Modeling of the catalytic core of Arabidopsis thaliana Dicer-like 4 protein and its complex with double-stranded RNA

https://doi.org/10.1016/j.compbiolchem.2016.11.003Get rights and content

Highlights

  • Plant Dicer-like proteins have Platform domain similar to Giardia and human Dicers.

  • Assembling of the AtDCL-dsRNA complex was performed using NMA and template-based approaches.

  • The arrangement of the Platform and PAZ domains, rather than the length of the Connector, are responsible for the size of Dicer products.

Abstract

Plant Dicer-like proteins (DCLs) belong to the Ribonuclease III (RNase III) enzyme family. They are involved in the regulation of gene expression and antiviral defense through RNA interference pathways. A model plant, Arabidopsis thaliana encodes four DCL proteins (AtDCL1-4) that produce different classes of small regulatory RNAs. Our studies focus on AtDCL4 that processes double-stranded RNAs (dsRNAs) into 21 nucleotide trans-acting small interfering RNAs. So far, little is known about the structures of plant DCLs and the complexes they form with dsRNA. In this work, we present models of the catalytic core of AtDCL4 and AtDCL4-dsRNA complex constructed by computational methods.

We built a homology model of the catalytic core of AtDCL4 comprising Platform, PAZ, Connector helix and two RNase III domains. To assemble the AtDCL4-dsRNA complex two modeling approaches were used. In the first method, to establish conformations that allow building a consistent model of the complex, we used Normal Mode Analysis for both dsRNA and AtDCL4. The second strategy involved template-based approach for positioning of the PAZ domain and manual arrangement of the Connector helix. Our results suggest that the spatial orientation of the Connector helix, Platform and PAZ relative to the RNase III domains is crucial for measuring dsRNA of defined length. The modeled complexes provide information about interactions that may contribute to the relative orientations of these domains and to dsRNA binding. All these information can be helpful for understanding the mechanism of AtDCL4-mediated dsRNA recognition and binding, to produce small RNA of specific size.

Introduction

RNA interference (RNAi) is a highly conserved mechanism that controls gene expression in almost all eukaryotes (Szweykowska-Kulińska et al., 2003, Wilson and Doudna, 2013). Important players of the RNAi machinery are small regulatory RNAs (srRNAs) that are responsible for selectivity of gene silencing (Borges and Martienssen, 2015, Elbashir et al., 2001). Several types of srRNAs could be distinguished depending on the origin and structure of their precursors, and modes of their functioning (Bologna and Voinnet, 2014). The most recognized srRNAs are: (i) microRNAs (miRNAs), which originate from single-stranded stem-loop RNA precursors, pre-miRNAs, encoded in the nuclear genome and (ii) small interfering RNAs (siRNAs) derived from long double-stranded RNA (dsRNA) precursors of endogenous or exogenous origin (Axtell et al., 2011, Mickiewicz et al., 2016, Voinnet, 2009). The mature miRNA or siRNA serves as a probe that guides the RNA induced silencing complex (RISC) to the evolutionary conserved sequences in a target mRNA. In animals, miRNA binding sites are usually located in the 3′ untranslated region (3′-UTR) of the targeted transcripts. In plants, miRNAs mostly have target sites within the coding sequences (Brodersen and Voinnet, 2009). Further, in animals, miRNAs predominantly repress translation by forming partially complementary duplexes with target mRNAs, whereas siRNAs trigger degradation of the transcripts that is favored by perfect or near perfect base pairing (Huntzinger and Izaurralde, 2011). In plants, the assistance of partnering dsRNA-binding proteins, DRB1 or DRB2, determines which form of regulation plays the dominant role − cleavage or translational inhibition of mRNA targets, respectively (Reis et al., 2015). In both animals and plants, the mRNA cleavage is carried out by the Argonaute-family protein (AGO), that is the catalytic engine of every RISC (Liu et al., 2004, Meister et al., 2004, Wang et al., 2009).

Among core components of srRNA biogenesis machinery are ribonuclease III (RNase III) type enzymes, termed Drosha and Dicer in animals, and Dicer-like (DCL) proteins in plants (Nicholson, 2014). Ribonuclease Drosha is responsible for the first step of miRNA biogenesis – it cleaves the primary transcript to release precursor, pre-miRNA (Gregory et al., 2004). Then, Dicer protein performs the second step of cleavage and releases the miRNA duplex from pre-miRNAs (Chendrimada et al., 2005, Kurzynska-Kokorniak et al., 2015). Dicer produces also siRNAs from long double-stranded RNA (dsRNA) precursors. In case of plants, at least four genes encoding DCL proteins have been identified (Margis et al., 2006). Moreover, duplications of these genes have also been detected (Kapoor et al., 2008, Tworak et al., 2016). Arabidopsis thaliana, a model plant, contains four DCL proteins (AtDCL1-4) (Gasciolli et al., 2005). Each of them recognizes different substrates and generates srRNAs of a specific size. They also play distinct functions in the cell. AtDCL1 is involved in biogenesis of 21-nucleotide (nt) miRNAs from primary transcripts that contain stem-loop structures. The produced miRNAs regulate gene expression in various biological processes; e.g., development and metabolism (Bartel et al., 2004, Voinnet, 2009). In contrast, AtDCL2 generates 22-nt siRNAs from viral-derived dsRNAs and, similarly as its insect homologs, is a part of the antiviral defense system (Mlotshwa et al., 2008, Xie et al., 2004). Next, AtDCL3 is required for excision of 24-nt siRNAs from the perfectly complementary endogenous dsRNAs produced by the plant encoded RNA-dependent RNA polymerase-2 (RDR2). AtDCL3-dependent siRNAs participate in chromatin modifications, hence they influence chromatin structure and preserve genome stability (Chan et al., 2004). Function of AtDCL4 remained unclear for the longest time among all four Arabidopsis DCL proteins. In 2005, Gasscioli and Allen reported independently the discovery of trans-acting siRNAs (ta-siRNAs) (Allen et al., 2005, Gasciolli et al., 2005) that were derived from the RNA-coding genes termed TAS. It has been shown that after a miRNA-triggered cleavage of TAS transcripts (Felippes and Weigel, 2009) their fragments serve as a template for a complementary-strand synthesis by plant-encoded RDR6 (Rajeswaran and Pooggin, 2012). Resultant dsRNAs are cleaved by AtDCL4 in a phased way into 21-nt duplexes (Xie et al., 2005, Yoshikawa et al., 2005). AtDCL4-produced ta-siRNAs, similar as miRNAs and other siRNAs, are incorporated into the RISC to direct silencing of cognate mRNAs (Brodersen et al., 2008, Vazquez and Hohn, 2013). DCL4 is also important in antiviral defense against RNA viruses (Bouché et al., 2006, Deleris, 2006, Jakubiec et al., 2012, Zhang et al., 2015). Nevertheless, when DCL4 is absent, DCL2 can produce 22-nt virus-derived srRNAs (Deleris, 2006, Wang et al., 2011). All small RNAs in plants are methylated at their 3′-ends by the RNA 2′-O-methyltransferase HEN1 (Ji and Chen, 2012).

Both animal Dicer and plant DCL proteins are large, multi-domain enzymes, that typically consist of DExD/H-box helicase, helicase-C, DUF283 (domain of unknown function), PAZ domain, two RNase III domains (RIIIA and RIIIB) and one or two dsRNA-binding domains (dsRBDs). The RIIIA and RIIIB domains form a dimer containing single processing center to cleave RNA substrates and to release products that carry characteristic 2-nt overhangs at the 3′-ends and phosphates on 5′-ends (Zhang et al., 2004). The PAZ domain has been shown to bind the 3′-end of the substrates, with a preference for 2-nt-long overhangs (Lingel et al., 2004, Ma et al., 2004, Song et al., 2003, Yan et al., 2003). A cleft within the PAZ domain, responsible for the recognition of the 3′-end of a substrate and its binding, was termed the 3′ pocket. Due to their complexity and size, the eukaryotic Dicers are difficult to crystallize. The crystal structure of an intact Dicer has only been established for parasite Giardia intestinalis. The Giardia Dicer (GiDCR), often termed “the minimal Dicer”, is composed of only PAZ domain, two RNase III domains, Connector helix and a stabilizing Platform domain, that all constitute the minimal core of Dicer shown to be sufficient to produce a short dsRNA of a defined length (MacRae and Doudna, 2007a, MacRae et al., 2006b). Importantly, very recently the structure of human Drosha (HsDrosha) has been solved. The HsDrosha structure turned out to be similar to the structure of GiDCR enzyme (Kwon et al., 2016). Both Dicer and Drosha act as “molecular rulers” that measure and cut RNA substrate at the defined length from the specific RNA structure. The structure of GiDCR reveals that the distance between the 3′ pocket in the PAZ domain and the cleavage sites in the RNase III domains corresponds to size of srRNA produced by Dicer. The structure of HsDrosha shows unique Bump helix that may act as a measuring tool to indicate the 11-base-pair distance from the basal junction of pri-miRNA (Kwon et al., 2016). Additionally, structures of the individual Dicer domains have been experimentally determined at the atomic level: (i) the RIIIB domain from human Dicer (HsDCR) (Takeshita et al., 2007), (ii) the RIIIB domain together with the dsRNA binding domain from mouse Dicer (MmDCR) (Du et al., 2008), (iii) a fragment spanning Platform-PAZ-Connector helix domains from HsDCR in complex with siRNA (Tian et al., 2014) and (iv) the N-terminal helicase from HsDCR (Wilson et al., 2015). The overall architecture of the whole HsDCR, including the experimental localization of the individual domains, has been proposed based on the observations from electron microscopy (EM) (Lau et al., 2012). For plant DCLs the structure is available only for domains outside of the catalytic core: (i) DUF283 from DCL4 (Qin et al., 2010), (ii) the second dsRBD from DCL1 (Burdisso et al., 2012).

Structural data concerning the interactions of Dicer with dsRNA substrates are also limited. Recently reported structures of HsDCR Platform-PAZ-Connector helix fragment in a complex with several siRNAs have visualized the location of 3′ and 5′ substrate ends (Tian et al., 2014). The cleft that binds the 5′-end of a substrate, termed the 5′ pocket, is located within the PAZ and Platform domains. Moreover, models of HsDCR-RNA complexes have been constructed at 29–31 Å resolution, based on cryo-EM data (Taylor et al., 2013). Additional information about interactions of the individual Dicer domains with RNA can be deduced from the crystal structures of homologues domains in a complex with RNA; i.e., the bacterial RNase III domain (Gan et al., 2006), eukaryotic RNase III domain (Liang et al., 2014) and PAZ domains from AGO and PIWI proteins (Ma et al., 2004, Tian et al., 2011). Thus far, there is no structural data for interaction of DCLs with their RNA substrates.

A lack of structural data limits the understanding of DCL protein-mediated processes, including their interactions with RNA substrates and the molecular basis of the generated products’ length variation. Thus, we have attempted to model a complex composed of the AtDCL catalytic core assembled with the dsRNA substrate in the cleavage-competent state. For the modeling we have chosen AtDCL4 that generates 21-nt siRNA products from perfect dsRNA substrates. The modeled fragment of AtDCL4 corresponded to the minimal Dicer from GiDCR and comprised the Platform, PAZ, Connector helix and two RNase III domains (DCL4 amino acid residues (a.a.) 784–1464) (Fig. 1A). Importantly, it has been previously shown that the lack of “auxiliary domains” of Dicer-type proteins, i.e. the helicase, DUF283 and dsRBD domains, does not influence the length of srRNAs produced in vitro by the domain-deletion mutant, as compared to the full-length enzyme, but rather the efficiency of cleavage performed by the individual Dicer mutant (Ma et al., 2008, Ye et al., 2007). Consequently, we have performed homology modeling based on the crystal structures of the respective homologues domains from various species mentioned above. Interestingly, we have noticed that despite the low sequence similarity, the predicted secondary structure of Platform domains from different Dicer and DCL proteins, reveals a characteristic ββββαβα fold. Then, by applying two different approaches, we built the AtDCL4-dsRNA complex in its catalytic form. As a result we have found that presumably not a length of the Connector helix linking the PAZ and the RNase IIIA domains, as previously suggested (MacRae et al., 2006a, MacRae et al., 2006b), but the spatial arrangement of PAZ and RNase IIIA contributes to the “measurement” of the cleaved dsRNAs. An important role in this arrangement is played by the conserved residues of RNase IIIA and Connector, and the presence of the Platform that surrounds the Connector. The modeled complexes provide the structural insight into the architecture of the AtDCL4 regions involved in the interactions with dsRNA. This information can be helpful in designing the AtDCL4 mutants disabled in RNA binding and, consequently, in understanding the molecular mechanism of AtDCL4-mediated dsRNA recognition, binding and cleavage.

Section snippets

Sequence analysis

The analysis involved three sets of protein sequences. The first set consisted of plant DCL protein sequences validated experimentally or predicted from the completely sequenced and assembled genomes. Protein sequences of all four AtDCLs were used to search homologous DCL sequences in plant genomes gathered in Plant Genome Database release 187 (http://www.plantgdb.org) (Dong et al., 2004), for each species separately. The final hits were then used in BLAST as query sequences to search GenBank

Sequence analysis of dicer and dicer-like proteins

The first step toward the homology modeling of DCL4 was multiple sequence alignment (MSA) of plant DCL proteins. A search through PlantGDB database for DCLs encoded in plant genomes, using an E-value cutoff of 1e-20, resulted in over 100 hits. We discarded the sequences that shared the strong similarity of short fragments only, because they suggested the presence of homological domains but not a whole Dicer-related protein. Remaining 75 records in most cases were not annotated as a specific

Conclusions

In this work we proposed models of catalytic core of DCL4 in a complex with dsRNA. AtDCL4 was modeled using homology modeling. First, the models of individual domains and blocks of domains were obtained and, next, the structure of the GiDCR was used to model architecture of the domains. The functional “measuring” of 21-nt RNA by AtDCL4 requires a somewhat different orientation of the PAZ domain in relation to the RNase III domains as compared to GiDCR. In order to build the AtDCL4-dsRNA complex

Acknowledgments

This work was supported by the grant N N519 405037 from the National Science Centre, Poland. This work was partially supported by the European Union Regional Development Fund within the PARENT-BRIDGE Programme of Foundation for Polish Science [Pomost/2011-3/5 to A.K.-K.]. This publication was also supported by the Polish Ministry of Science and Higher Education under the KNOW program.

Calculations were performed at the Poznań Supercomputing and Networking Center.

References (123)

  • E. Ma et al.

    Autoinhibition of human dicer by its internal helicase domain

    J. Mol. Biol.

    (2008)
  • E. Ma et al.

    Coordinated activities of human dicer domains in regulatory RNA processing

    J. Mol. Biol.

    (2012)
  • I.J. MacRae et al.

    Ribonuclease revisited: structural insights into ribonuclease III family enzymes

    Curr. Opin. Struct. Biol.

    (2007)
  • R. Margis et al.

    The evolution and diversification of Dicers in plants

    FEBS Lett.

    (2006)
  • G. Meister et al.

    Human Argonaute2 mediates RNA cleavage targeted by miRNAs and siRNAs

    Mol. Cell

    (2004)
  • C. Notredame et al.

    T-Coffee: a novel method for fast and accurate multiple sequence alignment

    J. Mol. Biol.

    (2000)
  • A. Sali et al.

    Comparative protein modelling by satisfaction of spatial restraints

    J. Mol. Biol.

    (1993)
  • T. Sasaki et al.

    Evolutionary conservation of a unique amino acid sequence in human DICER protein essential for binding to Argonaute family proteins

    Gene

    (2007)
  • B. Simon et al.

    Recognition of 2'-O-methylated 3'-end of piRNA by the PAZ domain of a Piwi protein

    Structure

    (2011)
  • S.F. Altschul et al.

    Gapped BLAST and PSI-BLAST: A new generation of protein database search programs

    Nucleic Acids Res.

    (1997)
  • M.J. Axtell et al.

    Vive la différence: biogenesis and evolution of microRNAs in plants and animals

    Genome Biol.

    (2011)
  • N.A. Baker et al.

    Electrostatics of nanosystems: application to microtubules and the ribosome

    Proc. Natl. Acad. Sci. U. S. A.

    (2001)
  • N.G. Bologna et al.

    The diversity, biogenesis, and activities of endogenous silencing small RNAs in Arabidopsis

    Annu. Rev. Plant Biol.

    (2014)
  • C.S. Bond et al.

    ALINE: A WYSIWYG protein-sequence alignment editor for publication-quality alignments

    Acta Crystallogr. Sect. D Biol. Crystallogr.

    (2009)
  • G.M. Boratyn et al.

    Domain enhanced lookup time accelerated BLAST

    Biol. Direct

    (2012)
  • F. Borges et al.

    The expanding world of small RNAs in plants

    Nat. Rev. Mol. Cell Biol.

    (2015)
  • N. Bouché et al.

    An antagonistic function for Arabidopsis DCL2 in development and a new function for DCL4 in generating viral siRNAs

    EMBO J.

    (2006)
  • T.A. Bozorov et al.

    DICER-like proteins and their role in plant-herbivore interactions in nicotiana attenuata

    J. Integr. Plant Biol.

    (2012)
  • P. Brodersen et al.

    Revisiting the principles of microRNA target recognition and mode of action

    Nat. Rev. Mol. Cell Biol.

    (2009)
  • P. Brodersen et al.

    Widespread translational inhibition by plant miRNAs and siRNAs

    Science

    (2008)
  • B.R. Brooks et al.

    CHARMM: A program for macromolecular energy, minimization, and dynamics calculations

    J. Comput. Chem.

    (1983)
  • B.R. Brooks et al.

    CHARMM: The biomolecular simulation program

    J. Comput. Chem.

    (2009)
  • P. Burdisso et al.

    Second double-stranded RNA binding domain of dicer-like ribonuclease 1: Structural and biochemical characterization

    Biochemistry

    (2012)
  • S.W. Chan et al.

    RNA silencing genes control de novo DNA methylation

    Science

    (2004)
  • T.P. Chendrimada et al.

    TRBP recruits the Dicer complex to Ago2 for microRNA processing and gene silencing

    Nature

    (2005)
  • C. Cole et al.

    The Jpred 3 secondary structure prediction server

    Nucleic Acids Res.

    (2008)
  • P. Comella et al.

    Characterization of a ribonuclease III-like protein required for cleavage of the pre-rRNA in the 3′ETS in Arabidopsis

    Nucleic Acids Res.

    (2008)
  • S. Cristobal et al.

    A study of quality measures for protein threading models

    BMC Bioinf.

    (2001)
  • A. Deleris

    Hierarchical action and inhibition of plant dicer-Like proteins in antiviral defense

    Science

    (2006)
  • E.J. Denning et al.

    Impact of 2′-hydroxyl sampling on the conformational properties of RNA: Update of the CHARMM all-atom additive force field for RNA

    J. Comput. Chem.

    (2011)
  • Q. Dong et al.

    PlantGDB, plant genome database and analysis tools

    Nucleic Acids Res.

    (2004)
  • Z. Du et al.

    Structural and biochemical insights into the dicing mechanism of mouse Dicer: a conserved lysine is critical for dsRNA cleavage

    Proc. Natl. Acad. Sci. U. S. A.

    (2008)
  • P. Dunoyer et al.

    DICER-LIKE 4 is required for RNA interference and produces the 21-nucleotide small interfering RNA component of the plant cell-to-cell silencing signal

    Nat. Genet.

    (2005)
  • S.M. Elbashir et al.

    RNA interference is mediated by 21- and 22-nucleotide RNAs

    Genes Dev.

    (2001)
  • F.F. Felippes et al.

    Triggering the formation of tasiRNAs in Arabidopsis thaliana: the role of microRNA miR173

    EMBO Rep.

    (2009)
  • M.E. Fitzgerald et al.

    Dicer-related helicase 3 forms an obligate dimer for recognizing 22G-RNA

    Nucleic Acids Res.

    (2014)
  • N. Foloppe et al.

    All-Atom empirical force field for nucleic acids: I. parameter optimization based on small molecule and condensed phase macromolecular target data

    J. Comput. Chem.

    (2000)
  • A. Fukudome et al.

    Specific requirement of DRB4, a dsRNA-binding protein, for the in vitro dsRNA-cleaving activity of Arabidopsis Dicer-like 4

    RNA

    (2011)
  • R. Fukunaga et al.

    Inorganic phosphate blocks binding of pre-miRNA to Dicer-2 via its PAZ domain

    EMBO J.

    (2014)
  • J. Gan et al.

    A stepwise model for double-stranded RNA processing by ribonuclease III

    Mol. Microbiol.

    (2008)
  • Cited by (0)

    1

    Contributed equally to this work.

    View full text