Phytopathology®  2022  112:608-619  https://doi.org/10.

1094/PHYTO-09-20-0398-R

Molecular and Physiological Plant Pathology

Exonic Circular RNAs Are Involved in Arabidopsis Immune Response

Against Bacterial and Fungal Pathogens and Function Synergistically with
Corresponding Linear RNAs
Lin Wang,1,2 Jiao Li,1,2 Baohuan Guo,1,2 Le Xu,1,2 Leyao Li,1,2 Xiaoning Song,1,2 Xiaoyan Wang,1,2 Xuebin Zeng,1,2
Lihua Wu,1,2 Dongdong Niu,1,2 Kai Sun,3 Xiaoyong Sun,3,† and Hongwei Zhao1,2,†
1
College of Plant Protection, Nanjing Agricultural University, Nanjing 210095, China
2
The Key Laboratory of Plant Immunity, Nanjing Agricultural University, Nanjing 210095, China
3
Big Data Research Center, College of Information Science, Shandong Agricultural University, Tai’an 271018, China
Accepted for publication 25 August 2021.

ABSTRACT

Circular RNAs (circRNAs) are a group of covalently closed RNAs, responses. We discovered that circR194 and circR4022 are involved in
and their biological function is largely unknown. In this study, we plant response against P. syringae infection, whereas circR11208 is
focused on circRNAs that are generated from exon back-splicing (exonic involved in response against Botrytis cinerea infection. Intriguingly, our
circRNAs). The linear RNA counterparts encode functional proteins so results indicate that these exonic circRNAs function synergistically with
that we can compare and investigate the relationship between circular their corresponding linear RNAs. Furthermore, circR4022 and circR11208
and linear RNAs. We compared circRNA expression profiles between also play substantial roles in Arabidopsis tolerance to salt stress. This study
untreated and Pseudomonas syringae-infected Arabidopsis and identified extends our understanding of the molecular functions of plant circRNAs.
and experimentally validated differentially expressed exonic circRNAs by
multiple approaches. We found that exonic circRNAs are preferentially Keywords: bacterial pathogens, disease resistance, fungal pathogens,
enriched in biological processes that associate with biotic and abiotic stress molecular, plant immune responses, plant stress and abiotic disorders

Circular RNAs (circRNAs) are covalently closed circular RNA circRNAs (Jeck et al. 2013). In the second scenario where the
molecules that are widely conserved both in eukaryotes and prokar- looping is mediated by base pairing, normally it is facilitated by
yotes (Lai et al. 2018; Li et al. 2018). CircRNAs form a “head-to- cis- and trans-elements. The cis-elements can be inverted repeat
tail” structure (Dragomir and Calin 2018; Ebbesen et al. 2017; elements (such as Alu elements), located in the upstream and
Memczak et al. 2013; Sanger et al. 1976) and lack a capped 59 end downstream introns (Jeck and Sharpless 2014; Zhang et al. 2014).
and polyadenylate 39 end (Suzuki and Tsukahara 2014). In contrast The trans-elements can be RNA binding proteins that bind to spe-
to linear RNAs that undergo a classic splicing procedure, circRNAs cific motifs in the flanking introns (such as the quaking protein
are generated through a completely different splicing mechanism HQK encoded by QKI) (Ashwal-Fluss et al. 2014; Conn et al.
called back-splicing, which refers to the looping of the splicing 2015; Greene et al. 2017; Li et al. 2018; Zhang et al. 2018b).
donor site of a downstream intron and the splicing acceptor site of Needed circRNAs can be classified into four categories, which
a upstream intron (Memczak et al. 2013). The looping happens nat- are ecircRNAs that contain only exons (Memczak et al. 2013), cor-
urally as an intermediate step of splicing, which may or may not be neas that contain only introns (Zhang et al. 2013), EIciRNAs that
mediated by base pairing between inverted repeat elements. In the contain both exons and introns (Li et al. 2015), and circRNAs
first scenario, exons undergo internal back-splicing and generate exon- whose biogenesis occurs through circulation of transfer RNA,
containing circRNA, called exon skipping. The regular intron lariats rRNA, or small nuclear RNA (Grabowski et al. 1981; Guarnerio
escape from debranching destination, generating intron-containing et al. 2016; Li et al. 2015; Zhang et al. 2018b). CircRNAs have
multiple characteristics that are associated with unique structural
† features. First, circRNAs are more resistant to RNase R exonucle-
Corresponding authors: H. Zhao; [email protected],
and X. Sun; [email protected] ase activity than their linear RNA counterparts, largely because of
their circular structure (Memczak et al. 2013; Suzuki and Tsukahara
Author contributions: X. Sun and H. Zhao designed the experiments; L. Wang,
J. Li, B. Guo, L. Xu, L. Li, X. Song, X. Wang, X. Zeng, L. Wu, and K. Sun per-
2014). Second, circRNAs are conserved among multiple species,
formed all the experiments; L. Wang, X. Sun, and H. Zhao analyzed the results; including Homo sapiens, Mus musculus, Danio rerio, Drosophila
L. Wang and H. Zhao wrote the manuscript; H. Zhao supervised the project. All melanogaster, Caenorhabditis elegans, Oryza sativa, and Arabidop-
authors read and approved the final manuscript. sis thaliana (Chen 2016; Jeck et al. 2013; Lai et al. 2018; Memczak
Funding: This work is supported by the Fundamental Research Funds for the Cen- et al. 2013; Wang et al. 2014; Xia et al. 2017). Third, circRNA
tral Universities (grant KYXK202009) to H. Zhao; the Top-notch Academic Pro- expression levels are associated with different tissues and develop-
grams Project of Jiangsu Higher Education Institutions (grant PPZY2015B157) to mental stages in the nucleus and cytoplasm (Dong et al. 2017;
X. Wang, X. Zeng and L. Wang; and the National Natural Science Foundation of
China (grant 31571306) to X. Sun.
Memczak et al. 2013; Salzman et al. 2012; Xia et al. 2017).
CircRNAs may function through diverse mechanisms. Some
*The e-Xtra logo stands for “electronic extra” and indicates there are supplemen- circRNAs can bind to microRNA, which blocks small RNA-
tary tables, four supplementary figures, and one supplementary file published online.
mediated gene silencing by “sponging” microRNA (Hansen et al.
The author(s) declare no conflict of interest. 2013; Memczak et al. 2013). Some circRNAs are able to regulate
the transcription of their linear RNA counterparts (Li et al. 2015;
© 2022 The American Phytopathological Society Senfter et al. 2015; Zhang et al. 2013), and others are known to

608 PHYTOPATHOLOGY®
regulate their translation. For example, some circRNAs suppress mutants was PCR amplified to identify homozygotes, with pri-
messenger RNA (mRNA) translation by sequestering or interfering mers LBb1.3/SALK_107769C-LP/SALK_107769C-RP, LBb1.3/
with the function of key translational factors (Holdt et al. 2016). SALK_045494C-LP/SALK_045494C-RP, and LBb1.3/SALK_
Some circRNAs themselves can be translated into proteins. For 093620C-LP/SALK_093620C-RP. Transgenic plants including
example, circ-ZNF609 can be translated into protein and functions circR194-OE, circR4022-OE, circR11208-OE, circR194-RNAi,
in myogenesis (Legnini et al. 2017). Some circRNAs are involved circR4022-RNAi, and circR11208-RNAi were in an Arabidopsis
in various biological processes such as stimulus response (Fischer ecotype Col-0 background and confirmed by PCR with con-
and Leung 2017), human and animal diseases (Du et al. 2017; Guo structed vector primers and real-time PCR (RT-PCR) assays of
et al. 2014), and plant response against biotic and abiotic stresses circRNAs expression levels (primers are listed in Supplementary
(Khraiwesh et al. 2012; Lai et al. 2018; Suzuki et al. 2014). In Table S5).
recent years, many plant circRNAs have been found to be associ- circRNA library construction. Arabidopsis Col-0 wild type
ated with the responses to multiple developmental and environ- (wt) plants were subjected to mock treatment (10 mM MgCl2) or
mental cues (Wang et al. 2018; Ye et al. 2015; Zuo et al. 2016). treated with P. syringae pv. tomato avrRpt2 as previously described.
For instance, maize infected with maize Iranian mosaic virus had Four mock-treated and five P. syringae pv. tomato avrRpt2-treated
about 160 circRNAs that had significant expression variations circRNA libraries were constructed and underwent the CircleSeq
(Ghorbani et al. 2018). Arabidopsis under heat stress had 1,583 protocol (Sun et al. 2016). RNA was isolated with an RNeasy RNA
circRNAs that were identified as heat stress-specific circRNAs extraction kit (Qiagen). DNA was removed with an on-column
(Pan et al. 2018). DNase digestion kit (Qiagen). Ribosomal RNA was removed with a
To identify circRNAs that may be involved in plant immune RiboMinus transcriptome isolation kit in six individual 10-lg reac-
responses against bacterial pathogen infection, we challenged Ara- tions per the standard protocol (Invitrogen) and repooled. Linear
bidopsis with mock and Pseudomonas syringae pv. tomato RNA was depleted with RNase R (Epicentre). In brief, the RNA
(avrRpt2) strains. We constructed nine circRNA libraries (four sample was set up in six 14.3-ll individual reactions. Diluted sam-
mock and five P. syringae pv. tomato avrRpt2) and compared dif- ples alone were briefly heated to 70 C to denature, then cooled to
ferent behaviors of circRNAs after P. syringae pv. tomato avrRpt2 40 C on a thermocycler. For each reaction, 1.7 ll of 10× RNase R
infection and found that Arabidopsis circRNAs could be generated buffer, 1 ll of water, and 1 ll of RNase R were added and mixed
through both canonical and noncanonical splicing. It is facilitated gently. The reactions were allowed to proceed at 40 C for 1 h. The
by complementary sequences that exist not only in introns but also produced RNA samples were fragmented, annealed to adapters,
in the sequences flanking splice sites (Sun et al. 2016). reversed transcribed, and constructed to circRNA libraries by
However, with the rising number of newly identified circRNAs Genergy per the standard protocol (Shanghai, China).
and a better understanding of how circRNAs function, it is interest- We also performed a search of the reported Arabidopsis circRNA
ing to investigate the biological role of circRNAs in plants, espe- data (Sun et al. 2016), which came from seven RNA sequencing
cially in responses to various biotic and abiotic stresses. Meanwhile, datasets by five independent studies. The seven RNA sequencing
it is unknown whether circRNAs function independently or cooper- datasets are SRR505743, SRR505744, SRR505745, SRR1170682,
atively with linear RNAs. Simultaneously investigating the function SRR944363, SRR352212, and SRR953400.
of circRNAs and the corresponding linear RNAs that originated Bioinformatic analysis. Bioinformatic analysis was done as
from the identical genomic loci will help us dissect the relationship described in our previous report (Sun et al. 2016). Briefly, raw RNA
between these different but related molecules. We specifically sequencing reads were aligned in TopHat version 2.0.6 with parame-
focused on Arabidopsis circRNAs originating from coding regions ters min-intron-length 20 and -max-multihits 1 against TAIR 10.
(called exonic circRNAs in this study) so that we can simultaneously Mapped reference alignments were sorted and indexed in SAMtools.
study the function of these exonic circRNAs and their linear RNA Reads that did not map to the reference genome were realigned to the
counterparts for their role against P. syringae pv. tomato or Botrytis genome in BWA-MEM, and low-quality mapping was filtered with a
cinerea infections. We discovered many stimulus-responding circR- threshold of 10. Screening of the paired chiastic clipping signals was
NAs by comparing expression profiles between untreated and P. done by extracting the proper reads and identifying the back-splicing
syringae pv. tomato-treated plants. The results show that exonic reads in R version 3.2.2. Additionally, junction reads were extracted
circRNAs are significantly enriched in biological processes associ- from the BMA files with TopHat mapping, and the two datasets were
ated with resistance or tolerance against biotic and abiotic stress combined based on the splice sites of the circRNAs.
responses. We identified and validated three exonic circRNAs that To identify the linear splicing junction reads, we extracted lin-
are associated with Arabidopsis resistance against P. syringae pv. ear splicing junction reads that overlapped with the back-splicing
tomato or B. cinerea infections. We also characterized two circRNAs reads within five nucleotides on the same strand by using Biocon-
that are involved in Arabidopsis tolerance against salt and osmotic ductor packages such as Biostrings, Genomic Alignments, and
stresses. Moreover, these exonic circRNAs function synergistically SplicingTypesAnno. Canonical and noncanonical splicing signals
with their linear RNAs. Our study reveals that circRNAs are actively were determined from the linear splicing junction reads. CircRNA
involved in plant responses against biotic and abiotic stresses, and was quantified from the reads supporting the back-splicing events
exonic circRNAs may be inherently associated with their linear RNA and was used as a metric for evaluating the expression of circRNAs.
counterparts. For cross-sample comparison, reads per million mapped reads was
used as a basis of quantification and normalization according to the
MATERIALS AND METHODS following formula: Normalized expression = circ_raw read counts/
total number of mapped reads × 106, where circ_raw read counts
Plant material and growth conditions. A. thaliana ecotype represent the read counts supporting back-splicing, and the total
Columbia (Col-0), T-DNA insertion mutants, and transgenic plants number of mapped reads are those mapped in one sample.
were grown in soil under controlled conditions with 8 h light/16 h Plant transgenic constructs. CircRNA overexpression was con-
dark at 22 to 24 C. Arabidopsis T-DNA insertion mutants of structed according to the method published by Conn et al. (2017).
AT3G29185 (SALK_107769C, T-DNA insertion at the 300- Specifically, the 60-bp intron upstream of exon 4 (AT3G29185),
UTR3), AT5G26340 (SALK_045494C, T-DNA insertion at the exon 4, and the reverse complement sequence of the 60-bp intron
third exon), and AT1G19860 (SALK_093620C, T-DNA insertion upstream of exon 4 were cloned together to overexpress circR194
at the 300-UTR5) were obtained from Arabidopsis Biological (circR194-OE). Similarly, the 60-bp intron upstream exon 2
Resource Center (The Ohio State University, Columbus, OH). (AT1G19860), second, third, and fourth exons, and the reverse com-
Genomic DNA extraction (CTAB method) of T-DNA insertion plement sequence of the 60-bp intron upstream exon 2 were cloned

Vol. 112, No. 3, 2022 609

together to overexpress circR4022 (circR4022-OE). The fractions. Then, the two RNA fractions (oligo[dT]+RNA and
circR11208-OE was constructed by the same procedures to the sec- oligo[dT]-RNA) were reextracted and resuspended in RNA-free
ond exon of AT5G26340. The two 60-bp completely complemen- water. The oligo(dT)+RNA and oligo(dT)-RNA were reverse
tary sequences positioned to the 59 and 39 ends of the exon/exons transcribed with a PrimeScript One Step RT-PCR Kit with ran-
were used to create hairpin structures for the primary transcript and dom primer. cDNAs were used for RT-PCR (with a ChamQ
promote back-splicing for generating circRNA (Conn et al. 2017). SYBR qPCR Master Mix kit; Vazyme Biotech, Piscataway, NJ)
For the circRNA RNAi constructs, specifically designed artificial with three biological replicates. All primers are listed in Supple-
microRNAs targeting the junction regions (11 bp in the left of mentary Table S5.
back-splice site and 10 bp in the right of back-splice site) of the
circRNAs were engineered. The constructs were designed according RESULTS
to a standard artificial microRNA expression protocol (http://wmd3.
weigelworld.org). All amplified sequences were cloned into the Exonic circRNAs are active ingredients of circRNAs. From
pENTR vector. the nine circRNA libraries (four mock and five P. syringae pv.
All constructed vectors were finally cloned into pEarleyGate202 tomato avrRpt2) sequenced in 2016, we identified 765 million reads
(35S promoter) binary plasmid and transformed into Arabidopsis by Illumina HighSeq 2000, in which 71,654 circRNAs and 15.6 M
ecotype Col-0 by the Agrobacterium-mediated (GV3101 strain) linear RNAs were identified, corresponding to 5.73× of the Arabi-
floral-dip transformation method. dopsis genome (Sun et al. 2016). In this study, we first investigated
P. syringae pv. tomato DC3000 and P. syringae pv. tomato the biogenesis and function of Arabidopsis circRNAs by aligning
avrRpt2 inoculation. Four-week-old Col-0 wt plants and the these circRNAs to the Arabidopsis genome according to sequence
transgenic plants were inoculated with 10 mM MgCl2, P. syringae matches (100%). We identified 803 circRNAs that originated from
pv. tomato DC3000, or P. syringae pv. tomato avrRpt2 strains the nuclear (93%) and the organelle genome (7%). Chromosomes 1
at 1 × 107 cfu/ml. After 10 h, plant samples were collected, and and 5 generate 26 and 25% of circRNAs, respectively; chromo-
circRNAs expression levels were detected by RT-PCR assay (three somes 2, 3, and 4 harbor 15, 15, and 12% of circRNAs, respec-
biological replicates). Hypersensitive response (HR) reaction was tively (Fig. 1A).
monitored for 12 h after the plants were inoculated with P. syringae We also searched for reported Arabidopsis circRNA data.
pv. tomato avrRpt2 strains at 1 × 107 cfu/ml (30 plants per sample, Together with circRNAs identified by our sequencing, we identified
five biological replicates). 11,283 unique circRNAs (Supplementary Table S1). These 11,283
B. cinerea challenge assay. B. cinerea strain B1301 was cul- circRNAs originate from different genomic loci, including inter-
tured on potato sucrose agar for 7 days. Spores were suspended in genic regions, introns, or exons. We found that about 41.61%
water and adjusted to 1 × 106 spores/ml. We inoculated 3-week-old (4,694) of circRNAs were generated from coding regions (Supple-
Col-0 wt plants and the transgenic plants by dropping 10 µl of B. mentary Table S2). These circRNAs associated with coding regions
cinerea spore suspension on the midvein of each half of the leaves. could be classified into three categories: circRNAs covering coding
After inoculation, all plants were grown in the growth room at 70% regions (5.95%), circRNAs overlapping with coding regions
relative humidity for 2 days (30 plants per sample, three biologi- (1.42%), and circRNAs residing within coding regions (34.24%).
cal replicates). Our results indicate that circRNAs associated with coding regions
Ion leakage measurement. We measured leaf cell death rate account for a large proportion of total circRNAs (Fig. 1B). We also
by examining the ion leakage rate. Five leaves from each sample examined the numbers of exons contained by each circRNA. We
(1 cm in diameter) were suspended in 5 ml of sterile water (in a found that the majority of Arabidopsis circRNAs contain only one
15-ml Falcon tube). Samples were kept at room temperature for 3 h exon, but circRNAs containing more than one exon were also fre-
before conductivity was measured (three biological replicates) quently observed. However, with the increasing number of exons,
(Bante 950, Bante Instruments, China). The conductivity of boiled the percentage of circRNAs in the total population dropped steadily,
samples (100 C for 25 min) was set at 100%. indicating that increased exon number (or length of RNA) reduces
Salt and mannitol stress. Four-day-old seedlings of Col-0 wt the chance for the 39 and 59 ends of the transcript to join. However,
plants, T-DNA insertion mutants, and transgenic plants germinated we did identify several circRNAs consisting of up to eight exons,
on Murashige and Skoog medium and then were transferred to Mura- indicating that back-splicing of multiple exons is possible (Fig. 1C).
shige and Skoog medium with 0, 100 mM NaCl or 0, 200 mM man- Hereafter, we designate these circRNAs as exonic circRNAs. The
nitol. Plants were grown in the incubator with a 12 h light/12 h dark results prompted us to study the function and operating mechanism
light cycle and were kept at 22 to 24 C. After 10 days, seedlings of of the exonic circRNAs.
circRNA expression levels were detected by RT-PCR assay, and Exonic circRNAs are preferentially involved in plant
root length was measured (30 plants per sample, three biologi- responses to stimuli. The circRNAs were aligned to their linear
cal replicates). RNA counterparts to identify candidates that have exact boundary
RNA extraction and RT-PCR. Total RNA was extracted by matches with linear RNA. These circRNAs were probably gener-
the TRIzol reagent according to the manufacturer’s protocol (Invi- ated through back-spliced exons that could be great candidates for
trogen, Waltham, MA). rRNA was removed with a Ribo-Zero studying the functional relationship between circular and linear
Magnetic Kit (Epicentre, Madison, WI). RNAs. We found 757 circRNAs that have at least one terminus
For each 20-ll RNase R treatment reaction, 6 lg of total RNA exactly matched to an exon boundary (gap = 0) (Supplementary
(rRNA removed) was added to 1 ll of 20 U/ll RNase R enzyme Table S3). To examine whether these circRNAs shared any com-
(Epicentre, Madison, WI), 2 ll of 10× RNase R buffer, and RNA- mon biological functions, we performed a Gene Ontology (GO)
free water and incubated for 30 min at 37 C. RNA was reextracted analysis of the gene information assigned to the coding gene of
and resuspended in 25 ll of RNA-free water. Then, the PrimeScript each specific genome locus. Surprisingly, GO analysis revealed
One Step RT-PCR Kit (the kit has DNase I; Takara Bio, Beijing, that these identified exonic circRNA-generating loci have a close
China) and random primer (NNNNNN) or oligo(dT) primer were link with plant responses to stimuli. In the top 30 GO categories of
used to generate complementary DNA (cDNA) from the RNA. The these 757 circRNAs, 18 were associated with plant response to
cDNA was the template of RT-PCR. stimuli (Table 1). We observed enrichment of biological processes
Total RNA (0.15 mg) without rRNA was used to enrich mRNA with responding to biotic stimuli, such as “response to fungus” (GO:
an mRNA enrichment kit (Thermo Scientific, Waltham, MA). RNA was 0009620), “response to bacterium” (GO: 0009617), “defense
passed through an oligo(dT) polyA+-binding column to separate poly- response to bacterium” (GO: 0042742), and “response to wounding”
A+(oligo[dT]+-RNA) and polyA-(remaining RNA/oligo[dT]-RNA) (GO: 0009611). We also observed a significant enrichment of

610 PHYTOPATHOLOGY®
biological processes associated with responses to abiotic stimuli, removes polyadenylated linear RNAs, divergent PCRs that specifi-
such as “response to temperature stimulus” (GO: 0009266), cally amplify circular but not linear RNAs, sequence-specific
“response to heat” (GO: 0009408), “response to inorganic sub- primer-directed reverse transcription, and Sanger sequencing of
stance” (GO: 0010035), “response to metal ion” (GO: 0010038), amplified junction regions (Supplementary Fig. S1). Thirteen out
and “response to water” (GO: 0009415) (Table 1). Taken together, of 15 circRNAs were successfully verified by divergent PCR fol-
our results suggest that these exonic circRNA-generating loci are lowed by Sanger sequencing (Fig. 2B). Sequencing results con-
involved in Arabidopsis response to stimuli. Therefore, these exonic firmed that these circRNAs were generated by back-splicing of
circRNAs’ function may be associated with their linear RNAs. exons as predicted (Supplementary File S1). Taken together, these
To validate the stimulus-related circRNAs, we selected 15 results demonstrate the existence of a substantial set of exonic
circRNAs from eight GO categories for experimental validation. circRNAs that are generated from genome loci associated with var-
These selected GO categories were “defense response” (GO: ious responses to stimuli. The detailed function of these circRNAs
0006952), “defense response to bacterium” (GO: 0042742), is worth further investigation.
“defense response to fungus” (GO: 0009817), “systemic acquired circRNAs and their corresponding linear RNAs are
resistance” (GO: 0009627), “response to salt stress” (GO: synergistically involved in response to P. syringae pv. tomato
0009651), “response to water” (GO: 0009415), “response to water and B. cinerea infection. To test whether these circRNAs are
deprivation” (GO: 0009414), and “response to abscisic acid” (GO: really associated with stimulus response, we examined the expres-
0009737). These GO categories are typical responses against biotic sion levels of these 13 exonic circRNAs that are generated from
and abiotic stimuli, some of which correspond closely to our P. genome loci associated with responses to stimulus. We challenged
syringae pv. tomato-treated samples (Fig. 2A). We used a set of Arabidopsis with P. syringae pv. tomato DC3000 and P. syringae
stringent criteria that guarantee the correctness of the validated pv. tomato avrRpt2. When the expression levels of circRNAs were
circRNAs. These validating strategies include RNase R treatment measured, we observed varied expression of eight circRNAs,
that specifically removes linear RNAs, oligo(dT) depletion that namely circR194, circR1998, circR2812, circR3177, circR3574,

Fig. 1. Summary of Arabidopsis circular RNAs (circRNAs). A, Distribution of circRNAs generating loci on Arabidopsis nuclear and organelle genomes. Top:
relative read abundance (%); bottom: read distribution normalized to genome size (reads per million). B, circRNA distribution on each genomic region. Cover,
circRNAs originating from regions covering coding regions (5.95%); Intergenic, circRNAs originating from intergenic regions (58.39%); Overlap, circRNAs
originating from regions overlapping with coding regions (1.42%); Within, circRNAs originating from regions residing within coding regions (34.24%). C, circRNA
exon composition.

Vol. 112, No. 3, 2022 611

circR4022, circR6142, and circR11208 (Fig. 3A). We found that with P. syringae pv. tomato avrRpt2, circR194-RNAi lines showed
circR194 was induced by P. syringae pv. tomato DC3000 but not stronger HR symptoms than the Col-0 wt plants, whereas the
by P. syringae pv. tomato avrRpt2 infection; circR2812 was dra- circR194-OE lines had almost no discernible phenotype. The
matically induced by P. syringae pv. tomato avrRpt2 infection but AT3G29185 T-DNA plants behaved similarly to the circR194-
not significantly by P. syringae pv. tomato DC3000; circR1998, RNAi plants (Fig. 4A). HR in circR4022-RNAi lines was obvious,
circR3177, circR3574, circR4022, and circR11208 were induced by as in Col-0 wt, but was much lower in the circR4022-OE lines.
both P. syringae pv. tomato DC3000 and P. syringae pv. tomato Again, the T-DNA plants behaved similarly to the circR4022-RNAi
avrRpt2 infections, and circR6142 was suppressed by both P. syrin- plants (Fig. 4A). In contrast, both circR11208-RNAi and -OE dem-
gae pv. tomato DC3000 and P. syringae pv. tomato avrRpt2 infec- onstrated a similar HR phenotype to Col-0 wt. When the
tions. The remaining five circRNAs did not show noticeable HR-inducing activity was measured by percentage ion leakage (Fig.
expression variations upon P. syringae pv. tomato infection. Altered 4B) (Block et al. 2005; Gao et al. 2011), the concordance between
expression upon a virulent (related to effector-triggered susceptibil- circRNAs and their linear RNAs was even more obvious. In plants
ity) or an avirulent infection (related to effector triggered immunity) treated with P. syringae pv. tomato avrRpt2, both circR194-RNAi
indicates that these circRNAs may be influenced by or involved in and circR4022-RNAi showed significantly higher ion leakage rates
the responses to P. syringae pv. tomato infections. We selected after 6 h postinoculation, which resembled the phenotype of their
circR194, circR4022, and circR11208, which specifically responded linear RNA T-DNA insertion mutants (Fig. 4B). In contrast, both
to effector-triggered susceptibility or effector triggered immunity circR194-OE and circR4022-OE plants demonstrated obviously
elicitors, for further study. lower ion leakage rate (Fig. 4B). Again, circR11208-RNAi, -OE,
We also examined the GENEVESTIGATOR database (https:// and T-DNA plants showed significantly higher ion leakage rate
genevestigator.com) to check whether the linear RNAs of these 13 than the Col-0 wt plants (Fig. 4B). Taken together, our results indi-
circRNAs could be associated with resistance to B. cinerea. The cate that circR194 and circR4022 have a negative influence on Ara-
expression of the linear RNAs of circR194, circR1387, circR6142, bidopsis immune response to P. syringae pv. tomato avrRpt2
and circR11208 showed altered patterns after B. cinerea infection infection, whereas circR11208 has no noticeable effect. We then
in the GENEVESTIGATOR database (Supplementary Table S4). tested whether these circRNAs are involved in resistance against B.
However, when we validated expression levels of these four circR- cinerea infection, as suggested by their linear RNAs. When chal-
NAs by RT-PCR, we found that circR11208 was significantly lenged with B. cinerea, the circR11208-OE plants demonstrated
induced by B. cinerea infection and circR194 was significantly significantly smaller lesions, and the circR11208-RNAi lines
reduced, and circR1387 and circR6142 showed no dramatic varia- showed more severe disease symptoms (Fig. 4C). The T-DNA
tion (Fig. 3B). Therefore, circR194 and circR11208 are potentially plants also showed more severe disease symptoms than the Col-0
involved in resistance against B. cinerea. wt plants (Fig. 4C). Neither circR194 nor circR4022 plants demon-
We created transgenic Arabidopsis plants that both overexpress strated a noticeable resistant phenotype to B. cinerea infection (Fig.
(-OE) and silence (-RNAi) the expression of circRNAs (Supple- 4C). Therefore, we found that only circR11208 is involved in resis-
mentary Fig. S2). We used HR as the immune response indicator tance against B. cinerea.
to determine whether these circRNAs are involved in immune circRNAs are also responsive to salt and osmatic stresses.
responses to P. syringae pv. tomato infection and necrotic symp- Because our GO analysis also significantly associated circRNAs
toms as an indicator of resistance to B. cinerea. When challenged with responses to abiotic stimuli, we tested whether these circRNAs

TABLE 1. Gene ontology (GO) analysis of circular RNA (circRNA) generating locia
GO Term Count % P
GO: 0009628 Response to abiotic stimulus 35 14.64 3.27E-07
GO: 0055114 Oxidation reduction 25 10.46 0.003306
GO: 0010035 Response to inorganic substance 19 7.95 3.53E-05
GO: 0006952 Defense response 19 7.95 0.040101
GO: 0009266 Response to temperature stimulus 17 7.11 2.32E-06
GO: 0046686 Response to cadmium ion 15 6.28 1.88E-05
GO: 0010038 Response to metal ion 15 6.28 0.000108
GO: 0006091 Generation of precursor metabolites and energy 15 6.28 0.000183
GO: 0009620 Response to fungus 13 5.44 0.003865
GO: 0044271 Nitrogen compound biosynthetic process 12 5.02 0.027297
GO: 0016052 Carbohydrate catabolic process 11 4.60 0.000101
GO: 0009409 Response to cold 11 4.60 0.000298
GO: 0009617 Response to bacterium 11 4.60 0.00043
GO: 0009416 Response to light stimulus 11 4.60 0.037203
GO: 0009314 Response to radiation 11 4.60 0.045043
GO: 0006811 Ion transport 11 4.60 0.051043
GO: 0009408 Response to heat 10 4.18 1.28E-05
GO: 0044275 Cellular carbohydrate catabolic process 10 4.18 3.69E-05
GO: 0009415 Response to water 10 4.18 0.000239
GO: 0033554 Cellular response to stress 10 4.18 0.050744
GO: 0009414 Response to water deprivation 9 3.77 0.00083
GO: 0042742 Defense response to bacterium 9 3.77 0.001113
GO: 0016051 Carbohydrate biosynthetic process 9 3.77 0.007921
GO: 0070727 Cellular macromolecule localization 9 3.77 0.020991
GO: 0006812 Cation transport 9 3.77 0.066879
GO: 0006970 Response to osmotic stress 9 3.77 0.080432
GO: 0009611 Response to wounding 8 3.35 0.001341
GO: 0019318 Hexose metabolic process 8 3.35 0.002101
GO: 0005996 Monosaccharide metabolic process 8 3.35 0.00486
GO: 0034637 Cellular carbohydrate biosynthetic process 8 3.35 0.005453
a
Proteins encoded by exonic circRNA generating loci undergo GO analysis. The top 30 biological processes are presented.

612 PHYTOPATHOLOGY®
Fig. 2. Characterization of Arabidopsis exonic circular RNAs (circRNAs). A, Fifteen circRNA-generating genes associated with responses against biotic and abiotic
stresses are presented. Genomic locus: the corresponding linear genes encoded by a circRNA-generating locus. Divergent PCR detection: (a to m) corresponding PCR
validation results in B. ND: not detected. circRNA ID: a numerical ID assigned to each circRNA. B, circRNAs are validated by divergent PCRs. Bands with expected
sizes are marked by arrows with size given. Single and multiple circles represent circRNAs and their rollover molecules. Panels a to m correspond to circRNAs pre-
sented in A (a: circR194, locus gene AT3G29185; b: circR659, locus gene AT1G30580; c: circR1387, locus gene AT1G54100; d: circR1998, locus gene AT1G79380;
e: circR2812, locus gene AT1G59870; f: circR3177, locus gene AT1G70520; g: circR3574, locus gene AT1G70740; h: circR4022, locus gene AT1G19860; i:
circR5998, locus gene AT2G39800; j: circR6142, locus gene AT3G63190; k: circR6773, locus gene AT1G08930; l: circR8398, locus gene AT1G59610; m:
circR11208, locus gene AT5G26340; n: Actin2, locus gene AT3G18780, as a control). The circRNA generating loci are listed to the top of each gel. Arrows indicate
the direction of primers. gDNA and cDNA, genomic and complementary DNAs as PCR templates; oligo(dT) + or −, cDNA was reverse transcribed from the RNA
enriched or not by the oligo(dT); RNase R (ribonuclease R) + or −, cDNA was reverse transcribed from the RNA with or without RNase R treatment.

Vol. 112, No. 3, 2022 613

Fig. 3. Circular RNAs (circRNAs) are differentially expressed upon Pseudomonas syringae pv. tomato and Botrytis cinerea infection. A, CircRNA expression is exam-
ined by real-time PCR 10 h after P. syringae pv. tomato infection of 4-week-old Col-0 wild-type (wt) Arabidopsis. Mock, 10 mM MgCl2 infiltration; DC3000, P. syrin-
gae pv. tomato DC3000 (virulent strain) at 1 × 107 cfu/ml; avrRpt2, P. syringae pv. tomato avrRpt2 (avirulent strain) at 1 × 107 cfu/ml. These data are representative of
three independent biological experiments. ABC: P < 0.01, Student’s t test, bars indicate SE. B, CircRNA expression is examined by real-time PCR 24 h after B. cinerea
infection in Arabidopsis 3-week-old Col-0 wt plant. Mock, water only; B. cinerea, B. cinerea strain B1301. Spores were suspended in water at 1 × 106 spores/ml. These
data are representative of three independent biological experiments. **P < 0.01, Student’s t test; bars indicate SE. NA, not significant.

614 PHYTOPATHOLOGY®
Fig. 4. Circular RNAs (circRNAs) are involved in defense response against Pseudomonas syringae pv. tomato and Botrytis cinerea. Arabidopsis Col-0 wild
type or transgenic plants (circR194-OE lines, circR194-RNAi lines, AT3G29185-T-DNA line, circR4022-OE lines, circR4022-RNAi lines, AT1G19860-T-
DNA line, circR11208-OE lines, circR11208-RNAi lines, AT5G26340-T-DNA line) were challenged with P. syringae pv. tomato or B. cinerea. A, Four-
week-old plants, hypersensitive response (HR) phenotype 12 h after P. syringae pv. tomato avrRpt2 at 1 × 107 cfu/ml infection. Five independent biological
experiments, n = 30. B, The HR-inducing activity was measured as percentage ion leakage in plants, treated with P. syringae pv. tomato avrRpt2 at 1 × 107
cfu/ml for the indicated times (0, 5, 6, 7, and 8 h). These data are representative of five independent biological experiments, n = 30. **P < 0.01, Student’s t
test; bars indicate SE. C, Three-week-old plants, disease symptoms after 24 h of B. cinerea infection. Three independent biological experiments, n = 30.

Vol. 112, No. 3, 2022 615

Fig. 5. Exonic circular RNAs (circRNAs) are associated with responses to abiotic stresses. Four-day-old seedling of Col-0 wild type and the transgenic plants
(circR194-OE lines, circR194-RNAi lines, AT3G29185-T-DNA line, circR4022-OE lines, circR4022-RNAi lines, AT1G19860-T-DNA line, circR11208-OE lines,
circR11208-RNAi lines and AT5G26340-T-DNA line) were challenged with NaCl or mannitol treatments. Plants germinated on Murashige and Skoog (MS) medium
were transferred to MS medium with 0, 100 mM NaCl or 0, 200 mM mannitol. After 10 days, root length of seedlings was measured. A, circR194; B, circR4022; and
C, circR11208. These data are representative of three independent biological experiments, n = 30. **P < 0.01, Student’s t test; bars indicate SE. NA, not significant.

616 PHYTOPATHOLOGY®
are involved in responses to salt and osmatic stresses. We first sequencing (Supplementary File S1), and confirmed that these
examined the expression of both circRNAs and their linear RNAs circRNAs contain back-splicing of downstream 39 ends to upstream
before and after NaCl or mannitol treatment (Supplementary Fig. 59 ends. Altogether, we have strong evidence that the molecules we
S3). Neither circR194 nor BFA1 (AT3G29185) showed noticeable studied in this research meet all the critical criteria for circRNAs
expression alterations after 100 mM NaCl treatment; in contrast, (Chen 2016).
circR194 expression was induced while BFA1 was suppressed by circRNA functions are closely associated with their linear
200 mM mannitol treatment. Expression of circR4022 was sup- RNAs. Our results show that circR194 has a negative on P. syringae
pressed by 100 mM NaCl treatment, whereas AT1G19860 expres- pv. tomato avrRpt2 infection (Fig. 4A). circR194 originates from
sion was induced. With 200 mM mannitol treatment, expression of AT3G29185, which encodes a chloroplast protein BFA1. BFA1 muta-
circR4022 remained unchanged, although the expression of tion may contribute to plant immunity by disrupting its interactions
AT1G19860 was induced. Both circR11208 and the linear RNA with the chloroplast ATP synthase subunits (Zhang et al. 2018a).
(AtSTP13: AT5G26340) were induced by 100 mM NaCl challenge, Therefore, circR194 may function by affecting its linear RNA func-
but with 200 mM mannitol treatment, only the linear RNA was tion. Similarly, we observed that circR4022 not only had a negative
induced (Supplementary Fig. S3). Our results suggest that circR194 role in Arabidopsis immune response against P. syringae pv. tomato
may be associated with tolerance to osmotic stress, and circR4022 avrRpt2 infection (Fig. 4A) but also influenced Arabidopsis tolerance
and circR11208 may play a part in tolerance to salt stress. to salt stress (Fig. 5B). circR4022 is generated at the chr1:6891710-
To validate our hypothesis, we challenged circR194-related 6893237 locus AT1G19860. AT1G19860 encodes a zinc finger pro-
plants (-OE, -RNAi, and T-DNA insertion plants) with 200 mM tein that is involved in intron retention in response to abiotic stress
mannitol or circR4022- and circR11208-related plants (-OE, (e.g., cold), in both Arabidopsis and wheat (Mastrangelo et al. 2005).
-RNAi, and T-DNA insertion plants) with 100 mM NaCl and mea- Intron retention and malformation are closely related to circRNA bio-
sured root length, which is a typical indicator of plant tolerance to genesis (Cheng et al. 2018). Therefore, our discovery not only is sup-
stresses. The circR194-OE, -RNAi, and BFA1 T-DNA insertion ported by previous study but also is extended to biological stress
plants exhibited significant reductions in root length in the normal (P. syringae pv. tomato infection) and other abiotic stress (salt).
condition; with 200 mM mannitol treatment, none of these mutant Our results show that circR11208 increases Arabidopsis resis-
plants showed aggravated or rescued root shortening phenotypes tance against B. cinerea infection (Fig. 4C) and tolerance to salt
(Fig. 5A and Supplementary Fig. S4), indicating that circR194 may stress (Fig. 5C). circR11208 is generated by the chr5:9245850-
not play a significant role in tolerance against osmotic stress. On 9246301 locus. The gene at this locus (AT5G26340) encodes the
the contrary, circR4022-RNAi and T-DNA insertion plants showed Arabidopsis Sugar Transport Protein 13 (AtSTP13), which demon-
suppressed root development under normal conditions; however, strates high hexose affinity and hexose-specific/H+ symporter activ-
after 100 mM NaCl treatment, the effect on root length decreased ity. AtSTP13 is a monosaccharide transmembrane transporter that
(Fig. 5B and Supplementary Fig. S4). circR11208 and its linear responds to abscisic acid, salt stress, and water deprivation.
RNA mutants showed significantly reduced root length under nor- AtSTP13 was found to be induced by abiotic stress, with low
mal conditions; interestingly, after 100 mM NaCl treatment, the expression under normal conditions. In roots under high salinity
root length difference between Col-0 wt and mutant plants conditions, the amount of STP13-dependent glucose uptake
became nonsignificant (Fig. 5C and Supplementary Fig. S4). increases. Furthermore, the amount of glucose efflux from stp13
These results may suggest that circR4022 and circR11208 play mutants was higher than that from wild-type plants under high-
roles in Arabidopsis tolerance to salt stress. salinity conditions. These results indicate that STP13 can reabsorb
the monosaccharides that are released by damaged cells under high-
DISCUSSION salinity conditions (Yamada et al. 2011). AtSTP13 is also associated
with defense against bacterial (Yamada et al. 2016) and fungal
Exonic circRNAs are confirmed by multiple evidence. The pathogens (Lemonnier et al. 2014). We found significant agreement
loop structure forms the basis of circRNA’s molecular character between the function of circR11208 and AtSTP13, indicating
and biological function, which is generated by back-splicing of a circR11208 function through its linear RNA.
downstream 39 end to an upstream 59 end (Jeck and Sharpless Exonic circRNAs may function synergistically with their
2014; Li et al. 2018). Therefore, it is important to validate the circu- linear RNAs. In this study, we showed that circR194, circR4022,
lar structure of circRNAs. In this study we used multiple experi- and circR11208 functions are associated with their linear RNAs.
mental procedures to validate these circRNAs (Supplementary Fig. However, the relationship between circRNAs and their linear
S1). First, we used divergent PCR that specifically amplifies circR- RNAs is unknown. For example, it is unknown whether these two
NAs but not the corresponding linear RNAs. Our validation ampli- kinds of molecules are associated through biogenesis, biological
fied bands with expected sizes from each candidate, indicating that activity, or interaction. Our results show that when plants were
these RNA molecules circulate and form junctions at predicted challenged with P. syringae pv. tomato or B. cinerea, all three
regions. Moreover, we observed that the PCR amplicons passed the circRNAs demonstrated synergistic patterns with their linear RNAs
original primer-annealing positions and rolled over into a second (Fig. 4). This synergy indicates that these exonic circRNAs work
round of amplification (double-circle marks in Fig. 2B). This find- collaboratively with their linear RNAs on similar or even identical
ing indicates that these molecules are circles. Second, we used biological processes. One mechanism by which circRNAs can col-
RNase R exonuclease that exclusively digests linear RNAs but not laborate with their linear RNAs is direct translation (Legnini et al.
the circRNAs. It demonstrates that these molecules were signifi- 2017). The translated short peptides by circRNAs may possess par-
cantly resistant to RNase R exonuclease. Third, we found that these tial or similar functions to their linear RNAs. Previous reports have
molecules were obviously enriched in the oligo(dT)-depleted frac- shown that some circRNAs are associated with translating ribo-
tion, indicating that these RNA molecules do not have poly(A) tails somes. These circRNAs use the start codon of the hosting mRNAs,
at the 39 end (a landmark of linear RNAs). Fourth, we showed that have evolutionarily conserved termination codons, and undergo
these detected molecules were not derived from aberrant amplifica- cap-independent translation. Interestingly, some circRNA transla-
tion of genomic DNA. Neither the identical divergent primers nor tions are responsive to stimuli such as starvation (Pamudurti et al.
the convergent primers could amply these bands when genomic 2017). Therefore, direct translation may explain the synergistic
DNA was used as template DNA. Fifth, we proved that these bands function between circR194, circR4022, and circR11208 and their
were not from accidental amplification of linear cDNA because linear RNAs.
these bands were absent in samples reverse-transcribed with an oli- There is another mechanism whereby these circRNAs affect the
go(dT) primer. Sixth, we examined each junction region by Sanger transcription activity of their linear RNAs. For example, Arabidopsis

Vol. 112, No. 3, 2022 617

circRNA derived from exon 6 of the SEPALLATA3 (SEP3) gene Guarnerio, J., Bezzi, M., Jeong, J. C., Paffenholz, S. V., Berry, K., Naldini,
can bind strongly to its cognate DNA locus, forming an RNA–DNA M. M., Lo-Coco, F., Tay, Y., Beck, A. H., and Pandolfi, P. P. 2016.
hybrid, which results in increased abundance of exon-skipped alter- Oncogenic role of fusion-circRNAs derived from cancer-associated chro-
mosomal translocations [published correction appears in Cell 166:1055-
native splicing variant, ultimately leading to floral homeotic pheno- 1056]. Cell 165:289-302.
types (Conn et al. 2017). In another example, EIciRNAs interact Guo, J. U., Agarwal, V., Guo, H., and Bartel, D. P. 2014. Expanded identification
with U1 small nuclear ribonucleoproteins and increase their parent and characterization of mammalian circular RNAs. Genome Biol. 15:409.
gene expression by prompting Pol II association (Li et al. 2015). Hansen, T. B., Jensen, T. I., Clausen, B. H., Bramsen, J. B., Finsen, B.,
Therefore, it is also likely that circR194, circR4022, and circR11208 Damgaard, C. K., and Kjems, J. 2013. Natural RNA circles function as
may regulate their parent gene splicing or expression. Although the efficient microRNA sponges. Nature 495:384-388.
Holdt, L. M., Stahringer, A., Sass, K., Pichler, G., Kulak, N. A., Wilfert, W.,
exact mechanism is not known yet, out results provide evidence that and Teupser, D. 2016. Circular non-coding RNA ANRIL modulates ribo-
circR194, circR4022, and circR11208 can collaborate with their cor- somal RNA maturation and atherosclerosis in humans. Nat. Commun. 7:
responding linear RNAs and contribute to Arabidopsis responses to 12429.
biotic or abiotic stimuli. Further study of the interplay between Jeck, W. R., and Sharpless, N. E. 2014. Detecting and characterizing circular
circRNAs and the corresponding linear RNAs will lead to a thor- RNAs. Nat. Biotechnol. 32:453-461.
ough and an in-depth understanding of this novel and mysterious Jeck, W. R., Sorrentino, J. A., Wang, K., Slevin, M. K., Burd, C. E., Liu, J.,
Marzluff, W. F., and Sharpless, N. E. 2013. Circular RNAs are abundant,
group of RNAs. Our study establishes the relevance of circRNA conserved, and associated with ALU repeats. RNA 19:141-157.
and plant immunity against bacterial and fungal pathogens. Our dis- Khraiwesh, B., Zhu, J. K., and Zhu, J. 2012. Role of miRNAs and siRNAs
covery will be important in elucidating the importance of circRNA in biotic and abiotic stress responses of plants. Biochim. Biophys. Acta
and exploring the utility of circRNAs in crop breeding and disease 1819:137-148.
resistance management. Lai, X., Bazin, J., Webb, S., Crespi, M., Zubieta, C., and Conn, S. J. 2018.
CircRNAs in plants. Adv. Exp. Med. Biol. 1087:329-343.
Legnini, I., Di Timoteo, G., Rossi, F., Morlando, M., Briganti, F., Sthandier,
ACKNOWLEDGMENTS O., Fatica, A., Santini, T., Andronache, A., Wade, M., Laneve, P., Rajewsky,
N., and Bozzoni, I. 2017. Circ-ZNF609 is a circular RNA that can be
We thank Gaurav Sablok for stimulating discussions. translated and functions in myogenesis. Mol. Cell 66:22-37.E9.
Lemonnier, P., Gaillard, C., Veillet, F., Verbeke, J., Lemoine, R., Coutos-
Thevenot, P., and La Camera, S. 2014. Expression of Arabidopsis sugar
LITERATURE CITED transport protein STP13 differentially affects glucose transport activity
and basal resistance to Botrytis cinerea. Plant Mol. Biol. 85:473-484.
Ashwal-Fluss, R., Meyer, M., Pamudurti, N. R., Ivanov, A., Bartok, O., Li, X., Yang, L., and Chen, L. L. 2018. The biogenesis, functions, and
Hanan, M., and Kadener, S. 2014. CircRNA biogenesis competes with challenges of circular RNAs. Mol. Cell 71:428-442.
pre-mRNA splicing. Mol. Cell 56:55-66. Li, Z., Huang, C., Bao, C., Chen, L., Lin, M., Wang, X., Zhong, G., Yu, B.,
Block, A., Schmelz, E., O’Donnell, P. J., Jones, J. B., and Klee, H. J. 2005. Hu, W., Dai, L., Zhu, P., Chang, Z., Wu, Q., Zhao, Y., Jia, Y., Xu, P.,
Systemic acquired tolerance to virulent bacterial pathogens in tomato. Liu, H., and Shan, G. 2015. Exon-intron circular RNAs regulate transcrip-
Plant Physiol. 138:1481-1490. tion in the nucleus [published correction appears in Nat. Struct. Mol. Biol.
Chen, L. L. 2016. The biogenesis and emerging roles of circular RNAs. Nat. 24:194]. Nat. Struct. Mol. Biol. 22:256-264.
Rev. Mol. Cell Biol. 17:205-211. Mastrangelo, A. M., Belloni, S., Barilli, S., Ruperti, B., Di Fonzo, N., Stanca,
Cheng, J., Zhang, Y., Li, Z., Wang, T., Zhang, X., and Zheng, B. 2018. A A. M., and Cattivelli, L. 2005. Low temperature promotes intron retention
lariat-derived circular RNA is required for plant development in Arabidopsis. in two e-cor genes of durum wheat. Planta 221:705-715.
Sci. China Life Sci. 61:204-213. Memczak, S., Jens, M., Elefsinioti, A., Torti, F., Krueger, J., Rybak, A.,
Conn, S. J., Pillman, K. A., Toubia, J., Conn, V. M., Salmanidis, M., Maier, L., Mackowiak, S. D., Gregersen, L. H., Munschauer, M., Loewer,
Phillips, C. A., Roslan, S., Schreiber, A. W., Gregory, P. A., and Goodall, A., Ziebold, U., Landthaler, M., Kocks, C., le Noble, F., and Rajewsky,
G. J. 2015. The RNA binding protein quaking regulates formation of N. 2013. Circular RNAs are a large class of animal RNAs with regulatory
circRNAs. Cell 160:1125-1134. potency. Nature 495:333-338.
Conn, V. M., Hugouvieux, V., Nayak, A., Conos, S. A., Capovilla, G., Pamudurti, N. R., Bartok, O., Jens, M., Ashwal-Fluss, R., Stottmeister, C., Ruhe,
Cildir, G., and Conn, S. J. 2017. A circRNA from SEPALLATA3 regulates L., Hanan, M., Wyler, E., Perez-Hernandez, D., Ramberger, E., Shenzis, S.,
splicing of its cognate mRNA through R-loop formation. Nat. Plants 3: Samson, M., Dittmar, G., Landthaler, M., Chekulaeva, M., Rajewsky, N., and
17053. Kadener, S. 2017. Translation of circRNAs. Mol. Cell 66:9-21.E7.
Dong, R., Ma, X. K., Chen, L. L., and Yang, L. 2017. Increased complex- Pan, T., Sun, X., Liu, Y., Li, H., Deng, G., Lin, H., and Wang, S. 2018. Heat
ity of circRNA expression during species evolution. RNA Biol. 14: stress alters genome-wide profiles of circular RNAs in Arabidopsis [pub-
1064-1074. lished correction appears in Plant Mol. Biol. 96:231]. Plant Mol. Biol. 96:
Dragomir, M., and Calin, G. A. 2018. Circular RNAs in cancer: Lessons 217-229.
learned from microRNAs [published correction appears in Front. Oncol. Salzman, J., Gawad, C., Wang, P. L., Lacayo, N., and Brown, P. O. 2012.
8:307]. Front. Oncol. 8:179. Circular RNAs are the predominant transcript isoform from hundreds of
Du, W. W., Yang, W., Chen, Y., Wu, Z. K., Foster, F. S., Yang, Z., Li, X., human genes in diverse cell types. PLoS One 7:e30733.
and Yang, B. B. 2017. Foxo3 circular RNA promotes cardiac senescence Sanger, H. L., Klotz, G., Riesner, D., Gross, H. J., and Kleinschmidt, A. K.
by modulating multiple factors associated with stress and senescence 1976. Viroids are single-stranded covalently closed circular RNA mole-
responses. Eur. Heart J. 38:1402-1412. cules existing as highly base-paired rod-like structures. Proc. Natl. Acad.
Ebbesen, K. K., Hansen, T. B., and Kjems, J. 2017. Insights into circular Sci. 73:3852-3856.
RNA biology. RNA Biol. 14:1035-1045. Senfter, D., Holzner, S., Kalipciyan, M., Staribacher, A., Walzl, A., Huttary,
Fischer, J. W., and Leung, A. K. 2017. CircRNAs: A regulator of cellular N., Krieger, S., Brenner, S., Jager, W., Krupitza, G., Dolznig, H., and
stress. Crit. Rev. Biochem. Mol. Biol. 52:220-233. Mader, R. M. 2015. Loss of miR-200 family in 5-fluorouracil resistant
Gao, Z., Chung, E.-H., Eitas, T. K., and Dangl, J. L. 2011. Plant intracellular colon cancer drives lymphendothelial invasiveness in vitro. Hum. Mol.
innate immune receptor resistance to Pseudomonas syringae pv. maculicola Genet. 24:3689-3698.
1 (RPM1) is activated at, and functions on, the plasma membrane [pub- Sun, X., Wang, L., Ding, J., Wang, Y., Wang, J., Zhang, X., Che, Y., Liu,
lished correction appears in Proc. Natl. Acad. Sci. USA 108:8915]. Proc. Z., Zhang, X., Ye, J., Wang, J., Sablok, G., Deng, Z., and Zhao, H. 2016.
Natl. Acad. Sci. USA 108:7619-7624. Integrative analysis of Arabidopsis thaliana transcriptomics reveals intuitive
Ghorbani, A., Izadpanah, K., Peters, J. R., Dietzgen, R. G., and Mitter, N. splicing mechanism for circular RNA. FEBS Lett. 590:3510-3516.
2018. Detection and profiling of circular RNAs in uninfected and maize Suzuki, H., and Tsukahara, T. 2014. A view of pre-mRNA splicing from
Iranian mosaic virus–infected maize. Plant Sci. 274:402-409. RNase R resistant RNAs. Int. J. Mol. Sci. 15:9331-9342.
Grabowski, P. J., Zaug, A. J., and Cech, T. R. 1981. The intervening Suzuki, N., Rivero, R. M., Shulaev, V., Blumwald, E., and Mittler, R. 2014.
sequence of the ribosomal RNA precursor is converted to a circular RNA Abiotic and biotic stress combinations. New Phytol. 203:32-43.
in isolated nuclei of tetrahymena. Cell 23:467-476. Wang, J., Yang, Y., Jin, L., Ling, X., Liu, T., Chen, T., Ji, Y., Yu, W., and
Greene, J., Baird, A. M., Brady, L., Lim, M., Gray, S. G., McDermott, R., Zhang, B. 2018. Re-analysis of long non-coding RNAs and prediction of
and Finn, S. P. 2017. Circular RNAs: biogenesis, function and role in circRNAs reveal their novel roles in susceptible tomato following TYLCV
human diseases. Front. Mol. Biosci. 4:38. infection. BMC Plant Biol. 18:104.

618 PHYTOPATHOLOGY®
Wang, P. L., Bao, Y., Yee, M. C., Barrett, S. P., Hogan, G. J., Olsen, M. N., Zhang, L., Pu, H., Duan, Z., Li, Y., Liu, B., Zhang, Q., Li, W., Rochaix,
and Salzman, J. 2014. Circular RNA is expressed across the eukaryotic J.-D., Liu, L., and Peng, L. 2018a. Nucleus-encoded protein BFA1
Tree of Life PLoS One 9:e90859. promotes efficient assembly of the chloroplast ATP synthase coupling factor
Xia, S., Feng, J., Lei, L., Hu, J., Xia, L., Wang, J., Xiang, Y., Liu, L., Zhong, S., 1. Plant Cell 30:1770-1788.
Han, L., and He, C. 2017. Comprehensive characterization of tissue-specific Zhang, X. O., Wang, H. B., Zhang, Y., Lu, X., Chen, L. L., and Yang, L.
circular RNAs in the human and mouse genomes. Brief. Bioinform. 18:984-992. 2014. Complementary sequence-mediated exon circularization. Cell 159:
Yamada, K., Kanai, M., Osakabe, Y., Ohiraki, H., Shinozaki, K., and Yamaguchi- 134-147.
Shinozaki, K. 2011. Monosaccharide absorption activity of Arabidopsis Zhang, Y., Zhang, X. O., Chen, T., Xiang, J. F., Yin, Q. F., Xing, Y. H.,
roots depends on expression profiles of transporter genes under high salinity Zhu, S., Yang, L., and Chen, L. L. 2013. Circular intronic long noncoding
conditions. J. Biol. Chem. 286:43577-43586. RNAs. Mol. Cell 51:792-806.
Yamada, K., Saijo, Y., Nakagami, H., and Takano, Y. 2016. Regulation of Zhang, Z., Yang, T., and Xiao, J. 2018b. Circular RNAs: Promising bio-
sugar transporter activity for antibacterial defense in Arabidopsis. Science markers for human diseases. EBioMedicine 34:267-274.
354:1427-1430. Zuo, J., Wang, Q., Zhu, B., Luo, Y., and Gao, L. 2016. Deciphering the roles
Ye, C. Y., Chen, L., Liu, C., Zhu, Q. H., and Fan, L. 2015. Widespread non- of circRNAs on chilling injury in tomato. Biochem. Biophys. Res.
coding circular RNAs in plants. New Phytol. 208:88-95. Commun. 479:132-138.

Vol. 112, No. 3, 2022 619