A Study On Transcriptional Corepressors Biology Essay

Published: Last Edited:

This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.

Amie Blinkhorn The Interactions of TPL/TPR Transcriptional Corepressors


Transcription factors are proteins that regulate the transcription of genes by forming complex protein-protein interaction networks in order to control developmental processes. Transcription factors function by binding directly to specific DNA sequences on the promoters of target genes to either activate or repress gene expression. Much focus on activiators has led to repressors being largely unstudied and not understood. TPL/TPR are a small family of transcriptional corepressors in plants. They are implicated in meristem maintenance through interaction with WUS and auxin signalling via AUX/IAA interaction. In this study yeast two-hybrid techniques are used to identify other proteins that interact with TPL, TPR4 and TRP3, common domain structures that enable these interactions, and how the phylogenetic relationships of this family relates to the types of proteins with which they interact. Through this experimentation it was found that TPR3 interacts predominantly with certain families of transcription factors. Many of these interacting transcription factors are known repressors and have some variant of the known repressive EAR motif. Most of these interaction partners were also found in an unpublished yeast two- hybrid screen with TPL. This and previous work provides exciting insights into where TPL/TPRs are crucial, for example, in the repression mechanisms used by JAZ proteins in JA signalling and ARFs in auxin-independent pathways. No link was found in this study with phylogeny and interaction specificity. Overall the TPL/TPR family are thought to be general developmental repressors which are involved in many aspects of plant development.


Plants are sessile organisms; they do not have the capacity to move around as animals do in order to attain resources and to respond to changes in the environment, therefore they must use other strategies in order to survive. They have the capacity to develop continuously by growing new organs, such as leaves, flowers and roots, to solve this problem (Schmid et al, 2005). The formation of these new organs are the result of the activity of meristems where pluripotent stem cells produce undifferentiated cells which eventually expand and differentiate to produce the plants organs. As more of these stem cells are produced the older ones are displaced to the periphery of the meristem where they differentiate. To maintain the stability of the meristem its various actions i.e. cell proliferation and organ initiation, must be regulated (Mayer et al, 1998). Plants rely heavily on transcriptional regulation to control gene expression within these meristems and to control their activities when changes to the environment occur (Krogan and Long, 2009).

Transcription factors are proteins which regulate the transcription of genes (Qu and Zhu, 2006). They form complex networks at a transcription level and through protein-protein interactions in order to regulate and control developmental processes (Reichmann and Ratcliffe, 2000). They are capable of controlling gene expression through many pathways, including chromatin remodelling and through the activities of the RNA polymerase II transcription-initiation complex (Singh, 1998).

The Arabidopsis genome contains 1510 to 1581 transcription factors which are split into large families. However, only a small proportion of these genes have been molecularly or genetically characterised (Qu and Zhu, 2006). By examining the expression patterns of these gene families it has become clear that groups of related transcription factors are involved in specific developmental processes such as flower development and different combinations of these genes are active at different times and areas of development (Schmid et al, 2005).

Transcription factors function by binding directly to specific DNA sequences on the promoters of target genes and in doing so either activate or repress the expression of the gene. This gives rise to specific regulation of the gene (Singh, 1998). The activation or repressive activities of different transcription factors allow them to be characterised into their functions as either repressors or activators. Coactivators and corepressors mediate the activities of repressors and activators through specific protein-protein interactions with binding domains, allowing them to also be specific to a promoter without themselves binding to DNA. These interactions result in the formation of an Enhanceosome (Singh, 1998) or a Repressosome (Courey and Jia, 2001) which are protein complexes formed when multiple proteins interact in order to regulate a specific gene and these can either activate or repress respectively (Singh, 1998). The composition of these complexes may change as the result of environmental or developmental signals (Singh, 1998).

In the past the study of transcriptional regulation has focused on the activation of genes, however, repression is just as important. In fact mechanisms of gene regulation through activation are constantly being shown to be impeded by transcriptional repressors (Cowel, 1994) and so the old view that without activation the gene would be fixed and silent without any other cues is fast becoming redundant.

There are two types of repressors; active and passive. Active repressors inhibit initiation of transcription directly through binding to the promoter of the gene. In contrast to passive repressors, which downregulate the activity of other transcription factors (Cowel, 1994). Active repressors contain defined repression domains that are sequence-specific enabling them to bind to corepressors and other regulators (Krogan and Long, 2009). These could be chromatin remodelling factors which result in chromatin tightening and therefore repression of the gene (Krogan and Long, 2009).

The Activities of the WUSCHEL Transcription Factor

The homeodomain transcriptional regulator WUSCHEL (WUS) is required to maintain the undifferentiated state of stem cells in the shoot apical meristem (Ikeda et al, 2009) and the floral meristem (Laux et al, 1996) and is an excellent example of a gene which has both activation and repression capabilities. These functions are exemplified as it activates AGAMOUS (AG) and LEAFY (LFY) therefore terminating stem cell maintenance in floral meristems (Lenhard et al, 2001) and also represses the expression of the type-A ARABIDOPSIS RESPONSE REGULATOR (ARR) genes, which in turn, repress cytokinin signalling which is involved in cell proliferation (Leifried et al, 2005).

The size of the pool of stem cells which the WUS gene creates is controlled by a negative feedback loop involving the CLAVATA (CLV) genes (Leifried et al, 2005). WUS activates the expression of the small peptide CLV3 which, in some way, acts as a ligand for the CLV1-CLV2 heterodimer which represses WUS expression, therefore maintaining the stem cell number through circular signals (Leifried et al, 2005).

Leifired et al (2005) performed a comparative microarray screen using plants with ethanol-inducible overexpression alleles of WUS and after treatment with ethanol, levels of ARR5, ARR6, ARR7 and ARR15 mRNA were decreased. By transiently repressing WUS through inducing CLV3 Leifried et al (2005) showed that ARR levels increased, suggesting that ARR genes are repressed by WUS. In connection with this ARR5, ARR6, ARR7 and ARR15 promoters were found to be active in the meristem, consistent with their interaction with WUS in this area of the plant.

The former description of WUSCHEL, as an activator of AG, was the result of analysis of the Arabidopsis WUSCHEL mutant, wus-1 (Figure 1B). In the wus-1 mutant stem cells were misspecified and instead of remaining pluripotent experienced differentiation thus resulting in the collapse of the shoot apical meristem (SAM) and the floral meristem (Mayer et al, 1998). The root apical meristem, however, is unaffected (Laux et al 1996). Wildtype Arabidopsis flowers have four sepals, four petals, six stamens and two fused carpels (Figure 1A); however, the wus-1 mutant does not have most of the central organs and ends in a single stamen (Laux et al 1996). wus floral meristems and SAMs were found to terminate prematurely (Laux et al 1996). WUS was characterised as specifying stem cell fate and not as a repressor of organ formation, as cells were only misspecified and not incorporated into organs (Mayer et al, 1998). It was confirmed that the WUS protein is expressed in a small number of meristem cells forming an organising centre just below the stem cells and not in the stem cells themselves (Mayer et al, 1998).

Picture1.pngFigure 1 - A comparison between an Arabidopsis wildtype flower and wus mutant flower shows A) Wildtype flower with four whorls of organs; four sepals, four petals, six stamens and two fused carpels, B) Wus Mutant which contains organs in the two outer whorls but ends in only a single stamen. Scale bars: 500 µm. Modified from Mayer et al, 1998.

In contrast to shoot apical mersitems (SAMs) which are indeterminate, floral meristems are determinate which means that once the flower structure is complete stem cell maintenance is stopped (Lenhard. M et al, 2001). As both these meristems contain an organizing centre to maintain the stem cells, as the result of WUS expression, it cannot be this that distinguishes between them. Meristems also express organ identity genes APETALA1 (AP1) and LEAFY (LFY) which are present in floral meristems only and therefore ensure the difference between the floral and shoot apical meristems. The determinacy of the floral meristem is the result of a negative autoregulatory mechanism which requires the activation of AG (a MADS domain transcription factor) by WUS and the subsequent repression of WUS by AG. Therefore, once AG is activated by WUS, AG represses the expression of WUS thus preventing the maintenance of the stem cells and preventing further growth (Lenhard et al, 2001). This mechanism is further complicated in the mutant lfy where AG is not sufficiently activated by WUS and therefore termination of the meristem cannot take place. It appears that WUS requires the addition of LFY to activate AG expression. As LFY is only expressed in the floral meristem and not in SAM, SAM remains indeterminate in the presence of both AG and WUS (Lenhard et al, 2001). Similarly, AG does not have the capacity to repress WUS alone and therefore must require other transcription factors in the same way as WUS requires LFY (Lenhard et al, 2001). A recent paper by Sun et al 2009 showed that AG controls the transcription of the gene KNUCKLES (KNU) and that KNU aids AG in the repression of WUS.

The activation and repressive activities of WUS are due to its possession of various conserved sequence motifs at the C-terminal end of the protein (Kieffer et al, 2006) (Figure 2). Excluding the homeodomain the WUS protein contains: an acidic domain, a WUS box and an EAR-like domain, each of which have specific functions in either the repression or activation of other proteins. The acidic domain (LEGHGEEEECGGDA) is a common domain conserved in plant transcription factors and it is the only activation domain present on the WUS protein (Kieffer et al, 2006). The EAR motif (L/FDLNL/F(x)P or LXLXL) is a repression domain that is conserved in plants (Ohta et al, 2001), it is required for the repressive function of many repressors such as: the class II ethylene-repressive element binding factors (ERFs), TFIIA-type zinc repressors and the AUX/IAA proteins (Kieffer et al, 2006). As the WUS protein possesses an EAR-like motif (ASLELTLN), we are able to suggest that it has a repressive function within the WUS protein through the evidence that it is present in other transcription factors with repressive function. However, mutation within this EAR-like motif does not interfere with the repressive activity of WUS leading to the idea that the WUS box (TLPLFPMH) is also a repression domain (Ikeda et al, 2009). The WUS Box is present in all the WUSCHEL-RELATED HOMEBOX (WOX) proteins, except WOX 13. Some members of this WOX family have also been found to control aspects of plant development (Kieffer et al, 2006). The WUS box domain remains conserved in most of the WUS/WOX proteins therefore it is speculated that it is important in the function of the entire family of WUS/WOX proteins. It is considered more important in conferring the function of the WUS/WOX proteins than the EAR domain as the EAR domain is only present on the WUS protein.

As previously mentioned WUS activates AG. It also has repressive function displayed by its capability to repress the activity of the type-A ARABIDOPSIS RESPONSE REGULATOR (ARR) genes: ARR5, ARR6, ARR7 and ARR15. Ikeda et al (2009) concluded that this repression is the result of the WUS box, as levels of these ARR genes were increased when both the WUS box defective mutant and the EAR-like and WUS box defective mutants were analysed. These results point towards the EAR-like motif having no involvement with this reaction as the same increase in ARR protein is found in both the WUS box defective mutant and the EAR-like and WUS box defective mutants (Ikeda et al, 2009).

The Regulation of WUSCHEL

Although the mechanism for the prevention of stem cell differentiation by WUSCHEL is not yet completely understood the conserved domain structures: the WUS box, the EAR-like domain and the acidic domain have also been found to be required for the interaction between WUS and two members of a family of Arabidopsis corepressors, TOPLESS (TPL) and TOPLESS-RELATED (TPR) 4 (Previously named WISP1 and WISP2 respectively)(Kieffer et al, 2006). These protein corepressors were found in a yeast two-hybrid screen using WUS as the bait (Kieffer et al, 2006) and were found to be expressed throughout the development of the plant by microarray analysis (Schmid et al., 2005). Interactions between WUS protein and TPL or TPR4 only occurred when the three conserved WUS domains were present (Kieffer et al, 2006). The test for this involved only either all the domain structures present or where none were present. It was not tested which of the domains were needed for the WUS-TPL/TPR4 interactions so it may be possible that only one of these domains, such as the WUS box, is vital for the WUS-TPL/TPR4 interaction and not all of the domains (Kieffer et al, 2006). TPL and TPR4 contain N-terminal LisH and CTLH domains, a pro-rich region, and two domains containing WD repeat motifs (Kieffer et al, 2006). The LisH domain is a dimerization domain present in all plant TPL/TPRs and LEUNIG (LUG) family proteins (Liu and Karmerkar, 2008). WD repeats provide a structure where protein-protein interactions can take place (Smith et al., 1999); however, Kieffer et al (2006) showed that the LisH and CTLH are required for WUS interaction and not these WD repeats. Many other unrelated transcriptional corepressors, such as LUG and the human TBL1 and TBLR1, contain an N-terminal LisH and these WD repeats suggesting these are common motifs in transcriptional corepressors. Therefore, this contributes to the confirmation that TPL and TPR4 are also corepressors. Also, it is possible that as the WOX family has a similar domain structure to WUS it is comprehensible that the WOX family of proteins may interact with the TPL/TPR family of proteins. This would result in a network of interactions between these two families in order to control plant meristem maintenance. What's more, as the TPL/TPR family have been found to be present in many areas of the plant, interactions between the WUS/WOX family and the TPL/TPR family may have some control over plant development in general (Kieffer et al, 2006).

A model was proposed by Kieffer et al (2006) in which WUS recruits the TPL and TPR4 corepressors in order to provide repression of genes required for the differentiation of stem cells in SAM. In the floral meristem this WUS-TPL complex would, somehow, be disrupted and result in WUS switching from a repressor to an activator of AG. Leading on from this it is possible that this disruption could be caused by LEAFY and as LEAFY is only present in the floral meristem and not SAM this disruption would cause only floral meristem maintenance to be disrupted. The mechanism for repression of stem cell differentiation genes by WUS-TPL/TPR is unknown but it is speculated that it may occur as a result of histone deacetylation which is shown to be mediated by other transcriptional repressors such as the Drosophila protein GROUCHO (Courey and Jia, 2001).

The GROUCHO gene encodes a transcriptional corepressor which also contains WD tandem repeats and a tetramerization domain, but does not have a LisH domain as in the aforementioned corepressors (Courey and Jia, 2001). The similarity between the GROUCHO WD repeats and those in the GROUCHO yeast homolog, TUP1 has led to the grouping of these genes into the GROUCHO/TUP1 superfamily. The GROUCHO/TUP1 superfamily has been revealed to cause chromosomes to become silent through histone deacetylation and therefore cause genes in that area to become repressed. This occurs through the interaction of GROUCHO/TUP1 proteins with Histone Deacetylases (HDACs).

A nulceosome consists of eight histone subunits; these subunits are composed of a globular C-terminal domain and an N-terminal tail which projects outward. The deacetylation by HDACs of this tail causes the chromatin structure to become more compact, resulting in either transcriptional machinery becoming unable to transcribe the genes due to their inaccessibility or the action of deacetylation could cause the recruitment of chromatin remodelling factors. The actual mechanism through which this occurs is still unresolved (Courey and Jia, 2001). Both histone deacetylases (HDACs) and histone acetyltransferases (HATs) were found to interact with TPL by genetic approaches as the mutant hdac19 had tpl like phenotype (Long et al, 2006). Long et al (2006) have revealed that TPL interacts with the HISTONE DEACETYLASE 19 (HDA19). Mutations in the HDA19 gene increases the mutant tpl-1 phenotype. It is therefore, suggested that it works with TPL to regulate the repression of genes through the recruitment of HDA19.

The Topless-1 Phenotype

A mutation in the Arabidopsis TOPLESS gene, topless-1 (tpl-1), results in the transformation of the shoot pole into a second root pole (Figure 3E) (Long et al, 2006). The structure of the apical root in the extreme phenotype is similar to the wild-type root, although it has more cells and is wider. It also demonstrates normal gravitropism and has a root cap (Long et al, 2002).This phenotype is dramatic as the only mutants to be isolated, which affect specific patterning of the plant thus far, have not been able to switch one organ into another. In some of these mutants (eg monopteros, bodenlos and axr6) there are disturbances to the root of the plant but the overall polarity of the entire plan is not disrupted (Long et al, 2002). This suggests that the TOPLESS gene works at a higher level of transcriptional regulation than these other mutants (Long et al, 2006). This extreme phenotype only occurs at the restrictive temperature of 29°C at lower temperature other phenotypes of this mutant occur. At lower temperatures the plant has no hypocotyls, cotyledons or shoot apical meristem (Figure 3D) (Long et al, 2002). Other characteristics may arise where the plant has a cup-shaped cotyledon (Figure 3BC) where as others may have no cotyledon at all and only make hypocotyls. A common characteristic of the mutant seedlings is the nonexistent SAM, however, in some seedlings a SAM can be formed after germination resulting in a relatively normal fertile plant (Long et al, 2002).

Photographs demonstrating the various phenotypes of the topless-1 mutant A) Wildtype seedling B) mutant with a single cotyledon and no SAM C) mutant with cup-shaped cotyledon and no SAM D) Mutant with no cotyledon and no SAM E) tpl-1mutant with both apical (a) and basal (b) roots. Scale bars: 1 mm. Taken from Long et al, 2002.

In the extreme topless-1 phenotype the mutant embryo begins morphologically indistinguishable from the wildtype. However, at the heart shaped stage they become more oblong in shape and do not have cotyledon primordia. In the apical half of the tpl-1 embryo there is reduced expression of genes normally expressed in that area and ultimately genes expressed in the basal half are expanded into the apical half resulting in the change in axis formation (Long et al, 2006). The suggestion that to begin with the axis formation of the embryo is correct and then is lost during the transition stage is shown by the original expression of WUS in the apical half of the embryo and then its loss (Long et al, 2006).

Therefore, Long et al (2006) suggest a model where, at the transition stage of embryogenesis, TPL and other TPR proteins repress this expansion of basal gene expression in the apical half of the embryo and that this repression involves the histone deacetylase, HDA19.

TOPLESS and the regulation Auxin

It is commonly recognised that Auxin and Cytokinin are associated with plant growth and development. In relation to this a treatment of high auxin to cytokinin levels given to plants results in encouraged root development, while high cytokinin to auxin levels results in encouraged shoot development (Skoog and Miller,1957).

As previously mentioned there are mutants in addition to tpl-1 which cause disruption to the root. One of these, the Monopteros mutant, does not develop roots or hypocotyls during embryogenesis. The MONOPTEROS gene encodes a member of the ARF (Auxin Response Factors) family of transcription factors which transduce auxin signals with the purpose of regulating root development (Long et al, 2002). ARF Transcription factors possess a DNA-binding domain (DBD) which is used to bind to TGTCTC auxin response elements within the promoter of target genes (Guilfoyle and Hagen, 2007).

ARFs have either activation or repression domains which confer its capacity to regulate. Of the 22 known ARFs which encode proteins there are five known to activate transcription (ARF 5-8 and 19), while the others are all repressors (Guilfoyle and Hagen, 2007). Repressive ARF transcription factors dimerize with the AUX/IAA (Auxin/Indole acetic acid) corepressors using a dimerization domain in order to regulate the transcription of auxin response genes (Figure 4). Without this dimerisation with AUX/IAA proteins repressive ARF proteins are unable to regulate genes in response to auxin (Guilfoyle and Hagen, 2007). The monopteros-7 encodes a mutant which is a truncated version of the wildtype protein but it does not contain this domain and it therefore exhibits the monopteros phenotype (Long et al, 2002).

AUX/IAAs are degraded in the presence of auxin and therefore auxin response genes seem to be regulated ultimately by the levels of auxin within the plant. AUX/IAA repressors bind to ARF activators which exist on the promoters of target genes resulting in the repression of the target gene. In the presence of high auxin levels AUX/IAAs are degraded and are removed from repressing the ARF activator, therefore causing the target genes to be activated (Figure 4) (Guilfoyle and Hagen, 2007). AUX/IAA proteins contain the EAR-motif (LXLXL) which is required for the binding of TPL corepressors. A direct interaction between the CTLH domain of TPL and the EAR domain of IAA12 suggests that TPL has some involvement in the AUX/IAA interaction with ARFs and that they use their corepressive function with AUX/IAA to repress ARF activators thus repressing target genes (Figure 4) (Szemenyei, H. et al, 2008).

Schematic diagram depicting the mechanism of repressing the auxin target gene through AUX/IAA corepression shows at low auxin levels AUX/IAA corepressors is free to repress the ARF bound to the target gene promoter with the aid of TPL. At high auxin levels the TPL leaves the AUX/IAA which is degraded and the ARF is then able to activate the transcription of the target gene. Taken from Causier (unpublished).

There are also known repressive ARFs, in a yeast two-hybrid screen with TPL as bait, TPL was found to interact with these ARFs without the need for an AUX/IAA intermediate (Causier, Unpublished).

The TOPLESS Family of genes

As previously mentioned the TOPLESS (TPL) gene is a member a family of TOPLESS-RELATED (TPR) genes which have considerable amino acid similarity with itself (Long, J.A et al, 2006). The four TOPLESS-RELATED genes have similar corepressive function as TPL this is because tpl-1 acts as a type of dominant negative allele for multiple TPR family members (Long et al, 2006). The TPR family members are thought to also be involved with WUS/WOX genes and ARF genes.

The phylogenic relationships between the TPL/TPR family members was then drawn through genetic analysis and it was found that TPR1 was the most related to TPL. The second most related TPR to TPL is TPR4. Which is hardly surprising as TPR4 was the first TPR to be pulled out of the WUS screen along with TPL (Kieffer et al, 2006). The two least related family members are TPR2 and TPR3 (Figure 5).

Aims and Objectives

This papers primary aim was to identify the interaction partners of TPR3 and compare these with the previously established interaction partners of TPL. Firstly, TPR3 was directly tested for its ability to interact with previously identified TPL interactors using a yeast two-hybrid mating matrix. In addition to this the ARF family and WUS/WOX family were used to compare interaction specificities of TPL and TPR3 using a similar yeast two-hybrid mating matrices. The ARF matrix was required to find a mechanism by which the TPL/TPRs are involved in auxin signalling pathways.

A yeast-two hybrid library screen was performed using TPR3 as bait in order to see the full range of potential interactions of TPR3. This experiment worked to recognise the different numbers and types of interactors which work with TPR3 in contrast to TPL. Ultimately a comparison of the interaction specificities between TPL and TPR3, two TPL - class corepressors at opposite ends of the phylogenic tree, was established.

Method and Materials

1. Amplification of TPR3 and TPR4 using Gateway™ technology att B PCR reactions

att B PCR primers were designed to amplify the coding region of TPR3 and TPR4, according to the Gateway Technology instructions (Invitrogen). PCR reactions (50µl) contained inflorescence first-strand cDNA, 200µM each dNTP, 0.2µM of appropriate forward and reverse attB PCR primers, 1 unit Phusion DNA polymerase (Finnzymes) and 1x Phusion HF buffer. PCR reactions were run using the following conditions: 98ºC / 1 min; 35 cycles of 98 ºC / 20 secs, 58 ºC/ 20 secs, 72 ºC / 2 mins; and a final step of 72 ºC / 5 mins. Following PCR, products were checked by agarose gel electrophorisis and purified using the QIA quick PCR Purification Kit, according to the manufacturer's instructions (Qiagen).

2. Gateway™ recombination cloning of attB PCR products into pGBKT7

Cloning of attB TPR3 and TPR4 PCR products into pGBKT7 was performed using the ‘one-tube' protocol according to the manufacturer's instructions (Invitrogen). Essentially, a BP reaction, consisting of PCR product, pDONR207 plasmid and BP Clonase II enzyme mix, was incubated at 25ºC for 4 hours. To this pGBKT7, LR Clonase II enzyme mix and NaCl (to a final concentration of 25mM) were added, and the reaction allowed to proceed at 25ºC overnight. Following proteinase K treatment, reaction products were transformed into E. coli and transformants selected on LB plated containing 50µg/ml Kanamycin (see Section11).

3. Transformation of pGBKT7 TPR3 and pGBKT7 TPR4 into yeast strains AH109 and Y187

Approximately 25µl AH109 and Y187 yeast cells were resuspended and washed in 1ml sterile distilled water and pelleted for 15 seconds at 15,000 rpm in a microcentrifuge. The cells were resuspended in 1ml of 0.1M LiAc, incubated at room temperature for 5 minutes and pelletted as above. To the pellet the following was added in order; 240µl 50% PEG, 36µl 1M LiAC, 50µl ssDNA (2mg/ml), 5µl plasmid miniprep and 29µl sterile distilled water. This mixture was vortexed and incubated at 42 ºC for 1 hour. The cells were pelleted and resuspended in 100µl sterile water and plated on SD -W plates (Section 13), and incubated for 3 days at 30 ºC.

4. Autoactivation test for AH109 BD -TPR3 yeast

AH109 BD-TPR3 yeast colonies were resuspended in 100µl sterile distilled water and 5 ml aliquots were spotted onto a SD -W plate (which provided the control) and onto separate SD-WH plates with 0mM, 2.5mM and 5mM concentrations of 3-AT, in triplicate. No yeast growth at 2.5mM 3-AT indicated that this was a suitable 3-AT concentration for the yeast two-hybrid library screen (Section 13) (Appendix 1).

5. Growth of bait strain for yeast two-hybrid library screening

A single colony of appropriate yeast bait strain was inoculated into a 50ml culture containing: 5ml 10x -W dropout, 2.5ml 40% glucose and 42.5 ml YSD broth (Section 13) and incubated at 28ºC at 200 rpm overnight. Yeast cell numbers were measured using a haemocytometer.

The 2.5x109 total number of cells was found which is higher than 1x109 as suggested by the manufacturer (Matchmaker 3 from Clontech) this constitutes that the TPR3-BD fusion was not toxic and that enough cells were available for mating.

6. Mating the library host strain with the bait strain

Library matings included 2 x 107 yeast cells (strain AH109) for the whole plant Arabidopsis library in pGADT7Rec and 2.5x109 bait cells (strain Y187 containing pBD-TPR3) in 50ml 2xYPAD (see Section 13). Matings were allowed to proceed overnight at 30ºC/50 rpm.

7. Select for yeast diploids expressing interacting proteins

The mating efficiency was determined by spreading 100µl of a 1:10,000, 1:1,000, 1:100 and 1:10 dilution of the mating mixture on the following three media (100mm plates) : SD/-L, SD/-W and SD/-WL. These were then incubated at 30ºC for 5 days. The remainder of the mixture was spread onto 25 15 cm plates with SD -WHLA + 3-AT medium, 400µl on each plate.

8. Yeast plasmid rescue

3ml of yeast overnight cultures were centrifuged at 14000rpm for 1 minute. The pellet was resuspended in 200µl lysis buffer (2% triton X-100, 1% SDS, 100mM NaCl, 10mM Tris pH8, 1 mM EDTA), 0.3 g of glass beads and 200µl chloroform were added. The tube was vortexed for 2mins and spun for 5 minutes. Plasmid DNA was recovered from 200µl of the supernatant by ethanol precipitation and resuspended in 20µl SDW. 5µl of plasmid solution was transformed into E.coli according to section 11, purified (Section 12) and analysed by diagnostic restriction enzyme digests (EcoRI and BamHI) and gel electrophoresis. Successful rescues were sent for DNA sequencing.

9. Cloning of ARF genes into pGAD424

Forward and reverse primers were designed (Appendix 3) to amplify the coding region of the following ARFs; ARF1, ARF2, ARF3, ARF7, ARF9, ARF16, ARF17 and ARF23. PCR reactions were performed essentially as described in Section 1. PCR products and vector (pGAD424) were digested with appropriate restriction enzymes (Appendix 2), the vector dephosphorylated and appropriate PCR-vector ligation reactions performed essentially as described (Sambrook & Russell, 2001). Ligation reactions were transformed into E.coli cells as Section11, purified as Section 12, were analysed by diagnostic restriction enzyme digests and gel electrophoresis. Correctly cloned ARFs were transformed into Y187 yeast as described in Section 3.

10. Matrix matings

All combinations of TPL/TPR3/TPR4 and AD-Proteins: TPL known interactors JAZ3, JAZ,6, JAZ10, ERF3, ERF4, ZAT10,LBD41, LBD37, IAA16, IAA2, IAA7, MYB32, TEM1, ARF2, ARF9 and unknown; ARF1, ARF9, ARF16 and ARF17 and WUS, WOX1, WOX4 and WOX5 were mated on YPAD plates. Each mating was patched across to SD-WL plates to recover the diploids. Each of the diploids were then patched across to SD-WHLA + 3-AT and SD-WL XαGAL plates in order to analyse the interactions between the proteins.

11. E.coli transformation

25 µl of Bioline Gold efficiency α-select chemically competent E.coli cells were thawed on ice. DNA solution was added to the thawed cells and left for 30 minutes, incubated for 42ºC for 45 seconds and left on ice for a further 2 minutes. 475µl LB was added to each tube and incubated for 1 hour at 37 ºC. Cells were pelleted, resuspended in 100µl LB and plated out on LB plates containing the appropriate antibiotic, and incubated at 37 ºC overnight.

12. Plasmid DNA Mini-Preps

Single E.coli colonies were inoculated into 5ml of LB containing the appropriate antibiotic and incubated at 37 ºC at 200 rpm overnight. 1.5ml of culture was pelleted and plasmids purified using the QIA prep Spin Miniprep Kit, according to the manufacturer's instructions (Qiagen).

13. Media

SD medium plates were made up by pouring plates from the following 1L mixture containing: 6.7g of yeast nitrogen base without amino acids, 20g Agar, 850ml water, 100ml amino acid drop-out solution (either -W, -L, -WL, -WLHA according to requirement) and 2.5mM 3AT or X-Gal (According to requirement), according to manufacturer's instructions (Clontech).

YPAD medium plates were made up by pouring plates from the following 1L mixture containing: 950 ml sterile water, 20 g/L Difco peptone, 10 g/L Yeast extract, 20 g/L Agar and add 15 ml of a 0.2% Alaninehemisulfate solution, according to manufacturer's instructions (Clontech).

LB - 10g/l tryptone, 10g/l NaCl, 5g/l yeast extract (= 15g/l agar for solid media).


The Interactions between the known TPL interactors and TPR3 and TPR4

In a previous library screen with TPL as the bait (Causier, Unpublished) many different types of transcription factors were pulled out. This experiment questions whether the same transcription factors also interact with other TPRs. Here, a matrix of transcription factors known to interact with TPL are mated with TPL, TPR3 and TPR4 in order to test this hypothesis; this was performed using yeast two-hybrid techniques. Two assays were performed, one of which involved the growth of the diploid yeast analysed through the yeasts ability to grow in a medium without vital amino acids: Tryptophan (W), Leucine (L), Histidine (H) and Alanine(A). Production of the nutritional reporter genes HIS3 and ADE1 produced Histidine and Alaninewhen an interaction between the bait and prey was made. The production of these vital amino acids therefore allowed the yeast to grow when a positive interaction was made in the presence of both plasmids (Tryptophan (W), Leucine (L)). The other assay required the diploid yeast to make the protein α-galactosidase, when the gene MEL1 was present, which would cause the yeast to turn a blue colour in the presence of the substrate X-α-gal.

TPL interacts with most of the proteins pulled from its library screen (Figure 6A). Surprisingly, it is shown to not interact with JAZ3 or JAZ10 in the histadine and Alanineassay (-WLHA assay) or JAZ6 or JAZ10 in the MEL1 assay (-WL+XαGal assay). As all three of these proteins were pulled from the TPL screen (Causier, Unpublished) one would expect a positive interaction between TPL and these proteins.

TPR3 was found to interact with most of the same interactors as TPL (Figure 6B), however positive weak interactions were also found in the MEL1 assay for JAZ6 and JAZ10 (Colour intensity 1). TPL interactors also found not to interact with TPR3 were ERF4 in the MEL1 assay and JAZ3, JAZ6, JAZ10, TEM1, ARF2 and Unknown in the -WLHA assay (Figure 6B). There is some difference in the intensity of interactions found for TPR3 in both assays. In the -WLHA assay weaker interactions were found for all in comparison to TPL with the exception of MYB32 which seems to interact with the same strength as the TPL interaction. In the MEL1 assay the only interactions found to the same strength as TPL were; ERF3, LBD37, IAA2, MYB32, TEM1 and ARF9. All these showed the same colour intensity of 5 as the TPL-TPL known interactor interations. All the other interactions of TPR3 shown in the MEL1 assay were weaker than those of TPL, with the exception of IAA7 which shows a colour intensity of 4 in TPR3 and a weaker interaction of 3 in TPL.

TPR4 interacts with very few of the same interactors of TPL and TPR3 (Figure 6C). In the -WLHA assay TPR4 is shown only to interact with LBD37, TEM1 and ARF9 and in the MEL1 assay JAZ6, LBD37, TEM1, Unknown and ARF9. In all three TPL/TPR family members JAZ3 is shown to interact with the same intensity in the MEL1 assay and not at all in the -WLHA. LBD37 interacts with all threes TPL/TPRs with the same intensity (Colour intensity 5) in the MEL1 assay and in varying intensities in the -WHLA assay. TEM1, Unknown and ARF9 interact with all three of the TPL/TPRs with varying intensities across the family and between the two different assays.

The X axis for Graphs A, B and C show a qualitative measure of colour intensity from 1- 5 of yeast grown in the MEL1 assay (-WL+XαGal ), the higher the number the stronger the interaction. Pictures A, B and C show the amount of yeast grown in the Histidine and Alanineassay (-WLHA), the higher the amount of yeast the stronger the interaction. Picture A and Graph A show that TPL interacts with most of the interaction partners tested, with the exception of JAZ3 and JAZ10 in the -WLHA assay and JAZ6 and JAZ10 in the MEL1 assay. There are weaker interactions in the MEL1 assay from JAZ3 (Colour Intensity - 2) and IAA7 (Colour Intensity - 3). Picture B and Graph B show that TPR3, also, interacts with most of the interaction partners, with the exception of JAZ3, JAZ6, JAZ10, TEM1, ARF2 and Unknown in the -WLHA assay and ERF4 only in the MEL1 assay. There is a range of strengths of interactions between the interactors in both assays in TPR3. Picture C and Graph C shows that TPR4 interacts with very few of the TPL interactors. TPR4 is show to interact with JAZ3, LBD37, TEM1, Unknown and ARF9 in the MEL1 assay and LBD37, TEM1 and ARF9 in the -WLHA assay.

In the growth assay of this matrix TPL is shown to interact with its known interactors with more intensity than in both TPR3 and TPR4, with the exception of JAZ3 and JAZ10. This is shown as there are much more yeast growing for the TPL positive interactions than for the TPRs positive interaction in the -WLHA assay. This indicates that the TPL interaction partners interact more readily with TPL than either TPR3 or TPR4. This, perhaps, is not surprising as all these interactors were pulled out of the TPL screen (Causier, Unpublished). In the TPR3 positive interactions yeast does not seem to grow as much as in the TPL interactions, this could be due to the phylogenic relationship between TPL and TPR3 (Figure 5).

In general this mating matrix has shown that TPR3 has similar interaction partners to TPL and TPR4 does not interact as readily with the protein transcription factors of TPL. This is surprising as the phylogenic relationships between the TPL/TPR family members would suggest the opposite to be true (Figure 5).

The Interactions between members of the WOX family and the TPL/TPR Family Members

A further assay which analysed the interactions between WUS/WOX family members and TPL, TPR4 and TPR3 allowed the additional understanding into how many and what types of proteins are capable of interacting with TPR3. As the TPL/TPR family is very small it is presumed that TPR3 would have some similar interactors to TPL (Figure 5). Here it is questioned whether these differences are due to the phylogenic relationships between the interactors. As TPL and TPR4 are closely related phylegenically one could assume that they have more similar WUS/WOX interaction partners than TPR3 . To prove this an experiment was performed in the same way as above, through the production of a matrix. The same two -WLHA and MEL1 assays were performed to result in the matrix shown in Figure 7 .

Figure 7- The Interaction matrix displaying the protein-protein interactions between the WUS/WOX family members (prey) and TPL, TPR4 and TPR3 (bait) The X axis for Graphs A, B and C show a qualitative measure of colour intensity from 1- 5 of yeast grown in the MEL1 assay (-WL+XαGal ), the higher the number the stronger the interaction. Pictures A, B and C show the amount of yeast grown in the Histidine and Alanine assay (-WLHA), the higher the amount of yeast the stronger the interaction. Picture A and Graph A show that TPL interacts with WOX1, WOX4 and WUS in both the MEL1 assay and the -WLHA assay. Picture B and Graph B show that TPR3 interacts with WOX4 and WUS in the MEL1 assay and WOX1 and WUS in the -WLHA assay. Picture C and Graph C shows that TPR4 interacts with the same interactors as TPL, WOX1, WOX4 and WUS in both the MEL1 assay and the -WLHA assay.

In this WUS/WOX interaction matrix TPL is shown to interact with all tested WUS/WOX proteins with the exception of WOX5. WOX1, WOX4 and WUS are all shown to interact with the same intensity (Colour intensity - 4) in the MEL1 assay and similar amounts of growth are seen for these proteins in the -WLHA assay. WOX5 does not show any colouration and minimal, if any, growth (Figure 7A).

TPR3 is shown to interact with WOX4 and WUS in the MEL1 assay at a colour intensity of 3 and in the -WHLA assay growth can be seen for WOX1, WOX4 and WUS. This shows that in the -WHLA assay TPL and TPR3 share the same WUS/WOX interactions.

TPR4 is shown in this matrix to have all the same interaction specificities as TPL including the levels of interaction intensities, as predicted.

None of the TPL/TPR family were found to interact with WOX5 in the MEL1 assay. The non-interaction of WOX5 and the TPL/TPR family members is in contrast to a previous matrix (Table 2) where WOX5 interacted with TPR4 (and TPR1). However, this work used the TPL/TPR family as the prey and not the bait, as in this assay, and therefore this could have an effect on the outcome. In this matrix all tested TPL/TPR family members were found to interact with WUS and this, again, is in contrast with the previous matrix (Table 2) where TPR3 (and TPR2) did not interact. Therefore, again, raising the question as to whether in general the orientation of binding domain and activation domain within yeast two-hybrid experiments affects the interaction specificities of the interaction partners. In both assays here WOX4 interacts with TPL, TPR3 and TPR4 but it was not found to interact with any TPL/TPR family members in the previous WUS/WOX matrix (Clerici, Unpublished). Finally, WOX1 interacts with TPL and TPR4 in the MEL1 assay but some growth can be seen corresponding to TPR3 thus questioning whether it also interacts. In the previous WUS/WOX matrix (Table 2) WOX1 was not found to interact with any of the TPL/TPR family.












Table 2 - The Interactions between the TPL/TPR family members (Activation Domain) and the WUS/WOX family members (Binding Domain) shows through a α-galactosidase assay no WUS/WOX family members interact with TPR3 or TPR2. WUS, WOX2 and WOX4 interact with TPL, TPR1 and TPR4. TPR1 and TPR4 interact with WOX5. ALL other WUS/WOX family members were tested and were not found to interact with any of the TPL/TPR family. Taken and Modified from Clerici, Unpublished.

Over all TPL and TPR4 were found to interact with the same WUS/WOX family members in this experiment. Despite previous works, this shows that the phylogenic relationships between these corepressors can have some effects on their interaction partners.

The TPR3 Library Screen


Number found

TPL Interactor

EAR-Like Domain (LXLXL)








































LXL (2)









False Positives







Table 3 - Positive results from the TPR3 library screen shows the genes encoding proteins which interact with the bait TPR3. It shows the number of times these proteins made positive interactions therefore alluding to how much the proteins interact naturally. This table also indicates whether these proteins interact with TPL.

To further analyse all proteins which interact with TPR3 an exhaustive library screen was performed with TPR3 as bait. By doing this proteins which do not interact with TPL but do interact with TPR3 could be found. As well as this proteins which were pulled out of the previous TPL library screen but not tested in the matrix. This would, therefore, allow a comprehensive understanding into proteins which interact with TPR3 and a comparison could be made between these proteins and proteins which interact with TPL. Identification of the similarities and differences between TPL and TPR3 interactors could be established as originally it was thought that TPR3 was more different to TPL than TPR4. The library screen tested 2x107 diploids in total and 200 positives were found on -WLHA + 3A.T 2.5. Of these positives, 57 gene sequences were processed and analysed (Table 3). Several false positives were found; 8 HSF4 heat shock proteins were identified and one copy of the RRM. This is unsurprising as the HSF4 heat shock protein can be found in all library screens this is due to its ‘sticky' nature. One copy of the RRM protein was pulled out which was in the wrong orientation, this is also a known to occur in these yeast two-hybrid experiments.

In this screen with TPR3 many similar transcription factors were pulled out as the previous TPL yeast two-hybrid screen. 141 positives were sequences as a whole in the TPL screen in contrast to the 57 sequenced so far in the TPR3 screen. The LBD family in the TPL screen was a large part of the whole TPL screen as 25% of all positives sequenced were from that family (Causier, Unpublished). In this case an even larger proportion (40%) of all positives sequenced were from the LBD family. In connection with this, LBD37 and LBD41 were the main LBDs pulled from the TPL screen, which is the same in this case. However, in the TPL screen LBD37 was pulled out only 5 times, whilst here it was pulled out 15 times. Conversely, LBD41 was pulled out in this screen 8 times and 19 times in the TPL screen. This questions whether the relationships between LBD41 and TPL, and LBD37 and TPR3 are due to some difference in the protein structure and/or the EAR domain, causing preferences to be made for LBD37 for TPR3 and LBD41 for TPL.

In the TPL screen the MYB family of transcription factors was 12% of all positives sequenced. Of these MYB32 was pulled out twice. In this screen MYB32 was pulled out 4 times in 57 positives sequences out of 200. Here, EPR1 was pulled out once; this is the same as in the TPL screen. In the TPL screen many more different MYB family members were pulled out however, in this screen, MYB44 was found. MYB44 was not found in the TPL screen.

The ZAT family transcription factors were another large proportion of the TPL screen (14%). In the TPR3 screen only ZAT10 was pulled out. Although it was positive 8 times in the screen suggesting a high interaction specificity with TPR3.

In the TPL screen JAZ5 was found once along with many other JAZ family members. In the TPR3 screen JAZ5 was positive twice although no other JAZ proteins were found in the 57 sequenced genes.

Another similar family of transcription factors, the ARF family of proteins, were found in a small proportion in the TPL screen. The members found were ARF2 and ARF9 and were found in the TPL screen 1 and 3 times respectively. In this TPR3 screen they were found once each. Although in the TPL screen these were not found until late on in the screen the ARFs found here were found in the first 20 proteins sequenced, suggesting that more are to be found in the other 143 positives in the TPR3 screen.

The E3 ligase ATARI1 was a positive in both screens. This protein is not in a large family of transcription factors.

In this, TPR3 screen, there were two proteins which were identified and not identified in the TPL screen. These were HAP8 and RAP2/TOE1 which were found once and 3 times respectively. Like ATARI1 neither of these proteins are in a large group of proteins already known to be associated with this family of transcriptional co repressors. However, they were pulled out early on in the screen suggesting that their interaction withTPR3 is relatively strong and therefore, may offer a unique dimension of functions for TPR3 only. Whether these interactions also occur in TPL and the other members of the family could be tested through a matrix similar to those represented in these experiments. However, it is possible that these proteins contain some specificity required for interaction in TPR3 only and none of the other TPL/TPRs. This would enable some understanding as to why the TPL/TPRs interact with different proteins.

The Interactions between members of the ARF family and the TPL/TPR Family Members

As ARF2 and ARF9 were found in both screens it was interesting to test repressive ARFs with TPL, TPR3 and TPR4. This would enable further understanding of the contrasting interactions of the TPL/TPR family. TPR3 was found to interact with none of the ARF family members in the MEL1 assay or the growth assay (Figure 8). This is in contrast to TPL which was found to be associated the ARF1, ARF16 and ARF17 in both the assays (Figure 8A). It was not, however, found to associate with ARF9 which was surprising as it was pulled out in both a previous yeast 2-hybrid screen with TPL (Causier, Unpublished) and the library screen with TPR3 as the bait. TPR4 interacts with ARF16 and ARF17 in both assays and appears to interact with ARF1 in the -WLHA assay (Figure 8B). This would the place TPL and TPR4 with the same ARF interactors, which would not be surprising due to their close phylogenic relationship.

Figure 8 - The Interaction matrix displaying the protein-protein interactions between the ARF family members (prey) and TPL, TPR4 and TPR3 (bait) The X axis for Graphs A, B and C show a qualitative measure of colour intensity from 1- 5 of yeast grown in the MEL1 assay (-WL+XαGal ), the higher the number the stronger the interaction. Pictures A, B and C show the amount of yeast grown in the Histadine and Alanine assay (-WLHA), the higher the amount of yeast the stronger the interaction. Picture A and Graph A show that TPL interacts with ARF1, ARF16 and ARF17 in both the MEL1 assay and the -WLHA assay. Picture B and Graph B show that TPR3 interacts with no TPL interactors in either the MEL1 assay or the -WLHA assay. Picture C and Graph C shows that TPR4 interacts with ARF16 and ARF17 in the MEL1 assay and ARF1, ARF16 and ARF17 in the -WLHA assay.

The EAR-Like Motifs of the TPR3 Interactors

As predicted all the positive interactors found from the TPR3 screen contained an LXLXL EAR-motif or a variation of it (Table 3). Both the LBDs found; LBD37 and LBD41 have the basic EAR domain LDLSL and LDLTLRL respectively. As it seems that TPR3 favours interaction with LBD37 over LBD41 in contrast to TPL it could be possible that it is something to do with the specificities of the EAR-motif.

In the retrieved ATARI1 and RAP2/ TOE 1 sequences there are no LXLXL motifs as the sequences are too short. However, there are two LXLXL motifs on the actual protein (LKL / LQLGI) in ATARI1 and a JAZ-like EAR-motif on the RAP2/TOE1 (IDLNL) (Table 3). As both the sequences are slightly shorter than in the retrieved fragment found on the gel it is possible that these motifs could be on their corresponding fragments but the sequence was not long enough to complete the whole protein. To rectify this, one would need to design primers which would sequence from the other side of the protein to be confident that these motifs were actually on the retrieved fragment and that these are, in fact, what is required for TPR3 interaction specificity.

The HAP8 protein contains two separate LXL motifs which seem to provide enough of an interaction with TPR3 (Table 3). All of the positive interactors of TPR3 found here, with the exception of ZAT10 and ARF9 contain at least one LXL within this motif, which suggests that it could be only this which is vital for an interaction with TPR3. This impression is strengthened by EPR1 as it only has an LSL domain. Therefore, suggesting that there could be another requirement for TPL/TPR interaction.


In this study the protein interactors of TPR3 were identified using yeast two-hybrid techniques. Yeast two-hybrid is deemed a reputable tool in the mapping of protein-protein interactions on a global scale and places uncharacterised and characterised genes in a functional context (Causier and Davies, 2002). In this case the yeast two-hybrid technique is used to understand the protein-protein interaction of the TPL/TPR family and ultimately will enable the understanding of the genetic networks of plant development as a whole.

In a previous yeast two-hybrid screen performed with TPL as the bait a number of different transcription factors were identified (Causier, Unpublished). Many of these transcription factors are known transcriptional repressors and have a range of different functions, for example, JAZ proteins were identified which have known functions to regulate the Jasmonate pathway (Staswick, 2008) and in contrast LBD proteins were also identified in the TPL screen have completely different functions e.g. LBD37 is involved in anthocyanin resistance pathways (Rubin. G et al, 2009). Therefore, it is suggested that TPL is a general transcriptional corepressor which works in varied developmental pathways within the plant. Here, we provide evidence that the closely related TPRs also have this role general and varied role.

The EAR Domain and the Interactors of TPR3

The interactors of TPR3 were found to have the common EAR motif (L/F DLN L/F(x)P). This motif was first recognised in three class II ERFs; NtERF3, AtERF3 and AtERF4 (Ohta et al , 2001), which have been found to be involved in plant growth and development and are responsive to signals induced by extracellular stress. These ERFs were also found to be active transcriptional repressors and this repression was due to the essential activity of the EAR domain, as mutations within this motif prevented the repressive capacity of the ERFs. The EAR motif has been found in a number of Zinc finger proteins, the AUX/IAAs and, now, in the interactors of the TPL/TPR family. It is suggested that this EAR motif confers the repressive activity of these proteins in a number of different functions within the plant (Ohta et al, 2001).

In a previous yeast two-hybrid screen using TPL as bait (Causier, Unpublished) ERF3 was found as an interactor, confirming the idea that the EAR domain is an important requirement for interactions with the TPL/TPR family. This coincides with the definition of the TPL/TPR family as transcriptional corepressors, and suggests that the active transcriptional repression of the ERFs requires interaction with the TPL/TPR family proteins. However, ERF3 was not found to be a positive interactor in the TPR3 library screen. Although it is possible that this due to some characteristic of TPL which TPR3 does not possess, for example the site of interaction or the actual globular nature of the protein, it is more likely that a positive interaction between these proteins was not found in the positives sequenced to date and that the interaction can still be found in the remaining positives. This was confirmed in the interaction matrix as ERF3 interacted with TPR3 and TPL. However, it did not interact with TPR4, therefore questioning the importance of the phylogenetic relationship between the TPL/TPRs (Figure 5) and the importance of the orientation of the binding and activation domains in the yeast two-hybrid screen. The interaction matrix (Figure 6) and the TPR3 library screen (Table 3) show that on the whole TPR3 and TPL interactions are similar, while TPR4 interacts with less of the same proteins as these two, therefore, phylogenic relationships do not necessarily predict interaction specificities or functional differences of TPL/TPR members.

Most of the positive protein interactors of TPR3 are known transcription factors. Arabidopsis has around 33,000 genes which encode proteins of which 1510 to 1581 are transcription factors (Qu and Zhu, 2006). In the TPR3 screen over 80% of positives identified to date are known transcription factors, therefore, transcription factors are vastly over represented in the screen. Most of the proteins found to interact with TPR3, like in the previous TPL library screen, contained either an EAR domain or a variation of it. Most of these proteins certainly contained at least an LXL structure, with the exception of ZAT10 and ARF9. However, ZAT10 and ARF9 do contain known repression domains that are similar to the EAR domain. An interaction between the TPL/TPRs and this domain or the LXL structure would explain why these proteins were identified in both the screens. So it can be deemed that, as they possess an EAR domain, they are transcriptional repressors and, as prior stated, this repression is aided by the corepressive activity of TPR3.

A large family of know transcription factors, the MYB family, are represented by the positives EPR1, MYB32 and MYB44 in this library screen. The MYB family can be activators, repressors or both of these and, like most of the genes discussed in this paper, have a modular structure (Stracke et al, 2001). This modular structure enables the classification of this large family. In Arabidopsis MYB proteins contain two related helix-turn helix motifs, the R2 and R3 repeats and are consequently known as the ‘R2R3-type MYB' factors (Kranz et al, 1998). As the MYB genes are expressed in different tissues and physiological conditions it is indicated that they are involved in many aspects of plant developmental and metabolic processes and, therefore, their interaction with TPR3 is justified as the TPL/TPR family are found in most areas of the plant.

In this case MYB32 was identified the most in this screen, out of those in this family. It was identified 4 times in this screen in comparison to MYB44, the only other MYB found in this screen, which was identified only once. MYB32 has a long EAR-Like motif of LDLNLEL (LXLXLXL); therefore, one could assume that this contributes more to its interaction as it is a more ‘stable' interaction. As MYB32 was also identified of the TPL screen it can be suggested that it is a common interactor of the TPL/TPR family and that its lengthened LXLXL motif contributes to this. MYB32 has been confirmed to be involved in many plant developmental processes such as its induction of two phenylpropanoid pathway genes: ANTHOCYANIDIN SYNTHASE (ANS) and DIHYDROFLAVANOL 4-REDUCTASE (DFR) (Wheeler et al, 2005) and in normal pollen grain development (Preston et al, 2004), so interaction with TPR3 is meaningful as the TPL/TPR family are suggested to be common corepressors and are present in many areas of the plant. However, in the interaction matrix MYB32 was not found to interact with TPR4, which is surprising due to the EAR domain of MYB32 and the close phenotypic relationship between TPR4 and TPL (Figure 5). Therefore, it maybe possible that the length of the EAR domain and/or the relationship status within the TPL/TPR family has no impact on the types and strength of interactions of the TPL/TPR proteins. This suggestion is supported as the MYB protein, MYB44 (LSLSL), was identified of this TPR3 screen and not from the TPL screen. Like any other MYB it is involved in a variety of functions including abiotic stress tolerance and it enhances stomatal closure (Jung et al, 2008). Therefore, there seems to be no functional reason for TPL not to interact with MYB44 as the TPL/TPR family are found in many areas of the plant. So the lengthened EAR domain must have no impact on the different interaction specificities of TPL and TPR3. Other regions of the protein, other than the EAR domain, will also confer interaction specificity.

Further evidence suggesting that the length of the LXLXL motif and that the relationship status of the TPL/TPR family is not a factor in understanding what types of protein interact with the TPL/TPR proteins is the interaction with EPR1. EPR1, which has only an LSL structure, was identified in both, TPR3 and TPL, screens. This enables the suggestion that it is only an LXL which is needed for an interaction with a TPL/TPR. However, neither ZAT10 nor ARF9 has an LXL structure within their EAR-like motif.

ZAT10 is a C2 H2 Zinc finger transcriptional repressor protein involved in salinity, heat and osmotic stress tolerance (Mittler et al, 2006). It was identified of this library screen 8 times which suggests that it is a strong interactor of TPR3. The ZAT10 EAR domain structure is FDLNI, which is a variation of the ERF EAR domain structure; L/F DLN L/F(x)P, as it has an Isoleucine at the end instead of a Leucine. FXLXF is also a well established repression motif and although it does not have an LXL, many interactors of TPL have this common variation repression motif.

Another large family of genes in Arabidopsis is represented in this screen by LBD37 and LBD41. The proteins in the LBD family contain a conserved region called the LOB domain which consists of a zinc finger and a leucine-zipper motif (Albinsky et al, 2010). LBD37 and LBD41 are induced by nitrogen or glutamine signalling. As nitrogen plays an important role in plant growth, LBD37 and LBD41, therefore, must have pivotal roles in the growth of plants (Albinsky et al, 2010). In particular LBD37 has been found to negatively regulate anthocyanin biosynthesis by repressing anthocyanin regulators PAP1 and PAP2 after their induction by Nitrogen and nitrate (Rubin et al, 2009). LBD37 is a known transcriptional repressor. There is no known biological function of LBD41 as of yet, however, there are some available insertion lines which may allow investigation in the future. LBD37 and LBD41 have the EAR domain structures, LDLSL and LDLTLRL, respectively. In this screen with TPR3 as the bait LBD37 was identified 15 times where as LBD41 was identified only 8, this is in contrast to the TPL screen were LBD41 was identified 19 times and LBD37 5 times (Causier, Unpublished). This could be due to the way in which each of the TPL/TPRs recognises the protein this could involve the EAR domains of each of the LBDs but it is more likely to be due to some characteristic of the TPL/TPRs themselves. In this case the interaction matrix shows that LBD37 interacts with all tested TPL/TPRs, including TPR4, however, LBD41 only interacted with TPL and TPR3. However, in TPR3 there was only a weak interaction, which is consistent with the idea that LBD41 interacts less with TPR3 from th