Plant material was collected from healthy and CLas-infected sweet orange (Citrus sinensis L. Osb), mandarin (C. reticulata Blanco), grapefruit (C. paradisi Macf.), Mexican lime (C. aurantifolia Swingle), and pummelo (C. maxima Burm.) from the Citrus Research and Education Center (CREC), University of Florida (Lake Alfred, FL, United States), the Citrus Center, Texas A&M University-Kingsville (Weslaco, TX, United states), and the USDA-ARS Citrus Quarantine Facility in Beltsville, MD, United States.

Mature leaves and bark tissues were collected from 1-year-old branches of greenhouse raised plants as well as field trees in citrus groves. Citrus leaves with the blotching symptoms and confirmed for CLas infection by TaqMan qPCR were considered “symptomatic” samples; whereas leaves without typical HLB symptom but with CLas being detected by qPCR were considered as “asymptomatic” samples. Additional healthy samples collected from citrus trees maintained in the quarantine greenhouse of the Citrus Clonal Protection Program (CCPP) at the University of California, Riverside were also included as negative controls.

Sec-translocon dependent extracytoplasmic protein analysis of the CLas was recently reported (Prasad et al., 2016). Here, we focused on the SDEs that are unique to CLas or Ca. Liberibacter spp. (Table 1) so that they could potentially be used as biomarkers for HLB. SDEs were predicted using signalP3.0 and signalP4.0 (Petersen et al., 2011). The SDE1 genes in eight CLas strains were obtained from NCBI and their sequences were confirmed by PCR-sequencing. The sequence conservation of SDE1 among the CLas strains was analyzed using clustalW (Thompson et al., 1994).

Expression of potential CLas SDEs in HLB-infected citrus was determined by semi-quantitative RT-PCR. Total RNA was isolated from leaf and bark tissues using TRIzol^® (Invitrogen, United States) following the manufacturer’s instruction. Total RNA (1 μg) was treated with 1 U RNAse-free DNase I (Invitrogen, United States) and used for reverse transcription. The RNA concentrations were determined by spectrophotometer NanoDrop2000c (Thermo Fisher Scientific Inc., United States) and their integrity was assessed by agarose gel electrophoresis. The first-strand complementary DNA (cDNA) was synthesized in a 15 μL reaction, containing 12.5 μL (1 μg) RNA, 1 μL oligo dT_15-18 primer (100 μM stock), and 1.5 μL dH₂O. The mixture was incubated at 70°C for 10 min in a thermal cycler (MyCycler^TM, Bio-Rad Laboratories, Inc.) and then immediately chilled on ice. The second strand was synthesized by adding 5 μL 5× First-Strand Buffer (250 mM Tris–HCl, 375 mM KCl, 15 mM MgCl₂), 1.25 μL dNTPs (10 μM), 0.5 μL M-MLV reverse transcriptase (200 U/μL), 0.625 μL RiboLock RNase Inhibitor (40 U/μL), and 2.625 μL dH₂O to the first strand reaction mix. The total mixture was incubated at 42°C for 1.5 h, then at 70°C for 15 min followed by immediate chilling on ice. PCR was carried out using gene-specific primer pairs (Table 2) for the SDE genes. The citrus COX gene was used as an internal control.

Quantitative RT-PCR was used to determine the expression profiles of SDE1 in different hosts (citrus, periwinkle, and psyllids) and at different disease stages of CLas-infected citrus tissues of Valencia sweet orange as described in Yan et al. (2013). Briefly, 15 fully expanded mature leaves (five leaves from each tree) from three symptomatic, three asymptomatic, and three healthy greenhouse raised citrus (5-year-old) were collected for RNA and DNA extraction. CLas titer was determined using the DNA extracts by TaqMan qPCR. The SDE1 expression levels were determined by qRT-PCR. Relative transcript abundances were expressed as fold changes using DNA gyrase subunit A of CLas (CLIBASIA_00325) as the internal standard. All PCR reactions were performed in triplicate and the PCR products were separated by electrophoresis to confirm the presence of the products and their sizes. Data from duplicated experiments (expression in different host) were analyzed by one-way analysis of variance, followed by the all pairwise multiple comparison procedures (Tukey’s HSD test at P = 0.05). Data from duplicated experiments (expression at different disease stages) were analyzed by Student’s t-test at P < 0.01. All data were analyzed using SigmaPlot 13.0 statistical software package (Systat Software, Inc.). This experiment was repeated twice.

Direct binding of the antibody with SDE1 was evaluated using indirect ELISA. One hundred microliters of the antigen solution at different concentrations (20, 200, and 2,000 ng/mL) were used to coat ELISA plates (Immulon^® 2 HB Flat Bottom MicroTiter^® Plates, Thermo Fisher Scientific Inc., United States) by incubation at 4°C overnight. Wells were washed three times, 3 min each, using a HydroFlex^® microplate washer (Tecan, United States), with 300 μL of phosphate-buffered saline (PBS)-T buffer (PBS buffer containing 0.1% Tween-20). Plates were then blocked with 200 μL of 1× PBS containing 3%w/v non-fat milk at 37°C for 1 h. Wells were washed again as described above and incubated with 100 μL of anti-SDE1 antibody at different concentrations (5, 20, 100, and 1,000 ng/mL) at 37°C for 1 h. Plates were washed again and incubated with the goat-anti-rabbit IgG-horseradish peroxidase-conjugated secondary antibody (80 ng/mL, 1:5,000) for 1 h at room temperature. For signal detection, 100 μL of Ultra-3,3′,5,5′-tetramethyl benzidine TMB)-ELISA substrate solution (Thermo Fisher Scientific, Inc.) was added to each well and incubated in dark at room temperature until color development (up to 15 min). One hundred microliters of 2 M H₂SO₄ was added to stop the reaction, and the absorbance at 450 nm was measured using Tecan Plate Reader M200Pro. All samples were tested in triplicates.

The binding specificity of the anti-SDE1 polyclonal antibody to SDE1 in citrus tissues was tested by western blotting. Healthy bark tissues were ground into fine powder in liquid nitrogen and then suspended in 1× PBS amended with 1× protease inhibitor cocktail (Sigma) at a ratio of 1:5 (i.e., 0.1 g tissue in 0.5 mL buffer). After being vortexed and incubated on ice for 30 min, supernatant was collected after a 20-min centrifugation at 13,000 × g at 4°C and then filtered through a 0.22 μm polyvinylidene difluoride (PVDF) syringe filter (EZFlow^®). The tissue extract was spiked with purified SDE1 proteins and western blotting was used to examine specific binding of the anti-SDE1 antibody to SDE1 in citrus extracts. For the western blots, total proteins were separated by 12% SDS–PAGE. The concentrations of the primary and secondary antibody were 1:1,000 (200 ng/mL) and 1:3,000 (80 ng/mL), respectively.

Western blotting was also performed using CLas-infected tissue. Phloem-rich tissues (bark and midribs) were excised from symptomatic and asymptomatic citrus seedlings graft-inoculated with budwoods from the same citrus branch that was previously tested positive for CLas via TaqMan qPCR. The tissue powder was suspended in 2× Laemmli buffer (Laemmli, 1970) and western blotting was performed as described above. Tissues from a healthy seedling were used as the control.

Young branches (1-year-old) were collected around the canopy of individual citrus trees. Stem tissues were cut with a steady motion to obtain a single plane cut surface using a sterile razor blade. The samples were “printed” by gently pressing the freshly cut cross-section of branches on nitrocellulose membranes (Plant Print Diagnostics S.L., Spain), leaving faint green-colored marks from the sap. The printed membranes were dried overnight at 4°C, then washed in TBS-T buffer (125 mM Tris–HCl pH = 7.4, 140 mM sodium chloride, and 3.0 mM potassium chloride, 0.05% Tween-20) at room temperature for 30 min to reduce non-specific binding. Membranes were blocked with 1× TBS-T containing 5% w/v non-fat milk at 4°C overnight and then incubated with the anti-SDE1 antibody (TBS-T containing 5% fat-free skim milk) and the antibody (200 ng/mL) in 1:1,000 dilutions, for 90 min at room temperature. The membranes were washed three times with TBS-T (5 min each time), and incubated with the secondary antibody (1:3,000 dilution, 80 ng/mL) for 1 h at room temperature. Signals were detected by SuperSignal^TM West Pico Chemiluminescent Substrate (Thermo Fisher Scientific Inc., United States) following the manufacturer’s instruction.

Stem samples from young branches were diced into small sections and ground to a fine powder using frozen stainless steel canisters in a stainless steel Kleco pulverizer (Kinetic Laboratory Equipment Company, Visalia, CA, United States). One gram of the tissue powder was suspended in 2 mL of extraction buffer containing 50 mM Tris–HCl, 150 nM NaCl, 1 mM EDTA, 1% Tween-20, 0.05% β-mercaptoethanol, and 10% glycerol. Each sample was vortexed for 5 s and incubated in ice for 30 min. Supernatants were obtained by two consecutive 25 min centrifugations at 13,800 × g. Fifty and 300 μL of the supernatants were then transferred to a clean tube and diluted 1:4 in carbonate coating buffer containing (100 mM sodium bicarbonate and 33.5 mM sodium carbonate, pH = 9.5). Diluted samples were applied in triplicates to a nitrocellulose membrane, pre-wetted in TBS-T buffer for 5–10 min, with the aid of a manifold apparatus (Schleicher & Schuell, Inc., Germany) under a vacuum. The spotted membranes were air-dried for 10 min at room temperature, and then processed as described in the DTBIA protocol except anti-SDE1 antibody concentration used was 1:400 (500 ng/mL).

Plant extracts were obtained as described above for the DBIA experiment. ELISA plates were coated with 250 μL of plant sample (1:4 dilutions in carbonate coating buffer) at 4°C overnight and then washed three times with the PBS-T buffer as described above. Each well was blocked with 200 μL PBS buffer containing 1% BSA and 0.1% Tween-20 and incubated at 37°C for 90 min. After blocking, plates were washed three times and incubated with 100 μL of anti-SDE1 antibody (200 ng/mL) for 1 h at 37°C. After three washes (5 min each time), the plate was incubated with 100 μL of secondary antibody (1: 1,500, 80 ng/mL) for 1 h at 37°C. Signals were detected as described above in the “binding affinity analysis of the polyclonal antibody” section. Data from duplicated experiments were analyzed by the Kruskal–Wallis one-way analysis of variance on rank, followed by the all pairwise multiple comparison procedures (Student–Newman–Keuls method; at P = 0.05). All data were analyzed using SigmaPlot 13.0 statistical software package (Systat Software, Inc.). This experiment was repeated twice.

All the citrus samples that were examined by the anti-SDE1 antibody were also tested by qPCR to determine the CLas bacterial titer. Total nucleic acid was extracted from rich phloem tissues (bark) using a procedure optimized from a previously reported protocol for citrus (Osman et al., 2012). The system utilized Cryo-station and Geno Grinder 2010 (SPEX SamplePrep, Metuchen, NJ, United States), the MagMAX^TM Express-96 (Life Technologies, Carlsbad, CA, United States), and the MagMAX-96 Viral RNA Isolation Kit (Life Technologies, Carlsbad, CA, United States). Briefly, 250 mg of plant tissue was placed in an Eppendorf tube and submerged in liquid N₂ for 30 s. Two 5/32″ stainless steel grinding balls were added in each tube along with 600 mL guanidine extraction buffer. The tubes were then placed into the cryo-blocks of Geno Grinder 2010 where the tissue was ground for 20–30 s at 1,750 RPM. The crude homogenate was centrifuged at 4°C for 30 min on a bench-top centrifuge at 14,000 RPM and the supernatant was then subjected to DNA extraction as described in Osman et al. (2012).

Two microliters of extracted DNA was used for quantitative TaqMan PCR (qPCR) with primers and probes described by Li et al. (2006) on a CFX96 Real-Time PCR System (Bio-Rad Laboratories, Inc.). Each reaction (20 μL) consisted of 250 nM (each) target primer (HLBas and HLBr), 150 nM target probe (HLBp), 300 nM (each) internal control primers (COXf and COXr), 150 nM internal control probe (COXp), and 2× iTaq Universal Probes Supermix (Bio-Rad Laboratories, Inc.). DNA extracted from healthy citrus tissues was used as the negative control.

Recent analysis on Sec-translocon dependent extracytoplasmic proteins of CLas (Prasad et al., 2016) revealed approximately 86 proteins that are potentially secreted by CLas through the Sec secretion system. We are interested in secreted proteins that are smaller than 20 kDa in size, potentially facilitating their movement in the infected trees along the photosynthate transport flow, and unique in CLas or Ca. Liberibacter so that they could be used as specific biomarkers for HLB detection. These genes were analyzed using semi-quantitative RT-PCR for their expression in CLas-infected tissues of citrus varieties with different HLB susceptibility levels, including Sun Chu Sha mandarin, Duncan grapefruit, Hamlin sweet orange, and Mexican lime. This analysis allowed us to narrow down to four CLas-specific SDEs (Table 1) from which we could detect expression in all four citrus varieties (Figure 1A).

We were particularly interested in CLIBASIA_05315 (hereafter SDE1), which showed the most consistent results in the semi-quantitative RT-PCR experiment. Further analysis on the expression profile of SDE1 using quantitative RT-PCR shows that it is expressed approximately 10-fold higher in infected citrus tissues than in psyllids (Figure 1B). SDE1 is also highly expressed in periwinkle (approximately 25-fold higher) compared to the expression level in the insect vector, suggesting that SDE1 proteins may accumulate to a high level in plant hosts. Furthermore, we were able to detect SDE1 transcripts from the asymptomatic citrus seedlings, which had a lower bacterial titer. Intriguingly, the relative expression of SDE1 (normalized by a housekeeping gene of CLas) was significantly higher (P = 0.01) in the asymptomatic citrus seedlings compared to that in the symptomatic trees (Figure 1C). These results indicate that SDE1 is expressed at the relatively earlier infection stages, before the development of disease symptoms. Finally, SDE1 is highly conserved among eight CLas strains whose genome sequences were available with 100% identity in nucleotide sequences (Figure 2). Taken together, these analyses support SDE1 as a promising biomarker for early HLB detection.

To develop HLB detection using SDE1 as a biomarker, we raised polyclonal antibodies in rabbits injected with purified SDE1 protein as the antigen. A DNA fragment encoding the mature SDE1 protein (i.e., excluding the predicted N-terminal signal peptide, 1–24aa) was cloned into the E. coli expression vector pET28a. The recombinant protein was purified using nickel resins and used to inject rabbits for antibody production. The polyclonal antibody was purified from rabbit serum using affinity chromatography and evaluated for binding affinity to the SDE1 antigen using indirect ELISA. The binding affinity was tested using different concentrations of the purified anti-SDE1 antibody and SDE1 (Figure 3A). From this analysis, we determined that 100 ng/mL of purified anti-SDE1 antibody was the optimal concentration to use in indirect ELISAs and there was a positive correlation between the ELISA signal (absorbance at 450 nm) and the antigen concentration (Figure 3B). These analyses confirm that the anti-SDE1 polyclonal antibody has a high binding affinity to SDE1.

We next examined whether the antibody could detect SDE1 proteins in citrus extracts with high specificity. Purified SDE1 proteins were spiked into healthy citrus phloem extracts of Rio Red grapefruit and Navel orange, which were then detected by the anti-SDE1 antibody in western blotting experiments. Our results show that the healthy tissues had a minimal non-specific binding background, whereas the spiked samples showed a specific signal at a position consistent with the predicted molecular weight of SDE1 (14.3 kDa), in a dosage-dependent manner (Figure 4A). Furthermore, comparing with the purified SDE1 protein in PBS buffer, signals from the spiked sample in citrus extracts were even stronger, suggesting that citrus extracts do not interfere with antibody binding to SDE1.

We further detected SDE1 in CLas-infected citrus tissues using the anti-SDE1 antibody by western blotting. The same tissues were also subjected to TaqMan qPCR to determine the bacterial titer. Total proteins were extracted from bark tissues of healthy and CLas-infected Mexican lime seedlings. We were able to detect positive signals from protein extracts of the CLas-infected seedlings, no matter they showed HLB symptoms or remained asymptomatic, although the signals from the asymptomatic tissues were much weaker (Figure 4B). On the contrary, positive detection of CLas DNA by TaqMan qPCR was only achieved from the symptomatic tissue, but not from the asymptomatic tissue. These results suggest that: (1) SDE1 proteins accumulate during disease progression and/or with increasing bacterial titer; (2) SDE1 proteins may be present in citrus tissues independent of CLas cells; and (3) SDE1 is likely produced at an early infection stage. Taken together, these tests support further development of serological HLB detection methods using the anti-SDE1 antibody.

We next pursued the development of antibody-based detection methods for HLB using the anti-SDE1 antibody. First, we tried the DTBIA, which has been successfully used for the detection of citrus tristeza virus (CTV). The advantage of DTBIA lies in its effectiveness in field surveys due to simple equipment requirement and sample preparation (Lin et al., 1990; Garnsey et al., 1993; Knapp et al., 1995; Cambra et al., 1997; Amari et al., 2001). For this assay, we imprinted young branches (1-year-old) of citrus trees from grapefruits and sweet oranges on nitrocellulose membranes. The printed membranes were incubated with the anti-SDE1 antibody and the signals were detected by chemiluminescent substrates. The antibody was able to successfully detect CLas-infected trees from greenhouse (Figure 5A) and the field (Figure 5B). Consistent with the western blotting data shown in Figure 4B, positive signals were observed from both symptomatic and asymptomatic branches with weaker signals generated by the asymptomatic tissues, probably due to the lower bacterial titers (Figure 5). Furthermore, the signals were present exclusively in the regions corresponding to the phloem-rich tissues (the inner bark regions), where the bacterium and, presumably, the SDE1 protein should be located.

Although the DTBIA results suggest that sensitive detection of CLas infection was possible, the requirement of using the chemiluminescent substrates was inconvenient for diagnostic labs in non-research institutions due to the uses of X-ray films, a darkroom, and a film developer. Unfortunately, the less sensitive chromogenic substrates (e.g., TMB) were unable to generate convincing positive signals (data not shown). Since the targeted biomarker SDE1 is likely in a low abundance in early infected asymptomatic tissues, we further pursued the development of detection methods that would allow the uses of a larger amount of plant tissues/extracts in the tests so that the more convenient TMB substrate is sufficient to generate positive signals. For this purpose, we employed indirect ELISA and a vacuum-based DBIA. Compared to the DTBIA method, where only the plane of the cut surface of young branches is probed, ELISA allows up to 250 μL of plant extract (collected from 50 μg of plant tissue) that could be tested per reaction in a high throughput 96-well plate setup, and vacuum-based DBIA allows the application of plant extracts up to 1,500 μL collected from 300 μg of plant tissue.

Indeed, our results using ELISA (Figure 6A) and vacuum-based DBIA (Figure 6B) suggest that the sensitivity of CLas detection was increased. In the ELISA, protein extracts of asymptomatic and symptomatic tissues from four trees gave positive signals that are statistically significant (P < 0.001) compared to the healthy tissue (Figure 6A). In addition, stronger signals were observed from the vacuum-based DBIA when a larger amount of extract (300 μL vs. 50 μL) was used (Figure 6B). These assays suggest that SDE1 is a useful marker for HLB detection and the anti-SDE1 antibody is suitable for the development of serological diagnosis using multiple platforms.