MORF9 increases the RNA-binding activity of PLS-type ... · PLS-type pentatricopeptide repeat protein in plastid RNA editing ... Reports have been shown that MORF proteins interact

MORF9 increases the RNA-binding activity ofPLS-type pentatricopeptide repeat protein inplastid RNA editingJunjie Yan1†, Qunxia Zhang1†, Zeyuan Guan1, Qiang Wang1, Li Li1, Fengying Ruan1, Rongcheng Lin2,Tingting Zou1 and Ping Yin1*

RNA editing is a post-transcriptional process that modifies the genetic information on RNA molecules. In flowering plants,RNA editing usually alters cytidine to uridine in plastids and mitochondria. The PLS-type pentatricopeptide repeat (PPR)protein and the multiple organellar RNA editing factor (MORF, also known as RNA editing factor interacting protein (RIP)) aretwo types of key trans-acting factors involved in this process. However, how they cooperate with one another remains unclear.Here, we have characterized the interactions between a designer PLS-type PPR protein (PLS)3PPR and MORF9, and found thatRNA-binding activity of (PLS)3PPR is drastically increased on MORF9 binding. We also determined the crystal structures of(PLS)3PPR, MORF9 and the (PLS)3PPR–MORF9 complex. MORF9 binding induces significant compressed conformationalchanges of (PLS)3PPR, revealing the molecular mechanisms by which MORF9-bound (PLS)3PPR has increased RNA-bindingactivity. Similarly, increased RNA-binding activity is observed for the natural PLS-type PPR protein, LPA66, in the presence ofMORF9. These findings significantly expand our understanding of MORF function in plant organellar RNA editing.

RNA editing is a post-transcriptional modification process inwhich the identity of an RNA molecule is altered by theaddition, deletion or conversion of nucleotides from the

genome-encoded sequence. Different types of RNA editing havebeen widely observed in distinct kinds of RNA from divergentorganisms1. These modifications play crucial roles in multipleaspects of developmental regulation2, signal transduction3, antivirusresponse4, hormone and stress response5, organelle biogenesis6,7 andcell division7. In humans, dysregulation of RNA editing has beenassociated with myriad diseases such as glioblastoma8 and cancer9.

In eukaryotes, the most prevalent RNA editing is exemplified bytwo specific base deaminations, adenine to inosine (A-to-I) andcytosine to uracil (C-to-U), which are catalysed by several familiesof deaminases. A-to-I RNA editing is carried out exclusively by theenzyme named adenosine deaminase acting on RNA (ADAR).ADAR family members harbour one to three amino (N)-terminaldouble-stranded RNA binding domain(s) (dsRBDs) that form con-tacts with dsRNA and a carboxy (C)-terminal catalytic domain thatdeaminates the target adenine10. The best characterized C-to-Uediting in humans occurs in the nucleus and is catalysed by thedeaminase APOBEC1 and the APOBEC1 complementation factor(ACF), which together form the editosome. APOBEC1 cannotfunction alone and requires ACF, which recognizes the sequencedownstream of the edited cytidine and presents the target site toAPOBEC1 for deamination11. Intriguingly, C-to-U RNA editing inplants mainly occurs in plastids or mitochondria12–16. More than600 cytidines that are modified to uracil in these organelles havebeen identified17; however, the underlying mechanisms for the vastmajority of C-to-U editing events remain largely unknown.

To date, a number of editing factors have been genetically ident-ified to participate in C-to-U RNA editing in plant plastids andmitochondria, two of which are the components of the hypotheticaleditosome. The first is the PLS-type PPR protein, which functions in

target RNA recognition in plant organelles. PPR mutants exhibitdecreased RNA editing efficiency, which results in various develop-mental defects in plants2,18. Characteristic PLS-type PPR proteinsconsist of an array of PLS triplets, where the P (the canonical) motifcontains 35 amino acids, the L (long) motif contains 35–36 aminoacids and the S (short) motif contains 31 amino acids19. A numberof studies have described the right-handed superhelical assembly ofcanonical P-type repeats20–24; nonetheless, how PLS-type repeats areorganized is still unknown.

The second is MORF (RIP)25,26. In Arabidopsis thaliana, thisfamily consists of ten members, all of which share a similarly con-served domain without any structural information. MORF2 andMORF9 are targeted to plastids, MORF1 and MORF3–7 arelocated in mitochondria25 and MORF8 are dual-targeted26.Although morf mutants result in decreased editing efficiency25,26,the molecular functions of MORF members are poorly understood.Reports have been shown that MORF proteins interact with PPRproteins in vivo25,26, yet how these two factors cooperate with eachother requires further investigation.

In this study, we characterized the interactions between an engin-eered three-PLS triplet, (PLS)3PPR, and MORF9. We also deter-mined the crystal structures of (PLS)3PPR, MORF9, and(PLS)3PPR in complex with MORF9. We observed that(PLS)3PPR undergoes notable conformational changes on MORF9binding, leading to an increase in (PLS)3PPR’s RNA-bindingactivity. This finding provides us new insights into the molecularfunction of MORF protein and the assembly of key componentsof the hypothetical RNA editosome in plant organellar RNA editing.

ResultsDesign of an artificial PLS-type PPR protein. We previouslyreported the superhelical structure of the P-type PPR proteinPPR10 with or without ssRNA20. Based on these structures and

1National Key Laboratory of Crop Genetic Improvement and National Centre of Plant Gene Research, Huazhong Agricultural University, Wuhan 430070,China. 2Key Laboratory of Photobiology, Institute of Botany, Chinese Academy of Sciences, Beijing 100093, China. †These authors contributed equally tothis work. *e-mail: [email protected]

ARTICLESPUBLISHED: 10 APRIL 2017 | VOLUME: 3 | ARTICLE NUMBER: 17037

NATURE PLANTS 3, 17037 (2017) | DOI: 10.1038/nplants.2017.37 | www.nature.com/natureplants 1

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

sequence conservation analysis, we successfully designed fourartificial P-type PPR proteins that specifically targeted A, U, G orC bases. Structures of the designer P-type PPR (dPPR) proteins incomplex with RNA exhibited nearly identical superhelicalassemblies compared with the natural PPR10 structure21. Similardesign strategies for the artificial P-type PPR proteins have alsobeen adopted by other groups23,24.

Our initial efforts to determine the structure of PLS-type PPR pro-teins based on screening dozens of natural PPR proteins expressed inEscherichia coli were unsuccessful, as all of the recombinant proteinswere prone to aggregation. Thus, a similar design strategy was appliedto the L- and S-type motifs of PLS-type PPR proteins. Briefly, all ofthe PLS-type PPR proteins in A. thalianawere collected and analysedusing Prosite. A total of 1,117 31-amino-acid sequences were ident-ified as S-type motifs (Supplementary Fig. 1a). Sequence alignmentbetween the P- and L-type motifs revealed obvious differences forthe 24th and 27th residues, with L-type motifs favouring H24 andP-type motifs favouring M27 (ref. 16). Based on this discrepancy, atotal of 263 sequences were identified as L-type motifs by restrictingthe 24th residue to histidine and the 27th residue to not methionine(Supplementary Fig. 1a). The P-type motif was used as previouslydescribed21. According to recent motif definitions27, our designedP-, L- and S-type motif belong to P1-, L1- and S1 types, respectively.To optimize the solubility and behaviour of the designed protein, an

N-terminal cap (RRQGVAPT) and a C-terminal solvating helix(ELTYRRVVESYCRAKRFE) from PPR10 were added. Here, the arti-ficial PLS-type PPR protein (PLS)3PPR was constructed with ninerepeats containing three PLS triplets capped with two termini(Fig. 1a and Supplementary Fig. 1b). As expected, the engineered(PLS)3PPR was well expressed in E. coli and purified to homogeneityin large quantities for crystallization (Supplementary Fig. 1c).

Crystal structure of the designer (PLS)3PPR. The designer(PLS)3PPR was crystallized in the P43 space group with twosymmetric PLS triplets in one unit cell. A complete (PLS)3PPRprotein could be assembled from the neighbouring threeasymmetric unit cells (Supplementary Fig. 2). The overallstructure of the designer (PLS)3PPR displays a superhelicalarchitecture similar to P-type PPR proteins, which have anapproximately 75 Å polar axis and a 45 Å equatorial diameter(Fig. 1b); the structure of the designer P-type PPR proteinexhibited a longer, 91 Å polar axis and a similar equatorialdiameter of 47 Å23 (Supplementary Fig. 3).

On closer inspection of the structure, each motif consists of twoantiparallel α-helices designated helix a and helix b (Fig. 1b).Helices a and b from each repeat constitute the concave and convexsurfaces of the superhelical assembly, respectively. Each type ofmotif displays identical helical conformations (SupplementaryFig. 4a–c). To evaluate their conformational variations, the L- andS-type motifs were each aligned with the P-type motif. In the L-typemotif, helix b clearly rotates outward from the convex surface byapproximately 10° (Fig. 1c). For the S-type motif (31 amino acids),one helical turn is reduced in helix b but not in helix a or the twoloops (Fig. 1d and Supplementary Fig. 4a,c), which reflects the lackof four residues compared with the P-type motif (35 amino acids).

Based on the structural observation on PPR10, we had proposednumbering the PPR motif from the first amino acid of the firsthelix20. According to this numbering criterion, the 5th and 35th resi-dues (corresponding to the residues 6 and 1′ in Barkan et al.28 and tothe residues 4 and ii in Yagi et al.29) were found to be specificallyinvolved in RNA base recognition21,27. To assess whether the L- andS-type motifs are involved in RNA base recognition, the distancebetween 5th and 35th (31st in S-type motif) residues in each motifwas calculated. This value is nearly two times larger in the L-typemotif than in the P-type motif (6.03 Å versus 3.02 Å) (Fig. 1e)because of the outward shifting of helix b (Fig. 1c), but it is similarto the S-type motif (3.32 Å versus 3.02 Å) (Fig. 1e), indicating thatthe reduced helical turn in helix b has little effect on the conformationof the last loop (Fig. 1d). These results suggest that the S-type motif iscapable of coordinating the RNAbase, whereas the L-typemotif is not,unless there have been substantial conformation changes.

Crystal structure of MORF9 protein. The ten MORF members inArabidopsis, which all contain an uncharacterized conserveddomain (Supplementary Fig. 5), are required for RNA editing inplastids and mitochondria25,26. To uncover the functions of MORFproteins, we sought to determine the structures of this family. Allof the MORF proteins were expressed, purified to homogeneity andsubjected to crystallization. After numerous crystallization trials, wefinally determined the crystal structure of MORF9 (residues 75–196,C85S/C187S) at a resolution of 2 Å. We traced the conservedMORF domain, which comprises six core antiparallel β-sheetsflanked by three α-helices and two long loops (Fig. 2a) with aβααββαβββ topology (Fig. 2b). A search for structural homologuesof MORF9 using DALI (ref. 30) revealed no significant matches,highlighting the structural novelty of the MORF family.

MORF9 increases the RNA-binding activity of (PLS)3PPR.Previous studies have demonstrated that MORF proteins couldassociate with PLS-type PPR proteins in vivo25,26,31. However,

Residue no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35

P V V

N P L S P L S P L S C

T Y N T L I D G L C K A G K L D E A L K L F E E M V E K G I K P DL E F T F S S V L K A C A R L G A L E L G K Q I H G Y V I K S G F E S NS V V V Y N A L I D M Y S K C G L L E E A R K V F D E M P E K D

75 Å 45 ÅN C

1a

2a3a

4a8a 8b

7b

5b4b

90°

10°

90°

a

b

c d

e

Helix aHelix b

Helix bHelix a

90°

35th residue (L type)

31st residue (S type)

35th residue (P type)5th residue

3.03 Å

3.32 Å

6.03 Å

6b6a

9a

9b

Figure 1 | Structural overview of the designer (PLS)3PPR. a, Domainorganization of (PLS)3PPR, which comprises nine repeats containing threePLS triplets fused to the N- and C-termini of PPR10. The residues from eachP-, L- and S-type motif are shown below the depiction of the domainorganization. b, The structure of (PLS)3PPR in two perpendicularorientations. The two helices within each repeat are designated helix a andhelix b. c, Structural alignment of the L- and P-type motifs. d, Structuralalignment of the S- and P-type motifs. e, Distance calculation between the5th and 35th (31st in the S-type motif) residues. The residues andsecondary structures of the P-, L- and S-type motifs are coloured yellow,pink and slate, respectively. All figures were prepared using PyMOL.

ARTICLES NATURE PLANTS

NATURE PLANTS 3, 17037 (2017) | DOI: 10.1038/nplants.2017.37 | www.nature.com/natureplants2

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

whether and how MORF proteins directly interact with PPRproteins remains obscure. We first assessed the interaction via sizeexclusion chromatography (SEC). The elution volume of MORF9and (PLS)3PPR were at approximately 16 and 15 ml, respectively(Fig. 3a, upper and middle panels). On co-incubation, MORF9and (PLS)3PPR were co-eluted at 12.5 ml, indicating the formation ofa (PLS)3PPR–MORF9 complex (Fig. 3a, lower panel). It is noteworthythat the stoichiometry of MORF9: (PLS)3PPR was approximately3:1, which is corroborated by analytical ultracentrifugation (AUC)described below.

We examined the RNA-binding affinity of (PLS)3PPR, MORF9,and the binary complex by electrophoretic mobility shift assay(EMSA). Our previous work had demonstrated that di-residuesND and SN at 5th and 35th residues in the P-type motif recognizeuracil (U) and adenine (A), respectively21. On the basis of the struc-tural alignment of the three motifs (Fig. 1e) and the other group’spredictions16, we speculate that both the S-type motif and the L-type

motif share a similar RNA base recognition mode as the P-typemotif. Thus, di-residues at 5th (N) and 31st (D) in the S-type motifshould also recognize U. As the distance of di-residues at the 5th (S)and 35th (N) in the L-type motif is notably larger than that of P-and S-type motif (Fig. 1e), there remain some uncertainties whetherthe L-type motif can efficiently recognize its targeted RNA base Aor not. Accordingly, we used these four kinds of RNA substrates(5′-(UNU)3-3′, N represents A, G, C or U) for EMSA.

Neither (PLS)3PPR nor MORF9 showed obvious RNA-bindingactivity even at high protein concentrations (Fig. 3b andSupplementary Fig. 6). Intriguingly, increasing band retardationwas detected with increasing amounts of the (PLS)3PPR–MORF9complex with a substrate preference. The most favoured RNAsubstrate of the complex is 5′-(UAU)3-3′ (the underlined lettersindicate the differences of distinct RNA substrate) with an estimateddissociation constant of approximately 5.0 µM (Fig. 3b); promiscuousRNA-binding bands was observed for 5′-(UGU)3-3′ (SupplementaryFig. 6a); and no RNA-binding activity was observed using either5′-(UCU)3-3′ or 5′-(UUU)3-3′ (Supplementary Fig. 6b,c). Theseresults suggest that the L-type motif shares the similar RNA baserecognition mode with that of the P-type and the S-type motifs.Moreover, isothermal titration calorimetry (ITC) experimentswere applied to analyse the RNA-binding affinity of those proteinstargeting to the RNA 5′-(UAU)3-3′. Consistently, only the(PLS)3PPR–MORF9 complex exhibits remarkable RNA-bindingactivity with a disassociation constant of approximately 1.4 µM(Supplementary Fig. 7). Thus, taken together with the result of theSEC and the AUC assays, it can be concluded that MORF9 is ableto interact with (PLS)3PPR and increase its RNA-binding activity.

L-type motifs play crucial roles in mediating (PLS)3PPR andMORF9 interactions. To elucidate the molecular basis of(PLS)3PPR recognition by MORF9, we attempted to determinethe crystal structure of the (PLS)3PPR–MORF9 complex. After aseries of rigorous explorations, we finally determined the complex

β1C

C

β2

β5

β6

α1

α2

α3N

a b

N

β3

β4

β4

β1

β3

β2

β5

β6

α1

α2

α3

Figure 2 | Crystal structure of MORF9. a, Overall structure of MORF9. Thesecondary structural elements are labelled. b, Topology diagram of MORF9,with the secondary structure coloured and labelled as in a.

MORF9

MORF9

MORF9

MORF9

(PLS)3PPR + MORF9

(PLS)3PPR + MORF9

(PLS)3PPR

(PLS)3PPR(PLS)3PPR

(PLS)3PPR

Elution volume (ml)

Lane no.

U

U

U

B

12 14 16 18 20

Elution volume (ml)

12 13 14 15 16 17 18

121110987654321 Lane no. 1098765432113 14

a b

Figure 3 | MORF9 interacts with (PLS)3PPR and increases its RNA-binding activity. a, Size exclusion chromatography analysis of MORF9, (PLS)3PPR andMORF9 in complex with (PLS)3PPR. Fractions at the same elution volume from individual injections were examined using SDS–PAGE. b, The RNA-bindingactivity of MORF9, (PLS)3PPR and the (PLS)3PPR–MORF9 complex, as revealed by EMSA. The final protein concentrations of MORF9 and (PLS)3PPR inlanes 1–10 are 0, 1.4, 2.1, 3.2, 4.7, 7.1, 10.7, 16, 24 and 36 µM, respectively. The protein concentrations of the (PLS)3PPR–MORF9 complex were 0, 0.9, 1.4,2.1, 3.2, 4.7, 7.1, 10.7, 16 and 24 µM. The RNA probe is 5′-UAUUAUUAU-3′. B, bound; U, unbound.

NATURE PLANTS ARTICLES

NATURE PLANTS 3, 17037 (2017) | DOI: 10.1038/nplants.2017.37 | www.nature.com/natureplants 3

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

structure at a resolution of 2.5 Å. After assigning most of the aminoacids, we clearly observed three MORF9 molecules bound to one(PLS)3PPR, with each MORF9 binding to a single PLS triplet(Fig. 4a), consistent with the observations from the SEC assay(Fig. 3a, lower panel). We therefore speculated that a PLStriplet alone is sufficient for MORF9 binding. To test this hypothesis,we constructed two additional PPR proteins containing either oneor two PLS triplet(s), designated (PLS)1PPR and (PLS)2PPR. Both(PLS)1PPR and (PLS)2PPR interacted with MORF9 in SEC assays(Supplementary Fig. 8a). The molar ratios of these three PPR proteinsand MORF9 were further explored by AUC. As expected, one, twoand three MORF9 molecules were observed to bind (PLS)1PPR,(PLS)2PPR and (PLS)3PPR, respectively (Supplementary Fig. 8b).Thus, the biochemical analysis corroborated the structural observations.

In the complex structure, most of the interactions occur on theconvex surface of (PLS)3PPR, which is composed of helices bfrom each repeat (Fig. 4a). The interaction of each MORF9–PLStriplet is primarily mediated by an L-type motif and includesone hydrogen bond and two networks of extensive hydrophobiccontacts. The hydrogen bond is formed between K29 from theL-type motif and D164 from MORF9 (Fig. 4b, the middle). Asingle mutation at either of these two residues completely abolishesthe interaction (Supplementary Fig. 8c). Four residues (I80, L82,W160 and L162) from MORF9 that buttress the hydrogen bondare embedded in two hydrophobic grooves formed by the P-type(L21, F24) and L-type motifs (L19, I23 and Y26) and the L-type(H24, I28) and S-type motifs (Y11, L16, V23) (Fig. 4b).

MORF9 induces substantial conformational changes of(PLS)3PPR. Structural superposition of MORF9 with each(PLS)3PPR-bound MORF9 reveals few conformational changesfor MORF9, with a root mean squared deviation (r.m.s.d.) valueof approximately 0.5 Å over 97 Cα atoms (Supplementary Fig. 9).Interestingly, the architecture of (PLS)3PPR undergoes strikingconformational changes upon MORF9 binding (Fig. 5a,b). Thesuperhelical assembly of MORF9-bound (PLS)3PPR is compressed

towards the centre, whereas both the axial length and diameterare reduced (Fig. 5b). Close structural inspection of the individualrepeats reveals the molecular basis (Fig. 5c–e): both the P- andS-type motifs exhibit few conformational changes (Fig. 5c,e);however, the L-type motif shows pronounced conformationalvariation, with helix b rotated inward by approximately 6°, placingthe 35th residue closer to the 5th residue (Fig. 5d). This differencemay be gradually amplified over an increasing number of L-typemotifs, ultimately leading to the prominent compression of thesuperhelix, which is likely to favour target RNA binding.

DiscussionRNA editing has received particular attention because it challengesthe central dogma of molecular biology by changing genetic infor-mation at the transcript level. RNA editing is a dynamic and pre-cisely regulated process. Recent progress, mostly via geneticstudies, has revealed two types of key players in C-to-U RNAediting in plant organelles as the two components of the hypotheti-cal RNA editosome: PLS-type PPR proteins6,32 and MORF pro-teins25,26. Here, we successfully designed an artificial PLS-typePPR protein, (PLS)3PPR (Fig. 1a), and reconstituted the interactionbetween (PLS)3PPR and MORF9 (Fig. 3a). Moreover, we identifiedthe obvious RNA-binding activity of the complex (Fig. 3b). We alsodetermined the crystal structures of (PLS)3PPR (Fig. 1b), MORF9(Fig. 2a) and the binary complex (Fig. 4a). Structural alignment offree (PLS)3PPR with MORF9-bound (PLS)3PPR and a RNA-bound dPPR21 revealed distinct degrees of compactness for eachindividual superhelical assembly. RNA-bound dPPR and (PLS)3PPRdisplayed the most and least compact assembly, respectively,whereas MORF9-bound (PLS)3PPR ranked in the middle (Fig. 6a).This structural disparity may explain why MORF9-bound (PLS)3PPRhas increased RNA-binding ability compared with free (PLS)3PPR.On the basis of these findings, we propose a working model for howMORF9 functions in plant organellar RNA editing. In the absence ofMORF9, PLS-type PPR protein alone retains weak RNA-bindingactivity. However, in the presence of MORF9, PLS-type PPR protein

L162

MORF9

MORF9

MORF9

MORF9

P S

L a

a a

b b b

MORF9

MORF9

MORF9

MORF9

P S

L a

a a

b b b

L21

F24

L19

Y26

I23

I80 L82

L21

F24

L19

Y26

I23

I80 L82D164

S165

Y174

K29

D164S165

Y174

K29

I28

H24

L7

Y11 V23

L16L162 W160

I28

H24

L7

Y11 V23

L16W160

a

b

Figure 4 | Crystal structure of (PLS)3PPR in complex with MORF9. a, The overall architecture of the (PLS)3PPR–MORF9 complex. Left, (PLS)3PPR andMORF9 are coloured slate and grey, respectively. Right, the L motif is embedded in a MORF9 groove. The P-, L- and S-type motifs are coloured yellow,pink and slate, respectively. Helix a and helix b are labelled on each motif. Areas that mediate protein–protein interactions are boxed with dotted lines.b, Interactions between the P-, L-, and S-type motifs and MORF9. Left, hydrophobic interactions between the P- and L-type motifs and MORF9. Middle,hydrogen bond interactions between the L-type motif and MORF9. Right, hydrophobic interactions between the L- and S-type motifs and MORF9. Residuesfrom the P-, L- and S-type motifs and MORF9 are coloured yellow, pink, slate and black, respectively.

ARTICLES NATURE PLANTS

NATURE PLANTS 3, 17037 (2017) | DOI: 10.1038/nplants.2017.37 | www.nature.com/natureplants4

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

undergoes compressed conformational changes; specifically, helix bfrom the L-type motif moves towards the RNA-binding face. Thischange results in increased RNA-binding activity of PLS-type PPRproteins (Fig. 6b).

Usually, PLS-type PPR proteins retain RNA-binding activity. Itis, however, interesting that the designer (PLS)3PPR does not bindto RNA obviously (Fig. 3b and Supplementary Fig. 6). On thebasis of structural alignment of the L-type motif with the P-typemotif (Fig. 1e), we speculate that, without other factors, the di-residues (the 5th and 35th) responsible for RNA base recognitionin the L-type motif of (PLS)3PPR could inefficiently coordinatethe base. And the remaining inconsecutive six repeats (threeP-type and three S-type motifs) in (PLS)3PPR might be not efficientfor RNA binding, which is reminiscent of the dPPR proteins23,33.

The MORF protein family comprises ten members inArabidopsis. Functional studies have revealed that different morfmutants result in distinct RNA editing deficiencies25, implyingunique roles for each MORF protein. Although the centrally con-served MORF domains in this family indicate their similar struc-tural folds (Supplementary Fig. 5) and similar roles in binding toPLS-type PPR proteins, whether individual MORF membersdisplay functional divergence remains to be investigated. Multiplesequence alignments using the designed PLS triplet and all chloro-plast PLS-type PPR proteins from Arabidopsis revealed that anumber of natural PLS-type PPR proteins featured prominentsequence conservation. Thus, these natural PLS-type PPR proteinsmay be able to interact with different MORFs.

To further validate this working model, we have screened a seriesof plastid PLS-type PPR proteins interacting with MORF9 viaco-expression strategy. After tedious trials, the LPA66–MORF9complex could be purified to homogeneity (Supplementary Fig. 10a)for further biochemical validation. The LPA66 is involved in theRNA editing event of psbF (ref. 34). Then we examined theRNA-binding activity of MORF9, LPA66 and LPA66-MORF9complex against psbF (Supplementary Fig. 10b) by EMSA. NoRNA-binding ability was observed for MORF9 (SupplementaryFig. 10c). Both LPA66 and LPA66-MORF9 complex exhibit RNA-binding ability, but the binding affinity of the LPA66–MORF9complex is notably stronger than that of LPA66 alone

(Supplementary Fig. 10d,e). Furthermore, the RNA-binding activityof LPA66 is strikingly improved with the increased amount of MORF9(Supplementary Fig. 10f). These results corroborate the workingmodelthat MORF9 functions in plant plastid RNA editing by increasing theRNA-binding ability of PLS-type PPR proteins.

Unlike their progenitor TPR (tetratricopeptide repeat) proteins,which are known to mediate protein–protein interactions, PPR pro-teins are generally recognized to be RNA-binding proteins. Thecomplex structure from this study reveals important new insightsinto PPR protein molecular architecture, particularly the interactionof PLS-type PPR proteins with MORF proteins through their convexsurface (Fig. 4a), which represents a novel mode of protein–proteininteraction. During the preparation of our manuscript, a studyreported that the P-type PPR protein from human mitochondria,LRPPRC, interacts with another protein, SLIRP, by means of resi-dues involved in RNA recognition at the concave surface35. Toour knowledge, these are the first two examples of interactionsbetween PPR proteins and other proteins. Here, the structuralstudy expands our knowledge that the concave surface of the PLS-typePPR proteins attach to the target RNA, and the convex surface isaccessible to protein–protein interactions. Previous studies havedemonstrated that other editing factors such as ORRM1/2/3 (refs 36and 37), PPO1 (ref. 38), OZ1 (ref. 39) affect the RNA editing andinteract with either MORF or PLS-type PPR proteins in plant plastidsor mitochondria. Thus, we speculate that PLS/PPR–MORF proteincomplex could potentially function as a scaffold to recruit otherediting factors facilitating the functional RNA editosome assemblyin vivo.

The identity of the enzyme that catalyses the C-to-U conversionin plant plastids and mitochondria has been a longstanding ques-tion. In Arabidopsis, nearly half of all PLS-type PPR proteinscontain C-terminal DYW domains. The DYW domain is thoughtto contribute to cytidine deaminase activity, as it contains ahighly conserved signature HxE(x)nCxxC motif that resembles thezinc-binding active site motif in known cytosine deaminase40.Functional studies revealed that the DYW domain is dispensablefor a number of PPR proteins such as CRR22 and CRR28(ref. 41), but is required for other PPR proteins, such as QED1and RARE1 (ref. 42). Moreover, DYW1, which consists of no

MORF9 unboundMORF9 bound

P type

MORF9 unboundMORF9 bound

S type

MORF9–(PLS)3PPR(PLS)3PPR

N

a b

c ed

N C

90°

MORF9 unboundMORF9 bound

L type

Figure 5 | Structural alignment of free (PLS)3PPR with MORF9-bound (PLS)3PPR. a, Conformational changes in (PLS)3PPR on MORF9 binding. MORF9,MORF9-bound (PLS)3PPR and free (PLS)3PPR are coloured grey, slate and yellow, respectively. b, Conformational changes in (PLS)3PPR upon MORF9binding. MORF9 has been removed for clarity. c–e, P-type motif (c), L-type motif (d) and S-type motif structural superimposition (e) upon MORF9 binding.

NATURE PLANTS ARTICLES

NATURE PLANTS 3, 17037 (2017) | DOI: 10.1038/nplants.2017.37 | www.nature.com/natureplants 5

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

more than a DYW domain without any PPR repeats, functions inRNA editing in conjunction with CRR4, which lacks the DYWdomain43. These cases suggest that DYW domains have evolveddivergently. Based on the observation that MORF proteins canform homodimers or heterodimers in vivo25 and interact withPLS-type PPR proteins in our study (Figs 3a and 4a), we speculatethat MORFs might mediate connections to different PLS-typePPR proteins to constitute the functional editosome. As a result,the potential catalytic DYW domains could be supplied in trans ifabsent from some PLS-type PPR proteins. Taken together, ourresults offer new insights into the molecular function of MORFproteins in plant organellar RNA editing.

MethodsSequence design and molecular cloning. All the PLS-type PPR proteins inArabidopsis were collected and submitted to Prosite online analysis. A total of1,117 sequences of 31-amino-acid sequences were identified as the S-type motif.Sequence alignment between the P- and L-type motifs revealed their distinctpreference at the 24th and 27th residues. Based on this discrepancy, we restricted the24th residue to histidine and the 27th residue to not methionine for L-type motif. Asa result, a number of 263 sequences were obtained. Sequence logos were constructedusing WebLogo44. The consensus sequence for the S-type motif isVVVYNALIDMYSKCGLLEEARKVFDEMPEKD; and for the L-type motif,EFTFSSVLKACARLGALELGKQIHGYVIKSGFESD; the P-type motif was used aspreviously described21: VVTYNTLIDGLCKAGKLDEALKLFEEMVEKGIKPD. Toimprove the solubility of the protein, an N-terminal cap comprising residuesRRQGVAPT and a C-terminal solvating helix comprising residuesELTYRRVVESYCRAKRFE from maize PPR10 were added. In this study, weconstructed one, two and three PLS triplets with the same two termini, designated(PLS)1PPR, (PLS)2PPR and (PLS)3PPR. The DNA coding sequences wereoptimized, synthesized by Genewiz (Suzhou, China) and sub-cloned into pET21vector (Novagen) for expression.

The full length of MORF9 was amplified from Arabidopsis cDNA and sub-cloned into pET21 (Novagen). Severe protein degradation was observed for the full-length MORF9 protein. Mass spectrometry was used to identify the enriched proteinfragment from the 59th residue of MORF9, which is far away from the signalcleavage site of residue 23–24 predicted by SignalP 3.0 (ref. 45). Thus, MORF9(residues 59–232) was sub-cloned into pET21 (Novagen) or no-tagged vectorpBB75. The mutant constructs were generated using the two-step PCR method.LPA66 (At5g48910.1) was amplified from genomic DNA without a signal peptidesequence and DYW domain (residues 35–515), and cloned into the modified vectorpET15-MBP, generating MBP-tagged protein. All the constructs were verified byDNA sequencing.

Protein expression and purification. The sequenced plasmid was transformed intoE. coli BL21 (DE3). Lysogeny broth medium (1 l) supplemented with 100 mg ml−1

ampicillin was inoculated with a transformed bacterial pre-culture and shaken at37 °C until the cell density reached an A600 of 1.0–1.2, protein expression wasinduced with 0.2 mM isopropyl-β-D-thiogalactoside at 16 °C for 12 h. The cells werecollected by centrifugation, homogenized in buffer A (25 mM Tris–HCl, pH 8.0,150 mM NaCl) and lysed using a high-pressure cell disrupter (JNBIO, China). Celldebris was removed by centrifugation at 20,000g for 1 h at 4 °C, and the supernatantwas loaded onto a column equipped with Ni2+ affinity resin (Ni-NTA, Qiagen),washed with buffer B (25 mM Tris–HCl, pH 8.0, 150 mM NaCl, 15 mM imidazole)

and eluted with buffer C (25 mM Tris–HCl, pH 8.0, 250 mM imidazole). Theprotein was then separated by anion exchange chromatography (Source 15Q,GE Healthcare) using a linear NaCl gradient in buffer A. The purified protein wasconcentrated and subjected to gel filtration chromatography (Superdex-200 Increase10/300 GL, GE Healthcare) in a buffer containing 25 mM Tris-HCl, pH 8.0, 150 mMNaCl, 5 mM dithiothreitol. Purity of the proteins was examined using sodiumdodecyl sulfate polyacrylamide gel electrophoresis (SDS–PAGE) and visualized byCoomassie blue staining through all purification processes. The peak fractions werepooled for crystallization immediately or stored at −80 °C. Selenomethionine-labelled (PLS)3PPR or MORF9 and the mutant proteins were purified similarly asthe wild-type proteins. For co-expression analysis, the pET15–MBP–LPA66 plasmidand pBB75–MORF9 plasmid were co-transformed into E. coli BL21 (DE3). Proteininduction and purification was performed as described above.

Size exclusion chromatography. To examine the interaction between the PLStriplets and MORF9 protein, the recombinant (PLS)1PPR, (PLS)2PPR, (PLS)3PPRwas individually incubated with MORF9 protein for an hour at 4 °C and furtherpurified by SEC (Superdex-200 Increase 10/300 GL, GE Healthcare) in a buffercontaining 25 mM Tris-HCl, pH 8.0, 150 mM NaCl, 5 mM dithiothreitol.Fractions at same elution volume from individual injections were examined bySDS–PAGE. The peak fractions of (PLS)1PPR–MORF9, (PLS)2PPR–MORF9 and(PLS)3PPR–MORF9 were subjected for AUC analysis.

Crystallization. Protein was concentrated to 10 mg ml−1 before crystallization trials.Crystallizations were performed using the sitting-drop vapour diffusion method at18 °C by mixing equal volumes (1 µl) of protein with reservoir solution. The rod-likecrystal of (PLS)3PPR protein appeared overnight and grew to full size withinthree days in the well buffer containing 0.1 M sodium HEPES, pH 7.0, 0.1 Mmagnesium chloride, 15% (v/v) polyethylene glycol 4,000. As our initial attempts tocrystallize MORF9 (residue 59–232) were unsuccessful, a series of truncations andcysteine mutations were performed. After several rounds of optimization, theneedle-shaped crystal of MORF9 (residue 75–196, C85S/C187S) appeared in the wellbuffer containing 0.1 M MES, pH 5.9, 0.1 M sodium bromide, 0.2 M sodiumthiocyanate, 22% (v/v) polyethylene glycol 3,350. The fusiform crystal for the(PLS)3PPR–MORF9 complex appeared in the well buffer containing0.1 M sodium acetate, pH 5.3, 1% (v/v) polyethylene glycol 4,000, 17% (v/v)2-methyl-2,4-pentanediol. These crystals were flash-frozen in liquid nitrogenand cryoprotected by adding glycerol to a final concentration of 10–20%.

Data collection and structure determination. All the diffraction data were collectedat Shanghai Synchrotron Research Facility (SSRF) on beamline BL17U or BL19U,integrated, and processed with the HKL2000 program suite and XDS packages46.Further data processing was carried out using CCP4 suit47. The structure of MORF9and (PLS)3PPR were solved via single-wavelength anomalous diffraction (SAD)method. The structure of (PLS)3PPR–MORF9 was solved by molecular replacementwith the newly solved apo (PLS)3PPR and MORF9 as the search model using theprogram PHASER48. All the structures were iteratively built with COOT49 andrefined with PHENIX program50. Data collection and structure refinement statisticsare summarized in Supplementary Table 1. All figures were generated using theprogram PyMOL (http://www.pymol.org/).

Electrophoretic mobility shift assay. The ssRNA oligonucleotides (5′-g(UNU)3uc-3′,N represents A, G, C or U) were radiolabelled at the 5′ end with [γ-32P]ATP(PerkinElmer), catalysed by T4 polynucleotide kinase (Takara). The labelled RNAwas purified by centrifugation through a 2 cm bed of G-25 size exclusion resinpacked in a mini-spin column (GE Healthcare) and centrifuged at 750g for 2 min.For EMSA, proteins were incubated with approximately 2 nM 32P-labelled probe in

RNA

N

a b

C

RNA-bound dPPR

(PLS)3PPR

MORF9-bound (PLS)3PPRStrong RNA-binding activity

Weak RNA-binding activity (PLS)3PPR

(PLS)3PPR Conformational changes

MORF9

RNA

RNA

Figure 6 | Proposed working model of MORF9 in plant plastid RNA editing. a, Structural alignment of free (PLS)3PPR, MORF9-bound (PLS)3PPR andRNA-bound dPPR (PDB ID: 5I9G). RNA, RNA-bound dPPR, free (PLS)3PPR and MORF9-bound (PLS)3PPR are coloured green, brown, yellow and slate,respectively. b, The proposed working model of the improved RNA-binding activity of (PLS)3PPR upon MORF9 binding.

ARTICLES NATURE PLANTS

NATURE PLANTS 3, 17037 (2017) | DOI: 10.1038/nplants.2017.37 | www.nature.com/natureplants6

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

final binding reactions containing 25 mM Tris-HCl, pH 8.0, 5 mM MgCl2, 40 mMNaCl, 50 ng ml−1 heparin and 10% glycerol for 30 min at 25 °C. Then the reactionswere resolved on 6% native acrylamide gels (37.5:1 acrylamide:bis-acrylamide) in0.5× Tris-glycine buffer under an electric field of 15 V cm−1 for 1.5 h. Gels werevisualized on a phosphor screen (Amersham Biosciences) using a Typhoon TrioImager (Amersham Biosciences). The results are representative of three independentexperiments. The dissociation constant (Kd) was determined from the concentrationof protein at which 50% of the RNA probe was bound.

Isothermal titration calorimetry. ITC experiments were performed at 25 °C usingAuto-iTC200 titration calorimetry (MicroCal). To assess the RNA-binding affinityof proteins, RNA (250 µM) was dissolved in reaction buffer containing 20 mMHEPES, pH 7.5, and 150 mM NaCl (40 µl) and titrated against 20 µM protein in thesame buffer (200 µl). To analyse the binding ability of (PLS)1PPR to MORF9,MORF9 or its mutant (100 µM) was dissolved in reaction buffer containing 25 mMTris-HCl, pH 8.0, and 150 mM NaCl (40 µl) and titrated against 10 µM (PLS)1PPRin the same buffer (200 µl). The first injection (0.5 µl) was followed by 19 injectionsof 2 µl. The heat of dilution values was measured by injecting RNA or MORF9 intobuffer alone in each experiment. The values were subtracted from the experimentalcurves before data analysis. The stirring rate was 750 r.p.m. The MicroCal ORIGINsoftware supplied with the instrument was used to determine the site-binding modelthat produced a good fit (low × 2 value) for the resulting data.

Analytical ultracentrifugation. The stoichiometry of MORF9 in complex with(PLS)1PPR, (PLS)2PPR or (PLS)3PPR were investigated by AUC experiments, whichwere performed in a Beckman Coulter XL-I analytical ultracentrifuge usingtwo-channel centrepieces. MORF9–(PLS)1PPR, MORF9–(PLS)2PPR and MORF9–(PLS)3PPR were in solutions containing 25 mM Tris-HCl, pH 8.0, 150 mM NaCl.Data were collected by absorbance detection at 18 °C for proteins at a concentrationof 0.8∼1 mg ml−1 at a rotor speed of 45,000 r.p.m. The SV-AUC data were globallyanalysed using the SEDFIT program and fitted to a continuous c(s) distributionmodel to determine the molecular mass of each complex.

Data availability. Coordinates and structure factors have been deposited in theProtein Data Bank under accession codes 5GI0 (MORF9), 5IZW ((PLS)3PPR) and5IWW ((PLS)3PPR–MORF9 complex).

Received 8 September 2016; accepted 20 February 2017;published 10 April 2017

References1. Takenaka, M., Zehrmann, A., Verbitskiy, D., Hartel, B. & Brennicke, A.

RNA editing in plants and its evolution. Annu. Rev. Genet. 47, 335–352 (2013).2. Liu, Y. J., Xiu, Z. H., Meeley, R. & Tan, B. C. Empty pericarp5 encodes a

pentatricopeptide repeat protein that is required for mitochondrial RNA editingand seed development in maize. Plant Cell 25, 868–883 (2013).

3. Huang, H. et al. RNA editing of the IQ domain in Cav1.3 channels modulatestheir Ca2+-dependent inactivation. Neuron 73, 304–316 (2012).

4. Chiu, Y. L. et al. Cellular APOBEC3G restricts HIV-1 infection in resting CD4+

T cells. Nature 435, 108–114 (2005).5. Zhu, Q. et al. The Arabidopsis thaliana RNA editing factor SLO2, which affects

the mitochondrial electron transport chain, participates in multiple stress andhormone responses. Mol. Plant 7, 290–310 (2014).

6. Chateigner-Boutin, A. L. et al. CLB19, a pentatricopeptide repeat proteinrequired for editing of rpoA and clpP chloroplast transcripts. Plant J. 56,590–602 (2008).

7. Sosso, D. et al. PPR2263, a DYW-Subgroup Pentatricopeptide repeat protein, isrequired for mitochondrial nad5 and cob transcript editing, mitochondrionbiogenesis, and maize growth. Plant Cell 24, 676–691 (2012).

8. Hwang, T. et al. Dynamic regulation of RNA editing in human braindevelopment and disease. Nat. Neurosci. 19, 1093–1099 (2016).

9. Chen, L. et al. Recoding RNA editing of AZIN1 predisposes to hepatocellularcarcinoma. Nat. Med. 19, 209–216 (2013).

10. Nishikura, K. Functions and regulation of RNA editing by ADAR deaminases.Annu. Rev. Biochem. 79, 321–349 (2010).

11. Mehta, A. & Driscoll, D. M. Identification of domains in apobec-1complementation factor required for RNA binding and apolipoprotein-BmRNA editing. RNA 8, 69–82 (2002).

12. Covello, P. S. & Gray, M. W. RNA editing in plant mitochondria. Nature 341,662–666 (1989).

13. Gualberto, J. M., Lamattina, L., Bonnard, G., Weil, J. H. & Grienenberger, J. M.RNA editing in wheat mitochondria results in the conservation of proteinsequences. Nature 341, 660–662 (1989).

14. Hiesel, R., Wissinger, B., Schuster, W. & Brennicke, A. RNA editing in plantmitochondria. Science 246, 1632–1634 (1989).

15. Hoch, B., Maier, R. M., Appel, K., Igloi, G. L. & Kossel, H. Editing of a chloroplastmRNA by creation of an initiation codon. Nature 353, 178–180 (1991).

16. Barkan, A. & Small, I. Pentatricopeptide repeat proteins in plants. Annu. Rev.Plant Biol. 65, 415–442 (2014).

17. Bentolila, S., Oh, J., Hanson, M. R. & Bukowski, R. Comprehensive high-resolution analysis of the role of an Arabidopsis gene family in RNA editing.PLoS Genet. 9, e1003584 (2013).

18. Pyo, Y. J., Kwon, K. C., Kim, A. & Cho, M. H. Seedling Lethal1, apentatricopeptide repeat protein lacking an E/E+ or DYW domain inArabidopsis, is involved in plastid gene expression and early chloroplastdevelopment. Plant Physiol. 163, 1844–1858 (2013).

19. Lurin, C. et al. Genome-wide analysis of Arabidopsis pentatricopeptide repeatproteins reveals their essential role in organelle biogenesis. Plant Cell 16,2089–2103 (2004).

20. Yin, P. et al. Structural basis for the modular recognition of single-stranded RNAby PPR proteins. Nature 504, 168–171 (2013).

21. Shen, C. et al. Structural basis for specific single-stranded RNArecognition by designer pentatricopeptide repeat proteins. Nat. Commun. 7,11285 (2016).

22. Ke, J. et al. Structural basis for RNA recognition by a dimeric PPR-proteincomplex. Nat. Struct. Mol. Biol. 20, 1377–1382 (2013).

23. Coquille, S. et al. An artificial PPR scaffold for programmable RNA recognition.Nat. Commun. 5, 5729 (2014).

24. Gully, B. S. et al. The design and structural characterization of a syntheticpentatricopeptide repeat protein. Acta Crystallogr. D 71, 196–208 (2015).

25. Takenaka, M. et al. Multiple organellar RNA editing factor (MORF) familyproteins are required for RNA editing in mitochondria and plastids of plants.Proc. Natl Acad. Sci. USA 109, 5104–5109 (2012).

26. Bentolila, S. et al. RIP1, a member of an Arabidopsis protein family, interactswith the protein RARE1 and broadly affects RNA editing. Proc. Natl Acad. Sci.USA 109, E1453–E1461 (2012).

27. Cheng, S. et al. Redefining the structural motifs that determine RNA binding andRNA editing by pentatricopeptide repeat proteins in land plants. Plant J. 85,532–547 (2016).

28. Barkan, A. et al. A combinatorial amino acid code for RNA recognition bypentatricopeptide repeat proteins. PLoS Genet. 8, e1002910 (2012).

29. Yagi, Y., Hayashi, S., Kobayashi, K., Hirayama, T. & Nakamura, T. Elucidation ofthe RNA recognition code for pentatricopeptide repeat proteins involved inorganelle RNA editing in plants. PLoS ONE 8, e57286 (2013).

30. Holm, L., Kaariainen, S., Rosenstrom, P. & Schenkel, A. Searching proteinstructure databases with DaliLite v.3. Bioinformatics 24, 2780–2781 (2008).

31. Glass, F., Hartel, B., Zehrmann, A., Verbitskiy, D. & Takenaka, M. MEF13requires MORF3 and MORF8 for RNA editing at eight targets in mitochondrialmRNAs in Arabidopsis thaliana. Mol. Plant 8, 1466–1477 (2015).

32. Kotera, E., Tasaka, M. & Shikanai, T. A pentatricopeptide repeat protein isessential for RNA editing in chloroplasts. Nature 433, 326–330 (2005).

33. Shen, C. et al. Specific RNA recognition by designer pentatricopeptide repeatprotein. Mol. Plant 8, 667–670 (2015).

34. Cai, W. et al. LPA66 is required for editing psbF chloroplast transcripts inArabidopsis. Plant Physiol. 150, 1260–1271 (2009).

35. Spahr, H. et al. SLIRP stabilizes LRPPRC via an RRM-PPR protein interface.Nucleic Acids Res. 44, 6868–6882 (2016).

36. Sun, T. et al. An RNA recognition motif-containing protein is required forplastid RNA editing in Arabidopsis and maize. Proc. Natl Acad. Sci. USA 110,E1169–E1178 (2013).

37. Shi, X., Hanson, M. R. & Bentolila, S. Two RNA recognition motif-containingproteins are plant mitochondrial editing factors. Nucleic Acids Res. 43,3814–3825 (2015).

38. Zhang, F. et al. Tetrapyrrole biosynthetic enzyme protoporphyrinogen IXoxidase 1 is required for plastid RNA editing. Proc. Natl Acad. Sci. USA 111,2023–2028 (2014).

39. Sun, T. et al. A zinc finger motif-containing protein is essential for chloroplastRNA editing. PLoS Genet. 11, e1005028 (2015).

40. Boussardon, C. et al. The cytidine deaminase signature HxE(x)nCxxC of DYW1binds zinc and is necessary for RNA editing of ndhD-1. New Phytol. 203,1090–1095 (2014).

41. Okuda, K. et al. Pentatricopeptide repeat proteins with the DYW motif havedistinct molecular functions in RNA editing and RNA cleavage in Arabidopsischloroplasts. Plant Cell 21, 146–156 (2009).

42. Wagoner, J. A., Sun, T., Lin, L. & Hanson, M. R. Cytidine deaminase motifswithin the DYW domain of two pentatricopeptide repeat-containing proteinsare required for site-specific chloroplast RNA editing. J. Biol. Chem. 290,2957–2968 (2015).

43. Boussardon, C. et al. Two interacting proteins are necessary for the editing of theNdhD-1 site in Arabidopsis plastids. Plant Cell 24, 3684–3694 (2012).

44. Crooks, G. E., Hon, G., Chandonia, J. M. & Brenner, S. E. Weblogo: a sequencelogo generator. Genome Res. 14, 1188–1190 (2004).

45. Bendtsen, J. D., Nielsen, H., von Heijne, G. & Brunak, S. Improved prediction ofsignal peptides: SignalP 3.0. J. Mol. Biol. 340, 783–795 (2004).

46. Otwinowski, Z. & Minor, W. Processing of X-ray diffraction data collected inoscillation mode. Methods Enzymol. 276, 307–326 (1997).

47. Winn, M. D. et al. Overview of the CCP4 suite and current developments.Acta Crystallogr. D 67, 235–242 (2011).

NATURE PLANTS ARTICLES

NATURE PLANTS 3, 17037 (2017) | DOI: 10.1038/nplants.2017.37 | www.nature.com/natureplants 7

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.

48. McCoy, A. J. et al. Phaser crystallographic software. J. Appl. Crystallogr. 40,658–674 (2007).

49. Emsley, P. & Cowtan, K. Coot: model-building tools for molecular graphics.Acta Crystallogr. D 60, 2126–2132 (2004).

50. Adams, P. D. et al. PHENIX: building new software for automatedcrystallographic structure determination. Acta Crystallogr. D 58,1948–1954 (2002).

AcknowledgementsWe thank the staff from BL17U/BL19U1/beamline of the National Centre for ProteinSciences Shanghai (NCPSS) at the Shanghai Synchrotron Radiation Facility for assistanceduring data collection, and research associates at the Centre for Protein Research,Huazhong Agricultural University, for technical support. This work was supported byfunds from the Ministry of Science and Technology (grant 2015CB910900), theFok Ying-Tong Education Foundation (grant 151021), the National Science Foundation ofChina (grant 91535301), the Fundamental Research Funds for the Central Universities(programme no. 2014PY026, no. 2015PY219 and no. 2014JQ001), Huazhong AgriculturalUniversity Scientific & Technological Self-innovation Foundation (programmeno. 2013RC013) and the China Postdoctoral Science Foundation (grant 2015M572163).

Author contributionsJ.Y., Q.Z., R.L., T.Z. and P.Y. designed all experiments. J.Y., Q.Z., L.L. and F.R. performedprotein expression, purification and crystallization. Z.G. determined all of the structures.J.Y., Q.Z. and Q.W. carried out biochemical assays. J.Y. and P.Y. wrote the manuscript.All authors discussed the results and commented on the manuscript.

Additional informationSupplementary information is available for this paper.

Reprints and permissions information is available at www.nature.com/reprints.

Correspondence and requests for materials should be addressed to P.Y.

How to cite this article: Yan, J. et al.MORF9 increases the RNA-binding activity of PLS-typepentatricopeptide repeat protein in plastid RNA editing. Nat. Plants 3, 17037 (2017).

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.

Competing interestsThe authors declare no competing financial interests.

ARTICLES NATURE PLANTS

NATURE PLANTS 3, 17037 (2017) | DOI: 10.1038/nplants.2017.37 | www.nature.com/natureplants8

© 2017 Macmillan Publishers Limited, part of Springer Nature. All rights reserved.