truncated promoters of AT5G64860, AT1G51090, AT3G10410, AT2G22450, and AT2G43550, which lacked the Dof, I-box, G-box, BS1, and MY3 (named as pHIS2-Dofp−, pHIS2-Iboxp−, pHIS2-G-boxp−, pHIS2-BS1p−, pHIS2-MY3−), respectively, to drive expression of HIS3, were generated as negative controls. The interactions of these sequences with the AtbZIP53 were studied using Y1H analysis. Tobacco transient activation assay To further verify the specificity of the bindings of AtbZIP53 to different motifs or the truncated promoter containing or lacking the studied motifs, DNA–protein interactions were further assayed in tobacco leaves. For reporter plasmids construction, three tandem copies of BRS1 “GTGCG” and its complete mutant “ACAAA”, Dof “AAAG” and its complete mutant “CCCA”, I-box “GATAA” and its complete mutant “ACCCC”, G-box “CACGTG” and its complete mutant “ACAAAA”, BS1 “AGCGGG” and its complete mutant “CAAAAA”, and MY3 “CGACG” and its complete mutant “AACAA”, were cloned into a reformed pCAMBIA1301 (35S: hygromycion had been delegated, and a 46 bp minimal promoter was inserted between the region of BglII site and ATG of GUS) to replace 35S promoter and be fused to the minimal 35S promoter (−46 to +1) for driving a GUS gene (designed as pCAM-BRS1, pCAM-Dof, pCAM-I-box, pCAM-G-box, pCAM-BS1, pCAM-MY3, pCAM-BRSM7, pCAM-DM7, pCAM-IM7, pCAM-GM7, pCAM-BM7, pCAM-MM7, respectively) (see Supporting Information Table S2 for primers). For binding assay of AtbZIP53 to the truncated promoters, the reporter vectors were constructed using the truncated promoters as in Y1H assay, and these truncated promoters were fused to the minimal 35S promoter (−46 to +1) to drive GUS expression in the reformed pCAMBIA1301 (designed as pCAM-BRS1p+, pCAM-BRS1p−, pCAM-Dofp+, pCAM-Dofp−, pCAM-I-boxp+, pCAM-Iboxp−, pCAM-G-boxp+, pCAM-G-boxp−, pCAM-BS1p+, pCAM-BS1p−, pCAM-MY3p+, pCAM-MY3p− respectively) (see Supporting Information Table S2 for primers). The effector was constructed by cloning the ORF of AtbZIP53 into pROKII under the control of the CaMV 35S promoter (named as pROKII-AtbZIP53). To compare the binding affinities of AtbZIP53 to different motifs, the effector construct pROKIIAtbZIP53 was cotransformed with each reporter constructs pCAM-BRS1, -Dof, -I-box, -G-box, -BS1, -MY3, C-box (“GACGTC”), or A-box (“TACGTA”) into tobacco leaves for GUS activity analysis. The transformation of a single reporter
Plant Mol Biol (2014) 86:367–380
plasmid and the empty effector vector were used as negative controls. The reformed pCAMBIA1301 was transformed and used as the positive control. To normalize transformation efficiency, the construct harboring a luciferase gene driven by the CaMV 35S promoter (35S:Luc) was also co-transferred. The transformation into tobacco was performed using the particle bombardment method (Bio-Rad, Hercules, CA, USA), and the procedure of transformation was followed the manual for the Biolistic® PDS-1000/He Particle Delivery System (Bio-Rad). The transformation conditions were 9 cm target distance, 1,100 psi helium pressure, and 2 times number of bombardment. The quantity ratios of reporter, effector and 35S:Luc were 2:2:1. Three independent biological repeats were performed, and GUS activity levels were determined according to Jefferson (Jefferson et al. 1987). ChIP analysis To further confirm the binding of AtbZIP53 to BRS1, Dof, I-box, G-box, BS1, and MY3 cis-acting elements, ChIP analysis was performed. The coding regions of AtbZIP53 without the termination codon were ligated in frame to the N-terminus of GFP driven by the CaMV 35S promoter to generate the AtbZIP53::GFP fusion gene. For ChIP analysis in tobacco, each of the reporter constructs used in the GUS activity analyses above were cotransformed with the AtbZIP53::GFP fusion gene into tobacco leaves using particle bombardment. Three days after transformation, the transgenic plants were used for the ChIP assay. The ChIP procedure followed the method of Haring et al. (2007). Briefly, protein and DNA were cross-linked using 3 % formaldehyde. The purified cross-linked nuclei were sonicated to shear the chromatin into 0.2–1.0 kb fragments, and 1/10 volume was saved as the input control. The remaining sonicated chromatin was divided into two aliquots, which were incubated with GFP antibody (ChIP+) or a rabbit anti-hemagglutinin (HA) antibody as a negative control (ChIP−), respectively. The antibody-bound complex was precipitated with protein A Agarose beads. The DNA fragments were released from the immunoprecipitated complexes by reversing the cross-linking at 65 °C for 3 h. Immunoprecipitated DNA was purified by chloroform extraction. For gel electrophoresis analysis, PCR was performed as follows: 94 °C for 2 min; 35 cycles of 94 °C for 12 s, 58 °C for 30 s and 72 °C for 30 s; and 72 °C for 5 min. The PCR productions were analyzed by gel electrophoresis. Real-time PCR was performed using SYBR Green Real-time PCR Master Mix (Toyobo) in a 20 μL reaction volume on a MJ Research OpticonTM2 instrument (BioRad, Hercules, CA, USA). The PCR cycling parameters was as follows: 94 °C for 2 min; 45 cycles of 94 °C for 20 s, 56 °C for 30 s, and 72 °C for 1 min; and 80 °C for 1 s for a plate reading. The sequence of α-tubulin was used as the internal control. Primers used are listed in the Supplemental
371
Table S5. The ChIP assays were performed three times, with similar results, and data are mean ± SD from three independent experiments. Analysis of the interaction of other Arabidopsis bZIP family members with AtbZIP53‑binding elements To study the interaction of these six motifs with other bZIPs, eight bZIPs from different subfamilies were selected for study, including At5g24800, At2g16770, At5g06960, At1g75390, At1g43700, At5g11260, At1g42990, and At3g56850. The ORF of these genes were cloned into pGADT7-Rec2 (designated as: pGADT7-At5g24800, -At2g16770, -At5g06960, -At1g75390, -At1g43700, -At5g11260, -At1g42990, -At3g56850) at the site between the SMART III Sequence and CDS III Sequence, using the infusion method, as effectors. Each of these effectors was cotransformed with the reporter vectors pHIS2-BRS1, -Dof, -I-box, -G-box, -BS1, or -MY3 into yeast Y187 cells. The bindings of these eight bZIPs to the motifs harbored by pHIS2 were analyzed using the Y1H system. Statistical analyses Statistical analyses were carried out using SPSS 16.0 (SPSSInc, Chicago, IL, USA) software. Data were compared using Student’s t test. Differences were considered to be significant if P
A novel method to identify the DNA motifs recognized by a defined transcription factor Xiaoyu Ji · Liuqiang Wang · Xianguang Nie · Lin He · Dandan Zang · Yujia Liu · Bing Zhang · Yucheng Wang
Received: 21 June 2014 / Accepted: 29 July 2014 / Published online: 10 August 2014 © Springer Science+Business Media Dordrecht 2014
Abstract The interaction between a protein and DNA is involved in almost all cellular functions, and is vitally important in cellular processes. Two complementary approaches are used to detect the interactions between a transcription factor (TF) and DNA, i.e. the TF-centered or protein–DNA approach, and the gene-centered or DNA– protein approach. The yeast one-hybrid (Y1H) is a powerful and widely used system to identify DNA–protein interactions. However, a powerful method to study protein– DNA interactions like Y1H is lacking. Here, we developed a protein–DNA method based on the Y1H system to identify the motifs recognized by a defined TF, termed TF-centered Y1H. In this system, a random short DNA sequence insertion library was generated as the prey DNA sequences to interact with a defined TF as the bait. Using this system, novel interactions were detected between DNA motifs and the AtbZIP53 protein from Arabidopsis. We identified six
motifs that were specifically bound by AtbZIP53, including five known motifs (DOF, G-box, I-box, BS1 and MY3) and a novel motif BRS1 [basic leucine zipper (bZIP) Recognized Site 1]. The different subfamily bZIP members also recognize these six motifs, further confirming the reliability of the TF-centered Y1H results. Taken together, these results demonstrated that TF-centered Y1H could identify quickly the motifs bound by a defined TF, representing a reliable and efficient approach with the advantages of Y1H. Therefore, this TF-centered Y1H may have a wide application in protein–DNA interaction studies. Keywords Cis-acting element · Gene expression regulation · Protein–DNA interaction · bZIP · Transcription factor · Yeast one-hybrid
Introduction Xiaoyu Ji and Liuqiang Wang have contributed equally to this work. Electronic supplementary material The online version of this article (doi:10.1007/s11103-014-0234-5) contains supplementary material, which is available to authorized users. X. Ji · Y. Wang (*) Key Laboratory of Biogeography and Bioresource in Arid Land, Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Üru¯mqi 830011, Xinjiang, China e-mail: [email protected] L. Wang · X. Nie · L. He · D. Zang · B. Zhang State Key Laboratory of Tree Genetics and Breeding (Northeast Forestry University), 26 Hexing Road, Harbin 50040, China Y. Liu College of Food Engineering, Harbin University of Commerce, 1 Xuehai Street, Harbin 150028, China
The interaction between a protein and DNA is vitally important in cellular processes, and is involved in almost all cellular functions, such as transcriptional regulation, chromosome maintenance, DNA replication and repair, and chromosome segregation at mitosis (Dey et al. 2012). The interaction between DNA and protein is also the fundamental regulatory mechanism of gene expression regulation, forming the backbone of gene regulatory networks. A study of gene regulatory networks can provide important information about the gene control principles and, thereby, reveal the mechanisms of organismal cellular processes, such as growth, development, differentiation, and the response to environmental change. Therefore, it is very important to identify the regulatory elements in promoters recognized by specific transcription factors (TFs). To detect the interactions between TFs and DNA, two complementary
13
368
approaches have been developed, i.e. the TF-centered or protein-to-DNA approach, and gene-centered or DNA-toprotein approach (Reece-Hoyes et al. 2011; Arda and Walhout 2010). In the TF-centered approach, a TF or set of TFs of interest were studied to determine the DNA sequences recognized by these TF(s), such as by systematic evolution of ligands by exponential enrichment (SELEX), chromatin immunoprecipitation (ChIP) and DNA adenine methyltransferase identification (DamID) (Kim and Ren 2007). On the other hand, in the gene-centered approach, one or more regulatory DNA elements were investigated to determine which TFs could bind to these DNA sequences (Walhout 2006), such as by the yeast one-hybrid (Y1H), and electrophoretic mobility shift assay (EMSA) methods. TFs play a vital role in various biological processes, either repressing or activating transcription by binding to cis-acting genomic regions to control expression of their target genes. Many TFs can both repress and activate transcription, depending on the cellular context, which provides the first level of gene control. Therefore, determination of the cis-acting elements recognized by a certain TF is important to understand the function of TFs and revealing the regulatory networks in which they are involved. Y1H is a gene-centered (DNA-to-protein) approach, which has the advantages of being simple and efficient and can detect protein–DNA interactions in vivo. These advantages of Y1H have made it one of the most widely used genecentered technologies, and has great potential to reveal TFs that interact with specific cis-acting elements and to delineate gene regulatory networks involving different genes (Mitsuda et al. 2010). However, this method can only determine the TFs that bind to a defined DNA motif, it cannot determine the type of DNA motifs that bind to a defined TF. Therefore, development of a TF-centered Y1H that retains the advantages of Y1H may have great potential to reveal the cis-acting elements recognized by a defined TF or set of TFs, and will have wide application in the study of the interactions between DNAs and proteins. Basic leucine zippers (bZIPs) TFs are present throughout the plant kingdom, and contain two structural features: a basic region that is responsible for DNA binding, and an adjacent heptad leucine repeat domain responsible for dimerization. These two structural features constitute the conserved basic domain in a bZIP protein (Vinson et al. 1989). There are other conserved domains present in bZIP proteins, such as proline-rich, glutamine-rich and acidic domains, which play roles in transcriptional activation (Lee et al. 2006). In plants, bZIPs are important regulators involved in many physiological processes, including morphogenesis, seed formation, plant senescence, photomorphogenesis, light signaling, and abiotic and biotic stress responses (Lee et al. 2006; Mallappa et al. 2006; Alves et al. 2013; Zou et al. 2008). To date, only a few
13
Plant Mol Biol (2014) 86:367–380
bZIPs have been investigated. In addition, plant bZIP proteins preferentially bind to ACGT core sequence such as G-boxes, C-boxes and A-boxes, and also bind certain nonpalindromic binding sites (Jakoby et al. 2002). However, the bZIP proteins might bind to other, as yet unknown, motifs, and knowledge of these DNA motifs bound by TFs will be important to reveal their functions. Therefore, study of the cis-acting elements recognized by bZIP will help to fully understand their functions. In the present study, we developed a method to determine the DNA sequences recognized by a defined TF, based on the Y1H system. Using this method, we identified six motifs that interact with bZIP proteins in Arabidopsis; among these motifs, five motifs were identified that were not known to interact with bZIP proteins. These results showed that this system is a simple, reliable and efficient method to identify the DNA sequence bound by TFs of interest, and may be widely applicable to study the function of TFs and reveal novel DNA motifs. In addition, the identified motifs recognized by bZIPs will aid our understanding of the function of bZIPs.
Materials and methods Construction of a random short DNA sequence insertion library Y1H assay was performed using Matchmaker One-Hybrid System of Clontech (Clontech™, Palo Alta, CA, USA). For insertion of a random DNA fragments in pHIS2, three single-stranded DNA sequences were synthesized, named as Y1, Y2 and Y3. Y1: “CTCACTATAGGGCGAATTCCCANNNNNNCGGGGAGCTCACGCGTTCGCGA”; Y2: “CTCACTATAGGGCGAATTCCCC(T)NNNNNNC GGGGAGCTCACGCGTTCGCGA”; and Y3: “CGCGA ACGCGTGAGCTC”. The underlined ‘Ns’ are random DNA sequences that were used to determine the cis-acting element motifs recognized by a certain TF. The flanking sequences of the underlined DNA sequences are same as the two flanking sequences of the Sma I site in pHIS2. PCR was performed using Y1 and Y2 as templates, and Y3 as the primer. The PCR reaction system included: 2 μL of Y1 or Y2 (10 μM), 0.5 μL of dNTP (10 mM each), 3 μL of Y3 (10 μM), 1 μL PCR buffer, and 0.5 U ExTaq (Takara, Dalian, China) with the volume of 10 μL. The PCR reaction conditions were as follows: 94 °C for 90 s, 55 °C for 15 min, 50 °C for 30 min (one cycle). The pHIS2 vector was Sma I (Promega) digested and the digested vector was purified by agarose gel electrophoresis (0.8 % agarose). For addition of a single ‘T’ base at the linear pHIS2 terminals, the following reagents were included in the PCR: 1 μg of purified pHIS2, 0.5 μL of
Plant Mol Biol (2014) 86:367–380
dTTP (10 mM), 1 μL 10× PCR buffer, and 0.5 U ExTaq (Takara), in 10 μL. The PCR was carried out at 74 °C for 30 min, and was purified by agarose gel electrophoresis, to obtain the T-vector of pHIS2. The T-A ligation system includes 0.2 μg of T-vector of pHIS2, 0.5 μL of PCR production (PCR products of Y1 + Y3 or Y2 + Y3), 1 μL of 10× ligation buffer, 3 U of T4 ligase (Promega), PEG6000 at final concentration of 10 % (w/v). The ligation conditions were 12 °C for 20 h, then 10 U Sma I (Promega) was added, and incubated at 25 °C for 4 h to linearize the empty pHIS2. The two ligations (Y1 + Y3 and Y2 + Y3) were then mixed and the mixture was transformed into DH5α Escherichia coli competent cells using heat shock. After incubation at 42 °C for 90 s, 1 mL of LB medium was added to the transformation mixture, which was incubated at 37 °C for 1 h, then 10, 50 and 100 μL of transformation mixture were plated to evaluate the transformation efficiency. LB medium (4 mL) was added to the remaining reactions. After incubation at 37 °C for 14 h, the culture was used for plasmid isolation. The isolated plasmids formed the random DNA insertion library, which was used for screening the DNA motifs recognized by a definite TF. The ORF of AtbZIP53 was cloned into pGADT7-Rec2 (designated as: pGADT7-AtbZIP53) at the site between the SMART III Sequence and the CDS III Sequence using the infusion method, following the procedures in the user manual (In-Fusion® HD Cloning Kit). Screening the random DNA insertion library The following were combined into a sterile 15-mL tube: 2 μg pGADT7-AtbZIP53, 1.5 μg of the random DNA insertion library, 10 μL of Herring Testes Carrier DNA. The mixture was added to 600 μL of competent Y187 yeast cells. The transformation and selection methods were performed according to the manufacturer’s protocol (BD Matchmaker™ Library Construction & Screening Kits User Manual). The positive clones were selected on the SD/-His/-Leu/-Trp (TDO) medium supplied with 30 mM 3-AT (3-Amino-1, 2, 4-triazole). The positive clones were further selected on the high stringency selection mediums (supplied with 80 mM 3-AT) to select the clones having high binding affinities to AtbZIP53. For all the Y1H assay in this study, the primary culture of yeast transformants were transferred into fresh medium for further culture, and were grown to about 0.6 at OD600 before spotting. The densities of yeast cultures were measured at OD600, and each transformants was adjusted to equal density for spotting. The positive transformants were further confirmed by spotting yeast cells with serial dilutions (1/1, 1/10, 1/100, 1/1,000) onto TDO medium supplied with 50 mM 3-AT.
369
Fig. 1 Serial deletion of the inserted sequences to determine the novel motif recognized by AtbZIP53. The insertion sequences that contain the sequences recognized by AtbZIP53, named BRS1, was studied. pGADT7-AtbZIP53/pHIS2-BRS1D1-8: the serially deleted insertion sequences (D1–D8) interacted with AtbZIP53. D1–D8: the sequences of serially deleted insertion D1–D8 (right panel). Positive control: pGADT7-p53 interacting with p53HIS2 (p53HIS2/pGADT7P53); Negative control: pGADT7-AtbZIP53 interacting with p53HIS2 (p53HIS2/pGADT7-AtbZIP53)
The transformants grown at SD/-Leu/-Trp (DDO) were used as growth positive controls. Analysis of the insertion sequences of positive clones The pHIS2 plasmids were rescued from the positive clones identified by TF-centered Y1H analysis, and sequenced. The insertion sequences were analyzed using PLACE (http:// www.dna.affrc.go.jp/PLACE/) and PlantCARE (http:// bioinformatics.psb.ugent.be/webtools/plantcare/html/) to identify whether they were known motifs. Determination of the novel motif recognized by AtbZIP53 The sequence “CAGTGCGC”, containing a novel DNA motif, was serially deleted to determine the exact DNA sequence recognized by AtbZIP53. The two flanking sequences of the insertion may also be part of a novel DNA motif; therefore, three bases of each flanking sequence, together with the insertion, i.e. “CCCCAGTGCGCGGG”, was used in serial deletion analysis. Three tandem copies of these deletions were cloned into pHIS2, and their interactions with AtbZIP53 assayed using Y1H analysis. The sequences of the serial deletions of the novel motif are shown in Fig. 1. The novel motif was designed as BRS1 (NRS: bZIP Recognized Sequence). To study the presents of BRS1 in the genome of Arabidopsis, the BRS1 sequence “GTGCG” was search on the promoter regions of all the Arabidopsis genes using Patmatch program with the patter search of “TAIR10 Loci Upstream Sequences—1,000 bp (DNA)” in the Tair database (http://www.arabidopsis.org/cgi-bin/patmatch/nph-patmatch.pl).
13
370
Y1H assay of the binding of AtbZIP53 to different motifs and the truncated promoters Three tandem copies of the BRS1 (“GTGCG”), Dof (“AAAG”), I-box (“GATAA”), G-box (“CACGTG”), BS1 “AGCGGG”, and MY3 motif (“CGACG”) together with their respective mutants “ATGCG”, “GCGCG”, “GTACG”, “GTGAG”, “GTGCA”, “GCAAG”, “ACAAA” (mutants of BRS1, named BRSM1, BRSM2, BRSM3, BRSM4, BRSM5, BRSM6 and BRSM7, respectively); “CAAG”, “ACAG”, “AACG”, “AAAA”, “CAAA”, “ACCG”, and “CCCA” (mutants of Dof, named DM1, DM2, DM3, DM4, DM5, DM6 and DM7, respectively); “AATAA”, “GCTAA”, “GACAA”, “GATCC”, “ACCAA”, “GCCCC” and “ACCCC” (mutants of I-box, named IM1, IM2, IM3, IM4, IM5, IM6 and IM7, respectively); “CCCGTG”, “CAAGTG”, “CACATG”, “CACGCG”, “CCCGCG”, “CCAACG”, and “ACAAAA” (mutants of G-box, named GM1, GM2, GM3, GM4, GM5, GM6 and GM7, respectively); “CGCGGG”, “AACGGG”, “AGAGGG”, “AGCAAA”, “CACGGG”, “CAAGGG”, and “CAAAAA” (mutants of BS1, named BM1, BM2, BM3, BM4, BM5, BM6 and BM7, respectively); “AGTCG”, “CATCG”, “CGCCG”, “CGTAG”, “CGTCA”, “CACAG”, and “AACAA” (mutants of MY3, named MM1, MM2, MM3, MM4, MM5, MM6 and MM7, respectively were cloned into pHIS2 (see Supporting Information Table S2 for primers), respectively. The above motifs were mutated following this principle, i.e. “A/T” was mutated to “C” and “C/G” was mutated to “A”. Y1H screening analysis was performed to determine whether AtbZIP53 was able to activate gene expression by interacting with the promoters containing these identified motifs. The sequences of BRS1, Dof, I-box, G-box, BS1, and MY3 were searched against the promoter regions of genes in Arabidopsis in the Tair database (http://www.arabidopsis.org/) to find the genes that were putatively regulated by AtbZIP53. The criteria for promoter selection were the promoter regions (0 to −1,000 bp) of genes contain fewer kinds of the motifs of BRS1, Dof, I-box, G-box, BS1, or MY3, but have more copies of the contained motif. The genes of AT1G51090, AT5G64860, AT1G51090, AT3G10410, AT2G22450, and AT2G43550, which contain BRS1, Dof, I-box, G-box, BS1, and MY3 motifs, respectively, in their promoter regions were selected for further study. Vector pHIS2 harboring the truncated promoter of AT1G51090 (without Dof, I-box, G-box, BS1, and MY3) containing BRS1 “GTGCG” (pHIS2-BRS1p+) or with the deleted “GTGCG” (pHIS2-BRS1p−) (as negative controls) driving HIS3 gene, were generated as reporter vectors (see Supporting Information Table S2 for primers). Similarly, the pHIS2 harboring the truncated promoters of AT5G64860, AT1G51090, AT3G10410, AT2G22450, and AT2G43550,
13
Plant Mol Biol (2014) 86:367–380
which contained the motif of Dof, I-box, G-box, BS1, and MY3, respectively, but lacked the other five corresponding motifs (named as pHIS2-Dofp+, pHIS2-I-boxp+, pHIS2G-boxp+ pHIS2-BS1p+ and pHIS2-MY3p+), were generated as reporter vectors (see Supporting Information Table S2 for primers). Meanwhile, pHIS2 harboring truncated promoters of AT5G64860, AT1G51090, AT3G10410, AT2G22450, and AT2G43550, which lacked the Dof, I-box, G-box, BS1, and MY3 (named as pHIS2-Dofp−, pHIS2-Iboxp−, pHIS2-G-boxp−, pHIS2-BS1p−, pHIS2-MY3−), respectively, to drive expression of HIS3, were generated as negative controls. The interactions of these sequences with the AtbZIP53 were studied using Y1H analysis. Tobacco transient activation assay To further verify the specificity of the bindings of AtbZIP53 to different motifs or the truncated promoter containing or lacking the studied motifs, DNA–protein interactions were further assayed in tobacco leaves. For reporter plasmids construction, three tandem copies of BRS1 “GTGCG” and its complete mutant “ACAAA”, Dof “AAAG” and its complete mutant “CCCA”, I-box “GATAA” and its complete mutant “ACCCC”, G-box “CACGTG” and its complete mutant “ACAAAA”, BS1 “AGCGGG” and its complete mutant “CAAAAA”, and MY3 “CGACG” and its complete mutant “AACAA”, were cloned into a reformed pCAMBIA1301 (35S: hygromycion had been delegated, and a 46 bp minimal promoter was inserted between the region of BglII site and ATG of GUS) to replace 35S promoter and be fused to the minimal 35S promoter (−46 to +1) for driving a GUS gene (designed as pCAM-BRS1, pCAM-Dof, pCAM-I-box, pCAM-G-box, pCAM-BS1, pCAM-MY3, pCAM-BRSM7, pCAM-DM7, pCAM-IM7, pCAM-GM7, pCAM-BM7, pCAM-MM7, respectively) (see Supporting Information Table S2 for primers). For binding assay of AtbZIP53 to the truncated promoters, the reporter vectors were constructed using the truncated promoters as in Y1H assay, and these truncated promoters were fused to the minimal 35S promoter (−46 to +1) to drive GUS expression in the reformed pCAMBIA1301 (designed as pCAM-BRS1p+, pCAM-BRS1p−, pCAM-Dofp+, pCAM-Dofp−, pCAM-I-boxp+, pCAM-Iboxp−, pCAM-G-boxp+, pCAM-G-boxp−, pCAM-BS1p+, pCAM-BS1p−, pCAM-MY3p+, pCAM-MY3p− respectively) (see Supporting Information Table S2 for primers). The effector was constructed by cloning the ORF of AtbZIP53 into pROKII under the control of the CaMV 35S promoter (named as pROKII-AtbZIP53). To compare the binding affinities of AtbZIP53 to different motifs, the effector construct pROKIIAtbZIP53 was cotransformed with each reporter constructs pCAM-BRS1, -Dof, -I-box, -G-box, -BS1, -MY3, C-box (“GACGTC”), or A-box (“TACGTA”) into tobacco leaves for GUS activity analysis. The transformation of a single reporter
Plant Mol Biol (2014) 86:367–380
plasmid and the empty effector vector were used as negative controls. The reformed pCAMBIA1301 was transformed and used as the positive control. To normalize transformation efficiency, the construct harboring a luciferase gene driven by the CaMV 35S promoter (35S:Luc) was also co-transferred. The transformation into tobacco was performed using the particle bombardment method (Bio-Rad, Hercules, CA, USA), and the procedure of transformation was followed the manual for the Biolistic® PDS-1000/He Particle Delivery System (Bio-Rad). The transformation conditions were 9 cm target distance, 1,100 psi helium pressure, and 2 times number of bombardment. The quantity ratios of reporter, effector and 35S:Luc were 2:2:1. Three independent biological repeats were performed, and GUS activity levels were determined according to Jefferson (Jefferson et al. 1987). ChIP analysis To further confirm the binding of AtbZIP53 to BRS1, Dof, I-box, G-box, BS1, and MY3 cis-acting elements, ChIP analysis was performed. The coding regions of AtbZIP53 without the termination codon were ligated in frame to the N-terminus of GFP driven by the CaMV 35S promoter to generate the AtbZIP53::GFP fusion gene. For ChIP analysis in tobacco, each of the reporter constructs used in the GUS activity analyses above were cotransformed with the AtbZIP53::GFP fusion gene into tobacco leaves using particle bombardment. Three days after transformation, the transgenic plants were used for the ChIP assay. The ChIP procedure followed the method of Haring et al. (2007). Briefly, protein and DNA were cross-linked using 3 % formaldehyde. The purified cross-linked nuclei were sonicated to shear the chromatin into 0.2–1.0 kb fragments, and 1/10 volume was saved as the input control. The remaining sonicated chromatin was divided into two aliquots, which were incubated with GFP antibody (ChIP+) or a rabbit anti-hemagglutinin (HA) antibody as a negative control (ChIP−), respectively. The antibody-bound complex was precipitated with protein A Agarose beads. The DNA fragments were released from the immunoprecipitated complexes by reversing the cross-linking at 65 °C for 3 h. Immunoprecipitated DNA was purified by chloroform extraction. For gel electrophoresis analysis, PCR was performed as follows: 94 °C for 2 min; 35 cycles of 94 °C for 12 s, 58 °C for 30 s and 72 °C for 30 s; and 72 °C for 5 min. The PCR productions were analyzed by gel electrophoresis. Real-time PCR was performed using SYBR Green Real-time PCR Master Mix (Toyobo) in a 20 μL reaction volume on a MJ Research OpticonTM2 instrument (BioRad, Hercules, CA, USA). The PCR cycling parameters was as follows: 94 °C for 2 min; 45 cycles of 94 °C for 20 s, 56 °C for 30 s, and 72 °C for 1 min; and 80 °C for 1 s for a plate reading. The sequence of α-tubulin was used as the internal control. Primers used are listed in the Supplemental
371
Table S5. The ChIP assays were performed three times, with similar results, and data are mean ± SD from three independent experiments. Analysis of the interaction of other Arabidopsis bZIP family members with AtbZIP53‑binding elements To study the interaction of these six motifs with other bZIPs, eight bZIPs from different subfamilies were selected for study, including At5g24800, At2g16770, At5g06960, At1g75390, At1g43700, At5g11260, At1g42990, and At3g56850. The ORF of these genes were cloned into pGADT7-Rec2 (designated as: pGADT7-At5g24800, -At2g16770, -At5g06960, -At1g75390, -At1g43700, -At5g11260, -At1g42990, -At3g56850) at the site between the SMART III Sequence and CDS III Sequence, using the infusion method, as effectors. Each of these effectors was cotransformed with the reporter vectors pHIS2-BRS1, -Dof, -I-box, -G-box, -BS1, or -MY3 into yeast Y187 cells. The bindings of these eight bZIPs to the motifs harbored by pHIS2 were analyzed using the Y1H system. Statistical analyses Statistical analyses were carried out using SPSS 16.0 (SPSSInc, Chicago, IL, USA) software. Data were compared using Student’s t test. Differences were considered to be significant if P
