Huanglongbing (HLB), or Yellow Dragon Disease, is considered the most devastating citrus disease worldwide (^{Ding et al., 2015}). The etiological agent is a gram-negative α-proteobacterium and three species are currently known to infect citrus: Candidatus Liberibacter asiaticus (CLas), Candidatus Liberibacter africanus (CLaf) and Candidatus Liberibacter americanus (CLam) (^{Achor et al., 2020}; ^{Andrade et al., 2020}). In Mexico, the most severe damage of this disease occurs in Mexican lime (Citrus aurantifolia) and Persian lime (Citrus latifolia), while in other regions of Asia and Africa and in countries such as Brazil, USA and Cuba, the damages are mainly reported in sweet orange (Citrus sinensis) (^{McCollum et al., 2016}).

The official confirmation of the bacterium in plants is done by real-time PCR, since it is a very sensitive method and is considered the most reliable tool for its detection. However, this method is expensive; therefore, a fast and reliable test is required to corroborate its presence in the field. Immunoenzymatic methods or immunoassays, such as the enzyme-linked-immuno sorbent assays (Elisa) technique, are based on the antigen-antibody (Ag-Ab) reaction, are economical, rapid, and specific to detect low concentrations of Ag or Ab and can be used for their sensitivity to detect phytopathogens (^{Fundora et al., 2013}). The results of its application have allowed the development of integrated disease management strategies, improving the quality and health of crops, as well as their competitiveness and profitability (^{Zherdev et al., 2018}).

There are reports of the presence of proteins that may be immunogenic and are located in the outer membrane of pathogens (^{Bastianel et al., 2005}; ^{Lu et al., 2013}). They are also identified as determinant antigenic or discrete sites that are recognized by B or T lymphocytes through their own specific receptors (^{Carrizo et al., 2009}). These proteins can be used for the generation of antibodies in warm-blooded animals, therefore, the objective of this research was to identify proteins located in the outer membrane of Candidatus Liberibacter asiaticus, with antigenic capacity that serve as a basis for the development of an efficient and low-cost immunoenzymatic method for the detection of this bacterium in citrus.

For the bioinformatic analysis of outer membrane proteins, the complete genome of CLas strain GX-1 (Genbank Access PRJNA158395) was used, which was reported by ^{Lin et al. (2013}), in the NCBI (http://www.ncbi.nlm.nih.gov/), later it was analyzed in the UniprotKB database (http://www.uniprot.org) and the amino acid sequences of the outer membrane proteins encoded in the CLas genome were obtained.

The prediction of outer membrane protein function was performed using the Predict Protein web server (www.predictprotein.org/) and the Blast2go program (Run Blast, Mapping, Annotation, Inter ProScan) (^{Conesa et al., 2005}). A Muscle alignment of the selected proteins based on codons was performed and a dendrogram was generated to determine their similarity by Neighbor Joining, using the MEGA X software. For epitope prediction, the protein B8Y674 was selected because it is repetitive in the CLas genome and the immune epitope database and analysis resource (Iedb) (http://tools.immuneepitope.org/tols/bcell/iedb-input) was used, by means of the Kolaskar and Tongaonkar antigenicity scale (Ktas) method.

The detection of CLas was performed by endpoint PCR from genomic DNA extracted from Mexican lime (Citrus aurantifolia) leaves with symptoms of HLB. DNA extraction was performed by the CTAB method optimized by ^{Rodríguez et al. (2010}) and in the PCR reactions, the primers designed on the sequence of the Omp gene encoding the protein B8Y674 were used (Rodríguez et al., 2018). The amplified fragments (two for each pair of primer used) were purified, cloned, and sequenced for online analysis through the NCBI database (http://www.ncbi.nlm.nih.gov/) in the Blast section.

Fifty-three outer membrane proteins were identified (data not shown) and based on their homology, motifs, and domains, 12 of these proteins were selected (Table 1), of which only C6XHX5 had previously been characterized as a protein of tolerance to organic solvent. With HhPred/HhSearch (^{Söding et al., 2005}), functional predictions were obtained with probabilities ranging from 62.5 to 99.2% and with acceptable E-values (3.3E^-05, 1.4E^-06, 2.5E^-13, 2.8E^-13, 3.5E^-15), obtaining alignments with significant results for 10 proteins. In the proteins C6XGP8 and C6XGU8, probabilities of 93.8 and 62.5%, as well as E-values of 0.0027 and 1.1 respectively, were recorded, so they were considered non-significant or unreliable results in their predictions (Table 1).

Annotations of functional domains indicated predictions of surface antigen for proteins B8Y674, B8Y671, B2KNJ1, J7H0I4, B8Y672, B8Y673, B8Y675, C6XGP8, C6XGU8 and C6XFB8, suggesting that they have antigenic capacity, C6XHX5 had prediction of tolerance to organic solvents and C6XF21 maintenance of cell envelope integrity. The multiple alignment of the proteins encoded by the Omp gene (B8Y671, B8Y672, B8Y673, B8Y674 and B8Y675) showed an identity among them of 97% (Figure 1).

Figure 1 Dendrogram generated by alignment of sequences of outer membrane proteins of CLas. 

The prediction of epitopes for the protein B8Y674 allowed obtaining eight B or T epitopes, with a specificity of 91% (Table 2). The IEDB database determined that the epitopes FSEVNISQLP ANDVSDYVILRVSVKQLSAGSVGIA, RHAVQNYTLSVEDPYFLGSPISAGFD, IPSIYTTLI EHG, and FSSHSISQSIIYNY, have been evaluated experimentally with identities of 90%, 80%, 80%, and 80%, respectively (http://www.iedb.org/). In addition, it revealed that the type of cell involved in most cases was the B cell. Some of the important functions of B cells are the production of antibodies against antigens and eventually transform into memory B cells after being activated by interaction with an antigen (^{Mauri and Bosma, 2012}).

Fragments of 1 121 bp and 904 bp respectively of the Omp gene of CLas encoding the protein B8Y674 were amplified from DNA extracted from Mexican lime leaves with symptoms of HLB using the primers reported by ^{Rodríguez et al. (2018}). The four sequences obtained from the amplified fragments presented 98% identity with CLas sequences previously reported in GenBank (CP019958.1, CP004005.1, KC473159.1, KC473156.1, JQ928886.1, JQ928883.1, JQ928882.1, JN049494.1).

It is important to note that the sequences used to date for detection by ELISA have been isolated from sweet citrus (^{Lu et al., 2013}) or psyllids carrying the bacterium (^{Yuan et al., 2016}), unlike the sequences of our study that were isolated from Mexican lime, finding five nucleotide variations, corroborating what was mentioned by ^{Tomimura et al. (2009}), who reported that the gene that codes for the outer membrane protein has nucleotide variations that allow differentiating between them. In 2010, Chen et al. pointed out that, at the OMP locus, the variation is limited to a few SNPs, validating the results of this research.

These data highlight the potential of the findings obtained in this study with the identification of surface antigens, from which specific antibodies that avoid cross-reactions with bacteria other than CLas can be obtained, in an immunoenzymatic method for the early detection of the causative agent of citrus HLB disease.