Molecular characterization of Arabidopsis thaliana PUF proteins – binding specificity and target candidates Carlos W. Francischini and Ronaldo B. Quaggio Departamento de Bioquı ´ mica, Instituto de Quı ´ mica, Universidade de Sa˜o Paulo, Brazil Introduction The translational control of RNA is an important reg- ulatory process in animal development. This regulation is accomplished by sequence-specific RNA-binding proteins that recognize cis-acting elements usually located in the 3¢ UTR. In recent years, and as a result of great efforts aiming to understand the mechanism of RNA control in animals, the function of a diverse number of RNA-binding proteins has been elucidated [1–4]. Despite this, translational control through the binding of RNA-binding proteins to 3¢ UTR tran- scripts has been poorly described in plants. PUF proteins are a large family of RNA-binding proteins found in all eukaryotes. These proteins reduce the expression of mRNA targets by binding in 3¢ UTR regulatory elements, thus controlling translation or mRNA stability [5]. Members of the PUF family have been implicated in diverse processes in development. In Drosophila, Pumilio binds to the Nanos response ele- ment (NRE) sequence within the 3¢ UTR of maternal hunchback mRNA and reduces its expression in the posterior pole of the embryo. This control is essential for abdomen formation [6]. In Caenorhabditis elegans hermaphrodites, the Pumilio homolog FBF binds to the 3¢ UTR of fem-3 mRNA, repressing its translation and controlling the sperm–oocyte switch [7]. Dictyoste- lium PufA represses pkaC mRNA and inhibits the Keywords Arabidopsis; PUF proteins; RNA-binding protein; three-hybrid system; translational control Correspondence R. B. Quaggio, Instituto de Quı ´ mica, Departamento de Bioquı ´ mica, Universidade de Sa˜o Paulo, Avenida Professor Lineu Prestes, 748, Sa˜o Paulo 05508-000, Brazil Fax/Tel: +55 11 3091 2171 E-mail: rquaggio@iq.usp.br (Received 24 March 2009, revised 15 July 2009, accepted 22 July 2009) doi:10.1111/j.1742-4658.2009.07230.x PUF proteins regulate both stability and translation through sequence-spe- cific binding to the 3¢ UTR of target mRNA transcripts. Binding is medi- ated by a conserved PUF domain, which contains eight repeats of approximately 36 amino acids each. Found in all eukaryotes, they have been related to several developmental processes. Analysis of the 25 Arabid- opsis Pumilio (APUM) proteins presenting PUF repeats reveals that 12 (APUM-1 to APUM-12) have a PUF domain with 50–75% similarity to the Drosophila PUF domain. Through three-hybrid assays, we show that APUM-1 to APUM-6 can bind specifically to the Nanos response element sequence recognized by Drosophila Pumilio. Using an Arabidopsis RNA library in a three-hybrid screening, we were able to identify an APUM- binding consensus sequence. Computational analysis allowed us to identify the APUM-binding element within the 3¢ UTR in many Arabidopsis tran- scripts, even in important mRNAs related to shoot stem cell maintenance. We demonstrate that APUM-1 to APUM-6 are able to bind specifically to APUM-binding elements in the 3¢ UTR of WUSCHEL , CLAVATA-1, PINHEAD ⁄ ZWILLE and FASCIATA-2 transcripts. The results obtained in the present study indicate that the APUM proteins may act as regulators in Arabidopsis through an evolutionarily conserved mechanism, which may open up a new approach for investigating mRNA regulation in plants. Abbreviations APBE, APUM-binding element; APUM, Arabidopsis Pumilio; IRP, iron regulatory protein; NRE, Nanos response element. 5456 FEBS Journal 276 (2009) 5456–5470 ª 2009 The Authors Journal compilation ª 2009 FEBS development of fruiting bodies [8], whereas, in yeast, both Puf3 and Puf5 (Mpt5) proteins promote the decay of COX17 and HO mRNA, respectively, through binding to their 3¢ UTR sequences [9,10]. Although members of this family of proteins have been shown to play distinct roles in different organ- isms, the maintenance and self-renewal of stem cells appears to be an ancestral function [5,11]. Drosophila Pumilio binds to a NRE-like sequence within the 3¢ UTR of cyclin B1, repressing its translation and promoting germline stem cell development [12–14]. C. elegans FBF also controls germline stem cell main- tenance by regulating gld-1 mRNA expression and sus- taining mitosis [15]. The Planaria PUF homolog DJPum is expressed in neoblasts, which are capable of self-renewal and differentiation during Planaria regen- eration. DJPum inactivation by dsRNA was found to cause a dramatic reduction in the number of neoblasts and impaired tissue regeneration [16]. In mammals, PUM2 is expressed in human germline stem cells [17], whereas the mouse homologs Pum1 and Pum2 are expressed in fetal and adult hematopoietic stem cells, as well as in fetal neural stem cells [11]. The canonical PUF domain comprises eight PUF repeats of approximately 36 amino acids each, arranged in tandem to form a single concave structure, usually located in the C-terminal region of the protein. Each repeat is formed by three a-helices that align with the equivalent helices in the adjacent repeat, forming three ladders of helices running through the domain [18,19]. The crystal structure of the human Pumilio ⁄ NRE complex demonstrated that each repeat of the PUF domain recognizes a single nucleotide in the RNA. Sequence-specific recognition is mediated by three conserved amino acids residues present at posi- tions 12, 13 and 16, located in the second helix of each repeat [20]. Recently, it was shown that these residues are also important for C. elegans FBF specificity, suggesting that PUF proteins of different organisms recognize RNA with the same modularity [21]. Although PUF proteins have been shown to regu- late distinct mRNA targets across species, the nucleo- tides recognized appear to be conserved because all known mRNAs regulated by these proteins contain a UGURN 1-3 AU(A ⁄ U) sequence [7–10,15,22–29]. In addition to its ability to bind RNA, the PUF domain was demonstrated to take part in the pro- tein–protein contacts necessary for RNA regulation [6,30,31]. In the present study, we report the first analysis of plant proteins possessing PUF repeats. Using compu- tational analyses and yeast three-hybrid assays, we found that at least six Arabidopsis thaliana proteins possess eight PUF repeats and can specifically recog- nize the NRE sequence of Drosophila hunchback mRNA. Through a yeast three-hybrid screening using an Arabidopsis RNA hybrid library, we identified mRNAs that may be target candidates of Arabidopsis Pumilio (APUM) regulation. The screen also allowed us to determine a consensus sequence recognized by the six APUM proteins that can bind to the NRE sequence (APUM-1 to APUM-6). Using this consen- sus, we show that APUM proteins are able to bind to the 3¢ UTR of transcripts related to self-renewal and stem cell maintenance in the shoot apical meristem. Moreover, the consensus sequence suggests that a great number of Arabidopsis transcripts are potential targets for regulation by the PUF family of proteins. The results obtained reveal a molecular conservation of PUF proteins in Arabidopsis thaliana and suggest that translational regulation via binding to 3¢ UTR in plants may have a role as important as that previously described in animals. Results PUF proteins in A. thaliana blast-p analysis of the Arabidopsis genome database (The Arabidopsis Information Resource – TAIR; http://www.arabidopsis.org) with Drosophila Pumilio was carried out to localize PUF proteins. Further pfam analyses using a cut-off E-value of 1.0 over the blast-p output identified 25 proteins containing PUF repeats, which is the largest number of putative PUF proteins found in a single organism to date. We named the putative A. thaliana Pumilio homologs APUM-1 to APUM-25 (Fig. 1). clustal w alignment of these protein sequences was used to generate a phylogenetic tree [32], indicating that they may be separated into four distinct groups of similar pro- teins, which we named groups I, II, III and IV (Fig. 1A). Only 3 out of 25 proteins were found to fall outside these groups. Within each group, some proteins show a degree of primary sequence identity, from 40% to 90% along their entire lengths, and from 63% to 96% among their PUF domains (Table 1). The analysis also showed that the PUF domain of the six proteins from group I are highly similar (approximately 50% identical and 75% simi- lar) to the Drosophila PUF domain (Fig. 2A and Table 2), whereas the six proteins from group II have lower levels of similarity (30% identical and 50% similar) with the Drosophila PUF domain (Fig. 2B and Table 2). Moreover, these 12 putative APUM proteins from groups I and II have eight PUF C. W. Francischini and R. B. Quaggio PUF proteins in Arabidopsis FEBS Journal 276 (2009) 5456–5470 ª 2009 The Authors Journal compilation ª 2009 FEBS 5457 repeats in the C-terminal region (Fig. 1B), equivalent to the number found in the well-characterized PUF proteins [5,11]. Proteins from group III, group IV and the three outsiders show more similarity among themselves than they do with the Drosophila PUF domain (data not shown). A B Fig. 1. Analysis of the 25 putative APUM proteins. (A) Phylogenetic tree constructed based on CLUSTAL W alignment of all putative APUM proteins and Drosophila Pumilio (accession number A46221). Numbers rep- resent the bootstrap analysis from 1000 trials. (B) Number of PUF repeats identified for each APUM in the PFAM analysis. Gray circles represent the localization of repeats in the protein and the numbers indicate the position of each repeat in the PUF domain. Black circles represent repeats identified in the PFAM that fall outside the C-terminal region. APUM proteins were named APUM-1 to APUM-25. PUF proteins in Arabidopsis C. W. Francischini and R. B. Quaggio 5458 FEBS Journal 276 (2009) 5456–5470 ª 2009 The Authors Journal compilation ª 2009 FEBS Prediction of APUM-binding specificity In the human Pumilio–NRE complex, residues 12 and 16 of each repeat make hydrogen bonds or Van der Waals contacts with a specific RNA base, whereas res- idue 13 makes stacking interactions [20]. An analysis of these residues for each PUF repeat of all putative APUM proteins showed that APUM-1 to APUM-6 have the amino acids in positions 12, 13 and 16, simi- lar to human and Drosophila Pumilio. On the other hand, the amino acids in these positions in APUM-7 to APUM-12 are more similar to those found in yeast Puf4 and Puf5 proteins (Table 3; data not shown). The remaining APUM proteins do not show conserva- tion of residues 12, 13 and 16 with any well-character- ized PUF homolog (data not shown) and possess less than eight PUF repeats in their PUF domains (Fig. 1B). The analysis allows us to suggest that the group I proteins (APUM-1 to APUM-6) share the same RNA-binding specificity of Drosophila and human Pumilio-1. Thus, we expected that these APUM proteins should bind to the NRE sequence within the 3¢ UTR of Drosophila Pumilio mRNA target hunchback. The group II proteins APUM-7 to APUM-11, which have a nonconservative Asn fi His substitution in residue 13 of repeat 7 (Table 3), would be expected to have binding to the second nucleotide in the UGU triplet impaired. Binding to this nucleotide is essential for RNA recognition [24,25,33–36]. Table 1. Amino acid identity between some putative Arabidopsis PUF proteins in the full-length and PUF domain. Gene ID Similar to: Full protein identity (%) PUF domain identity (%) At2g29190 At2g29140 ⁄ At2g29200 90 95 At3g20250 A4g25880 38 63 At1g78160 At1g22240 65 84 At1g35730 At1g35750 78 80 At5g43090 At5g43110 64 68 A B Fig. 2. CLUSTAL W alignment of the APUM proteins with the PUF domains most similar to Drosophila PUF domain. (A) APUM proteins of group I. (B) APUM proteins of group II. C. W. Francischini and R. B. Quaggio PUF proteins in Arabidopsis FEBS Journal 276 (2009) 5456–5470 ª 2009 The Authors Journal compilation ª 2009 FEBS 5459 Binding of APUM to NRE To test the predictions regarding the RNA-binding specificities of the putative APUM proteins, we investi- gated the capacity of APUM to bind to the NRE sequence of hunchback mRNA (Fig. 3B) [36]. We used the APUM-2 protein as a representative member of group I proteins and APUM-7 as a representative of group II APUM proteins (Table 3). Protein–RNA interactions was evaluated using yeast three-hybrid system, which was shown to be a reliable approach for identifying true interactions [25,33,37–39]. This system uses LexA ⁄ MS2 coat protein fusion to tether the RNA hybrid to the promoter of reporter genes. The RNA-binding protein is produced as a tran- scription activation fusion domain through which the reporter genes (HIS3 or LacZ) are transcribed when the RNA–protein interaction is established (Fig. 3A) [37]. The pYESTrp3 ⁄ APUM-2 and pYESTrp3 ⁄ APUM-7 vectors were transformed in the yeast YBZ-1 strain [40] together with the pRH5¢⁄NRE vector. After growth in selective medium, individual colonies were tested for LacZ reporter activation. The results showed that LacZ reporter was activated in yeast colonies transformed with APUM-2 and NRE, but not with APUM-7 and NRE (Fig. 3C). In the NRE fragment, two UGU sequences, named Box A and Box B (Fig. 3B), were shown to be essential for Drosophila Pumilio recognition because UGU resi- dues substitution in both boxes abolished the interaction with the protein [36]. To verify the specificity of APUM- 2 for the NRE, three mutant NREs with nucleotide sub- stitutions in the UGU sequence of Box A, NRE(A ) B + ); Box B, NRE(A + B ) ); and in both Box A and B, NRE(A ) B ) ), were used as baits in the yeast three-hybrid assay (Fig. 3B) [33]. The results of reporter activation indicated that APUM-2 interacts with NRE(A ) B + ), Table 2. The 12 Arabidopsis PUF domains most similar to the Dro- sophila PUF domain. Gene ID Similarity to Drosophila PUF domain (%) Identity to Drosophila PUF domain (%) Group I At2g29200 (APUM-1) 74 54 At2g29190 (APUM-2) 75 54 At2g29140 (APUM-3) 74 54 At3g10360 (APUM-4) 73 54 At3g20250 (APUM-5) 73 55 At4g25880 (APUM-6) 69 52 Group II At1g78160 (APUM-7) 56 29 At1g2240 (APUM-8) 55 30 At1g35730 (APUM-9) 51 29 At1g35750 (APUM-10) 53 29 At4g08840 (APUM-11) 57 29 At5g56510 (APUM-12) 53 31 Table 3. Alignment of the nucleotide binding residues of human Pumilio-1 and the corresponding residues in the APUM proteins. The well- characterized Drosophila Pumilio and yeast Puf4 are also shown. Amino acids at position 12, 13 and 16, respectively, of each repeat are boxed in gray. a Well-characterized Drosophila Pumilio and yeast Puf4 proteins included as comparison. b Preferential sequences recognized by Drosophila Pumilio [27]. c Preferential sequences recognized by yeast Puf4 [23]. PUF proteins in Arabidopsis C. W. Francischini and R. B. Quaggio 5460 FEBS Journal 276 (2009) 5456–5470 ª 2009 The Authors Journal compilation ª 2009 FEBS whereas no interaction was observed with NRE(A + B ) ) and NRE(A ) B ) ) (Fig. 3C). Quantitative analysis of LacZ expression showed that the binding affinity of APUM-2 for NRE(A ) B + ) was not significantly altered with respect to the wild-type NRE sequence, whereas the interaction with NRE(A + B ) ) and NRE(A ) B ) ) was fully abolished (Fig. 3D). Furthermore, assays using APUM-7 as prey did not interact with the wild-type or any of the mutant NREs (Fig. 3C). To confirm that the result of binding specificity observed between APUM-2 and NRE can be extended to the remaining group I proteins, we tested the interac- tion of APUM-1, APUM-3, APUM-4, APUM-5 and APUM-6 with wild-type and mutant NREs. Qualitative (data not shown) and quantitative analysis of LacZ activity (Fig. 3D) revealed that all five APUMs tested recognized the NRE and NRE(A ) B + ) sequences, but did not bind to NRE(A + B ) ) or NRE(A ) B ) ). Together, these results confirmed our predictions regarding the binding specificity of the subset of group I APUM proteins, showing that A. thaliana has at least six PUF proteins with conserved RNA-binding and similar specificity. The group I APUM proteins recog- nize Box B within the NRE sequence because UGU substitutions in Box A did not abolished the interac- tion. The Box B sequence presents a trinucleotide AUA downstream of the UGU motif (Fig. 3B), indicating that APUM proteins could recognize the sequence UGUANAUA, as do PUF proteins of other organisms (Table 3) [11]. These observations allow us to speculate that, although NRE is not the natural RNA target in Arabidopsis, the Box B sequence should mimic the authentic Arabidopsis targets. On the other hand, APUM-7 was unable to bind to NRE, possibly as a result of the nonconservative substitution at repeat 7. Influence of the Asp fi His substitution on the APUM-7 RNA-binding capacity The APUM-7 protein has the same binding residues as yeast Puf4 and Puf5 (Table 3; data not shown), except for an Asp fi His substitution at repeat 7. The yeast proteins have been shown to recognize sequences similar to the NRE Box B sequence (Fig. 3B and Table 3) [23]. We considered that, if APUM-7 did not bind to NRE because of the nonconservative substitution at residue 13 of the repeat 7, then changing this back to Asp may restore APUM-7 binding to NRE. In a simi- lar manner, if this Asp is critical for interaction, its substitution for a His would be expected to abolish binding of APUM-2 to NRE. To evaluate these hypotheses, we tested the interaction of APUM-2 ⁄ N fi H (APUM-2 with the Asp fi His A B C D E Fig. 3. Interaction analysis between APUM and the NRE transcript. (A) Schematic representation of the yeast three-hybrid system. (B) Sequence of the wild-type NRE transcript (WT) and NRE mutants with nucleotides substitutions in Box A, NRE(A ) B + ); Box B, NRE(A + B ) ); and in both Box A and B, NRE(A – B – ). (C) Qualitative analysis of LacZ reporter activation in the interaction of APUM-2 and APUM-7 with NRE WT and NRE mutants. The iron responsive element RNA and the IRP protein were used as positive controls for the interaction. (D) Quantitative analysis of LacZ reporter activa- tion in the interaction between APUM-1, APUM-2, APUM-3, APUM-4, APUM-5 and APUM-6 with NRE WT and mutants. (E) Interaction assay of APUM-2 with Asn to His substitution in the residue 13 of the repeat 7 (APUM-2 ⁄ N fi H) and the protein APUM-7 with His fi Asn substitution in the same position (APUM- 7 ⁄ H fi N) with the NRE transcript. C. W. Francischini and R. B. Quaggio PUF proteins in Arabidopsis FEBS Journal 276 (2009) 5456–5470 ª 2009 The Authors Journal compilation ª 2009 FEBS 5461 substitution at residue 13 of repeat 7) and of APUM- 7 ⁄ H fi N (APUM-7 with the His fi Asp substitution at residue 13 of repeat 7) with the NRE transcript in the three-hybrid system. The results obtained showed that APUM-2 ⁄ N fi H continued to recognize the NRE, whereas APUM-7 ⁄ H fi N did not (Fig. 3E), indicating that the failure to bind to NRE is not a result of Asp fi His substitution. Because APUM-8 to APUM-11 proteins share the same substitution in repeat 7, it is expected that they will behave as APUM-7 does (i.e. they will not bind to NRE). Similarly, the APUM-12 protein has exactly the same amino acids binding residues as APUM- 7 ⁄ H fi N, which suggests that they may exhibit simi- lar binding behaviors. Yeast three-hybrid screen to identify APUM-binding RNA To identify putative mRNA targets of APUM pro- teins, we used a yeast three-hybrid screen, which was shown to be a useful and reliable approach for profil- ing mRNAs that bind directly to a specific RNA-bind- ing protein [41–43]. Accordingly, we generated an Arabidopsis RNA hybrid library of small fragments (50–150 bp) and used this as prey in a three-hybrid screen with APUM-2 as bait (Fig. 4A). From approximately eight million independent Ara- bidopsis RNA sequences screened, 189 positive interac- tions derived from 63 distinct sequences were isolated (Fig. 4B). Of these 63 clones, 27 (43%) were insert cloned in antisense position. The other 36 clones (57%) were sense sequences, with five (14%) 3¢ UTR transcripts (Fig. 4C and Table 4). Computational analysis of RNA sequences identified in the yeast three-hybrid screen Although only five of 63 transcripts identified by the three-hybrid screens were derived from 3¢ UTR regions, all of them (sense and antisense) bound to bait specifically, suggesting the existence of a consensus motif within these 63 distinct transcripts recognized by APUM-2. We therefore analyzed these sequences using multiple expectation maximization for motif elicitation (meme) as a motif discovery tool [44] (http://meme. nbcr.net/meme/intro.html). The analysis identified an eight nucleotide motif present in all 63 transcripts (Fig. 5A). The consensus possesses a UGUR tetranu- cleotide sequence, which has been reported to be pres- ent in all targets of the PUF family [5,11]. In addition, a(A⁄ U)(U ⁄ G)(A ⁄ U ⁄ C) sequence located one nucleo- tide downstream of the UGUR motif is highly similar to the trinucleotide AUA and AUU present in the target consensus of many other PUF members [21,23,27,39,45]. In some transcripts, these last three nucleotides were AGA and AGC, which have not been described for any other PUF protein to date. On the basis of these results, we were able to iden- tify two NRE Box B-like consensus sequences, which we named the APUM-binding elements (APBE) (Fig. 5B). The APBE of the 3¢ UTR sequences identi- fied in the screening is shown in Table 4. Evaluation of the APBE identified by means of yeast three-hybrid screen Because the deduced binding consensus is very small, it must be present in a large number of Arabidopsis transcripts. Indeed, a search for the APBE motif in all 5¢ UTR, 3¢ UTR and ORFs annotated at the TAIR database showed that approximately 56% of all ORF A B C Fig. 4. Screen of an Arabidopsis RNA hybrid library to identify RNA bound by APUM-2. (A) Scheme of the three-hybrid strategy used in the screen. (B) Number of colonies identified in each step of the screen. (C) Distribution of the 63 distinct sequences in relation of their position in the Arabidopsis transcriptome. PUF proteins in Arabidopsis C. W. Francischini and R. B. Quaggio 5462 FEBS Journal 276 (2009) 5456–5470 ª 2009 The Authors Journal compilation ª 2009 FEBS sequences and 43% of all 3¢ UTR sequences have at least one binding consensus for APUM proteins, whereas, in 5¢ UTR, its occurrence is significantly lower (Table 5). As a result of the high occurrence of the APUM binding sites in the plant genome, we decided to focus in the binding of APUM consensus to 3¢ UTR tran- scripts expressed in the tissue related to plant meris- tems because the regulation of transcripts related to stem cell maintenance is considered to be an ancestral function of PUF proteins in animals. Thus, a 32 nucle- otide region of the 3¢ UTR of CLAVATA-1(CLV-1) (At1g75820), ZWILLE ⁄ PINHEAD (ZLL) (At5g43810), WUSCHEL (WUS) (At2g17950) and FASCIATA-2 (FAS-2) (At5g64630) transcripts was cloned in the pRH5¢ vector and tested with APUM-2 protein in the three-hybrid system (Fig. 5C). These four transcripts have been described to code for proteins involved in diverse developmental processes, including shoot meri- stem organization, stem cell maintenance and mainte- nance of cellular organization of apical meristems [46– 50]. The LacZ reporter was activated in all assays tested (Fig. 5D), indicating that the APBE motif is sufficient for APUM-2 recognition. The APUM-1, APUM-3, APUM-4, APUM-5 and APUM-6 proteins also interacted with these transcripts, whereas APUM- 7 did not (data not shown). These results confirm that the APBEs can be recognized by proteins of group I and also indicate that these consensus can be useful to identify putative mRNAs targeted by APUM-1 to APUM-6. Group I APUM proteins requires nucleotides in both 5¢ and 3¢ of the APBE motif In the computational analysis used to identify a con- sensus binding motif, no biases towards nucleotides outside the APBE were identified (Fig. 5A,B). How- ever, the interactions between APUM-2 with the NRE transcript and with the four 3¢ APBE UTR sequences chosen by bioinformatics analysis showed distinct val- ues of LacZ reporter activation (Figs 3D and 5E). These data suggest that binding affinity may be influ- enced either by nucleotides outside of the consensus motif or by small variations within the consensus. The interaction of APUM-2 with the FAS-2 tran- script was the strongest among the interactions tested in the three-hybrid system (Figs 3D and 5E). The FAS-2 transcript used in the binding assay differs from that of WUS, CLV-1 and ZLL sequences in both APBE and flanking nucleotides (Fig. 5C), whereas its binding core element is exactly the same as that of Box B present in the NRE transcript (Fig. 3B). Because APUM-2 binds to FAS-2 approximately five- fold more strongly than to NRE (Fig. 5E), we can sug- gest that specific nucleotides flanking the core element of FAS-2, which are not present in the NRE sequence, may contribute to APUM-2 binding. To examine the contributions of flanking nucleotides in the affinity between APUM-2 and FAS-2, we pro- duced double mutations in nucleotides upstream and downstream of the APBE (Fig. 6A). Quantification analysis of b-galactosidase activity showed that several substitutions reduced the binding affinity to different degrees (Fig. 6B,C). Most significantly, mutations at positions )1 ⁄ )2 abolished the interaction with APUM-2 (Fig. 6B). The interaction of APUM-1, APUM-3, APUM-4, APUM-5 and APUM-6 proteins with the FAS-2 transcript was also abolished when the nucleotides at positions )1 ⁄ )2 were substituted (Fig. 6D). These results demonstrate that nucleotides upstream and downstream of the binding consensus are critical for interaction with APUMs from group I. We can therefore consider the APBE as the core binding element, whereas other flanking nucleotides contribute to the accomplishment of strong or weak inter- actions. Table 4. 3¢ UTR transcripts identified in the yeast three-hybrid screening. Upper case letters and boxed sequences indicate the presumptive APUM binding sites. Information about each gene product was obtained from the TAIR database. Gene ID (number of times isolated) Coding for: Sequence identified (5¢-to3¢) At3g63500 (7) Protein containing PHD domain; unknown function ugcgucugaca UGUACAGCcccugccaaauuuuaauaggcaat AGUAAAUAaauaacgacaagaagcaaaugg At5g24490 (1) Ribosomal protein; unknown function cucaucucuccuuacaguuuaccuguguaggaguuaggguucuuga auaaacaaugcaacaaagauuguagaagucag UGUACAUA At4g36040 (1) Protein containing DNAJ domain; unknown function cuacgucggacggaacugggaaaccgaucaguguugguagugaguuaa cucggugaccgaguuaguagaacgaguuaauuag UGUAAAUAcgaagcca At4g39090 (1) ‘Embryo defective’ (RD19); response to physiological stress uuuaucucugcuucuugcu UGUAAAUAaa At3g47470 (1) Chlorophyll a ⁄ b-binding protein cuccaugaacaaauuuggaaucuucaa UGUACAGA C. W. Francischini and R. B. Quaggio PUF proteins in Arabidopsis FEBS Journal 276 (2009) 5456–5470 ª 2009 The Authors Journal compilation ª 2009 FEBS 5463 Discussion Multiple PUF members in A. thaliana Currently, the largest number of PUF proteins found in a single organism was in C. elegans, which has eleven homologs, whereas yeast has six; human and mouse possess two; and Drosophila and Dyctiostelium have only one member [5]. Recently, new studies have revealed the presence of ten, two and one homologs in Trypanosome, Plasmodium and Planaria, respec- tively [16,33,51]. In the present study, we showed that the A. thaliana genome may contain the largest number of putative PUF proteins described to date (Fig. 1). Functional characterizations of different homologs have shown that a single PUF protein may be associ- ated with several distinct developmental processes. Moreover, PUF proteins in the same organism may have overlapping and independent functions. In C. ele- gans, FBF-1 and FBF-2, which share 90% sequence identity, act redundantly in sperm–oocyte switch and germ stem cell maintenance [7,15]. However, these two proteins show distinct patterning functions in the distal germ line, independently affecting the number of cells in the mitotic region [29]. Also in C. elegans, the lack of PUF-8, which is more similar to Drosophila Pumilio than to FBF, causes germ line dedifferentiation and the formation of fast growing tumors [52]. In Drosoph- ila, the single Pumilio has been related to many inde- pendent processes [14,53–56], and five of the six yeast PUF homologs, which are significantly divergent in sequence, appear to have predominately distinct func- tions [23]. In A. thaliana, we have identified three highly con- served gene families that account for 22 of 25 putative PUF proteins. The three remaining proteins can be divided into a closely-related pair and a single outsider (Fig. 1A). The large number of copies of highly similar proteins (Table 1) could be an indicative of redundant functions in the plant. However, these functions might be specific to each group of duplicated genes. We A B C D E Fig. 5. Identification and evaluation of a common sequence motif in the mRNA obtained from yeast three-hybrid screen. (A) Eight nucleotide motif found by MEME analysis in all 63 distinct clones. (B) Deduced APBE. (C) Computational identification of an APBE (boxed sequences) in the 3¢ UTR region of transcripts FASCIATA-2 (FAS-2), WUSCHEL (WUS), CLAVATA-1 (CLV-1) and ZWILLE ⁄ PIN- HEAD (ZLL). The sequences shown are the 3¢ UTR regions used in the yeast three-hybrid assays. (D) Qualitative analysis of LacZ repor- ter activation in the interaction between APUM-2 and the tran- scripts FAS-2, WUS, CLV-1 and ZLL. The NRE sequence (Fig. 2B) was used as a positive control. (E) Quantitative analysis of LacZ activity in the interactions shown in (D). Table 5. Occurrence of the APBE in the A. thaliana transcriptome. Consensus 5¢ UTR a 3¢ UTR a ORF a UGURNAKH 1831 7630 16803 UGURNUUA 377 1881 3844 a Known and putative sequences in the A. thaliana database (TAIR). A total of 21835 3¢ UTR sequences, 20 564 5¢ UTR sequences and 36 690 ORFs were analyzed separately and independently of length. PUF proteins in Arabidopsis C. W. Francischini and R. B. Quaggio 5464 FEBS Journal 276 (2009) 5456–5470 ª 2009 The Authors Journal compilation ª 2009 FEBS therefore predict that various PUF of A. thalina may be involved in many different processes in the plant. RNA-binding capacity of the APUM proteins pfam analysis of all putative APUM proteins showed that the six APUM group I proteins, all highly similar to Drosophila Pumilio (Figs 1A and 2A and Table 2), have the eight conserved repeats characteristic of the PUF family of proteins (Fig. 1B). These six homologs have the same residues necessary to confer RNA speci- ficity in human Pumilio-1 (Table 3) and can bind to the NRE sequence specifically (Fig. 3). Six group II APUM proteins (APUM-7 to APUM-12) (Fig. 1B) also possess eight PUF repeats, some of which do not show conservation in residues directly involved in nucleotide recognition (Table 3). Through site-directed mutagenesis and interactions assays, we showed that this substitution is not responsible for the APUM-7 binding impairment (Fig. 3E). Although PUF proteins have been shown to recog- nize RNA through a UGUR tetranucleotide followed by an AU(A ⁄ U) sequence, the number of nucleotides between these two sequences is variable among different homologs. For example, C. elegans FBF recognize RNA that have the UGUR and AUA sequence sepa- rated by two nucleotides, whereas C. elegans PUF-8, Drosophila and human Pumilio and yeast Puf3 recog- AB CD Fig. 6. Analysis of binding affinity between APUM-2 and the FAS-2 3¢ UTR transcripts with nucleotide substitutions upstream and down- stream of the APBE. (A) Double substitutions in flanking nucleotides of APBE (lower case). Bold letters in the wild-type sequence indicate the APBE. The first nucleotide of the motif is numbered base one. The individual adenine to guanine substitution at nucleotide four was per- formed to confirm the deduced APBE, which admits a guanine in this position (Fig. 5). (B) Quantitative analysis of LacZ reporter activation in the interactions between APUM-2 and the FAS-2 transcripts with substitutions in nucleotides upstream of the APBE. (C) Quantitative analy- sis of LacZ reporter activation in the interactions between APUM-2 and the FAS-2 transcripts with substitutions in the nucleotides down- stream of the APBE. (D) Quantitative analysis of LacZ activation in the interactions of APUM-1, APUM-3, APUM-4, APUM-5 and APUM-6 with the FAS-2 transcript wild-type (WT) and FAS-2 transcript with substitutions at positions )2 ⁄ )1. C. W. Francischini and R. B. Quaggio PUF proteins in Arabidopsis FEBS Journal 276 (2009) 5456–5470 ª 2009 The Authors Journal compilation ª 2009 FEBS 5465 [...]... role of post-transcriptional operon, in which mRNA encoding functionally-related proteins should be coordinately regulated by specific mRNP components [59] Systematic identification of the mRNA targets for five of six yeast PUF proteins showed that each homolog interacts with specific subpopulations of mRNA Moreover, Puf3 , Puf4 and Puf5 were shown to bind, respectively, to 56%, 26% and 49% of all known and. .. yeast and Drosophila, APUM proteins may be involved in the regulation of many aspects of growth and development Interaction between APUM proteins and transcripts possessing the APBEs In the present study, we have shown that APUM-1 to APUM-6 proteins are evolutionarily conserved PUF proteins Moreover, we have presented two binding consensus motifs that allow the efficient identification of target candidates. .. plant and regulate their translation and ⁄ or turnover A large number of APUM target candidates: is this plausible? RNA -binding proteins have been associated with diverse aspects of post-transcriptional gene regulation, including RNA processing, export, localization, degradation and translational regulation [57,58] Recent studies have shown that specific RNA -binding proteins associate with large and distinct.. .PUF proteins in Arabidopsis C W Francischini and R B Quaggio nize the same two sequences separated by only one nucleotide [21,23,25,27] In the RNA targets of yeast Puf5 , three nucleotides separate the UGUR and AUA trinucleotides, whereas the separation is only two nucleotides in the RNA targets of Puf4 [23] Thus, we speculate that, if APUM-7 to APUM-12 proteins can bind to RNA, the UGUR and AUA... APUM proteins, in which residues 12, 13 and 16 of each repeat are not conserved (data not shown), either bind to RNA targets that deviate from targets of typical PUF proteins or do not bind to RNA The results obtained provide strong evidence that at least six APUM proteins must function as translational regulators in a manner similar to that of other wellcharacterized members of PUF family These proteins. .. 5¢-TCCCCCGGGGG-3¢ and extended with Klenow fragment of Escherichia coli DNA polymerase After extension, the double-stranded DNA were digested PUF proteins in Arabidopsis with XbaI and SmaI and ligated to an AvrII-SmaI-digested pRH5¢ vector The sequences of primers used are given in Tables S1 and S2 Single amino acid change in APUM-2 and APUM-7 proteins The amino acid substitution to generate pYESTrp3APUM2 ⁄ N fi H and. .. Xcat-2, and a cytoplasmic polyadenylation element -binding protein J Biol Chem 276, 2094 5– 20953 23 Gerber AP, Herschlag D & Brown PO (2004) Extensive association of functionally and cytotopically related mRNAs with Puf family RNA -binding proteins in yeast PLoS Biol 2, E79 24 Jackson JS Jr, Houshmandi SS, Lopez Leban F & Olivas WM (2004) Recruitment of the Puf3 protein to its mRNA target for regulation of. .. by Mpt5, a yeast homolog of Pumilio and FBF EMBO J 20, 55 2–5 61 11 Spassov DS & Jurecic R (2003) The PUF family of RNA -binding proteins: does evolutionarily conserved structure equal conserved function? IUBMB Life 55, 35 9–3 66 12 Forbes A & Lehmann R (1998) Nanos and Pumilio have critical roles in the development and function of Drosophila germline stem cells Development 125, 67 9– 690 13 Asaoka-Taguchi... sequence motifs bound by PUF- 5 and PUF- 6 proteins from C elegans Using a threehybrid assay, they screened a random small fragment RNA library to define the binding motif sequence When the motif was used to search the C elegans mRNA database, they found a high number of mRNAs that have the binding motif on their 3¢ UTR sequence Some of the mRNAs were shown to be potential targets of the PUFs homologs in a... mRNA–protein complexes using a yeast three-hybrid system Methods 26, 12 3–1 41 Bernstein D, Hook B, Hajarnavis A, Opperman L & Wickens M (2005) Binding specificity and mRNA targets of a C elegans PUF protein, FBF-1 RNA 11, 44 7–4 58 Hook B, Bernstein D, Zhang B & Wickens M (2005) RNA–protein interactions in the yeast three-hybrid system: affinity, sensitivity, and enhanced library screening RNA 11, 22 7–2 33 . Molecular characterization of Arabidopsis thaliana PUF proteins – binding specificity and target candidates Carlos W. Francischini and Ronaldo. binding specificity of the subset of group I APUM proteins, showing that A. thaliana has at least six PUF proteins with conserved RNA -binding and similar specificity.