Candidatus Liberibacter asiaticus (CaLas), affects the citrus-producing areas in Mexico (^{DGSV-SENASICA and Mora-Aguilera, 2012}); it reduces by 18.6 % the volume of juice and by 17.3 % the weight of fruits in Persian lime in Yucatan, Mexico, causing losses in production of 2.4 t.ha^-1 (^{Flores-Sánchez, et al., 2015}). The infection induced by HLB alters the export of photoassymilates caused by the retention of starch, which generates the expression of symptoms in the tree (^{Kim et al., 2009}; ^{Koh et al., 2012}).

CaLas causes biochemical and structural changes in citrus plants to prevent the spread of the bacteria, by activating response proteins (^{Albrecht and Bowman, 2008}). In response to the infection of citrus plants with HLB, carbohydrate metabolism involving α-amylase is altered (^{Albrecht and Bowman, 2008}; ^{Etxeberria et al., 2009}).

Phenylalanine ammonia lyase (PAL) intervenes in response to the attack of the pathogens (^{Almario et al., 1994}). In citrics, the flavedo has the ability to respond to attacks by pathogens, by the increase of the PAL transcript levels, as well as of its activity (^{Ballester et al., 2006}). The enzyme peroxidase (POD) protects plants from the damage caused by the free radicals or ROS generated by different types of stress and also participates in the lignification of the cell wall and in the degeneration of indolacetic acid (^{Robinson, 1991}). Phenols have proven to play an important part in the defense of plants to different biotic and abiotic factors (^{Lu et al., 2015}). They can be modified by effectors secreted by the bacteria Ca. Liberibacter inside the host, and that alter genes related to defense; CaLas codifies salicylate hydroxylase as a mechanism to evade the plants defenses (^{Aritua et al., 2013}). ^{Mora-Aguilera et al., 2014}, indicate that HLB control must consider the vulnerability of the citrus species and the load of the inoculant. Understanding the enzymatic behavior of trees infected with CaLas in different species could help develop alternatives for the protection or defense of the plant. This study was developed with the objective of determining differences in the antioxidant enzyme activity and the concentration of phenols in three citrus species in Bustamante, Tamaulipas, Mexico.

Samples were taken in Felipe Angeles, Bustamante, Tamaulipas, Mexican lime (Citrus aurantifolia (Christm.) Swingle), Persian lime (Citrus latifolia Tanaka) and sweet orange trees (Citrus sinensis (L.) Osbeck) aged 10 to 12 years were sampled under rainfed conditions, in the phenological stage of fruit set. Samples were taken from two trees with symptoms per species, and one without symptoms. The material was gathered in each one of the cardinal points of the tree (north, south, east, and west), three repetitions of five leaves were taken from each point and kept in liquid nitrogen.

For DNA extraction, the technique quoted by ^{Almeyda-León et al. (2001)} was modified for its use: 100 mg of tissue were ground in liquid nitrogen, 1 mL was added of 2 ME/CTAB extraction solution, preheated at 65 °C. It was mixed in a vortex for 30 sec and incubated at 65 °C for 45 min. Next, 500 μL of isoamyl alcohol chloroform was added 24:1, the mixture was stirred inversely and centrifuged at 12,000 rpm for 10 min. The aqueous phase was extracted and an equal volume of isoamyl alcohol chloroform was added 24:1; the centrifuge cycle was repeated. The top layer was taken and 0.6 volume of isopropanol was added to precipitate the DNA at -20 °C for 24 h. It was centrifuged at 12,000 rpm for 20 min, the pellet was washed with ethanol at 70 %, centrifuged at 12,000 rpm for 15 min, left to dry and resuspended in 50 μl injectable water (Pisa). The presence of bacteria was determined according to the protocol by ^{SENASICA-SAGARPA, (2010)} used in the National Plant Epidemiology and Health Station (SENASICA-ENECUSaV), in the state of Querétaro, Mexico.

The raw extract for the quantification of total proteins, phenylalanine ammonia lyase, peroxidase, and α-amiylase, was taken from ^{Díaz et al. (2010)}, and modified: 1 g of grinded sample was placed in liquid nitrogen, in a sodium phosphate buffer solution 100 mM, pH 7. The samples were centrifuged at 12,000 rpm for 20 minutes at a temperature of 4 °C; later, the supernatant was collected and stored at -20 °C. For the extraction of phenols, methodology B, used by ^{Kähkönen et al. (1999)} was used here with some modifications; 250 mg of the grinded sample were placed in a 2 mL eppendorf tube and 1 mL of methanol at 80 % was added; it was shaken in a vortex for one minute and centrifuged at 10,000 rpm for 15 minutes. The supernatant was collected in a 1.5 mL amber eppendorf tube. It was re-extracted with 500 μL of absolute methanol, repeating the process and stored at -20 °C until use.

Protein concentration was determined according to the technique by ^{Bradford (1976)}, which consisted in mixing 1 mL of Bradford reagent with 100 μL of raw extract. The total protein absorbance values were determined at a λ=595 nm in a thermo spectronic Biomate 3. The protein content was expressed in mg of protein per mg of fresh tissue. The pattern curve was obtained with the albumen of cow serum, as per ^{García and Vázquez (1998)}.

Phenylalanine ammonia lyase (PAL) activity was determined according to the technique described by ^{Rodríguez-Pedroso et al., (2006)}. The 0.0174 M^-1cm^-1 molar extinction coefficient was used as a reference to calculate enzyme activity (^{Trotel-Aziz et al., 2008}).

Peroxidase activity was determined using the methodology used by ^{Ruttimann et al., (1992)}. Enzyme activity was expressed as millimoles of phenol red oxidized per gram of fresh tissue per minute (Yedidia et al., 1999).

For the quantification of α amylase, we used the ^{SIGMA-ALDRICH protocol, (2015)}. The standard curve was produced with various concentrations of maltose. The α amylase activity was defined as µmol.min^-1 of maltose released per µg of protein (^{Menéndez et al., 2006}).

The content of phenols was determined with the Folin-Ciocalteu method (^{Singleton et al., 1999}). A mixture was prepared using 200 μL of the methanolic extract, 100 μL Folin-Ciocalteu reagent, and 200 μL of sodium carbonate at 20 %. It was shaken and left to rest for 30 min. After this time, total phenol absorbance was measured at one λ=760 nm. A standard curve was produced with various concentrations of gallic acid.

The SAS 9.0 statistical package was used to perform an analysis of contrasts between the trees with positive and negative diagnoses for CaLas. The variables analyzed were: total proteins, PAL, POD, α-amylase, and total phenols.

The sampled trees with symptoms displayed a general yellowness of the canopy and they appeared dehydrated; leaves presented asymmetric, less visible spots. Symptomatic samples were positive for CaLas, and asymptomatic samples were negative. For the detection of CaLam all samples were negative.

The analysis of contrasts between positive and negative trees displayed significant differences between species (P<0.0001). Protein concentration for Mexican lime displayed an average of 6.1 mg of proteins per g of fresh tissue for infected trees, whereas for negative ones, it was 5.46 mg of proteins per g of fresh tissue with a significance of P=0.0238. In sweet orange, there is a significant difference of P=0.0002 with an average of 5.27 and 6.37 mg of proteins per g of fresh tissue in positive and negative trees, respectively (Table 1; Figure 1). Determining total proteins in the plan helps find changes caused by different types of biotic and abiotic stress (^{Casado, 2004}), as shown in this study when finding differences in the concentration of proteins between positive and negative trees for CaLas.

Table 1. Analysis of contrasts between trees with negative and positive diagnoses for Ca. Liberibacter asiaticus in the state of Tamaulipas. 

Figure 1. Comparison of averages of total proteins (1A), phenylalanine ammonium lyase (PAL) (1B), peroxidase (1C) α-amylase (1D), and total phenols (1E), for the sources of variations of location, species, and branch orientation evalu ated in trees, both negative and positive for Ca. Liberibacter asiaticus. 

Determining PAL did not display significant differences between species, although a higher concentration of PAL was observed in Persian lime and Mexican line (21.75, 16.09 U.mg of protein^-1) for negative trees. In contrast, sweet orange trees infected by CaLas showed a higher concentration (17.98 U.mg of protein^-1) (Table 1; Figure 1). The results of this study coincided with ^{Chenyang et al. (2001)}, who mention that the enzyme activity of PAL is modified by biotic and abiotic factors in the plant, since despite having no significant differences, a higher PAL concentration was observed in sweet oranges affected by HLB. The results in Persian and Mexican lime coincided with ^{Martinelli et al. (2016)}, who point out that the regulation of PAL dropped due to the infection of CaLas in both species, one considered moderately tolerant, Volkameriana lime (Citrus × volkameriana), and the other, Navel orange, highly vulnerable.

The levels of peroxidase activity presented a significant difference for Persian lime (P=0.0341). A higher concentration of the enzyme was observed in trees not infected with CaLas with averages of 1.96 U.mg of protein^-1 for Persian lime and sweet orange, and for Mexican lime, 1.53 U.mg of protein ^-1 (Table 1; Figure 1). In this investigation, we oberved a higher enzyme peroxidase activity in negative citrus trees; however, symptoms of central and secondary veins degeneration in leaves leaves infected with HLB, produced by the induced lignification by this enzyme, as mentioned ^{Robinson (1991)}.

In the enzyme quantification of α-amylase, significant differences were presented between species and between positive and negative sweet orange trees (P<0.0001). Negative Persian lime and Mexican lime trees displayed a higher concentration of α-amylase (1.31, 0.99 U.mg of protein^-1), respectively; in positive trees, the highest concentration was found in Persian lime (1.26 U.mg of protein^-1) (Table 1; Figure 1). In the study by ^{Martinelli et al. (2016)}, α-amylase was expressed in Volkameriana lime (Citrus × volkameriana), indicated as moderately tolerant, and β-amylase in Navel orange, a highly vulnerable material. In this study, the stress on the tree caused the α-amylase activity to increase in Persian lime, followed by sweet orange.

The levels of total phenols were displayed in a minimum range of 18.5 mg of gallic acid per g of fresh weight registered for the samples of Persian lime infected with CaLas. Sweet orange presented the maximum value (21.10 mg of gallic acid per g of fresh weight) in negative trees. In the analysis of contrasts, the determination of total phenols did not show significant differences between species (P=0.5045) (Table 1; Figure 1). Coinciding with Yedidia et al. (1999), who point out that when suppressing the peroxidase activity, the production of phenolic compounds is also reduced, the results do not display a significant difference in peroxidase activity or in phenol compounds.