Assessment of the Effect of Thermotherapy on ‘Candidatus Liberibacter asiaticus’ Viability in Woody Tissue of Citrus via Graft-Based Assays and RNA Assays

‘Candidatus Liberibacter asiaticus’ (CLas) is a Gram-negative, phloem-limited α-proteobacterium associated with Huanglongbing (HLB), which is one of the most devastating citrus diseases worldwide (Jagoueix et al. 1994). Since the discovery of CLas in Florida, there has been a substantial decline in citrus production (USDA 2020), resulting in substantial economic losses and downsizing of the citrus industry (Singerman and Rogers 2020). Therefore, there is an immediate need to introduce management practices that can reduce damage caused by HLB in field trees. One of the potential treatment strategies is the use of heat (thermotherapy) on field trees because it has been used to manage HLB for more than 50 years (da Graça 1991). Treating citrus budwood with vapor-saturated hot air at a temperature of 48 to 58°C for 35 min to 1 h reportedly eliminated yellow shoot (renamed HLB) disease without viability loss (Lin and Lo 1965; Lin and Zheng 1964). Similarly, treating symptomatic citrus plants with moist, hot air at a temperature of 49 to 50°C for 50 to 60 min restored tree health (Weiwen et al. 1981). More recently, a study by Hoffman et al. (2013) showed that heat treatment reduced CLas titer under controlled environmental conditions in citrus seedlings. The whole seedlings continuously exposed to 40 to 42°C for a minimum of 48 h up to 7 to 10 days under fluorescence lamps showed a significant reduction in CLas titer and remained disease-free for more than 2 years in the absence of subsequent infection. Heavy defoliation occurred after thermotherapy; however, new flush was observed 2 months after treatment. Hoffman et al. (2013) also observed that there were no significant differences among the different temperatures and times. Another study compared the results of heat treatment sessions in temperatures of 45 and 48°C for 4 h repeated once per week for 3 consecutive weeks to determine their ability to eliminate disease symptoms and CLas from greenhouse citrus plants (Fan et al. 2016). Most of the plants showed a decrease of >90% in CLas titer in new flush after 8 weeks of treatment and a decrease of 56 to 60% in titer in mature leaves. No tests after 8 weeks have been reported; therefore, the long-term dynamics of the CLas population cannot be concluded based on these results. However, speculation of the recovery of CLas has been reported. The aforementioned thermotherapy studies were performed in a controlled greenhouse environment, which showed promising results regarding the efficacy of heat as a potential HLB management technique. However, follow-up of these experiments involving lower temperatures and longer exposure times under field conditions is impractical because of the size, age, and number of HLB-affected trees, as well as the equipment required for treatment in the field. Therefore, in this study, we experimented with higher temperatures and shorter exposure times. Furthermore, we used steam as a source of rapid heat so that thermotherapy was a viable solution.

Research of HLB is hampered by many difficulties that impede the development of control strategies. The most relevant impediment is the difficulty culturing HLB-associated bacterium, CLas, in artificial media. Very limited success has been achieved when culturing CLas in vitro (Davis et al. 2008; Ha et al. 2019; Parker et al. 2014). All studies were able to co-cultivate CLas in the presence of other microorganisms, but an axenic culture of CLas was not established. Without the ability to isolate pure cultures or use a reliable culturing technique, in vitro evaluation of chemicals and other management practices against CLas is difficult.

To evaluate any treatment effect on a pathogen in vitro, the growth of the pathogen is tested against the treatment in pure culture. Bacterial viability, in particular, is commonly determined by a direct colony count, determining cellular integrity, studying the metabolic activities, and nucleic acid-based analyses (Keer and Birch 2003). However, CLas has not been cultured yet; therefore, difficulties arise when using conventional methods to assess its viability. Grafting treated scions can be performed to evaluate the effect of a treatment on the CLas population in scions, and the presence of CLas can help to determine whether the treatment was a success. Assessing the presence of nucleic acids is a suitable option for studying the presence of fastidious pathogens such as CLas (Keer and Birch 2003), and DNA-based methods such as polymerase chain reaction (PCR) and quantitative PCR (qPCR) are used widely to detect CLas in plant samples. A DNA assay cannot differentiate between live and dead cells; however, when quantifying bacterial viability, only live cells are of interest.

Quantitative analysis of total DNA using qPCR allows detection of DNA from any samples if the targeted nucleic acid sequences are intact; therefore, there is a high probability of overestimating the viable bacterial titer because of undegraded DNA persisting in a sample. DNA from dead cells of Listeria monocytogenes and Escherichia coli persisted and were PCR-detectable even after exposure to temperatures up to 100°C (Masters et al. 1994). It has been reported previously that genomic DNA (gDNA) from nonviable E. coli remained stable in water samples and were PCR-detectable, even after 16 weeks (Josephson et al. 1993). In a recent study by Etxeberria et al. (2019), HLB eradication caused by heat treatment was reported; however, residual CLas DNA remained qPCR-detectable for 5 months after heat treatment involving 40°C for 48 h, and the treated trees were HLB-negative for more than 2 years. Additionally, leaf discs grafted onto healthy plants did not transmit HLB.

In addition to total DNA quantification, several studies have attempted to quantify viable bacterial cells using ethidium monoazide (EMA) as well as propidium monoazide (PMA). EMA and PMA are DNA-intercalating dyes that diffuse through the compromised cell membrane of bacterial cells (Nocker et al. 2006; Nogva et al. 2003). These dyes consist of a photo-inducible azide group that allows EMA and PMA to be covalently cross-linked to DNA of dead cells on exposure to a bright light source with a maximum absorbance at 460 nm. This cross-linking renders the DNA insoluble and is selectively removed during DNA purification, thus preventing PCR amplification (Nocker et al. 2006). However, the peak absorption of chlorophyll a ranges between 450 and 475 (blue light), and that of chlorophyll b ranges between 650 and 675 nm (orange-red) (Ballschmiter and Katz 1968), likely interfering with EMA and PMA activity when using leaf samples. Trivedi et al. (2009) developed an EMA qPCR method to quantify live CLas cells in citrus and periwinkle. They reported approximately a 72% lower viable CLas population in EMA-treated citrus bark samples compared to the population detected in nontreated samples. However, using EMA in the qPCR was reported to be a poor measure of cell viability (Flekna et al. 2007) because viable cells were underestimated in heat-treated cultures of L. monocytogenes without the complication of plant tissue. Hu et al. (2013) used PMA qPCR to quantify live CLas cells in citrus tissues because PMA has greater selectivity than EMA (Nocker et al. 2006) when entering ruptured/compromised cell membranes, but not the intact live cells; they found no interference of the plant tissues with the PMA qPCR and reported a correlation between the microscopically quantified CLas population and the results of PMA qPCR of psyllids. In another study by Vazquez (2015), heat treatment of citrus leaf samples, followed by the PMA qPCR, revealed an inconsistent effect on the quantification of CLas. Additionally, Barbau-Piednoir et al. (2014) reported decreased qPCR signal when living Salmonella enterica cells were at a low concentration, resulting in an underestimation of viable cells; however, increasing the PMA concentration did not reduce the qPCR signal of dead cells. These findings mirrored work performed by a collaborating laboratory that was unable to reliably quantify CLas or Xanthomonas citri subsp. citri with PMA or EMA in citrus samples (Supplementary Fig. S1). Therefore, we resorted to developing an RNA-based assay to quantify viable CLas from the bark tissue of the treated and untreated samples.

RNA is a highly labile molecule with a shorter half-life than DNA; therefore, it can be a marker of viability of bacterial cells after stress treatments (Deere et al. 1996; Keer and Birch 2003; Masters et al. 1994). Van der Vliet et al. (1994) investigated the effects of rifampin and ofloxacin on the viability Mycobacterium smegmatis. Their results showed that after a few days of exposure to those drugs, no viable bacterial cells were detected, as determined by quantifying colony-forming units (CFUs) on media, and no expression of 16S rRNA was detected. However, they detected DNA throughout the course of exposure to both drugs even after no CFUs were observed in culture. A positive relationship between RNA and the viability of bacteria was established; however, the presence and viability of DNA were not related (Van der Vliet et al. 1994).

A range of molecular targets could be evaluated based on their expression patterns to understand the effects of thermotherapy at the molecular level. mRNA and rRNA are suitable markers of the bacterial viability assay because of the short half-life of RNA (McKillip et al. 1998; Villarino et al. 2000). However, choosing a suitable gene transcript as a target is important for validating the bacterial viability (Keer and Birch 2003).

In this study, we assessed the viability of CLas to examine the effects of thermotherapy on CLas titer in woody tissue. Therefore, the primary goals of this study were to examine the optimal time–temperature combination that can achieve significant thermal kill of CLas, to reduce the level of detection with no long-term bacterial regrowth observed in woody tissue, and to assess the viability of CLas via RNA-based assays in the woody tissues.

MATERIALS AND METHODS

Sample collection and steam treatment.

Field samples used in this study were from select trees in a ‘Valencia’ orchard close to the Citrus Research and Education Center (CREC). The trees were under standard management practices.

Three- to four-month-old bud sticks ≈0.6 to 0.8 cm in diameter that were still green, yet hardened, were collected from HLB-affected ‘Valencia’ sweet orange trees in the field. Leaves and apical tips were removed. The remaining branches were cleaned with soapy water and blotted dry. For the grafting experiment, 13 branches per treatment were used; the whole experiment was conducted three times. For the RNA-based assay, three branches per treatment were used for the sampling of bark tissues; the experiment was repeated twice. The source trees of bud sticks sampled for these experiments were confirmed as CLas-positive trees via qPCR before the assays. A steam chamber consisting of steam release valves, temperature sensors (K-type thermocouples), and a data logger (12-channel temperature recorder; Omega Engineering, Stamford, CT) connected to thermocouples was used for the steam treatment of bud sticks (Fig. 1). There were 12 sensors in the steam chamber that were held in position by three vertical bars. Each vertical bar had four sensors arranged from top to bottom such that the sensors at each level would be at the same height as the sensors on the other two bars. The data logger recorded temperatures from the sensors every second. The temperatures for every treatment were graphed to visualize whether the temperatures in the steam chamber were maintained over the required duration during the treatments. The steam chamber had a horizontal metal rack for bud stick placement during the treatment. The branches were positioned horizontally in the steam chamber in such a way to avoid a temperature gradient over the length of the bud stick. Because the metal rack was leveled at the height of sensors 3 and 7, the samples were arranged in such a way that they would be close to them, and only the temperatures logged from those sensors were used to determine the temperature consistency maintained during the duration of treatment. The following time–temperature combinations were used for treatments: 50°C for 60 s; 55°C for 0 s, 30 s, 60 s, 90 s, and 120 s; 60°C for 30 s; and an untreated control (UTC). The temperature within the chamber was maintained by manually controlling the valve (required temperature ±3°C). The treatment involving 55°C for 0 s was used to test whether the ramp time needed to achieve 55°C affected the health of the bud stick or the CLas titer. After steam treatment, the branches were stored for 1 to 5 days in airtight plastic bags with moist paper towel at 4°C, followed by either grafting or RNA extraction.

Download as PowerPoint

Grafting and CLas detection.

Grafting was performed 24 h after the steam treatment because of time constraints. Healthy HLB-free rootstocks [US-812; Sunki mandarin (C. sunki) × Benecke Trifoliate (Poncirus trifoliata L.)] of ≈1.5 cm diameter were used for the grafting. The modified side grafting technique was used to graft the bud sticks to the rootstocks (Fig. 2) (Kumar 2011). First, a long diagonal cut of ≈2 cm was made on smoother side of the rootstock at a height ≈15 to 20 cm from the topsoil. Then, the basal end of the scion was cut into an ≈2-cm tapered wedge, with one side of the wedge slightly shorter so it would fit into the cut on the rootstock. Care was taken to ensure that the cuts were smooth. Then, the scion was inserted in the rootstock and secured with Parafilm strips. The grafted plants were kept in the greenhouse until further observations and sample collection. While in the greenhouse, the plants were regularly watered approximately three times per week (four times during hot and dry weather). They were fertilized with slow-release fertilizer (Harrell’s) 18-5-10 NPK (nitrogen/phosphorus/potassium) and liquid fertilizer (Peters Professional water-soluble fertilizer; 20–10–20 NPK diluted at 1:100 in water) every 2 weeks in summer and every 4 weeks in winter. Pest infestation was managed using an alternation of Triple Action Neem oil, M-Pede, and other pesticides (Carbaryl, Agmectin, Conserve, and Delegate) as necessary. After 3 to 4 weeks, the plants were evaluated for graft success. Leaf samples were collected from the scions that survived at 6, 9, and 12 months for CLas detection.

Download as PowerPoint

One to three leaves (depending on the availability) were randomly collected from the grafted bud sticks. Midribs from the leaves of each plant were excised, pooled, and cut into very small pieces. Approximately 100 mg of the finely chopped midribs was collected in 2-ml screw cap tubes with 5-mm stainless steel beads (Qiagen, Valencia, CA). Samples were stored at −20°C before processing. The samples processed for total DNA extraction using the DNeasy Plant Mini Kit (Qiagen) according to the manufacturer’s instructions. Total DNA samples were stored at −20°C until assayed by qPCR to detect CLas in the samples.

All the qPCR runs for CLas detection were performed in a 96-well plate using a CFX96 Real-Time PCR Detection System (Bio-Rad). Primers and a probe targeting the β-operon region of CLas, CQULA04F-CQULAP10-CQULA04R, were used (Wang et al. 2006). The primer pair and probe were designed from the target sequence of CLas ribosomal protein genes. The qPCR conditions were programmed as 95°C for 10 min, followed by 49 cycles of 95°C for 15 s and 60°C for 45 s. The final reaction volume was 25 μl, consisting of 1× PCR buffer, 0.2 mM of dNTP, 0.8 μM of each primer, 0.2 μM of probe, 1 unit of HotStart Taq Plus polymerase (Qiagen, Valencia, CA), and 1 μl of total genomic DNA (gDNA). Water was used as the nontemplate control. The standard curve was generated from a series of dilutions of plasmid pLBA2 (Trivedi et al. 2009). Each individual sample was replicated twice for verification. Any detection level was considered positive for the presence of CLas; however, it is worth mentioning that the samples with cycle threshold (C_t) values more than 40 are calculated to have fewer than one copy number of CLas. Factors that could contribute to this result at very low copy numbers are stochastic effects in the PCR reaction or PCR inhibitors slowing amplification, leading to detection but not reliable quantification.

Furthermore, fibrous roots from the grafted plants with a negligible amount of CLas DNA in the midrib were sampled after 12 months of grafting to confirm CLas detection. Root DNA was extracted using the Qiagen Power Soil kit following the methods of Johnson et al. (2014).

RNA extraction and cDNA synthesis.

Approximately 1.5 cm of every bud stick was cut from the stick, and 50 to 60 mg of the bark sample was weighed and collected in 2-ml tubes for RNA extraction before and after steam treatment and then flash-frozen in liquid nitrogen before being stored at −80°C. The RNeasy Plant Mini Kit (Qiagen) was used for RNA extraction; the manufacturer’s protocols were followed with slight modifications during the elution step. To elute the RNA, the spin column was placed in a new 1.5-ml tube and 50 μl of RNase-free water was added to the membrane. The tube was incubated for 2 min at room temperature and centrifuged for 1 min at 10,000 rpm. Then, the eluate was again added to the membrane in the spin column and centrifuged to obtain a higher RNA concentration. Following RNA extraction, the total concentration and purity of RNA were determined from the ratio of absorbance reading at 260/280 nm and 260/230 nm using the ND-1000 Spectrophotometer (NanoDrop Technologies, Wilmington, DE). The RNA integrity was tested by following the bleach gel protocol described by Aranda et al. (2012). Then, RNA was treated with DNase (Turbo DNA free kit, Invitrogen) to remove any gDNA contamination in the final RNA eluate according to the instructions of the manufacturer. Complementary DNA (cDNA) was synthesized from 1 μg of total DNase-treated RNA in 20 μl of the final volume with random hexamers as well as gene-specific primers using SuperScriptIV reverse transcription (Invitrogen, Life Technologies) according to the manufacturer’s protocols. Reverse-transcription qPCR (RT-qPCR) was performed.

Selection of reference genes.

FBOX, SAND, GAPDH, GAPC2, COX, WRKY70, EF-1α, ACT2, and TUB genes (Table 1) in citrus were analyzed for their constitutive expression before and after heat treatment. Because all of these citrus genes were shown to be reliable internal controls for gene expression studies under different experimental conditions, they could be used as an endogenous reference for RNA expression analysis (Mafra et al. 2012). The geNorm algorithm was used to determine the most suitable reference gene for the treatments and was used for the normalization and validation of the expression study. To test the primer efficiency, a template of known concentration was serially diluted and performed in duplicate or triplicate with the primer pairs. A melt curve analysis was also performed following the standard curve analysis to examine the specificity of the primer pairs. If the standard curve had reaction efficiency close to 100%, and if there was a single peak in the melt curve, then the primers were used for further experiments.

TABLE 1. Primers used in this study for reference citrus gene selection and RNA viability assay

Primer ID	Sequence (5′-3′)	Target	Size (bp)	Reference
FBOX-F	TTGGAAACTCTTTCGCCACT	F-BOX family protein	112	Mafra et al. 2012
FBOX-R	CAGCAACAAAATACCCGTCT
SAND-F	AGGTTGTCAGCCTTGTTGGT	SAND family protein	145	Mafra et al. 2012
SAND-R	GCCGTGGGATTATATCTTGG
GAPDH-F	GCGCTGACAATGAGCTGAAA	Glyceraldehyde-3-phosphate dehydrogenase	129	Yan et al. 2012
GAPDH-R	ACAGTACCCTCTGCTGTACC
GAPC2-F	TCTTGCCTGCTTTGAATGGA	Glyceraldehyde-3-phosphate dehydrogenase C2	80	Carvalho et al. 2010
GAPC2-R	TGTGAGGTCAACCACTGCGACAT
COX-Fa	GTATGCCACGTCGCATTCCAGA	Cytochrome oxidase gene	68	Li et al. 2006
COX-Ra	GCCAAAACTGCTAAGGGCATTA
WRKY70-F	AAGGAGCTTGTGAAGGGTAAG	WRKY70 transcription factor	146	Carvalho et al. 2010
WRKY70-R	GTACCGAGCACAGAAAGAGTC
ACT2-F	CCCTTCCTCATGCCATTCTTC	Actin-2	105	Mafra et al. 2012
ACT2-R	CGGCTGTGGTGGTAAACATGT
EF-1α-F	TCAGGCAAGGAGCTTGAGAAG	Elongation factor 1-alpha	80	Carvalho et al. 2010
EF-1α-R	GGCTTGGTGGGAATCATCTTAA
TUB-F	TTCCTCTCCCTTCTGTGCTCAA	Beta tubulin	120	Yan et al. 2012
TUB-R	TTACAAAAGCCAGTCGGTAGAC

^a Reference gene selected to be used for the RNA viability assays.

View as image HTML

Selection and primer design for target bacterial genes.

In our study, we selected several mRNAs as well as ribosomal RNA targets to determine the viability of CLas post-steam treatment. CLas 16S rRNA, β-operon (rpoB), nrdB, Omp2, SDE1, groEL, and dnaK were selected to analyze the effect of steam treatment on CLas in woody tissues (Table 2). Primers for CLas 16S rRNA, rpoB, nrdB, Omp2, and SDE1 were obtained from already published primer sequences (Table 2). For other selected genes, their sequences were obtained from the GenBank database. Gene-specific primer pairs were designed based on their sequence using primer design software from the National Center for Biotechnology Information (NCBI). Primer efficiency was tested as described before using the primers to determine relative expression of the target genes.

TABLE 2. Primers for ‘Candidatus Liberibacter asiaticus’ (CLas) gene targets used in this study for the RNA viability assay

Primer ID	Sequence (5′-3′)	Target	Size (bp)	Reference
RPO-F	CGTAGAACAGGTTGATGCGA	rpoB	143	Merfa et al. 2018
RPO-R	CCAATGGACGCTGGTTGACAC
SDE1-F	AGGAAATATCGTGCGTAAAA	CLIBASIA_05315	143	Pagliaccia et al. 2017
SDE1-R	GCTCCAACATTTTTCTATGG
16S-F	TCGAGCGCGTATGCGAATACG	16S rRNA	76	Li et al. 2006
16S-R	GCGTTATCCCGTAGAAAAAGGTAG
Omp2-F	ACTCGCCTCTTTCTGTTCTGCTC	omp	474	Nageswara-Rao et al. 2013
Omp2-R	AATCCTCCGACTTCAACACCGACA
groEL-F	TCAGCGCTTCAACCGTTACT	groEL	147	This study
groEL-R	CACTGCAAGAACTGGCAACC
dnaK-F	ATGGGAGCGGCTATACAAGC	dnaK	156	This study
dnaK-R	TCGGCGGCTGTTGAGAATAC
RNR-F	CATGCTCCATGAAGCTACCC	nrdB	80	Zheng et al. 2016
RNR-R	GGAGCATTTAACCCCACGAA

View as image HTML

Reverse-transcription qPCR.

The RT-qPCR was performed on a BioRad Thermocycler in a 96-well reaction plate. SYBR Green PCR master mix (Qiagen) was used for all reactions. The final reaction volume was 20 μl, consisting of 10 μl of the SYBR Green Master Mix, 1.2 μl of forward and reverse primer each (final concentration of 0.3 μM each), 5.6 μl of nuclease-free water, and 2 μl of template cDNA. The RT-qPCR parameters were programmed at 95°C for 10 min, followed by 49 cycles at 95°C for 15 s and the adjusted annealing temperature for each primer pair for 45 s. The melting curve analysis of the primer pairs was also performed to ensure the specificity of the primers. The reaction plate also included nontemplate control (NTC) containing water instead of template cDNA and non-reverse-transcription control (NRT) for all the samples. There were two technical replicates and three biological replicates for each gene.

Gene expression analysis.

To calculate the relative fold change in the target gene expression caused by treatment, the 2-ΔΔC_t method was used (C_t, is the baseline at which the fluorescence signal crosses the threshold line or background level).

For pretreatment and posttreatment samples, ΔC_t was calculated as follows:

ΔΔC_t was calculated as follows for the corresponding target genes:

Finally, fold change was calculated as the expression of the target genes relative to the internal control in treated samples when compared with the untreated control (Livak and Schmittgen 2001),

Statistical analysis.

To test for significance in the expression of target genes under different treatment combinations, PROC GLIMMIX in SAS 9.4 was used. LSMEANS was used to compute the least square means of the fixed effects of treatments and time points for multiple comparisons of the treatment effects on the target genes. Samples were treated as random effects in the analysis.

RESULTS

Scion survival.

The temperatures for all treatments were consistent throughout the treatment–time duration (required temperature ±3°C). As the severity of the thermotherapy treatment increased, the graft survival rate decreased. The survival of the scions (bud-sticks) was indicated by the presence of flush on them (Table 3 and Figure 3B). Some of the scions did not flush until 9 months after grafting despite being green; therefore, they were considered alive, but they were not sampled during the 6-month period. Visual observation of some scions indicated they were dead (Fig. 3); however, after 6 or 9 months, flush was observed. Late flushing in the treated scions masked graft survival at 6 and 9 months, resulting in an apparent increased survival rate during the 12-month observation after grafting. The total number of scions that survived across the treatments at 6, 9, and 12 months after grafting is shown in Table 3. Six months after grafting, only 26 of the UTC scions out of 39 had flushed and had green bark. For the most extreme treatment, 60°C for 30 s, only one scion survived and a leaf sample was collected for the CLas detection; however, by the 9-month sampling period, the bud stick was dead. For the less severe treatments at 50°C for 60 s and 55°C for 0 s, the scion survival rate was comparable to that of the UTC; however, as the treatment duration and temperature increased, the incidence of dead scions also increased. Some of the scions did not flush until 9 months after grafting in treatments with reduced scion survival, primarily because of bud damage during the treatment.

TABLE 3. The number of treated scions that survived after grafting and detected positive for ‘Candidatus Liberibacter asiaticus’ (CLas) DNA by quantitative polymerase chain reaction at 6, 9, and 12 months after treatment for the three repetitions combined

Treatments	Total grafts	6 months		9 months		12 months
		CLas-positive/surviving grafts	Average log CLas titer ± SEa	CLas-positive/surviving grafts	Average log CLas titer ± SE	CLas-positive/surviving grafts	Average log CLas titer ± SE
Untreated control	39	26/26	4.2 ± 0.2	25/29	3.4 ± 0.3	29/30	3.6 ± 0.2
50°C for 60 s	39	18/25	2.6 ± 0.3	18/24	2.7 ± 0.3	22/24	3.4 ± 0.2
55°C for 0 s	39	25/32	3.4 ± 0.3	26/32	2.9 ± 0.2	29/32	3.2 ± 0.2
55°C for 30 s	39	7/17	1.7 ± 0.4	6/17	1.1 ± 0.3	8/17	1.5 ± 0.3
55°C for 60 s	39	2/11	0.7	7/12	1.4 ± 0.4	3/11	0.9 ± 0.4
55°C for 90 s	39	1/9	0.4	2/9	0.5	0/10	0
55°C for 120 s	39	0/11	0	0/12	0	2/12	0.03
60°C for 30 s	39	0/1	0	0/0	0	0/0	0

^a When there were fewer than three positive samples, the standard error (SE) was not calculated for the CLas average log titer.

View as image HTML

Download as PowerPoint

CLas detection in grafted plants.

Bud stick source trees were PCR-positive for CLas presence. In our study, we found 96.7% of the grafted scions were CLas-positive in the untreated control 1 year after grafting (Fig. 4). After steam treatment of bud-sticks at different time and temperature combinations, CLas detection in the scions differed. Because the number of grafts that survived varied, the CLas-positive samples are presented as proportions for each treatment. For treatments with lower temperatures and times (i.e., 50°C for 60 s and 55°C for 0 s), the graft survival rate and number of CLas-positive grafts were comparable to that of the UTC; with the increase in time at 55°C, the graft survival rate decreased. Treatment with 60°C for 30 s proved fatal to the grafts (Table 3 and Fig. 4).

Download as PowerPoint

At 12 months after grafting, the final percentages of CLas-positive scions were 47.1, 27.3, 0, and 16.7% with the treatment temperature of 55°C for 30, 60, 90 and 120 s, respectively, although the graft survival rate decreased as well. For the one sample treated with 60°C for 30 s, when collected after 6 months, CLas detection was negative. Therefore, CLas detection was less frequent with the increasing severity of the treatment; in other words, exposure to higher temperatures for a longer time period and steam treatments at 55°C for 60, 90, and 120 s either reduced CLas titer significantly or eliminated it from the treated bud sticks (Table 3 and Fig. 4). The CLas titers for the treatment involving 55°C for 120 s detected in two of the grafted scions were very low (Table 3). Root DNA from those specific grafted plants with low CLas DNA titer was extracted to confirm the presence of CLas. Only one sample was detected with CLas DNA at a titer of 1.0 CLas genome per milligram of root (C_t value ≈41).

Response of target genes to steam treatment.

To investigate the viability of CLas in the phloem of the bark of young branches after exposure to treatments at different time–temperature combinations, the expressions of various reference genes from citrus and target genes in CLas were analyzed using qRT-qPCR. Several potential reference genes, including GAPC2, GAPDH, SAND, FBOX, COX, WRKY70, and EF-1α, were screened using the geNorm algorithm, and the gene with the lowest average expression stability values (M) was chosen as the reference gene (Vandesompele et al. 2002). When evaluated, the responses of GAPC2, GAPDH, and COX were the most stable reference genes immediately after treatment (Supplementary Fig. S2). The COX gene was subsequently selected as the reference gene for normalization of target gene expression, although the RNA of the COX gene also showed degradation 5 days after treatment with severe treatments; the plant tissue started to deteriorate (Fig. 5). This may have led to an underestimate of the reduction of CLas gene expression in these treatments. The bacterial genes rpoB, 16S rRNA, nrdB, SDE1, Omp2, dnaK, and groEL were evaluated for their transcriptional activity in response to the treatments. These genes were selected to represent cellular housekeeping genes (16S rRNA, rpoB, nrdB), a membrane-associated gene (Omp2), the conserved CLas Sec-delivered effector encoding gene (SDE1), and genes associated with heat-shock proteins (dnaK and groEL). Among most of the selected genes, there was not enough transcriptional activity measured to detect differential expression. There was either minimal or no detection of transcriptional activities of rpoB, nrdB, Omp2, dnaK, groEL, and SDE1 in the bark samples, which made it difficult to reliably quantify the expression patterns of these genes (Fig. 6), even in the pretreatment samples (Supplementary Fig. S3). Using gene-specific primers during the cDNA synthesis process did not have much effect on the expression activity of the respective genes but did improve detection. However, an effect of thermotherapy on the transcriptional activity of 16S rRNA was observed among treatments compared with pretreatment after calibration via the COX gene. Therefore, CLas 16S rRNA was used for further analysis of the treatment effect on CLas in the woody tissues.

Download as PowerPoint

Download as PowerPoint

After 1 day posttreatment (DPT), the transcriptional activity of 16S decreased 0.15-fold, 0.59-fold, and 0.05-fold in the samples of UTC, 50°C for 60 s, and 55°C for 0 s compared with pretreatment (Fig. 7), suggesting that the CLas cells in the UTC and the less severe treatments were viable and functioning with minimal damage. However, the expression analysis showed a decrease in CLas activity at 55°C for 30 s, 55°C for 60 s, 55°C for 90 s, 55°C for 120 s, and 60°C for 30 s by 2-fold, 2.2-fold, 2.6-fold, 2.2-fold, and 2-fold, respectively, at 24 h after treatment. During further sampling at 5 DPT), there was a 0.64-fold (Fig. 7) increase in the CLas 16S rRNA in the UTC, and there were no significant changes recorded for 50°C for 60 s and 55°C for 0 s, with a slight decrease in the 16S activities by 0.88-fold and 0.74-fold, respectively. The mean C_t values and se are also provided for reference in Supplementary Table S1. Significant differences were found for more severe treatments at 5 DPT (P ≤ 0.0001) compared with the UTC. For treatments at 55°C for 90 s, 55°C for 120 s, and 60°C for 30 s, the COX expression in the plant appeared to decline (Fig. 5). There was no significant difference in the C_t values of COX within the treatment; however, at 5 DPT, C_t values for COX were significantly higher (P ≤ 0.0001), showing damage to plant tissue. Compared with pretreatment samples, the 16S transcriptional activity of 5 DPT samples decreased by 3.4-fold, 3.4-fold, and 2.3-fold with 55°C for 90 s, 55°C for 120 s, and 60°C for 30 s, respectively. The bacterial activity reduction is probably underestimated because of the reduction in reference gene activity, with the greatest underestimate for samples treated at 60°C for 30 s, for which physical degradation of plant tissue was evident. In contrast, we observed either recovery or stability of the 16S expression with treatments of 55°C for 30 s and 55°C for 60 s; decreased expression levels of 1.8-fold and 2.3-fold were observed after 5 DPT, similar to those at 1 DPT. This may indicate that CLas cells in the bark were only heat-injured and started recovering after a while; however, with severe treatments, CLas was heat-killed and did not recover.

Download as PowerPoint

The RNA-based assay indicated that CLas 16S rRNA activity decreased after 24 h of treatment, particularly for the treatments involving 55°C for 30, 60, 90, and 120 s and 60°C for 30 s. After 5 DPT, RNA activity further decreased with the treatments involving 55°C for 90 s, 55°C for 120 s, and 60°C for 30 s; however, titers were stable or recovering in bark tissue following treatments involving 55°C for 30 s and 55°C for 60 s.

DISCUSSION

This study indicated there was a direct effect of thermotherapy on the CLas population in the bark tissue of citrus as well as physical damage to the bud sticks. Grafting experiments showed that CLas titer as well as graft survival decreased with an increase in treatment severity. Treatments at 55°C for 30 and 60 s eliminated CLas from more than half of the scions that survived (Table 3 and Fig. 4). Treatments at 55°C for 90 and 120 s caused the CLas populations to decline below the detection limit in the treated scions. However, in 2 of 12 surviving grafts that underwent treatment involving 55°C for 120 s, very low levels of CLas were detected 12 months after grafting, which might have been residual DNA of dead CLas present in the plant. A low level of CLas was detected in only one of the root samples of plants grafted with twigs treated at 55°C for 120 s. As CLas rapidly moves into the root system and multiplies (Johnson et al. 2014), we can infer from the low incidence of CLas in our root samples 1 year after grafting that the treatments of 55°C for 90 and 120 s can eliminate a transmissible CLas population. Whether the detected DNA was from nonviable cells moving through the phloem to new tissues or from damaged cells that moved and subsequently died is difficult to determine; however, in either case, no replication appeared to have occurred.

One of the disadvantages of DNA-based methods is that DNA does not distinguish between alive and dead bacterial cells (Birch et al. 2001). It was shown that CLas DNA remains detectable for long periods of time (4 to 9 months) via PCR, even after the bacteria are dead (Etxeberria et al. 2019). At the most severe treatment of 60°C for 30 s, host tissue survival was compromised, and none of the scions survived with that treatment severity. Hoffman et al. (2013) also reported serious tissue damage to the citrus plants with continuous exposure to dry heat at temperatures of 42 and 45°C for 10 and 3 days, respectively; however, citrus trees treated at 42°C for 10 days had vigorous growth 1 to 2 months after treatment. Our observations of the efficacy of thermotherapy and the effect of high temperatures on the CLas population paralleled the findings that CLas can be managed with thermotherapy under greenhouse conditions.

Thermotherapy was shown to effectively eliminate CLas in potted citrus plants for more than 2 years when treated with temperatures of 40 to 42°C for 7 to 10 days in the absence of subsequent infection (Hoffman et al. 2013). Fan et al. (2016) also reported that when graft-inoculated citrus plants were heat-treated at 45 and 48°C for 4 h and this was repeated once per week for 3 consecutive weeks, there was a decrease in the CLas titer in the citrus plants and the new flush had no HLB symptoms. The decrease in CLas titer was observed only after 8 weeks of heat treatment. These studies treated small plants with lower temperatures and longer exposure times; however, this is not feasible or representative of what occurs with field treatments. Using solar heat to treat field trees did not achieve CLas eradication, even when the temperatures reached higher than 40°C for 3 to 8.5 h daily on most days (Doud et al. 2017). High temperatures for a short duration of time may provide a practical solution for HLB thermotherapy in the field; therefore, the thermotherapy experiments reported in this study focused on evaluating steam treatments at high temperatures of 50, 55, and 60°C for multiple time periods to determine what was needed to eliminate CLas in green woody tissues.

To corroborate the efficacy of thermotherapy in eliminating CLas from the bark tissues, a further understanding of its impact on the CLas viability is important. Various target genes were evaluated during this study to assess CLas viability. CLas 16S rRNA was the only target gene tested that had sufficient expression to detect and quantify differential transcriptional activity corresponding to the treatments (Fig. 7). The transcript levels of other target genes studied were sporadically expressed with different treatments without clear changes in their expressions (Fig. 6). This is not surprising when considering the uneven distribution of the pathogen within the plant (Tatineni et al. 2008). Additionally, bark tissues from sweet orange have lower CLas populations compared with midrib (Kunta et al. 2014).

An intriguing observation of our results was the stability of RNA transcripts over time. Immediately after steam treatment, no significant difference in the gene expression pattern was found across the treatments (data not shown). Initially, it was presumed to be caused by insufficient time for the RNA from dead cells to degrade after treatment. However, even at 24 h after treatment, CLas 16S rRNA, along with other genes, was detected in the samples. In a study of E. coli and Staphylococcus aureus, the stability of 16S rRNA was monitored to determine its suitability as an indicator of bacterial viability (McKillip et al. 1998). They showed that although rRNA was a suitable target for assessing the bacterial viability under extreme heat treatment (i.e., 121°C), when heat inactivation of bacteria at 80°C was used, rRNA was amplified for up to 48 h after cell death. Sheridan et al. (1998) also showed the stability of 16S rRNA after 16 h of incubation at room temperature. Therefore, in our study, we decided to examine the expression of CLas 16S rRNA after longer intervals after treatment. Our results were consistent with the findings of McKillip et al. (1998), who reported that rRNA is stable for longer time periods; however, we could see a gradual decrease in the 16S rRNA over time for some of the severe treatments that indicate cell death and inability to recover at those treatment severities. Heat injury could explain the limited changes in the transcriptional activity across the moderate treatments. Collins-Thompson et al. (1973) demonstrated that injured S. aureus after sublethal heat treatment at 52°C for 15 min could recover and repair.

At 5 DPT, the expression of plant gene COX was not stable with 55°C for 90 s, 55°C for 120 s, and 60°C for 30 s treatments; damage to plant tissue was observed. Additionally, the RNA-based results correlated with the results of the graft experiment in which there was no CLas detection with higher temperatures and time combinations. Failed recovery and eventual death of the CLas cells resulted in no CLas detection in most of the grafts at the severe treatment levels of 55°C for 90 and 120 s. The 16S rRNA can be used as an indicator of bacterial viability; however, optimization of the experiment is necessary to achieve accurate results.

Furthermore, the aforementioned treatments can be used to eliminate CLas from the propagative materials in the nursery and produce disease-free scions. Vidalakis et al. (2010) reported thermotherapy as one of the procedures to eliminate pathogens from citrus propagative materials. However, the therapy process described in their article involves bud grafting and preconditioning the grafted seedlings by placing them in a hot greenhouse with temperatures at 28 to 40°C/25°C (daytime/nighttime) for 30 days, followed by placing the preconditioned seedlings in a temperature-controlled chamber at 40°C (16 h) and 30°C (8 h) for 3 months. This process requires a long time to achieve disease-free propagative materials. Alternatively, steam-generated thermotherapy can be a quick, viable, and practical solution to create HLB-free grafting materials for the nursery and maintain clean commercial citrus germplasm and breeding materials. Preconditioning the grafting materials can be tested to improve the graft viability when using steam for heat treatment.

A recent proteomic analysis performed to understand the molecular processes associated with thermotherapy on CLas-infected citrus plants by Nwugo et al. (2016) showed strong upregulation of bacterial heat shock proteins. Protein extraction was from grapefruit leaves that were exposed to 40°C for 6 days. Contrary to those results, when expressions of the bacterial heat shock genes, groEL and dnaK, were evaluated for our study, detection of these genes from the bark samples was inconsistent, and there were no differences in the C_t values observed across the various heat treatments (data not shown). However, the time and temperature treatments were different between studies. This study used a quick exposure of branches to high temperatures rather than prolonged heat treatment, and bark was sampled instead of leaves, as in the study by Nwugo et al. (2016). The contradictory results can raise questions about the thermal conductance of the bark compared to that of the midribs or leaves, which has not been explored much.

Bark is generally considered a heat insulator (Pásztory and Ronyecz 2013); therefore, it is possible that the treatment regimes in our study might not have increased the phloem tissue to the required temperatures, particularly the regimens with lower temperatures and time durations. This may be why these treatments had either minimal or no effect on the bacterial population in the phloem of bark. Bark thickness and moisture content also affect the heat conductance and influence the efficacy of thermotherapy for woody tissues (Wesolowski et al. 2014). These factors make the application of thermotherapy challenging in the field, where many branches and trunks have much larger diameters than the 0.6- to 0.8-cm bud sticks used in this experiment. Trees will also have varying levels of water stress in the field depending on rainfall and irrigation patterns that could affect the ability of a thermotherapy system to sufficiently increase the temperature on large-diameter scaffold branches and trunks. However, causing heat stress may be sufficient to harm CLas cells because two nearly identical prophages, SC1 and SC2, were found in CLas strains (Zhang et al. 2011). Studies have suggested that the SC1 lytic cycle can be induced, providing a foundation for HLB management (Fleites et al. 2014; Zhang et al. 2011). The replicative form of SC1 was found in citrus and periwinkle (alternative host of CLas), but not in psyllids. There is no substantial evidence for lytic cycle activation in CLas-infected citrus, and the mechanism of lytic cycle activation in artificially inoculated periwinkle is unknown; however, SC1 is highly replicative in infected periwinkle. Heat treatment can induce a lytic cycle in bacteriophage (Lieb 1964), and a lytic burst would trigger bacterial cell death. This mechanism should be further explored to understand whether this is a mechanism of CLas cell death caused by heat treatment, although it would be challenging to discern CLas viability from a prophage study.

Using the grafting experiments and viability study, we demonstrated that thermotherapy with steam as a heat source could eliminate CLas from most of the surviving bud sticks when treated at 55°C for 90 and 120 s. This combination of temperature and time may be further optimizable to achieve higher graft survival rates as well. CLas RNA shows more stability than expected in plant tissue, but it can be used as an early indicator of treatment efficacy. Better gene targets with high expression and reduced stability may exist beyond the limited set tested in this study. RNA activity could be used to study the thermotherapy efficacy for larger branches with thicker bark on field trees that cannot be grafted into a greenhouse bioassay.

ACKNOWLEDGMENTS

We thank Mayara Murata for providing invaluable assistance with the RNA assay; the laboratory of Dr. Nian Wang at the Citrus Research and Education Center, University of Florida, for providing plasmid pLBA2; Tony McIntosh, Etelvina Aguilar, Roy Sweeb, and Eric Ramjit for their excellent technical assistance in the greenhouse and during steam treatment of the samples; and Tonya Weeks for providing high-quality versions of the images used in the manuscript.