Robust Survival-Based RNA Interference of Gene Families ...blots, is stymied. Furthermore, certain...

Breakthrough Technologies

Robust Survival-Based RNA Interference of GeneFamilies Using in Tandem Silencing ofAdenine Phosphoribosyltransferase1[OPEN]

Robert G. Orr,a Stephen J. Foley,b Catherine Sherman,a,c Isidro Abreu,d Giulia Galotto,a Boyuan Liu,a

Manuel González-Guerrero,d and Luis Vidalia,b,2,3

aDepartment of Biology and Biotechnology, Worcester Polytechnic Institute, Worcester, Massachusetts 01609bDepartment of Chemistry and Biochemistry, Worcester Polytechnic Institute, Worcester, Massachusetts 01609cBioinformatics and Computational Biology Program, Worcester Polytechnic Institute, Worcester,Massachusetts 01605dCentro de Biotecnología y Genómica de Plantas (UPM-INIA), Universidad Politécnica de Madrid, Madrid28223, Spain

ORCID IDs: 0000-0002-6281-4822 (S.J.F.); 0000-0002-1647-9883 (I.A.); 0000-0002-1014-9509 (B.L.); 0000-0001-7334-5286 (M.G.-G.);0000-0002-5835-4587 (L.V.)

RNA interference (RNAi) enables flexible and dynamic interrogation of entire gene families or essential genes without the needfor exogenous proteins, unlike CRISPR-Cas technology. Unfortunately, isolation of plants undergoing potent gene silencingrequires laborious design, visual screening, and physical separation for downstream characterization. Here, we developed anadenine phosphoribosyltransferase (APT)-based RNAi technology (APTi) in Physcomitrella patens that improves upon themultiple limitations of current RNAi techniques. APTi exploits the prosurvival output of transiently silencing APT in thepresence of 2-fluoroadenine, thereby establishing survival itself as a reporter of RNAi. To maximize the silencing efficacy ofgene targets, we created vectors that facilitate insertion of any gene target sequence in tandem with the APT silencing motif. Wetested the efficacy of APTi with two gene families, the actin-dependent motor, myosin XI (a,b), and the putative chitin receptorLyk5 (a,b,c). The APTi approach resulted in a homogenous population of transient P. patensmutants specific for our gene targetswith zero surviving background plants within 8 d. The observed mutants directly corresponded to a maximal 93% reduction ofmyosin XI protein and complete loss of chitin-induced calcium spiking in the Lyk5-RNAi background. The positive selectionnature of APTi represents a fundamental improvement in RNAi technology and will contribute to the growing demand fortechnologies amenable to high-throughput phenotyping.

Loss-of-function studies have long served as buildingblocks of our understanding of biological processes.

Currently, the researcher is confronted with an ever-growing toolbox to perturb gene function at the gene,transcript, or protein level, with each presenting itsunique challenges and insights (Housden et al., 2017).CRISPR-Cas technology has facilitated targeted geneticknockouts in previously intractable systems, but itsefficacy is limited by accessibility of the genetic locusand is subject to off-target effects (Horlbeck et al., 2016;Jensen et al., 2017; Verkuijl and Rots, 2019). Further-more, genetic knockouts cannot isolate essential genesor reversibly interrogate developmental, tissue, or cell-specific functions of a given gene or splice variant.Importantly, recent work has questioned the lastingdogma that genetic knockouts are the “gold standard”loss-of-function approach (Rossi et al., 2015; Smits et al.,2019). These studies demonstrate genetic compensationor residual protein expression, signifying that geneticalterations alone are insufficient to directly infer genefunction. Therefore, reliable inferences of loss-of-func-tion studies will require integration of multiple inde-pendent approaches (Deans et al., 2016), ideally in ahigh-throughput manner to maximize confidence.RNA interference (RNAi) is a popular gene-silencing

strategy that does not depend upon exogenous proteins,

1This work was supported by the National Science Foundation(grant no. NSF-MCB 1253444 to L.V.), the National Institutes ofHealth (grant no. 1R15GM134493 to L.V.), and Fulbright Scholar Pro-gram (to L.V.), and a Juan de la Cierva-Formación postdoctoral fel-lowship from Ministerio de Ciencia e Innovación, Ministerio deUniversidades (grant no. FJCI–2017–33222 to I.A.).

2Author for contact: [email protected] author.The author responsible for distribution of materials integral to the

findings presented in this article in accordance with the policy de-scribed in the Instructions for Authors (www.plantphysiol.org) is:Luis Vidali ([email protected]).

R.G.O. and L.V. designed and supervised the research; L.V. con-ceived the project; S.J.F. created the APTi constructs; R.G.O. devel-oped and optimized the APTi microscope assay and immunoblots;G.G. and B.L. contributed to the initial optimization of microscopeassay; I.A., M.G.-G., and L.V. participated in creating the GCaMP lineand contributed with C.S. in Lyk5-RNAi construct creation; C.S. per-formed and analyzed the Lyk5-RNAi experiments; R.G.O. wrote themanuscript with contributions from S.J.F., C.S., G.G., and L.V.

[OPEN]Articles can be viewed without a subscription.www.plantphysiol.org/cgi/doi/10.1104/pp.20.00865

Plant Physiology�, October 2020, Vol. 184, pp. 607–619, www.plantphysiol.org � 2020 American Society of Plant Biologists. All Rights Reserved. 607

https://plantphysiol.orgDownloaded on June 2, 2021. - Published by Copyright (c) 2020 American Society of Plant Biologists. All rights reserved.

https://orcid.org/0000-0002-6281-4822https://orcid.org/0000-0002-6281-4822https://orcid.org/0000-0002-1647-9883https://orcid.org/0000-0002-1647-9883https://orcid.org/0000-0002-1014-9509https://orcid.org/0000-0002-1014-9509https://orcid.org/0000-0001-7334-5286https://orcid.org/0000-0001-7334-5286https://orcid.org/0000-0002-5835-4587https://orcid.org/0000-0002-5835-4587https://orcid.org/0000-0002-6281-4822https://orcid.org/0000-0002-1647-9883https://orcid.org/0000-0002-1014-9509https://orcid.org/0000-0001-7334-5286https://orcid.org/0000-0002-5835-4587http://crossmark.crossref.org/dialog/?doi=10.1104/pp.20.00865&domain=pdf&date_stamp=2020-09-29http://dx.doi.org/10.13039/100000001http://dx.doi.org/10.13039/100000057http://dx.doi.org/10.13039/100000057mailto:[email protected]://www.plantphysiol.orgmailto:[email protected]://www.plantphysiol.org/cgi/doi/10.1104/pp.20.00865https://plantphysiol.org

such as Cas9, and enables reversible reduction of proteinlevels through targeted degradation of mRNA (Small,2007). RNAi’s ease of use and flexibility lends itself asan indispensable complement to genetic editing tech-niques. Traditionally, RNAi in plants is induced throughexpression of long inverted repeats that self-base pair toform double-stranded RNA (dsRNA), which is thenprocessed into multiple small interfering RNAs (siR-NAs) and targeted to complementary sequences withinmRNA (Chuang and Meyerowitz, 2000; Hannon, 2002;Baulcombe, 2004). Importantly, a single dsRNA target-ing one gene can simultaneously silence multiple geneswith sufficient similarity, or a single dsRNA can begenerated that contains multiple gene targets in tandemfor simultaneous silencing (Vidali et al., 2007; Li et al.,2013). The expression of dsRNAs can be specificallymodulated, either through induction or unique pro-moters, which allows developmental and cell-type-specific reduction of even essential gene products(Byzova et al., 2004; Nakaoka et al., 2012; Miki et al.,2015; Liu and Yoder, 2016). Despite these advantages,RNAi is hindered by variable efficacy of target genesilencing and potential off-target effects (Xu et al., 2006).However, some argue the prevalence of off-target ef-fects is overstated (Zimmer et al., 2019), and impor-tantly an ideal RNAi experiment should demonstraterescue of the gene silencing phenotype (Vidali et al.,2009, 2010; Ding et al., 2018). Work using artificialmicroRNAs (amiRNAs) attempts to circumvent thelimitations of traditional dsRNA-based RNAi by engi-neering a single siRNA (Schwab et al., 2006; de laGutiérrez-Nava et al., 2008). Although amiRNA tech-nology ameliorates possible off-targets derived fromthe initial dsRNA, evidence in Physcomitrella patensshows generation of additional siRNAs upon cleavageof the amiRNA target, potentially negating the speci-ficity of the amiRNA (Khraiwesh et al., 2008). Further-more, amiRNAs display variable silencing efficiency,thereby necessitating screening of multiple amiRNAsand limiting experimental throughput (Li et al., 2013;Zhang et al., 2018). To date, no RNAi method addressesanother potential source of variability of silencing:transcriptional silencing of the RNAi transgene itself,likelymediated byDicer-like3 (DCL3;Morel et al., 2000;Fusaro et al., 2006; Small, 2007).

Generating and characterizing loss-of-function mu-tants using gene silencing methods is fundamentally atwo-step process: (1) the implementation of a specificgene-silencing technique; and (2) the identification andisolation of the target(s) undergoing gene silencing forfurther characterization. Recent advancements in gene-silencing technologies have focused on enhancing theflexibility and robustness of step 1 (Hauser et al., 2013;Zhang et al., 2018) but fail to address the practicallimitations imposed by step 2. Identification of activelysilencing mutants has been eased by coupling the si-lencing target of interest to a reporter, such as a nuclear-localized fluorescent protein (Bezanilla et al., 2005;Vidali et al., 2007, 2010; van Gisbergen et al., 2018;Zhang et al., 2018). When paired with an automated or

semiautomated image acquisition and analysis pipelinethe burden of identifying silencing mutants is sub-stantially mitigated (Wu and Bezanilla, 2012; Galottoet al., 2019). Nevertheless, the tedium of manually iso-lating silencing plants remains. This limitation isa consequence of traditional gene-silencing constructdesign, whereby the silencing module is regulated in-dependently from the selectable marker. A typicalgene-silencing experiment will contain a heterogeneouspopulation of actively silencing plants, presumably aresult of the plant silencing the exogenous silencingmodule to rescue itself (Fusaro et al., 2006; Khraiweshet al., 2010). This transcriptional-based silencing notonly increases variability of target silencing, but whencoupled with visual screening, it exacerbates themanual labor required to isolate mutants. Therefore,downstream characterization of silencing plants, such asreverse transcription quantitative PCR and immuno-blots, is stymied. Furthermore, certain reporter-basedsilencing limits the experimental scope to testing onlywithin established reporter lines (Bezanilla et al., 2003;Nakaoka et al., 2012). For example, using a fluorescentreporter to infer silencing could complicate the subse-quent use of other fluorescent outputs, such as biosen-sors or fluorescent fusions for protein localization, tocharacterize mutant function.

Here, we generated a modular, Gateway-basedRNAi construct that couples silencing any gene(s) ofinterest in tandem with silencing of adenine phosphori-bosyltransferase (APT-interference/APTi). This approachresults in near-undetectable levels of background (non-silencing) transformants, thereby trivializing the iden-tification and isolation of actively silencing plants to thesimple observation and harvesting of all living plants.Unlike traditional antibiotic positive selection followedby visual screening, our new construct simultaneouslyselects for transformation and silencing through pos-itive selection alone. We achieved this by exploitingthe ubiquitous selectable marker system APT, whichconverts purine analogs to cytotoxic nucleotides(Schaff, 1994). The APT loss-of-function selectionsystem has been successfully used in plants (Moffattand Somerville, 1988; Charlot et al., 2014), mammals(Schaff et al., 1990), and bacteria (Levine and Taylor,1982), but to our knowledge, all experiments with APTinvolve stable genetic mutants. To quantitativelyevaluate the performance and robustness of our APTisystem, we separately targeted two gene families,myosin XI (a,b) and Lyk5 (a,b,c), using the plant modelorganism P. patens. Targeting the myosin XI familyusing APTi resulted in an exclusively mutant surviv-ing population, displaying the characteristic loss-of-growth phenotype (Vidali et al., 2010) and a .90%reduction of target myosin XI protein abundance.Additionally, APTi enabled rapid functional analysisof the previously uncharacterized P. patens Lyk5 chitinreceptor family, which resulted in complete desensi-tization of P. patens to chitin. Together, APTi-basedsilencing efficacy far surpasses other dsRNA-basedmethods (Vidali et al., 2007, 2010; Nakaoka et al., 2012;

608 Plant Physiol. Vol. 184, 2020

Orr et al.


https://plantphysiol.org

Guo et al., 2019), is comparable to silencing efficiencies foroptimized amiRNAs (Zhang et al., 2018), and simplifiesdownstream analysis for functional genomics.

RESULTS

Development of a Positive Selection RNAi MethodologyBased on Endogenous APT Activity

Current RNAi methods produce a range of pheno-type severity due to variable silencing efficiencies. Thisvariability necessitates optimization experiments thatscreen fluorescent reporters to maximize silencing bygene sequence targets (Li et al., 2013; Zhang et al., 2018).We sought to simultaneously improve RNAi silencingefficiency and streamline characterization of RNAimutants by coupling silencing of the gene target with asurvival advantage. Previous work in P. patens directlycoupled the sequence of a stably integrated, constitu-tively expressed reporter, such as a fluorescent proteinand/or GUS, to the gene target sequence in invertedrepeats (Bezanilla et al., 2003, 2005; Nakaoka et al.,2012). Therefore, expression would result in dsRNAformation and coreduction of the intracellular reporterand the coupled target gene.We reasoned that couplingof a lethal reporter sequence in tandem with any othergene sequence would result in maximal cosilencing toensure silencing of the lethal reporter, thereby pro-moting survival.The APT gene has been frequently used as a reporter

to evaluate gene-targeting efficiency (Schaefer, 2001;Charlot et al., 2014). Functional APT converts adenineanalogs, such as 2-fluoroadenine (2-FA), to cytotoxicnucleotides (Schaff, 1994). Therefore, sufficient reduc-tion of APT activity will impart resistance to 2-FA, butthis has only been demonstrated in genetic knockouts.To test if silencing of PpAPT (Pp3c8_16590) effectivelyconferred survival to plants grown on 2-FA, we inser-ted an APT-targeting sequence into a previously de-veloped RNAi vector that also contains a reportertargeting sequence (Bezanilla et al., 2005). To maximizesilencing efficacy, we generated an APT targeting se-quence consisting of the 59 untranslated region (UTR;179 bp) and first 210 bp of the APT gene. The APT si-lencing construct (APT-RNAi) conferred resistance towild-type P. patens cultured on 1.25 mg mL21 2-FA,whereas no plants survived on 2-FA when transformedwith a control plasmid lacking the APT silencing se-quence (control-RNAi; Fig. 1A). This result clearlyestablishes survival on 2-FA paired with APT targetingas a conspicuous phenotypic reporter of activesilencing.Our previous APT silencing experiment demon-

strated feasibility, but further development of the tech-nique was constrained by the available construct. Asmentioned above, the first iteration of APTi was insertedinto a vector created specifically for RNAi (Bezanillaet al., 2005). This construct included a pair of invertedGateway sites coupled to a target sequence (GUS) for an

internal reporter of active silencing. The reporter se-quence targeted a nuclear-localized GFP:GUS fusionprotein, thereby requiring the use of a special transgenicline for RNAi experiments (Bezanilla et al., 2003). Inprinciple, our APTi would not require any specific mossline and instead would permit the researcher to performRNAi experiments in any genetic background. There-fore, we replaced the GUS reporter target sequence withthe APT target sequence while maintaining the invertedGateway cassettes and loop region and named the con-struct pGAPi (Fig. 1B). This construct allows straight-forward insertion of any gene sequence and ensuresfusion to the APT target, thus permitting inference ofgene silencing within surviving plants. Importantly,entire gene families can be targeted by simple insertionof a conserved sequence of ;400 bp or concatenation ofindividual target sequences followed by cloning into ourpGAPi (Fig. 1B). Additionally, we created another APTiconstruct, named pAPi, which lacks the internal Gate-way cassettes to serve as a “positive control” for anyRNAi experiment (Fig. 1B). Together, these constructsserve as the foundation for the APTi system and ad-vance positive selection as an effective reporter of genesilencing.

APTi Experimental Design forHigh-Throughput Phenotyping

Although our APTi system clearly selects for silenc-ing plants cultured over a 2-week period by visual in-spection (Fig. 1A), we sought to establish a rapid,semiautomated microscopy assay using APTi for plantphenotyping. A fundamental attribute of any high-throughput assay is the effective and automated dis-crimination between objects of interest and backgroundthat co-occupy the same space. Our APTi system sim-plifies the problem of automated separation. Unlikefluorescence reporter systems where the reporter signalis continuous and exhibits natural variation, APTi re-duces separation of silencing and nonsilencing plants tothe binary decision of alive or dead. We reasoned thatchlorophyll autofluorescence could function as a proxyfor plant survival, thereby enabling automated detec-tion of alive, and therefore actively silencing, plants. Inprinciple, all plants successfully transformed with aconstruct silencing APT, pAPi (Fig. 1B), will survive on2-FA medium, whereas plants transformed with anRNAi construct silencing a nonexistent reporter, pUGi(Bezanilla et al., 2005), will die.We tested this by using the experimental design il-

lustrated in Figure 2. P. patens protoplasts were trans-formedwith either pAPi or pUGi, allowed to regeneratefor 4 d, then transferred to growth medium supple-mented with 1.25 mg mL21 2-FA. Importantly, duringthe optimization phase of this assay, we observedsubstantial variability in the outcome of 2-FA selection.We empirically determined that starting the selec-tion 4 d posttransformation and making the 2-FA se-lection plates fresh on the day of selection mitigated

Plant Physiol. Vol. 184, 2020 609

APT-Based RNA Interference Enables Positive Selection



essentially all experimental variability. We stronglysuggest first optimizing the 2-FA selection concentra-tion when applying the APTi system, as our resultswere all obtained using a single lot of 2-FA from Oak-wood Chemical.

Following 4 d on 2-FA-supplemented growth me-dium, cultureswere removed from the growth chamberfor protein extraction andmicroscopy analysis (Fig. 2B).To facilitate high-throughput phenotyping and removehuman bias, we used an epifluorescent microscopeequipped with an automated stage integrated withimage tiling and stitching software. This enabled largeregion-of-interest (ROI) acquisition, with the size of theROI only constrained by the memory available to thecomputer. For our experiments, every composite imagewas constructed from a 12 3 12 grid of single images,with a 15% overlap, which corresponds to an ROI sur-face area of approximately 1.8 cm2.

To simplify downstream image segmentation, allplants were stained with calcofluor to label the cellwall before imaging. Each individual image contained

two channels, the chlorophyll autofluorescence andthe calcofluor signal. Visualization of the chlorophyllsignals of APT-RNAi and control-RNAi clearly re-veals the efficacy of the APTi system: not a singlecontrol-RNAi plant survives and does not markedlygrow beyond the initial regeneration size (Fig. 2A).This result is reproducible, as the chlorophyll signalacross three independent experiments for control-RNAi plants consistently clustered below a charac-teristic intensity (Fig. 2A). We used chlorophyll in-tensity parameter as a threshold, from which wefunctionally partitioned a mixture of plants on a plateinto alive (silencing) and dead (nonsilencing) plants(Fig. 2A). The alive or dead classification ensured thatthe morphometric parameters extracted from the im-age analysis pipeline were confined to only alive,and therefore silencing, plants. Therefore, it stands toreason that when using our APTi system any statis-tically supported observed difference between controland treatment plants is directly attributable to si-lencing of the targeted gene(s).

Figure 1. Proof of principle and construction ofAPT-based RNAi (APTi) vectors to enable posi-tive selection of actively silencing plants. A,Illustration of the APT-interference positive se-lection principle: plants are transformed with avector that creates a long dsRNA hairpin tar-geting the APT gene, thereby reducing endog-enous APT levels and subsequent production ofcytotoxic nucleotides when supplementedwith 2-FA. Targeting APT using RNAi is suffi-cient for transformed plants to grow on stan-dard PpNH4 medium supplemented with1.25 mg mL21 2-FA. B, Schematics of the APT-based RNAi vectors. The pAPi (plasmid APTRNAi) and pGAPi (plasmid Gateway APT RNAi)constructs were created using the pUGi andpUGGi vectors, respectively, from Bezanillaet al. (2005) as templates. The thin black ar-rows indicate the direction of the open readingframe, and the inverted repeat regions of bothconstructs are flanked by a constitutive maize(Zea mays) ubiquitin promoter (thick black ar-row) and a NOS terminator sequence (blackrectangle). Both constructs target the 59UTR(blue rectangle, 179 bp) and CDS (green rect-angle, 210 bp) of the APT gene. pAPi containsonly the loop (red rectangle, 402 bp in pAPi,392 bp in pGAPi) region within the inverted re-peat, whereas pGAPi contains inverted Gatewaysites to facilitate insertion of target sequence. Thetarget may be unique to gene X, or if the targetsequence is conserved, the user can simulta-neously silence multiple targets in tandem withAPT silencing, thereby enabling survival ofsilencing plants.


Orr et al.



The APTi System Silences the Myosin XI Family withHigh Efficacy

We demonstrated that silencing of APT permits po-tent positive selection of actively silencing plants(Fig. 2A), but we sought to extend APTi to mutantanalysis. We reasoned that insertion of a target se-quence into our APTi vector (Fig. 1B) would produce intandem silencing of APT and the target gene. To testthis hypothesis, we exploited the well-characterizedtransient myosin XI(a1b) RNAi mutant that producesa dramatic loss of polarized growth phenotype (Vidaliet al., 2010). Furthermore, the myosin XI(a1b) mutantwas previously generated using a fluorescent reporter-based RNAi strategy, allowing a direct comparison ofmethodology.In P. patens, the myosin XI gene family contains two

functionally redundant isoforms, XIa and XIb, whichare both expressed in protonemata (Vidali et al., 2010).To simultaneously silence both myosin XI genes andallow for rescue experiments, we created an APTiconstruct that contains a concatenated 59UTR sequencederived from both isoforms of myosin XI. We namedthis construct “myoUTi(a1b),” and it resulted in astriking recapitulation of the myosin XI phenotype(Fig. 3A). Impressively, nearly every surviving plant

manifested the characteristic “bunch of grapes” mor-phological phenotype (Supplemental Fig. S1). This isexemplified by the relatively narrow distributions ofthe APTi myosin XI knockdown in the twomorphologyparameters, solidity and area (Fig. 3B). We speculatethe survival advantage inherent to our APTi methodcould reduce phenotypic variability sometimes ob-served in RNAi experiments.We inquired whether this phenotype could be res-

cued in the APTi system by cotransforming themyoUTi(a1b) constructwith a plasmid expressing onlythe coding sequence of myosin XIa, “XIa CDS,” therebyevading the silencing construct that targets the 59 UTR.We observed near-complete rescue of the myosin XIphenotype, demonstrating an absence of off-target ef-fects and, more importantly, highlighting the rapidphenotyping utility of the APTi system when coupledwith P. patens. Within 8 d, our APTi system isolatedthrough positive selection a relatively homogenouspopulation of mutant plants, which were amenable torescue. Our APTi system produced average myosin XImutant morphological parameters similar to a myosinXI mutant derived from RNAi using an internalGFP reporter system (Vidali et al., 2010) but withoutnonsilencing background plants. Together, these resultssupport our prediction that aberrant morphologies

Figure 2. Experimental strategy and timelinefor the APTi system. A, Optimized experimentalpipeline for transient transformation of P. patensprotoplasts and high-throughput acquisition andanalysis for the APTi system. Images are a sub-section of the entire plate area and demonstratethe characteristic difference in survival betweenconditions that do (pAPi) or do not (pUGi) si-lence APT. The scatter plot determines thebackground threshold value, which is basedupon the maximal observed chlorophyll signalin the control-RNAi condition (pUGi, blue dots).Each point corresponds to an individual plant,where the points are pooled from three inde-pendent experiments. Scale bar 5 1 mm. B,Timeline of events for an APTi assay. Typically,each condition is plated in at least duplicate foradequate sample abundance for both imagingand molecular analyses.




http://www.plantphysiol.org/cgi/content/full/pp.20.00865/DC1https://plantphysiol.org

observed using the APTi system are caused by silencingof the conjugated target, in our case myosin XI.

As our APTi approach removes essentially all back-ground, we asked if the previously intractable problemof reliable protein quantification could be trivialized toharvesting all material present on the 2-FA plate. Toaccurately quantifymyosin XI protein abundance usingour APTi method, we first established the linear rangeof our antibodies. To reflect our RNAi experimentalconditions, we transformed wild-type moss with pAPi,grew it on medium supplemented with 2-FA, thenharvested the entire plate at 8 d posttransformation.Total protein was determined using a Bradford micro-plate microassay (Bio-Rad), then a range of total protein(1–20 mg) was probed using both an in-house devel-oped antibody against myosin XIa’s coiled-coil tail(CCT) region and a publicly available anti-a-tubulinantibody (DSHB: AA4.3). This approach revealed anapproximate linear range for both antibodies from 5 to

20 mg total protein (Supplemental Fig. S2A). Impor-tantly, under equivalent conditions, our limit of de-tection for endogenous, wild-type myosin XI wasapproximately 1 mg total protein (Supplemental Fig.S2A). This step is essential, as it allows for confident,semiquantitative estimation of the extent of proteinreduction. Without this, the wild-type protein of inter-est could be loaded at or very near the limit of detection,resulting in a complete absence of protein signal in theRNAi condition, even if the true reduction is modest.

We performed an APTi experiment for myosinXI, as shown in Figure 3, and harvested at 8 d post-transformation. Every condition was implemented in atleast duplicate to allow for both harvesting of plantsand our phenotyping assay (Fig. 2). In this way, theresults of our protein analysis directly reflected theinternal protein abundance and corresponding mor-phologies we observed in the phenotyping assay(Fig. 3A). We observed a dramatic decrease of myosin

Figure 3. The APTi system robustly silences themyosin XI family in P. patens. A, Representativeimages of 8-d-old plants regenerated fromprotoplasts transformed with APTi vectors tar-geting APT alone (pAPi), APT in tandem with aconcatenated 59 UTR sequence for myosin XIaand XIb (myoUTi), targeting a nonexistent GUSsequence (pUGi), or myoUTi cotransformedwith a construct overexpressing myosin XIa’scoding sequence. All images are cropped fromcomposite images that capture a large samplearea for an individual condition, as shown inFigure 2A. For every condition, each imagecorresponds to an independent experiment.Chlorophyll autofluorescence is colored ma-genta, and calcofluor signal is colored green.Scale bar 5 100 mm. B, Quantification of mor-phometric parameters solidity and area fromthree independent experiments. Area is normal-ized to the mean area of the pAPi condition,which represents near wild-type growth mor-phology. pAPi, n 5 270; myoUTi, n 5 125;myoUTi1 XIa CDS, n5 178. Lowercase lettersindicate statistical difference (P , 0.001) be-tween groups as determined by one-wayANOVA with post-hoc Tukey test. C, Immuno-blots demonstrating the reduction and restora-tion of myosin XI protein levels when using theAPTi system. Each experiment represents anindependent transformation and subsequentplant harvesting and immunoblotting. Ten mi-crograms of total protein was loaded per lane.Myosin XI was probed using a polyclonalantibody generated against the myosin XIa CCTfragment from P. patens, and a-tubulin wasused as a loading control. D, Densitometryof immunoblot signals was performed usingImageJ and shows an ;90% reduction of my-osin XI and a complete rescue of myosin XIprotein levels when compared to the pAPicontrol.


Orr et al.


http://www.plantphysiol.org/cgi/content/full/pp.20.00865/DC1http://www.plantphysiol.org/cgi/content/full/pp.20.00865/DC1http://www.plantphysiol.org/cgi/content/full/pp.20.00865/DC1https://plantphysiol.org

XI in the “myoUTi” condition, which when normalizedto a-tubulin results in amaximal 93% reduction relativeto the control (“pAPi”). Additionally, the rescued mu-tant morphology precisely corresponds to an almostwild-type restoration of myosin XI levels (Fig. 3, C andD). Of note, these results were consistent across inde-pendent experiments and resulted in an average si-lencing efficiency of 90% (Fig. 3, C and D). We exploredthe longevity target silencing using APTi by probingmyosin XI at 2 weeks posttransformation. Althoughless potent than at 8 d posttransformation (Fig. 3),myosin XI levels were still substantially reduced(Supplemental Fig. S2, B and C), opening the possibilityfor long-term phenotyping. At the longer time, therescue condition was noticeably weaker than at theshort time point. We attribute this to loss of the myosinXI expression plasmid, as it is not under selection.Taken together, these data establish that aberrant phe-notypes observed using our APTi system are directlyattributable to reduction of target protein abundance.

APTi-Based Silencing of the Lyk5 Family Eliminates P.patens Perception to Chitin Oligosaccharides

Despite the success of the APTi system when appliedto myosin XI, we could not discount the possibility thatmyosin XI was particularly amenable to APTi and al-ternative gene targets would yield less favorable re-sults. To address this concern, we reasoned that theLyk5 gene family of P. patens presents an excellentcandidate to test the robustness and potential forfunctional discovery using APTi. In Arabidopsis (Ara-bidopsis thaliana), LYK5 is a member of a gene family oflysin-motif-containing proteins and functions as a re-ceptor for the fungal polysaccharide chitin, likely inconcert with chitin elicitor receptor kinase1 (CERK1;Cao et al., 2014). LYK5, together with LYK4 in Arabi-dopsis, functions redundantly in chitin perception, withthe lyk4 lyk5 double mutant losing all chitin sensitivity(Cao et al., 2014). Like Arabidopsis, application of chitinto P. patens elicits calcium transients (Galotto et al.,2020), resulting in the up-regulation of a downstreamdefense response (Oliver et al., 2009; Alvarez et al.,2016); however, the mechanism of chitin perception inP. patens is unknown. Previous work has identified theputative ortholog of the coreceptor CERK1 in P. patens(Bressendorff et al., 2016); however, the ortholog ofLYK5 chitin receptor is still unidentified. Phylogeneticanalysis revealed a monophyletic group of three putativeLyk5 genes in P. patens (Pp3c7_26350, Pp3c9_5820,Pp3c1_6050) thatwere homologous at the amino acid levelto the LYK2/4/5 clade in Arabidopsis (Supplemental Fig.S3). Importantly, LYK2 expression is extremely low inArabidopsis, thus limiting our functional interpretationsto LYK4/5 (Liang et al., 2014). Our gap in understandingof how P. patens senses pathogenic fungi coupled withsequence homology between the LYK2/4/5 gene familyof Arabidopsis and the Lyk5a/b/c family of P. patenspresents an opportune test for our APTi system.

To simultaneously target all three P. patens Lyk5genes, we concatenated three separate sequences de-rived from the coding sequences of Lyk5a,b,c andcloned this into our pGAPi vector, producing the ‘Lyk5-RNAi’ construct. Based on our phylogenetic analy-sis, we hypothesized the Lyk5 family in P. patens isfunctionally homologous to LYK4/5 in Arabidopsis.Therefore, we predicted depletion of all three Lyk5isoforms using APTi will result in a population of sur-viving plants that fail to elicit calcium spikes whenstimulated with chitin oligosaccharides. To test this,we transiently transformed plants expressing the cal-cium sensor GCaMPf6 (Nakai et al., 2001; Chenet al., 2013; Galotto et al., 2020) with our Lyk5-RNAiconstruct, transferred plants to selection 4 d post-transformation, then randomly selected plants at 8 dposttransformation for chitin treatment and imaging(Fig. 4A). Plants only targeting APT displayed calciumspikes indistinguishable from wild type when treatedwith chitin (Fig. 4B), and the quantified parameters(Fig. 4, C and D) were consistent with previous results(Galotto et al., 2020), suggesting APT-RNAi alone hasno influence on the output of our assay. When treatedwith Lyk5-RNAi, plants were rendered entirely insen-sitive to chitin (Fig. 4B). Furthermore, the characteristicdose dependence of calcium signal on external chitinwas abolished (Fig. 4C), and calcium oscillations werenonexistent when targeting the entire Lyk5 family(Fig. 4D). Our preliminary experiments silencing onlytwo of the three Lyk5 genes showed calcium spikes inresponse to chitin, suggesting we are targeting all threegenes with our Lyk5-RNAi construct. Additionally, asonly the double knockout lyk4 lyk5 mutant in Arabi-dopsis loses all chitin sensitivity (Cao et al., 2014), weconclude our APTi-based silencing of the Lyk5 family isfunctioning with high efficiency to phenocopy a doubleknockout mutant in Arabidopsis. Importantly, theseresults elucidate functional homology of the LYK4/5genes between Arabidopsis and P. patens, wherebyLyk5 functions at the level of chitin perception. Previ-ous work has demonstrated that P. patens CERK1 res-cues the Arabidopsis cerk1 mutant (Bressendorff et al.,2016), strongly supporting a conserved mechanism ofchitin sensing. All together, our findings demonstratethe robustness and versatility of our APTi system formutant isolation and characterization.

DISCUSSION

RNAi offers an invaluable complement to traditionalgene knockout studies. However, substantive ad-vancements in RNAi methods are trailing the explosionof CRISPR/Cas-based technologies. Here, we estab-lished the first survival-based RNAi methodology thatrobustly selects for actively silencing plants. We ac-complished this by engineering vectors that elicit aprosurvival response when processed by the orga-nism’s endogenous RNAi machinery. Using the previ-ously characterized myosin XI mutant in P. patens as an




http://www.plantphysiol.org/cgi/content/full/pp.20.00865/DC1http://www.plantphysiol.org/cgi/content/full/pp.20.00865/DC1http://www.plantphysiol.org/cgi/content/full/pp.20.00865/DC1https://plantphysiol.org

initial case study, we showed that in tandem fusion of amyosin XI target with the prosurvival sequence resul-ted in potent selection of morphologically mutantplants. We demonstrated that surviving plants activelytargeting myosin XI through our novel vectors con-tained approximately 7% of normal myosin XI proteinabundance. Additionally, using APTi facilitated thesilencing of the Lyk5 family and resulted in mossplants completely insensitive to chitin-induced calciumspikes. To our knowledge, this is the first evidencedemonstrating the function of Lyk5 in P. patens andsupports the notion of conservation in fungal percep-tion between vascular and nonvascular plants. OurAPTi technology represents a dramatic improvement insilencing efficacy and experimental implementationover previous RNAi methods.

RNAi has been extensively employed in the modelmoss P. patens for both discovery and validation of genefunction (Vidali et al., 2007, 2010; Augustine et al., 2008,

2011; Prigge et al., 2010; Wu et al., 2011; Miki et al.,2015; Bascom et al., 2019). We attribute the popularityof dsRNA-based RNAi in P. patens to a method thatuses an internal fluorescent reporter of RNAi to ob-tain results in 1 week (Bezanilla et al., 2003). Wesought to fundamentally improve upon RNAi re-porters by creating a reporter where survival itself isindicative of active silencing. We demonstrated thefeasibility and effectiveness of this approach by si-lencing the APT gene in the presence of 2-FA in themedium. Furthermore, we engineered two plasmidsthat facilitate insertion of any target of interest intandem with the APT-silencing sequence. We callthis approach APT-based RNAi, or APTi. To achieverapid phenotyping using APTi, all experiments wereperformed transiently. Modifications of the APTiplasmids to promote stable integration and inductionof the silencing cassette represent an important areaof future work. Both APTi plasmids and the myosin

Figure 4. Simultaneous silencing of the Lyk5gene family with APTi abolishes chitin-inducedcalcium transients. A, Schematic of APTi-basedfunctional assay of Lyk5. Control RNAi plantstreated with chitin elicit calcium transients, vi-sualized with GCaMP. Simultaneous depletionof the Lyk5 receptor family (a, b, c) using APTidesensitizes the plants to chitin, resulting in lossof calcium spikes. B, Eight-day-old plants trea-ted with 20 mg mL21 chitin oligosaccharides.Time stamps aremin:s. Dark blue indicates plantautofluorescence, where aqua to red reflects anincrease of calcium signal. Wild-type representsGCaMPf6 in the Gransden background. C, Doseresponse of P. patens’ calcium signal to chitinapplication for control RNAi plants and plantstransformed with an APTi construct targeting allthree Lyk5 genes. Time-averaged mean grayvalue (left plot) and the SD of the time-averagedmean gray value (right plot) are normalized by(F 2 F0)/F0 for both conditions. All data pointsare the mean of at least three plants pooled fromthree independent experiments. D, Characteri-zation of calcium transients within a 30-minobservation period, as in A, across multipleconcentrations of chitin. All points represent anindividual plant, pooled across three indepen-dent experiments. Bars represent the mean 6the SEM. Each condition was compared againstitself across concentrations, with an asteriskrepresenting a statistical difference between thatcondition and its corresponding water control.Statistics were performed with a Kruskal-Wallisone-way nonparametric ANOVA, followed bya Dunn’s multiple comparisons test and indi-cated by asterisks (*P , 0.001).


Orr et al.



XI-RNAi plasmid are publicly available from theplasmid repository Addgene.We obtained potent positive selection of actively si-

lencing plants in P. patens by engineering RNAi vectorsthat exploit the function of the APT gene. However, thephysiological consequences of APT silencing remain inquestion. Although it is a salvage enzyme, APT’sfunction presents a more energetically efficient meansof nucleotide production than de novo synthesis(Ashihara et al., 2018). Consequently, knockouts of APTin vascular plants demonstrate severe defects in pollengermination and pollen tube growth, presumably aresult of impairing the energy-intensive fast growth ofthe pollen tube (Moffatt and Somerville, 1988; Zhouet al., 2006). Interestingly, an alternative mutant alleleof APT that results in partial reduction of APT activ-ity imparts enhanced growth and stress tolerance(Sukrong et al., 2012). We suspect this hypomorphicallele better represents the internal state of our APT-silenced plants. Taken together with our observationsof APT-silencing plants, reproducing results achievedwith an orthogonal RNAi system andAPT-RNAi plantsdisplaying calcium spikes consistent with wild-type,we conclude the reduction of APT results in no clearphysiological defects within the scope of our assays.We expect our APTi strategy to be applicable to other

organisms given the ubiquity of the APT gene (Schaff,1994). Like other organisms, such as in humans, P.patens contains only one copy of APT, making it espe-cially amenable to the APTi strategy as we showed.Interestingly, the vascular plant Arabidopsis containsfive APT genes (Allen et al., 2002). We suspect APTicould be applied in Arabidopsis by constructing anAPT-silencing module comprised of concatenated se-quences targeting specific isoforms. Additional work isnecessary to translate to other systems, but we submitthat our efforts establishing APTi in P. patens willgreatly benefit the community in understandingfundamental and conserved biological processes(Orr et al., 2020).Previous work using the fluorescent reporter-based

RNAi has identified silencing plants based on loss offluorescence (Bezanilla et al., 2003). However, slightreduction in fluorescence confounded interpretationbecause it could be attributed to natural variation in thereporter signal or reflect a real, but modest, silencingeffect. Furthermore, we observed spontaneous loss ofreporter signal in the moss reporter line over long pe-riods of continuous propagation. Without careful ob-servation and subcloning to remove chimeric reportercultures, a researcher could inadvertently concludesilencing when none is occurring. Our APTi systemsimultaneously removes the requirement of a dedi-cated moss reporter line and dismisses any ambiguityinherent to a continuous reporter signal. We dem-onstrated the utility of this advancement in our in-vestigation of Lyk5 by transforming our APTi-basedconstruct into amoss line expressing the calcium sensorGCaMP6f (Galotto et al., 2020). This saved consider-able time, as the previous RNAi methodology would

require creation of a new transgenic line that containsthe RNAi GFP reporter and a non-GFP-based calciumsensor.Interestingly, prior RNAi methods can result in

background plants that survive antibiotic selection forthe plasmid containing the RNAi transgene but do notsilence the reporter. This is likely a result of transcrip-tional gene silencing, whereby the plant rescues itselffrom RNAi transgene expression, but expression of theindependently regulated antibiotic resistance is un-modified (Morel et al., 2000; Fusaro et al., 2006; Small,2007). We hypothesize our method enhances silencingefficiency by engineering a fitness punishment forthe organism to silence the RNAi transgene. To this end,with APTi survival on 2-FA is directly coupled to ex-pression of the APT-RNAi transgene. Therefore, theorganism cannot survive if it silences the expression ofthe RNAi cassette, thus ensuring expression of the APT1 gene target hairpin. Although we did not determinethe extent of APT silencing, we know the reduction issufficient to promote survival on 2-FA and produce a90% reduction of a fused target, myosin XI, or completeloss of detectable calcium signal when all three Lyk5genes are targeted.We speculated that the enhanced fitness benefit im-

posed by our APTi systemwill result in more consistentand potent silencing efficiencies of target genes. Con-sistent with this, we observed a homogenous popula-tion of mutants evidenced by a 90% reduction ofendogenous myosin XI and total loss of chitin sensi-tivity when targeting myosin XI(a,b) and Lyk5(a,b,c),respectively. APTi offers a noticeably higher silencingefficacy when compared to previous reports usingdsRNA and GFP-based reporters in P. patens (Vidaliet al., 2007, 2010; Nakaoka et al., 2012). Importantly,previous reports could only estimate protein silencingbased on a small subset of individual plants deliberatelychosen by the experimenter, which may not accuratelyreflect the average silencing effect (Vidali et al., 2007,2009, 2010; Augustine et al., 2008). This microscope-based methodology was required because mutantsresulting in small morphologies failed to yield adequateplantmaterial for immunoblots andwere surrounded bynonsilencing background plants. We demonstrated thatAPTi’s positive selection enables simple harvesting ofthe entire plate that can be easily scaled for reproducibleprotein quantitation. Based on our observed myosin XIknockdown, ourAPTi silencing efficacy is comparable tothe most effective amiRNAs (Zhang et al., 2018) butwithout the need for prior engineering and screening ofmultiple amiRNAs. Furthermore, performing our Lyk5functional assay with prior GFP-based RNAi would re-quire an additional time-consuming and tedious pre-screening step, where actively silencing plants arephysically isolated from the nonsilencing background.WithAPTi, any surviving plant can be chosen at randomfor functional analysis, greatly enhancing the through-put and likely the consistency of the results.We showed that the APTi system is well suited for

high-throughput phenotyping. We fully anticipate this





area to be iteratively improved, not just with respect tothe volume of acquisition but with increased compu-tational sophistication. The large obtainable data setsare ripe for both classic exploratory data methods andcutting-edge deep-learning techniques. For example,we analyzed living plants by first segmenting the im-ages by traditional thresholding. We then filtered andclassified the hundreds of plants based on a character-istic biological feature, chlorophyll autofluorescence,which we derived from the control-RNAi dying pop-ulation. Although less intuitive, deep learning offersthe possibility of automating image segmentation andclassification, perhaps resulting in discovery of novelmutant features (Moen et al., 2019).

CONCLUSION

This work represents a fundamental transition fromvisual screening for RNAi plants to positive selection ofactively silencing plants. We achieved this by engi-neering vectors that produce a single hairpin RNAtargeting the APT gene and any other genes of interestin tandem. This results in effective isolation of all sur-viving plants undergoing RNAi of the target genewhengrown in the presence of 2-FA. Importantly, the efficacyof gene silencingwas consistently greater with the APTisystem, maximal 93% reduction of target protein, thanprevious reports silencing the same myosin XI genesusing a fluorescent screening method. Additionally,with APTi, we simultaneously silenced the Lyk5 genefamily (a,b,c) and demonstrated its requirement forperception of chitin oligosaccharides in P. patens. Webelieve our APTi system provides a flexible, fast, andeffective platformwith unprecedented low backgroundand variability to facilitate high-throughput character-ization for loss-of-function mutants.

MATERIALS AND METHODS

Plant Materials and Culture Conditions

Three Physcomitrella patens lines were used in this study: (1) NLS4 (Bezanillaet al., 2003) in Figure 1A; (2) GCaMP6f, called wild-type in Figure 4 (Galottoet al., 2020); and (3) wild-type Gransden (Ashton et al., 1979) in all other ex-periments. All lines were cultured as previously described (Vidali et al., 2007).In brief, tissue was propagated weekly by homogenization and transferred tosolid PpNH4 medium overlaid with cellophane. Cultures were grown at 25°Cunder long-day light (90 mmol m22 s21) conditions (16 h:8 h, light:dark).

APT-Based RNAi Construct Design

The APT transcript fragment (Phytozome: Pp3c8_16590) containing the 59UTR and CDS was amplified by PCR with forward (APT_BSK_F) and reverse(APT_BSK_R) primers and cloned into pBlueScript K1 using restriction en-zymes SacI/EcoRV (generating the APT pBSK1 plasmid). The pUGGi Gatewaycassette and loop domain lacking the GUS regions was amplified in two piecesusing PCR with primer sets Gateway1_F/R and Gateway2_F /R (both reac-tions at 60°C, 2 min elongation), and both fragments were ligated individuallyinto pBlueScript K1 (generating the Gateway F pBSK1 and Gateway R pBSK1

plasmids). The pUGGi loopwas amplified by PCRwith primers Loop_F/R andinserted into pBlueScript K1 (generating the Loop pBSK1 plasmid). The APTand Loop pBSK1 constructs were transformed into DH5a Escherichia coli,

whereas both Gateway pBSK constructs were transformed into ccdB Echerichiacoli. All constructs were blue-screened for successful clones by plating trans-formants on LB1Carb1Chlor plates and adding 40 mL each of isopropylthio-b-galactoside and X-Gal. Next, using the SacI/XhoI restriction sites, the GatewayF fragment was inserted into the Gateway R pBSK1 backbone, generating thecomplete Gateway pBSK1 plasmid. Importantly, this construct contains theentire loop region derived from pUGGi, as the two intermediate Gatewayconstructs discussed above each contained half of the loop region. It was nec-essary to isolate the pUGGi loop in its own plasmid to create an additionalcontrol plasmid lacking the Gateway cassettes.

The first APT fragment was then excised from APT pBSK1 using HindIII/SwaI, then transferred into the Gateway pBSK1 plasmid, and cut with HindIII/PmeI (generating Gateway-1APT pBSK1). Next, the second APT fragment wasexcised from APT pBSK1 and transferred into the Gateway-1APT pBSK1

plasmid using SacI/SwaI sites (generating Gateway-2APT pBSK1). TheGateway-2APT fragment was digested and then ligated into the pUGGibackbone using SacI/KpnI sites, generating the pGAPi plasmid. To generate ourpositive control plasmid without Gateway sites, two APT fragments wereinserted into the Loop pBSK1 plasmid using the same procedure as describedabove—first by HindIII/SwaI and HindIII/PmeI sites, then by SacI/SwaI sites—producing the Loop-2APT pBSK1 plasmid. Then the Loop-2APT fragment wascloned into the pUGGi backbone using SacI/KpnI sites, generating the pAPiplasmid. At all points in this process where new plasmids were made byrestriction-based cloning, those plasmids were confirmed by restriction andsequence analysis.

The myoUTi:pGAPi construct was created by extracting the previouslypublished, concatenated 59UTR targeting sequences (Vidali et al., 2010) from itsRNAi destination vector via a Gateway BP reaction, then inserted into our newpGAPi construct via a Gateway LR reaction. The transient myosin XI-RNAiphenotype was rescued by expressing the myosin XIa coding sequence(Vidali et al., 2010). All three APTi plasmids are publicly available fromAddgene (pGAPi, #127547; pAPi, #127548; myoUTi:pGAPi, #127549), and se-quences are also available at GenBank (pGAPi, MK975250; pAPi, MK975251;myoUTi:pGAPi, MK975252).

APTi Phenotyping Assay

One-week-old moss was protoplasted and transformed as described (Liuand Vidali, 2011), with transformed protoplasts being suspended in liquidplating medium and plated at 1.4 3 105 protoplasts per 100-mm petri dish ofprotoplast regeneration medium for the bottom layer. Each condition wasplated in at least duplicate, with themyoUTi condition being plated in triplicateto allow enough plant material of the mutant plants to be harvested for im-munoblotting and imaging. Four days posttransformation, the cellophanes ofeach plate were transferred to growth medium (PpNH4) supplemented with1.25 mg mL21 2-FA from a 5 mg mL21 in dimethyl sulfoxide stock (OakwoodChemical). As mentioned previously, each lab should first optimize the 2-FAselection concentration, as our results were all obtained using a single lot of2-FA from Oakwood Chemical

Eight days posttransformation plants were stained with 10 mg mL21-calcofluor from a 1 mg mL21 in water stock (Fluorescent Brightener 28, Sigma)and imaged with a 103 A-Plan (0.25 NA) objective of an epifluorescent mi-croscope (Zeiss Axiovert 200M) coupled to a CCD camera (Zeiss AxioCamMRm). This microscope was equipped with an automated stage and integratedwith the AxioVision software through the MosaicX module, enabling preciseacquisition, tiling, and stitching of individual images to create a large com-posite. Our composite images contained 12 3 12 individual images, acquiredwith a 15% overlap. Each individual image consisted of two channels, thecalcofluor and chlorophyll signals. The chlorophyll channel was acquired witha 480/40 bandpass excitation, a 505 long-pass dichroic mirror, and a 510 long-pass emission filter cube and with a fixed 150-ms exposure for all experiments.The calcofluor signal was acquired with a standard DAPI filter and automati-cally adjusted for each experiment to maximize contrast.

Stitched images were first processed and segmented using a custom ImageJmacro (available upon request). Our macro only discarded segmented objectsapproximately the size of two adjoining protoplasts or smaller, as to not biasagainst the discovery of small mutants and remove nonsurviving protoplasts.Each segmented image was visually inspected for overlapping or truncatedplants, and if present they were discarded from further analysis. Subsequentfiltering of alive plants and visualization of the data were performed usingMATLAB (MathWorks). Statistical testing (one way ANOVA-Tukey) wasperformed with GraphPad Prism.


Orr et al.



Analysis of Myosin XI Protein Abundance

Eight days posttransformation, all tissue was harvested by scraping, flash-frozen in liquid nitrogen, and stored at 280°C. To create protein extracts, thefrozen tissue was ground to a powder in liquid nitrogen, then resuspended inextraction buffer (250mM Suc, 20 mM EGTA, 50mM PIPES, 150mMNaCl, 60 mMMgCl2, and 1% [w/v] casein) supplemented with fresh dithiothreitol (2 mMfinal) and protease inhibitors. The powder from the pAPi condition wasresuspended in 300 mL, whereas the myoUTi and rescue conditions were al-ways resuspended in 200 mL extraction buffer. Extracts were vortexed for 15 s,then placed on ice for 30 s, with this repeated twice more. Extracts were thenspun at 13,000 rpm for 10 min at 4°C, followed by removal of 175 mL of theclarified supernatant; 120 mL of extract was immediately combined with SDSloading buffer and boiled, then stored at 280°C. The remaining extract wasused for total protein determination using the Bradford microplate microassayprocedure (Bio-Rad).

ToprobemyosinXIprotein directly, an antibodywasdevelopedagainst a 63His fusion of the P. patens myosin XIa-CCT (Capralogics). An antibody againsta-tubulin was used as a loading control (DSHB: AA4.3). The approximate linearrange of both antibodies was determined by first loading 1 to 20 mg of totalprotein from a pAPi-treated moss extract on a 4 to 12% (w/v) Bis-Tris SDS-PAGE gel (Thermo Fisher). Protein was then transferred to nitrocelluloseovernight at 4°C, followed by blocking with 5% (w/v) milk in Tris-bufferedsaline, 0.1% (v/v) Tween 20 at room temperature for 1 hr. The nitrocellulosewas cut at the 80-kD marker, with the higher Mr piece incubated with myosinXIa-CCT primary antibody (1:10,000) and the a-tubulin (0.5 mg mL21 final)incubated with the lowerMr fragment for 1 hr at room temperature. Followingprimary incubation, blots were washed three times in Tris-buffered saline, 0.1%(v/v) Tween 20, incubated in secondary antibody (goat anti-rabbit for XIa-CCT,goat anti-mouse for a-tubulin) at 1:100,000 dilution for 1 hr at room tempera-ture, then washed a final three times. Blots were developed using homemadeenhanced chemiluminescence reagent, and chemiluminescent images wereacquired using an Azure 600 (Azure Biosystems). Densitometry was performedusing ImageJ (Schneider et al., 2012) to allow comparison of relative proteinabundance.

Lyk5-RNAi Construct Design and Functional Assay

To disrupt the expression of Lyk5 genes in P. patens, an RNAi plasmid wasconstructed. The first 414 bp of Lyk5a and 418 bp of Lyk5bwere amplified usingspecific primers (Supplemental Table S1) with the Phusion polymerase (ThermoFisher). Then, both fragments were joined by overlapping PCR, and the 0.8-kbfragmentwas gel purified (NucleoSpin gel and PCR clean-up,Macherey-Nagel)and inserted in the entry vector pDONR207 via a BP reaction (Thermo Fisher).To insert the remaining Lyk5 gene, primers were designed for Lyk5c(Supplemental Table S1) that amplified a fragment of exon 1. The corre-sponding 0.4-kb amplicon was gel purified (Zymogen clean gel DNA recoverykit) and inserted into pDONR207-Lyk5a,b via BamHI. Successful pENT-Lyk5a,b,c clones were identified through restriction analysis and sequencing.The concatenated Lyk5a,b,c target sequence was inserted into pGAPi via an LRreaction (Thermo Fisher) between pGAPi and pENT-Lyk5a,b,c, yielding Lyk5-RNAi. The Lyk5-RNAi construct was sequence verified. Transient transfor-mation and APT selection were performed as described above.

Calcium imaging and analysis were performed as previously described(Galotto et al., 2020). In brief, plants were mounted on an agar pad containingserial dilutions of chitin oligosaccharides, starting at 20 mg mL21 (TokyoChemical Industry), or water and imaged with a Zeiss Axio Observer.A1 mo-torized microscope using a 103 objective lens (NA 0.25). The internal calciumsensor was excited with a mercury lamp X-Cite series 120 PC EXFO, and in-dividual plant images were acquired every 30 s for 30 min within 10 min ofinitial chitin exposure. Analysis of the calcium time series data was performedusing ImageJ and RStudio. Microscope drift was corrected in ImageJ using the‘StackReg’ plugin (Thévenaz et al., 1998). Next, the area around the plant wasselected and the background was cropped. Then a macro applied, for eachframe in the series, a thresholding function that equalized the images andmasked the area around the plant; the masked images were used to analyze theCa21 peaks in chitin-treated plants. In addition, the ImageJ macro recorded themean gray value of each image. The mean gray values for a time series wereaveraged (calcium levels) and the SD of the average calculated (calcium fluc-tuations). The dose response was calculated in GraphPad Prism by plotting thetime-averaged values and SD for each concentration of chitin oligosaccharides.To estimate the number of calcium peaks and their frequency, the masked

images produced by our ImageJ macro were analyzed using MATLAB. Thescript used a rolling 3 3 3 pixel grid to detect the position of the plantthroughout the video. For each frame, the grid scanned through the image andrecorded the pixel value of a 33 3 pixel area. Only the gridswith positive valueswere considered as being part of a plant. When all the positive pixels wereidentified, the MATLAB script computed the mean of the pixel values andcalculated the SD of that mean. Once the mean of the pixel values was obtained,the script found the Ca21 peaks. A signal was considered a peak when its in-tensity was greater than one SD multiplied by 0.345, a constant, from the meangray value of the plant. Once the peak detection was completed, the periodbetween peaks was measured with a built-in MATLAB function ‘findpeaks.’This analysis was repeated for each plant.

Accession Numbers

All main-text sequence data were extracted from the Physcomitrella patensv3.3 assembly found on Phytozome (phytozome.jgi.doe.gov): APT(Pp3c8_16590); myosin XIa (Pp3c9_25360); myosin XIb (Pp3c15_24240); Lyk5a(Pp3c7_26350); Lyk5b (Pp3c9_5820); and Lyk5c (Pp3c1_6050). All three APTiplasmids are publicly available fromAddgene (pGAPi, #127547; pAPi, #127548;and myoUTi:pGAPi, #127549), and sequences are also available at GenBank(pGAPi, MK975250; pAPi, MK975251; and myoUTi:pGAPi, MK975252).

Supplemental Data

The following supplemental materials are available.

Supplemental Figure S1. The APTi system reduces variability in observedmutant phenotype.

Supplemental Figure S2. Determination of the linear range of used anti-bodies and long-term reproducibility of immunoblots using the APTisystem.

Supplemental Figure S3. Phylogenetic tree of Lyk genes from P. patens andother selected plant and alga species.

Supplemental Table S1. List of primers used for vector construction andsequence verification.

ACKNOWLEDGMENTS

We thank the members of L.V.’s group at Worcester Polytechnic Instituteand Mary Munson’s group at the University of Massachusetts Medical Schoolfor helpful discussions. We also thank Victoria Huntress for her management ofthe Microscopy core facility at Worcester Polytechnic Institute.

Received June 30, 2020; accepted July 15, 2020; published August 6, 2020.

LITERATURE CITED

Allen M, Qin W, Moreau F, Moffatt B (2002) Adenine phosphoribosyl-transferase isoforms of Arabidopsis and their potential contributions toadenine and cytokinin metabolism. Physiol Plant 115: 56–68

Alvarez A, Montesano M, Schmelz E, Ponce de León I (2016) Activation ofshikimate, phenylpropanoid, oxylipins, and auxin pathways in Pecto-bacterium carotovorum elicitors-treated moss. Front Plant Sci 7: 328

Ashihara H, Stasolla C, Fujimura T, Crozier A (2018) Purine salvage inplants. Phytochemistry 147: 89–124

Ashton NW, Grimsley NH, Cove DJ (1979) Analysis of gametophytic de-velopment in the moss, Physcomitrella patens, using auxin and cytokininresistant mutants. Planta 144: 427–435

Augustine RC, Pattavina KA, Tüzel E, Vidali L, Bezanilla M (2011) Actininteracting protein1 and actin depolymerizing factor drive rapid actindynamics in Physcomitrella patens. Plant Cell 23: 3696–3710

Augustine RC, Vidali L, Kleinman KP, Bezanilla M (2008) Actin depo-lymerizing factor is essential for viability in plants, and its phosphor-egulation is important for tip growth. Plant J 54: 863–875

Bascom C Jr., Burkart GM, Mallett DR, O’Sullivan JE, Tomaszewski AJ,Walsh K, Bezanilla M (2019) Systematic survey of the function of ROPregulators and effectors during tip growth in the moss Physcomitrellapatens. J Exp Bot 70: 447–457




http://www.plantphysiol.org/cgi/content/full/pp.20.00865/DC1http://www.plantphysiol.org/cgi/content/full/pp.20.00865/DC1http://phytozome.jgi.doe.govhttp://www.plantphysiol.org/cgi/content/full/pp.20.00865/DC1http://www.plantphysiol.org/cgi/content/full/pp.20.00865/DC1http://www.plantphysiol.org/cgi/content/full/pp.20.00865/DC1http://www.plantphysiol.org/cgi/content/full/pp.20.00865/DC1https://plantphysiol.org

Baulcombe D (2004) RNA silencing in plants. Nature 431: 356–363Bezanilla M, Pan A, Quatrano RS (2003) RNA interference in the moss

Physcomitrella patens. Plant Physiol 133: 470–474Bezanilla M, Perroud PF, Pan A, Klueh P, Quatrano RS (2005) An RNAi

system in Physcomitrella patens with an internal marker for silencingallows for rapid identification of loss of function phenotypes. Plant Biol(Stuttg) 7: 251–257

Bressendorff S, Azevedo R, Kenchappa CS, Ponce de León I, Olsen JV,Rasmussen MW, Erbs G, Newman MA, Petersen M, Mundy J (2016)An innate immunity pathway in the moss Physcomitrella patens. PlantCell 28: 1328–1342

Byzova M, Verduyn C, De Brouwer D, De Block M (2004) Transformingpetals into sepaloid organs in Arabidopsis and oilseed rape: Im-plementation of the hairpin RNA-mediated gene silencing technology inan organ-specific manner. Planta 218: 379–387

Cao Y, Liang Y, Tanaka K, Nguyen CT, Jedrzejczak RP, Joachimiak A,Stacey G (2014) The kinase LYK5 is a major chitin receptor in Arabidopsisand forms a chitin-induced complex with related kinase CERK1. eLife 3:e03766

Charlot F, Chelysheva L, Kamisugi Y, Vrielynck N, Guyon A, Epert A, LeGuin S, Schaefer DG, Cuming AC, Grelon M, et al (2014) RAD51Bplays an essential role during somatic and meiotic recombination inPhyscomitrella. Nucleic Acids Res 42: 11965–11978

Chen TW, Wardill TJ, Sun Y, Pulver SR, Renninger SL, Baohan A,Schreiter ER, Kerr RA, Orger MB, Jayaraman V, et al (2013) Ultra-sensitive fluorescent proteins for imaging neuronal activity. Nature 499:295–300

Chuang CF, Meyerowitz EM (2000) Specific and heritable genetic inter-ference by double-stranded RNA in Arabidopsis thaliana. Proc Natl AcadSci USA 97: 4985–4990

Deans RM, Morgens DW, Ökesli A, Pillay S, Horlbeck MA, KampmannM, Gilbert LA, Li A, Mateo R, Smith M, et al (2016) Parallel shRNA andCRISPR-Cas9 screens enable antiviral drug target identification. NatChem Biol 12: 361–366

de la Luz Gutiérrez-Nava M, Aukerman MJ, Sakai H, Tingey SV,Williams RW (2008) Artificial trans-acting siRNAs confer consistent andeffective gene silencing. Plant Physiol 147: 543–551

Ding X, Pervere LM, Bascom C Jr., Bibeau JP, Khurana S, Butt AM, OrrRG, Flaherty PJ, Bezanilla M, Vidali L (2018) Conditional geneticscreen in Physcomitrella patens reveals a novel microtubule depolyme-rizing-end-tracking protein. PLoS Genet 14: e1007221

Fusaro AF, Matthew L, Smith NA, Curtin SJ, Dedic-Hagan J, Ellacott GA,Watson JM, Wang MB, Brosnan C, Carroll BJ, et al (2006) RNAinterference-inducing hairpin RNAs in plants act through the viral de-fence pathway. EMBO Rep 7: 1168–1175

Galotto G, Abreu I, Sherman C, Liu B, Gonzalez-Guerrero M, Vidali L(2020) Chitin triggers calcium-mediated immune response in the plantmodel Physcomitrella patens. Mol Plant Microbe Interact 33: 911–920

Galotto G, Bibeau JP, Vidali L (2019) Automated image acquisition andmorphological analysis of cell growth mutants in Physcomitrella patens.Methods Mol Biol 1992: 307–322

Guo XY, Li Y, Fan J, Xiong H, Xu FX, Shi J, Shi Y, Zhao JQ, Wang YF, CaoXL, et al (2019) Host-induced gene silencing of MoAP1 confers broad-spectrum resistance to Magnaporthe oryzae. Front Plant Sci 10: 433

Hannon GJ (2002) RNA interference. Nature 418: 244–251Hauser F, Chen W, Deinlein U, Chang K, Ossowski S, Fitz J, Hannon GJ,

Schroeder JI (2013) A genomic-scale artificial microRNA library as atool to investigate the functionally redundant gene space in Arabidopsis.Plant Cell 25: 2848–2863

Horlbeck MA, Witkowsky LB, Guglielmi B, Replogle JM, Gilbert LA,Villalta JE, Torigoe SE, Tjian R, Weissman JS (2016) Nucleosomesimpede Cas9 access to DNA in vivo and in vitro. eLife 5: e12677

Housden BE, Muhar M, Gemberling M, Gersbach CA, Stainier DY,Seydoux G, Mohr SE, Zuber J, Perrimon N (2017) Loss-of-functiongenetic tools for animal models: Cross-species and cross-platform dif-ferences. Nat Rev Genet 18: 24–40

Jensen KT, Fløe L, Petersen TS, Huang J, Xu F, Bolund L, Luo Y, Lin L(2017) Chromatin accessibility and guide sequence secondary structureaffect CRISPR-Cas9 gene editing efficiency. FEBS Lett 591: 1892–1901

Khraiwesh B, Arif MA, Seumel GI, Ossowski S, Weigel D, Reski R,Frank W (2010) Transcriptional control of gene expression by micro-RNAs. Cell 140: 111–122

Khraiwesh B, Ossowski S, Weigel D, Reski R, Frank W (2008) Specificgene silencing by artificial microRNAs in Physcomitrella patens: An al-ternative to targeted gene knockouts. Plant Physiol 148: 684–693

Levine RA, Taylor MW (1982) Mechanism of adenine toxicity in Escherichiacoli. J Bacteriol 149: 923–930

Li JF, Chung HS, Niu Y, Bush J, McCormack M, Sheen J (2013) Com-prehensive protein-based artificial microRNA screens for effective genesilencing in plants. Plant Cell 25: 1507–1522

Liang Y, Tóth K, Cao Y, Tanaka K, Espinoza C, Stacey G (2014) Lip-ochitooligosaccharide recognition: An ancient story. New Phytol 204:289–296

Liu SM, Yoder JI (2016) Chemical induction of hairpin RNAi molecules tosilence vital genes in plant roots. Sci Rep 6: 37711

Liu Y-C, Vidali L (2011) Efficient polyethylene glycol (PEG) mediatedtransformation of the moss Physcomitrella patens. J Vis Exp 50: e2560

Miki T, Nishina M, Goshima G (2015) RNAi screening identifies the ar-madillo repeat-containing kinesins responsible for microtubule-dependent nuclear positioning in Physcomitrella patens. Plant CellPhysiol 56: 737–749

Moen E, Bannon D, Kudo T, Graf W, Covert M, Van Valen D (2019) Deeplearning for cellular image analysis. Nat Methods 16: 1233–1246

Moffatt B, Somerville C (1988) Positive selection for male-sterile mutantsof Arabidopsis lacking adenine phosphoribosyl transferase activity. PlantPhysiol 86: 1150–1154

Morel JB, Mourrain P, Béclin C, Vaucheret H (2000) DNA methylation andchromatin structure affect transcriptional and post-transcriptionaltransgene silencing in Arabidopsis. Curr Biol 10: 1591–1594

Nakai J, Ohkura M, Imoto K (2001) A high signal-to-noise Ca(21) probecomposed of a single green fluorescent protein. Nat Biotechnol 19:137–141

Nakaoka Y, Miki T, Fujioka R, Uehara R, Tomioka A, Obuse C, Kubo M,Hiwatashi Y, Goshima G (2012) An inducible RNA interference systemin Physcomitrella patens reveals a dominant role of augmin in phrag-moplast microtubule generation. Plant Cell 24: 1478–1493

Oliver JP, Castro A, Gaggero C, Cascón T, Schmelz EA, Castresana C,Ponce de León I (2009) Pythium infection activates conserved plantdefense responses in mosses. Planta 230: 569–579

Orr RG, Cheng X, Vidali L, Bezanilla M (2020) Orchestrating cell mor-phology from the inside out: Using polarized cell expansion in plants asa model. Curr Opin Cell Biol 62: 46–53

Prigge MJ, Lavy M, Ashton NW, Estelle M (2010) Physcomitrella patensauxin-resistant mutants affect conserved elements of an auxin-signalingpathway. Curr Biol 20: 1907–1912

Rossi A, Kontarakis Z, Gerri C, Nolte H, Hölper S, Krüger M, Stainier DY(2015) Genetic compensation induced by deleterious mutations but notgene knockdowns. Nature 524: 230–233

Schaefer DG (2001) Gene targeting in Physcomitrella patens. Curr Opin PlantBiol 4: 143–150

Schaff DA (1994) The adenine phosphoribosyltransferase (APRT) select-able marker system. Plant Sci 101: 3–9

Schaff DA, Jarrett RA, Dlouhy SR, Ponniah S, Stockelman M, StambrookPJ, Tischfield JA (1990) Mouse transgenes in human cells detect specificbase substitutions. Proc Natl Acad Sci USA 87: 8675–8679

Schneider CA, Rasband WS, Eliceiri KW (2012) NIH Image to ImageJ:25 years of image analysis. Nat Methods 9: 671–675

Schwab R, Ossowski S, Riester M, Warthmann N, Weigel D (2006) Highlyspecific gene silencing by artificial microRNAs in Arabidopsis. Plant Cell18: 1121–1133

Small I (2007) RNAi for revealing and engineering plant gene functions.Curr Opin Biotechnol 18: 148–153

Smits AH, Ziebell F, Joberty G, Zinn N, Mueller WF, Clauder-Münster S,Eberhard D, Fälth Savitski M, Grandi P, Jakob P, et al (2019) Biologicalplasticity rescues target activity in CRISPR knock outs. Nat Methods 16:1087–1093

Sukrong S, Yun KY, Stadler P, Kumar C, Facciuolo T, Moffatt BA,Falcone DL (2012) Improved growth and stress tolerance in the Arabi-dopsis oxt1 mutant triggered by altered adenine metabolism. Mol Plant 5:1310–1332

Thévenaz P, Ruttimann UE, Unser M (1998) A pyramid approach tosubpixel registration based on intensity. IEEE Trans Image Process 7:27–41


Orr et al.



van Gisbergen PAC, Wu SZ, Chang M, Pattavina KA, Bartlett ME,Bezanilla M (2018) An ancient Sec10-formin fusion provides insightsinto actin-mediated regulation of exocytosis. J Cell Biol 217: 945–957

Verkuijl SA, Rots MG (2019) The influence of eukaryotic chromatin stateon CRISPR-Cas9 editing efficiencies. Curr Opin Biotechnol 55: 68–73

Vidali L, Augustine RC, Kleinman KP, Bezanilla M (2007) Profilin is es-sential for tip growth in the moss Physcomitrella patens. Plant Cell 19:3705–3722

Vidali L, Burkart GM, Augustine RC, Kerdavid E, Tüzel E, Bezanilla M(2010) Myosin XI is essential for tip growth in Physcomitrella patens. PlantCell 22: 1868–1882

Vidali L, van Gisbergen PAC, Guérin C, Franco P, Li M, Burkart GM,Augustine RC, Blanchoin L, Bezanilla M (2009) Rapid formin-mediated actin-filament elongation is essential for polarized plant cellgrowth. Proc Natl Acad Sci USA 106: 13341–13346

Wu SZ, Bezanilla M (2012) Transient RNAi assay in 96-well plate formatfacilitates high-throughput gene function studies in planta. MethodsMol Biol 918: 327–340

Wu SZ, Ritchie JA, Pan AH, Quatrano RS, Bezanilla M (2011) Myosin VIIIregulates protonemal patterning and developmental timing in the mossPhyscomitrella patens. Mol Plant 4: 909–921

Xu P, Zhang Y, Kang L, Roossinck MJ, Mysore KS (2006) Computationalestimation and experimental verification of off-target silencing duringposttranscriptional gene silencing in plants. Plant Physiol 142: 429–440

Zhang N, Zhang D, Chen SL, Gong BQ, Guo Y, Xu L, Zhang XN, Li JF(2018) Engineering artificial microRNAs for multiplex gene silencingand simplified transgenic screen. Plant Physiol 178: 989–1001

Zhou CJ, Li J, Zou JC, Liang FS, Ye CJ, Jin DM, Weng ML, Wang B (2006)Cloning and characterization of a second form of the rice adeninephosphoribosyl transferase gene (OsAPT2) and its association withTGMS. Plant Mol Biol 60: 365–376

Zimmer AM, Pan YK, Chandrapalan T, Kwong RWM, Perry SF (2019)Loss-of-function approaches in comparative physiology: is there a futurefor knockdown experiments in the era of genome editing? J Exp Biol 222:jeb175737




Robust Survival-Based RNA Interference of Gene Families ...blots, is stymied. Furthermore, certain...

Documents

Transcript of Robust Survival-Based RNA Interference of Gene Families ...blots, is stymied. Furthermore, certain...