Rapid differentiation of citrus Hop stunt viroid variants by real-time RT-PCR and high resolution melting analysis

Molecular and Cellular Probes 27 (2013) 221e229

Contents lists available at ScienceDirect

Molecular and Cellular Probes journal homepage: www.elsevier.com/locate/ymcpr

Rapid differentiation of citrus Hop stunt viroid variants by real-time RT-PCR and high resolution melting analysis Giuliana Loconsole a, Nuket Önelge b, Raymond K. Yokomi c, Raied Abou Kubaa d, Vito Savino a, Maria Saponari e, * a

Dipartimento di Scienze del Suolo, della Pianta e degli Alimenti, Università di Bari “Aldo Moro”, Via Amendola 165/A, 70126 Bari, Italy Çukurova University Agriculture Faculty, Plant Protection Department, 01330 Balcali Adana, Turkey USDA, Agricultural Research Service, USDA-ARS San Joaquin Valley Agricultural Sciences Center, 9611 South Riverbend Avenue, Parlier, CA 93648-9757, USA d Department of Plant Protection, Ministry of Agriculture and Agrarian Reform, Damascus, Syria e Istituto di Virologia Vegetale, Consiglio Nazionale delle Ricerche, UOS Bari, Via Amendola 165/A, 70126 Bari, Italy b c

a r t i c l e i n f o

a b s t r a c t

Article history: Received 23 May 2013 Received in revised form 19 July 2013 Accepted 19 July 2013 Available online 8 August 2013

The RNA genome of pathogenic and non-pathogenic variants of citrus Hop stunt viroid (HSVd) differ by five to six nucleotides located within the variable (V) domain referred to as the “cachexia expression motif”. Sensitive hosts such as mandarin and its hybrids are seriously affected by cachexia disease. Current methods to differentiate HSVd variants rely on lengthy greenhouse biological indexing on Parson’s Special mandarin and/or direct nucleotide sequence analysis of amplicons from RT-PCR of HSVdinfected plants. Two independent high throughput assays to segregate HSVd variants by real-time RTPCR and High-Resolution Melting Temperature (HRM) analysis were developed: one based on EVAGreen dye; the other based on TaqMan probes. Primers for both assays targeted three differentiating nucleotides in the V domain which separated HSVd variants into three clusters by distinct melting temperatures with a confidence level higher than 98%. The accuracy of the HRM assays were validated by nucleotide sequencing of representative samples within each HRM cluster and by testing 45 HSVd-infected field trees from California, Italy, Spain, Syria and Turkey. To our knowledge, this is the first report of a rapid and sensitive approach to detect and differentiate HSVd variants associated with different biological behaviors. Although, HSVd is found in several crops including citrus, cachexia variants are restricted to some citrus-growing areas, particularly the Mediterranean Region. Rapid diagnosis for cachexia and noncachexia variants is, thus, important for the management of HSVd in citrus and reduces the need for bioindexing and sequencing analysis. Ó 2013 Elsevier Ltd. All rights reserved.

Keywords: Citrus Viroid Cachexia HRM Real time

1. Introduction Recent advances in viroid detection have given rise to reports of new viroids or variants, and up to now, at least seven species and different variants are known to infect Citrus spp. However, only two viroid-induced diseases, namely cachexia and exocortis, have been shown to seriously affect crop yield [1e3]; while many viroid variants on sensitive rootstocks induce different degrees of stunting and can be of horticultural value as dwarfing agents to decrease tree size and increase planting tree densities with a concomitant increase in yield per hectare [4,5]. Citrus cachexia affects mandarin (Citrus reticulata Blanco), clementine (Citrus clementine Hort. ex Tan.), satsuma (Citrus unshiu * Corresponding author. Tel.: þ39 (0)805443068; fax: þ39 (0)805443608. E-mail address: [email protected] (M. Saponari). 0890-8508/$ e see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.mcp.2013.07.003

(Macf.) Marc.) and several mandarin hybrids when grafted on trifoliate and trifoliate-hybrid rootstocks; while other commercial cultivars or citrus species remain symptomless. The most common symptoms in sensitive hosts are bark gumming, undulating stem pitting or bumps and projections in the bark corresponding to depressions in the wood [2]; due to these alterations, trees generally decline which can progress to death of the infected tree. The causal agent of cachexia disease was reported to be a specific variant of Hop stunt viroid (HSVd) [6] in the late 1980s. Currently, at least two biologically distinct variants of HSVd are known in citrus: “pathogenic variants” that induce the cachexia disease in sensitive hosts and “non-pathogenic variants” that infect the same hosts without inducing symptoms [7]. Indeed, the non-pathogenic HSVd variants have been used beneficially to reduce vegetative growth (dwarfing agent) while increasing yield and fruit size in sweet orange cultivars [4,5,8].

222

G. Loconsole et al. / Molecular and Cellular Probes 27 (2013) 221e229

The two HSVd groups in this study associated with cachexia and non-cachexia are similar in size: non-cachexia is 302 nucleotides (nt); cachexia ranges between 295 and 299 nt and both share a sequence identity of 98% [9]. Variants of HSVd that express cachexia symptoms were shown to have a “cachexia expression motif” consisting of five to six conserved nt changes in the “variable domain” (V) region located to the right of the conserved domain (C) [10e14]. High sequence homology and similar size makes it difficult for sequential polyacrylamide gel electrophoresis (sPAGE) and conventional reverse transcription-PCR (RT-PCR) to distinguish these two groups apart. Although several molecular protocols have been developed for the generic detection of HSVd in citrus [15e18], only few reports describe strain-specific oligoprobes and/or primers for variant differentiation by molecular hybridization or RT-PCR [14,19e21]. High-resolution melting (HRM) analysis, coupled with real-time (q) PCR, provides an alternative rapid approach compared to direct DNA sequencing for the detection of singleenucleotide polymorphisms (SNPs) [22]. HRM analysis provides simultaneous detection of multiple mutations in one assay, reducing time and costs of tests [23]. The qPCR genotyping approach is based on the unique pattern of amplicon melting curves that develop at different temperatures (Tm) which allow discrimination of variants associated with nucleotide change(s). The assay consists of a qPCR reaction containing either a DNA-binding fluorescent dye (generally SYBR or EvaGreen) or dually-labeled TaqMan probes. The qPCR products are overheated and the melting behavior is subsequently monitored by plotting changes in fluorescence during denaturation of the double-strand DNA. Recent applications of qPCR and HRM analysis in plant pathology have been limited to EvaGreen-based assays for differentiating genotypes of bacteria and virus strains [24,25]; while that for complementary fluorescent TaqMan probes have only been reported to differentiate mutants of human pathogens [26e28]. In this report, a procedure was developed to differentiate HSVd variants by subjecting amplicons generated from two independent RT-qPCR assays to high resolution temperature and analyzing its characteristic melt curve profile. This allowed simultaneous detection and identification of pathogenic and nonpathogenic HSVd variants by either of two reporters. The first procedure utilized EvaGreen DNA-binding dye and included a universal primer pair designed to target the HSVd genomic region encompassing the “cachexia expression motif” and included three discriminating nt. The second assay targeted the same genomic region but used a cyanine 5- (CY5)-labeled TaqMan probe to detect the three nt changes associated with sequences of the cachexia and non-cachexia variants. Both assays proved more sensitive than conventional RT-PCR and less time consuming than biological indexing or sequence analysis. Thus, the new assays can serve as important tools to assist in the management and control of cachexia.

2. Materials and methods 2.1. Hop stunt viroid variants Three genetically distinct citrus HSVd variants were selected based on V domain sequences (Fig. 1) for the RT-qPCR and HRM analyses. HSVd-a contains the non-cachexia motif; HSVd-b has the cachexia motif; and HSVd-h has a V domain intermediate between the cachexia and non-cachexia sequences, hereafter, referred to as the “hybrid variant”. Full-length genomes of the three variants were amplified from field infected sources, cloned and sequenced. Recombinant clones harboring the variants were used to establish parameters for the qPCR assays and HRM analyses. To validate procedures, samples from field trees from California, Italy, Spain, Syria and Turkey known to be infected by HSVd based on molecular hybridization tests, were tested (Table 1) along with appropriate controls. California and Italian sources were from symptomless sweet orange and clementine grafted on C35 or sour orange rootstocks. Samples from Turkey were Satsuma mandarin trees either symptomless or showing brown-gum stain on the trunk typical of severe cachexia (Fig. 2AeB) grafted on sour orange. Syrian samples were from Clementine trees either symptomless or showing pronounced bark scaling (Fig. 2C) grafted on sour orange. No biological information was available on samples from Spain. 2.2. Assay A e EvaGreen-based real time RT-PCR and HRM analysis The primer pair HSVd-hrm-f (50 -CTCTTCTCAGAATCCAGCG-30 ) and HSVd-hrm-r (50 GAAGCCTCTACTCCAGAGCA-30 ) based on the conserved region contiguous to the cachexia expression motif, were used for the RT-qPCR and HRM analysis. Depending on the HSVd variant, the amplified target region consisted of 62e64 nt encompassing the three discriminating nucleotides (Fig. 1). Real time PCR templates consisted of (i) plasmid DNA purified from an overnight culture of recombinant clones of HSVd-a, HSVd-b, and HSVd-h; or (ii) cDNA synthetized from 0.5 mg of total nucleic acid (TNA) extracted from natural HSVd-infected citrus plants. TNA were prepared from leaf petioles harvested from mature leaves using the NucleoMag 96 RNA Kit (MachereyeNagel Inc., Bethlehem, PA, USA) on the Freedom EVOÒ150 extraction robot (Tecan, Männedorf, CH) (17). Complementary DNA was synthetized by 1 h incubation at 42  C, of 0.5 mg of pre-denatured TNA, in 20 ml of 1 MMLV buffer (Invitrogen Life Technologies, Paisley, UK) with 0.5 mg random primers, 1.25 ml of 2.5 mM dNTPs, 1 ml of RNasin (40 U/ml), 1.2 ml of DTT 0.1 M and 0.4 ml of M-MLV-RT (200 U/ml). Amplification was carried out on a CFX96 thermocycler (Bio-Rad Laboratories, Hercules, CA, USA) in 10 ml of 1 Precision Melt Supermix (Bio-Rad Laboratories) containing 0.3 mM of each primer. Cycling conditions were one cycle at 98  C for 2 min, followed by 40 cycles at 98  C for 5 s and 58  C for 15 s. Melting curves of PCR amplicons were obtained with temperatures ranging from 65  C to 95  C. Data

Fig. 1. Multiple sequence alignment of the region targeted by the real-time RT-PCR assays. The three differential nucleotides are marked in boxes. Cachexia variants: X701 (AF213486) and HSVd-b (this study); non-cachexia variants: CC-A2 (FJ716171) and HSVd-a (this study); and intermediate variants: CC-A4 (FJ716173) and HSVd-h (this study). Primers are indicated by black arrows. Gray double-ended arrow corresponds to the region targeted by the TaqMan probes.

G. Loconsole et al. / Molecular and Cellular Probes 27 (2013) 221e229

223

Table 1 Results of the melting curve and HRM analyses on a panel of field Hop stunt viroid-infected trees of different origin. ID citrus plants

Origin

Assay “A”

Assay “B”

Mean Tm values

HRM cluster

Confidence (%)

Mean Tm values and DTma

HRM cluster

Confidence (%)

67.0 (0.0) 66.5 (0.5) 66.0 (1.0) 67.0 (0.0) 67.0 (0.0) 66.5 (0.5) 66.5 (0.5) 66.5 (0.5) 66.5 (0.5) 66.5 (0.5) 67.0 (0.0) 66.5 (0.5) 66.0(1.0) 66.0 (1.0) 66.5 (0.5) 66.5 (0.5) 66.5 (0.5) 66.5 (0.5) 66.0 (1.0) 66.0 (1.0) 66.0 (1.0) 66.0 (1.0) 66.5 (0.5) 66.5 (0.5) 66.5 (0.5) 66.5 (0.5) 66.5 (0.5) 67.0 (0.0) 67.0 (0.0) 67.5 (þ0.5) 67.0 (0.0) 67.0 (0.0) 66.5 (0.5) 62.0 (5.0) 62.0 (5.0) 62.0 (5.0) 62.0 (5.0) 61.0 (6.0) 61.5 (5.5) 61.5 (5.5) 61.5 (5.5) 63.0 (4.0) 62.5 (4.5) 66.5e53.0 (0.5;14.0) 66.5e52.5 (0.5; 14.5)

A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A A B B B B B B B B B B D

98.9 98.6 99.5 98.8 98.8 98.6 99.3 99 99 99 98.6 98.9 98.6 98.7 98.7 98.8 98.4 99.7 98.8 99.3 99.1 98.7 98.2 99.3 99.3 99.3 98.2 98.3 98.8 98.8 99.1 99.3 99.8 99.2 98.0 98.5 98.1 98.5 98.5 98.0 98.5 98.5 98.6 100

D

100

IT-A IT-B IT-C IT-D IT-E TK-3 TK-9 TK-10 TK-11 TK-12 IT-P1 IT-P2 IT-P3 IT-P4 IT-P5 IT-P6 IT-P7 IT-P8 IT-P9 IT-P10 IT-P11 IT-P12 IT-Lau Ca-890 Ca-866 Ca-TB-1 SYR-FSC3 SP-L1 SP-L2 SP-L4 SP-L7 SP-L9 SP-L8 SP-L10 TK-1 TK-2 TK-6 Syr-A1 Syr-A3 Syr-A4 SP-L3 SP-L5 SP-L6 Tk4

Italy Italy Italy Italy Italy Turkey Turkey Turkey Turkey Turkey Italy Italy Italy Italy Italy Italy Italy Italy Italy Italy Italy Italy Italy California California California Syria Spain Spain Spain Spain Spain Spain Spain Turkey Turkey Turkey Syria Syria Syria Spain Spain Spain Turkey

81.6 81.6 81.8 81.6 81.6 81.7 81.6 81.6 81.7 81.7 81.7 81.7 81.7 81.7 81.7 81.7 81.7 81.7 81.7 81.6 81.6 81.7 81.7 81.7 81.7 81.7 81.7 81.5 81.6 81.6 81.5 81.5 81.5 82.2 82.1 82.2 82.1 82.2 82.1 82.1 82.2 82.2 82.2 82.6

1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 3

99.7 99.6 99.7 98.5 99.2 98.9 99.3 99.6 97.9 99.8 99.5 99.5 99.5 99.8 99.8 99.7 99.6 99.6 99.6 99.3 99.1 99.7 99.7 99.7 99.9 99.9 99.5 99.7 97.5 99.2 98.6 99.7 97.6 99.2 99.4 98.3 98.5 99.3 99.5 99.2 98.6 99.5 99.3 99.1

TK-5

Turkey

82.6

3

99.1

a

In parenthesis is reported the temperature deviation (DtM) from the reference non-cachexia variant HSVd-a.

acquisition was performed for every 0.2  C temperature with 10 s intervals. For each variant, three replicates of each recombinant clone were tested in 2 independent assays. Similarly two batches of TNA were extracted from 45 HSVd infected field samples and tested in 2 independent assays. HRM analysis software (Bio-Rad Laboratories) automatically clustered the samples according to its melting profiles and assigned a confidence score to each sample. As mentioned earlier, each experiment included extracts from healthy citron Etrog (Citrus medica L.), the non-template controls and the DNA of the three reference variants. To investigate discrimination of HSVd variants in mixed infections, artificial DNA mixtures were generated by mixing equivalent amounts of DNA from recombinant clones harboring each one of the HSVd variants. The detection limit of the new assay was determined by comparison with conventional RT-PCR assays using six serial 10-fold dilutions prepared by mixing TNA from HSVd source IT-P1, with the TNA from healthy citron. Each dilution was used to generate a random primed cDNA. For conventional RT-PCR, the cDNA was

amplified with primers HSVd-For78 and HSVd-Rev83 following the standard amplification protocol reported by Sano et al. [29]. 2.3. Assay B e TaqMan-based asymmentrical real time RT-PCR and HRM analysis The TaqMan-based assay utilized CY5-labeled probes HSVd-prC (50 CCC TCT CTC CAC GCC CCC G 30 ) and HSVd-prNC (50 CCC TCT CTC CTA CGC CTC TCG3’) in asymmetrical qPCR reactions using primers HSV-a [18] and HSVd-hrm-r (this work). HSVd-prC was fully complementary to the cachexia motif sequence; whereas HSVd-prNC was complementary to the to the non-cachexia motif. Real-time RT-PCR templates for Assay B were the same as for Assay A (EvaGreen-based assay). The reaction mix was prepared in a final volume of 10 ml. To avoid probe hydrolysis during the elongation steps, GoTaq DNA polymerase (Promega, Madison, WI, USA), which lacks 50 to 30 exonuclease activity was used. Asymmetric amplification was performed with a final concentration of the forward HSV-a primer twentyfold (1 mM) higher than the reverse

224

G. Loconsole et al. / Molecular and Cellular Probes 27 (2013) 221e229

Fig. 2. Symptoms recovered on the trunks of field trees of Satsuma mandarin (isolate Tk5) from Turkey (A and B) and of Clementine (C) from Syria (isolate A3).

primer HSVd-hrm-r (0.05 mM). Cycling conditions were one cycle at 98  C for 1 min, followed by 40 cycles at 98  C for 5 s, 55  C for 20 s and 72  C for 20 s. The melting program was 30 s at 95  C, followed by continuous detection of fluorescence from 45  C to 90  C, with a temperature increase rate of 0.5  C/step at 10 s intervals. The resultant melting curves were then analyzed to determine the deviation of the melt peak Tm with respect to the non-cachexia reference variant. Additionally, the HRM software was used to determine the HRM clusters. Experiments included positive and negative controls, artificial mixtures and the reference target DNA as specified earlier. 2.4. HRM-Based clusters and phylogenetic analysis The accuracy of the differentiation of HSVd variants by the RTqPCR HRM was confirmed by cloning and sequencing amplicons obtained from 8 selected field samples from representatives of each HRM-based cluster. For each sample, at least 10 HSVd-recombinant clones were screened by qPCR assays using both Assay A and Assay B. One representative recombinant clone from each of the three distinct HRM clusters was selected and sequenced. Multiple alignments of the resultant nucleotide sequences, encompassing the C, V and the terminal right domains (ca. 120bp) was used to generate an UPGMA phylogenetic tree using MEGA (Version 5) to show the resultant phylogenetic groups related to the HRM clusters from both assays. The analysis included sequences of the three reference variants found in this study as well as sequences of isolates X-701 (AF213486) and Ca905 (AF131250), known to be variants of cachexia and isolate CC-A2 (FJ716171), known to be a non-cachexia variant. Sequences for the three latter isolates were used as reference points for comparison. 3. Results 3.1. Assay A Assay A successfully detected HSVd in all recombinant clones containing the reference variants (HSVd-a, -b and -h) (Table 2). Moreover, the assay was found to be approximately 100-fold more sensitive than conventional RT-PCR (Supplementary Fig. S1). Standard melt curve analysis confirmed that the new primers amplified the specific target region which originated from each variant which had a single melt peak and predicted Tm. HRM analysis unequivocally identified three HRM clusters (1, 2, 3) (Fig. 3A). The Tm associated to each variant was 81.6  C for HSVd-a, 82.2  C for HSVd-h and 82.6  C for HSVd-b. The three variants were

assigned to three distinct clusters with a confidence level higher than 99% (Table 2). The Tm deviation for the variants associated with HSVd-b and HSVd-h differed by 0.6  C and 1.0  C, respectively, from the Tm of non-cachexia variant (HSVd-a). Melt curves from artificial mixtures of HSVd variants showed a single peak with the Tm of the mixed amplicons shifted toward the Tm of the cachexia variant, which had the highest Tm of the variants tested, rather than double or triple melt peaks. Because amplicons from the non-cachexia variants had the lowest Tm, its presence in the mixture was not observed by melt curve analysis in any of the combinations tested. Although this was a disadvantage, cachexia variants were still clearly identified in the mixture of HSVd variants. Thus, the method successfully identified the “cachexia motif” whenever present. 3.2. Assay B The two-probe system using HSVd-prNC and HSVd-prC specifically detected the single-strand target amplicon generated from the asymmetric RT-qPCR assay. The probe HSVd-prNC detected the three variants tested and generated similar quantification cycle (Cq) values and end-point fluorescence intensities. HSVd-prC, on the other hand, reacted with the homologous HSVd-b variant but not with HSVd-a or HSVd-h variants and was abandoned from further testing. The melt curve analysis of the hybridized HSVdprNC probe clearly differentiated the cachexia (HSVd-b) from the non-cachexia variants (HSVd-a) by its Tm deviations (Table 2). The fully hybridized dsDNA of “HSVdprNC:HSVd-a” gave the highest Tm; while the dsDNA from the hybridization of the “HSVdprNC:HSVd-b” had a 15  C lower Tm and that of the dsDNA “HSVdprNC:HSVd-h” was even 5.5  C lower. When the melt curves were examined with the HRM software, three separate clusters (A, B, C) (Fig. 3B) were generated. Thus, melt curves from Assay B were in agreement with that of Assay A and indicated both assays successfully differentiated the three reference HSVd variants. Melt curves from the artificial mixtures of variants showed presence of multiple peaks (Fig. 4). For all combinations the highest peak of the derivative plot of melting curves corresponded to the non-cachexia component. 3.3. Application of EvaGreen and TaqMan-based melting temperature analysis to natural infected citrus field trees Assays A and B both detected HSVd in all 45 HSVd-infected field samples with Cq between 18 and 28 (data not shown) regardless if field source tree was symptomless or symptomatic (brown-gum

G. Loconsole et al. / Molecular and Cellular Probes 27 (2013) 221e229

225

Table 2 Schematic representation of the sequence and secondary structure of the “cachexia expression motif” and corresponding melting temperature (Tm) obtained in real time RTPCR using EvaGreen (assay A) or the CY5-labeled TaqMan probe (assay B). Intra and interassay standard deviations are reported in parenthesis. Hop stunt viroid

Assay “A”

Varianta

Tm (SD)c ( C)

Assay “B”

HSVd-ab

81.60 (0.05)

0.0

67.00 (1.0)

0.0

HSVd-h

82.20 (0.10)

þ0.6

61.50 (0.5)

5.5

HSVd-b

82.60 (0.08)

þ1.0

52.00 (0.5)

15.0

Tm deviationd ( C)

Tm (SD) ( C)

Tm deviation ( C)

a

HSVd-a contains the non-cachexia motif, HSVd-b the cachexia motif and HSVd-h an intermediate sequence between the two variants. Nucleotide positions represented in red indicate the location of the three deletions characteristic of the cachexia motif; whereas in HSVd-h only deletion at nt position 115 occurs. c Average of the standard deviations within and between the runs. d Temperature deviation from the reference non-cachexia variant HSVd-a. b

Fig. 3. Melting curve analysis in duplicate wells using 3 variant-specific plasmid DNA in real-time PCR HRM analysis performed using the EvaGreen (A) or the CY5-labeled TaqMan probe (B). Graphs on the left indicate the negative derivative of fluorescence over temperature (dRFU/dT) plots displaying different peaks. Graphs on the right indicate the difference plots; HRM generated clusters are indicated with a number or letter. Different color are used to indicate distinct profiles, red curves correspond to HSVd-a variant, blue curves to HSVd-b and green curves to HSVd-h variants. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

226

G. Loconsole et al. / Molecular and Cellular Probes 27 (2013) 221e229

Fig. 4. Derivative melting curves (dRFU/dT) generated by the CY5-labeled TaqMan probe on the artificial mixes of the three HSVd variants. Reaction mix contained HSVd-a, -b and e h (A); HSVd-a and -b (B); HSVd-a and -h (C).

stain). HRM analysis in assay A separated field samples into the same three relative clusters (1, 2, 3) as the three reference HSVd variants (Table 1, Fig. 5). Isolates from Italy and California clustered with the HSVd-a reference variant associated as non-cachexia variants (cluster 1). The Spanish and Turkish isolates were separated into the same three HRM groups as the reference variants and indicated that the variants contained the non-cachexia motif (cluster 1), the cachexia motif (cluster 2) and the “intermediate” motif (cluster 3). The Syrian isolates clustered with either the noncachexia (HSVd-a) or with the hybrid (HSVd-h) variants. For assay B, isolates were categorized based on the Tm deviations (DTm) from the non-cachexia variant. Results confirmed field isolates were separated into three groups (Table 1); the first group included the HSVd-a reference variant and isolates with DTm equal to 0.0  C; the second group contained the reference variant HSVd-h and isolates with DTm of 5.5  C; the third group contained two isolates (Tk4 and Tk5) showing a double melt peak, with one peak with a DTm of 0.0  C and the second peak with a DTm of 15.0  C. This second DTm corresponded to that of the reference variant HSVd-b (cachexia). Similar variant separation was achieved

when the melting data were analyzed with the HRM software: field isolates formed three clusters A, B and D, respectively. The isolates Tk4 and Tk5 (HRM clustrer D) had melt peaks with DTm typical of the cachexia variant. HRM software placed these isolates in a distinct HRM cluster due to the presence of the double melt profile rather than in cluster C with the reference cachexia variant (HSVd-b). The results from assays A and B were in agreement for all samples. For the isolates Tk4 and Tk5, both assays confirmed presence of the cachexia variant, but assay B also showed the mixture contained the non-cachexia variant. These results were reproducible by repeating two independent tests for each assay with resultant small standard deviations (data shown). 3.4. Correlation between HRM data and phylogenetic analysis A minimum of two samples selected within each HRM cluster were tested by conventional RT-PCR and products cloned and sequenced. The clones were initially screened by HRM qPCR

Fig. 5. Discrimination of Hop stunt viroid (HSVd) variants in field samples by real-time RT-PCR and HRM analysis based on EvaGreen. Normalized (A) and difference plots (B) obtained on a panel of HSVd-infected samples from California, Italy, Spain, Syria and Turkey. Different colors are used to indicate distinct profiles, corresponding to the HRM-based clusters indicated with a different number. Red curves indicate HSVd-a variant, blue curves the HSVd-b and green curves the HSVd-h variants. Green and red columns represent preand post-melting normalization regions. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

G. Loconsole et al. / Molecular and Cellular Probes 27 (2013) 221e229

analysis. Recombinant clones within each of the following isolates 26D, Tk1, FSC3, A3, Tk2, Tk3, L3 and L8 generated a single HRM cluster associated with the reference variant used in this study. Whereas, for the isolate Tk4 and Tk5, 9 of 10 recombinant clones clustered with the reference HSVd-b variant and 1 clone (Tk4-2 and Tk5-10) clustered with the HSVd-a variant. The nucleotide sequences obtained from the selected clones 26D-1, FSC3-21, Tk1-1, Tk2-20, Tk3-8, Tk4-2, Tk4-3, Tk5-10, Tk5-11, Tk5-12, A3-23, L3-8 and L8-18 were then used to generate a multiple alignment to determine phylogenetic relationships with the reference HSVd variants. The UPGMA tree showed that the sequences from isolates in HRM-cluster 1 and A were in the same phylogenetic clade along with the HSVd-a reference variant and with the non-cachexia reference isolate CC-A2 (Fig. 6). Isolates categorized in HRM cluster 2 and B contained sequences homologous to variant HSVd-h. Isolates Tk4 and Tk5 were from the HRM-cluster “3” and “D” and had sequences in the same clade with the reference variant HSVd-b and with the cachexia isolates X-701 and Ca905. In addition, they also contained sequences genetically related to reference isolate CC-A2 which confirmed results from Assay B. Thus, in comparison with the sequence data, all variants were detected and differentiated with 100% accuracy for sensitivity and specificity. 3.5. Intra and interassay reproducibility The average Tm for each variant showed very small standard deviations within and between runs (Table 2) indicating consistency and reproducibility. The three melt peaks were separated by Tm differences greater than the standard deviations of each variant (Table 2, Fig. 3). With field validation tests, the average standard deviation amongst the samples in 2 runs were in the range of 0.06e 0.12  C for assay “A” and 0.5e1.0  C for assay “B” (data not shown).

227

HRM analysis is sensitive enough to detect Tm changes as little as 0.2  C. Thus, differentiation of the HSVd variants in the unknown samples was achieved by RT-qPCR and HRM analysis using both assays. The advantage of Assay B was that Tm deviations (15  C) between the cachexia and non-cachexia variants, were much higher than the standard deviations within and between runs (1.0  C); thus, the presence of these two variants was clearly seen in the HSVd-infected samples even without the HRM software. 4. Discussion HSVd is recognized as a viroid with a broad host range resulting in a large collection of variants. The biology of the citrus HSVd variants either as a disease-causing agent or as transmissible molecules which improves commercial citrus productivity, illustrates the importance for discrimination of these closely-related variants. Pathogenic and non-pathogenic variants differ by a cachexia expression motif of five to six nt located in the variable domain of the genome. Based on the sequence of this motif, HSVd variants can be differentiated in three groups: a) cachexia; b) noncachexia and c) intermediate variants [11,13,14]. Molecular markers, such as oligoprobes and primers were developed since the late 1990s to target this motif and differentiate cachexiainducing variants. However, annealing temperature, MgCl2 concentration and stringency conditions were found to be critical for the high specificity of the hybridization and RT-PCR assays. Because of the high sequence identity between the pathogenic and non-pathogenic variants, conventional strain-specific assays still remains as a challenge to replace biological indexing or sequence analysis for the discrimination of the cachexia-inducing variants. HRM has emerged as fundamental technique for the rapid and sensitive detection of a single nucleotide mutation providing informative results to genotype and categorize closely related

Fig. 6. UPGMA dendrogram constructed using the nucleotide sequences encompassing the cachexia motif of Hop stunt viroid. On the right of the dendrogram, the variant and the HRM-based clusters are reported. Bootstrap values (1000 replicates) are indicated on the corresponding nodes. Genetic similarity among variants was estimated by p-distance method, computed on MEGA 5.

228

G. Loconsole et al. / Molecular and Cellular Probes 27 (2013) 221e229

variants. In this report, we document successful development of two independent RT-qPCR approaches using melting curves of amplicons for the simultaneous detection and differentiation of HSVd variants. The key element was selecting primers and probes based on the nucleotide sequence of the motif involved in the pathogenicity [12]. Both assays differentiated cachexia and noncachexia variants as well as an intermediate variant in infected field samples. Both assays relied on a single universal primer set, using either EvaGreen DNA binding dye or a single TaqMan probe. These approaches have several advantages in comparison to a standard strain-specific TaqMan assay. Only one primer set was necessary to differentiate the three HSVd variants; and, similarly, only one TaqMan probe was used in the assay “B” to differentiate the three variants. Thus, both assays can be used for detection in single fluorescent channel or run on single channel instruments. HRM analysis can detect small Tm shifts as low as 0.2  C but require advanced real time PCR instruments with a high temperature resolution and uniformity. The second approach developed in this report achieved separation of HSVd variants using TaqMan probes with melting temperature differences greater than 5  C. The advantage of this assay was that differences in melting temperature were large enough to be detected by most real-time PCR instruments without need of a more sophisticated thermocycler developed for HRM applications. The other advantage of Assay B was its capacity to detect multiple infections of HSVd variants; whereas Assay A was only able to detect the pathogenic variant when mixed infections contained the pathogenic and nonpathogenic variants. Since our data showed high discriminatory capacity of both assays, the choice of assay would depend on what thermocyler was available or, in the case of co-infection of viroid mixtures (e.g. cachexia and non-cachexia variants) and the detection of the non-pathogenic variant is needed. Data from both assays were found to be highly reproducible and as sensitive as conventional RT-PCR. Moreover, the assays were rapid and can be completed in less than 2 h. HSVd discrimination was achieved for 45 samples with 100% agreement based on sequence analysis of representative samples along with its symptomology. The presence of the pathogenic variant was successfully detected in both assays, either in single or mixed infections as shown from the two source trees of Satsuma mandarin from Turkey severely affected by cachexia disease. Our procedures will be useful for epidemiological studies spread of severe cachexia variant spread as has reported in Turkey and China [14,30]. Very little is known about the pathogenicity of the “intermediate” HSVd variant. Although a few reports indicates that it causes symptomless infections on indicator plants [28,30], our results showed that this variant was present in the samples from Turkey collected from field trees showing symptoms resembling cachexia. We should mention that these trees were also co-infected by other viroids including Citrus exocortis viroid, Citrus viroid (CVd) I and IV (data not shown). The same variant was detected in the isolate A3 from Syria which came from a clementine tree with strong bark scaling on the trunk. In this case, the source tree was found to be coinfected with Citrus tristeza virus, CVdI, CVdIV and CEVd but not psorosis (data not shown). Further investigations on the biological properties of these HSVd-h variants are now underway, particularly because bark scaling symptoms of the Syrian A3 were similar to that associated with concave gum for which the causal agent is still unknown. In conclusion, the assays developed in this study can facilitate high-throughput screening for economically important HSVd variants. Since HRM analysis is rapid, economical and convenient, it could be used for routine identification of the HSVd cachexia variant in nursery propagative materials in a high-throughput manner.

Acknowledgments Some of the equipment and partial support for this work was provided by Regione Puglia e Rete di Laboratori Pubblici SELGE, Specialty Crops Block Grant SCB 10005 from the California Department of Food and Agriculture and from the Italian Ministry of Economy and Finance to the National Research Council for the project “Innovazione e Sviluppo del MezzogiornoeConoscenze Integrate per Sostenibilità ed Innovazione del Made in Italy AgroalimentareeLegge n.191/2009”. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture. USDA is an equal opportunity provider and employer. Appendix A. Supplementary data Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.mcp.2013.07.003. References [1] Duran-Vila N, Semancik JS, Broadbent P. Viroid diseases, cachexia and exocortis. In: Timmer LW, Garnsey SM, Graham JH, editors. Compendium of citrus diseases. 2nd ed. St. Paul, MN: APS; 2000. p. 51e4. [2] Roistacher CN. Graft-transmissible diseases of citrus. In: Handbook for detection and diagnosis. Rome: IOCV and FAO; 1991. [3] Vernière C, Perrier X, Dubois C, Dubois A, Botella L, Chabrier C, et al. Citrus viroids: symptom expression and effect on vegetative growth and yield of clementine trees grafted on trifoliate orange. Plant Dis 2004;88:1189e97. [4] Broadbent P, Forsyth JB, Bevington KB, Hutton RJ. Citrus tree size control with dwarfing agents. Calif Citriculture 1986;71:8e10. [5] Broadbent P, Gillings M, Gollnow BI. Graft-transmissible dwarfing in Australian citrus. In: Proc conf int org citrus virol 1988; 10th, 219e225. [6] Semancik JS, Roistacher CN, Rivera-Bustamante R, Duran-Vila N. Citrus cachexia viroid, a new viroid of citrus: relationship to viroids of the exocortis disease complex. J Gen Virol 1988;69:3059e68. [7] Reanwarakorn K, Semancik JS. Correlation of hop stunt viroid variants to cachexia and xyloporosis diseases of citrus. Phytopathol 1999;89:568e74. [8] Semancik JS, Rakowski AC, Bash JA, Gumpf DJ. Application of selected viroids for dwarfing and enhancement of production of Valencia orange. J Horticult Science 1997;72:563e70. [9] Levy L, Hadidi A. Direct nucleotide sequencing of PCR amplied DNAs of the closely related citrus viroids IIa and IIb (cachexia). In: Proc conf int org citrus virol 1993, 12th. p. 180e6. [10] Palacio-Bielsa A, Romero-Durban J, Duran-Vila N. Characterization of citrus HSVd isolates. Arch Virol 2004;149:537e52. [11] Reanwarakorn K, Semancik JS. Regulation of pathogenicity in hop stunt viroid related group II citrus viroids. J Gen Virol 1998;79:3163e71. [12] Serra P, Gago S, Duran-Vila N. A single nucleotide change in Hop stunt viroid modulates citrus cachexia symptoms. Virus Res 2008;138:130e4. [13] Velázquez K, Soto M, Peréz R, Peréz JM, Rodriguez D, Duran-Vila N. Biological and molecular characterization of two isolates of citrus viroids recovered from Cuban plantations. In: Proc conf int org citrus virol 2002, 15th. p. 258e63. [14] Wang XF, Zhou Y, Li ZG, Tang KZ, Liu YQ, Cao MJ, et al. Molecular, biological and phylogenetic analysis of Chinese isolates of Hop stunt viroid associated with citrus cachexia disease. J Phytopathol 2010;158:372e7. [15] Bernad L, Duran-Vila N. A novel approach for detection and characterization of citrus viroids. Mol Cell Probes 2006;20:105e13. [16] Ito T, Ieki H, Ozaki K. Simultaneous detection of six citrus viroids and apple stem grooving virus from citrus plants by multiplex reverse transcription polymerase chain reaction. J Virol Methods 2002;106:235e9. [17] Saponari M, Loconsole G, Liao H, Jiang B, Savino V, Yokomi RK. Validation of high-throughput real time polymerase chain reaction assays for simultaneous detection of invasive citrus pathogens. J Virol Methods 2013 [accepted for publication]. [18] Tessitori M, Rizza S, Reina A, Catara V. Real-time RT-PCR based on SYBR-Green I for the detection of citrus exocortis and citrus cachexia diseases. In: Proc conf int org citrus virol 2005, 16th. p. 456e9. [19] Ito T, Furuta T, Ito T, Isaka M, Ide Y, Kaneyosh J. Identification of cachexia inducible Hop stunt viroid variants in citrus orchards in Japan using biological indexing and improved reverse transcription polymerase chain reaction. J Gen Plant Pathol 2006;72:378e82. [20] La Rosa R, Tessitori M, Albanese G, Catara A, Davino M. Diagnosis of citrus exocortis and hop-stunt-homologous citrus viroids by synthetic probes. In: Proc conf int org citrus virol 1993, 12th. p. 435e7. [21] Reanwarakorn K, Semancik JS. Discrimination of cachexia disease agents among citrus variants of hop stunt viroid. Ann Appl Biol 1999;135:481e7.

G. Loconsole et al. / Molecular and Cellular Probes 27 (2013) 221e229 [22] Liew M, Pryor R, Palais R, Meadows C, Erali M, Lyon E, et al. Genotyping of single-nucleotide polymorphisms by high-resolution melting of small amplicons. Clin Chem 2004;50:1156e64. [23] Montgomery J, Wittwer CT, Kent JO, Zhou L. Scanning the cystic fibrosis transmembrane conductance regulator gene using high-resolution DNA melting analysis. Clin Chem 2007;53:1891e8. [24] Bester R, Jooste AEC, Maree HJ, Burger JT. Real-time RT-PCR high-resolution melting curve analysis and multiplex RT-PCR to detect and differentiate grapevine leafroll-associated virus 3 variant groups I, II, III and VI. Virol J 2012;9:219. [25] Gori A, Cerboneschi M, Tegli S. High-resolution melting analysis as a powerful tool to discriminate and genotype Pseudomonas savastanoi pathovar and strains. PLo SONE 2012;7:e30199. http://dx.doi.org/10.1371/journal.pone.0030199. [26] Liu Q, Luo T, Li J, Mei J, Gao Q. Triplex real-time PCR melting curve analysis for detecting Mycobacterium tuberculosis mutations associated with resistance to

[27]

[28]

[29]

[30]

229

second-line drugs in a single reaction. J Antimicrob Chemother 2013;68: 1097e103. http://dx.doi.org/10.1093/jac/dks509. Luo T, Jiang L, Sun W, Fu G, Mei J, Gao Q. Multiplex real-time PCR melting curve assay to detect drug-resistant mutations of Mycobacterium tuberculosis. J Clin Microbiol 2011;49:3132e8. Soliman HR, Othman AA. Evaluation of DNA melting curve analysis real-time PCR for detection and differentiation of Cryptosporidium species. Parasitologists United J 2009;2:47e54. Sano T, Mimura R, Ohshima K. Phylogenetic analysis of hop and grapevine isolates of hop stunt viroid supports a grapevine origin for hop stunt disease. Virus Genes 2001;2:53e9. Onelge N, Cinar A, Szchowski JA, Vidalakis G, Semancik JS. Citrus viroid II variants associated with “Gummy Bark” disease. Eur J Plant Pathol 2004;110: 1047e52.