Introduction

Spiroplasma citri is a species of spiroplasmas in the Spiroplasmataceae family. This bacterium does not have a cell wall but is contained by a plasma membrane, is motile, has a distinctive helical morphology and resides in the phloem sieve tubes. S. citri is the causal agent of citrus stubborn disease (Saglio et al. 1973). This bacterium is disseminated by infected planting material and by phloem feeding leafhoppers (Nejat et al. 2011). Citrus stubborn disease is preferentially found in hot and dry climate such as the Mediterranean basin and the Middle East. It is also present in parts of the USA, Mexico and northern Africa (EPPO 2018).This disease causes a decline in yields, with abnormally small and malformed fruits (Bové et al. 1988). It affects all citrus, particularly orange trees, but can also infect many species of herbaceous plants, often without apparent symptoms (Nejat et al. 2011).

S. citri is considered as a quarantine pest following the regulation (EU) 2016/2031 of the European council of 26 October 2016 and is also a quarantine pest in other parts of the world. As a result, the introduction in Europe of plants of Citrus L., Fortunella Swingle, Poncirus Raf., and their hybrids, other than fruit and seeds is controlled, among other things, to prevent accidental introduction of the pathogen. In order to meet the regulatory needs, it is necessary to have efficient and reliable detection methods.

In post-entry quarantine, the infected plants are generally asymptomatic, bacterial titer is low and the bacteria are unevenly distributed in the infected plant (Shi et al. 2014). S. citri is one of the few plant pathogenic mollicutes to have been cultured. Routine detection of S. citri from plants can be implemented by culturing in SP4 or M1A medium followed by examination of the organism for helical morphology and motility by dark field microscopy (Lee and Davis 1983). However, this method is time consuming and generally less sensitive than serological and molecular methods. Enzyme-linked immunosorbent assay (ELISA) using polyclonal antisera is also available for the detection of S. citri (Clark et al. 1978). Bové et al. (1987) have found that the sensitivity of ELISA and culture are of the same order of magnitude. However, the tests are not sensitive enough to reliably detect S. citri in symptomless trees (Bové et al. 1987) and, in their conditions of use, Bové and collaborators (Bové et al. 1983) pointed out that ELISA is not specific as it produces false positive results with other spiroplasmas of group I.

Currently, detection methods for bacteria are mainly polymerase chain reaction (PCR)-based (Palacio-Bielsa et al. 2009). There are many methods of gene amplification for the detection of S. citri targeting the genome of the bacteria itself, or phage-related sequences, such as plectrovirus SpV1, ubiquitously contained in the genome of S. citri. At least, 14 different conventional PCR assays targeting different areas of the genome are described including the detection of the 16S rDNA (Gasparich et al. 2004; Lee et al. 2006), the putative adhesin P89 gene (Yokomi et al. 2008), the putative adhesin P58 gene (Yokomi et al. 2008), the integrated virus (Lee et al. 2006), the spiralin gene (Foissac et al. 1996; Yokomi et al. 2008; Khanchezar et al. 2010) and the pE gene (Khanchezar et al. 2012). Real-time PCR allows accurate detection of plant pathogenic bacteria with higher sensitivity levels when compared to conventional PCR (Palacio-Bielsa et al. 2009). Seven real-time PCR assays are described targeting different areas of the S. citri genome, including the putative adhesin P58 gene (Yokomi et al. 2008), the integrated SpV1 virus (Wang et al. 2015) and the spiralin gene (Shi et al. 2014; Yokomi and Sisterson 2011). However, there have been no studies to compare the performance of these different methods.

To prevent the spread of the pathogen and disease outbreak, the early detection of latent infections in planting material, especially in the context of post-entry quarantine, is an important principle in plant disease control. In this context, rapid, sensitive and specific techniques to detect S. citri are essential. This study aims to evaluate the performance of different PCR methods for the detection of S. citri and provide robust validation data for these tests.

Materials and methods

Design of the study

Based on the literature and available validation data, a selection of methods for the detection of S. citri was made. The selection criteria were: i) easy and fast to use in routine analysis; ii) high analytical sensitivity; iii) high diagnostic sensitivity; and iv) no false positive results on healthy host tissue. These criteria resulted in selecting mainly real-time PCR protocols but conventional PCRs were kept for the evaluation if they targeted different parts of the S. citri genome. During a preliminary study in the Plant Health Laboratory of ANSES (LSV-UBVO), one DNA extract of citrus contaminated by S. citri and one DNA extract of healthy citrus were amplified two times with twelve different PCR methods. If unexpected results were obtained (i.e. false negative or false positive results), the related PCR protocols were removed from this study. This preliminary study allowed selection of four different PCR methods for the detection of S. citri (data not shown).

Then, in a first step, the LSV-UBVO compared and characterized the analytical specificity, the analytical sensitivity and the repeatability of these four PCR protocols and selected the three most effective. In a second step, an inter-laboratory test performance study (TPS) was conducted to confirm the results and to evaluate the reproducibility of the selected protocols.

Tested methods

For the first step, four different PCR protocols were tested: one conventional and one real-time PCR developed by Yokomi et al. (2008) targeting the P58 putative adhesin multigene of S. citri, one real-time PCR developed by Wang et al. (2015) targeting the SpV1-ORF3 integrated virus of S. citri and one real-time PCR developed by Yokomi and Sisterson (2011) the spiralin gene of S. citri. Details (amplification conditions) are provided on each method in Table 1.

Table 1 Amplification conditions of PCR methods evaluated during the study
Full size table

Performance criteria

Performance criteria and validation procedures were established following PM 7/76 (4) and PM7/98 (2) EPPO standards (European Plant Protection Organization 2014a, 2017) and International Organization for Standardization ISO 16140:2003 (2003). In particular, analytical specificity with diagnostic specificity (DSP), diagnostic sensitivity (DSE), analytical sensitivity (ASE), repeatability (DA) and reproducibility (CO) were assessed. The definitions and the calculations of these performance criteria are detailed in Chabirand et al. (2017).

In addition, for each method tested during TPS, a probability of detection was calculated per dilution level, according to the following equation x/n where x corresponds to the number of positive results and n corresponds to the number of results obtained for a given dilution level.

Intra-laboratory characterization

To evaluate DSP, DNA extracts from twenty-one S. citri infected plant samples, representative of the diversity of the bacterium, were analyzed two to three times per protocol (Table 2).

Table 2 List of Spiroplasma citri strains used in the intra-laboratory study
Full size table

To evaluate DSE, fifteen samples consisting of healthy citrus plants known as susceptible to S. citri, four closely related Spiroplasma species and nine plant pathogens of citrus (Table 3). These samples were analyzed two to three times per protocol.

Table 3 List of DNA samples used as non-target in the intra-laboratory study
Full size table

To evaluate ASE and DA, 10-fold serial dilutions of two S. citri strains (16/47.1 and 16/47.3) spiked in DNA extracts of healthy citrus were tested six times each.

Inter-laboratory test performance study

The TPS involved five partner laboratories in four countries (Table 4). For the evaluation of the different methods, the participating laboratories had to test an identical series of blinded samples according to the three working protocols and data reporting sheets provided.

Table 4 List of participants involved in the inter-laboratory trials
Full size table

A panel of fifty coded samples of frozen DNA extracts was sent to each participant. The detailed composition of each panel is presented in the Table 5. Preliminary tests aiming to verify the homogeneity and stability of the different samples were performed (data not shown).

Table 5 Type of samples used to evaluate performance criteria of methods for detection of Spiroplasma citri during TPS
Full size table

From these samples, the TPS participants were to carry out the different methods under the normal working conditions of the laboratory and in the same manner as other samples which are usually analyzed in the laboratory. PCR reagents and controls (positive and negative controls) were not provided by the organizer.

All participants were involved in the evaluation of each method.

Results

Analytical specificity

The analytical specificity of the evaluated methods ranged from 95 to 99% during the intra-laboratory study.

During the intra-laboratory study, the PCR protocols developed by Yokomi et al. (2008) allowed the detection of all the tested isolates of S. citri and the diagnostic sensitivity of those protocols was 100%. On the other hand, the diagnostic sensitivity of the other real-time PCR protocols ranged from 92 to 96% (Table 6). Those protocols did not allow the detection of two S. citri isolates (16/47.6 and 16/47.7). The diagnostic sensitivity of all tested protocols was 100% on condition of reproducibility (TPS) confirming the ability of the tested protocols to detect a diversity of S. citri isolates.

Table 6 Results for the characterization of the PCR methods for the detection of Spiroplasma citri: diagnostic sensitivity, diagnostic specificity, analytical sensitivity, repeatability, reproducibility
Full size table

The real-time PCR developed by Yokomi and Sisterson (2011) targeting the spiralin gene was the most specific method with a diagnostic specificity of 100%. The two protocols targeting the P58 putative adhesin multigene of S. citri (Yokomi et al. 2008) provided a false positive result with the DNA extract of Spiroplasma phoeniceum. The real-time PCR targeting the SpV1-ORF3 integrated virus of S. citri repeatedly provided false positive results with the DNA extracts of S. phoeniceum and Spiroplasma kunkelli. During TPS, with the real time PCR developed by Yokomi et al. (2008), one laboratory returned a false weak positive result with a high Ct value (38.74 cycles) and was obtained in a non-repetitive mode for sample LXX-47. This sample was a DNA extract of Citrus medica which did not produce a cross reaction during intra-laboratory testing.

Analytical sensitivity and probability of detection

During the intra-laboratory study, the best analytical sensitivity was obtained with both PCRs targeting the putative adhesin multigene P58 (Yokomi et al. 2008). Those PCR protocols were able to detect S. citri at 100% down to a 1 × 10−2 dilution of DNA extracts of infected plants with S. citri. The other real-time PCR protocols were 10 to 100 times less sensitive (Table 6).

The samples corresponding to the two last levels of dilution (1 × 10−4 and 1 × 10−5) of the target S. citri using for TPS cannot be considered as homogenous and/or stable (data not shown), the results of those samples were excluded from data analysis. If the results of these dilutions (1 × 10−4 and 1 × 10−5) of the target S. citri were taken into account, the real-time PCR protocol targeting the putative adhesin multigene P58 (Yokomi et al. 2008) provided a better analytical sensitivity (Fig. 1). Furthermore, the results of analytical sensitivity obtained during the TPS confirmed the results of the intra-laboratory study.

Fig. 1
figure 1

Probability of detection of the methods evaluated for the detection of Spiroplasma citri based on dilution level of a DNA extract of citrus infected by Spiroplasma citri in a DNA extract of healthy citrus. On the left, the two last levels of dilution were taken into account even if those samples were considered not homogeneous and/or not stable. On the right, only results with homogeneous and stable samples were included

Full size image

Repeatability and reproducibility

In the first step of the study at LSV-UBVO, the best repeatability was obtained with the real-time PCR targeting the SpV1-ORF3 integrated virus of S. citri (Wang et al. 2015). In TPS and among all the laboratories, the best repeatability was obtained by the PCR protocols (Yokomi et al. 2008) targeting the putative adhesin multigene P58 with a 100% repeatability while the real-time PCR (Yokomi and Sisterson 2011) targeting spiralin gene had a repeatability ranging from 72 and 100% depending on the laboratory (Table 6).

The reproducibility of the evaluated methods ranged from 98 to 100%. The best reproducibility is obtained with the PCR protocols targeting putative adhesin multigene P58 (Yokomi et al. 2008) (Table 6).

Discussion

S. citri is a plant pathogen that has been reported to cause a significant economic impact on citrus production (Bové 1986; EPPO datasheet). For this reason, the introduction of this pathogen is prohibited in many countries and post-entry quarantine is required for imported citrus plants. Therefore, reliable and highly sensitive detection methods are required to test plants in post-entry quarantine. This paper presents a study that evaluates the test parameters of a number of PCR protocols for the detection of S. citri.

During the preliminary study in the LSV-UBVO, many PCR protocols were removed from further evaluation and the TPS because these protocols provided false negative results from citrus samples containing S. citri (data not shown). Surprisingly, false negative results were also detected for the real-time PCR developed by Wang et al. (2015) that target the multicopy prophage gene, SpV1 ORF1, which had showed improved detection of S. citri when compared to other Spiroplasma genes (Spiralin, P58, SpV1-ORF3) (Wang et al. 2015). SpV1 replicative form (RF) has already shown a structural instability and the primer sequences may not be present in all isolates (Marais et al. 1996). Moreover, Wang et al. (2015) did use different test samples compared to those used in our study. Since such false negative results are unacceptable in testing plants in post-entry quarantine, those methods were not rated.

During our intra-laboratory study, the best performances were obtained with the PCR protocols developed by Yokomi et al. (2008) targeting the putative adhesin multigene P58. The targeted region of the S. citri genome allowed the specific detection of all the tested strains. The PCR protocols targeting the SpV1 ORF3 integrated virus (Wang et al. 2015) and the spiralin gene of S. citri (Yokomi and Sisterson 2011) produced false negative results in two DNA extracts. However, the high Ct values obtained from those samples with the real-time PCR of Yokomi et al. (2008) indicated that the target concentration was low. The failure of detection of those samples can be due to a problem of analytical sensitivity more than a problem of diagnostic sensitivity. This hypothesis is supported by the fact that some other isolates from the same area (California, USA) were detected by those PCR protocols. The best diagnostic specificity was obtained with the PCR protocol targeting the spiralin gene (Yokomi and Sisterson 2011) with 100% diagnostic specificity. The protocols targeting the putative adhesin multigene P58 (Yokomi et al. 2008) cross reacted with Spiroplasma phoeniceum. This was in accordance with the results obtained by Yokomi et al. (2008). S. phoeniceum is a spiroplasma isolated from symptomatic periwinkles in Syria between 1983 and 1984 but which since has never been described in a sanitary or epidemic context. As S. phoeniceum has never been found in citrus and is potentially a plant infectious microorganism, it’s co-detection in post-entry quarantine is not considered a real problem. Likewise, S. phoeniceum and S. kunkelli were detected by the real-time PCR targeting the SpV1-ORF3 of S. citri (Wang et al. 2015). Wang et al. (2015) did not include those spiroplasmas in their study. Both PCRs targeting the putative adhesin multigene P58 (Yokomi et al. 2008) were 10 to 100 times more sensitive than the other PCR protocols.

The second part of the study was designed as a TPS which provides additional value to the validation process as it provides more information on repeatability and reproducibility (EPPO 2014b). As the real-time PCR targeting the SpV1-ORF3 of S. citri (Wang et al. 2015) gave the worst diagnostic specificity in the first part of the study, this method was not rated in the TPS. The methods developed by Yokomi et al. (2008), which provided the best performance, were assessed and compared to the PCR protocol targeting the spiralin gene (Yokomi and Sisterson 2011). This allowed evaluation of a different method targeting another part of the S. citri genome. The results of the TPS confirmed the results previously obtained in the intra-laboratory study and showed that the PCR protocols targeting the putative adhesin multigene P58 (Yokomi et al. 2008) were the most efficient methods to detect S. citri. However, it also highlighted a risk of micro-contaminations of 2% with the real-time PCR of Yokomi et al. (2008). As for all highly sensitive methods, measures should be taken to avoid this type of contamination. The legitimacy to use these methods for the detection of S. citri in post-entry quarantine is reinforced by the high reproducibility (98–100%) obtained during the TPS for the evaluated methods using different reagents and real-time PCR machines.

As real-time PCR is faster and more sensitive than conventional PCR, the real-time PCR targeting the putative adhesin multigene P58 (Yokomi et al. 2008) can be efficiently used in post-entry quarantine. In case of a positive detection and in critical cases, the other evaluated protocols can be used to confirm the results obtained by real-time PCR. For testing laboratories following the standard ISO 17025 (International Organization for Standardization 2017), the methods used should be characterized (point 7.02.2 of the standard). That is why, in a context of quality assurance management system, the study reported in this paper provides essential validation data for the detection of S. citri.