Cytokinin-overinduced transcription factors and thalianol cluster genes in CARBOXYL-TERMINAL DOMAIN PHOSPHATASE-LIKE 4-silenced Arabidopsis roots during de novo shoot organogenesis

ABSTRACT

Cytokinin (CK) is one of key phytohormones for de-differentiation and de novo organogenesis in plants. During the CK-mediated organogenesis not only genes in CK homeostasis, perception and signal transduction, but also factors regulating basic transcription, splicing and chromatin remodeling contribute to coordinate a sequence of events leading to formation of new organs. We have found that silencing of RNA polymerase II CTD-phosohatase-like 4 (CPL4_RNAi) in Arabidopsis induces CK-oversensitive de novo shoot organogenesis (DNSO) from roots, partly by early activation of transcription factors such as WUSCHEL and SHOOT MERISTEMLESS during pre-incubation on callus induction media. Here we show that a cluster of thalianol-biogenesis genes is highly expressed in the CPL4_RNAi during DNSO, implying involvement of CPL4 in transcriptional regulation of the thalianol pathway in DNSO.

Plant cells have remarkable plasticity in that they can de-differentiate (callus formation) and re-differentiate into a whole plant through de novo organogenesis of shoot and root. This regenerative capacity has been widely explored in the field of biotechnology and agriculture. De novo shoot/root organogenesis is orchestrated by an interplay between two major phytohormones auxin and cytokinin (CK).^1–3 Incubating excised plant tissue (explant) on media with auxin induces de novo root formation. Presence of CK together with high concentration of auxin (i.e. callus induction media, CIM) results in formation of callus through the same pathway for lateral root primordia development.⁴ Incubating explants on media with high CK/auxin ratio (i.e. shoot induction media, SIM) after pre-incubation on CIM triggers de novo shoot organogenesis (DNSO). Consequently, many genes involved in CK-sensing,^5,6 signaling,^6,7 and CK-responsive transcription factors,^8-12 have been implicated in the regulation of DNSO. In addition, genes involved in basic transcriptional and post-transcriptional regulation have been identified in callus formation and DNSO processes;^13,14 for example, TATA-binding protein complex associated factor 12B (TAF12B) as CYTOKININ HYPERSENSITIVE 1 (CKH1);¹⁵ CHD3-type SWItch/Sucrose Non-Fermentable (SWI/SNF) chromatin remodeling factor, chromatin remodeler PICKEL (PKL) as CYTOKININ HYPERSENSITIVE 2 (CKH2);¹⁶ a DEAH-box RNA helicase involved in pre-mRNA splicing as ROOT INITIATION DEFECTIVE 1 (RID1);^17,18 and a small nuclear RNA (snRNA) activator protein complex subunit SNAP50 homolog as SHOOT REDIFFERENTIATION DEFECTIVE 2 (SRD2).^19,20

RNA polymerase II CTD-phosphatase-like 4 (CPL4), a homolog of yeast TFIIF-interacting CTD-phosphatase FCP1, is an essential regulator of Pol II-CTD phosphoregulation in Arabidopsis. We have shown that CPL4 suppresses xenobiotic stress responsive genes²¹ and mediates snRNA transcription termination/3ʹ-end formation.²² In our recent report,²³ we showed that silencing of CPL4 induces CK-oversensitive DNSO from Arabidopsis roots. DNSO from CPL4 knock-down (CPL4_RNAi) root explants are more vigorous than wild type (WT) in standard SIM condition. Moreover, CPL4_RNAi root explants can show DNSO even with 1/10 concentration of CK (0.05 mg/L 2iP, low-CK) in SIM, while WT cannot. RNA-seq analysis on WT (W) and CPL4_RNAi (C4) root explants on SIM with (+) or without (-) low-CK at 3-day-post-transfer identified 2,838 genes significantly up-regulated (> 1.5-fold) in CPL4_RNAi on low-CK SIM [C4(+)] compared to WT on SIM without CK [W(-)] (referred to as C4(+)UT, C4(+)-Upregulated Transcripts). We further classified them into four classes based on expression pattern in each sample; class I genes require both CPL4_RNAi and low-CK to be up-regulated; class II genes can be up-regulated solely by CPL4_RNAi; class III genes are low-CK-responsive; class IV genes can be up-regulated either by CPL4_RNAi or low-CK. Transcription factors essential for DNSO, such as WUSCHEL (WUS), SHOOT MERISTEMLESS (STM) and CUP-SHAPED COTYLEDON 2 (CUC2), were all found in class II, indicating that these gene expression in CPL4_RNAi did not rely on CK in SIM. Class II genes were most abundant (1,594 out of 2,383) and enriched with gene ontologies related to DNA replication and cell cycle regulation, suggesting that CPL4_RNAi root explants after pre-incubation on CIM are highly proliferative even without CK in the SIM. RT-qPCR on root explants during pre-incubation on CIM confirmed that indeed WUS, STM and CUC2 are up-regulated by CIM-preincubation, presumably by oversensitivity to low-CK (0.05 mg/L kinetin) in the CIM.

During the course of the analysis, we found that number of genes were induced in C4(-) and in W(+), but induced to higher levels in the C4(+) condition. Because qualitative classification adapted in the above analysis does not highlight this behavior, we also evaluated quantitative behavior of each gene and identified 248 genes (61, 71, 28, 88 genes from class I, II, III and IV, respectively) expressed higher in C4(+) than any other conditions thus possibly contributing to low-CK DNSO. Among them, 26 are transcription factors (GO:0003700), and 14 of them, including ENHANCER OF SHOOT REGENERATION 1 (ESR1),⁸ are associated with a developmental process GO (Table 1; GO0037502).

Table 1.

Twenty-six transcription factors in the 248 genes highly expressed in low-CK CPL4_RNAi.

AGI	Symbol	Class	Gene description	FPKM
				W(-)	C4(-)	W(+)	C4(+)
DNA binding transcription factor activity (GO:003700); Developmental Process (GO:0032502)
AT5G65640	bHLH093	I-a	beta HLH protein 93	27.19	40.65	33.08	61.60
AT1G75390	bZIP44	IV	basic leucine-zipper 44	11.19	23.93	21.48	45.74
AT1G10480	ZFP5	IV	zinc finger protein 5	9.27	17.16	17.42	38.20
AT5G64060	NAC103	IV	NAC domain containing protein 103	1.24	14.53	3.32	23.98
AT2G21650	MEE3	IV	Homeodomain-like superfamily protein	1.16	3.93	3.95	23.09
AT4G37750	ANT	IV	Integrase-type DNA-binding superfamily protein	3.83	9.68	14.78	22.70
AT5G03790	HB51	II	homeobox 51	0.56	3.70	1.18	8.21
AT3G50060	MYB77	II	myb domain protein 77	1.42	3.24	2.22	5.45
AT3G12720	MYB67	IV	myb domain protein 67	1.88	2.99	3.14	4.88
AT1G12980	ESR1	III	Integrase-type DNA-binding superfamily protein	0.40	0.77	1.39	4.54
AT1G02250	NAC005	III	NAC domain containing protein 5	0.88	0.37	2.65	3.98
AT5G03680	PTL	IV	Duplicated homeodomain-like superfamily protein	0.50	2.01	1.09	3.60
AT1G01060	LHY	II	Homeodomain-like superfamily protein	1.01	2.24	0.74	3.44
AT4G17785	MYB39	I	myb domain protein 39	1.31	0.69	1.66	2.83
DNA binding transcription factor activity (GO:0003700)
AT5G25190	ESE3	IV	Integrase-type DNA-binding superfamily protein	12.76	41.18	28.78	77.40
AT1G14600	-	II	Homeodomain-like superfamily protein	3.02	15.34	4.37	23.96
AT3G07340	-	IV	basic helix-loop-helix (bHLH) DNA-binding superfamily protein	3.87	11.56	8.46	19.64
AT2G17770	BZIP27	II	basic region/leucine zipper motif 27	1.62	3.10	2.52	7.42
AT4G39410	WRKY13	II	WRKY DNA-binding protein 13	0.52	4.82	0.69	7.41
AT2G46130	WRKY43	II	WRKY DNA-binding protein 43	1.53	3.45	2.82	6.72
AT3G60390	HAT3	I	homeobox-leucine zipper protein 3	1.81	1.62	2.39	4.94
AT5G11060	KNAT4	II	homeobox protein knotted-1-like 4	0.85	1.91	1.19	3.50
AT1G80590	WRKY66	I	WRKY DNA-binding protein 66	1.54	1.38	2.10	3.46
AT2G47520	ERF71	II	Integrase-type DNA-binding superfamily protein	0.48	1.71	0.57	3.19
AT1G75240	HB33	III	homeobox protein 33	0.23	0.11	1.10	2.10
AT5G03720	HSFA3	II	heat shock transcription factor A3	0.45	0.94	0.50	1.67

W(-), wild type on SIM without 2iP; W(+), wild type on SIM with 0.05 mg/L 2iP; C4(-); CPL4_RNAi on SIM without 2iP; C4(+), CPL4_RNAi on SIM with 0.05 mg/L 2IP.

Co-expression analysis by ATTED-II identified only one co-expression network consisting of more than 10 genes. The network includes three genes from a gene cluster for synthesizing and metabolizing thalianol, a tricyclic triterpene compound synthesized from 2,3-oxidosqualene (Figure 1a). The genes are oxidosqualene cyclase thalianol synthase 1 (THAS1), cytochrome P450 thalianol Hydroxylase (THAH) and a BADH acyltransferase (ACT), which are the three of four thalianol synthesis/metabolism genes tandemly arranged in an operon-like organization on chromosome 5²⁴ (Figure 1b). The missing gene, thalian-diol desaturase (THAD), does not pass the new criteria but shows significantly higher expression in low-CK CPL4_RNAi over low-CK WT by 1.38-fold (Figure 1b). Because of the unique gene structure, we also checked expression behavior of 34 potential operon-like gene clusters involving 308 genes in total.²⁵ Only five genes passed the criteria described above; three are from the thalianol cluster, and two are each from independent clusters (AT1G63480 and AT3G16430). Therefore, the thalianol cluster is the sole operon-like cluster over-induced in CPL4_RNAi during low-CK DNSO. THAS1 catalyzes the conversion of oxidosqualene to thalianol; THAH is responsible for converting thalianol into thalian-diol, which can be desaturated by THAD.²⁴ Overaccumulation of thalianol or thalian-diol results in shoot dwarf phenotype.²⁴ The thalianol cluster genes have been identified as CK-responsive.^26-32 Additionally, ACT is directly bound and regulated by a chromatin remodeler PICKLE/CYTOKININ-HYPERSENSITIVE 2 (PKL/CKH2)³³ Mutations in PKL/CKH2 result in CK-hypersensitive greening of callus, indicating a potential interplay between chromatin regulation and CK-sensitivity.^14,16

Figure 1.

Thalianol biosynthesis-related co-expression network observed among the 248 genes. (a) A microarray-based (Ath-m) co-expression cluster comprised of more than 10 genes among the 248 genes highly expressed in CPL4_RNAi during DNSO, detected by NetworkDrawing tool in ATTED-II ver.9.2 (http://atted.jp/). Genes with a red asterisk show up in the same network based on RNA-seq (Ath-r). An associated class is indicated for each gene. The three genes from a thalianol-biosynthesis operon-like regulon on chromosome 5 are circled in red. The chemical structure of thalianol (C30H50O) is shown, drawn by PubChem Sketcher V2.4, based on the CID25229600. Full name of each gene is described in Abbreviations. (b) FPKM values of the four thalianol biosynthesis genes arranged in an operon-like manner on chromosome 5. C4(+)/W(+) fold change values and associated q-values from Cuffdiff are shown. (c) Expression level of each gene relative to GAPDH [2^{(Cp of GAPDH – Cp of target)}] was measured by RT-qPCR. Roots from 10-day-old seedlings (labeled as “R”) were subjected to CIM pre-incubation for 5 days, then moved to SIM containing 0.05 mg/L 2iP (“S”) for 3 days. Root samples right after the CIM pre-incubation were harvested and denoted as “C”. Bars represent standard error of means from biological triplicates. *p < 0.05, **p < 0.01 by one-tailed Student’s t-test between wild type (Col gl1) and each CPL4_RNAi line in each condition.

We also examined expression levels of the thalianol cluster genes before and during pre-incubation on CIM. In roots before treatments, expression levels of all four thalianol cluster genes were lower in CPL4_RNAi than in wild type (Figure 1c). During CIM pre-incubation, there was substantial decrease in ACT, THAD and THAH expression both in wild type and CPL4_RNAi, compared to intact roots. Expression levels of ACT and THAH during low-CK SIM incubation were higher in CPL4_RNAi than wild type (Figure 1c), confirming RNA-seq analysis data (Figure 1b). Between CIM pre-incubation and low-CK SIM, increase of ACT and THAH expression was observed in CPL4_RNAi but not in wild type (Figure 1c). Such CPL4_RNAi -specific re-activation of thalianol pathway genes might play a role in low-CK DNSO, which is also specific to CPL4_RNAi. Thalianol pathway intermediates differentially affect roots and shoot development.²⁴ High levels of thalianol in thah1 and thad1 mutants as well as in THAS overexpressing plants resulted in production of longer roots, whereas THAS overexpressing plants showed shoot dwarfism.²⁴ The shoot phenotype was further enhanced by co-expressing THAS and THAD, suggesting thalian-diol and/or desaturated thalian-diol are more active than thalianol.²⁴ Although no functional link between thalianol and DNSO has been made, the inhibitory effect of thalianol to shoot growth could be related to reduction of cell proliferation during DNSO, for example, decline in cycB expression in CPL4_RNAi after explants were transferred to low-CK-containing SIM medium.²³ Alternatively, thalianol may function as a negative feedback mechanism for DNSO. At molecular level, how CPL4 is involved in thalianol cluster regulation is still unknown, as CPL4_RNAi did not show notable transcription initiation or termination defects within the locus. PKL directly binds ACT³³ and with PICKLE RELATED 2 (PKR2) it positively regulates expression of ACT .³⁴ Also, PKL has been shown to associate with elongating Pol II co-purified with a transcription elongation factor SPT4.³⁵ It is tempting to speculate that CPL4 as a global Pol II-CTD phosphatase might regulate PKL recruitment to the thalianol cluster via modulating Pol II-CTD phosphorylation of elongating Pol II during DNSO. Taken together, these analyses identified the thalianol biosynthesis cluster as a potential regulator of low-CK DNSO in CPL4_RNAi. It will be an attractive target for further dissection of DNSO, as it might mediate interplay of CK, chromatin remodeling and Pol II-CTD phosphoregulation mediated by CPL4.

Funding Statement

This work was supported by the National Science Foundation [MCB0950459]; Texas A and M University [Genomics of Plant Water Use Seed Grant]; U.S. Department of Agriculture [2010-34402-20875]

Acknowledgments

The authors thank Dr. Charles D. Johnson and Dr. Richard Metz (Texas A&M AgriLife Genomics and Bioinformatics Service) for their support in the RNA-seq experiment. This work was supported by the National Science Foundation (MCB0950459, to H.K.), USDA-CSREES (2010-34402-20875 to H.K.) “Designing Food for Health”, and Genomics of Plant Water Use Seed Grant (Texas AgriLife Research to H.K.). A.F. was supported by the Student Exchange Support Program (Scholarship for long-term study abroad from Japan Student Services Organization) and Dissertation Fellowship from Office of Graduate and Professional Studies, Texas A&M University. The authors declare no conflict of interest.

Abbreviations

CPL4

RNA polymerase II CTD-phosphatase-like 4

CTD

carboxyl-terminal domain

CK

cytokinin

DNSO

de novo shoot organogenesis

SIM

shoot induction medium

CIM

callus induction medium

2iP

2-isopentenyladenine

GO

gene ontology

MLP34

major latex protein-like protein 34

HAD

haloacid dehalogenase

CIMT1

Cd-inducible methyltransferase 1

NAMT1

nicotinic acid methyltransferase 1

JAL22

jacalin-related lectin 22

DAO2

dioxygenase for auxin oxidation 2

SAMT-like

S-adenosyl-L-methionine-dependent methyltransferase-like

NPF5.12

nitrate transporter 1/peptide transporter family 5.12

DUF914

solute carrier family 35 protein

ROSY1

interactor of synaptotagmin-1

APY3

apyrase 3

cycB

cyclin B